The Type IV Secretion System (T4SS) Mediates Symbiosis between Bradyrhizobium sp. SUTN9-2 and Legumes

ABSTRACT

There has been little study of the type IV secretion system (T4SS) of bradyrhizobia and its role in legume symbiosis. Therefore, broad host range Bradyrhizobium sp. SUTN9-2 was selected for study. The chromosome of Bradyrhizobium sp. SUTN9-2 contains two copies of the T4SS gene, homologous with the tra/trb operons. A phylogenetic tree of the T4SS gene traG was constructed, which exemplified its horizontal transfer among Bradyrhizobium and Mesorhizobium genera. They also showed similar gene arrangements for the tra/trb operons. However, the virD2 gene was not observed in Mesorhizobium, except M. oppotunistum WSM2075. Interestingly, the orientation of copG, traG, and virD2 cluster was unique to the Bradyrhizobium genus. The phylogenetic tree of copG, traG, and virD2 demonstrated that copies 1 and 2 of these genes were grouped in different clades. In addition, the derived mutant and complementation strains of T4SS were investigated in representative legumes Genistoids, Dalbergioids, and Millettiods. When T4SS copy 1 (T4SS₁) was deleted, the nodule number and nitrogenase activity decreased. This supports a positive effect of T4SS₁ on symbiosis. In addition, delayed nodulation was observed 7 dpi, which was restored by the complementation of T4SS₁. Therefore, T4SS plays an important role in the symbiotic interaction between Bradyrhizobium sp. SUTN9-2 and its leguminous hosts.

IMPORTANCE SUTN9-2 is a broad host range strain capable of symbiosis with several legumes. Two copies of T4SS clusters belonging to the tra/trb operon are observed on chromosomes with different gene arrangements. We use phylogenetic tree and gene annotation analysis to predict the evolution of the tra/trb operon of rhizobia. Our finding suggests that the gene encoding the T4SS gene among Bradyrhizobium and Mesorhizobium may have coevolution. In addition, Bradyrhizobium has a uniquely arranged copG, traG, and virD2 gene cluster. The results of T4SS₁ gene deletion and complementation revealed its positive effect on nodulation. Therefore, T4SS seems to be another determinant for symbiosis. This is the first report on the role of T4SS in Bradyrhizobium symbiosis.

INTRODUCTION

Typically, rhizobia stimulate the nodulation process using Nod factors (NF), which communicate with Nod factor receptors (NFRs) in host plants and lead to nodule formation. However, some rhizobia lack nodulation (nod) gene encoding NF production; they use an alternative mechanism to form nodules known as secretion systems (1). Many pathogenic and symbiotic bacteria translocate compounds such as virulence genes, effector proteins, DNA, and DNA complexes into their host via secretion systems with similar signaling molecules (2). The secretion system involves bacterial membrane-embedded nanomachines translocating effector proteins into the cytoplasm of eukaryotic cells (3). Rhizobia genera such as Rhizobium, Mesorhizobium, Sinorhizobium, and Bradyrhizobium contain genes encoding secretion systems located on both their chromosome and plasmid, depending on the species. The type of secretion system in rhizobia involved in symbiosis has been identified as type III (T3SS), IV (T4SS), and VI (T6SS) secretion systems (4, 5). In rhizobia-legume symbiosis, T3SS is important for plant immune response and nodulation (4, 6). T3SS has both positive and negative effects on symbiotic interactions, depending on the host (6, 7). T3SS and T4SS are macromolecule transporter systems related to symbiotic interactions in several rhizobia. T4SS is well characterized in Agrobacterium and is a standard model for T4SS localization, structural components, and gene arrangement in bacteria (8). Typically, A. tumefaciens C58 transfers T-DNA, effector proteins, and virulence genes to plant cells via T4SS machinery in the pTi (tumor-inducer) plasmid. The pTi plasmid carries several T4SS genes, including virulence (vir), avirulence homolog (avh), and trb genes. The trb genes are annotated for conjugal transfer, while vir and avh are essential for T-DNA transfer (5, 9).

Gene encoding T4SS in Rhizobium, Mesorhizobium, Sinorhizobium, and Bradyrhizobium have been found on chromosomes and plasmids. Most rhizobia contain virB/D4 and tra/trb operons that encode the putative proteins involved in secretion apparatus production. In addition, T4SS is important to symbiotic relationships and host specificity. For example, the vir gene mutant strains of M. loti R7A and MAFF303099 displayed decreased and increased nodule formation depending on the host (10) as the same in S. meliloti KH46c and S. medicae M2 (5). The S. meliloti KH46c T4SS mutant strains showed less competition for nodulation when coinoculated with wild-type strain in equal cell numbers on Medicago truncatula (11). In contrast, the T4SS of S. meliloti 1021 is unnecessary for symbiosis but required for conjugation (12). Interestingly, most bradyrhizobia containing T4SS components harbored tra/trb operon on the chromosome, such as Bradyrhizobium sp. SUTN9-2 (7), B. diazoefficiens USDA110, and B. japonicum USDA6 (13), while vir operon was found on symbiotic plasmids of Bradyrhizobium sp. DOA9 (14) and Bradyrhizobium sp. BTAi1 (15).

Here, we focus on Bradyrhizobium sp. SUTN9-2 that was isolated from Aeschynomene americana. Bradyrhizobium sp. SUTN9-2 is a broad host range strain establishing symbiotic nodule with several legume genera. Based on 16S rRNA and housekeeping gene (dnaK, recA, and glnB) phylogenetic trees, Bradyrhizobium sp. SUTN9-2 belongs to the same clade as B. yuanmingense (16). Furthermore, the secretion systems of Bradyrhizobium sp. SUTN9-2, including T3SS and T4SS, were observed on the chromosome. T3SS of Bradyrhizobium sp. SUTN9-2 plays a negative effect on Vigna radiata and Macroptilium atropurpureum symbiosis while having no effect on the original host, A. americana (17). This phenomenon implied that T3SS is one of the symbiosis determinants for Bradyrhizobium sp. SUTN9-2 (7). In contrast, the information on T4SS in most species of Bradyrhizobium, including Bradyrhizobium sp. SUTN9-2, is still unclear. In this study, two copies of T4SS were identified on the chromosome of Bradyrhizobium sp. SUTN9-2 with different location. Both T4SS belong to the tra/trb operon. The T4SS copies 1 and 2 (T4SS₁ and T4SS₂) of Bradyrhizobium sp. SUTN9-2 showed high similarity with B. diazoefficiens USDA110 and Bradyrhizobium sp. BTAi1, respectively (17). This study focused on T4SS₁ since Bradyrhizobium sp. SUTN9-2 is a nonphotosynthetic bradyrhizobia that is closely similar to B. yuanmingense, the main rhizobia associated with V. unguiculata and V. radiata in the subtropical region (18).

Here, we present evidence for the evolution of T4SS using gene comparison and phylogenetic tree analyses. Moreover, the role of T4SS in the symbiotic interaction of Bradyrhizobium sp. is elucidated in this study for the first time. The function of T4SS in Bradyrhizobium sp. SUTN9-2 was explored by gene mutation and complementation, and then investigated for nodulation efficiency in representative legumes belonging to Dalbergioids (Aeschynomene americana), Genistoids (Crotalaria juncea), and Millettiods (V. radiata cv. SUT4).

RESULTS

Gene annotation of the type IV secretion system (T4SS) of the tra/trb operon in Rhizobia and Agrobacterium.

The tra/trb operon was identified in several genera in the rhizobia group, such as Bradyrhizobium, Rhizobium, Sinorhizobium, Mesorhizobium, and Agrobacterium. The tra/trb complex could be observed in both chromosome and plasmid depending on bacterial strain. The tra/trb operon was found on the chromosome of Bradyrhizobium sp. SUTN9-2, B. diazoefficiens USDA110 and Bradyrhizobium sp. BTAi1, which is the same as in Mesorhizobium. The gene annotation of T4SS in the rhizobia group was compared with A. tumefaciens C58, which is widely recognized as a plant crown gall disease causative agent via T4SS. Interestingly, bradyrhizobia’s traG gene is usually and uniquely flanked with the copG and virD2 genes. Two copies of gene encoded tra/trb operon on chromosome with different gene arrangement was observed in Bradyrhizobium sp. SUTN9-2 (Fig. 1a). Copy 1 was inserted in gene fragment sizes ~56 kb, between trb operon and traG gene. While both M. loti R7A and M. loti MAFF303099 lacked virD2 gene on their chromosomes. Although the tra/trb operon of Rhizobium and Sinorhizobium were found on their plasmids, as in Agrobacterium. This suggests that the tra/trb operon in rhizobia differs from the conjugative function of the tra/trb operon in Agrobacterium.

FIG 1.

Genetic organization and phylogenetic tree of T4SS clusters in strain Bradyrhizobium: (a) T4SS genes organization of Bradyrhizobium sp. SUTN9-2, Bradyrhizobium diazoefficiens USDA110 and Bradyrhizobium sp. BTAi1; Mezorhizobium: Mesorhizobium japonicum R7A and M. japonicum MAFF 303099; Rhizobium: Rhizobium etli CFN 42 (plasmid p42a), R. etli bv. mimosae str. Mim1 (pRetNIM1c); Sinorhizobium: Sinorhizobium fredii HH103 (pSfHH103), S. fredii NGR234 (pNGR234a); Agrobacterium: Agrobacterium tumefaciens C58 (pTiC58). Orthologous T4SS structural genes are in the same color and in white are the surrounding genes. (b) Maximum-likelihood phylogenetic tree based on traG gene sequences of Bradyrhizobium sp. SUTN9-2 with other reference strains of Bradyrhizobium, Rhizobium, and Agrobacterium bacteria. Cupriavidus taiwanensis LMG19424 was used as an outgroup. The phylogeny was inferred using the maximum likelihood method. Scale bar indicates 0.1 substitutions per site.

The phylogenetic tree of the T4SS structural gene (traG/trb operon).

To examine the evolutionary history of the two copies of traG/trb operon, the phylogenetic relationships of traG/trb operon homologous were constructed with various genera in the rhizobiaceae family, including Bradyrhizobium, Rhizobium, Sinorhizobium, Mesorhizobium, and Agrobacterium. Cuprividus taiwanensis LMG19424 was used as the outgroup strain. The phylogenetic tree could be divided into two clades (Fig. 1b). In clade 1 Bradyrhizobium, T4SS structural genes were genetically closely related to their homologous Mesorhizobium members. Both Bradyrhizobium sp. SUTN9-2 T4SS₁ and T4SS₂ were also located in this clade. Interestingly, at least one copy of the T4SS structural genes were found in bradyrhizobial strains, but Bradyrhizobium sp. SUTN9-2 contains two copies, which is unusual in Bradyrhizobium. Copy 1 showed similar sequences with B. yuanmingense BRP09 (copy 3), closely related to B. japonicum, B. diazoefficiens, and B. elkanii USDA61 (B1 clade). B. elkanii SEMIA938 and B. yuanmingense BRP09 (copy 1) were closely related to copy 2. Rhizobium, Sinorhizobium, and Agrobacterium members were included in cluster 2. Of note, Bradyrhizobium and Mesorhizobium were separated from other genera in the rhizobiaceae family, which implied that both T4SS of bradyrhizobia and mesorhizobia may have originated from the same ancestor. Thus, we hypothesized that the T4SS gene in bradyrhizobia evolved together with mesorhizobia upon horizontal gene transfer.

The phylogenetic tree of bradyrhizobia regarding copG, traG, and virD2.

Phylogenetic tree analysis of concatenated sequences of T4SS structural gene (traG) and T4SS-related genes (copG and virD2) of Bradyrhizobium sp. SUTN9-2 was constructed and compared with other T4SS-containing Bradyrhizobium strains. The phylogenetic trees of each traG, copG, and virD2 combination gene was generated using the same set of strains (Fig. 2). The T4SS of Bradyrhizobium sp. SUTN9-2 was identified in two clusters on the chromosome, and two copies of each gene were analyzed. As expected, the phylogenetic trees based on sequences of the traG, copG, and virD2 combination gene showed similar topologies. The phylogenies were separated into two groups with different copies of the T4SS gene of Bradyrhizobium sp. SUTN9-2. The traG, copG, and virD2 in copy 1 revealed a close evolutionary relationship with similar genes found in B. yuanmingense BRP09 (copy 3), B. japonicum E109, B. japonicum J5, B. elkanii USDA61, B. diazoefficiens USDA110, USDA122, and SEMIA5080. This copy was closely related to B. yuanmingense BRP09 (copy 3) with high bootstrap values. Copy 2 was closer to Bradyrhizobium sp. MOS004, B. yuanmingense P10 130, B. elkanii SEMIA938, Bradyrhizobium sp. Rc2d, B. ottawaense L2, and closely similar to B. yuanmingense BRP09 (copy 1). These results reflect that Bradyrhizobium has higher T4SS gene sequence similarity among Bradyrhizobium than other genera in the rhizobia group.

FIG 2.

Phylogenetic relationship of copG, traG and virD2 genes cluster. The tree was constructed by the maximum-likelihood method using MEGA10 software. The accession numbers are shown in brackets. Mesorhizobium oppotunistum WSM2075 was used as an outgroup. Bars: 0.10 estimated substitutions.

Bradyrhizobium sp. SUTN9-2 contains chromosomal integrative conjugative elements (ICEs).

This study aimed to identify the putative ICEs on chromosome of Bradyrhizobium sp. SUTN9-2.

The 243 proteins were explored in putative ICE, which contains integrase, T2SS, T4SS, transposase, and relaxase genes, among others. The putative ICE was located between nucleotides 744 and 167,491 (length 166,748 bp) with a GC content of approximately 59.45%. The putative ICE was located between 774 and 167,491 and its insertion was facilitated by the site-specific DNA recombinase (SUTN92_v1_10001). The ICE was flanked with the conserved attachment (att) site, including attL (location: 774-758) and attR (location: 167,477-167,491) (Fig. 3), Interestingly, the transposase, relaxase, and integrase were not observed T4SS₂ area. In addition, the putative ICE with different patterns and numbers of elements were observed in other bradyrhizobia, including B. yuanmingense CCBAU10071, B. yuanmingense CCBAU05623, B. yuanmingense 3051, B. diazoefficiens USDA 110, Bradyrhizobium sp. TSA, Bradyrhizobium sp. BTAi1, and Bradyrhizobium sp. OSR285 (Table S1).

FIG 3.

Integrative and conjugative element (ICE) identification on chromosome of Bradyrhizobium sp. SUTN9-2.

The T4SS of Bradyrhizobium sp. SUTN9-2 is involved in symbiotic interaction.

To examine the pivotal role of the T4SS in promoting nodulation in various leguminous plants, the nodulation-related organogenesis organs were determined at 7,14, and 21 dpi. The nodule formation of Shyleaf (original host) by Bradyrhizobium sp. SUTN9-2 was observed in the early stage of the nodulation process (7 dpi), while ΔT4SS₁ formed nodule-like structures (Fig. 4a), and the nitrogenase activity was lower than WT. In the case of mung bean, the plant inoculated with ΔT4SS₁ showed drastically retarded nodulation (primordia formation) at 7 dpi (Fig. 5a), and nodule formation was recovered at 14 dpi (Fig. 5b). However, the nodule number and nitrogenase activity was decreased compared with WT. As in mung bean, brown hemp generated nodules were quantified after inoculation with ΔT4SS₁ at 14 dpi (Fig. 6b). The WT generated the highest nodule number while ΔT4SS₁ showed lower numbers of nodules; the nitrogenase activity level at 21 dpi was not much different among all treatments (Fig. 6i).

FIG 4.

The symbiotic phenotype of ΔT4SS₁ complementation (T4SS_1compl). Plant growth, nodule phenotype, and bacteroid viability determined by live/dead staining of nodule section and confocal microscopy (a, b, c) nodule number (d, e, f) and nitrogenase activity (g, h, i) 7, 14, and 21 days postinoculation (dpi) of Aechynomen American (Thai). Scale bar white = 200 μm (10×) and yellow = 50 μm (40×). Error bar, mean (n = 5) ± standard deviation. Different letters above bars indicate significant differences (P < 0.05) using Tukey’s test.

FIG 5.

The symbiotic phenotype of ΔT4SS₁ complementation (T4SS_1compl). Plant growth, nodule phenotype, and bacteroid viability determined by live/dead staining of nodule section and confocal microscopy (a, b, c), nodule number (d, e, f), and nitrogenase activity (g, h, i) 7, 14 and 21 days postinoculation (dpi) of Vigna radiata cv.SUT4. Scale bar white = 200 μm (10×) and yellow = 50 μm (40×). Error bar, mean (n = 5) ± standard deviation.

FIG 6.

The symbiotic phenotype of ΔT4SS₁ complementation (T4SS_1compl). Plant growth, nodule phenotype, and bacteroid viability determined by live/dead staining of nodule section and confocal microscopy (a, b, c) nodule number (d, e, f) and nitrogenase activity (g, h, i) 7, 14, and 21 days postinoculation (dpi) of Crotalaria juncea. Scale bar white = 200 μm (10×) and yellow = 50 μm (40×). Error bar, mean (n = 5) ± standard deviation.

Interestingly, the numbers of nodules and ARA activity were decreased under inoculation of ΔT4SS₁ compared to the WT strain in all plants tested. The dead bacteroid cells of ΔT4SS₁ in the middle symbiosome at 7 dpi were observed, and the bacterial infection was lower than WT. Nevertheless, the derivative bacteroides derived by ΔT4SS₁ began to multiply and increase live cells at 14 dpi. This result demonstrated that T4SS₁ represents a strong positive effect on nodulation, nitrogen fixation, and cell survival.

Previous results showed ΔT4SS₁ decreased both nodule numbers and nitrogenase activity in plants tested. To ensure the function of T4SS₁, the T4SS₁ complementation (T4SS_1compl) was constructed and inoculated into legume plants (Fig. 4, 5, and 6). Nodulation efficiency was restored by T4SS_1compl, resembling those observed in WT depending on the host plant. In shyleaf at 7 dpi (Fig. 4a), nodules were observed, and the nodule number of T4SS_1compl was similar to WT. At 14 (Fig. 4b) and 21 dpi (Fig. 4c), nodule numbers descending order as follows WT, T4SS_1comp, and ΔT4SS₁. However, the nitrogenase activity of T4SS_1compl was detected at 14 dpi (Fig. 4h), which is near to WT, while the activity of nitrogenase was lower than WT at 21 dpi (Fig. 4i). Young nodules were formed on mung bean inoculated with T4SS_1compl at 7 dpi (Fig. 5a), and the nodule number was higher than that of plants nodulated by ΔT4SS₁ but lower than WT. The efficiency of nitrogenase activity was decreased in this complemented strain. It was the same trend at 7, 14, and 21 dpi (Fig. 5g to i). As in shyleaf and mung bean, nodule formation of T4SS_1compl was observed at 7 dpi in brown hemp, while ΔT4SS₁ formed the pseudonodules (Fig. 6a). Nitrogenase activity was detected at 14 dpi (Fig. 6h), which was higher than ΔT4SS₁ but still lower than WT. However, at 21 dpi (Fig. 6i), the nitrogenase activity level was the lowest. Since the T4SS_1compl generates effective nodules on the host plant similar to WT, the ARA activity of each legume corresponded to the nodule phenotype. Confocal microscopy was performed to detect live/dead cells of bacteroids inside nodules. The results showed more living cells of T4SS_1compl than dead cells in symbiosome compared with ΔT4SS₁ at 7 dpi (Fig. 4a, 5a, and 6a). Dead cells were present in mature nodules of T4SS_1compl but still less than in T4SS₁ inoculation. In order to determine whether the secretion occur in the beginning when Nod factor is produced or later time when nodule is developed. The gene expression of nodulation genes (nodA and nodC) and T4SS gene (traG) in Bradyrhizobium sp. SUTN9-2 was conducted. The result showed that nod genes were expressed under genistein induction at 24 h (Fig. S2a and b). However, traG was not induced under genistein induction but it was also expressed at 24 h (Fig. S2c). In contrast, the expression of nod genes in ΔT4SS₁ was less expressed than those of WT under genistein induction (Fig. S2a and b). Therefore, it confirms that both nod and T4SS genes were expressed at the early stage of nodulation. These results again indicate that T4SS in Bradyrhizobium sp. SUTN9-2 plays an important role in symbiotic interaction.

DISCUSSION

The genes encoding T3SS and T4SS of Bradyrhizobium sp. SUTN9-2 are located on the chromosome, whereas T6SS components were unidentified. We further investigated the role of T4SS of in Bradyrhizobium sp. SUTN9-2 controlling symbiosis with various leguminous plants. Typically, the pTi plasmid of Agrobacterium is the best study model for T4SS (19). The pTiC58 plasmid of A. tumefaciens C58 contains three types of T4SS genes, vir, avh, and trb, which are required for effector protein translocation and DNA conjugal transfer (5, 20). The virB/D4 and tra/trb operons are important for protein transfer and conjugation processes that enable translocating protein, DNA, or protein DNA complex into eukaryotic host cells (8, 21, 22). These operons are associated with symbiotic interaction effector protein translocation, limiting immune response and host specificity in eukaryotic host plants (23). Besides the Agrobacterium genus, T4SS has been identified in some rhizobia (4, 10, 14–16). Moreover, the effector proteins of T4SS have been reported in some rhizobia that represent the effect on symbiosis, for example, the Msi059 and Msi061effector proteins of M. loti R7A (10). Likewise, TfeA is a similar protein in Sinorhizobium spp. that exhibits 59% amino acid homology with the Msi061 protein from M. loti R7A and 25% identity with VirF protein of A. tumefaciens C58 (11). Previous works reported that the T4SS belonging to virB/D4 operon was regulated by NodD, which is the main regulator of nod gene expression (10, 24). This assumes that T4SS is associated with the early steps of the legume-rhizobia symbioses (2, 25), but nodD expression was not observed on the tra/trb operon (26). The tra/trb operon was also observed on the pTiC58 plasmid of A. tumefaciens C58 (27), and the gene organization was closely similar to tra/trb operon on a symbiotic plasmid (pSym) of Rhizobium and Sinorhizobium. The trbH and trbK genes were not observed on the T4SS cluster of Mesorhizobium, and trbH was also not found in Bradyrhizobium (Fig. 1a). Moreover, several Bradyrhizobium possess T4SS in the tra/trb operon on the chromosome (28). Bradyrhizobium with the tra/trb operon is often found on a broad range of host strains, such as Bradyrhizobium sp. SUTN9-2 (17), B. diazofficiens USDA110 (13), B. japonicum CPAC 15, and B. diazoefficiens CPAC 7 (29). In contrast, virB/D4 operon was observed only in the plasmid of Bradyrhizobium strains without virD2 (28), such as Bradyrhizobium sp. BTAi1 (15) and Bradyrhizobium sp. DOA9 (14).

Interestingly, the gene arrangement of T4SS structural (traG) and T4SS-related genes (copG and virD2) presented a unique characteristic of the Bradyrhizobium genus (Fig. 1a). Although the gene arrangement of the T4SS structural gene of Bradyrhizobium sp. SUTN9-2 are conserved well with Bradyrhizobia, Mesorhizobia, Rhizobia, Sinorhizobia, and Agrobacteria, but T4SS-related genes (copG and virD2) represent a unique characteristic of Bradyrhizobium genes (Fig. 1a). However, copG and virD2 genes arrangement inside the T4SS cluster was slightly diverse among bradyrhizobial group. The T4SS apparatus appears to have the same evolutionary origin as other rhizobia. Therefore, a unique T4SS gene organization (copG, traG, and virD2) may have evolved independently.

Likewise, the phylogenetic tree of copG, traG, and virD2 gene cluster indicated that the coevolution of T4SS structural and T4SS-related genes might be specific to bradyrhizobia (Fig. 2). Based on 16s rRNA gene sequence similarity, Bradyrhizobium sp. SUTN9-2 was phylogenetically closely related to B. yuanmingense (16, 30). Our results from the phylogenetic analysis of T4SS (copG, traG, and virD2) in copy 1 were mainly congruent with the phylogenetic tree relationship of the 16s rRNA gene. T4SS₁ was similar to B. yuanmingense BRP09 (copy 3) while T4SS₂ was similar to B. yuanmingense BRP09 (copy 1) with different clade. The Mesorhizobium sp. traG gene is a substrate receptor for T4SS expressed in early nodule formation generated by M. mediterraneum Ca36^T. Perhaps, the T4SS belonging to tra/trb operon is important for the mutualistic and symbiotic interaction (26). In addition, it was also found that the evolution of rhizobia involved lifestyle diversity that depended on bacterium-host interaction and eukaryotic host diversity. The genes associated with bacterium-host interactions, such as T3SS, T4SS, and T6SS, were derived in a lineage-specific form (31).

As in the Mesorhizobium genus, the traG gene could horizontally be transferred between strains because they could establish nodules with the same plant. The TraG protein is also associated with the efficiency of symbiosis gene transfer among strains in this genus (26). Based on the T4SS structural genes phylogenetic tree of the rhizobia group, including Bradyrhizobium, Mesorhizobium, Sinorhizobium, Rhizobium, and Agrobacterium demonstrated that the T4SS structural genes of Bradyrhizobium may have evolved from the same ancestor with Mesorhizobium, while Sinorhizobium and Rhizobium were closely similar with Agrobacterium (Fig. 1b). These results correspond to the hypothesis that Bradyrhizobium might be the ancestor of all rhizobia genera. This hypothesis has been confirmed by a molecular study, which concluded that Bradyrhizobium’s large genome acted as a broad host range strain responsible for effective nodules in several legumes (32). Moreover, Bradyrhizobium also have the ability to generate nodules on primitive nonlegume Parasponia plants (33). Similarly, Mesorhizobium has the ability to nodulate the chickpea, a primitive legume, and limit the plant host that restricts rhizobia infection (34). Thus, the presence of T4SS in Bradyrhizobium sp. SUTN9-2 could be a tool to overcome plant host inhibition and effectively nodulate and fix N₂ in several legumes.

Integrative and conjugative elements (ICEs) are mobile genetic elements that can integrate into the bacterial chromosome and transfer themselves to other bacteria through horizontal gene transfer. In the Gram-negative bacteria conjugation machinery encoded by T4SS (35), the ICEs consist of the core gene associated with conjugation and excision components (36). Moreover, the ICEs are able to spread large gene clusters, which provide competitive advantages to a particular group (37). Normally, the symbiosis genes were observed on a large symbiotic plasmid of Rhizobium and Sinorhizobium, while in Mesorhizobium, Azorhizobium and Bradyrhizobium were found in ICEs called ICEsym (38). In rhizobia species, ICEs were first described in Mesorhizobium spp. carried genes required for symbiosis, such as M. loti R7A, was explored ICEMlSym^R7A. ICEMlSym^R7A was transferred to nonsymbiotic mesorhizobia in New Zealand soils and can convert nonsymbiotic into symbiotic Lotus corniculatus (39, 40). In this study, we identified one putative ICE on a chromosome of Bradyrhizobium sp. SUTN9-2 (ICE^SUTN9-2) with low CG content containing 243 proteins, such as T2SS, T4SS, relaxase, transposase, and integrase proteins, essential for gene transfer (Fig. 3). Interestingly, T4SS₁ was a part of ICE^SUTN9-2 which is a core conjugative element. At least one of the large ICEs was found in the agent of bradyrhizobia species (Table S1). Based on copG, traG, and virD2 gene cluster phylogenetic tree (Fig. 2), B. yuanmingense was closely similar to Bradyrhizobium sp. SUTN9-2 and ICEs in a similar pattern were found on the chromosomes of both strains. Therefore, these results demonstrated that the horizontal gene transfer of Bradyrhizobium sp. SUTN9-2 is most likely performed by an ICE specific to B. yuanmingense group. Therefore, T4SS₁ was transferred by horizontal gene transfer together with ICE.

Plants inoculated with Bradyrhizobium sp. SUTN9-2 ΔT4SS₁ at 7 dpi showed delayed nodulations in mung bean and brown hemp while shyleaf formed small nodules on root (Fig. 4 to 6). This displayed similar patterns as in the T4SS (vir cluster) mutant of M. loti R7A with L. corniculatus compared with WT that consist of the secreted proteins facilitating host infection (10). In addition, the dead cells in the ΔT4SS₁ symbiosome were observed even in the original host, indicating the survival of cells within host plants associated with T4SS Bradyrhizobium sp. SUTN9-2. Although, when ΔT4SS₁ was restored, nodules were generated after 7 dpi but the nitrogenase activity was lower than WT. T4SS of M. loti, which suggests a symbiotic relationship with the eukaryotic host. Moreover, the M. loti R7A T4SS mutant positively affected L. corniculatus and negatively affected L. leucocephala (10). Taken together, these findings support that T4SS₁ had a positive effect via symbiotic interactions with legume plants. However, the effect of T4SS on symbiosis depends on the species of the host plant. Moreover, T4SS of Bradyrhizobium sp. SUTN9-2 may affect cells' survival during infection, which is important for host specificity of the symbiotic interaction. As in M. mediterraneum Ca36^T, the traG gene was expressed in early infection and symbiotic interaction (10).

The different gene organizations of T4SS₁ and T4SS₂ were explored. The T4SS₁, trb operon, and traG genes were inserted with ~56 kb fragment consisting of 84 genes (Fig. S1), such as putative T2SS, transposase, peptidase, permease, tyrosine-type recombinase/integrase, and hypothetical proteins (Table S1). In addition, the replication A (repA) gene was found in the upstream copG, traG, and virD2 cluster, while T4SS₂ contains complete tra/trb operon gene organization. Interestingly, a lower GC content and fewer transposase elements were found in the T4SS clusters, especially in the T4SS₁ cluster. This is one possible reason for the variation of genes in this region being more diverse than other regions and making different gene organization of both T4SS. To ensure the efficiency of T4SS₁ on symbiosis, we constructed T4SS₁ complement strain (T4SS_1compl) and tested it as in previous experiments (Fig. 4 to 6). This result confirmed that, besides nod factors and T3SS, Bradyrhizobium sp. SUTN9-2 used the T4SS pathway to establish symbiosis with several host plants. However, many aspects of T4SS’s involvement in symbiotic interactions remain unknown. For example, the function of the copG, traG, and virD2 genes warrants further investigation in legumes.

Conclusions.

Prior to our understanding the role of T4SS in Bradyrhizobium-legume symbiois, phylogenetic trees of the gene involved in T4SS were constructed. A unique arrangement of the copG, traG, and virD2 gene cluster was found in T4SS specific to the Bradyrhizobium group. The phylogenetic trees implied that the horizontal transfer of T4SS genes was restricted to the rhizobia group. The symbiotic interactions between Bradyrhizobium sp. SUTN9-2 and legume plants were performed using T4SS deletion strains. The results showed ΔT4SS₁ reduced nodule number, nitrogenase activity, and delayed nodulation in V. radiata cv. SUT4 and C. juncea but not in A. americana. In addition, dead cells were observed during the early stage of infection with ΔT4SS₁. The result indicated that T4SS may be important for cell survival within plant host that associated with host specificity. This is the first report to demonstrate that the T4SS of Bradyrhizobium sp. SUTN9-2 represents another mechanism for symbiotic interaction with leguminous plants.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The bacterial strains used in this study are listed in Table 1. Bradyrhizobium sp. SUTN9-2 was cultured on Arabinose-Gluconate (AG) medium (41) at 28°C, and Escherichia coli strains were cultured in Luria-Bertani (LB) medium at 37°C. The media were supplemented with antibiotics at following concentrations (in μg/mL): streptomycin (Sm), 200, nalidixic acid (Nal), 20, kanamycin (Km), 50 and cefotaxime (Cefo), 20.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristics	Reference or source
Strain
Bradyrhizobium sp.
SUTN9-2	A. americana nodule isolate (paddy crop)	(16)
ΔT4SS1	SUTN9-2 derivative, T4SS1 (copG, traG, and virD2)::Cefo; Cefor	This study
Escherichia coli
DH5α	supE44 ΔlacU169 hsdR17 recA1 endA1 gyrA96 thi-1 relA1	Toyobo Inc.
Plasmid
pRK2013	ColE1 replicon carrying RK2 transfer genes; Kmr; Helper plasmid	(43)
pNTPS129	Cloning vector harboring sacB gene under the control of the constitutive npt2 promoter; Kmr	(42)
pNTPS129-ΔT4SS1	pNTPS129-npt2-sacB containing the flanking region of T4SS copy 1	This study
pMG103-npt2-cefo	Complementation vector harbors a cefotaxime resistance gene under the control of the constitutive npt2 promoter; Cefor	Giraud E, IRD, France
pMG103-npt2-sm/sp	Complementation vector harbors a Ω cassette; Smr, Spr resistant genes under the control of constitutive npt2 promoter; Smr, Spr modified from pMG103-npt2-cefo (Giraud E, IRD, France), insert Ω cassette at HindIII site	This study
pMG103-npt2-sm/sp-T4SS1	pMG103-npt2-sm/sp-npt2-gfp carring 5,069 bp T4SS copy1 fragment; Smr, Spr	This study

Plasmid construction, gene deletion, and complementation.

All primers used in this study are listed in Table 2. The scheme for T4SS mutation and complementation of Bradyrhizobium sp. SUTN9-2 is displayed in Fig. 7. The construction of T4SS deletion mutants was achieved by double-crossing over. Briefly, the fragments corresponding to the upstream and downstream flanking region of T4SS₁ (up-1049 bp and down-804 bp) were amplified and merged by overlapped extension PCR and cloned into the pNTPS129 plasmid harboring the sacB gene (42), namely, pNTPS129-T4SS₁. Then, a cefotaxime and streptomycin/spectinomycin resistance cassette was introduced between the upstream and downstream flanking regions at the HindIII and BamHI sites of pNTPS129-T4SS₁. The resulting plasmids were transferred into Bradyrhizobium sp. SUTN9-2 via triparental mating using pRK2013 as a helper plasmid (43). The single recombinant mutants were selected on AG agar supplemented with Nal, 20, Cefo, 20 and verified by PCR using two primer pairs amplifying the upstream and downstream flanking target genes. The double-crossed mutants were selected on AG medium containing 10% sucrose supplemented with Cefo, 20 for T4SS₁. The transconjugant colonies appearing on selective media were verified by PCR using the primers described in Fig. S1. The mutant was screened on antibiotic plates (Cefo, 20) and verified by PCR (Table 2). The complementation of ΔT4SS₁ was performed by introduced a 5,069-bp fragment contained T4SS₁ and its candidate promoter into pMG103-npt2-Sp/Sm at the XbaI and EcoRI sites. The constructed plasmid was transferred into ΔT4SS₁ mutant using triparental mating and the transconjugant was confirmed by antibiotic resistant and PCR verification. The derivative mutant and complementation strains were further investigated for nodulation efficiency in various legumes.

TABLE 2.

Primers and cloning strategies in this study

Target genes	Primer names	Primer sequence (5′→3′)	References
Deletion
Upstream fragment of the ΔT4SS1	Up.F.T4SS1.XbaI	ACC CAG TCT AGA GAT AAC GCT CGA CCA ACT CTC A	This study
	Up.R.T4SS1.overl	CTG CGT AAG TTC GAA GCT TTC TCT TCA TCC GCT TCT GGC T	This study
Downstream fragment of the ΔT4SS1	Dw.F.T4SS1.overl	GCG GAT GAA GAG AAA GCT TCG AAC TTA CGC AGT TCG AC	This study
	Dw.R.T4SS1.BamHI	CAG AAT GGG ATC CAT CAT TTC CGC TTC ATC GTC TC	This study
Complementation
T4SS1 complementation (T4SS1comp)	Compl.T4SS1.XbaI.F	CGC TGG TCT AGA ATC AGA TCC TCC GTC GCT GCT	This study
	Compl.T4SS1.EcoRI.R	ACG ACA GGA ATT CTC GCA GCC ATC GTC CCT TT	This study
Gene expression
Nodulation (nod) gene	S9-2.nodA.F	GTT CAA TGC GCA GCC CTT TGA G	This study
	S9-2.nodA.R	ATT CCG AGT CCT TCG AGA TCC G	This study
	S9-2.nodC.F	ATT GGC TCG CGT GCA ACG AAG A	This study
	S9-2.nodC.R	AAT CAC TCG GCT TCC CAC GGA A	This study
Structural T4SS1 gene	S9-2.traG.F	TTC TCG ATC TGG TTC AGC GAC TG	This study
	S9-2.traG.R	TTG ACC GAG GAT CTT CAG GCC A	This study

FIG 7.

Scheme of T4SS mutation of Bradyrhizobium sp. SUTN9-2. The gray area represents the fragment of Δ T4SS₁ (copy 1), and the dark blue bars under copG, traG, and virD2 genes show the amplified region to complementation ΔT4SS₁ mutant. The candidate promoter of T4SS copy 1 is indicated with a red arrow.

Nodulation test and acetylene reduction assay (ARA).

The legume plants used in this study include Aeschhynomene americana (Shyleaf), Vigna radiata cv. SUT4 (Mung bean), and Crotalaria juncea (Brown hemp) for Dalbergioids, Millettiods, and Genistoids, respectively. All seeds were sterilized as previously described (44), and then rinsed and soaked with sterilized water overnight at room temperature. The germinated seeds were placed on 0.85% water agar plates and incubated at 28°C for 1 to 2 days. After 1 week, seedlings were transferred into Leonard’s jars containing sterilized vermiculite and inoculated with 1 mL bacterial cultures adjusted with sterilized distilled water to OD₆₀₀ = 0.8. Plants were harvested 7, 14, and 21 days postinoculation (dpi) for nodule number evaluation and Acetylene Reduction Assay (ARA). The nodules were incubated in 50 mL (Mung bean and Brown hemp) and 20 mL (Shyleaf) test tubes. Then, the air was withdrawn and replaced with pure acetylene at 10% (vol/vol). The samples were incubated for 1 h at 25°C. The conversion of acetylene (C₂H₂) to ethylene (C₂H₄) was measured by injecting 1 mL into Gas Chromatography (GC) using a PE-alumina-packed column at an injection temperature of 150°C and oven temperatures of 200°C and 50°C for flame ionization detection (FID) (45). The nitrogenase activity is presented as nmol ethylene/h/plant dry weight (46).

Microscopy.

Nodule phenotypes were investigated under a stereomicroscope (LEIGA EZ4). The Nikon Inverted Eclipse Ti-E Confocal Laser Scanning Microscope investigated nodule development and live/dead cell bacteroids in symbiosomes. For this purpose, nodules were embedded in agarose (5%) and then sectioned at 40 to 50 μm by VT1000S vibratome (Leica Nanterre, France). Nodule sections were imaged under a compound microscope (CARL ZEISS/PRIMO STAR HD) and then stained with fluorophore SYTO9 (5 μM)/PI (30 μM) of live/dead cells staining for 30 min and stained with 1× calcofluor white stain in 1× PBS buffer for 20 min. Calcofluor was detected at 460 to 500 nm emission, while SYTO9 and PI were detected at 510 to 570 nm and 600 to 650 nm, respectively.

Gene annotation, phylogenetic tree construction, CGView map, and integrative conjugative element (ICE) analysis.

Two phylogenetic trees, including traG/trb operon and copG, traG and virD2 genes cluster were constructed by the Maximum Likelihood method at a confidence level of 1,000 replicates using MEGA X,10.0.5 versions (47). Bradyrhizobium sp. SUTN9-2 genomes sequence and other bacterial strains were retrieved from the NCBI database (https://www.ncbi.nlm.nih.gov) and Genoscope (https://mage.genoscope.cns.fr). The circular CGView map of Bradyrhizobium sp. SUTN9-2 was generated by Proksee Genome Analysis (https://proksee.ca) (48). Integrative and conjugative elements (ICEs) were identified with ICEfinder online program (https://bioinfo-mml.sjtu.edu.cn/ICEfinder/ICEfinder.html) (35).

Bacterial induction.

Midlog-phase bacterial cultures of SUTN9-2 (WT), and ΔT4SS₁ strains (OD₆₀₀ = 0.4) were induced by 20 μM genistein (Sigma–Aldrich, USA) at 28°C for 24 h. Bacterial cells were collected by centrifugation (4,000 × g, at 4°C) for further total RNA isolation.

RNA isolation and quantification of gene expression by qRT–PCR.

Total bacterial RNA was extracted from cells using the FavorPrep Tissue Total RNA minikit (Favorgen, Hong Kong), according to the manufacture’s instruction. Total RNA was treated with RNase-free DNase I (NEB, USA) for 30 min at 37°C. Complementary DNA (cDNA) was synthesized from 500 ng of total RNA using iScript reverse transcription Supermix (Bio–Rad, USA) according to the manufacturers’ protocols. Fifty nanograms of cDNA were subjected to real-time PCR using specific primers for N-acyltransferase nodA, N-acetylglucosaminyltransferase nodC and conjugal transfer gene traG. All primer sets used in the expression analysis are listed in Table 2.

For qRT–PCR, cDNA from each sample was briefly mixed with Luna Universal qPCR Master Mix (NEB, USA) following the manufacturer’s protocol, and thermal cycling was conducted in a QuantStudio 3 real-time PCR system (Thermo Fisher, USA). qPCR amplification was performed under the following cycling conditions: hold stage: 95°C for 2 min; PCR stage: 40 cycles of 95°C for 15 s and 60°C for 30 s; melting curve stage: 95°C for 15 s and 60°C for 1 min; and a final step of 95°C for 15 s. Relative gene expression was analyzed by the comparative Ct method (-ΔΔCT) and normalized to the endogenous 16S rRNA housekeeping gene. Three biological replicates were pooled and analyzed. At least three replicates of PCR amplification were performed for each sample.

Statistical analysis.

For statistical analyses, one-way analysis of variance (ANOVA) followed by post hoc tests (Tukey’s tests at P ≤ 0.05) were performed using IBM SPSS Statistics 22.0 software.