About Us | Help Videos | Contact Us | Subscriptions
 

The Plant Genome - Original Research

CaDMR1 Cosegregates with QTL Pc5.1 for Resistance to Phytophthora capsici in Pepper (Capsicum annuum)

 

This article in TPG

  1. Vol. 7 No. 2
    unlockOPEN ACCESS
     
    Received: Mar 10, 2014
    Published: June 30, 2014


    * Corresponding author(s): avandeynze@ucdavis.edu
 View
 Download
 Alerts
 Permissions
Request Permissions
 Share

doi:10.3835/plantgenome2014.03.0011
  1. William Z. Rehriga,
  2. Hamid Ashrafia,
  3. Theresa Hilla,
  4. James Princeb and
  5. Allen Van Deynze *a
  1. a Seed Biotechnology Center, Univ. of California, Davis, CA 95616
    b College of Natural Sciences, California State Univ.-Chico, Chico, CA 95929

Abstract

A major problem for the pepper (Capsicum annuum) industry is the root rot disease caused by Phytophthora capsici (Pc), to which all commercial varieties suffer yield losses despite good management practices and available landraces with high levels of resistance. A high-density map with 3887 markers was generated in a set of recombinant inbred lines (RIL) derived from the highly resistant Capsicum annuum accession Criollo de Morelos-334 and Early Jalapeño. These lines have been systematically screened for Pc resistance against a set of isolates collected from Mexico, New Mexico, New Jersey, California, Michigan and Tennessee. Quantitative trait loci (QTL) associated with effective resistance across isolates have been identified and validated with SNP markers across additional segregating populations. By leveraging transcriptomic and genomic information, we describe CaDMR1, a homoserine kinase (HSK), as a candidate gene responsible for the major QTL on chromosome P5 for resistance to Pc. SNP markers for the resistant allele were validated to facilitate gene pyramiding schemes for recurrent selection in pepper.


Abbreviations

    AUDPC, area under the disease progress curves; BLASTN, basic local alignment search tool nucleotide; DOWNY MILDEW RESISTANT 1 (DMR1); DP, diversity population; EST, expressed sequence tag; GO, gene ontology; HSK, homoserine kinase; IGA, Illumina genome assembly; Ipcr, inhibitor of P. capsici resistance; LOD, maximum log-likelihood; NM, F7 RIL population C. annuum ‘Early Jalapeño’ × C. annuum ‘CM334’; Pc, Phytophthora capsici; PM, F2 biparental population C. annuum ‘PI 201234’ × C. annuum ‘Maor’; QTL, quantitative trait loci; RIL, recombinant inbred lines; SNP, single nucleotide polymorphism

Globally the pepper market in 2011 was valued at $29.1 billion, with a 40-fold increase in consumption over the last 30 yr driven mainly by hot chile types. The oomycete pathogen Pc is a major challenge in pepper production worldwide. Management for this pathogen is hampered by pesticide-resistant races and long persistence in infected fields. Genetic plant resistance is therefore the ideal management practice for this. Several sources of genetic resistance to Pc have been identified in peppers. One source of resistance is the well-studied C. annuum landrace accession ‘Cdriollo de Morelos-334’ (CM334) which has been reported to be completely resistant to multiple races of Pc (Glosier et al., 2008; Oelke et al., 2003; Thabuis et al., 2003; Truong et al., 2012). However, the genes underlying this resistance remain elusive to plant breeders. Early research suggested resistance was due to a single or a handful of dominant alleles based on complete resistance found in the F1 and segregation ratios of the F2 (Gil Ortega et al., 1992; Monroy-Barbosa and Bosland, 2008; Saini and Sharma, 1978; Smith et al., 1967; Sy et al., 2005; Walker and Bosland, 1999). Conversely, other studies report a polygenic inheritance based on multimodal distributions that range between the susceptible and resistant parents and higher order epistasis effects in resistance (Barksdale et al., 1984; Bartual et al., 1991; Bartual et al., 1993; Bonnet et al., 2007; Lefebvre and Palloix, 1996; Minamiyama et al., 2007; Ogundiwin et al., 2005; Palloix et al., 1988; Pochard and Daubeze, 1980; Reifschneider et al., 1992; Thabuis et al., 2003; Truong et al., 2012). The dissimilarities found between these studies may be due to environmental factors such as the use of different isolates, breeding lines, and the criteria used for scoring (Oelke et al., 2003). Moreover, added complexity in the inheritance of resistance is due to different genes conferring resistance to infection in different parts of the plant such as foliar, fruit, and root (Naegele et al., 2013; Walker and Bosland, 1999). Quantitative trait locus (QTL) studies provide important information for breeders by locating regions of the genome that explain the phenotypic variance for the trait of interest and defining their inheritance. With this information, plant breeders can identify molecular markers associated with these regions and use them to make indirect selections for Pc resistance.

There are numerous reports of QTL associated with resistance to Pc in pepper (Bonnet et al., 2007; Kim et al., 2008; Lefebvre and Palloix, 1996; Minamiyama et al., 2007; Ogundiwin et al., 2005; Pflieger et al., 2001; Sugita et al., 2006; Thabuis et al., 2003; Truong et al., 2012). However, many of these studies have been limited by the number of polymorphic markers employed and the use of segregating F2 and F3 populations that are limited in use for broad testing across environments. Additionally, these studies generally examine resistance to one or two Pc isolates. The selection of a given isolate is a large component of the environmental variance which influences heritability and the power to detect QTL. Indeed, previous research has found variability in different Pc collections. Using a set of host differential lines, Oelke et al. (2003) demonstrated nine different races of Pc collected across pepper growing regions of the United States, Italy, Turkey, and South Korea. This work was further expanded using RIL as host differentials to identify 13 Pc races (Candole et al., 2010; Sy et al., 2008). To better account for the variance contributed by these different races, resistant QTL studies would benefit from screening materials against an array of Pc isolates.

In the present study we evaluate Pc resistance across 20 isolates collected from various pepper growing regions. Using a high-density marker linkage map derived from 66 RIL, we report QTL conferring resistance and describe the identification of C. annuum DOWNY MILDEW RESISTANT 1 (CaDMR1) as a candidate gene for resistance to Pc in pepper.


Materials and Methods

Sixty-six F7 RIL derived from a cross between a commercial susceptible cultivar, Early Jalapeño, and the resistant landrace, CM334, were used to map QTL linked to resistance. We designate this population abbreviation as “NM” as it was created and developed by New Mexico State University and a subset has been used previously as host differentials to identify Pc races (Sy et al., 2008). The roots of these lines along with differential controls were inoculated with Pc and screened for resistance. To validate linkage disequilibrium between resistance loci and SNPs in other genetic backgrounds, two additional pepper populations were scored for resistance and genotyped. The first of these is an F2 biparental population (PM) of 192 individuals derived from a cross between Pc resistant PI 201234 and susceptible bell pepper C. annuum ‘Maor’. Second, a diversity population (DP) representing a mix of different C. annuum open-pollinated varieties, landraces, and wild relatives was selected to assess genetic diversity found within resistant loci. The genetic diversity and structure of the DP population have been characterized by Hill et al. (2013).

Phytophthora capsici Isolates and Assays

Phytophthora capsici isolates used in the present study, and the plant hosts and location where they were collected, are shown in Table 1. Transportation of isolates was under the USDA/APHIS permit P526P-11-04100, P526-11-04099, and P526P-11-04021 for NM, TN, and Mexico, respectively. All experiments were conducted in the greenhouse under UC Davis Biological Use Authorization 964 and California permit 2853. Before inoculation, Pc cultures were maintained on V8-agar media. Zoospores have been demonstrated to be the primary mode of pathogen transmission in the field and greenhouse (Larkin et al., 1995; Ristaino et al., 1992; Stanghellini et al., 1996). For this reason, plants were inoculated with Pc zoospores of each isolate at the two-true-leaf stage, or around 15 d after planting. Methods for root and crown inoculation were a modified version of those published by Bosland and Lindsey (1991). For each isolate, agar plugs were placed in a sterile Petri plate, submerged in sterile water and incubated at 25°C for 72 h, after which plates were incubated at 10°C for 90 min, then 25°C for 30 min. After incubation, zoospores were counted using a haemocytometer and adjusted to a concentration of 3000 zoospores per milliliter. To screen root Pc resistance, each plant was subjected to approximately 10,000 zoospores, which was immediately followed by constant water saturation throughout the duration of the experiment. This number of zoospores has previously been shown to prevent susceptible escapes and has given consistent reaction types on genetically resistant and susceptible plants (Glosier et al., 2008; Oelke et al., 2003).


View Full Table | Close Full ViewTable 1.

List of Phytophthora capsici isolates.

 
Isolate name Plant host Field collection site Source Relative Mean†‡
PWB-128 chile pepper New Mexico Paul Bosland 3.59a
PWB-54 chile pepper New Jersey Steve Johnson 3.17ab
Iso-9 chile pepper Mexico, Aguascalientes, San Francisco Jose de Jesus Luna Ruiz 2.68bc
Iso-7 chile pepper Mexico, Aguascalientes, Jesus Maria Jose de Jesus Luna Ruiz 2.58c
Iso-16 bean Connecticut Kurt Lamour 2.22cd
PWB-106 chile pepper New Mexico Paul Bosland 1.98de
PWB-53 chile pepper New Jersey Steve Johnson 1.77def
Iso-1 tomato Davis, CA Rick Bostock 1.76def
Iso-5 pumpkin Mexico, Aguascalientes, San Francisco Jose de Jesus Luna Ruiz 1.76def
Iso-18 chile pepper Peru Kurt Lamour 1.69ef
Iso-2 tomato CA delta region Van Deynze 1.68ef
Iso-10 chile pepper Mexico, Aguascalientes, San Francisco Jose de Jesus Luna Ruiz 1.64ef
PWB-85 chile pepper New Mexico Rincon 1.6ef
Iso-17 chile pepper New Mexico Kurt Lamour 1.52efg
Iso-6 chile pepper Mexico, Aguascalientes, Cosio Jose de Jesus Luna Ruiz 1.51efgh
PWB-89 chile pepper New Mexico PSRC 1.5efgh
Iso-8 sweet pepper Mexico, Aguascalientes, San Francisco Jose de Jesus Luna Ruiz 1.44fgh
Iso-4 pumpkin Mexico, Aguascalientes, San Francisco Jose de Jesus Luna Ruiz 1.36fgh
Iso-11 chile pepper Mexico, Aguascalientes, San Francisco Jose de Jesus Luna Ruiz 1.29fgh
Iso-13 squash Michigan Kurt Lamour 1.01gh
Iso-15 cucumber Michigan Kurt Lamour 1h
Relative mean of area under disease progress curve. Higher numbers represent increased pathogen virulence and plant susceptibility.
Means not sharing same letter are significantly different at p < 0.01.

Once the susceptible controls began to exhibit disease symptoms (typically 5 to 10 d postinoculation) all plants were scored for disease severity. The time between the first and final evaluation was based on the severity of the disease on the susceptible controls. The disease scale is based on the following criteria: 0 = no symptoms/healthy plant; 1 = leaf yellowing and no stem necrosis; 2 = minor stem girdling at base of the plant; 3 = moderate stem girdling and wilting; 4 = severe stem girdling and severe wilting with signs of leaf necrosis, but not yet dead; 5 = dead plant. All plants were rated over a time gradient of every 2 d to generate an area under disease progress curve (AUDPC; Shaner and Finney, 1977). The formula for AUDPC is calculated as AUDPC = Σi = 1 to n – 1[(Yi+1 + Yi)/2] × [Xi+1Xi], where Yi is score of plants (0–5) at the ith observation (i = 1 being the first observation at time zero), Xi is time at the ith observation, and n is the total number of observations. To obtain an average disease severity rating over time for each RIL, the AUDPC was divided by T, where T is the time (in days) of disease scoring.

The PM and DP populations were screened against the moderately aggressive isolate, PWB-53, because preliminary studies suggested the highly virulent PWB-128 overwhelmed the genetic resistance of the parent (PI 201234) of the PM population (data not shown) and made it difficult to identify subtle differences in the DP population.

Experimental Design and Analysis

Each Pc isolate trial was set up in a randomized complete block design where each treatment (RIL) was blocked by replicate. Three replicates of each treatment contained up to six plants each, depending on seed germination. Additionally, each tray of plants contained one technical control of the resistant parent (CM334), an intermediate resistant control (Paladin), and susceptible parent (Early Jalapeño) to ensure consistent disease scoring. Analysis was conducted using JMP10 ANOVA and Tukey-Kramer HSD for means comparison (SAS Institute, Cary, NC).

Genetic Map and QTL Analysis

DNA extraction and genotyping of the NM RIL was done using the Pepper Affymetrix chip (Hill et al., 2013). A 1 to 2 cM genetic map of expressed sequence tag (EST) contig markers was constructed for this population using the software program RECORD, and distances calculated with JoinMAP v. 4 (Stam, 1993; Van Os et al., 2005). The linkage group assignment was based on common markers between this population and the genetic map of RIL population C. frutescens cv. BG2814–6 × C. annuum cv. NuMexRNaky v. 03 (Yarnes et al., 2013). The order and distance between markers for the map was generated de novo, with 3887 contig markers and 1270 genetic bins representing the NM RIL of the Early Jalapeño and CM334 cross. However, due to high levels of heterozygosity, three RIL lines were left out of the map and QTL analysis. The map used for QTL mapping had <2% missing data and had proportionate AA and BB genotypes (51.6 and 48.4, respectively). Phenotypic scores of resistance were evaluated for QTL using composite interval mapping (CIM) analysis with QTL IciMapping (Li et al., 2007). Similar analyses were done using R/QTL (Arends et al., 2010; Broman et al., 2003; R Core Team, 2012) and WinQTL Cartographer (Wang et al., 2012; data not shown). Although WinQTL Cartographer and R/QTL showed several other significant resistance loci, we only report and validated the QTL consistent across all three computation platforms. A 1000-permutation test was performed to estimate significance level for each QTL.

Gene Ontology Annotation of Sanger-EST, IGA Assemblies, and Pepper Genome Sequence

To identify genomic regions under each QTL, the Sanger-EST and the Illumina genome assembly (IGA) contigs described by Ashrafi et al. (2012) that colocalized under each resistance QTL were aligned to the pepper genome sequence (v. 1.5; Kim et al., 2014) using BLASTN (basic local alignment search tool nucleotide, e–20). The scaffolds that had the highest similarity with Sanger-EST or IGA contigs were selected from the genome sequence Fasta file. For functional annotation the Blast2GO software (Conesa et al., 2005) was used to annotate the gene models with parameters described by Ashrafi et al. (2012). Models were verified by comparing EST and IGA transcriptomic data.

Marker Development and Validation

Sanger-EST contig markers associated with resistance QTL were used to identify single nucleotide polymorphisms (SNP) in close proximity to the resistance loci based on Ashrafi et al. (2012). The Sanger-EST contigs were aligned to Early Jalapeño and CM334 IGA contigs using BLASTN (e–20). The two transcriptomes were then compared for SNPs. The resistance QTL on P1 had three SNPs with one at the peak maximum log-likelihood (LOD) score and two SNPs upstream and downstream. Because the two QTL found on P5 were in close proximity to each other, SNPs were assayed by mapping every 2 cM, when possible, spanning the entire area of the resistance loci (19–29 cM see Fig. 1 and Table S1). Additionally, several SNPs for candidate resistance genes located under QTL on P5 were included for the analysis.

Figure 1.
Figure 1.

Chromosome P1 and P5 in C. annuum Early Jalapeno × C. annuum CM334 map with quantitative trait loci (QTL) results. The QTL results of seven Phytophthora capsici isolates. Isolate-specific resistance QTL found on P1 while five isolates mapped to a common region on P5. PWB-53 mapped to an adjacent region on P5. Average 1000 permutation threshold (maximum log-likelihood, LOD = 3.2) shown as dotted line. Single nucleotide polymorphisms (SNP) were chosen in region flanking and within resistance QTL.

 

Flanking sequences and SNP ID were submitted to LGC Genomics (Beverley, MA) for KASP genotyping assays. Genomic DNA of the NM RIL used in the original QTL analysis was assayed to verify SNP association to resistance alleles. About 50 to 100 mg of leaf tissue was collected from each of the lines and DNA was extracted using QIAxtractor and DX Reagents (Qiagen, Valencia, CA).

Upon receiving SNP genotype data for the NM RIL, allele calls were added to the original haplotype map used for the QTL analysis. With this added information, a new linkage map was constructed using IciMapping to determine if SNP markers will realign to the appropriate linkage group and marker order (Li et al., 2007). The newly constructed map was then used to reanalyze resistance data to validate previous QTL results. Likewise, SNP markers were mapped using IciMapping to the PM population independently for marker order and validation of trait effects.

Multiple Sequence Alignments of Candidate Genes CaDMR1

Tomato (Solanum lycopersicum), potato (Solanum tuberosum), and Arabidopsis thaliana sequences most similar to Capsicum HSK were identified using BLAST searches against the nonredundant protein sequences database, organism Arabidopsis, and organism Solanaceae (Altschul et al., 1997, 2005). A single homolog was identified in each species. The tomato proteins (ITAG release 2.30) database at the Solanaceae Genomics Network was also queried and no additional homologs were identified (Bombarely et al., 2011). Genbank accessions used for the multiple alignment include: Solanum tuberosum (XP_006360852.1), S. lycopersicum (XP_004236876.1), and A. thaliana (AAD33097.1). The alignment was generated using the MUSCLE application within the MEGA 6 program, default settings (Tamura et al., 2013). Conserved residues were highlighted using BOXSHADE 3.21 (http://www.ch.embnet.org/software/BOX_form.html, verified 12 May 2014). The three positions having the highest probability for transit peptide cleavage were determined using the ChloroP 1.1 server (Emanuelsson et al., 1999).


RESULTS

Screening NM RIL for Resistance to Pc Isolates

Sixty-six F7 RIL from Early Jalapeño × CM334 segregating for resistance were inoculated and scored for resistance to Pc using the AUDPC. The Levene test (Levene, 1960) on the 20 different Pc isolates from various locations and plant-host indicated heterogeneity of variance among isolates (P < 0.0001). For this reason, AUDPC and QTL detection were analyzed separately for each isolate. However, there were equal sampling sizes in every test to allow further QTL analysis.

Separate from the QTL analysis, the relative mean AUDPC of each isolate was calculated using Agricolae package in R (Mendiburu, 2012; R Core Team, 2012) and is presented in Table 1. Two isolates (Iso-13 and Iso-15) had very low virulence across RIL and susceptible controls. A majority (11) of the isolates used had a small effect on the RIL population despite strong disease symptoms among susceptible controls (see supplemental Fig. S1). For this reason, only seven isolates were further examined for QTL analyses (Fig. 2). The distribution of resistance among RILs was between the parental means, therefore showing no transgressive segregation. Isolate PWB-128 from New Mexico was the most virulent isolate.

Figure 2.
Figure 2.

Distribution of relative area under the disease progress curve for the seven most virulent Phytophthora capsici isolates. Higher numbers on the x-axis are associated with greater pathogen virulence and plant susceptibility.

 

QTL Analysis

The results of the QTL analysis of resistance to the seven most virulent Pc isolates are shown on Table 2. QTL IciMapping CIM analysis identified a common disease resistance locus on chromosome 5 (P5) for five of the seven isolates (Fig. 1). This region also explained much of the phenotypic variance seen in the experiments (R2 = 0.29–0.58). Resistance to Pc isolate PWB-53 mapped to a nearby region, while Iso-16 did not locate any significant QTL. In addition to the QTL on P5, a QTL located on chromosome 1 (P1) with modest effect (R2 = 0.12) was identified with isolate PWB-106 (Fig. 1). While a few other isolates also showed similar effects of this locus, the LOD scores were not significant (data not shown). Disease resistance was contributed by the resistant parent CM334 for all significant loci. These results verify resistant loci on P5 that contribute broad resistance across multiple Pc isolates as well as suggest that isolate-specific QTL may have a role in enhancing the effect of major resistance gene action.


View Full Table | Close Full ViewTable 2.

Quantitative trait loci (QTL) detected in association with root rot resistance using area under the disease progress curves of seven different Phytophthora capsici isolates.

 
Trait Name Chromosome Position (CI)† LOD‡ R2(%)§ Additive¶
PWB-106 P1 130 (128.7–133) 3.8 12.3 0.3
PWB-106 P5 28 (27.8–28) 11.1 43.8 0.6
PWB-128 P5 29 (26–29.3) 8.0 44.7 0.5
PWB-53 P5 23 (22–28) 3.6 20.6 0.2
PWB-54 P5 28 (27–28) 11.3 58.5 0.8
Iso-7 P5 29 (27.8–30) 6.5 29.4 0.5
Iso-9 P5 29 (27.8–28) 10.3 44.8 0.7
Map units in centimorgans based on linkage map along with confidence interval (CI) where resistance loci most likely to be found.
Maximum log-likelihood value of QTL.
§Percentage variation explained by the QTL.
Additive contribution of the QTL. Positive values are contributions of the donor parent (CM334).

SNP Marker Validation

Based on the QTL found on chromosomes P1 and P5, 12 SNPs were developed for QTL validation. Of all NM RIL genotyped, one failed across all markers due to insufficient DNA to assay. Additionally, CA_0004189 SNP failed validation and was dropped from the assay. Table S2 summarizes the allele calls for the remaining 11 SNPs across the three pepper populations. Several of the markers developed from the NM population were found to be monomorphic in populations with different parentage. However, Fig. 3 shows CA_0029291, CA_0045807, CA_028409, and CA_036024 have alleles that are unique to CM334 in all lines tested (with the exception of two DP lines that were heterozygous for CA_0045807). The SNP allele calls were compared with the original pepper GeneChip marker genotypes at the same map positions. Of the 682 SNP assays (62 RIL [three were dropped due to heterozygosity and one line failed SNP assay] × 11 SNPs), 52 (8%) had a different genotype than the GeneChip; however, 49 (94%) of these were due to heterozygous call differences and missing data. Heterozygotes are not well resolved with the chip assay (Hill et al., 2013). The remaining 3 (6%) differences all occurred for SNP marker CA_011264, which suggests recombination between this and the contig markers that were originally mapped in a single position.

Figure 3.
Figure 3.

Genotypes of CA_0029291, CA_004482, CA_045807, CA_028409, and CA_036024 in pepper diversity population. Square = nonpungent, circle = pungent/small hot type.

 

Table 3 summarizes QTL results of the newly constructed NM map with the added SNP markers and the marker validation population PI 201234 × Maor (PM). As expected, SNP markers colocated with the corresponding Sanger-EST-based markers (GeneChip) that were originally selected from the NM population. However, a slight rearrangement in marker order was found with CA_004482. Despite this, the QTL scan showed similar LOD peaks on P1 and P5, thus suggesting linkage between marker and resistance loci. Moreover, SNP markers CA_011264 and CA_004482 were shown to be highly correlated (R2 = 42 to 49.4%) with Pc resistance in the PM F2 validation population (see Fig. 4). Conversely, SNP markers for the resistance loci on P1 did not contribute to resistance in the PM population.


View Full Table | Close Full ViewTable 3.

Quantitative trait loci (QTL) analyses with SNP markers across populations.

 
Trait Name Pop† Pop type Pop size Chrm Position‡ Left Marker Right Marker LOD§ R2(%)¶ Add# Dom††
Isolate 106 NM RIL 66 P5 48 AQIP CA_011264 9.0 51.2 0.6
Isolate 128 NM RIL 66 P5 48 AQIP CA_011264 8.2 47.0 0.5
Isolate 53 NM RIL 66 P5 37 CA_041448 ADMR 4.1 16.0 0.1
Isolate 54 NM RIL 66 P5 48 AQIP CA_011264 11.5 60.5 0.8
Isolate 7 NM RIL 66 P1 14 CA_028409 AIEB 5.4 21.1 0.4
Isolate 7 NM RIL 66 P5 48 AQIP CA_011264 6.7 29.9 0.5
Isolate 9 NM RIL 66 P5 48 AQIP CA_011264 10.3 44.3 0.7
Isolate 53 PM F2 192 P5 2 CA_011264 CA_004482 25.5 47.4 0.7 0.2
NM = C. annuum ‘CM334’ × ‘Early Jalapeño’, PM = C. annuum ‘Maor’ × PI 201234.
Map units in centimorgans relative to each population.
§Maximum log-likelihood value of QTL.
Percentage variation explained by the QTL.
#Additive contribution of the QTL. Negative values are contributions of the donor parent (CM334).
††Dominant contribution of QTL. Negative values are contributions of the donor parent (PI 201234).
Figure 4.
Figure 4.

Mean of area under the disease progress curve and single nucleotide polymorphism (SNP) genotypes in C. annuum ‘Maor’ × C. annuum PI 201234 population. Error bars = standard deviation of mean. Lower mean scores associated with greater plant resistance.

 

Table 4 summarizes the relative AUDPC of the pepper breeding lines with diverse backgrounds (DP) to Pc isolate PWB-53. While the majority of these displayed high susceptibility, a few showed moderate degrees of Pc resistance. Additionally, only the few resistant lines had the CM334 allele for markers CA_041448, CA_036100, CA_026692, CA_028982, CA_004482, and CA_011264. A separate, independent observation showed higher frequency of the CM334 allele for these markers in pungent types (Fig. 3, Table S3).


View Full Table | Close Full ViewTable 4.

List of diverse pepper varieties and species used to estimate frequency of CM334 resistant alleles.

 
Classification Fruit shape Name Pungency† Relative AUDPC Mean‡
C. annuum var. annuum
Bell blocky Yolo Wonder NP 4.03
blocky Rumba NP 3.92
blocky Violetta NP 4.26
blocky Ariane NP 3.92
blocky Jupiter NP 4.06
blocky King of the North NP NA
blocky Dempsey NP 3.87
blocky Infante NP NA
blocky Grande de Reus NP 4.00
blocky Grosso de Nocera NP NA
triangular Dolmalik NP NA
Long wax triangular Midal NP 4.04
triangular Sweet Banana NP 4.19
triangular Long Yellow Marconi NP 4.14
triangular Corno di Toro NP 3.76
Other triangular Doux des Landes§ NP 3.85
other Carre d’Asti§ P 3.89
elongate Lange Westlandse Rode§ P NA
elongate Sivri Biber§ P 4.02
elongate Erjintiao¶ P 2.83
elongate Jeju¶ P 3.33
elongate Cheongsong¶ P 3.52
elongate Milyang¶ P 3.60
Cayenne elongate Carolina Cayenne P 4.06
Anaheim chile triangular NuMex R Naky P 4.10
triangular NuMex Joe E Parker P 3.49
Jalapeño elongate Jalapeño-M P NA
elongate Early Jalapeño P 2.45
Ancho triangular Ancho 101 P NA
Small hot elongate CM 334 P 1.01
elongate PI 201234 P 1.36
elongate Charleston P NA
elongate Perennial P 2.85
elongate Pusa Jwala P 3.77
elongate G-4 P 1.97
elongate PSP-11 P 3.91
elongate Thai Bird P 3.96
C. chinense elongate PI 159234 P NA
C. frutescens almost round 2814-6 P 4.13
C. pubescens almost round Rocoto P NA
Pungency is indicated by NP (nonpungent) and P (pungent).
Relative area under the disease progress curve (AUDPC) to Phytophthora capsici isolate PWB-53 calculated using agricolae package in R (Mendiburu, 2012, R Core Team, 2012). Higher numbers associated with greater susceptibility. NA = not available.
§Other classification, European origin.
Other classification, East Asian origin.

Gene Ontology Annotations of Sequence Found under Resistant QTL

Further investigation of the QTL found on P5 was performed using Sanger-EST, IGA transcriptomes (Ashrafi et al., 2012), and recent pepper genome assemblies to identify possible candidate resistance genes (Kim et al., 2014). Table S4 lists promising annotated, resistance gene candidates found under or nearby QTL peak on P5. Of these, a SNP (Ca_011264) was designed for a gene encoding HSK and used for the validation studies described above. This gene is orthologous to the DOWNY MILDEW RESISTANT 1 genes in Arabidopsis thaliana (DMR1) and tomato (SlDMR1; Huibers et al., 2013; Lee and Leustek, 1999; van Damme et al., 2009). Therefore, we designate this gene CaDMR1. Further investigation of this SNP revealed no amino acid change to the putative protein sequence. However, a second SNP in CaDMR1 413nt from Ca_011264 at Scaffold 5871 position 386 resulted in a nonsynonymous A to G mutation in the CM334 allele converting a threonine to an isoleucine at residue 28 of the putative translation product (Fig. 5). This amino acid lies within the predicted transit peptide required for chloroplast import of HSK (Fig. 5; Lee and Leustek, 1999; van Damme et al., 2009).

Figure 5.
Figure 5.

Multiple sequence alignment of Early Jalapeno, CM334, Solanum tuberosum, S. lycopersicum, and Arabidopsis thaliana DMR1. Conserved residues were highlighted with solid black and gray boxes, indicating identical consensus and conserved residues, respectively. The red inverted triangle indicates the position of the amino acid substitution in CM334 resulting from SNP Ca_011268. The motif most proximal to the n-terminous (motif1) that is conserved among GHMP superfamily kinases is indicated with an orange box above the sequence. The three most probable positions for transit peptide cleavage between the n-terminous and motif1 are indicated by arrows (blue = Arabidopsis, green = Capsicum, purple = Solanum) with numbers indicating the rank of probability values. The green box indicates the most likely position of cleavage for the AtDMR1 based on predicted probability and proteomic data.

 


DISCUSSION

The distribution of disease severity among isolates found in the current study is of interest, and demonstrates the difficulties of working with pathogens such as Phytophthora. Thirteen out of 20 isolates tested had low virulence, thus skewing disease distribution towards the resistant parent. In the present work, five of the low-virulent isolates came from crops other than pepper (Table 1). However, host-plant origin does not always explain pathogen virulence, as seen in this work where moderately-aggressive Iso-16 was collected from beans. Indeed, recent work by Yin et al. (2012) showed strong virulence patterns among a variety of isolates regardless of host plant origin. Moreover, Ristaino (1990) demonstrated that cucurbit isolates were just as virulent on pepper as pepper isolates. Geography also does not appear to influence virulence. In this work, the five most aggressive isolates originated from New England, New Mexico, and Mexico. This is supported by research done in South Korea demonstrating aggressive isolates collected in France, the Netherlands, Bulgaria, Italy, New Mexico, and South Korea (Kim and Hwang, 1992).

We report here the use of multiple isolates to search and validate QTL conferring resistance to Pc in chile peppers. Thabuis et al. (2003) were among the first to report a resistance locus on P5 that confers the strongest resistance on different organs and at different stages of infection using three unrelated resistant pepper accessions. One of these accessions, CM334, was used as a parent in a different population to investigate race-specific resistance QTL (Truong et al., 2012). That study found a disease resistance locus on P5 that conferred resistance to two Pc races. More recently, a metaanalysis on QTL studies with different sources of Pc resistance in pepper elegantly summarized the P5 linkage map (Mallard et al., 2013) and concluded that there were three resistant loci, Pc5.1, Pc5.2, and Pc5.3. Pc5.1 has been reported to lie between tomato genetic markers C2_At3 g51010 and C2_At1 g33970 (Mallard et al., 2013). Based on the alignments between our Sanger-EST contigs and these tomato markers (data not shown), the major resistance factor found on P5 in our research is Pc5.1, whereas the second minor resistance factor is Pc5.3. Pc5.1 (mapped at 29 cM, see Fig. 1) is reported to have the largest effect on resistance and to condition a broad resistance across multiple isolates. This is in agreement with our results where the major resistance factor on P5 was found across 7 isolates from different geographical origins. Moreover, Pc5.2 and Pc5.3 are reported to be derived from three tolerant accessions (CM334, Vania, and Perennial) but with a smaller overall effect on Pc resistance. Our data support these findings with the second QTL on P5 found only with isolate PWB-53 and with a smaller percent variation explained. We speculate this to be Pc5.3 due to relative map position (mapped at 23 cM) and similar loci effects (Mallard et al., 2013). We did not find a third resistance factor on P5.

Previous research suggests disease resistance loci on chromosomes P1, P3, P4, P6, P8, P9, P10, P11, and P12 (Bonnet et al., 2007; Kim et al., 2008; Ogundiwin et al., 2005; Thabuis et al., 2003; Truong et al., 2012) that are either broad resistance like the locus found on P5 or specific to Pc isolates. However, in the present study there was no indication of these resistance QTL except for the QTL found on P1 (Naegele et al., 2013; Thabuis et al., 2004). The differences in detected QTL between the present study and prior results can be attributed to different parents, population type, and population sizes used. Smaller populations such as the one used in the present study tend to have a higher rate of false negatives, particularly when the heritability of a trait is low (Collard et al., 2009). Moreover, the QTL analysis used here was set to be more stringent by using a consensus of three different QTL software programs to avoid false positives. Conversely, the density of markers used in our map allowed for all recombinations to be accounted for, assigning transcripts to genetic bins. Differences in inoculation and scoring methods can also alter QTL results. Previous research has found variability in different Pc collections and has used a set of host differentials to distinguish Pc collections into different races (Oelke et al., 2003; Sy et al., 2008). Although we report differences in virulence among Pc isolates, we did not find any additional resistance QTL that may contribute to the variability in Pc collections aside from those on P1 and P5. This may be due to the large effect on P5 overshadowing the subtle differences of other resistance factors. For this reason, future Pc studies in peppers may benefit from examining resistance in populations where QTL Pc5.1 is fixed, although these types of studies would not detect interactions.

Results of a QTL study can often be limited in application to the specific populations in which they are found. Therefore, we validated the SNP markers found in the vicinity of the resistance QTL in the NM population in two distinct pepper populations. The SNP assays colocated with the corresponding Sanger-EST-based markers (GeneChip) that were originally selected from the NM population. Majority of the differences were due to heterozygous calls or missing data which are not resolved with chip assays (Hill et al., 2013). These two factors explain the slight rearrangement found here for marker CA_004482. The results demonstrate high linkage disequilibrium between the two markers CA_011264 and CA_004482 and Pc resistance (R2 = 42 to 49.4%). This supports earlier research by Thabuis et al. (2003) which proposed a common resistance factor between pepper accessions PI 201234 (resistant progenitor of Vania) and CM334. This study provides tightly linked markers for breeders to facilitate introgression of resistance loci in different pepper populations. Moreover, the majority of SNPs showed normal segregation patterns in the NM and PM populations (see Table S2). Additionally, we report here several SNPs unique to the resistant accession CM334 across a diverse array of pepper populations. Of these, CA_0029291 explains 35% of the resistance variation in the NM population. It is also of interest to note that while not exclusive to CM334, CA_004482 and CA_011264 share common alleles between accessions Perennial, PI201234, and G4, which have all been reported to have moderate Pc resistance (see Fig. 3; Glosier et al., 2008; Thabuis et al., 2003). However, these alleles are shared by accessions Midal and Jeju which exhibit extreme susceptibility towards Pc (see Table 4). A recent study suggested that a resistance-inhibitor gene, Ipcr, may mask the effects of resistance loci (Reeves et al., 2013). It may be speculated that although these accessions have the resistance factor at Pc5.1, they display a susceptible phenotype because they also have an active Ipcr gene. However, further research is needed to confirm this hypothesis. The QTL in P1 and the second one on P5 (likely Pc5.3) were not validated in other populations due lack of polymorphism or assaying with a different isolate than was used within the NM population. The validation study screening the PM population to isolate PWB-53 disagrees with the shifted position originally found in the NM population. This may be attributed to some background effect from the parent Early Jalapeño, which has been shown to have low levels of resistance compared with highly susceptible Maor (Ogundiwin et al., 2005).

Results of the Sanger-EST and IGA contig alignments to the pepper genome suggest Pc5.1 lies between 22 and 32 Mb, while Pc5.3 lies between 6 and 8 Mb positions of the chromosome P5 pseudomolecule PGA v. 1.5 (Kim et al., 2014). The gene ontology (GO) annotations suggest 33 candidate genes that may regulate resistance under or nearby the Pc5.1 region including transcription factors, kinases, and R2-like resistance genes (Table S4). Each of these genes was examined for transcriptomic evidence in resistant and susceptible parents, genetic and physical map position, presence of nonsynonymous SNP, and R2 for disease severity. It should be noted that both Sanger-ESTs and IGA assemblies were derived from tissues exposed to Pc. Some of these genes were discounted due to no expression in either parent. Other genes were ignored due to lack of nonsynonymous SNPs. Although a few genes had functional SNPs, they mapped to a genetic bin that had a lower R2. However, the possibility of variation existing in the promoter or gene intron regions cannot be dismissed. Furthermore, it is possible the expression of these genes is tissue specific or has posttranscription modifications that regulate protein functionality.

The highest R2 values for genes under the resistance QTL across two mapping populations mapped to an ortholog of Arabidopsis thaliana DOWNY MILDEW RESISTANT 1 (DMR1) gene which encodes a HSK. Although not aligned to P5 in the genome (found in unassigned PGA v. 1.5.scaffold5871), our genetic maps, transcriptomic data, and BLAST data show that the HSK gene maps between the 24364694:24366808 genomic positions of P5. van Damme et al. (2009) characterized this gene in Arabidopsis while working with downy mildew (Hyaloperonospora arabidopsidis). Their research strongly suggests that DMR1 has an important role in resistance to this oomycete pathogen. There is a nonsynonymous change in amino acid (threonine to isoleucine, see Fig. 5) conserved to resistant lines tested, CM334 and PI201234, in the HSK transit peptide domain suggesting that HSK chloroplast localization may be disrupted. This change is due to a SNP from the functional allele of guanine in susceptible lines to adenine in resistant lines at position 386 on PGA v. 1.5.scaffold5871. It also refers to SNP CA_001128 on MGMT_Contig7759 (Ashrafi et al., 2012). It has been demonstrated that elevated homoserine levels in the chloroplast lead to Arabidopsis resistance to the oomycete H. arabidopsidis by an as yet undescribed mechanism (van Damme et al., 2009). In these experiments, modification of transit peptides did not complement mutations in HSK as the full transcript did. Disrupting HSK function has also been shown to confer resistance to downy mildew in lettuce (Latuca sativa L.) and onion (Allium cepa L., van Damme and van den Ackerveken, 2012). This gene was also characterized in tomato, in which the silencing of SIDMR1 resulted in resistance to downy mildew (Huibers et al., 2013). Interestingly, in the tomato experiment, expression of HSK resulted in small fruit and leaf size. This may explain why it has been challenging to transfer CM334-based Pc resistance to large blocky type peppers. Pc resistance from CM334 has been reported by some to be dominantly inherited (Bosland and Lindsey, 1991; Reeves et al., 2013). However, there are reports of epistatic and semidominance effects of resistance to Pc (Bartual et al., 1993; Lefebvre and Palloix, 1996; Thabuis et al., 2003). The results of the SNP effects in the F2 PM population suggest a dosage response (see Fig. 4). In the tested populations, the missense mutation found in CM334 and PI 201234 HSK behaves in a loss of function (NM RIL) or nonadditive manner (PM F2). The molecular effect of the CM334 Cadmr1 allele and the role of HSK in Pc resistance (or susceptibility) remain to be verified using gain and loss of function gene assays in peppers. Our results confirm a strong correlation between a SNP (Ca_011264) in CaDMR1 and resistant phenotypes in three separate pepper populations. Additionally, this allele is conserved in CM334 and other resistant accessions when compared with other pepper cultivars of diverse background (see Table 4, Fig. 3).


Conclusions

Our results support existing evidence that Pc virulence is not dependent on host or geographical location. Additionally, the broad resistance behavior of the resistance factor on P5 suggests that it contains a critical component of pepper Pc resistance. This, along with other minor resistance QTL such as the one found on P1, may account for CM334’s durable resistance to a diverse array of Pc isolates. Here we report several SNP molecular markers in close association with Pc resistance from CM334 and demonstrate the applicability of these markers in multiple pepper cultivars for breeding purposes. We propose that the CaDMR1 gene encoding a HSK, highly associated here with the major resistance QTL on P5 and consistent with results in Arabidopsis, tomato, and lettuce, is a strong candidate to be responsible for this Pc resistance trait. Further work on gene cloning, functional assays, and silencing experiments will test this hypothesis.

Supplemental Information Available

Supplemental information is included with this article. A video showing phenotypes in time lapse is available at http://youtu.be/oL8CoZMYil4.

Figure S1. Distribution of relative area under the disease progress curve for 13 nonvirulent Pc isolates. Higher numbers on x-axis associated with greater pathogen virulence and plant susceptibility.

  • Table S1. List of SNPs under Phytophthora capsici resistance QTLs.

  • Table S2. Summary of SNP allele frequencies and variance accounted for.

  • Table S3. Allele frequencies of SNPs among different pepper backgrounds.

  • Table S4. Candidate genes found under or nearby QTL peak of chromosome P5.

Acknowledgments

We would like to thank Dr. Jose de Jesus Luna Ruiz of the Universidad Autonoma de Aguascalientes, Mexico, Dr. Kurt Lamour, University of Tennessee, and Dr. Paul Bosland, New Mexico State Univ. for the Phytophthora cultures; Dr. Paul Bosland for the NM RIL population; Xiangyang Zheng, Magnum Seed Inc. for inoculation methodology; and Dr. Kent J. Bradford, University of California Davis, for critiquing and editing the manuscript.

 

References

Footnotes


Comments
Be the first to comment.



Please log in to post a comment.
*Society members, certified professionals, and authors are permitted to comment.