GENETIC AND TRANSCRIPTOMIC ANALYSES OF THE
PSEUDOPERONOSPORA CUBENSIS-CUCUMIS SATIVUS INTERACTION
By
Elizabeth Ann Savory

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Plant Pathology
2012

ABSTRACT
GENETIC AND TRANSCRIPTOMIC ANALYSES OF THE
PSEUDOPERONOSPORA CUBENSIS-CUCUMIS SATIVUS INTERACTION
By
Elizabeth Ann Savory
Pseudoperonospora cubensis [(Berkeley & M. A. Curtis) Rostovzev] is the causal agent
of cucurbit downy mildew, a foliar disease responsible for devastating losses worldwide
of cucumber, cantaloupe, pumpkin, watermelon, and squash.

In the United States,

cucurbit downy mildew is the major threat to cucumber (Cucumis sativus) production
and has been a significant limiting factor since the 2004 and 2005 growing seasons.
Prior to 2004, host resistance had been an effective means of controlling the disease
and, as such, limited research had been conducted to study the biology, genetics, or
virulence of Ps. cubensis. The first step to expand the knowledge base of Ps. cubensis
was the

generation of a 64.4 Mbp genomic assembly of the MSU-1 isolate and

prediction of 23,519 loci and 23,522 gene models. Similar to other oomycete plant
pathogens, Ps. cubensis utilizes RXLR and RXLR-like effector proteins, which can
function as either virulence or avirulence determinants during the course of host
infection. Using in silico analyses, 271 candidate effector proteins were identified with
variable RXLR motifs, including 20 different amino acids at position R1. Of these, only
109 (41%) had a putative ortholog in Phytophthora infestans and evolutionary rate
analysis of these orthologs shows that they are evolving at a significantly faster rate
than most other genes. One Ps. cubensis effector protein, RXLR protein 1 (PscRXLR1)
was characterized in detail. PscRXLR1 was shown to be up-regulated during the early

stages of host infection, and elicits a cell death response in Nicotiana benthamiana.
PscRXLR1 was also demonstrated to be a product of alternative splicing, marking this
as the first example of an alternative splicing event in plant pathogenic oomycetes
transforming a non-effector gene into a functional effector protein.

We present the first large-scale global gene expression analysis of Ps. cubensis
infection of a susceptible C. sativus cultivar, ‘Vlaspik’, and identification of both
pathogen and host genes involved in infection and the defense response, respectively.
Using mRNA-Seq, we captured differential expression of 2383 Ps. cubensis genes in
sporangia and at 1, 2, 3, 4, 6, and 8 days post-inoculation (dpi). Co-expression analyses
identified distinct modules of Ps. cubensis genes that were representative of early,
intermediate, and late infection stages. Additionally, the expression of 15,286 cucumber
genes was detected, of which 14,476 were expressed throughout the infection process
from 1 dpi to 8 dpi.

The rapid induction of key defense related genes, including

catalases, chitinases, lipoxygenases, peroxidases, and protease inhibitors was detected
within 1 dpi, suggesting recognition by C. sativus of the initial stages of Ps. cubensis
infection.

Co-expression network analyses revealed transcriptional networks with

distinct patterns of expression including down-regulation at 2 dpi of known defense
responsive genes suggesting coordinated suppression of host responses by the
pathogen.

In total, the work described herein presents an in-depth analysis of the

interplay between host susceptibility and pathogen virulence in an agriculturally
important pathosystem.

I would like to dedicate this dissertation to my family.

iv

ACKNOWLEDGEMENTS
I would like to thank my advisor, Dr. Brad Day for his unwavering support, advice, and
guidance over the past 5 years.

I would also like to thank my committee members Drs. C. Robin Buell, Raymond
Hammerschmidt, and Rebecca Grumet for their support, collaboration, and advice over
the course of my dissertation research.

I would like to thank the Michigan State University Project GREEEN, Pickle Packers
International, Inc, National Science Foundation East Asia and Pacific Summer Institutes
and the United States Department of Agriculture National Institute of Food and
Agriculture, Agriculture and Food Research Initiative for funding this reasearch.

Finally, I would like to thank my lab mates, Caleb Knepper, Katie Porter, Alyssa
Burkhardt, Patricia Santos, and Masaki Shimono who always made coming to lab
worthwhile, who were always there whether the data looked good or bad, and without
whom I would not have made it through. Thank you.

v

TABLE OF CONTENTS

LIST OF TABLES...........................................................................................................viii
LIST OF FIGURES...........................................................................................................ix
CHAPTER 1 The cucurbit downy mildew pathogen Pseudoperonospora cubensis
Summary................................................................................................................2
Introduction............................................................................................................4
Taxonomy and Morphology…………………………………………………………….5
Symptoms and Signs.............................................................................................7
Dispersal and Survival.........................................................................................10
Infection Mechanisms..........................................................................................12
Pathogenicity and Virulence................................................................................13
Disease Management..........................................................................................16
Future Prospects..................................................................................................19
Acknowledgements..............................................................................................21
.
References...........................................................................................................22
CHAPTER 2 Alternative splicing of a multi-drug transporter from Pseudoperonospora
cubensis generates an RXLR effector protein that elicits a rapid cell death
Abstract................................................................................................................32
Introduction..........................................................................................................33
Results.................................................................................................................38
Discussion............................................................................................................57
Materials and Methods.........................................................................................63
Acknowledgements..............................................................................................72
Appendix………………………………………………………………………………..73
References...........................................................................................................87
CHAPTER 3 mRNA-Seq Analysis of the Pseudoperonospora cubensis transcriptome
during cucumber (Cucumis sativus L.) infection
Abstract…………………………………………………………………………………95
Introduction……………………………………………………………………………..96
Results and Discussion……………………………………………………………...100
Conclusions………………………...…………………………………………………120
Materials and Methods………………………………………………………………121
Appendix………………………………………………………………………………127
References……………………………………………………………………………132

vi

CHAPTER 4 Expression profiling of Cucumis sativus in response to infection by
Pseudoperonospora cubensis
Abstract………………………………………………………………………………..141
Introduction……………………………………………………………………………142
Results and Discussion……………………………………………………………...145
Conclusions………………..………………………………………………………….162
Materials and Methods………………………………………………………………164
Acknowledgements…………………………………………………………………..169
Appendix………………………………………………………………………………170
References……………………………………………………………………………177
CHAPTER 5 Conclusions and Future Directions
Conclusions.......................................................................................................185
Future Directions................................................................................................191
References.........................................................................................................195

vii

LIST OF TABLES
Table 3.1

Number of differentially expressed genes between each combination of
time points and sporangia…………………………………………………...111

Table 4.1

Number of genes differentially expressed between different time
points……………………………………………………………………….…159

viii

LIST OF FIGURES
Figure 1.1

Cucurbit downy mildew symptoms caused by Pseudoperonospora
cubensis………………………………………………………………………….8

Figure 1.2

Morphology of Pseudoperonospora cubensis.............................................9

Figure 1.3

Life cycle of Pseudoperonospora cubensis………………………………..11

Figure 2.1

Strength of purifying selection on Pseudoperonospora cubensis
effectors…………………………………….…………………………………..41

Figure 2.2

PscRXLR1 encodes a RXLR-containing effector protein with homology to
a non-effector protein in Phytophthora infestans…………………………..47

Figure 2.3.

Functional characterization of PscRXLR1 and PITG_17484……………..50

Figure 2.4.

PscRXLR1 mRNA expression is up-regulated during Pseudoperonospora
cubensis infection of cucumber………………………………………………54

Figure 2.5.

PscRXLR1 is a splice variant of Psc_781.4………………………………..56

Figure S2.1 Signal peptide distribution among ortholog pairs…………………………..74
Figure S2.2 Relationship between PscRXLR1 and oomycete orthologs……………...75
Figure S2.3 Heterologous expression of PscRXLR1 specifically results in cell death in
Nicotiana benthamiana………………………………………………………..81
Figure S2.4 Multiple sequence alignments of splice variant isoforms…………………82
Figure S2.5 Heterologous expression of Psc_781.4 in Nicotiana benthamiana……...86
Figure 3.1

Experimental design and sample collection…………………………… 102

Figure 3.2

Symptoms and microscopy images of Ps. cubensis infected Cucumis
sativus cultivar 'Vlaspik' of time points used for transcriptome
analysis………………………………………………………………………..104

Figure 3.3

Number of total RNA-seq reads, reads mapped, and number of genes
expressed……………………………………………………………………..106

Figure 3.4

Correlation matrix of Pseudoperonospora cubensis expression profiles
throughout a time course of Cucumis sativus infection…………………..109

ix

Figure 3.5

Candidate effectors expressed at different timepoints…………………..113

Figure 3.6

CAZymes in Pseudoperonospora cubensis expressed during infection on
Cucumis sativus………………………………………………………………114

Figure 3.7

Comparison of ribonucleic acid sequencing (RNA-seq) and microarray
expression patterns…………………………………………………………116

Figure 3.8

Heat map of the eigengenes representing each gene module..………..119

Figure S3.1 Number of total RNA-seq reads, reads mapped, and number of genes
expressed at different timepoints…………………………………………..128
Figure S3.2 Concordance of FPKM values of the genes expressed in two biological
replicates of the Pseudoperonospora cubensis transcriptome………….129
Figure S3.3 Trend plots of the normalized gene expression values for each gene from
six identified gene co-expression modules………………………………..131
Figure 4.1

Symptoms of Pseudoperonospora cubensis infection on susceptible
Cucumis sativus cv. ‘Vlaspik’………………………………………………147

Figure 4.2

Experimental design and sample collection………………………………149

Figure 4.3

Comparison of total mRNA-Seq reads, reads mapped and number of
genes expressed……………………………………………………………..151

Figure 4.4

Correlation matrix of Cucumis sativus expression profiles during infection
by Pseudoperonospora cubensis…………………………………………..154

Figure 4.5

Comparison of orthologous gene expression in Cucumis sativus and
Arabidopsis thaliana in a compatible interaction………………………..157

Figure 4.6

Heat map of eigengenes representing each gene module………………161

Figure 4.7

Trend plots of the normalized gene expression values from six identified
gene co-expression modules………………………………………………163

Figure S4.1 Concordance of expression values in two biological replicates of Cucumis
sativus during infection by Pseudoperonospora cubensis………………171
Figure S4.2 Trend plots for all 11 modules………………………………………………173

x

CHAPTER 1

The cucurbit downy mildew pathogen Pseudoperonospora cubensis

This review was previously published in Molecular Plant Pathology.
Savory EA, Granke LL, Quesada-Ocampo LM, Varbanova M, Hausbeck M, and Day B.
2011. The cucurbit downy mildew pathogen Pseudoperonospora cubensis. Mol Plant
Pathol 12:217-226. © 2011 BSPP and Blackwell Publishing Ltd.

1

SUMMARY

Pseudoperonospora cubensis [(Berkeley & M. A. Curtis) Rostovzev], the causal agent
of cucurbit downy mildew, is responsible for devastating losses worldwide of cucumber,
cantaloupe, pumpkin, watermelon, and squash. Although downy mildew has been a
major issue in Europe since the mid-1980s, in the United States, downy mildew on
cucumber was successfully controlled for many years through host resistance.
However, since the 2004 growing season, host resistance is no longer effective, and as
a result, the control of downy mildew on cucurbits now requires an intensive fungicide
program. Chemical control is not always feasible due to the high costs associated with
fungicides and their application. Moreover, the presence of pathogen populations
resistant to commonly used fungicides limits the long-term viability of chemical control.
This review summarizes the current knowledge of taxonomy, disease development,
virulence, pathogenicity and control of Ps. cubensis.

In addition, topics for future

research that aim to develop both short- and long-term control measures of cucurbit
downy mildew are discussed.

Taxonomy:

Kingdom Straminipila; Phylum Oomycota; Class Oomycetes; Order

Peronosporales; Family Peronosporaceae; Genus Pseudoperonospora; Species
Pseudoperonospora cubensis.

Disease Symptoms: Angular chlorotic lesions bound by leaf veins on the foliage of
cucumber. Symptoms vary on different cucurbit species and varieties, specifically in

2

terms of lesion development, shape, and size. Infection of cucurbits by Ps. cubensis
impacts fruit yield and overall plant health.

Infection process: Sporulation on the underside of leaves results in the production of
sporangia that are wind-dispersed.

Upon arrival to a susceptible host, sporangia

germinate in free water on the leaf surface, producing biflagellate zoospores that swim
to and encyst on stomata where they form germ tubes. An appressorium is produced
and forms a penetration hypha, which enters the leaf tissue through the stomata.
Hyphae grow through the mesophyll and establish haustoria, specialized structures for
the transfer of nutrients and signals between host and pathogen.

Control: Management of downy mildew in Europe requires the use of tolerant cucurbit
cultivars in conjunction with fungicide applications. In the U.S., an aggressive fungicide
program with sprays every five to seven days for cucumber, and every seven to ten
days for other cucurbits, has been necessary to control outbreaks and prevent crop
loss.

Useful websites: http://www.daylab.plp.msu.edu/pseudoperonospora-cubensis/ (Day
Lab website with research advances in downy mildew), http://veggies.msu.edu/
(Hausbeck Lab website with downy mildew news for growers), http://cdm.ipmpipe.org/
(Cucurbit downy mildew forecasting homepage), http://ipm.msu.edu/downymildew.htm
(Downy mildew information for Michigan’s vegetable growers).

3

INTRODUCTION

Cucurbit downy mildew (caused by Pseudoperonospora cubensis) is one of the most
important foliar diseases of cucurbits, causing significant yield losses in the U.S.,
Europe, China and Israel [1]. The pathogen has a wide geographic distribution and has
been reported in over 70 countries, including environments ranging from semi-arid to
tropical. In addition, Ps. cubensis has a wide host range, infecting approximately 20
different genera of cucurbits [2,3].

The cucurbit crops grown in the U.S. that are

susceptible to this aerially dispersed oomycete pathogen are valued at more than
$246.2 million [4].

Control of downy mildew relies on application of fungicides and the use of host
resistance.

Nonetheless, fungicide-resistant Ps. cubensis populations have been

documented throughout the world [3,5,6,7,8], and host resistance is no longer sufficient
to control the disease as it once was in the U.S. [9]. The additional cost of fungicides,
coupled with potential yield losses of up to 100% caused by downy mildew, threaten the
long-term viability of cucurbit crop production [3,8,10].

Detailed knowledge of Ps. cubensis epidemiology, infection processes, and population
genetics is currently lacking, but is necessary to guide future efforts in developing new
resistant varieties and fungicides, as well as preventing pathogen populations from
overcoming host resistance and chemical control. Studies of pathogen epidemiology
and global population genetics, evaluation of fungicides for disease control, and

4

development of a high-coverage draft genome sequence will assist in our understanding
of the pathogen, as well as in developing effective diagnostics and control measures for
Ps. cubensis. The aim of this review is to briefly summarize what is currently known
about the cucurbit downy mildew pathogen, Ps. cubensis, including taxonomy, disease
development, virulence, pathogenicity and management.

TAXONOMY AND MORPHOLOGY

Ps. cubensis is the type species of the genus Pseudoperonospora, which includes five
accepted species: Ps. cubensis, Ps. humuli, Ps. cannabina, Ps. celtidis and Ps. urticae,
[11]. In addition, there are reports of a sixth species, Ps. cassiae, which, while rare,
may also be a true species of Pseudoperonospora [12]. Originally named Peronospora
cubensis when discovered in Cuba by Berkeley and Curtis [13], Ps. cubensis was
reclassified in 1903 after further observations of sporangia germination [14].
Pseudoperonospora species have true sporangia that germinate via cytoplasmic
cleavage to produce zoospores (Figure 1.2c); whereas, species of Peronospora have
sporangia that germinate directly via a germ tube [2,14].

Recent work has shown that there are no significant morphological differences between
Ps. cubensis and the hop (Humulus spp.) pathogen Ps. humuli; nonetheless, there is no
evidence that Ps. humuli can infect cucurbits, and limited support for Ps. cubensis
pathogenicity on hops [15]. Molecular evidence also shows conflictive results. Internal
transcribed spacer (ITS) region sequences of both pathogens are highly similar, which

5

suggests that Ps. humuli could be a taxonomic synonym of Ps. cubensis [11]. However,
a recent study using single nucleotide polymorphisms (SNPs) indicates that two nuclear
and one mitochondrial gene support the separation of Ps. cubensis from Ps. humuli
[15].

In addition, host range studies have demonstrated pathogenic differences

between Ps. cubensis and Ps. humuli that further supports separation of these species
[16].

Overall, these genetic, phenotypic and physiological characterizations of Ps.

cubensis and Ps. humuli provide support for the distinction between these species.
Further studies including evidence from hundreds of loci would be helpful to fully resolve
the phylogeny of these closely related species.

Morphological characters may not provide enough information for characterization of
Ps. cubensis isolates or even for differentiating species of Pseudoperonospora [17]. Ps.
cubensis sporangiophore morphology can vary with temperature, while sporangia
dimensions are influenced by cucurbit host [12,18]. Recent work with a single isolate of
Ps. cubensis inoculated onto six different cucurbit species has shown that the host cell
matrix plays a role in influencing five morphological criteria, including sporangiophore
length, length of ultimate branchlets, sporangial length and width, and the ratio between
sporangial length and width [17].

The differences among these morphological

characteristics were more obvious in phylogenetically unrelated hosts. These results
indicate that it is desirable to include information from genetic markers when resolving
phylogenetic relationships in species of Pseudoperonospora.

6

SYMPTOMS AND SIGNS

Downy mildew of cucurbits is a foliar disease and is easily recognizable by the
development of chlorotic lesions on the adaxial leaf surface, sometimes with necrotic
centers. These lesions can be restricted by the leaf vein, as in cucumber, giving them
an angular appearance (Figure 1.1). In other cucurbits, the symptoms may vary slightly
in terms of shape and color. For example, in both melon and watermelon, foliar lesions
are less defined than those on cucumber, and are not always bound by leaf veins [1].
As infection progresses, the chlorotic lesions expand and may become necrotic [19],
with necrosis occurring more quickly in hot, dry weather [20]. Leaves colonized by Ps.
cubensis undergo changes in temperature and transpiration rates, which vary during the
course of infection and over the leaf surface [19,21]. Low temperatures can delay
symptom development while still promoting colonization of the leaf tissue, while higher
temperatures result in faster lesion chlorosis that may inhibit pathogen growth [22]. As
the downy mildew disease progresses, entire leaves may die within days following the
initial infection, as lesions expand and coalesce [1].

A reduced canopy leads to

cessation of fruit development and increased sun exposure of the fruit, allowing for sun
scald and secondary rots [23].

Ultimately, crop yield and fruit quality are affected

(Figure 1.1).

When temperatures are below those that allow lesion formation and relative humidity is
≥90%, sporulation, the eponymous “downy” appearance on the lower leaf surface, may
be the first sign of disease (Figure 1.1) [24]. Hyaline sporangiophores (180-400 µm)

7

Figure 1.1

Figure 1.1 Symptoms caused by Pseudoperonospora cubensis. (A) Yellow
angular lesions on a cucumber leaf. (B) Typical “downy” appearance on abaxial leaf
surface caused by sporulation. Yields from untreated (C) and treated (D) cucumbers
infected with Ps. cubensis. E. Field showing typical disease symptoms. Severe
symptoms on untreated (F) versus treated (G) cucumbers. For interpretation of the
references to color in this and all other figures, the reader is referred to the electronic
version of this dissertation.
bearing papillate, lemon-shaped, gray-purple sporangia (20-40 x 14-25 µm) on
sterigmata emerge in groups of 1-6 from stomata on the abaxial surface of infected
leaves (Figure 1.2) [11,25]. While leaf wetness is prohibitory for sporangium production,
a period of near-saturated relative humidity must occur for ≥6 hours to induce
sporulation [26].

Sporulation, as in other downy mildews, is dependent on the diurnal

cycle, and is enhanced by longer photoperiods [27].

Differentiation of sporangia

requires a minimum 6 hour dark period [28]. The optimum temperature for sporangia
production is 15-20°C, but sporangia may form on cucumber at temperatures from 5 to

8

Figure 1.2

Figure 1.2 Morphology of Pseudoperonospora cubensis. (A) Sporangiophore
(bar = 50 µm). (B) Sporangia attached to distal end of sporangiophores. (C)
Sporangia germinating via cytoplasmic cleavage. (D) Left panel: zoospores. Right
panel: encysted zoospore with germ tube. (E) Intercellular growth: S, stomata; H,
haustorium; IH, intercellular hyphae. (F) Scanning electron micrograph (SEM) of
sporangiophore (bar = 20 µm). (G) SEM of multiple sporangiophores emerging
through stomata (bar = 20 µm). Bars = 25 µm, except where noted.

30°C [1]. Other factors such as the host species, cultivar, host nutritional status, and
host age may also affect sporulation [26].

9

DISPERSAL AND SURVIVAL

Ps. cubensis cannot overwinter in geographic locations with killing frosts. Instead, the
pathogen overwinters in areas with mild winter temperatures that permit cucurbit hosts
to be grown year round [29] or in greenhouses [1]. It was recently demonstrated that
Ps. cubensis could infect a perennial member of the Cucurbitaceae, Bryonia dioica, in
the laboratory, and the pathogen could potentially overwinter on this host in Central and
Northern Europe [30]. However, this has not been supported with observations in the
field [31], and it is unknown if B. dioica plays an important role in the life cycle of Ps.
cubensis [30]. While oospores have been observed in both temperate and tropical
regions

including

India,

Japan,

Austria,

Russia,

China,

Italy,

and

Israel

[2,32,33,34,35,36], the production of oospores is very rare [1,2]. The rare occurrence of
thick-walled resting structures, ie., oospores, limits Ps. cubensis survival in the absence
of a living host. It is currently unknown if oospores play an important role in the disease
cycle (Figure 1.3).

As an obligate biotroph Ps. cubensis requires live host tissue for reproduction and
dispersal. Copious asexual sporangia are produced on infected foliage, which may be
liberated to the air following a reduction in relative humidity when hygroscopic twisting
movements of sporangiophores actively release sporangia into air currents [37]. Hence,
airborne Ps. cubensis sporangia concentrations are greater in the morning and early
afternoon when changes in relative humidity and leaf wetness tend to occur [27]. The
distance sporangia travel depends upon where in the canopy the sporangia are

10

Figure 1.3

Figure 1.3 Life cycle of Pseudoperonospora cubensis. (A) Aerially dispersed
lemon-shaped gray-purple sporangia land on the leaf surface and germinate in free
moisture to form biflagellate zoospores. (B) Zoospores swim to and encyst in
stomata, then penetrate the leaf surface via a germ tube. (C) Hyphae colonizes the
mesophyll layer, establishing clavate-branched haustoria within plant cells. (D) The
diurnal cycle triggers sporulation and up to 6 sporangiophores emerge through each
stomate, bearing sporangia at their tips.
Sporangia are dislodged from
sporangiophores by changes in hydrostatic pressure and are picked up by wind
currents that carry them to their next host. (E) Chlorotic, angular lesions bound by
leaf veins are a symptom of Ps. cubensis infection visible on the adaxial leaf surface.
On the lower leaf surface, sporulation is visible (inset). (F) The role of the sexual
stage of Ps. cubensis is unknown.

11

produced, as well as wind conditions as they become airborne [38]. Like other downy
mildews [39,40], Ps. cubensis sporangia may be wind-dispersed over long distances
[1]. As such, it has been proposed that in the U.S., Ps. cubensis overwinters in frostfree areas of the southern states (e.g., Florida and Texas) and spreads northward each
growing season via wind currents [41,42,43]. Likewise, sporangia that infect cucurbits
in Central Europe originate from year-round production areas in Southeast Europe, and
are transported via wind currents [44].

Subsequent local transport of secondary

inoculum occurs primarily via wind, but sporangia also may be dispersed by rain splash
or physical transfer on equipment within a field [1].

Sporangial survival during transport is limited to 1 to 16 days after dispersal [26,45]
depending on temperature, relative humidity [1] and solar radiation [46].

Once a

sporangium lands on a host plant, that sporangium must survive until environmental
conditions are favorable for infection. Sporangial survival is favored by conditions of low
relative humidity, lower temperature and cloudy days [45,46].

INFECTION MECHANISMS

While liberation and dispersal of sporangia occurs under conditions of low leaf moisture,
leaf wetness is necessary for the pathogen to successfully germinate and infect the host
plant. At 15°C, the optimum temperature for infection, at least 2 hours of leaf wetness
are required for infection when high levels of inoculum are present. Sporangia may also
germinate and cause infection at temperatures ranging from 5-28°C [26], but longer

12

periods of leaf wetness are required [22,24] under these conditions or when less
inoculum is present [26]. The incubation period depends on temperature, photoperiod,
inoculum concentration, and leaf wetness duration, and can range from 4 to 12 days
[1,22].

Sporangia germinate via cytoplasmic cleavage, resulting in the release of 2-15 motile,
biflagellate zoospores [2], which preferentially swim to open stomata where they encyst
[47] (Figure 1.2; Figure 1.3). Germ tubes form from encysted zoospores and produce
appressoria. A penetration hypha develops from the appressorium and enters through
the stomatal aperture into the leaf tissue. Hyaline coenocytic hyphae subsequently form
and grow intercellularly through the mesophyll and palisade tissues. Clavate-branched
haustoria are established within mesophyll cells where they invaginate the plant cell
membrane [48,49] (Figure 1.2; Figure 1.3). These specialized structures are the site of
nutrient uptake by the pathogen and delivery of effector proteins that could potentially
function to redirect host metabolism and suppress defense responses [50,51].

PATHOGENICITY AND VIRULENCE

Ps. cubensis has a broad host range, infecting over 49 species in 20 genera within the
Cucurbitaceae, including 19 species in the genus Cucumis [2,52]. Cucumber (Cucumis
sativus L.), melon (Cucumis melo L.), watermelon (Citrullus lanatus Matsum. and
Nakai), and squash (Cucurbita spp.) are the four major food crops that are hosts to Ps.
cubensis. Other cucurbits infected by Ps. cubensis include loofa (Luffa acutangula (L.)

13

Roxb.), bottle gourd (Lagenaria siceraria (Molina) Stand.), wax gourd (Benincasa
hispida (Thunb.) Cogn.) and bitter melon (Momordica charantia L.) [2].

Ps. cubensis isolates show differences in virulence and pathogenicity depending on the
cucurbit variety. To date, six physiological races have been identified in the U.S., Israel
and Japan, and additional evidence suggests that many more exist in Europe
[36,53,54]. Using Citrullus, Cucumis, and Cucurbita spp., Thomas et al. [55] identified
five distinct physiological races of Ps. cubensis: 1 and 2 from Japan, 3 from Israel, and
4 and 5 from the U.S.. In 2003, Cohen et al. identified a sixth physiological race in
Israel based on its pathogenicity to a wider range of susceptible cucurbits compared to
race 3.

All six physiological races that have been described are pathogenic on

cucumber and muskmelon (C. melo var. reticulatus) but show differences in
pathogenicity on watermelon, squash or pumpkin.

Subsequently, Lebeda and

Widrlechner (2003) developed a set of differential taxa that included 12 representatives
from 6 genera, Benincasa, Citrullus, Cucumis, Cucurbita, Lagenaria and Luffa, which
represent natural hosts of Ps. cubensis. Using this set of hosts, the authors evaluated
the differences in pathogenicity of 22 additional isolates from the Czech Republic,
Spain, France and the Netherlands, which were classified as 13 physiological races
based on their virulence [52,53]. Collectively, these studies were the first to describe
differences in virulence and pathogenicity of Ps. cubensis in Europe in detail; however,
the genetic basis for differences among physiological races has not been established.

14

Differences in effector content could be a potential explanation for differences among
physiological races. Effector proteins have been shown to play roles as both virulence
and avirulence determinants in other oomycetes [56,57,58,59].

Oomycete effector

proteins contain the RXLR motif that is located downstream of the signal peptide and
are under diversifying selection at the C-terminal domain [60]. Preliminary sequence
data of the Ps. cubensis genome has yielded 61 putative effector proteins [61,62]. Of
these, 32 were secreted proteins containing the RXLR motif typical of previously
identified oomycete effector proteins, while the remaining 29 had an R to Q substitution
at the first amino acid (ie., QXLR). A family of QXLR containing effectors, PcQNE (Ps.
cubensis QXLR nuclear localized effectors), was shown to localize to the nucleus and
the C-terminal domain was under diversifying selection, as has been observed for
RXLR effectors [61]. Understanding the diversity and role of effector proteins is key to
understanding pathogenicity and the genetic basis for virulence differences between
isolates.

The pathogenic and genetic diversity of Ps. cubensis has been shown to vary
temporally and geographically [2,55,63].

In a study by Lebeda and Urban [3], Ps.

cubensis isolates were collected over a three-year period in the Czech Republic, and a
general population shift from isolates with low pathogenicity to those with high
pathogenicity was observed. This work also demonstrated that the variability of the
population decreased from 33 different physiological races in 2001 to only 13
physiological races in 2003 [3]. An increased representation of a highly pathogenic
physiological race that was capable of causing infection on all cucurbits studied was

15

also observed over the course of the experiments, indicating a population shift to high
pathogenicity isolates [3]. Recently, using a combination of amplified fragment length
polymorphism (AFLP) and ITS sequence analyses, molecular polymorphisms were
identified among populations of Ps. cubensis from Greece, the Czech Republic, the
Netherlands and France [54]. While there was no clear grouping of isolates based on
their pathogenicity, AFLP analysis indicated genetic differentiation between the Greek
isolates and those from central and western Europe [54]. The work by Sarris and
colleagues (2009) suggested that the clustering of Ps. cubensis isolates from Greece
corresponded to their geographical distribution rather than their pathogenicity or
virulence on cucurbit hosts. Nevertheless, this study was supported on SNP evidence
from one single loci and AFLP evidence with error rates of 1-4% due to band scoring. A
similar study with information from several loci, a reliable polymorphism scoring system,
and a more extensive sampling of regions of interest is needed to support or disprove
what Sarris and colleagues (2009) proposed in their paper.

DISEASE MANAGEMENT

Ps. cubensis outbreaks over the past several decades have been responsible for
annual yield losses of up to 80%, and as a result, cucurbit downy mildew is currently the
most destructive disease of cucumbers for both field and greenhouse production in
Europe [3,44]. Before 1984, downy mildew was not a major issue for Central [64,65]
and Northern Europe [66,67].

However, from ~1985 onward, epidemics of downy

mildew have been a challenge for cucurbit production in Europe [65]. In Europe, it has

16

been suggested that tolerant cultivars should be used in conjunction with fungicide
applications when conditions are favorable for downy mildew [68,69]. In the U.S., host
resistance introduced in the 1950s was effective in limiting losses caused by downy
mildew without the use of fungicides. A resurgence of the disease was observed in
many states along the Eastern Seaboard and the upper Midwest in 2004 and 2005,
respectively [9]. This loss in resistance has led to the management of downy mildew
through a spray program, with recommended fungicides applied every five to seven
days on cucumbers [8,70] and every seven to ten days on other cucurbits [71].

An aggressive spray program is essential, as plants must have a protective barrier of
fungicide prior to sporangium deposition to avoid yield losses. However, additional
fungicide applications to control downy mildew greatly increase the cost of production.
In Michigan, for example, major outbreaks of the disease have been observed since
2005, and the cost of additional fungicide sprays is over 6 million dollars annually
(Hausbeck, unpublished data).

Since fungicide applications are expensive, downy

mildew sporangia trapping and forecasts can be useful tools to alert growers to when
airborne sporangia are present or likely to be present so they can make an informed
decision about when to initiate fungicide applications [43]. Delaying the initiation of a
fungicide spray program may reduce management costs for growers and the amount of
fungicides in the environment.

The efficacy of chemical control measures may be diminished if Ps. cubensis
populations develop resistance to key fungicides; Ps. cubensis was the first oomycete

17

with documented resistance to metalaxyl and reduced sensitivity to mancozeb [63,72].
In addition, populations of Ps. cubensis resistant to strobulurin fungicides have been
described [73].

In the Eastern U.S. field and greenhouse fungicide trials, products

containing mefenoxam and strobilurins as the active ingredients failed to provide
adequate downy mildew disease control, indicating that resistance is widespread in this
region [5,8,70,74]. Since Ps. cubensis has a high potential of developing resistance to
fungicides [75], it is important that populations are carefully monitored for resistance to
currently registered products and that new active ingredients are tested.

While fungicide applications are currently necessary for adequate disease control [76],
resistant varieties and cultural techniques are important components of a management
strategy.

The full genetic parameters controlling resistance to downy mildew in

cucumber are unknown. The original source of host resistance (i.e., the recessive dm1
gene) was identified in cucumber accession PI 197087 and first described in India in
1954 [77].

The resistance response governed by dm1 is characterized by sparse

pathogen sporulation, small necrotic lesions, tissue browning, and rapid cell death,
indicative of the classical hypersensitive response (HR) type resistance.

Since the

1950s, resistance conferred by dm1 has been widely used in commercial cultivars for
cucumber production in the U.S. and was sufficient to prevent losses due to downy
mildew until 2004 [9].

Cultivars containing the dm1 gene still show some level of

resistance; unfortunately, the high level of resistance once observed is now lost.
Additionally, susceptible cultivars without the dm1 gene become infected earlier in the
season, and exhibit more severe damage than was previously observed [43].

18

While the dm1 remains useful in a disease management program, a robust source of
resistance is highly desirable.

Current breeding research for resistance to downy

mildew in cucumber is focused on the identification of resistant germplasm(s) and
cultivars via large scale screening trials [78,79]. Tolerant and high yielding germplasm
has been identified in these studies, but a source of complete resistance to downy
mildew in cucumber has been elusive, likely due to limited genetic diversity for Ps.
cubensis resistance in cucumber [52,79]. Other Cucumis spp., such as melon, may be
more relevant for identifying effective sources of resistance [52]. Wild melon line PI
124111F [PI], for example, has been shown to be resistant to the 6 physiological races
of Ps. cubensis via two constitutively expressed glyoxylate aminotransferase-encoding
genes, At1 and At2 [80]. These two genes are known as enzymatic resistance (eR)
genes which when expressed at high levels in either wild type or transgenic plants,
confer complete resistance to infection by Ps. cubensis.

FUTURE PROSPECTS

The re-emergence of cucurbit downy mildew in the U.S. and its persistence across
much of Europe and Asia, represents a significant threat to cucurbit production
worldwide. While the disease was successfully managed for decades within the U.S.
using host resistance (i.e., the dm-1 locus), severe epidemics have occurred since
2004. Whether this is due to a change in pathogen populations or a change in the
environment is currently unknown. To this end, we need to investigate the changes in

19

pathogen populations, environmental factors, and how the pathogen-environment
interaction affects the host-pathogen interaction and disease development.
Studies have investigated some aspects of the basic biology of Ps. cubensis and
its interactions with various cucurbit hosts. However, additional research is needed to
further clarify the taxonomy, variations in virulence and pathogenicity among
physiological races, and the pathogenicity determinants of the pathogen.

A better

understanding of each of these components will ultimately facilitate the development of
durable host resistance. To this end, the forthcoming genome sequence will provide
molecular tools for gene discovery and the development of molecular markers, which
may then be used to investigate population and evolutionary biology of the pathogen.
Such studies will yield information about possible migration events or evolutionary
changes within pre-existing U.S. pathogen populations that resulted in strains with
increased virulence.

An integrated research approach that includes all factors affecting disease development
(pathogen, host, and environment) is essential to control and predict future cucurbit
downy mildew epidemics. First, it would benefit growers if new fungicides that are more
economical and provide effective control were identified. Second, screening of breeding
lines and wild germplasm will help identify durable sources of genetic resistance that
would be preferable to an aggressive spray program. Finally, studies to determine the
effects of environment on inocula and disease development will serve as a first step in
developing rapid, and highly specific forecasting systems. In summary, research efforts
contributing to the development of sustainable management strategies such as durable

20

host resistance are a priority to ensure the long-term viability of the cucurbit production
industry.

ACKNOWLEDGEMENTS

The authors thank members of the Day lab for critical reading of the manuscript.
Cucurbit downy mildew research in the Day lab is funded by the Michigan Agricultural
Experiment Station (MAES), Project GREEEN (Award numbers GR06-0099D and
GF07-077), the Michigan State University Office of the Vice President for Research and
Graduate Studies, the National Science Foundation (Award number IOS-0641319), and
a joint grant awarded to MKH and BD from the Agricultural Research Fund of Pickle
Packers International Inc.. Work in the Hausbeck lab is supported by the Pickle and
Pepper Research Committee of MSU, Fresh Vegetable Growers of Ontario, North
Central IPM Center (Sub award 2003-51120-02111 S4256), and Project GREEEN
(Award Numbers GR07-077 and GR06-0099D).

21

REFERENCES

22

REFERENCES
1. Thomas CE (1996) Downy mildew. In: Zitter TA, editor. Compendium of cucurbit
diseases. Ithaca, NY: Cornell University Press. pp. 25-27.
2. Palti J, Cohen Y (1980) Downy mildew of cucurbits (Pseudoperonospora cubensis):
the fungus and its hosts, distribution, epidemiology, and control. Phytoparasitica
8: 109-147.
3. Lebeda A, Urban J (2007) Temporal changes in pathogenicity and fungicide
resistance in Pseudoperonospora cubensis populations. Acta Hort 731: 327-336.
4. Anonymous (2009) National Online Statistics. US Dep Agric Natl Agric Stat Serv.
5. Colucci SJ, Holmes GJ (2007) Fungicide insensitivity and pathotype determination of
Pseudoperonospora cubensis, causal agent of cucurbit downy mildew.
Phytopathology 97: S24.
6. Zhu S, Liu P, Liu X, Li J, Yuan S, et al. (2008) Assessing the risk of resistance in
Pseudoperonospora cubensis to the fungicide flumorph in vitro. Pest
Management Science 64: 255-261.
7. Mitani S, Araki S, Yamaguchi T, Takii Y, Ohshima T, et al. (2001) Biological
properties of the novel fungicide cyazofamid against Phytophthora infestans on
tomato and Pseudoperonospora cubensis on cucumber. Pest Manag Sci 58:
139-145.
8. Hausbeck MK, Cortright BD (2009) Evaluation of fungicides for control of downy
mildew of pickling cucumber, 2007. PDMR 3: V112.
9. Holmes GJ, Thomas CE (2006) The history and re-emergence of cucurbit downy
mildew. Phytopathology 99: S171.
10. Holmes G, Wehner T, Thornton A (2006) An old enemy re-emerges. American
Vegetable Grower Feb.: 14-15.
11. Choi YJ, Hong SB, Shin HD (2005) A re-consideration of Pseudoperonospora
cubensis and P. humuli based on molecular and morphological data. Mycol Res
109: 841-848.
23

12. Waterhouse GM, Brothers MP (1981) The taxonomy of Pseudoperonospora.
Mycological Papers 148: 1-18.
13. Berkeley MS, Curtis A (1868) Peronospora cubensis. J Linn Soc Bot 10: 363.
14. Rostovzev SI (1903) Beitrage zur Kenntnis der Peronosporeen. Flora 92: 405-430.
15. Mitchell MN, Ocamb C, Gent D (2009) Addressing the relationship between
Pseudoperonospora cubensis and Pseudoperonospora humuli by multigenic
characterization and host specificity. Phytopathology 99: S87.
16. Gent DH, Mitchell MN, Holmes GJ (2009) Genetic and pathogenic relatedness of
Pseudoperonospora cubensis and P. humuli: Implications fo detection and
management. Phytopathology 99: S171.
17. Runge F, Thines M (In Press) Host matrix has major impact on the morphology of
Pseudoperonospora cubensis. Eur J Plant Pathol: 1-10.
18. Iwata Y (1942) Specialization in Pseudoperonospora cubensis (Berk. et Curt.)
Rostov. II. Comparative studies of the morphologies of the fungi from Cucumis
sativus L. and Cucurbita moschata Duchesne. Ann Phytopathol Soc Japan 11:
172-185.
19. Oerke EC, Steiner U, Dehne HW, Lindenthal M (2006) Thermal imaging of
cucumber leaves affected by downy mildew and environmental conditions. J Exp
Bot 57: 2121-2132.
20. Cohen Y, Rotem J (1971) Rate of lesion development in relation to sporulating
potential of Pseudoperonospora cubensis in cucumbers. Phytopathology 61:
265-268.
21. Lindenthal M, Steiner U, Dehne HW, Oerke EC (2005) Effect of downy mildew
development on transpiration of cucumber leaves visualized by digital infrared
thermography. Phytopathology 95: 233-240.
22. Cohen Y (1977) The combined effects of temperature, leaf wetness, and inoculum
concentration on infection of cucumbers with Pseudoperonospora cubensis. Can
J Bot 55: 1478-1487.
24

23. Keinath AP, Holmes GJ, Everts KL, Egel DS, Langston DB (2007) Evaluation of
combinations of chlorothalonil with azoxystrobin, harpin, and disease forecasting
for control of downy mildew and gummy stem blight on melon. Crop Protection
26: 83-88.
24. Rotem J, Cohen Y, Bashi E (1978) Host and environmental influences on
sporulation in vivo. Ann Rev Phytopathol 16: 83-101.
25. Palti J (1975) Pseudoperonospora cubensis (Berk & M. A. Curtis) Rost. C M I Descr
Pathog Fungi Bact 457: 1-2.
26. Cohen Y (1981) Downy mildew of cucurbits. In: Spencer DM, editor. The Downy
Mildews. London: Academic Press. pp. 341-354.
27. Cohen Y, Rotem J (1971) Field and growth chamber approach to epidemiology of
Pseudoperonospora cubensis on cucumbers. Phytopathology 61: 736-737.
28. Cohen Y (1977) Growth and differentiation of sporangia and sporangiophores of
Pseudoperonospora cubensis on cucumber cotyledons under various
combinations of light and temperature. Physiol Plant Pathol 10: 93-103.
29. Bains SS, Jhooty JS (1976) Over wintering of Pseudoperonospora cubensis causing
downy mildew of muskmelon. Indian Phytopathol 29: 213-214.
30. Runge F, Thines M (2009) A potential perennial host for Pseudoperonospora
cubensis in temperate regions. Eur J Plant Pathol 123: 483-486.
31. Lebeda A, Cohen Y (In Press) Cucurbit downy mildew (Pseudoperonospora
cubensis)—biology, ecology, epidemiology, host-pathogen interaction and
control. Eur J Plant Pathol: 1-36.
32. Bains SS, Sokhi SS, Jhooty JS (1977) Melothria maderaspatana - a new host of
Pseudoperonospora cubensis. Indian J Mycol Plant Pathol 7: 86.
33. Hiura M, Kawada S (1933) On the overwintering of Peronoplasmopara cubensis Jap
J Bot 6: 507-513.

25

34. D'Ercole N (1975) La peronoospora del cetriolo in coltura protetta. Inftor Fitopath 25:
11-13.
35. Bedlan G (1989) First detection of oospores of Pseudoperonospora cubensis (Berk
et Curt.) Rost. on glasshousse cucumbers in Austria. Pflantzenschutzberichte 50:
119-120.
36. Cohen Y, Meron I, Mor N, Zuriel S (2003) A new pathotype of Pseudoperonospora
cubensis causing downy mildew in cucurbits in Israel. Phytoparasitica 31: 458466.
37. Lange L, Eden U, Olson LW (1989) Zoosporogenesis in Pseudoperonospora
cubensis, the causal agent of cucurbit downy mildew. Nord J Bot 8: 497-504.
38. Aylor DE (1990) The role of intermittent wind in the dispersal of fungal pathogens.
Ann Rev Phytopathol 28: 73-92.
39. Aylor DE, Taylor GS (1982) Aerial dispersal and drying of Peronospora tabacina
conidia in tobacco shade tents. Proc Natl Acad Sci USA 79: 697-700.
40. Wu BM, van Bruggen AHC, Subbarao KV, Pennings GGH (2001) Spatial analysis of
lettuce downy mildew using geostatistics and geographic information systems.
Phytopathology 91: 134-142.
41. Doran WL (1932) Downy mildew of cucumbers. Massachusetts Agr Exp Sta Bull
283.
42. Van Haltern F (1933) Spraying cantaloupes for the control of downy mildew and
other diseases. Georgia Exp Sta Bull.
43. Holmes GJ, Main CE, Keever ZT (2006) Cucurbit downy mildew: a unique
pathosystem for disease forecasting. In: Spencer-Phillips PTN, Jeger M, editors.
Advances in downy mildew research. Dordrecht, The Netherlands: Kluwer
academic publishers.
44. Lebeda A, Schwinn FJ (1994) The downy mildews - an overview of recent research
progress. J Plant Dis Protection 101: 225-254.

26

45.

Cohen Y, Rotem J (1971) Dispersal and viability of
Pseudoperonospora cubensis. Trans Br Mycol Soc 57: 67-74.

sporangia

of

46. Kanetis L, Holmes GJ, Ojiambo PS (2009) Survival of Pseudoperonospora cubensis
sporangia exposed to solar radiation. Plant Pathology 59: 313-323.
47. Iwata Y (1949) Studies on the invasion of cucumber plants by downy mildew. (In
Japanese, with English summary). Ann Phytopathol Soc Japan 13: 60-61.
48. Voglmayr H, Riethmuller A, Goker M, Weiss M, Oberwinkler F (2004) Phylogenetic
relationships of Plasmopara, Bremia and other genera of downy mildew
pathogens with pyriform haustoria based on bayesian analysis of partial LSU
rDNA sequence data. Mycol Res 108: 1011-1024.
49. Fraymouth J (1956) Haustoria of the Peronosporales. Trans Br Mycol Soc 39: 79107.
50. Hahn M, Mendgen K (2001) Signal and nutrient exchange at biotrophic plant-fungus
interfaces. Curr Opin Plant Biol 4: 322-327.
51. Whisson SC, Boevink PC, Moleleki L, Avrova AO, Morales JG, et al. (2007) A
translocation signal for delivery of oomycete effector proteins into host plant cells.
Nature 450: 115-118
52. Lebeda A, Widrlechner MP (2003) A set of Cucurbitaceae taxa for differentiation of
Pseudoperonospora cubensis pathotypes. J Plant Dis Protection 110: 337-349.
53. Lebeda A, Gadasova V (2002) Pathogenic variation of Pseudoperonospora
cubensis in the Czech Republic and some other European countries. Acta Hort
588: 137-141.
54. Sarris P, Abdelhalim M, Kitner M, Skandalis N, Panopoulos N, et al. (2009)
Molecular polymorphisms between populations of Pseudoperonospora cubensis
from Greece and the Czech Republic and the phytopathological and phylogenetic
implications. Plant Pathology Doi: 10.111/j.1365-3059.2009.02093.x.
55. Thomas CE, Inaba T, Cohen Y (1987) Physiological specialization in
Pseudoperonospora cubensis. Phytopathology 77: 1621-1624.
27

56. Schornack S, Huitema E, Cano LM, Bozkurt TO, Oliva R, et al. (2009) Ten things to
know about oomycete effectors. Molecular Plant Pathology 10: 795-803.
57. Hogenhout SA, Van der Hoorn RAL, Terauchi R, Kamoun S (2009) Emerging
concepts in effector biology of plant-associated organisms. Mol Plant-Microbe
Interact 22: 115-122.
58. Oliva R, Win J, Raffaele S, Boutemy L, Bozkurt TO, et al. Recent developments in
effector biology of filamentous plant pathogens. Cellular Microbiology 12: 705715.
59. Thines M, Kamoun S (2010) Oomycete-plant coevolution: recent advances and
future prospects. Curr Opin Plant Biol 13: 427-433.
60. Win J, Morgan W, Bos J, Krasileva K, Cano L, et al. (2007) Adaptive evolution has
targeted the C-terminal domain of the RXLR effectors of plant pathogenic
oomycetes. The Plant Cell 19: 2349-2369.
61. Tian M, Win J, Savory EA, Kamoun S, Day B (2011) 454 genome sequencing of
Pseudoperonospora cubensis reveals effector proteins with a putative QXLR
translocation motif. Mol Plant-Microbe Interact 24.
62. Savory EA, Tian M, Win J, Kamoun S, Day B (2009) Genome characterizarion and
discovery of novel QXLR effector motif in the cucurbit downy mildew pathogen
Pseudoperonospora cubensis 14th International ISM-MPMI Congress. Quebec,
Canada.
63. Thomas CE, Jourdain EL (1992) Host effect on selection of virulence factors
affecting sporulation by Pseudoperonospora cubensis. Plant Dis 76: 905-907.
64. Lebeda A (1986) Epidemic occurrence of Pseudoperonospora cubensis in
Czechoslovakia. Temperate Downy Mildews Newsletter 4: 15-17.
65. Lebeda A, Schwinn FJ (1994) The downy mildews--an overview of recent research
progress. J Plant Dis Protection 101: 225-254.
66. Forsberg AS (1986) Downy mildew-Pseudoperonospora cubensis in Swedish
cucumber fields. Vaxtskyddsnotiser 50: 17-19.
28

67. Tahvonen R (1985) Downy mildew of cucurbits found for the first time in Finland.
Vaxtskyddsnotiser 49: 42-44.
68. Chaban VS, Okhrimchuk VN, Sergienko VG (2000) Optimization of chemical control
of Pseudoperonospora cubensis on cucumber in Ukraine. EPPO Bulletin 30:
213-215.
69. Urban J, Lebeda A (2006) Fungicide resistance in cucurbit downy mildewmethodological, biological, and population aspects. Annals of applied biology
149: 63-75.
70. Gevens AJ, Hausbeck MK (2006) Control of downy mildew of cucumbers with
fungicides, 2005. F & N Tests: V062.
71. Hausbeck MK. Downy mildew and Phytophthora control in vine crops; 2009;
Syracuse, NY. pp. 193-195.
72. Reuveni M, Eyal H, Cohen Y (1980) Development of resistance to Metalaxyl in
Pseudoperonospora cubensis Plant Disease 64: 1108-1108.
73. Heaney SP, Hall AA, Davis SA, Olaya G. Resistance to fungicides in the QoI-STAR
cross-resistance group: current perspectives; 2000. pp. 755-762.
74. Keinath AP, DuBose VB, Lassiter AW (2008) Evaluation of fungicides to manage
downy mildew on pickling cucumber in Charleston, South Carolina. Plant
Disease Management Reports 2: V024.
75. Russell PE (2004) Sensitivity baselines in fungicide resistance research and
management. FRAC Monograph 3.
76. Gisi U (2002) Chemical control of downy mildews. In: Spencer-Phillips PTN, Gisi U,
Lebeda A, editors. Advances in downy mildew research. Dordrecht, The
Netherlands: Kluwer academic publishers. pp. 119-159.
77. Barnes WC, Epps WM (1954) An unreported type of resistance to cucumber downy
mildew. Plant Disease Rep 38: 620.

29

78. Wehner TC, Shetty NV (1997) Downy mildew resistance of the cucumber
germplasm collection in North Carolina field tests. Crop Science 37: 1331-1340.
79. Shetty NV, Wehner TC, Thomas CE, Doruchowski RW, Vasanth Shetty KP (2002)
Evidence for downy mildew races in cucumber tested in Asia, Europe, and North
America. Scientia Horticulturae 94: 231-239.
80. Taler D, Galperin M, Benjamin I, Cohen Y, Kenigsbuch D (2004) Plant eR genes
that encode photorespiratory enzymes confer resistance against disease. Plant
Cell 16: 172-184.

30

CHAPTER 2

Alternative splicing of a multi-drug transporter from Pseudoperonospora
cubensis generates an RXLR effector protein that elicits a rapid cell death

This chapter was originally published in PLoS ONE.
Savory EA, Zou C, Adhikari BN, Hamilton JP, Buell CR, Shiu S-H, and Day B (2012)
Alternative splicing of a multi-drug transporter from Pseudoperonospora cubensis. PLoS
ONE 7(4): e34701. doi:10.1371/journal.pone.0034701.
© 2012 Savory et al. This is an open-access article distributed under the terms of the
Creative Commons Attribution License, which permits unrestricted use, distribution, and
reproduction in any medium, provided the original author and source are credited.
Author Contributions:
Conceived and designed the experiments: EAS, S-HS, CRB, and BD. Performed the
experiments: EAS, CZ, and BNA. Analyzed the data: EAS, CZ, BNA, JPH, S-HS, CRB,
and BD. Contributed reagents/materials/analysis tools: EAS, CZ, BNA, and JPH. Wrote
the paper: EAS and BD.

31

ABSTRACT

Pseudoperonospora cubensis, an obligate oomycete pathogen, is the causal agent of
cucurbit downy mildew, a foliar disease of global economic importance. Similar to other
oomycete plant pathogens, Ps. cubensis has a suite of RXLR and RXLR-like effector
proteins, which likely function as virulence or avirulence determinants during the course
of host infection. Using in silico analyses, we identified 271 candidate effector proteins
within the Ps. cubensis genome with variable RXLR motifs. In extending this analysis,
we present the functional characterization of one Ps. cubensis effector protein, RXLR
protein 1 (PscRXLR1), and its closest Phytophthora infestans ortholog, PITG_17484, a
member of the Drug/Metabolite Transporter (DMT) superfamily. To assess if such
effector-non-effector pairs are common among oomycete plant pathogens, we
examined the relationship(s) among putative ortholog pairs in Ps. cubensis and P.
infestans. Of 271 predicted Ps. cubensis effector proteins, only 109 (41%) had a
putative ortholog in P. infestans and evolutionary rate analysis of these orthologs shows
that they are evolving significantly faster than most other genes. We found that
PscRXLR1 was up-regulated during the early stages of infection of plants, and
moreover, that heterologous expression of PscRXLR1 in Nicotiana benthamiana elicits
a rapid necrosis. More interestingly, we also demonstrate that PscRXLR1 arises as a
product of alternative splicing, making this the first example of an alternative splicing
event in plant pathogenic oomycetes transforming a non-effector gene to a functional
effector protein. Taken together, these data suggest a role for PscRXLR1 in
pathogenicity, and in total, our data provide a basis for comparative analysis of

32

candidate effector proteins and their non-effector orthologs as a means of
understanding function and evolutionary history of pathogen effectors.

INTRODUCTION

The identification and characterization of secreted effector proteins from plant
pathogens has anchored the recent evolution of molecular plant pathology [1,2,3]. As
components of many pathogenic microorganisms’ secretomes, effector proteins
represent a key component of phytopathogenicity, contributing to both the virulence and
avirulence capacity of the invading pathogen. Numerous studies have identified and
characterized the activities of secreted effector proteins from a broad range of
phytopathogens [reviewed in 1,4,5]. Collectively, these works have revealed two
primary functions for pathogen effector molecules. First, as virulence molecules, effector
proteins can enhance a pathogen’s ability to cause disease, likely through abrogating
host processes that would otherwise block pathogen infection, growth, and proliferation
within the host [5,6]. Secondly, as avirulence determinants, effector proteins function to
modulate the activation of host defense responses by perturbing the activity of host
resistance (R) proteins [5,6].

For infection, colonization, and subsequent propagation within their hosts to occur,
pathogens must dampen multiple layers of plant defense responses. Often described as
basal resistance, the initial perception and elicitation of defenses requires the
recognition of pathogen associated molecular patterns (PAMPs; e.g., chitin, flagellin,

33

LPS) [5,6], highly conserved molecules essential for the lifestyle and survival of the
microorganism. The recognition of PAMPs, which are highly specific elicitors, occurs
through receptors on the host membrane surface, and following initiation of this
receptor-ligand interaction, a rapid first response known as PAMP-triggered immunity
(PTI) is elicited [5,6]. Overall, the PTI response provides a basal level of resistance
against a wide range of microorganisms, often utilizing conserved signaling pathways
such as the up-regulation of the mitogen-activated protein kinase (MAPK) pathway, the
generation of reactive oxygen species, and the induction of defense-related genes [5,6].
To overcome PTI, phytopathogens, including bacteria and oomycetes, rely on the
delivery and activity of secreted effector proteins to abrogate this initial basal level of
defense, as well as to further promote virulence [3,5,6]. In response, pathogen effectors
can be recognized by R (resistance) proteins, leading to the activation of effectortriggered immunity (ETI) [5,6] best illustrated as an amplified and sustained layer of
defense. ETI is a robust response that is often associated with the activation of a
specific type of programmed cell death referred to as the hypersensitive response (HR)
[5,6]. Over time, as this cycle of subversion and recognition evolves, host specificity and
subsequent interactions between pathogen and host are modulated by the interplay
between the activity and recognition of secreted pathogen effector molecules and their
host counterparts.

Oomycetes are a phylogenetically distinct eukaryotic lineage within the Stramenopiles,
which as a group, are among the best-studied and most economically important plant
pathogens. In recent years, the genomes of several agriculturally important oomycete

34

pathogens have been sequenced, including Phytophthora infestans, Phytophthora
ramorum, and Phytophthora sojae, the causal agents of late blight of potato and tomato,
sudden oak death, and soybean root rot, respectively [7,8]. The genomes of two other
oomycete pathogens, Pythium ultimum, which causes damping off and root rot on a
wide range of hosts, and Hyaloperonospora arabidopsidis, a pathogen of Arabidopsis
thaliana, have also been sequenced [9,10]. These investigations, through the analysis
of genome content and structure, have provided a wealth of information, both towards
understanding the nature of the host-pathogen interaction (e.g., host specificity,
virulence strategies), as well as insight into the evolution of the interaction itself. Central
to the analysis of phytopathogen genomes, the identification and characterization of
oomycete effector proteins has moved swiftly into the forefront in the field of plantpathogen interactions, due in large part to the aforementioned available genomic
resources. At a primary level, the identification of a highly conserved N-terminal
translocation motif (i.e., RXLR; Arg-X-Leu-Arg, where “X” is any amino acid)
demonstrated to be necessary for effector delivery into host cells, has been a seminal
discovery in the field of plant-oomycete interactions [11,12]. Similar in function to
phytopathogenic bacterial effector proteins, oomycete RXLR-containing effectors have
been demonstrated to suppress PTI [13], as well as to activate ETI [11,14,15,16,17,18].
Structurally, oomycete effector proteins display a modular organization, consisting of a
N-terminal signal peptide, a conserved RXLR translocation motif, followed by a variable
C-terminal effector domain [3]. It is the function and activity of the variable C-terminal
effector domain that drives the activity of these molecules [3,4].

35

Alternative splicing (AS) of pre-mRNA drives the generation of multiple protein isoforms
through assembly of different combinations of splice sites within a single gene. In total,
this process represents a conserved mechanism found in eukaryotes which drives
proteome complexity within organisms with a finite number of genes [19]. In oomycetes,
there are few reports of intron processing [20,21,22], and to date, these analyses has
been strictly in silico [20,22], with little functional validation [21]. Costanzo et al. [21]
characterized alternative processing in P. sojae family 5 endoglucanases revealing the
generation of both coding and non-coding RNA isoforms. Additionally, based on their
large-scale analysis of intronic structure and alternative splicing in P. sojae, Shen and
colleagues [22] validated splice variants leading to premature translation termination.

Ps. cubensis is an obligate biotrophic oomycete pathogen of cucurbits (i.e., cucumber,
melon, squash, watermelon, etc.), and is the causal agent of cucurbit downy mildew, an
economically important foliar disease [23]. Capable of rapid defoliation of fields in short
periods of time (i.e., <10 days), Ps. cubensis is the primary factor limiting cucurbit
production in the United States. Despite obvious economic importance, very little is
known about the genetic determinants of virulence and pathogenicity of Ps. cubensis,
as well as the molecular-genetic basis of resistance in the cucurbits.

Similar to related oomycete pathogens of plants, Ps. cubensis possesses a suite of
effector proteins that likely function to promote virulence and suppress host defense
responses [3,24]. Recent work by Tian et al. [24] identified and characterized a
preliminary set of effector proteins from a draft genome sequence of Ps. cubensis

36

obtained using 454 pyrosequencing. In brief, this set of 61 candidate effectors included
a large class of variants with sequence similarity to the canonical RXLR motif found in
other oomycete plant pathogens [24]. Specifically, this work characterized the function
of a QXLR-containing effector, designated PcQNE, which was shown to be a member of
a large family of Ps. cubensis QXLR nuclear-localized effectors, up-regulated during
infection of cucumber. Additionally, internalization of PcQNE was shown to require the
QXLR-EER motif, thereby establishing a basic homology to the well-characterized
Phytophthora spp. effector proteins.

In the current study, we describe the identification and evolutionary potential of the Ps.
cubensis effector repertoire. First, through characterization of a RXLR effector protein,
PscRXLR1, we investigated the localization and in planta activity, and similarly to some
oomycete effector proteins described to date, PscRXLR1 induces a rapid cell death
response when delivered into plant cells. Additionally, using whole transcriptome
sequencing analyses, as well as RT-PCR, we show that PscRXLR1 is a product of
alternative splicing of the Psc_781.4 gene which encodes a putative multi-drug
transporter. Coupled with the induction and expression of PscRXLR1 mRNA during Ps.
cubensis infection of cucumber, as well as a complement of bioinformatic, cell biology
and in vivo analyses, we provide evidence suggestive of a virulence role for PscRXLR1.
Finally, we used PscRXLR1 as template for assessing the conservation and
evolutionary potential of oomycete effector proteins from Ps. cubensis, identifying and
analyzing orthologous pairs of Ps. cubensis effector proteins and P. infestans noneffector proteins. Using more robust methods, we identified additional candidate

37

effectors from Ps. cubensis for these analyses and showed that, like other oomycete
effectors, they tend to be influenced by positive selection. Assessment of evolutionary
rate and conservation of secretion signals between orthologous pairs revealed that Ps.
cubensis effectors are undergoing adaptive evolution and conservation of signal
peptides are similar among effector and non-effector proteins in Ps. cubensis. Overall,
our study provides support for the investigation of relationships among oomycete
effectors and their non-effector orthologs, and in total, the analysis presented herein
establishes a foundation for understanding the evolution of effector repertoires and
host-pathogen specificity.

RESULTS

Genome sequencing of Ps. cubensis

Next generation sequencing with the Illumina Genome Analyzer II platform was used to
generate an assembly of the Ps. cubensis MSU-1 genome. A total of 4.5 Gb of cleaned
paired end reads from two libraries were used to generate the assembly using Velvet, a
de novo short read assembler [25]. The final assembly contains 35,546 contigs with an
N50 contig size of 4.0 Kbp representing 64.4 Mbp. Protein coding genes in the draft
assembly were annotated using MAKER [26] which incorporated ab initio gene
predictions, protein evidence, and transcript evidence from other sequenced oomycete
genomes. In total, 23,519 loci and 23,522 gene models were predicted.

38

Identification of the Ps. cubensis effector repertoire

Our initial analysis of the effector complement of Ps. cubensis in an earlier draft
assembly [24] identified 61 sequences containing the conserved RXLR, or novel QXLR,
motif found in known oomycete effector proteins. This number is significantly less than
the effector count predicted for other plant pathogenic oomycetes (i.e., 563 effectors in
P. infestans, 396 in P. sojae, 374 in P. ramorum, and 134 in H. arabidopsidis; [7,10,27]),
and is likely the result of limited coverage generated from an initial 454 pyrosequencing
[24]. Generation of genomic sequences using the Illumina Genome Analyzer platform
and their subsequent assembly generated a more comprehensive dataset. Using this
resource, 269 additional sequences were identified as putative effector proteins.
Interestingly, the putative Ps. cubensis effectors showed more variation at the R1
position of the RXLR motif than previously shown [24], with 18 amino acids predicted at
the R1 position, in addition to R and Q (Table S2.1; Available at www.plosone.org,
e34701). Moreover, we have evidence for expression for at least one predicted effector
with any one of 19 amino acids (except Y, Tyr) at position R1, during the course of
infection on a susceptible cucumber cultivar [28], supporting the hypothesis of an
expanded translocation motif repertoire in Ps. cubensis. In total, including the previously
characterized PcQNE, the current predicted effector complement of Ps. cubensis
contains 271 members.

39

Nature of selection on Ps. cubensis paralogs

Based on comparative genomic analyses of several oomycete plant pathogens, positive
selection has been postulated to act disproportionately on effectors gene[s] compared
to other genes in the genome [27,29]. To this end, we examined the strength of
selection acting upon the predicted effector complement of Ps. cubensis by estimating
ω, the ratio of the non-synonymous substitution rate (Ka) to the synonymous
substitution rate (Ks). Among all Ps. cubensis effector paralogs, the median ω is 0.54,
which is significantly higher than that of Ps. cubensis paralogous genes in general (ω =
0.24, Wilcox Rank Sum Test, p < 2.2e-16). For comparison, we also examined P.
infestans effectors and arrived at the same conclusion (Wilcox Rank Sum Test, p <
4.0e-14). Because more recent duplicates tend to have elevated ω, we examined if the
higher ω values among effector paralogs can be attributed to recent duplication. We
found that, using Ks as a proxy of time, the ω values for effector pairs are in general
significantly higher than other paralogs in Ps. cubensis in a Ks bin (Figure 2.1A). Thus,
the elevated ω values among effectors are not exclusively due to relaxation of selection
among recent duplicates. The results for P. infestans effectors are similar (Figure 2.1B),
although the ω values of P. infestans are in general higher than those in Ps. cubensis.

Taken together, Ps. cubensis effectors either have experienced a significantly lower
degree of selective constraints, or tend to be positively selected. Consistent with the
latter, 6.3% of Ps. cubensis effector paralog pairs have ω > 1, compared to 3.2% of all
other paralogous gene pairs. In parallel to our observations in Ps. cubensis, 4.6% of P.

40

Figure 2.1

41

Figure 2.1 (cont'd) Strength of purifying selection on Pseudoperonospora
cubensis effectors. Comparison of selective constraints on effector paralogs (blue)
and all other genes (red) in (A) Ps. cubensis and (B) P. infestans. Frequency
distributions (C) of ω, the ratio between the nonsynonymous substitution rate (Ka) and
the synonymous substitution rate (Ks) of Ps. cubensis and P. infestans sequence pairs.
Distributions of Ks values (D) of Ps. cubensis and P. infestans sequence pairs. Green
symbols indicate Ps. cubensis effector – P. infestans non-effector ortholog pairs. Blue
symbols represent other orthologous gene pairs between the two species.

infestans effector paralog pairs have ω > 1, compared to 3.2% for all other paralogs.
Although there is no clear evidence suggesting that most effectors are subjected to
positive selection, it is interesting that even among relatively ancient effector duplicates,
the rate of evolution among effectors is significantly higher than most genes. Given that
older duplicates that survive for tens to hundreds of millions of years tend to be
subjected to substantially stronger selective constraints than young duplicates [30], this
would suggest that, perhaps, effectors function in a way that do not require as strong a
constraint on their primary sequence. Alternatively, it is also possible that pathogen
effectors, even those having undergone ancient duplication events, experience some
degree of continuous positive selection.

Relationship between Ps. cubensis effectors and their P. infestans orthologs

Subsequent in silico analysis of candidate Ps. cubensis effectors and comparisons to
annotated genes in P. infestans revealed that there were a number of orthologs
between Ps. cubensis effector proteins and both effector and non-effector genes in P.
infestans. We hypothesized that this scenario (i.e., effector with non-effector ortholog)
42

may provide a foundation for the analysis of the evolution of effectors from non-effector
proteins. Therefore, we identified orthologous pairs of predicted Ps. cubensis effector
proteins and their non-effector counterparts in P. infestans considering sequence
similarity and synteny (see Materials and Methods). With this approach, 11,601
orthologous gene pairs were identified between Ps. cubensis and P. infestans for
comparison. Of 271 Ps. cubensis effector sequences, 109 had a predicted ortholog in P.
infestans (TABLE S2.2; Available at www.plosone.org, e34701). As shown in Figure
2.1C, the Ps. cubensis effector P. infestans (PscE-Pi) ortholog pairs have significantly
higher ω values as compared to the baseline pairs (Kolmogorov-Smirnov test, p < 7.9e06), consistent with what was found with the effector paralogs (Figure 2.1A).
Additionally, the distribution of ω for the PscE-Pi pairs appears multi-modal. Given that
the first effector ortholog peak (at ω ~0.15) is mostly overlapping with that of the other
orthologs, these effector paralogs are more highly conserved. The second peak at ω
~0.3 likely indicates the presence of a group of effectors that are more quickly evolving
(Figure 2.1C). However, we cannot rule out the possibility that these peaks are present
simply due to the small effector ortholog sample size.

To determine if the overall higher ω value among effector orthologs is an artifact due to
mis-identification of orthologous genes, we examined if putative effectors, as well as the
other orthologs, have similar "age". As shown in Figure 2.1D, the distributions of Ks
values for the effector and the other orthologs are highly similar and are statistically
indistinguishable. Thus, mis-identified orthologous pairs likely do not significantly impact
our findings.

43

Signal peptide conservation among ortholog pairs

Signal peptides are essential components of oomycete effector proteins, as they are
required for translocation of the protein from the pathogen haustorium to the
extrahaustorial matrix prior to uptake by the host cell membrane [3]. As such, all 109 of
the Ps. cubensis effector sequences in the PscE-Pi dataset are predicted to have signal
peptides (Figure S2.1). However, only 71 (65%) of the corresponding P. infestans
orthologs were predicted to be secreted proteins. For comparison, predictions of signal
peptides were made for 10,383 of the 11,601 ortholog pairs. Of these, there were 688
(6.63%; Psc-sec/Pi-sec) ortholog pairs where both members were predicted to have
signal peptides, 428 (4.12%; Psc-sec/Pi-non) pairs where the Ps. cubensis protein was
predicted to be secreted and the P. infestans ortholog was not, and 622 (5.99%; Pscnon/Pi-sec) where the Ps. cubensis sequence did not have a predicted signal peptide
and its corresponding P. infestans sequence was predicted to be secreted. Additionally,
there were 8,645 (83.3%; Psc-non/Pi-non) ortholog pairs where neither member was
predicted to be secreted. For statistical analysis, the Psc-sec/Pi-sec and Psc-sec/Pi-non
datasets from the Ps. cubensis effector-P. infestans orthologs were compared to their
respective genome-wide datasets. Using the Fisher’s exact test, no significant
difference (p = 0.5354) was identified between the two datasets, indicating that
presence of signal peptide prediction is not a good indicator of potential selection for
effector peptide evolution.

44

Identification of Pseudoperonospora cubensis effector PscRXLR1

Using the RXLR effector identification pipeline [29], we previously identified 61
candidate effector protein sequences from a draft genome assembly of Ps. cubensis
[24]. Initial analysis of these sequences using the Basic Alignment Analysis Search Tool
(BLAST) against the proteome of P. infestans revealed that only 7 of these sequences
aligned with annotated proteins within the P. infestans genome database; moreover,
only 1 of these was predicted to be a secreted RXLR effector. Of these sequences, one
(contig01871_F1) had 75% amino acid identity to P. infestans protein PITG_17484, a
putative member of the drug/metabolite transporter (DMT) superfamily (CLO184; 2.2).
Additional cloning via 3' RACE PCR and subsequent analysis revealed that the Ps.
cubensis coding sequence, hereafter referred to as Ps. cubensis RXLR protein 1
(PscRXLR1), appeared significantly shorter (i.e., 132 amino acids), compared to its P.
infestans ortholog PITG_1784 (i.e., 332 amino acids). This apparent truncation in
PscRXLR1 results in a protein coding sequence lacking the EamA functional domain
(PF00892; formerly called DUF6) found in members of the DMT family [31].

SignalP analysis of PscRxLR1 identified a putative 26 amino acid signal peptide at the
N-terminus of the protein (Figure 2.2). Based on the conserved features and domain
organization of oomycete effectors, the presence of a canonical RXLR motif was
identified at amino acid position 70 (Figure 2.2). However, unlike several previously
characterized oomycete effector proteins, PscRXLR1 does not contain an EER motif,
which has also been implicated in oomycete effector translocation into the host cell

45

[12,24,27]. The P. infestans ortholog, PITG_17848, while not having a predicted signal
peptide, does contain an RXLR-like motif (i.e., RFMR; Figure 2.2A). To eliminate the
possibility that PITG_17484 was mis-annotated and did in fact contain a signal peptide
upstream of the predicted coding sequence, the region 500 bp upstream of the
annotated PITG_17484 sequence was examined and a canonical signal peptide
sequence was not identified. We therefore concluded that PITG_17484 is not an RXLR
effector protein.

The absence of a predicted signal peptide in PITG_17848 suggests that PscRXLR1
may have evolved this function independently. To address this possibility, and to further
explore the ancestral function of these proteins, orthologous PscRXLR1 sequences in
additional plant oomycete pathogen species were identified, including P. sojae, P.
ramorum, and Py. ultimum. Not surprisingly, the sequences from P. sojae and P.
ramorum were more similar to PscRXLR1 than those from Py. ultimum (76% and 72%,
respectively, compared to 59%; Figure S2.2). Additionally, while none of these orthologs
had predicted signal peptides, they did contain EamA functional domains, indicating that
they were also members of the DMT superfamily (Figure S2.2). Phylogenetic analysis to
infer evolutionary relationships between PscRXLR1 and orthologs from P. sojae, P.
ramorum, and Py. ultimum supported these observations (Figure 2.2C).

46

Figure 2.2

47

Figure 2.2 (cont'd)

Figure 2.2
PscRXLR1 encodes a RXLR-containing effector protein with
homology to a non-effector protein in Phytophthora infestans. (A) Schematic
representations of PscRXLR1, Psc_781.4 and PITG_17484 from P. infestans. (B)
Multiple sequence alignment of PscRXLR1, Psc_781.4 and PITG_174984. Protein
sequences were aligned using CLUSTALW and shaded for consensus using
BOXSHADE. The asterisk indicates a stop codon. (C) The full length protein
sequences of PscRXLR1, Psc781.4 and their orthologs from P. infestans
(PITG_17484), Phytophthora ramorum (P. ramorum 96701 0 3435), Phytophthora
sojae (P. sojae 156165) and Pythium ultimum (PYU_T005955) were aligned using
Muscle and evolutionary history was inferred by using the Maximum Likelihood
method based on the JTT matrix-based model [1] using MEGA5[2]. 500 bootstrap
runs were performed.

Functional validation of the PscRXLR1 signal peptide

A primary characteristic of oomycete effector proteins is signal peptide-mediated
secretion from the haustoria into the extrahaustorial matrix (EHM) prior to translocation
into the host cell [3]. PscRXLR1 contains a 26 amino acid signal peptide as predicted by
SignalP 3.0 (HMM Probability, 0.966), whereas its closest P. infestans ortholog,
PITG_17484, does not have a predicted signal peptide. To determine if the predicted
48

signal peptide from PscRXLR1 is functional, we used a yeast signal trap assay based
on the requirement of invertase secretion for yeast growth on media with raffinose as
the sole carbon source [32]. This assay has been used previously to confirm predicted
signal peptide sequences in candidate effector proteins from both P. infestans and Ps.
cubensis [24,33]. As shown in Figure 2.3A, pSUC2-PscRXLR1 (column 4) is able to
grow on medium containing only raffinose (YPRAA), indicating that the signal peptide of
PscRXLR1 is sufficient for secretion of invertase. As a second confirmation of signal
peptide function, we monitored the reduction of 2,3,5-triphenyltetrazolium chloride (TTC)
to the red-colored compound triphenylformazan. Again, pSUC2-PscRXLR1 (column 4)
was confirmed as having a functional signal peptide. In contrast, neither the control
yeast strains (i.e., YPK12 [column 1] or pSUC2 [column 2]), nor the pSUC2PITG_17484 (column 3) yeast strain containing a PITG_17484-invertase fusion
construct, grew on YPRAA, nor were they able to reduce TTC. Our positive control,
PcQNE-SP1 (5) was, as shown previously [24], both able to grow on YPRAA medium
and reduce TTC. These data support the annotation of PscRXLR1 as a secreted RXLR
effector protein and confirm that its P. infestans ortholog is a non-secreted protein.

PscRXLR1 and PITG_1784 localize to the plant plasma membrane

In planta localization of effector proteins has been successfully used to guide functional
analysis and to infer the function itself [34,35,36,37]. To identify a possible role for
PscRXLR1 in the pathogenicity of Ps. cubensis, and to provide clues as to its putative
role in planta, we investigated the localization of PscRXLR1. To this end, a C-terminal

49

Figure 2.3

50

Figure 2.3 (cont'd) Functional characterization of PscRXLR1 and PITG_17484.
(A) PscRXLR1 has a functional signal peptide validated by the yeast signal trap assay.
Yeast strains were grown on media with raffinose (YPRAA) as the sole carbon source
or in the presence of 2,3,5-triphenyltetrazolium chloride (TTC). Yeast strains YTK12
and YTK12:pSUC2 EV both lack a functional invertase gene and cannot grow on
YPRAA medium or reduce TTC to red formazan. The functional signal peptide of
PscRXLR1, when fused in-frame to the mature yeast invertase (pSUC2-PscRXLR1),
allows for secretion of invertase, resulting in growth on YPRAA medium, as well as
reduction of TTC to red formazan. PITG_17484, as predicted, does not have a
functional signal peptide (pSUC2-PITG_17484). (B) Both PscRXLR1_CFP (top row)
and PITG_17484_CFP (bottom row) co-localize with a plasma membrane-specific
AtPIP2A_YFP marker. Left panels: C-CFP fusion protein only. Center panels:
AtPIP2A_YFP. Right panels: Merge of CFP and YFP images. Scale bar = 10 µm. (C,
D) Heterologous expression of PscRXLR1 induces chlorosis and necrosis in Nicotiana
benthamiana. Agrobacterium tumefaciens C58-C1 expressing PscRXLR1_CFP,
PITG_17484_CFP, or an AtPIP2A-YFP construct were infiltrated into N. benthamiana.
The chlorosis and necrosis phenotype of PscRXLR1_CFP infiltrated leaves (C) was
photographed at 2, 3, and 7 days post-inoculation (dpi). Leaf areas infiltrated with
either PITG_17484_CFP or AtPIP2A-YFP (D) as designated by the dash line circles
did not shown any phenotype at 3 dpi.

CFP-fusion protein (i.e., PscRXLR1-CFP) was transiently expressed in Nicotiana
benthamiana, and protein localization visualized using laser scanning confocal
microscopy. Based on the similarity of PscRXLR1 with members of the DMT
superfamily in Phytophthora spp. and Py. ultimum, PscRXLR1 was predicted to localize
to the plant plasma membrane, despite the absence of a predicted transmembrane
domain when analyzed using InterProScan. To confirm this, PscRXLR1-CFP was coexpressed with a YFP-tagged construct encoding the aquaporin gene AtPIP2A, a
marker for plasma membrane localization [38]. As predicted, PscRXLR1-CFP colocalized with AtPIP2A-YFP (Figure 2.3B), confirming that PscRXLR1 localizes to the
plasma membrane in planta. Additionally, a C-terminal CFP fusion was made to the P.

51

infestans ortholog PITG_17484, which was also confirmed to be plasma membrane
localized (Figure 2.3B).

PscRXLR1, but not its P. infestans ortholog, elicits a rapid cell death response
when expressed in Nicotiana benthamiana

The obligate nature of a plant pathogen often presents challenges towards functional
characterization of virulence and susceptibility within their respective host(s). To
circumvent this limitation, transient heterologous systems have been developed and
have proved valuable in their functional characterization [34,39,40,41,42]. To investigate
the activity of PscRXLR1 in planta, we utilized heterologous expression in N.
benthamiana as means to characterize and determine the function of PscRXLR1. As
shown in Figure 2.3C, expression of PscRxLR1 resulted in leaf chlorosis throughout the
entire infiltration zone by 2 dpi, followed by browning and initiation of necrosis at 4 dpi.
By 7 dpi, the zone of infiltration was completely dehydrated. In comparison, neither
infiltration with PITG_17484, nor pMDGFP, resulted in any detectable cell death-type
phenotype in N. benthamiana leaves at 4 dpi (Figure 2.3D).

PscRXLR1 expression is induced during Ps. cubensis infection of cucumber

The function of pathogen effector molecules is to enhance the virulence of the pathogen
during its lifecycle, as well as to dampen host defense responses activated during
infection. In this regard, the temporal expression of effector molecules during infection

52

and pathogen development often signals critical stages in the host-pathogen interaction.
Expression of PscRXLR1 mRNA was measured using quantitative reverse transcription
(qRT)-PCR following infection of Ps. cubensis on the susceptible cucumber cultivar
‘Vlaspik’. As shown in Figure 2.4, expression of PscRXLR1 was significantly (p < 0.001)
induced during infection, beginning at 1 dpi and continuing through 4 dpi, as compared
to the basal expression level in sporangia. Induction of gene expression at 1 dpi
corresponds with zoospore encystment in the stomata, the first stage of pathogen entry
into the host (Figure 2.4B, left panel). Subsequent expression observed through 4 dpi
corresponds with hyphal penetration through the stomata, growth throughout the
mesophyll, and formation of haustoria (Figure 2.4B, center and right panels). This
pattern of expression supports a potential role for PscRXLR1 in initial establishment of
the infection possibly through dampening host defense responses. Additionally, this
pattern is consistent with the expression patterns observed in other oomycete plant
pathogen effectors, further supporting the prediction of PscRXLR1 as an effector protein
with a role in infection and disease development.

PscRXLR1 is a splice variant of Psc_781.4

Automated annotation of the Illumina-generated Ps. cubensis assembly described in
this study resulted in Psc_781.4, a gene model at the PscRXLR1 locus that more
closely mirrored PITG_17484 than our prediction for PscRXLR1 and what was obtained
through molecular cloning (Figure 2.5), with the primary difference between the two
predictions being that intron 1 is either spliced in Psc_781.4, or retained in PscRXLR1

53

Figure 2.4

Figure 2.4 PscRXLR1 mRNA expression is up-regulated during
Pseudoperonospora cubensis infection of cucumber. (A) Expression levels of
PscRXLR1 in sporangia (SP) and at 1, 2, and 4 days post-inoculation (dpi).
Expression is displayed as fold sporangia expression and all time points are
significantly different compared to SP control (*** indicates p < 0.001) using Two-way
ANOVA. Error bars represent the standard error of the mean of 2 technical replicates
from each of 2 biological replicates. (B) Differential interference (DIC) microscopy
images of stages of Ps. cubensis infection on cucumber where e = encysted
zoospore; a = appressorium; and h = haustorium. Scale bars = 25 µm.

54

(Figure 2.5A). Empirical whole transcriptome sequence data (RNA-seq) from Ps.
cubensis sporangia (unpublished results) provides support for both isoforms at this
locus. When the first intron is retained, a stop codon is also brought into frame, yielding
a truncated transcript (i.e., putative effector PscRXLR1), and subsequently, a smaller
protein, which, as described in Figure 2.2, lacks the EamA functional domain (Figure
2.5A).

Based on our in silico predictions we confirmed which gene model, or both, was
represented in vivo. Using an RT-PCR-based approach, we were able to amplify both
splice variants from purified sporangia (SP), as well as from infected leaf material
harvested at 4 and 8 days post-inoculation (dpi) (Figure 2.5B), suggesting that both
isoforms are present throughout the infection process. As an added control, the
resultant PCR products were cloned and sequenced to confirm that they corresponded
to the appropriate splice variant (Figure S2.4). Additional functional analysis of
Psc_781.4 confirmed transient expression in N. benthamiana does not elicit a cell death
response in planta, indicating that it likely has no virulence function in Ps. cubensis
(Figure S2.5). In total, these independent methods confirm our conclusion that
Psc_781.4 is alternatively spliced leading to generation of a functional RXLR effector
protein.

55

Figure 2.5

Figure 2.5
PscRXLR1 is a splice variant of Psc_781.4. (A) Schematic
representation of intron and exon structures of PITG_17484, Psc_781.4, and
PscRXLR1. (B) RT-PCR analysis of alternative splicing in Psc_781.4 in sporangia (SP)
and at 4 and 8 days post-inoculation (dpi). RT-PCR products were subcloned into the
pGEM vector and DNA was amplified by PCR using intron spanning primers as shown
in (A). S, transcript with spliced intron 1. NS, non-spliced transcript.

56

DISCUSSION

In this study, we describe a candidate RXLR-type effector from Ps. cubensis that results
from a splice variant of a putative multi-drug transporter protein, and additionally
expands the scope of our understanding of the function and evolutionary history of the
Ps. cubensis effector repertoire. While Ps. cubensis is an oomycete pathogen of
worldwide economic importance, insight into the mechanism(s) underlying its virulence
and pathogenicity remain limited [43]. A recent study has provided a foundation for
investigating the genetic basis for virulence and pathogenicity in Ps. cubensis through
generation of a large scale genomic dataset [24]. We build upon this previous work
using a combination of in silico analyses, gene expression studies, and cell biology to
functionally characterize PscRXLR1 and establish a potential role in promoting Ps.
cubensis infection and proliferation.

Alternative splicing has been previously described in oomycetes pathogens of plants;
specifically related to the family 5 endoglucanases (EGL5) from P. sojae [21], as well as
in gene families such as Crinklers (CRNs), protein kinases, and transcription factors
[22]. In P. sojae, EGL5 proteins have a role in infection of soybean and are highly upregulated during the early stages of infection. As part of these analyses, four different
mechanisms of alternative splicing were described: intron skip, exon skip, alternative
donor site, and alternative acceptor site, with intron skip, where the intron is retained,
being the most commonly observed mechanism [22]. In agreement with this previous
observation, we propose that the Psc_RXLR1 transcript is generated via a retained

57

intron from Psc_781.4, yielding an RXLR-type effector. From an evolutionary
standpoint, alternative-splicing functions to expand the capacity of an organism's
proteome, thus enabling the generation of multiple functional isoforms from a single
coding sequence. Over time, new isoforms may be maintained if they have a beneficial
function [44], or lost, if their function is not critical to the lifecycle of the organism. In the
case of plant pathogens, this process could potentially serve an adaptive role to allow
for generation of isoforms of "housekeeping" type genes that gain new function(s),
potentially to promote virulence or infection. Alternatively, the PscRXLR1 splice variant
could represent a step in evolutionary time, as it is generated from the same coding
sequence as Psc_781.4 and maintained in the coding repertoire, but has not been
duplicated or retained as a separate sequence.

Like other oomycete effectors characterized to date, PscRXLR1 has a functional 26
amino acid signal peptide necessary for secretion from the haustorium into the
extrahaustorial matrix prior to translocation into the host cytoplasm. Interestingly,
PscRXLR1 is also localized to the host plasma membrane, despite the lack of a
predicted transmembrane domain or localization signal (Figure 2.3). To further examine
this, we surveyed the genomes of additional oomycete plant pathogens for orthologous
sequences. We hypothesize that if functional characterization data for orthologs in any
of these better characterized species (e.g., P. infestans, P. ramorum and P. sojae)
existed, it might provide insight into both the function and conservation of PscRXLR1.
Through BLAST analysis of the P. infestans, P. sojae, P. ramorum and Py. ultimum
genomes, we identified orthologous sequences in each of the pathogens, all of which

58

were annotated as members of the Drug/Metabolite Transporter (DMT) superfamily [31].
The DMT superfamily encompasses 19 families; the orthologs described here are
members of the EamA family, named for the O-acetylserine/cysteine export gene in
Escherichia coli [45]. While PscRXLR1 is lacking the EamA transmembrane domains
that are characteristic of these transporter proteins, our data clearly demonstrated
plasma membrane localization.

Monitoring the expression of both pathogen and host genes during infection can provide
insight into the interplay between resistance and susceptibility [46]. Using qRT-PCR, we
demonstrated that expression of PscRXLR1 mRNA is up-regulated during the early
stages of infection of cucumber. While we were unable to distinguish between isoforms,
expression

was induced nearly 10-fold at 1, 2 and 4 days post-inoculation (dpi),

corresponding with zoospore encystment, appressorium formation and penetration, and
proliferation and haustoria establishment, respectively (Figure 2.5). Several effector
proteins from P. infestans have also been demonstrated to have distinct temporal
patterns of expression, and are often expressed during the pre- and early stages of
infection, representative of the biotrophic phase of the P. infestans life cycle [12]. Based
on the robust induction of PscRXLR1 mRNA during the early stages of infection, as well
as the aggressive nature of the necrosis-inducing activity observed in N. benthamiana,
we hypothesize that the expression pattern of PscRXLR1 could support a role in the
initial infection process, possibly through dampening of host defense responses.
Indeed, effectors from other oomycete plant pathogens, including PcQNE from Ps.
cubensis and members of the CRN family from P. infestans, have also been shown to

59

elicit similar phenotypes when transiently expressed in N. benthamiana [3,24,47],
supporting the classification of PscRXLR1 as an effector protein with a putative role in
virulence.

In the current study, ongoing analysis of the Ps. cubensis genome has expanded the
candidate effector complement of Ps. cubensis to 271 sequences, revealing significant
variation in the conserved translocation motif. While previous analyses revealed a near
equal distribution of RXLR:QXLR motifs in Ps. cubensis, our current work, based on a
higher coverage draft genome sequence and predicted protein sequences rather than
open reading frames (ORFs), predicts sequences with 20 different amino acid
possibilities at the R1 position of the XXLR motif. Of these 20 predicted R1
substitutions, 19 have expression support (Table S2.1). While all 20 R1 substitutions
have yet to be functionally validated, it is not surprising that Ps. cubensis effectors may
in fact utilize a more diverse set of translocation motifs compared to the Phytophthora
spp., given its obligate lifestyle and relatively narrow host range. Among Phytophthora
spp., the conservation of the RXLR motif is well-established, yet there are additional
classes of oomycete effectors, such as the CRN family, that appear to utilize disparate
translocation motifs [34,47]. Moreover, analysis of the Py. ultimum genome has
identified an additional predicted translocation motif, YxSL[RK] [9]. Indeed, divergence
of transport signal sequences is even more pronounced between oomycetes and the
true fungi, which have no obvious conserved motifs that could function in transport and
show high degrees of variation even within the same species [48]. For example, the
effectors AvrM and AvrL567 from Melampsora lini, an obligate rust fungi with a similar

60

lifestyle to Ps. cubensis, rely on unique N-terminal sequences for uptake [48]. These
sequences, while different in regards to sequence similarity from the RXLR motif
observed in Phytophthora spp., are similar in that they feature positively charged
residues, implying that secondary protein structure may be a factor contributing to
uptake of these proteins. Both M. lini and Ps. cubensis are obligate biotrophs with
specific host ranges, which may have influenced the evolution of their effector
repertoires to select for unique translocation motifs compared to those found in
Phytophthora spp.

Preliminary analysis of the Ps. cubensis effector repertoire reveals minimal orthology
with annotated effector proteins from P. infestans, similar to what has been observed
when comparing the effector complement from P. infestans with P. ramorum, P. sojae
or H. arabidopsidis [8,27,29]. Through extensive analysis using both evolutionary and
comparative genomics, Phytophthora RXLR effector genes have been shown to be
undergoing accelerated rates of birth and death evolution as well as both widespread
gene duplication and loss events [7,8]. As such, among the predicted RXLR effector
genes from P. infestans, P. sojae, and P. ramorum, there are few genes with high
degrees of sequence similarity or 1:1:1 orthology [7]. Similarly, the same phenomenon
was observed in comparing candidate effector sequences from Ps. cubensis to those of
P. infestans. Of 271 predicted Ps. cubensis sequences, less than half (41%) of these
had significant similarity (e-value < 1e-5) to predicted P. infestans proteins, and only 3 of
these sequences had similarity to annotated P. infestans RXLR effector proteins. These
results indicate that the effector repertoire Ps. cubensis utilizes to promote its virulence

61

and pathogenicity on its hosts is, as could be predicted, very different than that utilized
by P. infestans, and likely the other Phytophthora spp. as well. This is likely due to
differing selective pressures on Ps. cubensis resulting from host specificity as well as
differences in lifestyle between the two pathogens (i.e., obligate vs. non-obligate).
In this study, we identified minimal conservation between the predicted Ps. cubensis
effector complement and effectors from P. infestans. We hypothesize that the
identification and analysis of effector to non-effector relationships among oomycete
plant pathogens is a valid measure to assess conservation and rates of evolution.
Additionally, with the identification of PscRXLR1, a splice variant of a non-effector gene,
we posit that these types of analyses as well as a more thorough analysis of alternative
splicing may provide a preliminary baseline to not only investigate evolutionary
differences among oomycete plant pathogens, but to also infer the relationship between
effector repertoire and the host-pathogen specificity and lifestyle. We have used several
criteria (i.e., prediction of selection pressure, secretion, etc.) to identify and analyze the
relationship between predicted Ps. cubensis effectors and their orthologs in P.
infestans. As observed for other effector proteins, some Ps. cubensis effectors may
have experienced stronger positive selection than most other proteins within the
genome. Interestingly, in addition to varying significantly from the genome average, the
distribution of ω for the PscE-Pi pairs has two distinct peaks, representing groups of
effectors under different levels of selection pressure. Thus, it appears that aside from
acting as effectors during infection, some of these slower evolving genes may have
additional, "housekeeping" roles that are yet to be uncovered. Despite computational

62

evidence indicating that these slower evolving genes are likely effectors, their role(s) in
pathogenesis remain to be established.

MATERIALS AND METHODS

Ps. cubensis culture and growth conditions

Ps. cubensis was maintained on Cucumis sativus cv. ‘Vlaspik’ as previously described
[24]. Cucumber plants were grown at 22 °C with a 12 h light/dark photoperiod. For Ps.
cubensis inoculation, sporangia were collected from heavily sporulating leaves by
washing with cold sterile distilled water and collecting sporangia in a centrifuge tube.
Sporangia were enumerated with a hemocytometer and suspended to a concentration
of 1 x 105 sporangia/ml in sterile distilled water. The underside of fully expanded 2 nd or
3rd true leaves of 4-week-old cucumber plants were spray-inoculated, until run-off, with
the suspension, and incubated for 24 h at 100% humidity in the dark. After 24 h,
inoculated plants were moved to a growth chamber (22 °C with a 12 h light/dark
photoperiod).

DNA and RNA extraction

Genomic DNA of Ps. cubensis was isolated from sporangia of isolate MSU-1 using the
DNeasy Plant Mini kit (Qiagen, Germantown, MD) with slight modifications. Sporangia
were washed from heavily sporulating leaves with sterile distilled water and filtered

63

through a 40 m nylon cell strainer to remove residual plant debris. The resultant
sporangia suspension was centrifuged, and the supernatant decanted. Sporangia were
suspended in buffer AP1 containing RNase and 5 l of Proteinase K and incubated at
37 °C for 20 min. 50 l of 425-600 µm acid washed beads were added to the sporangia
suspension and sporangia disrupted for 3 min using a vortex. Subsequent DNA
extraction steps were performed according to manufacturer’s instructions.

Ps. cubensis total RNA was isolated as follows: sporangia were collected as described
above for DNA isolation, yet re-suspended in buffer RLT from the RNeasy Plant Mini Kit
(Qiagen, Germantown, MD) and disrupted as above. RNA isolation was performed
according to the manufacturer’s instructions. RNA samples were treated with DNase
(Promega, Madison, WI) prior to use.

Sequence, assembly, and annotation of the Ps. cubensis genome

Genomic DNA was isolated from Ps. cubensis MSU-1 as described above and libraries
constructed using the Illumina Genomic DNA Sample kit (Illumina, San Diego, CA). Two
separate paired-end libraries were end sequenced using an Illumina Genome Analyzer
II (Illumina, San Diego, CA) at the UC-Davis Genome Center. The first library was
sequenced with 84bp reads and an insert size of 180bp yielding 7.8 Gbp of sequence.
The second library was sequenced with 100bp reads and an insert size of 480bp
yielding 5.5 Gbp of sequence. Illumina reads were trimmed to 51 bp to remove low
quality regions at the 3’ end of the reads. Reads with more than one N base or a base
64

with a quality score less than 20 were removed. The reads were then searched against
the Cucumber genome assembly [49] with Bowtie v0.12.7 [50] and matching reads were
removed; 4.5 Gbp of sequence was retained following trimming and cleaning the reads.
The trimmed and cleaned reads were assembled using Velvet v1.0.14 [25]. Three
Velvet runs were performed with hash lengths of 31, 41, and 51 and coverage cutoffs of
7, 3.6, and 2, respectively. A minimum contig size cutoff of 200bp was used for all the
assembly runs. The contigs from each Velvet run were then merged into one assembly
using

the

Minimus2

(http://sourceforge.net/apps/mediawiki/amos/index.php?title=Minimus2).

pipeline
Contaminant-

containing and mitochondrial contigs were removed; the final assembly contains 35,546
contigs with an N50 contig size of 4.0 Kbp; the total assembly is 64.4 Mbp. Reads were
deposited in the Sequence Read Archive at the National Center for Biotechnology
Information under study number SRP011018. The assembly is available at NCBI under
accession

AHJF00000000

and

via

http://www.daylab.plp.msu.edu/wp-

content/uploads/psc_merged_contigs.fasta.zip. The annotation can be downloaded at
http://www.daylab.plp.msu.edu/wp-content/uploads/psc_merged_contigs.gff3.zip.

The assembly was annotated using the MAKER [26] annotation pipeline. The
FGENESH gene finder [51] was used with the Phytophthora matrix to produce the initial
gene calls for the pipeline. All transcript and protein sequences from sequenced
oomycete genomes were provided to MAKER to improve the quality of the annotation.
In total, 23,519 loci and 23,522 gene models were predicted. Putative functional

65

annotation was assigned by searching the gene models against UniRef100 using
BLASTX (cutoff: 1E-5) and transferring the first hit with informative annotation.

Identification and cloning of PscRXLR1, Psc_781.4 and PITG_17484

Amplification of the coding sequence of PscRXLR1 was performed using DNA primers
that correspond to the open reading frame of PscRxLR1 (Figure 2.5). Subsequent
isolation and cloning of PscRxLR1 was performed by PCR using gene-specific primers
and genomic DNA from Ps. cubensis sporangia. Resultant amplicons were cloned into
the TA cloning vector pGEM-T-Easy (Promega), generating pGEM_PscRXLR1. To
ensure identification of a complete coding sequence, as well as to verify the absence of
introns in the sequence, 3 RACE (Rapid Amplification of cDNA Ends) was performed
using the SMARTer RACE cDNA Amplification Kit (Clontech, Mountain View, CA).
Amplification of Psc_781.4 was performed using 3' RACE as described above and the
final coding sequence was amplified using gene specific primers (Figure 2.5). Fidelity of
all sequences was confirmed by DNA sequencing using the ABI 3730 Genetic Analyzer
(Applied Biosystems, Foster City, CA).

P. infestans clone PITG_17484 was amplified from cDNA of P. infestans by PCR using
gene-specific primers. Amplicons were subcloned into pGEM-T-Easy and sequences
confirmed by sequencing.

66

DNA cloning and construct preparation

To validate the predicted signal peptide (SP) of PscRXLR1, the yeast signal trap assay
was used [24,32,33]. A DNA fragment corresponding to the predicted 26 amino acid
signal peptide (including start codon) was amplified by PCR using gene specific primers
(Table S3) modified to include 5 EcoR1 and 3 XhoI restriction sites. Resultant
amplicons were cloned into TA cloning vector pGEM-T-Easy (Promega), to yield pGEMPscRXLR1-SP. The plasmid pGEM-PscRXLR1-SP was digested with EcoR1 and XhoI,
the 84 bp SP fragment was gel purified, and ligated into the EcoR1/XhoI sites of the
yeast signal trap vector pSUC2T7M13ROI [32], generating pSUC2-PscRXLR1.
The SP sequence of PITG_17484 was similarly amplified and cloned, generating
pSUC2-PITG_17484. Plasmids pSUC2T7M13ROI, pSUC2-PscRXLR1 and pSUC2PITG_17484 were transformed into the yeast SUC2 minus strain YTK12 using the
Frozen-EZ Yeast Transformation IITM kit (Zymo Research, Orange, CA) following the
manufacturer’s instructions. Transformants were selected on CMD-W plates (0.67%
yeast nitrogen base without amino acids, 0.075% tryptophan dropout supplement, 2%
sucrose, 0.1% glucose, and 2% agar) for 3 days at 28 C. To confirm signal peptide
function, YTK12 containing the pSUC2 constructs were grown on raffinose-containing
YPRAA plates (1% yeast extract, 2% peptone, 2% raffinose, 2 g/L Antimycin A, and
2% agar). Yeast strains were replicated onto YPDA plates (1% yeast, 2% peptone, 2%
glucose, 0.003% adenine hemisulfate, and 2% agar) and CMD-W plates as equal
viability controls. TYK12 without pSUC2 was used as a negative control. The detection

67

of the secreted invertase activity with 2,3,5-triphenyltetrazolium chloride (TTC) was
performed as described by Tian et al. [24].

For construction of plasmids used for localization and phenotype studies, the open
reading frames of PscRXLR1, Psc_781.4 and PITG_17484 (minus stop codons) were
amplified and cloned into the Gateway entry vector pENTR/D-TOPO (Invitrogen,
Carlsbad, CA), yielding pENTR-PscRXLR1-GW, pENTR-Psc_781.4-GW, and pENTRPITG_17484-GW, respectively. The destination vector pVKH18En6gw-cCFP [24] was
used to create the C-terminal CFP fusions, using heterologous recombination via LR
Clonase, as per the manufacturer’s instructions (Invitrogen).

Transient expression and localization in N. benthamiana

Infiltration and transient expression in N. benthamiana using A. tumefaciens was
performed on 4-6 week old plants as described in Tian et al. [24]. A. tumefaciens strains
were grown overnight at 28 °C on Luria-Bertani (LB) plates containing 50 µg/mL
rifampicin and 25 g/mL kanamycin. A. tumefaciens clones were re-suspended in
induction buffer (10 mM MES, pH 5.6, 10 mM MgCl2, 150 m acetosyringone) and
incubated at room temperature, shaking in the dark, for 2 hours prior to infiltration. A.
tumefaciens suspensions were infiltrated at a final concentration of OD600 = 0.8.

A. tumefaciens-mediated transient expression in N. benthamiana for localization of
PscRXLR1-CFP, Psc_781.4-CFP, and PITG_17484-CFP with AtPIP2A-YFP [38] was

68

performed as described above. Visualization of fluorescently tagged proteins was
observed using an Olympus Fluoview 1000 laser scanning confocal microscope.
Images were adjusted for contrast in Canvas X (ACD Systems).

Quantitative real time PCR

First strand cDNA was synthesized from 1 g total RNA using the first-strand cDNA
synthesis kit (USB, Cleveland, OH). Quantitative RT-PCR was performed using a
Mastercycler ep Realplex real-time PCR (Eppendorf, Hamburg, Germany) using
HotStart SYBR Green qPCR Master Mix (2x) (USB), as previously described [52]. For
amplification of PscRXLR1 transcripts, gene specific primers were designed to amplify a
50 bp fragment (Forward: 5'-TGCGTAGCATCGCCAACCGA-3 and Reverse: 5'TCTTGCCAGCTGCATCGCGA-3'). Primers specific for the Ps. cubensis internal
transcribed spacer (ITS) region were used as an endogenous control (Table S3). Cycle
parameters were as follows: 95 °C for 2 min, followed by 40 cycles of: 95 °C (15 sec),
60 °C (15 sec) and 72 °C (45 sec). Fold expression was calculated based on expression
in sporangia. Data were analyzed by two-way ANOVA using Prism 4 (GraphPad
Software).

Splice variant analysis

Primers spanning the region of intron 1 were used (Figure S2.4) to amplify RT-PCR
products from SP, 4, and 8 dpi cDNA samples and resultant products were cloned into

69

the TA cloning vector pGEM-T-Easy. Fidelity of all sequences was confirmed by DNA
sequencing as described above.

Ortholog identification and sequence analysis

Candidate effector proteins were identified from the predicted proteome of Ps. cubensis
generated from the draft genome assembly using a modified RXLR effector
identification pipeline [29]. Orthologs of PscRXLR1 were identified in P. infestans
(http://www.broadinstitute.org/),

P.

sojae

(http://genome.jgi-psf.org),

P.

ramorum

(http://genome.jgi-psf.org) and Py. ultimum (http://pythium.plantbiology.msu.edu) with
BLAST

[53].

Signal

peptides

were

http://www.cbs.dtu.dk/services/SignalP/)

predicted
and

protein

using

SignalP

motifs

3.0

identified

([54],
using

InterProScan ([55], http://www.ebi.ac.uk/Tools/pfa/iprscan/). Amino acid alignments
were generated using ClustalW2 ([56], http://www.clustal.org), and resultant figures
generated using BoxShade v3.21 (http://www.ch.embnet.org/software/BOX_form.html).
P. infestans effector sequences from Haas et al. [7] were used for analyses and P.
infestans protein models were obtained from the Phytophthora infestans Sequencing
Project (http://www.broadinstitute.org/).

Defining paralogs and orthologs and evolutionary rate estimates

Synonymous and non-synonymous substitution rates (Ks and Ka, respectively) were
determined using the yn00 program in PAML [57]. Protein sequences were aligned first

70

and “back-translated” to coding sequence alignments. Very few pairs had run errors
(e.g., NAN in PAML output), and those with run errors were excluded. Sequence pairs
that were too similar (Ks ≤ 0.005) or too divergent (Ks > 3) were also excluded from
further analysis. For each Ps. cubensis or P. infestans effector protein, the closest
paralogous genes were identified using within-species BLAST searches and used for
rate calculation. Rates between putative orthologs were calculated as well. Putative
orthologs were identified globally between Ps. cubensis, P. infestans or P. ultimum by
first determining pairwise sequence similarities between all genes in these species. For
each Ps. cubensis protein, X, a protein in a second species, Y, is considered an
ortholog if the following two conditions are met: 1) X is the reciprocal best match of Y
and 2) X is located in a syntenic block where Y is found. Syntenic regions were
established using Multiple Collinearity Scan [58], with 1e-5 as an alignment significance
threshold, match size ≥5, and average intergenic distance.

Molecular phylogenetic analysis

The full-length protein sequences of PscRXLR1 and its orthologs were aligned using
default parameters with MUSCLE [59]. The multiple sequence alignment was used to
infer phylogenetic relationships between PscRXLR1 and its orthologs using the
Maximum Likelihood method, based on the JTT matrix-based model [60] with MEGA5
[61]. Bootstrap values (based on 500 replicates) for each node are given if >25%.

71

ACKNOWLEDGEMENTS

We thank members of the Day lab for critical reading of the manuscript. Dr. Joe Win
(The Sainsbury Laboratory) is gratefully acknowledged for his contributions in the
preliminary analysis of genomic sequences from Ps. cubensis.

72

APPENDIX

73

Figure S2.1

Figure S2.1 Signal peptide distribution among ortholog pairs. (A) Distribution of
secreted or non-secreted proteins in the Pseudoperonospora cubensis –
Phytophthora infestans ortholog baseline dataset. (B) Distribution of P. infestans
orthologs of Ps. cubensis effectors that are predicted to be secreted. Psc-sec = Ps.
cubensis secreted protein. Psc-non = Ps. cubensis non-secreted protein. Pi-sec = P.
infestans secreted protein. Pi-non = P. infestans non-secreted protein.

74

Figure S2.2
PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

1
1
1
1
1
1

MVWLQLKKSGLGFTMSLSAVYGAVYAAAN-SVPAGKIDSGKKAMRHLENLPLLLASDSLE
--------------MSLSAVYGAVYAAAN-SVPAGKIDSGKKAMRHLENLPLLLASDSLE
--------------MPLPAVYGAVYATATSSAPAGKIDSGKKAMRHLENLPLLVASDSLE
--------------MPLPAVYGAVYASATSSAPAG-----KKALRHLENLPLLVASDSLE
--------------MPLPAVYGAVYATATRSAPAGKIDSGKKAMHHLENLPLLIASDSLD
--------------MPAPAAVGIAVAAAA------NTDSGKKAMKQLEELPLLAVRHASD

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

60
46
47
42
47
41

SVSTEGKWLPRFLRQAIMRSIANRIAGIILVSTSAFLASCIATLVKEDAVKLAPVEILFW
SVSTEGKWLPRFLRQAIMRSIANRIAGIILVSTSAFLASCIATLVKEDAVKLAPVEILFW
SMSTDGKWLPRFMRRAVVRSVANRIAGLILVGTSAFLASCIATLVKDDAFKLSTVETLFW
SMTTEGKWLPRFMRRAVVRSVANRIAGLILVGTSAFLASCITTLVKDDTIKLSAIEALFW
SMSTEGKWLPRFMRRAVVRSVANRIAGLILVGTSAVLASCIATLTKDADFKLSPVETLFW
AAIRTGSRLPKLVQKWYSKLSATQLEGVVLVAASAFTFSLLSTLIKYASQSMPSMETVFW

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

120
106
107
102
107
101

RSLVSWLLTLVSS*---------------------------------------------RSLVSWLLTLVAITTTGVKTRLKKEYYRPIVLRSFTGCIATTLTIIMLQELAVSNAIAIT
RSLVSWLLTLAAIAATGVKMRVKKEFQRPLLLRCFTGCIALTLTVLVIQKLEVSNATAIT
RSLVSWFLTVAALATTSTKMRVKKEFNRPLTLRCVFGCISTTLTIGVLEKLAVSNATAVT
RSLVSWLLTLAAIAATGIRMRVKKEFHRPLVLRCVTGCVAMTLTLLVLQTLAVSNATAIT
RSFVAWLLNLVAVWQ-----------------------------------MVLADASVLI

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

166
167
162
167
126

-----------------------------------------------------------YFSPLLAFAMAAKFLKEKPKLFAVACSVMCVIGAVLVVRPVFLFGKSGSTDASWYRRSMT
YLSPLLAFAMAAFFLKEKPGAFTLACSALCVVGAVQVVRPAFVFGKNGSTDAKWYRRSMT
YVSPLLAFAMATFFLKEKPGVFTAVCSALCVAGAVLVVRPAFLFGKSGSTDAKWYHRSMA
YFSPLLAFAMAALFIKEKPDIFTVACSVVCVAGAILVVRPAFLFGKDGSTDAKWYRRSMA
FTSPVMTFLLGAMVLKEKIDPVNMGYALFSFVGVICVVRPSFIFGNDHTTAG-------75

Figure S2.2 (cont'd)
PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

226
227
222
227
178

-----------------------------------------------------------SFVTSYLFGESLAIGCAIIVVFMQAGAYVSLRSLQKVPHLVVMHYYLVTTTLVSLP*--SFVTSYLFGESLAIGCAVVTVVMQAGAYVSLRSLQKVPHLVVMNYFLLTMTLVSLIAILV
SFVTSNLFGESLAVGCAVVVAFMQAGAYVSLRSLHKVEYLVVMQYYLFTMTLAALAAMLG
SFVTSSLFGESLAIGCAAVVVFMQAGAYVSLRSLQKVPHLVVMHYFLLSMTLVSLVAVLV
------TDGSVFAIMCALLGAAAQAVAYVSMRRLQQVNYLVVINYFLLTSSVMSALSLLL

287
282
287
232

----------------------------------------------------------------------------------------------------------------------VQHGKFKAGLSVETWGAILGTGALAFAEQLFLTRGFQFDGAGVLAATRLLHVGYEFVWGV
IQHGKFKAGTSLETWGAIVGTGALAFVEQLFLTRGFQFDGAGVLAATRLLHVSCEFAWGV
VQHGKFKTDLSLGTWSAILGTGALAFAEQLFLTRGFQFDGAGVLAATRLLHVGCEFAWGV
VQR-KFVIKMSLDVWLAVLGTGFLGFIGQLFLTRGFQLESAGTASVMRYLDVVFVFVWDI

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

347
342
347
291

----------------------------------------------------------------------------------------------------------------------ALLGTALNPWSASGAAATAAGVLFLALRRTARSREALAPHKNAYLH--------FSHARR
ILLGTALNPWSAGGAGVTAAGVLFLALRRVHTHWAARRSLRRILQ*-------------ILLGTALNPWSASGAAATAAGVLFLALRRVHTHWAARRSLRRMAAKPHKNAYLHFSSARR
TLLHERINAWSAVGALIICGSAIAIAIRKMQS*---------------------------

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

----------------------------------------------------------------------------------------------------------------------399 DELAAENPSWSVQQVSAELGRQWKALSAAERKPWVELAQFDKARFHTEAHHHV-NQQQSD
-----------------------------------------------------------407 EQLAEANPAWSVQQVSAELGRQWKSLAAVERKPWVELAQFDKARYHTEAHQHMRQQTDEQ
------------------------------------------------------------

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum96701_0_3435
PYU1_T005955

76

Figure S2.2 (cont'd)
PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

----------------------------------------------------------------------------------------------------------------------458 EQPEQAPPKRKKQSNEPRQPDTAYICFWKSQRPEVVAANPFLAAPLVSKEVGRQWRALSD
-----------------------------------------------------------467 PERPHLPTKRKKRPNEPRQPDTAYICFWKSRRPEVIAENPLLAAPSVSREEIARFQP--------------------------------------------------------------

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

----------------------------------------------------------------------------------------------------------------------518 DERQPTLAATTPDMLSALKTPLKDPFAPKPAKTAFQLFMSHNRESFMLLNMTINEFRAEM
-----------------------------------------------------------524 ------ALVATSEIPPALKAPLKDPFAPKPAKTGFQLFMSHNRESFTLLNMTINEFRTEM
------------------------------------------------------------

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

----------------------------------------------------------------------------------------------------------------------578 SQLWKRLSDADKAEWHELAKEDQRRYDTEMNAYKPPAYMDLVVQRSHKRMEELRRLARED
-----------------------------------------------------------578 SQLWKRLSDADKNEWYELAKLDERRYETEMNAYKPPAYMESAVQRAHKRLDELRRLARRD
------------------------------------------------------------

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

----------------------------------------------------------------------------------------------------------------------638 SAAPRLPMNAYNCYLSAKRQELVDRRPGRKNPEIMREIGVTWKALSDDERAVYQRKADED
-----------------------------------------------------------638 AAAPRLPMNAYNCYLSKERQELAVQRPDLKNPEIMREIGVTWKALSEDERASFQRKAEDD
-----------------------------------------------------------77

Figure S2.2 (cont'd)
PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

----------------------------------------------------------------------------------------------------------------------698 VERFRAEMEAHIAKKNEEEAANPLTKRRPRKRKEFDDEEELVKTPTVPRKKRKSGPPRRP
-----------------------------------------------------------698 VERFRADMEAYLTQQEEQRAVQEEQDVEPEVVREVVHEVVKEP--VVGRKKRKSVSPRRP
------------------------------------------------------------

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

----------------------------------------------------------------------------------------------------------------------758 KTAYNLMYMSKRTELLSTYQMSHNECSALCGKLWRQMSEAEREPYKRMAAEDKHRYEAEL
-----------------------------------------------------------756 KTAYNLMYMSKRAELLSTYQMSHNECSALCGRLWRQMSEEEREPYKRMAAEDKRRYETEM
------------------------------------------------------------

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

----------------------------------------------------------------------------------------------------------------------818 QVYNAQQEEANNKTLRDSAGFRHFLEAKRRENEAISSDEAAAIWQEMTEPHQLLWTELAR
-----------------------------------------------------------816 EIYNAEVDAANKKTLRESAGFSYFLEAKRRENEQISEGEAAAIWRDMLAPHQMLWTELAS
------------------------------------------------------------

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

----------------------------------------------------------------------------------------------------------------------878 DNKHKTSVERTAVDVLDTLL*-------------------------------------------------------------------------------------------------876 DTPAPASSNATHQKGQTGMNGGRWTEQEHQSFLAGLRLYGREWKKVAAKIKTRTSAQIRS
-----------------------------------------------------------78

Figure S2.2 (cont'd)
PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------936 HAQKYFAKLARDDEMRKHSGLSMIMAGSIGYFSDGGSSVAQNSGDDDAEASDASRQMARA
------------------------------------------------------------

PscRXLR1
Psc_781.4
P. sojae 156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------996 RSAGQSKGTAAILIAPMGSAVSGLYKQTTGATKKRARAAVTGFDGQLEMGAATSSFPYKL
------------------------------------------------------------

PscRXLR1
Psc_781.4
Psojae_156165
PITG_17484
P. ramorum 96701_0_3435
PYU1_T005955

--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------QKRQNDATRVEYLPSQEELLAKASPNLRHRLSSLIEAELCALQVLSCYAMLQQQEQISAP
------------------------------------------------------------

PscRXLR1
--------------------------------Psc_781.4
--------------------------------Psojae_156165
--------------------------------PITG_17484
--------------------------------P. ramorum 96701_0_3435 1116 RQKTKRQGSAKASTLGLPMLSTEQMPPTSSIY*
PYU1_T005955
--------------------------------79

Figure S2.2 (cont'd) Relationship between PscRXLR1 and oomycete orthologs. Alignment of PscRXLR1, Psc_781.4,
PITG_17484 (P. infestans), PYU_T005955 (Py. ultimum), P. ramorum 96701_0_3435, and P. sojae 156165 amino acid
sequences were generated using ClustalW and represented with BoxShade. PscRXLR1 signal peptide is boxed in blue.
The RXLR or RXLR-like domains are boxed in red. The green boxes represent the EamA domains found in each protein
sequence. Stop codons are represented by asterisks.

80

Figure S2.3

Figure S2.3 Heterologous expression of PscRXLR1 specifically results in cell death in Nicotiana benthamiana.
Infiltration of PscRXLR1_CFP with or without the plasma membrane marker construct AtPIP2A-YFP results in chlorosis
and necrosis 4 days post-inoculation (dpi). Circles mark the infiltration zones, visible at 0 dpi. Infiltration with AtPIP2A-YFP
alone does not result in any observable phenotype in N. benthamiana leaves.

81

Figure S2.4
A
781_4 exon 1(last 100 bp)
781_4 exon1 F primer
4 dpi #6-8
8 dpi #4-12
SP #5-9
INTRON 1
781_4 intron 1 span R primer
781_4 exon 2 (first 100 bp)

TGCCAGCTGC
----------------------------------------------------------------

ATCGCGACGT
----------------------------------------------------------------

TGGTGAAAGA
----------------------------------------------------------------

GGACGCCGTT
----------------------------------------------------------------

781_4 exon 1(last 100 bp)
781_4 exon1 F primer
4 dpi #6-8
8 dpi #4-12
SP #5-9
INTRON 1
781_4 intron 1 span R primer
781_4 exon 2 (first 100 bp)

CCGTGGAGAT
----------------------------------------------------------------

TTTGTTTTGG
------TTGG
------TTGG
------TTGG
------TTGG
----------------------------

CGCTCACTCG
CGCTCACTCG
CGCTCACTCG
CGCTCACTCG
CGCTCACTCG
----------------------------

TGTCTTGGCT GCTAACGCTT
TGTCTTG-------------TGTCTTGGCT GCTAACGCTT
TGTCTTGGCT GCTAACGCTT
TGTCTTGGCT GCTAACGCTT
---------- ------------------- ------------------- ----------

781_4 exon 1(last 100 bp)
781_4 exon1 F primer
4 dpi #6-8
8 dpi #4-12
SP #5-9
INTRON 1
781_4 intron 1 span R primer
781_4 exon 2 (first 100 bp)

------------------GTAAGCTCTT
GTAAGCTCTT
GTAAGCTCTT
GTAAGCTCTT
-------------------

------------------GACCCACTGA
GACCCACTGA
GACCCACTGA
GACCCACTGA
-------------------

------------------TATTGTACGA
TATTGTACGA
TATTGTACGA
TATTGTACGA
-------------------

------------------TGATTGCCTA
TGATTGCCTA
TGATTGCCTA
TGATTGCCTA
-------------------

82

AAATTAGCAC
----------------------------------------------------------------

------------------ACAAATTCTT
ACAAATTCTT
ACAAATTCTT
ACAAATTCTT
-------------------

Figure S2.4 (cont'd)
781_4 exon 1(last 100 bp)
781_4 exon1 F primer
4 dpi #6-8
8 dpi #4-12
SP #5-9
INTRON 1
781_4 intron 1 span R primer
781_4 exon 2 (first 100 bp)

------------------GGTGATTGAT
GGTGATTGAT
GGTGATTGAT
GGTGATTGAT
-------------------

------------------TACAAGGTTG
TACAAGGTTG
TACAAGGTTG
TACAAG------------G
------GTTG

------------------CAATCACGAC
CAATCACGAC
CAATCACGAC
---------CAATCACGAC
CAATCACGAC

------------------TACTGGCGTT
TACTGGCGTT
TACTGGCGTT
---------TACTGGCGTT
TACTGGCGTT

------------------AAG------AAG------AAG---------------AAG------AAGACGCGCT

781_4 exon 1(last 100 bp)
781_4 exon1 F primer
4 dpi #6-8
8 dpi #4-12
SP #5-9
INTRON 1
781_4 intron 1 span R primer
781_4 exon 2 (first 100 bp)

---------------------------------------------------------------TGAAGAAAGA

---------------------------------------------------------------GTATTATCGC

---------------------------------------------------------------CCAATCGTGC

---------------------------------------------------------------TTCGATCATT

---------------------------------------------------------------CACGGGTTGC

781_4 exon 1(last 100 bp)
781_4 exon1 F primer
4 dpi #6-8
8 dpi #4-12
SP #5-9
INTRON 1
781_4 intron 1 span R primer
781_4 exon 2 (first 100 bp)

----------- --------------- --------------- --------------- --------------- --------------- --------------- -----ATCGCCACGA CACTTA

83

Figure S2.4 (cont'd)
B
781_4 exon1 (last 100 bp)
781_4 exon1 F primer
4 dpi #3-1
8 dpi #4-7
SP #2-3
781_4 intron 1 span R primer
781_4 exon 2 (first 100 bp)

TGCCAGCTGC
-------------------------------------------------------

ATCGCGACGT
-------------------------------------------------------

TGGTGAAAGA
-------------------------------------------------------

GGACGCCGTT
-------------------------------------------------------

AAATTAGCAC
-------------------------------------------------------

781_4 exon1 (last 100 bp)
781_4 exon1 F primer
4 dpi #3-1
8 dpi #4-7
SP #2-3
781_4 intron 1 span R primer
781_4 exon 2 (first 100 bp)

CCGTGGAGAT
-------------------------------------------------------

TTTGTTTTGG
------TTGG
------TTGG
------TTGG
------TTGG
-------------------

CGCTCACTCG
CGCTCACTCG
CGCTCACTCG
CGCTCACTCG
CGCTCACTCG
-------------------

TGTCTTGGCT
TGTCTTG--TGTCTTGGCT
TGTCTTGGCT
TGTCTTGGCT
-------------------

GCTAACGCTT
---------GCTAACGCTT
GCTAACGCTT
GCTAACGCTT
-------------------

781_4 exon1 (last 100 bp)
781_4 exon1 F primer
4 dpi #3-1
8 dpi #4-7
SP #2-3
781_4 intron 1 span R primer
781_4 exon 2 (first 100 bp)

------------------GTTGCAATCA
GTTGCAATCA
GTTGCAATCA
---GCAATCA
GTTGCAATCA

------------------CGACTACTGG
CGACTACTGG
CGACTACTGG
CGACTACTGG
CGACTACTGG

------------------CGTTAAG--CGTTAAG--CGTTAAG--CGTTAAG--CGTTAAGACG

------------------------------------------------------CGCTTGAAGA

------------------------------------------------------AAGAGTATTA

84

Figure S2.4 (cont'd)
781_4 exon1 (last 100 bp)
781_4 exon1 F primer
4 dpi #3-1
8 dpi #4-7
SP #2-3
781_4 intron 1 span R primer
781_4 exon 2 (first 100 bp)

------------------------------------------------------TCGCCCAATC

------------------------------------------------------GTGCTTCGAT

-------------------------------------------------------

-------------------------------------------------------

-------------------------------------------------------

Figure S2.4. Multiple sequence alignments of splice variant isoforms. (A) Alignment representing non-spliced,
PscRXLR1 isoform. (B) Alignment representing spliced isoform indicative of Psc_781.4.

85

Figure S2.5

Figure S2.5 Heterologous expression of Psc_781.4 in Nicotiana benthamiana.
Infiltration and expression of Psc_781.4 does not result in any observable phenotype in
N. benthamiana leaves at 4 days post-inoculation (dpi). Circles mark the infiltration
zones, visible at 9 dpi.

86

REFERENCES

87

REFERENCES
1. De Wit PJGM, Mehrabi R, Van Den B, Harrold A, Stergiopoulos I (2009) Fungal
effector proteins: past, present and future. Mol Plant Pathol 10: 735-747
2. Mansfield JW (2009) From bacterial avirulence genes to effector functions via the hrp
delivery system: an overview of 25 years of progress in our understanding of
plant innate immunity. Mol Plant Pathol 10: 721-734
3. Schornack S, Huitema E, Cano LM, Bozkurt TO, Oliva R, et al. (2009) Ten things to
know about oomycete effectors. Mol Plant Pathol 10: 795-803
4. Phytopathology TPCRi (2006) A catalogue of the effector secretome of plant
pathogenic oomycetes. Ann Rev Phytopathol 44: 41-60.
5. Chisholm ST, Coaker G, Day B, Staskawicz BJ (2006) Host-microbe interactions:
shaping the evolution of the plant immune response. Cell 124: 803-814.
6. Jones JD, Dangl JL (2006) The plant immune system. Nature 444: 323-329.
7. Haas BJ, Kamoun S, Zody MC, Jiang RH, Handsaker RE, et al. (2009) Genome
sequence and analysis of the Irish potato famine pathogen Phytophthora
infestans. Nature 461: 393-398.
8. Tyler BM (2006) Phytophthora genome sequences uncover evolutionary origins and
mechanisms of pathogenesis. Science 313: 1261-1266
9. Levesque CA, Brouwer H, Cano L, Hamilton JP, Holt C, et al. (2010) Genome
sequence of the necrotrophic plant pathogen Pythium ultimum reveals original
pathogenicity mechanisms and effector repertoire. Genome Biol 11: R73.
10. Baxter L, Tripathy S, Ishaque N, Boot N, Cabral A, et al. (2010) Signatures of
adaptation to obligate biotrophy in the Hyaloperonospora arabidopsidis genome.
Science 330: 1549-1551.
11. Rehmany AP, Gordon A, Rose LE, Allen RL, Armstrong MR, et al. (2005)
Differential recognition of highly divergent downy mildew avirulence gene alleles
by RPP1 resistance genes from two Arabidopsis lines. Plant Cell 17: 1839-1850.
88

12. Whisson SC, Boevink PC, Moleleki L, Avrova AO, Morales JG, et al. (2007) A
translocation signal for delivery of oomycete effector proteins into host plant cells.
Nature 450: 115-118
13. Bos JI, Kanneganti TD, Young C, Cakir C, Huitema E, et al. (2006) The C-terminal
half of Phytophthora infestans RXLR effector AVR3a is sufficient to trigger R3amediated hypersensitivity and suppress INF1-induced cell death in Nicotiana
benthamiana. Plant J 48: 165-176.
14. Allen RL (2004) Host-parasite coevolutionary conflict between Arabidopsis and
downy mildew. Science 306: 1957-1960
15. Armstrong MR, Whisson SC, Pritchard L, Bos JI, Venter E, et al. (2005) An
ancestral oomycete locus contains late blight avirulence gene Avr3a, encoding a
protein that is recognized in the host cytoplasm. Proc Natl Acad Sci U S A 102:
7766-7771.
16. Shan W, Cao M, Leung D, Tyler BM (2004) The Avr1b locus of Phytophthora sojae
encodes an elicitor and a regulator required for avirulence on soybean plants
carrying resistance gene Rps 1b. Mol Plant-Microbe Interact 17: 394-403.
17. Qutob D, Tedman-Jones J, Dong S, Kuflu K, Pham H, et al. (2009) Copy number
variation and transcriptional polymorphisms of Phytophthora sojae RXLR effector
genes Avr1a and Avr3a. PLoS ONE 4: e5066
18. Dong S, Qutob D, Tedman-Jones J, Kuflu K, Wang Y, et al. (2009) The
Phytophthora sojae avirulence locus Avr3c encodes a multi-copy RXLR effector
with sequence polymorphisms among pathogen strains. PLoS ONE 4: e5556
19. Matlin AJ, Clark F, Smith CWJ (2005) Understanding alternative splicing: towards a
cellular code. Nat Rev Mol Cell Biol 6: 386-398.
20. Win J, Kanneganti TD, Torto-Alalibo T, Kamoun S (2006) Computational and
comparative analyses of 150 full-length cDNA sequences from the oomycete
plant pathogen Phytophthora infestans. Fungal Genet Biol 43: 20-33.
21. Costanzo S, Ospina-Giraldo M, Deahl K, Baker C, Jones R (2007) Alternate intron
processing of family 5 endoglucanase transcripts from the genus Phytophthora.
Current Genetics 52: 115-123.
89

22. Shen D, Ye W, Dong S, Wang Y, Dou D (2011) Characterization of intronic
structures and alternative splicing in Phytophthora sojae by comparative analysis
of expressed sequence tags and genomic sequences. Can J Microbiol 57: 84-90.
23. Savory EA, Granke LL, Quesada-Ocampo LM, Varbanova M, Hausbeck MK, et al.
(2011) The cucurbit downy mildew pathogen Pseudoperonospora cubensis. Mol
Plant Pathol 12: 217-226.
24. Tian M, Win J, Savory E, Burkhardt A, Held M, et al. (2011) 454 Genome
sequencing of Pseudoperonospora cubensis reveals effector proteins with a
QXLR translocation motif. Mol Plant-Microbe Interact 24: 543-553.
25. Zerbino DR, Birney E (2008) Velvet: Algorithms for de novo short read assembly
using de Bruijn graphs. Genome Res 18: 821-829.
26. Cantarel BL, Korf I, Robb SMC, Parra G, Ross E, et al. (2008) MAKER: An easy-touse annotation pipeline designed for emerging model organism genomes.
Genome Res 18: 188-196.
27. Jiang RHY, Tripathy S, Govers F, Tyler BM (2008) RXLR effector reservoir in two
Phytophthora species is dominated by a single rapidly evolving superfamily with
more than 700 members. Proc Natl Acad Sci U S A 105: 4874-4879.
28. Savory EA, Adhikari BN, Hamilton JP, Vaillancourt B, Buell CR, et al. (2012) mRNASeq Analysis of the Pseudoperonospora cubensis Transcriptome During
Cucumber (Cucumis sativus L.) Infection. PLoS ONE 7: e35796.
29. Win J, Morgan W, Bos J, Krasileva KV, Cano LM, et al. (2007) Adaptive evolution
has targeted the C-terminal domain of the RXLR effectors of plant pathogenic
oomycetes. Plant Cell 19: 2349-2369.
30. Lynch M, Conery JS (2000) The evolutionary fate and consequences of duplicate
genes. Science 290: 1151-1155.
31. Jack DL, Yang NM, Saier MH, Jr. (2001) The drug/metabolite transporter
superfamily. Eur J Biochem 268: 3620-3639.

90

32. Jacobs KA, Collins-Racie LA, Colbert M, Duckett M, Golden-Fleet M, et al. (1997) A
genetic selection for isolating cDNAs encoding secreted proteins. Gene 198:
289-296.
33. Oh SK, Young C, Lee M, Oliva R, Bozkurt TO, et al. (2009) In planta expression
screens of Phytophthora infestans RXLR effectors reveal diverse phenotypes,
including activation of the Solanum bulbocastanum disease resistance protein
Rpi-blb2. Plant Cell 21: 2928.
34. Schornack S, van Damme M, Bozkurt TO, Cano LM, Smoker M, et al. (2010)
Ancient class of translocated oomycete effectors targets the host nucleus. Proc
Natl Acad Sci U S A 107: 17421-17426.
35. Gurlebeck D, Jahn S, Gurlebeck N, Szczesny R, Szurek B, et al. (2009)
Visualization of novel virulence activities of the Xanthomonas type III effectors
AvrBs1, AvrBs3 and AvrBs4. Mol Plant Pathol 10: 175-188.
36. Jin P, Wood MD, Wu Y, Xie Z, Katagiri F (2003) Cleavage of the Pseudomonas
syringae type III effector AvrRpt2 requires a host factor(s) common among
eukaryotes and is important for AvrRpt2 localization in the host cell. Plant Physiol
133: 1072-1082.
37. Shan L, Thara VK, Martin GB, Zhou JM, Tang X (2000) The Pseudomonas AvrPto
protein is differentially recognized by tomato and tobacco and is localized to the
plant plasma membrane. Plant Cell 12: 2323-2338.
38. Nelson BK, Cai X, Nebenführ A (2007) A multicolored set of in vivo organelle
markers for co-localization studies in Arabidopsis and other plants. Plant J 51:
1126-1136.
39. Chaparro-Garcia A, Wilkinson RC, Gimenez-Ibanez S, Findlay K, Coffey MD, et al.
(2011) The receptor-like kinase SERK3/BAK1 is required for basal resistance
against the late blight pathogen Phytophthora infestans in Nicotiana
benthamiana. PLoS ONE 6: e16608.
40. Win J, Kamoun S, Jones AM (2011) Purification of effector-target protein complexes
via transient expression in Nicotiana benthamiana. Methods Mol Biol 712: 181194.

91

41. Liu T, Ye W, Ru Y, Yang X, Gu B, et al. (2011) Two host cytoplasmic effectors are
required for pathogenesis of Phytophthora sojae by suppression of host
defenses. Plant Physiol 155: 490-501.
42. Halterman DA, Chen Y, Sopee J, Berduo-Sandoval J, Sanchez-Perez A (2010)
Competition between Phytophthora infestans effectors leads to increased
aggressiveness on plants containing broad-spectrum late blight resistance. PLoS
ONE 5: e10536.
43. Savory EA, Granke LL, Quesada-Ocampo LM, Varbanova M, Hausbeck MK, et al.
(2011) The cucurbit downy mildew pathogen Pseudoperonospora cubensis.
Molecular Plant Pathology 12: 217-226.
44. Keren H, Lev-Maor G, Ast G (2010) Alternative splicing and evolution:
diversification, exon definition and function. Nat Rev Genet 11: 345-355.
45. Franke I, Resch A, Dassler T, Maier T, Bock A (2003) YfiK from Escherichia coli
promotes export of O-acetylserine and cysteine. J Bacteriol 185: 1161-1166.
46. Chandran D, Tai YC, Hather G, Dewdney J, Denoux C, et al. (2009) Temporal
global expression data reveal known and novel salicylate-impacted processes
and regulators mediating powdery mildew growth and reproduction on
Arabidopsis. Plant Physiol 149: 1435-1451.
47. Torto TA (2003) EST mining and functional expression assays identify extracellular
effector proteins from the plant pathogen Phytophthora. Genome Res 13: 16751685.
48. Rafiqi M, Gan PH, Ravensdale M, Lawrence GJ, Ellis JG, et al. (2010)
Internalization of flax rust avirulence proteins into flax and tobacco cells can
occur in the absence of the pathogen. Plant Cell 22: 2017-2032.
49. Huang S, Li R, Zhang Z, Li L, Gu X, et al. (2009) The genome of the cucumber,
Cucumis sativus L. Nat Genet 41: 1275-1281.
50. Langmead B, Trapnell C, Pop M, Salzberg SL (2009) Ultrafast and memory-efficient
alignment of short DNA sequences to the human genome. Genome Biol 10: R25.

92

51. Salamov AA, Solovyev VV (2000) Ab initio Gene Finding in Drosophila Genomic
DNA. Genome Res 10: 516-522.
52. Knepper C, Savory EA, Day B (2011) Arabidopsis NDR1 is an integrin-like protein
with a role in fluid loss and plasma membrane-cell wall adhesion. Plant Physiol
156: 286-300.
53. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, et al. (1997) Gapped
BLAST and PSI-BLAST: a new generation of protein database search programs.
Nucleic Acids Res 25: 3389-3402.
54. Emanuelsson O, Brunak S, von Heijne G, Nielsen H (2007) Locating proteins in the
cell using TargetP, SignalP and related tools. Nature Protocols 2: 953-971.
55. Quevillon E, Silventoinen V, Pillai S, Harte N, Mulder N, et al. (2005) InterProScan:
protein domains identifier. Nucleic Acids Res 33: W116-120.
56. Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, et al. (2007)
Clustal W and Clustal X version 2.0. Bioinformatics 23: 2947-2948.
57. Yang Z (1997) PAML: a program package for phylogenetic analysis by maximum
likelihood. Comp App Biosci 13: 555-556.
58. Tang H, Wang X, Bowers JE, Ming R, Alam M, et al. (2008) Unraveling ancient
hexaploidy through multiply-aligned angiosperm gene maps. Genome Res 18:
1944-1954.
59. Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and
high throughput. Nucleic Acids Res 32: 1792-1797.
60. Jones DT, Taylor WR, Thornton JM (1992) The rapid generation of mutation data
matrices from protein sequences. Comp App Biosci 8: 275-282.
61. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, et al. (2011) MEGA5:
Molecular Evolutionary Genetics Analysis using maximum likelihood,
evolutionary distance, and maximum parsimony methods. Mol Biol Evol doi:
10.1093/molbev/msr121.

93

CHAPTER 3

mRNA-Seq Analysis of the Pseudoperonospora cubensis transcriptome
during cucumber (Cucumis sativus L.) infection

This chapter was originally published in PLoS ONE.
Savory EA*, Adhikari, BN*, Hamilton JP, Vaillancourt, B, Buell CR, and Day B (2012)
mRNA-Seq Analysis of the Pseudoperonospora cubensis transcriptome during
cucumber (Cucumis sativus L.) infection.
PLoS ONE 7(4): e35796.
doi:10.1371/journal.pone.0035796.
© 2012 Savory et al. This is an open-access article distributed under the terms of the
Creative Commons Attribution License, which permits unrestricted use, distribution, and
reproduction in any medium, provided the original author and source are credited.
*

These authors contributed equally to this work

Author Contributions:
Conceived and designed the experiments: EAS, BNA, CRB, and BD. Performed the
experiments: EAS, BNA, JPH, and BV. Analyzed the data: EAS, BNA, JPH, CRB, and
BD. Contributed reagents/materials/analysis tools: EAS, BNA, JPH, CRB, and BD.
Wrote the paper: EAS, BNA, CRB, and BD.

94

ABSTRACT
Pseudoperonospora cubensis, an oomycete, is the causal agent of cucurbit downy
mildew, and is responsible for significant losses on cucurbit crops worldwide. While
other oomycete plant pathogens have been extensively studied at the molecular level,
Ps. cubensis and the molecular basis of its interaction with cucurbit hosts has not been
well examined. Here, we present the first large-scale global gene expression analysis of
Ps. cubensis infection of a susceptible Cucumis sativus cultivar, ‘Vlaspik’, and
identification of genes with putative roles in infection, growth, and pathogenicity. Using
high throughput whole transcriptome sequencing, we captured differential expression of
2383 Ps. cubensis genes in sporangia and at 1, 2, 3, 4, 6, and 8 days post-inoculation
(dpi). Additionally, comparison of Ps. cubensis expression profiles with expression
profiles from an infection time course of the oomycete pathogen Phytophthora infestans
on Solanum tuberosum revealed similarities in expression patterns of 1,576-6,806
orthologous genes suggesting a substantial degree of overlap in molecular events in
virulence between the biotrophic Ps. cubensis and the hemi-biotrophic P. infestans. Coexpression analyses identified distinct modules of Ps. cubensis genes that were
representative of early, intermediate, and late infection stages. Collectively, these
expression data have advanced our understanding of key molecular and genetic events
in the virulence of Ps. cubensis and thus, provides a foundation for identifying
mechanism(s) by which to engineer or effect resistance in the host.

95

INTRODUCTION
The phytopathogenic oomycete Pseudoperonospora cubensis, the causative agent of
cucurbit downy mildew [1,2], infects a wide range of cucurbits, including cucumber
(Cucumis sativus L.), squash (Cucurbita spp.), and melon (Cucumis melo L.). As an
obligate biotroph, Ps. cubensis is dependent on its host for both reproduction and
dispersal, and as such, has evolved a highly specialized host range limited to members
of the Cucurbitaceae. At present, downy mildew is the most important foliar disease of
cucurbits, affecting cucurbit production throughout the world [1,2]. Under favorable
conditions, Ps. cubensis is capable of infecting and defoliating a field in less than two
weeks, and as a result, is responsible for devastating economic losses. For more than
50 years, control of downy mildew on cucumber in the U.S. was maintained through
genetic resistance; however, since 2004, the likely introduction of a new pathotype into
U.S. pathogen populations has resulted in a loss of this resistance [1]. While minimal
knowledge of the genetic variation within Ps. cubensis exists - specifically related to
virulence, pathogenicity, and host specificity among physiological races - the genetic
basis of these processes, and the underlying mechanism(s) associated with infection
have not been elucidated [1,2,3,4].

To date, analyses of the Ps. cubensis-C. sativus interaction have been limited to the
identification of the aforementioned physiological races, and have largely focused on
the utilization of variation in host specificity for the identification and classification of
pathotypes [3,5]. To this end, six physiological pathotypes, or races, have been
identified within populations in the U.S., Israel, and Japan, as well as additional races
96

throughout Europe [1,2,4]. In the U.S., increased disease pressure on cucumber
production since 2004 is hypothesized to be the result of the introduction of a new, more
virulent pathotype, capable of overcoming the downy mildew resistance gene dm-1, that
has been widely incorporated into commercial cucumber varieties since the 1940’s [6].
While genetic analyses such as Amplified Fragment Length Polymorphism have been
used to differentiate these physiological races [4] and some effort has been made to
refine the species within Pseudoperonospora [6,7], there is limited information available
about pathogenicity or virulence genes in Ps. cubensis or the molecular-genetic basis of
resistance to this pathogen in the cucurbits.

Recent work generated the first sequence assembly of the Ps. cubensis genome and
subsequent in silico analysis has identified candidate effector proteins that may have
either virulence or avirulence roles in Ps. cubensis infection [8,9]. Structurally, oomycete
effector proteins display a modular organization, consisting of a N-terminal signal
peptide, a conserved RXLR (Arg-X-Leu-Arg, where “X” is any amino acid) translocation
motif, followed by a variable C-terminal effector domain [10]. In short, it is the function
and activity of the variable C-terminal effector domain that drives the activity of these
molecules [10,11]. A set of 61 candidate effectors were identified in the first draft of the
Ps. cubensis genome [8] and included a large class of variants with sequence similarity
to the canonical RXLR motif. Specifically, the function of a QXLR-containing effector,
designated PcQNE, was characterized and shown to be a member of a large family of
Ps. cubensis QXLR nuclear-localized effectors, which was up-regulated during infection
of cucumber [8]. Additionally, internalization of PcQNE into the host cell was shown to

97

require the QXLR-EER motif, thereby establishing a basic functional homology with the
well-characterized Phytophthora spp. effector proteins [8]. While this work serves as a
substantial development in understanding the genetic basis for pathogenicity in Ps.
cubensis, additional work is needed to identify and characterize additional effectors and
other proteins involved in establishment of infection and pathogen proliferation.

The accessibility of oomycete pathogen genome sequences, combined with gene
expression data from both pathogen and host throughout the course of infection, can
serve as a basis for identification and curation of genes that may have important roles in
both virulence and avirulence [12,13,14,15]. To date, oomycete RXLR effectors have
been demonstrated to suppress basal host resistance [16,17], as well as to activate
effector-triggered immunity (ETI) [18,19,20,21]. In addition to the RXLR class, other
cytoplasmically-localized effectors have been identified in Phytophthora spp. [11]. The
Crinkler (CRN) family, for example, has a conserved LXLFLAK motif necessary for
translocation into the host cytoplasm and subsequent import into plant nuclei where
they elicit a rapid cell death response [22,23]. Finally, oomycete effectors have also
been shown to function within the host apoplast, including functions as enzyme
inhibitors [24,25,26,27,28], small cysteine-rich proteins [22,29,30], the Nep1-like family
of proteins [31,32], and CBEL (Cellulose Binding, Elicitor, and Lectin-like) proteins
[33,34].

The initial stages of pathogen infection of a plant host involve adhesion, penetration,
and invasive growth within the host cell tissue. As such, cell wall degrading enzymes

98

(CWDE), such as endoglucanases, cutinases, cellulases, and β-glucanases have
evolved as essential components of an oomycete’s repertoire for cell wall penetration
[20]. Numerous CWDE have been identified computationally from the genomic
sequences of several plant pathogenic oomycetes, including Phytophthora sojae,
Phytophthora ramorum, Hyaloperonospora arabidopsidis, and Pythium ultimum
[30,35,36]. In P. sojae, members of the family 5 and family 12 endoglucanases have
been shown to be up-regulated during early stages of infection [37,38]. However, in H.
arabidopsidis, which causes downy mildew of Arabidopsis thaliana, CWDE-encoding
mRNAs are reduced [36]. This could indicate an adaptation in downy mildew pathogens
for evasion of recognition by their host, as break-down products from plant cell wall
components can function as elicitors of defense responses [39].

Recent advancements in sequencing technologies have led to an explosive growth in
the

analysis

of

in

planta-expressed

genes

of

biotrophic

plant

pathogens

[12,40,41,42,43,44]. In the current study, we present the first global gene expression
analysis of the infection stages of cucumber by the obligate oomycete pathogen Ps.
cubensis, the causal agent of cucurbit downy mildew. Through the analysis of a
susceptible cucumber cultivar interaction, we describe the identification of genes with
putative roles in infection, growth and pathogenicity. Using next-generation sequencing
technology, we assessed gene expression in Ps. cubensis in sporangia and at six time
points of infection. By combining visual assessment of symptoms with light microscopy
to monitor infection stages as well as minimizing collection of non-inoculated tissues, we
were able to capture expression of 7,821 Ps. cubensis genes ranging from 159 genes at

99

1 days post inoculation (dpi) to 7,698 at 8 dpi. In total, this work represents a
comprehensive examination of the key infection stages of Ps. cubensis growth and
development. In total, the work described herein provides a foundation for further
dissection of genes relevant to virulence in this obligate phytopathogen.

RESULTS AND DISCUSSION

Characterization and sampling of Ps. cubensis infection stages

While Ps. cubensis is a major pathogen of cucumber and other cucurbits, limited
resources describing the infection process and/or virulence determinants of this obligate
oomycete are available. In the current study, we sought to identify Ps. cubensis gene
expression from both purified sporangia, as well as from a time course of infected
cucumber tissues, representing a wide range of infection stages from 1 to 8 dpi. In total,
our goal was to gain a broad perspective of in planta gene expression during infection of
a susceptible cucumber host and to correlate this expression with the development of
outwardly visible symptoms, as well as the development of microscopic pathogen
infection structures. Like other phytopathogenic downy mildews and biotrophic fungi,
Ps. cubensis is non-culturable, and proliferates and reproduces only on a susceptible
cucurbit host. As with previously published reports on analyzing gene expression in
biotrophic phytopathogens, optimization of sampling techniques is key to maximize
pathogen tissue compared to host, particularly at early stages of infection (Figure 3.1)
[12,43,44,45].

100

Plants were inoculated on the abaxial leaf surface with purified Ps. cubensis sporangia,
and samples were collected using a cork borer, minimizing the amount of non-infected
tissue in each sample (Figure 3.1). Initial symptoms of downy mildew infection can be
observed on the abaxial leaf surface at 1-3 dpi as water soaking at the site of
inoculation, while no visual symptoms are apparent on the upper leaf surface (Figure
3.2). At 1 dpi, zoospores were encysted upon stomata on the lower leaf surface, and by
2 dpi, appressoria and initial penetration hyphae were visible beneath stomata. The
yellow angular lesions typical of cucurbit downy mildew were apparent on the upper leaf
surface by 4 dpi, and over time, became more chlorotic and necrotic as the infection
progressed. By 3 to 4 dpi, multiple haustoria formed within the mesophyll layer.

mRNA-Seq data analyses

Expression profiling of Ps. cubensis sporangia, as well as infection stages at six time
points of cucumber infection, were performed using mRNA-Seq. For each time point,
two biological replicates were sequenced. The total number of reads produced for each
time point ranged from 55 to 59 million reads, with a median of 57 million reads. Reads
were mapped to the Ps. cubensis genome which was generated by assembly of
Illumina next generation reads; in total the Ps. cubensis genome encompasses 67.9
Mb, with 23,519 protein coding genes and 23,522 gene models (AHJF00000000;

101

Figure 3.1

102

Figure 3.1 (cont'd) Experimental design and sample collection. A 1 x 10e5
sporangia/ml solution of Pseudoperonospora cubensis was used to inoculate the
abaxial leaf surface of cucumber cultivar ‘Vlaspik’. Samples were collected using a #3
cork borer to minimize uninfected tissue (black circles) at 1, 2, 3, 4, 6, and 8 days
post-inoculation (dpi). Leaf disks were used for microscopic analysis of infection
stages or pooled for RNA extraction. RNA-Seq libraries were made for each time
point from 2 biological replicates. Within a biological replicate, libraries were
barcoded and sequenced in multiple lanes. The sporangia-only library (SP) was not
barcoded and was sequenced on its own.

[9]). Of the total reads generated, for each time point, approximately 1.6 to 6.4 million
(3-12% of the total; Figure 3.3A) mapped to the Ps. cubensis genome. In turn, a
majority of reads in each sample were of host origin, and mapped to the cucumber
genome (see accompanying paper, [46]) (Figure S3.1). Through this analysis, we found
that there was no significant difference in the total number of combined reads from
different time points (p > 0.80); however, the number of Ps. cubensis genes expressed
at each time point was significantly different for all time point comparisons (p < 0.05;
Figure 3.3A). To assess the experimental variation attributable to biological variation,
we compared the gene expression pattern of the genes expressed in both of our
biological replicates. In total, our experiments showed very high levels of correlation for
biological replicates (in all cases examined, Pearson’s Correlation Coefficient (PCC) >
0.94; Figure S3.2), indicating that our sampling, assay, and analysis methods are
robust.

To evaluate the effect of sampling depth on gene expression detection and to assess
whether we have adequately sampled the mixed mRNA-Seq read pool for Ps. cubensis
transcripts, subsets of five to 30 million reads were randomly selected from the total

103

Figure 3.2

Figure 3.2 Symptoms and microscopy images of Ps. cubensis infected Cucumis sativus cultivar 'Vlaspik' of
time points used for transcriptome analysis. Symptom images were collected of the adaxial (top row) and abaxial
(middle row) at 1, 2, 3, 4, 6, and 8 days post-inoculation (dpi). Microscopy (bottom row) to assess stages of Ps. cubensis
invasion were collected from the same time points using ethanol-cleared, trypan blue stained samples. Scale bars at 1-4
dpi are 25 um. Scale bars at 6 and 8 dpi are 50 um. Dotted lines represent position of stomata relative to the pathogen
structure. e = encysted zoospore. s = stomate. h = haustorium.
104

read pool from each time point and mapped to the Ps. cubensis genome. The
simulation experiment showed a clear positive relationship between sampling depth and
number of expressed genes at the lower to medium sequencing depth (5 to 20 million
reads; Figure 3.3B). With the exception of 1 dpi in which few genes are expressed, after
20-25 million reads, the number of expressed genes begin to plateau, corresponding to
the minimum sampling depth of all libraries sequenced in this study.

mRNA-Seq transcriptome profiles

In concordance with our visual assessment (Figure 3.2) of pathogen growth throughout
the time course, our analyses showed a diversity of transcriptional changes in Ps.
cubensis, as well as a correlation between gene expression levels and similar stages of
pathogen growth. In support of this, we identified 7,821 genes expressed at different
time points of infection (Table S3.1, available at www.plosone.org, e35796) and 129 of
those genes (Table S3.2, available at www.plosone.org, e35796), mostly housekeeping,
were expressed throughout all time points. Analyses of the top 20 highly expressed
genes showed that genes expressed at earlier time points have substantially higher
FPKM (fragments per kilobase pair of exon model per million fragments mapped) values
than the genes expressed at later time points, consistent with the fewer numbers of
genes expressed in the early stages of expression and saturation of detection of Ps.
cubensis

expression

with

our

sampling

depth

(Table

S3.3,

available

at

www.plosone.org, e35796). For all time points analyzed, the minimum FPKM value was
zero but the maximum FPKM values ranged from 8,528 at 8 dpi to 270,121 at 1 dpi

105

Figure 3.3

106

Figure 3.3 (cont'd)

Figure 3.3 Number of total RNA-seq reads, reads mapped, and number of genes expressed. (A). Total number of
reads, number of reads mapped to the Ps. cubensis genome, and number of genes expressed by Ps. cubensis at
different time points are shown. Reads were mapped using Bowtie version 0.12.5 and TopHat version 1.2.0. Fragments
per kilobase pair of exon model per million fragments mapped (FPKM) values were calculated using Cufflinks version
0.9.3. Genes were considered expressed if the FPKM and 95% confidence interval lower boundary FPKM value was
greater than 0.001 and zero, respectively. dpi = days post-inoculation. (B) Comparison between number of expressed
genes detected and sampling depth. For all time points 5, 10, 15, 20, 25, and 30 million reads were randomly selected
from the total pool of reads from different time points. Read mapping and expression abundances are as described in
panel 3A.
107

(Table S3.3, available at www.plosone.org, e35796). The differences in transcriptome
profiles are clearly visible in correlation and cluster analyses between the sampled time
points. The Pearson Correlation Coefficient (PCC) values for comparisons of different
time points ranged from 0.26 (1 dpi vs. 8 dpi) to 0.79 (3 dpi vs. 4 dpi) (Figure 3.4).
Corresponding with our visual assessment of pathogen infection stages showing similar
growth at 2 and 3 dpi (i.e., penetration and initial hyphal growth, Figure 3.2), gene
expression patterns at 2 dpi are strongly associated with that of 3 dpi. Similarly, genes
expressed at 6 dpi showed high correlation with the genes from 8 dpi, which represent
comparable stages of pathogen growth and proliferation in the mesophyll (Figure 3.2).
Additionally, at 1 dpi when encystment of zoospores is occurring, we observed a poor
correlation (PCC ranged from 0.26 to 0.45) between expression at that time point with
any other time point, likely due to the unique set of genes that would be involved in this
process. Interestingly, gene expression at 3 dpi was highly correlated with other time
points (PCC ranged from 0.45 to 0.79), suggesting that events occurring at 3 dpi may
represent a transition phase between early and late infection.

Differential gene expression

To identify genes specifically involved in distinct stages of pathogen infection and
development, and to assess gene expression pattern changes over the course of
infection, we next evaluated differentially expressed genes between all time points. To
provide context for the differential expression, we included expression data from mRNASeq analysis of sporangia in which 8,254 Ps. cubensis genes were expressed (Table

108

Figure 3.4

Figure 3.4 Correlation matrix of Pseudoperonospora cubensis expression
profiles throughout a time course of Cucumis sativus infection. Normalized
transcript abundances for 7,821 genes were calculated in units of fragments per
kilobase pair of exon model per million fragments mapped (FPKM) with Cufflinks
version 0.9.3. Pearson product-moment correlations (PCC) of log2 FPKM values were
calculated for all pair-wise combinations using R. PCCs were clustered using
hierarchical clustering with a Pearson correlation distance metric and average linkage
using Multiple Experiment Viewer Software version 4.5. The bootstrap support values
shown on tree nodes were obtained from 1000 bootstrap replicates. dpi = days postinoculation.
109

S3.4, available at www.plosone.org, e35796). In concordance with the similarities
observed during our visualization of pathogen growth, comparisons with the least
number of differentially expressed genes are those between early time points, with only
147 (2%) genes differentially expressed between 1 and 2 dpi (Table 3.1). Additionally, 1
and 2 dpi had fewer differentially expressed genes compared to sporangia than the later
time points. Of all the combinations tested, 1 dpi had the highest percentage of
differentially expressed genes across all pair-wise comparisons, despite having the
lowest number of genes expressed, indicating that the events occurring at this stage of
infection are unique among the time course. This corresponds both with our cluster
analysis above (Figure 3.4) and our microscopic analysis of infection (Figure 3.2).
Interestingly, despite the high correlation between expression patterns at 2 and 3 dpi
seen using cluster analysis, there were a large number of genes differentially expressed
between 2 and 3 dpi. This is additionally supportive of our hypothesis that 3 dpi is a
transition phase between early and late infection. Not surprisingly, the highest number
of differentially expressed genes were observed in comparison of all other time points
and sporangia with 8 dpi, suggestive of an advanced stage of the infection process and
a likely transition to processes involved in sporulation.

Expression of genes involved in virulence and pathogenicity

As an obligate biotroph, Ps. cubensis must evade and/or overcome basal plant defense
responses, as well as effector-triggered immunity, in order to establish growth,
proliferate and reproduce within its host. This, as in other phytopathogens, is likely

110

Table 3.1 Number of differentially expressed genes between each combination of
time points and sporangia. Differential expression analysis was conducted using the
CuffDiff program in Cufflinks version 0.9.3 using the Pseudoperonospora cubensis
annotation with a false discovery rate of 0.01.

1 dpi
2 dpi
3 dpi
4 dpi

2 dpi†
147
(50%)§

3 dpi
193
(58%)
848
(28%)

4 dpi
189
(57%)
329
(10%)
560
(14%)

6 dpi
8 dpi
†

6 dpi
175
(60%)
306
(7%)
301
(7%)
342
(7%)

8 dpi
192
(59%)
898
(19%)
891
(17%)
820
(16%)
644
(10%)

Sporangia
177
(50%)
246
(16%)
391
(13%)
425
(14%)
559
(15%)
1,301
(32%)

dpi = days post-inoculation.
Numbers in parenthesis indicate the percent of significantly different
tests out of the total number of tests that could be performed for each
pairwise comparison.
§

achieved through the secretion of specific effector proteins that function within the host
apoplast to interfere with extracellular plant defense responses, such as the activity of
glucanases and proteases or cytoplasmically to suppress defense responses. Thus, the
identification and characterization of the temporal expression of pathogen-associated
genes throughout the course of infection can assist in the identification of secreted
effectors that allow for both the promotion of disease, as well as the avoidance of host
recognition. In support of this, we identified a suite of 271 candidate RXLR-type
effectors within the Ps. cubensis genome [9] with 20 possible amino acid substitutions
at position R1, including R and Q. In the current study, we analyzed the expression
distribution of all 271-candidate effectors, as well as predicted Crinkler (CRN) effectors,
111

over our time course of infection. As shown in Figure 3.5, the greatest number of
expressed candidate effectors have an ASLR (Ala-Ser-Leu-Arg) motif, and are
expressed at 2-8 dpi; candidate effectors with RXLR or QXLR motifs were expressed at
every time point. Finally, as noted above, the expanded repertoire at the conserved
RXLR motif in Ps. cubensis, with a total of 20 amino acids represented at the R1
position, represents a diversity in RXLR-type effectors previously undescribed. The
simplest explanation for this expansion at R1 is supported by the hypothesis that RXLRtype effectors may play a role in host range, and that an expanded effector repertoire
may impart plasticity. Moving forward, an extensive functional characterization of these
RXLR-type effectors will provide insight into both pathogen virulence and host range
specificity. Nonetheless, our data suggest that Ps. cubensis possesses a potentially
highly expanded virulence capacity, of which, we have determined the expression of
271 RXLR-type effectors over an extensive time-course of susceptibility and disease
elicitation in cucumber.

Gene families encoding host-targeted hydrolytic enzymes acting on plant proteinases,
lipases, and several sugar-cleaving enzymes (carbohydrate active enzymes; CAZymes)
were highly expressed in Ps. cubensis at 4 to 8 dpi, suggesting a possible role during
infection and proliferation (Figure 3.6). Comparison of glycoside hydrolase (GH),
glycosyltransferases (GT), polysaccharide lyase (PL), pectin esterase (PE), and
carbohydrate esterase (CE) encoding genes revealed significant differences in number
that were expressed as well as diversity across different time points. In total, 178 GH,
135 GT, 2 CE, and 15 PE were expressed throughout all the time points sampled

112

Figure 3.5

Figure 3.5 Candidate effectors expressed at different timepoints. From inner to
outermost circles; 2 dpi, 3 dpi, 4 dpi, 6 dpi and 8 dpi. CRN = Crinkler effectors. dpi =
days post-inoculation.

(Figure 3.6). GH was the most represented family, with expression of 30-78 members
followed by GT (17-27 members expressed). The most represented GH families
identified were GH3 and GH5, while GT20 and GT48 were the most represented among
all GTs. Additionally, substantial differences were observed in the types of CAZymes
expressed across different time points. For example, several members of GH (GH
family 7, 12, and 31), GT (GT family 1), CE (CE family 5), and PL family were absent in
113

Figure 3.6

Figure 3.6 CAZymes in Pseudoperonospora cubensis expressed during
infection on Cucumis sativus. The CAZymes coding genes in the Ps. cubensis
genome were annotated using CAZymes Analysis Toolkit- CAT according to the
CAZy database in combination with protein family domain analyses. Gene families
absent in at least one time point are underlined. CBM = carbohydrate binding
module. CE = carbohydrate esterase. GH = glycoside hydrolase. GT = glycosyl
transferase. PE = pectin esterase. PL = polysaccharide lyase. dpi = days postinoculation.
114

early time points (i.e., 2 and 3 dpi), yet were expressed at 4 to 8 dpi, suggesting a
possible role during the later stage of infection. GH family 12 endoglucanases as well as
CE family 5 cutinases have been previously implicated as having a role in infection by
Phytophthora spp. [38,47,48].

Comparison to genes induced during P. infestans infection of potato

The comparison of gene expression patterns between pathogens during infection of
their susceptible hosts can allow for identification of common genes that are specifically
involved in pathogenesis, as well as enable the discovery of genes unique to either
species. To this end, we chose to compare the gene expression pattern of Ps. cubensis
during infection to that of another economically important oomycete pathogen, P.
infestans, during the infection of potato, Solanum tuberosum. Using clustering analysis
of protein coding genes from both pathogens, we identified 7,374 single copy
orthologous genes between these two oomycetes. We then compared the gene
expression values obtained from our study (PCU) with those from microarray-based
expression profiling [10] experiments with P. infestans-S. tuberosum (PITG). Spearman
rank correlation coefficients (SCCs) of log2 expression values were calculated between
the single copy orthologs at all time points in the two datasets; between 1,576 and
5,581 genes were included in the pair-wise comparisons (Figure 3.7). The SCC values
among all comparisons ranged from 0.12 to 0.76 (Table S3.5, available at
www.plosone.org, e35796). Comparisons between time points reflecting similar stages
of pathogen infection showed higher overall correlations (0.29 to 0.76) as compared to

115

Figure 3.7

Figure 3.7 Comparison of ribonucleic acid sequencing (RNA-seq) and
microarray expression patterns. Microarray expression profiles were obtained from
time course analyses of genes expressed in P. infestans (PITG) during infection on
S. tuberosum. Single copy orthologous genes between P. infestans and Ps. cubensis
(PCU) were identified using OrthoMCL. Log2 transformed expression values of single
copy orthologous genes present in both the Ps. cubensis (log2 of fragments per
kilobase pair of exon model per million fragments mapped [FPKM]) and P. infestans
(log2 intensity) datasets are shown as scatter plots. SCC Spearman correlation
coefficient. dpi = days post-inoculation
116

comparisons between dissimilar time points (0.12 to 0.47). The most highly correlated
comparisons were those between genes expressed in Ps. cubensis at 4 dpi and P.
infestans at 4 dpi (SCC = 0.76). In P. infestans infection on potato, days 2-4 correspond
to haustoria formation [10]; likewise, extensive formation of haustoria by Ps. cubensis
was observed at 4 dpi (Figure 3.2). Correspondingly, genes expressed at 4 dpi include
a haustorium-specific membrane protein, secreted RXLR proteins, as well as an amino
acid transporter, which could possibly be involved in nutrient uptake via the haustorium
to Ps. cubensis. Gene expression was also highly correlated (SCC = 0.64) between Ps.
cubensis 6 dpi samples and P. infestans 5 dpi samples. At 6 dpi, symptoms of Ps.
cubensis infection on cucumber manifest as chlorotic yellow lesions (Figure 3.2).
Similarly, at 5 dpi, Ph. infestans has entered the mycelial necrotrophic growth stage,
showing a similar chlorotic phenotype on its host [10].

Gene co-expression network analyses

Correlation analyses in which associations between gene expression patterns are
identified are valuable for inferring common function and/or regulatory relationships [49].
In this study, we were primarily interested in identifying genes that are involved in both
establishment and maintenance of Ps. cubensis infection, as well as those specifically
involved in virulence. To this end, we constructed gene modules to identify highly coexpressed genes, where all members of a module are more highly correlated with each
other than to genes outside the module. Using a Coefficient of Variation (CV) cutoff of
1.0, 4,195 genes from an initial total of 7,821 expressed genes were retained for

117

downstream analyses. Using Weighted Gene Correlation Network Analysis (WGCNA)
[50], 3,146 genes were assigned to six different gene modules (Modules 1 to 6)
containing 107 to 1,312 genes; 1,049 genes were not assigned to any module (Table
S3.6, available at www.plosone.org, e35796; Figure S3.3). Eigengenes [51] were
calculated for each module and displayed in a heat map (Figure 3.8) revealing discrete
gene expression patterns across different time points.

As described above, the 1 dpi sample represents both an important initial stage in the
infection process, as well as a unique gene expression profile among the infection time
points analyzed. This is additionally reflected in Module 1, which contains 146 genes
that are highly expressed at 1 dpi, including genes involved in pathogenesis and
transport (Figure 3.8, Figure S3.3, Table S3.6, available at www.plosone.org, e35796).
The genes in this module are also expressed at 3 dpi, indicating that there may be
some similarities between processes involved in zoospore adhesion and encystment,
and initiation of haustoria formation. Genes expressed in Module 2 could represent
those processes involved in the transition from early to late stages of infection and that
are involved in the initial suppression of host defenses and establishment. This module,
which contains 508 genes, expressed at 2, 3, and 4 dpi, represents gene expression
occurring during initial penetration through the stomata into the host tissue, hyphal
growth, and initiation of haustoria formation. It includes genes such as candidate RXLRtype effectors, glucanase inhibitors, CRNs, and a haustorium-specific membrane protein
(Table S3.6, available at www.plosone.org, e35796), similar to what has been observed

118

Figure 3.8

Figure 3.8 Heat map of the eigengenes representing each gene module. The
columns in the heat map represent time points, and the rows represent eigengenes
for each of the six identified co-expression modules. The numbers of genes in each
module are given in parentheses. The cells in the heat map show eigengene values
between 0 and 1, indicators of relative expression levels of all genes in the module
(see Materials and Methods). dpi = days post-inoculation.

to be up-regulated during P. infestans infection on potato [10]. Our WGCNA analyses
additionally identified genes that are co-expressed during the later stages of infection,
specifically in Modules 4, 5, and 6 (Figure 3.3, Figure S3.2). Module 6, in particular,
119

represents genes most highly expressed at 8 dpi possibly indicative of those involved in
a shift to the reproductive phase and sporulation.

Transcription factors (TFs) are reported to play a key role in the regulation of many
biological processes including roles in oomycete pathogenesis [52,53] and within the
predicted Ps. cubensis proteome 27 transcription factor-related domains in 440 genes
were identified (Table S3.7, S3.8, available at www.plosone.org, e35796). A total of 247
of these were expressed throughout the infection process (Table S3.7, available at
www.plosone.org, e35796). We also identified genes encoding transcription factorrelated Pfam domains in all six co-expression modules. Two modules, 2 and 4, with
genes co-expressed across different time contained the majority of the transcription
factors. The transcription factor-related genes within those modules could play
important role in regulation of genes involved in pathogenesis. The bZIP and Myb, DNAbinding transcription factor, which play an important role in oomycete pathogenesis
[54,55], were the most abundant transcription factor-related domains expressed during
infection.

CONCLUSIONS

In this study, we present an extensive characterization of the gene expression analysis
of the obligate oomycete cucurbit pathogen Ps. cubensis during a compatible
interaction. This data set represents the first global gene expression profile of a cucurbit
pathogen. Using mRNA-Seq, we analyzed the differential expression of pathogen genes

120

across a time course of infection of cucumber, correlating expression with pathogen
infection structures, development, and the onset of disease symptoms. Our study
provides a comprehensive examination of the key infection stages of Ps. cubensis
growth and development and through clustering and co-expression network analyses,
describes genes that are specifically expressed during these stages. In addition, our
work has identified an expanded effector repertoire, represented by a unique diversity at
the canonical RXLR motif. Overall, the work described herein will significantly enhance
our understanding of the regulation of infection of oomycete phytopathogens, as well as
a baseline for identifying important virulence determinants in Ps. cubensis.

MATERIALS AND METHODS

Ps. cubensis inoculation and sample collection

Ps. cubensis MSU-1 was maintained on Cucumis sativus cultivar 'Vlaspik' as described
previously [8]. Four-week-old cucumber plants were inoculated on the abaxial surface of
the first fully-expanded leaf with a 1 x 105 sporangia/ml solution with 20-30 10 μl
droplets. Inoculated plants were maintained at 100% relative humidity in the dark for 24
hours and then transferred to growth chambers maintained at 22 °C with a 12 h
light/dark photoperiod. Samples were collected at 1, 2, 3, 4, 6, and 8 dpi with a #3 cork
borer to collect tissue at the site of inoculation. Samples for RNA extraction were frozen
in liquid nitrogen and stored at - 80 °C until use. Samples collected for microscopy were
cleared in 95 % ethanol and stored at room temperature.

121

Histological assessment of Ps. cubensis growth

Cleared infected leaf discs were stained in a solution of 250 μg/ml trypan blue in equal
parts lactic acid, water, and glycerol to visualize infection structures. Microscopy was
performed using an Olympus IX71 inverted light microscope. Images were captured
using an Olympus DC71 camera and were processed for contrast using Canvas X
(ACD Systems International, Inc., Seattle, WA).

Library preparation and sequencing

Sporangia were washed from the abaxial surface of heavily sporulating leaves, filtered
through a 40 μm nylon cell strainer, and pelleted via centrifugation. For RNA extraction
(RNeasy Mini Kit, Qiagen, Valencia, CA), sporangia were resuspended in 450 μl RLT
buffer with ~50 μl 425-600 μm acid-washed beads and vortexed for 3 minutes to break
cells. Additional extraction steps were followed according to the manufacturer's
instructions. RNA concentration and quality was determined using the Bioanalyzer 2100
(Agilent Technologies, San Diego CA). The sporangia library was sequenced in two
lanes at the UC DNA Sequencing Facility at University of California, Davis (Davis, CA).
RNA samples from the infection time course were processed as described in Adhikari et
al. (accompanying paper [46]). In brief, RNA was isolated using the RNeasy Mini Kit
(Qiagen, Germantown, MD), treated with DNase (Promega, Madison, WI) and barcoded
libraries constructed with the Illumina mRNA-seq kit (Illumina, San Diego CA). Libraries

122

were sequenced with the Illumina Genome Analyzer II platform generating 35-42 bp
single-end reads. Reads from biological replicates were pooled prior to expression
abundance measurements. Reads were deposited in the National Center for
Biotechnology

Information

Sequence

Read

Archive

under accession

number

SRP009350.

mRNA-Seq read mapping and transcript abundance estimation

The

assembled

and

annotated

Ps.

cubensis

MSU-1

genome

sequence

(AHJF00000000; [9]) was used to estimate transcript abundances. mRNA-Seq reads for
each time point and control (sporangia) were mapped to the 67.9 Mb Ps. cubensis
reference genome using the quality aware alignment algorithms, Bowtie version 0.12.7
[56] and TopHat version 1.2.0 [57]. The single-end reads from different time points were
aligned in single-end mode while the paired-end reads from the control were aligned in
paired-end mode. The minimum and maximum intron length was set to 5 and 50,000
bp, respectively and the insert size for paired-end mode was set to 140 bp.

The aligned read files produced by TopHat were processed by Cufflinks v0.9.3 [58]. A
reference annotation of the Ps. cubensis genome (23,519 loci and 23,522 gene models)
was provided and the maximum intron length was set to 50,000 bp. Normalized gene
expression levels were calculated and reported as FPKM. The quartile normalization
option was used to improve differential expression calculations of lowly expressed
genes [58]; all other parameters were used at the default settings. A gene was

123

considered expressed in a specific sample if the FPKM value and FPKM 95%
confidence interval lower boundary was greater than 0.001 and zero, respectively.

Pearson product-moment correlation analyses of log2 FPKM values among mRNA-Seq
libraries were performed using R (http://cran.r-project.org/), with all log2 FPKM values
less than zero set to zero. Only tests significant at p = 0.05 are shown. Correlation
values depicted as a heat map were clustered with hierarchical clustering using a
Pearson correlation distance metric and average linkage. The bootstrap support values
were calculated from 1000 replicates using Multiple Experiment Viewer Software (MeV)
v4.5 [59]. To understand variability among biological replicates, Pearson correlation
coefficients were calculated for the log2 transformed FPKM values of the genes
expressed in both replicates at a particular time point.

Identification of differentially expressed genes

Once transcript abundance estimation was calculated, differential expression analysis
was conducted using the Cuffdiff program within Cufflinks version 0.9.3 [58] utilizing the
read alignment files described above. The expression testing was done at the level of
genes. Quartile normalization [60] and a false discovery rate of 0.01 after BenjaminiHochberg correction for multiple testing were used. The Ps. cubensis genome and the
annotation files were provided as input parameters. All other parameters were used at
the default levels. Cuffdiff was used to perform pairwise comparisons of six time points
and sporangia.

124

mRNA-Seq and microarray comparative analyses
Gene expression data from a P. infestans-S. tuberosum time course experiment [10]
was used to assess gene expression pattern similarities/differences in two ooymcete
pathogens. The data set included P. infestans gene expression over a five-day (2-5
days) time course of a potato infection. Raw data was downloaded from the Gene
Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) (GSE14480) [61]. The probe
intensities were normalized using Robust Multichip Analyses method [62]. For the
mRNA-Seq to microarray comparative analysis, single copy orthologous genes were
identified using OrthoMCL [63,64] with default parameters. Clustering of 23,522 and
18,140 protein-coding genes from Ps. cubensis and P. infestans, respectively, yielded
7,374 clusters with single copy genes from both species. Only single copy orthologous
genes were used for the analyses. FPKM and probe intensity values were log2
transformed, and Spearman correlation coefficients were calculated using R
(http://cran.r-project.org/).

Functional analysis

Functional annotation for all Ps. cubensis genes were generated from searches of the
UniProt databases (Uuniref100) [65] with BLAST and combined with Pfam [66] protein
families assignment performed using HMMER3 [67]. Functional annotations for Ps.
cubensis sequences were taken from the best possible UniRef sequence match, but if
there was no UniRef sequence match, functional annotations were made based on the
best Pfam domain alignment. Transcription factors were identified based on PFAM
125

domains.

Gene co-expression network analysis

Gene co-expression network analysis was done according to the methods described by
Childs et al. [68] with some modifications. First, the FPKM gene expression values were
log2 transformed and FPKM values less than 1 were transformed to zero. Second,
genes showing no variation across time points were filtered out using a coefficient of
variance (CV) cutoff (1.0). Third, the β and treecut parameters were 7 and 0.6,
respectively. Eigengenes were calculated using the WGCNA package [51]. The heat
map of eigengenes for each gene module was constructed using R (http://cran.rproject.org/). Genes assigned to co-expression modules were annotated based on the
Ps. cubensis functional annotation.

126

APPENDIX

127

Figure S3.1

Figure S3.1 Number of total RNA-seq reads, reads mapped, and number of genes expressed at different
timepoints. Total number of reads, number of reads mapped and number of genes expressed in Cucumis sativus
mapped,
(Cusa) and Pseudoperonospora cubensis (Pcu) at different time-points are shown. Reads were mapped using Bowtie
version 0.12.5 and TopHat version 1.2.0. Fragments per kilobase pair of exon model per million fragments mapped
(FPKM) values were calculated using Cufflinks version 0.9.3. Genes were considered expressed if the 95% confidence
interval lower boundary FPKM value was greater than zero. dpi = days post-inoculation.
128

Figure S3.2

129

Figure S3.2 (cont'd) Concordance of FPKM values of the genes expressed in two biological replicates of the
Pseudoperonospora cubensis transcriptome. Reads from different timepoints were mapped to Ps. cubensis genome
using Bowtie version 0.12.5 and TopHat version 1.2.0. Fragments per kilobase pair of exon model per million fragments
mapped (FPKM) values were calculated using Cufflinks version 0.9.3 and Ps. cubensis genome annotations. For each
time point Log2 transformed FPKM values of equal number of genes from both replicates are plotted. Pearson Correlation
Coefficient
(PCC)
was
calculated
using
R.
dpi,
days
post-inoculation.

130

Figure S3.3

Figure S3.3 Trend plots of the normalized gene expression values for each gene from six identified gene coexpression modules. Modules consisting of genes expressed modules 1, 2, 3, 4, 5, and 6 are shown.
131

REFERENCES

132

REFERENCES
1. Savory EA, Granke LL, Quesada-Ocampo LM, Varbanova M, Hausbeck MK, et al.
(2011) The cucurbit downy mildew pathogen Pseudoperonospora cubensis. Mol
Plant Pathol 12: 217-226.
2. Lebeda A, Cohen Y (2011) Cucurbit downy mildew (Pseudoperonospora cubensis)—
biology, ecology, epidemiology, host-pathogen interaction and control. Eur J
Plant Pathol 129: 157-192.
3.

Thomas C, Inaba T, Cohen Y (1987) Physiological
Pseudoperonospora cubensis. Phytopathol 77: 1621-1624.

specialization

in

4. Sarris P, Abdelhalim M, Kitner M, Skandalis N, Panopoulos N, et al. (2009) Molecular
polymorphisms between populations of Pseudoperonospora cubensis from
Greece and the Czech Republic and the phytopathological and phylogenetic
implications. Plant Pathol 58: 933-943.
5. Lebeda A, Widrlechner MP (2003) A set of Cucurbitaceae taxa for differentiation of
Pseudoperonespora cubensis pathotypes. J Plant Dis Prot 110: 337–349.
6. Runge F, Choi Y-J, Thines M (2011) Phylogenetic investigations in the genus
Pseudoperonospora reveal overlooked species and cryptic diversity in the P.
cubensis species cluster. Eur J Plant Pathol 129: 135-146.
7. Choi Y (2005) A re-consideration of Pseudoperonospora cubensis and P. humuli
based on molecular and morphological data. Mycological Res 109: 841-848.
8. Tian M, Win J, Savory E, Burkhardt A, Held M, et al. (2011) 454 Genome sequencing
of Pseudoperonospora cubensis reveals effector proteins with a QXLR
translocation motif. Mol Plant-Microbe Interact 24: 543-553.
9. Savory EA, Zou C, Adhikari BN, Hamilton JP, Buell CR, et al. (2012) Alternative
Splicing of a Multi-Drug Transporter from Pseudoperonospora cubensis
Generates an RXLR Effector Protein That Elicits a Rapid Cell Death. PLoS ONE
7: e34701.

133

10. Haas BJ, Kamoun S, Zody MC, Jiang RH, Handsaker RE, et al. (2009) Genome
sequence and analysis of the Irish potato famine pathogen Phytophthora
infestans. Nature 461: 393-398.
11. Kamoun S (2006) A catalogue of the effector secretome of plant pathogenic
oomycetes. Ann Rev Phytopathol 44: 41-60.
12. Cabral A, Stassen JH, Seidl MF, Bautor J, Parker JE, et al. (2011) Identification of
Hyaloperonospora arabidopsidis transcript sequences expressed during Infection
reveals isolate-specific effectors. PLoS ONE 6: e19328.
13. Sierra R, Rodriguez RL, Chaves D, Pinzon A, Grajales A, et al. (2010) Discovery of
Phytophthora infestans genes expressed in planta through mining of cDNA
libraries. PLoS ONE 5: e9847.
14. Torto-Alalibo TA, Tripathy S, Smith BM, Arredondo FD, Zhou L, et al. (2007)
Expressed sequence tags from Phytophthora sojae reveal genes specific to
development and infection. Mol Plant-Microbe Interact 20: 781-793.
15. Randall TA, Dwyer RA, Huitema E, Beyer K, Cvitanich C, et al. (2005) Large-scale
gene discovery in the oomycete Phytophthora infestans reveals likely
components of phytopathogenicity shared with true fungi. Mol Plant-Microbe
Interact 18: 229–243.
16. Bos JI, Kanneganti TD, Young C, Cakir C, Huitema E, et al. (2006) The C-terminal
half of Phytophthora infestans RXLR effector AVR3a is sufficient to trigger R3amediated hypersensitivity and suppress INF1-induced cell death in Nicotiana
benthamiana. Plant J 48: 165-176.
17. Fabro G, Steinbrenner J, Coates M, Ishaque N, Baxter L, et al. (2011) Multiple
Candidate Effectors from the Oomycete Pathogen Hyaloperonospora
arabidopsidis Suppress Host Plant Immunity. PLoS Pathog 7: e1002348.
18. Armstrong MR, Whisson SC, Pritchard L, Bos JI, Venter E, et al. (2005) An
ancestral oomycete locus contains late blight avirulence gene Avr3a, encoding a
protein that is recognized in the host cytoplasm. Proc Natl Acad Sci U S A 102:
7766-7771.
19. Allen RL (2004) Host-Parasite Coevolutionary Conflict Between Arabidopsis and
Downy Mildew. Science 306: 1957-1960
134

20. Money NP, Davis CM, Ravishankar JP (2004) Biomechanical evidence for
convergent evolution of the invasive growth process among fungi and oomycete
water molds. Fung Gen Biol 41: 872-876.
21. Dong S, Qutob D, Tedman-Jones J, Kuflu K, Wang Y, et al. (2009) The
Phytophthora sojae avirulence locus Avr3c encodes a multi-copy RXLR effector
with sequence polymorphisms among pathogen strains. PLoS ONE 4: e5556
22. Torto TA (2003) EST mining and functional expression fssays identify extracellular
effector proteins from the plant pathogen Phytophthora. Genome Res 13: 16751685.
23. Schornack S, van Damme M, Bozkurt TO, Cano LM, Smoker M, et al. (2010)
Ancient class of translocated oomycete effectors targets the host nucleus. Proc
Natl Acad Sci U S A 107: 17421-17426.
24. Damasceno CM, Bishop JG, Ripoll DR, Win J, Kamoun S, et al. (2008) Structure of
the glucanase inhibitor protein (GIP) family from Phytophthora species suggests
coevolution with plant endo-beta-1,3-glucanases. Mol Plant-Microbe Interact 21:
820-830.
25. Rose JK, Ham KS, Darvill AG, Albersheim P (2002) Molecular cloning and
characterization of glucanase inhibitor proteins: coevolution of a counterdefense
mechanism by plant pathogens. Plant Cell 14: 1329-1345.
26. Tian M, Benedetti B, Kamoun S (2005) A Second Kazal-like protease inhibitor from
Phytophthora infestans inhibits and interacts with the apoplastic pathogenesisrelated protease P69B of tomato. Plant Physiol 138: 1785-1793.
27. Tian M, Huitema E, Da Cunha L, Torto-Alalibo T, Kamoun S (2004) A Kazal-like
extracellular serine protease inhibitor from Phytophthora infestans targets the
tomato pathogenesis-related protease P69B. J Biol Chem 279: 26370-26377.
28. Tian M, Win J, Song J, van der Hoorn R, van der Knaap E, et al. (2007) A
Phytophthora infestans cystatin-like protein targets a novel tomato papain-like
apoplastic protease. Plant Physiol 143: 364-377.
29. Liu Z, Bos JI, Armstrong M, Whisson SC, da Cunha L, et al. (2005) Patterns of
diversifying selection in the phytotoxin-like scr74 gene family of Phytophthora
infestans. Mol Biol Evol 22: 659-672.
135

30. Levesque CA, Brouwer H, Cano L, Hamilton JP, Holt C, et al. (2010) Genome
sequence of the necrotrophic plant pathogen Pythium ultimum reveals original
pathogenicity mechanisms and effector repertoire. Genome Biol 11: R73.
31. Fellbrich G, Romanski A, Varet A, Blume B, Brunner F, et al. (2002) NPP1, a
Phytophthora-associated trigger of plant defense in parsley and Arabidopsis.
Plant J 32: 375-390.
32. Cabral A, Oome S, Sander N, Kuefner I, Nürnberger T, et al. (2012) Non-toxic
Nep1-like proteins of the downy mildew pathogen Hyaloperonospora
arabidopsidis; repression of necrosis-inducing activity by a surface-exposed
region. Mol Plant-Microbe Interact.
33. Gaulin E, Drame N, Lafitte C, Torto-Alalibo T, Martinez Y, et al. (2006) Cellulose
binding domains of a Phytophthora cell wall protein are novel pathogenassociated molecular patterns. Plant Cell 18: 1766-1777.
34. Gaulin E, Jauneau A, Villalba F, Rickauer M, Esquerré-Tugayé MT, et al. (2002)
The CBEL glycoprotein of Phytophthora parasitica var-nicotianae is involved in
cell wall deposition and adhesion to cellulosic substrates. J Cell Sci 115: 4565.
35. Tyler BM (2006) Phytophthora genome sequences uncover evolutionary origins and
mechanisms of pathogenesis. Science 313: 1261-1266
36. Baxter L, Tripathy S, Ishaque N, Boot N, Cabral A, et al. (2010) Signatures of
adaptation to obligate biotrophy in the Hyaloperonospora arabidopsidis genome.
Science 330: 1549-1551.
37. Moy P, Qutob D, Chapman BP, Atkinson I, Gijzen M (2004) Patterns of gene
expression upon infection of soybean plants by Phytophthora sojae. Mol PlantMicrobe Interact 17: 1051-1062.
38. Costanzo S, Ospina-Giraldo MD, Deahl KL, Baker CJ, Jones RW (2006) Gene
duplication event in family 12 glycosyl hydrolase from Phytophthora spp. Fung
Gen Biol 43: 707-714.
39. Yamaguchi Y, Huffaker A (2011) Endogenous peptide elicitors in higher plants. Curr
Opinion Plant Biol 14: 351-357.

136

40. Fernandez D, Tisserant E, Talhinhas P, Azinheira H, Vieira ANA, et al. (2011) 454pyrosequencing of Coffea arabica leaves infected by the rust fungus Hemileia
vastatrix reveals in planta-expressed pathogen-secreted proteins and plant
functions in a late compatible plant–rust interaction. Mol Plant Pathol 13: 17-37.
41. Joly DL, Feau N, Tanguay P, Hamelin RC (2010) Comparative analysis of secreted
protein evolution using expressed sequence tags from four poplar leaf rusts
(Melampsora spp.). BMC Gen 11: 422.
42. Miranda M, Ralph SG, Mellway R, White R, Heath MC, et al. (2007) The
transcriptional response of hybrid poplar (Populus trichocarpa x P. deltoides) to
infection by Melampsora medusae leaf rust involves induction of flavonoid
pathway genes leading to the accumulation of proanthocyanidins. Mol PlantMicrobe Interact 20: 816-931.
43. Duplessis S, Hacquard S, Delaruelle C, Tisserant E, Frey P, et al. (2011)
Melampsora larici-populina tanscript profiling during germination and timecourse
infection of poplar leaves reveals dynamic expression patterns associated with
virulence and biotrophy. Mol Plant-Microbe Interact 24: 808-818.
44. Mosquera G, Giraldo M, Khang CH, Coughlan S, Valent B (2009) Interaction
transcriptome analysis identifies Magnaporthe oryzae BAS1-4 as biotrophyassociated secreted proteins in rice blast disease. Plant Cell 21: 1273-1290.
45. Polesani M, Desario F, Ferrarini A, Zamboni A, Pezzotti M, et al. (2008) cDNA-AFLP
analysis of plant and pathogen genes expressed in grapevine infected with
Plasmopara viticola. BMC Gen 9: 142 - 156.
46. Adhikari BN, Savory EA, Vaillancourt B, Childs KL, Hamilton JP, et al. (2012)
Expression Profiling of Cucumis sativus in Response to Infection by
Pseudoperonospora cubensis. PLoS ONE 7: e34954.
47. Ospina-Giraldo M, Griffith J, Laird E, Mingora C (2010) The CAZyome of
Phytophthora spp.: A comprehensive analysis of the gene complement coding for
carbohydrate-active enzymes in species of the genus Phytophthora. BMC Gen
11: 1-16.
48. Ospina-Giraldo M, McWalters J, Seyer L (2010) Structural and functional profile of
the carbohydrate esterase gene complement in Phytophthora infestans. Curr
Gen 56: 495-506.
137

49. Ihmels J, Bergmann S, Berman J, Barkai N (2005) Comparative gene expression
analysis by a differential clustering approach: Application to the Candida albicans
transcription program. PLoS Genet 1: e39.
50. Langfelder P, Zhang B, Horvath S (2008) Defining clusters from a hierarchical
cluster tree: the Dynamic Tree Cut package for R. Bioinformatics 24: 719 - 720.
51. Langfelder P, Horvath S (2007) Eigengene networks for studying the relationships
between co-expression modules. BMC Sys Biol 1: 54.
52. Iyer LM, Anantharaman V, Wolf MY, Aravind L (2008) Comparative genomics of
transcription factors and chromatin proteins in parasitic protists and other
eukaryotes. In J Parasitol 38: 1-31.
53. Wang Y, Dou D, Wang X, Li A, Sheng Y, et al. (2009) The PsCZF1 gene encoding a
C2H2 zinc finger protein is required for growth, development and pathogenesis in
Phytophthora sojae. Micro Path 47: 78-86.
54. Blanco FA, Judelson HS (2005) A bZIP transcription factor from Phytophthora
interacts with a protein kinase and is required for zoospore motility and plant
infection. Mol Micro 56: 638-648.
55. Judelson H, Ah-Fong A (2010) The kinome of Phytophthora infestans reveals
oomycete-specific innovations and links to other taxonomic groups. BMC Gen
11: 700.
56. Langmead B, Trapnell C, Pop M, Salzberg SL (2009) Ultrafast and memory-efficient
alignment of short DNA sequences to the human genome. Genome Biol 10: R25.
57. Trapnell C, Pachter L, Salzberg SL (2009) TopHat: discovering splice junctions with
RNA-Seq. Bioinformatics 25: 1105-1111.
58. Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, et al. (2010) Transcript
assembly and quantification by RNA-Seq reveals unannotated transcripts and
isoform switching during cell differentiation. Nat Biotechnol 28: 511-515.
59. Saeed AI, Bhagabati NK, Braisted JC, Liang W, Sharov V, et al. (2006) TM4
microarray software suite. Methods Enz 411: 134-193.
138

60. Bullard J, Purdom E, Hansen K, Dudoit S (2010) Evaluation of statistical methods
for normalization and differential expression in mRNA-Seq experiments. BMC
Bioinformatics 11: 94.
61. Edgar R, Domrachev M, Lash AE (2002) Gene Expression Omnibus: NCBI gene
expression and hybridization array data repository. Nucleic Acids Res 30: 207210.
62. Irizarry R, Hobbs B, Collin F, Beazer-Barclay Y, Antonellis K, et al. (2003)
Exploration, normalization, and summaries of high density oligonucleotide array
probe level data. Biostatistics 4: 249-264.
63. Li L, Stoeckert CJ, Jr., Roos DS (2003) OrthoMCL: identification of ortholog groups
for eukaryotic genomes. Genome Res 13: 2178-2189.
64. Chen F, Mackey AJ, Vermunt JK, Roos DS (2007) Assessing performance of
orthology detection strategies applied to eukaryotic genomes. PLoS ONE 2:
e383.
65. Suzek BE, Huang H, McGarvey P, Mazumder R, Wu CH (2007) UniRef:
comprehensive and non-redundant UniProt reference clusters. Bioinformatics 23:
1282-1288.
66. Bateman A, Birney E, Durbin R, Eddy SR, Howe KL, et al. (2000) The Pfam protein
families database. Nucleic Acids Res 28: 263-266.
67. Eddy SR (2009) A new generation of homology search tools based on probabilistic
inference. Genome Inform 23: 205-211.
68. Childs KL, Davidson RM, Buell CR (2011) Gene coexpression network analysis as a
source of functional annotation for rice genes. PLoS ONE 6: e22196.

139

CHAPTER 4

Expression profiling of Cucumis sativus in response to infection
by Pseudoperonospora cubensis

This chapter was originally published in PLoS ONE.
Adhikari, BN*, Savory EA*, Vaillancourt, B, Childs KL, Hamilton JP, Day B and Buell CR
(2012) Expression profiling of Cucumis sativus in response to infection by
Pseudoperonospora
cubensis.
PLoS
ONE
7(4):
e34954.
doi:10.1371/journal.pone.0034954
© 2012 Adhikari et al. This is an open-access article distributed under the terms of the
Creative Commons Attribution License, which permits unrestricted use, distribution, and
reproduction in any medium, provided the original author and source are credited.
*

These authors contributed equally to this work.

Author Contributions:
Conceived and designed the experiments: EAS, BNA, KLC, BD and CRB. Performed
the experiments: BNA, EAS, KLC, and JPH. Analyzed the data: BNA, EAS, KLC, and
JPH. Contributed reagents/materials/analysis tools: BNA, EAS, BV, KLC, JPH, BD and
CRB. Wrote the paper: BNA, EAS, BD and CRB.

140

ABSTRACT
The oomycete pathogen, Pseudoperonospora cubensis, is the causal agent of downy
mildew on cucurbits, and at present, no effective resistance to this pathogen is available
in cultivated cucumber (Cucumis sativus). To better understand the host response to a
virulent pathogen, we performed expression profiling throughout a time course of a
compatible interaction using whole transcriptome sequencing. As described herein, we
were able to detect the expression of 15,286 cucumber genes, of which 14,476 were
expressed throughout the infection process from 1 day post-inoculation (1 dpi) to 8 dpi.
A large number of genes, 1,612 to 3,286, were differentially expressed in pair-wise
comparisons between time points. We observed the rapid induction of key defense
related genes, including catalases, chitinases, lipoxygenases, peroxidases, and
protease inhibitors within 1 dpi, suggesting detection of the pathogen by the host. Coexpression network analyses revealed transcriptional networks with distinct patterns of
expression including down-regulation at 2 dpi of known defense response genes
suggesting coordinated suppression of host responses by the pathogen. Comparative
analyses of cucumber gene expression patterns with that of orthologous Arabidopsis
thaliana genes following challenge with Hyaloperonospora arabidopsidis revealed
correlated expression patterns of single copy orthologs suggesting that these two dicot
hosts have similar transcriptional responses in response to related pathogens. In total,
the work described herein presents an in-depth analysis of the interplay between host
susceptibility and pathogen virulence in an agriculturally important pathosystem.

141

INTRODUCTION

Cucumber (Cucumis sativus L.) is an economically important vegetable crop cultivated
in over 80 countries, with more than 66 million tons produced annually for both fresh
use and processing (http://faostats.fao.org). Cucumber has been utilized extensively as
a model system to study sex determination [1], vascular biology [2], and induced
defense responses [3,4]. Despite its extensive use as a model system in research, as
well as its obvious economic importance, genetic and genomic resources remain
limited. In recent years, the publication of both a genetic map [5] and genome
sequences [6,7] of cucumber, as well as generation of large-scale expression data sets
[8,9], have provided the first comprehensive resources for genetic and genomic based
inquiries into cucumber biology. Cucumber has limited genetic diversity, few wild
relatives, and only 7 pairs of chromosomes (2n = 14), whereas other Cucumis spp.,
such as melon (Cucumis melo), have 12 chromosomes, making interspecific breeding
difficult, if not impossible. As such, advances in breeding for important agronomic traits
such as increased yield, fruit quality, and disease resistance are slow.

Cucumber production is hindered by diseases caused by bacterial (e.g., Pseudomonas
syringae pv. lachrymans), viral (e.g., Cucumber mosaic virus), fungal (e.g.,
Sphaerotheca fulginea and Erysiphe cichoracearum), and oomycete (e.g., Phytophthora
capsici and Pseudoperonospora cubensis) pathogens [10,11]. Of these, the most
destructive is Ps. cubensis, the causal agent of cucurbit downy mildew, which threatens
cucumber production in nearly 80 countries and causes severe economic losses

142

[11,12,13]. Ps. cubensis is an obligate, biotrophic oomycete pathogen with a host range
limited to the Cucurbitaceae [11]. In recent years, a foundation has been established to
support advances in this area, including studies on epidemiology [14], host specificity
[15,16,17,18], pathogenic variation [19,20,21], and more recently, generation of a draft
genome sequence of Ps. cubensis [22,23].

The molecular and biochemical mechanisms associated with host resistance have been
investigated to a limited extent in cucumber and other cucurbits. In large part, signaling
of resistance is primarily associated with systemic acquired resistance (SAR) [24], for
which cucumber has historically been a model system [25,26]. In addition to SAR,
structural modifications (e.g., callose deposition), as well as the induction of defenserelated genes (e.g., peroxidases chitinases, and glucanases) are often associated with
the onset of resistance signaling following pathogen infection. Moreover, like other wellcharacterized plant-pathogen systems, the presence of nucleotide-binding site (NBS)containing genes encoding protein products that recognize cognate pathogen effector
proteins or their perturbations [27] have been postulated to play a role in disease
resistance in cucumber. To this end, analysis of the cucumber genome sequence has
identified 61 NBS-Resistance genes, considerably less than have been identified in
other plant genomes, such as Arabidopsis (200) or rice (600) [6]. However, of the 15
genes known to control disease resistance in cucumber, none have been cloned, nor
have they been associated through linkage maps with the 61 predicted NBS-Resistance
genes [28]. It is hypothesized that cucumber has an expanded lipoxygenase (LOX)

143

pathway which may provide an additional mechanism(s) to facilitate responses to biotic
stress [6].

While genetically conferred host resistance is the ideal means of disease control in crop
species, it has become ineffective in controlling cucurbit downy mildew, particularly in
the U.S., where introduction of a new, more virulent pathotype of Ps. cubensis is
responsible for economic losses in recent years [12,29]. To this end, control methods
for cucurbit downy mildew in both Europe and the U.S. require the use of fungicides,
coupled with a single host resistance locus, the recessive dm1 gene, which has been
incorporated into most commercial cucumber germplasm [11]. However, the
identification of the dm1 locus, as well as its functional role in resistance signaling
remains undefined. In addition to widespread incorporation of dm1, breeding of Ps.
cubensis resistance has focused mainly on genes from melon [30], as limited diversity
for resistance is available in cucumber or its wild relative, Cucumis hardwickii. Largescale screening trials to identify tolerant germplasm are in progress, but have yet to
identify a source of complete resistance to Ps. cubensis [16,31]. As such, new
resources must be explored to support development of improved cultivars and identify
new sources of resistance for breeding programs.

Next generation sequencing of the transcriptome (mRNA-Seq) permits deep, robust
assessments of transcript abundances and transcript structure [32]. When gene
expression profiling is applied to host-pathogen interactions of economically important
crops, insights into the mechanisms these pathogens use to suppress and subvert the

144

host defense response can be made [33,34,35]. In the current study, we performed
expression profiling, using the susceptible cucumber cultivar 'Vlaspik', over a time
course of infection with the downy mildew pathogen Ps. cubensis to identify genes,
pathways, and systems that are altered during a compatible interaction. Using this
technology, deep profiling of both the host and pathogen transcriptome (see
accompanying manuscript, [36]) was achieved, providing the first in-depth analysis of
this important plant-pathogen interaction. In this study, we cataloged the expression of
14,476 genes from cucumber through an 8-day time course of infection with a virulent
Ps. cubensis isolate. In total, this work identified major changes in gene expression in
cucumber at 1 days post-inoculation (dpi) continuing through 8 dpi, with up to 3,286
genes differentially expressed between time points. Comparative analyses revealed
correlated gene expression patterns in cucumber and Arabidopsis thaliana leaves
infected with downy mildew suggesting orthologous host responses in these two
dicotyledonous hosts. Through co-expression network analyses, modules of temporalspecific transcriptional networks were constructed that provide a framework to connect
transcription factors with defense response genes.

RESULTS AND DISCUSSION

Response of C. sativus leaves to pathogen infection

To correlate gene expression and host responses with observable disease symptoms
and pathogen invasion, the progression of infection in susceptible C. sativus cv.

145

'Vlaspik' was monitored at 1, 2, 3, 4, 6, and 8 dpi. As shown in Figure 4.1, the first
visible symptoms of Ps. cubensis infection were apparent at 1 dpi, in the form of water
soaking on the abaxial leaf surface at the inoculation site (Figure 4.1A). These
symptoms correspond to zoospore encystment and initial penetration through the
stomata into the host (see accompanying paper, [36]). In similar pathosystems systems
such as Hyaloperonospora arabidopsidis, the causal agent of downy mildew on A.
thaliana, analogous processes occur in the early stages of infection, with the exception
that H. arabidopsidis penetrates between anticlinal walls of epidermal cells rather than
utilizing natural openings like stomata [37]. While no symptoms are visible on the upper
leaf surface in cucumber, water soaking on the lower leaf surface can be seen as early
as 1 dpi, and remains present through 4 dpi, during which time hyphal growth through
the mesophyll of the host tissue occurs and haustoria formation begins (see
accompanying paper, [36]). Yellow angular lesions bound by leaf veins characteristic of
cucurbit downy mildew were first visible on the upper leaf surface at 4 dpi (Figure 4.1D),
and became more chlorotic and slightly necrotic at the centers by 8 dpi. These
symptoms are associated with extensive growth of hyphae through the plant mesophyll
(see accompanying paper, [36]).

Gene expression profiling in C. sativus

We performed mRNA-Seq profiling of C. sativus leaves following infection with Ps.

146

Figure 4.1

Figure 4.1 Symptoms of Pseudoperonospora cubensis infection on susceptible
Cucumis sativus cv. ‘Vlaspik’. Images were collected of the adaxial (left column) and
abaxial (right column) leaf surfaces at 1, 2, 3, 4, 6, and 8 days post-inoculation (dpi).
(A) 1 dpi, (B) 2 dpi, (C) 3 dpi, (D) 4 dpi, (E) 6 dpi, (F) 8 dpi.
cubensis over an 8-day period. Leaf disc samples were collected using a cork borer to

147

maximize the amount of infected tissue in each sample (Figure 4.2, black circles),
pooled within a given time point, and RNA was isolated. Two biological replicates of
each time point were sequenced, yielding 55 to 59 million reads from both replicates at
each time point. Additionally, a mock-inoculated C. sativus sample was collected and
sequenced, yielding 5.8 million reads. The number of reads that mapped to the C.
sativus genome ranged from 48 to 53 million (Figure 4.3A, 84-93% of the total reads)
per time point while the number of genes expressed at different time points ranged from
12,257 to 13,048 (Figure 4.3A).

The libraries were constructed from inoculated leaves and therefore, the reads
represent transcripts from the host (C. sativus) and the pathogen (Ps. cubensis). At the
early time points, nearly all of the reads obtained were of host origin, which is consistent
with our microscopy analysis revealing limited pathogen biomass. However, as we are
surveying a susceptible interaction, pathogen biomass increases throughout the time
course and consequentially, pathogen transcripts increase in the total read pool in the
later time points (see accompanying paper, [36]). However, even with the increased
percentage of pathogen reads in the later time points, we have saturated sampling of
the C. sativus transcriptome with our deep read coverage of the libraries. Randomly
selected subsets of reads, 5 to 30 million, from the total read pool were used to evaluate
the effect of sampling depth on gene expression detection. The simulation
demonstrates a positive relationship between sampling depth and numbers of
expressed genes at lower sequencing depths (5 to 20 million reads)

148

Figure 4.2

149

Figure 4.2 (cont'd) Experimental design and sample collection. Cucumber cv.
‘Vlaspik’ leaves were inoculated on the abaxial leaf surface with 10-30 10 l droplets
of a 1 x 105 sporangia/ml solution. Samples were collected at 1, 2, 3, 4, 6, and 8 days
post-inoculation (dpi) using a #3 cork borer to isolate tissue immediately around the
infection point (black circles). Samples from cucumber leaves mock-inoculated with
10 l droplets of dH2O were collected in the same manner at 1 dpi. Leaf disks from
each time point were pooled for RNA extraction. mRNA-Seq libraries were made for
each time point from two separate biological replicates. Within a biological replicate,
libraries were barcoded and sequenced in multiple lanes with the exception of the
mock-inoculated library, which was not barcoded.
(Figure 4.3B). The number of expressed genes, however, begins to plateau at
approximately 20 million reads, corresponding to the minimum sampling depth of all
libraries in this study. To study the repeatability of two biological replicates, pair-wise
scatter plots of gene expression values were generated. For biological replicates of
each time point, nearly all genes fell along the diagonal of plots, indicating no major
variation with correlation coefficients (R2) ranging from 0.97 to 0.98, thus providing
evidence for high reproducibility of biological replicates (Figure S4.1).

Host transcriptional changes in response to infection

Over the infection period, a total of 14,476 C. sativus genes were expressed (Table
S4.1, available at www.plosone.org, e34954), with 10,350 genes common to all time
points. For all data points analyzed, the minimum fragments per kilobase exon model
per million mapped reads (FPKM) value was zero, yet the maximum FPKM values
ranged from 8,869 at 1 dpi to 34,017 at 8 dpi (Table S4.1, available at
www.plosone.org, e34954). Interestingly, the highest up-regulated gene at 1 dpi was a
putative

galactinol

synthase

(Csa6M000080.1),

150

which

has

Figure 4.3

151

Figure 4.3 (cont'd)

Figure 4.3 Comparison of total mRNA-Seq reads, reads mapped and number of genes expressed. (A) Total
number of reads, total reads mapped, and number of genes expressed as determined from pooling of both biological
replicates are shown. Reads were mapped to the C. sativus genome [6] using Bowtie version 0.12.5 [55] and TopHat
version 1.2.0 [53]. Fragments per kilobase pair of exon model per million fragments mapped (FPKM) values for the
annotated C. sativus genes were calculated using Cufflinks version 0.9.3 [55]. Genes were considered expressed if the
FPKM value and 95% confidence interval lower boundary FPKM value was greater than 0.001 and zero, respectively.
(B) Effect of sampling depth on detection of expressed genes. For all time points 5, 10, 15, 20, 25, and 30 million reads
were randomly selected from the total pool of reads. Read mapping and expression abundance estimates were
performed as describe above. Solid lines indicate number of genes expressed and the dashed lines indicate number of
reads mapped at different time points. dpi, days post-inoculation.
152

been shown to be up-regulated in Cucumis melo (melon) in response to abiotic stress
[38] as well as in an inbred C. sativus line 'IL57' with a high level of resistance to downy
mildew [39]. The expression patterns of the top 20 highly expressed genes showed
expression of genes involved in defense responses including catalases, chitinases,
lipoxygenases, peroxidases, and protease inhibitors, beginning at 1 dpi and extending
through 8 dpi (Table S4.2, available at www.plosone.org, e34954). The detection of
defense-related genes within 1 dpi suggests that there is an active response by C.
sativus to early infection stages of Ps. cubensis, including zoospore encystment,
appressorium formation, and penetration via stomata (see accompanying paper, [36]).
Additionally, no pathogen defense-related genes are present within the top 20 highly
expressed genes in the mock-inoculated samples, which mainly consists of
housekeeping genes (Table S4.2, available at www.plosone.org, e34954).

Correlation and cluster analyses were used to identify similarities in transcriptome
profiles among the sampled time points. Pearson Correlation Coefficient (PCC) values
for time point comparisons ranged from 0.78 to 0.93, with tight clustering readily
apparent, revealing patterns that highlight the extent of transcriptional diversity
underlying early (1 dpi), intermediate (2, 3, and 4 dpi), and advanced (6 and 8 dpi)
stages of disease progression and the corresponding responses in host gene
expression (Figure 4.1). As described above, C. sativus defense-related genes are
expressed within 1 dpi of Ps. cubensis inoculation, and based on our correlation
analysis (Figure 4.4), these likely represent a distinct cluster of genes responding
specifically to initial recognition of sporangia, germination of zoospores, and zoospore

153

Figure 4.4

Figure 4.4 Correlation matrix of Cucumis sativus expression profiles during
infection by Pseudoperonospora cubensis. Tissue samples were collected from
C. sativus at different time points of infection with Ps. cubensis. Normalized transcript
abundances for 14,476 genes were calculated as fragments per kilobase pair of exon
model per million fragments mapped (FPKM) with Cufflinks version 0.9.3 [55].
Pearson correlation coefficient of log2 FPKM values were calculated for all pair-wise
comparisons using R. Hierarchical clustering was performed using Pearson
correlation distance metric and average linkage with the Multiple Experiment Viewer
Software version 4.5 [57]. The bootstrap support values shown on tree nodes were
obtained from 1000 bootstrap replicates. dpi, days post-inoculation.

154

encystment in stomata. Genes expressed at 2-4 dpi also cluster more with each other
than to 1 dpi or to later time points, which also reflects the similar stages of Ps. cubensis
infection apparent at days 2-4 dpi, indicating that these genes may be involved in host
responses to hyphal penetration, growth and haustorium formation. The clustering of
gene expression at later time points (6 and 8 dpi) likely corresponds to similar
symptoms observed (Figure 4.1) as well as plant responses to extensive Ps. cubensis
hyphal growth that is occurring at those time points (see accompanying paper, [36]).

Conserved host responses in C. sativus and A. thaliana in compatible
interactions as measured through expression profiling

Host responses to pathogen challenge have been well documented in the model
species A. thaliana [40], including those to the downy mildew pathogen H. arabidopsidis
[34,37]. To identify genes induced in a compatible interaction with a downy mildew
pathogen common to both C. sativus and A. thaliana, we identified single copy
orthologous genes in both plant genomes and analyzed their expression patterns. A
total of 7,396 clusters of single copy genes from both species were identified by
clustering 23,248 and 27,416 representative protein coding genes from C. sativus and
A. thaliana, respectively. Data from a previous microarray-based expression profiling
[41] experiment of a compatible A. thaliana-H. arabidopsidis interaction was compared
with our mRNA-Seq-based expression data. The H. arabidopsidis infection time points
were 0, 0.5, 2, 4, and 6 dpi, similar to the 1-8 dpi time points assayed in this study.
Spearman rank correlation coefficients (SCCs) of log2 expression values were
calculated between the single copy orthologs at all time points in the two datasets;
155

between 2,136 and 3,446 gene pairs were included in the pair-wise comparisons.
Among the six comparisons between similar time points, the SCC values ranged from
0.10 to 0.72 (Table S4.3, available at www.plosone.org, e34954). The correlations
between early time points were the lowest between the two interactions, possibly due to
the differences in penetration strategies between the two pathogens. The highest
correlations were observed between 6 dpi (SCC = 0.72) followed by 2 dpi (SCC = 0.66)
for both host-pathogen interactions (Figure 4.5). Overall, correlation coefficients
between analogous time points (0.65 to 0.72) were greater than comparisons between
dissimilar time points (0.10 to 0.33) indicating similar expression patterns for single copy
orthologous genes in C. sativus and A. thaliana.

Differential gene expression throughout the infection process

Differences in gene expression patterns between time points can provide insight into the
host response following pathogen perception and subsequent infection; thus, the
mechanism(s) through which pathogenicity occurs may be inferred [34,42,43,44]. For
example, a recent publication by Gaulin et al. [45] used a comparative approach
analyzing the transcriptomes of Aphanomyces euteiches and Phytophthora spp. to
identify novel pathogenicity factors and expanded repertoires for virulence. Through
comparison of gene expression patterns across time points, we identified genes
differentially expressed between all time points and the control sample, 1 dpi mock
inoculation; between 1,170 and 3,286 genes were differentially expressed in pair-wise

156

Figure 4.5

157

Figure 4.5 (cont'd) Comparison of orthologous gene expression in Cucumis
sativus and Arabidopsis thaliana in a compatible interaction. Microarray
expression profiles were obtained from time-course analyses of genes expressed in
A. thaliana (Arth) during infection by Hyaloperonospora arabidopsidis [39]. Single
copy orthologous genes between C. sativus (Cusa) and A. thaliana were identified
using OrthoMCL [60]. Log2 transformed expression values of single copy orthologous
genes expressed in the C. sativus mRNA-Seq dataset (log2 fragments per kilobase
pair of exon model per million fragments mapped [FPKM]) and A. thaliana microarray
dataset (log2 intensity) are shown as scatter plots. SCC, Spearman correlation
coefficient; dpi, days post-inoculation.
comparison between the mock inoculated and/or the inoculated time points (Table 4.1,
Table S4.4, available at www.plosone.org, e34954). In general, 12 to 31% of the genes
tested under different conditions were differentially expressed in the pair-wise
comparisons. In infected samples, comparisons of the three early to intermediate time
points (i.e., 1, 2, and 3 dpi) showed a higher number of differentially expressed genes
(2,214 to 3,286) than those between three intermediate to later (4, 6 and, 8 dpi) time
points (1,612 to 2,074), suggesting more correlated gene expression in later stages of
infection.

Gene co-expression pattern analyses

Using Weighted Gene Correlation Network Analysis (WGCNA) [46], we identified sets of
highly correlated genes and constructed modules where all members are more highly
correlated with each other than they are to genes outside the module [47]. Out of
15,286 genes expressed in the mock control or throughout the infection time course,
4,410 genes passed the Coefficient of Variance (CV) filter (0.4) and were retained for
downstream analyses. Of the 4,410 genes, a total of 2,169 were assigned to 11 gene

158

Table 4.1 Number of genes differentially expressed between different time
points.
MockControl
1 dpi
2 dpi
3 dpi

1 dpi¥
1,170
(12%)§

2 dpi
1,906
(19%)

3 dpi
2,596
(26%)

4 dpi
2,509
(25%)

6 dpi
1,713
(17%)

8 dpi
2,556
(26%)

3,286
(31%)

2,948
(30%)
2,214
(23%)

2,849
(27%)
1,736
(18%)
1,590
(16%)

2,010
(20%)
2,718
(27%)
1,840
(18%)
1,842
(18%)

3,014
(30%)
3,006
(29%)
2,218
(23%)
2,074
(21%)
1,612
(16%)

4 dpi
6 dpi

Differential expression analysis was conducted using the cuffdiff program in Cufflinks
version 0.9.3 [52], Cucumis sativus v2 annotation, and false discovery rate of 0.01.
¥
dpi, days post-inoculation.
§
Numbers in parenthesis indicate the percent of significantly different tests out of the
total number of tests that could be performed for each pair-wise comparison.

modules that contained between 50 and 428 genes (Table S4.5); 2,241 genes were not
assigned to any module. To visualize the relationship of the modules to each other with
respect to the progression of infection, eigengenes for each module were calculated
and displayed in a heat map [48]. As shown in Figure 4.6, some modules are
representative of genes with correlated co-expression at primarily a single or two time
points (Modules F, G, H, I, J, and K) whereas other modules represent genes that share
broader co-expression patterns across multiple time points (Modules A, B, C, D, E).
Examination of trend plots for the modules (Figure 4.7, Figure S4.2) reveals the
direction and magnitude of gene expression patterns. Genes within Module B are
expressed in the mock control and at 1 dpi, but are coordinately down-regulated at 2

159

dpi, remaining less abundant through 8 dpi (Figure 4.7). This set of 272 genes includes
a large suite of genes implicated in resistance including six lipoxygenase genes, four
cationic

peroxidases,

two

cinnamate

4-hydroxylases,

an

anthocyanidin

3-

glucosyltransferase, an anthocyanin 5-aromatic acyltransferase, a cysteine protease,
and the elicitor-inducible protein EIG-J7. All of these genes have been implicated in
defense responses in other plants [49,50,51,52,53,54], and in total, their coordinate
down-regulation at 2 dpi is suggestive of an active mechanism by the pathogen to alter
host defense responses. For example, the lipoxygenase pathway has been
hypothesized to be involved in defense responses in cucumber, due to its expansion
within the genome [6], and our data presented herein showing a rapid down-regulation
of these transcripts during infection supports this hypothesis. In addition, cinnamate 4hydroxylase, one of the primary enzymes in the phenylpropanoid pathway responsible
for the conversion of cinnamic acid to p-coumaric acid, was shown to be rapidly induced
in C. sativus in response to abiotic stress [51]. In this context, it is reasonable to

hypothesize that the down-regulation of its expression (observed in the present study),
as well as that of other genes within the phenylpropanoid pathway, are suggestive of an
active virulence mechanism to abrogate host responses associated with stress-induced
signaling. In addition to the direct correlation among defense gene expression noted
above, this module also includes the coordinated expression of 32 transcription factors
that may also have critical roles in regulating genes responsible for the induction of
defense signaling in the host. A total of 36 genes of no known function are also in this
module, and further examination of their roles in defense responses may provide new

160

Figure 4.6

Figure 4.6 Heat map of eigengenes representing each gene module. The mock
control and post-infection time points are represented in columns and the eigengenes
for each of the 11 identified co-expression modules [46] are presented in rows. The
numbers of genes in each module are given in parentheses. The eigengene values,
which range from 0 to 1, are a measure of centrality and indicate the relative
expression levels of all genes in the module (see Materials and Methods). dpi, days
post-inoculation.
insight into critical host genes modulated by virulent pathogens.

Modules F, G, H, I, J, and K all have discrete time points where genes are up-regulated
compared to the other sampled time points. As shown in Figure S4.2, genes in Module
F (162 genes total) have a peak of expression solely at 2 dpi. This module contains 20
transcription factors which could be key to regulating genes from Module G that have
peak expression at 3 dpi (Figure 4.7). Likewise, Module G (244 genes total) has 21

161

transcription factors that may regulate genes within Module H (89 genes total) that are
coordinately up-regulated at 4 dpi (Figure 4.7). Within all of these modules (F, G, H, I,
J, and K, 726 genes total) there are 199 genes with no known function and placement of
these genes in transcriptional networks provides a new functional annotation and
contextual information on their function.

CONCLUSIONS

The work described herein represents the first genome-scale analysis of the cucumberdowny mildew interaction in which we catalog gene expression throughout an 8 day
infection period and identify differentially expressed genes that could be correlated with
pathogen growth and development in planta. With expression profiles for nearly 15,000
genes during a compatible interaction, we have new insight into molecular events at the
host-pathogen interface including a suite of defense response-related genes that are
down-regulated early upon infection and transcriptional networks that respond in a
temporal manner throughout the infection cycle. Most intriguingly, these networks
include transcription factors and genes of no known function, which may have a role in
the host-pathogen interaction.

162

Figure 4.7

163

Figure 4.7 (cont'd) Trend plots of the normalized gene expression values from
six identified gene co-expression modules. Modules consisting of genes downregulated at 2 dpi (Module B), genes up-regulated at a single time point (Modules G,
H, I, K) and genes up-regulated at two time-points (Module J) are shown. The
number of genes in each module is shown in parentheses.

MATERIALS AND METHODS

Plant materials, inoculum, and pathogen infection

C. sativus cv. 'Vlaspik' was grown in growth chambers maintained at 22 °C with a 12 h
light/dark photoperiod. Ps. cubensis MSU-1 was maintained as previously described
[22]. The first fully expanded leaf of four-week-old cucumber plants was inoculated on
the abaxial surface with 20-30 10 μl droplets of a 1 x 105 sporangia/ml solution. After
inoculation, plants were kept at 100% relative humidity for 24 hours in the dark. Plants
were returned to growth chambers for disease progression.

Sampling and experimental design

Samples from two biological replicates were collected at 1, 2, 3, 4, 6, and 8 dpi at the
site of inoculation using a #3 cork borer. Additionally, a mock-inoculated sample (i.e.,10
μl dH2O), which was inoculated as described above and kept at 100% relative humidity
for 24 hours in the dark was collected at 1 dpi. Samples were frozen immediately in
liquid nitrogen and stored at -80 C until use.

164

Library preparation and sequencing

Total RNA was isolated from infected leaf discs using the RNeasy Mini Kit (Qiagen,
Germantown, MD) and treated with DNase (Promega, Madison, WI) per the
manufacturer's instructions. RNA concentration and quality was determined using the
Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA). The mRNA-Seq sample
preparation was done using the Illumina mRNA-Seq kit (Illumina, San Diego CA)
according to the manufacturer’s protocol. Parallel sequencing was performed using an
Illumina Genome Analyzer II (Illumina, Inc., San Diego, CA) at the Research
Technology Support Facility (RTSF) at Michigan State University. Each library within a
biological experiment was barcoded, six different barcodes for 6 time-points, pooled and
run on multiple lanes. Two biological replicates of each time-point were sequenced
multiple times and single-end reads between 35 and 42 bp were generated. Reads from
both biological replicates were pooled for determining expression abundances. The
mock-inoculated sample library was made as described above, but not barcoded and
run in a single lane.

Processing of mRNA-Seq data

mRNA-Seq reads obtained from Illumina Pipeline version 1.3 were quality evaluated on
the Illumina purity filter, percent low quality reads, and distribution of phred-like scores
at each cycle. Reads were deposited in the National Center for Biotechnology
Information Sequence Read Archive under accession number SRP009350. Reads in

165

FASTQ formats were aligned to the C. sativus [6] reference genome using TopHat
v1.2.0/Bowtie v0.12.5 [55,56]. A reference annotation of C. sativus (version 2) from the
Cucurbit Genomics Database (http://www.icugi.org/cgi-bin/ICuGI/misc/download.cgi)
was provided in which a representative isoform, the gene model with the longest CDS,
was used to estimate expression of the gene; a total of 23,248 gene models from a total
of 25,600 gene models were used. All other isoforms were discarded from the
annotation set. The minimum and maximum intron length was set to 5 and 50,000 bp,
respectively; all other parameters were set to the default values.

Normalized gene expression levels were calculated using Cufflinks v0.9.3 [57] using the
quartile normalization option [58] to improve differential expression calculations of lowly
expressed genes. The maximum intron length parameter was set to 50,000 bp. All other
parameters were used at the default settings. Sampling depth was evaluated on
expression estimates by randomly selecting 5, 10, 15, 25 and 30 million reads from the
total pool of reads at all time points. Pearson correlation coefficient analyses of log2
FPKM values were performed using R (http://cran.r-project.org/), in which all log2 FPKM
values less than zero were set to zero. Only tests significant at p = 0.05 are shown. The
correlation values were clustered with hierarchical clustering using a Pearson
correlation distance metric with average linkage and depicted as a heat map. For each
node, bootstrap support values were calculated from 1000 replicates using Multiple
Experiment Viewer Software (MeV) v4.5 [59]. To examine biological variation, PCC was
calculated for the log2 transformed FPKM values of the genes expressed in both
replicates at a particular time point.

166

Microarray data acquisition and processing

Comparative analyses of host gene expression responses during a compatible
interaction with the model species A. thaliana utilized microarry-based gene expression
data from an H. arabidopsidis-A. thaliana time course experiment [41]. The dataset was
comprised of A. thaliana genes expressed in response to H. arabidopsidis infection at 0,
0.5, 2, 4, and 6 dpi using the ATH1 Affymetrix platform. The probe intensity values were
downloaded

from

Gene

Expression

Omnibus

(http://www.ncbi.nlm.nih.gov/geo/;

GSE22274) [60], normalized using Robust Multichip Analysis method [61]. OrthoMCL
[62] was used to identify clusters of orthologs and close paralogs in cucumber and A.
thaliana. For the mRNA-Seq to microarray comparative analysis only single copy
orthologous genes were considered for further analyses. The FPKM and intensity
values were log2 transformed, and Spearman rank correlations of the single copy
orthologs in both hosts were calculated in R (http://cran.r-project.org/). Only tests
significant at p = 0.05 are shown.

Identification of differentially expressed genes

The Cuffdiff program within Cufflinks version 0.9.3 [57] was used to identify differentially
expressed genes using pair-wise comparisons of the six time points and the control.
The minimum number of alignments at a gene required to test was set to 100. Quartile
normalization and a false discovery rate of 0.01 after Benjamini-Hochberg correction for
multiple testing were used. Significance in the numbers of differentially expressed
167

genes between time points was tested using the two-sample t-test.

Functional analyses of differentially expressed genes

Functional annotation for all C. sativus genes were generated from BLAST searches of
the UniProt databases (Uniref100) [63] and combined with Pfam (version 24.0, [64])
protein families assignment performed using HMMER3 [65]. C. sativus sequences were
functionally annotated based on the best possible UniRef sequence match using a
minimum E value cutoff of 1E-5. If there was no UniRef sequence match, functional
annotations were assigned using Pfam domains. Transcription factors were annotated
based on PFAM domains assignment.

Gene co-expression network analysis

Gene modules of highly correlated genes were identified using the WGCNA method [47]
implemented in R. All gene FPKM expression values were log2 transformed and any
transformed FPKM value less than 1 was converted to zero. Genes without variation
across the mock sample and 6 time points were filtered out using a Coefficient of
Variation (CV = σ/μ) cutoff of 0.4. The β and treecut parameters for WGCNA were 13
and 0.9, respectively.

All other parameters were used with their default values.

Eigengenes for each gene module [48] were calculated and presented as a heat map.

168

ACKNOWLEDGEMENTS

We thank members of the Day lab for critical reading of the manuscript.

169

APPENDIX

170

Figure S4.1

171

Figure S4.1 (cont'd) Concordance of expression values in two biological
replicates of Cucumis sativus during infection by Pseudoperonospora
cubensis. Reads from different time points were mapped to the C. sativus genome
using Bowtie version 0.12.5 and TopHat version 1.2.0. Fragments per kilobase pair of
exon model per million fragments mapped (FPKM) values were calculated using
Cufflinks version 0.9.3 and the C. sativus genome annotations. For each time point,
log2 transformed FPKM values of equal number of genes from both replicates are
plotted. R2, correlation coefficient; dpi, day post-inoculation.

172

Figure S4.2

173

Figure S4.2 (cont'd)

174

Figure S4.2 (cont'd)

175

Figure S4.2 (cont'd)

Figure S4.2 Trend plots for all 11 modules. All 11 modules generated using
WGCNA are shown (Modules A through K). The number of genes in each module is
shown in parentheses.

176

REFERENCES

177

REFERENCES
1. Tanurdzic M, Banks JA (2004) Sex-determining mechanisms in land plants. Plant
Cell 16 Suppl: S61-71.
2. Lough TJ, Lucas WJ (2006) Integrative plant biology: role of phloem long-distance
macromolecular trafficking. Ann Rev Plant Biol 57: 203-232.
3. Hammerschmidt R (1999) PHYTOALEXINS: What have we learned after 60 years?
Ann Rev Phytopathol 37: 285-306.
4. Hammerschmidt R (1999) Induced disease resistance: how do induced plants stop
pathogens? Physiol Mol Plant Pathol 55: 77-84.
5. Ren Y, Zhang Z, Liu J, Staub JE, Han Y, et al. (2009) An integrated genetic and
cytogenetic map of the cucumber genome. PLoS ONE 4: e5795.
6. Huang S, Li R, Zhang Z, Li L, Gu X, et al. (2009) The genome of the cucumber,
Cucumis sativus L. Nat Genet 41: 1275-1281.
7. Wóycicki R, Witkowicz J, Gawroński P, Dąbrowska J, Lomsadze A, et al. (2011) The
genome sequence of the North-European cucumber (Cucumis sativus L.)
unravels evolutionary adaptation mechanisms in plants. PLoS ONE 6: e22728.
8. Guo S, Zheng Y, Joung JG, Liu S, Zhang Z, et al. (2010) Transcriptome sequencing
and comparative analysis of cucumber flowers with different sex types. BMC Gen
11: 384.
9. Ando K, Grumet R (2010) Transcriptional profiling of rapidly growing cucumber fruit
by 454-pyrosequencing analysis. J Am Soc Hort Sci 135: 291-302.
10. Olczak-Woltman H, Schollenberger M, Niemirowicz-Szczytt K (2009) Genetic
background of host-pathogen interaction between Cucumis sativus L. and
Pseudomonas syringae pv. lachrymans. J Appl Genet 50: 1-7.
11. Savory EA, Granke LL, Quesada-Ocampo LM, Varbanova M, Hausbeck MK, et al.
(2011) The cucurbit downy mildew pathogen Pseudoperonospora cubensis. Mol
Plant Pathol 12: 217-226.

178

12. Colucci SJ, Wehner TC, Holmes GJ (2006) The downy mildew epidemic of 2004
and 2005 in the Eastern United States. In: Holmes G, editor. Proc Cucurbitaceae
2006. Raleigh, NC: Universal Press.
13. Lebeda A, Pavelková J, Urban J, Sedláková B (2011) Distribution, host range and
disease severity of Pseudoperonospora cubensis on cucurbits in the Czech
Republic. J Phytopathol 159: 589-596.
14. Arauz LF, Neufeld KN, Lloyd AL, Ojiambo PS (2010) Quantitative models for
germination and infection of Pseudoperonospora cubensis in response to
temperature and duration of leaf wetness. Phytopathol 100: 959-367.
15.

Thomas C, Inaba T, Cohen Y (1987) Physiological
Pseudoperonospora cubensis. Phytopathol 77: 1621-1624.

Specialization

in

16. Shetty NV, Wehner TC, Thomas CE, Doruchowski RW, Vasanth Shetty KP (2002)
Evidence for downy mildew races in cucumber tested in Asia, Europe, and North
America. Scientia Hort 94: 231–239.
17. Lebeda A, Widrlechner MP (2003) A set of Cucurbitaceae taxa for differentiation of
Pseudoperonespora cubensis pathotypes. J Plant Dis Prot 110: 337–349.
18. Cohen Y, Meron I, Mor N, Zuriel S (2003) A new pathotype of Pseudoperonospora
cubensis causing downy mildew in cucurbits in Israel. Phytoparasitica 31: 458–
466.
19. Lebeda A (2002) Pathogenic variation of Pseudoperonospora cubensis in the Czech
Republic and some other European countries. II International Symposium on
Cucurbits 588. pp. 137–141
20. Lebeda A, Urban J (2004) Distribution, harmfulness and pathogenic variability of
cucurbit downy mildew in the Czech Republic. Acta Fytotech Zootech 7: 170-173.
21. Sarris P, Abdelhalim M, Kitner M, Skandalis N, Panopoulos N, et al. (2009)
Molecular polymorphisms between populations of Pseudoperonospora cubensis
from Greece and the Czech Republic and the phytopathological and phylogenetic
implications. Plant Pathol 58: 933–943.

179

22. Tian M, Win J, Savory E, Burkhardt A, Held M, et al. (2011) 454 Genome
sequencing of Pseudoperonospora cubensis reveals effector proteins with a
QXLR translocation motif. Mol Plant-Microbe Interact 24: 543-553.
23. Savory EA, Zou C, Adhikari BN, Hamilton JP, Buell CR, et al. (2012) Alternative
Splicing of a Multi-Drug Transporter from Pseudoperonospora cubensis
Generates an RXLR Effector Protein That Elicits a Rapid Cell Death. PLoS One
7: e34701.
24. Durrant WE, Dong X (2004) Systemic acquired resistance. Ann Rev Phytopathol 42:
185-209.
25. Phuntumart V, Marro P, Métraux J-P, Sticher L (2006) A novel cucumber gene
associated with systemic acquired resistance. Plant Sci 171: 555-564.
26. Sticher L, Mauch-Mani B, Metraux JP (1997) Systemic acquired resistance. Ann
Rev Phytopathol 35: 235-270.
27. Bent AF, Mackey D (2007) Elicitors, effectors, and R genes: The new paradigm and
a lifetime supply of questions. Ann Rev Phytopathol 45: 399-436.
28. Call AD, Wehner T (2010) Gene list 2010 for cucumber. Cucurbi Genet Coop Rep
28-29: 105-141.
29. Runge F, Choi Y-J, Thines M (2011) Phylogenetic investigations in the genus
Pseudoperonospora reveal overlooked species and cryptic diversity in the P.
cubensis species cluster. Eur J Plant Pathol 129: 135-146.
30. Taler D, Galperin M, Benjamin I, Cohen Y, Kenigsbuch D (2004) Plant eR genes
that encode photorespiratory enzymes confer resistance against disease. Plant
Cell 16: 172-184.
31. Criswell A (2008) Screening cucumber (Cucumis sativus) for resistance to downy
mildew (Pseudoperonospora cubensis). NCSU Thesis.
32. Wang ET, Sandberg R, Luo S, Khrebtukova I, Zhang L, et al. (2008) Alternative
isoform regulation in human tissue transcriptomes. Nature 456: 470 - 476.

180

33. Boddu J, Cho S, Muehlbauer GJ (2007) Transcriptome analysis of trichotheceneinduced gene expression in barley. Mol Plant-Microbe Interact 20: 1364-1375.
34. Huibers RP, de Jong M, Dekter RW, Van den Ackerveken G (2009) Diseasespecific expression of host genes during downy mildew infection of Arabidopsis.
Mol Plant-Microbe Interact 22: 1104-1115.
35. Gupta S, Chakraborti D, Sengupta A, Basu D, Das S (2010) Primary metabolism of
chickpea is the initial target of wound inducing early sensed Fusarium oxysporum
f. sp. ciceri race I. PLoS ONE 5: e9030.
36. Savory EA, Adhikari BN, Hamilton JP, Vaillancourt B, Buell CR, et al. (2012) mRNASeq Analysis of the Pseudoperonospora cubensis Transcriptome During
Cucumber (Cucumis sativus L.) Infection. PLoS One 7: e35796.
37. Coates ME, Beynon JL (2010) Hyaloperonospora arabidopsidis as a pathogen
model. Ann Rev Phytopathol 48: 329-345.
38. Volk GM, Haritatos EE, Turgeon R (2003) Galactinol synthase gene expression in
melon. J Am Soc Hort Sci 128: 8-15.
39. Li JW, Liu J, Zhang H, Xie CH (2011) Identification and transcriptional profiling of
differentially expressed genes associated with resistance to Pseudoperonospora
cubensis in cucumber. Plant Cell Rep 30: 345-357.
40. Knepper C, Day B (2010) From perception to activation: The molecular-genetic and
biochemical landscape of disease resistance signaling in plants. The Arabidopsis
Book: e012.
41. Wang W, Barnaby JY, Tada Y, Li H, Tor M, et al. (2011) Timing of plant immune
responses by a central circadian regulator. Nature 470: 110-114.
42. Miranda M, Ralph SG, Mellway R, White R, Heath MC, et al. (2007) The
transcriptional response of hybrid poplar (Populus trichocarpa x P. deltoides) to
infection by Melampsora medusae leaf rust involves induction of flavonoid
pathway genes leading to the accumulation of proanthocyanidins. Mol PlantMicrobe Interact 20: 816-831.

181

43. Polesani M, Bortesi L, Ferrarini A, Zamboni A, Fasoli M, et al. (2010) General and
species-specific transcriptional responses to downy mildew infection in a
susceptible (Vitis vinifera) and a resistant (V. riparia) grapevine species. BMC
Gen 11: 117.
44. Polesani M, Desario F, Ferrarini A, Zamboni A, Pezzotti M, et al. (2008) cDNA-AFLP
analysis of plant and pathogen genes expressed in grapevine infected with
Plasmopara viticola. BMC Gen 9: 142.
45. Gaulin E, Madoui MA, Bottin A, Jacquet C, Mathe C, et al. (2008) Transcriptome of
Aphanomyces euteiches: new oomycete putative pathogenicity factors and
metabolic pathways. PLoS ONE 3: e1723.
46. Langfelder P, Zhang B, Horvath S (2008) Defining clusters from a hierarchical
cluster tree: the Dynamic Tree Cut package for R. Bioinformatics 24: 719 - 720.
47. Zhang B, Horvath S (2005) A general framework for weighted gene co-expression
network analysis. Stat App Gen Mol Biol 4: Article 17.
48. Langfelder P, Horvath S (2007) Eigengene networks for studying the relationships
between co-expression modules. BMC Sys Biol 1: 54.
49. Takemoto D, Demirdöven A (2007) Lipoxygenase in fruits and vegetables: A review.
Enz Micro Tech 40: 491-496.
50. Hilaire E, Young SA, Willard LH, McGee JD, Sweat T, et al. (2001) Vascular
defense responses in rice: peroxidase accumulation in xylem parenchyma cells
and xylem wall thickening. Mol Plant-Microbe Interact 14: 1411-1419.
51. Varbanova M, Porter K, Lu F, Ralph J, Hammerschmidt R, et al. (2011) Molecular
and biochemical basis for stress-induced accumulation of free and bound pcoumaraldehyde in cucumber. Plant Physiol 157: 1056-1066.
52. Brader G, Djamei A, Teige M, Palva ET, Hirt H (2007) The MAP kinase kinase
MKK2 affects disease resistance in Arabidopsis. Mol Plant-Microbe Interact 20:
589-596.
53. Bernoux M, Timmers T, Jauneau A, Brière C, de Wit PJGM, et al. (2008) RD19, an
Arabidopsis cysteine protease required for RRS1-R-mediated resistance, Is
182

relocalized to the nucleus by the Ralstonia solanacearum PopP2 effector. Plant
Cell 20: 2252-2264.
54. Takemoto D, Doke N, Kawakita K (2001) Characterization of elicitor-inducible
tobacco genes isolated by differential hybridization. J Gen Plant Pathol 67: 8996.
55. Trapnell C, Pachter L, Salzberg SL (2009) TopHat: discovering splice junctions with
RNA-Seq. Bioinformatics 25: 1105-1111.
56. Langmead B, Trapnell C, Pop M, Salzberg S (2009) Ultrafast and memory-efficient
alignment of short DNA sequences to the human genome. Genome Biol 10: R25.
57. Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, et al. (2010) Transcript
assembly and quantification by RNA-Seq reveals unannotated transcripts and
isoform switching during cell differentiation. Nat Biotechnol 28: 511-515.
58. Bullard J, Purdom E, Hansen K, Dudoit S (2010) Evaluation of statistical methods
for normalization and differential expression in mRNA-Seq experiments. BMC
Bioinformatics 11: 94.
59. Saeed AI, Bhagabati NK, Braisted JC, Liang W, Sharov V, et al. (2006) TM4
microarray software suite. Methods Enzymol 411: 134-193.
60. Edgar R, Domrachev M, Lash AE (2002) Gene Expression Omnibus: NCBI gene
expression and hybridization array data repository. Nucleic Acids Res 30: 207210.
61. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, et al. (2003)
Exploration, normalization, and summaries of high density oligonucleotide array
probe level data. Biostat 4: 249 - 264.
62. Li L, Stoeckert CJ, Jr., Roos DS (2003) OrthoMCL: identification of ortholog groups
for eukaryotic genomes. Genome Res 13: 2178-2189.
63. Suzek BE, Huang H, McGarvey P, Mazumder R, Wu CH (2007) UniRef:
comprehensive and non-redundant UniProt reference clusters. Bioinformatics 23:
1282-1288.

183

64. Bateman A, Birney E, Durbin R, Eddy SR, Howe KL, et al. (2000) The Pfam Protein
Families Database. Nucleic Acids Res 28: 263-266.
65. Finn RD, Clements J, Eddy SR (2011) HMMER web server: interactive sequence
similarity searching. Nucleic Acids Res 39: W29-37.

184

CHAPTER 5

Conclusions and Future Directions

184

CONCLUSIONS

Cucurbit downy mildew, caused by the oomycete pathogen Pseudoperonospora
cubensis, emerged as a significant economic threat to cucumber production in the
United States during the 2004 and 2005 growing seasons and is currently the major
limiting factor of cucumber production [1]. Although downy mildew had been a major
issue in Europe since the mid-1980's, in the U.S., the disease on cucumber had been
successfully controlled since the 1940's through host resistance. As such, only limited
research was conducted to understand the biology, genetics, and virulence of the
pathogen. The focus of this dissertation was to expand knowledge of this economically
important oomycete pathogen using cell biology, genetic, genomic, and transcriptomic
approaches to better understand the molecular-genetic basis for the interaction between
cucumber (Cucumis sativus) and Ps. cubensis.

The establishment of genetic resources in the form of a draft genome assembly laid the
groundwork for additional analyses, including identification and characterization of
candidate effector proteins (Chapter 2) as well as providing a framework for mapping
mRNA-Seq reads from a large-scale gene expression study of infection (Chapter 3).
One RXLR-type effector protein, PscRXLR1 was characterized in detail and shown to
be the result of alternative splicing of a non-effector protein (Chapter 2). Additionally,
due to the obligate nature of Ps. cubensis and our sampling methods for the mRNA-Seq
study, we were able to measure host gene expression changes during C. sativus
infection with Ps. cubensis, providing the first in-depth analysis of this important plantpathogen interaction (Chapter 4). Overall, the research presented in this dissertation
185

greatly expands our knowledge of the C. sativus-Ps. cubensis interaction and provides
much-needed insight into the virulence of Ps. cubensis and the defense response of C.
sativus.

The establishment of genomic resources for major oomycete pathogens such as
Phytophthora infestans [2], Phytophthora sojae, Phytophthora ramorum [3], Pythium
ultimum [4], as well as for the Arabidopsis thaliana pathogen, Hyaloperonospora
arabidopsidis [5], and subsequent identification of RXLR and RXLR-like effector
proteins has anchored the recent evolution of molecular plant pathology [6]. The study
of oomycete effector proteins has been central to the study of oomycete pathogenicity
and virulence and the understanding of the interactions between these pathogens and
their hosts [7,8,9,10,11,12,13,14,15,16]. Therefore, the first step in identifying these
important pathogenicity and virulence determinants in Ps. cubensis was the generation
of a 64.4 Mbp genomic assembly of the MSU-1 isolate and prediction of 23,519 loci and
23,522 gene models (Chapter 2).

Due to their modular structure, including an N-

terminal signal peptide, an RXLR translocation motif, and a C-terminal functional
domain, RXLR effectors are readily identified via bioinformatics [17,18]. We identified
271 candidate effector proteins within the Ps. cubensis genome with variable RXLR
motifs, including the previously identified R and Q, predicting 20 possible amino acids at
position R1. Additionally, mRNA-Seq analysis of infection provided expression support
for 19 of the 20 possible R1 substitutions (except Y, Tyr; Table S2.1, Figure 3.5). While
candidate effectors with these R1 amino acid substitutions, with the exception of QXLR
[19], have yet to be functionally validated, their expression during a compatible

186

interaction supports their putative virulence function and role as bona fide effector
proteins.

This variation may provide clues to the diversity among Ps. cubensis

pathotypes in terms of their virulence and host specificity.

In-depth functional analysis of one Ps. cubensis effector protein, RXLR protein 1
(PscRXLR1) led to a surprising and interesting result - it was demonstrated that
PscRXLR1 arises as a product of alternative splicing (Figure 2.5), making this the first
example of an alternative splicing event in plant pathogenic oomycetes transforming a
non-effector gene into a functional effector protein.

Using a set of experiments

designed to validate effector protein function, it was shown that PscRXLR1 was upregulated during the early stages of infection and that heterologous expression of
PscRXLR1 in Nicotiana benthamiana elicits a rapid cell death phenotype (Figure 2.3,
Figure 2.4).

Additionally, characterization of the closest P. infestans ortholog,

PITG_17484, a member of the Drug/Metabolite Transporter (DMT) superfamily to
Psc_RXLR1 confirmed that it did not elicit the same cell death response (Figure 2.3).
The orthologous relationship between PscRXLR1 and PITG_17484 led us to question if
such effector-non-effector ortholog pairs are common among oomycete plant
pathogens. We subsequently examined the relationship(s) among putative ortholog
pairs in Ps. cubensis and P. infestans. Of 271 predicted Ps. cubensis effector proteins,
only 109 (41%) had a putative ortholog in P. infestans and evolutionary rate analysis of
these orthologs showed an evolutionary rate significantly faster than most other genes
(Figure 2.1). In total, these data provide a basis for comparative analysis of candidate
effector proteins and their non-effector orthologs as a means of understanding function

187

and evolutionary history of pathogen effectors. In addition, the discovery of an effector
protein arising from an alternative splicing event may indicate an adaptive mechanism
utilized by Ps. cubensis and potentially other pathogens to generate proteome
complexity during infection.

Next generation sequencing of the transcriptome (mRNA-Seq) permits deep, robust
assessments of transcript abundance and structure, and when applied to host-pathogen
interactions, enables insight into pathogen mechanisms to suppress and subvert host
defense responses [20,21,22,23].

In this dissertation, the first large-scale global

transcriptome analysis of Ps. cubensis infection of a susceptible C. sativus cultivar,
‘Vlaspik’ is presented, yielding new information about both pathogen and host gene
expression during a compatible interaction (Chapter 3, 4).

Sampling Ps. cubensis-

infected C. sativus tissue during an 8-day time course and then separating speciesspecific reads in silico enabled simultaneously collection and analysis of infection
transcriptome dynamics from both pathogen and host. As described herein, we were
able to detect the expression of 7,821 Ps. cubensis genes (Chapter 3) and 14,476
cucumber genes (Chapter 4) throughout the infection process from 1 day postinoculation (1 dpi) to 8 dpi, providing the first genome-scale analysis of the cucurbit
downy mildew interaction.

In Ps. cubensis, genes involved in virulence, including RXLR and RXLR-like effectors,
Crinkler (CRN) effectors, and host-targeted hydrolytic enzymes acting on plant
proteinases, lipases, and several sugar-cleaving enzymes were all differentially

188

expressed throughout the infection time course.

Corresponding with our visual

assessment of symptoms and infection structures, clustering and co-expression
analyses identified distinct modules of Ps. cubensis genes that were representative of
early, intermediate, and late infection stages. The 1 dpi time point, when encystment of
zoospores is occurring, represents a unique pattern of gene expression supported by
multiple modes of analysis. Gene expression at this time point poorly correlated with all
other time points (Figure 3.4), had the highest percentage of differentially expressed
genes across all pair-wise comparisons (Table 3.1), and was represented by a unique
module, Module 1 (Figure 3.8), when analyzed using Weighted Gene Correlation
Network Analysis (WGCNA). The genes represented here are likely those involved in
zoospore encystment, appressorium production, and the initial penetration of stomata.
Our data also supports an intermediate, transition stage of infection, represented mainly
by overlapping gene expression between the 2, 3, and 4 dpi time points. Module 2,
containing 508 genes including candidate RXLR-type effectors, CRNs, and haustoriumspecific proteins, represents expression during hyphal penetration, growth, and initiation
of haustoria formation, which occurs between 2-4 dpi. The late stage of infection, at 6
and 8 dpi, is characterized by extensive hyphal growth within the mesophyll (Figure 3.1)
and a transition to the expression of genes likely involved in sporulation (Module 6,
Figure 3.8). Overall, these expression data and analyses and their correlation with
pathogen growth have advanced our understanding of molecular and genetic events in
the infection of Ps. cubensis.

189

During Ps. cubensis infection, expression of defense-related genes, including catalases,
chitinases, lipoxygenases, peroxidases, and protease inhibitors, was highly upregulated in C. sativus within 1 dpi (Table S4.2), suggesting an active host defense
response to early infection by Ps. cubensis. These genes are coordinately downregulated at 2 dpi (Module B, Figure 4.7) and remain lowly expressed for the
continuation of the time course, suggesting that they are suppressed by the pathogen.
With expression profiles for nearly 15,000 genes during a compatible interaction, we
have new insight into molecular events at the host-pathogen interface including a suite
of defense response-related genes that are down-regulated early upon infection and
transcriptional networks that respond in a temporal manner throughout the infection
cycle. Most intriguingly, these networks include transcription factors and genes of no
known function, which may have a role in the host-pathogen interaction.

Overall, the work presented in this dissertation represents a substantial advance in the
understanding of the Ps. cubensis-C. sativus interaction. Using a variety of methods
ranging from cell biology to genetics to large-scale transcriptomic analysis, it serves to
advance knowledge of Ps. cubensis biology, genomics, virulence determinants, and
gene expression during infection, as well as corresponding host responses.

190

FUTURE DIRECTIONS

While substantial progress has been made over the course of this dissertation project to
establish genetic, genomic, and transcriptomic resources for the study of the Ps.
cubensis-C. sativus interaction, it only represents a start to the study of this important
and fascinating pathosystem.

To build on the work herein, further functional

characterization of additional candidate effector proteins, exploration of alternative
splicing events over the course of infection, and a more detailed analysis of the Ps.
cubensis transcriptome data set is needed to provide additional insight into both the
lifestyle and virulence of Ps. cubensis. For C. sativus, additional analysis of the mRNASeq data would increase our knowledge of host defense responses in cucumber, and
more transcriptome analyses of lines with resistance to Ps. cubensis could potentially
lead to discoveries of new sources of resistance which could be utilized in breeding of
resistant varieties for growers.

This dissertation and recent work [19] have identified 271 candidate effector proteins
with variable RXLR motifs and established a pipeline for their characterization using
both bioinformatics and molecular biology tools. The prediction of 20 possible amino
acid substitutions at the R1 position has expanded our previous knowledge of the
distribution of R1 domains in putative Ps. cubensis effector proteins, previously thought
to be equally distributed between RXLR and QXLR motifs [19]. While expression
support exists for 19 of these 20 substitutions, additional functional validation is
necessary and would add to both general knowledge of oomycete and fungal effector

191

proteins and provide insight into the possible mechanisms utilized by Ps. cubensis to
cause infection.

Functional characterization of the effector protein PscRXLR1 demonstrated that it arises
as a product of alternative splicing from Psc_781.4, making it the first example of an
alternative splicing event in plant pathogenic oomycetes transforming a non-effector
gene into a functional effector protein. This was substantiated using a combination of
RT-PCR and transcriptome analysis of Ps. cubensis gene expression during infection.
To build upon this work, analysis to quantify changes in expression levels of PscRXLR1
and Psc_781.4 over the course of infection should be completed. Using qRT-PCR, an
up-regulation of mRNA was identified at 1 dpi, extending through 4 dpi. However, due
to limitations with this technique, we were unable to distinguish between PscRXLR1 and
Psc_781.4. More in depth analysis of alternative splicing events in Ps. cubensis during
infection will allow us to distinguish between these isoforms, as well as enable us to
identify additional changes in splicing over the course of infection. In addition, such
analyses would enable the identification of global splicing events in Ps. cubensis, which
would expand general knowledge about alternative splicing changes during infection, an
as-yet understudied phenomenon.

The large-scale mRNA-Seq data sets generated in this dissertation provide a wealth of
knowledge about gene expression during Ps. cubensis infection on C. sativus, and
while extensive, the analyses and data presented herein only represent a small fraction
of what could be gleaned from them. The analyses for this project, focused on genes

192

involved in pathogenicity and infection, such as effector proteins, pathogenicity-related
genes, and CAZymes. Ps. cubensis is an obligate biotroph and is dependent on C.
sativus for completion of its life cycle. As such, this mRNA-Seq dataset would be ideal
for analyses focused on identifying genes involved in obligate biotrophy and the
establishment of this intricate relationship between pathogen and host. This would not
only yield insight into the nature of obligate biotrophic relationships which could be
extrapolated to other systems, but could also be useful in understanding the
pathogenicity of Ps. cubensis, potentially leading to improved control strategies.
Monitoring the expression of both pathogen and host genes during infection can provide
insight into the interplay between resistance and susceptibility. Initial analysis of the
mRNA-Seq data focused on studying genes and expression patterns that were either
Ps. cubensis or C. sativus specific. This data set is unique in that expression data from
both host and pathogen were collected from the same samples and as such, ideal for
correlative analyses that could look at the co-regulation of gene expression between
pathogen and host. This could potentially aide in the identification of pathogen effectors
that are down-regulating host defense responses or specific host genes that are upregulated in response to pathogen virulence determinants.
Looking specifically at the C. sativus mRNA-Seq data in more detail could increase our
knowledge of host defense responses in cucumber. We identified a suite of genes upregulated at 1 dpi and subsequently down-regulated at 2 dpi (Figure 4.7) - functional
characterization of a set of these genes could provide insight into the host defense
response of C. sativus, as one could hypothesize that they could potentially play a role
in defense if pathogen infection results in their down-regulation. Finally, the mRNA-Seq
193

data presented herein represents a compatible interaction between a virulent pathotype
of Ps. cubensis and a susceptible C. sativus cv., 'Vlaspik'. While ideal for studying
pathogen virulence determinants, to better study the host defense response, additional
transcriptome analyses of C. sativus lines with resistance to Ps. cubensis should be
done.

This could potentially lead to discoveries of new sources of resistance and

breeding of resistant varieties for growers.

The Ps. cubensis-C. sativus interaction, while in its infancy in terms of our
understanding of the genetic and biochemical processes that regulate pathogenicity and
resistance, presents an opportunity to develop a crop-based model system in which to
study oomycete-plant interactions, ultimately focused on identifying mechanisms of host
resistance that can be translated into field application.

The work presented in this

dissertation lays a substantial groundwork to achieve that goal by establishing genomic,
genetic and transcriptomic resources for both Ps. cubensis and C. sativus that can be
built upon to better understand the interplay between pathogen and host.

194

REFERENCES

195

REFERENCES
1. Savory EA, Granke LL, Quesada-Ocampo LM, Varbanova M, Hausbeck MK, et al.
(2011) The cucurbit downy mildew pathogen Pseudoperonospora cubensis. Mol
Plant Pathol 12: 217-226.
2. Haas BJ, Kamoun S, Zody MC, Jiang RH, Handsaker RE, et al. (2009) Genome
sequence and analysis of the Irish potato famine pathogen Phytophthora
infestans. Nature 461: 393-398.
3. Tyler BM (2006) Phytophthora Genome Sequences Uncover Evolutionary Origins
and Mechanisms of Pathogenesis. Science 313: 1261-1266.
4. Levesque CA, Brouwer H, Cano L, Hamilton JP, Holt C, et al. (2010) Genome
sequence of the necrotrophic plant pathogen Pythium ultimum reveals original
pathogenicity mechanisms and effector repertoire. Genome Biol 11: R73.
5. Baxter L, Tripathy S, Ishaque N, Boot N, Cabral A, et al. (2010) Signatures of
adaptation to obligate biotrophy in the Hyaloperonospora arabidopsidis genome.
Science 330: 1549-1551.
6. Birch PR, Armstrong M, Bos J, Boevink P, Gilroy EM, et al. (2009) Towards
understanding the virulence functions of RXLR effectors of the oomycete plant
pathogen Phytophthora infestans. J Exp Bot 60: 1133-1140.
7. Allen RL (2004) Host-Parasite Coevolutionary Conflict Between Arabidopsis and
Downy Mildew. Science 306: 1957-1960.
8. Shan W, Cao M, Leung D, Tyler BM (2004) The Avr1b locus of Phytophthora sojae
encodes an elicitor and a regulator required for avirulence on soybean plants
carrying resistance gene Rps 1b. Mol Plant-Microbe Interact 17: 394–403.
9. Rehmany AP, Gordon A, Rose LE, Allen RL, Armstrong MR, et al. (2005) Differential
recognition of highly divergent downy mildew avirulence gene alleles by RPP1
resistance genes from two Arabidopsis lines. Plant Cell 17: 1839-1850.
10. Armstrong MR, Whisson SC, Pritchard L, Bos JI, Venter E, et al. (2005) An
ancestral oomycete locus contains late blight avirulence gene Avr3a, encoding a
protein that is recognized in the host cytoplasm. Proc Natl Acad Sci U S A 102:
7766-7771.
196

11. Bos JI, Kanneganti TD, Young C, Cakir C, Huitema E, et al. (2006) The C-terminal
half of Phytophthora infestans RXLR effector AVR3a is sufficient to trigger R3amediated hypersensitivity and suppress INF1-induced cell death in Nicotiana
benthamiana. Plant J 48: 165-176.
12. Whisson SC, Boevink PC, Moleleki L, Avrova AO, Morales JG, et al. (2007) A
translocation signal for delivery of oomycete effector proteins into host plant cells.
Nature 450: 115-118.
13. Qutob D, Tedman-Jones J, Dong S, Kuflu K, Pham H, et al. (2009) Copy Number
Variation and Transcriptional Polymorphisms of Phytophthora sojae RXLR
Effector Genes Avr1a and Avr3a. PLoS ONE 4: e5066.
14. Dong S, Qutob D, Tedman-Jones J, Kuflu K, Wang Y, et al. (2009) The
Phytophthora sojae Avirulence Locus Avr3c Encodes a Multi-Copy RXLR
Effector with Sequence Polymorphisms among Pathogen Strains. PLoS ONE 4:
e5556.
15. Bos JI, Armstrong MR, Gilroy EM, Boevink PC, Hein I, et al. (2010) Phytophthora
infestans effector AVR3a is essential for virulence and manipulates plant
immunity by stabilizing host E3 ligase CMPG1. Proc Natl Acad Sci U S A 107:
9909-9914.
16. Cabral A, Stassen JH, Seidl MF, Bautor J, Parker JE, et al. (2011) Identification of
Hyaloperonospora arabidopsidis Transcript Sequences Expressed during
Infection Reveals Isolate-Specific Effectors. PLoS ONE 6: e19328.
17. Schornack S, Huitema E, Cano LM, Bozkurt TO, Oliva R, et al. (2009) Ten things to
know about oomycete effectors. Mol Plant Pathol 10: 795–803.
18. Kamoun S (2006) A catalogue of the effector secretome of plant pathogenic
oomycetes. Ann Rev Phytopathol 44: 41-60.
19. Tian M, Win J, Savory E, Burkhardt A, Held M, et al. (2011) 454 Genome
sequencing of Pseudoperonospora cubensis reveals effector proteins with a
QXLR translocation motif. Mol Plant-Microbe Interact 24: 543-553.
20. Wang E, Sandberg R, Luo S, Khrebtukova I, Zhang L, et al. (2008) Alternative
isoform regulation in human tissue transcriptomes. Nature 456: 470 - 476.
197

21. Boddu J, Cho S, Muehlbauer GJ (2007) Transcriptome analysis of trichotheceneinduced gene expression in barley. Mol Plant-Microbe Interact 20: 1364-1375.
22. Huibers RP, de Jong M, Dekter RW, Van den Ackerveken G (2009) Diseasespecific expression of host genes during downy mildew infection of Arabidopsis.
Mol Plant-Microbe Interact 22: 1104-1115.
23. Gupta S, Chakraborti D, Sengupta A, Basu D, Das S (2010) Primary metabolism of
chickpea is the initial target of wound inducing early sensed Fusarium oxysporum
f. sp. ciceri race I. PLoS ONE 5: e9030.

198