Genetic Mapping of Seven Kinds of Locus for Resistance to Asian Soybean Rust

Abstract

Asian soybean rust (ASR), caused by Phakopsora pachyrhizi, is one of the most serious soybean (Glycine max) diseases in tropical and subtropical regions. To facilitate the development of resistant varieties using gene pyramiding, DNA markers closely linked to seven resistance genes, namely, Rpp1, Rpp1-b, Rpp2, Rpp3, Rpp4, Rpp5, and Rpp6, were identified. Linkage analysis of resistance-related traits and marker genotypes using 13 segregating populations of ASR resistance, including eight previously published by our group and five newly developed populations, identified the resistance loci with markers at intervals of less than 2.0 cM for all seven resistance genes. Inoculation was conducted of the same population with two P. pachyrhizi isolates of different virulence, and two resistant varieties, ‘Kinoshita’ and ‘Shiranui,’ previously thought to only harbor Rpp5, was found to also harbor Rpp3. Markers closely linked to the resistance loci identified in this study will be used for ASR-resistance breeding and the identification of the genes responsible for resistance.

1. Introduction

Asian soybean rust (ASR), caused by Phakopsora pachyrhizi (Sydow and Sydow), is one of the most serious soybean diseases, especially in major tropical and subtropical soybean (Glycine max Merrill)-growing regions worldwide. The huge cost of fungicide control or severe yield loss is common in Asia, Africa, and South America [1,2,3]. Although the introduction of resistant varieties has advantages in terms of both cost and reduced environmental impact, proper selection of resistance genes is essential, considering the virulence of the ASR fungus.

The virulence of ASR varies at the field level, as well as with annual and regional variations. In Mexico, two types of ASR pathogens with largely different virulence characteristics have been reported to be present simultaneously in two regions of the country, approximately 1000 km apart [4]. In Bangladesh, the virulence of ASR fungi changed significantly and increased between 2016 and 2018 [5]. In Uruguay, the variability of ASR pathogens within soybean fields was high [6]. Therefore, ASR-resistant varieties must adapt to these pathogenic variations.

Pyramiding the resistance gene Rpps is an effective means of addressing such diverse and variable ASR virulence. In addition, Rpp-pyramiding confers strong synergistic resistance to soybean plants [7,8,9]. The synergistic effects of Rpp-pyramiding have been used to develop new ASR-resistant varieties, and their high resistance is effective at the field level [10]. To identify soybean lines that carry multiple Rpp genes without DNA selection, multiple P. pachyrhizi races are matched to the target genes, and a subsequent progeny test is required to confirm homozygosity/heterozygosity. In addition, it is sometimes difficult to determine the presence of resistance genes by phenotypic values since the effect of Rpp gene on the phenotype is influenced by the genetic background, except for Rpp gene [11]. Therefore, the selection of resistant soybean lines using DNA markers is essential for the effective and efficient breeding of resistant varieties through Rpp-pyramiding. The selection of soybean genotypes by two DNA markers that are closely linked to each other and sandwich a resistance locus contributes not only to variety development but also to the efficient generation of near-isogenic lines (NILs) [12].

In marker-assisted breeding, using DNA markers linked to the target gene, undesirable traits may be introduced along with the target gene owing to linkage drag [13]. Therefore, the DNA markers should be located closer to the target gene. Thus, we reanalyzed previously published linkage maps for the ASR resistance loci Rpp1, Rpp1-b, Rpp2, and Rpp3 [14,15,16,17], and identified DNA markers that are closely linked to these loci by additive marker positioning. In addition, we analyzed new segregants for Rpp1-b, Rpp4, Rpp5, and Rpp6 and identified DNA markers closely linked to these loci.

2. Results

2.1. Genetic Loci of Seven Rpps

Rpp1 was mapped using three mapping populations—Xiao Jin Huang × BRS184, BRS184 × Himeshirazu, and KS 1034 × BRS184—as described in previous studies [14,15]. Reconstruction of linkage maps by adding new simple sequence repeat (SSR) markers revealed that Rpp1 was mapped at the position of 10.8 cM from the top of the linkage group between BARCSOYSSR(SSR)18_1793 (5.6 cM) and SSR18_1854 (12.5 cM) in Xiao Jin Huang × BRS184, the position of 11.7 cM between Sct_187 (11.1 cM) and Sat_064 (12.3 cM) in BRS184 × Himeshirazu, and the position of 6.3 cM between Sct_187 (4.3 cM) and Sat_064 (8.7 cM) in KS 1034 × BRS184, respectively (Figure 1,

Table 1. Genetic maps of resistance genes to Asian soybean rust (ASR) constructed using DNA markers.

Pop.	Parental Cross	Target Gene	Chromo-some	Numbers of Markers	Numbers of Marker Loci	Total Length of Linkage Group	Position of Target Rpps 2	One of the Nearest Markers 3
1	Xiao Jin Huang × BRS184	Rpp1	Chr. 18	10	7	15.4 cM	10.8 cM	SSR18_1854
2	BRS184 × Himeshirazu	Rpp1	Chr. 18	9	7	14.1 cM	11.7 cM	Sct_187
3	KS 1034 × BRS184	Rpp1	Chr. 18	6	5	10.6 cM	6.3 cM	Sct_187
4	BRS184 × PI 594767A	Rpp1-b	Chr. 18	7	5	17.0 cM	16.4 cM	SSR18_1861
5	BRS184 × PI 587905	Rpp1-b	Chr. 18	6	4	5.6 cM	5.4 cM	SSR18_1861
6	BRS184 × PI 587855	Rpp1-b	Chr. 18	11	8	15.9 cM	13.3 cM	SSR18_1861
7	PI 587880A × No12-1-A 1	Rpp1-b	Chr. 18	17	8	11.1 cM	7.4 cM	SSR18_1861
8	Iyodaizu B × BRS184	Rpp2	Chr. 16	9	8	15.7 cM	8.7–9.4 cM	SSR16_0908
9	BRS184 × PI 416764	Rpp3	Chr. 6	20	11	9.0 cM	2.4–6.0 cM	Satt079 4
10	BRS184 × PI 459025	Rpp4	Chr. 18	15	7	15.5 cM	2.1–3.1 cM	SSR18_1568
11	BRS184 × Kinoshita	Rpp5	Chr. 3	8	8	17.0 cM	2.3–4.2 cM	SSR03_0929
		Rpp3	Chr. 6	14	11	11.1 cM	0.4–2.0 cM	Sat_263
12	BRS184 × Shiranui	Rpp5	Chr. 3	7	7	22.0 cM	5.2–8.5 cM	Sat_275
		Rpp3	Chr. 6	11	7	8.7 cM	1.5–3.3 cM	Sat_263
13	BRS184 × PI 567102B	Rpp6	Chr. 18	13	10	13.1 cM	1.5–3.5 cM	SSR18_0356

¹ Rpp2 and Rpp5 regions are not included in this table because the pathogen is virulent to these genes. ² Positions of Rpp loci identified as quantitative trait loci (QTLs) are shown as intervals. ³ For QTL, the marker closest to the peak LOD value. Other markers may be present and co-located with the listed marker or at the same distance from Rpp. ⁴ The marker closest to the SL peak within the interval was considered representative because the peaks of the LOD values for the three traits did not match.

). The peaks of logarithm of the odds (LOD) scores for numbers of uredinia per lesion (NoU), frequency of lesions with uredinia (%LU), and sporulation level (SL) by additional quantitative trait locus (QTL) analyses were located at the same interval as the SSR markers (Supplementary Figure S1). As a result, Rpp1 was found to be located in the genomic region between Sct_187 and SSR18_1854.

Rpp1-b was mapped using three mapping populations—BRS184 × PI 594767A, BRS184 × PI 587905, and BRS184 × PI 587855—which were used in previous studies [16,17], and a new mapping population, PI 587880A × No12-1-A. Linkage maps were reconstructed by adding new SSR markers, and Rpp1-b was mapped at the position of 16.4 cM from the top of the linkage group between SSR18_1861 (16.4 cM) and Sat_372 (17.0 cM) in BRS184 × PI 594767A, the position of 5.4 cM between SSR18_1861 (5.2 cM) and Sat_372 (5.6 cM) in BRS184 × PI 587905, the position of 13.3 cM between SSR18_1861 (13.1 cM) and SSR18_1862 (13.5 cM) in BRS184 × PI 587855, and the position of 7.4 cM between SSR18_1861 (7.4 cM) and SSR18_1862 (8.4 cM) in BRS184 × PI 587855, respectively (Figure 1, Table 1). Rpp1-b locus was considered to be positioned opposite Rpp1, across the marker Sat_064 (Figure 1). The peak positions of the LOD values of the QTL analysis for resistance-related traits were also in this vicinity in most cases (Supplementary Figure S1).

Rpp2 of Iyodaizu B was mapped as a QTLs for three resistance characters, NoU, %LU, and SL, by a reconstructed linkage map using the Iyodaizu B × BRS184 population (Figure 2). The LOD peaks of QTLs for these three characters were detected at the same position of 0.7 cM-interval between the markers: Satt620 (the position of 8.7 cM from the top of linkage group) and SSR16_0908 (9.4 cM) (Figure 2, Table 1).

Rpp3 of PI 416,764 was also re-mapped as QTLs for three ASR resistance characters using the BRS184 × PI 416764 population (Figure 3). However, the peak positions of the QTLs for these three characters were different: the position of 5.4 cM for NoU, 2.9 cM for %LU, and 3.6 cM for SL, respectively (Figure 3). Thus, the 3.6 cM interval between the markers SSR06_1466 (position: 2.4 cM) and SSR06_1516 (6.0 cM) containing three peaks of QTLs were selected as the candidate regions where Rpp3 was located (Table 1, Figure 3).

Rpp4 of PI 459025 was mapped as a QTL using the BC₃F₂ population developed in this study and derived from the BRS184 × PI 459,025 cross (Figure 4). The LOD peaks of three characters are located at the same position of 2.6 cM where three markers—Sc21_3420, SSR18_1557, and SSR18_1568—were located. The 1.0 cM interval between the markers SSR18_1551 (position of 2.1 cM) and SSR18_1572 (3.1 cM) containing the LOD peaks was identified as the candidate region where Rpp4 was located (Table 1, Figure 4).

Rpp5 was mapped using two newly developed mapping populations derived from crosses BRS184 × Kinoshita and BRS184 × Shiranui. By the infection of Brazilian BRP-2.5 isolate, in both populations, QTLs for NoU, %LU, and SL were detected in the region of Chromosome 3 where Rpp5 was previously mapped [18] (Figure 5). However, upon infection with the Japanese E1-4-12 isolate, QTLs for these traits were detected in the region of chromosome 6 where Rpp3 is located. No QTL was detected in the Rpp3 region after BRP-2.5 inoculation, and, conversely, no QTL was detected in the Rpp5 region at E1-4-12 inoculation (Figure 5). Rpp5 in Kinoshita and Shiranui were determined presenting 1.9 cM interval between the markers, Sat_275 (position: 2.3 cM) and SSR03_0929 (4.2 cM), and 3.3 cM interval between SSR03_0737 (5.2 cM) and SSR03_0929 (8.5 cM), respectively. Rpp3 in Kinoshita and Shiranui were located at 1.6 cM interval between the markers, Satt460 (position: 0.4 cM) and SSR06_1496 (2.0 cM), and 1.8 cM interval between Sat_263 (1.5 cM) and SSR06_1496 (3.3 cM), respectively. Rpp5 and Rpp3 were mapped to the same regions in both populations (Table 1, Figure 5).

Rpp6 of PI 567102B was mapped as a QTL using the F₂ population developed in this study and derived from the BRS184 × PI 567102B cross (Figure 6). The LOD peaks of the three characteristics were located at the same position (2.5 cM) from the top of the linkage group. The 2.0 cM interval between the markers SSR18_0356 (1.5 cM) and SSR18_0368 (3.5 cM) containing the LOD peaks was identified as the candidate region where Rpp6 was located (Table 1, Figure 6).

2.2. Genetic Effect of Seven Rpps

For the three traits: NoU, %LU, and SL for resistance, the contribution of each resistance gene to phenotypic variance (variance explained: VE, %) and genetic effects (additive and dominance effects) were calculated based on the genotype of the marker closest to each locus (

Table 2. Genetic effects and variance explained of resistance genes for the character trait, numbers of uredinia per lesion (NoU). Data in this table are based on the nearest marker for each resistance gene.

Pop.	Parental Cross	Target Gene	The Nearest Markers	Pathogen	Additive Effect 1	Dominance Effect 1	Degree of Dominance	Variance Explained (%)
1	Xiao Jin Huang × BRS184	Rpp1	SSR18_1854	T1-2	−0.69	−0.72	1.04	48.99%
2	BRS184 × Himeshirazu	Rpp1	Sct_187	E1-4-12	−0.76	−0.76	1.00	64.65%
3	KS 1034 × BRS184	Rpp1	Sct_187	E1-4-12	−0.93	−0.80	0.86	60.54%
4	BRS184 × PI 594767A	Rpp1-b	SSR18_1861	T1-2	−1.40	−1.40	1.00	55.93%
5	BRS184 × PI 587905	Rpp1-b	SSR18_1861	T1-2	−1.68	−1.55	0.92	64.18%
6	BRS184 × PI 587855	Rpp1-b	SSR18_1861	E1-4-12	−0.85	−0.85	1.00	65.41%
7	PI 587880A × No12-1-A	Rpp1-b	SSR18_1861	BRP-2.1	−0.86	−0.87	1.01	48.94%
8	Iyodaizu B × BRS184	Rpp2	SSR16_0908	E1-4-12	−0.81	−0.51	0.62	69.42%
9	BRS184 × PI 416764	Rpp3	Satt079	T1-2	−1.05	−0.22	0.21	65.73%
10	BRS184 × PI 459025	Rpp4	SSR18_1568	E1-4-12	−0.90	−0.41	0.46	70.59%
11	BRS184 × Kinoshita	Rpp5	SSR03_0929	BRP-2.5	−1.15	−0.03	0.02	82.60%
		Rpp3	Sat_263	E1-4-12	−0.81	−0.74	0.91	51.50%
12	BRS184 × Shiranui	Rpp5	Sat_275	BRP-2.5	−0.97	−0.09	0.09	52.29%
		Rpp3	Sat_263	E1-4-12	−0.82	−0.81	0.99	52.37%
13	BRS184 × PI 567102B	Rpp6	SSR18_0356	E1-4-12	−0.64	−0.61	0.96	59.07%

¹ Genetic effects of the resistant allele contrasted with the susceptible allele of BRS184 in all populations, except population 7. In population 7, the susceptible allele of Rpp1-b was the No12-1-A allele.

,

Table 3. Genetic effects and variance explained of resistance genes for the character trait, frequency of lesions with uredinia (%LU). Data in this table are based on the nearest marker for each resistance gene.

Pop.	Parental Cross	Target Gene	The Nearest Markers	Pathogen	Additive Effect 1	Dominance Effect 1	Degree of Dominance	Variance Explained (%)
1	Xiao Jin Huang × BRS184	Rpp1	SSR18_1854	T1-2	ns	ns	ns	ns
2	BRS184 × Himeshirazu	Rpp1	Sct_187	E1-4-12	−46.17%	−46.17%	1.00	66.58%
3	KS 1034 × BRS184	Rpp1	Sct_187	E1-4-12	−49.57%	−40.84%	0.82	59.62%
4	BRS184 × PI 594767A	Rpp1-b	SSR18_1861	T1-2	−23.21%	−23.06%	0.99	57.61%
5	BRS184 × PI 587905	Rpp1-b	SSR18_1861	T1-2	−49.47%	−37.37%	0.76	69.03%
6	BRS184 × PI 587855	Rpp1-b	SSR18_1861	E1-4-12	−47.62%	−47.42%	1.00	67.78%
7	PI 587880A × No12-1-A	Rpp1-b	SSR18_1861	BRP-2.1	−44.71%	−45.24%	1.01	62.11%
8	Iyodaizu B × BRS184	Rpp2	SSR16_0908	E1-4-12	−45.72%	−21.04%	0.46	72.25%
9	BRS184 × PI 416764	Rpp3	Satt079	T1-2	−33.63%	22.76%	−0.68	59.92%
10	BRS184 × PI 459025	Rpp4	SSR18_1568	E1-4-12	−38.64%	−1.12%	0.03	63.43%
11	BRS184 × Kinoshita	Rpp5	SSR03_0929	BRP-2.5	−40.19%	21.15%	−0.53	72.61%
		Rpp3	Sat_263	E1-4-12	−42.26%	−36.57%	0.87	55.69%
12	BRS184 × Shiranui	Rpp5	Sat_275	BRP-2.5	−29.98%	5.70%	−0.19	36.61%
		Rpp3	Sat_263	E1-4-12	−42.50%	−41.42%	0.97	55.93%
13	BRS184 × PI 567102B	Rpp6	SSR18_0356	E1-4-12	−40.00%	−38.23%	0.96	59.38%

¹ Genetic effects of the resistant allele contrasted with the susceptible allele of BRS184 in all populations, except population 7. In population 7, the susceptible allele of Rpp1-b was the No12-1-A allele.

and

Table 4. Genetic effects and variance explained of resistance genes for the character trait, sporulation level (SL). Data in this table are based on the nearest marker for each resistance gene.

Pop.	Parental Cross	Target Gene	The Nearest Markers	Pathogen	Additive Effect 1	Dominance Effect 1	Degree of Dominance	Variance Explained (%)
1	Xiao Jin Huang × BRS184	Rpp1	SSR18_1854	T1-2	−0.57	−0.61	1.07	45.87%
2	BRS184 × Himeshirazu	Rpp1	Sct_187	E1-4-12	−0.94	−0.94	1.00	66.32%
3	KS 1034 × BRS184	Rpp1	Sct_187	E1-4-12	−1.42	−1.23	0.87	60.54%
4	BRS184 × PI 594767A	Rpp1-b	SSR18_1861	T1-2	−1.46	−1.45	1.00	57.38%
5	BRS184 × PI 587905	Rpp1-b	SSR18_1861	T1-2	−1.48	−1.33	0.90	66.92%
6	BRS184 × PI 587855	Rpp1-b	SSR18_1861	E1-4-12	−1.00	−1.00	1.00	66.84%
7	PI 587880A × No12-1-A	Rpp1-b	SSR18_1861	BRP-2.1	−0.88	−0.90	1.01	59.73%
8	Iyodaizu B × BRS184	Rpp2	SSR16_0908	E1-4-12	−0.96	−0.52	0.55	71.16%
9	BRS184 × PI 416764	Rpp3	Satt079	T1-2	−1.03	0.04	−0.04	71.63%
10	BRS184 × PI 459025	Rpp4	SSR18_1568	E1-4-12	−0.99	−0.24	0.24	69.12%
11	BRS184 × Kinoshita	Rpp5	SSR03_0929	BRP-2.5	−1.10	0.08	−0.07	83.05%
		Rpp3	Sat_263	E1-4-12	−0.98	−0.88	0.90	51.71%
12	BRS184 × Shiranui	Rpp5	Sat_275	BRP-2.5	−1.04	−0.07	0.06	48.93%
		Rpp3	Sat_263	E1-4-12	−1.09	−1.07	0.99	52.55%
13	BRS184 × PI 567102B	Rpp6	SSR18_0356	E1-4-12	−0.80	−0.77	0.96	58.38%

¹ Genetic effects of the resistant allele contrasted with the susceptible allele of BRS184 in all populations, except population 7. In population 7, the susceptible allele of Rpp1-b was the No12-1-A allele.

). For trait NoU, the VE of the resistance genes in each population ranged from 48.99% to 82.60%, with the major genes, Rpps, determining most phenotypes. The gene with the highest VE was Rpp5 in Kinoshita, and the lowest was Rpp1-b in PI 587880A (Table 2). VE in %LU ranged from 0% (not significant) to 72.61%, with the highest value observed for Rpp5 in Kinoshita (Table 3). No significant QTLs were detected for the %LU trait in the Xiao Jin Huang × BRS184 population (Table 3, Supplemental Figure S1); therefore, the phenotypic values between genotypes of the nearest marker were not significantly different. The VE for SL ranged from 45.87% to 83.05%, with the highest value for Rpp5 in Kinoshita and the lowest of Rpp1 in Xiao Jin Huang (Table 4).

The additive effects in Table 2, Table 3 and Table 4 show the relative effects of one resistant allele relative to the susceptible allele for each resistance gene. Negative values indicate that the resistance allele reduced the number of uredinia, frequency of uredinia-forming lesions, and sporulation levels, with larger absolute values indicating a greater effect of the resistant allele. The biggest additive effects for the three traits NoU, %LU, and SL were observed in Rpp1-b of PI 587905 (−1.68), Rpp1 of KS1034 (−49.57%), and Rpp1-b of PI 587905 (−1.48), respectively. However, smaller additive effects for the three traits NoU, %LU, and SL were observed for Rpp6 of PI 567102B (−0.64), Rpp1 of Xiao Jin Huang (not significant), and Rpp1 of Xiao Jin Huang (−0.57). The results showed that the additive effect had a more than two-fold difference in effectiveness, depending on the resistance genes.

Regarding the dominance effect of NoU, all Rpps showed the same negative values as those of the additive effects (Table 2). However, the dominance effect of Rpp3 in PI 416764 was very small compared to the additive effect, and the dominance effect of Rpp5 in Kinoshita and Shiranui was negligible. On the other hand, all Rpp1 or Rpp1-b, Rpp3 in Kinoshita and Shiranui, and Rpp6 in PI 567102B showed a dominance effect equivalent to the additive effect, with the degree of dominance of resistance alleles being 0.86–1.04. Thus, these genes are considered responsible for the complete dominance of the resistant phenotype. The degree of dominance in Rpp2 of Iyodaizu B and Rpp4 of PI 459025 was 0.62 and 0.46, respectively, showing that the resistant phenotype is incomplete dominance. The dominance effect of SL was similar to that of the NoU. The degree of dominance in Rpp1 of three varieties, Rpp1-b of four varieties, Rpp3 of Kinoshita and Shiranui, and Rpp6 of PI 567102B ranged from 0.87–1.07, indicating a complete dominance of the resistance phenotype (Table 4). Since the dominance effect in Rpp3 of PI 416764, and Rpp5 of Kinoshita and Shiranui was hardly observed (−0.07–0.06), resistant and susceptible phenotypes in these genes were considered codominant. The degrees of dominance in Rpp2 of Iyodaizu B and Rpp4 of PI 459025 were 0.55 and 0.24, respectively, showing that the resistant phenotype is incomplete dominance. For %LU, all resistance genes except Rpp1 of Xiao Jin Huang, Rpp3 of PI 416764, Rpp4 of PI 459025, and Rpp5 of Kinoshita had the same degree of dominance as NoU and SL (Table 3). On the other hand, Rpp1 of Xiao Jin Huang was not significant for %LU, Rpp3 of PI 416764 and Rpp5 of Kinoshita were incompletely dominant for the susceptible phenotype (degree of dominance: −0.68, −0.53), and Rpp4 of PI 459025 was codominant.

3. Discussion

3.1. Position of Resistance Loci

Hyten et al. positioned Rpp1 of PI 200492 between Sat_187 and Sat_064 on chromosome 18 using a population crossing the Williams82 variety [19], and the position of Rpp1 concluded in this study, between Sct_187 and SSR18_1854, was consistent with this report. Chakraborty et al. mapped the resistance gene of PI 594538A to Rpp1-b between Sat_064 and Sat_372 on chromosome 18 [20]. Consistent with the conclusion that Rpp1-b is located between SSR18_1861 and SSR18_1862, we restricted the candidate locus to a narrower region.

Yu et al. constructed a detailed genetic map of Rpp2 in PI 230970 and showed that the gene is located between SSR06_0902 and SSR06_0908 on chromosome 16 [21]. Although this study used a different variety, Iyodaizu B, from PI 230970, the Rpp2 locus was in the same region as Satt620 (the same as SSR06_0901) and SSR06_0908.

Monteros et al. placed Rpp3 in Hyuuga between the SSR markers Satt460 and Satt307 on chromosome 6 [22]. Hyten et al. positioned Rpp3 of PI 462312 between Satt460 and Sat_263 [23]. Aoyagi et al. mapped the resistance genes of the resistant soybean varieties “COL/THAI/1986/THAI-80” and “HM 39” as Rpp3 close to the SSR marker Satt079 [14]. Because the Satt079 locus is located between Satt460 and Sat_263, the Rpp3 loci of these four varieties were considered identical. In this study, the peak positions of the LOD scores for the QTL analysis of the three resistance-related traits did not match, so the interval of Rpp3 presence (3.6 cM between SSR06_1466 and SSR06_1516) in PI 416764 could not be narrowed to a short interval. However, this region of Rpp3 matched the one identified in the aforementioned four varieties. In addition, the peak LOD score for SL was detected on Satt079, the nearest marker to Rpp3 (Figure 3). Rpp3 was located between Satt_460 and SSR06_1496 (1.6 cM) in the Kinoshita population and between Sat_263 and SSR06_1496 (1.8 cM) in the Shiranui population (Figure 5), indicating that these Rpp3 loci were also located in the same region, as previously reported. These results indicated that there was no discrepancy between the location of Rpp3 in the four varieties reported in the past and the loci of Rpp3 identified in this study in the three varieties, which are in very close proximity to the marker Satt079.

Meyer et al. mapped Rpp4 of PI 459025 to 4.29 cM interval between the SSR markers Sc21-3360 and Satt288 on chromosome 18, using the population derived from the same parental cross population as in this study [24]. The marker closest to Rpp4 in that report was Satt288, located at a distance of 1.19 cM. In this study, Rpp4 was placed at a 1 cM interval between SSR18_1551 and SSR18_1572. In the intermediate region between these markers, Sc21_3420, SSR18_1557, and SSR18_1568 shared the same locus, and the LOD scores for the three resistance traits were the highest for these three markers (Figure 4). Sc21_3420, which is the same locus as Rpp4 in this study, is 4.67 cM away from Rpp4 on the linkage map of Meyer et al. and the results of the two studies are clearly different. In addition to this marker, we mapped the same markers, Sc21_2716, Sc21_2922, Sc21_4808, and Satt288, as reported by Meyer et al., in this study, however, their positions did not match those of Meyer et al. Since we used the same parents and markers, no speculation can reasonably explain these locus discrepancies. The loci for resistance and DNA markers need to be reconfirmed using another Rpp4-segregating population with the same parental varieties.

Lemos et al. (2011) positioned Rpp5 in Kinoshita approximately 4 cM between the markers Sat_275 and Sat_280 on chromosome 3 by QTL analysis of five resistance-related traits using a population in which Rpp2, Rpp4, and Rpp5 segregate [11]. This Rpp5 position is the same as that of Rpp5 in PI 200456, PI 471904, and PI 200526 (Shiranui), as reported by Garcia et al. [18], and Rpp5 in Hyuuga, as reported by Kendrick et al. [25]. We also mapped Rpp5 in Kinoshita to these regions and narrowed the interval between the markers, where Rpp5 was present between Sat_275 and SSR03_0929 (Figure 5). In contrast, in Shiranui, the peak position of the LOD score for the three traits was Sat_275; therefore, the interval sandwiching Sat_275 between SSR03_0737 and SSR03_0929 was the interval for the presence of Rpp5, and included Kinoshita’s Rpp5 region (Figure 5). The Rpp5 loci identified in this study were consistent with those reported previously.

Li et al. located the rust resistance locus on PI 567102B as Rpp6 between the markers Satt324 and Satt394 on chromosome 18 [26]. In this study, the interval of Rpp6 was located at 2 cM between the two markers SSR18_0356 and SSR18_0368, and 13.1 cM between these markers (Figure 6). These results showed no discrepancies from previous reports for the six resistance loci other than Rpp4.

3.2. The Effect of Introducing Resistance Genes

Because soybean is a diploid self-pollinating crop, additive rather than dominance effects appear with the introduction of resistance genes in breeding. The Rpp1 allele of Xiao Jin Huang was the least effective of all resistance genes analyzed in this study, reducing NoU by 1.38 and SL by 1.14 when introduced as homozygosity (Table 2 and Table 4). No effect was observed on %LU (Table 3). In contrast, Rpp1 in KS1034 reduced NoU by 1.86, %LU by 99.14, and SL by 2.84. Similarly, there were 1.98-, 2.13-, and 1.68-fold differences in NoU, %LU, and SL between PI 587905, which showed the greatest effect, and the other Rpp1-b carrying varieties showed the smallest effect. The difference in effectiveness was more pronounced for different variety alleles at the same locus than among loci, suggesting that the selection of donor varieties will be important at least when Rpp1 or Rpp1-b is used for variety development.

ASR resistance Rpp genes are highly race-specific, and the virulence of ASR pathogens in many regions except Japan is diverse [1,4,5,27]. Under such circumstance, Rpp-pyramiding can lead to high resistance against P. pachyrhizi races that are virulent to the individual Rpp genes that have been pyramided [8]. The genetic effect of each resistance gene was revealed in this study. However, to develop useful rust-resistant soybean varieties, it would be desirable to determine the Rpp genes to be introduced based on the virulence of the ASR pathogen in the target region and the degree of synergistic effects of Rpp-pyramiding.

3.3. The Necessity of Marker-Assisted Selection for ASR Resistance and Utility of Markers

All Rpp1, Rpp1-b, and Rpp6, and Rpp3 of Kinoshita and Shiranui showed complete dominance of the resistance phenotype against the ASR pathogenic strains used in this study. When these resistance genes are introduced into soybean cultivars, it is challenging to distinguish between homozygous and heterozygous lines during selection without DNA markers. Using the DNA markers identified in this study, it was possible to select the presence or absence of resistance genes and homozygosity or heterozygosity in the selected generation without pathogen inoculation and progeny test.

Worldwide, ASR pathogens are diverse in their virulence, and no single resistance gene can guarantee stable resistance. There were highly virulent strains that were pathogenic to all seven resistance genes used in this study, such as the ASR pathogen found in Bangladesh [5]. Additionally, high resistance through the synergic effect of gene pyramiding is useful for developing ASR-resistant varieties [10]. Therefore, gene pyramiding is essential for the development of ASR varieties with a high degree of resistance and stability. For effective and efficient selection of individuals carrying all target genes at homozygosity and synergic effect from a population in which multiple resistance genes are segregated, DNA markers tightly linked to the target genes must be used. This study placed Rpp1, Rpp1-b, Rpp2, Rpp3, Rpp4, Rpp5, and Rpp6 between the DNA markers of 1.2 cM, 0.4 cM, 0.7 cM, 1.6 cM, 1.0 cM, 1.9 cM, and 2.0 cM in minimum, respectively (Table 1, Figure 1, Figure 2, Figure 3, Figure 4, Figure 5 and Figure 6). Because the probability of double recombination between markers below 2.0 cM is less than 0.04%, we can assume that a resistance gene is present if the two markers flanking the resistance locus show the genotype of the resistant parent. Therefore, these markers can be utilized for marker-assisted breeding for ASR resistance. DNA markers flanking these resistance loci can also be used for positional cloning of resistance genes. In addition, the markers around Rpp loci may be used to minimize probability of the introduction of undesirable genes other than Rpp by linkage drag when introducing resistance genes by cross-breeding.

3.4. Second Resistance Gene Rpp3 in Rpp5-Carrying Varieties, Kinoshita and Shiranui

Shiranui, carrying Rpp5, is included in the differential varieties of ASR and has been observed to be more frequently resistant to rust pathogens than other differential varieties with single resistance genes. It was resistant to Bangladeshi ASR samples, with the highest frequency of resistance among the seven genes in the 2016 and 2018 samples [5]. It also showed the highest frequency of resistance to Mexican ASR samples, with resistance in 97.5% of the samples evaluated [4]. It also showed the second-highest frequency of resistance to the South American ASR samples after Rpp1-b. In addition, Shiranui showed resistance to the Japanese ASR samples [27]. The resistance tendency of Shiranui may be due to the presence of both Rpp5 and Rpp3. Shiranui (and Kinoshita) showed resistant and immune (resistant) phenotypes to the Brazilian rust strain BRP-2.5, and the Japanese ASR strain E1-4-12 used in this study, respectively; however, only Rpp5 was resistant to BRP-2.5, and only Rpp3 was resistant to E1-4-12. In addition to the Japanese varieties Kinoshita and Shiranui, Hyuuga has been reported to possess both Rpp3 and Rpp5 [25]. It is not known why ASR-resistant varieties possessing both Rpp3 and Rpp5 are frequently found among ASR-resistant Japanese varieties. However, Walker et al. found that the response patterns of rust-resistant soybean varieties to rust were similar in each country of origin [28]. ASR-resistant common ancestor may be present in soybean varieties from southern Japan.

The fact that Kinoshita and Shiranui have Rpp3 in addition to Rpp5 and the identification of selection DNA markers for each resistance gene, are important findings of this study for resistance breeding. When these varieties are used as resistance donors, if only the known Rpp5 is introduced, the new improved varieties may not show resistance to ASR strains, such as E1-4-12, to which Shiranui and Kinoshita are resistant. If Kinoshita and Shiranui are to be used as resistance donors for breeding, both the Rpp5 and Rpp3 resistance genes must be efficiently and reliably introduced using markers. In addition, to properly monitor the reactions of soybean resistance genes with the virulence genes of pathogenic races, Rpp5 and Rpp3 NILs bred from these varieties must be used instead of Kinoshita and Shiranui. The Rpp5 and Rpp3 selection markers identified in this study will be highly effective in developing NILs.

4. Materials and Methods

4.1. Soybean Populations to Map ASR Resistance Loci

The soybean populations used in this study are listed in

Table 5. Information on soybean populations and Japanese and Brazilian inocula for mapping Rpp resistance loci against Asian soybean rust.

Pop.	Target Gene	Parental Cross	Population Structure	Numbers of Genotypes	Inoculation Target 3	Pathogenic Strain	Reference
1	Rpp1	Xiao Jin Huang × BRS184	F2	90	Plant	T1-2	[15]
2	Rpp1	BRS184 × Himeshirazu	F2	120	Plant	E1-4-12	[15]
3	Rpp1	KS 1034 × BRS184	F2	163	Leaf	E1-4-12	[14]
4	Rpp1-b	BRS184 × PI 594767A	F2	82	Plant	T1-2	[16]
5	Rpp1-b	BRS184 × PI 587905	F2	117	Plant	T1-2	[16]
6	Rpp1-b	BRS184 × PI 587855	F2	106	Plant	E1-4-12	[17]
7	Rpp1-b1	PI 587880A × No12-1-A	F2	96	Plant	BRP-2.1	-
8	Rpp2	Iyodaizu B × BRS184	F2	143	Plant	E1-4-12	[15]
9	Rpp3	BRS184 × PI 416764	F2	86	Plant	T1-2	[16]
10	Rpp4	BRS184 × PI 459025	BC3F2	97	Plant	E1-4-12	-
11	Rpp5, (Rpp3)	BRS184 × Kinoshita 2	F2	130	Leaf	BRP-2.5, E1-4-12	-
12	Rpp5, (Rpp3)	BRS184 × Shiranui	F2	138	Leaf	BRP-2.5, E1-4-12	-
13	Rpp6	BRS184 × PI 567102B	F2	100	Plant	E1-4-12	-

¹ Rpp2 and Rpp5 except for Rpp1-b are also segregated in this population since No6-12-A carries these genes. ² To develop this population, the resistant variety Kinoshita was crossed with the susceptible variety BRS184 as the germ or pollen parent to obtain F₁ individuals. ³ Plant: pathogen was inoculated on the leaves of the plants and the disease was incubated on the plants until evaluation; leaf: pathogen was inoculated on leaves detached from the plants and the disease was cultured on leaves until evaluation.

. DNA and phenotypic data from eight previously published mapping populations [14,15,16,17] were also used in this study. In addition to these populations, five new mapping populations were developed, derived from crosses using resistant varieties: PI 587880A, PI 459025, Kinoshita (PI 200487), Shiranui (PI 200526), and PI 567102B, for mapping Rpp1-b, Rpp4, Rpp5, Rpp5, and Rpp6, respectively. Although the F₂ mapping population derived from PI 587880A × No12-1-A was developed to obtain Rpp-pyramided lines—Py7-1-8-5 carrying Rpp1-b from PI 587880A and Rpp2 and Rpp5 from No6-12-1A [29]—we used this population to map Rpp1-b using the ASR pathogen that is avirulent to Rpp1-b but virulent to Rpp2 and Rpp5. Although the parental combination BRS184 × PI 459025 was used for mapping Rpp4 in a previous study [30], the BC₃F₂ mapping population used in this study was developed from an independent cross. Two kinds of Rpp5-mapping populations were also newly developed by crossing the ASR-susceptible variety, BRS184, and the known Rpp5-carrying resistant varieties, Kinoshita and Shiranui. The resistant variety Kinoshita was crossed with the susceptible variety BRS184 as the germ or pollen parent to obtain F₁ individuals. An F₂ mapping population was obtained using the cross BRS184 × PI 567102B while developing a NIL for Rpp6. The sizes of the 13 mapping populations are presented in Table 5.

These populations were grown under constant temperature and light conditions in a growth chamber, and the leaves of the plants or leaves detached from the plants were inoculated with P. pachyrhizi strains. DNA from the parental varieties and mapping populations was obtained from leaves using the modified CTAB method described in the manual [29].

4.2. Resistant Phenotyping in the Mapping Populations

For the analysis of the eight previously published mapping populations [14,15,16,17], the phenotypic data available at the time were used. However, only two resistance traits, the number of uredinia per lesion (NoU) and sporulation level (SL), have been used in previous studies [14,15,16,17]. In this study, the frequency of lesions with uredinia (%LU) was calculated in addition to NoU and SL. For the five mapping populations—PI 587880A × No12-1-A, BRS184 × PI 459025, BRS184 × Kinoshita, BRS184 × Shiranui, and BRS184 × PI 567102B—developed in this study, the first trifoliate were used for inoculation with P. pachyrhizi strains listed Table 1 by following the manual [29]. In the case of Rpp5-segregating Kinoshita and Shiranui mapping populations, two types of P. pachyrhizi strains, BRP-2.5 and E1-4-12, were used for inoculation because preliminary inoculations with these two strains revealed different segregation of resistance in the partial F₂ populations (data not shown). Each leaflet of the first trifoliate was inoculated with BRP-2.5 and E1-4-12. Inoculation of P. pachyrhizi and evaluation of the resistance-related characteristics NoU, %LU, and SL followed the manual [29]. P. pachyrhizi strains BRP-2.5, T1-2, and E1-4-12 used in this study were obtained from our previous studies [7,8,27]. Urediniospores of these P. pachyrhizi strains were inoculated at a concentration of 3.6–5.8 × 10⁴ urediniospores/mL.

4.3. Genotyping of DNA Markers on the Mapping Populations

For the eight mapping populations—Xiao Jin Huang × BRS184, BRS184 × Himeshirazu, KS 1034 × BRS184, BRS184 × PI 594767A, BRS184 × PI 587905, BRS184 × PI 587855, Iyodaizu B × BRS184, and BRS184 × PI 416764—previously published [14,15,16,17], additional SSR markers were newly analyzed in this study using the DNA solution preserved in −20 °C, respectively. For the five mapping populations—PI 587880A × No12-1-A, BRS184 × PI 459025, BRS184 × Kinoshita, BRS184 × Shiranui, and BRS184 × PI 567102B—the newly developed SSR markers around the target Rpp regions were chosen for analysis. For the PI 587880A × No12-1-A population, SSR markers around the target Rpp1-b locus [16], SSR markers around Rpp2 and Rpp5 [18,30] were also analyzed to confirm that these two loci did not contribute to the resistance phenotype in this mapping population for the P. pachyrhizi strain BRP-2.1. Regarding the mapping populations, BRS184 × PI 459025 for Rpp4 and BRS184 × PI 567102B for Rpp6 SSR markers around these two regions [26,30] were chosen and used for mapping.

In the case of two mapping populations—BRS184 × Kinoshita and BRS184 × Shiranui for Rpp5—a small subset (22 and 24 F₂ plants, respectively) of F₂ plants independent of the main mapping populations (130 and 138 F₂ plants, respectively; Table 5) were evaluated for their resistance phenotypes against E1-4-12 and genotyped using polymorphic and codominant SSR markers Sat_372, Satt380, Sat_263, Satt288, Sat_280, and Satt324 linked to Rpp1 (Rpp1-b), 2, 3, 4, 5 and 6, respectively. A significant association between NoU and marker genotypes was found in the E1-4-12 and Rpp3 markers, and fully mapped populations were analyzed by SSR markers around the regions of Rpp3 and Rpp5. The protocols for the PCR and genotyping of SSR markers are described in the manual [29].

4.4. Linkage Analysis of SSR Markers and Deciding Rpp Loci

MAPMAKER/EXP 3.0b [31] was used to determine the order of the SSR marker loci in the mapping populations. The Kosambi mapping function was used to convert the recombination values into map distances (cM). The minimum logarithm of the odds (LOD) score and the maximum genetic distance for linkage map construction were 3.0 and 37.2 cM, respectively. For the three mapping populations for Rpp1 and four mapping populations for Rpp1-b, F₂ plants were classified as resistant or susceptible by mapping resistance loci as a single resistance factor together with SSR markers, according to our previous studies [14,15,16,17].

In the other mapping populations for Rpp2-Rpp6, genomic regions significantly associated with NoU, %LU, and SL were determined using interval mapping (IM) and composite interval mapping (CIM) from the Windows QTL Cartographer v.2.5 [32]. Other parameters defined for the QTL analysis were 0.5 cM walk speed, 1000 permutations (permutation test), and a 0.01 significance level, following the methodology and parameters employed by [15]. For the CIM analysis, a standard model with five control markers, a window size of 10 cM, and forward- and backward-regression models were applied together with the same parameters as the IM in this study. The resistance loci identified by QTL analysis were determined in the marker interval where the maximum LOD scores for all three characteristics, NoU, %LU, and SL, were detected.

Phenotypic data of resistance and genotypic data of DNA markers for all 13 segregation populations used for mapping can be found in Supplementary Material File S1.