Introduction

The genus Cucumis, one of the 118 genera in the family Cucurbitaceae, is comprised of 32 species, of which cucumber (C. sativus L.) and muskmelon (C. melo L.) are the most important economically (Lebeda et al. 2006; Criswell 2008). There are about 45 different pathogens of Cucumis species. Because the fruits of cucumber and muskmelon are used primarily for fresh consumption, the application of pesticides is risky, as it may result in unacceptable levels of pesticide residue. Breeding for disease resistance is the preferred means of controling diseases in cucumber and muskmelon. Disease-resistance breeding efforts have been targeted predominantly against viral diseases, bacterial angular leaf spot (caused by Pseudomonas syringae pv. lachrymans), downy mildew (caused by Pseudoperonospora cubensis), and powdery mildew (caused by Sphaerotheca fuliginea and Erysiphe cichoracearum). Among those, downy mildew requires special attention, as it is the most destructive disease of the Cucumis species (Lebeda et al. 2006).

Downy mildew was observed to cause disease on cucumbers and melons as early as the 19th century, but it was not until the mid-1980s that it occurred on an economically significant scale (Colucci et al. 2006). A severe outbreak of downy mildew occurred on melons in France in 1984 (Pitrat and Blancard 1988). In 1985, the disease reached epidemic levels in cucumbers grown in Central-Eastern Europe. It became a serious problem in the USA starting in 2004 (Colucci et al. 2006). Nowadays, yearly downy mildew epidemics threaten cucumber production in up to 80 countries and muskmelon production in over 50 countries, causing significant economic losses (Lebeda and Urban 2004; Colucci et al. 2006).

The pathogen and pathogenesis

The causal organism of cucurbit downy mildew is P. cubensis (Berk. et Curt.) Rostov., an Oomycete that belongs to the kingdom Chromista, phylum Heterokontophyta, class Oomycetes, order Peronosporales, family Peronosporaceae, genus Pseudoperonospora (Göker et al. 2007; Kirk et al. 2008; Bouwmeester et al. 2009). P. cubensis is an obligate parasite that overwinters in warm regions. The dispersal of sporangia is aided by wind and rain (Lebeda and Urban 2004). The first symptoms appear on the adaxial surface of older leaves as light-green or yellow, angular in shape, pinpoint lesions that are restricted by the veins. As the disease progresses, the lesions become first chlorotic and later necrotic, causing a decrease in photosynthesis, and, ultimately, resulting in the desiccation and death of the infected leaf. Light greyish-blue sporangiophores may be observed on the abaxial surface of the infected leaf, especially during periods of high relative humidity. Severely infected plants produce small, misshapen fruit that may not mature, and, in the case of melons, the fruit has low sugar content (Pitrat et al. 1998).

The sporangiophores of P. cubensis are produced by intercellular mycelium and are extruded onto the leaf surface via stomata. At first, they are poorly branched and, later, they become dichotomous. Each branch of a sporangiophore terminates with a sporangium (zoosporangium), in which zoospores are produced. Zoospores are the primary source of infection. They are grey, yellowish, or purple in color, ellipsoidal in shape, and measure 20–40 μm × 14–25 μm. The zoospores are mobile by means of two flagella of unequal lengths. The zoospores swim in a film of water and enter leaf tissue via stomata to establish infection (Bouwmeester et al. 2009). The disease symptoms become visible 4–12 days after infection, depending on the light intensity, temperature, humidity, and inoculum concentration. Under optimal conditions (light ∼50 W × m−2, temperature ∼20°C day and ∼15°C at night, and at least 6 h of 100% relative humidity per day), symptoms appear as early as four days after inoculation (Lange et al. 1989).

Pathotypes

Initial studies of variability among isolates of P. cubensis were performed by Thomas et al. (1987), who reported the existence of five pathotypes among isolates collected in Israel, Japan, and the USA. These studies were based on compatible reactions with species of Cucumis, Citrullus, and Cucurbita. Cohen et al. (2003) reported a sixth pathotype, isolated in Israel. Shetty et al. (2002) found that European and North American pathotypes were more closely related, and that Asian pathotypes were more distinct. Lebeda (1990) observed that the geographic origin and year of isolation played a role in the pathogenicity of P. cubensis. In later studies, Lebeda and Gadasová (2002) used differential hosts representing five genera from among the Cucurbitaceae and found 13 different pathotypes in Central Europe. Lebeda and Widrlechner (2003) introduced a set of 12 different hosts (representing different genera, species, and cultivars) to identify pathotypes in P. cubensis. Such broad pathogen variability is typical for the Oomycetes and is also found in Phytophthora infestans and Bremia lactucae (Bouwmeester et al. 2009).

Recently, the genomes of three Oomycetes, P. ramorum, P. sojae (Tyler et al. 2006), and P. infestans (Haas et al. 2009), were sequenced. In P. cubensis, the 709-bp rDNA-ITS region was sequenced, showing that it can be divided into three distinct parts: 141 bp of rDNA-ITS1, 406 bp of rDNA-ITS2, with GC contents of 41% and 46%, respectively, and the relatively conserved 5.8 S coding region (Wang et al. 2008). The non-coding ITS regions (ITS1 and ITS2) are often used in phylogenetic studies of Oomycetes because of their relatively high sequence variability and because of the availability of polymerase chain reaction (PCR) primers. The rDNA-ITS sequence is almost the same in different isolates of P. cubensis, which makes it suitable for use as a molecular marker for species identification (Wang et al. 2008).

Resistance to downy mildew

Disease resistance can be broadly defined as the host’s ability to suppress or inhibit a pathogen’s activity (Ton et al. 2006). The most widely observed type of resistance is non-host resistance, which occurs when the plant is not infected because of a pathogen’s inability to establish infection and to cause disease following the initial contact. This kind of resistance is durable and is difficult to overcome by pathogens (Ton et al. 2006; Heath 2009). The second type of resistance is determined by resistance genes. Based on the genetic mechanism involved, this type of resistance can be divided into two categories: race-specific resistance, which displays a gene-for-gene relationship between the host and the pathogen, and race-nonspecific resistance, which usually depends on multiple genes of small individual effect (Ton et al. 2006). The third type of resistance is the induced systemic resistance, ISR (often referred to as systemic acquired resistance, SAR). The ISR is a phenomenon in which a biotic or an abiotic stimulus (e.g., a pathogen infection, activation of plant associated microorganisms, or the application of chemicals) causes an elevation of plant resistance to a specific pathogen, or a group of pathogens. This elevated response is initiated by the production of salicylic or jasmonic acids, or ethylene (Kuć 2006; Tuzun 2006).

Studies of the genetic bases of resistance to P. cubensis resulted in the identification of both single-gene-mediated and polygenically inherited resistance (van Vliet and Meysing 1974; Epinat and Pitrat 1994a). The identification of two types of resistance is likely due to the use of different plant materials. It also could be due to the use of different parameters to measure resistance when studying the host–pathogen relationship, such as the suppression of sporulation, or presence or absence of chlorosis/necrosis on the infected leaves. Thomas et al. (1987) measured sporulation as a means of identifying P. cubensis pathotypes. Sporulation was also used to identify resistance genes (van Vliet and Meysing 1974) and to evaluate host resistance (Lebeda and Prašil 1994; Lebeda et al. 2007). The percentage of necrotic leaf spots was used to characterize plant material by Neykov and Dobrev (1987) and to identify genes for resistance by Doruchowski and Łąkowska-Ryk (1992).

Other researchers combined these two methods and described resistance as the presence of small, chlorotic lesions with a low level of sporulation (Angelov and Krasteva 2000; Petrov et al. 2000). An investigation by Criswell et al. (2008) showed that the occurrence of necrosis and chlorosis were highly correlated, indicating that these two phenotypes may be a manifestation of the same genetic mechanism. In the same study, abundant sporulation was only moderately correlated with chlorosis and necrosis and, therefore, did not appear to be controlled by the same genetic mechanism. This observation did not hold true when a different pathotype was used, and correlations obtained by measuring chlorosis, necrosis, and sporulation indicated that resistance to that pathotype was controlled by the same genetic mechanism (Criswell et al. 2008). In addition, other researchers found anatomical and cytological factors that contribute to the level of host resistance. Habdas et al. (1996) found that the epidermis in all highly resistant cucumber accessions was covered with a thick cuticle that also partly covered the stomata.

Qualitative (vertical) resistance

The mode of inheritance of resistance to downy mildew in cucumber has been studied for the past 70 years (Criswell 2008). However, it was not until 1963 that it was elucidated for the first time by Shimizu et al. (1963), who reported three recessive resistance genes in the cucumber cultivar (cv.) ‘Aojihai’. Three recessive resistance genes were also reported in studies using different plant material by Doruchowski and Łąkowska-Ryk (1992), who designated them as dm1, dm2, and dm3. On the other hand, van Vliet and Meysing (1974) reported that resistance in cv. ‘Poinsett’ was determined by one recessive gene (dm). Cv. ‘Poinsett’ was selected from 197087 plant introduction of Indian origin. The single-gene resistance was later confirmed by Fanourakis and Simon (1987).

In C. melo, two complementary, incompletely dominant genes (Pc-1 and Pc-2) were described by Thomas et al. (1988) in the resistant line MR-1. However, using different plant material, Epinat and Pitrat (1989) reported that resistance in melon was controlled by a single dominant gene, designated Pc-3. The presence of a single dominant gene for resistance was also indicated in studies by Angelov and Krasteva (2000), who stated that, in a susceptible line, the expression of this gene was masked by an epistatic gene.

The research cited above indicates major differences in the understanding of the genetics of resistance to downy mildew in these two host species. In cucumber, resistance to downy mildew is likely to be determined by a recessive gene or genes, whereas in melon, it is likely to be determined by a dominant gene(s). Moreover, in cucumber, there is a correlation between the presence of genes for resistance to downy mildew (dm) and to powdery mildew (pm) (van Vliet and Meysing 1977; Pivovarov 1988). Linkage analysis studies indicate that the dm and pm genes are located either at the same locus or in closely linked loci (Fanourakis 1984). In contrast, resistance to these two diseases in muskmelon is inherited independently (Thomas et al. 1988).

Several studies have identified molecular markers that appear to be linked to cucumber resistance to downy mildew. Horejsi et al. (1999) found four polymorphic random amplified polymorphic DNA (RAPD) markers linked to the dm gene and converted them to sequence characterized amplified region (SCAR) markers. The same authors subsequently identified five additional markers (Horejsi et al. 2000). Recently, Śmiech et al. (2008) described eight RAPD markers that produced polymorphic amplicons in either susceptible or resistant lines of cucumber. In an earlier study, Danin-Poleg et al. (2000) developed 34 simple sequence repeat (SSR) markers for Cucumis and mapped several of them in both cucumber and muskmelon, demonstrating cross-homology between these two species and the potential utility of these markers for comparative mapping between melon and cucumber.

In 2009, the cucumber genome was sequenced independently by two research teams. A team led by Chinese researchers worked on ‘Chinese Long’ inbred line 9930 and used a combination of the traditional Sanger and next-generation Illumina GA sequencing methods (Huang et al. 2009). A team led by Polish scientists worked on ‘pickling’ inbred line B10 and used the 454 Titanium sequencing method (Wóycicki et al. 2009). The high co-linearity between the cucumber and melon genomes (Huang et al. 2009) has the potential to enable the sequence derived from cucumber to serve as a model system for other species in the Cucumis genus. Identification of map location of the molecular markers linked to downy mildew resistance will help in the identification of the gene sequence, and, in turn, may help to elucidate the mechanism of resistance and/or metabolic pathways activated by these genes. Development of high-resolution mapping families, with highly accurate phenotypic data on disease response, is the most important resource that will allow the application of genomic sequences in breeding for increased resistance.

Quantitative (horizontal) resistance

Resistant phenotypes that do not segregate into discrete categories of resistance are assumed to be under the control of multiple genes for resistance. In general, quantitative resistance that is under polygenic control is more durable than resistance conferred by a single dominant gene (Kelly and Vallejo 2006). However, quantitative inheritance of resistance as measured by disease severity is often characterized by low heritability and is under significant environmental influence (Olczak-Woltman et al. 2009). In fact, in the case of polygenic resistance to cucumber downy mildew, broad-sense heritability was found to be relatively low: 8% and 17% for two independent populations analyzed, respectively (Horejsi et al. 2000).

In cucumber, McFerson (1978) noticed that, in the breeding line GY14A, which was also developed using line PI 197097, resistance to downy mildew was polygenic. In muskmelon, the genetics of quantitative resistance to downy mildew was investigated by Epinat and Pitrat (1994a) with the use of a diallel cross with heterozygotic progenies of MR-1 and other lines of similar origin. The results indicated the presence of multiple alleles, a possible presence of double epistasis, and continuous variation for the ranking of resistance level varying from high resistance to high susceptibility. The authors concluded that resistance seemed to be controlled by numerous additive loci. In another study, the same authors found that 84–99% of variation in the resistance was due to additive gene action, while dominance or epistasis accounted only for 3% and 8% of the observed variation, respectively (Epinat and Pitrat 1994b). The quantitative resistance to downy mildew in muskmelon is positively correlated with increased peroxidase activity. Reuveni et al. (1990) observed that individuals with high peroxidase activity were resistant to P. cubensis when inoculated.

Regions that encode genes of partial effect are often referred to as quantitative trait loci (QTL). Some QTLs encode genes for race-specific resistance, whereas others for race-nonspecific resistance (Young 1996). The application of QTL analysis can be highly valuable in breeding for resistance, as it reveals the location of loci controling disease resistance and, ultimately, may allow marker-assisted selection for resistance without confounding effects of environmental factors (Kelly and Vallejo 2006). The first QTLs for resistance to downy mildew in muskmelon were identified by Perchepied et al. (2005), who mapped 11 QTL markers on six linkage groups on an unsaturated linkage map of this species.

Sources of resistance

Plant sources with the highest levels of resistance to downy mildew originate from the Far East. Highly resistant accessions of cucumber and muskmelon were found in India, China, and Japan (Pitrat et al. 1989; Staub et al. 1989).

The most resistant muskmelon lines, PI 124111 (as well as MR-1 and PI 124111 F, which were selected from PI 124111) and PI 124112, were reported to be resistant to all pathotypes of P. cubensis and to Alternaria sp., F. oxysporum f. sp. melonis, and S. fuliginea (Thomas et al. 1988; Balass et al. 1992; Pitrat et al. 1996). These accessions are from India; they and lines selected from them are vegetable melons (subsp. agrestis var. momordica). Their suite of inferior fruit quality traits poses a challenge for the development of dessert melons with resistance to downy mildew.

Resistance in the accession PI 124111 F is manifested by the induction of a hypersensitive response that leads to the deposition of callose along the inner surface of the host cell walls, resulting in an encasement of the Oomycete haustoria within callose-like deposits, and massive lignification of the host cell (Cohen et al. 1989). Balass et al. (1992) found that the resistant line PI 124111 F contained a 45-kD protein (P45) that was not present in the susceptible genotypes. Taler et al. (2004) conducted partial sequencing of the P45 protein and showed that it was a peroxisomal aminotransferase enzyme. Genes that encode this type of protein enhance the level of enzymatic activity of the host and, therefore, are referred to as enzymatic resistance (eR) genes. Transformation of a susceptible C. melo line with a gene construct carrying eR genes (At1 and At2) resulted in a line highly resistant to P. cubensis (Taler et al. 2004). Homologs of the At genes, which could be candidates for downy mildew resistance genes, have recently been identified in the cucumber genome (Huang et al. 2009).

Because high levels of resistance were generally not found in cucumber, Lebeda et al. (1996) made an attempt to transfer resistance from the melon line MR-1 to cucumber using a conventional crossing technique, but the interspecific cross was unsuccessful. The potential sources of resistance in cucumber are completely homozygous double haploid lines (Sztangret-Wiśniewska et al. 2006). Among other Cucumis species, C. hystrix displays a high level of resistance to P. cubensis (Chen et al. 2004). DNA homology studies indicate that C. melo, C. sativus, and C. hystrix belong to the same species cluster, that C. hystrix is either a progenitor species of C. sativus or that they share at least a common ancestral lineage (Chung et al. 2006). Recently, genetically stable introgression lines with resistance to downy mildew have been obtained from an interspecific hybridization of C. sativus and C. hystrix (Zhou et al. 2009).

Conclusions

Downy mildew is one of the most destructive diseases of cucumber and muskmelon. The control of this disease with fungicides, although necessary, often does not give satisfactory results. Therefore, breeding for resistance seems to be the most effective way to control this disease. It is easier to introduce resistance genes into the elite breeding lines, inbreds, and cultivars when the genetic basis of the resistance is understood. There is no doubt that the recent sequencing of the cucumber genome will provide the opportunity to acquire, in a short time frame, reliable and accurate information on resistance genes to downy mildew for the Cucurbitaceae, their location, construction, and function. The sequenced genome enables identification of the gene sequences, which, in turn, will enable the elucidation of the resistance mechanism and the metabolic pathways activated by these genes. This will make possible the determination of the conditions in which the genes are activated and of the metabolic pathways that are involved in the expression of resistance. However, because the genetic work on cucumber and melon that has been done so far is fragmentary, stronger co-ordination among various research and breeding centers is recommended to ensure optimal efficiency in future progress.