Introduction

Lettuce (Lactuca sativa) is one of the most widely consumed vegetable crops worldwide. It has an annual production of more than 3.6 billion tonnes with a value of more than $2.4 billion in the U.S. (USDA-NASS 2014). Downy mildew (DM) caused by the oomycete pathogen Bremia lactucae is the most important disease of lettuce that decreases quality of the marketable portion of the crop. The impact of this disease is often accentuated by postharvest losses that occur during transit and storage. DM can also increase the consequences of microbial contamination by human enteric pathogens (Simko et al. 2015a).

Strategies for control of DM include the combined use of resistant cultivars and fungicides as well as agronomic practices that reduce foliar humidity. The use of fungicides is constrained by high costs and the development of fungicide-resistant strains (Crute 1987; Schettini et al. 1991). Moreover, increasing restrictive regulations aimed at reducing pesticide applications are coming into force; in Europe several chemicals that are effective against B. lactucae will be redrawn from the market. The deployment of cultivars carrying dominant resistance genes (Dm genes) is the most effective method for controlling DM; however, pathogen variability has led to the rapid defeat of individual Dm genes (e.g. Ilott et al. 1987). Consequently, the search for new sources resistance to B. lactucae has been a continuous, long-term priority of lettuce breeding programs (Crute 1992; Lebeda et al. 2002, 2014). Over 50 genes for resistance have been reported so far and genetically characterized to varying extents (see below for references). In addition, many other sources of resistance have been identified in germplasm screens but have yet to be characterized genetically (e.g. Farrara and Michelmore 1987; Bonnier et al. 1994; Lebeda and Zinkernagel 2003; Beharav et al. 2006). As more resistance genes are characterized from these and other sources, it is likely that several hundred genes with efficacy against B. lactucae will be identified that will require a coordinated, rational nomination process.

This review compiles the knowledge of genetic resistance against B. lactucae, summarizing what is known of genes for resistance to B. lactucae that has been generated over more than 50 years of lettuce genetics and breeding (Supplemental Fig. 1). Resistance has been reported by multiple researchers leading to duplications in nominations as well as gaps in sequence. The extent of genetic characterization has varied. Genes characterized as single Mendelian loci were designated Dm genes, while those that were less well characterized genetically were often but not always termed R-factors. For this review we have rationalized the nomenclature of all of the Dm genes and R-factors reported so far and provide the foundation for future designation and use of Dm genes.

Historical overview of breeding for resistance to B. lactucae

Breeding for resistance to B. lactucae in cultivated lettuce has been carried out since the beginning of the last century. Initial breeding efforts utilized resistance identified in old lettuce cultivars (L. sativa). French traditional cultivars Gotte à Graine Blanche de Loos and Rosée Printanière were the first sources of resistance (referenced in Crute 1992). Subsequently, resistance was identified in several other cultivars such as Meikoningen, May Queen, Gotte à Forcer à Graine Noire, Bourguignonne Grosse Blonde d’Hiver and Blonde Lente à Monter (Jagger and Chandler 1933; Schultz and Röder 1938; Jagger and Whitaker 1940; Ogilvie 1944; Rodenburg et al. 1960). Later breeding efforts accessed resistance from wild Lactuca species. One of the first inter-specific crosses for this purpose was between L. sativa cv. Imperial D and DM–resistant L. serriola PI104854 (Whitaker et al. 1958). This resistance was used in breeding programs in California during the 50’s, that led to the crisphead cv. Calmar (cv. Great Lakes × USDA 45325; Welch et al. 1965) and subsequently cv. Salinas (cv. Calmar × Vanguard 75; Ryder 1979a, b) that is in the pedigree of many current crisphead cultivars (Mikel 2007, 2013). Contemporary breeding efforts are focused on introgression of new genes from wild species. L. serriola, the likely progenitor of cultivated lettuce, and, to a lesser extent, L. saligna have been used as donors of resistance genes (Jagger and Whitaker 1940; Crute and Johnson 1976a, b; Lebeda et al. 1980; Bonnier et al. 1994; Witsenboer et al. 1995; Maisonneuve et al. 1999; Jeuken and Lindhout 2002, Michelmore and Ochoa 2006, 2008; Mc Hale et al. 2009, Zhang et al. 2009). L. virosa also possesses race-specific resistance to B. lactucae (Lebeda and Boukema 1991; Lebeda et al. 2002, Maisonneuve et al. 1999), but its use in breeding programs has been restricted by infertility barriers between L. virosa and L. sativa (Maisonneuve 2003). Transfer of resistance from L. virosa to L. sativa has occasionally been enabled by embryo rescue (Maisonneuve 1987; Maisonneuve et al. 1995). Other wild species such as L. aculeata are potential sources of resistance that have yet to be used as donors in breeding programs (Jemelková et al. 2015). Currently, introgression of recently identified resistance from L. serriola, L. saligna and L. virosa through repeated backcrosses to L. sativa is being carried out by multiple public and commercial breeding programs resulting in a large increase in the number of resistance genes being deployed (e.g. Michelmore and Ochoa 2008; Michelmore et al. 2013a).

Genetics of resistance and nomenclature of resistance genes

Extensive classical genetic studies have been carried out for the resistance of lettuce to B. lactucae. The gene-for-gene interaction between L. sativa and B. lactucae was first elaborated by Crute and Johnson (1976a, b). This interaction is now one of the best characterized gene-for-gene plant-pathogen relationships (Farrara et al. 1987; Hulbert and Michelmore 1985; Ilott et al. 1987, 1989). At least 28 Dm genes are currently known that provide high levels of resistance against specific isolates of B. lactucae (Table 1). Although most Dm genes confer complete resistance, some Dm genes show incomplete resistance that varies depending on the environmental conditions and the isolate of B. lactucae. Low temperature decreases the effectiveness of several Dm genes (Judelson and Michelmore 1992). Cultivars carrying Dm6 and Dm7 may exhibit partial resistance associated with macroscopically visible hypersensitive necrosis (Crute and Norwood 1978). Additionally, gene dosage affects the resistance phenotype of some Dm genes, for example Dm17 under high pathogen pressure (Maisonneuve et al. 1994). Some lettuce cultivars such as Iceberg and Grand Rapids exhibit resistance at the adult plant stage in the field that cannot be attributed to Dm genes (Milbrath 1923; Verhoeff 1960; Crute and Norwood 1981). This quantitative resistance phenotype (Norwood and Crute 1985; Grube and Ochoa 2005) has so far been shown to be inherited polygenically (Simko et al. 2013); transgressive segregation resulting in elevated resistance has been observed in progeny from cvs. Grand Rapids × Iceberg (Grube and Ochoa 2005; Simko et al. 2013). Some cultivars such as Green towers and Cobham Green have no known Dm genes and are used as susceptible lines to grow B. lactucae isolates; however, the European isolate Serr84/99 is avirulent on Cobham Green and there is some evidence suggesting polygenic resistance in this cultivar (Maisonneuve 2011b). Most accessions of L. saligna are completely resistant to isolates of B. lactucae derived from L. sativa; therefore, this species has been proposed to be a non-host for downy mildew (Bonnier et al. 1992; Petrželová et al. 2011; van Treuren et al. 2013); this complete resistance is in part determined by several quantitative trait loci (QTLs) operating at different developmental stages (Jeuken and Lindhout 2002; Jeuken et al. 2008, Zhang et al. 2009; Den Boer et al. 2013). However, the mechanism of resistance in L. saligna is still unresolved. Stacking of these QTLs in 10 pairwise combinations hardly showed an increase in the level of resistance suggesting that epistatic interactions play a role (De Boer et al. 2014; Den Boer 2014).

Table 1 Compilation of information available for genes and genetic factors for resistance to Bremia lactucae in lettuce
Full size table

The genetic studies have resulted in the identification of numerous Dm genes and R-factors. In order to remove duplications in nomenclature and evaluate the genetic evidence for Dm genes and R-factors, we reviewed all the primary literature reporting resistance to B. lactucae in lettuce. When possible, this involved multiple rounds of consultation with the authors and with the lettuce genetics and breeding community at large. This resulted in the classification of 28 Dm genes and 23 R-factors that provide resistance to specific isolates of B. lactucae (Table 1). Resistance was assigned a Dm designation when supported by genetic evidence and mapped to a single locus in the lettuce genome. Resistances were designated as R-factors, when the resistance specificity as determined by reactions to isolates of B. lactucae indicated presence of new resistances genes; however, such resistances had not (yet) been shown to be monogenic or mapped. Over eighty percent of the Dm genes and R-factors were identified in wild Lactuca species collected in Europe (Table 1). Most of these resistances have been introgressed into cultivars of L. sativa as part of breeding programs in Europe and the USA. Parallel research and breeding efforts resulted in several duplicate designations for resistance from different sources. Seven resistances were therefore renamed to remove duplications and to fill in gaps in the sequence of designations; resistances identified in the same study were kept adjacent to the extent possible (Table 1). Fifteen major QTLs for resistance to B. lactucae have so far been identified (Table 2). The QTLs were renamed to be consistent with the convention for describing QTLs in lettuce, in which a QTL is prefixed with ‘q’ followed by capital letters indicating resistance to the disease (DMR in this case) and two numbers indicating the chromosomal linkage group followed by the number of the QTL on that linkage group.

Table 2 Compilation of information available for quantitative trait loci for resistance to Bremia lactucae in lettuce identified at the seedling and adult plant levels
Full size table

Some lettuce cultivars possess the same resistance specificity, despite the fact that their resistances were introgressed from different sources, sometimes even from different Lactuca species. Linkage analysis of Dm5 and Dm8 and parallel genetics of virulence in B. lactucae demonstrated that both resistances are controlled by the same gene (Norwood and Crute 1984; Hulbert and Michelmore 1985). These resistances were identified from different accessions of L. serriola collected from Turkey and Russia (Jagger and Whitaker 1940; Leeper et al. 1963; Lebeda et al. 1980) and have different molecular haplotypes (Witsenboer et al. 1995). Similarly, Dm38 and R24 cosegregate and share specificities (J. Schut, unpublished). Dm38 and R24 were introgressed from L. serriola sources from Czechoslovakia and Hungary, respectively (Bonnier et al. 1994; Maisonneuve et al. 1999). Similarly, Dm18 and R32 cosegregate and have the same specificity; both resistances were rendered ineffective simultaneously by a change in virulence in B. lactucae (Petrželová et al. 2013). Dm18 originated from L. serriola LS17, while R32 originated from L. saligna LJ81632, suggesting either conservation since the diversification of these Lactuca species or independent convergent evolution of these genes. Dm36 in cv. Ninja has been reported to be identical to Dm37 in cv. Discovery based on reactions to European isolates and had been named Rsal-1 (Maisonneuve 2007, 2011a); however, this conclusion is not supported by the reactions of Ninja and Discovery to Californian isolates (C. Tsuchida and L. Parra, unpublished). Both Dm36 and Dm37 were introgressed from accessions of L. saligna from Israel (B. Moreau, pers. comm.), but the identity of the donor for Dm36 is uncertain and both resistances may have originated from the same source. Resolution of the relationship of Dm36 to Dm37 awaits analysis at the sequence level.

The genetic location is known for 28 Dm genes. As in other plants, resistance genes are clustered in the lettuce genome. The known Dm genes are located in major resistance clusters (MRCs) along with genes determining resistance to other diseases (Table 1; Hulbert and Michelmore 1985; Farrara et al. 1987; Bonnier et al. 1994; Mc Hale et al. 2009; Christopoulou et al. 2015a, b). MRC1 contains Dm5/8, Dm10, Dm17, Dm25, Dm36, Dm37, Dm43, Dm45, as well as Tu and Mo2 for resistance to Turnip Mosaic Virus (TMV) and Lettuce Mosaic Virus (LMV) respectively, and qFUS1.1 and qFUS1.2 for resistance to wilt caused by Fusarium oxysporum f.sp. lactucae. MRC2 includes Dm1, Dm2, Dm3, Dm6, Dm14, Dm15, Dm16, Dm18, Dm50 and qDMR2.2, along with Tvr for resistance to Tomato Bushy Stunt Virus (TBSV), Ra for root aphid resistance, and qANT1 for resistance to anthracnose. MRC4 contains Dm4, Dm7, Dm11, Dm24, Dm38, Dm44 and Dm48 as well as qFUS4.1 for resistance to Fusarium wilt. MRC9A contains qDMR9.1, qDMR9.2 and qDMR9.3, and qVERT9.1 for resistance to wilt caused by Verticillium dahliae (Christopoulou et al. 2015a, b). Dm39 was initially mapped at a locus similar to MRC9A based on analysis of an interspecific F2 population derived from L. saligna CGN05271 × L. sativa cv. Olof (Jeuken and Lindhout 2002); however, this resistance phenotype turned out to be due to an interaction between a L. saligna-allele of Rin4 at MRC9A and the L. sativa-allele of Dm39 at MRC8C (Jeuken et al. 2009). In Arabidopsis Rin4 is a negative regulator of basal defense and known to be the target for three effectors of Pseudomonas syringae and guarded by two R-genes (Axtell and Staskawicz 2003; Mackey et al. 2002).

There has been only limited characterization of specific Dm genes at the molecular level. Each phenotypic MRC spans multiple megabases in the lettuce genome and encompasses complex clusters of genes encoding nucleotide binding-leucine rich repeat, receptor-like proteins (NLRs). Sequence analysis of 385 NLR-encoding genes in the reference lettuce genome identified 25 multigene families and 17 singletons of resistance gene candidates (RGCs) that could be classified as TNL- or CNL-encoding types, depending on the presence or absence of Toll interleukin 1 receptor domain (TIR) at the N-terminus Christopoulou et al. (2015b). Functional analysis of NLR-encoding genes that co-segregated with Dm phenotypes using RNAi demonstrated four NLR-encoding multigene families that were required for 13 Dm phenotypes (Table 1). Only two individual Dm genes have been cloned so far. The map-based cloning of Dm3, encoding a CNL type of NLR, was confirmed by transgenic complementation (Shen et al. 2002). Dm7 was identified on the basis of multiple EMS-induced mutations (Christopoulou et al. 2015a).

Implications for control of downy mildew

This review provides the foundation for naming Dm genes in future. Genetic dissection of R-factors into their Mendelian components will reveal the number and genomic position of the underlying Dm genes. Genetic dissection of QTLs will also reveal candidate genes, although they may not be of the NLR type. Germplasm screens will continue to identify many new sources of resistance that are likely to be conferred by new Dm genes. The International Bremia Evaluation Board (IBEB; http://www.worldseed.org/isf/ibeb.html) should be consulted in order to coordinate the naming of such new Dm genes. IBEB currently consists of representatives from Europe and the US who are knowledgeable of efforts to control DM in lettuce and genetics of resistance to B. lactucae. IBEB should therefore serve in an advisory capacity to avoid duplications and ensure sequential designation.

Genomic analyses show that the MRCs are complex clusters of multiple NLRs. One or more genes could be conferring a resistance phenotype depending on which isolate is used to detect it. Simple segregation analysis of the host alone does not reveal how many genes are effective at a single Mendelian locus. This can be revealed by segregation analysis of the virulence phenotype in B. lactucae; however, this is a slow and labor intensive process. The potential presence of multiple effective genes at a single locus has consequences; recombination at a MRC during backcross programs may result in loss of some Dm genes and parallel introgressions from the same source of resistance may result in different subsets of Dm genes being retained. There is some evidence for this occurring with Dm18 (Wroblewski et al. 2007). Detailed genetic analysis of MRCs may result in the identification of multiple Dm genes and require revision/splitting of current Dm designations.

Genome sequencing and assembly has revealed that all plant genomes contain many, usually hundreds, of NLR-encoding genes. Therefore, all plants have many resistance genes; even if active specificities have yet to be recognized. Consequently, avirulence factors recognized by additional Dm genes in the cultivars described in this paper may be identified in the future, particularly in isolates from L. serriola. Although these Dm genes are effective in limiting migration of isolates from L. serriola onto L. sativa, they are of marginal relevance to control of DM in cultivated lettuce; however, they will be relevant when introgression of a new resistance specificity from a wild species inadvertently replaces such Dm genes, and consequently introduces susceptibility to isolates from the wild species.

Resistance to DM can also be mediated by recessive genes. DMR6 in Arabidopsis is necessary for susceptibility to downy mildew; a recessive dmr6 allele derived by mutation results in resistance against Hyaloperonospora arabidopsidis (Van Damme et al. 2005, 2008). A DMR6 ortholog has been identified in lettuce, where its over-expression increases host susceptibility to B. lactucae (Stassen et al. 2012; Zeilmaker 2012). Natural variation in DMR6 that confers resistance to B. lactucae has yet to be identified.

The plethora of known resistance genes and those now in multiple public and commercial breeding pipelines provides the opportunity for rational deployment of resistance genes (Dm and QTLs; Michelmore et al. 2013b). Pyramids of resistance genes based the nomenclature proposed here that are effective against the diversity of B. lactucae should be generated so as to maximize the evolutionary hurdle required for B. lactucae to become virulent. Pyramids of dissimilar sets of resistance genes should be deployed in the different lettuce types so as to provide heterogeneity in the selection pressure acting on the pathogen population. This should result in more durable resistance to DM.