Marker-Assisted Gene Pyramiding and the Reliability of Using SNP Markers Located in the Recombination Suppressed Regions of Sunflower (Helianthus annuus L.)

Rust caused by the fungus Puccinia helianthi and downy mildew (DM) caused by the obligate pathogen Plasmopara halstedii are two of the most globally important sunflower diseases. Resistance to rust and DM is controlled by race-specific single dominant genes. The present study aimed at pyramiding rust resistance genes combined with a DM resistance gene, using molecular markers. Four rust resistant lines, HA-R3 (carrying the R4 gene), HA-R2 (R5), HA-R8 (R15), and RHA 397 (R13b), were each crossed with a common line, RHA 464, carrying a rust gene R12 and a DM gene PlArg. An additional cross was made between HA-R8 and RHA 397. Co-dominant simple sequence repeat (SSR) and single nucleotide polymorphism (SNP) markers linked to the target genes were used to discriminate between homozygotes and heterozygotes in F2 populations. Five pyramids with different combinations of rust resistance genes were selected in the homozygous condition through marker-assisted selection, and three of them were combined with a DM resistance gene PlArg: R4/R12/PlArg, R5/R12/PlArg, R13b/R12/PlArg, R15/R12, and R13b/R15. The pyramiding lines with the stacking of two rust and one DM genes were resistant to all known races of North American sunflower rust and all known races of the pathogen causing DM, potentially providing multiple and durable resistance to both rust and DM. A cluster of 12 SNP markers spanning a region of 34.5 Mb on chromosome 1, which co-segregate with PlArg, were tested in four populations. Use of those markers, located in a recombination suppressed region in marker selection, is discussed.


Introduction
Sunflower (Helianthus annuus L.) is cultivated globally and is highly valued as a source of edible oil rich in lineoleic or oleic acids with a high vitamin E content. However, sunflower crops can be infected by disease-causing bacterial, fungal, and viral pathogens, subsequently reducing yield and quality. Rust caused by the fungus Puccinia helianthi Schwein. and downy mildew (DM) caused by the obligate pathogen Plasmopara halstedii (Farl.) Berl. et. de Toni are the two of the most important sunflower diseases. Both are native to North America (NA) but have spread to nearly every sunflower growing region in the world; for review see [1,2]. Resistance to rust and DM is controlled by race-specific single dominant genes. There is a long history of the use of resistant varieties and hybrids to control rust and DM in sunflower production.
The first rust resistant cultivar was developed in Canada in 1954, and subsequently, two rust resistance genes (R genes), R 1 and R 2 , were discovered [3,4]. Since then, a total of 13 rust R genes, R 1 -R 5 , R 10 -R 15 , P u6 , and R adv , have been reported from sunflower and its wild relatives, which are summarized in Ma et al. [5]. Major gene resistance against biotrophic pathogens, such as rust, is generally unstable and nondurable, due to the emergence of virulent races in pathogen populations. In the early 1960s, only four NA P. helianthi races, 1, 2, 3, and 4, corresponding to races 100, 500, 300, and 700 of the coded triplet system were identified [6]. Over four decades and more, Gulya and Markell [7] reported 39 NA P. helianthi races from 300 rust isolates collected from fields in 2007 and 2008. Friskop et al. [8] identified 29 NA P. helianthi races from 238 single-pustule isolates collected from fields in 2011 and 2012. Among the 13 rust R genes, only seven, R 11 , R 12 , R 13a , R 13b , R 14 , R 15 , and R 16 , remain effectively resistant to all P. helianthi races identified in the USA [5,[9][10][11][12]. In some areas, although a single gene confers resistance to the existing pathogen population, large-scale use of this gene results in the breakdown of resistance. Pyramiding of more than one resistance gene in a single genotype is expected to considerably extend the durability and longevity of resistance due to the low probability of the pathogen being able to assemble multiple, rare virulence genes by mutation or recombination.
In traditional plant breeding, phenotypic selection of superior genotypes within segregating progeny obtained from crosses is a labor-intensive and time-consuming process. Advances in technology have changed agricultural practices over time. Development of modern plant molecular and quantitative genetics over the last three decades has made the integration of biotechnology and conventional breeding possible for many crops. The principles of gene pyramiding assume that parental lines containing target genes and markers linked to the target genes are available. Marker-assisted selection (MAS) can be used to pyramid several R genes into a single host genotype, which has been reported in rice [13][14][15], wheat [16][17][18][19][20], barley [21], soybean [22], and tomato [23], as well as in sunflower [24,25].
The rust R genes R 4 in HA-R3 and R 5 in HA-R2 confer resistance to 96.6% and 78.6% of 238 rust isolates tested, respectively, in the USA in 2011 and 2012 [8], while R 12 in RHA 464, R 13b in RHA 397, and R 15 in HA-R8 confer resistance to all P. helianthi races identified in the USA [10]. These genes have been mapped to different sunflower chromosomes corresponding to linkage groups with linked markers-R 5 on chromosome 2, R 15 on chromosome 8, R 12 on chromosome 11, and R 4 and R 13b on chromosome 13 [5,10,[26][27][28][29].
DM is a seedling disease initiated by soil borne oospores of P. halstedii or infected seeds. The pathogen infects plants through the roots, eventually becoming systemic. There is no rescue treatment once the disease manifests. The inbred line RHA 464 carries a rust R gene, R 12 , as well as a broad-spectrum DM R gene, Pl Arg , which is resistant to all P. halstedii races [29][30][31][32][33]. Pl Arg has recently been genetically and physically mapped using high-density single nucleotide polymorphism (SNP) markers on chromosome 1, and the 12 diagnostic SNP markers co-segregating with Pl Arg span a physical distance of 34.5 Mb, due to suppressed recombination [25]. The distribution of the recombination being population-dependent was reported in wheat [34]. In this study, we report pyramiding of four rust R genes, R 4 , R 5 , R 13b , and R 15 , with R 12 and Pl Arg from RHA 464, as well as R 13b and R 15 , using MAS to promote rust resistance efficiency and durability. Meanwhile, we examined the possible segregation of cluster markers linked to Pl Arg in the four distinct F 2 populations to narrow down the physical interval of Pl Arg .

Parents and Populations for Gene Pyramiding
Five sunflower lines, HA-R2, HA-R3, HA-R8, RHA 397, and RHA 464, were used in the present study. HA-R2 (PI 650753, carrying the R 5 gene) and HA-R3 (PI 650754, carrying R 4 ) were released in 1985; these are selections from Argentinian open-pollinated cultivars [35]. RHA 397 (PI 597974) and HA-R8 (PI 607511) were released in 1997 and 2001, respectively [36,37]. The R 13b gene in RHA 397 originated from a South African line RO-20-10-3-3-2, while R 15 in HA-R8 was derived from a sunflower landrace of PI 432512 collected from Arizona, USA. The RHA 464 (PI 655015) line was released by the Genes 2020, 11, 10 3 of 14 USDA-ASR with the North Dakota Agricultural Experiment Station in 2010 and possesses the rust and DM R genes R 12 and Pl Arg , respectively [38]. R 12 originated from the wild H. annuus PI 413047 and Pl Arg from the wild H. argophyllus PI 468651. The sunflower inbred line HA 89, which is susceptible to all DM and rust races was used as a susceptible control in the present study.

Marker Selection
DNA markers used in the present study are listed in Table 1. The co-dominant nature of the polymorphisms exhibited by the selected markers enables discrimination between homozygotes and heterozygotes in F 2 populations. In addition, 14 SNP markers diagnostic for Pl Arg were used for cluster marker analysis. Among them, 12 SNP markers co-segregate with Pl Arg and physically span a region of 34.5 Mb on chromosome 1 ( Table 2) [25]. Marker-assisted selections were performed in all F 2 generations. For each F 2 population, initial selection was conducted using one marker per gene, and selected multi-R plants were further confirmed with additional markers.  Genomic DNA from each population, along with their parental lines, was extracted from the lyophilized tissues using the DNeasy 96 Plant Kit (Qiagen, Valencia, CA, USA), following the manufacturer's instructions. DNA quantity and quality were determined using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA). Genotyping of the selected homozygous F 2 plants from Pop2 and Pop3, along with the SNP markers, were conducted in BioDiagnostics Inc. (River Falls, WI, USA) in 2014. Genotyping of simple sequence repeat (SSR) and PCR-based SNP markers was performed by following the methods described by Qi et al. [9,39]. PCR products were diluted 20-160 times, depending on their yield and were detected using an IR2 4300/4200 DNA Analyzer with denaturing polyacrylamide gel electrophoresis (LI-COR, Lincoln, NE, USA).

New SSR Marker Development for R 15
The rust resistance gene R 15 was recently mapped to the upper end of sunflower chromosome 8 [5]. Unfortunately, SNP markers linked to R 15 did not have any polymorphisms between HA-R8 and RHA 464. We extracted a sequence from the sunflower reference genome HA412-HO using the flanking markers SFW05824 (physical position 10,040,792 bp) and NSA_008457 (11,391,650 bp). SSRs were identified from the extracted sequences using the SSR Identification Tool from the Gramene. Out of the 13 designed SSR markers, two SSRs, SUN398 and SUN406, that mapped close to R 15 and were used to screen R 15 in Pop4 derived from the HA-R8/RHA 464 cross and the Pop5 derived from HA-R8/RHA 397. The SSR primer sequences were as follows (5 -3 ): SUN398 F-ATCCAACCCGACTTCTTCGG, R-TGACAAACAGCCGCCTCTC and SUN406 F-CTCACTGGAAGCAGCCTCTC, R-TTCCATGTGCATCAATGTGGC.

Disease Evaluation
In addition to DNA marker selection in the F 2 generations, both DM and the rust-resistant tests of these F 2 -derived F 3 families that had two or three genes of interest in the homozygous state were conducted under greenhouse control conditions. The whole seedling immersion method was applied to test reactions to DM in sunflower seedlings using the NA P. halstedii race 734 [39,40], a new virulent race identified in the USA in 2010 [41]. Sunflower seedlings infected with DM display typical leaf chlorosis with white sporulation on the underside of their cotyledons and true leaves. A plant was scored as susceptible (S) if sporulation was observed on the cotyledons and true leaves, and was scored as resistant (R) if no sporulation was observed.
Twenty-four to seventy-two individual seedlings of each F 3 family were first inoculated with P. halstedii race 734, and the resistant plants were then transferred to 36 cell plastic flats (each cell 4.6 cm × 5.4 cm) filled with Sunshine SB 100B potting mixture (SunGro Horticulture, Bellevue, WA, USA). After 10-12 days, seedlings at the four-leaf stage were inoculated with P. helianthi race 336, as described by Qi et al. [9]. The infection types (ITs) of rust were recorded on a scale of 0-4 [42], combined with the percentage of leaf area covered in pustules (severity), as described by Gulya et al. [43], after 12-14 days post inoculation. IT 0, 1, and 2 along with pustule coverage of 0 to 0.5% were recorded as resistant, while IT 3 and 4 with pustule coverage larger than 0.5% were considered susceptible.

Marker Selection and Disease Evaluation of Homozygous Multi-Resistant Plants
Three DNA markers, ORS316, NSA_001392, and NSA_002798, linked to genes R 4 , R 12 , and Pl Arg , respectively, were first used to screen Pop1, which was derived from a cross between HA-R3 and RHA 464, for genotypes being homozygous for the three genes. Out of the 376 F 2 individuals screened, four plants were selected as homozygotes at all three marker loci and were further confirmed by the additional six markers, SFW05240 and SFW01497 for R 4 , NSA_00064 and NSA_001570 for R 12 , and NSA_002851 and NSA_006530 for Pl Arg (Figure 1, Table 1).

Disease Evaluation
In addition to DNA marker selection in the F2 generations, both DM and the rust-resistant tests of these F2-derived F3 families that had two or three genes of interest in the homozygous state were conducted under greenhouse control conditions. The whole seedling immersion method was applied to test reactions to DM in sunflower seedlings using the NA P. halstedii race 734 [39,40], a new virulent race identified in the USA in 2010 [41]. Sunflower seedlings infected with DM display typical leaf chlorosis with white sporulation on the underside of their cotyledons and true leaves. A plant was scored as susceptible (S) if sporulation was observed on the cotyledons and true leaves, and was scored as resistant (R) if no sporulation was observed.
Twenty-four to seventy-two individual seedlings of each F3 family were first inoculated with P. halstedii race 734, and the resistant plants were then transferred to 36 cell plastic flats (each cell 4.6 cm × 5.4 cm) filled with Sunshine SB 100B potting mixture (SunGro Horticulture, Bellevue, WA). After 10-12 days, seedlings at the four-leaf stage were inoculated with P. helianthi race 336, as described by Qi et al. [9]. The infection types (ITs) of rust were recorded on a scale of 0-4 [42], combined with the percentage of leaf area covered in pustules (severity), as described by Gulya et al. [43], after 12-14 days post inoculation. IT 0, 1, and 2 along with pustule coverage of 0 to 0.5% were recorded as resistant, while IT 3 and 4 with pustule coverage larger than 0.5% were considered susceptible.

R4/R12/PlArg Homozygous Plants
Three DNA markers, ORS316, NSA_001392, and NSA_002798, linked to genes R4, R12, and PlArg, respectively, were first used to screen Pop1, which was derived from a cross between HA-R3 and RHA 464, for genotypes being homozygous for the three genes. Out of the 376 F2 individuals screened, four plants were selected as homozygotes at all three marker loci and were further confirmed by the additional six markers, SFW05240 and SFW01497 for R4, NSA_00064 and NSA_001570 for R12, and NSA_002851 and NSA_006530 for PlArg ( Figure 1, Table 1). To confirm the presence of DM and rust resistance and to observe the effects of stacking rust genes, DM and rust tests were performed on selected F 2 -derived F 3 families. The susceptible line HA 89 and the parental line HA-R3 showed the expected susceptibility reactions to NA P. halstedii race 734 exposure, after seedling inoculation, while 197 F 3 plants from four F 2 -derived F 3 families exhibited resistance to race 734 (Table 3). Subsequent rust tests revealed that F 3 plants with two gene stacks, R 4 and R 12 , were highly resistant, showing hypersensitive fleck without pustules after seedling inoculation with P. helianthi race 336, as compared to the parental lines HA-R3 and RHA 464, which exhibited an IT of 1 and a pustule coverage of 0.1% (Table 3). These results indicated an enhanced rust resistance in these plants. DM and rust tests confirmed that the entries were homozygous for R 4 , R 12 , and Pl Arg .  (Table 3). Comparing the effect of the different resistance sources for rust disease, the effect of the rust resistance gene in RHA 464 was higher than that in HA-R2. HA-R2 had an IT of 2 with a pustule coverage of 0.5%, while RHA 464 had an IT of 1 with a pustule coverage of 0.1% (Table 3). The selected plants with two gene stacks, R 5 and R 12 , had a reaction to rust infection similar to RHA 464.
3.1.3. R 13b /R 12 /Pl Arg Homozygous Plants Initial screens of Pop3, derived from the RHA 397/RHA 464 cross, were conducted using three markers, ORS316 for R 13b , NSA_001392 for R 12 , and ORS610 for Pl Arg , and three of the 758 F 2 plants tested were homozygous at all three marker loci. The selected triple R-plants were further confirmed with six additional SNP markers, three NSA_000187, NSA_005565, and NSA_006846, one NSA_001570, and two NSA_002851 and NSA_006530 linked to the three targeted genes, respectively ( Table 1).
As expected, the susceptible control HA 89 and the parental RHA 397 lines were susceptible to downy mildew, while the 180 F 3 individuals from the three F 2 -derived F 3 families, along with their resistant donor RHA 464, were resistant to inoculation with P. halstedii race 734. All F 3 plants carrying R 13b and R 12 proved to be free of rust infection after inoculation with P. helianthi race 336 (Table 3). This demonstrated that the combination of R 13b and R 12 exerted an additive effect on the degree of resistance to rust.

R 15 /R 12 Homozygous Plants
Two new SSR markers were developed in the present study due to a lack of polymorphic SNP markers linked to R 15 in the HA-R8/RHA 464 F 2 population. Linkage analysis with 186 F 2 segregating plants derived from the HA 89/HA-R8 cross previously used to map the R 15 gene [5] indicated that SSRs SUN398 and SUN406 co-segregated with a previously mapped SNP marker, SFW05824, distal to R 15 at a genetic distance of 0.4 cM ( Table 1).
Unexpectedly, 92 F 3 plants from the two F 2 -derived F 3 families exhibited homozygous susceptibility to DM, similar to the susceptible parent HA-R8, after inoculation with P. halstedii race 734 (Table 3). Pl Arg in RHA 464 co-segregated with 12 SNP markers that spanned a region of 34.5 Mb on chromosome 1 ( Table 2) [25]. The two SNP markers used in the above screening, NSA_002798 and NSA_001835, were located on the lower end of the marker cluster (Table 2). One possibility is that recombination occurred between the clustered markers and the gene during line development, which altered the linkage phase between the markers and Pl Arg . To confirm this hypothesis, we tested these 14 SNP markers in all selected F 2 individuals from the four F 2 populations, from which the F 3 families were derived (see below).
Because the F 3 families tested were susceptible to DM and died, we regrew 44 and 48 pyramiding individuals from each of the two F 3 families carrying R 12 and R 15 for rust evaluation. No segregation was detected in the rust phenotypic assessment, indicating the homozygous state of the selected F 3 families. All F 3 plants exhibited hypersensitive fleck without pustules after seedling inoculation with P. helianthi race 336, indicating an increased resistance to rust, as compared to both parents (Table 3).

R 13b /R 15 Homozygous Plants
Two SSR markers, ORS316 and SUN398, targeting the rust R genes R 13b and R 15 , respectively, were used to screen Pop5, derived from the RHA 397 and HA-R8 cross. Twenty plants from 376 F 2 individuals tested were homozygous at both marker loci, which was confirmed by an additional two markers, HT382 for R 13b and SUN406 for R 15 (Figure 2).
A total of 192 F 3 plants from the four selected F 3 families were evaluated for rust resistance using the P. helianthi race 336, along with the susceptible control HA 89 and both parents, RHA 397 and HA-R8. All F 3 plants exhibited a hypersensitive fleck without pustules, compared to the susceptible control HA 89, which had an IT of 4, and more than 40% of the leaves were covered with pustules, and both parents, which had an IT of 1% and 0.1% of leaves, were covered with pustules (Table 3).

R13b/R15 Homozygous Plants
Two SSR markers, ORS316 and SUN398, targeting the rust R genes R13b and R15, respectively, were used to screen Pop5, derived from the RHA 397 and HA-R8 cross. Twenty plants from 376 F2 individuals tested were homozygous at both marker loci, which was confirmed by an additional two markers, HT382 for R13b and SUN406 for R15 (Figure 2).
A total of 192 F3 plants from the four selected F3 families were evaluated for rust resistance using the P. helianthi race 336, along with the susceptible control HA 89 and both parents, RHA 397 and HA-R8. All F3 plants exhibited a hypersensitive fleck without pustules, compared to the susceptible control HA 89, which had an IT of 4, and more than 40% of the leaves were covered with pustules, and both parents, which had an IT of 1% and 0.1% of leaves, were covered with pustules (Table 3).

Detection of Recombination in the Marker Cluster Linked to PlArg
A total of 14 SNP markers that are diagnostic for RHA 464 marker alleles linked to PlArg were used to detect recombination among cluster markers in the multi-resistant F2 plants selected from the four different F2 populations. Of the 14 SNP markers selected, 12, spanning a region of 34.5 Mb on chromosome 1 physical map, co-segregated with PlArg, and two were proximal to PlArg, with genetic distances of 0.31 and 0.83 cM ( Table 2) [25]. No recombination was detected in F2 plants derived from Pop1 of the HA-R3/RHA 464 cross

Detection of Recombination in the Marker Cluster Linked to Pl Arg
A total of 14 SNP markers that are diagnostic for RHA 464 marker alleles linked to Pl Arg were used to detect recombination among cluster markers in the multi-resistant F 2 plants selected from the four different F 2 populations. Of the 14 SNP markers selected, 12, spanning a region of 34.5 Mb on chromosome 1 physical map, co-segregated with Pl Arg , and two were proximal to Pl Arg , with genetic distances of 0.31 and 0.83 cM ( Table 2) [25]. No recombination was detected in F 2 plants derived from Pop1 of the HA-R3/RHA 464 cross or Pop2 of HA-R2/RHA 464 (Table 4). In Pop3 of RHA 397/RHA 464, one (14-21-129) of the three F 2 plants exhibited recombination between markers NSA_008037 and NSA_007595, based on their physical positions ( Figure 3, Table 4). The 14-21-129 F 2 plant had heterozygous alleles at three SNP loci, NSA_007595, NSA_001835, and NSA_006530, while the F 2 -derived F 3 family exhibited homozygous resistance to DM, indicating that the recombination did not involve the Pl Arg locus (Table 3).  (Table 4). In Pop3 of RHA 397/RHA 464, one (14-21-129) of the three F2 plants exhibited recombination between markers NSA_008037 and NSA_007595, based on their physical positions ( Figure 3, Table 4). The 14-21-129 F2 plant had heterozygous alleles at three SNP loci, NSA_007595, NSA_001835, and NSA_006530, while the F2-derived F3 family exhibited homozygous resistance to DM, indicating that the recombination did not involve the PlArg locus (Table 3). In Pop4, derived from the HA-R8/RHA 464 cross, recombination was detected between SNP markers NSA_005063 and NSA_002851 based on their physical position in the two selected F2 plants, 16-46-202 and 16-46-329 (Figure 4, Tables 2 and 4). Five SNPs, NSA_002208, NSA_000630, NSA_004149, NSA_005423, and NSA_005063, physically located in a region between 106 and 123 Mb, were homozygous for the HA-R8 14-21-129. 14-21-129 was homozygous at the NSA_005423 and NSA_008037 loci but heterozygous at the NSA_007595 and NSA_001835 loci.  In Pop4, derived from the HA-R8/RHA 464 cross, recombination was detected between SNP markers NSA_005063 and NSA_002851 based on their physical position in the two selected F 2 plants, 16-46-202 and 16-46-329 (Figure 4, Tables 2 and 4). Five SNPs, NSA_002208, NSA_000630, NSA_004149, NSA_005423, and NSA_005063, physically located in a region between 106 and 123 Mb, were homozygous for the HA-R8 SNP alleles, while the remaining nine SNPs physically located in a region between 124 and 145 Mb were homozygous for the RHA 464 SNP alleles. Two SNP markers, NSA_002798 and NSA_001835, which were used for the initial selection of F 2 plants, belonged to the latter group. Although the two selected F 2 plants had RHA 464 alleles at both NSA_002798 and NSA_001835 loci, the F 2 -derived F 3 families exhibited homozygous susceptibility to DM, indicating that the genetic linkage between the markers and the Pl Arg gene was broken, altering the linkage phase of Pl Arg with the markers. Combined phenotyping and genotyping data placed Pl Arg in a position close to the first five SNP markers in the cluster ( Figure 5, Table 4).

Discussion
Marker-assisted gene pyramiding has previously been successfully used in plant breeding, especially when selecting for disease and insect resistance controlled by major genes; for review see [44]. In the present study, we developed five pyramids with different rust R gene combinations, three of which were combined with a DM R gene: PlArg, R4/R12/PlArg, R5/R12/PlArg, R13b/R12/PlArg, R15/R12, and R13b/R15. Accumulating major genes for resistance in an elite genotype by conventional breeding is laborious and time-consuming when one or more of the genes are effective against all known isolates of the pathogen. Due to a lack of P. helianthi race to differentiate the R12 gene from the other four rust R genes, selection for plants having multiple genes using molecular markers is extremely important. The co-dominant nature of both SSR and SNP markers used in this study made it possible to select homozygous pyramids in the F2 generation. Rust evaluation of the F2-derived F3 families indicated that the pyramids generally showed enhanced resistance to the P. helianthi pathogen, compared to the parental lines, demonstrating a complementary effect of the two R genes when present together. The pyramids carrying R4/R12/PlArg, R13b/R12/PlArg, R15/R12, and R13b/R15 proved  Table 2. Lower case letters represent the HA-R8 SNP allele and the upper case letters represent the RHA 464 SNP allele. Pl Arg co-segregated with the first five markers.

Discussion
Marker-assisted gene pyramiding has previously been successfully used in plant breeding, especially when selecting for disease and insect resistance controlled by major genes; for review see [44]. In the present study, we developed five pyramids with different rust R gene combinations, three of which were combined with a DM R gene: Pl Arg , R 4 /R 12 /Pl Arg , R 5 /R 12 /Pl Arg , R 13b /R 12 /Pl Arg , R 15 /R 12 , and R 13b /R 15 . Accumulating major genes for resistance in an elite genotype by conventional breeding is laborious and time-consuming when one or more of the genes are effective against all known isolates of the pathogen. Due to a lack of P. helianthi race to differentiate the R 12 gene from the other four rust R genes, selection for plants having multiple genes using molecular markers is extremely important. The co-dominant nature of both SSR and SNP markers used in this study made it possible to select homozygous pyramids in the F 2 generation. Rust evaluation of the F 2 -derived F 3 families indicated that the pyramids generally showed enhanced resistance to the P. helianthi pathogen, compared to the parental lines, demonstrating a complementary effect of the two R genes when present together. The pyramids carrying R 4 /R 12 /Pl Arg , R 13b /R 12 /Pl Arg , R 15 /R 12 , and R 13b /R 15 proved to be free of rust infection (Table 3). These lines, once released, will serve as valuable germplasms for the breeder to use in breeding programs. As a hybrid crop, sunflower breeders can transfer the different gene combinations into the cytoplasm male sterile (CMS) and male fertility restorer (Rf) lines, respectively, and the resulting F 1 hybrids from the crosses of CMS/Rf will carry multiple rust R genes and a DM R gene, effectively suppressing the emergence of virulent isolates of the rust pathogen and potentially providing wider spectra and durable resistance to rust.
The minimum population size for successful recovery of a desirable genotype can be calculated in a three-unlinked gene pyramiding project. To obtain one F 2 individual that is homozygous for resistance alleles at all three gene loci with a 99% probability of success, 293 individuals must be evaluated [45]. In the present study, we screened the four F 2 populations, with sizes ranging from 376 to 758 F 2 individuals, recovering three-gene pyramids from Pop1 (4/376), Pop2 (4/756), and Pop3 (3/758); Pop4 was an exception. As the recombination occurred between the flanking markers of Pl Arg , two selected F 2 plants from 470 F 2 individuals had lost the Pl Arg gene (Table 4).
Pl Arg was originally transferred from a wild H. argophyllus into cultivated sunflower in 1989 with no reports of resistance breakdown for more than 25 years [31,33,46]. Molecular mapping placed Pl Arg in a region with highly suppressed recombination on sunflower chromosome 1 [30,47]. Qi et al. [25] reported that 78 SNP markers co-segregated with Pl Arg in an F 2 population derived from the cross of HA 89/RHA 464, and 12 of them were diagnostic for Pl Arg , which spanned a physical distance of 34.5 Mb (between 106.0 Mb and 140.5 Mb) in the HA412-HO genome assembly (Table 2). In the present study, the recombination events were observed in the marker cluster in the two F 2 populations of RHA 397/RHA 464 and HA-R8/RHA 464. Crossover occurred between NSA_008037 and NSA_007595 in plant 14-21-129 in Pop3 of RHA 397/RHA 464 but did not change the linkage phase of Pl Arg with the markers (Table 4). Plant 14-21-129 displayed heterozygous alleles in three marker loci, NSA_007595, NSA_001835, and NSA_006530, while the F 3 family derived from 14-21-129 exhibited homozygous resistance to DM. However, the observed recombination between NSA_005063 and NSA_002851 in Pop4 of HA-R8/RHA 464 did change the linkage phase of the Pl Arg with markers. Of the 14 SNP markers tested, selected F 2 plants showed HA-R8 SNP alleles in the first five SNP loci and RHA 464 SNP alleles in the latter nine SNP loci (Tables 2 and 4). The F 2 -derived F 3 families were all susceptible to DM, indicating that Pl Arg is close to the first five SNPs in the marker cluster ( Figure 5). This finding narrows down the Pl Arg -harboring region from 34.5 Mb to 17.3 Mb-between 106.0 Mb and 123.3 Mb (Table 2).
A high-density SNP map for Pl Arg was constructed using a biparental F 2 population, as is the case for most genetic maps generated where one cycle of meiosis provided all recombination events in the population [25]. Whereas, a breeding program might involve more than two parents and multiple crosses, as a result, greatly increasing the recombination rate in breeding population and decreasing the reliability of the marker-based selection because of the increasing cycles of meiosis. Although the present study placed Pl Arg within a five-SNP cluster that co-segregated with the gene, these markers still spanned a physical distance of 17.3 Mb. With potential increases in recombination in breeding populations, there is a chance that crossover occurred between Pl Arg and the five SNP-cluster. Using SNP markers selected from a five SNP-cluster combining the closest flanking marker, NSA_002851, is recommended in MAS for breeding programs, which would greatly increase the reliability of the markers for predicting phenotypes. Funding: This research was supported by the USDA-ARS CRIS Project No. 3060-2100-043-00D. The mention of trade names or commercial products in this report is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the US Department of Agriculture. The USDA is an equal opportunity provider and employer.