Geographical distribution of Bradyrhizobium spp. isolated from Korean soils
Among 279 of strains isolated from soybean nodules, 254 isolates were identified as Bradyrhizobium with allocating into 6 species of B. diazoefficiens, B. elkanii, B. japonicum, B. ottawaense, B. guangxiense, and B. lianoningense (Additional file 1: Table S1). As shown in Fig. 1A, B. diazoefficiens (48.1%) was the most dominant species, followed by B. elkanii (18.2%), B. japonicum (13.8%) and B. ottawaense (8.6%). Figure 1B shows the regional distribution patterns of Bradyrhizobium isolates in South Korea. B. diazoeffciens occupied most part of the nation with except for JN_MA, JB_JU, CB_JC, GW_PC, and CN_TA. B. elkanii was dominant species in JN_MA and JB_JU at 94.4% and 88.9%, respectively. B. ottawaense was also found to be major Bradyrhizobium species in CB_JC (66.7%), and KW_PC (38.9%). B. japonicum was the most common strain in CN_TA with 61.1%. Overall, B. diazoeffciens, to which the most commonly used nitrogen fixing biofertilizer type strain USDA110 belongs, was found in the all sampling sites as a mostly dominant colonizer. Interestingly, B. elkanii species was found in the warm western part of the Korean peninsula, and occupied southwestern part as a major Bradyrhizobium species. However, B. ottawaense was found in the mountainous sites with relatively low temperature. The geographical distribution of indigenous Bradyrhizobium spp in Japan and the United States have shown similar results that B. elkanii was mainly dominant in the southern regions, while B. japonicum and B. diazoefficiens were dominant in the northern regions.
NMDS analysis also revealed that the distribution of Bradyrhizobium spp. was thought to be affected by the annual average temperature (Fig. 2). Additionally, PERMANOVA test revealed that the community of Bradyrhuzobium spp. of soil samples was significantly different by temperature level (R = 0.1806, p = 0.0399). Otherwise, the annual temperature is thought as a potential environmental factor to classify the geographical regions by annual temperature as high temperature (13–14 °C), moderate (11.5–13 °C) and low (10–11.5 °C) (Fig. 2), supporting the different composition of Bradyrhizobium spp. as shown in Fig. 1B. Yuichi Saeki and Sokichi Shiro reported that the indigenous Bradyrhizobium community of Japan and the United State correlated with latitude, and the community could be influenced by the soil temperature associated with the latitude of the particular geographical location or the diversity of acclimatized host plants to the climate [9]. Suzuki et al. describes that B. elkanii is dominant at high temperature, whereas B. japonicum, is dominant at low temperature under the inoculation experiment condition [22]. In Japan, Kyoto (in which the latitude is similar to that in Jeollado provinces and similar temperature to that in Uiryeong, Gyeongsangnamdo) showed the high distribution of the various strains of B. japonicum and the B. diazoefficiens USDA110T. However, B. elkanii USDA76T occupied the largest proportion of isolates in Okinawa located in the southernmost of Japan with higher temperature than Kyoto [7]. In the USA, B. japonicum was predominant in Ohio region (of which the annual average temperature similar to that of Gangwondo provinces). Furthermore, B. elkanii USDA76 was predominant in Kentucky where annual average temperature and latitude are similar to those in Muan, Jeollanamdo [8]. Depending on the study of Ethiopia, the clusters group included in B. ottawaense were dominant in Borcha that have similar temperature and latitude with the northern region of Korea [23]. Those studies support the pattern of distribution of Korean indigenous Bradyrhizobium spp. with similar tendency to those of other countries affected by temperature, geographical, and physiochemical characteristics. The characteristics of geographical distribution of soil samples was also explained using principal component analysis (PCA) using temperature, precipitation, total nitrogen, total phosphate, total organic carbon, carbon to nitrogen ratio, cation exchange capacity, soil texture (Additional file 3: Fig. S1) of soil samples, showing the separation of JN and JB regions from other regions. Eigenvalues of environmental factors obtained from PCA analysis were included in Additional file 4: Table S3.
The abundance of Bradyrhizobium spp. from soybean cultivars is shown in Fig. 3. B. diazoefficiens was predominated in most of the soybean cultivars. Exceptionally, B. elkanii was particularly dominant in the Taekwang soybean cultivar collected from Jeollanamdo province. Devine et al. suggested that certain Bradyrhizobia strains produce the substance named rhizobitoxine, and it could be induced foliar chlorosis into the host plant. The result of the study has been described that B. elkanii USDA 76 T has the highest chlorosis score among 25 Bradyrhizobium strains [24]. The foliar chlorosis affected to postproduction of soybeans by inhibiting the production of flowers and fruits as well as disturbing the initial growth [25]. In this study, most of the B. elkanii identified were classified as strain USDA 76T. In Korea, the distribution of Bradyrhizbium spp. is spatially various, and environmental conditions such as temperature associated with climate and soybean cultivars are thought to affect the distribution structure.
Genetic diversity of Bradyrhizobium spp. in strain level
Dendrograms of each dominant Bradyrhizobium spp. were represented the genetic similarity of respective DNA fragments extracted from each strain. A total of 42 B. elkanii strains were grouped as 4 major clusters based on 70% similarity (Additional file 5: Fig. S2). Cluster A includes the most strains from Jeongeup, Jeollabukdo and cluster B was grouped with the strains obtained from Muan, and Jeollanamdo. Both clusters C and D included strains from Gimje, and Jeollabukdo. The strains were regionally distinguished with the similarity of more than 80%. A total of 129 B. diazoefficiens strains were distinguished into 4 clusters based on 50% similarity (Additional file 6: Fig. S3). Cluster A and D were grouped strains from northern regions, while cluster C was mainly comprised of strains from southern regions. Although cluster B included strains obtained from both the north and south regions, they were divided into two subgroups, which were distinguished by latitudinal variation depending on 55.6% similarity. A total of 31 B. japonicum strains were grouped as 3 clusters based on 50% similarity (Additional file 7: Fig. S4). Most strains isolated from southern regions were classified in cluster C. However, the strains isolated from central and northern regions were clustered in cluster A. A total of 20 strains identified as B. ottawaense were grouped as 3 clusters based on 70% similarity (Additional file 8: Fig. S5). Strains obtained from both northern regions and central regions were clustered in cluster A, and the regions were classified with a similarity of more than 86.4%. Overall, the strains were genetically clustered by geographical location, showing specific strains are thought to have indigenous specificity to regions. In particular, the strains B. diazoefficiens and B. japonicum from southern regions were classified as the genetically different groups from central and northern regions, even though they have over 98.5% similarity of 16S rRNA-based DNA sequence. Exceptionally, the strains isolated from Gimje, Jeollabukdo have represented the high similarity with both strains isolated from central and northern regions and isolated from the southern regions.
Because of horizontal gene transfer of symbiotic chromosomal genes, rep-PCR has been used as an enhanced genomic fingerprinting method to identify the higher taxonomic level of Bradyrhizobium [26]. Some previous studies suggested that the genetic diversity of Bradyrhizobium strains could be influenced by physicochemical soil characteristics such as the proportion of clay in the soil, pH and salt condition [26,27,28]. Elboutahiri et al. verified the high level of genetic changeability under the high salinity and drought conditions in their study through the increasing profile of rep-PCR results using Sinorhizobium species. They proposed that exposure to stressful environments enabled the evaluation of the genome by exchange and acquisition through horizontal gene transfer [29]. In addition, the previous study of Japan described that the genotype of soybean cultivars changes the genetic composition of the Bradyrhizobial community [30].
In conclusion, a high level of diversity at the strain level was confirmed between the strains from different regions. This result may suggest that the Bradyrhizobium strains, symbionts of soybean, were genetically different depending on geographical location and associated environmental differences. It could be explained by the horizontal gene transfer between Bradyrhizobium spp. that occurred for adjusting to environmental condition [31]; however, further studies should be required to prove the direct cause of genetic diversity. Taken together, this study surveyed the distribution and genetic difference of the indigenous Bradyrhizobium community in South Korea for the sustainable agriculture techniques with easing global warming problem.