1. Introduction
Downy mildew (DM) is a devastating foliar disease of cucurbits with a global distribution. The causal agent,
Pseudoperonospora cubensis (Berk. and Curt.) Rost. (Oomycota, Peronosporaceae), is an obligate biotrophic oomycete pathogen that attacks over 40 host plant species belonging to 20 genera of the
Cucurbitaceae [
1,
2]. Typical symptoms in cucumber consist of chlorotic irregular lesions with sporulation on the lower leaf surface. Several review articles provide basic information on the biology, epidemiology, and control of the disease [
1,
2,
3,
4,
5]. Infection may take place if free leaf moisture is available for ≥2 h at an appropriate temperature [
6]. The sporangia release biflagellate zoospores that swim towards the stomata where they encyst, germinate, and penetrate. Hyphae grow into the intercellular space, colonize the mesophyll tissue, and establish intracellular haustoria for nutrient uptake. Haustoria also deliver effector proteins to facilitate the establishment and/or combat the host plant’s defense response system [
3,
7]. At ≥4 days post-inoculation, hyaline sporangiophores emerge from stomata bearing dark sporangia at their tip. Sporangia are dispersed by wind or rain and continue the asexual disease cycle. Lesion development is strongly affected by temperature and light. Sporulation occurred in darkness at a moisture-saturated atmosphere. Sporangial yield is strongly influenced by the availability of photosynthate. Labeled C
14-CO
2 it was supplied to infected cucumber plants during the day and found in carbohydrates in the sporangia that were produced the following night [
8].
Current control measures of downy mildew in cucurbits rely on fungicide applications. However, the frequent appearance of insensitive isolates to the current chemistries rendered them ineffective. The novel fungicide oxathiapiprolin that targets oxysterol binding proteins provides a relief due to its excellent systemic activity against DM in cucumber [
9,
10].
In the past two decades, major changes in the population structure of
P. cubensis occurred: new genotypes, races, pathotypes, and mating types were reported from around the globe [
4,
5,
11,
12,
13]. Possible mechanisms may involve the cultivation of large acreages of monocultures, the introduction of new cultivars, changes in climatic conditions (e.g., global warming), and the migration, mutation, and sexual recombination of the pathogen. By using ISSR (Inter-Simple Sequence Repeats) and SRAP (Sequence Related Amplified Polymorphism) markers, Polat et al. [
14] discovered remarkable genetic diversity within and among isolates of
P. cubensis in Europe and Asia. While isolates from Turkey and the Czech Republic exhibited uniform genetic background, the isolates from Israel were clearly distinguished from the others, probably due to migration and/or frequent sexual reproduction of the pathogen in Israel. Wallace et al. [
15] used 10 SSR (Single Sequence Repeats) markers to show that in the USA,
P. cubensis has two distinct, host-adapted clades at the cucurbit species level. Clade 1 isolates preferentially infect
Cucurbita pepo, Cucurbita maxima, Cucurbita moschata, Citrullus lanatus, and wild hosts
Momordica charantia, and
Momordica balsamina, while clade 2 isolates preferentially infect
Cucumis sativus, Cucumis melo, and the wild host
Lagenaria siceraria. Clade 1 showed random mating and evidence of recombination and clade 2 non-random mating and no evidence of recombination. In Israel, the A1 isolates preferably infect
Cucumis sativus, Cucumis melo, and
Lagenaria sp. while the A2 isolates preferably infect
Cucurbita pepo, Cucurbita maxima, and Cucurbita moschata [
16].
Several introductions of wild cucumber were reported to carry resistance genes/QTLs (Quantitative Trait Loci) against
P. cubensis, including PI 197085, PI 197087, PI 197088, PI 330628, Chinese Long, TH118FLM, and
Cucumis hystrix [
17]. Overall, many QTLs associated with resistance to DM have been identified across seven chromosomes [
18]. However, researchers in different countries reported on a different number of genes or QTLs that confer resistance against the disease, probably because they worked in different environments, used different isolates of the pathogen, and/or used different evaluation methods. PI 197087 was reported to carry one, two, or three recessive genes, or two or three partially dominant genes for resistance. PI 197088 was reported to have 3–14 QTLs residing on chromosomes 1–7.
Table 1 summarizes the genetic data available in the literature on genes/QTLs conferring resistance of cucumber against DM caused by
P. cubensis.
Despite the extensive screening and breeding efforts that were done to identify sources of resistance and to incorporate them into commercial cultivars [
9], no cultivars currently offer a high level of resistance to the populations of
P. cubensis that occur in different parts of the world. The reasons for the lack of resistant cultivars may derive from the heterozygosity of the resistant sources used for breeding, the continuous changes in the population structure of the pathogen, and the difficulty to pyramid the number of genes/QTLs in one cultivar.
The most promising current sources of resistance are PI 197088 and PI 330628 [
30]. However, no data are available on the magnitude of their resistance against different isolates of
P. cubensis from different parts of the world.
The objectives of this study were to: (i) stabilize PI 197088 and PI 330628 for resistance against multiple isolates of P. cubensis from different parts of the world. (ii) study the mechanism of resistance of PI 197088 and PI 330628 against P. cubensis. (iii) determine the mode of inheritance of resistance in PI 197088 and PI 330628 against multiple isolates of P. cubensis from different parts of the world.
4. Discussion
The population of
P. cubensis in the field may consist of many isolates, pathotypes, or races with varying degrees of pathogenicity or virulence thus rendering host resistance ineffective. Indeed, combating downy mildew (DM) in cucumber through host plant resistance or fungicide applications has become more complex in the past two decades due to the emergence of new pathotypes, races, and mating types of the causal agent
P. cubensis. Old cucumber cultivars resistant to DM succumbed to the new pathotypes, and the old fungicidal chemistries lost activity due to the prevalence of resistant isolates of the pathogen [
4,
13]. Breeding cucumber for DM resistance is a long and laborious task due to the lack of stable, multi-race resistant sources and the complex mode of resistance inheritance.
Here we identified two sources of wild cucumber with multi-race/pathotype resistance. We characterized the mechanism of their resistance and determined the way they inherit resistance to their progeny plants. Because the resistance of accession to a local isolate of P. cubensis does not necessarily mean that it will be resistant to isolates that prevail in other locations, we used a large collection of isolates from different parts of the world to screen resistance. We developed a detached leaf bioassay in which we could determine the resistance of a single plant to multiple isolates of P. cubensis. We thus were able, for the first time, to study the mode of inheritance of resistance to multiple isolates and predict the performance of the resistant pedigrees in other countries.
Of the six
Cucumis sativum genotypes known to exhibit resistance against
P. cubensis [
4], only PI 197088 and PI 330628 [
28,
30] exhibited multiple-isolate resistance. They were self-pollinated for three generations to bring their multiple-isolate resistance to homozygosity. The stabilized lines were used for the inheritance studies reported here.
When grown in the field under natural epiphytotic conditions, no disease was observed on the leaves of PI 197088 or PI 330628. However, when artificially inoculated in the field or in growth chambers, a few necrotic lesions did appear. Microscopic observations revealed that PI 197088 and PI 330628 exhibit similar responses to artificial inoculation with
P. cubensis. The pathogen ceased developing at a relatively late stage after penetration and developed some initial hyphae and haustoria. The haustoria formed were encased with callose, which probably inhibits the intake of nutrients into the mycelium, while the infected cells accumulated lignin-like, phloroglucinol-positive materials. A similar structural mode of resistance was observed in melons resistant to
P. cubensis [
33,
34]. These defense compounds still allowed some deteriorated sporangiophores to emerge from the stoma but almost totally prevented sporangial production. We show here that unlike the resistance in melon which breaks down at 14 °C [
34], the resistance in PI 197088 and PI 330628 remained effective at a low colonization temperature of 14 °C.
We used a double visual scoring system (percent infected leaf area and sporulation intensity) to determine the level of resistance to DM in detached leaves (
Figure 1). We observed that leaves taken from F2 plants of the cross PI 197088 × SMR-18 or 330628 × SMR-18 segregated in their phenotypic responses to infection with
P. cubensis ranging from complete resistance to high susceptibility. The pattern of segregation depended on the isolate used for inoculation. The differential pattern of response to different isolates indicated that a number of genes might be involved in resistance. The Mendelian analysis was employed to the segregated populations after categorical classification into S: R or R: MR: S. The analysis of two categories R and S indicated that resistance in PI 197088 or PI 330628 is controlled by either 1 dominant, 1 recessive, or 2 recessive genes, depending on the isolate used for inoculation. Analysis with three categories of S, MR, and R did not fit, in most cases, any Mendelian model of segregation.
F3 plants, derived from either susceptible or resistant F2 plants of the cross PI 197088 × SMR-18, showed a continuous pattern of variable resistance to DM in the field. When detached leaves were inoculated with different isolates, F3 plants segregated R:S 3:1 or 1:3, depending on the isolate used for inoculation, reaffirming that inheritance of resistance to DM is isolate-dependent.
Interestingly, field-grown F2 plants of the cross between the two resistant genotypes PI 197088 × PI 330628 were all fully resistant at the end of the season (no DM symptoms). However, two out of 75 plants in 2017 and one out of 130 plants in 2019 showed a few DM lesions, consisting of about 5% infected leaf area. Neighboring SMR-18 plants showed about 90% leaf area infected. This suggests that PI 197088 and PI 330628 differ in at least one gene for resistance. On the other hand, they share one QTL dm4.1 as suggested by Wang et al. [
28].
Isolate-dependent inheritance of disease resistance is a rarely reported phenomenon [
36]. Lapin et al. [
37] showed that unlike most natural
Arabidopsis thaliana accessions that are susceptible to one or more isolates of the downy mildew pathogen
Hyaloperonospora arabidopsidis, accession C24 is resistant to all isolates tested. The resistance of C24 was found to be a multigenic trait with complex inheritance. Many identified resistance loci were isolate-specific and located on different chromosomes. Among the C24 resistance QTLs, there were dominant, codominant, and recessive loci. Interestingly, none of the identified loci significantly contributed to resistance against all three tested isolates.
Unlike wild cucumbers, resistance of the wild melon (
Cucumis melo L) PI 124111F against
P. cubensis is broad-spectrum but not isolate-specific [
38]. That resistance was controlled genetically by two partially dominant, complementary loci [
39]. Unlike other plant disease resistance genes, which confer an ability to resist infection by pathogens expressing corresponding avirulence genes, the resistance of PI 124111F to
P. cubensis is controlled by enhanced expression of the enzymatic resistance (eR) genes
At1 and
At2. These constitutively expressed genes encode the photorespiratory peroxisomal enzyme proteins glyoxylate aminotransferases. The low expression of
At1 and
At2 in susceptible melon lines is regulated mainly at the transcriptional level. This regulation is independent of infection with the pathogen. Transgenic melon plants overexpressing either of these eR genes displayed the enhanced activity of glyoxylate aminotransferases and remarkable resistance against
P. cubensis [
35,
40]. Our attempts to transfer
At1 and
At2 to cucumber did not succeed (Cohen,
unpublished data).
The results presented here corroborate with other studies in which multiple QTLs for resistance against
P. cubensis were identified in PI 197088 and PI 330628. (
Table 1). Wang et al. [
30] reported QTL mapping results for DM resistance with F2:3 families from the cross between DM-resistant inbred line PI 330628 (WI7120) and susceptible ‘9930’. Four QTLs,
dm2.1,
dm4.1,
dm5.1, and
dm6.1 were consistently and reliably detected across at least three of the four environments which together could explain 62–76% phenotypic variations. Among them,
dm4.1 and
dm5.1 were major effect QTL and
dm2.1 and
dm6.1 had moderate and minor effects, respectively.
Wang et al. [
28] used recombinant inbred lines from a cross between PI 197088 and the susceptible line ‘Coolgreen’. Phenotypic data on responses to natural DM infection were collected in three years and five locations from replicated field trials in North Carolina. The observed ratings followed a normal distribution that covered a large range of ratings at each environment and date. The interaction effects of genotype-by-location and genotype-by-year were significant at all ratings. QTL analysis identified 11 QTL for DM resistance harbored on chromosomes 1–6, accounting for more than 73.5% total phenotypic variance. Among the 11 DM resistance QTLs,
dm5.1,
dm5.2, and
dm5.3 were major effect contributing QTL whereas
dm1.1,
dm2.1, and
dm6.2 conferred susceptibility. The QTL
dm4.1 which had a moderate effect was likely the same as the major-effect QTL
dm4.1 detected in PI 330628 [
30]. Three DM QTLs
dm2.1,
dm5.2, and
dm6.1, were co-localized with powdery mildew (PM) QTLs,
pm2.1,
pm5.1, and
pm6.1, respectively, which was consistent with the observed linkage of PM and DM resistances in PI 197088.
Katz et al. [
29] reported on nine QTLs associated with resistance of PI 197088 against each of the seven isolates of
P. cubensis. They examined for two years the response of a segregating F2 family (PI-197088 × SMR-18,
n = 170) to seven isolates in growth chambers and the field. NGS (Next-Generation Sequencing) was performed for genotyping, and polymorphic SNPs were obtained from the same populations in both years. QTLs obtained for isolate 23C- resided on chromosomes 4 and 5; for isolate Pol.1- on chromosomes 1, 4, and 5; for isolate Pol.4- on chromosome 7; for isolate US-506- on chromosomes 1 and 2; for isolate 81C- on chromosomes 4 and 5; for isolate 88C- on chromosomes 3 and 6; for isolate 90C- on chromosomes 1, 4, and 6; for field isolate 2016, on chromosomes 3 and 5, and for field isolate 2017- on chromosomes 4 and 5. These authors concluded that the inheritance of resistance against DM in PI 197088 was isolate-dependent.
Tian et al. [
41] sequenced 14% of the genome of one isolate of
P. cubensis and identified 32 putative RXLR effector proteins and 29 secreted peptides with high similarity to RXLR effectors. They suggested that these effectors might play pivotal roles in pathogen fitness and pathogenicity. Sexual reproduction of the pathogen [
12,
42] may result in recombinant isolates which carry various combinations of effector proteins. It might, therefore, occur that isolates evolved in different parts of the world and therefore belong to different races, pathotypes, and mating types, each carries a unique set of effectors. Of this set of effectors, some might be secreted while others may not. Of the secreted effectors, some may recognize certain R genes in the host while others will not. This will make some host genotypes resistant to some genotypes of the pathogen.
The isolate-dependent inheritance of resistance of cucumber against P. cubensis may indicate that each isolate secretes a different battery of effectors that ignite a unique set of R genes in the host, thus making the inheritance of resistance isolate-dependent.