Genetic Analysis of Resistance to Tomato Late Blight in Solanum Pimpinellifolium Accession PI 163245

Genetic analysis of resistance to tomato late blight in Solanum pimpinellifolium accession PI 163245

E RIK W. OHLSON and M AJID R. FOOLAD* Department of Plant Science and the Intercollege Graduate Degree Program in Genetics, The Pennsylvania State University, University Park, PA 16802, USA; *Corresponding author, E-mail: [email protected] With 3 figures and 1 table Received October 9, 2015 / Accepted February 6, 2016 Communicated by M. Havey

Abstract and Goodwin 1997, Mayton et al. 2001, Smart and Fry 2001). Late blight (LB), caused by Phytophthora infestans, is one of the most However, many strains of P. infestans have developed resistance devastating diseases of tomato (Solanum lycopersicum) worldwide. Due to phenylamides, one of the few effective systemic treatments to the emergence of new and aggressive P. infestans isolates, identifying for controlling LB infection (Gisi and Cohen 1996, Goodwin new genetic resistance to LB is a priority in tomato breeding. Recently, et al. 1996). While application of preventative control measures we reported the identification of several Solanum pimpinellifolium acces- such as mefenoxam or copper-based fungicides can be effective, sions with strong LB resistance. In this study, we investigated the utility frequent applications have financial and environmental conse- of resistant accession PI 163245 for tomato breeding by examining heri- 2 quences and may only prevent infection by certain genotypes of tability (h ) of resistance and the response to selection for resistance. the pathogen (McGrath et al. 2013). Identifying new sources of Estimates of h2 based on F :F and F :F parent : offspring correla- 2 3 3 4 LB resistance and incorporating resistance into elite breeding tion analyses averaged 0.79 and 0.94, respectively, suggesting the heritable nature of LB resistance in PI 163245. Analysis of response to lines and hybrid cultivars is an effective strategy to counter new 2 aggressive strains of P. infestans and reduce fungicide applica- selection for resistance from F2 to F4 generations indicated a realized h of 0.63, confirming the utility of this resistance in tomato breeding. Two tions (Foolad et al. 2008, Nowicki et al. 2013). methods of estimating the minimum number of loci involved indicated Three LB resistance genes are currently known for use in the presence of one major resistance locus. Currently, genetic mapping tomato breeding: Ph-1, Ph-2 and Ph-3. Ph-1, located on chromo- and breeding efforts are underway to further confirm the viability of this some 7, is a dominant resistance gene initially identified in accession for improving tomato LB resistance. accessions West Virginia 19 and 731 (Bonde and Murphy 1952, Ph-1 Key words: disease resistance — parent–offspring regression Gallegly and Marvel 1955, Peirce 1971). While provides P. infestans analysis — qualitative resistance — resistance breeding — resistance to race T-0, it is ineffective against the Solanum lycopersicum L. — vertical resistance predominant race T-1, and as a result, the gene is no longer useful in breeding for LB resistance (Peirce 1971, Foolad et al. 2014). The resistance genes Ph-2 and Ph-3 are currently the most effective forms of LB resistance in tomato. Ph-2 confers The cultivated tomato, Solanum lycopersicum L., is the most partial resistance and was first identified in the tomato accession popular and economically important vegetable crop in the world, West Virginia 700 (Gallegly and Marvel 1955), and subse- valued at almost $60 billion annually (faostat3.fao.org). It is sec- quently mapped to the long arm of chromosome 10 (Moreau ond only to potato in volume consumed and accounts for a sig- et al. 1998). While Ph-2 is not considered a particularly strong nificant portion of vitamins and minerals in many diets form of LB resistance (Goodwin et al. 1995, Black et al. 1996, (faostat3.fao.org). However, tomato is susceptible to more than Foolad et al. 2014), a very high level of resistance is obtained 200 diseases caused by fungi, bacteria, viruses and nematodes when it is combined with Ph-3 (Gardner and Panthee 2010a,b, (Lukyanenko 1991). As such, breeding for disease resistance is Panthee and Gardner 2010). Ph-3 was identified in the Solanum vital for improving yield and ensuring adequate production to pimpinellifolium accession L3708 (a.k.a. LA1269 and PI meet worldwide demand. One disease of particular importance is 365957) (AVDRC, 1994), located on the long arm of chromo- late blight (LB), caused by the oomycete Phytophthora infestans some 9 (Chunwongse et al. 1998), and recently fine mapped and (Mont.) de Bary, which is responsible for approximately 7% of cloned (Zhang et al. 2013, 2014). Furthermore, a second minor tomato yield losses annually in the United States and accounts LB resistance QTL was recently identified in L3708 and mapped for similar losses elsewhere in the world (Nowicki et al. 2012). to chromosome 2 (Chen et al. 2014). Despite the strong LB Identification and exploitation of new sources of genetic resis- resistance conferred by Ph-3, tomato cultivars containing this tance to LB is a priority in tomato breeding due to the occur- resistance gene alone have displayed susceptibility to LB when rence of new, more aggressive strains of the pathogen (Foolad exposed to intense disease pressure or particularly aggressive et al. 2008, 2014). P. infestans isolates (Chunwongse et al. 2002, Foolad et al. Phytophthora infestans, the causal agent of the Irish Potato 2008; R. G. Gardner, personal communication). Consequently, Famine, is one of the most destructive diseases of tomato and identification and introgression of additional forms of genetic potato (Solanum tuberosum L.) worldwide, and capable of resistance into new tomato cultivars is expected to be desirable destroying susceptible tomato and potato crops within 7–10 days and may increase the durability of LB resistance. of first colonization (Foolad et al. 2008, Nowicki et al. 2012). Additional LB resistance genes/QTLs have been identified and Until the late 1970s, LB was effectively controlled using fre- studied in several wild species of tomato, but have yet to be quent fungicide application and proper cultural practices (Fry effectively utilized in breeding. For example, race non-specific 392 E. W. OHLSON AND M. R. FOOLAD resistance QTLs were reported in Solanum habrochaites progeny, which were used for disease evaluation and estimation of accession LA2099 on all 12 chromosomes (Brouwer et al. heritability of LB resistance (described below). In all disease-screening 2004). The three strongest QTLs were fine mapped to chromo- experiments, in addition to the parental lines and filial populations, somes 4, 5 and 11 (Brouwer and St. Clair 2004). Resistance was several inbred lines, including NC 84173 (LB susceptible), New Yorker also reported in the S. habrochaites accession BGH6902 and as (containing Ph-1), NC 63EB (Ph-2), NC 870 (Ph-3) and NC 03220 (Ph-2 + Ph-3), were used as control genotypes. Seed of the control lines many as 28 QTLs were detected to contribute to the observed was kindly provided by R.G. Gardner, North Carolina State University, resistance (Abreu et al. 2008). However, the quantitative nature Mills River, NC. of the resistance reported in these accessions has limited their usefulness in tomato breeding. Traditional breeding strategies Inoculum preparation: The P. infestans isolate RS2009T1, belonging and marker-assisted selection (MAS) have been unable to utilize to the US clonal lineage US-23, race T-1 and mating type A1, was used these LB resistance QTLs without introducing undesirable char- in all disease evaluations. RS2009T1 was originally collected from a acteristics from S. habrochaites, including small fruit size, poor commercial tomato field in Rock Springs, PA, in 2009. This clonal growth habit, and late maturity due to linkage drag (Brouwer lineage was selected for its aggressive infection rates in tomato and its and St. Clair 2004). Additionally, effects of these QTLs were widespread natural occurrence throughout the north-eastern USA often relatively small and had low heritability (Abreu et al. (Gugino and Foolad 2013, Foolad et al. 2014, Gugino et al. 2014). The 2008). Resistance to LB has also been reported on chromosome P. infestans isolate was established on susceptible tomato leaflets placed 9 6 in the wild tomato species Solanum pennellii, although this abaxial (lower) side up in 100 15 mm Petri dishes. The Petri dishes contained a thin layer of 1.7% water agar to maintain high humidity. QTL is linked to the SP allele and it has been suggested this The infected tomato leaflets were maintained in an incubator at 14–16°C, resistance is potentially a result of indeterminate growth habit 100% humidity and 12-h photoperiod for 7–11 days. The inoculum was rather than a true resistance gene (Smart et al. 2007). Recently, prepared by placing the leaflets in 4°C distilled water and incubating at new LB resistance genes were identified and mapped to chromo- 4°C for 1 h to facilitate zoospore release. After briefly vortexing the somes 1 and 10 in S. pimpinellifolium accession PI 270443, inoculum to dislodge sporangia, the suspension was filtered through which displays strong LB resistance comparable to Ph-2 + Ph-3 cheesecloth to remove leaf debris, and subsequently adjusted to 10 000 combined (Merk et al. 2012). sporangia/ml using a light microscope and haemocytometer. Solanum pimpinellifolium is the most closely related wild = tomato species to the cultivated tomato (Miller and Tanksley F2 disease evaluation: The F2 population (n 560), parental lines and 1990, Peralta and Spooner 2000) and is the most well-utilized control genotypes were grown in 72-cell seedling flats in an isolated, source of LB resistance in tomato breeding. In order to identify environmentally controlled greenhouse section. For each of the parental additional sources of LB resistance, a screening of 67 and control lines, four replications of six plants were grown and placed on opposite ends and sides of the greenhouse section. After six weeks, S. pimpinellifolium accessions was undertaken to evaluate the greenhouse temperature was adjusted to 16–18°C and high-pressure response to LB infection under field and greenhouse conditions foggers were initiated to increase relative humidity to 95–100%. Clear (Foolad et al. 2014). From this study, several accessions were plastic was draped around each greenhouse bench preventing plants from identified that exhibited strong resistance to LB. The accession being directly exposed to the high-pressure fog. Additionally, ambient PI 163245 displayed very high LB resistance under different lighting was reduced using blackout curtains to suppress hypersensitive field conditions and resistance to eight P. infestans isolates from and salicylic acid plant defence responses (Griebel and Zeier 2008, three US clonal lineages in greenhouse trials. Furthermore, Roden and Ingle 2009). After six hours, the high-pressure foggers were molecular markers associated with Ph-2 and Ph-3 resistance turned off and the plants were misted with water. Thirty minutes later, indicated LB resistance exhibited by PI 163245 might not be the P. infestans inoculum was uniformly sprayed on the plants using a due to either of these previously utilized resistance genes (Foo- volume of 1 l per 1000 plants. A second similar application of inoculum was sprayed 30 min after the first spray. Following inoculation, the high- lad et al. 2014), although genetic mapping is needed to confirm pressure foggers were reinitiated. One day later, blackout curtains were the novelty of this resistance. To determine the viability of LB raised to allow for natural light (with no supplemental light). resistance conferred by PI 163245 in tomato breeding, this study Approximately seven days after inoculation, the plants were evaluated was undertaken to estimate the heritability of LB resistance and for foliar disease severity (% DS) on a scale of 0–100%, where 0% indi- examine response to selection for resistance. cated no tissue was affected by LB, and 100% indicated no remaining

healthy tissue. Each individual F2 plant was assigned a separate % DS score, while each six-plant replicate of control, parental and F1 progeny Materials and Methods was assigned a mean % DS score by visually estimating the average Plant material: The LB-resistant S. pimpinellifolium accession PI score of all six plants in each replicate. Following disease evaluation, 63 < 163245 was hybridized (as staminate parent) with Fla. 8059, a LB- of the most resistant (% DS 20) and 36 of the most susceptible F2 susceptible tomato breeding line. PI 163245 is an inbred accession that plants (% DS > 85) were selected, grown to maturity and self-pollinated produces small yellow fruit and has exhibited strong resistance to LB to produce F3 progeny families. under both field and greenhouse trials (Foolad et al. 2014). Original seed of PI 163245 was obtained from the USDA, ARS Plant Genetic F3 disease evaluation: The LB disease screenings of the F3 progeny Resources Unit (PGRU), Geneva, NY. Fla. 8059 is an inbred line that families were conducted four different times and were considered four produces firm large fruit with high lycopene content and good flavour, separate experiments (I, II, III and IV). Efforts were made to use the and has been used as a parent in the production of a commercial F1 same F3 families in all four experiments; however, due to insufficient hybrid cultivar (Scott et al. 2008). Original seed of Fla. 8059 was seed at the time of experiments, not all families were included in every received from J.W. Scott, University of Florida, Gulf Coast Research & experiment. Briefly, of the total of 99 selected F3 families, 31 were Education Center, Wimauma, FL. Hybrid F1 progeny were self- included in all four experiments, 26 in three experiments, 18 in two pollinated to produce F2 seed. A large F2 population (n = 560) was experiments and 24 in one experiment. Experiment I consisted of 45 screened for LB resistance under greenhouse conditions, and extremely resistant and 19 susceptible F3 families, which were evaluated resistant and susceptible individuals, composing the two tail ends of the approximately six days after disease inoculation. Experiment II was response distribution, were identified. The selected individuals were composed of 39 resistant and 20 susceptible F3 families, evaluated seven grown to maturity and self-pollinated to produce F3 and subsequently F4 days after inoculation. Experiment III contained 52 resistant and 24 Genetics of late blight resistance in tomato 393 susceptible families and experiment IV contained 49 resistant and 14 and environmental variances, respectively, and n is the number of indi- susceptible families. Experiments III and IV were both evaluated viduals in each of the F3 or F4 progeny families. The standard errors 2 approximately five days after inoculation. In each of the F3 experiments (SE) of the h estimates were calculated using the formulae 2 2 1/2 2 2 I, II and III, two replications totalling 10–12 F3 plants per family were h (F2:3)SE = [(1 r F2 : F3)/(n2)] and h (F3:4)SE = [(1r F3 : 1/2 2 grown and evaluated in two adjacent and similar sections of the F4)/(n2)] , where n is the number of families used to estimate h . The 2 greenhouse to accommodate the large number of individuals in each realized heritability (h R) for LB resistance was measured based on experiment. In experiment IV, both replicates of all F3 families were response to selection from F2 to F4 generations using the following grown in one section of the greenhouse, each replicate on opposite sides equation: of the greenhouse section. Parental and control genotypes (except NC þ 84173) were included in all experiments. NC 84173 and F1 progeny 2 ¼ RF2 : F3 RF3 : F4 hR were included only in experiments I and IV. All F3 families, along with SF2 : F3 þ SF3 : F4 the control genotypes, parental lines and F1 progeny, were inoculated with the pathogen and evaluated for disease response as described above where R is the selection response [change in mean % DS between respec- for the F2 population. After assigning a % DS score for each individual tive parental (e.g. F2 or F3) and progeny generations (e.g. F3 or F4)], and F3 plant, a mean % DS score was calculated for each F3 family by S is the selection differential (difference between the mean of selected- averaging individual values within the family. resistant individuals and overall mean of the parental population before

In experiment IV, the most resistant individuals within each of the selection) (Falconer 1989). For the F3 and F4 population screenings, only 49 F3 resistant families were selected and grown to maturity and self-pol- resistant-class F3 progeny families and resistant-class F4 progeny families 2 linated to obtain 49 resistant-class F4 progeny families. Similarly, the were considered for calculating h R. most susceptible individuals within each replicate of each of the 14 F3 The minimum number of loci involved in the observed LB resistance susceptible families (for a total of 28 plants) were selected; however, was estimated by two methods. First, the proportion of F2 individuals only 24 plants survived, which were grown to maturity and self- exhibiting % DS similar to the homozygous susceptible parent was com- n pollinated to produce 24 susceptible-class F4 progeny families. The pared to the expected Mendelian ratio of (¼) , where n is the minimum selected-resistant and susceptible F4 progeny families were used to esti- number of loci contributing to resistance in the F2 population. The sec- mate F3 :F4 heritability, examine the response to selection and measure ond method estimated the number of loci as follows: realized heritability, as described below. ðm m Þ2 n ¼ 1 2 F4 disease evaluation: The selected F4 progeny families (obtained from 8ðVF2 VF1Þ F3 experiment IV) were grown and screened for disease resistance at two different times and were considered two separate experiments (F 4 where m1 and m2 are the means of the % DS for the two parental lines, experiments I and II). Experiment I consisted of 47 resistant-class and 24 Fla. 8059 and PI 163245, and VF2 and VF1 are the variances in disease susceptible-class F families, and experiment II contained 30 resistant- 4 response in the F2 and F1 generations, respectively (Wright 1952, Fal- class and 22 susceptible-class F4 families. Two resistant-class families coner 1989). were not included in experiment I due to poor germination, and 21 (19 resistant class and 2 susceptible class) families were excluded from experiment II due to thrips infestation. For each F4 family, two Results replications totalling 10–12 individuals were grown on opposite benches Disease response of the parental and control lines, and F2 in a greenhouse section. Inoculation and disease evaluation in each F4 population experiment was conducted as previously described for the F3 experiments. Experiment I was evaluated seven days following The susceptible parent, Fla. 8059, displayed a high level of sus- inoculation, and experiment II was evaluated eight days after inoculation. ceptibility to LB in all experiments, with % DS averaging more Parental, F1 progeny and control genotypes were included in both than 90% and ranging from 50% to 100%. The resistant parent, experiments. The plants from parental, F1 progeny and control genotypes PI 163245, displayed a good level of resistance, with % DS aver- were evaluated individually and averaged for each replication. aging 16.5% and ranging from 5% to 40%. The observed resis- Individuals in each F4 progeny family were assigned a separate % DS, and the average % DS was calculated for each family. tance in PI 163245 was statistically similar to NC 63EB (Ph-2) and NC 870 (Ph-3), but not as strong as NC 03220 (Ph-2 + Ph- 3) (Table 1). The % DS in the susceptible control lines NC 84173 Data analysis: The F2,F3 and F4 disease evaluation data were used to estimate heritability of disease resistance and the number of disease and New Yorker (Ph-1) averaged 94.5% and 93.4%, respectively, resistance loci involved. Heritability (h2) of LB resistance was estimated while that in the resistant lines NC 63EB, NC 870 and NC 03220 by F2 :F3 and F3 :F4 parent–offspring (P–O) correlation analyses, using averaged 19.6%, 10.3% and 4.6%, respectively. The average % the following equations: DS of the F1 progeny (50.4%) was similar to the mid-parent value (54.4%) and statistically higher than that of the resistant parent 2ð Þ¼ ¼ CovF3;F2 h F2 : 3 rF2:F3 = (16.5%), suggesting the resistance conferred by PI 163245 is ðV V Þ1 2 F3 F2 likely additive or incompletely dominant in nature (Table 1). VA þ 1=2VD ¼ The % DS in the F2 population (n = 560) averaged ½ð þ = þ = Þð þ þ Þ 1=2 VA 1 4VD 1 nVE VA VD VE 51.9 30.2%, and ranged from 0% to 100% (Table 1). The Shapiro–Wilk test for normality indicated that the F2 population and was not normally distributed (P < 0.001) for disease response and slightly skewed towards susceptibility (skewness = 0.194; 2 CovF4;F3 h ðF : Þ¼r : ¼ 3 4 F3 F4 ð Þ1=2 Fig. 1). The representation of % DS from 0% to 100% was fairly VF4VF3 consistent across the F2 population, although the population = V þ = V ¼ 3 2 A 3 8 D resembled a nearly binomial distribution with resistant and sus- ½ð = þ = þ = Þð = þ = þ Þ 1=2 3 2VA 3 16VD 1 nVE 3 2VA 3 4VD VE ceptible tails being represented with somewhat higher frequencies than the rest of the population (Fig. 1). Of the 560 F2 individuals, where r is the correlation coefficient, Cov is the covariance between par- 319 individuals displayed disease responses greater or equal to ent and progeny generations, VA, VD and VE are the additive, dominance 50% and 241 individuals displayed % DS less than 50% (Fig. 1). 394 E. W. OHLSON AND M. R. FOOLAD

Table 1: Late blight disease severity (% defoliation SD) in tomato for resistant (PI 163245) and susceptible (Fla. 8059) parents, F1 progeny, control lines, and F2,F3 and F4 populations

Genotype Number of individuals or families % DS1,2 Range h2 (parent : offspring)

a P1 (Fla. 8059) 240 92.4 8.6 50–100 cd P2 (PI 163245) 221 16.5 9.4 5–40 b F1 100 50.4 12.7 25–80 NC 84173 120 94.5 4.5a 85–100 New Yorker (Ph-1) 240 93.4 5.9a 70–100 NC 63EB (Ph-2) 236 19.6 11.2c 5–60 NC 870 (Ph-3) 207 10.3 4.5de 5–20 NC 03220 (Ph-2 + Ph-3) 211 4.6 2.6e 0–10 F2 population 560 51.9 30.2 0–100 F2 :F3 Experiment I F2 selected individuals (resistant class) 45 7.9 4.8 0–20 F2 selected individuals (susceptible class) 19 91.2 2.0 90–95 F3 progeny families (resistant class) 45 30.0 24.9 2–100 F3 progeny families (susceptible class) 19 65.8 29.6 5–100 0.76 0.08 F2 :F3 Experiment II F2 selected individuals (resistant class) 39 7.5 4.5 0–20 F2 selected individuals (susceptible class) 20 90.4 2.5 85–95 F3 progeny families (resistant class) 39 30.5 28.5 2–100 F3 progeny families (susceptible class) 20 80.9 25.6 5–100 0.78 0.08 F2 :F3 Experiment III F2 selected individuals (resistant class) 52 7.7 4.7 0–20 F2 selected individuals (susceptible class) 24 90.0 2.9 85–97 F3 progeny families (resistant class) 52 28.9 28.8 1–100 F3 progeny families (susceptible class) 24 85.67 23.6 3–100 0.81 0.07 F2 :F3 Experiment IV F2 selected individuals (resistant class) 49 8.0 4.9 0–20 F2 selected individuals (susceptible class) 14 90.0 2.3 85–97 F3 progeny families (resistant class) 49 28.9 23.9 0–99 F3 progeny families (susceptible class) 14 85.7 26.6 5–100 0.81 0.07 F3 :F4 Experiment I F3 selected individuals (resistant class) 47 4.4 2.6 2–15 F3 selected individuals (susceptible class) 24 92.5 4.6 85–100 F4 progeny families (resistant class) 47 13.2 19.6 0–100 F4 progeny families (susceptible class) 24 74.7 28.9 5–100 0.91 0.05 F3 :F4 Experiment II F3 selected individuals (resistant class) 30 4.9 3.4 2–15 F3 selected individuals (susceptible class) 22 92.4 4.6 85–100 F4 progeny families (resistant class) 30 16.3 15.8 1–100 F4 progeny families (susceptible class) 22 83.1 21.4 5–100 0.97 0.04

1Mean comparisons of parental and control lines were determined using Tukey’s HSD test and are denoted as by superscript (a–e). 2In all disease evaluation experiments, the P. infestans isolate RS2009T1, belonging to the US clonal lineage US-23, race T-1 and mating type A1, was used.

80 phenotypically similar to the susceptible parent (% DS = 80– 70 60 100%). A comparison of these percentages to the expected Men- 50 delian ratio of (¼)n suggested the involvement of 1–2 resistance 40 30 loci with no dominance effects, consistent with the intermediate 20 disease response observed in the F progeny (Table 1). In the Frequency 1 10 second method, the number of resistance loci was estimated as 0 follows: 29 39 -74 -59 0-4 5-9 100 20-24 60-64 65-69 85-89 30-34 35- 70 75-79 80-84 10-14 15-19 40-44 45-49 50-54 55 90-94 25- 2

95- ¼ð Þ = ð Þ Disease severity (% DS) n m1 m2 8 VF2 VF1 ¼ð92:4 16:5Þ2=8ð914:6 299:3Þ¼1:17; Fig. 1: Frequency distribution of percent disease severity (% DS) for an = F2 population (n 560) derived from a cross between Solanum lycoper- also suggesting the involvement of 1–2 resistance loci. sicum (Fla. 8059) and Solanum pimpinellifolium (PI 163245). Foliar % DS was measured as a percentage of whole plant infection/defoliation and evaluated approximately 7 days after late blight infection under controlled greenhouse conditions Heritability estimations based on F2 :F3 generations The % DS of the selected-resistant parental F2 individuals (the resistant class) for the four F3 progeny experiments averaged Estimation of the minimum number of resistance loci between 7.5% and 8.0%, and % DS of the selected-susceptible Two methods were employed to estimate the minimum number of F2 parental individuals (the susceptible class) averaged between resistance loci involved. In the F2 population, approximately 29% 90.0% and 91.2% (Fig. 2). The F3 progeny families from the of the individuals displayed disease response similar to the selected-resistant and susceptible parental F2 individuals were resistant parent (% DS = 0–25%) and approximately 30% were screened for LB resistance in four separate experiments, and Genetics of late blight resistance in tomato 395

generally, similar results were obtained in all experiments. The Heritability estimations based on F3 :F4 generations average % DS of the F3 progeny families of the resistant class varied from 19.7% to 30.5%, and those from the susceptible The F4 progeny families from the selected F3 parental individu- class varied from 65.8% to 85.7% across the four experiments als were screened for LB resistance in two separate experiments (Table 1; Fig. 2). As the correlation coefficients were high (I and II), and generally, similar results were obtained in both experiments. The % DS of the resistant and susceptible F paren- between replications within all F3 experiments (r > 0.67, 3 P < 0.001), replications within each experiment were pooled tal individuals averaged 4.4% and 92.5%, respectively, for F4 2 experiment I, and 4.9% and 92.4% for F experiment II. The % before estimating h . In each experiment, the total F3 progeny 4 distribution (resistant + susceptible classes) was skewed towards DS of the resistant-class F4 progeny families averaged 13.2% in the resistant and susceptible ends of the disease spectrum experiment I, and 16.3% in experiment II (Table 1). The % DS (Fig. 2b, d, f, h); this was expected given the distribution of the of the susceptible-class F4 progeny families averaged 74.7% in experiment I, and 83.1% in experiment II. In the F generation, selected individuals in the F2 population (Fig. 2a, c, e, g). The 4 individual family performance ranged from 3.6% to 99% and % DS of the F3 progeny families generally resembled the % DS from 5.8% to 97% in experiments I and II, respectively. In each of the corresponding parental F2 individuals. However, approxi- F experiment, the correlation between replications was high mately 13% of the resistant-class F3 families exhibited higher- 4 than-expected % DS (>60%) and approximately 19% of the sus- (r > 0.91, P < 0.001), and thus, replications within each experiment were pooled before estimating h2. The F progeny families ceptible-class F3 families displayed less-than-expected % DS 4 (<60%) in one or more of the experiments; this was likely a were skewed in disease response towards both susceptible and result of either disease escape or an area of higher disease pres- resistant ends of the disease spectrum (Fig. 3b, d), as was 2 expected based on the distribution of selected-resistant and sus- sure in the greenhouse. The h estimates based on the F2 :F3 correlation analyses ranged between 0.76 and 0.81 with an aver- ceptible F3 parental individuals (Fig. 3a, c). Almost all resistant- 2 = age of 0.79, suggesting moderately high h of LB resistance class F4 progeny families (n 49) resembled the F3 parental derived from PI 163245 (Table 1). phenotype; however, one family in experiment I exhibited higher

Resistant (a)30 F2 Suscepble (b) 200 F3 25 150 20 15 100 10 Frequency Frequency 50 5 0 0 54 74 -64 5-9 0-4 5-9 0-4 60 65-69 70-74 75-79 15-19 20-24 10-14 15-19 20-24 25-29 40-44 45-49 85-89 30-34 35-39 40-44 45-49 50-54 10-14 55-59 60-64 70- 75-79 80-84 85-89 90-94 25-29 30-34 35-39 50- 55-59 90-94 80-84 65-69 95-100 Disease severity (% DS) 95-100 Disease severity (% DS)

(c)F (d) F 30 2 200 3 25 150 20 15 100 10 Frequency Frequency 50 5 0 0 -9 64 49 19 -69 -29 0-4 0-4 5 5-9 100 - 90-94 85-89 90-94 75-79 85-89 70-74 75-79 65 80-84 70-74 80-84 55-59 60- 65-69 60-64 35-39 40-44 50-54 55-59 50-54 45- 40-44 45-49 25-29 20-24 30-34 35-39 15- 20-24 15-19 25 30-34 10-14 10-14 Disease severity (% DS) 95 Disease severity (% DS) 95-100 (e) (f) 30 200 F2 F3 25 150 20 15 100 10 Frequency

Frequency 50 5 0 0 4 -9 -9 89 34 -94 -74 -89 - -69 -64 -39 -44 - -24 5 -29 -14 0-4 0- 5 90 90-94 80-84 85 85 70 75-79 75-79 60 65-69 80-84 70-74 55-59 60-64 65 45-49 40 45-49 50-54 50-54 55-59 35 35-39 40-44 30 30-34 20 20-24 25-29 25 10 15-19 10-14 15-19 95-100 Fig. 2: Frequency distribution of Disease severity (% DS) Disease severity (% DS) 95-100 percent disease severity (% DS) for selected resistant and susceptible (g) (h) 30 F2 200 F3 tomato F2 individuals and their 25 respective F3 progeny families for 150 experiment I (a, b), experiment II (c, 20 d), experiment III (e, f) and 15 100 experiment IV (g, h). Foliar % DS 10

Frequency Frequency 50 was measured as a percentage of 5 whole plant infection/defoliation and 0 0 4 - -9 84 74 49 39 – 34 -89 -69 -64 -59 -49 - - -24 -29 -14 0 0-4 5 evaluated 5 7 days after late blight 5-9 90-94 85 90-94 80- 85-89 75-79 75-79 70- 60 80-84 65-69 70-74 50-54 55-59 65 55 60-64 45 50-54 40-44 40-44 30 35-39 45- 30-34 35 25 10 15-19 20-24 25-29 15-19 20 infection under controlled 10-14 95-100 95-100 greenhouse conditions Disease severity (% DS) Disease severity (% DS) 396 E. W. OHLSON AND M. R. FOOLAD

F F (a)3 Resistant (b) 4 30 Suscepble 200 25 20 150 15 100 10 Frequency Frequency 50 5 0 0 -4 24 19 84 89 0 -14 5-9 5-9 -54 0-4 - 100 - 35-39 20- 65-69 15- 40-44 50 60-64 80-84 85-89 25-29 70-74 75-79 10-14 15-19 10 45-49 55-59 90-94 30-34 20-24 25-29 35-39 40-44 90-94 65-69 75-79 30-34 45-49 50-54 60-64 85- 70-74 80 55-59 95-100 Disease severity (% DS) 95 Disease severity (% DS) Fig. 3: Frequency distribution of (c)F (d) F percent disease severity (% DS) for 30 3 4 200 selected resistant and susceptible 25 tomato F3 individuals and their 20 150 respective F4 progeny families for 15 100 experiment I (a, b) and experiment 10 II (c, d). Foliar % DS was measured Frequency Frequency 50 5 as a percentage of whole plant 0 0 infection/defoliation and evaluated 94 84 54 54 – -69 -79 -64 -64 - -29 -19 5-9 0-4 5-9 0-4 5 8 days after late blight infection 100 - 85-89 90- 80- 85-89 90-94 70-74 75-79 50- 65-69 80-84 65 60 60 70-74 75 40-44 55-59 55-59 35-39 45-49 35-39 50 30-34 40-44 45-49 25-29 15 20-24 30-34 15-19 20-24 10-14 25 10-14 95 95-100 under controlled greenhouse Disease severity (% DS) Disease severity (% DS) conditions

%DS(>60%) than expected. Of the 24 susceptible-class F4 over 40% of the population exhibited intermediate resistance families, five families in experiment I and one family in experi- (25 ≤ %DS≤ 80) and there was no significant skewness ment II exhibited lower % DS (<60%) than expected. Estimates towards resistance (Fig. 1). Both methods of estimating the mini- 2 of h using F3 :F4 correlation analyses were high in both exper- mum number of resistance loci suggested involvement of 1–2 iment I (h2 = 0.91) and experiment II (h2 = 0.97), indicating a loci, which is also consistent with the distribution of disease highly heritable nature of the LB resistance derived from PI response in the F2 population (Fig. 1). However, the actual num- 163245. ber of resistance loci should be confirmed by genetic mapping studies. The h2 of LB resistance conferred by PI 163245 was esti- Calculation of realized heritability based on F2 :F4 mated/measured by two methods, using data from several repli- generations cated experiments. Parent : offspring (P : O) correlation analyses 2 2 The realized heritability (h R) was calculated using data obtained were conducted to estimate h based on both F2 :F3 and F3 :F4 from the F3 experiment IV, and the corresponding F4 experi- generations. Heritability estimates obtained based on P : O ments. Because the F4 progeny were evaluated twice (F4 experi- regression (or correlation) analysis are generally close to the nar- 2 2 ments I and II), two measures of h R were obtained, 0.66 row-sense h and more reliable than estimates obtained based on (experiment I) and 0.61 (experiment II). variance components analysis and analysis of variance (Dudley and Moll 1969, Foolad and Jones 1992, Foolad et al. 2002, Merk and Foolad 2012). Regression or correlation analyses are Discussion generally free of assumptions normally made when conducting Accession PI 163245 exhibited strong LB resistance across all analysis of variance and variance components analysis, including experiments and was statistically similar to NC 63EB (Ph-2) and normality of distribution and similarity of environmental NC 870 (Ph-3) but not as resistant as NC 03220 (Ph-2 + Ph-3) variances across populations/generations (Vogel et al. 1980, (Table 1). PI 163245 also displayed high levels of LB resistance Casler 1982, Foolad and Jones 1992). Further, in the present under field conditions, where it exhibited higher LB resistance study, correlation rather than regression analysis was employed (% DS = 0.5 0.0) than L3708 (% DS = 10.7 10.0), the because it provides a more accurate estimate of h2 when there S. pimpinellifolium original source of Ph-3 (Foolad et al. 2014) are scalar differences between parental and progeny populations and tomato breeding lines containing Ph-2 or Ph-3 alone (RG due to unintended variation in growing and evaluation conditions Gardner, personal communication). It is unknown whether the (Frey and Horner 1957, Dudley and Moll 1969, Foolad and resistance in PI 163245 is conferred by Ph-2, Ph-3 or other Jones 1992). Additionally, experiments were conducted in multi- genes, although in a previous study it was determined that PI ple generations (i.e. F2 :F3 and F3 :F4) to obtain an assessment 163245 did not have the molecular markers known to be associ- of the significance of potential dominance genetic variance ated with either Ph-2 or Ph-3 (Foolad et al. 2014). Currently, affecting trait expression. For example, in the presence of domi- molecular mapping studies are underway to identify and map nance, estimates based on F3 :F4 generations are more reliable genes conferring LB resistance in PI 163245 (see below). How- than those based on F2 :F3 generations as they are less biased ever, in previous studies PI 163245 had displayed resistance to by any potential dominance effect. Moreover, multiple experi- several P. infestans isolates of different US clonal lineages in ments were run for each progeny generation to reduce undesired field and greenhouse experiments (Foolad et al. 2014), suggest- environmental effects and sampling error of the selected F3 or F4 ing the potential utility of this resistance resource in tomato progeny families. breeding. Although all disease-screening experiments were conducted in The F1 progeny exhibited an intermediate resistance (average an environmentally controlled greenhouse, which provided fairly 50.4% DS) between the two parents, suggesting an additive or similar conditions throughout all experiments, the conditions did incompletely dominant nature of this resistance. This conclusion vary slightly based on uncontrollable outdoor factors, including is also supported by the distribution of the F2 generation, where light, temperature and humidity. The presence of such variation, Genetics of late blight resistance in tomato 397 combined with potential genetic variation caused by sampling of resistance genes were not responsible for the observed resistance the progeny populations (e.g. sampling of F3 or F4 progeny fam- in PI 163245 (Foolad et al. 2014). However, genetic mapping ilies), necessitated repeating experiments multiple times and studies are currently underway to identify the genomic regions obtaining multiple estimates of h2. However, the estimates of h2 responsible for LB resistance in PI 163245 (Ohlson et al., were generally similar across experiments, suggesting reliability unpublished data). Preliminary analysis suggests the presence of of the estimates. The estimates were moderately high, ranging two major LB resistance QTLs in PI 163245, one on each of from 0.76 to 0.81 in F2 :F3 generations and 0.91 to 0.97 in chromosomes 2 and 10. While the latter QTL may be an allele F3 :F4 generations (Table 1), suggesting the highly heritable of Ph-2, or a separate resistance gene located in the vicinity of nature of this resistance. The comparatively higher h2 estimates Ph-2, the resistance QTL on chromosome 2 may be novel. Fur- in F3 :F4 than in F2 :F3 generations could be due to fixation of ther, no resistance QTL was identified on chromosome 9, where resistance and susceptibility genes in the F3 generation due to Ph-3 is located. Therefore, PI 163245 may harbour at least one directional selections made in the F2 generation, resulting in new LB resistance gene/QTL, which may be useful for tomato higher covariance between F3 and F4 generations. The moderate- breeding. Breeding efforts are currently underway to incorporate to-high estimates of h2 suggest that the LB resistance in acces- resistance from PI 163245 into desirable tomato breeding lines. sion PI 163245 is highly transmittable and potentially could be Further testing of PI 163245 against other P. infestans isolates transferred to the cultivated tomato genetic background via phe- and clonal lineages would be useful for determining the durabil- notypic selection and breeding. ity and utility of this resistance. To confirm the heritable nature of this resistance, two cycles of selection for high resistance in F2 and F3 generations were Acknowledgements made and the response to selection (R) in both F3 and F4 progeny generations was determined. Using the breeder’s equa- The authors graciously thank Dr. Matthew Sullenberger and PhD student tion (a.k.a response-to-selection equation), R=h2S, the realized Mengyuan Jia for reviewing the manuscript before submission and mak- 2 ing useful comments and suggestions. The authors also thank all Penn heritability (h R) of resistance was calculated by evaluating the response to selection (actual improvement in disease resistance) State staff, graduate and undergraduate students who helped with these from F to F generations. The F progeny population was eval- experiments, Dr. John (Jay) Scott (University of Florida) for providing 2 4 4 seed of Fla. 8059, Dr. Randolph Gardner (NC State University) for pro- uated in two separate experiments, and thus, two measures of 2 viding seeds of the control genotypes and the USDA PGRU for provid- h R were obtained. The two measures were similar, averaging ing the original seed for PI 163245. This research was supported in part 0.63. This is slightly lower than estimates obtained based on by the Pennsylvania Vegetable Marketing and Research Program, Mid- 2 P : O correlation analysis, which is expected because P : O h America Food Processors Association, the College of agricultural estimates are potentially inflated by the effects of dominance Sciences and Department of Plant Science at the Pennsylvania State 2 (see equations in the Materials and Methods section) while h R University. is a better measure of narrow-sense h2 (i.e. h2 based on additive genetic variance). Furthermore, measures of h2 based on response-to-selection equation are sensitive to changes in envi- References ronmental variances across parent and progeny generations, Abreu, F. B., D. J. H. da Silva, C. D. Cruz, and E. S. G. Mizubuti, resulting in scalar differences (Falconer 1989), which may 2008: Inheritance of resistance to phytophthora infestans (peronospo- account for the lower measures of h2 obtained compared to esti- rales, pythiaceae) in a new source of resistance in tomato (solanum sp. R (formerly lycopersicon sp.), solanales, solanaceae). Genet. Mol. Biol. mates of h2 obtained from P : O correlation analyses. It is 3, 493—497. expected that in later filial generations (e.g. F6 or later), the two AVDRC, 1994: 1993 Progress Report, 201-203. Asian Vegetable estimates would approach each other. Research and Development Center, Shanhua, Tainan, Taiwan. 2 The various estimates/measures of h were similar and moder- Black, L. L., T. C. Wang, P. M. Hanson, and J. T. Chen, 1996: Late ately high for all experiments, suggesting relatively simple blight resistance in four wild tomato accessions: effectiveness in diverse genetic control of the LB resistance in accession PI 163245 and locations and inheritance of resistance. Phytopathology 86, S24. consistent with the estimates of 1–2 genetic loci affecting the Bonde, R., and E. F. Murphy, 1952: Resistance of certain tomato vari- resistance (discussed above). This is also congruent with obser- eties and crosses to late blight. Maine Agric. Expt. Sta. Bull. 497,5— vations of simple genetic control of LB resistance in other acces- 15. sions of S. pimpinellifolium (Bonde and Murphy 1952, Gallegly Brouwer, D. J., and D. A. St. Clair, 2004: Fine mapping of three quantitative trait loci for late blight resistance in tomato using near isogenic and Marvel 1955, Peirce 1971, Chunwongse et al. 2002, Merk lines (NILs) and sub-NILs. Theor. Appl. Genet. 108, 628—638. and Foolad 2012, Merk et al. 2012). The strong level of LB 2 Brouwer, D. J., E. S. Jones, and D. A. St. Clair, 2004: QTL analysis of resistance in PI 163245 and high h observed through multiple quantitative resistance to Phytophthora infestans (late blight) in tomato generations and experiments suggest this resistance can be useful and comparisons with potato. Genome 47, 475—492. for breeding in tomato. Identifying and implementing new LB Casler, M. D., 1982: Genotype 9 environment interaction bias to parent- resistance genes in tomato breeding is imperative as Ph-2 and offspring regression heritability estimates. Crop Sci. 22, 540—542. Ph-3 are the only widely utilized LB resistance genes in com- Chen, A.-L., C.-Y. Liu, C.-H. Chen, J.-F. Wang, Y.-C. Liao, C.-H. mercial tomato breeding. Additionally, even Ph-3, the strongest Chang, M.-H. Tsai, K.-K. Hwu, and K.-Y. Chen, 2014: Reassessment utilized LB resistance gene in tomato, has been overcome by of QTLs for late blight resistance in the tomato accession L3708 using some P. infestans isolates (Chunwongse et al. 2002, R. G. Gard- a restriction site associated DNA (RAD) linkage map and highly aggressive isolates of Phytophthora infestans. PLoS ONE 9, e96417. ner, personal communication), emphasizing the significance of Chunwongse, J., C. Chunwongse, L. Black, and P. Hanson, 1998: Map- identifying new sources of LB resistance. PI 163245 has dis- ping of Ph-3 gene for late blight from L. pimpinellifolium L3708. Rpt. played high levels of resistance against all tested isolates of Tomato Genet. Coop. 48,13—14. P. infestans, including eight from clonal lineages US-13, US-14 Chunwongse, J., C. Chunwongse, L. Black, and P. Hanson, 2002: and US-23 (Foolad et al. 2014). Previous examination of molec- Molecular mapping of Ph-3 gene for late blight resistance in tomato. ular markers associated with Ph-2 and Ph-3 suggested that these J. Hort. Sci. Biotechnol. 77, 281—286. 398 E. W. OHLSON AND M. R. FOOLAD

Dudley, J. W., and R. H. Moll, 1969: Interpretation and use of estimates McGrath, M. T., W. E. Fry, and C. D. Smart, 2013: Recent occurrences of heritability and genetic variances in plant breeding. Crop Sci. 9, of late blight on tomatoes in the Northeastern USA. Acta Hort. 1069, 257—261. 321—326. Falconer, D. S., 1989: Introduction to Quantitative Genetics, 3rd edn. Merk, H. L., and M. R. Foolad, 2012: Parent-offspring correlation esti- Longman Scientific and Technical, London. mate of heritability for late blight resistance conferred by an accession Foolad, M. R., and R. A. Jones, 1992: Parent-offspring regression esti- of the tomato wild species Solanum pimpinellifolium. Plant Breed. mates of heritability for salt tolerance during germination in tomato. 131, 203—210. Crop Sci. 32, 439—442. Merk, H. L., H. Ashrafi, and M. R. Foolad, 2012: Selective genotyping Foolad, M. R., P. Subbiah, and G. Ghangas, 2002: Parent-offspring cor- to identify late blight resistance genes in an accession of the tomato relation estimate of heritability for early blight resistance in tomato, wild species Solanum pimpinellifolium. Euphytica 187,63—75. Lycopersicon esculentum Mill. Euphytica 126, 291—297. Miller, J. C., and S. D. Tanksley, 1990: RFLP analysis of phylogenetic Foolad, M. R., H. L. Merk, and H. Ashrafi, 2008: Genetics, genomics relationships and genetic variation in the genus Lycopersicon. Theor. and breeding of late blight and early blight resistance in tomato. Crit. Appl. Genet. 80, 437—448. Rev. Plant Sci. 27,75—107. Moreau, P., P. Thoquet, J. Olivier, H. Laterrot, and N. Grimsley, 1998: Foolad, M. R., M. T. Sullenberger, E. W. Ohlson, and B. C. Gugino, Genetic mapping of Ph-2, a single locus controlling partial resistance 2014: Response of accessions within tomato wild species, Solanum to Phytophthora infestans in tomato. Mol. Plant Microbe Interact. 11, pimpinellifolium to late blight. Plant Breed. 133, 401—411. 259—269. Frey, K. J., and T. Horner, 1957: Heritability in standard units. Agron. J. Nowicki, M., M. R. Foolad, M. Nowakowska, and E. U. Kozik, 2012: 49,59—62. Potato and tomato late blight caused by Phytophthora infestans:an Fry, W. E., and S. B. Goodwin, 1997: Re-emergence of potato and overview of pathology and resistance breeding. Plant Dis. 96,4—17. tomato late blight in the United States. Plant Dis. 81, 1349—1357. Nowicki, M., E. U. Kozik, and M. R. Foolad, 2013: The tomato late Gallegly, M. E., and M. E. Marvel, 1955: Inheritance of resistance to blight. In: R. K. Varshney, and R. Tuberosa (eds), Translational Geno- tomato race-0 of Phytophthora-infestans. Phytopathology 45, 103— mics for Crop Breeding: Vol. I, Biotic Stress, 1st edn, 241—265. John 109. Wiley & Sons, Inc., USA. Gardner, R. G., and D. R. Panthee, 2010a: ‘Plum Regal’ Fresh-market Panthee, D. R., and R. G. Gardner, 2010: ‘Mountain Merit’: a late blight- plum tomato hybrid and its parents, NC 25P and NC 30P. HortScience resistant large-fruited tomato hybrid. HortScience 45, 1547—1548. 45, 824—825. Peirce, L. C., 1971: Linkage tests with Ph conditioning resistance to race Gardner, R. G., and D. R. Panthee, 2010b: NC 1 CELBR and NC 2 0, Phytophthora infestans. Rpt. Tomato Genet. Coop. 21, 30. CELBR: early blight and late blight resistant fresh market tomato Peralta, I., and D. Spooner, 2000: Classification of wild tomatoes: a breeding lines. HortScience 45, 975—976. review. Kurtziana 28,45—54. Gisi, U., and Y. Cohen, 1996: Resistance to phenylamide fungicides: a Roden, L. C., and R. A. Ingle, 2009: Lights, Rhythms, Infection: the role case study with Phytophthora infestans involving mating type and race of light and the circadian clock in determining the outcome of plant– structure. Annu. Rev. Phytopathol. 34, 549—572. pathogen interactions. Plant Cell 21, 2546—2552. Goodwin, S. B., L. S. Sujkowski, and W. E. Fry, 1995: Rapid evolution Scott, J. W., E. A. Baldwin, H. J. Klee, J. K. Brecht, S. M. Olson, J. A. of pathogenicity within clonal lineage of the potato late blight disease Bartz, and C. A. Sims, 2008: Fla. 8153 Hybrid Tomato; Fla. 8059 and fungus. Phytopathology 85, 669—676. Fla. 7907 Breeding Lines. HortScience 43, 2228—2230. Goodwin, S. B., L. S. Sujkowski, and W. E. Fry, 1996: Widespread dis- Smart, C. D., and W. E. Fry, 2001: Invasions by the late blight patho- tribution and probable origin of resistance to metalaxyl in clonal geno- gen: renewed sex and enhanced fitness. Biol. Invasions 3, 235—243. types of Phytophthora infestans in the United States and western Smart, C. D., S. D. Tanksley, H. Mayton, and W. E. Fry, 2007: Resis- Canada. Phytopathology 86, 793—800. tance to Phytophthora infestans in Lycopersicon pennellii. Plant Dis. Griebel, T., and J. Zeier, 2008: Light regulation and daytime dependency 91, 1045—1049. of inducible plant defenses in arabidopsis: phytochrome signaling con- Vogel, K. P., F. A. Haskins, and H. J. Gorz, 1980: Parent-progeny trols systemic acquired resistance rather than local defense. Plant Phys- regression in indiangrass: inflation of heritability estimates by environ- iol. 147, 790—801. mental covariances. Crop Sci. 20, 580—582. Gugino, B. K., and M. R. Foolad, 2013: Update on late blight and use Wright, S., 1952: The genetics of quantitative variability. In: E. C. R. of host resistance for disease management Mid-Atlantic Fruit and Reever, and C. H. Waddington (eds), Quantitative Genetics, 4—41. Vegetable Convention, 95-97, Hershey, PA. H.M. Stat. Office, London. Gugino, B. K., I. A. Huerta, M. R. Foolad, and W. Fry, 2014: The inte- Zhang, C., L. Liu, Z. Zheng, Y. Sun, L. Zhou, Y. Yang, F. Cheng, Z. grated management of late blight on tomato Mid-Atlantic Fruit and Zhang, X. Wang, S. Huang, B. Xie, Y. Du, Y. Bai, and J. Li, 2013: Fine Vegetable Convention, 118-121, Hershey, PA. mapping of the Ph-3 gene conferring resistance to late blight (Phytoph- Lukyanenko, A. N., 1991: Disease resistance in tomato. In: G. Kalloo thora infestans) in tomato. Theor. Appl. Genet. 126, 2643—2653. (ed), Genetic Improvement of Tomato, 99—119. Springer-Verlag, Ber- Zhang, C., L. Liu, X. Wang, J. Vossen, G. Li, T. Li, Z. Zheng, J. Gao, lin Heidelberg. Y. Guo, R. G. F. Visser, J. Li, Y. Bai, and Y. Du, 2014: The Ph-3 Mayton, H., G. Forbes, E. Mizubuti, and W. E. Fry, 2001: The roles of gene from Solanum pimpinellifolium encodes CC-NBS-LRR protein three fungicides in the epidemiology of potato late blight. Plant Dis. conferring resistance to Phytophthora infestans. Theor. Appl. Genet. 85, 1006—1012. 127, 1353—1364.