CopG₁, a Novel Transcriptional Regulator Affecting Symbiosis in Bradyrhizobium sp. SUTN9-2

1. Introduction

A complex exchange of signals between legumes and rhizobia bacteria is the key factor in the successful symbiosis between two partners. The establishment of nitrogen-fixing nodules is ensured by this intricate communication system, which is crucial for legumes to thrive in nitrogen-limited environments [1,2]. Two crucial mechanisms are necessary for the nodulation process of many rhizobia: (i) the classical mechanism of perception of the Nod factor (NF) and (ii) the type III secretion system (T3SS) which deliver effector proteins into plant cells [3,4]. NFs are well-studied while, we are just beginning to understand the roles of effectors/secretion systems. The secreted effector proteins from T3SS (T3Es) can have either a positive or negative impact on the symbiosis efficiency, depending on the plant species [5]. For example, Bradyrhizobium vignae ORS3257 contains multiple effector proteins crucial for modulating symbiotic properties in different Vigna species. NopT and NopAB play essential roles in nodulation in V. unguiculata and V. mungo. Whereas, NopP2 displayed incompatibility with V. radiata [6]. Beside the T3SS, others secretion systems have been identified in several rhizobia genera, including type IV and VI secretion systems (T4SS and T6SS) and these systems have also reported as the key determinant to the symbiotic interactions during infection processes. For instance, the T6SS of Rhizobium etli Mim1 and Bradyrhizobium sp. LmicA16 (A16) exhibit a positive effect on nodulation with their host [8,9]. In the case of T4SS, this secretion machinery is found in various bacteria, including Rhizobium. It functions as a molecular channel, allowing bacteria to transport diverse molecules across their cell envelope [10]. Various rhizobia, including Bradyrhizobium, Rhizobium, Sinorhizobium, and Mesorhizobium, possess T4SS which belongs to the tra/trb operon and could be found either in chromosome or plasmid depending on the bacterial strain [11,12,13,14]. Interestingly, most of bradyrhizobia harbor the tra/trb operon on the chromosome. Few studies about the role of T4SS of Bradyrhizobium in the symbiotic process have been reported. Previously, we found two clusters of T4SS (T4SS₁ and T4SS₂) located on the chromosome of Bradyrhizobium sp. SUTN9-2 with different gene arrangements (Figure S1). Additionally, a specialized gene arrangement consisting of copG, traG, and virD2 genes was observed in the T4SS gene cluster of Bradyrhizobium species. The T4SS evolutionary analysis of the Rhizobiales order, encompassing Bradyrhizobium, Rhizobium, Sinorhizobium, and Mesorhizobium through the traG gene phylogenetic tree, demonstrates a co-evolutionary trend between Bradyrhizobium and Mesorhizobium. Upon phylogenetic examination of the copG, traG, and virD2 combination genes in Bradyrhizobium, two copies of these clusters were divided into two clades within the bradyrhizobia group [15]. Moreover, copy 1 of copG, traG, and virD2 genes exhibits a close evolutionary association with B. yuanmingense BRP09, which are main rhizobia associated with cowpea and mung bean in the subtropical region of China [16], as well as with B. diazoefficiens USDA110, soybean inoculant [17]. Interestingly, T4SS₁ mutant (deleted copG₁, traG₁, and virD2₁ fragment) retard nodulation in V. radiata cv. SUT4 and Crotalaria juncea and both the number of nodules and nitrogenase activity were decreased compared with the wild type. The results indicated that T4SS₁ has a positive effect on the symbiotic interaction with tested plants [15]. To understand the role of T4SS₁ in the symbiotic process, individual deletion mutants were generated in copG₁, traG₁ and virD2₁ genes (encoding a putative transcriptional factor gene, T4SS structural gene, and relaxase gene, respectively), and examined for their roles in the symbiotic interaction with V. radiata cv. SUT4 [18,19,20]. In addition to V. radiata cv. SUT4, the loss of nodulation ability was observed in different plant species of Genistoids, Dalbergioids, and Millettioids upon infection with ΔcopG₁ (Table S1).

2. Materials and Methods

2.1. Bacterial Strains, Plasmids, and Growth Conditions

The bacterial strains and plasmids used in this study are listed in Table 1. Bradyrhizobium sp. SUTN9-2 was grown in arabinose-gluconate (AG) medium [21]. The derivative mutants ΔcopG₁, ΔtraG₁, and ΔvirD2₁ were supplemented with streptomycin (sm) at 200 μg/mL. Escherichia coli strains were grown at 37°C in Luria–Bertani (LB) medium. Antibiotic was added in to the medium as required at the following concentrations: 50 μg/mL kanamycin (km), 30 μg/mL nalidixic acid (nal), and 200 μg/mL streptomycin (sm).

2.2. Plasmid Construction and Gene Deletion

The deletion mutants of copG₁, copG₂, traG₁, and virD2₁ genes in Bradyrhizobium sp. SUTN9-2 (GeneBank accession number LAXE00000001) were obtained as follows. The upstream and downstream regions of copG₁ (up: 575 bp, dw: 841 bp), copG₂ (up: 874 bp, dw:1043 bp), traG₁ (up: 1060 bp, dw:921 bp), and virD2₁ (up: 944 bp, dw: 738 bp) genes were obtained by PCR using primers listed in Table 2. The target deletion genes in SUTN9-2 were obtained by double crossover. PCR fragments corresponding to the upstream and downstream flanking regions of the gene of interest were merged by overlap extension and introduced into a pNTPS129 plasmid harboring the sacB gene [23]. Then, an Ω cassette fragment (spectinomycin/streptomycin resistance genes) from pHP45 (omega) [24] was introduced between upstream and downstream flanking regions which were already cloned into pNTPS129. The restriction site for antibiotic insertion was HindIII for copG₁ and BamHI for copG₂, traG₁, and virD2₁. The resulting plasmids were transferred into SUTN9-2 by triparental mating using pRK2013 as a helper plasmid [22] as described previously [15]. A single recombinant clone was obtained by antibiotic selection and PCR verification. Double recombinant clones were selected by culture on AG medium supplemented with 10% sucrose and 200 μg/mL sm. Candidate clones were verified for the loss of the sacB gene from pNTPS129, and the replacement of the Ω cassette was verified by PCR. All mutant strains were further investigated for nodulation efficiency in V. radiata cv. SUT4.

2.3. Nodulation Test and Acetylene Reduction Assay (ARA)

V. radiata cv. SUT4 seeds were surface-sterilized and germinated as previously described [26] and placed on 0.85% water agar at 28°C overnight. One-day-old germinated seedlings were transferred into Leonard’s jars containing sterilized vermiculite and liquid buffered nodulation media (BNM) [27]. At seven days after gemination, seedlings were inoculated with a bacterial suspension of Bradyrhizobium sp. SUTN9-2 or derivative mutants (1 ml per seedling; adjusted to OD₆₀₀ = 0.8). Five plants per treatment were selected for nodule counting. The symbiotic phenotypes and nitrogen activity were measured at 7, 14, and 21 dpi.

Acetylene reduction assays (ARAs) were used to evaluate the nitrogenase activity. The root samples were transferred into test tubes, which were closed with a plastic stopper. Then, the samples were incubated with 10% (v/v) pure acetylene instead of air, which was withdrawn for 1 h at room temperature. A 1 ml sample was examined using gas chromatography (GC) with a PE-alumina-packed column to measure the conversion of acetylene (C₂H₂) to ethylene (C₂H₄). Detection was done at an injection temperature of 150°C and oven temperatures of 200°C and 50°C for flame ionization detection (FID) [28]. The experiment was conducted with five biological replicates per treatment. The nitrogenase activity is presented in nmol ethylene/h/plant dry weight [29].

The symbiotic profiles of ΔcopG₁ were compared to SUTN9-2 using a growth pouch test and several leguminous plants, including Genistoids (Crotalaria juncea), Dalbergioids (Aeschynomene americana cv. Thai, Arachis hypogaea cv. Thainan 9 and A. hypogaea cv. Khonkaen 5), and Millettioids (Indigofera tinctoria, Macroptilium atropurpureum, V. radiata cv. SUT1, V. radiata cv. CN72, V. radiata cv. KUML4, V. radiata cv. CN36, V. radiata cv. KPS1, V. mungo cv. U thong 2 and, V. subterranean). Seeds were surface-sterilized and germinated as previously described [26,30]. Pouches were prepared [29] and supplemented with BNM medium. Seedlings were grown at two plants per pouch (Five pouch replicates per treatment) and inoculated with 1 ml per plant of a suspension containing OD₆₀₀ = 0.8. Plants were grown under the conditions mentioned above.

2.4. Bacterial Induction, RNA Isolation, and qRT-PCR Analysis of Gene Expression

For bacterial induction, the mid-log phase of bacteria cultures including Bradyrhizobium sp. SUTN9-2 and copG₁ mutant strains (OD₆₀₀ = 0.4) were induced by 20 μM genistein at 28°C for 24 hours. Then, bacterial pellets were collected by centrifugation (4,000×g, at 4°C) for total RNA isolation. Total RNA was isolated from bacterial pellets using an RNeasy® Protect Cell Mini Kit (Qiagen, Chatsworth, CA, U.S.A.) according to the manufacturer’s instructions. Total RNA was treated at 37°C for 30 min with RNase-free DNase I (NEB, USA).

cDNA was synthesized using iScript™ Reverse Transcription Supermix for RT-qPCR (Bio-Rad Laboratories, Inc.) according to the manufacturers’ protocols. A cDNA concentration of 50 ng/µL was subjected to real-time PCR using specific primers (Table 2) for nodulation genes (nodA, nodB, nodC, nodD1, and nodD2), transcriptional regulator of T3SS (ttsI), T4SS structural genes (traG₁ and trbE₁), and other genes. qRT-PCR reactions were performed with Luna® Universal qPCR Master Mix (NEB, USA) by following the manufacturer’s protocol, and thermal cycling was conducted in a CFX Opus 96 Real-Time PCR System (Bio-Rad Laboratories, Inc.). The reactions were done in triplicate for each of the three biological replicates. Relative gene expression was analyzed by the comparative Ct method 2(-ΔΔCT), and 16s rRNA (accession number: JN578804) was used as internal control [11]. Three biological replicates were analyzed.

2.5. Protein Preparation, Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE) Analysis, and Protein Identification

Wild type (WT) and ΔcopG₁ Bradyrhizobium sp. SUTN9-2 were grown in AG medium with shaking at 200 rpm and 30°C until they reached an OD₆₀₀ of 1. One percent (v/v) of each starter was inoculated into 100 ml of AG medium with and without 20 µM genistein induction. Then, cultures were incubated at 30°C until they reached an OD₆₀₀ of approximately 1.0. The bacterial supernatants were harvested by centrifugation at 4,000xg and 4°C for 1 h, followed by 8,000xg for 30 min. One milliliter of 1 M dithiothreitol (DTT) and 7.5 ml of phenol solution (equilibrated with 10 mM Tris HCl at pH 8.0 with ESTA) were added into 25 ml of fresh supernatant. The solution was vigorously mixed with a vortex before centrifugation at 8,000xg and 4°C for 30 min. The water phase was discarded, and the phenol phase was added into another 25 ml of supernatant, followed by vigorous mixing and centrifugation at 8,000xg and 4°C for 30 min. Next, 20 ml of methanol containing 300 µl of 8 M ammonium acetate and 400 µl of 1 M dithiothreitol were added to remove the phenol phase. The secreted protein was precipitated for overnight at -20°C. The solution was then centrifuged at 8,000xg and 4°C for 1 h, and the supernatant was discarded. After precipitation, the protein was washed with chilled 70% (v/v) ethanol and air-dried in a laminar flow before being dissolved in phosphate-buffered saline (PBS). Protein concentrations were determined using a plate reader and the manufacturer’s protocol (PanReac) according to the method of Bradford [31]. A standard calibration curve was constructed using 0 to 2 µg of bovine serum albumin (BSA). Denaturing SDS-PAGE was carried out according to the method of Laemmli [32], in which 10 µg each lane of protein were analyzed on a 12% SDS-PAGE gel. The protein samples were mixed with loading buffer containing β-mercaptoethanol and heated for 10 min before loading. Protein bands were stained with colloidal Coomassie brilliant blue R-250 for visualization of the expression of secreted protein. The protein bands which were observed on WT lane but not observed on ∆T4SS₁ lane were cut protein identification by mass spectrometry. Breiftly, the protein bands were performed in-gel digestion by 12.5 ng/µL trypsin (mass spectrometry grade; Promega, USA). The extacted peptides were collected and dried in the Nitrogen Evaporator (Organomation, USA). Then, the peptide were reconstituted in 15 µl of 0.1% formic acid (FA) for LC/MS analysis. The LC–MS/MS system consists of a liquid chromatography part (Dionex Ultimate 3000, RSLCnano System) in combination with an captivespray ionization / mass spectrometer (Model Q-ToF Compact, Bruker, Germany) at the Proteomics Services, Faculty of Medical Technology, Mahidol University (Salaya Campus), Nakonpathum, Thailand. Mass spectral data from 300 to 1500 m/z was collected in the positive ionization mode. The most abundant peptide ions were analysed using MS/MS, to determine the peptide sequence. The peptide sequence was searched on the Uniprot database using the Mascot daemon version 2.6.0 (Matrix Science, London, UK) searching engine. The search parameters in Mascot daemon MS/MS Ions search included carbamidomethyl at cysteine residues as a fixed modification and oxidation on methionine as variable modifications. The peptide tolerance was set at ±1.6 Da, and the MS/MS fragment tolerance was set at ±0.8 Da. Protein hits were selected with a p-value of ≤ 0.05. The obtained results were examined against protein-NCBI database to identify and annotate proteins.

2.6. Microscopy

The nodule phenotypes and cross sections of representative nodules generated by the wild type (WT) or mutants were examined under stereomicroscope (LEIGA EZ4). For in-situ live/dead cell staining, the nodules were harvested and embedded with 5% agarose [33]. Three plants per treatment were selected for nodule sections with thickness of 40-50 µm were prepared with a VT1000S vibratome (Leica Nanterre, France) and incubated with live/dead staining solution (5 µM SYTO9 and 30 µM propidium iodide (PI) in PBS pH 7.0 buffer) for 30 min, followed by staining with 1X calcofluor white stain for 20 min. Sections were washed to remove the staining solution and mounted in 10% glycerol in 1X PBS buffer. After staining, nodules were observed by confocal microscopy using a Nikon Inverted Eclipse Ti-E Confocal Laser Scanning Microscope. Calcofluor was detected with emission at 460-500 nm, while SYTO9 and PI were detected at 510-570 nm and 600-650 nm, respectively [15]. Three nodules were randomly selected for nodule imaging and bacteroid observation.

2.7. Bioinformatics

Bradyrhizobium sp. SUTN9-2 genome sequences were obtained from the NCBI database (https://www.ncbi.nlm.nih.gov) and Genoscope (https://mage.genoscope.cns.fr) [35]. Multiple sequence alignments were determined by CLUSTALW (2.1) (https://www.genome.jp/tools-bin/clustalw). Domain architecture analysis was performed with the Simple Modular Architecture Research Tool (SMART) (https://smart.embl.de) [36] and InterPro (https://www.ebi.ac.uk/interpro) [37]. The annotation features and whole genome sequences were analyzed by SnapGene software (www.snapgene.com).

2.8. Statistical Analysis

All data were obtained from experiments performed in triplicate. For statistical analyses, one-way analysis of variance (ANOVA) followed by Tukey’s honestly significant difference (HSD) test (Tukey’s tests at P ≤ 0.05) and Student’s t-tests (P ≤ 0.05) were performed using SPSS software (SPSS version 22.0 windows: SPSS Inc., Chicago, IL) and GraphPad Prism statistical software (Version 10.0.3).

3. Results

3.1. Symbiotic Properties of ΔcopG₁, ΔtraG₁, and ΔvirD2₁ in Vigna Radiata cv. SUT4 Symbiosis

Differences between the wild type and mutants of Bradyrhizobium sp. SUTN9-2 were observed in terms of nodulation and nitrogenase activity in V. radiata cv. SUT4 (Figure 1). ΔtraG₁ and ΔvirD2₁ induced a higher number of nodules on the plant tested than the wild type at 7 days post inoculation (dpi) (Figure 1Q), although the nodules produced displayed a white color (Figure 1I,M) instead of pink, indicating problems in nodule development. Moreover, there were higher numbers of dead cells in nodules inoculated with ΔtraG₁ and ΔvirD2₁ in the symbiosome area (Figure 1J,N). At 21 dpi, there was no difference in the number of nodules obtained with ΔtraG₁, ΔvirD2₁, and the wild type (Figure 1R).

In addition, there were different results of the nitrogenase activities in each mutant at 21 dpi, ΔtraG₁ was identical to that of the wild type, while low nitrogenase activity was obtained with ΔvirD2₁ (Figure 1T). Interestingly, ΔcopG₁ showed a significant effect on nodulation in that nodule formation was abolished (Figure 1E,G,Q-R). Although nodule organogenesis was not observed in the plant inoculated with ΔcopG₁, both live and dead cells were detected in the cortex and vascular tissue instead (Figure 1H). According to the results, ΔcopG₁ was able to infect plant cells, but it was no longer capable of surviving in host cells.

3.2. The copG Genes Are Involved in Nodulation Efficiency of Bradyrhizobium sp. SUTN9-2

The copG gene typically encodes the CopG protein, which is a transcriptional factor that consists of ribbon helix-turn-helix (RHH) motifs [18,34,38]. Bradyrhizobium sp. SUTN9-2 has two copies of copG genes located downstream of the traG and virD2 genes. These gene clusters are located in distinct locations on the SUTN9-2 chromosome. Although two copies of copG genes were present on the chromosome, they did not share the same gene sequences. The copG gene copies 1 and 2 (copG₁ and copG₂) revealed a low degree of similarity with 52.35% DNA sequence identity (Figure S2) and 51.77 % amino-acid sequence identity (Figure S3), which are different in both the N- and C-terminals. Domain architecture analysis identified CopG₁ as an unidentified domain, which shows similarities with B. yuanmingense BRP09, CCBAU05623, and, more distantly, in P10 130. While, CopG₂ was identified as a Pfam:RHH domain that is also found in B. diazoefficiens USDA110, B. diazoefficiens SEMIA5080, and B. japonicum J5 (Figure S4).

To gain a more complete understanding of the function of copG₁ and copG₂, we constructed ΔcopG₂, inoculated it into V. radiata cv. SUT4, and compared it with ΔcopG₁ and wild type strains (Figure 2). Contrary to ΔcopG₁ (Figure 2B), ΔcopG₂ was able to produce pink nodules (Figure 2C) that were smaller than those generated by the wild type (Figure 2A). At 14 and 21 dpi, ΔcopG₂ produced the highest number of nodules compared to other strains (Figure 2G,H). Despite the high number of nodules generated by ΔcopG₂, the nitrogenase activity was significantly lower than that of the wild type (Figure 2I,J). Confocal microscopic examination showed dead cells in nodules generated by ΔcopG₂ (Figure 2F) in the symbiosome, in contrast with the wild type (Figure 2D). These findings indicated both copG₁ and copG₂ genes are essential for the symbiotic relationship between SUTN9-2 and legume plants. copG₁ is necessary for nodulation, while copG₂ is crucial for nitrogenase efficiency. Lack of copG₂ leads to decreased nitrogenase activity, despite the presence of high nodule numbers.

3.3. The copG₁ Gene Plays a Crucial Role in the Expression of Nodulation (nod) Genes and Transcriptional Regulator TtsI (ttsI)

To examine if copG₁ affects the structuring of the NF backbone, transcript levels of nodABC and transcriptional activator nodD (nodD1 and nodD2) were examined with and without 20 μM genistein induction (Figure 3). The nodA, nodC, nodD1, and nodD2 genes were almost not expressed in ΔcopG₁ (Figure 3A,C,D,E), while the expression of nodB was not affected by a mutation in copG₁ (Figure 3B). The results indicated that copG₁ modulate, either directly or indirectly, the expression of nod genes, especially the nodD gene (Figure 3D,E), which is a transcriptional activator of NFs [39]. Beside NF biosynthesis, NodD1 also activated the transcriptional regulator TtsI (ttsI), gene encoding for T3SS secretion and synthesis [40]. Similar to the nod genes, the expression level of ttsI was not determined in ΔcopG₁ in all conditions (Figure 3F). The loss of nodule formation in ΔcopG₁ may be caused by the suppression of NF synthesis and also T3SS due to the absence of nodD expression.

3.4. Bradyrhizobium sp. SUTN9-2 copG₁ is Involved in the Repression of T4SS Structural Genes traG₁ and trbE₁

To understand the relationship of T4SS and copG₁ gene more clearly, the gene expression fold changes were examined. T4SS with trbE₁ and traG₁ genes showed high expression levels under non-symbiotic conditions when copG₁ was deleted (Figure 4A,B). These results indicated that copG₁ of SUTN9-2 may act as a suppressor of trbE₁ and traG₁ under non-symbiotic conditions. However, under mimicked symbiotic conditions with 20 μM genistein induction, the expression of the trbE₁ gene significantly decreased in ΔcopG₁ compared to the wild type. The expression of the traG₁ gene did not differ in both the wild type and the ΔcopG₁ (Figure 4C,D). However, it is noteworthy that under mimicked symbiosis conditions, CopG₁ can induce the expression of trbE₁ and traG₁, resulting in significantly higher levels than those observed under non-induction conditions. These results suggest that copG₁ might act as a synergistic regulator of the T4SS gene in SUTN9-2 under flavonoid induction (Figure 4C,D).

3.5. Effect of T4SS and copG₁ on Secreted Protein Pattern after 48 h With Genistein Induction

The secreted protein patterns of Bradyrhizobium sp. SUTN9-2, ∆T4SS₁, and ∆copG₁ with 20 µM genistein after 48 h induction were analyzed by SDS-PAGE. The results revealed distinct protein band patterns in the different conditions (Figure 5). The amino acid sequences of each selected band were examined using mass spectrometry (MS) with the MASCOT program. In the wild type with genistein induction (band 1) contain a protein matched with the T3SS translocon protein, NopX (27%), while this band was absent at the same position of ∆T4SS₁. Similarly, a protein band was observed in the wild type with genistein induction (band 6), a match was found with the T3SS host specificity protein, NopP (18%), this protein was not observed in ∆T4SS₁ (band 7) at the same position. While protein bands from the wild type (band 3) and ∆T4SS₁ (band 4) were found a protein matched with Dct; C₄-dicarboxylate ABC transporter (31%), but it was not observed in ∆copG₁ (band 5). Nevertheless, ∆copG₁ (band 5) was associated with amino acid ABC transporter substrate-binding protein glutamate/aspartate transporter subunit (38%).

qRT-PCR was performed to identify the gene expression of T3SS (nopP and nopX) and C₄-dicarboxylate transporter (dct), which was not in secreted protein from ∆copG₁. The results showed that the expression of T3Es (nopP and nopX) and C₄-dicarboxylate transporter (dct) under genistein induction was down-regulated in ∆copG₁ (Figure 6). This indicated that these genes required copG₁ to mediate the regulation under genistein induction.

4. Discussion

At an early nodulation stage of V. radiata cv. SUT4, ΔtraG₁ and ΔvirD2₁ generated a high number of nodules with smaller size compared to the wild type (Figure 1I,M,Q-R). The symbiosome space of ΔtraG₁ and ΔvirD2₁ infecting nodules revealed some dead cells that were not found in the wild type under confocal microscopy (Figure 1J,N). According to the findings, T4SS is beneficial in the early stages of symbiotic interaction between SUTN9-2 and legumes. Bradyrhizobia have a TraG/Trb operon on the chromosome in the symbiosis island that is similar to that of mesorhizobia based on the traG gene’s phylogenetic and gene organization [15]. Beside the structural protein, various bacteria containing T4SS also identified ATPase/Coupling protein, VirD4/TraG and relaxase, VirD2 [20]. The traG is commonly found in conjugative plasmids that are responsible for horizontal gene transfer between bacteria. The traG gene required for encode T4SS component served as ATPase to generate energy during secretion [19]. In addition, TraG also act as a substrate receptor of T4SS called coupling protein, a substrate receptor that mediate the substrate such as effector proteins, DNA or DNA-protein complex through T4SS channel [7,10,41,42]. In the Pfam prediction, TraG protein was matched with Pfam family T4SS-DNA_transfer (PF02534), TrwB_AAD_bound (PF10412), and TraG-D_C (PF12696) (Figure S8A). The C-terminal of this protein is able to interact with the relaxosome, which is essential for DNA transfer and conjugation in bacteria [19,43,44]. In mesorhizobia, traG plays an important role in an early stage of infection, and its expression was observed during induction with root exudate and early nodules generated by M. mediterraneum Ca36^T. Corresponding to mesorhizobia, traG₁ of SUTN9-2 may play a crucial role in the beginning of symbiotic interaction with legumes [7]. In Agrobacterium, VirD2 protein is a part of the relaxase family that plays a crucial in conjugating and mobilizing plasmids that are required for translocation and integration of T-strands into recipient plant cells [45,46]. The conjugative transfer of ICEMlSymR7A in M. loti R7A requires VirD2 relaxase to initiate the rolling-circle replication [47]. VirD2₁ of SUTN9-2 possesses a domain of unknown function (DUF), DUF3363, which is an uncharacterized protein (Figure S8B). Although ΔvirD2₁ had no effect on the number of nodules of V. radiata cv. SUT4, but the activity of nitrogen fixation was reduced (Figure 1S,T).

Nodules generated by ΔvirD2₁ showed many uninfected cells (Figure 1N,P). This finding showed that the communication between SUTN9-2 and the legume at the beginning of nodule organogenesis plays an important role in enhancing infection efficiency and nitrogenase activity after infection. These results strongly indicated that the traG₁ and virD2₁ genes may be necessary for symbiotic interaction during an early infection stage. Unlike ΔtraG₁ and ΔvirD2₁, ΔcopG₁ has an impact on the symbiotic interaction between SUTN9-2 and legumes because this mutant was unable to generate nodules with the tested plant (Figure 1E,G). Since, copG₁ was located downstream of traG₁ and virD2₁ within the same cluster, it is assumed that copG₁ may share a common promoter with traG₁ and virD2₁. This observation was supported by the previous study, where T4SS complementation successfully restored nodule formation [15]. Several bacteria, such as Pseudomonas aeruginosa [48], Streptococcus agalactiae [38], Vibrio cholerae [49], Bradyrhizobium sp., and Mesorhizobium sp. contain the copG gene in their genomes [15]. This gene encodes CopG protein, a small transcriptional repressor containing a helix-turn-helix motif domain, which is similar to that of regulatory repressors such as Mnt, Arc, and MetJ in Salmonella typhimurium bacteriophage P22 and Escherichia coli [18,50,51]. The CopG protein was first discovered in streptococcal plasmid pMV158 as a transcriptional repressor that interacts with RepB to control the copy number of the plasmid [34,38,52]. In addition, copper resistance was also demonstrated to be influenced by CopG in P. aeruginosa and V. cholerae [48,49].

In SUTN9-2, the CopG₁ protein was classified as an uncharacterized conserved protein, while CopG₂ was annotated as the Pfam;RHH_1 domain (PF01402), which may serve as a transcriptional regulator within the CopG family (Figure S4) [34]. The removal of both copG genes in SUTN9-2 resulted in distinct nodulation efficiency. Even without a nodule generated by ΔcopG₁, it can still infect plants because we could monitor both live and dead cells within plant tissues. Surprisingly, live cells were found mostly in the vascular bundle tissue, which is similar to how endophytic bacteria behave. These findings imply that copG₁ may be crucial for SUTN9-2 in protecting the survival of bacterial cells in the host plant. Bacteria can evolve and adapt to their environment through horizontal gene transfer, which is usually facilitated by conjugation. Conjugation is a significant biological process as it is the primary way to spread antibiotic resistance genes [53]. Integrative and conjugative elements (ICEs) are another essential mechanism that contributes to conjugation. ICEs are recognized as elements encoded for excision and transferred by conjugation and integration, regardless of the specific mechanisms involved [54]. The T4SS found in SUTN9-2 is classified as a tra/trb operon and is recognized for its crucial role in facilitating conjugal transfer. Although SUTN9-2 lacks a conjugation plasmid, ICE is still present on the chromosome. The genes encoding the T4SS₁ cluster are presented in this ICE, which is an alternative mechanism of genetic exchange in this bacterial strain [15]. In order to study the impaired nodulation phenotype of ΔcopG₁ in V. radiata cv. SUT4, we analyzed the expression of nod genes with and without genistein induction. Common nod genes in SUTN9-2 including nodA and nodC genes were not expressed even with a lack of copG₁ under flavonoid induction condition, but it did not affect nodB expression. Additionally, copG₁ acts as a stimulator for nodD1 and nodD2, which are the primary transcription factors responsible for NF production. In addition to nod genes, the expression level of ttsI, which was not determined in ΔcopG₁. TtsI protein is a transcriptional regulator (previously called y4xI) is activated by flavonoids and NodD1 that bind to conserved sequences called tts-boxes [40,55,56]. This protein controlled the genes responsible for T3SS synthesis and effector protein secretion which have an effect for symbiosis interaction between rhizobium-legumes interaction depends on plant host [6,57]. During symbiotic interaction, SUTN9-2 required copG₁ to mediate the expression of nod genes and ttsI under flavonoid induction. These results indicate that CopG₁ may positively mediate the expression of nod genes via NodD activation before stimulating NF production, nodule organogenesis and T3SS. Furthermore, copG₁ plays a role as a repressor in T4SS gene expression, suppressing trbE and traG gene expression under flavonoid stimulation (Figure 4C,D). In contrast, these genes were not affected by flavonoids in the absence of the copG₁ gene.

The protein expression profiles of SUTN9-2, ∆T4SS₁ and ∆copG₁ with genistein treatment were analyzed by SDS-PAGE. The results revealed distinct protein band patterns in the different conditions (Figure 5). ∆copG₁ exhibited a deficiency in producing nodules in various plant species and a striking increase in protein expression compared to the wild type. According to the analysis of copG₁ domain protein (Figure S4), CopG₁ was predicted to be a transcriptional regulator which might play a role in the regulation of gene expression. The ΔcopG₁ lane appears to have much more protein intensity overall because the proportion of protein in this lane might be less than other lanes. So, 10 µg might be shown higher band intensity. A comparative proteomic analysis of the whole secretome should be conducted further to identify additional target proteins involved in this interaction. It was found that several proteins were secreted, but the C₄-dicarboxylate transport system (dct) protein was not identified in ∆copG₁ and this result was corresponded to the down-regulated of dct gene quantified by qRT-PCR (Figure 6). The dct gene play a crucial role for symbiosis in numerous rhizobia [58]. For example, the dct mutant of S. meliloti and R. trifolii can generate ineffective nodules with the host legume [59,60]. Besides the dct gene, there are other genes that are expressed in the same pattern, including nopX and nopP, which are also essential for symbiosis (Figure 6). NopX is a component of T3SS as a translocation pore (translocon) apparatus that is important for host-specific interaction between the rhizobium and host plant. The NGRΔnopX has a significant effect on the nodule number because this mutant forms fewer nodules in all plant species tested, including Flemingia congesta, Tephrosia vogelii, Pachyrhizus tuberosus, and Lablab purpureus [61]. NopP is a T3SS effector protein that is phosphorylated by plant kinases [62]. A lack of NopP in Rhizobium sp. NGR234 reduces the capacity of nodule organogenesis in tropical legumes. This indicates a positive effect of NopP on symbiosis [61]. NopP of B. diazoefficiens USDA122 is necessary and causes Rj2-dependent incompatibility [63]. The T3SS of SUTN9-2 has no impact on the symbiotic relationship with V. radiata [11]. However, based on the protein secretion results of T3Es (NopP and NopX) and nodD gene expression, it is evident that copG₁ regulates the function of nodD and T3SS. Previous reports indicated that nodD controls the function of nod cluster by binding to the nod box region. Similarly, nodD can regulate the T3SS function by binding to ttsI [55]. Therefore, the results of this experiment confirm that CopG₁ controls the function of nodD, influencing the expression of the nod cluster genes and T3SS. Perhaps, CopG₁ is a crucial factor in the early stage of legume and SUTN9-2 communication. It is plausible that the regulatory system governing the expression of nod genes does not solely depend on the interaction between flavonoids and NodD. Another factor, CopG₁, also collaborates with flavonoids and NodD in regulating the expression of nod genes and T3SS. Carbon and nitrogen metabolism are the primary mechanisms that are necessary for the exchange of nutrients between plant and bacteria partners. The proteins secreted from ∆copG₁ matched with the periplasmic binding proteins of the glutamate/aspartate ABC transporter. Glutamate is a significant contributor to the total metabolite content, which plays an essential role in nitrogen metabolism, amino acid metabolism, transamination, and carbon sources [64,65]. During symbiosis, the main carbon source utilized by rhizobia is C₄-dicarboxylic acid [66,67]. In the mimicked symbiotic conditions, ∆copG₁ lost the ability to establish symbiotic interaction. Thereafter, the increasing of glutamate/aspartate ABC transporter may promote carbon and nitrogen uptake to support bacterial cell survival, but it is not necessary for symbiotic interaction. This again suggests that copG₁ may act as a regulator of nodD and T4SS gene expression under symbiosis conditions.

5. Conclusions

The T4SS₁ containing copG₁, traG₁, and virD2₁ has beneficial effects on symbiotic interactions with diverse legumes. The deletion of each gene in a unique cluster showed a distinctive effect on symbiosis. The early stage of infection and nodulation is influenced by traG₁ and virD2₁, while nodulation with legume hosts was not efficient in the absence of copG₁. While, copG₁ is necessary for nodulation, the ΔcopG₂ results in reduced nitrogenase activity, despite maintaining high nodule numbers. This underscores the essential role of copG₂ in nitrogenase efficiency. Overall, both copG₁ and copG₂ genes play crucial roles in the symbiotic relationship between SUTN9-2 and legume plants, contributing to effective nitrogen fixation during symbiosis. The expression of nodulation (nod) genes including common nodA, nodC, nodD1, and nodD2 and ttsI were not detected with and without genistein induction in ΔcopG₁, while the T4SS genes in ΔcopG₁ were highly expressed in non-induction conditions. Interestingly, the absence of copG₁ with the induction of flavonoid led to a low expression level of T4SS genes. The results suggest that copG₁ may play an important role in nodulation, which is a crucial for the symbiotic interaction between Bradyrhizobium sp. SUTN9-2 and legumes. Additionally, copG₁ served as a suppressor of T4SS genes under non-induction conditions and was required to stimulate the expression of T4SS genes through flavonoid induction. copG₁ also acted as a suppressor of secreted protein under flavonoid induction conditions. In addition, a lack of copG₁ led to suppressed expression of nopX, nopP, and dct, which are important for infection and nodulation during symbiosis. Thus, copG₁ is most likely responsible for regulation via functions in T3SS, nodD regulation, and the carbon and nitrogen exchange systems, which are significant for SUTN9-2 during symbiosis.