Near-isogenic soybean lines carrying Asian soybean rust resistance genes for practical pathogenicity validation

Asian soybean rust caused by the fungal pathogen Phakopsora pachyrhizi is the most devastating disease of soybean. The host cultivar specificity of the pathogen shows considerable differentiation depending on the area and season of its emergence. Although resistance genes for P. pachyrhizi (Rpp) have been reported in several soybean varieties, the genetic background of these varieties is highly differentiated. Furthermore, some of the varieties harbor unknown genetic factors in addition to Rpp that could influence resistance reactions against the pathogen. In order to gain a comprehensive understanding of Rpp–P. pachyrhizi interactions, homogenous plant material harboring Rpp genes is necessary. In this study, we bred Rpp-near isogenic lines (Rpp-NILs), which retained identical plant characters originating from a single genetic background, and accordingly showed low-variant compatible/incompatible reactions against the pathogen. These Rpp-NILs can be used as genetic resources for studying P. pachyrhizi epidemiology and elucidating resistance mechanisms. Compatible/incompatible relationships between the soybean rust resistance gene Rpp and isolates of the pathogen P. pachyrhizi are clearly distinguishable using the Rpp-NILs bred in this study.

www.nature.com/scientificreports/ currently hampering research efforts. In this regard, it is desirable to generate genetically uniform materials to compare results of inoculation tests. In this study, we sought to prepare genetically homogeneous plant materials harboring Rpps, for which we generated near-isogenic lines (NILs) of Rpps bred from a single genetic background. Given that these NILs retain an almost identical genome, the phenotypes of the different NILs only reflects the effect of the introduced Rpp gene. In the present study, we bred NILs for Rpp1, Rpp2, Rpp3, Rpp4, Rpp6, and Rpp1-b from the ASR-susceptible soybean variety BRS 184, which shows low valiant inoculation results compared with those from other Rpp donor varieties. These clear reactions are useful for assessing the diversity and host-specificity of pathogens based on inoculation tests. By using our leaf-culture inoculation method 22 and these Rpp-NILs, we are able to ascertain a pathogen's compatibility/incompatibility against each Rpp in only 2 weeks. Moreover, these NILs are potentially applicable for the study of gene-for-gene relationship to identify avirulence genes from P. pachyrhizi that correspond to Rpps.

Results and discussion
Development of Rpp-niLs. Nine Rpp-NILs (BC 5 F 2 lines) carrying single Rpp genes of donor ASR-resistant varieties were successfully developed by backcrossing and SSR marker screening. The results of markerscreening from generation F 1 to BC 5 F 2 are shown in Table S1. In the BC n F 1 from the backcross of BC n-1 F 1 , 50% of the screened plants should carry the target Rpp in the heterozygous state, whereas in the BC 5 F 2 generation, 25% of individuals should carry a homozygous target Rpp. In practice, we obtained frequencies of positive plants that were lower than these theoretical frequencies, owing crossing failure and recombination between markers of the target Rpp. In the most cases, however, an ideal number of positive plants were obtained (Table S1).
We observed that the Rpp-NILs retained uniform plant characters such as seed color and size, whereas the Rpp donor varieties are completely different from each other (Fig. S1). The Rpp-NILs shares more than 95% of the genomic region of BRS 184 by virtue of a single cross and five backcrosses. In order to compare the severity of infection with P. pachyrhizi isolates, these NILs and the Rpp donor varieties were subjected to an inoculation test. Before (Figs. 1, 2). For example, we observed differences in sporulation level on the Rpp1 donor variety PI 200492 and the NIL B1-8-3-2. Moreover, we observed that the donor variety infected with the Brazilian isolate BRP-2.5 showed differences in the levels of sporulation. However, in the case of NIL, all of evaluated lesion revealed that BRP-2.5 showed the highest level of sporulation, thereby indicating that compatible/incompatible relationships between pathogen and plant were more clearly identified using the NIL than the donor as host. In contrast, we observed that the Japanese P. pachyrhizi isolate E1-4-12 failed to form lesions on either NIL or donor (Fig. 2), which we accordingly interpret as indicating an immune response (see Table S2 for details). In the case of the Mexican isolate MRP-13.18, however, we observed lesions on the NIL, whereas no lesions could be detected on the donor, which tends to indicate the donor parental variety show a more complex resistance reaction based not only on Rpp but also other genes contributed to resistance against P. pachyrhizi.

Disease reactions of Rpp-niLs.
For Rpp2, Rpp3, and Rpp6, we bred two, three and one NILs respectively, and all of the NILs exhibited lower variant sporulation levels against pathogens than the differential varieties ( Fig. 2). However, in response to inoculation with MRP-13.18, Rpp3-NILs showed different levels of subsequent sporulation. The results obtained for Rpp1-NIL indicate that this NIL shows a weaker resistance reaction against the pathogens than the donor variety, which can be attributed to the fact that the genetic background of this NIL was P. pachyrhizi susceptible, and that the NIL lacks the benefits conferred by other genes contributed to resistance against ASR. Among the three Rpp3 donors, Hyuuga is relatively resistant to Brazilian isolates and has been reported to carry another resistance gene 18 . The stronger resistance of Hyuuga than its NIL and the other two Rpp3 donors is probably conferred by this additional resistance gene. One possibility explaining this variant reaction is the viability of the pathogen. The reactions of NILs reflect an effect of the introduced Rpp gene only, and pathogens compatible with the introduced Rpp will show clear viability. A similar response was observed for the Rpp4-NIL, for which sporulation varied by lesion. In contrast, we detected no differences between the NIL (B1b-4-2) and the differential variety (PI 594767A).
The uredinium is an essential structure required for pathogen sporulation, and the number of uredinia (NoU) and frequency of lesions with uredinia (%LU) are considered reliable indices for determining the ASR susceptibility/resistance of soybeans. As shown in Fig. S2, we observed no pronounced differences between the NoU on Rpp-NILs and the donor parents. However, the average of NoU on Rpp-NILs indicated a slightly weaker (susceptible) reaction against the pathogen compared with that of the donors (e.g., Rpp6, Fig. S2). As these observations were based on SL phenotype, a possible reason of this trend is the susceptible genetic background of the NILs. This tendency was also observed with respect to frequency of lesions with uredinia (%LU, Fig. S3).
On the basis of the presence of lesions, the presence of spores within lesions, SL, NoU, and %LU (Table S2) 22 , we categorized reaction type of the four P. pachyrhizi isolates on Rpp-NILs into five degrees. As shown in Table 1, Brazilian isolates (BRP-2.5 and BRP-2.6) were more aggressive than the other two isolates.
The genetic background of the NILs is that of the soybean variety BRS 184, which originates from Brazil 27,28 . Given that BRS 184 retains small plant size at the V2-3 growth stage, the Rpp-NILs bred in this study can be  www.nature.com/scientificreports/ maintained in growth chambers, and 3 weeks of growth after sowing is sufficient to the inoculation test. Furthermore, the leaf culture inoculation method used in this study is rapid and can be used to reliably distinguish the efficacy of Rpp genes against pathogens, with lesions appearing on leaves in less than 2 weeks after inoculation.
In combination with Rpp-NILs, this inoculation method represents a powerful tool for research on ASR epidemiology. Moreover, the recent release of genomic sequences of P. pachyrhizi will yield fundamental information relating to pathogen and homogeneous plant resources such as the Rpp-NILs, and will thus provide a strong incentive for further studies on ASR. In most cases, the Rpp-NILs showed clear susceptible/resistant reactions against the pathogens (Figs. 1, 2), whereas in some Rpp differentials, genetic background influences the phenotype related to ASR susceptibility/ resistance 19,23 . Given that, apart from the target Rpps, Rpp-NILs theoretically share more than 95% similarity with respect to genetic background, these homogenous genetic resources are applicable for studies on gene-for-gene relationships between soybean Rpp genes and the avirulence genes of P. pachyrhizi.  (Table S3). PI 200492, PI 594767A, PI230970, FT2, PI462312, PI459025, PI 567102B, and BRS 184 were provided by the Brazilian Agricultural Research Corporation (Embrapa) in Brazil. Iyodaizu B and Hyuuga were provided by the National Institute of Crop Science (NICS) in Japan. Six PI accessions and 2 Japanese varieties: Iyodaizu B and Hyuuga can be accessed via U.S. National Plant Germplasm System (https ://npgsw eb.ars-grin.gov/gring lobal /searc h.aspx) and NARO Genebank project (https ://www.gene. affrc .go.jp/index _en.php), respectively. The F 1 plants thus obtained were once again crossed with BRS 184 (backcross) to generate BC 1 F 1 plants. In order to achieve 95% identity between NILs and BRS 184 and among NILs, we performed a total of five backcrosses. In the process of recurrent backcrossing, BRS 184 was used as an ovule parent at least once to exclude genetic influences from cytoplasmic difference between donor parents and BRS 184. BC 5 F 1 plants were then selfed to obtain BC 5 F 2 plants (BCF 2 ). BC 5 F 3 and BC 5 F 4 plants were developed by single-seed decent (SSD) from BC 5 F 2 , Therefore, we consider that all NILs are BC 5 F 2 lines. A total of nine NILs (Table 1) were developed and BC 5 F 2 (for the NILs B1b-4-2 and B6-5-2) or BC 5 F 4 plants (for the other 7 NILs) were used for inoculation with pathogens and evaluation of their reactions. F 1 plants were assessed for hybridisms using one of the simple sequence repeat (SSR) markers polymorphic between the parents. The backcrossed progenies (BC 1 F 1 -BC 5 F 1 ) were also screened using SSR markers (Table S3) to determine whether they carry the target resistance genes of their donor parents in the heterozygous state. For each Rpp locus, we used at least two SSR markers polymorphic between the parents and the sandwiching target Rpp locus. BC 5 F 2 plants were screened using the same SSR markers used in BC 5 F 1 generation to obtain BC 5 F 2 plants carrying resistant Rpp alleles in the homozygous state. Given that BC 5 F 2 plants should possess Rpp genes in the homozygous state, all SSR markers used in this study were co-dominant between parents. DNA extraction, PCR amplification, and subsequent electrophoresis were performed following previously described procedures 22 .

Phakopsora pachyrhizi inoculation test on Rpp-niLs.
To evaluate the ASR compatibility/incompatibility of Rpp-NILs, P. pachyrhizi isolates from three countries, Japan (E1-4-12), Brazil (BRP-2.5 and BRP-2.6), and Mexico (MRP-13.18) were subjected to an inoculation test. Of the aforementioned isolates, E1-4-12, BRP-2.5, and BRP-2.6 were isolated in our previous studies 14,23 , whereas MRP-13.18 was isolated from MRP-13 29 by a single-lesion isolation method 22 in the present study. The Brazilian and Mexican P. pachyrhizi isolates are Import-prohibited articles in Japan where the work was conducted (Import Permit Numbers: 20Y157 and 27Y935 from the Yokohama Plant Quarantine Office of Japan for Brazilian and Mexican P. pachyrhizi populations, respectively). Spores of P. pachyrhizi collected from soybean leaves were dried and maintained at − 80 °C for long-term storage. Frozen spores were heated to 39 °C in a water bath for 1 min immediately prior to inoculation. Inoculation of P. pachyrhizi isolates was performed following an online manual 22 . Briefly, leaflets were detached from the lowest trifoliate leaf of V2-3 stage soybean plants, and gently hand-rubbed with autoclaved distilled water to remove trichomes on the leaf surface. Spores of P. pachyrhizi were soaked in 0.04% Tween 20 and adjusted to a concentration of 5.0 × 10 4 spores/mL. For each inoculation, spore suspensions were spread onto three soybean leaves using a paintbrush. Each of three leaves were sampled from different plants. Inoculated leaves were maintained at 21 °C in Petri dishes to maintain moist conditions. Immediately after inoculation, the Petri dishes were covered in aluminum foil to maintain darkness for 24 h. Two weeks after inoculation, we evaluated the average sporulation level (SL), average number of uredinia per lesion (NoU), and frequency of lesions with uredinia (%LU) in all assessed lesions for each of up to 30 lesions per inoculation. We initially evaluated SL, after which NoU was counted following the removal of spores from the inoculated leaves using a paintbrush. Statistical analysis of each value was performed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA).

Data availability
Near isogenic lines (NILs) used in this study are available at Japan International Research Center for Agricultural Sciences under Material Transfer Agreement.