Genetic Characterization of Soybean Rhizobia Isolated from Different Ecological Zones in North-Eastern Afghanistan

Seventy rhizobial isolates were obtained from the root nodules of two soybean (Glycine max) cultivars: Japanese cultivar Enrei and USA cultivar Stine3300, which were inoculated with different soil samples from Afghanistan. In order to study the genetic properties of the isolates, the DNA sequences of the 16S rRNA gene and symbiotic genes (nodD1 and nifD) were elucidated. Furthermore, the isolates were inoculated into the roots of two soybean cultivars, and root nodule numbers and nitrogen fixation abilities were subsequently evaluated in order to assess symbiotic performance. Based on 16S rRNA gene sequences, the Afghanistan isolates obtained from soybean root nodules were classified into two genera, Bradyrhizobium and Ensifer. Bradyrhizobium isolates accounted for 54.3% (38) of the isolates, and these isolates had a close relationship with Bradyrhizobium liaoningense and B. yuanmingense. Five out of the 38 Bradyrhizobium isolates showed a novel lineage for B. liaoningense and B. yuanmingense. Thirty-two out of the 70 isolates were identified as Ensifer fredii. An Ensifer isolate had identical nodD1 and nifD sequences to those in B. yuanmingense. This result indicated that the horizontal gene transfer of symbiotic genes occurred from Bradyrhizobium to Ensifer in Afghanistan soil. The symbiotic performance of the 14 tested isolates from the root nodules of the two soybean cultivars indicated that Bradyrhizobium isolates exhibited stronger acetylene reduction activities than Ensifer isolates. This is the first study to genetically characterize soybean-nodulating rhizobia in Afghanistan soil.

Soybean (Glycine max (L.) Merr.) originated in eastern Asia, and is now being cultivated worldwide under various climatic conditions. Soybean has significant agronomic and nutritional importance because of the high concentrations of protein and oil in its grains. In order to achieve optimum growth and a high yield, soybean has strong demands for nitrogen (N) to synthesize protein. However, nitrogen is one of the nutrients that most frequently limits plant growth in many soils. Soybean may obtain a large part of its nitrogen requirement by establishing N 2 -fixing symbiosis with rhizobia. Soybean-nodulating rhizobia have the ability to nodulate and establish effective symbiosis with soybean plants such as other legumes. They have been isolated and genetically characterized in different continents and climatic zones. These bacteria are Gram-negative and genetically diverse, and have been classified into different genera and species. Soybean is nodulated by fast-and slow-growing rhizobia. The major slow-growing soybean-nodulating bradyrhizobia are Bradyrhizobium japonicum (19), B. elkanii (23), B. liaoningense (45), and B. yuanmingense (3,31), while the fast-growing rhizobia are Ensifer fredii (also known as Sinorhizobium fredii) (10,28), Mesorhizobium albiziae (44), M. septentrionale, M. temperatum (15), and M. tianshanense (9). The genetic diversity and geographical distribution of soybean rhizobia are related to numerous factors such as pH (47) and climate (1). Among three well-known soybean symbionts (B. japonicum, B. elkanii, and B. liaoningense), B.
Afghanistan is a landlocked country located in the center of Asia and forms part of South Asia, Central Asia, and Greater Middle Eastern Asia. It is bordered by Pakistan in the south and east, Iran in the west, Turkmenistan, Uzbekistan, and Tajikistan in the north, and China in the far northeast, as shown in Fig. 1.
Agriculture is an important sector in Afghanistan, and its economy accounts for approximately one third of the gross domestic product (GDP) (20). Agriculture production is regarded as the key component of the economy and livelihood (20). Soybean cultivation was launched in 1881 (36), and followed with the last yield of 2.5 t ha -1 (18). The Afghanistan government and American NGOs in Afghanistan recently introduced new soybean varieties including Stine3300. In the present study, we attempted to evaluate the host specificities of Asian (Enrei) and American (Stine3300) soybean varieties with Afghanistan soybean microsymbionts. Six soil samples were collected from various fields in different agricultural-ecologicalclimatic zones, including the main cropping area. Samples were used as inoculants for two soybean cultivars, the Asiatic cultivar (Japanese c.v. Enrei) and USA cultivar (Stine3300), in order to isolate the root nodule bacteria associated with soybean. Rhizobial diversity was evaluated using several molecular approaches, including the 16S rRNA gene and symbiotic genes (nifD and nodD1). The most promising isolates based on their growth performance in inoculated soybean cultivars were selected as candidates to develop bio-fertilizers for soybean in Afghanistan.

Soil sampling
In order to isolate soybean-nodulating rhizobia, 6 soil samples were collected from various legumes fields (alfalfa, mung bean, and soybean) in different agricultural climatic regions at depths of 0 to 20 cm (Fig. 1, Table 1). Soil sampling was performed between August 15th, 2012 and September 10th, 2012. These samples were transported by air to Japan under strict control by the Yokohama Plant Quarantine office. After arriving at Narita in Japan, these soil samples were maintained at 4°C in a cold room. Rhizobium isolation using the samples imported from Afghanistan was conducted between October and November, 2012. The sampled sites of the soils had no history of soybean cultivation, except for Kabul (recently started soybean cultivation), and no bacterial inoculation. Therefore, the isolated strains were considered to be indigenous to Afghanistan.

Isolation of indigenous rhizobia
The seeds of two soybean cultivars: c.v. Stine3300 (USA variety) and c.v. Enrei (Japanese variety), were surface-sterilized by immer-sion in 70% ethanol for 30 s and then in 3% sodium hypochlorite solution for 3 min before being exhaustively washed with sterile water. Regarding germination, sterilized seeds were incubated on 0.9% agar medium at 25°C for 1.5 d. Five-fold dilutions (1 g 5 mL -1 ) of soil suspensions were used as inoculants. Each inoculant was applied to 300-mL glass jars (except the control) containing germinated seeds and sterilized vermiculite. The jars were transferred to a growth chamber and kept under controlled conditions (16-h light/8-h dark photoperiod, at 25°C/18°C day/night temperatures). Plant growth was supported by adding sterilized nitrogen-free nutrient solution (7) to the jar up to the 60% soil moisture level. After four weeks, whole plants were removed from the jars, washed in running tap water to remove vermiculite, and the root nodules were harvested. Root nodules were surface-sterilized by immersion in 70% ethanol for 30 s and in 3% sodium hypochlorite for 3 min, and then washed five times with sterile water. Each nodule was crushed in 100-200 μL glycerol solution [15% (v/v)] to obtain a turbid suspension. An aliquot (10 μL) of the suspension was streaked onto yeast extract mannitol agar (YEM) (37) plates and incubated at 28°C for 3-7 d. The remaining suspension was maintained at -80°C. Single colonies were picked and checked for purity by repeated streaking onto fresh YEM. Isolates were transferred into 15% glycerol stocks and stored at -80°C for long-term maintenance and for short-term maintenance were transferred into slant stocks and stored at 4°C.

DNA extraction
Isolates were grown in 20 ml YEM broth medium at 25°C for 4-7 d. Prior to genomic isolation, cells were harvested and washed twice with equal volumes of TNE buffer (10 mM Tris, 0.1 M NaCl, and 1 mM EDTA, pH 8). Genomic DNA was extracted from isolates using the method described previously by Yokoyama et al. (49). The concentration and purity of DNA were checked using a Nano Drop 2000 UV-Vis Spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA).

DNA amplification and sequencing
The PCR amplification and sequencing of DNA fragments of the 16S rRNA, nifD, and nodD1 genes were performed as described by Habibi et al. (16). The primer sets used for PCR amplification and sequencing are shown in Table S1. PCR products were sequenced using an ABI PRISM 3500 genetic analyzer (Applied Biosystems) according to the manufacturer's protocols. DNA sequences were compared to the GenBank nucleotide sequence database using BLAST (NCBI). Sequence alignment and construction of the phylogenetic tree were performed using MEGA version 6.06 (41).

Nucleotide sequence accession numbers
The 16S rRNA, nodD1, and nifD sequences have been deposited in the DNA Data bank of Japan under the accession numbers AB901294-AB901363 for 16S rRNA, LC009304-LC009373 for nodD1, and AB982147-AB982216 for nifD.

Symbiotic performance
The same soybean cultivars (Enrei and Stine3300) were used to confirm compatibility for effective nodule formation by each isolate. Fourteen isolates were selected from different genotypes as representatives (Table 2) and grown in 15 mL YEM broth at 28°C for 3-7 d. In order to evaluate the cell numbers of the inoculants, bacterial cells were collected by centrifugation at 28,000×g at 4°C for 10 min, and then washed twice with 1× TNE solution. Colony-forming units (CFU) were counted using the dilution plate count method. Inoculant cells (1-1.5 mL) at a density of 10 8 CFU were applied to a seed in a growing jar. The jars were transferred to a growth chamber and kept under controlled conditions (16-h light/8-h dark photoperiod, at 25°C/18°C day/night temperatures). After four weeks, whole plants were removed from the jars, washed in running tap water to remove vermiculite, and the root nodules were harvested. This experiment was performed in a completely randomized block design that consisted of three replicates for each treatment. Different plant growth parameters such as shoot weight (fresh and dry), root weight (fresh and dry), root nodule weight (fresh and dry), and nodule numbers were assessed. Plant shoots and root nodules were dried at 80°C for 48 h in order to obtain dry weights. In the acetylene reduction assay, fresh roots that contained root nodules were placed in a 300-mL jar, the air in the jar was supplemented with 10% acetylene (v/v) for each treatment, and the jar was incubated at room temperature (25°C) for 1 h. The concentration of ethylene in the jar was measured using a gas chromatograph (Shimadzu 2014AF, Kyoto, Japan). B. japonicum USDA 110 (31) was included as a positive control in the symbiotic test. The significance of differences between inoculated and un-inoculated plants was assessed using Tukey's test (P<0.05) ( Table 2).

Root nodule numbers of Enrei and Stine3300 in combination with 6 different soils
Eighty nodules were harvested from two soybean cultivars (29 from Enrei and 51 from Stine3300). Two soil samples (Kabul and Bamyan) failed to form root nodules in either soybean cultivar (Table 1). We also did not observe any root nodules in soybean plants growing in Kabul Province at the time of soil sampling. No root nodule was found in the Enrei cultivar after inoculations with five of the soil samples (samples 2, 3, 4, 5, and 6 in Table 1). The only soil sample that produced a large number of root nodules in both cultivars was from Nangarhar (sample 1 in Table 1). The Stine3300 cultivar showed higher adaptability than the Enrei cultivar to the indigenous soybean rhizobia in three other soil samples (samples 3, 4, and 5 in Table 1). Seventy rhizobial isolates were obtained from the nodules (Table 3).

Phylogenetic analysis based on the 16S rRNA gene
The almost full length of the 16S rRNA gene (1331 bp) of 70 isolates was sequenced and their accession numbers (AB901294 to AB901363) are shown in Table 3. These isolates were divided phylogenetically into two groups (GI and GII) based on the 16S rRNA sequence analysis, as shown in Fig. 2. The GI group contained 38 isolates (54.3% of the * Value is significantly different from the control within each column in plant biomass production (P<0.05). 1) and 2) show the first and second values of each symbiotic performance of the tested isolates to soybeans. Phylogenetic analysis based on the nod D1 gene Based on differences in the nodD1 (658 bp) phylogenetic analysis, 70 isolates were divided into two major groups, as shown in Fig. 3 and Table 3. The GI group contained 39 isolates (55.7% of the total), while the GII group contained 31 isolates (44.3%). The GI group was divided into two subgroups: GIa, which contained 32 isolates and had a close relationship (98%) with B. yuanmingense; and GIb, which contained seven isolates and separated from B. yuanmingense T8 AB601727 and B. yuanmingense T1 (AB601720). All B. liaoningense isolated from Afghanistan soil shown in Fig. 2 had the nod genes of B. yuanmingense. The GII group had a close relationship (99%) with E. fredii and reference strains. Interestingly, one isolate (KS2) that was classified into the Ensifer group based on 16S rRNA sequences showed a close relationship (99%) with B. yuanmingense based on nodD1 sequences (Table 3).

Phylogenetic analysis based on the nifD gene
Seventy tested isolates formed two groups in the phylogenetic tree based on nifD sequences (677 bp), as shown in Fig.  4 and Table 3. The GI group contained 39 isolates (55.7% of the total), while the GII group contained 31 isolates (44.3%), as shown in Table 3. The GI group showed a close relationship (99%) with B. yuanmingense and was divided into two subgroups: GIa, which contained 34 isolates, and GIb, which contained five isolates. In the GIb subgroup, four isolates (PS2, PS5W, PS6, and PS8) were from the same region (Parwan Province), while the KS3 isolate was from Kunduz Province. All isolates in GIa and GIb belonged to Bradyrhizobium and had close relationships with the B. yuanmingense isolates. Furthermore, all B. liaoningense isolated in Afghanistan soil shown in Table 3 and Fig. 2 had the nif genes of B. yuanmingense. The KS2 isolate belonged to the Ensifer group based on 16S rRNA sequences and was classified into the GIb subgroup based on nifD sequences. The GII group showed a close relationship (99%) with E. fredii USDA 257 and E. fredii HH103.

Symbiotic performances
The symbiotic performances of the isolates belonging to different phylogenetic groups based on the 16S rRNA, nifD, and nodD1gene sequences are shown in Table 2. Fourteen representative isolates produced root nodules when inoculated onto the seeds of the Stine3300 and Enrei cultivars. The root nodules that developed in soybean by the Bradyrhizobium Fig. 3. Phylogenetic tree constructed using partial nucleotide nodD1sequences (658 bp) from 70 Afghanistan isolates. GenBank accession numbers are given in brackets. The numbers at the nodes indicate the level of bootstrap support based on a neighbor-joining analysis. One isolate (KS2) from a Ensifer species showed close similarities to Bradyrhizobium yuanmingense and was categorized into the GIa subgroup. isolates exhibited slightly stronger acetylene reduction activities (ARA) than the Ensifer isolates. However, no significant differences were observed in root nodule numbers in the two soybean cultivars among the Bradyrhizobium and Ensifer isolates.
Isolates in the GI group defined by the nodD1 sequence analysis belonged to the genus Bradyrhizobium, and KS3 and GE3 showed high root nodule numbers in both soybean cultivars. Of the eight isolates tested, the root nodule numbers of six (KS3, BgS4, PS3, GE3, KS12, and PS8) were higher in Enrei than in Stine3300. In the ARA, the PS2 and KS12 isolates, which belonged to the GIa and GIb groups, respectively, based on the phylogenetic tree constructed using the nifD sequences, exhibited strong ARA activities in the Stine3300 cultivar. In the root nodules of Enrei, GE3 and KS12 also exhibited strong ARA activities, with KS12 exhibiting particularly strong ARA activity in both cultivars. GE3 showed a higher biomass amount with the strongest ARA activity in Enrei, while KS3 increased the biomass of Enrei and had the highest root nodule number among all the isolates tested. No significant differences were observed in symbiotic performances among the three GI subgroups of Bradyrhizobium isolates. In the GII group based on the phylogenetic tree constructed using the nodD1 sequences, the isolates were E. fredii and generally showed higher numbers of root nodules in Enrei than in Stine3300 as well as stronger ARA activities (excluding the GE24W isolate). GE6W showed high nodulation adaptability in the two cultivars. In the GII group based on the phylogenetic tree constructed using the nifD sequences, the isolates in the root nodules obtained from Enrei exhibited strong ARA activities, with the exception of GE24W. GS4 showed relatively good symbiotic performance based on the high root nodule number and strong ARA activity. GE6W isolates showed high biomass amounts in the two cultivars, with high root nodule numbers; however, ARA values were not very high. The KS2 isolate was grouped with Bradyrhizobium based on the trees constructed using the nifD and nodD1 sequences ( Table 2, Part B). The symbiotic performance of KS2 was similar to that of GE24W, which belonged to the Ensifer type.

Isolation of root nodule bacteria
Six soil samples were collected from different sites (North to South-eastern) in Afghanistan (Fig. 1), inoculated into two soybean cultivars (Enrei and Stine3300) as trap hosts, and rhizobia were isolated from their root nodules. A soil sample collected in the Nangarhar Province showed a high frequency for the distribution of soybean rhizobia; this may be because of continuous mung bean cultivation in this region (4,51). In contrast, two soil samples obtained from Kabul and Bamyan provinces showed no distribution of soybean rhizobia. In the soil sample from Kabul Province, we were unable to confirm any soybean rhizobia despite soybean having been cultivated in this region. This result was consistent with our observation that no root nodule was observed in the soybean plants cultivated in Japan. The Japanese Enrei cultivar had a low level of host specificity; among the six soils tested, only the soil sample from Nangarhar province contained soybean rhizobia with the ability to nodulate with cv. Enrei.

Soybean rhizobia communities in Afghanistan soils
Based on 16S rRNA gene sequences, 70 Afghanistan isolates were categorized into two major clusters, the genus Bradyrhizobium and genus Ensifer (Fig. 2). Bradyrhizobium isolates were further divided into three subgroups: GIa, GIb, and GIc. GIa isolates shared almost identical 16S rRNA gene sequences with that of B. yuanmingense NBRC 100594 T , while GIb isolates were similar to, but also different from B. yuanmingense NBRC 100594 T . B. yuanmingense was originally described by Yao et al. (48) as a microsymbiont of Lespedeza sp.. Host-specificity studies reported ineffective root nodule formation in soybean by B. yuanmingense isolates (46,48); however, the effective nodulation capability of B. yuanmingense was recently demonstrated (3,31). The GIc group contained 28 isolates and had a close relationship with B. liaoningense, which was firstly reported in China (44) and then frequently described in other studies (2,3,17,25,34,43,46,47). Li et al. (25) reported that B. yuanmingense and B. liaoningense were the predominant soybean microsymbionts in alkaline and saline soils. The GII group contained Ensifer isolates, fast-growing bacteria isolated from the root nodules of Lablab purpureus in Papua New Guinea alkaline soil (30). E. (Sinorhizobium) fredii was initially isolated from an old Peking soybean variety (G. max) in China (21,22), and their effective nitrogen-fixing symbioses with Asian soybean cultivars (G. max and G. soja) have been reported (12,21,38). Subsequent studies showed that E. fredii also established effective nodules with some American soybean cultivars (5). Ensifer species were initially included in the genus Rhizobium and classified as Rhizobium fredii (35), they were then assigned as S. fredii by Chen et al. (10), and finally reassigned to the genus Ensifer (Young 2003). Ensifer species are acid-producing rhizobia (10), and are well adapted to saline-alkaline soils (3, 17 25, 26, 28, 34, 40); their ability to produce acid substances may contribute to the survival of rhizobia under alkaline conditions (33). Regarding E. fredii isolated from soil samples from Nangarhar and Kunduz provinces, soil pH were 7.66 and 7.65, respectively. E. fredii NGR234 has a wide host range in leguminous plants (112 genera), in which it may form either determinate or indeterminate nodules (30), as well as in the non-legume tropical tree Parasponia andersonii (24). E. fredii NGR234 has a phylogenetically close relationship with E. fredii USDA257, and both strains share most of their genomic backgrounds (29).

Symbiotic gene diversity of Afghanistan isolates
Among 38 Bradyrhizobium isolates from Afghanistan, 28 had a close relationship with B. liaoningense based on 16S rRNA gene sequences, as shown in Table 3 and Fig. 2 Table 2), implying that symbiotic gene transfer from B. yuanmingense to E. fredii occurred in Kunduz soil, as described by Sullivan et al. (39) for Lotus-nodulating rhizobia. Horizontal or lateral gene transfer has long been known to occur among prokaryotes (11), in which it plays a major role in prokaryote genome evolution (14,27). Barcellos et al. (6) and Djedidi et al. (13) also recently described horizontal gene transfer among different genera and species. In the present study, we found that the nodD1 and nifD genes of B. yuanmingense were transferred to KS2 of Ensifer (Tables 2 and 3). However, it currently remains unclear whether the remaining nod and nif genes in the symbiotic island of B. yuanmingense were also transferred to KS2. If all the nod and nif genes of B. yuanmingense were transferred to KS2 and were active, this may explain why the performance of nodulation by the KS2 isolate was similar to those of the isolates of B. yuanmingense (Table 2). However, based on the ARA values of the root nodules produced by KS2, the nitrogen fixation performance of KS2 was markedly lower than that of the B. yuanmingense nodules, while ARA values were similar to those of E. fredii. Further studies are required in order to elucidate the weakness of nif gene activities in KS2 isolates.
We found no clear relationship among the ARA values of the nifD groups, root nodule numbers in nodD1 groups, and plant dry weights shown in Table 2. In Ensifer, all the isolates tested had the same nodD and nifD ( Table 2, Part A) and belonged to the same phylogenetic group (GII); however, their nodule numbers and ARA values markedly varied. In Bradyrhizobium, all the tested isolates exhibited similar tendencies to those of Ensifer ( Table 2, Part C). This result showed that when effective soybean inoculants for Afghanistan soybean cultivation are obtained, characteristics such as the nodule numbers, ARA values, and plant promotion activities of the target candidates need to be checked.

Conclusion
Based on the symbiotic performances of the isolates, which were tested with the two soybean cultivars, we propose that the Ensifer GS4 and GE6W isolates and Bradyrhizobium GE3 isolate are potential candidates to improve soybean cultivation and production in Afghanistan. Soybean is a new legume crop in Afghanistan and its development will contribute greatly to food security in Afghanistan. However, increasing the performance of the soybean crop is a major challenge for Afghanistan farmers, and the efficacy of symbiotic nitrogen fixation may be an important factor for enhancing productivity through the successful management of the soybean and indigenous-rhizobia symbiosis. Our results will contribute to the development of an effective soybean inoculant in Afghanistan. Further studies are required in order to establish a suitable inoculation technology for soybean cultivation in Afghanistan.