Fine Mapping of Ur-3, a Historically Important Rust Resistance Locus in Common Bean

Bean rust, caused by Uromyces appendiculatus, is a devastating disease of common bean (Phaseolus vulgaris) in the Americas and Africa. The historically important Ur-3 gene confers resistance to many races of the highly variable bean rust pathogen that overcome other rust resistance genes. Existing molecular markers tagging Ur-3 for use in marker-assisted selection produce false results. Here, we describe the fine mapping of the Ur-3 locus for the development of highly accurate markers linked to Ur-3. An F2 population from the cross Pinto 114 (susceptible) × Aurora (resistant with Ur-3) was evaluated for its reaction to four different races of U. appendiculatus. A bulked segregant analysis using the SNP chip BARCBEAN6K_3 placed the approximate location of Ur-3 in the lower arm of chromosome Pv11. Specific SSR and SNP markers and haplotype analysis of 18 sequenced bean varieties positioned Ur-3 in a 46.5 kb genomic region from 46.96 to 47.01 Mb on Pv11. We discovered in this region the SS68 KASP marker that was tightly linked to Ur-3. Validation of SS68 on a panel of 130 diverse common bean cultivars containing all known rust resistance genes revealed that SS68 was highly accurate and produced no false results. The SS68 marker will be of great value in pyramiding Ur-3 with other rust resistance genes. It will also significantly reduce time and labor associated with the current phenotypic detection of Ur-3. This is the first utilization of fine mapping to discover markers linked to rust resistance in common bean.

Genetic resistance is the most cost-effective strategy to manage bean rust disease. Rust resistance in common bean is conditioned by single and dominant genes identified by the Ur-symbol (Kelly et al. 1996). To date, 10 genes have been named and tagged, mostly with RAPD or SCAR molecular markers (Miklas et al. 2002). Five genes  belong to the Middle American gene pool, while five genes  belong to the Andean gene pool (Augustin et al. 1972;Ballantyne 1978;Stavely 1984Stavely , 1990Grafton et al. 1985;Finke et al. 1986;Jung et al. 1998;Liebenberg and Pretorius 2004;Souza et al. 2011).
The Ur-3 gene present in the Middle American white-seeded common bean, Aurora, was reported by Ballantyne (1978). Since then, this gene has been used extensively as the source of rust resistance in a large number of dry bean cultivars from various market classes of the United States, as well as in fresh market and processing snap beans Stavely et al. 1997;Pastor-Corrales et al. 2007;Urrea et al. 2009;Osorno et al. 2010;Brick et al. 2011;Beaver et al. 2015). Ur-3 has also been used as a source of rust resistance in dry bean cultivars of South Africa (Liebenberg et al. 2005). In addition, Ur-3 has been the subject of different studies, including genetics (Grafton et al. 1985;Kalavacharla et al. 2000), molecular markers, and gene tagging (Haley et al. 1994). The Ur-3 is also present in Middle American cultivars Mexico 235, Ecuador 299, NEP 2, and 51052, in addition to other undefined rust resistance genes Miklas et al. 2000;Hurtado-Gonzales et al. 2016).
The Ur-3 gene confers resistance to 55 of 94 races of the bean rust pathogen maintained at Beltsville, MD ). More importantly, Ur-3 confers resistance to many races that overcome the resistance of all other named rust resistance genes in common bean. For example, the Ur-3 gene confers resistance to race 22-52 (previously known as race 108), the only race known to overcome the broad-spectrum resistance of the Ur-11 gene present in PI 181996 and PI 190078, and of the Ur-14 gene present in Ouro Negro (Stavely 1998;Alzate-Marin et al. 2004). The name of race 108 and of six other races (41, 47, 49, 53, 67, and 84) used in this study, was changed after these races were phenotyped on a new set of bean rust differential cultivars adopted for the characterization of races of U. appendiculatus and a binary system to name these races (Steadman et al. 2002;Pastor-Corrales and Aime 2004). The new and old names (in parentheses) of the races used in this study are: 15-1 (41), 15-3 (47), 22-6 (49), 31-1 (53), 31-22 (67), 37-1 (84), and 22-52 (108).
The Ur-3 gene also complements the broad-spectrum rust resistance in accessions PI 151385, PI 151388, PI 151395, and PI 151396, which are also only susceptible to race 22-52. Similarly, Ur-3 confers resistance to race 37-1, the only known race that overcomes the rust resistance in PI 260418 (Pastor-Corrales 2005). In addition, Ur-3 confers resistance to many races that overcome the Ur-4, Ur-5, Ur-6, Ur-7, Ur-9, Ur-12, and Ur-13 genes. Although Ur-3 is not resistant to all races of Mesoamerican origin, this gene confers resistance to most races of U. appendiculatus of Andean origin; that is, races isolated from common beans of the Andean gene pool. Thus, Ur-3 is a critical component of gene pyramiding of common bean cultivars with broad resistance to rust. The information above provides strong evidence of the historical importance and current relevance of Ur-3 for breeding dry and snap beans with broad and durable resistance to rust in the United States and other nations (Stavely 2000;Pastor-Corrales et al. 2001).
The resistant reaction of Ur-3 gene to U. appendiculatus is initially characterized by the production of small water-soaked chlorotic spots that subsequently become, in 48 hr, well-defined necrotic spots without sporulation. This resistant phenotype is classified as grade 2, 2+ and it is known as the hypersensitive reaction (HR) in the bean rust grading scale Stavely 1998).
The Ur-3 gene has been mapped on chromosome Pv11 of the common bean genome (Stavely 1998;Miklas et al. 2002). Inheritance of resistance and phenotypic data revealed that the Ur-3 gene was very closely linked to Ur-11 on the terminal position of chromosome Pv11 (Kelly et al. 1996). The close proximity between these two genes led to the naming of the rust resistance gene in PI 181996 as Ur-3 2 (Kelly et al. 1996). However, later reports demonstrated the independence of Ur-3 and Ur-3 2 and revealed that these two genes were linked in repulsion and different from each other (Stavely 1998). Thus, Ur-3 2 was renamed Ur-11 (Stavely 1998). The close proximity of Ur-3-Ur-11 may be one of the main reasons why it has been difficult to find DNA markers that are specific for the Ur-3 gene. There are other named rust resistance genes (Ur-6 and Ur-7) on Pv11, as well as two other unnamed genes (Ur-Dorado 53 and Ur-BAC 6), although these genes are not as tightly linked to Ur-3 as Ur-11 (Miklas et al. 2002;Kelly et al. 2003).
Four specific races of the bean rust pathogen have been reported as phenotypic markers that effectively identify rust resistance genes; race 31-1 identifies Ur-3, race 22-6 recognizes Ur-4, race 15-3 identifies Ur-6, and race 31-22 recognizes Ur-11 (Stavely 2000;Pastor-Corrales and Stavely 2002). These races identify the presence of these genes alone or in combination with other rust resistance genes. However, the phenotypic identification of these rust resistance genes is laborious, time consuming, and currently only performed at the Bean Project at Beltsville, MD. Moreover, the detection of multiple rust resistance genes in common bean using phenotypic markers is also often complicated by the presence of epistasis between rust resistance genes (Miklas et al. 1993;Pastor-Corrales and Stavely 2002). Furthermore, the current molecular markers (mostly RAPD and SCAR markers) linked to rust resistance genes in common bean that were published almost two decades ago, yield false positive and false negative results, as indicated by the authors that reported the currently available RAPD (OK14 620 ) and SCAR (SK14) markers linked to the Ur-3 locus (Haley et al. 1994;Nemchinova and Stavely 1998).
Several factors contributed to the false positive and false negative results when using the current molecular markers. Among these factors is the weak linkage of some molecular markers with the gene of interest. For instance, the RAPD marker (OK14 620 ) tagging Ur-3, was reported to be positioned 2.23 cM from this gene (Haley et al. 1994). Another constraint was the close proximity of rust resistance genes, as is the case with the Ur-3 and Ur-11 genes. Additionally, the lack of a reference n Table 1 Reaction of the common bean cultivars used in this study to races 15-1 (41), 31-1 (53), 37-1 (84), and 22-52 (108) of Uromyces appendiculatus, the causal agent of the bean rust disease
genome for common bean hindered the development of highly specific, tightly linked DNA markers. The publication of the common bean reference genome in 2014 (Schmutz et al. 2014), along with the development of high-throughput genotyping technologies for common bean, are making possible the identification of more effective molecular markers. Although the Ur-3 is a very important rust resistance gene in common bean, to date there is not a reliable molecular marker tagging Ur-3. Thus, to improve the durability of common bean cultivars to the highly variable bean rust pathogen, Ur-3 cannot be combined with other rust resistance genes using marker-assisted selection. As indicated earlier, at present, pyramiding Ur-3 with other rust resistance genes is only feasible using specific races of the rust pathogen, an activity that is reliable but highly laborious and time consuming. The objective of this study was to develop highly specific, tightly linked, effective molecular markers for the detection of the historically important and widely used Ur-3 rust resistance gene, either alone or in combination with other rust resistance genes of common bean.

MATERIALS AND METHODS
Population development and phenotypic evaluation of the bean rust disease A total of 129 F 2 plants were derived from the cross Pinto 114 · Aurora. Both are dry beans of the Middle American pool of common bean, where Pinto 114 was the susceptible parent and Aurora was the resistant parent containing the Ur-3 gene. The following cultivars with known rust resistance genes were included in the inoculation as internal controls of successful rust inoculation: Early Gallatin (Ur-4), Golden Gate Wax (Ur-6), and PI 181996 (Ur-11) ( Table 1). All F 2 plants, parents, and control cultivars were grown in 12.7-cm diameter pots containing two plants per pot. The primary (unifoliate) leaves of bean plants were inoculated 7 d after seeding, when the primary leaves were 35-65% expanded (Stavely 1984). To prepare the rust inocula, suspensions of frozen urediniospores of various races of U. appendiculatus were placed in a 25-ml solution of cold tap water and 0.01% Tween 20 in a 250-ml Erlenmeyer flask. The spore solutions were prepared with a concentration of 2 · 10 4 urediniospores per ml 21 . All 129 F 2 plants and the control cultivars were inoculated with races 15-1, 31-1, 37-1, and 22-52 of U. appendiculatus. Races 15-1, 31-1, 37-1, and 22-52 elicited the same resistance (HR) reaction on plants with Ur-3, as shown in Supplemental Material, Table S1. However, these races elicited a different type of reaction on PI 181996 (the control cultivar with Ur-11) and on cultivars with other rust resistance genes. Thus, one important reason for using four races to phenotype the F 2 population was to unequivocally ensure the phenotype of each F 2 plant, the parents, and of the control cultivars, which included plants with Ur-4, Ur-6, Ur-11, and other rust resistance genes. The F 2 plants were inoculated using a cotton swab to apply the spore solution of each of the races on the abaxial side of the primary leaves. After inoculation, the plants were transferred to a mist chamber (20 6 1°and relative humidity .95%) for 18 hr, under darkness. After this period, the plants were transferred to the greenhouse. Visible rust symptoms were observed on susceptible plants 10-12 d after inoculation (dai).
The F 2 population and parents were evaluated for their rust phenotype 12-14 dai using a 1-6 scale (Stavely and Pastor-Corrales 1989), scored as follows: 1 = no visible rust symptoms; 2 = necrotic or chlorotic spots without sporulation, ,0.3 mm in diameter (HR); n 2+ = necrotic spots without sporulation, 0.3-1.0 mm in diameter; 2++ = necrotic spots without sporulation, 1.0-3.0 mm in diameter; 2+++ = necrotic spots .3.0 mm in diameter; 3 = uredinia ,0.3 mm in diameter (tiny sporulating pustules); 4 = uredinia 0.3-0.5 mm in diameter (large sporulating pustules); 5 = uredinia 0.5-0.8 mm in diameter (large sporulating pustules); and 6 = uredinia .0.8 mm in diameter (very large sporulating pustules). Plants with grades 2 and 3 were classified as resistant, whereas those with rust grades of 4, 5, or 6 were classified as susceptible. Thereafter, the F 2 plants were maintained in the greenhouse to produce F 2:3 families by self-fertilization. A total of 281 F 3 plants from 12 selected F 2:3 families were inoculated with race 31-1 of U. appendiculatus. These families were inoculated using an Air Brush-Depot compressor, model TC-20, and an Iwata Airbrush, Revolution BCR, by applying the spore solution (concentration of 2 · 10 4 per ml 21 ) of race 31-1 on the abaxial side of the leaves. After spraying, plants were treated similarly to the F 2 plants, as described above. The reaction (rust phenotype) of the differential bean cultivars to all races of U. appendiculatus used in this study is presented in Table S1.
Bulk segregant analysis and single nucleotide polymorphism assay Newly emerged trifoliate leaves from each of the F 2 plants were collected and total genomic DNA was isolated using DNeasy 96 Plant Kit (Qiagen, Valencia, CA) according to manufacturer's instructions. Based on the rust reaction of each of the F 2 plants, three susceptible (rr) bulks were prepared. Each bulk consisted of DNA from eight different F 2 susceptible plants. Bulks of resistant F 2 plants were not prepared to avoid the inclusion of heterozygous-resistant (Rr) plants. These bulks were used for bulk segregant analysis (BSA) for identification of markers potentially associated with the Ur-3 gene (Michelmore et al. 1991). The DNA from susceptible bulks and two samples from each parent were analyzed with 5398 single nucleotide polymorphism (SNP) markers on the Illumina BeadChip BARCBEAN6K_3, following the Infinium HD Assay Ultra Protocol (Illumina, Inc., San Diego, CA). The results obtained on the BeadChip were visualized by fluorescence intensity using the Illumina BeadArray Reader and alleles were called using Illumina GenomeStudio V2011.1 (Illumina, Inc.). Allele calls were visually inspected and errors in allele calling were corrected manually. SNPs were considered to be associated with the Ur-3 locus when they were polymorphic between the Pinto 114 (susceptible) and Aurora (resistant) parents and the three susceptible bulks were homozygous and clustered tightly with the susceptible parent (Pinto 114).
Developing and evaluating simple sequence repeat markers linked to Ur-3 The sequence fragments containing SNPs associated with the Ur-3 locus were aligned to the common bean reference genome using Standalone Megablast (Morgulis et al. 2008) to identify the scaffolds in the reference genome. Scaffolds were downloaded from Phytozome (https:// phytozome.jgi.doe.gov/pz/portal.html), DOE, JGI (Department of Energy, Joint Genome Institute). The scaffolds were screened for the presence of simple sequence repeat (SSR) markers. Procedures for SSR identification, SSR screening, and primer design were previously described by Song et al. (2010).
The polymorphism and quality of the SSR markers were first tested using DNA from the Pinto 114 (susceptible) and Aurora (resistant) parents. Polymorphic SSR markers between Pinto and Aurora were then used to analyze the DNA of the F 2 population from the Pinto 114 · Aurora cross. Polymerase chain reaction (PCR) was performed with 5 ng of genomic DNA, 0.25 mM of forward and reverse primers, 1· PCR buffer [200 mM Tris-HCl (pH 8.0), 500 mM KCl, 2 mM each dNTP, 10% glycerol, 15 mM MgCl 2 , 20 ng/ml of single-stranded binding protein], and 0.1 unit of Taq DNA polymerase. The PCR thermocycling parameters were 3 min at 92°and 38 cycles of 50 sec at 90°, 45 sec at 58°, and 45 sec at 72°, followed by a 5 min extension at 72°and hold at 10°. PCR products were resolved on 2-3% agarose gels (Agarose SFR; Amresco, Dallas, TX) prepared with TBE 1· buffer (Tris-borate-EDTA) and stained with 1 mg/ml 21 ethidium bromide.
Developing and testing Kompetitive Allele Specific PCR markers A subset of SNPs positively associated with Ur-3 found using BSA were selected for genotyping the F 2 population from Pinto 114 · Aurora using Kompetitive Allele Specific PCR (KASP) assays. Additional SNPs used for KASP genotyping were retrieved from SNP chip tables found in Song et al. (2015). KASP primers were designed using the Primer-Express software and KASP reactions were conducted following the manufacturer's instructions. The 10 ml reaction contained 5 ml of 2· premade KASP master mix (LGC, Middlesex, UK), 0.14 ml of primers mix (Sigma-Aldrich, St. Louis, MO), and 20-40 ng of genomic DNA. PCR parameters were as described by LGC, on standard thermocycling n Fine mapping of the Ur-3 locus in F 3 plants using KASP markers F 3 families were selected based on the recombination between Ur-3 and the SSRs and KASPs molecular markers found in the F 2 population. A total of 10 F 3 families were selected for screening with KASP markers n Table 4 Physical position and primer sequences of KASP markers associated with Ur-3 rust resistance gene in common bean SS4 and SS6 flanking the Ur-3 locus. One homozygous-resistant family and one susceptible family were evaluated as internal controls. The number of plants per family varied from 22 to 32, according to the availability of seeds. A total of 281 F 3 plants were inoculated with race 31-1 of U. appendiculatus, as described in Materials and Methods. DNA from the F 3 plants was isolated according to Lamour and Finley (2006) and were genotyped with KASP markers SS4 and SS6. F 3 plants showing recombination between markers SS4 and SS6 were selected for additional genotyping with newly designed KASP markers in order to narrow the genomic region containing the Ur-3 locus.  Song et al. (2015) and used for the haplotype analysis. These lines were also inoculated with races 22-6, 31-1, 31-22, and 22-52 of U. appendiculatus. The four races were used to identify the presence of certain rust resistance genes in these cultivars; races 31-1 and 22-52 to identify the presence of Ur-3, race 31-22 to identify Ur-11, and race 22-6 to identify Ur-4.
Cultivars with HR to races 31-1 and 22-52 had the Ur-3 gene (Table S1). The sequence variants in the targeted genomic region of the 18 varieties and their phenotypes were used to identify haplotypes associated with resistance and susceptibility to race 31-1. All SNPs identified between KASP markers SS4 and SS6 were handled using Microsoft Excel and haplotypes were identified by visual inspection. At least two KASP markers were designed for each of the observed haplotypes. Whenever feasible, SNP markers were located every 10 kb across the 470,487 bp genomic region. When KASP markers were polymorphic between the Pinto 114 (ur-3) and Aurora (Ur-3) parents, they were used to genotype F 3 plants with recombination between the markers SS4 and SS6.
Validation of the markers linked to the Ur-3 locus A panel of 130 diverse bean cultivars that included all rust resistance genes in common bean were genotyped using KASP markers tightly linked with Ur-3. This was performed with the purpose of generating accurate Ur-3 markers useful in marker-assisted selection. In this panel, some cultivars had the Ur-3 gene alone, other cultivars had Ur-3 combined with other rust resistance genes, while others did not have any reported rust resistance genes. The cultivars in the panel were phenotyped before or during this study with multiple races of the bean rust pathogen, including race 31-1, the phenotypic marker for the Ur-3 gene.

Data availability
All data described in this manuscript related to bean rust phenotypes, Pinto 114 · Aurora F 2 genetic map, F 3 fine-mapping population, and haplotype analysis are available in Table S1, Table S2, Table S3, Table  S4, Table S5, Table S6, and Table S7.

Inheritance of rust resistance in common bean Aurora
A total of 129 F 2 plants from the Pinto 114 · Aurora cross were evaluated for their reaction to races 15-1, 31-1, 37-1, and 22-52 of U. appendiculatus (Table S2). Aurora was resistant to all four races and exhibited the same type of reaction that was characterized by necrotic spots without sporulation (grades 2, 2+). Pinto 114 was susceptible to the same four races, with a reaction characterized by large uredinia (grades 4, 5, and 6). Based on the reaction to all four races, the inheritance of rust resistance study of the 129 F 2 plants exhibited a segregation equal to 101 resistant and 28 susceptible plants, fitting a ratio of 3 resistant to 1 susceptible (x 2 = 0.747, P value = 0.38), confirming that the rust resistance in Aurora was conferred by the single and dominant Ur-3 gene (Table S2).

BSA and SNP genotyping using BARCBEAN6K_3 BeadChip
Based on the BSA, 28 SNPs were associated with Ur-3 ( Table 2). The alleles of these SNPs could separate the susceptible Pinto 114 and the three susceptible bulks from the resistant Aurora parent. According to the genetic linkage map created by Song et al. (2015), these 28 SNPs were distributed from 72.3 to 76.2 cM on the lower end of the common bean chromosome Pv11. The physical location of the associated 28 SNPs was between 46,437,627 bp (ss715647455) and 48,784,158 bp (ss715641910), a region spanning a total of 2.1 Mbp ( Table 2).

Mapping of the Ur-3 gene
The large portion of the genomic region containing the 28 SNPs associated with the Ur-3 rust resistance gene was targeted for SSR development. A total of 48 SSR markers located between 46,266,888 and 48,664,905 bp on Pv11 were developed. Thirteen of the 48 SSRs markers were polymorphic between the parents Pinto 114 (susceptible) and Aurora (resistant) parents (Table 3). These markers, which showed unequivocal allele separation in agarose gel, were used to map the Ur-3 locus in the F 2 population Pinto 114 · Aurora. Linkage analysis positioned the Ur-3 locus between markers BARCPVSSR14001 (46,535,562 bp) and BARCPVSSR14082 (47,291,606 bp), a 756,044 bp genomic region (data not shown). In addition, four positively associated SNPs from the BSA and two SNPs [retrieved from Song et al. (2015)] near the SSRs flanking the Ur-3 locus were selected and converted into KASP markers (Table 4). Five KASP markers (SS1, SS3, SS4, SS5, and SS6) showed clear separation of the three clusters (two homozygous and one heterozygous) and were polymorphic between the Pinto 114 and Aurora parents. These KASP markers were used to refine the Ur-3 gene map. Linkage analysis in the F 2 population genotyped with 13 SSRs and the five KASP markers showed that Ur-3 was flanked by KASP marker SS5 and SSR marker BARCPVSSR14007 between 46,667,862 and 46,865,194 bp, respectively, on chromosome Pv11 ( Figure 1A and Table S3). The distance of the Ur-3 locus to these flanking markers was 0.2 cM ( Figure 1A).
Analysis of recombination in F 3 and Ur-3 haplotype identification KASP markers SS4 and SS6 were mapped at 0.6 and 1.0 cM from the Ur-3 locus, respectively ( Figure 1A), in a 470,487 bp (470 kb) genomic region of chromosome Pv11, from 46,613,419 to 47,083,906 bp ( Figure 1B).
n Table 5 Major haplotypes identified between KASP markers SS4 and SS6 using SNP calls from 18 sequenced common bean varieties (Song et al. 2015) and G19833, the common bean reference genome landrace. These markers were chosen to genotype 12 selected F 3 families from the cross Pinto 114 · Aurora. Among the 12 families, four were derived from recombinant F 2 plants between KASP markers SS4 and SS6, six families were heterozygous between markers SS4 and SS6 flanking Ur-3, and two families were used as internal controls: one homozygous resistant and the other homozygous susceptible. In addition, these 12 families (281 F 3 plants) were inoculated with race 31-1 of U. appendiculatus. Genotyping the 281 F 3 plants resulted in 87 F 3 plants with recombination events between the SS4 and SS6 KASP markers (Table  S4). These 87 F 3 plants were selected for subsequent fine-mapping analysis with additional KASP markers (Table 4). SS5 (ss715647451 at position 46,667,862) was the only KASP marker derived from the BeanChip that was located between SS4 and SS6; thus, SS5 was also used to genotype the recombinant 87 F 3 plants.
We then mined the SNP sequence data from the 18 common bean varieties (Song et al. 2015) to search for additional SNPs between SS4 and SS6. Based on the whole genome sequence of the 18 common bean varieties, 6000 SNPs and small indels were found between SS4 and SS6 (Table S5). These SNPs were grouped into 10 major haplotypes ( Table 5). Each of these haplotypes were then tagged with one or two KASP markers and were examined for their polymorphism between Pinto 114 (ur-3), Aurora (Ur-3), Mexico 235 (Ur-3+), and PI 181996 (Ur-11). The KASP markers polymorphic between the Pinto 114 and Aurora parents were tested on the set of 87 F 3 recombinant plants identified previously with KASP markers SS4 and SS6. Analysis of the 87 F 3 recombinant plants positioned the Ur-3 gene between KASP markers SS17 and SS21, in the 83,198 bp genomic region ( Figure  1B and Table S7). Concurrently, a specific haplotype for Ur-3 was identified based on the reaction of the 18 sequenced varieties to race 31-1 of U. appendiculatus. Only the varieties C 20, Matterhorn, Stampede, T-39, and Sierra had a resistant phenotype (HR) to races 31-1 and 22-52, indicating that these cultivars have the Ur-3 gene (Table S7). The final genotyping analysis on the 87 recombinant plants mapped Ur-3 between KASP markers SS36 and SS21, in a specific genomic region of 46,563 bp, ranging from 46,967,787 to 47,014,350 bp of Pv11 (Table 6). Two F 3 plants, one resistant and the other susceptible, had the same recombination breakpoint, demonstrating that the Ur-3 gene was located in the region flanked by SS36 and SS21 ( Figure 1C and Table 6).
Subsequent genotyping of the 129 F 2 plants from the Pinto 114 · Aurora cross using KASP SS36 and KASP marker SS68, which was targeting the Ur-3 haplotype and only 200 bp downstream from SS36, showed that these markers were linked to the Ur-3 rust resistance gene, with no recombination observed between bean rust phenotype and genotype (Table S2). SNP for KASP marker SS68 (46,967,980 bp in Pv11) is a transversion nucleotide change from A to T, where A is susceptible and T is resistant. KASP marker SS68 effectively differentiated homozygous-resistant, homozygous-susceptible, and heterozygous plants (Figure 2). Conversely, the KASP marker SS36 did not always differentiate homozygous-resistant from heterozygous plants (data not shown). KASP marker SS68 is located proximal (500 bp) to the leucine-rich repeat-containing gene, Phvul.011G193100.

Validation of KASP marker SS68 linked to the Ur-3 gene
We used the SS68 KASP marker to genotype a panel of 130 common bean cultivars that included dry and snap beans. Some of these common beans possessed the Ur-3 gene alone, while others had Ur-3 in combination with other rust resistance genes. In addition, other cultivars had single or combinations of the other 10 rust resistance genes in common bean. The results of this validation showed that SS68 was highly accurate for the identification of the Ur-3 locus (Table 7). No false positives or false negatives were observed when comparing the genotypic n Table 6 Genotypes (AA, BB) at nine SNP loci (from SS4 to SS6), and the reaction to race 31-1 of Uromyces appendiculatus (evaluation with SS68 marker) and phenotypic (reaction to race 31-1) evaluations of these cultivars.

Development of accurate SNP markers linked to the Ur-3 locus
The historically important Ur-3 gene confers resistance to the pathogen that causes the rust disease of common bean. The effective incorporation of Ur-3 into dry and snap beans using molecular markers has been limited by the inaccuracy of the molecular markers linked to this gene (Haley et al. 1994;Nemchinova and Stavely 1998;Stavely 2000). The authors that reported the RAPD (OK14 620 ) and SCAR (SK14) markers linked to Ur-3 indicated that these markers produced both false negatives and false positives results (Haley et al. 1994;Nemchinova and Stavely 1998).
More recently, we have used BSA, SNP assay, and whole genome sequencing to discover SSR markers closely linked to the Ur-3 and other disease resistance genes. However, even the use of closely linked BARCPVSSR14007, an SSR marker reported in this study positioned at 0.2 cM from the Ur-3 locus, resulted in .3% false positive results when this marker was used on the panel of 130 common bean lines (data not shown). Additionally, as indicated earlier, the inability to find specific molecular markers linked to Ur-3 may have been exacerbated by the presence of the Ur-11 rust resistance gene that is closely linked to Ur-3 on the terminal position of chromosome Pv11. Currently, the most reliable method to monitor for the presence of the Ur-3 gene in dry and snap bean cultivars continues to be race 31-1 (53) of U. appendiculatus. Race 31-1 is used as a phenotypic marker that effectively identifies common bean plants with Ur-3 alone or in combination (Pastor-Corrales 2002). However, phenotypic evaluations under greenhouse conditions are very laborious and time consuming (21 d). Moreover, due mostly but not only to the biotrophic condition of the rust pathogen, most breeders of dry and snap beans do not have the option of using this methodology.   Given the importance of Ur-3, we determined to search for highly accurate molecular markers linked to Ur-3 using a fine-mapping approach. We employed a variety of technologies that included phenotyping with specific races of the bean trust pathogen, BSA coupled with high-throughput SNP genotyping using the BARCBEAN6K_3 BeadChip, SSR and KASP marker development, and local association analysis using SNPs from previous whole genome shotgun sequencing efforts. In summary, the combination of these technologies permitted the identification of KASP marker SS68, which was highly accurate in identifying the presence of Ur-3 in a panel of 130 common bean cultivars that included dry and snap beans with and without the Ur-3 gene. Marker SS68 was also tested on a mapping population of 184 F 2 genotypes from the cross between Pinto 114 · Mexico 235 (Ur-3+). No recombination was observed between phenotype and the genotype in this study (data not shown). These results confirm the accuracy and utility of the KASP marker SS68 even when this marker is used on mapping populations in which the origin of the Ur-3 gene is not the cultivar Aurora.
Survey of the SS68 KASP marker in a common bean diversity panel In this study, we determined the potential utility of the KASP SNP marker SS68 in a panel of common bean cultivars carrying different rust resistance genes and in bean lines representing the major common bean market classes in the United States. Marker SS68 reliably identified cultivars containing Ur-3, independent of the gene pool (Andean or Middle American), type of common bean (dry or snap), or market class of dry edible beans (pinto, great northern, navy, red kidney, black, and others). Additionally, SS68 effectively distinguished common bean lines carrying Ur-3 alone as well as lines combining the Ur-3 and Ur-11 genes that are closely linked on Pv11 (Table 7). Because Ur-3 gene is epistatic to Ur-11, it is difficult to combine these two genes using inoculations with races of the rust pathogen (Stavely 2000). Thus, using marker SS68 to identify Ur-3 when combined with Ur-11 avoids this problem.
The Ur-3 locus maps to a 46 kb region possessing candidate genes with resistant gene motifs Through haplotype analysis and KASP marker development, it was possible to determine a genomic region of 46,563 bp containing the Ur-3 locus and delimited by markers SS36 and SS21 on Pv11. Six candidate genes were identified within this 46.5 kb region in the P. vulgaris reference genome, obtained by sequencing the landrace G 19833 of Andean origin. Among the six candidate genes, there were three genes with NB-ARC LRR domains. Proteins containing NB-ABC LRR domains are known to be involved in plant resistance and activation of innate immune responses to various types of pathogens (Hammond-Kosack and Jones 1997;Jones and Dangl 2006). Similarly, protein kinases (also found in the 46.5 kb region) are known to play a central role in signaling during pathogen recognition and the subsequent activation of plant defense mechanisms (Xue et al. 2015). The genomic region containing the Co-4 gene on chromosome Pv08, conferring resistance to Colletotrichum lindemuthianum in common bean, has been characterized and known to contain an open reading frame coding for a serine threonine kinase (Oblessuc et al. 2015), a type of protein which has also been identified in our studies. Additionally, serine threonine protein kinase constitutes candidate genes for controlling angular leaf spot resistance in the Andean landrace G 5686 (Keller et al. 2015). Whether the phenotype of the Ur-3 locus is the result of the expression of one or more of the six candidate genes will be a matter of further investigation.
Sequence analysis of the Andean landrace G 19833, used to sequence the reference genome of common bean, revealed that the 46.5 kb genomic region containing the Ur-3 locus is highly duplicated ( Figure  S1), and it includes repetitive elements in the intergenic spaces. Additionally, this genomic region is AT-rich (33% vs. 16% for GC), which suggests that it is highly unstable. Sequence analysis comparing the Middle American Aurora common bean and the Andean landrace G 19833, will provide valuable insights into the structural differences and evolutionary history of the important Ur-3 rust resistance locus.

Conclusions
This study used a new approach to generate KASP SS68, the first highly accurate DNA marker linked to the Ur-3 rust resistance gene in common bean. We fine-mapped a 46.5 kb genomic region in chromosome Pv11, present in Middle American common bean cultivar Aurora. This was accomplished using the BARCBEAN6K_3 BeadChip, SSRs, KASP technology, and local association. The validation of this newly discovered KASP SS68 marker on a panel of 130 common bean lines revealed that SS68 was highly accurate in identifying Ur-3. This marker will be of value for combining Ur-3 with other Andean and Middle American genes with broad spectrum resistance to the highly variable bean rust pathogen. In addition, the utilization of the new marker SS68 will significantly reduce the time and labor associated with the transfer of the Ur-3 gene using inoculations of bean plants with specific races of the rust pathogen. The genomic region containing the Ur-3 locus included six genes annotated in the reference genome of P. vulgaris. These genes are likely candidates for the Ur-3 rust resistance gene. Gene expression analysis of these candidate genes and functional approaches will enhance our understanding of the mechanisms underlying the reaction of P. vulgaris to U. appendiculatus.

ACKNOWLEDGMENTS
The authors thank Rob Parry and Chris Pooley for their assistance with sequence analysis by installing computer hardware and software. This work was supported, in part, by funding from the Norman Borlaug Commemorative Research Initiative of the United States n Agency for International Development, project no. 0210-22310-004-96R, and by the United States Department of Agriculture, Agricultural Research Services, project no. 8042-22000-286-00D (M.A.P.-C.). This research was also financially supported by the National Council for Scientific and Technological Development and the Coordination for the Improvement of Higher Education Personnel . The contents of this publication do not necessarily reflect the views or policies of the United States Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the United States government.