ORIGINAL PAPER

Mapping quantitative trait loci associated with resistanceto bacterial spot (Xanthomonas arboricola pv. pruni) in peach

Nannan Yang & Gregory Reighard & David Ritchie &

William Okie & Ksenija Gasic

Received: 26 June 2012 /Revised: 21 September 2012 /Accepted: 1 November 2012 /Published online: 29 November 2012# Springer-Verlag Berlin Heidelberg 2012

Abstract Bacterial spot, caused by Xanthomonas arbori-cola pv. pruni (Xap), is a serious disease that can affectpeach fruit quality and production worldwide. This diseasecauses severe defoliation and blemishing of fruit, particular-ly in areas with high rainfall, strong winds, high humidity,and sandy soil. The molecular basis of its tolerance andsusceptibility in peach is yet to be understood. An F2 pop-ulation of 63 genotypes derived from a cross betweenpeaches “O’Henry” (susceptible) and “Clayton” (resistant)has been used for linkage map construction and quantitativetrait loci (QTL) mapping. Phenotypic data for leaf and fruitresponse to Xap infection were collected over 2 years at twolocations. A high-density genetic linkage map that covers agenetic distance of 421.4 cM with an average spacingbetween markers of 1.6 cM was developed using theInternational Peach Single Nucleotide PolymorphismConsortium (IPSC) 9K array v1. Fourteen QTLs with anadditive effect on Xap resistance were detected, includingfour major QTLs on linkage groups (LG) 1, 4, 5, and 6.Major QTLs, Xap.Pp.OC-4.1 and Xap.Pp.OC-4.2, on LG4were associated with Xap resistance in leaf; Xap.Pp.OC-5.1on LG5 was associated with Xap resistance in both leaf and

fruit, while Xap.Pp.OC-1.2 and Xap.Pp.OC-6.1 on LG1 andLG6, respectively, were associated with Xap resistance infruit. This suggested separate regulation of leaf and fruitresistance for Xap in peach as well as participation of genesinvolved in general plant response to biotic stress. Thepotential for marker-assisted selection for Xap resistance inpeach is discussed.

Keywords SNPmarkers . Bacterial leaf and fruit diseases .

Peach . Quantitative trait loci

Introduction

Bacterial spot, caused by Xanthomonas arboricola pv. pruni(Xap), is a serious disease that can affect nearly all cultivatedPrunus species and their hybrids (EPPO 1997). It was firstdescribed on plum in the USA (Smith 1903). Xap was alsoidentified on peach and other stone fruits (Rolfs 1915;Dunegan 1932). The most severe infections have beenreported on Japanese plum (Prunus salicina), Korean cherry(Prunus japonica) and plum hybrids, and on peach andnectarines (Prunus persica) and their hybrids, with over50 % infection (Ritchie 1995). Today, the pathogen is pres-ent and widespread in eastern USA, China, South Africa,and Uruguay, whereas local outbreaks have been reported inmany other countries such as Bulgaria, Romania, Moldova,Russia, Ukraine, India, Pakistan, Japan, Korea, Canada,Mexico, Argentina, Brazil, and Australia (EPPO 2006).Xap can affect leaves, twigs, and fruits. Severe infectionresults in premature leaf defoliation and tree weakening, aswell as reduced fruit quality to the point of unmarketableand thus lower yield (Ritchie 1995). Traditionally, antibac-terial sprays, such as oxytetracycline or copper-based com-pounds, are the primary control method (Ritchie 1995).However, with the environmentally conscious public, chem-ical control of Xap is under close scrutiny. Thus, interest in

Communicated by E. Dirlewanger

Electronic supplementary material The online version of this article(doi:10.1007/s11295-012-0580-x) contains supplementary material,which is available to authorized users.

N. Yang :G. Reighard :K. Gasic (*)SAFES, Clemson University, Clemson, SC 29634, USAe-mail: kgasic@clemson.edu

D. RitchieDepartment of Plant Pathology, North Carolina State University,Raleigh, NC 27695, USA

W. OkieARS-USDA, S.E. Fruit and Tree Nut Research Lab,Byron, GA 31008, USA

Tree Genetics & Genomes (2013) 9:573–586DOI 10.1007/s11295-012-0580-x

developing resistant peach cultivars has expanded in peachbreeding programs.

Peach cultivars vary greatly in susceptibility to Xap andthe most effective control is through the use of host plantresistance (Werner et al. 1986). Unfortunately, many resis-tant cultivars lack specific desirable fruit and marketingcharacteristics (Okie 1998). The breeding program inNorth Carolina was successful in developing a seriesof Xap-resistant cultivars, the most resistant of which were“Clayton” and “Candor” (Okie et al. 2008), through intro-gressing resistance from the cultivar Elberta into the popularcommercial cultivar J. H. Hale (Okie 1998). Althoughhighly resistant cultivars have been identified (Keil andFogle 1974; Simeone 1985; Werner et al. 1986), consider-able variation has been noticed in disease incidence fromyear to year, and under favorable conditions for infection, allcultivars show at least some symptoms. Integration of agenomics approach with traditional breeding facilitatesmore efficient introgression of Xap resistance to developnew peach cultivars. A molecular breeding approach viathe application of DNA markers, which tag the resistanceloci of interest, offers preselection of resistant individualsand, therefore, could accelerate the breeding process. Theapplication of marker-assisted breeding (MAB) requireswell-developed genetic resources. Peach is one of the bestcharacterized fruit tree species and serves as a model forgenetics studies in Rosaceae and other tree species(Dirlewanger et al. 2004; Shulaev et al. 2008). The availablePrunus reference map (Dirlewanger et al. 2004) along withrelease of peach genome sequence v1 (Sosinski et al. 2010;Arus et al. 2012) and recently developed single nucleotidepolymorphism (SNP) genotyping resources (Ahmad et al.2011; Verde et al. 2012) offers vast resources for markerdetection and MAB application.

The number of cultivars resistant to bacterial spot re-leased by eastern US breeding programs suggests involve-ment of dominant genes in Xap resistance (Sherman andLyrene 1981). However, the inconsistent levels of leaf andfruit resistance in peach indicate that separate genetic factorsmight regulate the leaf and fruit resistance (Werner et al.1986). The molecular mechanism of resistance/susceptibil-ity to Xap is not yet clear. Recently, there were severalattempts to understand the molecular basis of Xap resistancein Prunus (Yang et al. 2010; Socquet-Juglard et al. 2011).Our preliminary work (Yang et al. 2010) suggested a poly-genic nature of Xap resistance in peach. One putative quan-titative trait loci (QTL) region was detected on linkagegroup 4, but the low-density linkage map restricted theQTL analysis and discovery of other QTLs with majoreffects (Yang 2012). Additionally, Socquet-Juglard et al.(2011), using a low-density simple sequence repeat (SSR)linkage map (Dondini et al. 2007), identified four genomicregions related to Xap resistance in apricot and reported a

single QTL on linkage group 5 being of interest for marker-assisted selection. However, to date, no tightly linkedmarkers or isolation of genes associated with Xap resistancewere reported. Both studies in peach and apricot sufferedfrom low-density linkage maps and not enough diversitybetween parents to allow for more markers to be analyzed.Recently released peach genome sequence v1 (Sosinski etal. 2009; www.rosaceae.org) and growing single nucleotidepolymorphism marker resources (Ahmad et al. 2011; Verdeet al. 2012) facilitate the development of a high-densitylinkage map and discovery of QTLs associated with impor-tant agronomic traits.

Therefore, in the present study, we are reporting the useof the International Peach SNP Consortium (IPSC) 9Kpeach SNP array v1 (Verde et al. 2012) in developing ahighly saturated linkage map and mapping the QTLs re-sponsible for Xap resistance in peach. The mode of inheri-tance of Xap resistance in peach leaf and fruit and potentialfor marker-assisted selection will be discussed.

Materials and methods

Plant material

Peach cultivar O’Henry (highly susceptible to Xap) wascrossed with Clayton (highly resistant) in 2002 at theUSDA-ARS Southeastern Fruit and Tree Nut Lab inByron, GA. One of the resulting F1 plants, BY03p2388x,was enclosed in netting during bloom to produce self-pollinated F2 plants. These F2 plants were replicated asyoung seedlings (Okie 1984) and planted in three locations.This “O’Henry” × Clayton (OC) mapping population, con-sisting of 188 seedling clones, was used for phenotyping ofXap resistance in the field. Clayton is a yellow, melting,freestone peach selected from a “Pekin” × Candor cross inthe North Carolina peach breeding program and is one of themost highly resistant cultivars to bacterial spot (X. arbor-icola pv. pruni). O’Henry is a high-quality, yellow, melting,and freestone peach released in 1968 by Grant Merrill inRed Bluff, California, from an open pollination of “MerrillBonanza” (Okie 1998). O’Henry is one of the most suscep-tible peaches to Xap. The OC population also segregates forflower type and skin pubescence. Clayton has non-showyflowers (Sh/Sh) and is homozygous for pubescence (G/G);O’Henry has showy flowers (sh/sh) and is heterozygous forskin pubescence (G/g), so some of the F2s are nectarines(g/g). Showy flower (sh) and glabrous skin (g) are controlledby recessive genes (Blake 1932; Bailey and French 1949).The mapping population was maintained in two replicates attwo locations: Sandhill Research and Education Center,Pontiac, SC (Latitude 34.129711°, Longitude −80.862644°,Elevation 128.631 m above sea level) and Sandhills Research

574 Tree Genetics & Genomes (2013) 9:573–586

Station, Jackson Springs, NC (Latitude 35.18782°, Longitude−79.68437°, Elevation 190.5 m). Both locations have ahumid, subtropical climate with high humidity (up to 80 %),and mild temperatures in the spring (10–15 °C).

Assessment of Xap incidence

Bacterial inoculum was prepared by growing virulent iso-lates of X. arboricola pv. pruni (Xap) on agar medium(sucrose peptone, PDA, nutrient agar, and 1 % glucose orsucrose) for 36–48 h. The cultures were washed off from themedia with sterile water, and bacterial suspension with theoptical density of 1.0–1.5 or greater (600 nm) was prepared.The bacterial suspension was applied by spraying inoculumon 2-year-old trees in the early spring of 2008 in NC and2009 in SC from the late petal fall to shuck split stage toensure the presence of inoculum in each tree. Field responseto Xap infection on leaves and fruits in each cycle ofevaluation was assessed using the ordinal scale method from0 (no symptoms on leaves and fruits) to 5 (>50 % diseasedleaves or observed defoliation or >50 % fruit surface withspot lesions) described in Yang (2012). Symptoms on leaveswere visually evaluated according to the overall perfor-mance of intensity and distribution in the tree for eachreplicate at both locations. Symptoms on the fruits were alsovisually evaluated, but only scores of fruits with the highestseverity were recorded for QTL analysis. Leaf symptomswere evaluated once a month from May to July during twoseasons at two locations, NC (2008 and 2009) and SC (2009and 2011). In 2009 at the NC location, third cycle symptoms(July) were not scored due to inability to distinguish bacte-rial spot from other disease symptoms and mechanicaldamage. Fruit symptoms on each tree were evaluated inJune 2008 in NC and June 2011 in SC.

DNA isolation and SNP genotyping

SNP genotyping was performed on a subset of 63 progenyfrom the OC population, with the disease response in agree-ment between the two locations. Isolation of genomic DNAand a subsequent Infinium assay were performed asexplained in Verde et al. (2012). In short, genomic DNAwas isolated from fresh young leaves of 63 OC progenyusing the E-Z 96 Tissue DNA kit (Omega Bio-Tek, Inc.,Norcross, GA, USA), and quantitated with the Quant-iT™PicoGreen® Assay (Life Technologies, Grand Island, NY,USA) using the Victor multiplate reader (Perkin Elmer, Inc.,San Jose, CA, USA). Concentrations were adjusted to aminimum of 50 ng/μl in 5 μl aliquots and submitted to theResearch Technology Support Facility at Michigan StateUniversity (East Lansing, MI, USA) where the Infiniumassay was performed following the manufacturer’s protocol(Illumina, Inc.). After amplification, PCR products were

hybridized to VeraCode microbeads via the address se-quence for detection on a VeraCode BeadXpress Reader.SNP genotypes were scored with the Genotyping Module ofGenomeStudio Data Analysis software (Illumina, Inc.). AGenTrain score of >0.4 and a GenCall 10 % of >0.2were applied to remove most SNPs that did not cluster(homozygous) or had ambiguous clustering. SNPs homozy-gous for alternate alleles in two parents as well as SNPshomozygous in one and heterozygous in the other parentwere considered for mapping. F2 population type codeswere applied (Van Ooijen 2006).

Linkage map construction

Linkage analyses were performed using JoinMap 4.1 (VanOoijen 2006) and R/QTL package (Broman et al. 2003). Thedeviations from a Mendelian ratio were tested using a Chi-square goodness-of-fit test (P<0.05) available in JoinMap4.1. Polymorphic SNP and SSR markers from previouswork (Yang et al. 2011) were initially grouped byJoinMap. Each group was then compared to the peachgenome v1.0 (GDR, www.rosaceae.org) sequence and edi-ted for the SNP position. Then, each group was separatelyre-created by R/QTL, using a minimum 6.0 log of odds(LOD) and a maximum 0.35 recombination frequency. Theplotting of marker order in each group was accomplished by“plot.rf.” The final linkage map was constructed using “rip-ple” and “mapthis” functions (P<0.005). Marker orders thatconflicted with the physical map were adjusted and recalcu-lated based on LOD scores using the “switchorder” functionin R/QTL. The map distances were calculated usingKosambi’s mapping function (Kosambi 1944). Accuracyof the linkage map was iteratively checked and confirmedby calculating pairwise recombination fractions across thegenome and comparing marker order to the physical loca-tion on the peach genome v1.0.

Comparison of the position of the SNPs in the physicaland genetic map

The set of SNPs mapped in each linkage group were alignedwith their position on the peach genome using MapChart 2.2(Voorrips 2002), and colinearity among the linkage andphysical map was evaluated.

Statistical analysis

Mean and standard deviation were calculated, and the Xapresistance scores were tested for normality. Broad-senseheritability (H2) of genotypic mean values was estimated

using the formula H2 ¼ σ2g σ2

g þ σ2e

� �., where σg

2 is the

genotypic variance, and σe2 is the environmental variance as

Tree Genetics & Genomes (2013) 9:573–586 575

described in Rubio et al. (2010). Statistical analyses of thedatasets were performed using ANOVA in SPSS v.20(IBM). Significant differences were calculated using theTukey HSD test at P<0.05.

QTL analysis and mapping of Xap resistance

Xap incidence data, collected for leaf and fruit, were orga-nized in datasets, according to Rubio et al. (2010). Threemonthly data points collected for bacterial spot incidence onleaves for each accession replicate in each season and loca-tion and the maximum values for each data point, location,and year were organized into 36 leaf datasets (Supplementalfile SM 1). Two leaf datasets from the NC location, C3-NC09LEA and D3-NC09LEA, could not be created sincethe third cycle scores for 2009 were lacking (see theabove “Assessment of Xap incidence”). Three types ofmaximal scores were included for each location: maximalscores that combined three cycles of rating data pointsfor each replicate; maximal scores that combined tworeplicates and three cycles of rating data points for eachseason, and maximal scores that combined all the ratingdata points for each location. The bacterial spot inci-dence (most severe symptoms) on peach fruit for eachaccession replicate was rated once at each location. Inaddition, the maximal score for each tree was extracted,resulting in six fruit datasets. Due to tree death, notevery dataset contained ratings for all 63 trees. In total,42 datasets were constructed and used for QTL analysis(Table 1).

Phenotypic data were tested for the normality of distri-bution using Windows QTL Cartographer V2.5 (Wang et al.2007; http://statgen.ncsu.edu/qtlcart/WQTLCart.htm).Detection of putative QTLs was performed using compositeinterval mapping (CIM), with a 1,000-permutation test, asdescribed by Rubio et al. (2010). A nonparametric test basedon the Kruskal–Wallis (KW) (Kruglyak and Lander 1995)and multiple regression (MR) with the threshold of 0.5 %were conducted using the MapQTL 6.0 software (VanOoijen 2009) for data sets that departed from normality. Inaddition, a less stringent threshold of 5 % was applied incase no putative QTLs were detected by CIM, MR, and/orKW. MR analysis was used to estimate the percentage ofphenotypic variation (R2) explained for each individual QTLand for all QTLs (R2t).

Candidate gene mapping

Coding sequences were extracted from Xap.Pp.CO-1.2,Xap.Pp.CO-4.1, Xap.Pp.CO-4.2, Xap.Pp.CO-5.1, andXap.Pp.CO-6.1 QTL regions of peach genome v1 andannotated against the Arabidopsis database (TAIR; www.arabidopsis.org/) using BLAST2GO (Conesa et al. 2005).

Results

Phenotypic evaluation of resistance to Xap

Leaf symptoms on the highly resistant parent Clayton variedfrom “1” to “3” in different years and locations; however, nosymptoms on the fruit were detected. At the same time, thehighly susceptible parent, O’Henry, exhibited high leaf andfruit susceptibility to Xap in both locations and all seasons(score of ≥3) (data not shown). Phenotypic evaluation ofXap leaf incidence was obtained over three seasons, from2008 to 2011, at two locations, SC and NC (Table 1; sup-plemental figures SM 2 and 3). No significant differenceswere observed between datasets representing average Xapleaf incidence for replicates in the same evaluation cycles,year, and location. However, significant differences wereobserved between locations and evaluation years. Xap inci-dence on leaves was evaluated in both locations only during2009, and no significant difference between average symp-tom scores in the same evaluation cycle was observed be-tween the two locations (Table 1). The Xap incidence scoresin 24 of 36 leaf, and all six fruit datasets were close tonormal distribution at the 5 % level (Table 1). Four of thenon-normal sets used maximal scores (Table 1). Dataobtained from SC showed a higher average value of Xapincidence in 2011 (4.41) than in 2009 (1.83). A similar trendwas observed for NC data where Xap incidence was higherin 2008 (2.19) than 2009 (1.93) (Table 1). No individualaccession had consistent leaf resistance across the two eval-uation seasons and locations. Seven individual accessionsshowed low Xap incidence on fruit in both SC and NC, butonly one, 031, was scored “0” for both locations.

The mean values were generally lower in early evaluationstages, with lowest for A1-SC09LEA (0.73) and C1-NC09LEA (0.55) (Table 1; supplemental files SM 2 and 3).As expected, the highest mean values were scored in datasetsrepresenting the maximal disease symptoms with the highestin MaxA-SC11LEA (4.27) (Table 1). The range of symptomscores was wide in both locations and all years, with thenarrower scores observed in SC (0–1) and NC (0–2) in 2009and the widest in SC (1–5) and NC (0–5) in 2011 and 2008,respectively. Effects of environmental factors were evaluatedwith broad-sense heritability, which ranged from 0.15 (B1-SC09LEA) to 0.84 (MaxD-NC08LEA) in 36 leaf datasets,suggesting that important environmental factors were in-volved in leaf resistance to Xap (Table 1). Higher heritability(over 0.8) for the six fruit datasets, however, suggested minorenvironmental effects on fruit resistance to Xap infection.

SNP genotyping

The individual sample call rate was of ≥99 % for 63 indi-vidual samples and the two parents, except for no. 134 for

576 Tree Genetics & Genomes (2013) 9:573–586

Table 1 Summary of the statistics computed with the phenotypic data of leaves and fruit obtained from selected OC progeny

Datasets Population size Mean (SD) Range Skewness Kurtosis S test Heritability (H2)

Leaf

A1-SC09LEAab 51 0.73 (0.60) 0–3 0.17 0.77 22.37 0.31

A2-SC09LEAbcd 51 1.00 (0.20) 0–2 0.00 0.05 1348.2 –

A3-SC09LEAbcd 51 0.96 (0.20) 0–1 −0.04 0.04 1329.8 –

B1-SC09LEAabc 58 0.83 (0.38) 0–1 −0.10 0.09 35.24 0.15

B2-SC09LEAcde 58 1.31 (1.14) 0–4 1.33 5.26 7.66 0.81

B3-SC09LEAcdef 58 0.65 (0.81) 0–4 0.71 2.08 26.57 0.62

A1-SC11LEAjk 51 3.63 (0.72) 1–5 −0.48 1.61 33.66 0.52

A2-SC11LEAkl 52 3.96 (0.74) 2–5 −0.10 0.86 0.54 0.54

A3-SC11LEAkl 51 3.86 (0.53) 2–5 −0.02 0.29 1.12 0.28

B1-SC11LEAj 59 3.25 (0.82) 1–5 −0.28 1.75 4.29 0.63

B2-SC11LEAj 58 3.53 (1.08) 1–5 −0.88 4.60 5.07 0.79

B3-SC11LEAjk 59 3.73 (0.87) 2–5 −0.06 1.39 0.80 0.67

C1-NC08LEAab 44 0.84 (0.83) 0–3 0.33 1.26 2.63 0.64

C2-NC08LEAcdef 42 1.31 (1.02) 0–3 0.20 2.38 1.47 0.76

C3-NC08LEAfgh 43 1.84 (0.87) 1–4 0.37 1.36 2.95 0.67

D1-NC08LEAab 38 0.87 (1.09) 0–4 0.73 9.57 42.64 0.79

D2-NC08LEAbcd 37 1.19 (1.17) 0–5 1.84 9.22 13.15 0.82

D3-NC08LEAdefg 37 1.62 (1.09) 0–4 0.90 4.62 3.08 0.79

C1-NC09LEAa 42 0.55 (0.67) 0–2 0.60 0.58 4.97 0.44

C2-NC09LEAefgh 42 1.86 (1.03) 0–4 0.88 2.46 1.07 0.76

D1-NC09LEAab 36 0.69 (0.82) 0–3 0.69 1.60 5.96 0.63

D2-NC09LEAcdef 36 1.42 (1.05) 0–5 0.36 2.70 1.53 0.77

MaxA-SC09LEAbcd 51 1.06 (0.37) 0–3 0.16 0.38 747.25 0.14

MaxB-SC09LEAefgh 58 1.83 (1.05) 0–4 0.74 2.74 5.20 0.77

MaxA-SC11LEAl 52 4.27 (0.50) 3–5 0.42 0.16 2.71 0.24

MaxB-SC11LEAkl 59 4.08 (0.75) 2–5 −0.17 0.89 1.60 0.55

MaxC-NC08LEAgh 44 1.95 (0.83) 1–4 0.20 1.15 1.55 0.64

MaxD-NC08LEAefgh 38 1.66 (1.26) 0–5 1.90 9.30 6.58 0.84

MaxC-NC09LEAgh 42 1.88 (1.02) 0–4 −0.05 2.44 0.91 0.76

MaxD-NC09LEAefgh 36 1.42 (1.05) 0–3 0.36 2.70 1.53 0.77

Max-SC09LEAefgh 60 1.83 (1.01) 1–4 0.78 2.43 6.82 0.76

Max-SC11LEAl 63 4.41 (0.59) 2–5 −0.18 0.68 28.47 0.27

Max-SC LEAl 63 4.41 (0.59) 2–5 −0.18 0.68 28.47 0.27

Max-NC08 LEAhi 54 2.19 (1.05) 0–5 0.50 3.66 1.69 0.77

Max-NC09 LEAgh 54 1.93 (1.01) 0–4 −0.08 2.09 2.14 0.75

Max-NCLEAi 55 2.53 (0.96) 0–5 −0.19 2.94 0.92 0.73

Fruit

A1-SC11FRUa 43 1.77 (1.43) 0–5 2.31 11.59 4.57 0.88

B1-SC11FRUa 42 2.10 (1.14) 0–5 0.78 4.84 1.98 0.81

Max-SC11FRUa 50 2.22 (1.39) 0–5 0.81 8.56 1.57 0.87

C1-NC09FRUab 20 1.50 (1.36) 0–4 1.58 8.98 1.43 0.86

D1-NC09FRUb 28 0.93 (1.15) 0–3 1.18 4.23 3.22 0.81

Max-NC09FRUa 40 1.40 (1.26) 0–4 0.96 5.71 2.40 0.84

Each dataset name reflects replication (A, B, C, and D), evaluation (1, 2, and 3), location (SC and NC), year (2008, 2009, and 2011), and plantorgan (LEA and FRU). Maximal leaf and fruit phenotypic data were generated for each replicate from three cycles of rating data, for each seasonfrom two replicates and three cycles of rating data, and for each location from all the rating data. Those datasets that show normal distribution arebolded. The critical values for the rejection of normality are 5.99 and 9.21 at the 5 and 1 % levels, respectively. The mean difference is significant atthe 0.05 level. Datasets with the same letter are not significantly different. Leaf and fruit datasets are compared separately

SC South Carolina, NC North Carolina, LEA leaf, FRU fruit

Tree Genetics & Genomes (2013) 9:573–586 577

which genotyping was successful for 74.1 % of availableSNPs on the IPSC peach 9 K SNP v1 array. Out of 8,144SNP markers on the array, 5,317 (65 %) had GT>0.6 andwere considered for linkage analysis. Although polymor-phism between Clayton and O’Henry was observed in64 % of SNPs, only 33 % (1,764) of the polymorphicSNPs were informative in progeny and could be used inthe linkage analysis. The number of polymorphic/informa-tive SNPs was further reduced to 1,341 (25 %) by removingSNPs with more than 20 % missing data.

Map construction

A genetic linkage map was constructed using a subset of 63progeny. The 1,167 (87 %) SNPs were successfully mappedon 256 map positions in eight linkage groups (Table 2), with130 of them exhibiting segregation distortion. Two hundredand sixty-three SNP markers could not be mapped in the OCpopulation and were removed from further analysis.Approximately 78 % of the mapped SNPs shared the samemap positions, due to the absence of recombination causedby the small number of accessions genotyped. For clarity ofthe figures, a single SNP marker was selected for eachunique position and map figures produced (Fig. 1). In addi-tion, two SSR markers, ssrPaCITA16 and CPPCT006, werealso mapped in linkage groups (LG) 2 and 8, respectively.The average marker density considering 258 markers was1.63 cM/marker. Among mapped SNP markers, 38 deviatedsignificantly from the Chi-square expectations; 26 (10.1 %)and 12 (4.7 %) at the 5 and 1 % threshold, respectively. Thenumber of unique map positions, mapped on each linkagegroup, ranged from 15 in LG5 and LG6 to 63 in LG1, with amean of 27. The average marker density ranged from0.8 cM/marker in LG6 to 2.4 cM/marker in LG2 and LG5.The LGs length was variable, with LG1 being the largest,100.6 cM, and LG6 covering the shortest distance, 12.5 cM.Two gaps larger than 15 cM were observed in LG3 and LG5.

Comparison of the physical and genetic map

Linkage positions of the 95 % of SNP markers in the OClinkage map were in agreement with their positions on thepseudomolecules/scaffolds of peach genome v1.0. Sixregions in the OC map, involving six markers on LG1, sixon LG2, four on LG3, seven on LG7, and two markers onLG8, appeared inverted relative to the physical map(Table 2). Linkage groups 4, 5, and 6 exhibited high homol-ogy with the “dhLovell” physical map. The physical lengthof the OC linkage map was estimated to cover 63 % of thepseudomolecules/scaffolds of peach genome v1.0 (Table 2).The physical length was estimated with the largest coverageon scaffold 1 (96 %), and lowest on scaffold 6 (14 %)(Table 2). In addition, the estimated average coverageper marker on the pseudomolecules/scaffolds ranged from1/200 kb (LG6) to 1/800 kb (LG2) (Table 2).

QTL analysis

QTL analysis was performed for each of the 36 leaf and sixfruit datasets. A total of 14 regions associated with Xapresistance in OC map were detected with at least two inde-pendent analyses (KW, MR, and CIM) and the less stringentthreshold (5 %) for KWor MR. These QTLs were designatedas Xap.Pp.OC-1.1, Xap.Pp.OC-1.2, Xap.Pp.OC-1.3,Xap .Pp .OC-2 .1 , Xap .Pp .OC-2 .2 , Xap .Pp .OC-3 .1 ,Xap .Pp .OC-3 .2 , Xap .Pp .OC-4 .1 , Xap .Pp .OC-4 .2 ,Xap .Pp .OC-5 .1 , Xap .Pp .OC-6 .1 , Xap .Pp .OC-7 .1 ,Xap.Pp.OC-8.1, and Xap.Pp.OC-8.2, according to patho-gen, host species, population, linkage group, and positionfrom the top of the LG (Table 3). The locations and effectsof detected QTLs are summarized in Table 3 and theirlocations in the linkage groups in Fig. 1.

The phenotypic variation explained by the MR analysismodels fitting all the QTLs varied from 15.4 to 56.4 % inleaf datasets and ranged from 33 to 60.7 % in fruit datasets

Table 2 Comparison of the OC linkage map with the peach physical map

Group OC linkage map Coverage (%) Marker density Average coverage (kb/cM)

Marker no. (inverted) Physical length (Mb) Genetic distance (cM) kb cM

G1 63 (6) 45 100.6 96 700 1.6 447

G2 20 (6) 15 47.4 56 800 2.4 316

G3 32 (4) 21 64.2 95 700 2 327

G4 40 24 49.9 80 600 1.2 481

G5 15 6 36.3 33 400 2.4 165

G6 15 3 12.5 14 200 0.8 240

G7 41 (7) 17 63.5 77 400 1.5 268

G8 32 (1) 12 47 57 400 1.5 255

Only 256 SNP markers that were used to represent the map positions were considered for the calculation. One unit of OC genetic map represents1 cM, while 1 unit of the physical map represents 1 Mb

578 Tree Genetics & Genomes (2013) 9:573–586

(Table 3). The phenotypic variation associated only withleaf resistance to Xap ranged from 16.7 to 54.5 %(Table 2). Xap.Pp.OC-4.1 with the strongest effect(>45 %) was detected via KW, CIM, and MR analysismethods by one dataset from SC (2011) and two datasetsfrom NC (2009). Xap.Pp.OC-4.2 was detected by sevendatasets spanning all years and both locations, via KW andMR analysis methods (P<0.05) with phenotypic variationranging from 16.7 to 44.9 %. Xap.Pp.OC-1.1 andXap.Pp.OC-8.2 were detected only by two datasets fromSC (2011) with phenotypic variation ranging from 18.8 to31.4 %, and Xap.Pp.OC-1.3 was detected by four datasetsfrom NC (2008 and 2009) with phenotypic variation from16.5 to 54.5 %.

Xap.Pp.OC-3.1, Xap.Pp.OC-3.2, and Xap.Pp.OC-5.1were involved in both leaf and fruit resistances to Xap withphenotypic variance ranging from 15.4 to 18.1 %. Out ofthose, Xap.Pp.OC-3.1 was detected by three leaf datasetsfrom SC (2011), one leaf dataset from NC (2009), and onefruit dataset from NC (2009) with 15.4 to 18.1 % phenotypicvariance. The phenotypic variance of QTLs associated onlywith fruit resistance to Xap ranged from 33 to 60.7 %(Table 3). Only Xap.Pp.OC-1.2 was detected by threedatasets from both SC (2011) and one dataset from NC(2009) with phenotypic variance ranging from 33 to60.7 %. Although Xap.Pp.OC-6.1 was detected onlyby two datasets from SC (2011) using KW and MRanalysis, it was detected by all six fruit datasets usingCIM analysis with the LOD threshold lowered at 2.0(data not shown).

Additive effects were also calculated to speculate theorigins of resistance alleles (Table 3). Additive effects often QTLs, Xap.Pp.OC-1.1, Xap.Pp.OC-1.2, Xap.Pp.OC-1.3, Xap.Pp.OC-2.1, Xap.Pp.OC-2.2, Xap.Pp.OC-4.1,Xap.Pp.OC-5.1, Xap.Pp.OC-6.1, Xap.Pp.OC-8.1, andXap.Pp.OC-8.2, varied from 0.09 to 1.17. The positivevalues suggested that the resistance alleles originated fromthe resistant parent Clayton, whereas Xap.Pp.OC-7.1showed a negative additive value (−0.78), indicating thepossible contribution of resistance alleles from the suscep-tible parent O’Henry. However, the remaining three QTLsshowed both positive and negative additive effects andrequire further investigation to determine the origins ofresistant alleles.

Four QTLs, Xap.Pp.OC-4.1 associated only with leafresistance, Xap.Pp.OC-5.1 associated with resistance inboth leaf and fruit, and Xap.Pp.OC-1.2 and Xap.Pp.OC-6.1 associated only with fruit resistance, were consideredmajor, based on the size and stability of additive effect andprior knowledge (Yang 2012). In addition, the Xap.Pp.OC-4.2 QTL region has also been included in analysis since itwas detected by datasets originating from all experimentalyears of both SC and NC locations.

Co-location of candidate genes and bacterial spot QTLs

Major QTL regions were compared to the peach genomesequence (www.rosaceae.org) and annotated usingArabidopsis genome sequence to identify potential candi-date genes associated with resistance to bacterial spot. In

Fig. 1 QTLs mapped on the OC linkage map. QTLs are figured withan arrow on the right of the linkage groups. The QTL name reflectspathogen (Xap), species [Prunus persica (Pp)], population (OC), thelinkage group (LG) on which QTLs were identified, and a position

from the top of the LG. Empty circle indicates QTLs associated withleaf resistance; striped circle indicates QTLs associated with both leafand fruit resistance, and filled circle indicates QTL associated with fruitresistance to bacterial spot

Tree Genetics & Genomes (2013) 9:573–586 579

Tab

le3

Sum

maryof

theQTLsdetected

foreach

scoringdatasetby

Kruskal–W

allis

test,multip

leregression

,andcompo

site

interval

mapping

QTL

Datasetsa

LG

Closestmarkerb

KW

Pvalueb

MRpo

sit

Pvalueb

CIM

posit

LODc

LODtd

Add

eR2f

R2tg

Xap

.Pp.OC-1.1

A2-SC11LEA

1SNP_IGA_589

1<0.01

6.3

0.00

36.3

53.5

0.4

12.8

31.4

A3-SC11LEA

1SNP_IGA_1

7833

<0.01

13.1

0.00

5–

––

0.09

–18

.8

MaxA-SC11LEA

1SNP_IGA_5

891

<0.00

56.3

0.00

16.3

3.9

3.4

0.33

–23

.3

Xap

.Pp.OC-1.2

A1-SC11FRU

1SNP_IGA_343

06<0.05

230.01

231.2

5.6

3.6

0.85

30.6

43.6

B1-SC11FRU

1SNP_IGA_3

9717

<0.01

33.6

0.00

1–

––

0.69

21.9

44.1

Max-SC11FRU

1SNP_IGA_4

0295

<0.05

35.2

0.01

232

.34.3

3.4

0.8

20.7

33

D1-NC09

FRU

1SNP_IGA_6

3746

<0.05

43.5

<0.00

143

.64.1

40.85

17.3

60.7

Xap

.Pp.OC-1.3

D2-NC09

LEA

1SNP_IGA_1

0342

2<0.05

65.6

0.00

3–

––

0.6

36.6

54.5

MaxC-N

C08

LEA

1SNP_IGA_112

042

<0.05

75.3

0.00

3–

––

0.3

14.5

34.1

MaxD-N

C09

LEA

1SNP_IGA_11175

5<0.05

73.7

0.01

––

–0.59

36.6

50.8

Max-N

C08

LEA

1SNP_IGA_1

0702

9<0.05

70.4

0.00

8–

––

0.54

–16

.5

Xap

.Pp.OC-2.1

MaxC-N

C09

LEA

2SNP_IGA_1

3725

3<0.05

7.4

0.00

4–

––

0.45

32.6

47.9

Max-N

C09

LEA

2SNP_IGA_1

4035

2<0.00

58.2

0.00

28.3

3.8

3.5

0.53

–20

.6

Xap

.Pp.OC-2.2

Max-N

C09

FRU

2SNP_IGA_2

3807

7<0.01

30.6

0.00

4–

––

0.37

35.4

51.3

Xap

.Pp.OC-3.1

B2-SC09

LEA

3SNP_IGA_2

9543

3<0.05

40.00

1–

––

0.16

–15

.4

A2-SC11LEA

3SNP_IGA_3

0356

4<0.05

11.3

0.01

711.3

3.6

3.5

0.34

19.7

31.4

MaxB-SC09

LEA

3SNP_IGA_2

9543

3<0.01

40.00

3–

––

0.28

–18

.1

MaxC-N

C09

LEA

3SNP_IGA_3

0430

7<0.00

512

.1<0.00

1–

––

−0.58

17.3

47.9

B1-SC11FRU

3SNP_IGA_3

0085

1<0.01

8.9

0.00

1–

––

0.45

23.2

44.1

Xap

.Pp.OC-3.2

B1-SC09

LEA

3SNP_IGA_3

3956

8<0.05

30<0.00

1–

––

−0.08

10.6

34.5

C2-NC09

LEA

3SNP_IGA_3

2516

6<0.00

130

.8<0.00

1–

––

−0.72

–37

.3

D1-NC09

FRU

3SNP_IGA_325

166

–30.8

0.00

830.8

5.4

40.23

24.6

60.7

Xap

. Pp.OC-4.1

A1-SC11LEA

4SNP_IGA_4

0850

5<0.01

12.6

0.00

5–

––

0.23

46.3

56.4

D2-NC09

LEA

4SNP_IGA_4

1160

1<0.00

516

.70.03

916

.74.3

3.7

0.78

45.5

54.5

MaxD-N

C09

LEA

4SNP_IGA_4

1160

1<0.00

516

.7–

16.7

4.3

3.7

0.78

46.1

50.8

Xap

.Pp.OC-4.2

A3-SC09

LEA

4SNP_IGA_4

3918

6<0.00

538

0.00

1–

––

−0.14

–25

.7

B1-SC09

LEA

4SNP_IGA_4

5194

7<0.00

543

.60.00

5–

––

−0.19

21.2

34.5

B1-SC11LEA

4SNP_IGA_4

40116

<0.01

39.6

0.00

5–

––

0.54

–16

.7

D1-NC09

LEA

4SNP_IGA_4

2113

9<0.05

330.00

3–

––

0.57

24.4

45.7

D2-NC09

LEA

4SNP_IGA_4

2095

5<0.01

32.2

0.03

2–

––

0.86

44.9

54.5

MaxC-N

C08

LEA

4SNP_IGA_4

40116

<0.05

39.6

0.00

7–

––

−0.22

17.3

34.1

MaxD-N

C09

LEA

4SNP_IGA_4

2095

5<0.01

32.2

0.04

6–

––

0.86

41.6

50.8

Xap

.Pp.OC-5.1

D1-NC09

LEA

5SNP_IGA_5

9143

9<0.05

2.3

0.00

12.3

4.1

3.9

0.34

19.4

45.7

Max-N

C09

FRU

5SNP_IGA_5

9409

0<0.00

510

.5<0.00

111.3

4.1

3.7

0.94

26.2

51.3

Xap

.Pp.OC-6.1

A1-SC11FRU

6SNP_IGA_6

8253

1<0.01

4.7

<0.00

14.7

3.9

3.6

1.17

1843

.6

Max-SC11FRU

6SNP_IGA_6

8253

1<0.01

4.7

0.00

3–

––

0.92

15.6

33

580 Tree Genetics & Genomes (2013) 9:573–586

total, 37 candidate R genes were found (Table 4): 18 inXap .Pp .OC-1 .2 , four in Xap .Pp .OC-4 .1 , three inXap .Pp .OC -4 .2 , and s ix in Xap .Pp .OC -5 .1 andXap.Pp.OC-6.1 region, respectively. Thirty of them belongto the nucleotide-binding site leucine-rich repeat (NBS-LRR) gene family. Mildew resistance locus, O 11(ppa003706m) on scaffold 4 co-located with Xap.Pp.OC-4.1, a major QTL associated with Xap leaf resistance. Aprotein kinase (ppa015135m) identified in the Xap.Pp.OC-4.2 region could potentially be responsible for bacterial spotleaf resistance. Only one candidate, the NBS-LRR gene(ppb017370m), from the Xap.Pp.OC-5.1 region waspresumably associated with resistance in both leaf andfruit. Interestingly, the remaining five candidate genes(ppa011846m, ppa012062m, ppa016381m, ppa019022m,and ppa025746m) were all annotated as disease resistance-responsive family proteins, which are involved in generaldisease defense in plants. All the candidate R genes from thetwo QTL regions, Xap.Pp.OC-1.2 and Xap.Pp.OC-6.1,belong to NBS-LRR resistance genes. Out of 8,144 SNPsavailable on the IPSC 9K peach SNP array v1, 470 (5.7 %)reside in five major QTL regions on four peach chromo-somes/scaffolds, 118 on LG1, 315 on LG4, 21 on LG5, and16 on LG6, associated with bacterial spot resistance(Supplemental files SM 4–7). Among them, ten SNPmarkers mapped in OC population, SNP_IGA_411340(ppa003706m) and SNP_IGA_422191 (ppa014887m) onscaffold 4; and SNP_IGA_680747 (ppa021741m), SNP_IGA_680857, SNP_IGA_680882, SNP_IGA_680896,SNP_IGA_680901, SNP_IGA_680909, SNP_IGA_680953, and SNP_IGA_680959 (ppa024306m) on scaffold6 (Table 4), are anchored in coding sequences of candidateR genes. No mapped SNP markers in the OC populationwere found in coding sequences of candidate R genes onscaffold 1 or 5.

Discussion

OC genetic map

Development of SNP genetic linkage maps and SNP markerresources for peach have recently been published (Ahmad etal. 2011; Martinez-Garcia et al. 2012; Verde et al. 2012;Eduardo et al. 2012). Estimated SNP frequency of 1/100 innoncoding/intronic and 1/225 in coding/exonic genomeregions have been reported (Sargent et al. 2009; Illa et al.2011). The IPSC peach 9K SNP v1 array contains 8,144high-quality SNPs covering all eight peach chromosomeswith an average spacing of 26.7 kb between SNPs, whichwere all detected in exonic regions of peach genome (Verdeet al. 2012). In our genetic map, the estimated SNP markerdensity was ranging from 0.8 to 2.4 cM or from 1/165T

able

3(con

tinued)

QTL

Datasetsa

LG

Closestmarkerb

KW

Pvalueb

MRpo

sit

Pvalueb

CIM

posit

LODc

LODtd

Add

eR2f

R2tg

Xap

.Pp.OC-7.1

B2-SC11LEA

7SNP_IGA_7

4206

7<0.00

59.6

0.00

1–

––

−0.78

–21

.4

Xap

.Pp.OC-8.1

MaxB-SC11LEA

8SNP_IGA_8

4129

8<0.00

53

0.00

1–

––

0.23

–20

.2

Xap

.Pp.OC-8.2

A1-SC11LEA

8SNP_IGA_8

6779

4<0.00

0135

.8<0.00

1–

––

0.68

21.6

56.4

B3-SC11LEA

8SNP_IGA_871

727

<0.00

540.7

<0.00

140.7

7.1

3.6

0.62

–27.2

The

QTLnamereflectspathog

en(Xap

),ho

stspecies[Prunu

spersica(Pp)],po

pulatio

n(O

C),thelin

kage

grou

pon

which

QTLswereidentified,

andapo

sitio

nfrom

thetopof

theLG.M

ajor

QTLs

areun

derlined

LG

linkage

grou

p,SC

Sou

thCarolina,NCNorth

Carolina,LEAleaf,FRU

fruit

aEachdatasetnamereflectsreplication(A

,B,C,andD),evaluatio

n(1,2,

and3),locatio

n(SCandNC),year

(200

8,20

09,and20

11),andplantorgan(LEA

andfruit)

bClosestmarkerisgivenby

theKruskal–W

allis

test.Pvalueisthesign

ificance

oftheassociationbetweenthemarkerandtheQTL.Thresho

ldwas

setabov

e0.05

cLog

arith

mof

odds

scoreun

dercompo

site

interval

mapping

,thoseQTLsbetweenLOD1(95%)andLOD2(99%)confidence

interval

arebo

lded

dLOD

thresholdun

dercompo

site

interval

mapping

eAdd

itive

effects,ob

tained

from

MRanalysis

fIndividu

alcontribu

tionto

thevariance

accoun

tedforby

theQTL(inpercent),ob

tained

from

MRanalysis

gTotal

variance

explainedby

themod

el(inpercent),ob

tained

from

MRanalysis

Tree Genetics & Genomes (2013) 9:573–586 581

to 1/447 kb. In addition, similar map coverage (421 vs422 cM), number of mapped SNP markers (1,161 vs1,037), shared map positions (256 vs 298), and averagemarker density (1.42 vs 1.48 cM) was obtained in recentlypublished high-density “Pop-DF” SNP peach map(Martinez-Garcia et al. 2012). The accuracy of the high-resolution OC genetic map was confirmed through pairwiserecombination fractions analysis and comparison with thepeach genome assembly v1 (GDR, www.rosaceae.org).Several inversions of SNP marker order (<10 cM) wereobserved in LG1, LG2, LG3, LG7, and LG8 (Table 2).Sixty-seven mapped SNP markers exhibited different orien-tation when compared to peach genome v1.0, similar to 56SNP markers in Pop-DG map (Martinez-Garcia et al. 2012).When comparing the positions of anchor markers betweenT × E and 13 other Prunus maps, Dirlewanger et al. (2004)observed occasional divergences between maps of differentspecies and attributed it to the mapping of different dupli-cates of markers (RFLPs or SSRs) that have more than onecopy in different regions of the Prunus genome. Moreover,order inversions almost always affected pairs of loci that

are close together in the T × E map (~10 cM), suggestingthat they were rather caused by errors in the assignment ofmarker order than to inversion of chromosome fragments.Only one major chromosomal rearrangement has been docu-mented in peach, a reciprocal translocation between G6 andG8 which was demonstrated in the F2 progeny of almond(cv. Garfi) × peach (cv. Nemared) (Jauregui et al. 2001) andin the peach F2 cv. Akame × cv. Juseitou (Yamamoto et al.2001). The OC map also has one inverted region larger than15 cM on the upper part of LG2 (Table 2) that might be dueto the translocation of chromosome fragments. Genotypingof more progeny from this population is necessary to support ahypothesis of the chromosome fragment translocation.

Genetic basis of quantitative resistance to Xap

Our study indicates that Xap resistance in peach is aquantitative trait controlled by polygenic factors, which issupported by the evidence in the literature where cultivarsreported resistant have quite diverse pedigrees (Okie1998). Thus, the rating method applied in our research was

Table 4 Prediction of candidate Xap resistance genes in five QTL regions

QTL region Scaffold Predicted R gene accession no. Functional annotation Mapped SNPs in coding region

Xap.Pp.OC-1.2 1 ppa015521m, ppa016009m,ppa017163m, ppa017506m,ppa018004m, ppa018388m,ppa018885m, ppa019235m,ppa019783m, ppa019910m,ppa019915m, ppa020437m,ppa021678m, ppa023410m,ppa023526m, ppa023712m,ppa024868m, ppa025202m

NBS-LRR resistance gene family –

Xap.Pp.OC-4.1 4 ppa003706m Mildew resistance locus O 11 SNP_IGA_411340

ppa016517m, ppa021560m,ppa026627m

NBS-LRR resistance gene family –

Xap.Pp.OC-4.2 ppa014887m NBS-LRR resistance gene family SNP_IGA_422191

ppa015135m Protein kinase –

ppa026334m NBS-LRR resistance gene family –

Xap.Pp.OC-5.1 5 ppa011846m, ppa012062m,ppa016381m, ppa019022m,ppa025746m

Disease resistance-responsive familyproteins

–

ppb017370m NBS-LRR resistance gene family –

Xap.Pp.OC-6.1 6 ppa021741m NBS-LRR resistance gene family SNP_IGA_680747

ppa024306m NBS-LRR resistance gene family SNP_IGA_680857,SNP_IGA_680882,SNP_IGA_680896,SNP_IGA_680901,SNP_IGA_680909,SNP_IGA_680953,SNP_IGA_680959

ppa000373m, ppa016692m,ppa019076m, ppa1027168m

NBS-LRR resistance gene family –

The predicted R gene accession number is the name of transcript model found in GDR database (www.rosaceae.org)

NBS-LRR nucleotide-binding site leucine-rich repeat

582 Tree Genetics & Genomes (2013) 9:573–586

critical for QTL mapping, and a strong attempt was made toensure accuracy and precision of the applied score. In eachcycle of evaluation, visual symptoms were rated twice oneach tree to assess the whole tree performance using theordinal scale method, which is deemed more reliable andaccurate (Bardsley and Ngugi 2010). Recently, a PCRmethod for detection of a specific ABC transporter gene ofXap was developed to facilitate detection of disease in thefield (Pagani 2004; Palacio-Bielsa et al. 2011). Maximalscore, which was deemed as a good parameter in revealingpotential for disease severity development and genetic basisfor it in each genotype, was used to detect the potentialQTLs (Rubio et al. 2010). Datasets acquired from each cyclewere also used to capture additional QTLs in order toelucidate genetic control of Xap resistance in differentenvironments (Rubio et al. 2010).

Our findings suggest that the leaf and fruit resistance toXap in peach are regulated by different QTLs, which is inagreement with the report of Werner et al. (1986). In ourstudy, we detected a total of 14 QTLs involved in Xapresistance. This is higher than the number of genes associ-ated with bacterial spot resistance reported in pepper (six)and tomato (five) (Stall et al. 2009), but similar to rice (19)(Nino-Liu et al. 2006). The QTL Xap.Pp.OC-4.1, with themajor effects of R2>45 %, was associated only with leafresistance to Xap and co-localized with the QTL region onLG4 detected in our previous study (Yang 2012).Furthermore, this QTL region includes the marker AG8Aon LG4 of the Prunus resistance map, which is associatedwith powdery mildew resistance (Lalli et al. 2005). Anotherputative Xap.Pp.OC-4.2 co-localizes with the SSR markerBPPCT036 on LG4 of the Prunus resistance map and is alsoassociated with powdery mildew resistance (Lalli et al.2005). These findings suggest pleiotropic effects, indicatingthat this region of the peach genome harbors genes associ-ated with resistance to both bacterial spot and powderymildew in peach. In addition, Grube et al. (2000) suggestedthat highly similar R genes may confer resistance to differ-ent pathogen types, while highly similar pathogen races mayemploy different R genes. On the other hand, Xap.Pp.OC-3.1 and Xap.Pp.OC-5.1 were detected by both leaf and fruitdatasets and seem to co-localize with the QTLs on LG3 andLG5 also reported in apricot (Socquet-Juglard et al. 2011).Higher resolution maps and a set of shared markers betweenthe OC and apricot genetic maps are necessary to confirm ifit is indeed the same region in both species responsible for Xapresistance inPrunus. Additionally, twoQTLs,Xap.Pp.OC-1.2and Xap.Pp.OC-6.1, associated only with fruit datasets werealso detected in the OC population. These findings reveal thecomplexity of Xap resistance in peach and suggest existenceof different genes involved in leaf and fruit resistance as wellas general resistance genes that elicit a resistant response toboth leaf and fruit Xap infections in peach.

For seven of the putative QTLs identified in this study,Xap .Pp .CO-1 .1 , Xap .Pp .CO-1 .2 , Xap .Pp .CO-1 .3 ,Xap.Pp.CO-2.1, Xap.Pp.CO-4.1, Xap.Pp.CO-6.1, andXap.Pp.CO-8.2, favorable alleles conferring high resistancewere inherited from the resistant parent Clayton as expected,suggesting that breeders were successful in pyramiding re-sistance genes using only field phenotyping. However, oneQTL, Xap.Pp.OC-7.1 with a favorable allele for resistanceseems to originate from the susceptible parent O’Henry. Thepedigree analysis suggests that the resistant alleles fromO’Henry may originate from the grandparent “J. H. Hale,”which is a mildly susceptible cultivar. Resistant alleles fromsusceptible parents were indicated in previous reports forvarious plant–pathogen interactions (Young et al. 1993;Dirlewanger et al. 1994, 1996; Mestries et al. 1998; Kelleret al. 1999; Foulongne et al. 2003). Since O’Henry is highlysusceptible to Xap, leaf and fruit results suggest existence ofrecessive alleles for Xap resistance. Recessive alleles con-ferring resistance to pathogens have previously beenreported in other plant species, such as pepper and tomato(Stall et al. 2009) and rice (Nino-Liu et al. 2006). Among 63individuals studied, only two, no. 31 and no. 192, have thefavorable combination of the major QTLs for bacterial spotresistance in both leaf and fruit and fruit only, respectively.However, further characterization of SNP haplotypes pres-ent in the major QTL regions is necessary to help breedersidentify strategies for Xap marker-assisted selection (MAS)in peach.

Implications for MAB for bacterial spot in peach

Four types of candidate genes associated with resistance tobacterial spot were identified in our study. Thirty candidateR genes were annotated as NBS-LRR type in all the fiveQTL regions. NBS-LRR proteins are well-known to guardthe plant by recognizing pathogen effector proteins andinitiating hypersensitive reaction in “gene-for-gene” resis-tance response. The largest class of R genes involved indisease resistance contain NBS-LRR genes, i.e., 149 inArabidopsis, 317 in Populus, 480 in rice (Kohler et al.2008), and approximately 424 in peach (Yang 2012). Inaddition, candidate R gene ppa015135m, detected on scaf-fold 4, belongs to protein kinases, which are known totrigger the downstream signaling for the induction of raceand non-race-specific defense in plants (Romeis 2001).Several isolated R genes confer protein kinase activitiessuch as Pto in tomato (Martin et al. 1993), Xa21 in rice(Song et al. 1995), and FLS2 in Arabidopsis (Gomez-Gomez and Boller 2000). Another interesting candidategene, detected also on scaffold 4, is mildew resistance locusO 11 (MLO11) (ppa003706m) that belongs to the MLOgene family involved in plant defense against powderymildew infection reported in barley (Schweizer et al. 2000;

Tree Genetics & Genomes (2013) 9:573–586 583

Douchkov et al. 2005), Arabidopsis (Consonni et al. 2006),tomato (Bai et al. 2008), and grape (Feechan et al. 2008;Winterhagen et al. 2008). Moreover, recent reports of apepper MLO gene related to bacterial and oomycete propa-gation in pepper and Arabidopsis and involved in cell deathprocess (Kim and Hwang 2012) warrant further investiga-tion into the involvement of candidate gene ppa003706m inbacterial spot resistance in peach. Finally, five of the sixcandidate R genes detected within the region of Xap.Pp.OC-5.1, major QTL on scaffold 5 associated with resistance inboth leaf and fruit, co-located with general pathogen-responsive proteins, such as glucanases, chitinases, andphytoalexins.

Ten mapped SNP markers anchored in candidate Rgenes, one each in ppa003706m and ppa014887m on scaf-fold 4 and one in ppa021741m and six in ppa024306m onscaffold 6, are good candidates for MAS/MAB for bacterialspot resistance in Prunus. Further analysis of mean progenyphenotype and SNP haplotypes will provide stronger evi-dence of their involvement with resistance/susceptibility andsuitability for MAS.

Conclusions

Introgression of Xap resistance or tolerance has been initi-ated in many peach breeding programs. However, the poly-genic character of Xap resistance makes traditional breedingtime consuming and labor intensive. Five main QTLs wereconsidered for the marker development and future MAB,including Xap.Pp.OC-4.1 and Xap.Pp.OC-4.2 associatedonly with leaf resistance, Xap.Pp.OC-5.1 associated withboth leaf and fruit resistance, and two QTLs, Xap.Pp.OC-1.2 and Xap.Pp.OC-6.1, associated only with fruit resis-tance. Our study supports breeding strategies for thedevelopment of Xap-resistant peach cultivars based onmarker-assisted selection of favorable QTLs in advancedgenerations. It also suggests that an advisable strategyto ensure a stable level of Xap resistance in both leafand fruit would be to combine favorable alleles at thesefour QTLs in the same genotype. However, achievingthis combination solely through phenotypic selectionwill be difficult since it is hard to control the environ-mental conditions and pathogen population in the field.Therefore, development of markers associated with Xapresistance for application in MAB would be very usefulin that regard.

Acknowledgments This work was supported by NIFA/USDA underproject number SC-1700382 and technical contribution no. 6045 of theClemson University Experiment Station. This work was partiallyfunded by USDA’s National Institute of Food and Agriculture–Specialty Crop Research Initiative project, “RosBREED: enablingmarker-assisted breeding in Rosaceae” (2009-51181-05808).

Data Archiving Statement OC linkage map and QTL positions areavailable on Genome Database for Rosaceae (www.rosaceae.org).

References

Ahmad R, Parfitt DE, Fass J, Ogundiwin E, Dhingra A, Gradziel TM,Lin DW, Joshi NA, Martinez-Garcia PJ, Crisosto CH (2011)Whole genome sequencing of peach (Prunus persica L.) forSNP identification and selection. BMC Genomics 12:569

Arus P, Verde I, Sosinksi B, Zhebentyayeva T, Abbott A (2012) Thepeach genome. Tree Genet Genomes 8:531–547

Bai Y, Pavan S, Zheng Z, Zappel NF, Reinstadler A, Lotti C, DeGiovanni C, Ricciardi L, Lindhout P, Visser R, Theres K,Panstruga R (2008) Naturally occurring broad-spectrum powderymildew resistance in a Central American tomato accession iscaused by loss of mlo function. Mol Plant-Microbe Interact21:30–39

Bailey JS, French AP (1949) The inheritance of certain fruit andfoliage characters in the peach. Mass Agr Expt Sta Bull, pp 452

Bardsley SJ, Ngugi HK (2010) Reliability and accuracy of visualmethods used to quantify foliar symptoms of bacterial spot ofpeach and nectarine. Phytopathology 100:S11

Blake MA (1932) The JH Hale peach as a parent in peach crosses. ProcAm Soc Horti Sci 29:131–136

Broman KW, Wu H, Sen S, Churchill GA (2003) R/qtl: QTL mappingin experimental crosses. Bioinformatics 19:889–890

Conesa A, Gotz S, Garcia-Gomez JM, Terol J, Talon M, Robles M(2005) Blast2GO: a universal tool for annotation, visualizationand analysis in functional genomics research. Bioinformatics21:3674–3676

Consonni C, Humphry ME, Hartmann HA, Livaja M, Durner J,Westphal L, Vogel J, Lipka V, Kemmerling B, Schulze-Lefert P,Somerville SC, Panstruga R (2006) Conserved requirement for aplant host cell protein in powdery mildew pathogenesis. NatGenet 38:716–720

Dirlewanger E, Isaac PG, Ranade S, Belajouza M, Cousin R, DeVienne D (1994) Restriction fragment length polymorphism anal-ysis of loci associated with disease resistance genes and develop-mental traits in Pisum sativum L. Theor Appl Genet 88:17–27

Dirlewanger E, Pascal T, Zuger C, Kervella J (1996) Analysis ofmolecular markers associated with powdery mildew resistancegenes in peach [Prunus persica (L.) Batsch] × Prunus davidianahybrids. Theor Appl Genet 93:909–919

Dirlewanger E, Graziano E, Joobeur T, Garriga-Caldere F, Cosson P,Howad W, Arus P (2004) Comparative mapping and marker-assisted selection in Rosaceae fruit crops. PNAS 101:9891–9896

Dondini L, Lain O, Geuna F, Banfi R, Gaiotti F, Tartarini S, Bassi D,Testolin R (2007) Development of a new SSR-based linkage mapin apricot and analysis of synteny with existing Prunus maps.Tree Genet Genomes 3:239–249

Douchkov D, Nowara D, Zierold U, Schweizer P (2005) A high-throughput gene-silencing system for the functional assessmentof defense-related genes in barley epidermal cells. Mol Plant-Microbe Interact 18:755–761

Dunegan JC (1932) The bacterial spot disease of the peach and otherstone fruits. Tech Bull US Dep Agric 273:53

Eduardo I, Chietera G, Pirona R, Pacheco I, Troggio M, Banchi E,Bassi D, Rossini L, Vecchietti A, Pozzi C (2012) Genetic dissectionof aroma volatile compounds from the essential oil of peach fruit:QTL analysis and identification of candidate genes using dense SNPmaps. Tree Genet Genomes. doi:10.1007/s11295-012-0546-z

EPPO/CABI (1997) Xanthomonas arboricola pv. pruni. Quarantine pestsfor Europe, 2nd edn. CABI, Wallingford, UK, pp 1096–1100

584 Tree Genetics & Genomes (2013) 9:573–586

EPPO/OEPP (2006) Data sheets on quarantine pests. Xanthomonasarboricola pv. pruni. http://www.eppo.org/QUARANTINE/listA2.htm.

Feechan A, Jermakow AM, Torregrosa L, Panstruga R, Dry IB (2008)Identification of grapevine MLO gene candidates involved insusceptibility to powdery mildew. Funct Plant Biol 35:1255–1266

Foulongne M, Pascal T, Pfeiffer F, Kervella J (2003) QTLs forpowdery mildew resistance in peach × Prunus davidiana crosses:consistency across generations and environments. Mol Breed12:33–50

Gomez-Gomez L, Boller T (2000) FLS2: a LRR receptor-like kinaseinvolved in recognition of the flagellin elicitor in Arabidopsis.Mol Cell 5:1–20

Grube RC, Radwanski ER, Jahn M (2000) Comparative genetics ofdisease resistance within the solanaceae. Genetics 155:873–887

Illa E, Eduardo I, Audergon J, Barale F, Dirlewanger E, Li X, MoingA, Lambert P, Le Dantec L, Gao Z, Poessel J-L, Pozzi C, RossiniL, Vecchietti A, Arus P, Howad W (2011) Saturating the Prunus(stone fruits) genome with candidate genes for fruit quality. MolBreed 28:667–682

Jauregui B, de Vicente MC, Messeguer R, Felipe A, Bonnet A,Salesses G, Arus P (2001) A reciprocal translocation between‘Garfi’ almond and ‘Nemared’ peach. Theor Appl Genet102:1169–1176

Keil HL, Fogle HW (1974) Orchard susceptibility of some apricot,peach, and plum cultivars and selections to Xanthomonas pruni.Fruit Var J 28:16–19

Keller M, Keller B, Schachermayr G, Winzeler M, Schmid JE, StampP, Messmer MM (1999) Quantitative trait loci for resistanceagainst powdery mildew in a segregating wheat × spelt population.Theor Appl Genet 98:903–912

Kim, D, S, and Hwang, B, K, (2012) The pepper MLO gene,CaMLO2, is involved in the susceptibility cell-death responseand bacterial and oomycete proliferation. The Plant Journal.doi:10.1111/tpj.12003

Kohler A, Rinaldi C, Duplessis S, Baucher M, Geelen D, DuchaussoyF, Meyers BC, Boerjan W, Martin F (2008) Genome-wideidentification of NBS resistance genes in Populus trichocarpa.Plant Mol Bio 66:619–636

Kosambi DD (1944) The estimation of map distance from recombinationvalues. Ann Eugen 12:172–175

Kruglyak L, Lander ES (1995) A non-parametric approach formapping quantitative trait loci. Genetics 139:1421–1428

Lalli DA, Decroocq V, Blenda AV, Schurdi-Levraud V, Garay L, LeGall O, Damsteegt V, Reighard GL, Abbott AG (2005)Identification and mapping of resistance gene analogs (RGAs)in Prunus: a resistance map for Prunus. Theor Appl Genet111:1504–1513

Martin GB, Brommonschenkel SH, Chunwongse J, Frary A, GanalMW, Spivey R, Wu T, Earle ED, Tanksley SD (1993) Map-basedcloning of a protein kinase gene conferring disease resistance intomato. Science 262:1432–1436

Martinez-Garcia P; Parfitt D; Ogundiwin E; Fass J; Chan H; Ahmad R;Lurie S; Dandekar A; Gradziel T, Crisosto C (2012) High densitySNP mapping and QTL analysis for fruit quality characteristics inpeach (Prunus persica L.). Tree Genet Genomes:1–18,doi:10.1007/s11295-012-0522-7

Mestries E, Gentzbittel L, Tourvieille de Labrouhe D, Nicolas P,Vear F (1998) Analysis of quantitative trait loci associatedwith resistance to Sclerotinia sclerotiorum in sunflowers(Helianthus annuus L.) using molecular markers. Mol Breed4:215–226

Nino-Liu DO, Ronald PC, Bogdanove AJ (2006) Xanthomonas oryzaepathovars: model pathogens of a model crop. Mol Plant Pathol7:303–324

Okie WR (1984) Rapid multiplication of peach seedlings byherbaceous stem cuttings. HortSci 19:249–251

Okie WR (1998) Handbook of peach and nectarine varieties—performance in the southeastern United States and index ofnames. USDA/ARS Agriculture

Okie WR, Bacon T, Bassi D (2008) Fresh market cultivar develop-ment. In: Layne DR, Bassi D (eds) The peach: botany, productionand uses. CABI, Cambridge, pp 139–174

Pagani MC (2004) An ABC transporter protein and moleculardiagnoses of Xanthomonas arboricola pv. pruni causingbacterial spot of stone fruits. Raleigh, USA: University of NorthCarolina State PhD thesis. http://www.lib.ncsu.edu/theses/available/etd-10042004-232356

Palacio-Bielsa A, Cubero J, Cambra MA, Collados R, Berruete IM,Lopez M (2011) Development of an efficient real-time quantita-tive PCR protocol for detection of Xanthomonas arboricola pv.pruni in Prunus species. Appl Environ Microbiol 77:89–97

Ritchie DF (1995) Bacterial spot. In: Ogawa JM, Zehr EI, Bird GW(eds) Compendium of stone fruit diseases. APS Press, St. Paul,pp 50–52

Rolfs FM (1915) A bacterial disease of stone fruits. Coli Agric Mem,Cornell Univ Agric Exp Stn 8:381–436

Romeis T (2001) Protein kinases in the plant defence response. CurrOpin Plant Biol 4:407–414

Rubio M, Pascal T, Bachellez A, Lambert P (2010) Resistance ofalmond cultivars to Plum pox virus (sharka). Tree GenetGenomes 6:291–304

Sargent DJ, Marchese A, Simpson DW, Howad W, Fernandez-Fernandez F, Monfort A, Arus P, Evans KM, Tobutt KR(2009) Development of “universal” gene-specific markersfrom Malus spp. cDNA sequences, their mapping and usein synteny studies within Rosaceae. Tree Genet Genomes5:133–145

Schweizer P, Pokorny J, Schulze-Lefert P, Dudler R (2000) Technicaladvance. Double-stranded RNA interferes with gene function atthe single-cell level in cereals. Plant J 24:895–903

Sherman WB, Lyrene PM (1981) Bacterial spot susceptibility in lowchilling peaches. Fruit Var J 35:74–77

Shulaev V, Korban SS, Sosinski B, Abbott AG, Aldwinckle HS, FoltaKM, Iezzoni A, Main D, Arus P, Dandekar AM, Lewers K,Brown SK, Davis TM, Gardiner SE, Potter D, Veilleux RE(2008) Multiple models for Rosaceae genomics. Am Soc PlantBiol 147:985–1003

Simeone AM (1985) Study on peach and nectarine cultivars suscepti-bility to the main fungus and bacteria. Acta Hort 173:541–551

Smith EF (1903) Observation on a hitherto unreported bacterial disease,the cause of which enters the plant through ordinary stomata.Science 17:456–457

Socquet-Juglard D, Christen D, Gessler C, Duffy B, Patocchi A (2011)Mapping QTL resistance in apricot against Xanthomonasarboricola pv. pruni. XIII Eucarpia Symposium on FruitBreeding and Genetics, Warsaw

Song WY, Wang GL, Chen LL, Kim HS, Pi LY, Holsten T, Gardner J,Wang B, Zhai WX, Zhu LH, Fauquet C, Ronald P (1995) Areceptor kinase-like protein encoded by the rice disease resistancegene Xa21. Science 270:1804–1806

Sosinski B, Shulaev V, Dhingra A, Kalyanaraman A, Bumgarner R,Rokhsar D, Verde I, Velasco R, Abbott AG (2009) Rosaceaousgenome sequencing: perspectives and progress. In: Folta KM,Gardiner SE (eds) Genetics and genomics of Rosaceae, plantgenetics and genomics: crops and models 6. Springer, Science+Business Media, LLC, pp 601–615

Sosinski B, Verde I, Morgante M, Rokhsar D (2010) A first draft of thepeach genome sequence and its use for genetic diversity analysisin peach. 5th International Rosaceae Genomics Conference,Stellenbosch, South Africa

Tree Genetics & Genomes (2013) 9:573–586 585

Stall RE, Jones JB, Minsavage GV (2009) Durability of resistance intomato and pepper to xanthomonads causing bacterial spot. AnnuRev Phytopathol 47:265–284

Van Ooijen JW (2006) JoinMap® 4.0: software for the calculation ofgenetic linkage maps in experimental populations. Kyazma BV,Wageningen

Van Ooijen JW (2009) MapQTL® 6: software for the mapping ofquantitative trait loci in experimental populations of diploidspecies. Kyazma BV, Wageningen

Verde I, Bassil N, Scalabrin S, Gilmore B, Lawley CT, Gasic K,Micheletti D, Rosyara UR, Cattonaro F, Vendramin E, Main D,Aramini V, Blas AL, Mockler TC, Bryant DW, Wilhelm L,Troggio M, Sosinski B, Aranzana MJ, Arus P, Iezzoni A,Morgante M, Peace C (2012) Development and evaluationof a 9K SNP array for peach by internationally coordinatedSNP detection and validation in breeding germplasm. PLoSOne 7:1–13

Voorrips RE (2002) MAPCHART: software for the graphical presen-tation of linkage maps and QTLs. J Hered 93:77–78

Wang S, Basten CJ, Zeng ZB (2007) Windows QTL cartographer 25.Department of Statistics, North Carolina State University, Raleigh

Werner DJ, Ritchie DF, Cain DW, Zehr EI (1986) Susceptibility ofpeaches and nectarines, plant introductions, and other Prunusspecies to bacterial spot. HortSci 21:127–130

Winterhagen P, Howard SF, Qui W, Kovacs LG (2008) Transcriptionalupregulation of grapevine MLO genes in response to powderymildew infection. AMJ Enol Viticult:159–168

Yamamoto T, Shimada T, Imai T, Yaegaki T, Haji T, Matsuta N,Yamaguchi M, Hyashi T (2001) Characterization of morpholog-ical traits based on a genetic linkage map in peach. Breeding Sci51:271–278

Yang N (2012) Mapping quantitative trait loci associated withresistance to bacterial spot (Xanthomonas arboricola pv. pruni)in peach. Dissertation, Clemson University

Yang N, Righard GL, Ritchie D, Okie WR, Gasic K (2010)Construction of a genetic linkage map for identification ofmolecular markers associated with resistance to Xanthomonasarboricola pv. pruni in peach [Prunus persica (L.) Batsch][abstract]. HortSci 45:S304

Yang N, Reighard GL, Ritchie D, Okie WR, Gasic K (2011)Development of a genetic linkage map for identification ofmolecular markers associated with resistance to bacterial spot(Xanthomonas arboricola pv. pruni) in peach. XIII EucarpiaSymposium on Fruit Breeding and Genetics, September 11–15,2011 Warsaw, Poland

Young ND, Danesh D, Menancio-Hautea D, Kumar L (1993) Mappingoligogenic resistance to powdery mildew in mungbean withRFLPs. Theor Appl Genet 87:243–249

586 Tree Genetics & Genomes (2013) 9:573–586