Phenotypic Evaluation of Soybean Genotypes for Their Reaction to a Mississippi Isolate of Phakopsora pachyrhizi Causing Soybean Rust

Soybean rust (SBR) caused by Phakopsora pachyrhizi Syd. and P. Syd. is one of the most important foliar diseases of soybean. SBR has the potential to cause major economic damage to global and U.S. soybean production. Analysis of reactions of soybean genotypes to P. pachyrhizi is an important step towards breeding for resistance to SBR. Fifty-four diverse soybean genotypes with both known and unknown Rpp resistance genes were tested for their reactions to a Mississippi P. pachyrhizi isolate. PI 567102B (Rpp6) had a near-immune reaction with the lowest disease severity score and no sporulation. Among seventeen genotypes with resistant or incomplete resistant reddish-brown (RB) reactions, eight are improved breeding lines that are available to researchers through material transfer agreements (MTAs). Thirty-six genotypes had the susceptible TAN reaction. Four soybean lines (RN06-32-1(7-b, GC 00138-29, G01-PR16, and GC 84051-9-1) had RB reactions and significantly lower SBR severity and sporulation than three of the six resistant checks, PI 230970 (Rpp2), PI 462312 (Rpp3), and PI 459025B (Rpp4). G01-PR16 is a publicly released germplasm. This research provides new information about reactions of different soybean genotypes to a midsouthern USA isolate of P. pachyrhizi and thereby aids in breeding for resistance to SBR.


Introduction
Soybean rust (SBR) is one of the most economically important foliar diseases of soybean (Glycine max (L.) Merr.), occurring in many major soybean-producing countries. Significant soybean yield losses of up to 80% to 90% due to soybean rust have been reported in many countries, including Brazil and the U.S. [1,2]. These losses result primarily from a reduction in pods, seeds per pod, and seed weight [1] due to a decrease in photosynthesis in infected leaves and premature defoliation [3]. The soy market could be directly impacted by SBR. The reduction of grain due to SBR could lead to a decrease in the production of soy oil, protein, and related derivatives [4].
Soybean rust is caused by the obligate biotrophic fungal pathogen Phakopsora pachyrhizi Syd. and P. Syd. [5]. The pathogen was first described in Japan in 1902 by Hennings who named the pathogen Uredo sojae Henn. [6]. Later, Hans and Paul Sydow gave the fungus its current name, Phakopsora pachyrhizi Syd. and P. Syd, based on an isolate obtained from the leguminous host plant Pachyrhizus erosus (L.) Urb. (=Pachyrhizus angulatus) in Taiwan [7,8]. Since then, P. pachyrhizi has spread and can be found in many soybean-growing regions around the world. The first discovery of soybean rust in the United States was in Hawaii in 1994 [9]. The disease was first detected in the continental United States in a field near Baton Rouge, Louisiana on 6 November 2004 [10].
In general, soybean has three major types of reaction in response to P. pachyrhizi infection: (1) resistant immune (IM) reaction with no macroscopically visible lesions; (2) resistant or incomplete resistance with reddish-brown (RB) lesions; and (3) a susceptible reaction with tan (TAN) lesions [11,12]. The TAN lesion type is usually associated with high levels of sporulation of the pathogen, but two types of TAN lesions have been described: "few" or "many" uredinia per lesion [5]. The reddish-brown colored lesion type can be sporulating or non-sporulating, whereas the immune reaction is defined as a lack of visible lesions on leaves after being challenged with P. pachyrhizi [5,13]. The IM and RB reactions, with low sporulating or non-sporulating lesions, have been considered resistant reactions to the SBR causal pathogen.
Seven resistance genes (Rpp 1-7) to P. pachyrhizi [14][15][16][17][18][19][20][21][22] have been identified, but each of these Rpp genes conditions resistance to only a limited number of P. pachyrhizi isolates [23][24][25]. Identification of genotypic sources of resistance is the first step towards developing cultivars with resistance to soybean rust. Moreover, as no one Rpp gene has resistance to all isolates, there is a continuing need to identify new sources of soybean rust resistance.
Soybean [Glycine max (L.) Merr.] is one of the most important legume crops in the world. It is a major crop in Mississippi where it provides income to local U.S. economies and is an important source of high-quality plant protein and oil to the U.S. and the world [26,27]. Moreover, Mississippi was one of the first states in the USA where soybean rust was detected on soybean in Adams County on 16 November 2004 [28]. We hypothesized that a diverse group of soybean genotypes with both known and unknown Rpp resistance genes, including multiple improved breeding lines derived from the diverse sources, would have differential reactions to a domestic isolate of P. pachyrhizi from Mississippi. Soybean genotypes identified to have resistance to this southern P. pachyrhizi isolate could be utilized for the development of Rpp gene pyramids or crossed with other resistance resources to obtain more durable resistance to SBR and thereby aid in breeding for resistance to SBR.
Hence, the objectives of this research were to analyze the phenotypic reactions of soybean genotypes to SBR and determine their suitability for use in breeding programs. Rust lesion type, disease severity, and P. pachyrhizi sporulation for the genotypes were compared and analyzed. This research provides new information for the selection of breeding materials to aid in breeding resistance to SBR.

Results
Results of the analysis of variance (not shown) indicated that there werea significant (p ≤ 0.05) differences among soybean genotypes that were associated with the different sources of genotypic resistance.

Rust Severity
Statistically, there were significant (p ≤ 0.05) differences in soybean rust severity among the soybean genotypes tested. The mean soybean rust severity of all genotypes tested was 3.5 with a range of 1.2 to 4.7 based on a lesion density scale of "1", which was no visible lesions, to "5", which was the highest score of disease severity that featured prolific lesion development covering most of the leaf [29,30]. PI 567102B (Rpp6) had the lowest rust severity score of 1.2. Among 17 genotypes with RB reactions, 11 had severity scores of less than 3, where PI 200492 (Rpp1a) and PI 200487 (Rpp5) had low rust severity scores of 1.3 and 1.9, respectively. In contrast, Williams 82, NC-Roy, and PI 587905 (Rpp1e) had high severity scores of 4.6, 4.6, and 4.7, respectively ( Table 2). Rust severity analysis was also performed based on lesion type (RB and TAN). Differences in the severity scores among soybean genotypes producing an RB reaction type were significant (p ≤ 0.05). Among them, four lines with RB reactions (RN06-32-1(7-b, GC00138-29, GC84051-9-1, and GC84051-9-1) had significantly lower SBR severity than three of the six resistant checks, PI 230970 (Rpp2), PI 462312 (Rpp3), and PI 459025B (Rpp4). The severity scores for the RB group ranged from 1.3 to 3.6 with a mean of 2.8 ( Table 2). Among genotypes with the TAN lesion type, there were also significant (p ≤ 0.05) differences for soybean rust severity. The severity scores for the TAN group ranged from 3.1 to 4.7 with a mean of 4.0 (Table 2), which was higher than the mean severity of the RB group (2.8). The lower end of the TAN range for severity (3.1) overlaps with the upper end of the severity range (3.6) for the RB group.

Rust Severity
Statistically, there were significant (p ≤ 0.05) differences in soybean rust severity among the soybean genotypes tested. The mean soybean rust severity of all genotypes tested was 3.5 with a range of 1.2 to 4.7 based on a lesion density scale of "1", which was no visible lesions, to "5", which was the highest score of disease severity that featured prolific lesion development covering most of the leaf [29,30]. PI 567102B (Rpp6) had the lowest rust severity score of 1.2. Among 17 genotypes with RB reactions, 11 had severity scores of less than 3, where PI 200492 (Rpp1a) and PI 200487 (Rpp5) had low rust severity scores of 1.3 and 1.9, respectively. In contrast, Williams 82, NC-Roy, and PI 587905 (Rpp1e)

Correlation Analysis between Soybean Rust Severity and Sporulation
Results indicated that soybean rust severity over all genotypes tested was significantly correlated with soybean rust sporulation (r = 0.8049, p ≤ 0.0001), (Figure 2). For soybean genotypes with RB lesion types, soybean rust severity was positively correlated with soybean rust sporulation (r = 0.653, p ≤ 0.0001). For soybean genotypes producing TAN lesions, soybean rust severity was also significantly (r = 0.3656, p ≤ 0.0001) correlated with soybean rust sporulation.

Discussion
Soybean rust is one of the most economically important soybean foliar diseases [32][33][34]. Although yield losses have not been significant in the United States since the first discovery of soybean rust in the continental U.S. in 2004 [10], the disease still has the potential to cause major economic damage to U.S. soybean production. This is because the causal agent P. pachyrhizi has a broad host range [35], is airborne [36], and has a diverse genetic structure [37,38], complex virulence patterns, and a high level of pathotype diver- Figure 2. Correlation between soybean rust severity and sporulation of 54 soybean genotypes evaluated against a Mississippi isolate MS 10-6b of Phakopsora pachyrhizi. Soybean rust severity was evaluated using a 5-point scale based on lesion density of the percentage of infected area where 1 = no visible lesions, 2 = a few lesions (1-20% infected area), 3 = moderate lesion density (21-50% infected area), 4 = heavy lesion density (51-80% infected area), and 5 = very heavy lesion density over most of the leaf (81-100% infected area). Sporulation of P. pachyrhizi was rated based on the relative percentage of lesions producing urediniospores on each plant tested using a 1-to-5 scale where 1 = no sporulation, 2 = less than 25%, 3 = 26 to 50%, 4 = 51 to 75%, and 5 = 76 to 100% of the lesion sporulating [29][30][31].

Discussion
Soybean rust is one of the most economically important soybean foliar diseases [32][33][34]. Although yield losses have not been significant in the United States since the first discovery of soybean rust in the continental U.S. in 2004 [10], the disease still has the potential to cause major economic damage to U.S. soybean production. This is because the causal agent P. pachyrhizi has a broad host range [35], is airborne [36], and has a diverse genetic structure [37,38], complex virulence patterns, and a high level of pathotype diversity [15,[38][39][40][41]. The broad host range of the pathogen has facilitated the spread of P. pachyrhizi. The ability of P. pachyrhizi urediniospores to be spread by the wind for long distances increases the opportunity for widespread distribution worldwide and has the potential for causing severe yield losses in any soybean growing area [42,43]. Moreover, the possibility of continuing adaptation of the pathogen in the continental U.S. should not be underestimated. Currently, no soybean cultivars grown in the U.S. are reported to be resistant to all isolates of P. pachyrhizi. When weather conditions are favorable for disease development, serious epidemics could occur. Therefore, developing soybean cultivars with resistance to domestic isolates of P. pachyrhizi is needed to combat emerging potential threats of soybean rust in the U.S. Analysis of reactions of soybean genotypes to P. pachyrhizi is an important step towards breeding for resistance to SBR.
The USDA Soybean Germplasm Collection located at the University of Illinois (Urbana, IL) maintains over 21,810 accessions of the genus Glycine with over 19,626 plant introductions from 92 countries (https://www.ars-grin.gov/npgs/, accessed on 16 February 2023). The accessions are a collection of natural genetic diversity. Miles et al. [29] screened 16,595 soybean accessions from the USDA Germplasm Collection with 4 foreign P. pachyrhizi isolates from Brazil, Paraguay, Thailand, and Zimbabwe at the USDA, ARS, Foreign Disease-Weed Science Research Unit located in Maryland. Some potential SBR-resistant accessions that had low rust severities with an RB-resistant reaction were identified. In another study, Walker et al. [44] evaluated 118 soybean germplasm accessions for resistance to SBR at up to 5 locations in the southern United States. No accession was immune to soybean rust in all field trials at the five locations.
To identify new sources of resistance to domestic P. pachyrhizi isolates, one of our strategies was to use domestic isolates to evaluate the soybean lines that were previously identified as resistant to foreign isolates [30,45]. In our previous study, the 10 plant introductions that were reported as resistant in Paraguay [46] were selected and tested using P. pachyrhizi isolates from Mississippi [30]. This approach resulted in the successful identification of new sources of resistance to American and Paraguayan isolates of P. pachyrhizi, which led to the identification and mapping of the Rpp6 gene in PI 567102B and PI 567104 [19,20].
Resistance to soybean rust has often been identified primarily based on infection/lesion types. However, variations in the RB and TAN reaction types have been reported [12,25,30]. Not all rust lesion types are typical in relation to levels of severity and sporulation. It has been suggested that the RB and TAN reaction types should be subdivided based on the levels of sporulation [5]. An assessment of the resistance to soybean rust was conducted not only for lesion color but also for uredinia of P. pachyrhizi per lesion, the frequency of lesions that had uredia, and the frequency of open uredinia and sporulation level [47]. In another study, Paul et al. [25] found that lesion color was usually a reliable indicator of the average number of uredinia per lesion and that using lesion color to classify infection type could be much quicker than counting uredinia per lesion [25].
In our present study, lesion type was not strongly predictive of rust severity and sporulation level, although the mean severity and sporulation for RB lesion types were lower than those for TAN lesions. For example, the RB-classified breeding line 4014-242-341 had scores of 3.6 for both rust severity and sporulation, which were relatively high, whereas the mean rust severity and sporulation scores for all RB-classified genotypes were only 2.8 and 2.3, respectively. In contrast, PI 567099A produced TAN lesions, but its severity score was only 3.1, which was lower than that of six RB genotypes, including the resistant check PI 459025B (Rpp4). Breeding line 7092-x-1 derived from PI 567099A had an RB reaction type unlike its parent but had a similarly low severity score (2.9) as its parent. However, the sporulation score of 7092-x-1 was significantly lower (1.8) than that of its parent (3.3). The cause for the more resistant lesion type and lower sporulation level in the progeny (7092-x-1) versus the parent (PI 567099A) is uncertain. It is possible that 7092-x-1 inherited a more favorable genetic background from its other parent (LG01-5087-9) for a fuller expression of resistance of the rpp3 gene derived from PI 567099A. This would be an example of epistasis where the rpp3 gene expresses a more resistant phenotype in the progeny compared to the parent because of its differential interaction with the genotypic background of the progeny versus the parent. The genetic mechanisms of the Rpp genes and their interactions with diverse genotypic backgrounds may not be fully understood and need further investigation.
Multiple soybean lines with different resistance alleles at the Rpp1 locus on Chr 18 were tested against the Mississippi isolate of P. pachyrhizi (MS 06-1b) in this study. Interestingly, all soybean lines containing either Rpp1b, Rpp1c, Rpp1d, or Rpp1e had TAN lesions and were susceptible to the Mississippi isolate. This result was in agreement with the findings from another study [25]. It was reported that the "Rpp1b gene of PI 594538A and the Rpp1 allele in PI 587880A had not provided any resistance to P. pachyrhizi populations in the U.S." However, all of these lines were effective against some P. pachyrhizi isolates and populations in South America [48,49]. PI 200492 with the Rpp1a allele at the Rpp1 locus had a susceptible reaction to some of the U.S. P. pachyrhizi isolates [25], but in our study, PI 200492 had an RB-resistant reaction against the Mississippi isolate with low severity and sporulation scores of 1.3 and 1.2, respectively.
Soybean breeding line 6106-132-1-2 had a TAN reaction even though it was developed from DS-880 x PI 567102B. Although one of its parents (PI 567102B with Rpp6) had the highest level of resistance against the Mississippi isolate, 6106-132-1-2 had high severity (3.9) and sporulation (4.2) scores to go with its TAN lesion type. As 6106-132-1-2 did not inherit Rpp6 from PI 567102B but still had RB reaction types to multiple isolates used by Stone et al. [49], it likely has a resistance gene that is different from Rpp6 but was still likely inherited from PI 567102B, as DS-880 has no known rust resistance.
Different levels of resistance were found among 54 soybean genotypes to a Mississippi isolate of P. pachyrhizi in our present study. Although none of the soybean lines showed completed immunity, PI 567102B containing the Rpp6 gene had a near-immune reaction with only a few lesions on some leaves but no sporulation. PI 567102B was also resistant to four of five P. pachyrhizi isolates from Alabama, Florida, and Louisiana [25] and was resistant to 12 of 16 isolates used by Stone et al. [49]. PI 567102B could be useful for the development of unique Rpp gene pyramids to obtain more durable resistance to SBR.
Considering the diverse genetic structure of the pathogen populations and pathogenic/virulence variations among isolates, experiments are currently underway to test selected soybean breeding lines against multiple P. pachyrhizi isolates from the U.S., with the goal to develop durable resistance to soybean rust and make it available to researchers.  Table 1. Soybean seeds of the PIs, G01-PR16, and cultivars were obtained from the USDA Soybean Germplasm Collection in Urbana, IL and were increased at Stoneville, Mississippi, and Isabela, Puerto Rico. Seed of the GC lines was obtained from the World Vegetable Center and increased in Stoneville. The 26 unreleased improved breeding lines, along with CM422 and DS5-67, were developed and tested by USDA scientists at multiple locations and in collaboration with researchers in Paraguay from the Instituto Paraguayo de Tecnologia Agraria (IPTA) and the Camara Paraguayo de Exportadores de Cereales y Oleaginosas (CAPECO).

Pathogen Isolate, Purification, and Maintenance
A Mississippi P. pachyrhizi isolate (MS 06-1b) was used to test the reaction of all soybean genotypes. This pathogen was collected from urediniospores of naturally infected kudzu leaves in Jefferson County, Mississippi in August 2006. Three methods, including microscopy determination, enzyme-linked immunosorbent assay (ELISA), and polymerase chain reaction (PCR), were used to confirm the identity of the pathogen as previously described [28]. Increasing urediniospores were conducted by inoculating on leaves of a susceptible soybean cultivar Williams 82 grown in a growth chamber. The isolate was then purified as previously described [30]. Briefly, under an Olympus SZX12 dissecting microscope, a single uredinium from a lesion was picked with a small needle and reinoculated onto a leaf of Williams 82. After four such "inoculation-isolation" cycles, urediniospores from single-uredinium, isolate-infected leaves were collected using a Cyclone Surface Sampler (Burkard Manufacturing Co. Ltd., Rickmansworth, UK) connected to a vacuum pump. It was performed from 10 to 28 days after inoculation. Urediniospores and infected leaves were stored in a −80 • C freezer for further use.

Inoculum Preparation and Plant Inoculation
Inoculum was prepared using freshly collected urediniospores of MS 06-1b from Williams 82 plants in a growth chamber. As previously described [30], urediniospores were suspended in a solution containing 0.01% Tween-20 (vol/vol) in sterile distilled water and then filtered through a 100 mm cell strainer (BD Biosciences, Bedford, MA, USA) to remove clumped urediniospores and debris. Prior to inoculation, quantification of urediniospores in the suspension was determined using a hemocytometer and then diluted to a final concentration of 4 × 10 4 urediniospores/mL. Soybean entries were arranged in a randomized complete block design with two replications. Two seeds of each of the soybean lines were sown in each Jiffy poly-pak pot (Hummert, St. Louis, MO, USA). Plants in a pot were thinned to one plant per pot before inoculation. Sun Grow Metro Mix 360 soil (Sun Grow Horticulture Products, Belleview, WA, USA) was used to fill the pots. Pots were placed in a Conviron growth chamber under a 16 h photoperiod with a light intensity of 433 µEm −2 s −1 at 25 ± 2 • C. Plants were watered daily.
Inoculation was performed on 21-day-old seedlings at the V2-V3 growth stage [50] using a Preval sprayer (Younkers, NY, USA) at a rate of 1 mL of spore suspension per plant. A mock inoculation was carried out on two pots of Williams 82 plants to monitor infection of the same solution without urediniospores. After inoculation, plants were moved to a dew chamber in the dark at 22 • C overnight (approximately 16-18 h) and then placed back into the Conviron growth chambers where temperatures were maintained at 23 • C during the day and 20 • C at night under a 16 h photoperiod with a light intensity of 280 µEm −2 s −1 . The same experiment was performed three times.

Assessment of Lesion Types, Rust Severity, and Sporulation
The lesion types, soybean rust severity, and sporulation of P. pachyrhizi on lesions were assessed 14 days after inoculation (DAI). Rust severity was determined based on lesion density (percentage of infected area) on the first trifoliate leaves. A 5-point scale was used as previously described [29,30] where 1 = no visible lesions, 2 = a few lesions with approximately 1-20% of the area infected, 3 = lesion density with approximately 21-50% of the area infected, 4 = substantial lesion density with approximately 51-80% of the area infected, and 5 = very heavy lesion density with approximately 81-100% of the area infected. Lesion types on each soybean line were also recorded and classified as "TAN", "RB", or immune reaction as previously described [5,13]. The "TAN" lesion type had tan-colored lesions and was considered a susceptible reaction, whereas the "RB" type lesion exhibited reddish-brown colored lesions and was considered a resistant or partial/incomplete resistant reaction [12]. An immune reaction indicated a lack of obvious symptoms. Sporulation of P. pachyrhizi was rated based on the relative percentage of lesions producing urediniospores on each plant tested using a 1-to-5 scale where 1 = no sporulation, 2 = less than 25% sporulation, 3 = 26 to 50% sporulation, 4 = 51-75% sporulation, and 5 = 76 to 100% of the lesions sporulating [30,31].

Data Analysis
Data were analyzed statistically using the general linear mixed model procedure (PROC Glimmix) in SAS [version 9.1, SAS Institute, Cary, NC, USA]. Means of soybean rust severity and sporulation scores of soybean lines were compared with Fisher's protected least significant difference (LSD) at p ≤ 0.05 unless otherwise stated. A Pearson correlation coefficient analysis of soybean rust severity and sporulation was performed on rating scores from all genotypes, as well as separately on the soybean lines producing either RB or TAN reactions.

Conclusions
This research provides new information about the reactions of different soybean genotypes to a midsouthern USA isolate of P. pachyrhizi. PI 567102B (Rpp6) had a nearimmune reaction with the lowest disease severity score and no sporulation. Four soybean lines (RN06-32-1(7-b, GC 00138-29, G01-PR16, and GC 84051-9-1) had RB reactions and significantly lower SBR severity and sporulation than three of the six resistant checks, PI 230970 (Rpp2), PI 462312 (Rpp3), and PI 459025B (Rpp4). Those soybean lines could be utilized for the development of Rpp gene pyramids or crossed with other resistance resources to obtain more durable resistance to SBR and thereby aid in breeding for resistance to SBR.