Characterization and detection
of Puccinia horiana on chrysanthemum
for resistance breeding and sustainable control
Mathias De Backer
FACULTEIT BIO-INGENIEURSWETENSCHAPPEN
Voor Luka en Mirthe
Geen thesis zonder hun wijsheid!
Promotors:
Prof. dr. ir. Erik Van Bockstaele
Department of Plant production
Ghent University
Dr. ir. Kurt Heungens
Institute for Agricultural and Fisheries Research (ILVO)
Plant Sciences Unit - Crop Protection
Dean:
Prof. dr. ir. Guido Van Huylenbroeck
Rector:
Prof. dr. Paul Van Cauwenberge
Mathias De Backer
Characterization and detection
of Puccinia horiana on chrysanthemum
for resistance breeding and sustainable control
Thesis submitted in fulfillment of the requirements for the degree of
Doctor (PhD) in Applied Biological Sciences
Dutch translation of the title:
Karakterisering
en
detectie
van
Puccinia
horiana
op
chrysant
in
het
kader
van
resistentieveredeling en duurzame controle.
Cover picture: Symptoms of P. horiana on a heavily infected chrysanthemum and a field with
flowering chrysanthemums.
Please refer to this work as follows:
De Backer, M (2012). Characterization and detection of Puccinia horiana on chrysanthemum for
resistance breeding and sustainable control. PhD thesis, Ghent University, Belgium.
ISBN-number: ISBN 978-90-5989-540-9
Printed:..www.universitypress.be
The author and the promotors give the authorisation to consult and to copy parts of this work for
personal use only. Every other use is subject to the copyright laws. Permission to reproduce any
material contained in this work should be obtained from the author.
Woord van dank
Het werk dat voor je ligt is het resultaat van meer dan vijf jaar met volle overtuiging tegen een
wetenschappelijke muur lopen en er op tijd en stond eens over geraken. Gelukkig kon ik steeds rekenen op de
steun van mensen die al dan niet in hetzelfde wetenschappelijke schuitje zaten. En ook zij die zich bij
momenten ernstige vragen stelden bij zin en onzin van mijn experimenten hadden ergens ook gelijk (Schwartz
2008). Een woord van dank is dan ook zeker op zijn plaats.
Willen doctoreren is één ding, maar kunnen doctoreren is minder voor de hand liggend. Allereerst zou ik dan
ook mijn promotoren Prof. dr. ir. Erik Van Bockstaele en Dr. ir. Kurt Heungens alsook wetenschappelijk
directeur Dr. Tine Maes en hoofd van ILVO-Eenheid Plant Dr. ir. Kristiaan Van Laeke willen bedanken voor het
geloof dat ze steeds in mij en mijn project hebben gesteld. Kurt, dank je voor je mateloos enthousiasme, steun
en begrip in moeilijkere tijden en de “stamp onder mijn gat” die ik op tijd en stond nodig had om mij op het
juiste pad te houden. Wetenschappelijke problemen waren er om tot op het bot uit te spitten en als we iets
tekort kwamen was er steeds wel iemand in je netwerk die voor de nodige input kon zorgen. Het doet me
plezier hiervan de vruchten te kunnen plukken en ik hoop dat in de toekomst nog te kunnen doen. Dank je ook
voor de uren, dagen, weken die je in het nalezen, verbeteren en het formuleren van waardevolle opmerkingen
op deze tekst hebt gestoken. Hierbij hoort ook een bijzonder woord van dank voor Miriam omdat Kurt die tijd
erin mocht steken en voor Eva en Sara voor de tijd die zij niet met hun papa konden spelen.
Het vele werk dat aan de basis van mijn thesis ligt werd met de zeer gewaardeerde hulp van vele handen
geklaard. In een project rond een chrysantenziekten mogen veredelaars natuurlijk niet ontbreken. Dank je aan
Bas Brandwagt, Kees van’t Hoenderdal, Aike Post en Henk Dresselhuys voor de goede samenwerking in het
pathoype project met Plantum en het aanleveren van de nodige stekken en de interessante isolaten. Ook voor
Wim Declercq en medewerkers van Gediflora is een pluim weggelegd voor het zorgvuldig aanleveren van de
nakomelingen uit de kruisingsproef, doorgeven van weersgegevens, aanleveren van isolaten en het open staan
voor andere aspecten van het project. Een aantal weken werken in Wageningen was enkel mogelijk dank zij de
deskundige begeleiding van Theo van der Lee met Marga van Gent-Pelzer als gids in het PRI-labo. Op Plant-21
had ik steeds Isabel Roldan als baken voor alles wat de chrysantkant aanbelangde en kon ik rekenen op de zeer
gewaardeerde hulp in het labo van Nancy Mergan en Sabine Van Glabeke. In eigen huis kon ik steeds op verse
isolaten rekenen dank zij het zorgvuldig overzetten van de collectie door Dirk en ook Sabine inoculeerde en
evalueerde een paar duizend stekken. Bij Steve kon ik steeds terecht met vraagstukken over de optimalisatie
van moleculaire technieken en mijn thesisstudent Mikhail deed een bijdrage bij het optimaliseren van het
detectieprotocol. Dank je Kris dat ik de laatste mycologieanalyses tussen het virologiewerk mocht proppen.
Ook voor mijn vele bureaugenoten die de revue gepasseerd zijn een speciaal woord van dank. Ik werd
fantastisch ontvangen op het ILVO door de Ramorum-vrouwen Annelies, Isabel en Tineke om dan de gang over
te steken naar mijn definitieve stek bij Kris en Cinzia. De laatste maanden bracht ik een vedieping lager door bij
Brigitte, Annemie, Evelien en Sofie. Er kon steeds worden gelachen en ze hebben allemaal mijn bijzondere
waardering voor Microsoft moeten doorstaan. Natuurlijk eindigt het lijstje hier niet: Kubben met Bjorn, Sam en
Fran, roddels aanhoren van Shana, de filosofie van de wetenschap bepreken met Rachid, een propere bureau
dank zij Brigit of Mireille, de stilte van Bart doorbreken bij Jane en Joa… Kortom teveel namen om op te
noemen, maar wel een welgemeende dank je wel aan al mijn collega’s van gewasbescherming.
Dank je ook voor de leden van de examencommissie: Prof. dr. Godelieve Gheysen, Prof. dr. ir. Monica Höfte,
Prof. dr. ir. Marie-Christine Van Labeke, Prof. dr. Isabel Roldan-Ruiz, Dr. Theo van der Lee en Dr. ir. Bas
Brandwagt. Hun kritische opmerkingen betekenden een belangrijke bijdrage bij het tot stand komen van de
finale tekst van deze thesis.
Natuurlijk wil ik ook mijn vrienden bedanken die gelukkig steeds klaarstonden voor de nodige ontspanning
buiten het ILVO. Enen gaan drinken of een gezelschapspel spelen met Joachim, Sofie, Kristof, Sebastien, Babs,
Stijn en Gert of nog eens uithalen van of terugdenken aan straffe stoten met Martine, Gert, Klaas, Do, Gorik,
Esther, Klaas, Joke, … stond steeds garant voor de nodige ambiance. De honger werd van tijd tot tijd gestild
tijdens “koken met vriendjes” met Annelies, Hanneke, Yirina en Steven, tof dat die vriendjes eigenlijk (ex-)
collega’s zijn. Voor eten of een spelletje kon ik ook steeds bij Nancy en Kenny terecht en als het echt te druk
werd stond Pieter steeds klaar om de trossen los te smijten en het ruime sop te kiezen.
Papa en mama verdienen een speciaal woord van dank voor de steun aan hun eeuwige student. Dank je dat ik
op tijd en stond nog eens naar huis mocht komen voor lekker eten en een babbel. Mijn zusjes Babs en Lore wil
ik bedanken voor hun immer sceptische benadering van het concept doctoreren. Zij scherpten hiermee enkel
mijn koppigheid aan om door te zetten. Ook de rest van de familie zoals opa en oma, tantes en nonkels,
vake,… verdient hier een plaatsje voor hun blijvende interesse in de soms ver-van-hun-bed show die ik soms
opvoerde.
Het laatste woord van dank is er eerder één van liefde voor Maureen. Lieve zoet, het was vaker moeilijk dan
gemakkelijk, maar ik heb genoten van onze momenten samen; thuis en aan de andere kant van de wereld.
Luka kwam en ging, maar samen met ons liefste Mirthe zullen we in de toekomst met plezier naar deze
bijzondere periode terugkijken.
Mathias
Augustus 2012
Schwartz MA (2008). The importance of stupidity in scientific research. Journal of Cell Science 121, 1771.
Table of Contents
Table of Contents
List of Abbreviations
Problem statement and outline ............................................................................................................... 1
Chapter 1:
General introduction ........................................................................................................ 7
1.1
Chrysanthemum....................................................................................................................... 9
1.2
Rust fungi ............................................................................................................................... 16
1.3
Puccinia horiana ..................................................................................................................... 20
1.4
Detection of fungal pathogens .............................................................................................. 34
1.5
Molecular tools for resistance breeding and diversity studies .............................................. 38
Part I:
Phenotypic variation in P. horiana ......................................................................................... 43
Chapter 2:
Identification and characterization of pathotypes in Puccinia horiana, a rust pathogen
of Chrysanthemum x morifolium ........................................................................................................... 45
2.1
Introduction .............................................................................................................. ............. 47
2.2
Materials and methods .......................................................................................................... 49
2.3
Results .................................................................................................................................... 55
2.4
Discussion............................................................................................................................... 61
Chapter 3:
Fungicide resistance of Puccinia horiana ....................................................................... 65
3.1
Introduction .............................................................................................................. ............. 67
3.2
Material and methods ........................................................................................................... 69
3.3
Results .................................................................................................................................... 73
3.4
Discussion............................................................................................................................... 74
Part II:
Genetic variation in P. horiana and spore detection in air samples using molecular
techniques.............................................................................................................................................. 79
Chapter 4:
Genotypic diversity of Puccinia horiana based on newly identified SNP markers ........ 81
4.1
Introduction ........................................................................................................................... 83
4.2
Material and methods ........................................................................................................... 85
Table of Contents
4.3
Results .................................................................................................................................... 91
4.4
Discussion............................................................................................................................... 97
Chapter 5:
Molecular detection of Puccinia horiana in air samples .............................................. 103
5.1
Introduction ......................................................................................................................... 105
5.2
Material and methods ......................................................................................................... 107
5.3
Results .................................................................................................................................. 113
5.4
Discussion............................................................................................................................. 122
Part III:
Inheritance in chrysanthemum ........................................................................................ 127
Chapter 6:
Segregation of resistance to P. horiana in chrysanthemum ........................................ 129
6.1
Introduction ......................................................................................................................... 131
6.2
Material and methods ......................................................................................................... 134
6.3
Results .................................................................................................................................. 139
6.4
Discussion............................................................................................................................. 145
Chapter 7:
Segregation of AFLP markers in chrysanthemum ........................................................ 151
7.1
Introduction ......................................................................................................................... 153
7.2
Material and methods ......................................................................................................... 155
7.3
Results .................................................................................................................................. 158
7.4
Discussion............................................................................................................................. 166
Chapter 8:
General discussion and future perspectives ................................................................ 171
8.1
The Puccinia horiana toolbox............................................................................................... 174
8.2
Puccinia horiana as a diverse globetrotter .......................................................................... 177
8.3
Insights in the biology and epidemiology of Puccinia horiana ............................................ 178
8.4
Breeding for resistance to Puccinia horiana ........................................................................ 179
References ........................................................................................................................................... 181
Summary .............................................................................................................................................. 203
Samenvatting ....................................................................................................................................... 207
Curriculum Vitae .................................................................................................................................. 213
List of Abbreviations
A.D.
anno Domini
AFLP
amplified fragment length polymorphism
ANOVA
analysis of variance
ASM
acibenzolar-S-methyl
Avr
Avirulence
bp
base pair
bps
beats per second
BSA
bulked segregant analysis
BTH
benzo(1,2,3)thiadiazole-7-carbothioic acid-S-methyl ester
CAN
Comunidad Andina (formely "JUNAC" (Junta del Acuerdo de Cartagena))
cM
centimorgan
CRoPS
complexity reduction of polymorphic sequences
Ct
cycle treshold
CTAB
cetyl trimethylammonium bromide
Cv
cultivar
CYP51
cytochrome P450 (family 51) gene
cyt b
cytochrome b gene
DNA
deoxyribonucleic acid
dpi
days post inoculation
ELISA
Enzyme Linked Immunosorbant Assay
EPPO
European and Mediterranean Plant Protection Organization
EST
expressed sequence tag
gDNA
genomic DNA
HR
hypersensitive reaction
IAPSC
Inter-African Phytosanitary Council
ITS
internal transcribed spacers
LFD
lateral flow devices
LOD
limit of detection
LOD
logarithm of odds
LOQ
limit of quantification
MAS
marker-assisted selection
NAPPO
North American Plant Protection Organization
List of Abreviations
PCR
polymerase chain reaction
pDNA
plasmid DNA
QoI
quinone outside inhibitors
qPCR
quantitative PCR
R
resistance or resistant
RAPD
randomly amplified polymorphic DNA
rDNA
ribosomal DNA
RH
relative humidity
RNA
ribonucleic acid
rpm
revolutions per minute
S
susceptibility or susceptible
SD
standard deviation
SNP
single nucleotide polymorphism
SSR
simple sequence repeat
UPGMA
Unweighted Pair Group Method with Arithmetic Mean
USDA
United States Department of Agriculture
Problem statement and outline
Symptoms of Puccinia horiana on Chrysanthemum
Problem statement and outline
Chrysanthemums (Chrysanthemum x morifolium) are among of the most important ornamental
crops, produced worldwide as cut flowers or potted plants. The first reports of chrysanthemum
breeding were made more than 2000 years ago in Asia, from where the crop was exported to the
rest of the world. Nowadays, several companies in Europe, America and Asia are involved in
breeding and in the production of cuttings, cut flowers and potted plants.
One of the most important threats for chrysanthemum production is the fungal pathogen Puccinia
horiana, causing chrysanthemum white rust. This pathogen was first described at the beginning of
the 20th century by Hennings (Hennings 1901) and was reported for the first time in Europe in 1963
(Baker 1967). Currently, P. horiana is widespread in most areas where chrysanthemums are grown
and the pathogen is recognized as a quarantine pest in most continents (Anon. 2004). Although the
life cycle of the pathogen has been described and the optimal conditions for development have been
determined (Firman and Martin 1968), sustainable control of the pathogen is not always realized, in
part due to gaps in our knowledge of the pathogen’s biology. For example, little is known about the
initial inoculum sources. The pathogen is mainly controlled by regular preventive spraying with
fungicides such as triazoles and strobilurins, although fungicide resistant strains to both types of
compounds have been reported (Cook 2001). However, little is known about resistance to different
types of fungicides in different isolates. Recently breeding for resistance to P. horiana has gained
importance as a sustainable control method. Anecdotal reports of infected cultivars that were
previously regarded as resistant indicate the presence of different physiological races or pathotypes
of chrysanthemum white rust. These pathotypes are the result of gene-for-gene interactions. The
complexity of the pathosystems is determined by the number of genes involved. Knowledge about
the pathosystem and the geographical distribution of pathotypes as well as the inheritance of
resistance to different pathotypes in chrysanthemum will enhance sustainable control of the
pathogen by resistance breeding.
The major objective of this dissertation was to determine the phenotypic diversity and the genotypic
diversity of P. horiana isolates from different geographic origins and time periods. A better
knowledge of this diversity should allow optimization of fungicide application schedules and
breeding programs, and can provide information about the general evolution of the pathogen in a
geographical context. A second objective was the development of a technique for the detection of P.
horiana spores in air samples. Such a technique will aid in the determination of the primary
inoculum sources and the local spread of the pathogen. In the future, this technique could also be
implemented in warning systems. Finally, I also studied the inheritance of resistance to P. horiana as
well as the inheritance of dominant molecular markers in chrysanthemum. Since chrysanthemum
3
Problem statement and outline
has a complex hexaploid genome, a better knowledge about inheritance of resistance to P. horiana
and inheritance in general, will increase efficiency in breeding programs.
The first chapter (Chapter 1) of this thesis is a literature review giving an overview of the history and
biological limitations of chrysanthemum breeding and an introduction on the rust fungi. For P.
horiana, the disease symptoms, the life cycle and the nuclear cycle as well as the available control
measures are discussed in detail. Finally, the available techniques for the detection of fungal spores
in air samples and the molecular tools to study genetic variation are reviewed.
The next six chapters deal with the research I conducted from 2007 to 2011 and are divided in three
major parts, with two chapters each. In Part I, the phenotypic variation between different isolates of
P. horiana is discussed. Chapter 2 describes the different pathotypes that are present in a worldwide
collection of isolates. A bioassay for the high throughput screening of chrysanthemum cuttings was
developed and the methodology was used to determine the minimum number of resistance genes
and corresponding avirulence genes that are involved in the pathosystem. In Chapter 3 I investigate
the fungicide resistance of an international selection of isolates to five systemic fungicides.
In Part II I used molecular techniques to study the genetic variation in P. horiana and the molecular
detection of the pathogen. In Chapter 4 the genetic diversity within chrysanthemum white rust was
established on a worldwide selection of 45 isolates, including the isolates used in Chapter 2. Based
on a CRoPS™ analysis (van Orsouw et al 2007), we identified SNP loci and used them to genotype the
isolates and determine their phylogenetic relations. Chapter 5 focuses on the epidemiology and
biology of the pathogen using molecular detection of P. horiana in air samples by means of spore
sampling using two types of spore samplers and quantitative real time PCR. Using the optimized
techniques, the timing and abundance of spore release was determined at different distances from
infected plants.
Part III describes the segregation of resistance to P. horiana and the inheritance of AFLP markers in
chrysanthemum. In Chapter 6 we studied the segregation of disease resistance in crosses between
resistant and susceptible parent plants. Based on the resistance segregation in the progeny, we
determined the copy number of resistance genes for different pathotypes that are present in the
parent plants and linkage between the resistance to different pathotypes. In Chapter 7 we studied
the inheritance in Chrysanthemum in more detail using AFLP markers. Specifically, the mode of
inheritance and bivalent formation during meiosis was explored and linkage between molecular
markers and resistance to different pathotypes was checked.
4
Problem statement and outline
The final chapter in this thesis (Chapter 8) is the general discussion in which I summarize the most
important findings and contributions of this research project and highlight the prospects for future
research. The knowledge regarding fungicide resistance, pathotypes and segregation of disease
resistance will make it possible to optimize fungicide treatments and resistance breeding so as to
evolve to a sustainable control of the pathogen.
5
Chapter 1:
General introduction
Microscopic view of teliospores from a squashed pustule
General introduction
1.1 Chrysanthemum
1.1.1 History and economical value
Chrysanthemums belong to the family of Asteraceae with species native to Europe or Asia. The first
reports of cultivation of chrysanthemums for decorative and pharmaceutical purposes were made by
the Chinese philosopher Confucius around 500 B.C. (Dowrick 1953). From China, the cultivated forms
were spread to Korea and Japan in the fourth century A.D.. In these countries the plant became an
important cultural symbol. In Japan, breeding resulted in improved genotypes with great variation in
form and color, from which the current cultivars have been selected. It took till the end of the 17th
century before chrysanthemums were introduced in the Netherlands and later in Great Britain,
France and the United States. In the 19th and 20th century, the exchange of varieties and hybrids
continued, leading to the current diversity of cultivars (Dowrick 1953; Anderson 2006). Two main
types of gene pools, including garden and greenhouse gene pools, are used in chrysanthemum
breeding between which hybridization regularly is used for trait transfer (Anderson 2006).
The name chrysanthemum, indicated in Japanese or Chinese with the “⳥”-character, is derived from
the Greek words ‘chrysos’ (gold) and ‘anthos’ (flower) and therefore means golden flower
(Spaargaren 2002; Anderson 2006). One of the ancestors of the current chrysanthemum cultivars,
Chrysanthemum indicum, originated in Japan and China and was classified in the genus
Chrysanthemum by Lineaus because of the similarities with the small south European wildflower
Chrysanthemum coronarium. In a similar way, another Chinese ancestor of the current
Chrysanthemum cultivars, Chrysanthemum morifolium, was described by Ramatuelle (Anderson
1987; Spaargaren 2002). In the 1960’s, the Chrysanthemum genus contained more than 200 species
and was later split in 38 genera (Anderson 1987). Most of the commercially important species were
classified in the genus Dendrathema. As this reclassification caused a lot of confusion in many
countries, a proposal was made to the International Botanical Congress in 1995 to reclassify the
commercial chrysanthemums in the genus Chrysanthemum, with as type species Chrysanthemum
indicum. In 1999 this proposal was accepted and defined in a ruling of the International Code of
Botanical Nomenclature, restoring the economically most important hybrid, Chrysanthemum x
morifolium (syn. Dendrathema x grandiflorum) or florist chrysanthemum in the genus
Chrysanthemum (Brummitt 1997; Shea 2007). This hybrid was primary derived from the hexaploid
species C. indicum, but it includes ten or more hexaploid species including C. erubescens, C.
japonense, C. japonicum, C. ornatum and C. sinense, several of which are hosts of rust species
(Dowrick 1953).
9
Chapter 1
Worldwide, chrysanthemums are one of the most important floricultural crops in the cut flower as
well as the potted plant industries. Cut flowers are mainly grown in greenhouses or plastic tunnels
and the major producing countries are Japan, the Netherlands, Italy and Colombia with a production
area of 2950 ha, 753 ha, 655 ha and 500 ha, respectively (Spaargaren 2002). Potted plants are mainly
grown in greenhouses as disbudded flowers (one flower per stem) or spray flowers (several flowers
per stem), while garden mums (=multiflora mums) are in general grown outdoors. The main
producing countries are Italy, Belgium and Germany on an area of 524 ha, 472 ha and 290 ha,
respectively (Spaargaren 2002; Tierens 2007). The production of chrysanthemum cut flowers in Japan
was more than 2 billion stems per year in 1993 (Anderson 2006). In the main Dutch flower auction,
chrysanthemums ranked as the second most important cut flower (after roses) with a total of 1,453
million stems and a turnover of €332 million in 2008 while pot flowers are ranked 7th with a total of
34 million pots and a turnover of €30 million (Anon. 2009). Also in the United States
chrysanthemums are listed in the top ten of most sold horticultural crops with a wholesale value of
$11,769 million for cut flowers, $24,010 million for flowering potted plants and $116,063 million for
garden chrysanthemums (Anon. 2011). Belgium is one of the largest European producers of garden
mums with a yearly production of 11.5 million units and a production value of €14 million. Flanders is
responsible for 91% of this production where it represents up to 6 % of the production value of nonedible crops (Van Lierde et al. 1999).
1.1.2 Chrysanthemum breeding
Commercial chrysanthemums are perennial plants. Flower production is initiated when the daylight
period becomes less than approximately 11 hours. The time between flower induction and flowering
depends on the cultivar and ranges from 6 to 11 weeks (Spaargaren 2002). Several traits of the plant
have been attributed to specific genes, but these have not been mapped to chromosomes yet
(Anderson 2006). The traits that are mainly bred for are flower color and shape, plant architecture,
shorter flowering reaction times and disease resistance. Breeding is done by outcross pollination
because of self incompatibility and inbreeding depression (Anderson et al. 1992). Also, mutation
breeding using gamma or Röntgen radiation is performed to obtain cultivars that differ only in flower
color (Spaargaren 2002). Cultivars are vegetatively propagated.
Florists’ chrysanthemums are self-incompatible, allohexaploid hybrids showing an increasing
inbreeding depression after repeated self-pollination (Anderson et al. 1992). They are related to the
hexaploid Asian species while the European botanical chrysanthemum species are mostly diploid
(Dowrick 1952). The basic chromosome number in Chrysanthemum is nine (2n = 6x = 54), although
the somatic chromosome numbers in different cultivars can range from 2n = 47-63 between and
10
General introduction
within plants including both hexaploid and heptaploid types (Dowrick 1953). Most varieties have 5456 chromosomes with a peak at 2n = 54 (Dowrick 1953). This variation in chromosome number
between and within varieties is one of the factors playing a role in the different phenotypes and the
occurrence of visible mutations (“sports”) in certain cultivar (Dowrick 1953).
Although meiosis in hexaploid species is more complex than in diploids, chrysanthemums behave like
diploids as cells with 54 chromosomes consistently form 27 bivalents to form pollen with 27
chromosomes. Cells containing a supplementary chromosome do not form trivalents with a normal
bivalent (Dowrick 1952; Dowrick 1953). During meiosis, bivalent formation should be obtained by
preferential pairing of the homologous chromosomes (Watanabe 1981; Watanabe 1983). The
segregation of resistance to Puccinia horiana as studied by De Jong and Rademaker (1986) rather
fitted the expected ratios of preferential pairing, although the results were not absolutely convincing.
Another study on the segregation of carotenoid pigmentation suggested a random chromosome
pairing except for one cross in which preferential segregation was observed (Langton 1989). These
conflicting results can be an indication that a combination of these two pairing types occurs during
meiosis.
Since each somatic cell carries six copies of each chromosome, the number of copies of a given allele
at a given locus can vary from zero to six. The number of copies of each allele carried by the parent
plants and the way in which chromosomes pair during meioses determine the segregation ratios at
any single locus in progenies derived from single crosses. The hexaploid character of Chrysanthemum
complicates enormously the estimation and the comparison, of observed and expected segregation
ratios. The situation becomes even more complicated if we take into consideration that different
allelic variants can have different modes of action, ranging from complete dominance to complete
recessivity. In this PhD only situations were considered in which a particular allele (at an AFLP marker
locus or at a resistance gene locus) had a dominant mode of action, while all other allelic variants
were recessive. When more complex situations, including co-dominance or gradations in level of
dominance, should be taken into consideration, the number of different expected segregation
patterns would become untreatable due to the hexaploid character of Chrysanthemum.
For a particular trait determined by a dominant allele, segregation will only be observed in the
progeny of crosses between a parent containing one, two or three copies of the dominant allele with
a parent containing zero to three copies of the dominant allele. In all other situations, all progeny
plants will carry at least one copy of the dominant allele and no segregation will be observed for the
trait. Plants containing none, one, two or three copies of a particular allele at a locus are referred to
as nulliplex, simplex, duplex or triplex respectively for that particular locus (Figure 1.1).
11
Chapter 1
Figure 1.1: Illustration of the different numbers of dominant (z) and recessive ({) alleles that can have an influence on the
segregation of a dominant trait in chrysanthemum. In a nulliplex situation (a), each of the six chromosomes of a certain
homeologous group contains the recessive allele. In a simplex (b), duplex (c) or a triplex (d) situation, one, two or three
copies of a dominant allele are present respectively.
In allopolyploids, it is accepted that the pairs of homologous chromosomes that originate from one
of the ancestors of the hybrid, show preferential pairing during meiosis resulting in a disomic and
sexually stable segregation (de Silva et al. 2005). A well-known example of this situation is hexaploid
wheat (Triticum aestivum), with its three chromosome sets (A, B and D) (Petersen et al. 2006). A
diploid-like meiosis has also been described in hexaploid taxa of Chrysanthemum with preferential
formation of bivalents, while multivalents or univalents were rarely observed (Watanabe 1977).
Multivalent formation is probably prevented by restriction of pairing initiation to one single site or
zygomere (Watanabe 1983). According to this author, the extent of pairing between related
chromosomes in this genus depends on zygomere number and homology, but not on the
heterogeneity of structural genes throughout the entire chromosome. The degree of differentiation
of the DNA sequence of zygomeres between homologous and homeologous chromosomes
determines thus whether bivalent formation in a particular Chrysanthemum plant will be random or
preferential. This depends, in turn, on the plant’s origin and ancestry as this will determine the level
of sequence homology between zygomeres.
Taking this information into consideration, for the analysis of the crosses described in this study, we
assumed that only bivalents were formed during meiosis. As it is not yet clear for C. x morifolium
whether bivalent formation is preferential or random, both possibilities were considered. This has
been illustrated in Figure 1.2. In the case of preferential pairing only homologous chromosomes will
pair while homoeologous chromosomes will never form bivalents. As a consequence, gametes will
always carry one copy of each homologous chromosome (Figure 1.2). In the case of random pairing,
homoeologous and homologous chromosomes will be able to pair with each other resulting in up to
seven different gametic combinations carrying homologous and homoeologous chromosomes
(Figure 1.2).
12
General introduction
Figure 1.2: Illustration of possible combinations of homologous and homoeologous chromosomes that can be obtained
after preferential or random pairing. Homologous chromosomes are marked with a white, gray or black circle.
Homoeologous chromosomes are indicated with circles of different colors. In the case of preferential pairing, only
homologous chromosomes (same color) will form bivalents, resulting in gametes containing one copy of each homologous
chromosome that are homoeologous to each other (upper part of the figure). In gametes that are the result of random
pairing, seven combinations of homologous and homoeologous chromosomes can be found (lower part of the figure).
The different modes of bivalent formation result in different proportions of gametes carrying a
certain number of copies of the dominant allele at the locus of interest. In the case of a simplex
situation, no difference is expected between preferential and random pairing in the proportion of
gametes carrying the dominant allele (Table 1.1). However, when a duplex or a triplex parent are
considered, a clear difference between preferential and random pairing is expected for the
proportion of gametes carrying different numbers of copies of the dominant allele (Table 1.1). When
a dominant allele is present in a duplex or triplex conformation on chromosomes of the same
homologous group, all gametes will carry the dominant allele in the case of preferential pairing. As a
result, all the progeny plants of a cross using such a parent will be positive for the trait. This is also
the case when four or more copies of a dominant allele are present in the parent plant for both,
random and preferential pairing.
13
Chapter 1
Table 1.1: Expected proportion of gametes carrying different number of copies of the dominant allele, in function of type of
bivalent formation and the number of copies of the dominant allele carried by the parental plant. “A” denominates the
dominant allele; “a” denominates the recessive allele. The six chromosomes are presented as three homologous pairs. For
each situation the number of gametes carrying a given number of copies of the dominant allele on the total number of
different chromosome combinations is given as well as the percentage. The proportions mentioned for preferential pairing
of the duplex and triplex situation are based on the assumption that the dominant alleles are located on homoeologous
chromosomes. A duplex parent with dominant alleles located on homologous chromosomes will generate 100% simplex
gametes while a triplex parent will generate 50% simplex (Aaa) and 50% duplex (AAa) gametes. In these cases no
segregation of the trait will be observed in progenies derived from this parent.
Genetic constitution of gamete producing plant
Type of bivalent
formation
Preferential pairing
Gamete
constitution
Simplex (Aa aa aa)
Duplex (Aa Aa aa)
Triplex (Aa Aa Aa)
aaa
4/8 (50%)
2/8 (25%)
1/8 (12.5%)
Aaa
4/8 (50%)
4/8 (50%)
3/8 (37.5%)
2/8 (25%)
3/8 (37.5%)
AAa
AAA
Random pairing
1/8 (12.5%)
aaa
10/20 (50%)
4/20 (20%)
Aaa
10/20 (50%)
12/20 (60%)
9/20 (45%)
4/20 (20%)
9/20 (45%)
AAa
AAA
1/20 (5%)
1/20 (5%)
Based on the segregation of a dominant marker allele in the progeny of a pair cross, an estimation
can be made of the number of copies carried by the respective parents and the type of bivalent
formation that is involved during meiosis. This is done by comparing the observed and the expected
proportion of progeny plants carrying the dominant marker allele, for different possible scenarios.
1.1.3 Chrysanthemum pests and diseases
Due to the intensive cultivation of Chrysanthemum x morifolium, several pests and diseases can have
an important impact on the yield of the crop creating a certain threat. Important pests (Table 1.2) are
the western flower thrips (Frankliniella occidentalis), leafminers (e.g. Liriomyza trifolii), several aphid
species (e.g. Myzus persicae) and spider mites (Tetranychus urticae) that cause major damage to the
leaves and flowers (Spaargaren 2002; Visser et al. 2007). Control of these pests is mainly done by
chemical pesticide treatments or the introduction of predatory insects.
14
General introduction
Table 1.2: Most important chrysanthemum pests.
Pest (Common name)
Western flower thrips
Onion thrips
Greenhouse whitefly
Serpentine leafminer
Tomato leafminer
Pea leafminer
Green peach aphid
Cotton aphid
Chrysanthemum Aphid
Two spotted spider mite
Scientific name
Frankliniella occidentalis
Thrips tabaci
Trialeurodes vaporariorum
Liriomyza trifolii
L. bryoniae
L. huidobrensis
Myzus persicae
Aphis gossypii
Macrosiphoniella sanborni
Tetranychus urticae
The most important diseases on chrysanthemums are listed in Table 1.3, of which chrysanthemum
white rust (Puccinia horiana), powdery mildew (Erysiphe chichoracearum) and gray mold (Botrytis
cinerea) are the most important airborne pathogenic fungi. Development of powdery mildew and
gray mold can be controlled by regular fungicide treatments, but control of chrysanthemum white
rust can be more problematic as a few infected plants can have serious implications for export
shippings due to quarantine regulations.
Table 1.3: Chrysanthemum diseases.
Disease (Common name)
White rust
Common rust
Gray mold
Powdery mildew
Pink rot
Septoria leaf spot
Ascochyta ray blight
Ray blight
Verticillium wilt
Pythium root rot
Rhizoctonia stem rot
Fusarium wilt
Alternaria leaf spot
Bacterial leaf spot
Bacterial blight
Foliar nematode
Root-Knot nematode
Spotted wilt
Chrysanthemum stunt
Pathogen (Scientific name)
Puccinia horiana
Puccinia chrysanthemi
Botrytis cinerea
Erysiphe cichoracearum
Sclerotinia sclerotiorum
Septoria chrysanthemi
Ascochyta chrysanthemi
Phoma chrysanthemi
Verticillium albo-atrum; V. dahliae
Pythium sp.
Rhizoctonia sp.
Fusarium oxysporum
Alternaria sp.
Pseudomonas cichorii
Erwinia chrysanthemi
Aphelenchoides ritzemabosi
Meloidogyne spp.
Tomato spotted wilt virus
Chrysanthemum stunt viroid
15
Chapter 1
1.2 Rust fungi
Rust fungi are obligate plant parasites including species that can cause a major threat to field crops
and ornamentals. They form intracellular haustoria into the host cells for feeding and usually do not
kill their host (Voegele and Mendgen 2003). However, they can dramatically reduce plant health
resulting in a significant decrease of production of agricultural crops or, in case of floricultural crops,
in a decrease of the aesthetic value due to the presence of pustules (Agrios 2005). The common
name of these pathogens is derived from the typical orange or orange-red spore types that are
associated with many species within this group. The rust fungi are basidiomycetes that are classified
in the order of the Pucciniales (formerly Uredinales) in the class of Pucciniomycetes (formerly
Urediniomycetes) (Aime 2006; Aime et al. 2006). Within the Pucciniales, more than 7000 species are
distributed among more than 100 genera in 14 families. The genera Puccinia and Uromyces represent
4000 and 600 species, respectively, or approximately half of all known rust fungi (Cummins and
Hiratsuka 2003; Maier et al. 2003).
Rust fungi are host specific parasites that complete their complete life cycle either on one host
species (autoecious rusts), or on two unrelated host species on which specific spore forms are
produced (called heteroecious rusts) (Webster and Weber 2007). The life cycle of rust fungi can be
very complex compared to other fungi and can include up to five different spore types (spermatia,
aeciospores, urediniospores, teliospores and basidiospores) that are produced in a unidirectional
order (Hiratsuka and Sato 1982). In spring, the life cycle starts with the germination of the teliospores
which mainly serve as the survival stage. After germination of the teliospores, haploid basidiospores
are formed after meiosis. They are dispersed by the air and infect the same host in the case of
autoecious rusts or an alternative host in the case of heteroecious rusts. After successful inoculation,
spermatia (=pycniospores, produced in spermogonia or pycnia) will develop. Their function is the
fertilization of receptive hyphae of the compatible mating type in order to form an aecium with the
formation of dikaryotic aeciospores (Hiratsuka and Sato 1982). The diploid aeciospores are dispersed
by the air to infect the telial host on which urediniospores will develop. Urediniospores are dispersed
by the wind as well, but infect the same host species as the one on which they developed, and
multiple such cycles can occur per year. In autumn, teliospores are formed to survive the winter.
Macrocyclic rusts include all these spore forms in their life cycle, except for species with a brachyform life cycle in which the acial stage is deleted or the aeciospores are morphologically
indistinguishable from urediniospores (Petersen 1974; Ono 2002). Species that only produce a telial
and basidiospore stage with or without a spermogonial stage on a plant are said to be microcyclic
rusts (Petersen 1974; Hiratsuka and Sato 1982). During the evolutionary reduction of the life cycle of
a macrocyclic heteroecious rust to a microcyclic rust, the telial stage of the microcyclic rust can be
16
General introduction
found on the aecial host of the macrocyclic rust from which it evolved. This phenomenon is now
referred to as “Tranzschel’s law” (Tranzschel 1905).
Worldwide, rusts are one of the economically most important pathogens of vascular plants with host
ranging from ferns and gymnosperms to families of monocotyledons and dicotyledons (Hiratsuka and
Sato 1982; Webster and Weber 2007). From an economical point of view, the most important rusts
are those that have cereals and grasses as hosts. These include Puccinia graminis or stem rust on
wheat and other small grains, Puccinia striiformis or stripe rust on wheat, barley and ryegrass,
Puccinia triticina or leaf or brown rust on wheat and ryegrass and Puccinia coronata or crown rust on
oats and ryegrass. Also in other crops rust can have an important impact on yield, including Puccinia
stakmani on cotton, Phakopsora pachyrhizi on soybean, Uromyces appendiculatus on bean and
Hemileia vastatrix on coffee (Agrios 2005). Their economic impact resulted in some rust fungi being
recognized as quarantine organisms, to prevent their introduction in pathogen-free areas through
international trade. However, these quarantine restrictions, including a post entry quarantine of
several months, and eradication efforts can also have a significant impact on agricultural and
floricultural production (Wise et al. 2004). In addition, detection of rusts during their latent phase is
often problematic, allowing undesired spread of the pathogen with symptomless plant and
propagation material (Stebbins and Johnson 2001).
Rust fungi that are morphologically identical but that are specific to different host genera are called
formae speciales (Agrios 2005). Besides these special forms, also physiological races can be observed
that can only attack certain genotypes of a host species. These physiological races can only be
distinguished by a set of differential genotypes and are described in various rust fungi such as
Puccinia recondita (Pfender 2009), U. appendiculatus (Stavely et al. 1989; Ochoa et al. 2007) and P.
graminis tritici (Singh et al. 2006). The presence of these physiological races can be explained by the
gene-for-gene concept as described by Flor (1956) for the rust species Melampsora lini and its host
Linum usitatissimum (flax). This concept states that for every resistance gene in the host there is a
matching avirulence gene in the pathogen. The avirulence gene products, recently referred to as
elicitors, often play an important role in the infection process as effector molecules (Bent and
Mackey 2007) and are shown to be secreted in the host cells by developing rust haustoria
(Catanzariti et al. 2006). The exact role and identity of these molecules was recently described for the
rust fungi Melampsora laricipopulina and Puccinia graminis f. sp. tritici (Duplessis et al. 2011). When
an elicitor is recognized by a host defense receptor, coded by a resistance gene, the host will start up
a cascade of defense reactions leading to a hypersensitivity reaction (Staskawicz et al. 1995). This
hypersensitive response triggers local and systemic biochemical defense mechanisms against the
pathogen leading to resistance of the host (Heath 2000). Since multiple avirulence genes and
17
Chapter 1
corresponding resistance genes can be present in gene-for-gene pathosystems, a set of pathogen
genotypes (pathotypes) can interact with a set of host genotypes, resulting in differential reactions as
illustrated in Figure 1.3 (Flor 1956; Flor 1971). Within different well studied host-pathogen systems,
several resistance genes and corresponding avirulence genes have already been identified by genetic
and molecular analysis. In the P. graminis - Triticum aestivum (wheat) system, more than 40
corresponding gene combinations have been identified (Leonard and Szabo 2005). For the M. lini – L.
usitatissimum (flax) complex, 31 genes at 5 loci have been designated (Ellis et al. 2007). However,
rusts, like many other pathogens, have the ability to overcome resistance based on the loss of
elicitors and as it is difficult to determine whether all involved genes are present in the test
population, the number of genes involved in a pathosystem cannot be determined on the long term
(Jones and Dangl 2006).
Figure 1.3: Illustration of the gene-for-gene concept in an example with three elicitors/receptors. Different pathotypes
express a diversity of elicitors (illustrated in red) while different cultivars express a diversity of pathogen receptors
(illustrated in green). A plant will only be able to express resistance when at least one pathogen elicitor is directly or
indirectly recognized by its corresponding receptor protein (indicated by at least one green V sign). When no elicitor is
recognized, the pathogen will be able to grow and cause disease. Different elicitor combinations in different isolates of the
pathogen and different receptor combinations in the host cultivars result in differential virulence and resistance patterns in
pathogen and host, respectively.
1.2.1 Chrysanthemum rust diseases
Several rust species have been identified on plant species within the genus Chrysanthemum and
related genera (Table 1.4)(Hiratsuka 1957; Punithalingam 1968a; Punithalingam 1968b; Gardner and
18
General introduction
Hodges 1989). The only rust species of economical importance for the cultivation of florist
chrysanthemums are Puccinia horiana and to a lesser extent Puccinia chrysanthemi, both belonging
to the family of the Pucciniaceae. P. horiana is native to Japan and China and also P. chrysanthemi is
said to be very common in Japan )(Punithalingam 1968a; Punithalingam 1968b). P. horiana was first
observed in Japan in 1895 (Hiratsuka 1957) while the first report of P. chrysanthemi was made in
England in 1895 (Hiratsuka 1957; Punithalingam 1968a).
Table 1.4: Rust species described on Chrysanthemum and related host species
Rust species
Host range
Puccinia horiana Henn.
See Table 1.5
P. chrysanthemi Roze
Argyranthemum frutescens;
(= P. chrysanthemi-chinensis Roze and Uredo Dendranthema boreale; D. grandiflorum;
D. lavandulifolium
chrysanthemi Roze)
P. tanaceti DC.
A. frutescens, D. boreale; Tanacetum spp;
Artemisia spp.
P. baschmica Petr.
Chrysanthemum myriophyllum
P. crysanthemicola Camara, Oliveira & Luz
C. coronarium
P. leucanthemi Pass.
Leucanthemum vulgare
(=P. asteris var. chrysanthemi-leucanthemi
Massal.)
P. heeringiana Kleb
A. frutescens; C. parthenium; T. parthenium
P. pyrethri Rabenh.
Tanacetum corymbosum
(= P. seriata Syd.; P. proximella Syd. & P. Syd.)
P. gaeumanni Major
Tanacetum cinerariifolium
Phakopsora artemisiae Hirats.
(= Uredo autumnalis Diet.)
Uredo neocomensis Mayor
P. aecidia-leucanthemi Ed. Fisch. (= Aecidium
leucanthemi DC. and P. leucanthemi-vernae
Gäumann)
1.2.1.1 Puccinia chrysanthemi
Puccinia chrysanthemi Roze is the causal agent of chrysanthemum brown rust, also referred to as
chrysanthemum common rust or black rust, and is common in Japan, but it is also sporadically
observed in other chrysanthemum producing regions worldwide. It is an autoecious rust growing on
Chrysanthemum x morifolium with a brachy-form life cycle developing spermogonial, uredial, telial
and basidial stages (Harada et al. 1996). The disease generally occurs in the late summer, infecting
the leaves and sepals of the host (Horst and Nelson 1997). The optimal temperature range for
urediniospore germination is 15 to 21°C (Gibson 1904; Campbell and Dimock 1955). The first
symptoms appear as yellow flecks on the upper and lower surface of the leaves that develop to
typical brown powdery pustules on the underside of the leaves. In the case of persistent infection, a
ring of secondary pustules can develop around the original pustules, bearing teliospores which die
19
Chapter 1
and turn black (Horst and Nelson 1997). In case of severe infections, the plant can become
defoliated, produce less flowers and can show stunting, affecting the commercial value of the crop.
In Europe and America, the urediniospores can overwinter on fallen leaves or in the soil. Hence it can
overwinter in these regions without the need for resting spores (Arthur 1900; Gibson 1904;
Punithalingam 1968a). In Japan, the formation of brown to dark brown teliospores together with the
uredinial stage has been reported (Punithalingam 1968a). Although epidemic development of the
rust is occasionally reported, the impact on chrysanthemum cultivation is limited (Punithalingam
1968a).
1.3 Puccinia horiana
Puccinia horiana Henn. causes chrysanthemum white rust or Japanese rust on more than 10 different
chrysanthemum species including Chrysanthemum x morifolium (Table 1.5). Phylogenetically, P.
horiana clusters with the economically important macrocyclic rusts with telial host in the Poaceae,
unlike the other rusts on chrysanthemum that cluster with other rust that have Asteraceae as telial
host (Alaei et al. 2009a). Following Tranzschel’s law, P. horiana probably evolved from a macrocyclic
rust that had chrysanthemum as its aecial host (Tranzschel 1905).
The first known observation of P. horiana was made in Japan in 1895 (Hiratsuka 1957). The species
was described taxonomically by Hennings (1901). From Japan it first spread to China and South Africa
sometime before 1963, from where it rapidly spread to other chrysanthemum producing areas
(Water 1981; Priest 1995). The first infection in England was reported in 1963 on plants originally
imported from Japan (Baker 1967). In 1964, a second introduction in western Europe occurred on
plants imported from South Africa (Baker 1967; Water 1981; Priest 1995). In the following years, the
pathogen further spread to the rest of Europe with reports in Germany (Stark and Stach 1965),
Norway (Gjaerum 1964), Sweden (Nilsson 1964), Denmark and Belgium (Jorgensen 1964). Currently,
the pathogen is present in many European countries as well as in countries with a substantial trade
of cut flowers and cuttings (Figure 1.4)(Rattink et al. 1985; Whipps 1993). As even low levels of the
disease can result in an unmarketable crop with major economic consequences, this rust is described
as the most serious disease on chrysanthemums. Presently, chrysanthemum white rust is classified as
a quarantine pathogen in most continents that have chrysanthemum producing regions (EPPO 2004).
20
General introduction
Table 1.5: Known host plants of Puccinia horiana with their current and previous scientific names and their common names
(Shea 2007).
Accepted name
Chrysanthemum arcticum
C. boreale
C. indicum
C. japonense
C. japonicum
Chrysanthemum × morifolium
C. pacificum
C. shiwogiku
C. yoshinaganthum
C. zawadskii subsp. yezoense
C. zawadskii subsp. zawadskii
Leucanthemella serotina
Nipponanthemum nipponicum
Synonyms
Arctanthemum arcticum,
Dendranthema arcticum
C. indicum var. boreale, D. boreale
D. indicum
D. japonense,
D. occidentali-japonense
C. makinoi, D. japonicum
Anthemis grandiflorum,
A. stipulacea, C. sinense ,
C. stipulaceum, D. × grandiflorum, D.
× morifolium, Matricaria morifolia
Ajania pacifica, D. pacificum
Ajania shiwogiku, D. shiwogiku
D. yoshinaganthum
C. arcticum subsp. Maekawanum, C.
arcticum var. yezoense,
C. yezoense, D. yezoens,
Leucanthemum yezoense
C. sibiricum, D. zawadskii,
D. zawadskii var. zawadskii
C. serotinum, C. uliginosum,
Pyrethrum uliginosum
C. nipponicum,
Leucanthemum nipponicum
Common name
Arctic chrysanthemum,
arctic daisy
Noji-giku
Ryuno-giku
Florist's chrysanthemum,
chrysanthemum, Mum
Iso-giku
Shio-giku
Giant daisy, High daisy
Nippon daisy, Montauk daisy,
Nippon - chrysanthemum
Figure 1.4: Distribution map of P. horiana based on the map in the Invasive Species Compendium
(http://www.cabi.org/isc/?compid=5&dsid=45806&loadmodule=datasheet&page=481&site=144) and the EPPO distribution
map of P. horiana (http://www.eppo.org)
21
Chapter 1
1.3.1 Puccinia horiana disease symptoms and spore morphology
Primarily the younger leaves are infected by the pathogen. In case of severe infection, leaves will roll
up or twist and finally die although dead leaves remain attached to the stems. The first symptoms
appear as light green or yellow spots on the upper surface of the leaves which gradually enlarge to
approximately 5 mm in diameter. As the infection progresses, the spots on the underside of the
leaves become raised and buff, after which pinkish, waxy pustules are formed (2 to 5 mm in
diameter). The pustules consist of teliospores that eventually produce basidiospores (sporidia), giving
the pustules a whitish aspect. In addition to the teliospores developing on the underside of the
leaves, pustules can occasionally be found on the upper side of the leaves, or in the case of a severe
infection, also on the bracts, stipula, stems and flowers (Figure 1.5). Apart from the cultivars of the
florist’s chrysanthemum, which is widely cultivated throughout the EPPO region, several other
chrysanthemum species show symptoms and signs after artificial inoculation (Table 1.5). However,
variation in susceptibility and symptoms can be observed between species and cultivars (Yamada
1956; Hiratsuka 1957; Dickens 1968; Sugimura et al. 1998; Wojdyla 1999b).
Figure 1.5: Symptoms of chrysanthemum white rust. The pathogen usually infects the leaves. Infected leaves will show
chlorotic flecks on the upper side (A) and pustules on the lower side (B). In case of severe infection, pustules can also be
found on the upper side of the leaves (C), on stipula (D) and stems (C).
22
General introduction
The pustules at the underside of infected leaves contain the telia in which the hyaline teliospores
develop on a hyaline, persistent pedicel of up to 45 ђm long. The teliospores of 32-45 ђm by 12-18
ʅm large are bicellular, cylindrical, fusiform, oblong to oblong clavate (rarely elliptical) and slightly
constricted in the middle. They have a smooth pale yellow cell wall of 1-2 ђm at the sides, and 3-10
ђm at the apex. In situ one or both cells of the teliospore germinate with the formation of a
promycelium from which two round, kidney-shaped or slightly curved basidiospores (12-18 ʅm × 915 ʅm), showing a hilum at one side, are formed (Figure 1.6).
Figure 1.6: Spores of P. horiana. Teliospores develop in the pustules on the lower side of the leaves. Under high relative
humidity and cool temperature conditions they germinate with the formation of a promycelium and basidiospores (B).
Basidiospores are dispersed by the air and germinate on new host tissue by forming of a germ tube (C). Bs = Basidiospore,
Pm = Promycelium, Ts = Teliospore, Gt = Germ tube, scale bars = approx. 10 ђm.
1.3.2 Puccinia horiana life cycle
As a microcyclic autoecious rust, P. horiana only produces teliospores and basidiospores and
completes its life cycle on a single host (Figure 1.7). The teliospores do not need a period of
dormancy and germinate in situ under favourable conditions with the formation of basidiospores
(Firman and Martin 1968). The basidiospores are dispersed by the air and can infect other
chrysanthemum leaves under conducive conditions to complete the life cycle. Teliospores can
germinate and discharge basidiospores at a temperature ranging from 4°C to 24°C according to
Firman and Martin (1968)and from 6°C to 36°C according to Yamada (1956) with an optimum at 17°C.
For basidiospore germination the optimum was observed at temperatures between 10°C and 20°C
although for penetration, a temperature range from 4°C to 24°C, with an optimum of 17-24°C is
reported (Firman and Martin 1968). A high relative humidity (RH) and a water film on the leaf surface
are necessary to induce germination of both the teliospores and basidiospores (Uchida 1983). Under
these optimal conditions, basidiospore discharge from the pustules can start within 3 hours and once
released they can germinate immediately with the formation of a germ tube. This germ tube can
penetrate the upper or lower leaf surface within 2 hours, which means that under optimal conditions
a new infection can become established within 5 hours (Firman and Martin 1968).
23
Chapter 1
Figure 1.7: Schematic illustration of the life cycle of P. horiana on Chrysanthemum x morifolium. Ts = Teliospore; Bs =
Basidiospore; Pm = Promycelium; Gt = Germtube; App = Appresorium; H = Haustorium; Hm = Haustorial mothercell; Vm =
Vegetative mycelium.
After penetration of the leaf, abundant hyaline intercellular hyphae are produced with intracellular
haustoria (Firman and Martin 1968; EPPO 2004). Eight to 10 days after infection the first symptoms
appear as chlorotic flecks that develop telia-bearing pustules 4 to 6 days later so that it takes at least
14 days to complete the life cycle (Zandvoort et al. 1968a). As conditions generally deviate from this
optimum, completion of the life cycle can take 4 weeks. Wide temperature variations between night
and day and periods of temperatures higher than 30°C can even prolong the incubation time to 8
weeks (Zandvoort et al. 1968a).
Dispersal of the pathogen occurs by wind over distances of at least 700 m (Zandvoort 1968) although
short distance dissemination by water splashing may be possible. Basidiospores are very sensitive to
desiccation: at 90% RH they lose their infectivity after 1 hour and at 80% RH after 5 min (Firman and
Martin 1968). Therefore, moist weather conditions are necessary for long distance dispersal (Firman
and Martin 1968; Horst and Nelson 1997). As a result, long distance dispersal is mostly due to the
movement of infected plants.
The survival of the teliospores is highest at low relative humidity with a survival of up to 30 days at
32% RH as described by Yamada (1956) or a survival of seven to eight weeks at 50% RH as reported
by Firman and Martin (1968).When the teliospores are kept in more humid conditions or buried in
compost their survival is limited to 3 weeks (Firman and Martin 1968). As a result, carryover of the
24
General introduction
disease by infected plant debris seems not to be important. However, these survival experiments
were conducted at high temperatures. It is unknown if the pathogen can overwinter outdoors at
lower temperatures (Punithalingam 1968b).
1.3.3 Nuclear cycle
Eight variations in the nuclear behaviour during teliospore germination and basidiospore formation
have been described for rust fungi (Hiratsuka and Sato 1982). A more recent study with a focus on
microcyclic rusts even described 9 types with 11 variations in the nuclear cycle and basidiospore
development, including one that is believed to occur in P. horiana (Ono 2002).
Since no spermatial stage is present in the life cycle of P. horiana, it is assumed that it lost its
capability for sexual recombination (Ono 2002). During its nuclear cycle (Figures 1.8 and 1.9) two
binucleate basidiospores are formed per teliospore cell (Kohno et al. 1974). Immature teliospores
contain two haploid nuclei per cell. These nuclei undergo karyogamy during maturation, with the
formation of a diploid nucleus. This nucleus migrates into the promycelium as the teliospore
germinates after which a first meiotic division occurs. The two resulting nuclei are separated by a
septum and subsequently undergo a second meiotic division, after which no additional septum is
formed. Each promycelium cell gives rise to a basidiospore that receives the two haploid nuclei from
its corresponding promycelium cell (Kohno et al. 1974; Kohno et al. 1975). Unlike many other rusts
with binucleate basidiospores, the nuclei in P. horiana are haploid meiotic diads and not the result of
a mitotic division (Anikster 1983; Ono 2002). During germination, basidiospores undergo a first
mitotic division giving rise to a binucleate vegetative mycelium (Kohno et al. 1975; Ono 2002).
P. horiana can be considered a sexually heterokaryotic, secondarily homothallic rust with a bisexual
basidiospore containing the nuclei of both sexes (Buller 1941). Homothallism implies that
populations of P. horiana can be considered as genetically homogeneous populations with a reduced
frequency of outcrossing (Carlile 1987). This is also illustrated in Figure 1.8 where a homozygous
strain can be considered a clonal line as the basidiospores contain identical parental and progeny
genomes (Taylor et al. 1999). When two basidiospores of different genotypes can germinate in close
proximity, heterozygous conditions at any gene locus can be expected through heterokaryosis after
anastomosis resulting in the exchange of nuclei between different germ tubes (Ono 2002; Wang and
McCallum 2009). Also, within populations, chromosomal mutations in one or both nuclei of a
vegetative mycelium can result in two genetically different haploid nuclei (Ono 2002). Figure 1.9
illustrates how a heterozygous vegetative mycelium evolves into two homozygous genotypes.
25
Figure 1.8: Illustration of the normal nuclear cycle of Puccinia horiana starting at the bottom of the figure with mycelium containing two identical haploid nuclei per cell, which is considered the standard nuclear
condition. Homozygosity is represented by the same color of the nuclei. Per cell of each teliospore, two basidiospores can be formed that each contains two haploid nuclei, identical to those of the mycelial
source cell, even in case of crossing over. The different events are described next to the figure arrows. Within the text boxes, the information on the nuclear DNA content is given (number of chromosome copies
(c) and haploid (n) or diploid (2n) state of each nucleus). Cycle based on (Kohno et al. 1974; Kohno et al. 1975).
Figure 1.9: Illustration of the nuclear cycle of Puccinia horiana in case different haploid nuclei were established in the mycelium after mutation or anastomosis. Heterozygosity is represented by different color
nuclei. New basidiospores formed will be homozygous (= normal scenario, depicted for the upper teliospore cell), except in case of crossing over during meiosis (= scenario depicted for the lower teliospore cell).
The different events are described next to the figure arrows. Within the text boxes, the information on the nuclear DNA content is given (number of chromosome copies (c) and haploid (n) or diploid (2n) state of
each nucleus). Cycle based on (Kohno et al. 1974; Kohno et al. 1975).
Chapter 1
A heterozygous mycelium will give rise to teliospores that will have a heterozygous diploid nucleus
after karyogamy. The majority of those nuclei will undergo meiosis without crossing over, resulting in
basidiospores that will be homozygous for either one of the genotypes. The progeny will further
behave as a clonal line as described above until a new heterozygous condition occurs due to
mutation or anastomosis with another genotype. Only in a limited number of nuclei crossing over will
occur during meiosis maintaining the heterozygous condition. Although these processes result in
relatively homogeneous populations, they can maintain genetic diversity in homothallic microcyclic
rusts like P. horiana (Ono 2002).
1.3.4 Control measures
There are several control measures that can be applied for the management of chrysanthemum
white rust. A primary strategy is to avoid the introduction and spread of the pathogen in and
between chrysanthemum growing areas through national and international quarantine measures.
This especially relates to starting each growing period with pathogen-free cuttings. Within the
nurseries, cultural and physical techniques can be applied to minimize the presence of conducive
environmental conditions. Furthermore, chemical control by means of fungicide applications and
breeding for resistant cultivars are powerful tools to control the disease. Alternative control options
such as biological control and induction of systemic acquired resistance have also been attempted.
However, some of these strategies have economical, environmental or social drawbacks. The
different control options are discussed below.
1.3.4.1 Quarantine
As chrysanthemum white rust is mainly introduced in nurseries via infected plant material, P. horiana
is registered as a quarantine pest by the EPPO, IAPSC, CAN (previously JUNAC) and the North
American Plant Protection Organization (NAPPO) (EPPO 2004). Also within nurseries it is very
important to implement quarantine measures for produced and received cuttings to avoid
transportation of infected cuttings and import of rust containing cuttings, cut flowers and potted
plants. Despite the ease in diagnosing sporulating lesions on infected plants, plants bearing small
non-sporulating lesions or latently infected plants remain undetected. Especially in large shipments,
the detection of a small number of lesions by visual inspection is very difficult and can result in a false
negative report with consequent import of the pathogen. The implementation of molecular
diagnostic techniques for the specific detection of the pathogen in bulked samples should be very
powerful to enforce quarantine control. Specific conventional PCR and quantitative real-time PCR
assays for the detection of the pathogen in latently infected plant material have been described
(Pedley 2009; Alaei et al. 2009b).
28
General introduction
1.3.4.2 Cultural and physical control
Within nurseries, cultural control is an important strategy to avoid the development of the disease.
As an essential aspect of this cultural control concerns climate control, these strategies are mainly
applied in greenhouse cultivation. As high relative humidity and extended leaf wetness are crucial
factors in the survival and infection process of P. horiana, it is very important to control these
parameters (Hellmers 1964; Huber and Gillespie 1992). To avoid a high relative humidity and foliar
wetness, watering is recommended early in the day, to allow the leaves to dry quickly. Also the use
of drip- and sub-irrigation systems instead of overhead watering will have an impact on foliar
wetness (Grouet et al. 1981). Relative humidity can also be reduced by removing the blackout
screens during the night and the use of portable fans for forced air movement under the blackout
screens (Dickens and Potter 1983). Also additional heating can be used for this purpose, although
heating in combination with ventilation requires high amounts of energy (Korner and Challa 2003).
Due to the increasing energy costs, this approach is reduced in present chrysanthemum cultivation
(Brandwagt, pers. comm.).
Once the disease is present in a crop, physical destruction of symptomatic plants can be an effective
control strategy. Symptomless plants need to be followed up to avoid further development of the
disease (Baker 1967). Nevertheless, this eradication strategy has been implemented at a good costbenefit ratio in many countries of the world (Lelliott 1984; Klassen 1992). Also heat can be used to
eradicate the disease from infected stools, killing the pathogen with a treatment of 8 hours at 45°c or
48 hours at 35°C. However, the margin between effective disease control and death of the plants is
narrow (Zadoks et al. 1968). Also dipping of the plants for 5 minutes in a water bath at 45°C has been
reported to be successful for eradication of chrysanthemum white rust and was even more
successful than dipping the cuttings in a fungicide solution (Dickens 1978; Coutin and Grouet 1983).
Due to the technical difficulties related to those techniques, it can only be applied on individual
plants and on cultivars with a sufficient heat resistance (Whipps 1993).
1.3.4.3 Biological control and induction of systemic acquired resistance
A sustainable approach for the control of P. horiana can be the use of hyperparasitic organisms that
are able to infect the pathogen. Verticillium lecanii is a fungus that is available as a biological control
agent against aphids and white flies in greenhouse crops but also showed to be a hyperparasite of
rust species (Srivastava et al. 1985; Sheroze et al. 2003). It is suggested that this pathogen can be
used for an integrated insect and white rust control (Whipps 1993). The hyperparasite is able to
colonize up to 90% of the teliospores within 5 days (Srivastava et al. 1985) but other studies report
lower efficacies of around 50% control (Rodriguez-Navarro et al. 1996). The treatment of infected
29
Chapter 1
chrysanthemums with Cladosporium sp. also resulted in a reduction of the disease and subsequent
sporulation (García-Velasco et al. 2005). The difficulties in applying these hyperparasitic fungi are
that they require specific environmental conditions that can only be controlled in greenhouses. Most
contradictory, these conditions are similar to the optimal conditions for the development of P.
horiana (Whipps 1993). Given the large number of teliospores formed, the relatively limited activity
of these biocontrol organisms and the speed at which basidiospores are formed and infect, question
is how effective these biocontrol agents can be under practical conditions.
Alternatively, the disease can also be controlled by the induction of the systemic acquired resistance
of the plant with biologically active agents, like chitosan that serve as elicitors of the plant defense
response (El Ghaouth et al. 1992). This induction results in an accumulation of chitinases, synthesis of
proteinase inhibitors and lignification of some plant organs, inhibiting the penetration of the fungus
(El Ghaouth et al. 1992). Chitosan is shown to successfully protect chrysanthemums against P.
horiana resulting in a strong decrease of the number of infected leaves and the number of pustules
(Wojdyla 2004a). The bioactive agent acibenzolar-S-methyl (ASM) or benzo(1,2,3)thiadiazole-7carbothioic acid-S-methyl ester (BTH) (Bion®) is known to induce systemic acquired resistance against
rusts on wheat (Gorlach et al. 1996), and recently the efficacy of ASM for the treatment of
chrysanthemums against P. horiana has been showed (Guerrand et al. 2011).
1.3.4.4 Fungicides and fungicide resistance
Currently, P. horiana is mainly controlled by preventive spraying with fungicides as also suggested by
EPPO (2004), although it is costly (Water 1981), it can be hazardous for the environment and there is
a loss of social acceptance. Several studies on the chemical control of chrysanthemum white rust
have shown the efficacy of single target fungicides such as triazoles (e.g., myclobutanil,
propiconazole) or strobilurins (e.g., azoxystrobin, kresoxim-methyl) and multi target fungicides such
as chlorothalonil and mancozeb (Dirkse 1980; Rattink et al. 1985; Dickens 1990; Wojdyla and
Orlikowski 1999; Wojdyla 1999a; Wojdyla 2004a; Wojdyla 2005; Wojdyla 2006). However, there is a
variation in the degree of control and consequently different recommendations for use. Weekly
preventive spraying with triazole or strobilurin fungicides was found to control and eradicate this
pathogen (Lam and Lim 1993; Bonde et al. 1995; Oneill and Pye 1997).
The major drawback of regular spraying with single target fungicides is the risk of development of
resistant pathogen strains. More than 30 years ago, the first strains resistant to benodanil and
oxycarboxin were reported in France, Japan and the Netherlands (Abiko et al. 1977; Grouet et al.
1981; Dirkse et al. 1982). In the Netherlands, strains resistant to propiconazole, triforine and
30
General introduction
bitertanol were detected (Whipps 1993). Strains with resistance to both triazole and strobilurin
fungicides have been reported (Cook 2001).
Surprisingly, P. horiana is the only rust so far for which strobilurin resistance is reported although
wheat rusts are also treated with strobilurins on a regular basis. Strobilurins specifically inhibit the
mitochondrial respiration of fungi by binding to the cytochrome bc1 complex at the Qo site (Bartlett
et al. 2002). Mainly two mutations in the cytochrome b (cyt b) gene (G143A and F129L) confer
strobilurin resistance without affecting the enzyme activity in several plant pathogenic fungi (Bartlett
et al. 2002). However, in rusts the G143A mutation is lethal. The F129L mutation, nor other described
mutations linked to QoI resistance could be found in rusts indicating that alternative resistance
mechanisms such as alternative respiration or efflux transporters are involved in strobilurin
resistance in P. horiana (Grasso et al. 2006a; Grasso et al. 2006b; Fernandez-Ortuno et al. 2008).
Triazole fungicides are sterol demethylation inhibitors and act on sterol 14ɲ -demethylase (CYP51),
an essential enzyme in the biosynthetic pathway of ergosterol, the predominant sterol in plasma
membranes of most fungi. In several plant pathogenic fungi, overexpression of the CYP51 gene as
well as point mutations are reported to contribute to triazole resistance (Ma and Michailides 2005).
In rusts there is only limited information about these resistance mechanisms. However, a report on
the response of several isolates of P. triticina to the triazole fungicide epoxiconazole showed no
correlation between fungicide resistance and point mutations or gene expression, indicating
alternative mechanisms have to be involved (Stammler et al. 2009).
1.3.4.5 Host plant resistance and resistance breeding
From the first reports of P. horiana in Europe in the 60’s, several studies clearly showed that certain
species and cultivars are more resistant to P. horiana than others (Baker 1967; Dickens 1968; Martin
and Firman 1970; Yamaguchi 1981; Leu et al. 1982; Rademaker and Dejong 1985; Rademaker and de
Jong 1987; Hahn 1989; Sugimura et al. 1998). Four types of interaction between chrysanthemum and
white rust have been described: susceptibility, necrosis, incomplete resistance and complete
resistance (Rademaker and de Jong 1987). In susceptible cultivars, teliospore pustules, which can
vary in size and number depending on isolate and cultivar, develop and sporulate freely. In the case
of necrosis, the plant reacts with a retarded hypersensitivity response resulting in necrotic flecks on
the leaves but usually this response is too slow to completely prevent sporulation (Rademaker and
de Jong 1987). A plant showing incomplete resistance is able to slow down the whole infection
process although the pathogen can complete its life cycle and produce a limited number of small
sporulating pustules. Penetration into the leaves seems to be difficult and infection requires a large
amount of inoculum and it is believed that several genes are involved in this resistance process
31
Chapter 1
(Rademaker and de Jong 1987; Whipps 1993). When a plant is completely resistant, cells that
become infected by the pathogen immediately die resulting in the death of the biotrophic pathogen
as well. This process is also referred to as hypersensitivity and is regulated by a gene-for-gene
interaction (Keen 1990). Segregation experiments in the Netherlands revealed that a single dominant
gene should be involved in this type of interaction in the chrysanthemum-rust pathosystem (de Jong
and Rademaker 1986). As this type of resistance is most easy to breed for, it is mostly used in
chrysanthemum breeding programs, although breeding for horizontal resistance would be more
sustainable (Poland et al. 2009). Resistance to P. horiana seems not linked to other genetic
characteristics such as flower color and size, day length response, flowering reaction time, and plant
morphology (Hahn 1989).
Soon after P. horiana appeared as a major problem for chrysanthemum growers in Europe, the
availability of resistance in local host cultivars was studied in the United Kingdom (Baker 1967;
Dickens 1968; Martin and Firman 1970; Dickens 1971), Germany (Stark and Stach 1965), the
Netherlands (Boerema 1964), Sweden (Nilsson 1964), Denmark (Anon. 1964) and Norway (Gjaerum
1964). However, these studies often report discrepancies in the level of susceptibility of specific
cultivars (Table 1.6), suggesting the presence of pathotypes (physiological races), most likely due to
gene-for-gene interactions (Flor 1956; Stukenbrock and McDonald 2009). This is also supported by
more recent studies reporting similar differential reactions in resistance (Orlikowski et al. 1982;
Norman et al. 1995; Wojdyla 1998; Wojdyla 1999b). The first report of a systematic testing of
multiple isolates on different cultivars was done in Japan, involving 40 cultivars and six Japanese
isolates (Yamaguchi 1981). This study clearly revealed differential interactions in 14 cultivars and
confirmed the differential interactions that were noticed for cultivar Fred Shoesmith in previous
studies (Table 1.6). A more recent study in Mexico, in which 16 local isolates were included showed a
clear variability in aggressiveness and in virulence, although no differential interactions were
described (Velasco et al. 2007).
Despite the fact that resistance can be controlled by a single dominant gene (de Jong and Rademaker
1986) for a given isolate, the presence of pathotypes indicates that more than one gene is involved in
the pathosystem (Flor 1956; Ellis et al. 2007). The presence of pathotypes is one of the main
challenges in resistance breeding as it is difficult to predict which plant genes will be necessary to
offer resistance against the future pathotypes of the pathogen. This can be solved by the
implementation of an anticipatory breeding strategy as suggested for rust diseases in wheat by
McIntosh and Brown (McIntosh and and Brown 1997). Such a strategy requires a good knowledge of
the epidemiology of the pathogen, an annual pathogen survey to detect new pathotypes, knowledge
of the main resistance genes and a well coordinated system for screening of breeding material with
32
General introduction
pathotypes posing the greatest threat. Once resistance genes in the host are characterized by
molecular techniques, markers developed for these genes can be a powerful tool in marker-assisted
breeding (Kumar 1999).
Stark & Stach
1965
Gjaerum
1964
Baker 1967
Dickens 1968
Dickens 1971
Firman
&
Martin 1968
Martin
&
Firman 1970
Wojdyla 1998
Wojdyla 1999
Zamorski &
Mirzwa-mroz
1996
Orlikowski et
al. 1982
Norman et al.
1995
Cultivar
Brietner
Milestone
Princess Anne
Favourite (variants)
Rival variants
Mayford Perfection
Loveliness
Balcombe Perfection
Fred Shoesmith
Lilian Shoesmith
Royals
Sheena Selecta
Super Yellow
Cymbale jaune
Fiji White
Eskort Red
Miss Oakland
Dilana
Snowdon
Nilson 1964
Table 1.6: Chrysanthemum cultivars for which differential interactions have been reported (references in first row). Nonsusceptible interactions are marked with “R” and include interactions that are reported as immune, very resistant or
resistant. Susceptible interactions are marked with “S” and include interactions that are reported as susceptible or very
susceptible. Interactions that were not tested are marked with “-“.
S
S
-
S
S
S
S
S
S
-
S
-
S
S
S
R
R
R
R
R
-
R
R
S
R
R
R
R
R
-
S
S
R
R
S
S
S
R
-
R
S
R
-
R
R
R
S
S
R
R
R
S
R
S
-
R
R
R
R
S
R
R
S
S
S
S
S
R
-
S
R
S
S
-
S
-
1.3.5 Maintenance of Puccinia horiana isolates and host susceptibility assays
As P. horiana is an obligate biotrophic parasite, the pathogen can only be maintained on living host
tissue. This requires regular transfer of isolates to fresh cuttings, approximately 3-weekly, making the
procedure very labor intensive and time consuming. Several inoculation assays have been described,
using plants, fresh cuttings, or in vitro plants (Yamaguchi 1981; Takatsu et al. 2000). Inoculation can
be performed by spraying macerated leaves with teliospores over the plants or suspending
teliospore-bearing leaves above the test plants to disperse the basidiospores (Zandvoort et al. 1968c;
Yamaguchi 1981).
Some cereal rusts have been reported to grow on artificial media (Scott and Maclean 1969). In
almost all cases, these cultures are started from the urediniospores, which are not produced by P.
horiana (Hartley and Williams 1971; Katsuya et al. 1978). Also, the morphology of the spores as well
as the pathogenicity of the isolates can be affected in axenic culture (Hartley and Williams 1971;
Katsuya et al. 1978; Yamazaki and Katsuya 1987).
33
Chapter 1
For long term preservation, the best option should be a cryopreservation protocol, such as storage of
the isolates in liquid nitrogen. The spore stage in P. horiana that seems most suitable for
cryopreservation is the teliospore, as the basidiospores are relatively fragile. Different
cryopreservation protocols for obligate parasites and rusts in particular have been described (Holden
and Smith 1992; Ryan and Ellison 2003; Bardin et al. 2007). Cryopreservation of the microcyclic rust
Puccinia spegazzinii using teliospores embedded in leaf tissue resulted in viable isolates producing
basidiospores after thawing, although no infection of the host plants could be obtained (Ryan and
Ellison 2003). Nevertheless, an optimization of these protocols might offer a good basis for long term
storage of P. horiana. Host susceptibility assays use similar inoculation techniques as for
maintenance of the isolates. For the scoring of the disease, a variety of indexes has been proposed
with six to seven levels of infection (Yamaguchi 1981; Wojdyla 1999b; Takatsu et al. 2000; Barbosa et
al. 2006).
1.4 Detection of fungal pathogens
For quarantine pathogens such as white rust, specific and sensitive detection protocols are essential.
Traditionally, the first step in disease diagnosis is the interpretation of symptoms on the host and
signs of the pathogen. In the case of rust fungi, this generally includes a detailed identification based
on rust spore morphology using microscopy. Based on pustule and teliospore shape, size, color and
surface ornamentation, P. horiana can easily be distinguished from other rust species that are also
reported on chrysanthemums (Punithalingam 1968a; Punithalingam 1968b). Major drawbacks of
traditional detection protocols are that they require skilled staff and that latent infections remain
undetected. Since it can take several days to weeks before latent infections become visible,
traditional techniques for the detection of P. horiana are far from optimal for high throughput
diagnosis in quarantine stations. A method based on chlorophyll fluorescence image analysis has
proven to be successful in the study of latent development of the pathogen (Alaei 2008) but was not
useful as a routine detection tool.
Immunological detection methods such as ELISA (Enzyme Linked Immunosorbant Assay) and LFD
(Lateral Flow Devices) are based on the recognition of specific antigens and offer the potential for
rapid field analysis with limited sample preparation. However, the production and maintenance of
monoclonal or polyclonal antibodies is very expensive. For several rusts, immunologic assays using
monoclonal antibodies have been developed including white pine blister rust (Cronartium ribicola)
(Ekramoddoullah and Taylor 1997), wheat stripe/yellow rust (Puccinia striiformis f.sp. tritici)
(Skottrup et al. 2007) and soybean rust (Phakopsora spp.) (Harmon et al. 2007). Nevertheless, for P.
horiana no antibodies are available.
34
General introduction
Finally, pathogens can be detected and identified based on the detection of DNA or RNA fragments
that are specific for the target organism. Essential for such pieces of nucleic acid is their presence in
sufficient quantities in the target organism and their absence in non-target organisms. Detection of
the target sequence can be achieved by PCR or real-time PCR, hybridization (e.g., with labeled
probes) or a combination of these techniques. The wide applicability, the relatively low development
and application cost, and the specific detection of species in a background of closely related species
are the main advantages of molecular methods. Using PCR, millions of copies of a specific DNA
sequence can be generated in a thermocyclic process consisting of repetitive cycles of DNA
denaturation, annealing of (specific) primers and DNA extension with a thermo stable DNA
polymerase (Mullis and Faloona 1987). These amplified DNA fragments or PCR products are
traditionally visualized by agarose gel electrophoresis, but also fluorometric detection can be used.
PCR is a rapid and very sensitive technique that allows the detection of minute amounts of (fungal)
pathogen DNA in latently infected plant tissue and even the detection of as few as 10 fungal spores
(Williams et al. 2001; Calderon et al. 2002a). A higher sensitivity can be obtained by a nested PCR in
which a second PCR reaction is run using primers sharing a sequence within the DNA fragment that is
amplified during the first reaction (Williams et al. 2001; Ma et al. 2003).
In real-time quantitative PCR the amplified products are monitored each PCR cycle based on the light
emitted by fluorescent dyes that are added to the reaction. The amount of fluorescence that is
detected is proportional to the amount of accumulated PCR product. During the reaction the number
of PCR cycles that are required to reach a certain level of exponentially increasing fluorescence (Ct or
Cycle threshold) is determined. The number of cycles needed to reach this Ct is dependent on the
initial amount of target DNA present in the sample. Samples with a high initial DNA concentration will
need fewer cycles to reach the Ct than samples with lower initial DNA concentrations. Based on a set
of standard samples with known target DNA concentrations, the relative amount of starting material
in every sample can be determined (Schena et al. 2004; Kubista et al. 2006). For the detection of the
PCR product, two types of fluorescent dyes can be used. Non-specific DNA intercalating dyes such as
SYBR® green are inexpensive and relatively easy to use (Morrison et al. 1998). As all the DNA in the
sample will be bound by such dyes, very specific primers and optimized reaction conditions are
necessary to avoid detection of non-specific amplification products or primer dimers, which would
lead to false positive detection. The accuracy of the PCR reaction needs to be assessed with an
additional melting curve at the end of the reaction (Mackay 2004). The other type of fluorescent dyes
that can be used are the fluorescent probes that specifically bind to the target amplification product.
One such type of probes, called Taqman® probes, are 5’ labeled with a fluorophore which is
quenched by a 3’ labeled fluorogenic quencher. During the PCR cycle the probe hybridizes to the
35
Chapter 1
DNA target after which the probe is degenerated by the exonuclease activity of the DNA polymerase,
releasing the fluorescent signal from the quencher and producing the fluorescent signal (Holland et
al. 1991; Livak et al. 1995; Wittwer et al. 1997; Livak 1999). The development of fluorescent probes is
relatively expensive but the specific hybridization with the amplified DNA target limits the chance for
false positive signals due to unspecific amplification or primer dimers.
1.4.1 Molecular detection of Puccinia horiana in the field
Molecular detection and identification by means of PCR or real-time PCR is already described for
several economically important rust species such as the rust pathogens of cereals and grasses
(Barnes and Szabo 2007; Wang et al. 2007) and soybean rust (Frederick et al. 2002; Lamour et al.
2006). A primer pair for the detection of P. horiana that was developed by Sugimura (2001) was not
species specific (Alaei et al. 2009b). Recently, protocols for the specific detection of P. horiana in
chrysanthemum using PCR and real-time PCR technology were described (Pedley 2009; Alaei et al.
2009b).
Both studies are based on the specific amplification of the internal transcribed spacers (ITS1 and
ITS2) of the ribosomal RNA gene cluster (rDNA) that are located between the genes for the 18S, 5.8S
and 28S subunits (White et al. 1990). For several reasons the rDNA-ITS regions are particularly
suitable for the detection and identification of (rust) fungi. The relatively short ITS1-5.8S-ITS2 region
of 500-800 bp can easily be amplified by PCR. Universal primer pairs as well as fungi and rust specific
primers within the RNA subunit genes have been described (White et al. 1990; Gardes and Bruns
1993; Liu et al. 1993; Kropp et al. 1997). These regions also show little to no intraspecific variation
but a larger amount of interspecific variation (White et al. 1990; Lee and Taylor 1992). The
availability of universal primers and the interspecific variation of the region allow the development of
species-specific primers for identification and detection. Multiple copies of the rDNA-ITS regions are
present in the genome, increasing the sensitivity of detection and allowing amplification of the target
even in small or degraded DNA samples (Gardes and Bruns 1993; Borneman and Hartin 2000).
In the study done by Alaei et al. (2009b) two commercial DNA extraction techniques (Genelute Plant
genomic DNA kit and Invisorb spin plant mini kit) as well as a cetyltrimethylammonium bromide
(CTAB) extraction procedure were found to extract high amounts of total DNA and relatively high
amounts of P. horiana DNA from 100 mg of infected leaf tissue. DNA extracted with these protocols
did not contain noteworthy amounts of co-extracted PCR inhibitors. The developed primers were
highly specific and gave a detection limit of approximately 10 fg target DNA in a total of 10 ng DNA
(0,001%) per PCR reaction in real-time PCR. When using SYBR® green technology, as little as 10 target
copies could be detected. Since approximately 80 target copies are expected based on the estimated
36
General introduction
genome size of P. graminis (Backlund and Szabo 1993), this theoretically corresponds to less than
one basidiospore. Traditional PCR assays as well as a Taqman real-time PCR assay were also
described (Alaei et al. 2009b).
In the alternative assay proposed by Pedley (2009), DNA is extracted with a slightly different CTAB
DNA extraction method than the one used by Alaei et al. (2009b). Pedley (2009) describes a
conventional PCR and a real-time PCR based on Taqman® technology. Using conventional PCR, the
lowest detection limit that could be obtained was 1 ng of DNA. With this PCR the pathogen could be
identified in symptomatic tissue. However, in asymptomatic tissue no reliable detection could be
performed. With his real-time PCR assay, a 100 times higher sensitivity could be obtained, allowing
detection of the pathogen in latently infected plant tissue. Nevertheless, the sensitivity of the assay
described by Alaei et al. (2009b) was substantially higher, possibly due to differences in DNA
extraction methods, primer design, PCR conditions and source material (basidiospores versus
infected leaf tissue).
The sensitive detection assays that are described for the detection of P. horiana in infected plant
tissue could be very powerful tools for the detection of pathogen spores in air samples taken in the
field. After optimization, spore detection in air and rain water samples could be used for the study of
disease epidemiology and implementation of disease warning systems. This could finally lead to a
more targeted and therefore more reduced application of fungicides. Also, if pathotype-specific
molecular markers could be developed then the same aerial spore capture techniques could be used
in pathotype monitoring, which could be used in optimal regional deployment of resistance genes
and in future breeding programmes.
Several air sampling techniques in combination with PCR and real-time PCR approaches have proven
successful for the detection of fungal spores in the field (West et al. 2008). Frequently used is the
Burkard volumetric spore sampler, which samples airborne particles at a rate of 10 l/min by
impaction onto an adhesive tape. The tape is mounted on a drum that makes one revolution in 7
days, giving the possibility to sample during one week without the need of intervention for sample
collection and allowing time-based fragmentation and analysis of the sample. A study on the nonquantitative detection of airborne ascospores of Leptosphaeria maculans and Pyrenopeziza brassicae
reported the detection of a single spore of L. maculans and 10 spores of P. brassicae in a
conventional PCR corresponding to respectively 40 and 400 spores in the initial samples (Calderon et
al. 2002a). For Penicilium roqueforti, 10 spores can be detected by conventional PCR and as low as
one spore can be detected by nested PCR (Williams et al. 2001; Calderon et al. 2002b). Detection of
Sclerotinia sclerotiorum using this type of spore sampler and traditional PCR resulted in a detection
37
Chapter 1
limit of 50 ascospores (Freeman et al. 2002) while in combination with a quantitative real-time PCR
based on SYBR® green technology had a detection limit of 2 ascospores (Rogers et al. 2009).
Apart from the Burkard spore sampler, other types of volumetric spore traps have been used for
pathogen detection or for the study of pathogen epidemiology (Neumeister-Kemp et al. 2004; van
Niekerk et al. 2010). These include cyclone traps, rotating-arm samplers and the ionic spore trap.
These do not allow analysis of subsamples based on the time during which they were captured.
However, their main advantage is the higher quantity of sampled air, resulting in a lower detection
limit of the pathogen in case of similar capture efficiencies. Cyclone traps directly collect airborne
particles in a recipient (e.g., 1.5 ml microcentrifuge tubes) allowing immediate processing of the
sample without the need to remove the spores from an adhesive tape. Using this method in
combination with conventional PCR as few as 10 spores of Penicillium roquefortii could be detected
(Williams et al. 2001). Rotating-arm traps consist of two upright rods that are coated with an
adhesive film or tape and rotate at a speed of у50 km/h sampling 50 to 200 l/min. It is a relatively
simple and inexpensive type of sampler allowing good detection of fungal spores (Calderon et al.
2002b). In case of the ionic spore trap, every minute 200 to 600 liters of air pass through a highly
charged electric field ionizing the airborne particles. These particles are forcibly attracted to a stub
with the opposite charge, containing an adhesive tape. This relatively new technique has been
proved to detect a single spore of Phakopsora pachyrhizi (soybean rust) (Schneider et al. 2009).
1.5 Molecular tools for resistance breeding and diversity studies
For the study of the genetic variation within a species, a wide gamma of molecular marker
techniques is available and except for allozymes (allelic variants of enzymes), they are all DNA based
(Schlötterer 2004). Two main classes can be distinguished if we consider the information that can be
derived from genotyping data: dominant and co-dominant markers.
1.5.1 Dominant markers
Dominant markers behave in a dominant-recessive way and do not allow to distinguish between
individuals being homozygous or heterozygous at the marker locus investigated. In most cases,
dominant marker techniques allow simultaneous screening of multiple loci. Their main advantage is
that no DNA-sequence information of the organism is required. However, marker profiles generated
using dominant marker technologies do not provide information about the allelic composition of the
organism investigated. A variety of dominant marker techniques is available, but the most widely
used in plants are Random Amplified Polymorphic DNA (RAPD) and Amplified Fragment Length
Polymorphism (AFLP) (Williams et al. 1990; Vos et al. 1995). In the case of RAPDs, a PCR reaction is
done with several arbitrary short primers of 8 to 12 bp and genomic DNA (gDNA) as a template. PCR
38
General introduction
fragments will only be generated when forward and reverse primers can bind in the right direction
on complementary regions in the target DNA that are not too far apart. Sequence polymorphism at
the primer binding sites and insertions/deletions among these sites are detected by RAPD (Williams
et al. 1990). In the case of AFLP, gDNA is first digested with two restriction enzymes, a frequent
cutter (e.g., MseI) and a rare cutter (e.g., EcoRI). Double stranded adapters that correspond to the
restriction site of the respective restriction enzymes are ligated to the generated fragments and
serve as primer binding site for the subsequent amplification of the restriction fragments. At the 3’
ends of the PCR primers, one to four selective nucleotides are included so that only a subset of the
generated restriction fragments will be amplified. During PCR, preferential amplification of fragments
that are flanked by two different adaptors will occur, resulting in a relatively low number of short
fragments that can be detected by gel electrophoresis or by capillary electrophoresis (using
fluorescently labeled primers). To obtain useful AFLP fragments for complex genomes, a
preamplification step using primers with selective nucleotides is generally required before the
selective amplification. Polymorphisms leading to a change in the restriction site and insertion and
deletions within the resctriction fragments will result in a missing or an additional fragment, or
fragments with a different electrophoretic mobility (Vos et al. 1995). The higher reliability and
robustness of AFLP compared to RAPD makes it the preferential dominant marker (Mueller and
Wolfenbarger 1999).
1.5.2 Co-dominant markers
Unlike dominant markers co-dominant markers distinguish between homozygous and heterozygous
individuals. Since they are based on the PCR-amplification or hybridization to the locus under
investigation, their development requires DNA-sequence information, making the development of
co-dominant markers more complex in non-model organisms, for which no sequence data are
available. Microsatellites or Simple Sequence Repeats (SSRs) and Single Nucleotide Polymorphisms
(SNPs) are the most frequently used co-dominant markers in genetic studies of plant species. SSRs or
microsatellites are tandemly repeated sequences with a typical repeat region that is shorter than 100
bp (Schlötterer 2004) that can be amplified with a standard PCR. In plants, highly polymorphic SSR
regions are distributed over the genome making them popular for genetic mapping and population
genetics (Tautz 1989; Schlötterer 2004). As microsatellites have a very high mutation rate due to
DNA-replication slippage, a variety of different alleles can be found in a particular locus (Ellegren
2004). However, this variable, very complex mutation pattern may create difficulties in populationgenetic analysis (Schlötterer 2004). SNPs are widespread in coding and non-coding regions of most
genomes. They are usually bi-allelic and have a relatively low mutation rate per generation
(Brumfield et al. 2003; Morin et al. 2004). However, due to this biallelic character relatively more
39
Chapter 1
SNPs are needed to study the genetic variation between populations than in the case of
microsatellites (Morin et al. 2004). The easiest way to develop SNP and SSR assays is by screening
DNA and EST databases for polymorphisms or by partial or complete genome sequencing (Zane et al.
2002; Schlötterer 2004).
1.5.3 From dominant to co-dominant markers
Multi-locus dominant marker techniques such as AFLP are too complex and not cost efficient to be
used as an assay that renders information on the allelic composition at a single locus. Therefore, a lot
of effort is made to convert dominant markers in co-dominant ones in organisms for which no
sequence data are available. In general these strategies are based on the sequencing of AFLP or
RAPD fragments, although the strategy will slightly differ depending on the desired information (Zane
et al. 2002; Brugmans et al. 2003; Nicod and Largiad 2003). For the isolation of microsatellites, SSRspecific probes can be used to enrich AFLP fragments containing microsatellites which can be
sequenced for internal primer design (Zane et al. 2002). In case polymorphic fragments are the point
of interest, e.g., because they are differential for a particular trait in different cultivars, the AFLP
marker can be converted into a simple PCR marker. This can be done by sequencing the DNAfragment of interest (and in some cases also the surrounding genome regions) after excising it out of
the gel, followed by internal primer design. This allows the identification of polymorphisms useful for
screening of the locus of interest. Based on identified polymorphisms, locus specific primers can be
developed (Brugmans et al. 2003). A similar approach can be followed to isolate a relatively high
number of SNPs based on AFLP fragments. In this case AFLP fragments that are present in at least
two individuals need to be sequenced and evaluated for the presence of SNPs. When SNPs are
identified, they need to be validated after which they can be used for further study (Nicod and
Largiad 2003; Roden et al. 2009). However, when SNPs are identified from a panel of individuals or
strains that is limited in size and composition compared to the target samples, an ascertainment bias
will be introduced. This results in the preferential selection of SNPs with a relatively high frequency
while rare SNPs are less likely to be detected in the small panel (Brumfield et al. 2003).
The availability of next generation sequencing technology now offers the possibility to screen
genomes for polymorphisms in relatively short periods of time at continuously decreasing costs.
Several technologies have been described for the identification of polymorphisms in complex
genomes including CRoPS™ (Complexity Reduction of Polymorphic Sequences) technology (van
Orsouw et al. 2007) and RADSeq (restriction siteassociated DNA sequencing) (Davey and Blaxter
2010). CRoPS™ technology was used in this study. The CRoPS™ protocol uses the AFLP technique to
reduce the complexity of the genome and to generate AFLP fragment libraries with thousands of
40
General introduction
fragments from two or more genetically different individuals or strains. The AFLP fragments from the
different genotypes are labeled with specific tags, and sequenced. The resulting sequences are
aligned and mined for SNPs that can further be developed into SNP-specific markers as described
above.
1.5.4 Applications of marker technology
In agricultural research, molecular markers are widely used for the construction of linkage maps in
diverse crop species for the development of marker-assisted selection (MAS) strategies for traits of
particular interest (Kumar 1999; Varshney et al. 2005). This requires a segregating plant population
derived from parents that differ in one or more traits of interest with a population size ranging
between 50 and 250 individuals and even higher numbers for high-resolution mapping (Mohan et al.
1997; Collard et al. 2005). As markers should be tightly linked to target loci to obtain a reliable
prediction of the phenotype, high numbers of markers need to be generated (Collard and Mackill
2008). In diploid species, linkage analysis is relatively straight forward, but in polyploid species
mapping is more difficult as a high number of genotypes for each DNA fragment can be expected and
the exact mode of chromosome pairing cannot always be determined (Wu et al. 1992). If the
mapping population is phenotypically evaluated for the traits of interest, markers can be linked to
these traits (Mohan et al. 1997).
Alternatively, identification and mapping of markers linked to genes of interest (e.g., resistance
genes) can be done by bulked segregant analysis (BSA) (Michelmore et al. 1991). In a BSA, a
segregating population is phenotypically screened for a particular trait after which individuals
displaying extreme phenotypes are pooled. This results in two pools containing all the alleles
segregating in the population, except for the loci linked to the phenotypic trait on which the
selection is based. DNA of each pool is bulked and screened for differences using numerous markers
distributed across the genome. BSA is most effective for the identification of markers linked to
dominant traits or to QTL regions with large effects. Further investigation of the particular marker on
the individuals used to construct the bulks is needed to confirm the marker/trait linkage, after which
it can be converted into a locus-specific marker (Michelmore et al. 1991; Brugmans et al. 2003).
These locus-specific markers can eventually be used for MAS in breeding programs. For several rust
resistance genes in economically important crops, markers have been described. These include two
clusters of AFLP markers associated with resistance to crown rust (Puccinia coronate f. sp. lolii) in
ryegrass (Lolium perenne) (Muylle et al. 2005), markers for resistance against leaf rust (Puccinia
triticina) and stripe rusts (P. striiformis) (William et al. 2006), and genes associated to bean rust
(Uromyces appendiculatus) (Faleiro et al. 2000).
41
Chapter 1
Several studies have described the genetic variability of chrysanthemum, but no markers for disease
resistance have been described up to now. A first study based on RAPD markers showed a relatively
high inter- and intra-species variability, although the markers could be used for cultivar identification
(Wolff and Peters-van Rijn 1993). Similar results were obtained in a later study (Huang et al. 2000).
Only very recently, two preliminary genetic linkage maps of chrysanthemum based on dominant
markers have been published. The maps cover approximately 50% of the chrysanthemum genome
and offer a framework for the mapping of additional markers (Zhang et al. 2010; Zhang et al. 2011).
Besides mapping of genes and the development of markers related to particular traits, molecular
markers can also be used to study the genetic diversity within and between populations of an
organism. Basically all types of markers can be used, but there seems to be a tendency towards codominant markers, due to their specific advantages and the increasing feasibility to identify them
due to the availability of high throughput sequencing (Schlötterer 2004; Manel et al. 2010). However
an ascertainment bias has to be taken in account when SNPs are used for the study of genetic
variation. The stability of SNPs and their higher abundance in genomes makes them more suitable for
the study of long term genetic diversity than microsatellites (Brumfield et al. 2003; Morin et al.
2004).
42
Part I:
Phenotypic variation in P. horiana
Chapter 2:
Identification and characterization
of pathotypes in Puccinia horiana,
a rust pathogen of Chrysanthemum x morifolium
A version of this chapter was published in European Journal of Plant Pathology:
De Backer,M., Alaei,H., Van Bockstaele,E., Roldan-Ruiz,I., van der Lee,T., Maes,M., & Heungens,K.
(2011). Identification and characterization of pathotypes in Puccinia horiana, a rust pathogen of
Chrysanthemum x morifolium. European Journal of Plant Pathology, 130, 1-14.
High troughput bioassay for the resistance screening of large
numbers of chrysanthemum cuttings
Pathotypes in P. horiana
2.1 Introduction
Puccinia horiana Hennings is an autoecious microcyclic rust fungus (Pucciniales) that causes
Chrysanthemum white rust. The teliospores are normally formed on the lower side of the leaves and
can germinate in situ into a promycelium without a period of dormancy (Firman and Martin 1968;
Kapooria and Zadoks 1973). On every promycelium an average of two basidiospores are formed
(Kapooria and Zadoks 1973), which are the mobile and infective propagules. The optimal conditions
for development of the pathogen are a high relative humidity and cool temperatures (17-20°C)
(Firman and Martin 1968). Under these conditions, symptoms appear 7 to 10 days post infection as
chlorotic spots that develop teliospores after 14 to 18 days. Leaf wetness and a high relative
humidity are essential for basidiospore formation, survival, and infection (Firman and Martin 1968;
Zandvoort et al. 1968a).
P. horiana can infect more than 10 Chrysanthemum species (Hiratsuka 1957; Punithalingam 1968b),
but it is especially known as a pathogen of the commercially important Chrysanthemum x morifolium.
This species is grown for the production of cut flowers, potted plants and garden chrysanthemums, in
numerous varieties and flowering forms. The turnover of chrysanthemum on the Dutch flower
auctions in 2008 was €332 x 106 for cut flowers and €30 x 106 for potted plants, making it one of the
most important florist species (Anon. 2009). The cut flower varieties are mostly produced in Japan,
the Netherlands, Italy and Colombia and are usually grown in greenhouses or in plastic tunnels
(Spaargaren 2002). Multiflora chrysanthemums (garden mums) are usually grown outdoors (Tierens
2007).
Puccinia horiana was first detected in Japan in 1895 (Hennings 1901; Hiratsuka 1957), from where it
spread to China and South Africa (Priest 1995). Since 1963 it has been reported in England and other
European countries (Baker 1967). Currently, the pathogen has been reported in most
Chrysanthemum-growing areas (Whipps 1993; EPPO 2004), where it can cause significant economic
loss if not controlled properly. It is classified as a pathogen of quarantine importance for the
European and Mediterranean Plant Protection Organization (EPPO), the Interafrican Phytosanitary
Council (IAPSC), the Junta del Acuerdo de Cartagena (JUNAC) (currently “Comunidad Andina”, (CAN))
and the North American Plant Protection Organization (NAPPO) (EPPO 2004).
As soon as P. horiana was reported as a major problem for chrysanthemum growers in the UK and
Germany, the presence of resistance in local host cultivars was studied (Stark and Stach 1965; Baker
1967; Dickens 1968). Dickens (1968) evaluated the resistance of 37 cultivars with an isolate from cv.
Favourite. He observed resistance in the cvs. Princess Anne and Rival’s Rival, which were reported as
susceptible in the study of Baker (1967). Cv. Princess Anne was also reported as susceptible in
47
Chapter 2
Sweden (Nilsson 1964), Denmark (Anon. 1964) and Norway (Gjaerum 1964). Dickens (1971) showed
that cv. Favourite did not display any sign of infection after inoculation with an isolate from cv.
Mayford Perfection. An inoculation study on 270 cultivars by Martin and Firman (Martin and Firman
1970) showed comparable discrepancies with the previously described studies. Wojdyla (1999b) also
described some discrepancies in susceptibility between his study and other inoculation experiments.
A study in which six Japanese isolates were inoculated on 40 cultivars involved the first systematic
testing of multiple isolates on multiple hosts, and revealed differential interaction phenotype profiles
for several of the cultivars used (Yamaguchi 1981). A more recent study performed in Mexico on five
cultivars with 16 Mexican isolates from four different regions showed variance in virulence, but did
not clearly describe pathotypes (Velasco et al. 2007). These data all suggest the presence of
pathotypes (physiological races), most likely as a result of gene-for-gene type of interactions (Flor
1971; Stukenbrock and McDonald 2009; Anon. 2011).
Resistance to host-specific pathogens that can overcome the plant basal resistance, such as rusts,
depends on the presence or absence of avirulence (Avr) genes in the pathogen and the
corresponding resistance (R) genes in the host. These avirulence factors are often effector molecules
that play a role in the infection process (Bent and Mackey 2007). In such interactions, the Avr gene
product (recently referred to as an elicitor) can interact with the product of an R gene in the host to
start a cascade of host defense reactions, which often leads to an hypersensitive reaction (HRreaction) and ultimately, resistance (Staskawicz et al. 1995). In rusts, specific proteins secreted by
haustoria into the host cell have been shown to be elicitors (Catanzariti et al. 2007).
Flor (Flor 1956) showed that in gene-for-gene type of resistance interactions, multiple resistance
genes can be present in the host, whose products each react with the products of corresponding Avr
genes in the pathogen. In such ideal pathosystems with n R genes (and n Avr genes), 2 n different host
genotypes (cultivars) can produce differential reactions with 2n different pathogen genotypes
(pathotypes). Some pathosystems involving gene-for-gene relations have been studied in detail using
genetic and molecular analysis. The number of R and Avr genes present in these pathosystems has
been determined, pending the discovery of new pathotypes or differential cultivars (Ellis et al. 2007;
Bolton et al. 2008). In several cases, this number can be determined even if not all cultivars and
pathotypes are known, via analysis of the proportions of resistance-breaking pathotypes and the
proportions of susceptible cultivars in an incomplete set of cultivars and pathotypes (Person 1959).
Especially if more than three resistance genes are present, multiple isolates need to be tested on
multiple cultivars to obtain reliable proportions.
48
Pathotypes in P. horiana
Considering the increasing international importance of Chrysanthemum and its planting material and
the increasing restriction of chemical fungicide use in some regions, breeding for resistance to a
relevant set of pathotypes is becoming more important. Since most published inoculation studies
with P. horiana used only a limited number of isolates, we developed a robust and efficient screening
assay that is compact, easy, and reliable. Using this assay we tested an international collection of
isolates on a set of cultivars for which anecdotal reports of differential reactions had been noted by
some of the main chrysanthemum-breeding companies. These data are used to determine the
minimum number of Avr (and R) genes involved in this pathosystem and to identify isolates with
different specificity that can be exploited in R gene identification in resistance breeding programs.
2.2 Materials and methods
2.2.1 Fungal isolates
A total of 22 isolates of P. horiana were collected on commercially-grown Chrysanthemum x
morifolium plants between 2003 and 2009 (Table 2.1). The main selection criterion was a wide
geographic distribution, including isolates from different continents. For the Belgian isolates, the
geographic origin surveyed most intensively due to our location, we included collection year as an
extra factor. For isolates collected after 2005, breeders were asked to preferentially provide
symptomatic plant material from cultivars that were previously reported as resistant, which
increased the probability of encountering new pathotypes. Diseased plant material was used to
inoculate fresh cuttings of the susceptible cvs. Medonia, Taliedo or cv. 29 (see “Cultivars” section) as
described by Alaei et al. (Alaei et al. 2009a). Single pustule cultures of each isolate were established
before conducting pathotype tests. All the isolates were maintained by tri-weekly transfer onto fresh
and rust-free cuttings of cvs. Medonia and Taliedo as described by Alaei et al. (Alaei et al. 2009a),
except the Colombian isolates, which were maintained on cv. 29, but using the same transfer
schedule. Because of the quarantine status of P. horiana, the pathogen was only handled in
laboratories and growth chambers that fulfill the biosafety rules for quarantine plant pathogens.
49
Chapter 2
Table 2.1 Isolates of Puccinia horiana with isolate code, origin, collection year and source.
Isolate code
Origin
Collection year
a
Source
BE1
Belgium, region Ghent
2003
ILVO
BE2
Belgium, region Ghent
2003
ILVO
BE3
Belgium, region Ghent
2003
ILVO
BE4
Belgium, region Ghent
2005
ILVO
BE5
Belgium; region Mechelen
2007
Gediflora
BE6
Belgium; region Mechelen
2008
Gediflora
CO1
Colombia; region Boyaca
2008
ICA
CO2
Colombia; region Cundinamarca
2008
ICA
FR1
France
2007
BBV
FR2
France
2008
BBV
GB1
United Kingdom
2005
FERA
GB2
United Kingdom
2005
FERA
GB3
United Kingdom
2005
FERA
JP1
Japan, region Hiroshima
2008
Deliflor
JP2
Japan, region Hiroshima
2008
Deliflor
JP3
Japan, region Tochigi
2009
Deliflor
MY1
Malaysia, region Cameron Highlands
2008
Dekker Breeding
MY2
Malaysia
2008
Dekker Breeding
NL1
the Netherlands
2006
PD
NL2
the Netherlands
2008
RVZ
PL1
Poland
2006
INSAD
US1
USA (Massachusetts)
2008
USDA APHIS PPQ
a
Samples obtained via: ILVO: Institute for Agricultural and Fisheries Research, Merelbeke,
Belgium / Gediflora, Staden-Oostnieuwkerke, Belgium / ICA: Instituto Colombiano Agropecuario,
Bogota, Colombia / BBV: Bretagne Biotechnologie Végétale, Saint Pol de Léon, France / FERA
(formerly CSL): The Food and Environment Research Agency, York, United Kingdom / Deliflor,
Maasdijk, the Netherlands / Dekker Breeding, Hensbroek, the Netherlands / PD: Plant Protection
Service, Wageningen, the Netherlands / RVZ: Royal Van Zanten, Rijsenhout, the Netherlands /
INSAD: Research Institute of Pomology and Floriculture, Skierniewice, Poland / USDA APHIS PPQ:
United States Department of Agriculture - Animal and Plant Health Inspection Service - Plant
Protection and Quarantine.
2.2.2 Cultivars
A set of 36 Chrysanthemum test cultivars was selected based on the multi-year experience of four
major Chrysanthemum breeding companies: Deliflor (Maasdijk, NL), Dekker Chrysanten (Hensbroek,
NL), Fides (De Lier, NL), and Royal Van Zanten (Rijsenhout, NL). This set consisted of cultivars that are
usually reported as susceptible (positive controls), cultivars assumed to be resistant (using
inoculation tests with non-characterized isolates), and cultivars for which anecdotal reports of
infections were reported to the breeders. Details of the test cultivars are given in Table 2.2. Mother
plants of these cultivars were maintained by Fides. Upon request, three-week-old rooted cuttings of
10 to 15 cm height were prepared in Grodan SBS 36/77 rockwool blocks (Roermond, the
Netherlands) and transferred to our test facility where they were maintained in a greenhouse for
50
Pathotypes in P. horiana
maximum one week before inoculation. The cultivars can be requested from the respective
companies for research purposes.
Table 2.2 Cultivars used, including cultivar number, plant type and source.
Cultivar
number
Type
1
2
3b
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Multiflora
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Cut flower
Pot flower
Multiflora
Cut flower
Cut flower
Pot flower
Multiflora
Multiflora
Multiflora
Multiflora
Multiflora
a
Source
Royal van Zanten
Deliflor
Royal van Zanten
Fides
Deliflor
Deliflor
Deliflor
Deliflor
Deliflor
Deliflor
Deliflor
Deliflor
Dekker Chrysanten
Dekker Chrysanten
Dekker Chrysanten
Dekker Chrysanten
Dekker Chrysanten
Dekker Chrysanten
Dekker Chrysanten
Dekker Chrysanten
Fides
Fides
Fides
Fides
Fides
Fides
Fides
Fides
Royal van Zanten
Royal van Zanten
Royal van Zanten
Royal van Zanten
Royal van Zanten
Royal van Zanten
Royal van Zanten
Royal van Zanten
a
Royal van Zanten (RVZ), Rijsenhout, the Netherlands / Fides, De Lier,
the Netherlands / Deliflor, Maasdijk, the Netherlands / Dekker
b
Chrysanten, Hensbroek, the Netherlands. cv. 3 = Alec Bedser
51
Chapter 2
2.2.3 Bioassay
The first objective was to develop a compact and reliable screening assay. Given these criteria and
the need for leaf wetness and a high humidity for successful infection, we developed an assay that
could be conducted on relatively small plants inside closed plastic containers, several of which should
easily fit in a controlled environment such as a growth chamber. Expanding on previously-described
inoculation methods (Firman and Martin 1968; Yamaguchi 1981; Takatsu et al. 2000; Alaei et al.
2009a), different inoculation techniques (suspending infected leaves from the lid of the container
versus spraying of macerated leaves over the plants), different ways of suspending leaves (with tape
versus with agar), different plant sizes, different substrates (peat versus rockwool), and different
plastic containers of different sizes were tested in preliminary assays. Eventually, we chose plastic
containers of 46.5 cm length x 26.2 cm width x 26.0 cm height (Savic, Kortrijk, Belgium). Glass plates
cut to size were used as covers during inoculum production. The inoculum of basidiospores was
generated by approximately 36 heavily-infected whole leaves distributed evenly and stuck to the
covers with their telia pointing downwards using 1% water agar as adhesive. To obtain the heavilyinfected whole leaves used for the inoculation, 50 fresh cuttings placed in Grodan AO 36/40
rockwool slabs were used per plastic container. For pathotype screenings, which involved larger
cuttings, the plants were placed in plastic trays (9 by 4 wells, cut out of Grodan SBS 36/77 trays) on
the bottom of each container and covered by an inverted second container. The test chamber height
was increased to allow for larger plants (Figure 2.1). Only 18 plants were used per tray, using
alternate positions. In this way, we avoided shadow effects caused by larger leaves covering the
smaller plants that could possibly have resulted in false negative results. The space was high enough
to ensure homogeneous dispersal of the spores over the larger test plants and avoided that the
plants reached the cover during the incubation period. To ensure a high relative humidity and a
water film on the leaves, the cuttings, the inner sides of the plastic containers and the covers holding
the inoculum were misted with demineralized water using a Preval Sprayer (Yonkers; NY; USA). A
layer of approximately 1 cm of demineralized water was placed in the bottom of the plastic
container, providing water for the plants and further contributing to a high relative humidity. After
setup, the plastic containers were placed in a dark growth chamber at 17°C. Two days after the start
of the inoculation (dpi), fluorescent light (Gro-lux® F58W/GRO-T8, Osram Sylvania, MA, USA) was
provided during 16 hours per day. In general, symptoms were easy to evaluate at 21 dpi, but a final
evaluation was done at 28 dpi to include possible retarded symptoms.
To test the homogeneity and the dose of inoculum, three microscope slides were placed in each
plastic container at the level of the cuttings during two separate screenings (using isolates CO2 and
BE4). After 24 hours, the slides were removed and the number of basidiospores in one square mm
52
Pathotypes in P. horiana
was counted at three arbitrarily chosen locations on each slide. Data were analyzed using ANOVA in
Statistica 9.0 (Statsoft, OK, USA).
Before each assay, plastic containers and covers were always washed and subsequently sterilized
with 70% ethanol. Trays were washed and subsequently sterilized in 0.5% sodium hypochlorite
overnight.
Figure 2.1: Bioassay setup for inoculum production (left) and pathotype screening (right).
2.2.4 Pathotype screening
The 22 isolates were separately inoculated on the 36 test cultivars (Tables 2.1, 2.2). Inoculum was
prepared for each isolate of P. horiana using fresh cuttings of cvs. Medonia, Taliedo, and in case of
the Columbian isolates, cv. 29, as described above. The single-pustule cultures were increased to
approximately 150 infected cuttings during three to four three-week cycles to obtain a sufficient
number of infected leaves. Per isolate, three cuttings (replicates) of each of the 36 test cultivars were
inoculated. Each set of 36 test cultivars was randomly distributed over two plastic containers (18
cultivars per container). The screenings were conducted between July 2007 and December 2009. To
test the reproducibility of the results and account for possible effects due to differences in the
physiological status of the host, nine isolates were screened twice, usually separated by several
months.
53
Chapter 2
2.2.5 Disease scoring
Plants were scored according to Table 2.3. If no teliospores were observed, the interaction was rated
“0”. If plants contained more than 10 well-developed pustules per leaf, the interaction was rated “2”.
A score of “1” was given to interactions if only a few pustules on a few leaves were observed.
Interactions with a score of “1” included plants that only showed infection on the stipules, the leaf
tips, or the upper side of the leaf. In case macroscopic evaluation of a lesion was difficult, leaves
were checked microscopically for the presence of teliospores. The scores of the three replicate
interactions were added together, resulting in an overall interaction phenotype score per cultivar
ranging from 0 to 6 (Table 2.3). The interaction phenotypes were reduced to two groups to allow
more detailed analysis. Interaction phenotypes were rated “susceptible” (the given isolate was
considered “virulent” on the given cultivar) with a score of 4 or more or “resistant” (the given isolate
was considered “avirulent” on the given cultivar) with a score of 3 or less. For the nine isolates that
were screened twice on the set of cultivars, interaction phenotypes with a score of at least 4 in one
of the two replicate tests were rated “susceptible”. The choice of the cutoff value was validated by
analysis of frequency distribution of the interaction phenotype scores (See Results and Discussion
sections).
Table 2.3 Interaction phenotype scoring and interpretation
a
Interaction
Number of cuttings with infection score
phenotype
2
1
0
score
0
0
0
3
1
0
1
2
2
0
2
1
3
0
3
0
4
1
2
0
5
2
1
0
6
3
0
0
a
See materials and methods for scoring system
Interaction
phenotype
interpretation
Resistant
Resistant
Resistant
Resistant
Susceptible (S)
Susceptible (S)
Susceptible (S)
2.2.6 Analysis of interaction phenotype profiles
After classification of each interaction as susceptible versus resistant, the cultivar-by-isolate
interaction phenotype matrix was sorted based on decreasing virulence of the isolates (number of
cultivars they were able to infect) and increasing susceptibility of the cultivars (number of isolates
they are susceptible to). To determine the minimum number of resistance genes involved in this
pathosystem, the sorted matrix was analyzed according to two methodologies. The first analysis was
based on the fact that in an ideal gene-for-gene system with a total of n R genes (and n Avr genes),
54
Pathotypes in P. horiana
cultivars lacking x R genes can be infected by 2x pathotypes that contain up to x Avr genes. The
number of cultivars that can be infected by 2x pathotypes is dependent on the number of R genes in
the system and follows a binomial distribution (n!/((n-x)!*x!)) (Person 1959). The second analysis of
the pathosystem was based on the geometric rule (Person 1959), which implies that for any two
cultivars or pathotypes, the number of specific R genes or Avr genes missing in both cultivars or
pathotypes (y) results in 2y common susceptible interactions. Examination of the number of common
susceptible interactions therefore allows determination of the number of commonly missing Avr or R
genes. For example, if one cultivar is infected by eight pathotypes and the other by 16, and there are
four pathotypes that infect both cultivars, then the first cultivar lacks three R genes, the second one
lacks four R genes, and two specific R genes are lacking in both cultivars.
2.3 Results
The new bioassay allowed the simultaneous screening of 18 large cuttings on a surface area of
approximately 0.12 m2. Inoculation using infected leaves attached to the cover of the plastic
container produced consistent and uniform infection in preliminary experiments, in contrast to
inoculation with macerated leaf suspensions (data not shown). The use of rockwool as a substrate
allowed cleaner and easier handling of the plants, and resulted in few to no problems with infections
due to Botrytis compared to peat-based potting media. The basidiospore inoculum dose at the level
of the plants varied from 20 to 170 spores mm-2, with an average (± SD) of 78.5 ± 29.2 spores mm-2, a
lower quartile of 58 spores mm-2, and an upper quartile of 93.5 spores mm-2. No significant
difference (P = 0.32) was observed between the two isolates (71.6 ± 12.5 spores mm -2 for plastic
containers inoculated with CO2 versus 85.5 ± 29.4 spores mm-2 for plastic containers inoculated with
BE4). Within each isolate, the difference in spore dose between the plastic containers (6 per isolate)
was also not significant, except for the two most differing plastic containers of isolate BE4 (115.8 ±
39.5 spores mm-2 versus 34.1 ± 16.4 spores mm-2; P = 0.026). Within plastic containers, several
statistically significant differences were observed, but on average, the SD between the different
sampling points within a plastic container was only 19.2 spores mm-2.
The results of the screening are presented in Table 2.4. A total of 82.5% of the interaction
phenotypes showed either no symptoms (61.8% had a score of 0) or severe symptoms (20.7% had a
score of 6). The third most common interaction phenotype was a score of 3 (8.7%), resulting from a
consistent scoring of minor symptoms (1+1+1) on the three replicate plants. Such consistent minor
symptoms were mostly observed on specific cultivars: 61% of score 3 was given to 6 cultivars (cvs. 1,
4, 7, 15, 27, 28). Only 8.8% of the interaction phenotypes had variable scoring between replicates of
the same experiment, but a score of 0 as well as a score of 2 was never observed together.
55
Chapter 2
In general, the results of the nine isolates that were tested twice were consistent. Using the scoring
system of ч3 for a resistant and ш4 for a susceptible interaction phenotype, consistent classification
was obtained in 302 out of the 324 combinations (93.2%) that were repeated. With isolates BE4, BE6,
GB1, and JP1 larger variation was observed, frequently on the same cultivars, e.g., cvs. 1, 4, 17 and
30. If omitting these more variable isolates and cultivars, consistent classification was obtained in
169 out of 172 cases (98.3%). However, as these few variable interactions may represent a
biologically interesting phenomenon and as their phenotype did not affect the main conclusions (see
further), they were included in further analyses.
Eight cultivars (cvs. 6, 10, 11, 12, 26, 32, 33, 35) showed complete resistance against all isolates
tested and two cultivars (cvs. 2, 29) showed overall susceptibility. There were three other groups of
cultivars that showed the same interaction phenotype profile for all isolates: cvs. 9, 18 and 21, cvs. 8
and 34, and cvs. 13 and 36. None of the 22 isolates showed an identical infection profile on the set of
36 cultivars so each of them represents a different pathotype, and will be referred to as such.
A non-redundant matrix of 22 pathotypes by 24 cultivars (Table 2.5) was obtained after grouping of
the cultivars with the same profile and reduction of the interaction phenotypes to susceptible vs.
resistant (see Materials and Methods). A cultivar listed in this table is interpreted in the sense of
Person (Person 1959), namely, a differential cultivar. In Table 2.5, the isolates (columns) were sorted
based on decreasing number of virulent interactions and cultivars (rows) were sorted based on
increasing number of susceptible interactions. This allowed further analysis of the data based on
Person (Person 1959), assuming that resistance in this pathosystems is indeed based on gene-forgene type of interactions. Cultivars and isolates at the edges of this table are the most informative
about the presence and absence of R genes and Avr genes, respectively. Cv. 6 was the universally
resistant cultivar, resistant to all pathotypes used. On the other hand, cv. 2 was susceptible to all
tested pathotypes. No pathotype tested infected all cultivars (the universally virulent isolate).
Similarly, no isolate was observed that could only infect the universally susceptible cultivar (the
universally avirulent isolate).
NL1 was able to infect the largest number of differential cultivars (19 out of 24). Other isolates that
were able to infect a large number of cultivars were NL2, BE6 and MY1, with 14, 13 and 13
differential cultivars, respectively. Isolates BE1, BE5 and CO1 were the isolates that could infect the
smallest number of differential cultivars (4). The remaining isolates were virulent on 5 to 12 cultivars,
with a variable capacity to also produce minor symptoms (scores 1 to 3) on additional cultivars
(Tables 2.4, 2.5).
56
Pathotypes in P. horiana
We identified one cultivar (cv. 14) that is susceptible to 21 pathotypes. As the number of pathotypes
a cultivar can be susceptible to is 2x, this cultivar is susceptible to at least 32 pathotypes indicating
that it is missing at least five (25=32) R genes. Based on the same rationale, NL2, BE6, MY1 and GB1
are able to infect at least 16 cultivars and are missing at least four Avr genes. Pathotypes NL2, BE6
and GB1 have 13, 9 and 11 susceptible interactions in common with NL1, respectively. Based on the
geometric rule, this means they each have at least 16 susceptible interactions in common with NL1.
In each case, there are also bidirectional differential interactions. As different pathotypes that each
infect 16 cultivars can only have a maximum of eight susceptible interactions in common, pathotypes
NL2, BE6 and GB1 have to infect at least 32 cultivars and lack at least five Avr genes. Based on the
number of common susceptible interactions between MY1 and NL2 or GB1, we can also conclude
that MY1 has to infect at least 32 cultivars. This brings the number of isolates infecting at least 32
cultivars to five, and these isolates have six cultivars in common that they can all infect. This situation
can only be explained in an ideal system with seven or more R/Avr genes. Analysis of the number of
common susceptible interactions in the most susceptible cultivars results in the same conclusion.
With at least seven genes, the minimum number of possible pathotypes and cultivars is 128 (=2 7).
Most of the possible pathotypes and differential cultivars are lacking from our available set and
further determination of the number of R/Avr genes based on the geometric rule is therefore
speculative. As a result, it was not possible to designate specific R genes and Avr genes to most
cultivars and isolates, respectively, even though the minimum number of missing R genes and Avr
genes could be calculated (Table 2.5).
57
Chapter 2
Table 2.4: Interaction phenotype score (See Table 2.3) of 22 isolates of Puccinia horiana (columns; see Table 2.1) on 36
Chrysanthemum x morifolium cultivars (rows; see Table 2.2). Isolates with two sets of numbers (upper and lower) were
tested twice, at different time points.
Isolates
Cultivar
Comm.a
number BE1 BE2 BE3 BE4 BE5 BE6 CO1 CO2 FR1 FR2 GB1 GB2 GB3 JP1 JP2 JP3 MY1 MY2 NL1 NL2 PL1 US1
1
3
3
3
2
2
2
1
5
3
4
4
6
6
5
6
6
1
4
6
6
3
6
3
4
3
0
3
3
6
5
6
2
6
6
6
6
6
6
6
6
6
4
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
3
6
6
3
6
6
3
3
6
3
4
3
6
6
5
0
6
6
4
6
0
3
1
6
6
6
6
6
6
3
6
6
4
1
2
3
2
3
3
0
3
4
0
0
0
3
2
3
0
6
3
3
1
4
3
3
0
0
0
3
6
3
5
6
5
0
6
6
0
0
0
0
6
6
0
0
0
0
6
6
6
6
0
6
0
0
0
0
6
6
0
6
6
6
2
1
6
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7
3
1
3
3
3
2
3
3
3
0
0
1
3
4
2
3
3
2
3
0
0
0
1
6
6
6
6
0
3
0
0
8
0
0
0
0
0
0
0
5
5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
9
0
0
0
0
0
6
6
6
6
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
11
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
12
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
13
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6
0
0
0
0
14
6
6
6
6
6
6
6
6
6
5
5
0
0
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
15
4
0
3
3
3
2
0
5
5
0
0
1
3
3
6
2
5
2
3
4
6
6
3
0
3
0
6
0
3
3
3
W, Z
16
3
3
6
6
6
3
0
5
6
0
0
1
3
2
0
3
6
6
6
6
6
6
6
6
6
2
6
6
6
6
6
Z
17
3
3
6
3
3
0
0
0
4
0
0
0
0
0
0
0
4
3
2
6
6
6
3
0
0
0
6
6
2
2
6
Z
18
0
0
0
0
0
6
6
6
6
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
19
0
0
3
0
0
0
0
0
0
0
0
2
3
0
6
0
0
0
0
0
0
0
0
6
6
0
6
0
0
0
0
20
0
0
0
0
0
0
0
0
0
0
1
5
3
0
0
0
0
0
0
0
0
0
0
6
6
6
6
6
0
0
0
21
0
0
0
0
0
6
6
6
6
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
22
0
0
3
0
0
0
0
0
0
0
0
0
0
0
6
4
0
0
0
0
0
0
0
6
6
0
6
6
0
0
0
23
0
0
0
0
0
0
0
0
0
3
5
6
6
0
0
0
0
0
0
0
0
0
1
6
6
6
6
6
0
2
0
24
0
3
3
0
1
0
0
6
6
0
0
0
0
6
6
4
6
0
5
0
0
0
0
6
6
0
6
6
2
0
0
25
2
4
3
0
3
0
0
2
6
0
0
0
0
2
2
0
1
0
0
0
3
3
5
0
0
0
1
0
2
0
0
58
V, W
V, W
W, X
Y
Pathotypes in P. horiana
Table 2.4 (continued)
Isolates
Cultivar
Comm.a
number BE1 BE2 BE3 BE4 BE5 BE6 CO1 CO2 FR1 FR2 GB1 GB2 GB3 JP1 JP2 JP3 MY1 MY2 NL1 NL2 PL1 US1
26
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
27
3
3
6
6
3
3
3
5
6
0
0
3
3
3
0
5
4
6
3
0
0
0
0
6
6
6
6
0
4
0
0
28
0
0
0
0
0
0
0
0
0
0
0
2
2
0
0
3
3
0
3
0
3
1
3
0
3
3
6
3
0
5
6
29
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
30
3
0
0
6
1
0
0
0
0
0
0
0
0
0
0
0
1
6
0
6
6
6
3
0
0
1
6
6
0
2
6
31
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
3
0
0
0
6
6
0
0
0
0
6
6
0
3
3
32
0
0
0
0
0
2
3
2
3
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
33
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
34
0
0
0
0
0
0
0
2
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
35
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
W, X
Y
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
4
0
0
1
0
0
0
0
0
0
0
0
0
Comments: (V) infection on stipules; (W) infection on leaf tips; (X) infection on upper side of the leaf; (Y) symptoms limited to minor
puncture-like lesions, only visible in counter light, and no formation of teliospores; (Z) retarded symptom expression (teliospore formation
after 4 weeks instead of 2 to 3 weeks)
36
a
Some cultivars show a unique interaction with a specific pathotype. Pathotype BE6 was the only one
able to infect cv. 8, while only isolate NL1 was capable of infecting cv. 13. This indicates that cvs. 8
and 13 each lack a specific but different R gene, the products of which interact with corresponding
specific elicitors in these pathotypes. As these two pathotypes are both virulent on a large number of
cultivars, they are very interesting in that these specific Avr genes probably represent the only Avr
gene in the case of NL1 and one of the few Avr genes in the case of BE6. For NL1, this is because the
resistance to this isolate in cvs. 8, 9, 25 and 4 is probably generated by a single matching R gene, of
which the Avr gene is lacking in BE6 (Table 2.5). As a result, BE6 is able to cause disease on these four
specific cultivars. Pathotype CO2 is also very interesting in that it was the only one that could not
infect cv. 14. Therefore, cv. 14 contains a single R gene that corresponds to an Avr gene that is only
present in pathotype CO2. Two other Avr genes could be assigned specifically to isolates US1 and BE5
based on their differential interactions on cvs. 28 and 9, respectively. Cv. 9 could only be infected by
BE6 and BE5, suggesting this cultivar contains all available R genes except those for recognition of
isolates BE6 and BE5. In a similar way, cv. 28 contains all available R genes except those for
recognition of isolates NL1 and US1. Given the large number of possible combinations in a 128 by 128
matrix and the limited number of interactions available, further determination of the presence or
absence of specific R genes in specific cultivars, and specific Avr genes in specific isolates is not
possible using this approach.
59
Table 2.5: Pathotypes of P. horiana (columns) and representative differential cultivars of Chrysanthemum x morifolium (rows) that were identified based on the observed interaction phenotype scores (Table 2.4)
and the interaction phenotype interpretation (Table 2.3). Susceptible interactions are marked with “S”. Cultivars were sorted increasingly based on the number of susceptible interactions. Pathotypes were
sorted decreasingly based on the number of infecting cultivars. Presence (upper case letter) or absence (lower case letter) of specific Avr genes in the pathotypes and R genes in the cultivars is listed where
possible. Avr/R genes labeled in italics were added based on the assumption of a single Avr gene in NL1 (see results).
Avr genea
A/a
B/b
C/c
D/d
E/e
F/f
G/g
a
Cultivarb
6
8
13
9
28
19
25
7
31
4
20
22
23
15
30
17
24
27
5
1
16
3
14
2
A/a
A
A
a
A
a
a
A
a
a
A
a
a
a
a
a
a
a
a
a
a
a
a
a
a
B/b
B
b
R gene
C/c D/d E/e
C
D
E
F/f
F
NL2
a
c
BE6
a
B
c
MY1
GB1
a
US1
a
JP1
a
JP2
FR2
c
c
c
D
c
c
c
Pathotype
MY2 GB2 GB3
c
c
c
BE3
BE4
FR1
JP3
a
BE2
a
CO2
PL1
BE1
BE5
a
CO1
c
c
c
c
c
C
c
c
c
c
E
G/g
G
S
S
b
e
S
d
S
S
S
S
S
S
S
b
d
c
S
S
S
S
S
S
S
S
S
S
S
S
S
S
c
b
d
d
b
b
b
b
c
b
b
b
b
S
S
b
c
C
c
d
d
d
d
d
e
e
e
f
g
# Susc.c
Min # missing Avr genes
a
NL1
A
b
c
d
e
f
g
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
19
14
13
13
12
9
9
8
8
8
7
7
6
6
6
6
5
5
5
4
4
4
5
5
5
5
5
4
4
4
4
4
4
4
4
3
4
4
3
3
3
3
3
3
# Susc.c
0
1
1
2
2
3
3
4
4
5
5
5
6
7
7
8
8
9
11
14
14
16
21
22
Min. #
missing R
genes
0
1
1
2
2
2
2
3
3
3
3
3
4
4
4
4
4
4
4
5
5
5
5
6
The pathosystems is shown with seven genes, the minimum number involved based on the Person analysis of the data (see Results section).
b
Cvs. 6, 10, 11, 12, 26, 32, 33, and 35 had the same interaction phenotype profile. Only the first cultivar is listed in the table as the representative cultivar. This was also the case for cvs. 2 and 29, cvs. 9, 18 and 21, cvs. 8 and 34, and cvs.
13 and 36.
c
Number of susceptible interactions for each cultivar. Number of susceptible cultivars for each pathotype.
Pathotypes in P. horiana
2.4 Discussion
The above-described bioassay in closed plastic containers allowed testing at constant temperature,
humidity and light conditions. We found these factors to be critical for consistently successful
inoculation of the pathogen. Our average inoculum level of 78.5 basidiospores mm -2 is in accordance
with the circa 70 spores mm-2 described by Yamaguchi (1981). The relatively small differences in
inoculum level, the high inoculation density and the use of three replicates surely contributed to the
robustness of the inoculation method. The interaction phenotypes were consistent over the three
replicates in 91.2% of the cases. Also, results obtained with the nine isolates that were re-screened in
a different period of the year were consistent. For most interactions, the timing of symptom
development was as described previously (Firman and Martin 1968). However, for some cultivars it
took a few days longer to develop teliospores (Table 2.4), which is why an additional scoring at 28 dpi
was performed. The bioassay offered the possibility to simultaneously screen several isolates in three
replicates on a relatively small surface area, such as that provided in a biosecurity growth chamber.
At an evaluation rate of up to 216 cuttings per week we screened a total of 3348 cuttings, making it
relatively high throughput.
Our focus was on the initial gene-for-gene type of recognition events. Therefore, our scoring method
aimed to evaluate the recognition events and not the amount of disease expression. We used an
interaction phenotype scoring system that was relatively simple compared to the scoring scales that
focus more on the relative leaf area covered with teliospores (Yamaguchi 1981; Wojdyla 1999b;
Barbosa et al. 2006) or the disease indexes derived from the infection level of separate leaves
(Takatsu et al. 2000; Alaei 2008). For the analysis of the pathotypes involved, the interaction
phenotype scores were reduced to susceptible or resistant interactions based on a clear frequency
distribution. However, as in most bioassays, some intermediate phenotypes were observed. As soon
as a cultivar showed a full-blown infection on at least one of the three replicate plants (a score ш 4) in
at least one of the replicate tests, the cultivar was considered susceptible. This arbitrary threshold
was based on the assumption that in such cases, recognition of the pathogen failed. This level was
considered sufficiently stringent, as disease expression is not always complete, possibly due to the
physiological status of the plants (Walters and Bingham 2007; Bolton 2009) or because expression of
disease resistance is inherently variable in specific cultivars, as was observed with cvs. 4, 17 and 30.
Given the frequency distribution of the infection phenotypes, with few scores of 4, limited
differences in outcome were observed if a more stringent infection phenotype cutoff (a score ш 5)
was used. One exception is found for isolates JP1 and JP2 (data not shown). Replication in time also
showed limited variability (6.8%), and was in half the cases linked to cultivars with variable disease
expression. These variable reactions were also linked to specific isolates (especially BE6 and GB1).
61
Chapter 2
However, we suspect that this variability may be more related to differences in environmental and
host conditions between the experiments with these isolates than in inherent variability of the
isolates themselves. Reason is that the general level of susceptibility fluctuated slightly between
these replicate experiments. Under those conditions, a small infection phenotype is considered an
incomplete expression of disease in the absence of an Avr/R match instead of the result of an
incomplete expression of defense after an Avr/R match. Therefore, preference is given to the more
susceptible score, which is reflected in our scoring system. However, even if preference was here
given to the resistant score for all the variable interactions, the main conclusion about the minimum
number of genes involved in this system would hold, demonstrating the limited impact of these
variable results and showing the robustness of the system.
Our results confirm the presence of pathotypes in P. horiana that can be explained by the gene-forgene concept described by Flor (Flor 1956) and analyzed as described by Person (Person 1959). These
results should be confirmed by genetic studies, but preliminary analyses of resistance characteristics
in progeny of a cross of a resistant and susceptible cultivar using isolates BE5, BE6, JP1, and NL1
indicate Mendelian inheritance of different R genes (see Chapter 6). Although several studies have
demonstrated the existence of differential reactions in the pathosystem of P. horiana and
chrysanthemum, none have tested a comprehensive collection of isolates covering a broad
geographical range on a large set of cultivars and tried to assess the race complexity of this
pathosystem and the number of resistance genes involved (Baker 1967; Dickens 1968; Yamaguchi
1981; Wojdyla 1999b; Velasco et al. 2007). The pathogenic variability shown in those studies was
rather limited compared to the results obtained here, presumably due to the wider geographical
origin of our isolates, the selection of isolates on cultivars that were previously reported as resistant,
and the larger number of test cultivars used, including some on which anecdotal infections were
reported.
The 22 isolates we tested represented 22 different pathotypes. Based on the number of isolates
expressing susceptibility on a large number of cultivars and the number of cultivars commonly
infected by these isolates, we could conclude that a minimum of seven genes are involved in this
pathosystem. In a system with at least seven genes, five isolates infecting 32 cultivars or more can be
found that have six susceptible interactions in common.
Complex gene for gene interactions have been found in other plant/rust interactions. In the
flax/Melampsora lini pathosystem, 31 R alleles were located on 5 different loci, with two to 13 alleles
per locus (Ellis et al. 2007). For Puccinia triticina, the most common leaf rust on wheat, most of the
60 currently known R genes were mapped to separate loci, distributed over the genome (Bolton et
62
Pathotypes in P. horiana
al. 2008). The large number of pathotypes in the chrysanthemum/white rust system surprised us as
P. horiana is considered an asexual microcyclic rust.
The most virulent isolate in the present study was NL1. It may have adapted through mutation or via
genetic exchange between different isolates, each lacking specific elicitors. As P. horiana is a
microcyclic rust and microcyclic rust usually don’t produce spermagonia and don’t produce aecia,
genetic exchange may have been generated asexually by other modes of dikaryotization that have
been described (Ono 2002). Anastomosis of vegetative mycelia at an early stage of basidiospore
infection can occur (Lindfors 1924; Walker 1928). If a similar genetic exchange can take place through
fusion of basidiospore germ tubes or infecting mycelia in P. horiana, this may have resulted in the
virulence spectrum observed. Alternatively, the sexual cycle may be hidden, as was found for other
fungi such as Aspergillus (O'Gorman et al. 2009).
Although the data are limited due to the lack of many possible pathotypes and differential cultivars,
they indicate relations between different isolates. The interaction phenotype profile of isolate CO1
fits into that of the more virulent isolate MY2, which fits into the MY1 profile, which in turn fits into
the profile of NL1. Assuming that these observations can be extrapolated to the complete profile,
this represents a geometric series (Person 1959). If so, this geometric series may be the result of
stepwise and cumulative loss of Avr genes in their respective isolates, in which case these isolates
would have dispersed internationally. Alternatively, isolates may independently have lost these
specific Avr genes and become dominant due to selection pressure on large monocultures of
chrysanthemum cultivars carrying the same specific R genes. Genetic characterization of the different
isolates with neutral markers such as AFLP or microsatellites may help differentiate between these
hypotheses as it may reveal a common ancestry of specific isolates. Unfortunately, performing AFLP
is difficult with P. horiana, because it is not easy to obtain sufficient amounts of non-contaminated
DNA of this obligate pathogen and microsatellite markers have not yet been identified.
In general, most cultivars showed a clear presence or absence of symptoms. However, cvs. 1, 4, 7, 15,
27 and 28 almost never exhibited complete resistance and still developed a few pustules. This
indicates a quantitative aspect to the disease resistance in those cultivars, possibly due to suboptimal
downstream defense reactions. As noted in Table 2.4, infection on these cultivars sometimes occurs
exclusively on particular parts of the plant such as the leaf tips or the stipula. This could be due to
incomplete resistance activation in those plants parts. Cvs. 15 and 16 showed a slower development
of symptoms compared to the other cultivars. Quantitative resistance genes with a minor impact
may play a role in these cultivars (Poland et al. 2009). An interesting observation was also made
about accidentally injured plants. While injury-free replicates of these cultivars showed complete
63
Chapter 2
resistance, injured plants showed clear infection and that only distally from the injury. This may
indicate the requirement of an active vascular system in disease resistance.
Incorporation of resistance to P. horiana will gain importance in future chrysanthemum breeding due
to the decreasing number of registered fungicides for this disease and an increase in the number of
fungicide-resistant strains (see Chapter 3). Also, control of the pathogen via lowering of the relative
humidity is not always possible, such as in semi-covered chrysanthemum growing systems in the
(sub)tropics. Even in heated greenhouses, this approach has been reduced due to the high cost of
energy use (B. Brandwagt, personal communication).
The bioassay we developed allows relatively high-throughput screening of plantlets with one isolate
or a combination of specific isolates. Given the increasingly international trade in cut flowers and the
international and often trans-continental production of planting material, resistance breeding will
need to involve as many R genes as possible. Ideally, the cultivars should be tested with as many as
possible of the pathotypes containing a single Avr gene. Although the proportion of possible
resistance-breaking pathotypes becomes smaller with additional R genes, a specific resistancebreaking pathotype may become important quantitatively due to selection pressure. Timely
detection of all the pathotypes present in a region may help optimize R gene deployment. This could
be achieved with a “pathotype detection network”, consisting of non-fungicide-treated specific
chrysanthemum cultivars that lack single or very few R genes. These cultivars could be placed and
inspected at chrysanthemum growing facilities throughout the region in a similar way as is done for
wheat rust detection (Kolmer 2005). Alternatively, molecular characterization of specific pathotypes
via pathotype-specific markers would be an even faster way for pathotype detection and could make
use of a single universally susceptible cultivar or air-trapped spores. However, as the Avr genes or
linked markers have not been identified yet in P. horiana, this is not an option in the short term.
64
Chapter 3:
Fungicide resistance of Puccinia horiana
Bioassay to screen a particular isolate for fungicide resistance
Fungicide resistance of P. horiana
3.1 Introduction
Due to the potentially high impact of chrysanthemum white rust or Puccinia horiana on
chrysanthemum cultivation, a variety of control measures have been applied to avoid outbreaks or to
eradicate the disease. To avoid the spread of the pathogen by international trade of infected plant
material, chrysanthemum white rust is registered as a quarantine pest by several plant protection
organizations including the European and Mediterranean Plant Protection Organization (EPPO), the
Inter-African Phytosanitary Council (IAPSC), the Junta del Acuerdo de Cartagena (JUNAC) (currently
“Comunidad Andina”, (CAN)) and the North American Plant Protection Organization (NAPPO) (EPPO
2004). To prevent the development of the disease in nurseries, avoidance of leaf wetness and the
maintenance of a low relative humidity by climate control are essential conditions to inhibit the
growth and spread of the pathogen. However, as climate control based on heating and ventilation of
the greenhouses requires a lot of energy, this approach becomes very expensive with increasing
energy costs. Alternative strategies for the control of P. horiana based on hyperparasitic fungi such as
Verticillium lecanii (Srivastava et al. 1985; Whipps 1993) or Cladosporium sp. (García-Velasco et al.
2005) only have moderate effects although the use of bioactive agents for the induction of the
systemic acquired resistance of the plant has been shown recently (Guerrand et al. 2011).
Nowadays, preventive spraying with fungicides is the preferred method to prevent outbreaks of P.
horiana, as also suggested by EPPO (2004). In Belgium, 11 active substances in four chemical groups
are allowed to be used for the treatment of chrysanthemums against P. horiana (Table 3.1)
(www.fytoweb.fgov.be). The chloronitrile molecule chlorothalonil and the dithiocarbamate
molecules mancozeb, metiram and maneb are contact fungicides with multi-site action involved in
glutathione metabolism and inhibition of sulfhydryl groups in amino acids, respectively (Cox 1997;
Agrios 2005). The active substances in the strobilurin group (kresoxim-methyl and azoxystrobin) and
triazole group (myclobutanil, propiconazole, triadimenol, bitertanol and difenoconazole) are systemic
single-site inhibitors that interact with the mitochondrial respiration and ergosterol synthesis,
respectively (Ma and Michailides 2005). Strobilurins are Qo inhibitor (QoI) fungicides and inhibit the
mitochondrial respiration of fungi by binding to the cytochrome bc1 complex at the Qo-site (Bartlett
et al. 2002) while triazoles are demethylation inhibitors (DMIs) binding to the cytochrome P450
sterol 14 ɲ-demethylase enzyme (CYP51) inhibiting the sterol C-14 ɲ-demethylation of a precursor of
ergosterol in fungi (Brent and Hollomon 1995; Ma and Michailides 2005). The efficacy of several of
these fungicides in the control of P. horiana has been investigated, including: chlorothalonil (Oneill
and Pye 1997), mancozeb (Oneill and Pye 1997), maneb (Zandvoort et al. 1968b; Pei et al. 2005),
kresoxim-methyl (Wojdyla and Orlikowski 1999), azoxystrobin (Oneill and Pye 1997; Wojdyla and
Orlikowski 1999; Wojdyla 1999a), propiconazole (Dickens 1990; Dickens 1991), myclobutanil (Dickens
67
Chapter 3
1991; Bonde et al. 1995) and bitertanol (Dickens 1990). Although a variation in efficacy of certain
fungicides is observed, weekly spraying with strobilurin or triazole fungicides is recommended to
control and eradicate the disease (Bonde et al. 1995; Oneill and Pye 1997).
Table 3.1: Fungicides based on a single active substance with a Belgian registration for the control and treatment of rusts on
ornamental plants or P. horiana on chrysanthemum in particular (*). The composition and formulation are those used in the
commercial products. For every substance the prescribed dose for the control and treatment of rusts on ornamental plants
is given. Commercial products (and corresponding active ingredients) that are underlined were used in the resistance
screening. Source: www.fytoweb.fgov.be.
Active substance
Chlorothalonil
Mancozeb
a
Chemical group
Chloronitriles
Dithiocarbamates
b
Composition Formulation Dose
Commercial names
500 g/l
SC
0,3 l/hl
Bravo®, Daconil®
800 g/kg
WP
0,35 kg/hl
Dithane® M 45 / WG, Spoutnik®,
750 g/kg
WG
0,37 kg/hl
Astraman®, Liman® 75 WG
Metiram
Dithiocarbamates 800 g/kg
WG
0,35 kg/hl
Polyram® WG
Maneb
Dithiocarbamates 800 g/kg
WP
0,35 kg/hl
Trimangol® 80
750 g/kg
WG
0,37 kg/hl
Trimangol® WG
Kresoxim-methyl* Strobilurins
500 g/kg
WG
0,1 kg/hl
Candit®
Azoxystrobin
Strobilurins
250 g/l
SC
0,1 l/hl
Ortiva®, Amistar®
Myclobutanil*
Triazoles
200 g/l
EW
0,036 l/hl
Systhane® 20 EW
240 g/l
EC
0,03 l/hl
Systhane® 24 EC
c
Propiconazole
Triazoles
250 g/l
EC
0,08 l/hl
Bumper®, (Tilt®)
Triadimenol
Triazoles
50 g/l
EW
0,4 l/hl
Exact®, Bayfidan® Special
Bitertanol
Triazoles
500 g/l
SC
0,06 l/hl
Baycor® SC 500
Difenoconazole
Triazoles
250 g/l
EC
0,05 l/hl
Difcor® 250 EC, Geyser®, Plover®
a
EC = Emulsifiable concentrate; WP = Wettable powder; SC = Suspension concentrate; WG = Water dispersible granules; EW
b
c
= Emulsion, oil in water. Only a selection of commercial products is given. Not registered any more for the treatment on
ornamental plants.
Despite the widespread use of fungicides such as chlorothalonil and mancozeb, resistance to these
substances is rarely observed, presumably due to their multi-site effects. However, there is a
substantially higher risk for the selection of pathogen strains that are resistant to fungicides acting on
a single metabolic pathway (Brent and Hollomon 1995). The first reports of fungicide resistance in P.
horiana were made around 1980 with strains that were resistant to benodanil and oxycarboxin in
France, the Netherlands and Japan (Abiko et al. 1977; Grouet et al. 1981; Dirkse et al. 1982). Strains
resistant to propiconazole, triforine and bitertanol were observed in the Netherlands (Whipps 1993)
and more recently strains with resistance to both strobilurins and triazoles have been described in
the UK (Cook 2001). Resistance to single-site fungicides is mostly linked to mutations in the target
genes. In several plant pathogenic fungi, strobilurin resistance is conferred mainly by two mutations
(G143A and F129L) in the cytochrome b (cyt b) gene (Bartlett et al. 2002). Except for P. horiana,
resistance to QoI fungicides has not been observed in rusts yet and no point mutations in the cyt b
gene of resistant P. horiana isolates have been found (Grasso et al. 2006b). This suggests that
alternative resistance mechanisms such as bypass reactions or efflux via ATP-binding Cassette (ABC)transporters may be involved (Grasso et al. 2006a; Grasso et al. 2006b). Resistance to triazole
fungicides is often due to mutations in the CYP 51 gene, although overexpression of the gene has
68
Fungicide resistance of P. horiana
been reported as well (Ma and Michailides 2005). Also, a higher efflux or a lower uptake of the
fungicides via ABC-transporters may also play a role in triazole resistance (Gisi et al. 2000).
The intensive use of fungicides can be hazardous for the environment and is losing social acceptance.
Additionally, legal restrictions on the (non-integrated) use of these products are increasing in certain
geographic regions, including the EU. Regional differences in legislation on the application of
fungicides may result in the selection of isolates with resistance to particular active substances. The
spread of fungicide resistant isolates by international trade of chrysanthemums can cause difficulties
to control these isolates in regions were no fungicide resistant isolates were present. To assess the
presence of fungicide resistance in isolates from different geographic origins and different years, a
selection of 17 isolates were tested with 2 strobilurin and 3 triazole fungicides that are often used in
the treatment of chrysanthemums against white rust. Based on these results, the optimal use in
future fungicide applications is discussed.
3.2 Material and methods
3.2.1 Isolates of Puccinia horiana
A selection of 17 isolates of P. horiana based on geographic origin with isolates from Europe, North
and South America and Asia, was made. For the Belgian isolates, isolates from different years were
included in the assay (Table 3.2). Isolates were collected between 2003 and 2009 on commercially
grown chrysanthemum cultivars and used to inoculate healthy cuttings of cultivars Medonia, Taliedo
and cv. 29 (De Backer et al. 2011) as described by Alaei et al. (2009a). From every isolate, single
pustule cultures were established and maintained by three-weekly transfer on rust-free cuttings as
for the first inoculation. To obtain a sufficient amount of inoculum for fungicide resistance screening,
the single pustule isolates were propagated to approximately 25 infected cuttings in an additional
three-week cycle.
69
Chapter 3
Table 3.2: Name, origin, collection year and source of Puccinia horiana isolates used in the fungicide resistance screening.
Collection year
a
Isolate name
Isolate code
Origin
Source
Ph 301
BE1
Belgium, region Ghent
2003
ILVO
Ph 308
BE3
Belgium, region Ghent
2003
ILVO
Ph 522
BE4
Belgium, region Ghent
2005
ILVO
Ph 703
BE16
Belgium
2007
Gediflora
Ph 707
BE5
Belgium, region Mechelen
2007
Gediflora
Ph 715
BE17
Belgium, region Gent
2007
PCS
Ph 801
BE6
Belgium, region Mechelen
2008
Gediflora
Ph Colombia 1
CO3
Colombia, region Boyaca
2008
ICA
Ph Colombia 3
CO2
Colombia, region Cundinamarca
2008
ICA
Ph Japan 13
JP3
Japan, region Tochigi
2009
Deliflor
Ph Malaysia 1
MY3
Malaysia, region Cameron Highlands
2008
Dekker Breeding
Ph Malaysia 2
MY1
Malaysia
2008
Dekker Breeding
Ph NL 802
NL2
The Netherlands
2008
RVZ
Ph PD 20
NL3
The Netherlands
2006
PD
Ph01RE
GB1
United Kingdom
2005
FERA
Ph03SP
GB3
United Kingdom
2005
FERA
Ph USA 1
US1
USA (Massachusetts)
2008
USDA APHIS PPQ
a
Samples obtained via: ILVO: Institute for Agricultural and Fisheries Research, Merelbeke, Belgium / Gediflora, StadenOostnieuwkerke, Belgium / PCS: Proefcentrum voor Sierteelt, Destelbergen, Belgium / ICA: Instituto Colombiano
Agropecuario, Bogota, Colombia / BBV: Bretagne Biotechnologie Végétale, Saint Pol de Léon, France / FERA (formerly CSL):
The Food and Environment Research Agency, York, United Kingdom / Deliflor, Maasdijk, the Netherlands / Dekker Breeding,
Hensbroek, the Netherlands / PD: Plant Protection Service, Wageningen, the Netherlands / RVZ: Royal Van Zanten,
Rijsenhout, the Netherlands / USDA APHIS PPQ: United States Department of Agriculture - Animal and Plant Health
Inspection Service - Plant Protection and Quarantine.
3.2.2 Fungicides
Five systemic single-site fungicides (Table 3.1) were used in the resistance screening of the isolates
(Table 3.2). Two fungicides belong to the strobilurins: kresoxim-methyl (Candit®, BASF, Germany) and
azoxystrobin (Ortiva®, Syngenta, Switzerland), while the other three fungicides are triazoles:
myclobutanil (Systhane® 20 EW, Dow AgroSciences, IN, USA), propiconazole (Tilt®, Syngenta,
Switzerland) and triadimenol (Exact®, Bayer CropScience, Germany). In Belgium, kresoxim-methyl
and myclobutanil are specifically registered for the treatment of P. horiana on chrysanthemum. The
Belgian registration of azoxystrobin and triadimenol includes the application of rusts on ornamental
plants, and therefore includes P. horiana. The use of propiconazole for the treatment of ornamental
plants is no longer registered in Belgium, but it had been used extensively in the control of
chrysanthemum white rust since the 80’s (Dickens 1990) and was therefore also included in the trial.
Fungicide solutions were freshly prepared for each trial in separate Preval Sprayer recipients
(Yonkers, NY, USA), at the concentration listed in their Belgian registration (Table 3.1).
70
Fungicide resistance of P. horiana
3.2.3 Fungicide resistance screening
The bioassay used for the fungicide resistance screening is based on the technique to maintain
isolates as described by Alaei (2009a) and was performed in 1 liter transparent polystyrene
containers (Nalgene®, Thermofisher Scientific, MA, USA). Before each bioassay, the containers and
the lids were thoroughly washed and subsequently sterilized with 70% ethanol. On the bottom of
each container, four fresh chrysanthemum cuttings were inserted in Rockwool blocks AO36/40
(Grodan, the Netherlands) and placed in approximately 200 ml of demineralized water (Alaei et al.
2009a). For all the isolates two cuttings of cv. Medonia and two cuttings of cv. Taliedo were used per
container, except for the Colombian isolates for which four cuttings of cv. 29 were used as these
isolates do not grow on Medonia or Taliedo. The cuttings were inoculated by attaching five heavily
infected leaves with fresh telia to the lid of the containers (with the telia pointing downwards) using
1% water agar as adhesive. Cuttings were sprayed with the different fungicides, or with
demineralized water for the positive control, until drip-off using Preval Sprayers (Yonkers, NY, USA).
The leaves with inoculum on the lid were always misted with demineralized water to ensure a good
sporulation of basidiospores and to avoid possible damage of the inoculum by fungicides. Closed
containers were incubated in a growth chamber at 17°C with fluorescent light (Gro-lux® F58W/GROT8, Osram, MA, USA) during 16 hours per day. Symptoms were evaluated 21 days post inoculation. A
maximum of two isolates were screened simultaneously and per isolate 6 containers were used: 1 for
each fungicide and one for the positive control. Sufficient time was allowed between inoculations
with different isolates to prevent cross-contamination. The reproducibility of the results was tested
by screening five of the isolates twice, separated by at least four weeks.
The level of resistance of an isolate to each of the tested fungicides was classified using 4 categories
as explained below and illustrated in Figure 3.1. Isolates producing heavy symptoms on more than
one leaf per fungicide-treated cuttings, comparable to the symptoms that were observed in the
positive control, received a fungicide resistance score of 3. In case the number of infected leaves with
more than 10 pustules was limited to a maximum 1 leaf per cutting, the isolate received a score of 2
for the given fungicide. A score of “1” was given when symptoms caused by an isolate in a particular
treatment remained limited to a maximum of 10 pustules per leaf on the different cuttings. Finally, a
score of “0” was given when no symptoms were observed after inoculation of fungicide-treated
plants. Isolates with score 2 and 3 that are able to heavily infect the cuttings after treatment with a
particular fungicide are considered to be respectively resistant and highly resistant to that fungicide.
Isolates with a score of 0 and 1 were considered to be highly susceptible and susceptible to the
fungicide, respectively. For isolates that were tested twice, the lowest score was considered to be
most reliable given that the positive control always resulted in a score of 3. In such cases the higher
71
Chapter 3
score likely indicates an imperfect application of the fungicide, resulting in a false positive indication
of resistance.
Figure 3.1: Scoring categories used in the fungicide resistance bioassay. When all cuttings were heavily infected with more
than one leaf bearing more than 10 pustules per cutting, a score of “3” was given (A). When only the number of heavily
infected leaves was limited to one leaf per cutting, a score of “2” was given (B). Plants with less than ten isolated pustules
per leaf received a score of “1” (C). In case no symptoms were observed, plants were scored as “0” (D).
72
Fungicide resistance of P. horiana
3.3 Results
The results of the fungicide resistance screenings are presented in Table 3.3. All control treatments
received a score of “3”, indicating a successful inoculation in each resistance screening assay. The
isolates that were tested twice showed in general a consistent resistance to the applied fungicides. In
total only eight out of 25 interactions were inconsistent and the differences were limited to four
cases where the infection was classified as “0” versus “1” and four cases where the infection was
classified as “2” versus “3”. Five of the eight inconsistent scorings were on account of one isolate (Ph
Spalding UK) with one of the replicates of the fungicide treatments giving consistently higher scoring
results.
Table 3.3: Fungicide resistance scoring of 18 isolates of Puccinia horiana (rows, see Table 3.1) to 5 fungicides (columns, see
Table 3.2). Scores that indicate fungicide resistance are shaded.
no fungicide
control
kresoximmethyl
(Candit)
azoxystrobin
(Ortiva)
propiconazole
(Tilt)
myclobutanil
(Systhane)
triadimenol
(Exact)
Ph 301
3
0
2
2
1
1
Ph 308
a
3
0
0/1
2
1
2
Ph 522
3
0
0
0
0
1
Ph 703
3
0
1
3
3
3
Ph 707
3
1
0
2
1
2
Ph 715
3
2
2
3
3
3
Ph 801
3
0
2
3
2
3
Ph Colombia 1
3
0
0
2
0
2
Ph Colombia 3
3
0
0
3
0
2
Ph Japan 13
3
0
0
2
2
3
Ph Malaysia 1
a
3
0
2
2
2
2
Ph Malaysia 2
3
0
2
2
3
3
Ph NL 802
3
0
0
2
2
2
Ph PD 20
3
1
0
3
3
3
a
3
0
0
3
2
2/3
a
3
0/1
0/1
2/3
2/3
2/3
3
0
0
0
0
0/1
Ph01RE
Ph03SP
Ph USA 1
a
a
Isolates tested twice.
Resistance to the strobilurin fungicides kresoxim-methyl and/or azoxystrobin was only observed in
five isolates. Isolates Ph 301, Ph 801, Ph Malaysia 1 and Ph Malaysia 2 showed resistance to
azoxystrobin but not to kresoxim methyl and isolate Ph 715 was resistant to both kresoxim-methyl
and azoxystrobin. None of the isolates was resistant to kresoxim-methyl while being susceptible to
azoxystrobin. Most isolates were susceptible or highly susceptible to both strobilurin products while
being (highly) resistant to at least one of the triazole fungicides. None of the isolates was able to
73
Chapter 3
cause an infection on plants treated with strobilurin fungicides to the same extent as the infection
observed on the control plants.
Compared to the resistance against strobilurins, substantially more isolates were resistant to
triazoles. From the 17 isolates that were included in our study, 10 were resistant or highly resistant
to all triazole fungicides tested. The highest level of resistance was observed in isolates Ph 703, Ph
715 and Ph PD 20, which were highly resistant (score 3) to all the triazole fungicides included in our
study. In the other isolates with general resistance to the triazole fungicides, the average resistance
to at least one of the substances is lower than score 3. Isolates Ph 308, Ph 707, Ph Colombia 1 and Ph
Colombia 3 were susceptible or highly susceptible to the fungicide myclobutanil, although they
showed resistance or strong resistance to propiconazole and triadimenol. Isolate Ph 301 was only
resistant to propiconazole and was susceptible to myclobutanil and triadimenol.
Two isolates were susceptible or highly susceptible to all fungicides we applied in this study. Isolates
Ph 522 and Ph USA 1 in general were highly susceptible to both the strobilurins and triazoles, except
for a slightly lower susceptibility to triadimenol.
3.4 Discussion
Fungicide treatment with contact and/or systemic fungicides is currently the most important strategy
to prevent the outbreak of chrysanthemum white rust. However, due to intensive fungicide
applications, fungicide resistance to mainly systemic single site fungicides has been reported (Grouet
et al. 1981; Dirkse et al. 1982; Cook 2001). These observations are confirmed by the results obtained
in the present study.
The bioassay for fungicide resistance screening has proven to be reliable for the inoculation of
chrysanthemum cuttings with P. horiana (Alaei et al. 2009a). Using this assay, quite consistent results
were also obtained in fungicide resistance scoring. In none of the cases in which an inconsistent
scoring was obtained between replicates, a variation of more than one scoring category was
observed. Moreover, a switch between the resistant and the susceptible categories has not been
observed. Isolate Ph Spalding UK showed a consistently higher score in the fungicide treatments of
one of the replicates, suggesting a consistent variation was present for all fungicide treatments.
Possibly the fungicides were not applied in the same amount (e.g. not sprayed until drip off) or in a
lower concentration than planned. The lowest scores were considered to be the best representation
of the fungicide resistance, as they best represented the capacity of the fungicides for the given
isolate.
74
Fungicide resistance of P. horiana
A clear difference in fungicide resistance between strobilurin and triazole fungicides could be
observed in the isolates included in this study, with a substantially lower resistance to strobilurins.
This is probably due to the fact that triazole fungicides were first introduced in the 1970s (Ma and
Michailides 2005) while the strobilurins azoxystrobin and kresoxim-methyl were only first applied in
1996 (Bartlett et al. 2002). Combined with the recommendations of weekly preventive spraying
(EPPO 2004) and the high popularity of triazole fungicides in P. horiana control before the
introduction of strobilurins (Bonde et al. 1995), a selection for triazole resistant strains is logical.
Resistance to strobilurin fungicides was limited to four isolates resistant to azoxystrobin only and one
isolate being resistant to both azoxystrobin and kresoxim-methyl. Also, the lower infection level of
the plants treated with the fungicides compared to the plants in the control treatment suggests only
a partial resistance to these fungicides (Table 3.3). Resistance to strobilurins was only observed in
Belgian and Malaysian isolates but appeared already in 2003. Resistance to azoxystrobin was also
reported by Cook (2001) as early as 2001 while Wojdyla (2004b) still listed it as effective against the
isolates he used in 2004. Due to the common mode of action of strobilurins, cross resistance to
different strobilurins is generally observed (Gisi et al. 2000; Gisi et al. 2002). Nevertheless, strong
differences in intrinsic activities, systemic behavior and pathogen spectrum are observed for the
individual QoI fungicides (Gisi et al. 2000). The substantially larger number of isolates with resistance
to azoxystrobin than those resistant to kresoxim-methyl suggests that the efficiency of azoxystrobin
in the control of P. horiana is lower than that of kresoxim-methyl. It also suggests that isolates
resistant to azoxystrobin are not necessarily cross-resistant to kresoxim-methyl. The only isolate
resistant to kresoxim-methyl (Ph 715) also showed resistance to azoxystrobin, indicating a possible
cross-resistance. Remarkably, strobilurin resistance was never observed in other rusts than P.
horiana although strobilurins are used on a regular basis against wheat rusts. This can be explained
by the fact that the one of the most common mutations conferring strobilurin resistance (G143A) is
lethal in rusts (Grasso et al. 2006b). However, other mutations linked to QoI resistance could be
found in rusts neither. As a consequence, alternative resistance mechanisms such as alternative
respiration or efflux transporters should be involved in strobilurin resistance in P. horiana (Grasso et
al. 2006a; Grasso et al. 2006b; Fernandez-Ortuno et al. 2008). Further investigation of the crossresistance between azoxystrobin and kresoxim-methyl, or finding alternative pathways for strobilurin
resistance is necessary to substantiate these hypotheses.
Of the 17 isolates included in this study, 10 isolates are resistant to all triazoles we applied and 14 are
resistant to both propiconazole and triadimenol (Table 3.3). This large number of resistant isolates
illustrates a strong decrease of efficiency of triazole fungicides in the control of P. horiana. Despite
the original effectiveness of triazoles against P. horiana, resistance to propiconazole has been
75
Chapter 3
reported in the Netherlands (Whipps 1993) and resistance to propiconazole and myclobutanil has
been reported in the UK (Cook 2001). The efficacy of propiconazole, myclobutanil and triadimenol for
the prevention and even eradication of the disease has previously been shown in several studies
(Dickens 1990; Bonde et al. 1995; Oneill and Pye 1997; Wojdyla 2002), with a protective effect of
myclobutanil higher than that of propiconazole (Dickens 1991). This is also confirmed by our results
in which myclobutanil is able to control up to 7 isolates from which 4 isolates are resistant to both
propiconazole and triadimenol. The higher efficiency of myclobutanil compared to propiconazole or
triadimenol can be explained by the differences in spectrum that are observed between different
compounds of the DMIs (Gisi et al. 2002). A clear cross-resistance between propiconazole and
triadimenol was observed, but not all isolates that are resistant to those substances were resistant to
myclobutanil. However, isolates resistant to myclobutanil were always cross-resistant to
propiconazole and triadimenol. Cross-resistance is also expressed between different DMI compounds
for isolates of other sensitive pathogens such as Mycosphaerella graminicola (Gisi et al. 1997).
Only two isolates (Ph 522 and Ph USA 1) were susceptible to both triazoles and strobilurins used in
this study indicating that these isolates were not under a high selection pressure of DMI or QoI
fungicides. This observation demonstrates that some strains of P. horiana can still be controlled with
systemic single-site fungicides, although a good resistance management strategy must be followed to
avoid build up of resistance. One of the major components of such a strategy is to use a mixture of
fungicides with different modes of action or to alternate fungicide classes between different
treatments (Brent and Hollomon 2007). Due to the different modes of action of strobilurins and
triazoles, the chance for cross-resistance between those fungicide groups is very small, making them
suitable to be used in combination or in combination with broad spectrum fungicides. However, the
high level of resistance of several P. horiana isolates to triazoles has to be taken in account, and the
occurrence of resistance has to be monitored. From the available triazole fungicides, myclobutanil is
preferred, although some resistant strains are present. The very high level of resistance to
propiconazole and triadimenol makes these fungicides less suitable for the control of P. horiana.
Both strobilurins we tested can still be used in practice, although kresoxim-methyl is the preferred
compound due to its higher efficiency. These preferred compounds are currently the only
compounds having a specific registration for the treatment of P. horiana on chrysanthemum in
Belgium (www.fytoweb.be).
To support resistant management strategies, a fast detection system of resistant strains would be
necessary. This would involve the molecular characterization and detection of resistant strains based
on the mutations in the genes of the targets of these fungicides and the identification of alternative
biochemical pathways. In combination with a spore detection system, such a molecular detection
76
Fungicide resistance of P. horiana
system could lead to a guided application of fungicides and could help to reach a more sustainable
control of the pathogen.
77
Part II:
Genetic variation in Puccinina horiana
and spore detection in air samples
using molecular techniques
Chapter 4:
Genotypic diversity of Puccinia horiana
based on newly identified SNP markers
A version of this chapter will be submitted to Molecular Ecology
.
Typical alignment obtained after CRoPS™ analysis.
A candidate SNP marker is indicated in red.
Migration, survival and recombination of P. horiana
4.1 Introduction
Chrysanthemum white rust is caused by Puccinia horiana Hennings and is one of the most important
diseases in chrysanthemum production. The pathogen can infect more than 10 different
chrysanthemum species of which Chrysanthemum x morifolium is economically the most important
one. After the first detection of the pathogen in Japan in 1895 (Hennings 1901; Hiratsuka 1957), it
was also reported in China and South Africa some time before 1963 (Priest 1995). In 1963 P. horiana
was introduced in England on plants originating from Japan (Baker 1967), followed by introductions
in continental Europe in 1964 on cuttings originating from South Africa (Baker 1967; Firman and
Martin 1968; Priest 1995). Currently it is present in most chrysanthemum producing areas
throughout the world (Whipps 1993; EPPO 2004). P. horiana is classified as a quarantine pathogen in
Europe, Africa and America (EPPO 2004) and also listed as regulated pest by the International Plant
Protection Convention (IPPC).
At an optimal temperature between 17°C and 20°C symptoms appear as chlorotic spots, 7 to 10 days
post infection (dpi) and teliospores develop 14 to 18 dpi on the underside of the leaves (Firman and
Martin 1968). These teliospores can germinate without a period of dormancy to form a promycelium
on which two binucleate basidiospores are produced (Kapooria and Zadoks 1973; Kohno et al. 1974;
Kohno et al. 1975). Basidiospores are the mobile and infective propagules that are easily spread by
wind and can infect plants in neighboring fields (Zandvoort 1968). Since it is a self-fertilizing rust,
each basidiospores is genetically uniform unless mutation or anastomosis between different
genotypes occurs (Rademaker and de Jong 1987; Ono 2002).
The presence of different pathotypes of the pathogen was reported on several occasions (Dickens
1968; Yamaguchi 1981; Wojdyla 1999b). However only recently, the full complexity of the
pathosystem was shown in a comprehensive study in which a collection of 22 isolates, covering a
broad geographical range, was tested on a set of 36 cultivars (De Backer et al. 2011). Based on the
gene-for-gene concept (Flor 1956), at least 7 avirulence genes were postulated in this pathosystem.
To study the genetic diversity of a population, dominant and co-dominant markers can be used.
Dominant markers based on Amplified Fragment Length Polymorphism (AFLP) or random
amplification of polymorphic DNA (RAPD) allow the rapid screening of different loci within a genome
without the need of sequence data but they do not allow amplification in a background of
contaminating plant material and microorganisms, making them less suitable for marker
development in obligate biotrophes such as P. horiana. Co-dominant markers such as Single
Nucleotide Polymorphisms (SNPs) or Simple Sequence Repeat (SSR) markers can be used, which are
usually locus specific. SNPs have relatively low mutation rates compared to SSR markers, making
83
Chapter 4
them more suitable for population genetics in distantly related isolates (Brumfield et al. 2003; Morin
et al. 2004). However, because microsatellites have multiple alleles while SNPs are usually biallelic,
relatively more SNP markers are necessary for estimating genetic variation between isolates (Morin
et al. 2004). SNP markers can be species specific and next-generation high-throughput genome
sequencing technologies now allow the identification of SNPs at much lower cost (Ganal et al. 2009).
However, for non-model organisms such as P. horiana the collection of sufficient sequence data from
multiple isolates is not an option yet (Schlötterer 2004). In such cases, SNP markers can still be
developed via the sequencing of randomly distributed AFLP fragments (Nicod and Largiad 2003;
Morin et al. 2004). This approach combines the power of genome complexity reduction of AFLP with
high throughput next generation sequencing and has been described as Complexity Reduction of
Polymorphic Sequences (CRoPS™) (van Orsouw et al. 2007).
In this study, we developed a set of molecular-markers to determine the genetic variability within a
worldwide collection of isolates of P. horiana. In a first step, SNP mining was performed using the
CRoPS™ technology on a set of isolates representing different pathotypes. In a second step, specific
PCR primers were developed to amplify the SNP-containing fragments. These primers were used to
determine the genotypic variability of 45 isolates of P. horiana. These data were used to assess
diversity in the set of isolates in our worldwide collection and verify the presence and role of
recombination in the formation of new genotypes. Based on the phylogenetic relations, the
worldwide migration of the pathogen and the endemic or exotic character of isolates in different
regions were analyzed. The relationship between these genotypic data and the pathotype data
obtained by De Backer et al. (2011) were determined, including the possibility for the development
of pathotype-specific SNP markers.
84
Migration, survival and recombination of P. horiana
4.2 Material and methods
4.2.1 Puccinia horiana isolates
Forty nine isolates of P. horiana were collected from commercially-grown Chrysanthemum x
morifolium plants between 2003 and 2010 (Table 4.1). Isolates were mainly selected based on
geographic distribution including Europe, Asia and North and South America. In contrast to the other
isolates, the isolates that were sampled in the Netherlands cannot be considered field isolates as
they were sampled from a mixed culture in a biosecurity facility. This mixed culture included local
isolates as well as isolates from customs interceptions. Belgian isolates were collected during
multiple years (2003-2009). Some isolates collected after 2005 were selected based on potential
pathotype differences, as they were mainly collected from cultivars previously reported as resistant.
Single pustule isolates were established, after which isolates were maintained by three-weekly
transfer onto fresh cuttings as described previously (Alaei et al. 2009a; De Backer et al. 2011). The
pathotype information of 22 of these isolates was obtained from De Backer et al. (2011). The
pathotype of four isolates (see Table 4.1) was determined in this study, using the methods of De
Backer et al. (2011). From a total of 45 isolates sufficient amounts of DNA could be obtained for
genotyping (see further). From four of the isolates with a known pathotype, no genotypic data could
be obtained as the isolates were no longer available after the pathotyping study. Inoculations and
pathogen maintenance were conducted in laboratories and growth chambers with appropriate
biosafety levels.
85
Chapter 4
Table 4.1: Information on the isolates used, including their name, origin, collection year, source, availability of pathotype information, and the
parts of this study they were used in.
Isolate
code
BE1
BE2
BE3
BE4
BE5
BE6
BE7
BE8
BE9
BE10
BE11
BE12
BE13
BE14
BE15
CO1
CO2
CO3
CN1d
FR1
FR2
FR3
GB1
GB2
GB3
JP1
JP2
JP3
JP4
LK1
MY1
MY2
MY3
MY4
MY5
MY6
MY7
NL1c
NL2
NL3c
NL4c
NL5c
NL6c
PL1
US1
US2d
US3d
US4d
US5d
Isolate name
Origin
Ph 301
Ph 307
Ph 308
Ph 522
Ph 707
Ph 801
Ph 312
Ph 515
Ph 520
Ph 531
Ph 712
Ph 724
Ph 804
Ph 805
Ph 902
Ph Colombia 2
Ph Colombia 3
Ph Colombia 1
Ph China 1
Ph CWR 07.B18
Ph CWR 08.A4
Ph CWR 07.A20
Ph01RE
Ph02SG
Ph03SP
Ph Japan 2
Ph Japan 3
Ph Japan 13
Ph Japan 12
Ph Sri Lanka1
Ph Malaysia 2
Ph Malaysia 3
Ph Malaysia 1
Ph Malaysia 8
Ph Malaysia 4
Ph Malaysia 6
Ph Malaysia 7
Ph01H
Ph NL 802
Ph PD 20
Ph PD 13
Ph PD 28
Ph PD 36
Ph01P
Ph USA 1
Ph USA CA 1-10
Ph USA CT1-10
Ph USA MD1-10
Ph USA PA1-10
Belgium, Ghent region
Belgium, Ghent region
Belgium, Ghent region
Belgium, Ghent region
Belgium; Mechelen region
Belgium; Mechelen region
Belgium, Ghent region
Belgium, Ghent region
Belgium, Ghent region
Belgium, Ghent region
Belgium, Ghent region
Belgium, Ghent region
Belgium; Mechelen region
Belgium; Mechelen region
Belgium; Mechelen region
Colombia; Boyaca region
Colombia; Cundinamarca region
Colombia; Boyaca region
China
France
France
France
United Kingdom
United Kingdom
United Kingdom
Japan, Hiroshima region
Japan, Hiroshima region
Japan, Toshigi region
Japan, Hiroshima region
Sri Lanka, Nuwara Eliya region
Malaysia, Cameron Highlands region
Malaysia
Malaysia
Malaysia
Malaysia
Malaysia
Malaysia
The Netherlands
The Netherlands
The Netherlands
The Netherlands
The Netherlands
The Netherlands
Poland
USA; Massachusetts
USA; California
USA; Connecticut
USA; Maryland
USA; Pennsylvania
Collection
year
2003
2003
2003
2005
2007
2008
2003
2005
2005
2005
2007
2007
2008
2008
2007
2008
2008
2008
2010
2007
2008
2007
2005
2005
2005
2008
2008
2009
2009
2009
2008
2008
2008
2008
2008
2008
2008
2006
2008
2008
2008
2008
2008
2006
2008
2010
2010
2010
2010
a
Sourcea
ILVO
ILVO
ILVO
ILVO
Gediflora
Gediflora
ILVO
ILVO
ILVO
ILVO
PCS
PCS
Gediflora
Gediflora
Gediflora
ICA
ICA
ICA
Deliflor
BBV
BBV
BBV
FERA
FERA
FERA
Deliflor
Deliflor
Deliflor
RVZ
RVZ
Dekker Breeding
Dekker Breeding
Deliflor
Dekker Breeding
Dekker Breeding
Dekker Breeding
Dekker Breeding
PD
RVZ
PD
PD
PD
PD
INSAD
USDA APHIS PPQ
USDA ARS NAA
USDA ARS NAA
USDA ARS NAA
USDA ARS NAA
Pathotype
known
Xb
Xb
Xb
Xb
Xb
Xb
AFLP
CRoPS™
X
X
X
X
X
X
Xb
Xb
Xb
Xb
Xb
Xb
Xb
Xb
Xb
Xb
X
X
Xb
Xb
X
X
X
X
X
X
X
X
X
SNP profile
known
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Xb
Xb
X
X
Xb
Xb
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Samples obtained via: ILVO: Institute for Agricultural and Fisheries Research, Merelbeke, Belgium / Gediflora, Staden-Oostnieuwkerke,
Belgium / PCS: Proefcentrum voor Sierteelt, Destelbergen, Belgium / ICA: Instituto Colombiano Agropecuario, Bogota, Colombia / BBV:
Bretagne Biotechnologie Végétale, Saint Pol de Léon, France / FERA (formerly CSL): The Food and Environment Research Agency, York, United
Kingdom / Deliflor, Maasdijk, the Netherlands / Dekker Breeding, Hensbroek, the Netherlands / PD: Plant Protection Service, Wageningen, the
Netherlands / RVZ: Royal Van Zanten, Rijsenhout, the Netherlands / INSAD: Research Institute of Pomology and Floriculture, Skierniewice,
Poland / USDA APHIS PPQ: United States Department of Agriculture - Animal and Plant Health Inspection Service - Plant Protection and
Quarantine / USDA ARS NAA: United States Department of Agriculture - Agricultural Research Service - North Atlantic Area.
b
Isolates included in pathotype study by De Backer et al. (2011)
c
The geographic origin of the isolates obtained from PD is not known, but they included custom interceptions. As they were maintained in a
mixed culture with other isolates, there was an increased probability of recombination among these isolates.
d
Isolates for which only DNA is available (so no pathotype tests can/could be performed)
86
Migration, survival and recombination of P. horiana
4.2.2 CRoPS™ analysis
Eight isolates were selected for the CRoPS™ analysis, mainly based on their differential pathotype
profiles (Table 4.1). P. horiana genomic DNA (gDNA) was extracted from basidiospores instead of
pustules to minimize contaminating (plant) DNA. Basidiospores were collected using a modified
version of the harvesting setup described in Alaei et al. (2009a). Infected chrysanthemum leaves
were attached to the lid of 10 cm Petri dishes with the telia pointing downwards. The leaves were
misted with demineralized water using a Preval Sprayer (Yonkers, NY, USA) and incubated at 17°C in
the dark for a maximum of 14 hours. Basidiospore “prints” on the bottom of the Petri dishes were
verified microscopically at 100x magnification for contaminating spores such as those from
Malassezia restricta (Alaei et al. 2009a). For each Petri dish, basidiospores were collected using 1 ml
of 0.1% Igepal CA-630 (a surfactant) in water (Sigma-Aldrich, MO, USA). Spores were collected in 1.5
ml microcentrifuge tubes and concentrated by centrifugation and careful removal of most of the
supernatant. Basidiospores harvested from up to 10 Petri dishes per isolate were combined in a
single pellet and stored at -20°C until DNA extraction. Genomic DNA was extracted from the
basidiospores using a cetyltrimethylammonium bromide (CTAB) extraction protocol as described by
Alaei et al. (2009a). DNA was quantified using a Nanodrop spectrophotometer (Thermo Scientific, DE,
USA) and stored at -20°C until further analysis.
The CRoPS™ analysis was performed by Keygene NV (the Netherlands). The genomic DNA (10-50 ng)
was used to generate AFLP fragments with restriction enzymes EcoRI and MseI. Fragments were
selectively amplified with a single primer combination consisting of the EcoRI-primer without
selective bases and the MseI-primer with one selective base (MseI-G) as described by Vos et al.
(1995). The AFLP primers were extended with four-nucleotides-long sample identification tags for a
correct assignment of the sequences to the corresponding samples after 454 sequencing. A sequence
library of the AFLP fragments was prepared based on the Roche GS FLX Titanium protocol (Roche,
Germany) and sequencing on a GS FLX Titanium sequencing instrument (Roche).
4.2.3 SNP mining and marker development
Sequence reads from the CRoPS™ analysis were quality clipped and assembled in contigs which were
aligned to determine putative polymorphisms including SNPs, indels and microsatellites. Contigs
containing putative polymorphisms were selected for further analysis if (I) they had a minimum
length of 100 nucleotides, (II) 10 or more reads were included in the allignment, (III) sequences from
at least four of the eight isolates were included, (IV) at least two reads per sample contained the
polymorphism and (V) each allele was present in 10 to 90% of the reads. Candidate SNPs or other
polymorphisms had to be at least 18 nucleotides from the 5’ and 3’ end of the fragment to be able to
87
Chapter 4
develop primers flanking the polymorphism. Such primers were developed using CLC DNA
Workbench (CLCBio, Denmark). Primers were tested using two or three candidate differential isolates
and the optimum annealing temperature (Tan) in the range of 45 to 60°C was determined.
PCR amplification was performed in a GeneAmp PCR System 9700 PE thermo cycler (Applied
Biosystems, CA, USA). Fragment-specific primer sequences that were used in the PCR reactions as
well as their optimal annealing temperatures are listed in Table 4.2. PCR reactions were performed in
50 ђl containing PCR Reaction Buffer (50 mM Tris/HCl, 10 mM KCl, 5 mM (NH4)2SO4, 2.0 mM MgCl2,
pH 8.3), 0.2 mM of each dNTP, 0.2 ђM of forward and reverse primer and 0.02 U/ђl FastStart Taq
DNA Polymerase. Per PCR reaction, 5 ђl of 50-fold diluted whole genome amplification product (see
below) was added. The cycle parameters were: an initial preheating at 94°C for 2 min, 35 cycles of
denaturation at 94°C for 30 s, annealing at 46°C or 54°C (depending on primer combination; see
Table 4.2) for 45 s and extension at 72°C for 45 s. A final extension step at 72°C for 7 min finished the
reaction after which the samples were cooled to 4°C. PCR products were checked by loading 10 ђl on
an EtBr-containing agarose gel (2.0% agarose in 0.5x TAE-buffer). PCR products were prepared for
Sanger sequencing using the BigDye® Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems)
followed by an ethanol precipitation step. Samples were loaded on a 3730XL DNA sequencer
(Applied Biosystems) for capillary electrophoresis and genetic analysis.
4.2.4 Genotyping using SNP markers
The SNP and SSR markers developed using the CRoPS™ data were used to genotype the 45 P. horiana
isolates listed as such in Table 4.1. Preliminary assays determined that the primer pairs flanking each
marker only amplified the P. horiana targets when using DNA samples from infected plant samples.
This allowed the use of pustules instead of basidiospores as starting material for DNA preparation.
Fresh pustules, containing mostly teliospores, were collected from Chrysanthemum cuttings that had
been inoculated three weeks earlier. Five to 30 mg of fresh pustules were frozen in liquid nitrogen
and ground with a pestle before DNA extraction using the Invisorb® Spin Plant Mini Kit (Invitek,
Germany) following the manufacturer’s instructions but without the optional RNase step. DNA was
eluted in 50 ђl of the kit’s buffer D and stored at -20°C.
Considering a large amount of DNA was required to optimize and test all developed primer
combinations, a whole genome amplification step was introduced using the Illustra GenomiPhi V2
DNA Amplification Kit (GE Healthcare, UK) following the manufacturer’s instructions. The
amplification products were stored at -20°C until further processing.
88
Migration, survival and recombination of P. horiana
4.2.5 Phylogeny and cluster analysis
SNP and SSR data of the isolates were converted to binary data by assigning the value “0” to the
most frequent allele among all isolates and the value “1” to the least frequent allele (Table 4.2). From
the binary data, similarity matrices were calculated using the Simple Matching coefficient (Gower
1971). Cluster analysis was done with the Unweighted Pair Group Method using Arithmetic averages
(UPGMA). The cophenetic correlation coefficients between the dendrogram-derived similarities and
the matrix similarities were calculated for each cluster to verify the adjustment between both
matrices (Farris 1969).
For each genotype we determined if it could be a recombinant of any other two genotypes in our
collection. Given that the chance for repeated emergence of a given SNP is extremely low, a
genotype should contain all the markers that are shared by the parents and a combination of alleles
that are different between the parents to qualify as a potential recombinant. Preferentially it should
also share rare markers with each of its parents. The number of combinations in which each
genotype in the dataset can be a potential parental genotype of each other genotype was
determined computationally using an in-house generated routine.
For the isolates with pathotype data, a cluster analysis based on the binary interaction phenotype
profiles was performed. Calculation of similarity matrices, cluster analysis and cophenetic correlation
coefficients was conducted as for the genotype data. The linkage between specific pathotypes or
avirulence genes and isolate genotypes was assessed by comparing the genotypic and pathotypic
clustering. All calculations were performed using the Bionumerics software (Applied Maths, Belgium).
89
Chapter 4
Table 4.2: AFLP fragments containing polymorphisms and corresponding primers for PCR amplification of the region surrounding the
polymorphism. For each fragment, the CRoPS™ fragment number, the GenBank accession number and the forward and reverse primers and
optimal annealing temperature (Tan) are listed. Primers used for the sequencing reaction are underlined. For each polymorphism, the
identification code and the position of the polymorphism within the fragment are listed. Each polymorphism is presented in its binary state,
either as an empty square () for the value “0” or a filled square () for the value “1” as used in Figure 4.1.
AFLP fragment
Polymorphism
Forward Primer
CRoPS™ #
GenBank accession
#
ID
position
002
JY084340 JY084341
SNP 002
72
003
JY084342
JY084343
SNP 003-1
90
SNP 003-2
155
SNP 006
124
SNP 046
60
SNP 141
115
SNP 174
23
006
046
141
174
180
201
268
292
358
431
453
486
499
568
595
597
603
661
732
793
JY084345
JY084344
JY084347
JY084346
JY084349
JY084348
JY084350
JY084351
JY084353
JY084352
JY084355
JY084354
JY084357
JY084356
JY084359
JY084358
JY084361
JY084360
JY084364
JY084363
JY084362
JY084365
JY084366
JY084368
JY084367
JY084389
JY084388
JY084370
JY084369
JY084371
JY084372
JY084373
JY084374
JY084376
JY084375
JY084377
JY084378
JY084380
JY084379
JY084381
JY084382
JY084383
SNP 180-1
31
SNP 180-2
112
SNP 180-3
169
SNP 201
160
SNP 268
29
SNP 292
18
SNP 358
112
SNP 431-1
SNP 431-2
96
SNP 453
50
SNP 486-1
24
SNP 486-2
86
SNP 499
40
SNP 568
47
SNP 595
36-37
SNP 597
100
SNP 603
98
SNP 661
51
SNP 732
33
SNP 793-1
SNP 793-2
854
JY084384
JY084385
SNP 854-1
SNP 854-2
SNP 854-3
892
895
JY084390
JY084391
JY084386
JY084387
94
SSR 892
SNP 895
86
label
A
T
G
A
C
A
G
A
C
A
T
C
C
T
G
A
G
C
A
G
T
A
G
T
T
C
T
C
T
C
G
A
A
G
T
G
A
G
G
A
G
C
AC
CA
C
T
C
A
A
G
T
C
T
C
C
243
A
T
64
C
C
108
A
A
114
G
50-64 5xAAC
50-67 6xAAC
C
26
A
Reverse Primer
Tan
Primer name
Sequence (5'-3')
Primer name
Sequence (5'-3')
002 Fw-gen
ACAAAGTTCCGTAGCAAATCTC
002 Rv-gen
CTACCTGTTTACAAAGTTATGC
46°C
003 Fw-gen
ATTCGGGACCCAAGTTGG
003 Rv-gen
AGGGGCTACTCCTTATTGG
53°C
006 Fw-gen
GACTCATCAAGATGGCTCAC
006 Rv-gen
TGGAGCAAAGTGATTGGTTAC
53°C
046 Fw gen
ATTCAAAGGGAAATGGAGAGAAG
046 Rv-gen
ACTGAAAATTGAAGCAAAAATTGC
46°C
141 Fw-gen
CACTTCTTTTCAGTATCAC
141 Rv-gen
TTAAGAAGGGATGCACG
46°C
174 Fw-gen
CCTTTCAATTGATCCAGATCCT
174 Rv-gen
TCTTTGGTACAAGAATAACCACAATG
53°C
180 Fw-gen
GAAATCCTGTCTGCGCC
180 Rv-gen
AGGAAAATTCAGTCGGATG
53°C
201 Fw-gen
TCGGAGGGTCTCAGCATTG
201 Rv-gen
GACTGGGTCGAAATGTGAGT
46°C
268 Fw-gen
CAAAAAAATGGAGGGCG
268 Rv-gen
ACCTTGACGAGCATCTT
46°C
292 Fw-gen
AGCATCTTTCCTTTACC
292 Rv-gen
GAAGTTTAGAGGACGAG
46°C
358 Fw-gen
AAACGCGGAACATAGAAA
358 Rv-gen
GGAAAAAAAGTGATGGAGG
46°C
431 Fw-gen
ACAACAACTCAACACTCC
431 Rv-gen
AAGCCGAGATGATCGAA
46°C
453 Fw-gen
ATTCACCAAGCAGCCCATC
453 Rv-gen
AGCCAAGAGCACTATCCAAG
46°C
486 Fw-gen
TGACCTTGTTTCTTGGCA
486 Rv-gen
AAGTAGAAAATCTGGGAGC
53°C
499 Fw-gen
CTGAACTCATCTGGAATGTC
499 Rv-gen
CCTCGACAAGTTCCAAAGG
46°C
568 Fw-gen
GCTTGGGTCAGGAGAGAAG
568 Rv-gen
GAATTGACGAACTTTTCGGAG
46°C
595 Fw-gen
GCTAGAGAGGTCCTTTTATTC
595 Rv-gen
GAGAGTGAATGTTTGGAAGG
53°C
597 Fw-gen2
TAAGAAAGGCTAAGACCCTTTG
597 Rv-gen
GATGGTTTTCTCTGCGGCTG
53°C
603 Fw-gen
TTAAGTTGACTGCCACACCG
603 Rv-gen
TGCGTCAATTCCCAGGTCG
53°C
661 Fw-gen2
CCAACAAGCGGTTCACTGAC
661 Rv-gen
CCCAGATAACTCACATACAACC
53°C
732 Fw-gen
GGAAAAAGAAGCAAATGGG
732 Rv-gen
ATTCCTAGTCTGTGTTCCT
53°C
793 Fw-gen
ATTCAACTGTCTTTACCACAAC
793 Rv-gen
AGGATGAGGTGTTGAAAGAAC
53°C
854 Fw-gen
TAGAATCTGAAATCCCAGATTCC
854 Rv-gen
892 Fw-gen
TAAGACTGGGTTCGACA
892 Rv-gen
TTCATATCCGAACGTTG
46°C
895 Fw-gen
GCTGAGTAGATGAAGATTATGTTGG
895 Rv-gen
GAATTCAATGCTTGTCTCGACC
53°C
90
GAATTCTTTTAGAGAAATACTGGAGGAC 53°C
Migration, survival and recombination of P. horiana
4.3 Results
4.3.1 CRoPS™ analysis and SNP mining
A total of 308,410 high quality reads with an average read length of 144.6 base pairs (bp) were
obtained in the CRoPS™ analysis of the 8 isolates. Of these, 229,126 could be used in alignments for
the assembly of 16,196 contigs with an average of 14.2 reads per contig. 77,304 singleton reads and
1980 chloroplast and mitochondrial reads could not be used in contig assembly. In 184 of the 16,196
contigs (1.1%) corresponding to approximately 2,332,224 bp, at least 1 SNP, or 0.01% of all analyzed
nucleotides, could be found. There were 38 alignments with more than 10 reads from at least 4
different isolates and with a length of at least 100 bp. Many of these alignments showed repetitive
sequences and were AT rich, complicating primer development. In the resulting PCR reactions, 6
primer combinations gave no amplicon, even in a Tan gradient PCR from 45 °C to 60 °C. For another 7
primer combinations, amplicons of the expected size were generated but the putative SNP could not
be confirmed by Sanger sequencing. The remaining 25 fragments had an average length of 172 bp
representing a total of 4300 bp. Within these fragments 33 polymorphisms were detected. One
polymorphism was a Simple Sequence Repeat (SSR) (SSR 892) consisting of either five or six “AAC”
repeats. The remaining 32 polymorphisms were SNPs, which were present in binary form for all
isolates. From the 33 polymorphisms, 24 (23 SNPs and one SSR) gave a non-redundant SNP-profile
(Figure 4.1). In 20 fragments a single SNP was found, four fragments (CRoPS™# 003, 431, 486 and
793) contained two SNPs, and two fragments (CRoPS™# 180 and 854) contained three SNPs. The
different SNPs within a single fragment were linked and their combined profile also showed a binary
format, except for the SNPs in fragments 793 and 431. Also the SNPs in fragments 141, 201 and 732
and those in fragments 002 and 174 showed identical patterns in all 45 isolates (Figure 4.1). Of the 33
polymorphisms, 31 SNPs were predicted by the CRoPS™ analysis while two new SNP’s were identified
(SNP 180-3 and SNP 431-1) in 45 isolates. This would result in a frequency of 1 SNP in 100 Kb per
isolate (2/4300 per 45 isolates).
4.3.2 Genotyping
All isolates included in the genotyping were homozygous for the polymorphisms we determined.
Within the 45 isolates included in the study, a total of 25 genotypes could be found based on the
differential SNP patterns (Figure 4.1). Seven genotypes are represented by more than one isolate.
These seven genotypes represent 27 isolates, the majority of which can be found in two groups of
isolates with nine (BE4, BE7, BE9, BE11, BE12, CO2, CO3, GB2, PL1) and eight (BE3, BE5, BE6, BE8,
BE13, BE14, FR3, FR1) isolates, respectively (Figure 4.1). Other isolates with identical genotypes
91
Chapter 4
grouped per two: isolates CO1 and FR2, US4 and US5, MY4 and MY7, US1 and US3, and JP1 and JP2.
The remaining 18 isolates all had unique genotypes. Most of the clonal groups contain isolates that
were sampled in the same geographical region, but in different growing seasons. The large clonal
groups, representing clusters I and II, contain samples taken between 2003 and 2008. Isolates US1
and US3 were also sampled in different years (2008 and 2010).
The SNP profiles of 20 isolates represent potential recombinations of other profiles (Table 4.3). For
example BE1 shares the polymorphisms SNP 180 1-3 and SNP 006 with the isolates of cluster I. It
shares polymorphisms SSR 892, SNP 268, SNP 895 and SNP 292 with cluster II and also contains all
the alleles that are shared by these potential parental genotypes. Most recombinant genotypes can
be explained by several potential combinations of parental genotypes, with a maximum of 30
different combinations for the genotype of isolate GB1. However, when several combinations are
possible, one or a few related genotypes are usually repeatedly identified as one of the potential
parents, suggesting that these genotypes, or one closely related to it are likely to be involved in the
ancestry of the recombinant genotype. This is also illustrated in isolate GB1 for which parental
genotype BE10 is one of the parents in 15 of the 30 possible recombination options. Similarly, the
JP1/JP2 genotype is likely ancestral to the geographically related isolates JP3 and JP4 (Table 4.3). For
five genotypes, represented by seven isolates (CN1, MY3, US1/US3, US2 and JP1/JP2), no potential
ancestral genotype could be identified in the dataset even though they are important parental
genotype themselves.
92
Migration, survival and recombination of P. horiana
Table 4.3: Potential recombination events among the genotypes in our collection. For each genotype (column) the number
of combinations in which each other genotype in the dataset (rows) can be a potential parental genotype is listed at the
row by column intersection. A relatively high number in a column indicates an increased chance that the genotype in the
corresponding row is a possible ancestor of the recombinant genotype in the corresponding column. The largest value is
marked with a dark shading and the second largest value (other than 1) is marked with a lighter shading.
2
2
2
GB1
1
4
1
6
15 13 6
1
2
1
3
3 10 8
5
1
2
1
GB3
1
2
5
1
2 10 7
3
2
1
1
1
2
9
NL5
1
2
5
US4/US5
2
2
JP4
2
8
LK1
2
1
1
1
3
3
1
1
1
2
1
1
1
5
1
1
2
1
3
1
MY4/MY7
1
1
2
MY6
1
1
2
5
1
1
MY5
MY1
1
NL2
2
1
1
2
3
4
2
1
4
1
1
1
2
8
1
2
2
1
1
US1/US3
1
3
1
2
7
US2
1
3
5
1
2
JP1/JP2
# of potential
recombination 5
events
5
9 26
1
1
3
NL3
2
NL2
MY6
CN1
MY4/MY7
LK1
2
2
0
0
0
0
2
1
1
1
1
1
1
1
1
1
1
1
1
3
1
1
2
3
1
1
2
1
1
1
1
1
1
1
2
1
9
5
1
3
10
1
1
1
4
1
4
2
2
7
1
2
3
5
4
2
2
3
2
1
2
2
1
1
2
1
1
1
2
1
1
1
1
2
NL3
MY3
1
2
1
1
1
2
3
1
10 1
1
1
CN1
JP4
3
2
1
1
1
BE15
JP3
US4/US5
1
JP3
3
NL5
1
US2
2
2
1
JP1/JP2
NL4
1
1
1
US1/US3
4
3
5
Cl.
VIII
MY3
CO1/FR2
1
1
Cl.
VII
MY1
1
2
Cl.
VI
MY5
3
BE15
2
GB3
BE13 et al.
BE10
1
1
Cl.
V
GB1
7
NL4
2
BE1
Cl.
IV
BE10
CO1/FR2
BE11 et al.
Cl.
III
BE1
BE11 et al.
Potential
parental
genotypes
Cl.
II
BE13 et al.
Potential recombinant genotypes
Cl.
I
1
2
6
1
15 25 18 30 19 31 7
5
5
6
1
93
0
2
1
2
1
1
1
3
7 12 24 13 1
7
1
6
2
1
Chapter 4
4.3.3 Phylogeny and cluster analysis
The dendrogram showing the evolutionary relations of the isolates based on the Simple Matching
similarity coefficients and the UPGMA method is shown in Figure 4.1. Cophenetic correlation
coefficients of more than 85% were observed, indicating that the dendrogram is a good
representation of the similarity matrix.
Within this dendrogram, eight clusters containing one to seven genotypes can be discerned. Six of
these clusters correspond to the clonal groups of isolates described above. The two additional
clusters consist of three (cluster IV) or eight (cluster VI) isolates that share one to four rare alleles
respectively. Most clusters predominantly contain isolates from the same geographic origin. Clusters
I and II mainly contain European isolates. Isolates in cluster VI mostly originate from Malaysia and
isolates in clusters V and VII are from the USA. However, despite this general correlation between
geographic origin and phylogenetic clustering, some inconsistencies are also observed. Two
Colombian isolates group with the European isolates from cluster I and a Colombian and a French
isolate group in cluster III. Similarly, two isolates from the mixed culture from the Netherlands group
with the Malaysian isolates in cluster VI. Although the remaining isolates group together with
particular genotypes in the dendrogram, their genotype profiles show similarities with multiple
genotype clusters.
A cluster analysis based on the pathotype profiles of a subset of 26 isolates is shown in Figure 4.2.
The diversity in these phenotypic profiles is larger than in the genotypic profiles. None of the isolates
has the same pathotype and only two main pathotype clusters can be distinguished. The first
pathotype cluster contains 21 isolates, with all genotype clusters represented, except clusters V and
VI. In contrast, the second pathotype cluster, with the Malaysian isolate (MY1) and the Dutch isolates
(NL2 and NL3) seems to consist of the single genotype cluster VI. Also isolates NL1 and NL6, for which
no genotype profile could be determined, can be found in this second pathotype cluster. Marker SNP
793-1 is uniquely associated with the second pathotype cluster.
94
Migration, survival and recombination of P. horiana
Figure 4.1: An unweighted pair-group method with arithmetic averages (UPGMA) dendrogram of genetic relationships
among 45 isolates of P. horiana based on the Simple Matching similarity coefficients obtained using 32 SNPs and one SSR.
Per isolate, the allele for every SNP is presented as described in Table 4.2. Numbers within the tree indicate the cophenetic
correlation coefficients for the cluster nodes. SNPs are sorted right to left based on their presence in the genotypes as
sorted from top to bottom. Dutch samples have an increased probability of recombination as they were maintained in a
mixed culture at a biosecurity facility. This mixed culture, included infected material from custom interceptions.
95
Chapter 4
Figure 4.2: An unweighted pair-group method with arithmetic averages (UPGMA) dendrogram of relationships among 26
isolates of P. horiana based on the Simple Matching similarity coefficients obtained using the interaction phenotype profiles
on a set of 36 cultivars as described by De Backer et al. (2011). For each isolate the position in the genotypic dendrogram
(Figure 4.1) or non-clustered recombinant genotypes is listed next to the isolate code. Numbers within the tree indicate the
cophenetic correlation coefficients for the cluster nodes.
96
Migration, survival and recombination of P. horiana
4.4 Discussion
Determination of the genetic diversity of obligate pathogens is very difficult as the development of
molecular markers requires the availability of pure pathogen DNA, which is difficult to obtain in
sufficient quantities due to contaminating DNA from the host or from contaminating
microorganisms. For example, AFLP profiles obtained in this study using P. horiana pustule DNA
could not be resolved due to the bands resulting from variable amounts of contaminating plant DNA.
Even if a limited amount of pure DNA could be obtained for whole genome sequencing and SNP
mining, the cost of de novo sequencing of several isolates at the start of this study prevented such a
strategy. We partly resolved these contamination issues by obtaining pure DNA from a limited
number of isolates via a basidiospore harvesting technique that still showed some contamination by
other microorganisms but the sequencing and alignment of the contigs of resulting AFLP fragments
using the CRoPS™ technique (van Orsouw et al. 2007) largely eliminated non- P. horiana
contaminations in our analysis. Candidate SNPs were identified and developed into markers that
could subsequently be tested on non-pure DNA of a larger collection of isolates using P. horianaspecific PCR primers. In combination with a whole genome amplification step, the specific SNP
markers could be scored using only a few pustules even in the presence of contaminating plant DNA.
This allows direct screening of samples obtained during quarantine screening, without the need for a
labor intensive, time-consuming and costly propagation of the pathogen at an appropriate biosafety
facility. Although a whole genome amplification step can create an amplification bias (Pinard et al.
2006), we did not observe any mutations due to this step. The use of a whole genome amplification
also resulted in a good representation of the original genome in the rust pathogen Puccinia
striiformis f.sp. tritici (Wang et al. 2009).
Approximately 0.01% of the 2.35 x106 nucleotide positions that were analyzed during the
CRoPS™ analysis were polymorphic. This indicates that the SNP frequency of P. horiana is extremely
low compared to other fungi: rates of 1.20% in Candida albicans (Forche et al. 2004), 1.71% in
Tricholoma matsutake (Xu et al. 2007), 1.35% in Aspergillus fumigatus (Bain et al. 2007), and 0.25% in
Pandora neoaphidis (Fournier et al. 2010) have been described. As expected, there are strong
indications that the SNP markers are stable: the clonal European isolates were collected between
2003 and 2007 and except for marker SNP431/1, all markers were found in at least two isolates,
suggesting that they were not recently generated.
Isolates that appear to be clonal based on the obtained genotype profiles, were observed
within the geographic regions where the pathogen is considered endemic, as in Europe, Japan and
Malaysia. This is expected, as these clonal isolates represent relatively recent local dispersal. For
97
Chapter 4
example, the five Belgian isolates from the area of Mechelen are mostly clonal and all group together
in cluster II. The clonal isolates from clusters I and II were also found in France, Poland and the UK,
indicating that longer distance spread within Europe also occurred, either via commercial trade of
infected plants or less likely, via very long distance spore dispersal (Nagarajan and Singh 1990; Brown
and Hovmoller 2002). A perfect match of the two genotypes of the Colombian isolates with
genotypes co-occurring in Europe strongly suggests repeated international long-distance dispersal,
presumably via commercial trade of infected plants. In case of CO2 and CO3, which have an identical
genotype as the cluster I clone that is present in at least three European countries, the likely
direction of the migration is from Europe to Colombia. For CO1, whose genotype was also found in a
single French isolate, it is not possible to determine the direction of the migration. The Dutch isolates
NL2 and NL3 may also be examples of recent international migration: although their genotypes do
not perfectly match with any other genotype, their profiles cluster with isolates from Malaysia. These
Dutch isolates were possibly import interceptions (Van Rijswick, personal communication) and may
not have spread in the Netherlands.
Although clonal presence of isolates within most regions was clearly demonstrated, the genotypic
diversity in each of the regions was larger than expected. In Europe there were 12 genotypes among
25 isolates, in Malaysia there were five genotypes among the six isolates and in Japan there were
three genotypes among the four isolates. The marker profiles within a geographical region did not
follow a pattern of stepwise accumulation of mutations, typical for an organism with a presumed
asexual cycle. Particularly striking was that certain haplotypes appear to have recombinant patterns
of SNPs found in isolates that originate in different geographic regions. This is obvious for the
Japanese isolates which serve as a donor for markers present in isolates of clusters II, III and V (Table
4.3). These observations indicate recombination and migration of genotypes and suggest a
parasexual cycle is present, disproving the presumed asexual nature of this organism. The
recombination in P. horiana may be explained by heterokaryosis after anastomosis of germtubes or
vegetative mycelium resulting in heterozygous vegetative mycelium. After somatic meiotic division in
the telial stage, homozygous recombinant genotypes can be expected based on the microscopic
observations of somatic meiotic nuclear divisions in P. horiana (Kohno et al. 1974; Kohno et al. 1975)
(Figure 4.3) as we indeed observed in our results. Germtube anastomosis has recently been
described for uredinial germtubes in the rust species Puccinia triticina (Wang and McCallum 2009)
and Phakopsora pachyrhizi (Vittal et al. 2012). Anastomosis after basidial infection of a microcyclic
rust has been suggested by Ono (2002). Isolates BE1 and BE15 are most likely local recombinants of
the two main clonal genotypes in Belgium. In the other geographic regions, for which fewer isolates
were collected, there are also clear indications of recombination. In fact nearly all SNPs identified are
98
Migration, survival and recombination of P. horiana
found in various combinations which is in disagreement with simple identity by decent in strictly
clonal lineages. However, as in most areas the sampling is sparse and the dominant genotypes are
unknown, there is uncertainty as to the exact identity of the parents.
Figure 4.3: Schematic illustration of recombination in Puccinia horiana. (A) “Tip-to-tip” anastomosis as also described by
Wang and McCallum (2009) between vegetative mycelium or germ tubes from two different genotypes (illustrated with
black versus white nuclei) of P. horiana eventually results in heterozygous mycelium with two haploid nuclei. (B) During the
telial stage karyogamy occurs with a rearrangement of the chromosomes resulting in a recombinant genotype (illustrated
with gray nuclei). After the second somatic meiotic division in the promycelium, two identical haploid nuclei migrate in each
of the two basidiospores, which give rise to homozygous recombinant mycelium (Kohno et al. 1974; Kohno et al. 1975). The
nuclear status is indicated for each stage in the rectangular insets: c represents the number of chromosomes, n represents
the ploidy number.
The USA isolates are unique in that the pathogen is not considered endemic in this country (Shea
2007). Four of the five USA isolates group in two clusters and belong to three genotypes, which do
not relate directly as recombinants (Table 4.3). This indicates that at least three separate
introductions into the USA took place. Isolates US4 and US5 have the same genotype, and so do
isolates US1 and US3. As they were sampled in different states and in case of US1 and US3 also three
years apart, survival and spread of these isolates within the USA is possible, indicating P. horiana may
now be endemic in the USA. This is supported by preliminary observations of overwintering of P.
horiana outdoors on infected plants in Pennsylvania (Kim et al. 2011). Alternatively, repeated
introduction of isolates of the same genotype into the USA may have occurred more than once.
Collection of trace-back information on the source of imported plants as well as genotyping of more
USA isolates and isolates from the possible geographic origins of imported plant material will provide
further evidence for one of these hypotheses. All USA isolates clustered with isolates from oriental
origin, containing markers typical for Japanese and/or Malaysian isolates. Four out of five USA
isolates also share the rare allele SNP431/2, which probably resulted from a historic recombination
with the yet unidentified parental genotype that also donated this marker to the recombinants BE10,
99
Chapter 4
GB1 and GB3. Isolates with parental genotypes to isolates CN1, MY3 (cluster VI), US1 (cluster VII),
US2 and US3 (cluster VII) and JP1/2 (cluster VIII) were not identified (Table 4.3). We hypothesize that
these genotypes are potentially ancestral and donators of specific markers in recombination events.
The diversity based on the pathotype data was significantly higher than the genotypic diversity. This
larger diversity can be explained by the complex pattern resulting from the interaction between at
least seven resistance genes in Chrysanthemum with corresponding avirulence genes in the pathogen
(De Backer et al. 2011). Even within the clonal isolates clear differences in pathotypes were observed
(e.g., BE3 versus BE6). In contrast with the presumed neutral SNP and SSR markers, avirulence genes
experience independent selection pressure based on the presence of specific resistance genes in the
chrysanthemum cultivars used in a given region, which may explain the larger pathotype-based
diversity. Alternatively, a higher variation in pathotype interaction profiles could be explained by
epigenetic changes in disease resistance, which are more flexible than sequence mutations (Stokes et
al. 2002). A correlation between the clusters of the pathotype- and genotype-based dendrograms
was only observed for BE4 and GB2, JP1 and JP2 and especially for the isolates from genotype cluster
VI. As these isolates are the most virulent ones identified so far they probably contain few avirulence
genes, including one that corresponds to a frequently-used resistance gene (De Backer et al. 2011).
This resulted in a unique pathotype cluster. Except for the presumed quarantine interceptions in the
Netherlands, these isolates are confined to Malaysia. The number of recombination events with
exotic isolates in that region seems to have been limited, resulting in a clear genotypic cluster for
these isolates. There is currently no information on the location of the avirulence genes or their
physical proximity to each other within the genome of P. horiana. Avirulence genes in other rust
fungi have been identified in clusters as well as individually on different chromosomes (Leonard and
Szabo 2005; Bolton et al. 2008). Their actual level of proximity in P. horiana will determine the
chance for rearrangement of these avirulence genes during recombination, and therefore the
potential for the creation of new pathotypes. This is especially relevant if the parental isolates each
lack important avirulence genes.
Our results demonstrate that recombination occurs on a regular basis in a presumed asexual
organism like P. horiana, thus increasing the potential formation of genotypes with greater virulence.
This may have practical implications when inoculations with presumed asexual organisms are used as
a time saving strategy in resistance breeding programs. Since this may lead to the undesirable side
effect of generating a highly adapted pathogen in the process, such screenings should only be
performed in strict containment. Given the risk for recombination demonstrated in this work, coexistence of P. horiana isolates with different genotypes should be avoided in nurseries. For example,
potentially infected planting material from different locations should not be propagated in the same
100
Migration, survival and recombination of P. horiana
greenhouse. Similarly, outdoor epidemics should be controlled in regions where different pathotypes
are present.
The SNP and SSR markers we developed allow for the rapid genotypic characterization of isolates of
P. horiana. In this study we determined the nature of the SNPs and SSR using sequencing. After the
development of faster SNP detection methods, such as real-time PCR with High Resolution Melt
analysis (Luchi et al. 2011) or SNP detection using Luminex technology, these markers will become
fast and relatively inexpensive tools to screen many samples. Such tools would be suitable to trackand-trace isolates. These methods could also be used to identify possible recombinations and could
be exploited to monitor populations at a global level addressing questions relating to migration and
survival of the pathogen. Tracking-and tracing can also help to guide and evaluate effectiveness of
quarantine measures. Due to the strict quarantine measures applied today, shipments are destroyed
when infected plants are intercepted during customs inspections. Genotyping of the isolate at hand
would determine whether it belongs to a given genotype that is already present in the region of
import or not. This of course implies that the endemic population has been characterized, which will
provide extra information as to the relative frequencies of the genotypes present. Such genotyping
would help evaluate the possible risk of impact of the specific introduction. A first step towards of
the identification of high-risk genotypes may be the unique marker (SNP793-1) that was identified in
this study. However, novel loss of avirulence may also occur in resident genotypes or in exotic
genotypes that do not contain the specific marker. Such events would not be detected with
association markers. Until the sequence of the specific avirulence genes of P. horiana have been
determined and developed into markers, associated markers such as SNP793-1 are our best option
for fast detection of isolates with a potentially higher pathogenic potential.
101
Chapter 5:
Molecular detection
of Puccinia horiana in air samples
Top view of Burkard spore trap with rotating sample drum
removed and displayed
Detection of P. horiana
5.1 Introduction
Puccinia horiana Henn. (Hennings 1901) is the causal agent of chrysanthemum white rust or
Japanese rust. This microcyclic rust is one of the most important fungal pathogens of
Chrysanthemum x morifolium cultivars and has been reported in most countries where
chrysanthemums are grown as cut flowers or potted plants (Punithalingam 1968b). Infections often
appear suddenly and on a large scale, leading to complete loss of the crop. Due to the high
economical damage caused by these infections, the pathogen is recognized as a quarantine organism
in several continents including Europe, North and South America, Africa and Asia (EPPO 2004).
Under optimal conditions the fungus has a life cycle of less than three weeks (Zandvoort et al.
1968a). The first symptoms appear around 8 to 10 days post infection as chlorotic flecks on the upper
side of the leaves at the sites of infection. At the underside of the infected leaves, pustules consisting
of teliospores develop around 18 to 21 days post infection (Firman and Martin 1968). These
teliospores germinate in conditions of high relative humidity and cool temperatures around 17°C
with the formation of basidiospores. The basidiospores are dispersed by wind over distances of at
least 700 meters (Zandvoort 1968) and can infect other chrysanthemums when a water film is
present on the leaves (Firman and Martin 1968). Although the life cycle of the pathogen is known,
there is a lack of knowledge about the origin of initial infections and about the possible latent
survival of the pathogen. Although the pathogen can overwinter in infected plant tissue,
chrysanthemums are produced as annual crops: growers remove all plant material during winter and
start from new cuttings the next spring, purchased from specialized companies. Development and
spread of the disease is nowadays mainly controlled by cultural measures and calendar-based
preventive spraying with fungicides, although breeding for resistance to P. horiana is gaining
importance due to its more sustainable character. The application of cultural measurements that
focus on a decrease of the relative humidity and the period of leaf wetness is limited to greenhouses
and is becoming more expensive due to increasing cost of the fuel needed for heating associated to
such measures. The control by preventive fungicides will decrease due to the decreasing number of
registered fungicides in the EU and the increase in the number of strains with resistance to specific
fungicides (Whipps 1993; Cook 2001)(see Chapter 3). An effective method to monitor the presence
of airborne basidiospores should be a powerful tool to study the epidemiology of the pathogen and
to guide the timing of fungicide applications, reducing the number of such applications to the periods
when they are strictly needed. In combination with molecular characterization of the detected
isolates, sampling of airborne inoculum could also be used to monitor the pathotypes (De Backer et
al. 2011) present in the local population, once the sequences of the avirulence genes have been
105
Chapter 5
identified. This could help growers and breeders to use the most suitable cultivars in a specific
region.
Several systems for the collection of airborne particles are currently available as reviewed by
Martinez et al. (2004) and can be based on the impact and adhesion of spores on culture media or
tapes or can be based on filtration. The samplers that are most often used for sampling of fungal
spores are Hirst-type samplers (Hirst 1952; Hirst et al. 1967), rotating arm samplers (Perkins 1957; DiGiovanni 1998) and cyclone traps (Williams et al. 2001). Recently, a new type of spore trap, the Ionic
Spore trap, was introduced. It uses an electrostatic force to capture airborne particles on an adhesive
tape (Schneider et al. 2009). Analysis of the samples is traditionally done by culturing the organisms
of interest on selective media (Crook and Sherwood-Higham 1997) or by counting the number of
spores using light or scanning electron microscopy (Eduard et al. 1990). However, these methods are
time consuming, require experienced staff able to recognize the organisms and can be unreliable.
Alternative methods for pathogen detection can be based on immunological detection methods
including ELISA (Enzyme Linked Immunosorbant Assay) (McCartney et al. 1997; Ward et al. 2004), a
microtiter immunospore trapping device (MTIST) (Kennedy et al. 2000) or molecular detection using
PCR or real time PCR (McCartney et al 2004, West et al 2008). Molecular techniques are gaining
importance in the detection of airborne fungal spores (McCartney et al. 2003; West et al. 2008).
Several sensitive PCR assays have been developed for the detection of fungi in air samples. DNA from
as low as 10 Penicillium roqueforti spores could be detected by PCR and nested PCR assays in air
samples taken with cyclone air samplers (Williams et al. 2001), Hirst-type and rotating-arm samplers
(Calderon et al. 2002b). Using Hirst-type samplers, PCR assays with similar sensitivities are also
described for Leptosphaeria maculans and Pyrenopeziza brassicae (Calderon et al. 2002a) and for
Sclerotinia sclerotiorum (Freeman et al. 2002). Recently real time PCR assays using SYBR® Green
technology were developed to quantitatively detect spores of Monilinia fructicola (Luo et al. 2007)
and S. sclerotiorum (Rogers et al. 2009) in Hirst-type air samples with detection limits of 2 and 1.4
spores per PCR reaction, respectively. A real time PCR assay for the detection of Phakopsora
pachyrhizi in air samples taken by the Ionic Spore trap could detect up to a single spore (Schneider et
al. 2009). These results demonstrate the potential of PCR, and especially real-time PCR, for a very
sensitive and quantitative detection of airborne fungal spores.
Considering P. horiana is an obligate parasite, culturing cannot be used for the monitoring of the
pathogen. Identification by microscopy is very difficult as well due to the fact that basidiospores are
difficult to distinguish from other particles present in the air. For several rust species specific
antibodies are available for immunological identification (Harmon et al. 2007; Skottrup et al. 2007),
but not for P. horiana. However, specific primers for the molecular detection of P. horiana in infected
106
Detection of P. horiana
leaf tissue by PCR and real time PCR have been developed (Pedley 2009; Alaei et al. 2009b). The real
time PCR assay described by Alaei et al. (2009b) is based on SYBR Green Technology and uses highly
specific primers for amplification of selected regions in the rDNA internal transcribed spacers (ITS).
Using this assay, 5 fg of genomic DNA, 5 target copies of cloned target DNA or 8 spores could be
detected. A TaqMan® probe for the detection of the same amplification fragment was developed as
well, although it was less sensitive than the SYBR green assay (Alaei et al. 2009b). Pedley (2009) also
described a real-time PCR based on TaqMan® technology using ITS-primers, but the limit of detection
was higher than the one reported by the assay described by Alaei et al. (2009b).
The sensitive SYBR Green real time PCR assay described by Alaei et al. (2009b), though only described
for application in plant samples, offers perspectives for the detection of the pathogen in air samples
as well. The main objective of this study was to use this assay for the detection of P. horiana
basidiospores in air samples taken with a Hirst-type spore sampler (Burkard spore trap, Burkard
Manufacturing Co. Ltd., UK) and an Ionic Spore trap (DS Scientific, LA, USA). Considering the airborne
particles are captured on different matrices in these two types of spore samplers, DNA extraction
protocols for both types of matrices were developed and optimized and the sensitivity of the
detection was determined. The optimized DNA extraction protocols were used to study the presence
of P. horiana in the field in three chrysanthemum growing seasons, and the data were correlated
with the amount of rainfall. The insights we obtained are discussed in respect to the epidemiology
and biology of the pathogen.
5.2 Material and methods
5.2.1 Basidiospore collection and spore sampling
For the optimization of the DNA extraction protocols and to determine the sensitivity and
repeatability of the spore detection (see below), pure basidiospores of P. horiana were collected to
spike the samples. Basidiospores were collected using a modified set up described by Alaei et al.
(2009a). Heavily infected chrysanthemum leaves with recently developed pustules were stuck to the
lid of 10 cm Petri dishes with the telia pointing downwards using 1% water agar as adhesive (Figure
5.1. The leaves were misted with deionized water using a Preval sprayer (Chicago Aerosol, IL, USA),
while the bottom part of the Petri dish was kept dry. Leaves were incubated for 14 hours at 17°C in
the dark to induce sporulation. Per Petri dish, the resulting spore prints (Figure 5.1) were
microscopically checked for absence of spores of contaminating fungi and suspended in
approximately 1 ml of 0.1% Igepal CA-630 (Sigma-Aldrich, MO, USA) by scraping of the spores with a
107
Chapter 5
micropipette tip. The basidiospore suspension was transferred into a 2 ml microcentrifuge tube and
vortexed to bring individual spores into suspension. The concentration of the spore suspension was
determined microscopically using a haemocytometer and adjusted to 10 5 spores ml-1 from which tenfold serial dilutions were made to spike the tapes.
Figure 5.1: Collection of pure basidiospores of P. horiana. Infected leaves were stuck to the lid of a Petri dish with the telia
pointing downwards as illustrated on the right (= inverted lid). After the leaves were misted with distilled water, the lid was
replaced onto the bottom part and the closed Petri dish was incubated over night. After incubation, basidiospore prints are
visible on the bottom of the Petri dish as illustrated on the left.
Spore sampling in the field was performed using two types of spore samplers: a Burkard 7 day
recording volumetric spore trap (Burkard Manufacturing Co. Ltd., UK) and an Ionic Spore Trap (DS
Scientific, LA, USA)(Figure 5.2). In the Burkard spore trap, spores are captured on a (5:1) petroleum
jelly : paraffin wax-coated Melinex tape (Burkard Manufacturing Co. Ltd., UK) that is attached to a 7
day rotating drum, allowing recording of spore concentrations over a 7 day period. The trap
processed 10 liters of air per min and the orifice was positioned approximately 50 cm above ground
level. The Melinex tape was replaced every seven days and cut into sections of 24 mm representing
an exposure time of 12 hours. Sections of 12 to 48 mm, representing exposure times of 6 to 24
hours, respectively, were used in some experiments, as specified further. The sections were put in 2
ml microcentrifuge tubes and stored at -20°C until DNA extraction. In the Ionic Spore Trap, spores
were captured on 3M® Model 666 double-sided tape (3M, MN, USA) pieces cut to size and mounted
on round stubs with a diameter of one inch (2.54 cm) and the tape pieces were replaced every 12 to
48 hours. The ionic spore trap processes up to 600 liters of air per min with the air intake situated at
108
Detection of P. horiana
approximately 150 cm above ground level. After sampling, tape pieces were placed in small Petri
dishes (5.5 cm) with the capture side up and stored at -20°C until DNA extraction.
Figure 5.2: Illustration of the Burkard spore trap (A) and the Ionic spore trap (B, credit: (Schneider et al 2009)) we used in
the field experiments.
5.2.2 DNA extraction
Two DNA extraction procedures, each time with or without pretreatments as listed in Table 5.2, were
evaluated. Criteria were the reliability and sensitivity of the SYBR® Green real time PCR-mediated
detection of P. horiana. The limit of quantification (LOQ) and limit of detection (LOD) were
determined for each procedure. Eventually, the best DNA extraction was obtained with the DNeasy
Plant Mini Kit (Qiagen; Germany), which could be used for the tapes from the Burkard spore trap as
well as the Ionic spore trap. The following pretreatment step was necessary. Before extraction, the
tapes were placed in 2 mL microcentrifuge tubes to which 200 ђg of 0.5 mm zirconium beads
(BioSpec Products, OK, USA) and 400 ђL of lysis buffer (Buffer AP1) were added. The tubes were
placed in a Retch MM200 ball mill (Retsch, Germany) and ground for 30 seconds at 30 beats per
second (bps), after which the tubes were inverted and the grinding was repeated. Subsequently the
samples were incubated at 65°C for 10 min followed by the grinding step described above. The
heating and grinding steps were repeated two more times. Buffer AP2 (130 ђL) was added and after
5 min incubation on ice the mixture was centrifuged at 20,000 g for 5 min. The lysate without the
109
Chapter 5
zirconium beads was pipetted on the supplied shredder columns and centrifuged for 2 min at 20,000
g. Further steps in the DNA extraction protocol were performed as described by the manufacturer.
DNA was eluted in 50 ђl elution buffer (Buffer AE) and stored at -20°C until further use.
5.2.3 Real time PCR assay
Quantitative detection of P. horiana DNA was done by real time PCR based on SYBR® Green
technology, using the primers Ph 263F and Ph 264R as described by Alaei et al. (2009b). Reactions
were performed in 25 ђl containing 5 ђl of DNA extract, 0.3 ђM of each primer and SensiMix SYBR®
master mix (Bioline, UK). Reactions were carried out in optical Framestar 96 plates (4titude, Surrey,
UK) capped with flat optical caps using an ABI Prism 7900 HT (Applied Biosystems, CA, USA). The
thermocycling profile for all PCR reactions was 10 min at 95 °C and 40 cycles consisting of 15 s
denaturation at 95 °C and 60 s annealing and extension at 60 °C. quantification was done by including
a standard curve per real-time PCR run containing a 10-fold dilution series from 10-2 ng cloned
plasmid DNA (pDNA) (equivalent to 2,151,090 copies) to 10-7 ng cloned pDNA (equivalent to 22
copies). Per real-time PCR run we also included a no template control in which the DNA was
substituted by MilliQ water. For every reaction we constructed and analyzed the amplification curves
using the manufacturer’s Sequence Detection Software (SDS, Applied Biosystems). The threshold
cycle number (Ct) was calculated using the default settings. The specificity of the real-time PCR
reactions was checked based on the dissociation curves of the amplicons.
5.2.4 Sensitivity and repeatability of detection with the Burkard and Ionic spore trap tapes
and the optimized DNA extraction
The sensitivity of detection of P. horiana basidiospores using the Burkard spore trap tape and the
Ionic spore trap tape in combination with the DNA extraction and real-time PCR was tested by
analyzing tape pieces that were spiked with a tenfold dilution series of 10 5 basidiospores down to a
single basidiospore by pipetting spore suspensions evenly across pieces of Melinex tape and pieces of
3M® Model 666 double-sided tape. Samples spiked with MilliQ water were taken throughout the
procedure of DNA extraction and detection by real-time PCR and served as a negative control. Tapes
were dried to allow the spores to adhere to the tapes and used for DNA extraction and real-time PCR.
The tapes were cut to size in a way which was identical to the tapes processed after spore sampling
with the spore traps (see above). For each spore concentration and each type of tape, five
independent DNA extractions and real-time PCR reaction series were performed to assess the
accuracy and repeatability of the detection. The standard curves that were obtained for both the
Burkard spore trap and Ionic spore trap, were compared to the standard curves obtained with nine
110
Detection of P. horiana
series of 10-2 ng cloned pDNA (2,151,090 copies ) to 10 -7 ng cloned pDNA (22 copies), as described
above.
5.2.5 Field experiments
A total of 5 field trials were performed during this study as summarized in Table 5.1. The objective of
trial 1 was to determine the number of spores we can detect over time in a chrysanthemum field in
which symptoms of the pathogen were reported. The objectives of trial 2 were to determine whether
basidiospores could be detected in a chrysanthemum field in which no symptoms of the pathogen
were reported and if so, determine the possible association between basidiospore detection and
rainfall and relative humidity. This association was also the main objective of field trial 3, which was
conducted in the presence of a controlled inoculum source. During the 2011 growing season, two
more trials were conducted in the presence of a controlled inoculum source (trial 4 and trial 5), but
this time included both the Burkard spore trap and the Ionic spore trap. The objectives of trial 4 were
to study the release of basidiospores in relation to the period of the day and to compare the relative
capture efficiency of the Burkard and Ionic spore samplers. The objective of trial 5 was to determine
the relative detection levels at varying distances from the inoculum source.
111
Table 5.1: Overview of the field trials performed during this study. For each trial the relevant information is listed, including the sampling period (start date and end date), the type of field
trial, the presence or absence of disease symptoms during the trial, the types of the spore trap used, the sampling intervals, and the environmental parameters that were monitored during
the trial.
Trial 1
Trial 2
Trial 3
Trial 4
Trial 5
Year
2009
2010
2010
2011
2011
Start date
September 3
September 17
October 20
September 8
September 16
End date
Type of field trial
October 27
Commercial chrysanthemum
field
Yes:
x limited symptoms
reported
October 19
Semi-commercial
chrysanthemum field
No symptoms observed
November 25
Controlled field trial
September 16
Controlled field trial
October 31
Controlled field trial
Yes:
x 15 very symptomatic
plants placed in a circle
(radius 1 m) around the
spore trap
Yes:
x 15 very symptomatic
plants placed in a circle
(radius 1 m) around the
spore traps
Variable per day:
a
x distance cycles :15 very
symptomatic plants used
b
as a point source at 1 m , 5
c
m or 20 m upwind from
spore traps, followed by 2
days in a nearby closed
greenhouse
x Symptomatic plants first
completely removed and
b
then replaced
-1
Yes (10 liters min )
-1
Yes (200 liters min = 35% of
maximum power)
e
18 hours (Burkard spore
trap): 6 pm - 2 pm
24 hours (Ionic spore trap): 6
pm - 6 pm
Presence of plants with
disease symptoms
-1
-1
-1
Burkard spore trap
Yes (10 liters min )
Yes (10 liters min )
Yes (10 liters min )
Ionic spore trap
No
No
No
Sampling intervals
12 hours
x Day: 10 am - 10 pm
x Night: 10 pm - 10 am
12 hours
x Day: 11 am - 11 pm
x Night: 11 pm - 11 am
24 hours: 10 am - 10 am
-1
Yes (10 liters min )
-1
Yes (200 liters min = 35% of
maximum power)
3 hours: 9 am - noon
6 hours: noon - 6 pm and
6 pm - midnight
9 hours: midnight - 9 am
Environmental parameters
Rainfall, Relative humidity
Rainfall
Rainfall
Rainfall
monitored
a
The distance between the Burkard spore trap and the plants was varied on a daily basis during 6 consecutive cycles of 5 days, after which a single period of complete plant removal and a
period of plant replacement took place.
b
The Ionic spore trap was placed 1 meter downwind from the plants at each time point.
c
The position of the plants and the spore traps was in line with the prevailing wind and this position was adjusted on a daily basis if need be.
d
After the 6 consecutive cycles, the plants were removed from the sampling field and stored in a closed growth chamber for a period of nine days after which they were replaced for a
sampling period of seven days.
e
To avoid cross contamination from spores from the preceding sampling distance, the last 8 mm from every sample, representing 4 hours, was removed.
Detection of P. horiana
5.3 Results
5.3.1 Optimization of DNA extraction
The limit of quantification (LOQ) and limit of detection (LOD) for the different DNA extraction
procedures are listed in Table 5.2. As removal of basidiospores from the surface of the tape was not
reliable by only using lysis buffer supplied with the two DNA extraction kits, pretreatments based on
the protocol described by Calderon et al. (2002) for the extraction of spores on Melinex tape were
included. In this protocol, the spores are disrupted from the tape by bead beating in 200 ђl Igepal CA630 (formerly called Nonidet P-40) (Sigma-Aldrich; MO; US) (=pre-treatment A), followed by DNA
extraction of the suspension. This pretreatment in combination with the DNeasy Plant Mini kit
resulted in a LOQ of 10 spores and a LOD of as low as a single spore for extraction on Melinex tape.
For 3M666 tape a tenfold lower extraction efficiency was obtained using pre-treatment A. Also, in
combination with the QuickPick kit (Bio-Nobile, Finland) the recovery efficiency was lower.
Adaptation of the pre-treatment A method by using 400 ђl of DNeasy lysis buffer instead of 200 ђl
Igepal CA-630 in the bead beating step (=pre-treatment B) further increased the extraction efficiency
for the 3M666 tape to the same level as with the Melinex tape. This procedure (pre-treatment B +
DNeasy Plant Mini Kit) was used in further sensitivity and repeatability testing for both types of tape.
Table 5.2: Limits of quantification (LOQ) and limits of detection (LOD) (both in spore equivalents) observed using different
DNA extraction procedures for the detection of Puccinia horiana basidiospores on Melinex tape (used in the Burkard spore
sampler) and 3M666 double-sided tape (used in the Ionic spore sampler).
Melinex tape with a petroleum jelly :
paraffin wax (5:1) coating
3M666 double sided tape
LOQ
LOD
LOQ
LOD
ND
ND
100
10
Treatment A* + QuickPick
100
10
100
10
DNeasy Plant Mini
Not quantitative
10
1000
100
Treatment A* + DNeasy Plant Mini
10
1
100
10
Treatment B* + DNeasy Plant Mini
10
1
10
1
QuickPick
*Treatment A: Bead beating of the sample in Igepal CA-630 as adapted from Calderon et al. (2002a). Treatment B: Bead
beating of the sample in Lysis buffer supplied with the DNeasy Plant Mini kit. ND: not done.
5.3.2 Sensitivity and repeatability
With both types of tapes used in the different spore traps, a very sensitive detection down to a single
P. horiana basidiospore could be obtained. However, occasionally no P. horiana was detected when
using the lowest concentration. Reverse standard plots for both types of spore traps were calculated
based on the five replicate dilution series of spores that were spiked on the types of tape used in
respective spore traps (Figure 5.3). The reverse standard plot for pDNA was based on nine replicate
dilution series. For the separate real-time PCR assays that were performed with genomic DNA that
113
Chapter 5
was derived from dilution series of spores on Melinex tape and for the dilution series of pDNA, highly
similar standard curves could be obtained. For those samples, the coefficient of variance was smaller
than 5% at all template levels and highly linear relationships (R²>0.97) between the Ct values and the
log of the spore equivalents or DNA concentration of pDNA were observed in all replicates. The
standard curves we obtained with the gDNA derived from the dilution series of spores spiked on the
3M666 tape showed a coefficient of variance of 6.5% and also the linear relationship was less strong,
with a minimum R² of 0.79.
Figure 5.3: Reverse standard plots from real-time PCR using the primers described by Alaei et al 2009. (A) Reverse standard
plot showing the average Ct of nine tenfold dilution series of pDNA (0.01 ng of pDNA or 2,151,090 copies to 0.1 fg of pDNA
or 22 copies). The regression equation is. (B) Reverse standard plot showing the average Ct using extracted DNA from five
5
tenfold dilution series of basidiospores (1 to 10 spores) spiked onto Melinex tape as used in the Burkard spore trap. The
regression equation is. (C) Reverse standard plot showing the average Ct using extracted DNA from five tenfold dilution
5
series of basidiospores (1 to 10 spores) spiked onto 3M666 tape as used in the ionic spore trap. The regression equation is.
Error bars in each plot represent the standard deviation.
In real-time PCR assays with unknown samples, the Ct-value of each unknown sample can be
converted to pDNA equivalents using a pDNA serial dilution as a standard. Based on these pDNA
equivalents, the number of spore equivalents in samples from the Burkard spore trap can be
calculated using the equation “y = 0.9456x - 0.0003” in which “x” represents the log of the plasmid
copy number corresponding to the Ct of the unknown sample and “y” represents the log of the
number of spore equivalents. In a similar way the number of spore equivalents captured in a sample
by the Ionic spore trap can be calculated using the equation “y = 1.0427x - 0.559”.
5.3.3 Detection of P. horiana basidiospores in the field
During the 2009 and 2010 growing seasons, three field trials were performed using the Burkard spore
trap including two field trials in (semi-)commercial chrysanthemum fields. The results of the first trial
are presented in Figure 5.4. Given that symptoms were observed before this trial, detection of
basidiospores during the trial was expected. Two main events of sporulation were clearly observed. A
first event on nights 4 and 5 showed a sporulation of 21,567 spore equivalents and 12,105 spore
114
Detection of P. horiana
equivalents per 12 hours, respectively, corresponding to an average of 2995 and 1681 spore
equivalents per m³ of sampled air. During a second main event on day 33 and night 33, sporulation of
6627 and 4338 spore equivalents per 12 hours of sampling were recorded corresponding to an
average of 920 and 602 spore equivalents per m³. As illustrated in the insert of Figure 5.4,
basidiospores were continuously present in between those main sporulation events, usually at a level
of 5 to 500 spore equivalents per 12 hours. This experiment also showed a clear day-night cycle in
sporulation with a decrease during the day and an increase during the night.
During trial 2 (Figure 5.5), sporulation was detected despite the absence of symptoms in the field
either before or during the trial. However, the level of sporulation was much lower than in trial 1,
with a maximum of 223 spores in 12 hours or 31 spores per m³ of sampled air. Again, sporulation was
mainly observed during the night and in general appears during or shortly after a rain event. As
illustrated in Figure 5.5, sporulation is more abundant in the period from day 7 until day 21, which
contained several rain events, than in the dryer period from day 22 to day 27, when almost no
sporulation was recorded. The rain events on days 28 and 29 again induce sporulation.
During the trial 3, in which the Burkard spore trap was placed in the middle of a ring of infected
chrysanthemum plants, much higher sporulation rates were observed (Figure 5.6). Up to 5 main
sporulation events were recorded during this period. Specifically, maximum values were reached on
days 1, 5, 18, 22 and 24 with 23,120, 22,991, 26,152, 14,136 and 35,498 spores per 24 hours
corresponding to an average of 1606, 1597, 1816, 982 and 2465 spore per m³ sampled air,
respectively. Again, there was a striking correspondence between the days with rainfall and the days
with maximal sporulation. There were a couple of exceptions. The sporulation preceded the rain on
days 16 and 17. On the other hand sporulation decreased after day 24, even though rainfall persisted
up to day 26. This last event could be due to a decrease in the number of teliospores which could still
release basidiospores after the massive release of basidiospores on day 24.
115
Figure 5.4: Puccinia horiana spore equivalents detected using the Burkard spore trap (Y-axis) during a 54 days field trial from September 3 (day 1) until October 27, 2009 (trial 1). Samples
were taken in 12 hour intervals corresponding to day (D) and night (N) (X-axis). The insert shows the spore equivalent data on a logarithmic scale for the samples taken from day 3 to day 34
with the nights indicated in a shaded background.
Figure 5.5: Puccinia horiana spore equivalents (left Y-axis) detected using the Burkard spore trap during a 33 days field trial from September 17 to October 19, 2010 (trial 2). Samples were
taken in 12 hour intervals corresponding to day (D) and night (N) (X-axis). Relative humidity (left Y-axis) and rainfall (right Y-axis) data for the respective days are included.
Figure 5.6: Puccinia horiana spore equivalents detected using the Burkard spore trap (left Y-axis) and associated rainfall data (right Y-axis) during a 37 days field trial from October 20 to
November 25, 2010 (trial 3) . Samples were taken in 24 hour intervals (X-axis). The experiment was interrupted for 24 hours on day 8.
Detection of P. horiana
Trials 4 and 5 included two types of spore traps. In trial 4, sporulation during different periods of the
day was studied in more detail. A day-night cycle can be observed with the highest spore
concentrations during the night and lower concentrations during the day (Figure 5.7). This is in
accordance with the previous observations we made in trials 1 and 2. In four out of seven cases
where it could be compared, the highest spore release occurred during the second part of the night,
while in three out of seven cases it happened during the first part of the night. Higher numbers of
spore equivalents were observed with the Burkard spore sampler.
Figure 5.7: Puccinia horiana spore equivalents detected per hour using the Burkard and ionic spore samplers (Y-axis; log
scale) during different periods of the day (X-axis) over a period of 8 consecutive days (trial 4). The nights are indicated with
a shaded background. Sampling with the Ionic spore trap was interrupted for 12 hours during day 4 and for six hours during
day 5.
119
Chapter 5
The results of the first part of trial 5, in which the distance between the Burkard spore trap and the
plants was varied on a daily basis, are presented in Table 5.3. At a distance of 5 m from the infected
plants, we detected a relative amount of 15.4% of the spores detected at 1 m and this amount of
spores further decreased to 1.4% when the distance between the infected plants and the spore traps
was increased to 20 m. When the plants were placed in a nearby greenhouse for two days, there was
still detection of P. horiana DNA, to an extent of 0.35% after one day and 0.32% after two days when
compared to the average numbers at 1 m. Placing the plants in a closed growth chamber for a period
of 8 consecutive days resulted in a further gradual reduction in the number of observed spore
equivalents (Figure 5.8) down to a complete absence after 8 days removal. Replacement of the
infected plants at 1 meter from the spore trap restored the high spore concentrations measured
before the plants were removed. In this trial we again observed a lower level of detection with the
Ionic spore trap than with the Burkard spore trap.
Figure 5.8: Detection of Puccinia horiana spore equivalents (on Y-axis / log scale) after removal of infected plants from the
field for eight days, followed by replacing the plants at a distance of one meter of the spore traps (x-axis) for another 7 days
(trial 5).
120
Detection of P. horiana
Table 5.3: Number of spore equivalents detected in the first part of trial 5, in which the distance between the inoculum
source and the Burkard spore trap (BST) was varied on a daily basis. When infected plants were placed at 1, 5 or 20 m of the
inoculum source (the part above the dotted line), the numbers of spore equivalents observed with the Ionic spore trap (IST)
at 1 m (100%) were used to determine the relative numbers observed with the BST at increasing distances. This included a
correction for the difference in capture efficiency between the IST and the BST (=42.14, the average difference in spore
equivalents trapped with the two spore traps when both placed at 1 m). The average proportion captured is mentioned in
bold for each distance. When infected plants were absent (the shaded part of the table), the “proportion captured”
represents the average of the number of spore equivalents observed with the BST as a proportion of the average number of
spore equivalents observed with the BST at 1 m.
Date
Distance to BST (m)
Spore-equivalents
BST
Spore-equivalents
IST (1m)
Standardized
spore
equivalents IST
15/09/2011
1
411969
108183
7015273
5.87%
20/09/2011
1
4377
20
1265
345.92%
25/09/2011
1
283703
26501
1718501
16.51%
30/09/2011
1
51348
5676
368042
13.95%
5/10/2011
1
830
171
11077
7.49%
10/10/2011
1
13
144
9368
Proportion capture
(1 m =100%)
0.14%
100.00% ± 211.99%
16/09/2011
5
19280
16673
1081193
1.78%
21/09/2011
5
18622
2945
190953
9.75%
26/09/2011
5
55223
43835
2842547
1.94%
1/10/2011
5
4000
2294
148742
2.69%
6/10/2011
5
1063
39
2554
41.62%
11/10/2011
5
3170
2042
132392
2.39%
15.44% ± 24.27%
17/09/2011
20
1565
16006
1037905
22/09/2011
20
758
4691
304228
0.15%
0.25%
27/09/2011
20
1434
452
29329
4.89%
2/10/2011
20
1254
22618
1466687
0.09%
7/10/2011
20
1633
54464
3531783
0.05%
12/10/2011
20
2780
30533
1979930
0.14%
18/09/2011
greenhouse (1 day)
351
15
23/09/2011
greenhouse (1 day)
235
74
28/09/2011
greenhouse (1 day)
369
117
3/10/2011
greenhouse (1 day)
366
ND
8/10/2011
greenhouse (1 day)
93
54
13/10/2011
greenhouse (1 day)
1240
71
19/09/2011
greenhouse (2 days)
74
35
24/09/2011
greenhouse (2 days)
149
101
29/09/2011
greenhouse (2 days)
1982
752
4/10/2011
greenhouse (2 days)
86
32
9/10/2011
greenhouse (2 days)
62
21
14/10/2011
greenhouse (2 days)
83
0
1.43% ± 2.99%
0.35% ± 0.26%
0.32% ± 0.62%
121
Chapter 5
5.4 Discussion
In this study we developed a protocol for the detection of P. horiana basidiospores in air and showed
that it is sufficiently sensitive for the monitoring of this quarantine pathogen in the field. To obtain a
reliable detection we first had to optimize the DNA extraction procedure to be able to extract the
spores from the two types of tape used in the spore traps involved in this study. The most important
factor in the DNA extraction protocol was the preparation of the samples and particularly the bead
beating step. This preparation step was based on the procedure described by Williams et al. (2001),
who used Nonidet P-40 (now called Igepal CA-630) and bead beating to increase DNA extraction
efficiency when disrupting spores collected using a cyclone trap. Calderon et al. (2002a) and Calderon
et al. (2002b) further adapted this protocol for the extraction of fungal spores collected with a
Burkard spore trap. This procedure gave good results with the Melinex tape, while the LOD and the
LOQ were tenfold higher with the 3M666 tape. This can be explained by the fact that the adhesive on
the Melinex tape (and not the 3M666 tape) was completely washed off the tape bringing all particles
present on the tape in suspension. As this was not the case for the 3M666 tape, a substantial loss of
material was observed when the sample was transferred in the lysis buffer for further extraction.
Replacement of the Igepal CA-630 with the lysis buffer of the DNeasy plant mini kit further simplified
the procedure and gave good detection with both types of tape. The LOQ and LOD obtained with the
QuickPick kit did not meet the same sensitivity obtained with the DNeasy Plant Mini kit, probably due
to the lower quantities of lysis buffer used in this protocol. Using the primers described by Alaei et al.
(2009b) as low as 10 spores could be quantified in a reliable way and as low as a single spore could
be detected after spiking either type of tape. Nevertheless, no detection was sometimes observed
using tapes spiked with a single spore, probably due to the low concentration of spores in the spore
suspension we used for spiking these samples. The sensitivity we obtained is in accordance with
other reports of detection of fungal spores using the Burkard spore trap in combination with realtime PCR (Luo et al. 2007; Rogers et al. 2009). Compared to reports using regular or nested PCR for
the detection of spores on Melinex tape, a 400 times lower concentration of fungal spores present in
similar air samples was detected (Freeman et al. 2002; Calderon et al. 2002a; Calderon et al. 2002b).
A single fungal spore could also be detected when using the DNA extraction and real-time PCR on the
3M666 tape used in the ionic spore trap, similar to the sensitivity described by Schneider et al.
(2009).
Despite the use of a different master mix (Sensimix SYBR, Bioline vs. SYBR Green I PCR Master Mix,
Applied Biosystems), the standard curve using pDNA had the same slope (-3.32 ± 0.12) as the slope (3.30 ± 0.14) obtained by Alaei et al. (2009b) indicating that the efficiency (100.1%) is the same for
both kits. The standard plots obtained for the two types of tape deviate from the standard curve
122
Detection of P. horiana
based on pDNA. The standard curves for the two types of tapes are based on DNA obtained from
DNA extraction of samples spiked with a tenfold dilution of spores while the standard curve for pDNA
was calculated based on the tenfold dilution of pure pDNA. Therefore, the effect of the DNA
extraction and the matrix of the real test sample are taken into account for the calculation of the
reverse standard plots (Kubista et al. 2006). When dilutions of genomic DNA were to be used, no
significant difference between the slopes is expected, as described by Alaei et al. (2009b). The slope
of the standard curve for the Melinex tape (-3.47 ± 0.082) corresponds to a real-time PCR efficiency
of 94.2% while the slope of the standard curve for the 3M666 tape used in the Ionic spore trap (-3.15
± 0.20) indicates an efficiency of 107.7%. This is in line with what can be expected for biological
samples and the difference is probably due to a difference in DNA extraction efficiency on the two
types of tapes. As the substrate used to trap the spores on the Melinex tape was previously reported
as relatively inert and not significantly inhibiting DNA extraction or PCR assay (Calderon et al. 2002a),
inhibition of the Burkard samples will be limited. The effect of the 3M666 tape on DNA extraction
was not described yet, although the considerably lower repeatability of DNA extractions of this type
of samples suggest that further optimization might be necessary. The high robustness of the pDNA
standard plot makes it suitable to use it as a standard for the quantification of basidiospores in air
samples taken using the Burkard spore trap or the Ionic spore trap in real-time PCR. Using the
equations for the conversion of the log of pDNA equivalents to the log of spore equivalents, our
method can be widely used for the quantification of P. horiana spores on both types of tape using a
plasmid dilution series as standard.
Despite the similar LOQ and LOD for both types of samples when they are spiked, a remarkable
difference in performance between the two spore traps was noticed when they were used in the
field. Although a 20 to 60 times higher volume is sampled using the Ionic spore trap, a five to forty
times lower detection was recorded during the 2011 field trials (Figure 5.7 and 5.8, Table 5.3). These
lower numbers of captured spores are possibly due to the high volume of sampled air causing a fast
saturation of the tape and/or due to a much lower capture efficiency of this type of device.
With increasing distance to the inoculum a gradual decrease in detected spores was noticed (Table
5.3). This decrease best fitted the equation y = 1.1363x -1.412 (R² = 0.9874) with y being the proportion
remaining spores in % and x being the distance from the source (in m). This model allows estimation
of the proportion of spores remaining at a specific distance from the spore trap. For example, y
equals 0.117 (or 11.7%) at 5 m distance or 0.17% at 100 m. It also allows calculation of the distance
at which for example 1 spore is left, given an inoculum source of a given size. For example, if 100000
spores (equivalents) are released at 1 m, then 1 spore (equivalent) would still be observed at 3805 m.
However, it has to be stated that this equation is based on a limited number of data points, affecting
123
Chapter 5
its power. Additional factors such as wind speed should also be included to obtain a more reliable
model (Kuparinen 2006). As a consequence more observations at different distances and under
different climatic conditions have to be determined to develop a reliable predictive model. When the
infected plants were placed in a nearby greenhouse a significant decrease in sporulation was
observed, although it takes up to 8 days to obtain a gradual drop of detected spores to zero. The
gradual decrease we observed can be explained with the equation y = -0.3417x + 2.474 (R² = 0.7143)
in which “x” represents the days after removal of the sporulation source and y represents the log of
the Burkard spore equivalents. The decrease of spore equivalents can be the result of the presence
of released spores in the surroundings of the spore traps that remain present for some days from
where they can be reintroduced in the air by wind. However, as the survival time of released
basidiospores is limited to a few hours (Firman and Martin 1968), spores that are captured 1 day or
later after the infected plants were removed can be considered as non viable.
During trials 1, 2 and 3 only the Burkard spore trap could be used. The results of these trials show
that high levels of spores can be expected when infected plants are present in the field. Levels higher
than 20000 spores per 12 hours, corresponding to 2800 spores m3, were observed in trial 1 and trial
3 (Figures 5.4 and 5.6). During trial 1, symptomatic plants were removed to avoid further infection
and no new symptoms were observed during sampling. Nevertheless, two peaks with high spore
concentrations were observed with a month interval indicating that an inoculum source must have
been present in the test field or in the surrounding chrysanthemum fields. These two sporulation
events probably correspond to a completed infection cycle that takes 3 weeks under ideal
circumstances but can last longer under field conditions (Firman and Martin 1968; Zandvoort et al.
1968a). Similarly, during trial 3 two ‘waves’ of sporulation could be observed with the first events of
sporulation on day 1 and day 5 followed by sporulation on days 18, 22 and 24. Between two infection
cycles, sporulation is strongly reduced but not completely absent with sporulation levels between
one and 500 spores per 12 hours (Figure 5.4 insert). Similar levels of spores were detected during
trial 2 in a chrysanthemum field where no infection was present, even after intensive monitoring of
the plants (Figure 5.5). Therefore, spores are probably continuously released by inconspicuous
pustules causing a low but continuous infection pressure and can probably travel long distances of
several hundreds of meters to non-infected fields as also reported by Zandvoort (1968).
In general, sporulation clearly followed a day/night cycle with remarkably higher sporulation levels
during the night than during the day. This was first observed during the trial 1 (Figure 5.4) but also
during trial 2 (Figure 5.5), and the more detailed sampling in trial 4 (Figure 5.6). This cycle of
sporulation can be explained by the formation of dew on the leaves and the higher relative humidity
during the night. As survival of basidiospores is strongly related to relative humidity, with a drop in
124
Detection of P. horiana
survival time from 1 hour to 5 min when the relative humidity decreases from 90% to 80%,
basidiospores are probably only released at a relative humidity higher than 95% (Firman and Martin
1968). Nevertheless, relative humidity is not the only factor influencing sporulation as we observed
only limited sporulation during several nights with ideal relative humidity conditions during trial 2
(Figure 5.5). Rain events seem crucial for high levels of spore release. During trials 2 and 3, major
sporulation events occurred together with or shortly after rain events. When rain lasts for several
days, as in days 7 to 10 in the trial 2 (Figure 5.5) and days 24 to 27 in the controlled field trial (Figure
5.6), a fast decrease in sporulation is observed, probably due to exhaustion of the pustules after
intensive sporulation. Similar patterns of spore dispersal in relation to relative humidity and rainfall
have also been described for other fungi (van Niekerk et al. 2010).
Our results demonstrate that air sampling in combination with real-time PCR is suitable for the
quantitative detection of P. horiana basidiospores. However, the performance of the spore trap and
the matrix on which the spores are trapped have to be taken into consideration for the optimization
of the protocol. For sampling in the field, the Burkard spore trap is most suitable as it allows
independent sampling for longer periods (up to a week) without sample saturation and we now have
a robust, sensitive and repeatable DNA extraction protocol adapted to the sample matrix. Spore
detection can be used to monitor the presence of the pathogen in the field and allows study of the
epidemiology and biology of the fungus. In the future, this technique can be used for the
development of a warning network using less expensive spore traps (i.e. rotating arm samplers) using
the same type of capture matrix. To avoid saturation of the capture matrix, sampling time can be
controlled in function of the weather conditions to increase the chance of detection in a limited
sampling period. Such a warning network would reduce the number of fungicide applications by
decreasing the number of preventive fungicide treatments. Also, in combination with genotyping of
the spores, it has the potential to guide breeders in the deployment of the right resistance genes,
further reducing the impact of this quarantine pathogen.
125
Part III:
Inheritance in chrysanthemum
Chapter 6:
Segregation of resistance
to P. horiana in chrysanthemum
Shaving of a chrysanthemum flower in order to reach the stigmas for
controlled pollination
Segregation of resistance to P. horiana
6.1 Introduction
As soon infections with Puccinia horiana were observed in the chrysanthemum growing regions of
Europe, the pathogen and its relation with its host were extensively studied. The outbreaks of
chrysanthemum white rust in the UK between 1963 and 1967 were described in detail by Baker
(1967), who also recorded the cultivars on which infection was observed. In the same period,
outbreaks were reported in Germany (Stark and Stach 1965), Sweden (Nilsson 1964), Denmark
(Anon. 1964)and Norway (Gjaerum 1964). Since its introduction in Europe, the pathogen further
spread in countries with a substantial trade of cut flowers and cuttings, where it is nowadays
commonly present (Rattink et al. 1985; Whipps 1993). Further research on P. horiana focused mainly
on the development of control systems including cultural measures (Whipps 1993), fungicide
application (Dickens 1990; Dickens 1991; Cook 2001) and biological control (Whipps 1993; GarcíaVelasco et al. 2005). However, the most sustainable method to control the spread and development
of a pathogen is the use of resistant cultivars. This implies the identification and selection of
resistant genotypes in Chrysanthemum breeding programs (Hammond-Kosack and Jones 1997;
McIntosh and and Brown 1997).
The availability of resistance to P. horiana in commercial cultivars has been the focus of several
studies. In a study carried out by Dickens (1968) more than 40 florist chrysanthemum cultivars were
artificially inoculated with P. horiana from which 37 were found to be resistant. He also included
other chrysanthemum species in his research, of which C. yezoense, C. koreanum and C. indicum
showed very high levels of infection, while in C. indicum variation in susceptibility was observed.
More than 20 other Chrysanthemum species and their cultivars remained uninfected under
conditions favorable for infection, although it is not clear whether all these species are hosts for P.
horiana. A similar study including more than 200 florist chrysanthemum cultivars classified the
plants in six susceptibility classes ranging from susceptible over resistant to immune (Martin and
Firman 1970). However, the reaction of specific cultivars to P. horiana infection in different studies
was frequently inconsistent, suggesting the existence of different pathotypes of the pathogen and
specific resistance genes in the plant (Dickens 1971).
The variation in resistance to different strains of the pathogen was studied in detail by Yamaguchi
(1981), who inoculated 40 cultivars with six different strains of P. horiana. The presence of different
pathotypes and the differential resistance response of the cultivars that were tested were
demonstrated. This was further supported by a more recent study in Poland on the susceptibility of
50 cultivars using natural infection (Wojdyla 1999b). Similar results were obtained by Velasco et al
(2007) using isolates originating from different regions of Mexico. Recently, the presence of
131
Chapter 6
pathotypes was confirmed once more in an extensive study showing differential interaction
phenotype profiles after inoculation of 36 differential chrysanthemum cultivars with 22 pathotypes
of P. horiana (De Backer et al. 2011).
Next to knowledge on the pathosystem, crucial factors for the design of efficient breeding programs
for resistance are knowledge about the number of resistance genes involved, their mode of action
and their segregation. However, little is known about the patterns of genome recombination and
segregation in chrysanthemum, and about the inheritance of resistance to P. horiana in particular.
The only study on the inheritance of resistance to P. horiana in chrysanthemum suggested
segregation by preferential pairing rather than by random pairing. However the segregation was not
always convincing, suggesting a mix of two pairing systems (de Jong and Rademaker 1986). Another
study on the inheritance of a carotenoid inhibitor gene in chrysanthemum, inhibiting the synthesis of
carotenoid causing white or pink flowers instead of yellow or bronze, supported random
chromosome pairing, but conflicting observations were made as well (Langton 1989).
Mainly three types of resistance to white rust have been described in chrysanthemum with different
cultivars expressing different types (Rademaker and de Jong 1987). Complete resistance to a
particular pathotype of P. horiana is regulated by a single dominant resistance gene (de Jong and
Rademaker 1986), making it easy to use in breeding programs. However, due to natural selection
pathogens can easily break through this type of resistance (Jones and Dangl 2006) after which
alternative resistance genes have to be found and incorporated in breeding programs. Complete
resistance occurs when the gene product of the resistance gene (receptor) is capable to interact with
an elicitor, the gene product of a pathogen avirulence gene. Recognition of pathogen elicitors by
plant resistance receptors triggers the host defense mechanism resulting in a hypersensitivity
response that arrests the development of the pathogen (Staskawicz et al. 1995). In such a gene-forgene resistance, recognition of different pathotypes of a pathogen requires different resistance
genes resulting in complex pathosystems (Flor 1956; Flor 1971). Based on the interaction phenotype
profiles of P. horiana isolates and resistance profiles of chrysanthemum cultivars, it was determined
that at least seven resistance and corresponding avirulence genes are present in the P. horiana –
chrysanthemum system (De Backer et al. 2011). Also in other rust pathosystems complex gene-forgene interactions have been described. For example, in the flax/Melampsora lini system, 31 R genes
located on 5 different loci have been described (Ellis et al. 2007). More than 40 and more than 60 R
genes have been described respectively against the wheat rusts Puccinia graminis and P. triticina
(Leonard and Szabo 2005; Bolton et al. 2008).
132
Segregation of resistance to P. horiana
A second type of resistance that has been observed in chrysanthemum is incomplete resistance. For
infection, high spore pressure and optimal environmental conditions are required due to a more
difficult penetration of the leaf. Although the fungus can complete its life cycle, usually only a few
pustules develop which are rather small and mature slowly (Rademaker and de Jong 1987). This
quantitative type of resistance can be regulated by multiple genes playing a role in plant anatomy /
morphology or in different plant defense pathways (Poland et al. 2009). Although this type of
resistance is not pathotype specific and is more difficult to break by simple mutations in the
pathogen, its multigenic nature makes it more difficult to use in breeding strategies. A final type of
resistance that is occasionally observed in chrysanthemum is extensive necrosis around growing rust
colonies due to a late form of hypersensitivity. This reaction is generally too slow to completely
prevent sporulation and is not inherited in a gene-for-gene way (Rademaker and de Jong 1987).
For breeding purposes, complete resistance based on dominant genes is preferred above incomplete
resistance since they are easy to incorporate in breeding strategies and the qualitative interactions
can be phenotyped efficiently (Pink 2002), although it has been suggested that a combination of
both types of resistance will provide the most durable solution (Poland et al. 2009).
Although the presence of resistance to P. horiana in chrysanthemum cultivars has been studied in
the past, the knowledge about the inheritance of resistance to the pathogen is limited and mostly
based on the reaction to a single isolate. Since the P. horiana-chrysanthemum pathosystem is very
complex with at least 7 Avr and R genes involved (De Backer et al. 2011), the segregation of the
resistance to different pathotypes can be variable, depending on the genomic organization of these
genes. The objective of this study was thus to determine the mode of inheritance of resistance to
different pathotypes in progenies derived from pair-crosses between resistant and susceptible
Chrysanthemum plants. A set of 12 resistant and 2 susceptible genotypes were assembled in
different combinations of one resistant and one susceptible genotype to generate 20 progenies. In a
first step, these 20 progenies were screened for resistance against a mix of European isolates. In a
second step, one selected progeny was screened more in detail with four pure isolates representing
different pathotypes (according to De Backer et al. (2011)). Based on the observed segregation
ratios, the number of copies of the resistance gene present in the corresponding resistant parent
(simplex, duplex, triplex) was estimated, as well as the way of bivalent formation (preferential vs.
random).
133
Chapter 6
6.2 Material and methods
6.2.1 Plant materials
Plant materials for the resistance screening experiments described below were kindly provided by
the chrysanthemum breeding company ‘Paraty breeding’ bvba. They performed the crosses and
delivered the progenies for resistance screening.
A set of 12 resistant and 2 susceptible chrysanthemum cultivars were selected after a preliminary
screening with undefined isolates of a larger set of plants, and based on the knowledge and
experience of ‘Paraty breeding’ bvba. The resistance status of the parents was checked at ILVO using
a mixed isolate of P. horiana composed of 6 European isolates (see further). In addition, cultivars R6,
R7 and S1 were screened with 4 pure isolates from different pathotypes, while the remaining 10
resistant cultivar were screened with three of these pure isolates (see further).
The 14 selected cultivars were used to carry out pair-crosses. Most crosses were strictly outcrosses,
although the progeny of some crosses had an inbreeding coefficient between 0.25 and 0.015 (Figure
6.1; Table 6.1). A total of 20 progenies was generated by crossing each resistant cultivar with the
susceptible cultivar S1 and by crossing 8 resistant cultivars (R1-R5 and R10-R12) with susceptible
cultivar S2 (Table 6.1). All crosses were performed in two directions as seed yield often depends on
the direction of the cross. The direction of the crosses that was used for the resistance screening and
their respective seed yields are presented in Table 6.1. In 15 of the 20 progeny, the resistant parent
was used as seed plant as there was no or only very limited seed set in the reciprocal crosses (data
not shown).
The 20 progenies listed in Table 6.1 were screened using a mixed P. horiana isolate. To guarantee
sufficient statistical power, all progeny plants that could be screened were included in the resistance
screening. From January till May 2008, seeds obtained from the different crosses were sequentially
sown in potting soil on a weekly basis and grown up in a greenhouse. Seven weeks after sowing the
seedlings were pruned to induce branching and ten weeks after sowing the young plants were
delivered at ILVO for resistance screening. At ILVO the plants were maintained in the greenhouse
between 10°C (night) and 20°C (day) and 16 hours light. For resistance screening, we used unrooted
cuttings. To assure that for a maximum of progeny plants three cuttings (= replicates) could be
screened, we chose the most vigorously growing seedlings. However, for some plants only 2 or 1
cuttings could be obtained for testing. The cuttings were labeled, planted in rock-wool blocks
(Grodan AO 36/40) and placed in trays for the bioassay. The individual progeny plants were labeled
and kept in the greenhouse until the results of the screening were known.
134
Segregation of resistance to P. horiana
Figure 6.1: Pedigree of the cultivars used for crosses (represented in squares). Cultivars represented in rectangular boxes
a
b
are ancestral plants for the generation of these parents. Ancestral plants with known resistance to P. horiana. Ancestral
c
plants susceptible to P. horiana. Ancestral plants showing differential resistance to isolates Ph PD 20 and Ph 707 (Wim
Declercq, personal communication).
The parental plants and the progeny of cross 17 (Table 6.1) were planted in 1.5 liter pots of 15 cm
diameter and placed in a greenhouse to produce several branches that could be used in detailed
resistance screening using single pustule isolates, as described below. In addition, from every
progeny plant of cross 17, young leaf material was collected, lyophilized and vacuum sealed for
molecular research (see Chapter 7).
135
Chapter 6
Table 6.1: Overview of the crosses between resistant and susceptible cultivars used in this study. For every cross the
respective seed plant and pollen plant are mentioned, as well as the inbreeding coefficient for the resulting progeny. Also
the seed yield for every cross is given.
Number
progeny
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Seed plant
Pollen plant
R10
R11
R12
R10
R11
R12
R1
R2
S1
S1
S1
R1
R2
R3
R4
R5
S1
R7
R8
S1
S1
S1
S1
S2
S2
S2
S1
S1
R3
R4
R5
S2
S2
S2
S2
S2
R6
S1
S1
R9
Inbreeding
coefficient
0.12500
0.25000
0.03125
0
0
0.08984
0
0.03125
0.12500
0
0
0
0
0
0
0
0
0.01563
0.07813
0
Seed yield
(number of seeds)
237
464
491
136
157
161
684
1084
227
622
226
275
321
704
295
523
121
119
104
303
6.2.2 Puccinia horiana isolates
The set of 14 parental genotypes and the 20 progenies were screened using a mixed isolate (referred
to as Ph Mixed), based on European isolates with similar pathotypes and composed of isolates Ph
301, Ph 304, Ph 313, Ph SGP UK, Ph Spalding UK and Ph Poland. The culture was started up by
inoculating healthy cuttings of the susceptible cultivars Medonia and Taliedo in a single recipient
with three to four heavily infected leaves obtained from the respective single pustule isolates.
The resistant parent plants (R1-R12) and the susceptible parent S1 were also screened with single
pustule cultures of isolates Ph Japan 2, Ph PD 20 and Ph 707. Cultivars R6, R7 and S1 were
additionally screened with the fourth single pustule isolate Ph 801. These single pustule isolates
belong to different pathotypes (De Backer et al. 2011).
Based on the results obtained, cross 17 was selected for a more detailed analysis. Progeny plants
were inoculated with single pustules cultures of isolates Ph Japan 2, Ph PD 20, Ph 707 and Ph 801.
Cultures of P. horiana were maintained on healthy cuttings of the cultivars Medonia and Taliedo by
three-weekly transfer as described by Alaei et al. (Alaei et al. 2009a) and were treated separately to
avoid cross contamination.
136
Segregation of resistance to P. horiana
6.2.3 Bioassay and resistance screening
The resistance screening was done using the bioassay described in chapter 2. In brief, 50 fresh
cuttings of the cultivars Medonia and Taliedo placed in Grodan AO 36/40 Rockwool slabs for
inoculum production or up to 36 cuttings placed in plastic trays (9 by 4 wells, cut out of Grodan SBS
36/77 trays) for resistance screening were placed on the bottom of a plastic container (Savic Super
Large, Savic, Belgium) with glass plates cut to size as covers. Plastic containers, glass covers and trays
were washed and sterilized before each assay. Plastic containers and glass covers were sterilized
with 70% ethanol and trays with 0.5% hypochlorite. For the production of basidiospores, around 36
heavily infected leaves were stuck to the cover of the recipient with 1% water agar, the telia pointing
downwards. Basidiospore production and germination were induced by spraying the cuttings, the
inner side of the plastic tank and the inoculum with demineralized water using a Preval Sprayer
(Yonkers; NY; USA). This misting ensures a high relative humidity and a water film on the leaf
surfaces and was repeated 24 hours post inoculation. At the bottom of the recipients, a 1 cm-deep
water layer contributed further to the high relative humidity and provided water for the plants. After
inoculation, the plants were placed in a dark growth chamber at 17°C for two days after which
fluorescent light (Gro-lux® F58W/GRO-T8, Osram Sylvania, MA, USA) was provided during 16 hours
per day. Three weeks after inoculation, symptoms were scored.
For the initial screening of the parental cultivars and for the screening of the progeny of the 20
crosses, Ph Mixed was increased to approximately 200 infected cuttings of cultivars Medonia and
Taliedo by three to four three-week cycles. This allowed the screening of up to 300 cuttings (3
replicates of 100 plants) and the maintenance of the inoculum for subsequent weekly resistance
screenings. During inoculum build up, the Ph Mixed culture was split up into three parallel lines to be
able to perform resistance screenings on a weekly basis with the same culture of P. horiana. This was
done by transferring a part of the culture to healthy cuttings at 18 dpi and the other part at 21 dpi
after the first cycle. After the second cycle, the first line was again transferred after 18 days and the
second line again after 21 days resulting in two lines with one week difference. This procedure was
repeated with the first line to obtain three lines that can be used for screening in three subsequent
weeks.
For the resistance screening of plants R1-R12, S1 and the progeny of cross 17, single pustule isolates
of Ph Japan 2, Ph PD 20, Ph 707 and Ph 801 were increased by three to four weekly cycles.
For the preliminary screening of cultivars, 10 cultivars (times 3 replicates) were screened per plastic
container. This was also the case for the confirmation of unclear disease scoring of some particular
progeny of cross 17 after the first screening with the single pustule isolates.
137
Chapter 6
For the resistance screening of the progeny of the crosses we used nine containers in three sets of
three. Per plastic container 33 to 34 cuttings were screened and 2 to 3 cuttings of cultivar Medonia
were added as positive control. For every progeny plant, three replicates were placed in separate
plastic containers. Per screening of 100 progeny, 3 replicates, each consisting of 3 plastic containers
were inoculated. For the screening of a maximum of ten plants three replicates could be screened in
one plastic container. This was the case for screening of the parent plants or for confirmation of
unclear disease scoring of some particular progeny of cross 17 after the first screening with the
single pustule isolates.
The scoring was based on the scoring system described in chapter 2. Each cutting was assigned a
disease score of 2 (highly infected), 1 (slight infection) or 0 (no infection). For each plant, the number
of cuttings in each scoring class was used to calculate an overall interaction phenotype score in
which the number of highly infected cuttings was multiplied by two, the number of slightly infected
cuttings multiplied by one and the number of healthy cuttings multiplied by zero. This results in a
score ranging from 0 (complete resistance) to 6 (very susceptible) when three replicates are used
(see chapter 2). Plants with an interaction phenotype score of 4 or higher were considered as
susceptible; plants with a score of less than 4 were considered as resistant. If more or less than three
replicates were available, the scores were standardized to the corresponding score based on three
replicates. I.e. score 420 includes 6 replicates from which 4 heavily infected and 2 with low infection
rates. This corresponds to 2 heavily infected plants and 1 intermediate resistant plant in the case
only three replicates should be scored giving an overall interaction score of 5 as explained above.
6.2.4 Analysis of segregation data
For the progenies of the 20 different crosses that were screened with Ph Mixed and for the
segregation of resistance to the single pustule isolates in cross 17, the number of resistant and
susceptible progeny plants was counted. The observed values were compared to the theoretical
expectations assuming different numbers of resistance genes carried by the resistant parent.
Observed and expected values were compared using chi²-tests (P<0.05). The following possibilities
were considered:
x
x
x
Susceptible cultivar (rrrrrr) x simplex resistant cultivar (Rrrrrr)
Susceptible cultivar (rrrrrr) x duplex resistant cultivar (RRrrrr)
Susceptible cultivar (rrrrrr) x triplex resistant cultivar (RRRrrr).
For each of these situations the possibility of random or preferential bivalent formation was
considered and the respective expected segregation ratios were estimated. The choice of
138
Segregation of resistance to P. horiana
segregation ratios to test was based on the fact that resistance is determined by dominant genes. As
a result the resistant parent of each cross possesses at least one copy of the gene that confers
resistance to the isolate used in the screening, while the susceptible parent does not carry this
particular gene. Nevertheless it has to be noticed that for duplex or triplex configurations, it will be
difficult to distinguish between preferential and random pairing due to the small differences in
expected segregation proportions (Table 6.3 and Table 6.5).
Cross 17 was selected for screening using single pustule isolates, since the resistance to Ph Mixed
followed a simplex segregation, and differential interactions to different pure isolates were observed
in the parent plants (see results section). For this cross, we also considered the possibility that both
parents carried one to three resistance genes with random or preferential bivalent formation. We
calculated the Pearson correlation coefficient between the resistance scores obtained for the
progeny plants using Ph Mixed and the score assigned using each of the four single pustule isolates.
When a correlation coefficient higher than 0.75 was achieved, the segregations were considered to
be highly correlated.
All statistical analyses were performed with Statistica 9.0 (Statsoft, OK, USA).
6.3 Results
6.3.1 Resistance of parent plants
All parents used in the crosses, except S2, were screened with Ph Mixed and the single pustule
isolates Ph PD 20, Ph Japan 2 and Ph 707 (Table 6.2). Parents R6, R7 and S1 were additionally
screened with the single pustule isolate Ph 801. The susceptible parent S2 was only screened with Ph
Mixed. The results of these tests are summarized in Table 6.2. As expected, after inoculation with Ph
Mixed all resistant cultivars remained healthy and both susceptible parents were highly infected.
Slightly different results were obtained when these plants were screened using single pustule
isolates. The cultivars R1, R3 and R9 showed slight infection when inoculated with isolates Ph Japan
2, Ph 707 or Ph PD 20, respectively. Three cultivars (R7, R11 and R12) showed a high susceptibility to
isolate Ph 707 and cultivar R7 showed also a high susceptibility to Ph 801. The susceptible cultivar S1
showed a high susceptibility to the single pustule isolates Ph PD 20 and Ph Japan 2, but remained
completely uninfected after inoculation with isolate Ph 707. It showed only a very limited infection
after inoculation with isolate Ph 801.
139
Chapter 6
Table 6.2: Interaction of the 12 resistant (R) and 2 susceptible (S) parent plants after inoculation with five different isolates.
Susceptible parent plant S2 was only screened with Ph Mixed and isolate Ph 801 was only used to screen parents R6, R7
and S1. For the remaining interactions no data are available (ND). Numbers represent the susceptibility scores ranging from
0 (complete resistance) to 6 (very susceptible).
a
Isolates P. horiana
Ph PD 20
Ph Japan 2
Ph 707
Ph 801
Parent plants
Ph Mixed
(NL3)
(JP1)
(BE5)
(BE6)
R1
0
0
3
0
ND
R2
0
0
0
0
ND
R3
0
0
0
1
ND
R4
0
0
0
0
ND
R5
0
0
0
0
ND
R6
0
0
0
0
0
R7
0
0
0
6
6
R8
0
0
0
0
ND
R9
0
1
0
0
ND
R10
0
0
0
0
ND
R11
0
0
0
6
ND
R12
0
0
0
6
ND
S1
6
4
4
0
1
S2
6
ND
ND
ND
ND
a
Isolate codes between brackets correspond to the codes used in Chapter 2 and Chapter 4.
6.3.2 Segregation of resistance in the 20 progenies
Depending on the number of seeds obtained and their germination capacity, between 65 and 222
progeny plants could be screened per cross (Table 6.3) using Ph Mixed. For 16 crosses 100 progeny
plants or more could be screened. In crosses 1, 4, 6 and 19 it was not possible to generate 100
seedlings and only 74, 65, 66 and 86 seedlings were used, respectively.
For every cross, the observed number of resistant and susceptible progeny plants was compared to
the expectation, assuming different hypotheses: the resistant parent carried either one, two or three
copies of the resistance gene (Table 6.3). In four cases (crosses 3, 10, 13 and 15) the observed
proportion of resistant and susceptible plants did not fit any of the possibilities tested. For 16
crosses, the observed segregation ratios fitted one of the ratios tested. The segregation proportions
observed for crosses 5, 14, 16, 17, 18 and 20 could be explained by the presence of a single
resistance gene in the resistant parent. In this case the mode of bivalent formation cannot be
determined since random or preferential bivalent formation both give the same expected
segregation proportions. For eight crosses, the observed segregations could be explained by R
parents carrying two copies of the resistance gene (duplex). In five of these crosses (2, 7, 9, 11 and
19) the observations indicated preferential pairing, while in three other crosses (6, 8 and 12) the
data could be explained by preferential as well as by random pairing.
140
Segregation of resistance to P. horiana
The segregation of resistance in the progeny of cross 1 only fitted with the expected segregations of
a triplex parent with preferential bivalent formation. Finally, for cross 4 it could not be determined
whether the resistant parent carried two or three copies of the resistance gene since the observed
segregation fits the expected segregations of a duplex situation with random or preferential pairing
as well as with a triplex situation with preferential pairing. These ambiguous results are due to the
small differences in expected segregations among those categories and could only be solved by
analyzing a much larger number of progeny plants.
141
Table 6.3: Observed and expected number of resistant and susceptible plants observed in the different crosses, assuming different segregation models. For every cross, the total number of progeny plants
screened as well as the number of resistant and susceptible plants is given. Five possible situations were considered, depending on the number of copies of the resistance gene carried by the resistant parent and
the type of bivalent formation. Observed and expected values were compared using Chi²-tests. The P-value of the Chi² tests performed is shown. P values above the threshold value of 0.05 (indicating no
significant difference between observed and expected proportions) are underlined.
Genes in parents
Pairing
Simplex x nulliplex
Rand./ pref
Cross 1
Cross 2
Cross 3
Cross 4
Cross 5
Cross 6
Cross 7
Cross 8
Cross 9
Resistant
66
93
137
52
51
52
97
78
80
30
72
74
66
54
39
57
53
42
61
Susceptible
8
36
85
13
49
14
36
22
36
70
28
26
34
46
61
43
47
58
25
56
Total
74
129
222
65
100
66
133
100
116
100
100
100
100
100
100
100
100
100
86
100
50
Resistant
50%
37
64,5
111
32,5
50
33
66,5
50
58
50
50
50
50
50
50
50
50
50
43
Susceptible
50%
37
64,5
111
32,5
50
33
66,5
50
58
50
50
50
50
50
50
50
50
50
43
50
1,56
E-11
5,21
E-07
4,83
E-04
1,32
E-06
0,841
2,90
E-06
1,23
E-07
2,14
E-08
4,40
E-05
6,33
E-05
1,08
E-05
1,59
E-06
1,37
E-03
0,424
2,78
E-02
0,162
0,549
0,110
1,04
E-04
0,230
80
Chi² P-value
Random
(RrRrrr x rrrrrr)
Resistant
80%
59,2
103,2
177,6
52
80
52,8
106,4
80
92,8
80
80
80
80
80
80
80
80
80
68,8
Susceptible
20%
14,8
25,8
44,4
13
20
13,2
26,6
20
23,2
20
20
20
20
20
20
20
20
20
17,2
20
4,81
E-02
2,48
E-02
9,61
E-12
1.000
4,17
E-13
0,806
4,16
E-02
0,617
2,97
E-03
7,47
E-36
4,55
E-02
0,134
4,65
E-04
8,03
E-11
1,18
E-24
8,92
E-09
1,48
E-11
2,10
E-21
3,55
E-02
2,26
E-19
75
Chi² P-value
a
Preferential
Resistant
75%
55,5
96,75
166,5
48,75
75
49,5
99,75
75
87
75
75
75
75
75
75
75
75
75
64,5
Susceptible
25%
18,5
32,25
55,5
16,25
25
16,5
33,25
25
29
25
25
25
25
25
25
25
25
25
21,5
25
4,82
E-03
0,446
4,82
E-06
0,352
2,98
E-08
0,477
0,582
0,488
0,133
2,69
E-25
0,488
0,817
3,77
E-02
1,24
E-06
9,27
E-17
3,23
E-05
3,76
E-07
2,52
E-14
0,383
8,12
E-13
95
Chi² P-value
Triplex x nulliplex
Random
(RrRrRr x rrrrrr)
Resistant
95%
70,3
122,55
210,9
61,75
95
62,7
126,35
95
110,2
95
95
95
95
95
95
95
95
95
81,7
Susceptible
5%
3,7
6,45
11,1
3,25
5
3,3
6,65
5
5,8
5
5
5
5
5
5
5
5
5
4,3
5
2,18
E-02
7,54
E-33
1,21
E-114
2,88
E-08
1,23
E-90
1,51
E-09
1,67
E-31
6,18
E-15
7,03
E-38
1,91
E-195
4,91
E-26
5,66
E-22
2,13
E-40
6,01
E-79
1,34
E-145
4,43
E-68
9,42
E-83
1,26
E-130
1,29
E-24
4,23
E-121
87,5
Chi² P-value
Preferentiala
Resistant
87,5%
64,75
112,875
194,25
56,875
87,5
57,75
116,375
87,5
101,5
87,5
87,5
87,5
87,5
87,5
87,5
87,5
87,5
87,5
75,25
Susceptible
12,5%
9,25
16,125
27,75
8,125
12,5
8,25
16,625
12,5
14,5
12,5
12,5
12,5
12,5
12,5
12,5
12,5
12,5
12,5
10,75
12,5
0,660
1,22
E-07
3,33
E-31
0,067
2,55
E-28
3,23
E-02
3,78
E-07
4,07
E-03
1,58
E-09
1,05
E-67
2,78
E-06
4,46
E-05
7,98
E-11
4,09
E-24
1,08
E-48
2,91
E-20
1,77
E-25
4,57
E-43
3,38
E-06
1,63
E-39
Chi² P-value
a
44
expected
(Rrrrrr x rrrrrr)
Duplex x nulliplex
Cross 10 Cross 11 Cross 12 Cross 13 Cross 14 Cross 15 Cross 16 Cross 17 Cross 18 Cross 19 Cross 20
expected ratio’s assuming that the resistance genes are located on homoeologous chromosomes, otherwise all progeny would be resistant.
Segregation of resistance to P. horiana
6.3.3 Segregation of resistance to different pathotypes in cross 17
The progeny plants of cross 17 (S1 x R6) responded differentially to isolates of different pathotypes.
The correlation between the interaction phenotype scores was calculated for Ph Mixed and for three
pathotypes (Table 6.4). A high correlation was found between the response to Ph Mixed at one side
and the response to infection with the single pustule isolates Ph 707 and Ph 801 at the other side,
whereas the correlation of the segregation to Ph Mixed with the other single pustule isolates was
rather low. The resistance to isolate Ph 707 and to isolate Ph 801 are strongly correlated as it is also
the case for the segregation of resistance to isolates Ph PD 20 and Ph Japan 2.
Table 6.4: Correlation coefficients for resistance scores in the progeny of cross 17 after resistance screening with different
isolates of P. horiana. Correlation values higher than 0.75 have been shaded.
Correlation
Mixed isolate
Ph PD 20
Ph Japan 2
Ph 707
Ph 801
*P < 0.05
Ph Mixed
Ph PD 20
Ph Japan 2
Ph 707
Ph 801
1,0000*
0,2485*
0,1670
0,7697*
0,7694*
1,0000*
0,8344*
0,4629*
0,5282*
1,0000*
0,3960*
0,4036*
1,0000*
0,9256*
1,0000*
The mode of segregation of the resistance to the isolates representing pathotypes was tested with a
Chi²-test to the expected segregations as described above (Table 6.5). Since the parents of cross 17
(R6 and S1) both appeared to be resistant to isolates Ph 707 and Ph 801, we also considered the
situation in which both parents possess one to three copies of the resistance gene, and segregation
occurred by either random or preferential bivalent formation. Resistance to isolates Ph Mixed, Ph PD
20 and Ph Japan 2 could be explained by a simplex configuration of the resistance gene in parent R6.
The segregation of the resistance to isolates Ph 707 and Ph 801 could be explained as the result of a
cross between two simplex parents, what means that both parents possess one single copy of the
corresponding resistance gene for these isolates. The data observed could also be explained by a
duplex x nulliplex cross with preferential pairing of homologous chromosomes, however, this does
not agree with both parents being resistant.
143
Chapter 6
Table 6.5: Segregation of resistance in cross 17 to Ph Mixed and four single pustule isolates. Isolate, the total number of
progeny plants screened is given as well as the number of resistant and susceptible progeny. Every observed segregation
was tested with a Chi²-test to 16 situations of segregation depending on the number of copies of the resistance gene in one
or both parent plants and the type of bivalent formation during meiosis. For every situation corresponding P-value of the
Chi² test is given. P-values of observations that show no significant difference with an expected situation are underlined.
The most probable explanation for every cross is marked with a shaded field.
Resistant
Susceptible
Total
Expected
resistant/susceptible
Simplex x nulliplex
Rand. / pref.
(Rrrrrr x rrrrrr)
Duplex x nulliplex
Random
(RrRrrr x rrrrrr)
a
Preferential
59
29
88
Chi²
50%/50%
P
0,549
1
0,458
2,11E-3
1,38E-3
80%/20%
P
1,48E-11
6,3E-13
7,85E-10
1,22E-3
2,38 E-3
3,76E-07
3,06E-08
4,43E-06
0,0578
0,0848
9,42E-83
2,73E-87
1,08E-70
1,88E-35
2,41E-33
87,50%/12,50%
P
1,77E-25
1,5E-27
1,02E-21
1,45E-09
6,56E-09
75%/25%
P
3,76E-07
3,06E-08
4,43E-06
0,0578
0,0848
90%/10%
P
5,99E-35
1,89E-37
8,13E-30
8,97E-14
7,09E-13
Preferential
87,50%/12,50%
P
1,77E-25
1,5E-27
1,02E-21
1,45E-09
6,56E-09
Random
97,50%/2,50%
P
1,1E-178
3,3E-187
9,7E-153
2,54E-79
8,47E-75
Preferentiala
93,75%/6,25%
P
1,36E-63
2,52E-67
2,61E-54
1E-26
4,23E-25
Random
96,00%/4,00%
P
1E-106
2,9E-112
3,39E-91
2,12E-46
1,09E-43
a
Preferential
93,75%/6,25%
P
1,36E-63
2,52E-67
2,61E-54
1E-26
4,23E-25
Random
99,00%/1,00%
P
0
0
0
3,7E-211
2,1E-199
Preferential
96,88%/3,13%
P
2,6E-140
3,1E-147
6,5E-120
9,37E-62
3,38E-58
Random
99,75%/0,25%
P
0
0
0
0
0
Preferentiala
98,44%/1,56%
P
7,3E-294
3,9E-307
3,1E-251
4,7E-132
1,2E-124
a
a
Ph 801
59
30
89
P
a
Triplex x triplex
(RrRrRr x RrRrRr)
Ph 707
48
41
89
P
Simplex x simplex
Rand. / pref.
(Rrrrrr x Rrrrrr)
Simplex x duplex
Random
(Rrrrrr x RrRrrr)
Duplex x triplex
(RrRrrr x RrRrRr)
Ph Japan 2
46
46
92
95%/5%
Preferentiala
Duplex x duplex
(RrRrrr x RrRrrr)
Ph PD 20
53
47
100
75%/25%
Triplex x nulliplex
Random
(RrRrRr x rrrrrr)
Simplex x triplex
(Rrrrrr x RrRrRr)
Ph Mixed
Expected ratios assuming that the resistance genes are located on homoeologous chromosomes, otherwise all progeny
would be resistant.
144
Segregation of resistance to P. horiana
6.4 Discussion
6.4.1 Reaction of parent plants to P. horiana infection
Several studies have demonstrated the existence of pathotypes in P. horiana (Dickens 1971;
Yamaguchi 1981; De Backer et al. 2011). Also after inoculation of the different parent plants of the
crosses used in this study, differential reactions were observed (Table 6.2). The parents were
classified as resistant or susceptible primarily based on the results after inoculation with a mixed
isolate (Ph Mixed) containing European isolates. Despite this classification, three resistant plants
showed light infection after inoculation with three single pustule isolates, and three other plants
were heavily infected by isolate Ph 707. On the other side, the susceptible plant S1 remained
(practically) uninfected when inoculated with isolates Ph 707 or Ph 808. This is in accordance with
the results obtained by De Backer et al. (2011) where differential interactions were observed on a set
of test cultivars using these isolates.
Infection with isolate Ph PD 20 was expected to be much more pronounced since this isolate was
identified as one of the most virulent and aggressive ones by Backer et al. (2011). The fact that
several cultivars with a broad resistance spectrum (resistant to Ph Mixed) were susceptible to isolate
Ph 707, and less susceptible to the aggressive Ph PD 20 can be explained by selection for Ph PD 20
specific resistance genes by Paraty breeding (Figure 6.1)(Wim Declercq, personal communication).
The slightly lower infection of the susceptible parent S1 to isolates Ph PD 20 and Ph Japan 2 is
possibly due to the fact that the cuttings used for the screening with these isolates came from older
plants. It is known that older leaves can be infected, but the number of lesions is lower because they
are more difficult to be penetrated by the fungus (Firman and Martin 1968; Panter and Jones 2002).
6.4.2 Segregation of resistance in the 20 progenies
For most crosses, around 100 progeny plants could be obtained. However, for four crosses this was
not possible. This can be explained by the fact that the seed yield in crosses 4, 6 and 19 was rather
low compared to the other crosses, with a maximum of 161 seeds produced (Table 6.1). As a result,
no additional seeds could be sown in the case of incomplete germination. Three of these crosses (1, 6
and 19) were also partially inbred possibly influencing seed production, germination and seedling
fitness by inbreeding depression (Anderson et al. 1992). In some cases, due to poor branching of
some of the seedlings after pruning, we could not take 3 replicates for screening. The number of
seedlings from which only one replicate could be taken appeared to be lower for non-inbred crosses
than for inbred crosses with 5.5% and 32.6% of the screened seedlings, respectively. This observation
is possibly due to a lower fitness of the progeny of inbred crosses (Anderson et al. 1992).
145
Chapter 6
For 16 of the 20 progenies investigated a plausible genotypic composition of the parents and
segregation pattern could be identified. For four crosses, the segregation observed did not fit any of
the possibilities explored. One has to take into account that observed segregation ratios can differ
from expectations since unbalanced chromosome numbers causing aneuploidy are often observed in
chrysanthemum hybrids (Dowrick 1953). To determine whether random chromosome pairing or
preferential chromosome pairing is involved during meiosis, only crosses in which the resistant
parent is duplex or triplex for the resistance gene can be used. Though, in the case of preferential
pairing the assumption has to be made that the different copies of a particular resistance gene are
positioned on non-pairing chromosomes, otherwise all progeny would be resistant and no
segregation would be observed.
Nevertheless, in the 16 progenies for which the observed data fitted a particular segregation pattern,
we could not discard the hypothesis of preferential pairing (Table 6.3). A clear preferential pairing
was observed in five progenies from a duplex x nulliplex cross (crosses 2, 7, 9, 11 and 19) and one
progeny from a triplex x nulliplex cross (cross 1). For the six progenies showing a simplex segregation
(crosses 5, 14, 16, 17, 18 and 20), we cannot differentiate between preferential and random bivalent
formation since expected segregations are equal. Also for progenies 6, 8 and 12, indicating a duplex x
nulliplex situation, we could not differentiate between preferential and random pairing. In the case
of cross 4, it was not possible to differentiate between duplex or triplex situation in the resistant
parent, nor between preferential or random pairing. For the duplex x nulliplex and triplex x nulliplex
crosses a much larger number of seedlings will have to be screened to be able to differentiate
between random and preferential pairing, as this would increase the power of the test. Our
observations are in accordance with the observations on inheritance of resistance to P. horiana made
by De Jong and Rademaker (1987) suggesting preferential pairing, although random chromosome
pairing was suggested by Langton (1989).
Although relatively clear segregation ratios were observed in most crosses, the conclusions that can
be made about the number of resistance genes in the resistant parents are conflicting (Table 6.1 and
Table 6.3). In crosses 1 and 2, the susceptible plant S1 was crossed with the resistant cultivars R10
and R11 respectively. The observed segregations indicate a triplex situation for cultivar R10 and a
duplex situation for cultivar R11. In crosses 4 and 5, the same resistant cultivars were crossed with
the susceptible cultivar S2 resulting in a segregation indicating respectively a triplex or duplex and a
simplex situation in the resistant cultivars. A similar observation is made for, at one side, crosses 9
and 11 in which respectively parents R3 and R5 are crossed with susceptible plant S1 all giving a
duplex segregation. At the other hand, a simplex segregation was observed in crosses 14 and 16 in
which the same resistant cultivars were crossed with susceptible cultivar S2. Hence, in these crosses,
146
Segregation of resistance to P. horiana
those with susceptible cultivar S2 show a substantially higher ratio of susceptible progeny than those
with susceptible cultivar S1, but there is no easy explanation for these results.
The apparent discrepancies observed when some of the resistant cultivars were crossed with one or
the other susceptible cultivar is possibly due to Ph Mixed used for the screening of the parents and
the progeny of the crosses. The single pustule isolates that are included in Ph Mixed show a
moderate virulence, but they belong to different European pathotypes with similar infection profiles
(De Backer et al. 2011). As a result it is not clear whether the resistance reaction observed in the
progeny is due to a single gene conferring resistance to all pathotypes or to multiple, pathotypespecific, linked resistance genes that segregate as one unit. Also a combination of these two
hypotheses, with a gene conferring broad spectrum resistance supplemented with pathotype-specific
resistance genes, is possible. Since the expression of more general resistance genes is epistatic over
pathotype specific resistance genes, the pathotype specific resistance genes remain undetected
when they are inherited as a single segregation unit (Crute and Pink 1996). Cultivars that are
resistant must be able to recognize all isolates included in Ph Mixed and as a consequence must
possibly carry more than one resistance gene. If more than one resistance gene is needed to remain
resistant to the different isolates included in Ph Mixed, it is possible that one of the susceptible
parent plants carries a resistance gene to one or some of the isolates included in Ph Mixed without it
was noticed during the resistance screening. The presence of such unrecognized resistance gene in
susceptible parent S2 can result in a shift towards less resistant progeny in the crosses with this
parent compared to susceptible parent S1. Nevertheless, when more than one resistance gene is
needed to confer resistance to Ph Mixed, the differences in expected segregations of crosses
between resistant parents and susceptible parents without resistance gene and susceptible parents
with multiple copies of unrecognized resistance genes is too low to explain the shift from triplex to
duplex situations or from duplex to simplex situations.
On the other hand it is known that, besides dominant resistance genes conferring race specific
resistance, there are single recessive resistance loci that can modulate resistance to a wide range of
pathogens in several plant species (Panstruga 2003; Kiraly et al. 2007) such as the Mlo gene
conferring resistance to powdery mildew (Erysiphe graminis) in barley (Hordeum vulgare) (Büschges
et al. 1997). Other genes can be up-regulated by the pathogen upon compatible interaction to
suppress defense pathways and cell death. For the flax gene fis1 this up regulation has been
described after infection with the rust Melampsora lini and several homologues of this rust-induced
gene are present in many plants (Ayliffe et al. 2002). The presence of a modulator system can explain
the deviant segregation between crosses with the same resistant parent, but with different
susceptible parents. A modulator in susceptible plant S2 possibly alters the susceptibility of the
147
Chapter 6
progeny resulting in a higher amount of susceptible progeny. The presence of a type of modulator in
resistant cultivar R4 can also explain the high proportion of susceptible progeny plants observed in
crosses 10 and 15, in which more susceptible than resistant progeny plants were observed (Table
6.3). A proportion of more than 50% susceptible progeny cannot be explained by a resistance system
in which only dominant resistance genes are involved, although with a modulator gene a higher
number of susceptible progeny can be explained. However, the observed decrease of resistant
progeny in crosses with parent S2 compared to the crosses with parent S1 cannot be explained by
the presence of a negative modulator gene.
Since the observed segregations, as one should expect in the case of a duplex or triplex segregation
of a dominant resistance gene, cannot be explained by the presence of an unrecognized resistance
gene or by the activity of a negative modulator, the segregations in the crosses with the susceptible
cultivar S1 are probably the best to determine the number of resistance genes. In that case, the
resistant cultivars R6, R7 and R9 involved in crosses 17, 18 and 20 respectively can be considered to
carry the resistance gene(s) in a simplex configuration while the resistant cultivar R8, involved in
cross 19, can be considered as a duplex. However, if we consider a modulator gene to be present in
susceptible parent S2, the observed segregation in cross 12 conflicts with the segregation in cross 7.
6.4.3 Segregation of resistance to different pathotypes in progeny 17
The observed segregations in the progeny of cross 17 indicate a simplex situation in the resistant
parent R6 for the resistance genes conferring resistance to the single pustule isolates as well as to Ph
Mixed (Table 6.5). Parent S1 is probably carrier of a single copy of the gene(s) conferring resistance
to isolates Ph 707 and Ph 801. The resistance segregation to the single pustule isolates as well as to
Ph Mixed clearly shows a differential segregation of resistance to different pathotypes, although a
correlation in the response to some pathotypes was observed (Table 6.4). The strong correlation
between the responses observed to isolates Ph 707 and Ph 801 were expected since those two
pathotypes can commonly infect certain key-cultivars, distinguishing them from other currently
known isolates (De Backer et al. 2011). The reaction to isolates Ph PD 20 and Ph Japan 2 were highly
correlated as well, although they clearly belong to pathotypes with different interaction phenotype
profiles (De Backer et al. 2011).
The high correlation of the resistance to isolates Ph 707, Ph 801 and Ph Mixed at one hand and the
isolates Ph PD 20 and Ph Japan 2 on the other hand, suggests that resistance to P. horiana is
conditioned by at least two loci containing closely linked resistance genes. Since the correlation is
never 100%, it can be assumed that the resistance to those isolates is conditioned by different genes.
Due to recombination these genes can be separated resulting in differential interactions to those
148
Segregation of resistance to P. horiana
isolates in different progeny plants as also observed in the progeny of cross 17 and explain why the
correlation coefficient is not close to 100%. Nevertheless, a low correlation can also be the result of
experimental variation due to differences in environmental and host conditions between the
screenings with the different isolates. The fact that different linked genes, rather than allelic variants,
are involved is also indicated by the positive correlations that are found (Table 6.4). In the case of
allelism, negative correlations should be expected. As a result, in parent R6 resistance to isolates Ph
707, Ph 801 and Ph Mixed is organized in one cluster while resistance to isolates Ph PD 20 and Ph
Japan 2 is organized in another cluster. However, as explained for Ph Mixed, it cannot be determined
how the resistance to these isolates in parent R6 is exactly organized. Resistance can be due to a
common resistance gene that conditions resistance to Ph Mixed as well, linked pathotype-specific
genes or to resistance genes for Ph Mixed or pathotype specific resistance genes that are hidden by
more general resistance genes. In parent S1 it is clear that resistance to isolates Ph 707 and Ph 801 is
coded by specific resistance genes since this parent is resistant to both isolates but is susceptible to
Ph Mixed (Table 6.2). These specific resistance genes can be other resistance genes on a different
locus than those conferring resistance to both isolates in parent R6 or they can be just a part of the
linkage group that is present in parent R6. More crosses with cultivars that are well characterized
regarding to resistance to P. horiana are needed to study the mode of segregation more in detail as
described in other pathosystems. By studying the pathotype specific resistance in the Melampsora
lini – flax pathosystem, the presence of 5 resistance loci (K, L, M, N, P) has been demonstrated (Islam
and Shepherd 1991; Ellis et al. 2007). In this pathosystem, the L locus consist out of 13 allelic variants
whereas the M, N and P loci are more complex containing multiple tandemly arranged paralogues
resistance genes (Ravensdale et al. 2011).
The results obtained clearly illustrate that the inheritance of resistance to P. horiana in
chrysanthemum is much more complex than was concluded by the study performed by de Jong and
Rademaker (de Jong and Rademaker 1986). Probably preferential pairing dominates during meiosis
although random pairing cannot completely be excluded. This causes difficulties in estimating the
expected segregation ratios. In breeding programs, it is very important to know the resistance
characteristics of the parents and determine which pathotypes will be used in resistance screening.
The possible involvement of modulator genes is an important factor to take into account, but it is
difficult to determine the presence of these factors by means of bioassays and further research is
needed to clarify their mode of action. Resistance to different pathotypes is conditioned by several
dominant resistance genes that are located on at least two loci, resulting in different resistance
segregations for different pathotypes. Although we showed the relation of resistance genes for 4
different pathotypes, it is not known how resistance genes for other pathotypes are linked to each
149
Chapter 6
other. The presence of clusters of resistance genes can have important implications on the coevolution of host and pathogens in that they ease the development of new resistant-susceptibility
phenotypes by rearrangement of resistance genes (Friedman and Baker 2007). However this implies
a continuous follow up of new emerging pathotypes and resistant cultivars which have to be
incorporated in bioassays to improve selection for resistance. Molecular techniques will be a valuable
tool to further investigate the different parameters involved in resistance in this pathosystem. The
development of resistance gene specific markers could improve the selection of cultivars carrying
particular resistance genes and allow for an optimal resistance breeding in chrysanthemum.
150
Chapter 7:
Segregation of AFLP markers
in chrysanthemum
Commercial breeding of chrysanthemum
Segregation of AFLP markers
7.1 Introduction
DNA markers offer the possibility to study inheritance in crop species and to identify genome regions
involved in the determination of monogenic and polygenic traits. Once markers tightly linked to traits
of agronomical relevance have been identified, they can be further transformed into molecular tools
useful for marker-assisted selection (MAS) (Ribaut and Hoisington 1998). The implementation of MAS
approaches in plant breeding programs, generally involving large number of plants, can significantly
increase the efficiency and effectiveness of breeding programs (Collard et al. 2005). At certain steps
in the selection process, MAS may be easier than conventional phenotypic screenings saving time,
resources and effort. MAS can be carried out at seedling stage giving the opportunity to quickly
eliminate undesired genotypes. In addition, single plants can be selected based on their
homozygosity for the desired gene or QTLs, combination of multiple genes (gene pyramiding) or
absence of undesired genes linked to the gene of interest (linkage drag) (Collard et al. 2005; Collard
and Mackill 2008). Nevertheless, when MAS is taken into consideration, a cost-benefit analysis has to
be made for the development and implementation of the markers. MAS will be highly beneficial
when phenotypic evaluation is labor-intensive and/or difficult (Dreher et al. 2003) or when a trait is
influenced by QTLs for which it is difficult to determine which particular favorable genomic region is
carried by particular plants using solely phenotypic screening (Kumar 1999).
The choice of DNA marker technology to use in a particular study depends on the availability of
genetic information available for the species studied. DNA marker technologies that generate
dominant markers such as Random Amplified Polymorphic DNA (RAPD) or Amplified Fragment
Length Polymorphism (AFLP) (Williams et al. 1990; Vos et al. 1995) are often used when no genetic
framework with well characterized markers is available. The high reliability of AFLP makes it the
method of choice for initial marker generation (Mueller and Wolfenbarger 1999). Co-dominant
markers give the opportunity to discriminate between the different alleles of the genes or QTLs of
interest and are based on the detection of polymorphisms such as microsatellites or I (Collard and
Mackill 2008), but the development of marker assays for the generation of SSR and/or I marker
information can be laborious and costly. However, nowadays for numerous crop species large sets of
SSR and/or I markers are readily available.
These DNA-markers can be used to identify the genetic location of genes or QTLs of interest in
segregating pedigrees. Mapping in 153olyploidy species such as Chrysanthemum is known to be
more complicated than in diploid species since for a particular locus a higher number of genotypes
per locus (either DNA marker or gene) are possible than in diploids, and these genotypes can be
difficult to determine by phenotyping or banding pattern (Wu et al. 1992; Ripol et al. 1999). Also it is
153
Chapter 7
not always clear for every 154olyploidy species whether disomic or polysomic inheritance (or a
combination of both) is the rule. This is also the case for chrysanthemum in which conflicting reports
have been published regarding the segregation pattern of different traits (de Jong and Rademaker
1986; Langton 1989). Nevertheless, a recent study based on sequence-related amplified
polymorphism (SRAP) markers reported that the vast majority of loci followed a diploid-like
inheritance, with only a few following hexasomic inheritance (Zhang et al. 2011). Several computer
programs for linkage analysis and construction of linkage maps of polyploidy species are available (He
et al. 2001; Hackett and Luo 2003; Hackett et al. 2007), but they are not suitable for the analysis of
segregation patterns in hexaploid plants.
Literature reports on molecular analysis and the use of markers for genetic improvement of
chrysanthemum are limited due to the complexity of its genome with a high level of heterozygosity
(Anderson et al. 1992; Anderson and Ascher 2000). RAPD markers, inter-simple sequence repeat
(ISSR) and RFLP (restriction fragment length polymorphisms) markers have been used to determine
the genetic relationships among cultivars (Wolff and Peters-van Rijn 1993; Huang et al. 2000;
Chatterjee et al. 2006) or related species (Wolff and Peters-van Rijn 1993; Dai et al. 1998), and for
cultivar identification in the context of the protection of plant breeder’s rights (Wolff et al. 1995).
Markers can be used for the identification of chrysanthemum cultivars, since these are propagated
vegetatively, keeping the same DNA pattern even after many years of cultivation (Wolff et al. 1995;
Huang et al. 2000).
At the start of this project, no reports discribing the uses of molecular markers for the study of
inheritance in chrysanthemum were available. Recenty, two chrysanthemum linkage maps have been
published (Zhang et al. 2010; Zhang et al. 2011). In both cases the same segregating population of
142 F1 progeny plants was used. Zhang et al. (2010) generated 567 RAPD, ISSR and AFLP markers, of
wich 336 could be mapped. Zhang et al. (2011) generated 896 SRAP markers, 611 of which were used
for linkage map construction. Despite the lower number of markers included in the first study, the
number of linkage groups obtained was lower (44 and 33 linkage groups) (Zhang et al. 2010) than in
the second study (57 and 55 linkage groups) (Zhang et al. 2011). Nevertheless, these numbers are
still remarkably higher than the 27 linkage groups that can be expected based on the haploid
chromosome number. These studies also revealed the presence of hotspots of recombination in the
genome. Due to the higher number of markers included in the second study, the parental maps had a
size of 1912.8 cM and 1887.9 cM (covering ؆ 65% and ؆ 66% of the genome, repectively) (Zhang et
al. 2011), which is much higher than the coverage of 1034 cM and 1095 cM (covering ؆ 51% and ؆
55% of the genome length, repectively) reported by Zhang et al. (2010). In none of these studies, the
linkage maps of the parents had enough markers in common to combine them to a single integrated
154
Segregation of AFLP markers
map. The SRAP-based map was further used for QTL mapping of inflorescence-related traits. This led
to the identification of 12 QTL regions acting on three inflorescence traits, distributed over 11 linkage
groups. This demonstrates that this map can be used as framework for gene tagging.
The main objective of this chapter is to use AFLP markers to study the patterns of chromosome
segregation in the progeny of cross 17 (see Chapter 6). The segregation of AFLP markers, generated
using 10 primer combinations, in the F1 progeny was analyzed and the mode of inheritance in this
particular cross was determined. Since the parent cultivars have differential phenotypes for
resistance to P. horiana, we also checked whether the inheritance of disease resistance could be
linked to any of the identified molecular markers.
7.2 Material and methods
7.2.1 Mapping population
A total of 99 progeny plants of cross 17 (seed plant S1 x pollen plant R6), from which the resistance
to five different isolates of chrysanthemum white rust was determined as described in Chapter 6,
were used for AFLP marker analysis. Healthy, young leaf material of these plants was collected
before artificial infection tests, frozen in liquid nitrogen, lyophilized and stored under vacuum
conditions until DNA-extraction.
7.2.2 DNA extraction
About 100 mg of lyophilized plant material was grinded with a Retch MM200 ball mill (Retsch,
Germany) and genomic DNA was extracted using a cetyltrimethylammonium bromide (CTAB)
extraction procedure adapted from Doyle and Doyle (1990). Per sample, 1 ml extraction buffer (In
brief, 200 ʅl extraction buffer (1,4 M NaCl (Sigma), 100 mM Tris-HCL (pH 8.0)(UCB), 0,02 M Na-EDTA
(Invitrogen), 2% (vol/vol) ɴ-mercaptoethanol (Sigma), 1% polivinylpyrolidone (Sigma) and 2% CTAB
(Sigma)) was added. The reaction mixture was vortexed and incubated on a thermoshaker
(Eppendorf, Germany) at 55°C for 1 hour. After incubation 400 ђl of chloroform:isoamyl alcohol (24:1
(vol:vol)) was added to the sample and incubated for 2 minutes during which the samples were
gently shaked. Samples were centrifuged at 13000 rpm for 10 minutes at 4°C. The clear supernatant
was transferred to a new 1.5 ml reaction tube. The DNA was then precipitated with isopropanol (20°C) (Novolab, Belgium) and incubated for 1 hour in the freezer at -20°C. After incubation, samples
were centrifuged at 13000 rpm for 10 minutes at a temperature of 4°C and the supernatant was
removed carefully without disturbing the pellet. The DNA pellet was washed with 1 ml ice-cold (20°C) 76% ethanol and subsequently centrifuged at 13000 rpm for 10 minutes at a temperature of
155
Chapter 7
4°C. The ethanol was removed without disturbing the pellet which was finally dried at room
temperature, resuspended in 100 ђl of MilliQ water (Millipore, MA, USA) and stored at -20°C.
7.2.3 AFLP analysis
The AFLP analysis was performed as described by Vos et al. (1995) with the restriction enzymes EcoRI
and MseI. 250 to 300 ng of DNA was digested for 2 h at 37°C in a final volume of 25 ђl (10 mM TrisHCl (pH 7,5) (Invitrogen, CA, USA), 10 mM MgAc (VWR, PA, USA), 50 mM Kac (VWR, PA, USA), 2,5 U
MseI (Invitrogen, CA, USA) and 2,5 U EcoRI (Invitrogen, CA, USA). MseI and EcoRI specific adaptors
were ligated to the restriction fragments by adding 25 ђl adaptor ligation mix (5 pmol EcoRI-adaptor
(Invitrogen, CA, USA), 50 pmol MseI-adaptor (Invitrogen, CA, USA), 1 U T4-DNA-ligase (Invitrogen, CA,
USA), 1 mM ATP (Roche Diagnostics, Switzerland) in 10mMTris-Hac (Invitrogen, CA, USA) 7.5, 10 mM
MgAc (VWR, PA, USA) and 50 mM Kac (VWR, PA, USA) and incubating the mixture for two hours at
37°C. The resulting ligation mixtures were ten times diluted.
A preamplification step with primers specific to the EcoRI and MseI adaptors with two selective 3’
nucleotides was carried out. PCR reactions were performed in 50 ђl volumes containing 5 ђl of the
diluted ligation mixture and 45 ђl PCR mix (5 ђl 10x PCR buffer (Applied Biosystems, CA, USA), 0.2
mM of each dNTP (Invitrogen, CA, USA), 25 ng of each primer (Invitrogen, CA, USA) and 1,25 U Taqpolymerase (Applied Biosystems, CA, USA). The PCR amplifications were performed in a GeneAmp
9700 PE (Applied Biosystems, CA, USA) thermocycler using 25 cycles consisting of 30 sec at 94°C, 60
sec at 56°C and 60 sec at 72°C. Products from the preamplification PCR were 10-fold diluted to be
used in the selective amplification.
Selective amplification with primers carrying 4 selective nucleotides each was done in 20 ђl volumes
containing 3 ђl of diluted pre-amplification product, 2 ђl 10x PCR buffer (Applied Biosystems, CA,
USA), 0,2 mM of each dNTP (Invitrogen, CA, USA), 50 nM of fluorescent (HEX and FAM) labeled EcoRI
primers (Eurofins, Belgium), 250 mM MseI primer (Invitrogen, CA, USA) and 0,6 U Taq-polymerase
(Applied Biosystems, CA, USA). The following parameters were used during the selective
amplification: 1 cycle consisting of 2 min at 94°C, 30 sec at 65°C and 2 min at 72°C, followed by 9
cycles with 1 min at 94°C, 30 sec at 65°C decreasing to 56°C at a rate of 1°C per cycle and 2 min at
72°C and finally followed by 23 cycles of 1 min at 94°C, 30 sec at 56°C and 2 min at 72°C. From the
PCR product, 1 ђl was mixed with 13.5 ђl of Hi-Di™ Formamide (Applied Biosystems, CA, USA) and 0,5
ђl of GeneScan™-500 Rox Size Dye (Applied Biosystems, CA, USA). Denaturation of the products was
done by heating the samples for 3 min at 95°C. Capillary electrophoresis and fragment detection
were performed on an ABI Prism 3130xl Genetic Analyzer (Applied Biosystems, CA, USA).
156
Segregation of AFLP markers
7.2.4 Marker scoring
The AFLP fingerprints were scored using Genemapper 4.0 (Applied Biosystems, CA, USA) and only
fragments with a length ranging from 50 bp to 500 bp were considered. For the parent plants (S1 and
R6) two independent DNA extractions were carried out. These two extractions were handled as
independent samples, and in a first step, only markers for which consistent results were obtained in
the parents (either present or absent in both replicates) were taken into consideration. Three
categories of polymorphic makers were scored based on their presence or absence in the parent
plants: (i) markers present in S1 and absent in R6; (ii) markers present in R6 and absent in S1; (iii)
markers present in S1 and R6 and segregating in the progeny. In a second step, also markers with
ambiguous results in the two replicates of the parent plants were considered.
7.2.5 Data analysis
For each AFLP marker the number of progeny plants carrying the marker and the number of progeny
plants in which the marker was absent were counted. These observations were compared using chi²tests (ɲ = 0.05) to expected proportions when different genetic configurations in the parents were
assumed (nulliplex, simplex, duplex or triplex). In each case, the possibility of preferential or random
chromosome pairing was considered (Table 7.1).
For parent-specific markers, the following situations were tested: simplex (SP) x nulliplex (NP), duplex
(DP) x nulliplex (NP) and triplex (TP) x nulliplex (NP), considering both random and preferential
pairing. It should be noticed however that as a result of the small differences in expected segregation
ratios between crosses of DPxNP with preferential pairing (75%-25%) and random pairing (80%-20%),
it is difficult to distinguish between those situations. The same problem appears for TPxNP
segregations with expected segregation ratios of 95%-5% for random pairing and 87.5%-12.5% for
preferential pairing.
For markers that were present in both parents, the following situations were tested: SPxSP, SPxDP,
SPxTP, DPxDP, DPxTP and TPxTP, considering both random and preferential bivalent formation.
However, the expected segregations for crosses in which duplex or triplex parents are involved are
very close to each other ranging from 7:1 ratio for a SPxDP cross with preferential pairing to 399:1 for
a TPxTP cross with random pairing (Table 7.1). For markers for which ambiguous results were
obtained for one or both parents, the segregation results were compared to situations in which the
markers were present in one of the parents or in both.
Based on the proportion of simplex to multiplex markers, the type of bivalent formation in cross 17,
i.e. random or preferential, was determined as described by Silva (1993) and Kriegner et al. (2003). In
157
Chapter 7
this approach, the probability of transmission of an absent allele for duplex and triplex markers are
summed to calculate the expected frequencies for non-single dose polymorphisms under the
assumption of either autohexaploidy (1/5 + 1/20 = 0.25) or allohexaploidy (1/4 + 1/8 = 0.375). If
random pairing is the dominant mode of bivalent formation, 20% (duplex) + 5% (triplex) = 25% of all
segregating markers can be considered to be multiplex and 75% simplex. When preferential bivalent
formation is the rule, 25% (duplex) + 12.5% (multiplex) = 37.5% of all segregating markers can be
considered to be multiplex and 62.5% simplex. Only parent-specific markers and markers for which
ambiguous results were obtained in the parents were used in this analysis since it is difficult to
determine marker dose when the marker is present in both parents.
Table 7.1: Expected segregation ratios (present : absent) assuming either random or preferential bivalent formation in
hexaploids. SP = Simplex, DP = Duplex, TP = Triplex, NP = Nulliplex.
Mode of bivalent formation
Parental configuration
Random
Preferential
SP x NP
1:1
1:1
DP x NP
4:1
3:1
TP x NP
19:1
7:1
Present in both parents
SP x SP
3:1
3:1
SP x DP
9:1
7:1
SP x TP
39:1
15:1
DP x DP
24:1
15:1
a
DP x TP
99:1
31:1
a
a
TP x TP
399:1
63:1
a
The power to test those hypotheses is limited due to the number of plants analyzed
Marker category
Parent-specific
In a final step we also determined whether any of the AFLP markers was associated to resistance to
P. horiana. Kruskal-Wallis tests were performed to determine the association between the presence
or absence of AFLP markers and the resistance to the 5 isolates of P. horiana that are described in
chapter 6. Statistical analysis was performed with Statistica 9.0 (Statsoft, OK, USA). Joinmap 4.0
(Kyazma, the Netherlands) and MapQTL5 (Kyazma, the Netherlands) were used for linkage analysis
and to test for associations between marker alleles and resistance.
7.3 Results
7.3.1 Patterns of marker segregation in cross 17
The 10 AFLP primer combinations used, rendered a total of 182 polymorphic markers (Table 7.2). A
total of 77 markers (42.3%) were specific for one of the parents (40 markers present in S1 and 37
markers present in R6). 70 markers (38.5%) were present in both parents and segregated in the
progeny. 70 of the markers showing segregation in the progeny were present in both parents. Finally,
35 markers (19.2%) showed ambiguous results in the replicate samples of at least one of the parents.
158
Segregation of AFLP markers
Table 7.2: Overview of the number of AFLP-markers that could be scored for the different primer combinations and the
number of markers in each category.
Ambiguous
20
Marker category
Non-parent
specific
13
0
33
Allele 1.1 – Allele 1.33
11
7
5
23
Allele 2.1 – Allele 2.23
PC 3
7
4
2
13
Allele 3.1 – Allele 3.13
PC 4
7
11
3
21
Allele 4.1 – Allele 4.21
PC 5
8
15
7
30
Allele 5.1 – Allele 5.30
PC 6
3
8
8
19
Allele 6.1 – Allele 6.19
PC 7
3
2
9
14
Allele 7.1 – Allele 7.14
PC 8
4
5
1
10
Allele 8.1 – Allele 8.10
PC 9
6
1
0
7
Allele 9.1 – Allele 9.7
PC 10
Markers per
category
8
4
0
12
Allele 10.1 – Allele 10.12
77
70
35
182
AFLP primer
combinations
Parent-specific
PC 1
PC 2
Markers per
primer
combination
Marker ID’s
Of the 77 parent-specific markers, 37 were only present in R6 and 40 were only present in S1 (Table
7.3). Five markers, 2 specific for parent R6 and 3 specific for parent S1, showed distorted
segregations and did not fit any of the segregation ratios tested. A total of 45 markers (58%) showed
a clear simplex segregation (1:1) with 26 and 19 markers specific for R6 and S1, respectively. In this
case, no distinction can be made between random and preferential pairing, as the expected
proportions of progeny plants carrying the marker is the same in both cases.
For the 27 multiplex markers, the exact segregation could not always be determined, as in many
cases the results fitted several of the hypotheses tested. Of the 9 multiplex markers that are specific
for parent R6, two showed a clear DPxNP segregation with preferential pairing whereas the
segregation of the seven remaining markers could be explained by a DPxNP segregation with either
random or preferential pairing. A clear multiplex segregation could also be observed for 6 markers
specific for parent S1, with two markers showing DPxNP segregation with preferential pairing, two
markers showing TPxNP segregation with preferential pairing and two markers showing TPxNP
segregation with random pairing. The segregation of the 12 remaining multiplex markers (all specific
for parent S1) could be explained by two or three of the segregation ratios tested (Table 7.3).
159
Chapter 7
Table 7.3: Different types of segregation observed for markers that were present in only one of the parents. For each
parental marker type, the total number of markers, as well as the number of markers with distorted, simplex or multiplex
segregation are shown. For the multiplex markers, the number of times that the observed segregation ratios can be
explained by the different expected segregation ratios (P-chi² > 0.05) is given. NP =Nulliplex, DP = Duplex, TP = Triplex, Dist.
= Distorted segregation. Rand. = Random chromosome pairing, Pref. = Preferential chromosome pairing.
Parental
Total number
Dist.
marker type of markers
S1:R6
0:1
37
2
Multiplex
Simplex
Rand.
1:1
Total
26
9
Per ratio DPxNP (4:1)
2
7
1:0
40
3
19
18
Pref.
TPxNP
(19:1)
a
DPxNP (3:1) TPxNP (7:1)
X
X
X
2
X
2
X
2
a
X
a
X
5
3
Total
a
77
5
45
X
2
a
2
a
X
X
X
X
X
X
X
27
More than two ‘X’ in a row mean that all the markers fitted those possibilities.
Of the 70 polymorphic markers that were present in both parents, 19 fitted only one of the
segregation ratios tested. 17 of these fitted a SP x SP pattern of segregation, and it was therefore
impossible to distinguish between random or preferential pairing (in both cases a 3:1 ratio expected).
Two markers fitted a SP x DP segregation with preferential pairing.
For the remaining 51 markers several hypotheses could explain the observed segregation ratios.
These were markers that were present in at least 81 of the 99 progeny plants studied (Table 7.4). In
that case, the number of progeny plants analyzed in this study does not provide sufficient power to
distinguish between the different situations. Except for a segregation ratio of 83:16 (present :
absent), only fitting a SP x DP segregation with preferential pairing.
160
Segregation of AFLP markers
Table 7.4: Different types of segregation observed for markers that were present in both of the parents of the cross. The
total number of markers as well as the number of markers with simplex or multiplex segregation are mentioned. For the
multiplex markers, the number of time that the observed segregation ratios can be explained by the different expected
segregation ratios (P-chi² > 0,05) are given. SP = Simplex, DP = Duplex, TP = Triplex,
Random chromosome pairing
Total number of
markers
Per
ratio
SP x SP
3:1
70
17
X
SP x DP SP x TP
39:1
9:1
Preferential chromosome pairing
DP x
DP x TP TP x TP SP x DP SP x TP
DP
a
a
99:1 399:1
15:1
7:1
24:1
2
2
DP x
DP
15:1
X
b
X
X
b
X
b
X
b
X
8
5
2
b
2
X
b
X
4
X
X
X
b
1
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
11
X
X
X
b
X
X
X
b
9
b
7
DP x TP TP x TP
a
31:1 63:1
X
X
X
X
X
X
X
X
X
a
The power to test those hypotheses is limited due to the number of plants analyzed
b
More than two ‘X’ in a row mean that all the markers fitted those possibilities.
Of the 35 markers that showed ambiguous results in the replicate samples of the parent plants, 5
displayed distorted segregation, and a SPxNP segregation was observed for 19 of them (Table 7.5).
The segregation of 11 markers could be explained by a multiplex situation in one of the parents and
nulliplex in the other parent. Two markers fitted a DPxNP segregation with preferential pairing. One
and three markers fitted a TPxNP segregation with random and preferential pairing, respectively. For
the remaining markers we could not distinguish between random or preferential pairing, with three
markers segregating as DPxNP, one marker as a TPxNP and one marker fitting three of the tested
hypotheses. Since the same or highly similar segregation ratios are expected in SPxSP and SPxDP
crosses, they can also explain the observed segregations of some of the markers that appear to be
multiplex considering a cross with a NP (data not shown).
161
Chapter 7
Table 7.5: Different types of segregation observed for markers giving ambiguous results in the parents of the cross. The
total number of markers as well as the number of markers with distorted, simplex or multiplex segregation are mentioned.
For the multiplex markers, the number of time that the observed segregation ratios can be explained by the different
expected segregation ratios (P-chi² > 0,05) are given. DP = Duplex, TP = Triplex, Dist. = Distorted segregation. Rand. =
Random chromosome pairing, Pref. = Preferential chromosome pairing.
Total number of
markers
35
Dist.
5
Multiplex
Simplex
Rand.
1:1
Total
19
11
Per ratio
DPxNP (4:1)
Pref.
TPxNP
(19:1)
2
DPxNP (3:1) TPxNP (7:1)
X
1
X
3
a
3
a
1
a
1
a
X
X
X
X
X
X
X
X
More than two ‘X’ in a row mean that all the markers fitted those possibilities.
7.3.2 Mode of chromosome segregation in cross 17
Based on the comparison between the proportion simplex and multiplex (duplex and triplex) markers
that segregated as expected with random and preferential bivalent formation, the mode of
inheritance in chrysanthemum was assessed (Table 7.6). Among the parent-specific markers, 45
simplex and 27 multiplex markers were observed (Table 7.3). When the results of the markers with
ambiguous results in the parents are added (Table 7.5), the segregation of 64 markers fitted a
simplex pattern, while 38 fitted a multiplex pattern. The observed ratio of simplex to duplex markers
highly supports a preferential pairing hypothesis. The inclusion of ambiguous markers in this analysis
did not affect the conclusions.
Table 7.6: Chi² goodness-of-fit analysis at a 0.05 significance level for mode of bivalent formation in chrysanthemum based
on the observed number of simplex versus multiplex markers. Observed numbers were based on the segregations observed
in the parent specific marker category only or including the segregations observed in the ambiguous marker category
(between brackets).
Expected markers
Random pairing
Preferential pairing
Marker type
Number
%
Number
%
54
45
Simplex
75%
62.5%
(76.5)
(63.75)
18
27
Multiplex
25%
37.5%
(25.5)
(38.25)
72
72
Total
100%
100%
(102)
(102)
a
0.014
1
Chi² P-value
a
(0.004)
(0.959 )
a
No significant difference at the 0.05 significance level for H0: observed ratio = expected ratio.
Observed markers
Number
%
45
62.5%
(64)
(62.75%)
27
37.5%
(38)
(37.25%)
72
100 %
(102)
(100 %)
162
Segregation of AFLP markers
7.3.3 Association AFLP-marker allele / resistance
No linkage programs are available to determine linkage relationships between markers (or to test for
association between markers and traits) with the complicated segregation patterns that are expected
in hexaploid organisms when more than one copy of the marker is carried by the parents. Therefore,
we only used simplex markers, and a marker-by-marker analysis (Kruskal-Wallis tests) to test for
eventual associations with P. horiana disease resistance scores in the progeny of cross 17. This
approach is justified, as in this cross preferential pairing seems to be the norm (see previous section).
A significant (p<0.05) association with resistance to at least one P. horiana isolate was observed for
nine parent-specific AFLP-markers (Table 7.7). For six of these markers, the parental configuration
was 0:1 (S1:R6) and for three markers 1:0. The highest level of significance was observed between
presence of marker Allele 1.34 and resistance to isolates Ph Mixed (p<0.005), Ph 707 (p<0.05) and Ph
801 (p<0.05). Also for marker Allele 8.5 strong association was observed with resistance to isolates
Ph PD 20 and Ph Japan 2. The association between non-parent specific markers and resistance was
significant for five markers. For three of those markers a highly significant association (p<0.005) was
observed with resistance to one to three isolates. Marker Allele 1.42 is associated with resistance to
Ph 801 (p<0.0005), Ph 707 (p<0.001), Ph Mixed (p<0.005) and Ph PD 20 (p<0.05). Marker Allele 10.13
and Allele 5.18 are associated with resistance to Ph Mixed (p<0.005) and with resistance to Ph Japan
2 (p<0.005), Ph PD 20 (p<0.05) and Ph 707 (p<0.1). Only one ambiguous marker, segregating as a
SPxNP in the parents, was associated with resistance to isolates Ph 707 (P<0.05) and Ph 801 (p<0.1).
For linkage analysis, the AFLP markers were separated in two sets. Each set contained the testcross
markers segregating from one parent and the intercross markers present in both parents, i.e. 1:0 +
1:1 for parent S1, and 0:1 + 1:1 for parent R6. Linkage analysis was then carried out according to the
double pseudo-testcross mapping strategy (Grattapaglia and Sederoff 1994). A strong linkage (LODscore = 18.99) was found for the markers 1.31, 1.36 and 2.22, present in parent R6. The genetic
distance (estimated using the Kosambi function) between marker alleles 1.31 and 2.22 was 7 cM. The
distance between 2.22 and 1.36 was 6 cM. In parent R6, these markers seem to be in repulsion with
the locus that confers resistance to Ph Mixed, as the presence of this marker is associated with a
lower level of resistance in the progeny plants (Figure 7.1b, d and e). Similarly these alleles seem to
be in phase with the locus that confers resistance to Ph Japan 2 and Ph PD 20 (Figure 7.1b, d and e).
163
Chapter 7
Table 7.7: AFLP markers from which the segregation is associated with the resistance segregation of least one of the
pathotypes. For each marker the significance level of the Kruskal-Wallis test statistic is given. Only markers with a
significance level of p<0.05 for at least one isolate are shown. Significance levels: *=p<0.1; **=p<0.05; ***=p<0.01;
****=p<0.005; *****=p<0.001; ******=p<0.0005. Isolate codes between brackets correspond to the codes used in Chapter
2 and Chapter 4.
Marker ID
Parental
marker type
S1:R6
Ph Mixed
Allele 1.4
Allele 1.31
Allele 1.34
Allele 1.36
Allele 2.22
Allele 5.19
Allele 5.30
Allele 8.5
Allele 8.8
0:1
0:1
0:1
0:1
0:1
1:0
1:0
0:1
1:0
**
**
****
**
*
**
-
Allele 1.42
Allele 5.1
Allele 5.18
Allele 10.4
Allele 10.13
1:1
1:1
1:1
1:1
1:1
****
****
0:1 / 1:0
-
Ph PD 20
Ph Japan 2
Ph 707
(NL3)
(JP1)
(BE5)
Parent-specific markers (simplex x nullliplex)
*
**
-
*
*
**
***
-
Ph 801
(BE6)
**
**
**
**
**
*
Non-parent specific markers (simplex x simplex)
**
**
**
-
**
****
**
-
*****
*
-
******
**
*
-
Ambiguous markers (simplex x nulliplex)
Allele 2.5
-
-
**
*
No strong linkages were found among other marker pairs. The parent-specific marker allele 1.4,
present in the resistant parent, seems to be in phase with the locus that confers resistance to isolate
Ph Mixed in parent R6 (Figure 7.1a). Marker alleles 1.34 and 8.5 on the other hand seem to be in
repulsion with the locus/loci that confer(s) resistance to isolates Ph Mixed, Ph 707 and Ph 801 and
isolates Ph PD 20 and Ph Japan 2, respectively.
Among the alleles present in parent S1, alleles 5.19 and 8.8 seem to be in phase with resistance to
respectively isolates Ph Mixed and Ph 801 and isolates Ph 707 and Ph 801 (Figure 7.2a and c). Marker
allele 5.30 is in repulsion with resistance to isolate Ph 707 in this parent (Figure 7.2b). From the non
parent-specific parents, the presence alleles 4.1, 5.18 and 10.4 is associated with increased
resistance, while the presence of marker allele 5.1, 5.18 and 10.4 is associated with increased
susceptibity (Figure 7.3).
164
Segregation of AFLP markers
Figure 7.1: Box-whisker plots for the AFLP markers present in the resistant parent R6 and for which a significant association
with resistance to at least one of the pathotypes tested was detected (Kruskal-Wallis, p<0.1). The line represents the mean
disease score of progeny plants with (1) and without (0) the marker. Boxes represent the standard error and whiskers
represent the 95% confidence interval.
Figure 7.2: Box-whisker plots for the AFLP markers present in the susceptible parent S1 and for which a significant
association with resistance to at least one of the pathotypes tested was detected (Kruskal-Wallis, p<0.1). The line
represents the mean disease score of progeny plants with (1) and without (0) the marker. Boxes represent the standard
error and whiskers represent the 95% confidence interval.
165
Chapter 7
Figure 7.3: Box-whisker plots for the AFLP markers present in both parents and for which a significant association with
resistance to at least one of the pathotypes tested was detected (Kruskal-Wallis, p<0.1). The line represents the mean
disease score of progeny plants with (1) and without (0) the marker. Boxes represent the standard error and whiskers
represent the 95% confidence interval.
7.4 Discussion
7.4.1 Patterns of marker segregation in cross 17
The few studies available on literature on the inheritance in chrysanthemum when this study was
initiated were based on the segregation of dominant traits and reported conflicting results on the
mode of bivalent formation (de Jong and Rademaker 1986; Langton 1989). Studies in which the
segregation of molecular markers, including AFLP markers, was used for the construction of a
preliminary linkage map were only published recently (Zhang et al. 2010; Zhang et al. 2011).
The number of markers we could score using 10 AFLP primer combinations averaged 18.2 markers
per primer combination, which is higher than the number of markers that is generally generated
using RAPDs (Martín et al. 2002; Lema-Ruminska et al. 2004; Chatterjee et al. 2006). However, it is
relatively low compared to the number of markers that could be obtained by the other studies in
chrysanthemum. In the study by Zhang et al. (2010) 353 markers were generated with 8 primer
combinations, or an average of 44 markers per primer combination, certainly taking into account
166
Segregation of AFLP markers
that in this study only reproducible simplex markers present in one or both parents were considered.
A possible explanation for this higher number of marker per primer combination is that the parents
used by Zhang et al. (2010) are more diverse than the parents used in our study, resulting in a higher
number of polymorphic markers. An alternative explanation for the lower number of fragments
obtained in our study is that four selective bases that were added to the AFLP primers compared to 3
selective bases in the study by Zhang et al. (2010).
The 77 parent-specific markers were evenly distributed over both parent plants. Their relatively high
number compared to the 70 markers that were not polymorphic between the parents points to a
high genetic differentiation between the cultivars used in the study. This variation is similar to the
variation between chrysanthemum cultivars that has been reported by studies based on RAPD
markers and can be expected for an outcrossing plant species (Wolff and Peters-van Rijn 1993;
Martín et al. 2002). 35 fragments were not reproducible in the parent plants, with a clear variation in
presence between the primer combinations that were used. Since the segregation in the progeny
could supply additional information about the segregation of the markers, we decided to include
these markers in the data analysis.
From the 182 segregating markers a total of 81 markers (44.5%) were present as a simplex marker or
so called single dose markers (Wu et al. 1992). This includes the 45 and 19 simplex markers from the
parent specific category and ambiguous category, respectively, as well as the 17 markers that appear
to be present as simplex markers in both parents. 91 markers (50%) were present as multiplex
markers in at least one of the parents. For two markers that were present in both parents the marker
dose in the parents could not be determined since there segregation could be explained by both
simplex and multiplex segregations. A total of 10 markers (5.5%) showed distorted segregation. In
comparison to other studies, the proportion of simplex markers (44.5%) was low compared to the
proportions of multiplex or distorted markers (55.5%). In the two mapping studies in chrysanthemum
using AFLP and SRAP markers, only markers segregating as single dose markers were considered and
multiplex markers were categorized as distorted (Zhang et al. 2010; Zhang et al. 2011). Nevertheless,
respectively 39,9% (Zhang et al. 2010) and 22,9% (Zhang et al. 2011) of the markers in those studies
were categorized as distorted which is substantially lower than the combination of the multiplex and
distorted markers in our study. Also in studies using AFLP to assess the segregation in 167olyploidy
species much higher amounts of simplex markers were observed. A study on the octaploid
strawberry generated up to 82% simplex markers and a very limited number of distorted markers
(3.2%) (Lerceteau-Köhler et al. 2003). In the hexaploid sweetpotato (Ipomoea batatas) 63.9 %
simplex markers could be obtained (Kriegner et al. 2003), but the number of markers showing
distorted segregation (12.9%) was similar to that observed in our study. 64 of the simplex markers
167
Chapter 7
(78%) were test-cross markers segregating in a 1:1 ratio and present in one of both parents. 45 could
be assigned to the parent plants, with 19 markers for the seed plant S1 and 26 markers for the pollen
plant R6, respectively, whereas the remaining 19 markers were in the category showing ambiguous
presence in the parent plants making it difficult to assign them to a particular parent plant. 17 from
the simplex markers (21%) were inter-cross markers that segregated in a 3:1 ratio and were present
in both parent plants. The proportions of test-cross and inter-cross markers we obtained slightly
differ from those described by Zhang et al. (2010) who found 60% of AFLP markers to be test-cross
and 40% to be inter-cross in chrysanthemum, although another study in chrysanthemum using SRAPmarkers provided 88% test-cross markers and 12% inter-cross markers (Zhang et al. 2011). Simplex
markers segregating in a 1:1 ratio offer a framework for genetic mapping in polyploids since they
segregate in a diploid way, while simplex marker segregating in a 3:1 ratio that are linked to 1:1
markers can be used to identify homology between parental linkage groups (Wu et al. 1992).
However, for the construction of preliminary maps between 100 and 200 markers are generally used
(Mohan et al. 1997). As a result, it was not possible to determine linkage groups due to the low
number of test-cross markers we obtained.
7.4.2 Mode of chromosome segregation in cross 17
To assess whether the mode of bivalent formation in cross 17 was random or preferential, the
observed segregations were tested by chi² analysis against the expected proportions that should be
observed according to different modes of bivalent formation. In the marker category of parent
specific markers, simplex, duplex and triplex markers could generally be distinguished from each
other. However, within duplex and triplex markers an overlap between random and preferential
bivalent formation was present (Table 7.3). In the non-parent specific category, simplex markers
could easily be distinguished from multiplex markers, except for two markers, but it was not possible
to determine the dose of the multiplex markers in the parents with markers showing segregations
that could be explained in up to seven different segregation scenarios (Table 7.4). The segregations
that were observed for the markers in the ambiguous marker category show a similar distribution in
simplex to multiplex markers as observed in the parent specific marker category, with a high amount
of simplex markers segregating in a 1:1 ratio (Table 7.5). Based on this observation, the assumption
was made that the markers in this category are parent specific markers, although one has to notice
that the segregation of some multiplex markers can be explained by expected segregation patterns
with markers in both parents as well. The overlapping segregations that observed in the multiplex
markers are the result of the small differences in the expected segregation ratios and can be
overcome by increasing the population and make it very difficult to assess the mode of bivalent
formation in polyploids by testing observed segregations to expected segregation ratios. However,
168
Segregation of AFLP markers
based on the ratio of simplex to multiplex markers the mode of bivalent formation can be predicted
as already described in other species such as the hexaploid Ipomoea batatas (Ukoskit and Thompson
1997; Kriegner et al. 2003) and the octaploid Saccharum spontaneum (Silva et al. 1993). In this
analysis the markers in the parent specific and ambiguous category were included since they allowed
a clear distinction between strict simplex and multiplex markers, whereas in the non-parent specific
category segregations between a simplex parent and multiplex parents were present making it
difficult to determine the exact number of simplex and multiplex markers in this category. The
proportions of simplex and multiplex markers observed in this study support the hypothesis of
preferential pairing. The simplex-multiplex ratios are very close to the expected ratios for
preferential pairing with a high goodness-of-fit. These data indicate preferential pairing, in
agreement with the findings reported by Zhang et al. (Zhang et al. 2010; Zhang et al. 2011). The type
of chromosome pairing can also be investigated based on the ratio of markers that are linked in
repulsion phase to the markers that are linked in coupling phase. A ratio of repulsion to coupling
linkage from 1:1 indicates allopolyploidy with preferential bivalent formation compared to a ratio of
0:1 for autopolyploids with random bivalent formation (Wu et al. 1992; Qu and Hancock 2001).
However, since our data are too limited to determine linkage, this analysis could not be done.
Due to the limited number of AFLP markers that is included in this study, it was impossible to carry
out a detailed linkage analysis of AFLP markers and resistance to P. horiana. Nevertheless, we tried to
estimate the association between marker presence and resistance to 5 different isolates of P.
horiana. A significant association between marker segregation and resistance to at least one isolate
was observed for 15 AFLP markers. Among the parent-specific markers present in the resistant
parent R6, three markers alleles (1.31, 1.36 and 2.22) were closely linked. However, the level of
significance of their association to the resistance for different pathotypes was relatively low (Table
7.7). The presence of these markers was in phase with resistance to isolate Ph Japan 2 and Ph PD 20,
but in repulsion with Ph Mixed. This is in conflict with the observations reported in Chapter 6, were
we found a low correlation in resistance segregations between Ph Mixed at one side and isolates Ph
Japan 2 and Ph PD 20 at the other side. A possible explanation could be that resistance to Ph Mixed is
conferred by several R genes as this isolate is actually a mixed culture of pure isolates. This is also
supported by the fact that marker allele 1.34, that is strongly associated with resistance to Ph Mixed,
Ph 707 and Ph 801, could not be linked to other markers suggesting that it belongs to a different
locus. This confirms the hypothesis that resistance to Ph Japan 2 and Ph PD 20 is organized in a
different cluster of genes than resistance to Ph 707 and Ph 801. Our data do not allow drawing
conclusions on the R genes conferring resistance to Ph Mixed. As resistance to P. horiana is described
as a dominant trait (de Jong and Rademaker 1986), association of markers segregating in the
169
Chapter 7
susceptible parent with resistance indicate the presence of a modulator gene rather than a R gene.
However, the marker effects identified here are not strong enough to confirm these hypotheses.
In conclusion, our data confirm that AFLP is a powerful tool to generate dominant markers in
chrysanthemum, but more markers need to be generated for a complete linkage analysis. A more
dense genetic linkage map of this population would allow determining the exact way of chromosome
segregation and tagging of genes conferring resistance to P. horiana. The different segregations of P.
horiana resistance observed in cross 17, demonstrate that this segregating population is appropriate
for mapping and identification of genomic regions associated with P. horiana resistance in this
species. Other segregating populations, as the one described by Zhang et al. (2010; 2011), are
probably less suitable for this purpose, as the parents might not be contrasting for resistance.
However, the main purpose of this research was to determine the mode of inheritance of AFLP
markers in cross 17, and the identification of AFLP-markers putatively associated with resistance. The
generation of a linkage map was out of the scope of this PhD.
170
Chapter 8:
General discussion
and future perspectives
Commercial production of chrysanthemum cuttings
General discussion
Chrysanthemum culture and breeding has a history of more than 2000 years and finds its origin in
Asia, from where it spread to western Europe and the United States over the last three centuries.
Nowadays chrysanthemum is one of the most important floricultural crops in the world and includes
the production of cut flowers, potted plants and garden chrysanthemums. Several pests and diseases
can have an important impact on the yield and value of this crop. One of the most important diseases
in chrysanthemum is white rust caused by the micro cyclic rust Puccinia horiana. It also originates in
Asia, where it was first observed in Japan at the end of the 19 th century (Hiratsuka 1957). It took until
1963 to spread to South-Africa and Western Europe (Baker 1967; Priest 1995). Currently, the
pathogen is present in most chrysanthemum producing regions in the world and is classified as a
quarantine pathogen due to its potentially very serious economic consequences (EPPO 2004). Under
favorable conditions, the life cycle takes approximately three weeks and includes teliospores and the
aerially dispersed basidiospores (Firman and Martin 1968). The conditions for the formation of these
spores have been identified but despite this knowledge, little is known about the initial inoculum
source in the case of serious disease outbreaks. As a consequence, the pathogen is mainly controlled
by regular preventive fungicide applications. The development of fungicide resistance (Cook 2001)
and decreased public acceptance of fungicide applications reduce the usefulness of this control
strategy. A more targeted use of fungicides is necessary, based on a better understanding of the
remaining activity of specific fungicides and the actual presence of inoculum in the field. The last
decade, breeding for resistance is gaining importance as an alternative and sustainable control
method. However, discrepancies between different studies, in which inconsistent resistance
characteristics were observed when using the same cultivars, as well as anecdotal reports of resistant
cultivars that are occasionally found symptomatic, suggests the presence of pathotypes. A better
understanding of the pathogenic variation and evolution within the P. horiana population and
advanced resistance breeding techniques are necessary for optimal resistance breeding.
173
Chapter 8
The main goal of this dissertation was to increase our knowledge of the pathogen and its host to help
solve the problems listed above and eventually help obtain sustainable control of pathogen.
Specifically, the objectives were the following.
x
x
x
x
To determine the phenotypic diversity of the pathogen, including the existence of
pathotypes and the extent of fungicide resistance
To determine the genotypic diversity of the pathogen
To develop a technique for the detection of basidiospores in air samples and apply it to start
unravel its characteristics of aerial spread
To study the mode of segregation of resistance and AFLP markers in the host
To obtain these objectives, I optimized a high throughput bioassay, identified SNP markers for P.
horiana, characterized a collection of isolates in terms of fungicide resistance and pathotypes,
developed and applied a protocol for the detection of the pathogen in air samples and characterized
the progeny of crosses of chrysanthemum at the phenotypic and/or genotypic level. This research
revealed a considerable amount of phenotypic and genotypic diversity in a worldwide collection of
isolates, which, together with the results of the aerial detection assays resulted in new insights into
the migration, aerial spread and fungicide control options of P. horiana. The information on the
pathological diversity of the pathogen together with the insights on the segregation of disease
resistance in chrysanthemum provides a basis for sustainable control of P. horiana. These aspects will
be discussed in this chapter.
8.1 The Puccinia horiana toolbox
8.1.1 High throughput bioassay
In order to investigate the presence of pathotypes of P. horiana in a worldwide collection of isolates,
a bioassay that allowed the simultaneous screening of more than 200 cuttings a week using single
pustule isolates had to be developed. This bioassay was adapted from the bioassay described by
Alaei et al (2009a) and optimized to allow the screening of larger cuttings. Cuttings were inoculated
using single pustule isolates in closed plastic containers which allowed control of factors critical for a
successful inoculation including constant temperature, humidity, leaf wetness and light conditions. In
combination with the use of at least three replicates, the bioassay was very robust. The scoring
method we developed was accurate for the evaluation of gene-for-gene recognition events and is
relatively simple compared to previously described scoring scales or disease indexes (Yamaguchi
1981; Takatsu et al. 2000; Alaei 2008).
174
General discussion
We used this bioassay for the identification and characterization of pathotypes in P. horiana as
described in Chapter 2 and to study the segregation of resistance in the progeny of different crosses
of susceptible and resistant chrysanthemums in Chapter 6. Furthermore, this bioassay was used to
generate inoculum of basidiospores for the inoculation experiments in Chapter 3 and to obtain
relatively large numbers of basidiospores for DNA extraction in Chapters 4 and 5. The robustness and
relatively easy scoring makes the bioassay suitable for high throughput resistance screening of
candidate cultivars in resistance breeding programs. This bioassay can also be used to study the
effect of recombination on the formation of new pathotypes by inoculating a differential set of
cultivars with a mix of two isolates which belong to different pathotypes. Rare pustules that occur on
cultivars that are resistant to both pathotypes may be an indication that a new pathotype developed
due to recombination between the original isolates. Care has to be taken that inoculum which is the
result of mixed inoculations is kept in strict containment to avoid the spread of new pathotypes.
The inoculation method and scoring scale was adjusted for the screening of isolates for fungicide
resistance (Chapter 3). This modified assay was conducted in 1 liter recipients, as also described by
Alaei et al. (2009a). It also proved to be robust and allowed simultaneous evaluation of resistance to
a set of fungicides. However, due to the long incubation time required for the analysis and the need
for uniform inoculation, this assay is not suitable for use in combination with a warning system for
guided fungicide applications.
8.1.2 SNP markers for the genotyping of P. horiana isolates
The study of genetic diversity requires relatively large amounts of pure gDNA with a minimal amount
of contaminating DNA to obtain molecular markers. However, for obligate plant parasites such as P.
horiana this is very difficult. It was achieved via controlled harvesting of large numbers of
basidiospores (Alaei 2008), the only type of spore of this fungus that is not attached to plant, in
combination with the complexity-reducing molecular technique referred to as CRoPS™ (van Orsouw
et al. 2007). Using this technology, molecular markers could be obtained using pure gDNA from a
limited number of isolates. CRoPS™ combines AFLP with next generation DNA sequencing to identify
polymorphisms. P. horiana-specific PCR primers were developed for a total of 33 polymorphisms in
25 AFLP fragments, including 32 SNPs and 1 SSR. These can now be used for the genotyping of
isolates without the need to harvest basidiospores (Chapter 4). In combination with a whole genome
amplification step the specific SNP markers could be scored using only a few pustules, even in the
presence of contaminating plant DNA. As this allows the application of the technique without the
need of a labor-intensive propagation of the pathogen and appropriate biosafety facilities, these
markers can be easily applied during quarantine screenings. For more detailed genotyping of isolates,
175
Chapter 8
the resolution can still be increased by including additional markers in the analysis. This is shown by
the fact that isolates that clearly represent different pathotypes, and as a consequence have
different genotypes, appear to be clonal based on the set of markers we developed. Additional
markers can be determined using the CRoPS™ protocol with other AFLP primers or via whole genome
sequencing. Markers developed for other rust fungi showing cross-amplification in other Puccinia
species can be used (Dracatos et al. 2006; Keiper et al. 2006). It has not been determined whether
the markers we developed show a similar cross-amplification with other related rust species. If so,
they can be very helpful in genotyping related rust species. As species specificity of the new markers
has not yet been determined, species specific primers as developed by Alaei et al. (2009b) remain
crucial for the detection and identification of P. horiana, especially as they target the multi-copy
rDNA ITS region, enhancing their sensitivity.
8.1.3 Molecular detection of airborne spores
To detect and quantify airborne P. horiana basidiospores in the field, protocols for the sampling and
DNA extraction were developed (Chapter 5). DNA extraction using the Qiagen DNeasy Plant Mini kit
allowed reliable and robust detection of spores in samples taken with two types of spore traps
(Burkard and Ionic) when the extraction was preceded by an intensive bead beating step. DNA
quantification was performed using a SYBR® green real-time PCR assay as developed by Alaei et al
(2009b) and resulted in a LOQ of 10 spores and a LOD of 1 spore for tape pieces from either spore
trap. However, when the spore traps were used in field trials, the Burkard spore trap showed a
remarkably higher performance than the Ionic spore trap. Together with the possibility of
independent sampling for longer periods (up to a week) without sample saturation, this spore trap is
most suited for the study of the epidemiology and biology of the pathogen. In the future this type of
spore tap, or a type of spore trap using a similar capture matrix (e.g. rotating arm samplers), can be
applied in a warning network for P. horiana. It also has to be determined whether the molecular
markers that were developed for the genotyping of P. horiana can be used to determine the
genotype of the pathogen in airsamples. However, the presence of other rust species in the
airsamples, possibly showing cross-amplification with the molecular markers we developed has to be
taken into consideration. Also, in case more than one pathotype is present in the sample, it will be
difficult to determine which marker belongs to what isolate. The sampling and DNA extraction
procedures we propose are most likely suitable for the detection of other fungal spores, as long as a
real-time PCR assay is available for the species of interest.
176
General discussion
8.1.4 Phenotypic and genetic characterization of an outcrossing population in
chrysanthemum
The number of studies describing the segregation of phenotypic traits in chrysanthemum are limited
and their conclusions on the mode of chromosome pairing are conflicting (de Jong and Rademaker
1986; Langton 1989). Recently, two articles were published describing a linkage map derived from a
single segregating F1 population (Zhang et al. 2010; Zhang et al. 2011), but in these studies no
reference was made to resistance against P. horiana. The segregating population we characterized
using 5 P. horiana isolates, representing 4 pathotypes (Chapter 6), offers the opportunity to map
resistance genes to these pathotypes. The genetic characterization (Chapter 7) of this population
based on a reduced set of AFLP markers is too limited for a complete linkage analysis, although it can
be used as a backbone for more detailed mapping in the future. This requires the generation of
additional progenies of this cross which would have to be screened with an extended set of AFLP
primers in order to construct a dense genetic map on which resistance genes can be mapped.
8.2 Puccinia horiana as a diverse globetrotter
Although P. horiana is considered an asexual micro cyclic rust, a surprisingly high diversity in
pathotypes and genotypes was observed. The pathotype study presented in this dissertation is the
first in which a comprehensive collection of isolates was tested on a large set of differential cultivars.
A total of 26 isolates (Chapters 2 and 4) originating from four continents and six different sampling
years were tested on a set of 36 cultivars. All isolates represented a different pathotype, illustrating
that pathotype diversity is present between different continents as well as within a continent or
geographic region. The complexity of this pathosystem is further confirmed by the minimum number
of seven resistance genes and corresponding avirulence genes necessary to explain the observed
interaction phenotype profiles.
The SNP based genotypic diversity was lower than the pathogenic diversity, but still remarkable
(Chapter 4). Although isolates with identical genotypes were present in most regions where the
pathogen is considered endemic, a larger genotypic diversity was observed than expected for an
asexual organism. As clonal isolates are an indication for recent dispersal of the pathogen, we could
conclude that not only regional dispersal occurred but also international long distance dispersal. This
is most likely as the result of international trade of infected plants as the survival time of
basidiospores is too short for international dispersal (Firman and Martin 1968). This is in contrast to
other rust species that can travel long distances due to the robustness of urediospores (Brown and
Hovmoller 2002). The presence of isolates with identical genotypes sampled in different growing
seasons suggest that the pathogen is able to survive multiple years in a particular region.
177
Chapter 8
Despite the assumed asexual nature of this pathogen, the genetic diversity and marker patterns we
observed are almost certainly the result of recombination. Recombination events between isolates
probably also contributed to the high level of diversity in virulence we observed, resulting in a low
correlation between pathotype and genotype. However, for the Malaysian isolates, the number of
recombination events with isolates of other geographic regions seems to be limited, which resulted
in a unique pathotype cluster. The founder population of the isolates in this group probably lost an
Avr gene for which the number of chrysanthemum cultivars containing the corresponding resistance
gene is large, resulting in the high level of virulence among this group. Compared to the isolates in
the other clusters, which tend to have a more widespread distribution, the isolates in this cluster are
limited to Malaysia and some customs interceptions in the Netherlands. A decrease in the quarantine
measures for isolates not related to those in this cluster may be warranted. Although it has to be
investigated in more detail, the economic impact of quarantine measures will probably be higher
than the introduction of a P. horiana strain related to strains that are already present. In combination
with a faster SNP detection method, the SNP markers (Chapter 4) we developed for the genotyping
of P. horiana allow the distinction of the isolates of this more virulent cluster and can be very useful
for the identification of these isolates with a potentially higher impact.
8.3 Insights in the biology and epidemiology of Puccinia horiana
To optimize control measures and avoid outbreaks of P. horiana, a good understanding of the biology
and epidemiology of the disease is crucial. Since the appearance of the disease in western Europe in
the 1960’s the life cycle and optimal growing conditions of this micro cyclic rust were determined
(Firman and Martin 1968). Teliospores have a survival time of at least several weeks, but for
basidiospores a high relative humidity is an important factor for a successful dispersal and infection
as the basidiospores are very sensitive to desiccation with an infectious period of only 1 hour at even
90% RH (Firman and Martin 1968). Basidiospores are easily dispersed by wind over distances of at
least 700 meters (Zandvoort 1968), while dispersal over longer distances is suspected to occur mostly
by transport of infected plants. Nevertheless, little is known about the survival and initial inoculum
sources and the timing and frequency of basidiospore release.
Using the molecular detection protocol for basidiospores in air samples (Chapter 5), we were able to
show the presence of the pathogen in commercial chrysanthemum fields even when no symptoms of
P. horiana were reported. When symptoms were reported, high spore concentrations were detected
with an interval of several weeks corresponding to the life cycle of the pathogen. However, between
these peaks, a continuous presence at low spore concentrations was observed. These spore
concentrations follow a day-night cycle with an increase in the number of spores during the night,
178
General discussion
probably related to an increase of relative humidity and leaf wetness (dew) during the night. This
day-night cycle was also confirmed in a more detailed field trial. Also when no symptoms were
reported during a field trial, low spore concentrations were detected, mostly during a rain event or
the night following a rain event, further confirming the need of high relative humidity and leaf
wetness for optimal spore dispersal.
These insights form the basis for further research regarding the introduction and dispersal of P.
horiana in nurseries. One of the strategies would be to monitor the presence of the pathogen in
several nurseries year round to determine the first occurrence of the pathogen and its abundance
throughout the season. To limit the overall cost and number of samples that have to be processed,
one can consider the use of cheaper spore trapping devices (e.g. rotating arm samplers) that are only
switched on during periods with an increased chance for detection, such as nights or following rain
events. Furthermore, it is important to determine the minimal spore concentration that can be
considered as a threat for chrysanthemums. These epidemiological data could allow the
development of a warning system that can be used as an instrument for guided fungicide
applications. In combination with application of fungicides which are still effective (Chapter 3), this
will allow a more sustainable control and a decreased risk for the development of fungicide
resistance.
8.4 Breeding for resistance to Puccinia horiana
Control of P. horiana by preventive fungicide applications is hampered by an increasing number of
fungicide resistant strains and a loss of social acceptance of this practice. Also alternative disease
control strategies based on the reduction of the relative humidity in greenhouses are limited in their
applicability due to the high cost of heating needed in this process. As a consequence, the most
sustainable strategy for the control of P. horiana in the future will be the incorporation of resistance
genes in chrysanthemum cultivars via plant breeding. As evidence was provided for the international
migration of isolates and given the trans-continental production of planting material, a maximum
number of R genes needs to be incorporated through resistance breeding. This requires a screening
and selection procedure in which isolates infecting a maximum number of differential cultivars are
involved. The most reliable selection for particular R genes is obtained by using pathotypes
containing a single Avr gene during the screening process, but such isolates are difficult to obtain due
to the complexity of the pathosystem. As the incorporation of additional R genes in cultivars
increases the selection pressure for resistance-breaking pathotypes, a timely detection of these new
pathotypes is essential for an optimal R gene deployment. The installation of a “pathotype detection
network”, including cultivars with a limited number of R genes, but covering the spectrum of R genes,
179
Chapter 8
as well as some cultivars with complete resistance to all known pathotypes, would allow detection of
known and new pathotypes. An alternative approach for pathotype detection based on the
molecular characterization of specific isolates using Avr-gene-specific markers is not an option in the
short term as no such markers have been identified yet (Chapter 4). However, due to the decreasing
costs of next generation sequencing, full genome sequencing of the most important pathotypes of P.
horiana will allow identification and characterization of the most important Avr genes and study
other aspects of the host-pathogen interaction. Recently the genomes of the rust species Puccinia
striiformis f. sp. tritici (Cantu et al. 2011), Puccinia graminis f. sp. tritici and Melampsora laricispopulina (Duplessis et al. 2011) have been sequenced. These genomes can possibly serve as a
backbone for future research on other rust species, including P. horiana.
Nevertheless, breeding for resistance against this pathogen is complex as we showed that there are
at least seven R and corresponding Avr genes involved (Chapter 2). We were able to identify some
specific pathotypes, together infecting a maximal number of cultivars in our set of differentials. Using
this set of isolates in our high throughput bioassay, we are able to identify cultivars that are resistant
to all pathotypes presently known. However, the genes conferring resistance to these pathotypes are
located on at least two loci consisting of clusters of closely linked resistance genes (Chapter 6 and 7).
This complicates resistance breeding, as a different resistance segregation will be observed for
different pathotypes, resulting in a decreased number of progeny with all desired resistance genes.
An alternative is the development of complex gene-pyramiding crossing and selection schemes. As a
consequence, more candidate cultivars will have to be screened to obtain sufficient numbers of
cultivars for commercialization. In the long term, MAS can provide a solution for the high throughput
selection of resistant cultivars using markers linked to resistance genes (Collard and Mackill 2008),
although it will not reduce the number of candidate cultivars that have to be screened. The
characterized outcross population we described offers a platform to identify markers linked to
resistance, although a more detailed genetic study based on more molecular markers, and probably
the analysis of multiple segregating populations is necessary to obtain information on multiple loci
and sufficient resolution for MAS. Additional research on the identification and characterization of
the resistance genes in chrysanthemum could allow the generation of new cultivars with multiple
resistance genes via cisgene transgenic approaches (Jacobsen and Schouten 2009).
180
References
References
Anon. (1964). Chrysanthemumsorters modtagelighed over for hvid chrysanthemumrust (tidligere:
Japansk chrysanthemumrust) (Puccinia horiana)(Forelobig meddelelse). Meddr.St.ForsVirks PlKult.
737.
Anon. (2004). Diagnostic protocols for regulated pests:Puccinia horiana. Bulletin of European and
Mediterranean Plant Protection Organization , 34, 209-211.
Anon. (2009). Kengetallen 2008. Flora Holland.
Anon. (2011). Floriculture Crops: 2010 Summary. United States Department of Agriculture - National
Agricultural Statistics Service (USDA-NASS).
Abiko,K., Kishi,K., & Yoshiko,A. (1977). Occurrence of oxycarboxin tolerant isolates of Puccinia
horiana P. Hennings in Japan. Annals of the Phytopathological Society of Japan,43, 145-150.
Agrios,G.N. (2005). Plant Pathology. New York. Academic Press
Aime,M.C. (2006). Toward resolving family-level relationships in rust fungi (Uredinales).
Mycoscience,47, 112-122.
Aime,M.C., Matheny,P.B., Henk,D.A., Frieders,E.M., Nilsson,R.H., Piepenbring,M. et al. (2006). An
overview of the higher level classification of Pucciniomycotina based on combined analyses of
nuclear large and small subunit rDNA sequences. Mycologia,98, 896-905.
Alaei, H. (2008). Puccinia horiana Henn. on chrysanthemum: phylogeny, molecular detection, latent
development, and screening for disease resistance. ISBN 978-90-5989-242-2
Alaei,H., De Backer,M., Nuytinck,J., Maes,M., Hofte,M., & Heungens,K. (2009a). Phylogenetic
relationships of Puccinia horiana and other rust pathogens of Chrysanthemum x morifolium based on
rDNA ITS sequence analysis. Mycological Research,113, 668-683.
Alaei,H., Baeyen,S., Maes,M., Höfte,M., & Heungens,K. (2009b). Molecular detection of Puccinia
horiana in Chrysanthemum x morifolium through conventional and real-time PCR. Journal of
Microbiological Methods,76, 136-145.
Anderson,N.O. (1987). Reclassifications of the genus Chrysanthemum L. HortScience,22, 313.
Anderson,N.O. (2006). Flower Breeding and Genetics: Issues, Challenges and Opportunities for the
21th Century. Dordrecht, The Netherlands. Springer
Anderson,N.O. & Ascher,P.D. (2000). Fertility Changes in Inbred Families of Self-incompatible
Chrysanthemums (Dendranthema grandiflora). Journal of the American Society for Horticultural
Science,125, 619-625.
Anderson,N.O., Ascher,P.D., & Widmer,R.E. (1992). Inbreeding depression in garden and glasshouse
chrysanthemums: germination and survivorship. Euphytica,62, 155-169.
Anikster,Y. (1983). Binucleate basidiospores - A general rule in rust fungi. Transactions of the British
Mycological Society,81, 624-626.
Arthur,J.C. (1900). Chrysanthemum rust. Bulletin of the Indiana Agricultural Experiment Station,85,
143-150.
183
References
Ayliffe,M.A., Roberts,J.K., Mitchell,H.J., Zhang,R., Lawrence,G.J., Ellis,J.G., & Pryor,T.J. (2002). A
Plant Gene Up-Regulated at Rust Infection Sites. Plant Physiology,129, 169-180.
Backlund,J.E. & Szabo,L.J. (1993). Physical characteristics of the genome of the phytopathogenic
fungus Puccinia graminis. Current Genetics,24, 89-93.
Bain,J.M., Tavanti,A., Davidson,A.D., Jacobsen,M.D., Shaw,D., Gow,N.A.R., & Odds,F.C. (2007).
Multilocus Sequence Typing of the Pathogenic Fungus Aspergillus fumigatus. Journal of Clinical
Microbiology,45, 1469-1477.
Baker,J.J. (1967). Chrysanthemum White Rust in England and Wales 1963-66. Plant Pathology,16,
162-166.
Barbosa,M.A.G., Michereff,S.J., & Mora-aguilera,G. (2006). Elaboration and validation of
diagrammatic key to evaluate white rust severity in chrysanthemum. Summa phytopatológica,32, 5762.
Bardin,M., Suliman,M.E., Sage-palloix,A.M., Mohamed,Y.F., & Nicot,P.C. (2007). Inoculum
production and long-term conservation methods for cucurbits and tomato powdery mildews.
Mycological Research,111, 740-747.
Barnes,C.W. & Szabo,L.J. (2007). Detection and identification of four common rust pathogens of
cereals and grasses using real-time polymerase chain reaction. Phytopathology,97, 717-727.
Bartlett,D.W., Clough,J.M., Godwin,J.R., Hall,A.A., Hamer,M., & Parr-Dobrzanski,B. (2002). The
strobilurin fungicides. Pest Management Science,58, 649-662.
Bent,A.F. & Mackey,D. (2007). Elicitors, effectors, and R genes: The new paradigm and a lifetime
supply of questions. Annual Review of Phytopathology,45, 399-436.
Boerema,G.H. (1964). Puccinia horiana de Japanse roest ook in Nederlandse chrysantenbedrijven.
Vakblad van de Bloemist,19, 697.
Bolton,M.D. (2009). Primary Metabolism and Plant Defense-Fuel for the Fire. Molecular PlantMicrobe Interactions,22, 487-497.
Bolton,M.D., Kolmer,J.A., & Garvin,D.F. (2008). Wheat leaf rust caused by Puccinia triticina.
Molecular Plant Pathology,9, 563-575.
Bonde,M.R., Peterson,L., Rizvi,S.A., & Smilanick,J.L. (1995). Myclobutanil As A Curative Agent for
Chrysanthemum White Rust. Plant Disease,79, 500-505.
Borneman,J. & Hartin,R.J. (2000). PCR primers that amplify fungal rRNA genes from environmental
samples. Applied and Environmental Microbiology,66, 4356-4360.
Brent,K.J. & Hollomon,D.W. (1995). Fungicide Resistance in Crop Pathogens: How Can It Be
Managed? Fungicide Resistance Action Committee 2007
Brent,K.J. & Hollomon,D.W. (2007). Fungicide Resistance: The Assessment of Risk Fungicide
Resistance Action Committee 2007
Brown,J.K.M. & Hovmoller,M.S. (2002). Aerial Dispersal of Pathogens on the Global and Continental
Scales and Its Impact on Plant Disease. Science,297, 537-541.
184
References
Brugmans,B., van der Hulst,R.G.M., Visser,R.G.F., Lindhout,P., & van Eck,H.J. (2003). A new and
versatile method for the successful conversion of AFLP(tm) markers into simple single locus markers.
Nucleic Acids Research,31, e55.
Brumfield,R.T., Beerli,P., Nickerson,D.A., & Edwards,S.V. (2003). The utility of single nucleotide
polymorphisms in inferences of population history. Trends in Ecology & Evolution,18, 249-256.
Brummitt,D. (1997). Chrysanthemum Once Again. The Garden,122, 662-663.
Buller,A.H.R. (1941). The diploid cell and the diplodisation process in plants and animals, with special
reference to the higher fungi. Botanical Review,16, 355-431.
Büschges,R., Hollricher,K., Panstruga,R., Simons,G., Wolter,M., Frijters,A. et al. (1997). The Barley
Mlo Gene: A Novel Control Element of Plant Pathogen Resistance. Cell,88, 695-705.
Calderon,C., Ward,E., Freeman,J., Foster,S.J., & McCartney,H.A. (2002a). Detection of airborne
inoculum of Leptosphaeria maculans and Pyrenopeziza brassicae in oilseed rape crops by polymerase
chain reaction (PCR) assays. Plant Pathology,51, 303-310.
Calderon,C., Ward,E., Freeman,J., & McCartney,A. (2002b). Detection of airborne fungal spores
sampled by rotating-arm and Hirst-type spore traps using polymerase chain reaction assays. Journal
of Aerosol Science,33, 283-296.
Campbell,C.E. & Dimock,A.W. (1955). Temperature and the geographical distribution of
chrysanthemum rust. Phytopathology,45, 644-648.
Cantu,D., Govindarajulu,M., Kozik,A., Wang,M., Chen,X., Kojima,K.K., Jurka,J., Michelmore,R.W., &
Dubcovsky,J. (2011). Next Generation Sequencing Provides Rapid Access to the Genome of Puccinia
striiformis f. sp. tritici, the Causal Agent of Wheat Stripe Rust. Plos One,6, e24230.
Carlile,M.J. (1987). Genetic Exchange and Gene Flow: Their Promotion and Prevention. In
Rayner,A.D.M., Brasier,C.M., & Moore,D. (Eds), Evolutionary Biology of the Fungi. (pp. 203-214).
Cambridge.Cambridge University Press.
Catanzariti,A.M., Dodds,P.N., & Ellis,J.G. (2007). Avirulence proteins from haustoria-forming
pathogens. Fems Microbiology Letters,269, 181-188.
Catanzariti,A.M., Dodds,P.N., Lawrence,G.J., Ayliffe,M.A., & Ellis,J.G. (2006). Haustorially Expressed
Secreted Proteins from Flax Rust Are Highly Enriched for Avirulence Elicitors. The Plant Cell,18, 243256.
Chatterjee,J., Mandal,A.K.A., Ranade,S.A., da Silva,J.A.T., & Datta,S.K. (2006). Molecular
systematics in Chrysanthemum x grandiflorum (Ramat.) Kitamura. Scientia Horticulturae,110, 372378.
Collard,B., Jahufer,M., Brouwer,J., & Pang,E. (2005). An introduction to markers, quantitative trait
loci (QTL) mapping and marker-assisted selection for crop improvement: The basic concepts.
Euphytica,142, 169-196.
Collard,B.C.Y. & Mackill,D.J. (2008). Marker-assisted selection: an approach for precision plant
breeding in the twenty-first century. Philosophical Transactions of the Royal Society B: Biological
Sciences,363, 557-572.
185
References
Cook,R.T.A. (2001). First report in England of changes in the susceptibility of Puccinia horiana, the
cause of chrysanthemum white rust, to triazole and strobilurin fungicides. Plant Pathology,50, 792.
Coutin,R. & Grouet,D. (1983). La protection phytosanitaire du chrysantheme. Phytoma,351, 31-37.
Cox,C. (1997). Fungicide Factsheet Chlorothalonil. Journal of Pesticide Reform,17, 14-20.
Crook,B. & Sherwood-Higham,J.L. (1997). Sampling and assay of bioaerosols in the work
environment. Journal of Aerosol Science,28, 417-426.
Crute,I.R. & Pink,D.A.C. (1996). Genetics and Utilization of Pathogen Resistance in Plants. The Plant
Cell Online,8, 1747-1755.
Cummins,G.B. & Hiratsuka,Y. (2003). Illustrated Genera of Rust Fungi. St. Paul. APS Press
Dai,S., Chen,J., & Li,W. (1998). Application of RAPD analysis in the study on the origin of Chinese
cultivated Chrysanthemum. Acta Botanica Sinica,40, 1053-1059.
Davey,J.W. & Blaxter,M.L. (2010). RADSeq: next-generation population genetics. Briefings in
Functional Genomics,9, 416-423.
De Backer,M., Alaei,H., Van Bockstaele,E., Roldan-Ruiz,I., van der Lee,T., Maes,M., & Heungens,K.
(2011). Identification and characterization of pathotypes in Puccinia horiana, a rust pathogen of
Chrysanthemum x morifolium. European Journal of Plant Pathology,130, 1-14.
de Jong,J. & Rademaker,W. (1986). The Reaction of Chrysanthemum Cultivars to Puccinia horiana
and the Inheritance of Resistance. Euphytica,35, 945-952.
de Silva,H.N., Hall,A.J., Rikkerink,E., McNeilage,M.A., & Fraser,L.G. (2005). Estimation of allele
frequencies in polyploids under certain patterns of inheritance. Heredity,95, 327-334.
Di-Giovanni,F. (1998). A review of the sampling efficiency of rotating-arm impactors used in
aerobiological studies. Grana,37, 164-171.
Dickens,J. & Potter,R. (1983). Spraying for white rust. Grower,100, 35-37.
Dickens,J.S.W. (1968). The Resistance of Various Cultivars and Species of Chrysanthemum to White
Rust (Puccinia horiana Henn.). Plant Pathology,17, 19-22.
Dickens,J.S.W. (1971). Further Observations on the Resistance of Chrysanthemum Cultivars to White
Rust (Puccinia horiana Henn.). Plant Pathology,20, 27-28.
Dickens,J.S.W. (1978). Treatment of Chrysanthemum Stools Against White Rust (Puccinia horiana
Henn). Plant Pathology,27, 118-119.
Dickens,J.S.W. (1990). Studies on the chemical control of chrysanthemum white rust caused by
Puccinia horiana. Plant Pathology,39, 434-442.
Dickens,J.S.W. (1991). Evaluation of Some Newer Fungicides in Comparison with Propiconazole
against Chrysanthemum White Rust Puccinia-Horiana. Annals of Applied Biology,118 (Supplement),
32-33.
Dirkse,F.B. (1980). Control of chrysanthemum white rust. Vakblad voor de Bloemisterij,35, 30-31.
186
References
Dirkse,F.B., Dil,M., Linders,R., & Rietstra,I. (1982). Resistance in white rust (Puccinia horiana P.
Hennings) of chrysanthemum to oxycarboxin and benodanil in the Netherlands. Mededelingen
Faculteit Landbouwwetenschappen Rijksuniversiteit Gent,47, 793-800.
Dowrick,G.J. (1952). The chromosomes of Chrysanthemum, I: The species. Heredity,6, 365-375.
Dowrick,G.J. (1953). The chromosomes of Chrysanthemum, II: Garden Varieties. Heredity,7, 59-72.
Doyle,J.J. & Doyle,J.L. (1990). Isolation of plant DNA from fresh tissue. Focus,12, 13-15.
Dracatos,P.M., Dumsday,J.L., Olle,R.S., Cogan,N.O.I., Dobrowolski,M.P., Fujimori,M. et al. (2006).
Development and characterization of EST-SSR markers for the crown rust pathogen of ryegrass
(Puccinia coronata f.sp. lolii). Genome,49, 572-583.
Dreher,K., Khairallah,M., Ribaut,J.M., & Morris,M. (2003). Money matters (I): costs of field and
laboratory procedures associated with conventional and marker-assisted maize breeding at CIMMYT.
Molecular Breeding,11, 221-234.
Duplessis, S., Cuomo, C. A., Lin, Y. C., Aerts, A., Tisserant, E., Veneault-Fourrey, C. et al. (2011).
Obligate biotrophy features unraveled by the genomic analysis of rust fungi. Proceedings of the
National Academy of Sciences , 108, 9166-9171.
Eduard,W., Lacey,J., Karlsonn,K., Palmgren,U., Ström,G., & Blomquist,G. (1990). Evaluation of
Methods for Enumerating Microorganisms in Filter Samples from Highly Contaminated Occupational
Environments. American Industrial Hygiene Association Journal,51, 427-436.
Ekramoddoullah,A.K.M. & Taylor,D. (1997). Monoclonal antibodies to white pine blister rust fungus
Cronartium ribicola. Free Patents Online,United States Patent (5616470).
El Ghaouth,A., Arul,J., Asselin,A., & Benhamou,N. (1992). Antifungal activity of chitosan on postharvest pathogens: induction of morphological and cytological alterations in Rhizopus stolonifer.
Mycological Research,96, 769-779.
Ellegren,H. (2004). Microsatellites: simple sequences with complex evolution. Nat Rev Genet,5, 435445.
Ellis,J.G., Dodds,P.N., & Lawrence,G.J. (2007). Flax Rust Resistance Gene Specificity is Based on
Direct Resistance-Avirulence Protein Interactions. Annual Review of Phytopathology,45, 289-306.
EPPO (2004). Puccinia horiana. Bulletin of the European and Mediterranean Plant Protection
Organization (EPPO),34, 209-211.
Faleiro,F.G., Vinhadelli,W.S., Ragagnin,V.A., Correa,R.X., Moreira,M.A., & de Barros,E.G. (2000).
RAPD markers linked to a block of genes conferring rust resistance to the common bean. Genetics
and Molecular Biology,23, 399-402.
Farris,J.S. (1969). On the Cophenetic Correlation Coefficient. Systematic Biology,18, 279-285.
Fernandez-Ortuno,D., Tores,J.A., De Vicente,A., & Perez-Garcia,A. (2008). Mechanisms of resistance
to Qol fungicides in phytopathogenic fungi. International Microbiology,11, 1-9.
Firman,I.D. & Martin,P.H. (1968). White rust of chrysanthemums. Annals of Applied Biology,62, 429442.
187
References
Flor, H. H. (1956). The complementary genetic systems in flax and flax rust. Advances in Genetics , 8,
29-54.
Flor,H.H. (1971). Current Status of the Gene-For-Gene Concept. Annual Review of Phytopathology,9,
275-296.
Forche,A., Magee,P.T., Magee,B.B., & May,G. (2004). Genome-Wide
Polymorphism Map for Candida albicans. Eukaryotic Cell,3, 705-714.
Single-Nucleotide
Fournier,A., Widmer,F., & Enkerli,J. (2010). Development of a single-nucleotide polymorphism (SNP)
assay for genotyping of Pandora neoaphidis. Fungal Biology,114, 498-506.
Frederick,R.D., Snyder,C.L., Peterson,G.L., & Bonde,M.R. (2002). Polymerase chain reaction assays
for the detection and discrimination of the soybean rust pathogens Phakopsora pachyrhizi and P.
meibomiae. Phytopathology,92, 217-227.
Freeman,J., Ward,E., Calderon,C., & McCartney,A. (2002). A polymerase chain reaction (PCR) assay
for the detection of inoculum of Sclerotinia sclerotiorum. European Journal of Plant Pathology,108,
877-886.
Friedman,A.R. & Baker,B.J. (2007). The evolution of resistance genes in multi-protein plant
resistance systems. Current Opinion in Genetics & Development,17, 493-499.
Ganal,M.W., Altmann,T., & Röder,M.S. (2009). SNP identification in crop plants. Current Opinion in
Plant Biology,12, 211-217.
García-Velasco,R., Zavaleta-Mejía,E., Rojas-Martínez,R.I., Leyva-Mir,S.G., Kilpatrick Simpson,J., &
Fuentes-Dávila,G. (2005). Antagonismo de Cladosporium sp. Contra Puccinia horiana Henn. Causante
de la Roya blanca del crisantemo (Dendranthema grandiflora Tzvelev). Revista Mexicana de
Fitopatología,23, 79-86.
Gardes,M. & Bruns,T.D. (1993). ITS primers with enhanced specificity for Basidiomycetes: application
to the identification of mycorrhizae and rusts. Molecular Ecology,2, 113-118.
Gardner,D.E. & Hodges,Jr.C.S. (1989). The rust fungi (Uredinales) of Hawaii. Pacific Science, 43, 4155.
Gibson,C.M. (1904). Notes on infection experiments with various Uredineae. New Phytologist,3, 184191.
Gisi,U., Chin,K.M., Knapova,G., Küng Färber,R., Mohr,U., Parisi,S., Sierotzki,H., & Steinfeld,U.
(2000). Recent developments in elucidating modes of resistance to phenylamide, DMI and strobilurin
fungicides. Crop Protection,19, 863-872.
Gisi,U., Hermann,D., Ohl,L., & Steden,C. (1997). Sensitivity profiles of Mycosphaerella graminicola
and Phytophthora infestans populations to different classes of fungicides. Pesticide Science,51, 290298.
Gisi,U., Sierotzki,H., Cook,A., & McCaffery,A. (2002). Mechanisms influencing the evolution of
resistance to Qo inhibitor fungicides. Pest Management Science,58, 859-867.
Gjaerum, H. B. (1964). Krysantemum-hvitrust funnet i Norge. Saertv.Gartneryrket , 12-13.
188
References
Gorlach,J., Volrath,S., KnaufBeiter,G., Hengy,G., Beckhove,U., Kogel,K.H. et al. (1996).
Benzothiadiazole, a novel class of inducers of systemic acquired resistance, activates gene expression
and disease resistance in wheat. The Plant Cell,8, 629-643.
Gower,J.C. (1971). A General Coefficient of Similarity and Some of Its Properties. Biometrics,27, 857871.
Grasso,V., Palermo,S., Sierotzki,H., Garibaldi,A., & Gisi,U. (2006a). Cytochrome b gene structure and
consequences tor resistance to Qo inhibitor fungicides in plant pathogens. Pest Management
Science,62, 465-472.
Grasso,V., Sierotzki,H., Garibaldi,A., & Gisi,U. (2006b). Characterization of the cytochrome b gene
fragment of Puccinia species responsible for the binding site of QoI fungicides. Pesticide Biochemistry
and Physiology,84, 72-82.
Grattapaglia,D. & Sederoff,R. (1994). Genetic Linkage Maps of Eucalyptus grandis and Eucalyptus
urophylla Using a Pseudo-Testcross: Mapping Strategy and RAPD Markers. Genetics,137, 1121-1137.
Grouet,D., Montfort,F., & Leroux,P. (1981). Mise en evidence, en France, d'une souche de Puccinia
horiana resistante a l'oxycarboxine. Phytiatr (Phytopharm),30, 3-12.
Guerrand,J., Aljabal,M., Stapel,O., & Hallier,S. (2011). Control of chrysanthemum white rust. AFPP Quatrieme Conference International sur les Méthodes Alternatives en Protection des Cultures - 8, 9 et
10 mars 2011, 451-459.
Hackett,C.A. & Luo,Z.W. (2003). TetraploidMap: Construction of a Linkage Map in Autotetraploid
Species. Journal of Heredity,94, 358-359.
Hackett,C.A., Milne,I., Bradshaw,J.E., & Luo,Z. (2007). TetraploidMap for Windows: Linkage Map
Construction and QTL Mapping in Autotetraploid Species. Journal of Heredity,98, 727-729.
Hahn,W. (1989). Prüfung von chrysanthemensorten auf anfalligkeit gegen weissen rost, Puccinia
horiana Henn. Arch. Züchtungsforschung,19, 178-192.
Hammond-Kosack,K.E. & Jones,J.D.G. (1997). Plant disease resistance genes. Annual Review of Plant
Physiology and Plant Molecular Biology,48, 575-607.
Harada,Y., Iwama,T., & Fukuda,T. (1996). Life cycle of Puccinia chrysanthemi Roze, the rust fungus
on Chrysanthemum morifolium Ramat. Annals of the Phytopathological Society of Japan,62, 199-201.
Harmon,C., Byrne,L.J.M., Hershman,D.E., Schwartz,H., Pierzynski,J., & Putnam,M.l. (2007). Soybean
rust: detection and diagnosis in the US. National Plant Diagnostic Network Meeting 2007.
Hartley,M.J. & Williams,P.G. (1971). Morphological and cultural differences between races of
Puccinia graminis f. sp. tritici in axenic culture. Transactions of the British Mycological Society,57,
137-148.
He,Y., Xu,X., Tobutt,K.R., & Ridout,M.S. (2001). Polylink: to support two-point linkage analysis in
autotetraploids. Bioinformatics,17, 740-741.
Heath,M.C. (2000). Hypersensitive response-related death. Plant Molecular Biology,44, 321-334.
Hellmers,E. (1964). The control of Japanese chrysanthemum rust. Gartner Tidende,80, 350-352.
189
References
Hennings,P. (1901). Einige neue japanische Uredineen [Some new Japanese rusts]. Hedwigia,40, 2526.
Hiratsuka,N. (1957). Three species of chrysanthemum rust in Japan and its neighboring districts.
Sydowia,2, 34-44.
Hiratsuka,Y. & Sato,S. (1982). Morphology and Taxonomy of Rust Fungi. The Rust Fungi (pp. 135).Academic Press.
Hirst,J. (1952). An automatic volumetric spore trap. Annals of Applied Biology,39, 257-265.
Hirst,J., Stedman,O.J., & Hogg,W.H. (1967). Long-distance Spore Transport: Methods of
Measurement, Vertical Spore Profiles and the Detection of Immigrant Spores. Journal of General
Microbiology,48, 329-355.
Holden,A.N.G. & Smith,D. (1992). Effects of Cryopreservation Methods in Liquid-Nitrogen on
Viability of Puccinia abrupta Var Partheniicola Urediniospores. Mycological Research,96, 473-476.
Holland,P.M., Abramson,R.D., Watson,R., & Gelfand,D.H. (1991). Detection of specific polymerase
chain reaction product by utilizing the 5'-3' exonuclease activity of Thermus aquaticus DNA
polymerase. Proceedings of the National Academy of Sciences of the United States of America,88,
7276-7280.
Horst,R.K. & Nelson,P.E. (1997). Compendium of Chrysanthemum Diseases. St. Paul, USA. The
American Phytopathological Society Press
Huang,S.C., Tsai,C.C., & Sheu,C.S. (2000). Genetic analysis of Chrysanthemum hybrids based on
RAPD molecular markers. Botanical Bulletin of Academia Sinica,41, 257-262.
Huber,L. & Gillespie,T.J. (1992). Modeling leaf wetness in relation to plant disease epidemiology.
Annual review of plant phytopathology,30, 553-577.
Islam,M.R. & Shepherd,K.W. (1991). Present status of genetics of rust resistance in flax.
Euphytica,55, 255-267.
Jones,J.D.G. & Dangl,J.L. (2006). The plant immune system. Nature,444, 323-329.
Jorgensen,H.A. (1964). Japanese chrysanthemum rust, a dangerous disease discovered in Denmark.
Gartner Tidende,80, 235.
Kapooria,R.G. & Zadoks,J.C. (1973). Morphology and cytology of the promycelium and the
basidiospore of Puccinia horiana. European Journal of Plant Pathology,79, 236-242.
Katsuya,K., Kakishima,M., & Sato,S. (1978). Axenic culture of two rust fungi, Puccinia coronata f.sp.
avenae and P. recondita f. sp. tritici. Ann. Phytopathol. Soc. Jpn.,44, 606-611.
Keen,N.T. (1990). Gene-For-Gene Complementarity in Plant-Pathogen Interactions. Annual Review of
Genetics,24, 447-463.
Keiper,F.J., Haque,M.S., Hayden,M.J., & Park,R.F. (2006). Genetic Diversity in Australian Populations
of Puccinia graminis f. sp. avenae. Phytopathology,96, 96-104.
190
References
Kennedy,R., Wakeham,A.J., Byrne,K.G., Meyer,U.M., & Dewey,F.M. (2000). A New Method To
Monitor Airborne Inoculum of the Fungal Plant Pathogens Mycosphaerella brassicicola and Botrytis
cinerea. Applied and Environmental Microbiology,66, 2996-3003.
Kim,S., Nikolaeva,E., Olson,T., & Kang,S. (2011). Overwintering of Chrysanthemum white rust
caused by Puccinia horiana in Pennsylvania and challenges in its management. Phytopathology,101,
S91-S92.
Kiraly,L., Barna,B., & Kiraly,Z. (2007). Plant Resistance to Pathogen Infection: Forms and
Mechanisms of Innate and Acquired Resistance. Journal of Phytopathology,155, 385-396.
Klassen,P. (1992). White rust dogs mum growers. Greenhouse Grower,10, 39-42.
Kohno,M., Nishimura,T., Ishizaki,H., & Kunoh,H. (1974). Teliospore germination and nuclear
behaviors in Chrysanthemum white rust (Puccinia horiana P. Hennings). The Bulletin of the Faculty of
Agriculture,Mie University,47, 1-9.
Kohno,M., Nishimura,T., Ishizaki,H., & Kunoh,H. (1975). Nuclear behaviors during the process from
teliospore stage through sporidial stage in two short-cycled rusts, Kuehneola japonica and Puccinia
horiana. The Bulletin of the Faculty of Agriculture,Mie University, 21-29.
Kolmer,J.A. (2005). Tracking wheat rust on a continental scale. Current Opinion in Plant Biology,8,
441-449.
Korner,O. & Challa,H. (2003). Process-based humidity control regime for greenhouse crops.
Computers and Electronics in Agriculture,39, 173-192.
Kriegner,A., Cervantes,J.C., Burg,K., Mwanga,R.O.M., & Zhang,D. (2003). A genetic linkage map of
sweetpotato [Ipomoea batatas; (L.) Lam.] based on AFLP markers. Molecular Breeding,11, 169-185.
Kropp,B.R., Hansen,D.R., Wolf,P.G., Flint,K.M., & Thomson,S.V. (1997). A Study on the Phylogeny of
the Dyer's Woad Rust Fungus and Other Species of Puccinia from Crucifers. Phytopathology,87, 565571.
Kubista,M., Andrade,J.M., Bengtsson,M., Forootan,A., Jonak,J., Lind,K. et al. (2006). The real-time
polymerase chain reaction. Molecular Aspects of Medicine,27, 95-125.
Kumar,L.S. (1999). DNA markers in plant improvement: An overview. Biotechnology Advances,17,
143-182.
Kuparinen,A. (2006). Mechanistic models for wind dispersal. Trends in Plant Science,11, 296-301.
Lam,C.H. & Lim,T.K. (1993). Efficacy of hexaconazole for the control of white rust on chrysanthemum
and powdery mildew on roses. International Journal of Pest Management,39, 156-160.
Lamour,K., Habera,L., Snover-Clift,K., Stack,J., Pierzynski,J., Hammerschmidt,R. et al. (2006). Early
detection of asian soybean rust using PCR. Plant Health Progress,24 May 2006.
Langton,F.A. (1989). Inheritance in Chrysanthemum morifolium Ramat. Heredity,62, 419-423.
Lee,S.B. & Taylor,J.W. (1992). Phylogeny of five fungus-like protoctistan Phytophthora species,
inferred from the internal transcribed spacers of ribosomal DNA. Molecular Biology and Evolution,9,
636-653.
191
References
Lelliott,R.A. (1984). Cost/benefit analysis as used in the United Kingdom for eradication campaigns
against alien pests and diseases. Bulletin OEPP/EPPO Bulletin,14, 337-341.
Lema-Ruminska,J., Zalewska,M., & Sadoch,Z. (2004). Radiomutants of chrysanthemum
(Dendranthema grandiflora Tzvelev) of the Lady group: RAPD analysis of the genetic diversity. Plant
Breeding,123, 290-293.
Leonard,K.J. & Szabo,L.J. (2005). Stem rust of small grains and grasses caused by Puccinia graminis.
Molecular Plant Pathology,6, 99-111.
Lerceteau-Köhler,E., Guérin,G., Laigret,F., & Denoyes-Rothan,B. (2003). Characterization of mixed
disomic and polysomic inheritance in the octoploid strawberry (Fragaria x ananassa) using AFLP
mapping. TAG Theoretical and Applied Genetics,107, 619-628.
Leu,L.S., Kao,C.W., Yang,H.C., & Lin,H.S. (1982). Puccinia horiana: Occurrence in Taiwan, release and
germination of sporidia, fungicide trial and screening for chrysanthemum resistance to white rust.
Plant Protection Bulletin (Taiwan, R. O. C. ) .,24, 9-18.
Lindfors,T. (1924). Studien über den Entwicklungsverlauf bei einigen Rostpilzen aus zytologischen
und anatomischen Gesichtpunkten. Svensk Botanisk Tidskrift,18, 1-84.
Liu,Z., Szabo,L.J., & Bushnell,W.R. (1993). Molecular cloning and analysis of abundant and stagespecific mRNAs from Puccinia graminis. Molecular Plant-Microbe Intractions,6, 84-91.
Livak,K.J., Flood,S.J., Marmaro,J., Giusti,W., & Deetz,K. (1995). Oligonucleotides with fluorescent
dyes at opposite ends provide a quenched probe system useful for detecting PCR product and nucleic
acid hybridization. Genome Research,4, 357-362.
Livak,K.J. (1999). Allelic discrimination using fluorogenic probes and the 5' nuclease assay. Genetic
Analysis: Biomolecular Engineering,14, 143-149.
Luchi,N., Pratesi,N., Simi,L., Pazzagli,M., Capretti,P., Scala,A., Slippers,B., & Pinzani,P. (2011). HighResolution Melting Analysis: a new molecular approach for the early detection of Diplodia pinea in
Austrian pine. Fungal Biology,115, 715-723.
Luo,Y., Ma,Z., Reyes,H., Morgan,D., & Michailides,T. (2007). Quantification of airborne spores of
Monilinia fructicola in stone fruit orchards of California using real-time PCR. European Journal of
Plant Pathology,118, 145-154.
Ma,Z., Luo,Y., & Michailides,T.J. (2003). Nested PCR assays for detection of Monilinia fructicola in
stone fruit orchards and Botryosphaeria dothidea from pistachios in California. Journal of
Phytopathology,151, 312-322.
Ma,Z.H. & Michailides,T.J. (2005). Advances in understanding molecular mechanisms of fungicide
resistance and molecular detection of resistant genotypes in phytopathogenic fungi. Crop
Protection,24, 853-863.
Mackay,I.M. (2004). Real-time PCR in the microbiology laboratory. Clinical Microbiology and
Infection,10, 190-212.
Maier,W., Begerow,D., Weiss,M., & Oberwinkler,F. (2003). Phylogeny of the rust fungi: an approach
using nuclear large subunit ribosomal DNA sequences. Canadian Journal of Botany,81, 12-23.
192
References
Manel,S., Joost,S., Epperson,B.K., Holderegger,R., Storfer,A., Rosenberg,M.S., Scribner,K.T.,
Bonin,A., & Fortin,M.J. (2010). Perspectives on the use of landscape genetics to detect genetic
adaptive variation in the field. Molecular Ecology,19, 3760-3772.
Martín,C., Uberhuaga,E., & Pérez,C. (2002). Application of RAPD markers in the characterisation of
Chrysanthemum varieties and the assessment of somaclonal variation. Euphytica,127, 247-253.
Martin,P.H. & Firman,I.D. (1970). Resistance of chrysanthemum cultivars to white rust (Puccinia
horiana). Plant Pathology,19, 180-184.
Martinez,K., Rao,C., & Burton,N. (2004). Exposure assessment and analysis for biological agents.
Grana,43, 193-208.
McCartney,H.A., Fitt,B.D.L., & Schmechel,D. (1997). Sampling bioaerosols in plant pathology. Journal
of Aerosol Science,28, 349-364.
McCartney,H.A., Foster,S.J., Fraaije,B.A., & Ward,E. (2003). Molecular diagnostics for fungal plant
pathogens. Pest Management Science,59, 129-142.
McIntosh and,R.A. & Brown,G.N. (1997). Anticipatory Breeding for Resistance to Rust Diseases in
Wheat. Annual Review of Phytopathology,35, 311-326.
Michelmore,R.W., Paran,I., & Kesseli,R.V. (1991). Identification of Markers Linked to DiseaseResistance Genes by Bulked Segregant Analysis - A Rapid Method to Detect Markers in Specific
Genomic Regions by Using Segregating Populations. Proceedings of the National Academy of Sciences
of the United States of America,88, 9828-9832.
Mohan,M., Nair,S., Bhagwat,A., Krishna,T.G., Yano,M., Bhatia,C.R., & Sasaki,T. (1997). Genome
mapping, molecular markers and marker-assisted selection in crop plants. Molecular Breeding,3, 87103.
Morin,P.A., Luikart,G., Wayne,R.K., & the SNP workshop group (2004). SNPs in ecology, evolution
and conservation. Trends in Ecology & Evolution,19, 208-216.
Morrison,T.B., Weis,J.J., & Wittwer,C.T. (1998). Quantification of low-copy transcripts by continuous
SYBR green I monitoring during amplification. BioTechniques,24, 954.
Mueller,U.G. & Wolfenbarger,L.L. (1999). AFLP genotyping and fingerprinting. Trends in Ecology &
Evolution,14, 389-394.
Mullis,K.B. & Faloona,F.A. (1987). Specific synthesis of DNA in vitro via a polymerase-catalyzed chain
reaction. Methods in Enzymology,155, 335-350.
Muylle,H., Baert,J., Van Bockstaele,E., Moerkerke,B., Goetghebeur,E., & Roldan-Ruiz,I. (2005).
Identification of molecular markers linked with crown rust (Puccinia coronata f. sp lolii) resistance in
perennial ryegrass (Lolium perenne) using AFLP markers and a bulked segregant approach.
Euphytica,143, 135-144.
Nagarajan,S. & Singh,D.V. (1990). Long-Distance Dispersion of Rust Pathogens. Annual Review of
Phytopathology,28, 139-153.
193
References
Neumeister-Kemp,H., Maxwell,A., Kemp,P., Dell,B., & Hardy,G. (2004). An advanced slit-type
volumetric spore trap for monitoring bioaerosols; new methods for identifying fungal spores.
Australasian Plant Pathology,33, 393-400.
organisms. Nucleic Acids Research,31, e19.
Nilsson, L. (1964). Vit Krysanthemumrost i Sverige. Växtskyddnotiser , 28, 99-108.
Norman,M.T., Garcia,F.A., Sandoval,R.F.R., Vazquez,G.L.M., Aquino,M.J., Corona,R.M.C., &
Pedral,M.E. (1995). Comparative evaluation of the resistance of 18 chrysanthemum (Dendranthema
grandiflora Anderson) varieties to white rust (Puccinia horiana Henn.). Revista Chapingo. Serie
Horticultura,1.
O'Gorman,C.M., Fuller,H.T., & Dyer,P.S. (2009). Discovery of a sexual cycle in the opportunistic
fungal pathogen Aspergillus fumigatus. Nature,457, 471-474.
Ochoa,J., Cruz,E., Murillo,A., & Danial,D.L. (2007). Variation in virulence and resistance in the beanbean rust pathosystem in Ecuador. Euphytica,153, 313-319.
Oneill,T.M.O. & Pye,D. (1997). Evaluation of fungicides for control of chrysanthemum white rust
(Puccinia horiana). Annals of Applied Biology,130, 8-9.
Ono,Y. (2002). The diversity of nuclear cycle in microcyclic rust fungi (Uredinales) and its ecological
and evolutionary implications. Mycoscience,43, 421-439.
Orlikowski,L.B., Wojdyla,A.T., & Skrzypczak,C.Z. (1982). Ocena podatnosci odmian zlocieni na rdze
biala. Ogrodnictwo,8, 217-219.
Panstruga,R. (2003). Establishing compatibility between plants and obligate biotrophic pathogens.
Current Opinion in Plant Biology,6, 320-326.
Panter,S.N. & Jones,D.A. (2002). Age-Related Resistance to Plant Pathogens. Advances in Botanical
Research (pp. 251-280).Academic Press.
Pedley,K.F. (2009). PCR-Based Assays for the Detection of Puccinia horiana on Chrysanthemums.
Plant Disease,93, 1252-1258.
Pei,M.H., Bayon,C., & Ruiz,C. (2005). Phylogenetic relationships in some Melampsora rusts on
Salicaceae assessed using rDNA sequence information. Mycological Research,109, 401-409.
Perkins, W. A. (1957). The rotorod sampler. Standford University.
Person, C. (1959). Gene-for-gene Relationships in Host : Parasite Systems. Canadian Journal of
Botany-Revue Canadienne de Botanique , 37, 1101-1130.
Petersen,G., Seberg,O., Yde,M., & Berthelsen,K. (2006). Phylogenetic relationships of Triticum and
Aegilops and evidence for the origin of the A, B, and D genomes of common wheat (Triticum
aestivum). Molecular Phylogenetics and Evolution,39, 70-82.
Petersen,R. (1974). The rust fungus life cycle. The Botanical Review,40, 453-513.
194
References
Pfender,W. (2009). Demonstration of Pathotype Specificity in Stem Rust of Perennial Ryegrass.
Phytopathology,99, 1185-1189.
Pinard,R., de Winter,A., Sarkis,G., Gerstein,M., Tartaro,K., Plant,R., Egholm,M., Rothberg,J., &
Leamon,J. (2006). Assessment of whole genome amplification-induced bias through highthroughput, massively parallel whole genome sequencing. BMC Genomics,7, 216.
Pink,D. (2002). Strategies using genes for non-durable disease resistance. Euphytica,124, 227-236.
Poland,J.A., Balint-Kurti,P.J., Wisser,R.J., Pratt,R.C., & Nelson,R.J. (2009). Shades of gray: the world
of quantitative disease resistance. Trends in Plant Science,14, 21-29.
Priest,M.J. (1995). Chrysanthemum white rust in new south wales. Australasian Plant Pathology,24,
65-69.
Punithalingam,E. (1968a). Puccinia chrysanthemi. CMI descriptions of pathogenic fungi and
bacteria,175, 1-2.
Punithalingam,E. (1968b). Puccinia horiana. CMI descriptions of pathogenic fungi and bacteria,176,
1-2.
Qu,L. & Hancock,J.F. (2001). Detecting and mapping repulsion-phase linkage in polyploids with
polysomic inheritance. TAG Theoretical and Applied Genetics,103, 136-143.
Rademaker,W. & de Jong,J. (1987). Types of resistance to Puccinia horiana in chrysanthemum. Acta
Horticulture,197, 85-88.
Rademaker,W. & Dejong,J. (1985). Japanse roest: hoe vatbaar of resistent is de chrysant? V de
Bloemisterijakblad voor,40, 45-49.
Rattink,H., Zamorski,C., & Dil,M.C. (1985). Spread and control of white rust (Puccinia horiana) on
chrysanthemums on artificial substrate. Mededelingen van de Faculteit Landbouwwetenschappen
Rijksuniversiteit Gent,50, 1243-1249.
Ravensdale,M., Nemri,A., Thrall,PH., Ellis,J.G., & Dodds,P.N. (2011). Co-evolutionary interactions
between host resistance and pathogen effector genes in flax rust disease. Molecular Plant
Pathology,12, 93-102.
Ribaut,J.M. & Hoisington,D. (1998). Marker-assisted selection: new tools and strategies. Trends in
Plant Science,3, 236-239.
Ripol,M.I., Churchill,G.A., da Silva,J.A.G., & Sorrells,M. (1999). Statistical aspects of genetic mapping
in autopolyploids. Gene,235, 31-41.
Roden,S.E., Dutton,P.H., & Morin,P.A. (2009). AFLP Fragment Isolation Technique as a Method to
Produce Random Sequences for Single Nucleotide Polymorphism Discovery in the Green Turtle,
Chelonia mydas. Journal of Heredity,100, 390-393.
Rodriguez-Navarro,J.A., Zavaleta-Mejia,E., & Alatorre-Rosas,R. (1996). Epidemiology and
management of chrysanthemum ( Dendranthema grandiflora Tzvelev) white rust ( Puccinia horiana
P. Henn.). Fitopatologia,31.
195
References
Rogers,S.L., Atkins,S.D., & West,J.S. (2009). Detection and quantification of airborne inoculum of
Sclerotinia sclerotiorum using quantitative PCR. Plant Pathology,58, 324-331.
Ryan,M.J. & Ellison,C.A. (2003). Development of a cryopreservation protocol for the microcyclic rustfungus Puccinia spegazzinii. Cryoletters,24, 43-48.
Schena,L., Nigro,F., Ippolito,A., & Gallitelli,D. (2004). Real-time quantitative PCR: a new technology
to detect and study phytopathogenic and antagonistic fungi. European Journal of Plant
Pathology,110, 893-908.
Schlötterer,C. (2004). The evolution of molecular markers: just a matter of fashion? Nat Rev Genet,5,
63-69.
Schneider,R., Haudenshield,J., Hartman,G., & Mahaffee,W. (2009). Development and use of
fluorescent antibody and qPCR protocols for the electrostatic spore trap. Phytopathology,99, S115.
Scott,K.J. & Maclean,D.J. (1969). Culturing of Rust Fungi. Annual Review of Phytopathology,7, 123146.
Shea,K. (2007). Cut flowers from countries with chrysanthemum white rust: final rules. Rules and
Regulation Federal Register, 72, 15805-15812.
Sheroze,A., Rashid,A., Shakir,A.S., & Khan,S.M. (2003). Effect of Bio-Control Agents on Leaf Rust of
Wheat and Influence of Different Temperature and Humidity Levels on Their Colony Growth.
International Journal of Agriculture & Biology,5, 83-85.
Silva,J.A., Sorrells,M.E., Burnquist,W.L., & Tanksley,S.D. (1993). RFLP linkage map and genome
analysis of Saccharum spontaneum. Genome / National Research Council Canada = Genome / Conseil
national de recherches Canada,36, 782-791.
Singh,R.P., Hodson,D.P., Jin,Y., Huerta-Espino,J., Kinyua,M.G., Wanyera,R., Njau,P., & Ward,R.W.
(2006). Current status, likely migration and strategies to mitigate the threat to wheat production
from race Ug99 (TTKS) of stem rust pathogen. CAB Reviews: Perspectives in Agriculture, Veterinary
Science, Nutrition and Natural Resources,1, 13.
Skottrup,P., Frokiaer,H., Hearty,S., O'Kennedy,R., Hejgaard,J., Nicolaisen,M., & Justesen,A.F.
(2007). Monoclonal antibodies for the detection of Puccinia striiformis urediniospores. Mycological
Research,111, 332-338.
Spaargaren,J.J. (2002). Jaarrondchrysanten: Teelt Rn Achtergronden. Aalsmeer. Spaargaren
Srivastava,A.K., Defago,G., & Kern,H. (1985). Hyperparasitism of Puccinia horiana and other
microcyclic rusts. Journal of Phytopathology,114, 73-78.
Stammler,G., Cordero,J., Koch,A., Semar,M., & Schlehuber,S. (2009). Role of the Y134F mutation in
cyp51 and overexpression of cyp51 in the sensitivity response of Puccinia triticina to epoxiconazole.
Crop Protection,28, 891-897.
Stark, C. & Stach, W. (1965). Der weisse chrysanthemenrost. Gartenweld, Hamburg , 23, 515-517.
Staskawicz,B.J., Ausubel,F.M., Baker,B.J., Ellis,J.G., & Jones,J.D. (1995). Molecular genetics of plant
disease resistance. Science,268, 661-667.
196
References
Stavely,J.R., Steadman,J.R., & Mcmillan,R.T. (1989). New Pathogenic Variability in Uromyces
appendiculatus in North-America. Plant Disease,73, 428-432.
Stebbins,T. & Johnson,D. (2001). Regulatory Control. In Jones,R.K. & Benson,D.M. (Eds), Diseases of
Woody Ornamentals and Trees in Nurseries (pp. 457-458). St. Paul, MN.American Phytopathological
Society.
Stokes,T.L., Kunkel,B.N., & Richards,E.J. (2002). Epigenetic variation in Arabidopsis disease
resistance. Genes & Development,16, 171-182.
Stukenbrock,E.H. & McDonald,B.A. (2009). Population Genetics of Fungal and Oomycete Effectors
Involved in Gene-for-Gene Interactions. Molecular Plant-Microbe Interactions,22, 371-380.
Sugimura,T. (2001). Nucleotide sequence of nuclear rDNA-ITS regions and design of genus specific
primers for the regions of several plant pathogens of strawberry and chrysanthemum. Bulletin of the
Nara Prefectural Agricultural Experiment Station,32, 9-17.
Sugimura,T., Yoshida,M., Nishizaki,M., & Okayama,K. (1998). Susceptibility difference of
chrysanthemum cultivars against rust fungus isolates (Puccinia horiana P. Henn) and chemical control
of the disease by inoculation with the isolates. Bulletin of the Nara Agricultural Experiment Station
(Japan),29, 9-14.
Takatsu,Y., Ohishi,K., Tomita,Y., Hayashi,M., Nakajima,M., & Akutsu,K. (2000). Use of
chrysanthemum plantlets grown in vitro to test cultivar susceptibility to white rust, Puccinia horiana
P. Hennings. Plant Breeding,119, 528-530.
Tautz,D. (1989). Hypervariabflity of simple sequences as a general source for polymorphic DNA
markers. Nucleic Acids Research,17, 6463-6471.
Taylor,J.W., Jacobson,D.J., & Fisher,M.C. (1999). The Evolution of Asexual Fungi: Reproduction,
Speciation and Classification. Annual Review of Phytopathology,37, 197-246.
Tierens, K. (2007). België, een belangrijke eindproducent van potchrysanten in Europa en een
wereldspeler op gebied van veredeling. Verbondsnieuws , 18, 44.
Tranzschel,W.A. (1905). Beiträge zur biologie der Uredineen. Travaux du Musée botanique de
l'Académie imperiale des sciences de St. Petérsbourg,2, 64-80.
Uchida,T. (1983). Studies on epidemiology and control of chrysanthemum white rust, caused by
Puccinia horiana P. Henn. Bulletin of the Yamanashi Agricultural Experimentaml Station, Futaba,
Yamanashi, Japan, 1-105.
Ukoskit,K. & Thompson,P.G. (1997). Autopolyploidy versus Allopolyploidy and Low-density
Randomly Amplified Polymorphic DNA Linkage Maps of Sweetpotato. Journal of the American Society
for Horticultural Science,122, 822-828.
Van Lierde, D., Taragola, N. & Delattre, B. (1999). De Belgische tuinbouw in nationaal en Europees
perspectief .
van Niekerk,J.M., Calitz,F.J., Halleen,F., & Fourie,P.H. (2010). Temporal spore dispersal patterns of
grapevine trunk pathogens in South Africa. European Journal of Plant Pathology,127, 375-390.
197
References
van Orsouw,N.J., Hogers,R.C.J., Janssen,A., Yalcin,F., Snoeijers,S., Verstege,E. et al. (2007).
Complexity Reduction of Polymorphic Sequences (CRoPS™): A Novel Approach for Large-Scale
Polymorphism Discovery in Complex Genomes. Plos One,2.
Varshney,R.K., Graner,A., & Sorrells,M.E. (2005). Genomics-assisted breeding for crop improvement.
Trends in Plant Science,10, 621-630.
Velasco,R.G., Zavaleta-Mejia,E., Martinez,R.I.R., Mir,S.G.L., Davila,G.F., & Simpson,J. (2007).
Pathogenic and molecular variability of Puccinia horiana Henn. isolates. Revista Fitotecnia
Mexicana,30, 223-234.
Visser,P.B., de Maagd,R.A., & Jongsma,M.A. (2007). Chrysanthemum. In Pua,E.C. & Davey,M.R.
(Eds), Transgenic Crops VI (pp. 253-272).Springer Berlin Heidelberg.
Vittal,R., Yang,H.-C., & Hartman,G. (2012). Anastomosis of germ tubes and migration of nuclei in
germ tube networks of the soybean rust pathogen, Phakopsora pachyrhizi. European Journal of Plant
Pathology,132, 163-167.
Voegele,R.T. & Mendgen,K. (2003). Rust haustoria: nutrient uptake and beyond. New
Phytologist,159, 93-100.
Vos,P., Hogers,R., Bleeker,M., Reijans,M., Vandelee,T., Hornes,M. et al. (1995). AFLP - A New
Technique for DNA-Fingerprinting. Nucleic Acids Research,23, 4407-4414.
Walker,R.I. (1928). Cytological studies of some of the short-cycled rusts. Transactions of the
Wisconsin Academy of Sciences Arts and Letters,23, 567-582.
Walters,D.R. & Bingham,I.J. (2007). Influence of nutrition on disease development caused by fungal
pathogens: implications for plant disease control. Annals of Applied Biology,151, 307-324.
Wang,X., Zheng,W., Buchenauer,H., Zhao,J., Han,Q., Huang,L., & Kang,Z. (2007). The development
of a PCR-based method for detecting Puccinia striiformis latent infections in wheat leaves. European
Journal of Plant Pathology,120, 241-247.
Wang,X. & McCallum,B. (2009). Fusion Body Formation, Germ Tube Anastomosis, and Nuclear
Migration During the Germination of Urediniospores of the Wheat Leaf Rust Fungus, Puccinia
triticina. Phytopathology,99, 1355-1364.
Wang,Y., Zhu,M., Zhang,R., Yang,H., Wang,Y., Sun,G., Jin,S., & Hsiang,T. (2009). Whole genome
amplification of the rust Puccinia striiformis f. sp. tritici from single spores. Journal of Microbiological
Methods,77, 229-234.
Ward,E., Foster,S.J., Fraaije,B.A., & McCartney,H.A. (2004). Plant pathogen diagnostics:
immunological and nucleic acid-based approaches. Annals of Applied Biology,145, 1-16.
Watanabe,K. (1977). Control of Diploid-Like Meiosis in Polyploid Taxa of Chrysanthemum(Compositae). Japanese Journal of Genetics,52, 125-131.
Watanabe,K. (1981). The Control of Diploid-Like Meiosis in Polyploid Taxa of Chrysanthemum.
Japanese Journal of Genetics,56, 649.
198
References
Watanabe,K. (1983). Studies on the Control of Diploid-Like Meiosis in Polyploid Taxa of
Chrysanthemum .4. Colchiploids and the Process of Cytogenetical Diploidization. Theoretical and
Applied Genetics,66, 9-14.
Water,J.K. (1981). Chrysanthemum white rust. Bulletin of the European and Mediterranean Plant
Protection Organization (EPPO),11, 239-242.
Webster,J. & Weber,R.W.S. (2007). Introduction to Fungi. Cambridge. Cambridge University Press
West,J.S., Atkins,S.D., Emberlin,J., & Fitt,B.D.L. (2008). PCR to predict risk of airborne disease.
Trends in Microbiology,16, 380-387.
Whipps,J.M. (1993). A review of white rust (Puccinia horiana Henn) disease on chrysanthemum and
the potential for its biological control with Verticillium lecanii (Zimm) Viegas. Annals of Applied
Biology,122, 173-187.
White,T.J., Bruns,T., Lee,S., & Taylor,J.A. (1990). Amplification and Direct Sequencing of Fungal
Ribosomal RNA Genes for Phylogenetics. In Innis,M.A., Gelfand,D.H., Shinsky,J.J., & White,T.J. (Eds),
PCR Protocols: a Guide to Methods and Applications (pp. 315-322). San Diego.Academic Press.
William,H.M., Singh,R.P., Huerta-Espino,J., Palacios,G., & Suenaga,K. (2006). Characterization of
genetic loci conferring adult plant resistance to leaf rust and stripe rust in spring wheat. Genome,49,
977-990.
Williams,J.G.K., Kubelik,A.R., Livak,K.J., Rafalski,J.A., & Tingey,S.V. (1990). DNA polymorphisms
amplified by arbitrary primers are useful as genetic markers. Nucleic Acids Research,18, 6531-6535.
Williams,R.H., Ward,E., & McCartney,H.A. (2001). Methods for integrated air sampling and DNA
analysis for detection of airborne fungal spores. Applied and Environmental Microbiology,67, 24532459.
Wise,K.A., Mueller,D.S., & Buck,J.W. (2004). Quarantines and ornamental rusts. APSnet Online
http://www.apsnet. org/online/feature/quarantine/.
Wittwer,C.T., Herrmann,M.G., Moss,A.A., & Rasmussen,R.P. (1997). Continuous fluorescence
monitoring of rapid cycle DNA amplification. BioTechniques,22, 130-&.
Wojdyla,A.T. (1998). Zwalczanie rdzy bialej zlocienia. Haslo Ogrodnicze,11, 54.
Wojdyla,A.T. (1999a). Azoxystrobin in the control of Puccinia horiana on chrysanthemum. Progress
in Plant Protection,39, 831-834.
Wojdyla,A.T. (1999b). Susceptibility of chrysanthemum cultivars to Puccinia horiana. Folia
Horticulturae, 11, 115-122.
Wojdyla,A.T. (2002). Influence of fungicides in the control of Puccinia horiana. Progress in Plant
Protection,42.
Wojdyla,A.T. (2004a). Chitosan (biochikol 020 PC) in the control of some ornamental foliage
diseases. Communication Agricultural and Applied Biological Sciences,69, 705-715.
Wojdyla,A.T. (2004b). Development of Puccinia horiana on chrysanthemum leaves in relation to
chemical compounds and time of their application. Journal of Plant Protection Research,44.
199
References
Wojdyla,A.T. (2005). Tolylfluanide + tebuconazole in the control of some ornamental foliage
diseases. Communication Agricultural and Applied Biological Sciences,70, 231-240.
Wojdyla,A.T. (2006). Falcon 460 EC in the control of foliage diseases of some ornamentals.
Communication Agricultural and Applied Biological Sciences,71, 1071-1079.
Wojdyla,A.T. & Orlikowski,L.B. (1999). Strobilurin compounds in the control of rust, powdery mildew
and black spot on some ornamental plants. Mededelingen - Faculteit Landbouwkundige en
Toegepaste Biologische Wetenschappen Universiteit Gent,64, 539-545.
Wolff,K., Zietkiewicz,E., & Hofstra,H. (1995). Identification of chrysanthemum cultivars and stability
of DNA fingerprint patterns. TAG Theoretical and Applied Genetics,91, 439-447.
Wolff,K. & Peters-van Rijn,J. (1993). Rapid detection of genetic variability in chrysanthemum
(Dendranthema grandiflora Tzvelev) using random primers. Heredity,71, 335-341.
Wu,K.K., Burnquist,W., Sorrells,M.E., Tew,T.L., Moore,P.H., & Tanksley,S.D. (1992). The detection
and estimation of linkage in polyploids using single-dose restriction fragments. TAG Theoretical and
Applied Genetics,83, 294-300.
Xu,J., Guo,H., & Yang,Z.L. (2007). Single nucleotide polymorphisms in the ectomycorrhizal
mushroom Tricholoma matsutake. Microbiology,153, 2002-2012.
Yamada,S. (1956). Experiments on the epidemiology and control of chrysanthemum white rust,
caused by Puccinia horiana P. Henn. Annals of the Phytopathological Society of Japan,20, 148-154.
Yamaguchi,T. (1981). Chrysanthemum breeding for resistance to white rust. Japanese Journal of
Breeding,31, 121-132.
Yamazaki,S. & Katsuya,K. (1987). Axenic cultures of Cronartium quercuum and their pathogenicity.
Ann. Phytopathol. Soc. Jpn.,53, 643-646.
Zadoks,J.C., Groenewegen,C.A.M., & Zandvoort,R. (1968). On heat treatment of Chrysanthemum
morilifolium against Puccinia horiana. European Journal of Plant Pathology,74, 25-27.
Zandvoort,R. (1968). Wind dispersal of Puccinia horiana. European Journal of Plant Pathology,74,
124-127.
Zandvoort,R., Groenewegen,C., & Zadoks,J. (1968a). On the incubation period of Puccinia horiana.
European Journal of Plant Pathology,74, 128-130.
Zandvoort,R., Groenewegen,C.A.M., & Zadoks,J.C. (1968b). Chemical control of Puccinia horiana in
Chrysanthemum morifolium. European Journal of Plant Pathology,74, 170-173.
Zandvoort,R., Groenewegen,C.A.M., & Zadoks,J.C. (1968c). Methods for the inoculation of
Chrysanthemum morifolium with Puccinia horiana. European Journal of Plant Pathology,74, 174-176.
Zane,L., Bargelloni,L., & Patarnello,T. (2002). Strategies for microsatellite isolation: a review.
Molecular Ecology,11, 1-16.
Zhang,F., Chen,S., Chen,F., Fang,W., Chen,Y., & Li,F. (2011). SRAP-based mapping and QTL detection
for inflorescence-related traits in chrysanthemum (Dendranthema morifolium). Molecular
Breeding,27, 11-23.
200
References
Zhang,F., Chen,S., Chen,F., Fang,W., & Li,F. (2010). A preliminary genetic linkage map of
chrysanthemum (Chrysanthemum morifolium) cultivars using RAPD, ISSR and AFLP markers. Scientia
Horticulturae,125, 422-428.
201
Summary
Summary
Cultivation of chrysanthemum (Chrysanthemum x morifolium) originates more than 2000 years ago in
Asia. Chrysanthemum is currently one of the most important ornamental crops in different parts of
the world, where it is either grown as cut flowers or as potted plants. A major threat in commercial
chrysanthemum production is chrysanthemum white rust, caused by the quarantine fungus Puccinia
horiana, which can cause devastating damage in susceptible cultivars if not controlled properly.
Outbreaks of the disease are mainly prevented by fungicide treatments, but due to the development
of fungicide resistant strains and the decreasing social acceptance of the excessive use of fungicides,
more sustainable control strategies are required. However, a lack of knowledge about the spread and
the pathogenic diversity of the pathogen as well as the complex segregation in chrysanthemum
hamper the deployment of alternative control strategies such as a guided fungicide treatment and
resistance breeding. In this context, the objectives of this thesis were to determine the phenotypic
and genotypic diversity of the pathogen, to develop and apply a molecular detection technique for P.
horiana spores in air samples and to study the segregation of molecular markers and the resistance
to the pathogen in chrysanthemum.
In the first part of this dissertation the focus was on the phenotypic diversity of a worldwide
collection of P. horiana isolates. In this part, emphasis was placed on the identification and
characterization of pathotypes. Using an optimized bioassay, 22 isolates originating in different
continents were screened on a specific set of 36 cultivars. All isolates showed different interaction
phenotype profiles and infected between 4 and 19 differential cultivars. Based on the Person analysis
of these profiles, the presence of a minimum of seven resistance genes and corresponding avirulence
genes was demonstrated in this pathosystems, which illustrates its complexity. A selection of 17
isolates was also tested for fungicide resistance to two strobilurins and three triazoles. Most isolates
were highly resistant to one or several triazole fungicides, while resistance to strobilurins was
considerably lower. Only two isolates were still susceptible to all fungicides included in the test.
In the second part, the biology of the fungus was studied in more detail using molecular techniques.
CRoPS™ technology, based on next generation sequencing of AFLP fragments, was used to develop
32 SNP markers and one SSR marker in P. horiana. These markers were used to genotype a
worldwide collection of 45 isolates. In most cases, phylogenetic clustering of isolates was related to
the geographic origin. In cases where the isolates were collected over multiple years and in several
parts of the region, this indicates the establishment of the pathogen in such regions. Nevertheless,
evidence for migration was also observed, including migration between Europe and South America
and between South-East Asia and Europe. A strong indication of recombination between different
genotypes was observed, disproving the presumed clonal propagation of the pathogen. The
correlation between genotype and pathotype was poor, although a molecular marker to one
205
Summary
important pathotype group could be developed. In the second part the development of protocols for
the molecular detection of P. horiana basidiospores in air samples using two types of spore traps is
described. Optimization of the techniques resulted in a limit of detection of a single spore and a limit
of quantification of 10 spores. The capture efficiency of the two spore traps differed. Using these
techniques, the presence of the pathogen was surveyed during five field trials. These showed that
sporulation of the pathogen especially occurs at night and is related to a high relative humidity and
especially to rainfall events. Large volumes of spores can be released and these can easily be
dispersed over the distance of a commercial field.
A third and final part of this thesis is dedicated to the segregation of resistance to P. horiana and the
inheritance of AFLP makers in chrysanthemum. The segregation of resistance to an isolate mixture of
P. horiana in the progeny of 20 crosses between resistant and susceptible parents suggest that
preferential chromosome pairing occurs in chrysanthemum. However, it was not possible to
determine the exact number of resistance genes in the resistant parents. A more detailed study on
the inheritance of resistance to different pathotypes of P. horiana in a single cross demonstrated that
resistance to different pathotypes is located on different loci each containing a cluster of R genes.
The segregation of 182 AFLP markers confirmed the hypothesis of preferential chromosome pairing
observed earlier. Some of the AFLP markers also cosegregated with resistance to specific pathotypes
of P. horiana. The association between several of these markers also confirmed the presence of
separate clusters of R genes.
The insights obtained during this research offer a basis for resistance breeding to particular
pathotypes of P. horiana. The bioassay developed for pathotype screening is also suitable for high
throughput resistance screening of candidate cultivars. By using a combination of isolates that can
overcome specific resistance genes, cultivars resistant to all known pathotypes can be developed.
The segregating population of progeny from one of the crosses can be used for the identification and
characterization of loci that are associated with resistance to P. horiana in chrysanthemum. In the
future this could offer the possibility to select resistant cultivars using marker-assisted selection.
Using the optimized spore detection protocol, the biology and epidemiology of the pathogen can be
studied more in detail. This technique can be developed into a warning system giving the opportunity
to decrease the use of fungicides. In combination with a well thought-out use of fungicides this can
help prevent the development of fungicide resistance. Using the SNP markers developed for the
characterization of P. horiana, the introduction, spread and local survival of P. horiana in a particular
region can be analyzed. The markers can also be used to identify higher risk exotic genotypes among
intercepted isolates. Together, implementation of the knowledge obtained during this research can
form the basis for a more sustainable control of P. horiana.
206
Samenvatting
Samenvatting
De teelt van chrysant (Chrysanthemum x morifolium) vindt zijn oorsprong meer dan 2000 jaar
geleden in Azië. Tegenwoordig is chrysant één van de belangrijkste sierteeltgewassen in
verschillende werelddelen, waar ze wordt geteeld als snijbloem of als potplant. Een belangrijke
bedreiging voor de commerciële chrysantenteelt is Japanse roest, die veroorzaakt wordt door de
quarantaine schimmel Puccinia horiana en die enorme schade kan aanrichten in vatbare cultivars
indien de ziekte niet naar behoren wordt beheerst. Uitbraken van de schimmel worden hoofdzakelijk
vermeden door middel van preventieve fungicidenbehandelingen, maar door het verschijnen van
fungicidenresistente stammen en een afname van het sociaal draagvlak voor het overmatig gebruik
van fungiciden, zijn duurzamere controlestrategieën vereist. Een gebrek aan kennis over de
verspreiding en de pathotypische diversiteit van P. horiana evenals de complexe overerving bij
chrysant staan een goede ontplooiing van duurzamere controle strategieën zoals begeleide
fungicidenbehandeling en resistentieveredeling echter in de weg. In het kader hiervan waren de
objectieven van deze scriptie om de fenotypische en genotypische diversiteit van de pathogeen te
bepalen, om een moleculaire techniek voor de detectie van sporen van de schimmel in luchtstalen te
ontwikkelen en toe te passen en om de overerving van moleculaire merkers en resistentie in
chrysant diepgaander te bestuderen.
In het eerste deel van deze scriptie lag de focus op de fenotypische diversiteit van een wereldwijde
collectie isolaten van P. horiana. Hierbij lag de nadruk op het identificeren en karakteriseren van
pathotypes. Met behulp van een geoptimaliseerde biotoets werden 22 isolaten getoetst op een
specifieke set van 36 cultivars. Alle isolaten hadden een uniek interactieprofiel en infecteerden
tussen 4 en 19 differentiële cultivars. Op basis van de Person analyse van deze profielen werd de
aanwezigheid van minimum zeven resistentiegenen en overeenkomstige avirulentiegenen in dit
pathosysteem aangetoond, wat de complexiteit ervan illustreert. Een selectie van 17 isolaten werd
ook getest op fungicidenresistentie tegen twee strobilurines en 3 triazolen. De meeste isolaten
vertoonden een hoge resistentie tegen één of meerdere triazolen, terwijl de resistentie tegen
strobilurines aanzienlijk lager was. Slechts twee isolaten waren nog vatbaar voor alle geteste
fungiciden.
In het tweede deel werd de biologie van de schimmel meer in detail bestudeerd met behulp van
moleculaire technieken. Door middel van de CRoPS™ technologie, gebaseerd op next generation
sequencing van AFLP fragmenten, werden 32 SNP merkers en één SSR merker in P. horiana
ontwikkeld. Deze merkers werden vervolgens gebruikt voor de genotypering van een wereldwijde
collectie van 45 isolaten. In de meeste gevallen was de fylogenetische clustering van isolaten
gerelateerd aan de geografische oorspong. Voor die gevallen waarin de isolaten werden verzameld in
verschillende jaren en in verschillende delen van een regio, geeft dit aan dat de pathogeen zich lokaal
209
Samenvatting
heeft gevestigd. Er werden echter ook verschillende gevallen van migratie van de pathogeen
vastgesteld, onder andere tussen Europa en Zuid-Amerika en tussen Zuid-Oost Azië en Europa. Een
sterke indicatie van recombinatie tussen verschillende genotypes werd waargenomen, wat de
veronderstelde klonale vermeerdering van de pathogeen weerlegt. De correlatie tussen genotype en
pathotype was beperkt, hoewel een moleculaire merker voor één belangrijke pathotype groep kon
worden ontwikkeld. In het tweede deel wordt ook de ontwikkeling van protocols voor de moleculaire
detectie van P. horiana basidiosporen in twee types luchtstalen beschreven. Optimalisatie van de
technieken resulteerde in een detectie limiet van één enkele spore en een kwantificeringslimiet van
10 sporen. De vangstefficiëntie van de twee sporenvallen verschilde significant. Met behulp van de
geoptimaliseerde technieken werd de aanwezigheid van de pathogeen gedurende vijf
veldexperimenten opgevolgd. Deze toonden aan dat sporulatie van de pathogeen vooral ’s nachts
plaats grijpt, dat sporulatie sterk gerelateerd is met hoge luchtvochtigheid en in het bijzonder met
regenval. Grote volumes sporen kunnen worden vrijgesteld en deze kunnen gemakkelijk over de
afstand van een commercieel veld worden verspreid.
Een derde en laatste deel beschrijft de segregatie van roestresistentie en de overerving van AFLPmerkers bij chrysant. De segregatie van resistentie tegen een mengisolaat van P. horiana in 20
kruisingen tussen resistente en vatbare ouders suggereert dat preferentiële chromosoom paring de
norm is in chrysant. Het was echter niet steeds mogelijk om het exacte aantal resistentiegenen in de
ouders te bepalen. Een meer gedetailleerde studie van de overerving van resistentie tegen
verschillende pathotypes van P. horiana in één enkele kruising toonde aan dat resistentie tegen
verschillende pathotypes op verschillende loci met elk een cluster van R genen liggen. De uitsplitsing
van 182 AFLP merkers bevestigende de hypothese van preferentiële chromosoomparing die eerder
was waargenomen. Enkele van deze merkers vertoonden ook cosegregatie met resistentie tegen
specifieke pathotypes van P. horiana. De associatie tussen verschillende van deze merkers bevestigt
eveneens het voorkomen van afzonderlijke clusters van R genen.
De inzichten die tijdens deze scriptie werden verworven bieden de basis voor een gerichte
resistentieveredeling tegen P. horiana. De biotoets die ontwikkeld werd voor de pathotypescreening
leent zich ook voor de resistentiescreening van kandidaat cultivars. Door gebruik te maken van een
combinatie van isolaten die specifieke resistentiegenen doorbreken, kunnen cultivars die resistent
zijn tegen alle gekende pathotypes worden ontwikkeld. De segregerende populatie van
nakomelingen van één van de kruisingen kan worden gebruikt voor de identificatie en karakterisering
van loci die geassocieerd zijn met resistentie tegen P. horiana in chrysant. Dit biedt de mogelijkheid
om in de toekomst resistente cultivars te selecteren met behulp van merker-geassisteerde selectie.
Het geoptimaliseerde protocol voor detectie van sporen in luchtstalen leent zich perfect voor het
210
Samenvatting
verder bestuderen van de epidemiologie van de pathogeen. In de toekomst kan deze techniek
worden gebruikt voor het ontwikkelen van een waarschuwingssysteem waardoor het gebruik van
fungiciden kan worden beperkt. In combinatie met een doordacht gebruik van fungiciden moet dit de
verdere ontwikkeling van fungicidenresistentie helpen tegen gaan. Door gebruik te maken van de
SNP-merkers voor de karakterisering van onderschepte isolaten kan de introductie, verspreiding en
lokale overleving van P. horiana worden geanalyseerd. De merkers kunnen ook worden gebruikt voor
de identificatie van exotische genotypes met een verhoogd risico in onderschepte isolaten.
Samengevat kan implementatie van de kennis die tijdens dit onderzoek werd vergaard een basis
vormen voor een duurzamere controle van P. horiana.
211
Curriculum Vitae
Curriculum vitae
Personal information
Name
Mathias De Backer
Date of birth
December 7, 1982
Place of birth Wilrijk
Nationality
Belgian
Contact home Robiniadreef 33
9050 Gentbrugge
Belgium
+32(0)485 369 748
Contact work Burgemeester van Gansberghelaan 96 bus 2
9820 Merelbeke
Belgium
+32(0)9 272 2463
Education
2007 – 2012
PhD training, Doctoral schools of Bioscience engineering, Ghent University
2004 – 2005
MSc in Environmental sciences and Technologies, faculty of Bioscience
Engineering, Ghent University
2000 – 2004
MSc in Biomedical Sciences; faculty of Medicine and Health Sciences, Ghent
University
1994 – 2000
Secondary school, Science – Math, Sint-Lodewijkscollege, Lokeren
Professional experience:
2011 – present Scientist at the Institute for Agricultural and Fisheries Research (ILVO), Plant
Science Unit, Crop protection, Virology
2007 – 2011
PhD Student at the Institute for Agricultural and Fisheries Research (ILVO),
Plant Science Unit, Crop protection (PhD scholarship granted by ILVO)
2006 – 2007
Scientific responsible, Laboratoires Pharmaceutiques Trenker, Brussels
Publications
Hossein Alaei, Mathias De Backer, Jorinde Nuytinck, Martine Maes, Monica Höfte and Kurt
Heungens (2009). Phylogenetic relationships of Puccinia horiana and other rust pathogens of
Chrysanthemum x morifolium based on rDNA ITS sequence analysis. Mycological Research
113: 668 – 683
215
Curriculum vitae
Mathias De Backer, Hossein Alaei, Erik Van Bockstaele, Isabel Roldan-Ruiz, Theo van der Lee,
Martine Maes and Kurt Heungens (2011). Identification and characterization of pathotypes in
Puccinia horiana, a rust pathogen of Chrysanthemum x morifolium. European Journal of
Plant Pathology, 130: 325 – 338
Mathias De Backer, Peter Bonants, Kerry Pedley, Martine Maes, Erik Van Bockstaele, Kurt
Heungens and Theo van der Lee (2013). Migration, survival and recombination of Puccinia
horiana based on newly identified SNP markers. Target journal: Molecular ecology (in
preparation)
Mathias De Backer, Martine Maes, Erik Van Bockstaele and Kurt Heungens (2013). Aerial
biology of Puccinia horiana. (in preparation)
Conference contributions
Mathias De Backer, Erik Van Bockstaele, Martine Maes and Kurt Heungens. Pathotypes of
Puccinia horiana, causal agent of chrysanthemum white rust. 62nd International Symposium
on Crop Protection, May 18, 2010, Gent, Belgium. (oral presentation)
Mathias De Backer, Martine Maes, Erik Van Bockstaele and Kurt Heungens. Pathotype
analysis in Puccinia horiana. International Mycological Conference (IMC9), August 5, 2010,
Edinburgh, UK. (oral presentation)
Mathias De Backer, Martine Maes, Erik Van Bockstaele, Theo van der Lee and Kurt Heungens.
Genotypic and phenotypic characterization of Puccinia horiana, a quarantine rust pathogen
on Chrysanthemum. 17th Symposium on Applied Biological Sciences, February 10, 2012,
Leuven, Belgium. (poster presentation)
Mathias De Backer, Martine Maes, Erik Van Bockstaele, Theo van der Lee and Kurt Heungens.
Indications of genetic recombination, international migration and local survival of Puccinia
horiana based on new SNP markers. 64nd International Symposium on Crop Protection, May
22, 2010, Gent, Belgium. (oral presentation)
Mathias De Backer, Martine Maes, Erik Van Bockstaele, Theo van der Lee and Kurt Heungens.
New SNP markers for the genetic characterization of the quarantine pathogen Puccinia
horiana and their application in pathogen migration analysis. QBOL-EPPO Conference on
DNA barcoding and diagnostic methods for plant pests, May 24, 2012 (oral presentation)
Conference attendance without contribution
59 th International Symposium on Crop Protection, Gent, May 22, 2009
BeSCroP meeting, fytosanitaire Balans van het Teeltjaar 2006-2007, Louvain-La-Neuve,
december 5, 2007
BeSCroP meeting, Quarantaine Organismen in de Landbouw, Gembloux, April 9, 2008.
60th International Symposium on Crop Protection, Gent, May 20, 2008.
216
Curriculum vitae
14th PhD Symposium on Applied Biological Sciences, Gent, September 15, 2008.
61th International Symposium on Crop Protection. Gent, May 19, 2009.
63d International Symposium on Crop Protection. Gent, May 24, 2011.
Belgian Plant Biotech Association: 5th symposium. Louvain-la-Neuve, November 18, 2011
Courses and workshops
Geographic Information Systems (GIS), prof. P. De Maeyer, Ghent University
Applied Genetics, by prof. D. Reheul, Ghent University
Crop Protection, by prof. W. Steurbaut, prof. M. Höfte, prof. P. De Clerq and dr. B. De
Cauwer, Ghent University
Presentation techniques: Effective Scientific Communication, by Jean-Luc Doumont,
Principae
Vijfde Workshop Phytopathology: Inoculatiemethoden en protocollen. Plantum, NAK
Tuinbouw. February 2, 2012
International 3 week training in SNP detection at Plant Research International (PRI),
Wageningen, November 2009
Supervision of undergraduate students
Mikhail Claeys. Moleculaire detectie van Puccinia horiana met real time PCR. Professional
Master thesis in applied Industrial Sciences: Biochemistry. Academic Year 2009-2010.
Hogeschool Gent, Departement Toegepaste Ingenieurswetenschappen.
217
Mathias De Backer