Comparison of chili pepper breeding populations for agronomic traits and polygenic resistance to Phytophthora blight

Belonging to the Oomycete class, Phytophthora capsici has wide range of host profile and is responsible for many devastating diseases in many countries. In addition to time consuming problem for transferring resistance to susceptible varieties, backcrossing method causes losing of genes providing resistance to susceptible varieties. In this study transferring P. capsici resistance genes to susceptible chili pepper lines was aimed during the extensive breeding period and resistant lines were confirmed by marker assistance. Two different breeding populations from CM334 and PM217 were compared by stem inoculation test to determine receptivity, inducibility, and stability resistance component. CM334 was found more effective for transferring all resistance components while PM217 was found suitable for keeping agronomic traits along with two important resistance component inducibility and stability. These two resistant components were found highly correlated to length of stem necrosis. C-29 and C-18 have been improved from CM334 as resistant as CM334; P-73 and P-77 have been improved satisfactorily resistant and yielded lines from PM217. Marker assisted selection proved that resistance of lines differentiated phenotypically despite the genotypes have the same genes.

In addition to time consuming problem for transferring resistance to susceptible varieties, backcrossing method causes loss of genes providing resistance to susceptible varieties (Palloix et al., 1990). Specific race host interaction has been referred by Sy et al. (2005) to manage Phytophthora resistance breeding program in consideration of gene to gene theory (Monroy- Barbosa & Bosland, 2011).
O v e r t h e p a s t y e a r s , m a n y quantitative trait loci (QTL) have been detected and several molecular markers have been reported related to resistance to P. capsici in pepper (Thabuis et al., 2004a;Quirin et al., 2005;Mallard et al., 2013). Thabuis et al. (2004b) have determined the QTL (Phyto 6.1.) on 6 th chromosome in the populations improved by recurrent selection using the markers ASC 012 and ASC 014. Bonnet et al. (2007) have identified 8 QTL on 1 st , 4 th , 5 th , 6 th , and 11 th chromosomes and observed that 4 chromosomes affect many of the resistance components.
ASC037 on P5 and ASC035p on P10 markers used in this study are significantly correlated to receptivity and stability components while ASC031 has a weak correlation to root rot index and stability on P2. CAPS and SCAR markers developed by Thabuis et al. (2004a) were used in this study to confirm the resistance of improved chili pepper lines because of digenic interaction on both markers and association with root rot index resistance. Transferring P. capsici resistance to susceptible chili pepper lines was aimed during extensive breeding period. Forty-five improved lines were evaluated for their resistance to P. capsici and agronomical traits. During breeding program, three resistance genes could be transferred to Sena chili pepper cultivar widely grown for dried pepper production.

I s o l a t i o n , p ro d u c t i o n a n d conservation of P. capsici isolates
Open field chili pepper cultivation areas were surveyed in Kahramanmaraş and pepper stem samples which had Phytophthora blight symptoms were collected. Eleven of the fifty-five fields infected by the pathogen widespread were surveyed. Five isolates were cultured and tested for aggressiveness on CM334 and Sena chili pepper variety which was commercial and susceptible to the pathogen (data not shown). One aggressive isolate from Doganlıkarahasan was used as inoculum.

Stem inoculation tests
The mycelium discs were placed on the cutting of the stem as described by Pochard & Daubèze (1980). An aluminium sheet was wrapped on the top of the stem in which was plugged a mycelium disc to prevent drying of the inoculum. The progression of fungal necrosis from top of the stem to base was measured with digital caliper 3, 10, 14 and 21 days after stem inoculation. Receptivity was measured in early infection process 3 rd day post inoculation (DPI) inducibility was measured between 3 rd and the 10 th DPI and stability was the average speed of the stem necrosis measured as mm day -1 between the 14 th and the 21 st DPI (Lefebvre & Palloix, 1996).

Breeding population
The lines were improved from two breeding populations. One of the breeding populations was derived from crosses between the resistant donor Criollo de Morelos 334 (CM334) and KM211 which had been selected as resistant to P. capsici from Kahramanmaraş chili pepper population. The other breeding population was obtained by crossing PM217=PI201234 and KM211. PM217 (abbreviated as PM male) with KM211 (abbreviated as K female) and CM334 (abbreviated as CM male) with KM211 (female) were crossed independently and 106 individuals were self-pollinated for three times during three years. Two genotypes were selected from CMK and PMK populations. These two new resistant lines were independently crossed by Sena which was the susceptible chili pepper variety. Two independent F 1 progenies were not backcrossed to susceptible recipient but self-pollinated and submitted to screening tests. The most resistant individuals were backcrossed to susceptible recipient Sena variety and self-pollinated again. CMKSeF 3 and PMKSeF 3 populations were used as new resistant sources through 86 individuals according to stem inoculation test results. After two backcrossings to Sena and self-pollination alternately, 45 lines were selected and subjected to stem inoculation tests. Improved lines originated from PM217 were indicated with P letter and those of CM334 indicated with C letter at the beginning of the line numbers. (Figures 1, 2).

Marker-assisted selection of improved lines
Total DNAs were purified from pieces of leaves (0.1 g) of improved 45 chili pepper lines backcrossed to Sena by both resistance sources originated from PM217 and CM334 with the Tri-Reagent kit (Molecular Research Centre Inc.) described as protocols. CAPS markers (ASC037 and ASC031) and SCAR marker (ASC035p) were used to assign resistance to Phytophthora blight (Thabuis et al., 2004a;Lefebvre et al., 1995). EcoRI and HaeIII endonucleases were used to determine codominance for ASC037 and ASC031 respectively. PCR and digested products were analyzed by capillary electrophoresis. The base sizes were determined by using Qiaxcel Advanced System with AM320 method and DNA scanning cartridge at the electrophoretic analysis of PCR products.

Evaluation of agronomic traits
Forty-five BC 2 S 4 lines improved from both resistance sources were evaluated for yield and fruit characteristics in 2016 under field conditions from May to September in Kahramanmaraş placed South East Anatolian Region of Turkey. The field was drip irrigated and fertilized (160 kg ha -1 N, 20 kg ha -1 P 2 O 5 and 160 kg ha -1 K 2 O) during experiment. Lines were evaluated in augmented design of 3 blocks with 20 plants. Seven control varieties Sena, Maraş-1, S. Demre, Carliston, H46, BT 46 and PR 90 were repeated in each block to calculate variance. The average fruit weight (g), fruit length (mm), fruit width (mm) and the fruit flesh thickness (mm) were measured. All plots were harvested two times when the fruits were matured, and yield values were transformed to yield per hectare.

Statistical analysis
Speed of stem necrosis and necrosis length of the CMKSeBC 2 S 4 and PMKSeBC 2 S 4 populations were performed by Analysis of Variance ( A N O VA ) a f t e r c h e c k i n g t h e variance homogeneity. Means were compared by LS Means Differences Tukey HSD multiple comparison tests. Data obtained from agronomic traits were also performed by ANOVA and means were compared by LS Means Differences Student's t test. Two breeding populations were compared by pairing means with t-test. JMP statistical software version 5.0.1 was used to calculate and compare means.

Resistance of the breeding populations
Three resistance components differed by improved lines derived from different resistant sources CM334 and PM217. The means of BC 2 S 4 population improved from CM334 and PM217 were significantly dissimilar for receptivity, inducibility and stability and it was verified by t-test. CM334 was a more effective genotype to transfer all three components of P. capsici resistance to its generations. Population derived from CM334 especially resisted to the pathogen at the inducibility and stability stage (Table 1).
Plants were arranged in four phenotypic groups as high level resistant, resistant, moderately resistant and low level resistant according to their resistance component receptivity, inducibility, and stability. The first phenotypic group was classified as high level resistant including C-18, C-29 and CM334 which had low speed of stem necrosis related to three resistance components corresponding to receptivity, inducibility and stability stage ( Figure 2). These three phenotypes which had the shortest length of stem necrosis after 21 days of inoculation and lines were separated from other phenotypes according to principal component analysis. Transferring the P. capsici resistance genes from PM217 depends on host pepper genotypes and aggressiveness of the isolates (Bartual et al., 1991). Pochard et al. (1983) declared that PM217 has low level resistance corresponding to inducibility component. PM217 has high level resistance with the inducibility component but low level resistance at receptivity stage in early infection process to Doganlikarahasan isolate. High level susceptibility of Sena variety to the pathogen has decreased the resistance of PM217 progenies despite using complex breeding program for introgression of resistance genes to the progenies. On the other hand, CM334 displayed the highest level of resistance for the four resistance components determining nine additive  Comparison of chili pepper breeding populations for agronomic traits and polygenic resistance to Phytophthora blight regions and is more effective to improve high level resistant progenies. CM334 has three resistance components in addition to root rot index; bell pepper genotype Vania and pungent Indian genotype Perennial have different level of resistance at receptivity, inducibility and stability stage (Thabuis et al., 2003).
The chromatogram of C-18 and C-29 showed that three regions related to the resistance conferring different levels of inducibility, receptivity and stability could be transferred by the breeding program. Thirteen improved lines produced expected size amplicons  by all three markers and ten lines were codominant related to ASC031 while only one was corresponding to ASC037 ( Figure 3). Fourteen lines improved from different sources possessed one or two resistance components. C-5, P-73 and P-77 were classified as resistant in response to the pathogen at inducibility, receptivity and stability stage. C-27, C-28, C-34, C-37, PM217 and P-2 did not resist to the pathogen at the first three days of the inoculation and they were accepted as moderately resistant with high level resistance corresponding to inducibility and stability components. Other genotypes, except for H46, Sena, P74, P74-1 and P6, straggled between resistant and susceptible genotypes. Control line H46 and variety Sena were quietly separated along with P74, P74-1 and P6 and considered as susceptible genotypes (Figure 4). After extensive breeding program including crossing, self-pollination and testing process new high-level resistant lines C-29 and C-18 have been improved from CM334 parent as resistant. Also, P-73 and P-77 have been improved from PM217 and KM211 (selected resistant line from local population) as resistant and more yielding lines than Sena commercial chili pepper variety cultivated widespread in the region. Improving resistant lines or varieties to Phytophthora capsici is more complex and requires new approaches such as local inspirations (Oelke et al., 2003). Carvalho et al. (2017) (Mo et al., 2014) and Japan (Sugita et al., 2006). Receptivity, inducibility and stability components significantly and positively correlated with each other and length of stem necrosis at the 21 st day post inoculation. Response of the genotypes at the inducibility, deceleration of the necrosis length stage highly correlated with their resistance. Stability component also affected the length of stem necrosis of the improved lines. In consideration of the pairwise correlation of the speed and length of the stem necrosis, resistance of the genotypes at the inducibility and stability components was more effective in resisting to the P. capsici (Box 1).
Inducibility and stability components are more effective than receptivity at early infection stage to resist the P. capsici. Thabuis et al. (2004a) have indicated that resistance at stability is highly correlated with resistance whereas receptivity is less correlated. Lefebvre & Palloix (1996) have observed a weak correlation between inducibility and other resistance component. In this study, inducibility has been found highly correlated with stability and length of stem necrosis. Mallard et al. (2013) proposed a new QTL on chromosome 5 related to broad-spectrum resistance in which are significantly interacted to inducibility and stability.
Amplicons of C-18, C-29 and CM334 were close to the expected size with all three markers. Base sizes of C-5, P-73, P-77, C-27, C-28, C-34, C-37, PM217 and P-2 lines were distinct from 11 to 58 base pairs (bp) for ASC 035, 69 to 165bp for ASC037 and 22 to 72bp for ASC031. C30 and C35 lines were amplified with ASC035 markers on 10 th chromosome and ASC037 on 5 th chromosome. P1, P6, P18 and P41 lines had only one of the resistance components by matching with ASC031 marker in the 2 nd chromosome. The C-32 and C-4 lines  and 16.84 t ha -1 fresh chili pepper yields respectively. The lines were significantly different by their fruit weight, fruit length and fruit width values individually (Table 2). The two breeding populations for agronomic components showed that there were no differences between BC 2 S 4 populations of CM334 and PM217 except for yield values. Mean differences related to yield between two improved populations originated from CM334 and PM217 was 4.20 t ha -1 . PM217 was found as more suitable resistance source to improve P. capsici resistant chili pepper lines than CM334 for transferring yield component traits. Fruit width, fruit length, fruit flesh thickness and fruit weight were not affected from resistance source ( Table  1).
Local genotype KM211 having resistance to the pathogen featured in breeding program has enabled to improve lines as resistant as CM334 and satisfying resistant lines from PM217. More yielding than registered varieties and satisfying resistant lines could be improved using PM217. Resistant alleles have been originated more frequently from the resistant parent, but they occasionally have been originated from the susceptible parent. Susceptible parents can carry resistance gene and resistance can be transferred from 3 rd , 5 th and 11 th chromosomes (Thabuis et al., 2003).
Turkey is one of the countries having most aggressive P. capsici isolates in the World (Oelke et al., 2003). Two breeding populations, including local resistance source KM211 to Phytophthora blight from CM334 and PM217, have been compared by three resistance components (receptivity, inducibility, stability) and fruit and yield characteristics. CM334 was a more effective genotype to transfer all three component of P. capsici resistance to its generations. PM217 could be useful for improving high yielding, suitable for spice processing and satisfactorily resistant chili lines to the pathogen. Improving new varieties to Phytophthora capsici not only requires local resistant genotypes but also complex breeding strategies markers have confirmed the resistance of the improved chili pepper lines selected after inoculation tests.

Agronomic traits
With this breeding program, new resistant chili pepper lines have been improved as resistant as CM334 and suitable for condiment producing. These lines can produce fresh chili pepper presenting almost the yield of Sena variety registered for the purpose of spice pepper production. PM217 has generated more yielding lines but lower resistance level than progenies of CM334. The most yielding line improved from PM217 resistance source was P-74 with 26.90 t ha -1 fresh chili yield while the highest yield (22.99 t ha -1 ) was harvested from C-22, improved from CM334. The longest fruit (121.49 mm) was observed on Carliston variety and the shortest fruits (40.87 mm) were harvested from line P-47. C-22 line had the longest fruit among the improved lines with 83.70 mm length. The highest fruit width (23.66 mm) was observed in C-36. S. Demre, and P-47 genotypes presented the narrowest fruits with 12.17 mm and 15.63 mm, respectively. Fruit weight varied between 12.73 g (Carliston) and 4.14 g (H46).
The highly resistant lines C-18 produced 15.81 t ha -1 and C-29 produced 15.78 t ha -1 fresh chili pepper yields. The fresh chili pepper harvested from resistant lines C-5, P-73 and P-77 yielded 16.05, 20.71 and 21.05 t ha -1 respectively. Moderately resistant lines C-27, C-28, C-34, C-37 and P-2 showed 17.23, 18.01, 17.95, 18.87 were the genotypes that had the only resistance gene indicated by the marking of ASC037 on the 5 th chromosome. C36 line had the resistance genes on the 2 nd and 5 th chromosomes coexist. Many lines derived from PM217 had only the resistance gene on the 10 th chromosome. Hat46, Sena, P6, P74 and P74-1 lines did not possess any resistance genes related to molecular markers.
Molecular markers and stem inoculation tests have proved that resistance of lines can be different despite the genotypes have same alleles. Seven CM334 and three PM217 originated lines have produced amplicons by using three markers but these lines have placed in three different groups according to principal component analysis related to their speed of stem necrosis ( Figure  4). Marker-assisted selection provides many advantages for plant breeding especially determining polygenic characters. It is accepted as a promising tool for breeding quantitative resistance (Thabuis et al., 2004a). However applications of molecular markers in diverse germplasms are generally limited because of phenotype and genotype mismatch (Barchenger et al., 2018). Epistasis and additive effect among resistance genes used in this study and other genes providing resistance may have affected on response of improved lines to pathogen. Lawson et al. (1997) indicated that genetic background can decrease molecular assisted selection (MAS) efficiency and phenotypic reactions arise under different gene effect. However in this study molecular including self-pollination, backcrossing and combination of different resistance sources.