Plants rely on a complex innate immune system to ward off pathogen attacks. Perception of pathogen-associated molecular patterns (PAMPs) by plant cell surface distributed receptors elicits the first line of defense-PAMP triggered immunity (PTI). To succeed in infection pathogens secrete effectors to the apoplast or inside plant cells to suppress PTI and reprogram host physiology in favor of disease, a process often referred to as effector-triggered susceptibility (ETS) (Jones and Dangl, 2006; Toruno et al., 2016). Plants have evolved a second line of defense, also known as effector-triggered immunity (ETI), by encoding diverse resistance (R) proteins to detect such effectors and activating high levels of immune responses such as hypersensitive cell death (Jones and Dangl, 2006). R gene-mediated resistance is typically qualitative and effective to restrict pathogen strains carrying specific effectors (avirulence proteins, Avr); in absence of a successful R-Avr interaction, however, quantitative disease resistance reduces the growth and spread of virulent pathogens, thereby protecting plants from further infection (Dangl and Jones, 2001; Poland et al., 2009).

The phenomena of quantitative disease resistance, also referred to as partial disease resistance or basal resistance, has been widely observed in many plant pathosystems, in which the distribution of resistance values in a genetically segregating population is typically continuous and does not follow a Mendelian pattern of inheritance (Poland et al., 2009). Genetic dissection of quantitative disease resistance in many crop plants with quantitative trait locus (QTL) mapping has uncovered genetic loci conferring resistance to a wide range of diseases (St. Clair, 2010). Recently, a number of genes mediating quantitative disease resistance have been successfully cloned from major crop plants including wheat (Fu et al., 2009; Krattinger et al., 2009), rice (Fukuoka et al., 2009; Hayashi et al., 2010), maize (Hurni et al., 2015; Zuo et al., 2015) and soybean (Cook et al., 2012). Notably, many of these genes provide durable resistance in the field and encode proteins distinct to typical R proteins, indicating that quantitative disease resistance and R-gene mediated resistance may recruit different components to deploy.

The model plant Arabidopsis has been used intensively to investigate the genetic and molecular basis of disease resistance to virulent pathogens with diverse lifestyles. Major progress has been achieved on how the signal molecules including salicylic acid (SA), jasmonic acid, and ethylene orchestrate plant defenses (Pieterse et al., 2012). In addition, multiple pattern recognition receptors have been uncovered in the plant to elicit PTI in response to various PAMPs (Gómez-Gómez and Boller, 2000; Zipfel et al., 2006; Miya et al., 2007; Albert et al., 2015). Nevertheless, most of these findings are based on experiments with frequently used Arabidopsis accessions such as Col-0. Several studies have demonstrated wide quantitative variation in disease resistance across Arabidopsis accessions collected from wild populations at different geographical locations, and many accessions are much less susceptible than Col-0 to pathogen infection, indicating the presence of important unknown mechanisms mediating disease resistance in these accessions (Kover and Schaal, 2002; Fan et al., 2008). Therefore, further investigation on the basis of this variation will not only broaden our knowledge on the complex genetic architectures of plant defense networks but also may help us design crops with durable resistance to a broad range of phytopathogens.

The bacterial effector HopW1-1, also known as HopPmaA, has been found in several strains of Pseudomonas syringae pathovars (pv) maculicola and phaseolicola, but is absent from the strain P. syringae pv tomato DC3000 (DC3000) (Guttman et al., 2002; Lindeberg et al., 2006). When delivered into plant cells by P. syringae, HopW1-1 targets the actin cytoskeleton and inhibits endocytosis, thereby promoting bacterial virulence in susceptible Arabidopsis plants (Kang et al., 2014). However, in some accessions such as Ws, hopW1-1 restricts bacterial growth and elicits resistance responses including the accumulation of SA (Lee et al., 2008). Three HopW1-1-interacting (WIN) proteins have been identified from Arabidopsis. Overexpression or downregulation of WIN proteins has diverse effects on plant disease resistance induced by HopW1-1 or other bacterial effectors, indicating that WIN proteins may have different roles in modulating bacteria-plant interactions (Lee et al., 2008). Variation in resistance among Arabidopsis accessions to bacterial strains carrying hopW1-1 has been revealed in several studies (Lee et al., 2008; Rant et al., 2013), but the genetic region underlying plant resistance specific to HopW1-1 remains to be revealed.

In this study, Arabidopsis accessions Aa-0 and Gie-0 were found to support different levels of bacterial growth compared with Col-0 plants. Two populations of recombinant inbred lines (RILs) derived from Aa-0 × Col-0 and Gie-0 × Col-0 crosses were constructed and used for genetic dissection of the variance in disease resistance to bacterial strain P. syringae pv maculicola ES4326 (ES4326) and DC3000, respectively. QTL analyses of the levels of bacterial growth continuously distributed across these two RILs showed that genetic variation in Aa-0 and Gie-0 at a locus on chromosome III, designated as Qpm3.1, preferentially restricted the growth of bacterial strain ES4326. Further screen for the genetic factor in ES4326 that limited bacterial growth on these Arabidopsis accessions successfully identified the effector gene hopW1-1. These findings highlight the role of pathogen effectors in quantitative natural variation in plant resistance to bacterial infection. The possible mechanism is also discussed on how Qpm3.1-hopW1-1 interaction led to enhanced resistance.

In this study, we identified several QTLs mediating quantitative disease resistance to the lux-tagged DC3000 and ES4326 strains, respectively (Figure 3 and Table 1). The patterns of QTL distribution apparently differed in association with the strains used for inoculation. Among Qpm loci, the prominent effect of Qpm3.1 was detected from both Col × Aa and Col × Gie RILs challenged with ES4326 but was absent when inoculated with DC3000. This locus also coincides with the major QTL QRps.JIC-3.1 that has been identified during infection of Col × Ler RILs with the strain ES4326 (Rant et al., 2013), indicating that these two QTLs may be the same and potentially strain-specific and broadly exist in Arabidopsis accessions.

In contrast, the distribution of Qpt loci seemed largely distinct between the two RILs although marginal overlap of genetic region exists between Qpt5.1 and Qpt5.2 (Figure 3 and Table 1), suggesting that the genetic basis dominating variance in resistance of Aa-0 and Gie-0 against virulent DC3000 differed substantially. The strain DC3000 has been used in other studies for dissection of natural variation in disease resistance in Arabidopsis, and multiple major QTLs have been uncovered to restrict the size of the bacterial population in planta. The genetic location of Qpt5.1, one of the large effect QTLs detected in this study, is close to that of PRP-Ps2, a major QTL identified with the population of Bayreuth × Shahdara RILs (Perchepied et al., 2006), but is distant from another major QTL previously identified at the bottom of chromosome V that explains most of the variation in resistance between accessions Columbia and San Feliu-2 (Kover and Cheverud, 2007). The location of Qpt1.1, another major QTL identified in this study determining the variation in susceptibility between Col-0 and Gie-0, is clearly distinct from the QTL previously identified on the same chromosome (Fan et al., 2008). These studies demonstrate the complexity of natural variation in Arabidopsis resistance to strain DC3000. We expect that more major QTLs could be uncovered when other accessions or alternative inoculation methods are used for analysis.

Notably, we also found that genetic regions closely overlap between Qpt4.1 and Qpm4.1, and between Qpt5.1 and Qpm5.1 as well (Table 1), implying the existence of QTLs effective against both bacterial strains. However, it remains to be elucidated whether the same genes in the overlapped QTLs mediate disease resistance to both strains. The diversity of QTLs detected in disease resistance, the sensitivity of Qpm3.1 to higher bacterial inocula (Supplementary Figure S6), and the epistasis between Qpm3.1 and Qpt1.1 (Figure 7) revealed in this study collectively indicate that the genetic architecture of the quantitative disease resistance is highly complex and further molecular dissection is necessary to unravel the underlying mechanisms.

In Arabidopsis, the molecular basis of the quantitative natural variation in disease resistance to bacterial infection is still poorly understood since few of the identified QTLs has been cloned. It has been proposed that PTI, ETI, and ETS collectively determine the plant resistance to a virulent pathogen (Jones and Dangl, 2006). Hence, the genetic variation modulating the outcome of any of these processes may lead to the quantitative variation in disease resistance. It has been shown that natural allelic variation at the locus ACCELERATED CELL DEATH 6 causes marked differences in accumulation of SA and in resistance to infections by a panel of phylogenetically distinct pathogens including P. syringae pv tomato DC3000 (Todesco et al., 2009). Other studies have shown that disruption of genes required for ETI compromises resistance in Arabidopsis against both virulent and non-adapted of P. syringae strains, indicating that ETI protects the plant during these seemingly distinct interactions (Zhang et al., 2010). In addition, the genetic variation at locus FLS2, which encodes the flagellin receptor that mediates PTI responses, has also been shown to account for the variation in Arabidopsis resistance to non-adapted P. syringae pv phaseolicola (Forsyth et al., 2010). A more recent study discovered that natural variation in Arabidopsis responses to the type III effector HopAM1 secreted by disarmed DC3000 is quantitatively determined although the HopAM1-induced cell death and chlorosis do not affect bacterial growth in planta (Iakovidis et al., 2016). Our study clearly demonstrated that the locus Qpm3.1 was effective to restrict the growth of bacterial strains carrying the effector gene hopW1-1, a reminiscent of specificity commonly exhibited in R-gene mediated resistance. Several genes around the Qpm3.1 locus are indeed annotated to encode R-like proteins, including one TIR-NBS-LRR gene (At3g04220), one TIR-NB gene (At3g04210); and four receptor-like kinase/protein genes (At3g05360, At3g05370, At3g05650, and At3g05660); several major R-genes have been shown to be responsible for race-specific QTLs conditioning quantitative resistance to Xanthomonas bacteriain Arabidopsis and rice plants (Li et al., 1999; Debieu et al., 2015). However, the narrow spectrum of Qpm3.1 does not necessarily mean that a “defeated” or “weak” R-gene is responsible for the observed quantitative resistance (see discussion below) as diverse molecular mechanisms other than R-avr interaction have been found to underpin the quantitative resistance in multiple pathosystems (Poland et al., 2009; French et al., 2016). In fact, apart from the R-like genes mentioned above, 12 more genes are annotated to confer defense responses among the predicted 384 genes within the Qpm3.1 locus. It is interesting to observe that higher inocula (OD₆₀₀ 0.03) compromised Qpm3.1s mediated resistance (Supplementary Figure S6), and the exact underlying mechanism is still unknown. A possible explanation is that an increased amount of effectors suppressing plant defense were delivered into host cells by the type three secretion system at a higher inoculum (Alfano and Collmer, 2004), which reduced directly or indirectly the effect of Qpm3.1 and hopW1-1 interaction on the restriction of bacterial growth. This finding also implies that the interaction may be effective at the early stage of infection when the bacterial load is still low. Further cloning of the allele responsible for Qpm3.1 will help elucidate the phenomenon.

The discovery of effector-associated QTLs highlights the role of bacterial effectors in delineating natural variation in plant quantitative disease resistance. Since many agriculturally important bacterial pathogens are sequenced and effector coding genes can be efficiently predicted (da Silva et al., 2002; Guttman et al., 2002; Salanoubat et al., 2002), effector-based approaches can be thus designed to scan Arabidopsis accessions for common or rare genetic variations with large effect on pathogenesis of bacterial disease, thereby providing new insights on the fundamental bacterial effector biology. The study on Qpm3.1 also indicated that RILs and HIFs are particularly useful for QTL identification and mendelization, which will markedly facilitate subsequent positional cloning of the responsible gene.

The effector gene hopW1-1, first identified from strain ES4326, is narrowly distributed in a few P. syringae strains (Guttman et al., 2002). This effector disrupts actin cytoskeleton and inhibits endocytosis to promote bacterial virulence when delivered into plant cells (Kang et al., 2014), but elicits resistance responses and limits bacterial growth on some Arabidopsis accessions including Ws (Lee et al., 2008). Infection of Ws plants with the strain DC3000 carrying hopW1-1 specifically triggers an enhanced accumulation of SA at the early stage of interaction; a number of genes involved in SA pathways or encoding the HopW1-interacting proteins are required for the effector-triggered resistance in Ws; two loci, on chromosome III and V, have been identified using Col × Ws F2 population (Lee et al., 2008). Nevertheless, none of these described genes, loci locate at the identified Qpm3.1 locus, and no gene product was annotated in the region as actin-associated linking with the proposed function of HopW1-1; therefore, additional gene(s) conditions natural variation in Arabidopsis response to HopW1-1. Interestingly, by measuring the growth of DC3000 in planta, Dobon et al. have identified four QTLs underpinning natural variation in exogenous SA-induced resistance across Arabidopsis accessions, among which SAQ3, detected in Col-0xLer RILs, locates closely to Qpm3.1 (Dobón et al., 2011). Further investigation on the possible linkage between these two loci may shed new light on the role of SA-dependent pathways in effector-triggered immunity against plant bacterial pathogens.

JF and Y-LP designed the research; QL, W-WL, and K-DP performed the research, collected and analyzed the data; QL, JF, and Y-LP wrote the manuscript. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.