Transcription Regulation of Cell Cycle Regulatory Genes Mediated by NtrX to Affect Sinorhizobium meliloti Cell Division

The cell division of the alfalfa symbiont, Sinorhizobium meliloti, is dictated by a cell cycle regulatory pathway containing the key transcription factors CtrA, GcrA, and DnaA. In this study, we found that NtrX, one of the regulators of nitrogen metabolism, can directly regulate the expression of ctrA, gcrA, and dnaA from the cell cycle pathway. Three sets of S. meliloti ntrX mutants showed similar cell division defects, such as slow growth, abnormal morphology of some cells, and delayed DNA synthesis. Transcription of ctrA and gcrA was upregulated, whereas the transcription of dnaA and ftsZ1 was downregulated in the insertion mutant and the strain of Sm1021 expressing ntrXD53E. Correspondingly, the inducible transcription of ntrX activates the expression of dnaA and ftsZ1, but represses ctrA and gcrA in the depletion strain. The expression levels of CtrA and GcrA were confirmed by Western blotting. The transcription regulation of these genes requires phosphorylation of the conserved 53rd aspartate in the NtrX protein that binds directly to the promoter regions of ctrA, gcrA, dnaA, and ftsZ1 by recognizing the characteristic sequence CAAN2-5TTG. Our findings suggest that NtrX affects S. meliloti cell division by regulating the transcription of the key cell cycle regulatory genes.


Introduction
Caulobacter crescentus is a model strain of α-proteobacteria in molecular cell biology [1]. It takes advantage of one cell division to produce two cells with different shapes and sizes [2]. In recent years, a complex cell cycle regulatory network has been revealed in this species. This network consists of multiple histidine kinases such as CckA, DivL, DivJ, and PleC, a histidine phosphotransfer protein ChpT, response regulators DivK and CpdR, and transcription regulators CtrA, GcrA, DnaA, SciP, and MucR [1][2][3][4][5][6]. Among these cell cycle regulators, CtrA and GcrA negatively regulate cell division, which is opposite to DnaA. Although this network has been reported to possibly mediate nutritional signals for regulating bacterial growth and proliferation, the exact molecular mechanism is currently unclear.
S. meliloti is a model strain of rhizobia that can infect the host plant and form nitrogenfixing nodules. During symbiosis, the cell division of S. meliloti on the surface of host plant alfalfa roots, at the front ends of extended infection threads and in the infection zones of root nodules, is stringently controlled [7], but how cell division is regulated is not very clear. Although NCR (nodule cysteine-rich) peptides secreted by host plants are known to induce terminal differentiation of bacteroids in host plant cells [8][9][10][11], several legumes do not produce NCR peptides. Therefore, there may be other mechanisms that

Bacterial Cell Synchronization
Bacterial cell synchronization was done according to the method of De Nisco et al. [33]. S. meliloti colonies were selected from an agar plate, inoculated into 5 mL LB/MC broth, and shaken cultured at 28 • C, 250 rpm/min overnight. Then, 100 µL of the bacterial culture was transferred into 100 mL LB/MC broth and shaken cultured overnight until OD 600 = 0.1-0. 15. The cells were collected by centrifugation (6500 rpm, 5 min, 4 • C), washed twice with sterilized 0.85% NaCl solution, resuspended in MOPS-GS synchronization broth, and shaken cultured for 270 min. After centrifugation, the cells were washed twice with sterilized 0.85% NaCl solution, resuspended in LB/MC broth, and grown at 28 • C.

RNA Extraction, Purification and qRT-PCR
The cells from 20 mL of bacterial cultures were collected by centrifugation (6000 rpm, 5 min, 4 • C), and washed twice with DEPC-treated water. RNA extraction was performed using 1 mL of Trizol (Thermo Fisher, Shanghai, China). Total RNA was treated with genomic DNA Eraser (Takara, Beijing, China) to remove any remaining genomic DNA, and then transcribed to cDNA using a PrimeScript RT Reagent Kit [31]. The qPCR reaction system included the following: SYBR ® Green Real-time PCR Master Mix, 4.75 µL; cDNA or DNA, 0.25 µL; 10 pmol/µL primers, 0.5 µL; ddH 2 O, 4.5 µL. The reaction procedure is as follows: 95 • C, 5 min; 95 • C, 30 s; 55 • C, 30 s; 72 • C, 1 kb/min. The selected reference gene was SMc00128. The 2 −∆∆CT method was applied to analyze gene expression levels. All primers are listed in Table S3.

Chromatin Immunoprecipitation (ChIP)
ChIP was performed as described by Pini [13] using rabbit anti-NtrX polyclonal antibodies prepared by Wenyuange, Shanghai [36]. In brief, Sm1021 cells (2 mL, OD 600 of 0.8) were cross-linked in 10 mM PBS (pH7.6) containing 1% formaldehyde at room temperature for 10 min, and incubated on ice for 30 min. The cells were washed three times with PBS, treated with lysozyme, sonicated (EpiShear TM , Active Motif, Tokoyo, Japan) on ice using 15 bursts of 30 s (50% duty) at 40% amplitude. Lysates were diluted in 1 mL of ChIP buffer and pre-cleared with 50 µL of protein-A agarose and 80 µg BSA. Anti-NtrX polyclonal antibodies were added to the supernatant (1:1000 dilution), incubated overnight at 4 • C with 50 µL of protein-A agarose beads pre-saturated by BSA, and washed with low, high salt and LiCl buffer once and twice with TE buffer. The protein-DNA complexes were eluted using 200 µL freshly prepared elution buffer (1% SDS, 0.1 M NaHCO 3 ) supplemented with NaCl to a final concentration of 300 mM, and incubated for 6 h or overnight at 65 • C Genes 2022, 13, 1066 4 of 17 to reverse the crosslinks. DNA was purified by a MinElute kit (QIAGEN, Shanghai, China) and resuspended in 40 µL of Elution Buffer. DNA sequencing was completed using Illumina HiSeq 2000 in BGI (Beijing Genomics Institute, Beijing, China). PCR was performed as the same as above qRT-PCR, and SMc00128 was used as an internal reference to normalize the data.

Flow Cytometry
De Nisco's flow cytometry protocol was used [33]. The cells from 4 mL of S. meliloti cultures were collected by centrifugation (6000 rpm, 5 min, 4 • C), and washed twice with a 0.85% NaCl solution (stored at 4 • C). Then, 250 µL of cell suspension was mixed with 1 mL of 100% ethanol to fixation. The fixed cells were collected by centrifugation (6000 rpm, 3 min), and incubated in 1 mL of 50 mM sodium citrate buffer containing 4 µg/mL RNase A at 50 • C for 1.5 h. Then, 1 µL of 10 µM SYTOX Green dye (Sigma, Shanghai, China) was added to each sample. Each sample was assessed using a MoFlo XDF (Beckman Coulter, Suzhou, China) flow cytometer, and the results were analyzed by Summit 5.1 software (Beckman Coulter, Suzhou, China).
Synchronized S. meliloti cells (Sm1021, SmLL1, Sm1021/pntrX, and Sm1021/pntrX D53E ) were subcultured in LB/MC broth containing 1 mM IPTG or not for 1 to 3 h. The cells from 1 mL culture were pelleted, resuspended in the buffer of 10 mM Tris-Cl, pH 7.5 and 4% SDS, incubated at room temperature for 5 min, mixed the loading dye, boiled for 10 min, and then loaded into the wells of Phos-tag™ acrylamide gels. Western blots were performed as described by Tang [37], with rabbit anti-NtrX (1:10,000) antibodies [36]. Chemiluminescent detection was performed using an ECL fluorescence colorimetric kit Genes 2022, 13, 1066 5 of 17 (Tiangen, Beijing, China) and fluorescent signals were visualized using a Bio-Rad Gel Doc XR. Band intensities were evaluated by Image J [38].
To determine the protein levels between Sm1021 and SmLL1, synchronized cells were subcultured in 100 mL of LB/MC broth at 28 • C for half to three hours. The ntrX depleted cells and the synchronized cells of Sm1021/pntrX and Sm1021/pntrX D53E were subcultured in 100 mL of LB/MC broth containing 1 mM IPTG for one to three hours.~10 8 cells were collected by centrifugation every half or one hour for each strain. Then, 1 mg His-fused CtrA and GcrA proteins were purified through Ni + columns from supernatant of E. coli BL21 lysates. Rabbit anti-CtrA and anti-GcrA polyclonal antibodies were prepared by Hua'an Biotech, Hangzhou.

Microscopy
A 5 µL aliquot of fresh S. meliloti culture (OD 600 = 0.15) was placed on a glass slide and covered with a cover glass. The slide was slightly baked near the edge of the flame of an alcohol lamp for a few seconds, and observed under a phase contrast microscope (Zeiss). The cells carrying pHC60 [39] were observed in GFP mode, and the images were acquired using a CCD camera Axiocam 506 color (Zeiss). The exposure time was set to 10 ms in order to capture bacterial morphology. Scanning electron microscopy was performed as described by Wang [31] to further observe cell shapes of S. meliloti at the mid-log phase.

DNA Sequencing and Analysis
ChIP-Seq was performed by BGI [40]. DNA library was prepared including DNA-end repair, 3'-dA overhang, the ligation of methylated sequencing adaptor, PCR amplification, and size selection (usually 100-300 bp, including adaptor sequence). Bioinformatics analysis was performed as follows. The ratio of N was over 10% in the whole read. Removed reads of unknown bases are more than 10%. The ratio of the base whose quality was less than 20 was over 10%. Clean parameter: SOAP nuke filter -l 15 -q 0.5 -n 0.01 -Q 2 -c 21. After filtering, the clean data were then mapped to the reference genome by SOAP aligner/SOAP2 (Version: 2.21t, BGI, Beijing, China). BWA (Burrows-Wheeler Aligner, Version: 0.7.10, University of Washington) was also used to do genome alignment after evaluating its performance. Align parameter: soap_mm_gz -p 4 -v 2 -s 35. MACS (model-based analysis for ChIP-Seq, version: MACS-1.4.2): the candidate Peak region was extended to be long enough for modeling. Dynamic Poisson distribution was used to calculated p-value of the specific region based on the unique mapped reads. The region would be defined as a peak when p-value < l × 10 −5 . Peak calling parameter: macs14 -s 50 -g 6,691,694 -p 1 × 10 −5 -w -space 50 -m 10, 30. UCSC (University of California Santa Cruz) Genome Browser was used for reading peaks.

Analysis of NtrX 3D Structure
The 3D structure of S. meliloti NtrX was reconstructed in the server of Swiss-Model using the 4d6y template from B. abortus in PDB [41]. The 3D structures of NtrX were analyzed by the software Pymol (Delano Scientific, Berkeley, CA, USA).

Defects of Cell Division Resulting from ntrX Mutation in S. meliloti
We previously constructed a plasmid insertion mutant of the ntrX gene in S. meliloti 1021, called as SmLL1 [31]. This mutant grew slowly in LB/MC medium compared to wildtype Sm1021 [31]. According to the determined growth curve, the doubling time of bacterial cell proliferation was calculated to be 180 min for SmLL1 compared to 160 min for the wildtype strain [33]. Microscopic observation revealed that 5% to 10% of SmLL1 cells grown in the LB/MC broth up to the logarithmic phase exhibited morphological abnormalities (such as cell elongation, Y-shaped or V-shaped), whereas Sm1021 had almost no abnormally shaped cells ( Figure 1A,B). To determine whether the appearance of abnormally shaped cells of the SmLL1 strain is associated with the synthesis and segregation of genomic DNA, we synchronized the S. meliloti cells, inoculated them in LB/MC broth, grew them for Genes 2022, 13, 1066 6 of 17 180 min, and then harvested the cells for flow cytometric analysis. The results showed that most of the Sm1021 cells were haploids, with only a few diploids, whereas the majority of SmLL cells were diploid ( Figure 1B), indicating a deceleration of replication and segregation of their genomic DNA in the SmLL1 cells as compared to those of the wild type. These observations indicate that the SmLL mutant has cell division defects. cells grown in the LB/MC broth up to the logarithmic phase exhibited morphological ab normalities (such as cell elongation, Y-shaped or V-shaped), whereas Sm1021 had almo no abnormally shaped cells ( Figure 1A,B). To determine whether the appearance of ab normally shaped cells of the SmLL1 strain is associated with the synthesis and segregatio of genomic DNA, we synchronized the S. meliloti cells, inoculated them in LB/MC broth grew them for 180 min, and then harvested the cells for flow cytometric analysis. Th results showed that most of the Sm1021 cells were haploids, with only a few diploid whereas the majority of SmLL cells were diploid ( Figure 1B), indicating a deceleration o replication and segregation of their genomic DNA in the SmLL1 cells as compared to thos of the wild type. These observations indicate that the SmLL mutant has cell division de fects. Because the deletion of ntrX may be fatal, the deletion mutant in S. meliloti 1021 wa not yet successfully obtained. Therefore, we constructed a depletion strain where the ntr gene on the genome was deleted, but carried an IPTG inducible-expression ntrX gen from a plasmid (ΔntrX/pntrX) to verify the above results. Optical microscopic observatio  Red arrows, abnormal cells; bars, 2 µm. Sm1021 and SmLL1 cells were grown in LB/MC broth to logarithmic phase in (A,B). The depletion cells (∆ntrX/pntrX) carrying pHC60 [39] were grown in LB/MC broth with or without 1 mM IPTG for two hours in (D). The cells of Sm1021/pntrX and Sm1021/pntrX D53E carrying pHC60 were grown in LB/MC broth containing 1 mM IPTG for two hours. (C,F,I) Genomic DNA content of ntrX mutants was determined by flow cytometry. 1C, haploid; 2C, diploid. Synchronized cells of Sm1021, SmLL1, Sm1021/pntrX, and Sm1021/pntrX D53E were grown in LB/MC broth for three hours in (C,I). The latter two strains were also incubated with 1 mM IPTG in (I). The depletion cells were grown in LB/MC broth with or without 1 mM IPTG for one hour in (F). (E,H) Growth curves of the ntrX depletion strain and Sm1021/pntrX D53E in LB/MC broth with 1 mM IPTG. Error bars, ±SD.
Because the deletion of ntrX may be fatal, the deletion mutant in S. meliloti 1021 was not yet successfully obtained. Therefore, we constructed a depletion strain where the ntrX gene on the genome was deleted, but carried an IPTG inducible-expression ntrX gene from a plasmid (∆ntrX/pntrX) to verify the above results. Optical microscopic observation showed that more than 30% of the ntrX depleted cells in LB/MC broth without IPTG displayed abnormal shapes (such as elongation and T-shaped), while in LB/MC broth with IPTG, almost no abnormal cells were observed ( Figure 1D). The depletion strain barely proliferated in LB/MC broth without IPTG, whereas it duplicated slowly with IPTG induction ( Figure 1E), indicating that ntrX gene expression is required for the cell division of S. meliloti. Flow cytometric analysis showed that three peaks were detected in the depletion cells, including haploid and diploid in LB/MC broth without IPTG induction ( Figure 1F). After the one-hour induction of IPTG, the peaks were similar to the wild-type cells ( Figure 1F), indicating that genomic DNA replication and segregation of S. meliloti requires the expression of the ntrX gene.
NtrX, as a regulator of nitrogen metabolism, is composed of a REC domain and a DNA-binding domain [30]. The phosphorylated NtrX has been reported in C. crescentus and B. abortus [28,30], and the putative phosphorylation site is predicted as the conserved 53rd aspartate residue (D53) on the REC domain ( Figure 4A,B). If the NtrX protein is indeed involved in the regulation of S. meliloti cell division, as described above, then the mutation of the conserved D53 residue would affect its regulatory function. To test this hypothesis, we tried to construct the substitutions of D53 (replaced by A, N, or E) of NtrX from the genome of S. meliloti 1021, but were unable to successfully obtain the mutants. As a result, we cloned the mutation gene into the expression vector pSRK-Gm [34] and then introduced the recombinant plasmid into Sm1021. On the LB/MC/IPTG plate, we found that the strain expressing NtrX D53A or NtrX D53N almost did not form visible colonies with IPTG induction; however, the strain expressing NtrX D53E or NtrX formed many colonies in the same condition ( Figure S1). GFP-labeled S. meliloti cells (S. meliloti strain carry pHC60 that constitutively expresses a gfp gene) [39] cultured in LB/MC/IPTG broth up to the logarithmic phase were observed under a fluorescence microscope, and more than 20% of Sm1021/pntrX D53E cells had abnormal morphology, while Sm1021/pntrX cells were almost normal ( Figure 1G). The growth curve determination also showed that the growth of Sm1021/pntrX D53E in LB/MC/IPTG broth was apparently slower than that of Sm1021/pntrX ( Figure 1G). Synchronized S. meliloti cells were subcultured into LB/MC/IPTG broth and grown for 180 min for flow cytometric analysis. The results showed that only one sharp peak (haploid) was found in Sm1021/pntrX cells, whereas three peaks were detected in Sm1021/pntrX D53E cells, including haploid and diploid ( Figure 1I). These results indicate that the D53 residue of the NtrX protein is important for cell division of S. meliloti.

Transcription of Cell Cycle Regulated Genes under the Regulation of NtrX in S. meliloti
Since NtrX affects cell division of S. meliloti, does it regulate the transcription of cell cycle regulatory genes? To test this possibility, we performed a preliminary RNA-Seq analysis between Sm1021 and SmLL1 cells. The results indicated that the differential expression of many cell cycle regulatory genes such as chpT, sciP, dnaA, ftsZ1, ccrM, podJ1, cckA, cbrA, pleD, divK, cpdR1, mucR, and clpP between the mutant strain and the wild-type strain were observed ( Figure S2 and Table S1), suggesting that NtrX may regulate the transcription of key cell cycle regulatory genes.
To confirm the above results in detail, we applied quantitative RT-PCR to analyze the transcript levels of cell cycle regulatory genes in S. meliloti cells. The qRT-PCR results showed that the transcript level of the ntrX gene increased first in Sm1021, then decreased, and reached the maximum value in the cells cultured for 90 min, displaying a trend of cyclical changes, while that of the ntrY gene was not obvious (Figure 2A). Known cell cycle regulatory genes, such as ctrA, gcrA, and dnaA, also exhibited a cyclical transcription trend (Figures 2A and S3A). Compared to the wild-type cells, the transcripts of the ntrX gene significantly decreased in the SmLL1 cells grown at different times, but the cyclical trend was unchanged, and cell cycle regulatory genes such as dnaA, ftsZ1, pleC, chpT, and cpdR1 showed similar downregulation trend (Figures 2A and S3A). Contrary to these results, the transcripts of ctrA showed an increasing trend in the SmLL1 cells relative to Sm1021, and transcription levels of gcrA, ccrM, and ntrY significantly increased in the mutants (Figures 2A and S3A). These findings suggest that the NtrX protein may repress the transcription of ctrA and gcrA and activate the transcription of dnaA and ftsZ1. gene transcripts were detected (Figures 2B and S3B). The transcript levels of cell cy regulatory genes were altered to varying degrees in depleted cells with IPTG for 2 or compared to the cells without IPTG: the transcription of ctrA, gcrA, and ccrM exhibite downregulated trend; the transcription of dnaA, ftsZ1, pleC, chpT, and cpdR1 indicated upregulated trend (Figures 2B and S3B). These results are almost consistent with th from SmLL1.  We analyzed the transcripts of cell cycle regulatory genes in cells of the depletion strain to verify the above results. The qRT-PCR results showed that depleted cells cultured in LB/MC broth without IPTG had extremely low levels of ntrX transcripts, while transcripts of most cell cycle regulatory genes were high-level detected (Figures 2B and S3B). After culturing the depleted cells in LB/MC broth with 1 mM IPTG for 1 h, numerous ntrX gene transcripts were detected (Figures 2B and S3B). The transcript levels of cell cycle regulatory genes were altered to varying degrees in depleted cells with IPTG for 2 or 3 h compared to the cells without IPTG: the transcription of ctrA, gcrA, and ccrM exhibited a downregulated trend; the transcription of dnaA, ftsZ1, pleC, chpT, and cpdR1 indicated an upregulated trend ( Figures 2B and S3B). These results are almost consistent with those from SmLL1.
The transcript levels of some cell cycle regulatory genes in Sm1021/pntrX D53E and Sm1021/pntrX cells were examined by qRT-PCR. The results indicated that transcripts of the ntrX gene significantly increased in cells cultured in LB/MC broth containing IPTG for 2 h compared to the cells without IPTG treatment; meanwhile, the transcripts of ctrA, gcrA, and ntrY showed a decreasing trend, while the transcripts of dnaA and ftsZ1 genes increased ( Figure 2C). Contrary to the above results, as transcripts of the ntrX D53E gene increased under IPTG induction, transcripts of ctrA, gcrA, and ntrY also increased significantly, while transcripts of dnaA and ftsZ1 genes decreased ( Figure 2C). These results are almost consistent with those from SmLL1 and the depletion strain, suggesting that the D53 residue on NtrX is important for transcription regulation.
To determine whether the expression of the key cell cycle regulatory genes ctrA, gcrA, and dnaA is regulated by NtrX in heterogeneous cells, the promoter-uidA fusions were co-transformed with pntrX or the empty vector (pSRK-Gm) into E. coli DH5α, respectively. Quantitative analysis of GUS activities showed that the weaker activities of the promoter of ctrA or gcrA in the cells carrying pntrX than those cells with pSRK-Gm were observed ( Figure S3C). In contrast, the activity of the dnaA promoter is significantly elevated in the cells co-expressing ntrX compared with those cells carrying pSRK-Gm ( Figure S3C). These data supported the conclusion that NtrX negatively controls transcription of ctrA and gcrA, but positively regulate the transcription of dnaA.

Levels of CtrA and GcrA Were Affected by NtrX in S. meliloti
To determine whether the levels of CtrA and GcrA are affected by the ntrX mutation, we first expressed His-tagged NtrX, CtrA, and GcrA proteins in E. coli. After purification via nickel columns, rabbit polyclonal antibodies were prepared for immunoblotting assays [36]. The results showed a varying trend of increasing first and then decreasing NtrX protein levels and a maximum occurring in the synchronized cells subcultured for 1.5 h ( Figure 3A). Unlike Sm1021, the total NtrX protein level in SmLL1 cells tended to decrease at most time points ( Figure 3A). Contrary to the NtrX protein, the change trend of CtrA and GcrA protein levels in Sm1021 first decreased and then increased. The levels of these two proteins tended to increase in SmLL1 cells compared to Sm1021 cells ( Figure 3A). These results are almost consistent with qRT-PCR data from SmLL. We evaluated the NtrX protein level in cells of the depletion strain grown in LB/MC broth by immunoblotting and found that cells cultured in the broth containing IPTG for 1 to 3 h expressed a high level of NtrX protein ( Figure 3B). CtrA and GcrA proteins were expressed in cells cultured in LB/MC broth without IPTG, whereas their levels tended to decrease in cells cultured in broth containing IPTG for 1 to 3 h ( Figure 3B). These results prove that NtrX protein levels are negatively correlated with CtrA and GcrA proteins, almost consistent with those in SmLL1.
To determine whether the D53 residue on the NtrX protein affects the levels of CtrA and GcrA, we performed immunoblot assays of lysates from Sm1021/pntrX D53E and Sm1021/pntrX cells. The results showed that the protein levels of NtrX and NtrX D53E tended to increase when cultured in LB/MC broth containing IPTG for 2-3 h ( Figure 3C). Under the same culture conditions, the protein levels of CtrA and GcrA tended to reduce in Sm1021/pntrX cells, while they tended to increase in Sm1021/pntrX D53E cells ( Figure 3C). These results confirm that the D53 residue in the NtrX protein is important for the transcription regulation of CtrA and GcrA.  The depletion cells were grown in LB/MC broth containing 1 mM IPTG for one to three hours, while the cells were incubated without IPTG for one hour as a control in (B). Synchronized cells of Sm1021/pntrX and Sm1021/pntrX D53E were grown in LB/MC broth containing 1 mM IPTG for one to three hours in (C). CB, total proteins stained by Coomassie brilliant blue. N/TP, C/TP, or G/TP, the blot intensity of NtrX, CtrA or GcrA (the larger one) divided the intensity of the strongest band of the total proteins stained by Coomassie brilliant blue. The intensity data were collected using Image J [38].

The 53rd Aspartate Residue as a Phosphorylation Site of S. meliloti NtrX
The homologous NtrX proteins in α-proteobacteria are composed of REC and DNA binding domains. The three-dimensional structure of the NtrX protein from B. abortus has been partially resolved [29,30]. Using this as a template, we reconstructed the 3D structure of the NtrX protein from S. meliloti and found that there were 5 α-helices and 5 β-sheets connected by loops in the REC domain ( Figure 4A,B). The conserved D53 is located at the end of the third β-sheet and predicted as a phosphorylated residue by PFAM.
From the report of B. abortus, the NtrY histidine kinase can phosphorylate NtrX in vitro [29,30]. We expressed and purified the NtrY kinase domain and NtrX protein of S. meliloti in E. coli for in vitro phosphorylation assays. Through Phos-Tag Gel assays, we found that the NtrY kinase domain was autophosphorylated, and phosphorylated the NtrX protein in vitro ( Figure 4C). To further determine whether NtrX is phosphorylated in vivo, Western blotting assays were performed using anti-NtrX antibodies after separating phosphorylated proteins of S. meliloti cells by Phos-Tag Gel. The results showed that more phosphorylated NtrX proteins were detected in Sm1021 cells than those in SmLL1 cells as the same as the unphosphorylated protein ( Figure 4D). To further verify that the D53 residue is the phosphorylation site of the NtrX protein, we applied the same method to analyze the phosphorylated NtrX protein level of Sm1021/pntrX D53E cells cultured in LB/MC/IPTG broth. The results showed that the ratio of phosphorylated NtrX protein compared to unphosphorylated protein in Sm1021/pntrX cells tended to increase, whereas the ratio in Sm1021/pntrX D53E cells tended to decrease ( Figure 4D). These results suggest that the D53 residue of NtrX is the phosphorylation site.   D53E were grown in LB/MC broth with 1 mM IPTG for 1 to 3 h. Then,~1 µg of total protein was loaded into each well. NP/N is the intensity of phosphorylated proteins divided the intensity of non-phosphorylated proteins. The intensity data were collected using Image J [38].
From the report of B. abortus, the NtrY histidine kinase can phosphorylate NtrX in vitro [29,30]. We expressed and purified the NtrY kinase domain and NtrX protein of S. meliloti in E. coli for in vitro phosphorylation assays. Through Phos-Tag Gel assays, we found that the NtrY kinase domain was autophosphorylated, and phosphorylated the NtrX protein in vitro ( Figure 4C). To further determine whether NtrX is phosphorylated in vivo, Western blotting assays were performed using anti-NtrX antibodies after separating phosphorylated proteins of S. meliloti cells by Phos-Tag Gel. The results showed that more phosphorylated NtrX proteins were detected in Sm1021 cells than those in SmLL1 cells as the same as the unphosphorylated protein ( Figure 4D). To further verify that the D53 residue is the phosphorylation site of the NtrX protein, we applied the same method to analyze the phosphorylated NtrX protein level of Sm1021/pntrX D53E cells cultured in LB/MC/IPTG broth. The results showed that the ratio of phosphorylated NtrX protein compared to unphosphorylated protein in Sm1021/pntrX cells tended to increase, whereas the ratio in Sm1021/pntrX D53E cells tended to decrease ( Figure 4D). These results suggest that the D53 residue of NtrX is the phosphorylation site.

Direct Binding of the NtrX Protein to the Promoter DNA of Some Cell Cycle Regulatory Genes
To determine whether the NtrX protein of S. meliloti directly regulates the expression of cell cycle regulatory genes, we used anti-NtrX polyclonal antibodies with high specificity ( Figure S4) to perform chromatin immunoprecipitation experiments. Sequencing results showed that a total of 82 DNA fragments were specifically precipitated from Sm1021 cells, 60 of which were derived from the chromosome, while the other 22 fragments originated from the plasmids SymA and SymB ( Figure 5A). After sequence analysis in detail, we found that the promoter DNA fragments of cell cycle regulatory genes ctrA, dnaA, mucR, and cpdR1 were specifically enriched ( Figure 5B and Table S2). Due to of the recognition sites (CAAN 3-5 TTG) of NtrX on the ntrY gene promoter reported in B. abortus [30], we searched for them in the precipitated DNA fragments, and found that nine of possible motifs are located in the promoter regions of cell cycle regulatory genes ctrA, dnaA, gcrA, and ftsZ1 ( Figure S5). To verify the ChIP-Seq results, we applied quantitative PCR to evaluate the level of genomic DNA fragments precipitated by anti-NtrX polyclonal antibodies. The results showed that the promoter regions of ctrA, dnaA, gcrA, and ftsZ1 genes were enriched to varying degrees ( Figure 5C), indicating that the NtrX protein in Sm1021 cells can bind directly to the promoter regions of the aforementioned cell cycle regulatory genes. Swiss-Model. Possible electrostatic interactions associated with the 53rd aspartate residue are labeled via Pymol. (C) In vitro NtrX phosphorylation catalyzed by the NtrY kinase domain. NtrY-Kr, His-NtrY kinase domain fusion protein (1, 3 and 10 μg); NtrXr-P, phosphorylated NtrXr (10 μg); Ac-Pi, acetyl phosphate (2 mM). (D,E) In vivo phosphorylation of S. meliloti NtrX. Phosphorylated NtrX proteins from S. meliloti cells were separated by Phos-Tag gel and detected by Western blotting of anti-NtrX polyclonal antibodies. S. meliloti 1021 cells carrying pntrX or pntrX D53E were grown in LB/MC broth with 1 mM IPTG for 1 to 3 h. Then, ~1 μg of total protein was loaded into each well. NP/N is the intensity of phosphorylated proteins divided the intensity of non-phosphorylated proteins. The intensity data were collected using Image J [38].

Direct Binding of the NtrX Protein to the Promoter DNA of Some Cell Cycle Regulatory Genes
To determine whether the NtrX protein of S. meliloti directly regulates the expression of cell cycle regulatory genes, we used anti-NtrX polyclonal antibodies with high specificity ( Figure S4) to perform chromatin immunoprecipitation experiments. Sequencing results showed that a total of 82 DNA fragments were specifically precipitated from Sm1021 cells, 60 of which were derived from the chromosome, while the other 22 fragments originated from the plasmids SymA and SymB ( Figure 5A). After sequence analysis in detail, we found that the promoter DNA fragments of cell cycle regulatory genes ctrA, dnaA, mucR, and cpdR1 were specifically enriched ( Figure 5B and Table S2). Due to of the recognition sites (CAAN3-5TTG) of NtrX on the ntrY gene promoter reported in B. abortus [30], we searched for them in the precipitated DNA fragments, and found that nine of possible motifs are located in the promoter regions of cell cycle regulatory genes ctrA, dnaA, gcrA, and ftsZ1 ( Figure S5). To verify the ChIP-Seq results, we applied quantitative PCR to evaluate the level of genomic DNA fragments precipitated by anti-NtrX polyclonal antibodies. The results showed that the promoter regions of ctrA, dnaA, gcrA, and ftsZ1 genes were enriched to varying degrees ( Figure 5C), indicating that the NtrX protein in Sm1021 cells can bind directly to the promoter regions of the aforementioned cell cycle regulatory genes.  In Sm1021, the promoter DNA of the ntrY gene can directly interact with the NtrX protein ( Figure 5C), which is similar to the report in B. abortus [30]. To further confirm the above results, we synthesized a biotin-labeled probe of ntrY promoter DNA (containing two CAAN 3-5 TTG motifs: CAACACCGTTG and CAATGCGTTG) for gel retardation assays. The results showed that phosphorylated NtrX specifically bound to it, forming two protein-DNA complexes ( Figure 6A). To determine whether the D53 residue of the NtrX protein is involved in the protein-DNA binding reaction, we replaced the phosphorylated NtrX protein with the NtrX D53E protein. The gel retardation results showed almost no protein-DNA complex formation ( Figure 6A), suggesting that the phosphorylation of D53 is essential for the binding of NtrX to the ntrY promoter region. The same method was used to analyze the binding ability between the phosphorylated NtrX protein and the biotinlabeled probe of dnaA promoter DNA (containing the CAAAACCCTTG motif) and found that they bound specifically to form a protein-DNA complex ( Figure 6B). We mutated the CAAAACCCTTG motif of the DNA probe to CGGAACCCCCG and found that the mutant probe virtually did not bind to the phosphorylated NtrX protein ( Figure 6B), suggesting that the base composition of the recognition site is important for NtrX binding reaction. We also used gel retardation assays to confirm whether the phosphorylated NtrX protein can specifically bind to biotin-labeled probes of ctrA, gcrA, and ftsZ1 promoter DNA (each containing a CAAN 3-5 TTG motif: CAACCTTG, CAAACCTTG, and CAATGGCTG), and found that at least one protein-DNA complex was formed, respectively ( Figure 6C-E).
These results indicate that the phosphorylated NtrX protein can bind specifically to the promoter regions of ctrA, gcrA, dnaA, and ftsZ1 in vitro.
Genes 2022, 13, x FOR PEER REVIEW 13 of 17 In Sm1021, the promoter DNA of the ntrY gene can directly interact with the NtrX protein ( Figure 5C), which is similar to the report in B. abortus [30]. To further confirm the above results, we synthesized a biotin-labeled probe of ntrY promoter DNA (containing two CAAN3-5TTG motifs: CAACACCGTTG and CAATGCGTTG) for gel retardation assays. The results showed that phosphorylated NtrX specifically bound to it, forming two protein-DNA complexes ( Figure 6A). To determine whether the D53 residue of the NtrX protein is involved in the protein-DNA binding reaction, we replaced the phosphorylated NtrX protein with the NtrX D53E protein. The gel retardation results showed almost no protein-DNA complex formation ( Figure 6A), suggesting that the phosphorylation of D53 is essential for the binding of NtrX to the ntrY promoter region. The same method was used to analyze the binding ability between the phosphorylated NtrX protein and the biotinlabeled probe of dnaA promoter DNA (containing the CAAAACCCTTG motif) and found that they bound specifically to form a protein-DNA complex ( Figure 6B). We mutated the CAAAACCCTTG motif of the DNA probe to CGGAACCCCCG and found that the mutant probe virtually did not bind to the phosphorylated NtrX protein ( Figure 6B), suggesting that the base composition of the recognition site is important for NtrX binding reaction. We also used gel retardation assays to confirm whether the phosphorylated NtrX protein can specifically bind to biotin-labeled probes of ctrA, gcrA, and ftsZ1 promoter DNA (each containing a CAAN3-5TTG motif: CAACCTTG, CAAACCTTG, and CAATGGCTG), and found that at least one protein-DNA complex was formed, respectively ( Figure 6C-E). These results indicate that the phosphorylated NtrX protein can bind specifically to the promoter regions of ctrA, gcrA, dnaA, and ftsZ1 in vitro. . Probe PdnaAs is the DNA probe PdnaA that CAAAACCCTTG was replaced by CGGAACCCCCG in (B). D/P complex, DNA-protein complex; competitor, the DNA probe without biotin labeling. Then, 0, 3, 6, and 15 ng His-NtrX proteins mixed with each probe (2 nM), respectively. TSS, transcriptional start sites from the literature [42]. Blue bars, probes for EMSA; red balls, the putative recognition site of NtrX, CAAN2-5TTG; green balls, the binding sites of CtrA [12,13].

Discussion
In symbiotic nitrogen-fixing bacteria, rhizobia, the level of combined nitrogen as a signal not only regulates the expression of nitrogen-fixing genes, but also affects cell . Probe PdnaAs is the DNA probe PdnaA that CAAAACCCTTG was replaced by CGGAACCCCCG in (B). D/P complex, DNA-protein complex; competitor, the DNA probe without biotin labeling. Then, 0, 3, 6, and 15 ng His-NtrX proteins mixed with each probe (2 nM), respectively. TSS, transcriptional start sites from the literature [42]. Blue bars, probes for EMSA; red balls, the putative recognition site of NtrX, CAAN 2-5 TTG; green balls, the binding sites of CtrA [12,13].

Discussion
In symbiotic nitrogen-fixing bacteria, rhizobia, the level of combined nitrogen as a signal not only regulates the expression of nitrogen-fixing genes, but also affects cell growth and division. However, whether the regulators of nitrogen metabolism are associated with bacterial cell division is unclear. This work first revealed in S. meliloti that the nitrogen metabolism regulator NtrX directly regulates the transcription of cell cycle regulatory genes ctrA, gcrA, dnaA, and ftsZ1 by specifically interacting with the promoter regions to affect cell division.
NtrX is involved in the cell cycle regulation of S. meliloti. Previous studies have suggested that NtrX is a regulator of nitrogen metabolism in bacterial cells because its mutants affect the utilization of nitrogen sources and NtrX is able to regulate amino acid metabolism and nitrogen oxidation [19][20][21][22]25,27]. Decreased utilization of nitrogen sources would inevitably lead to weakened nucleic acid and protein synthesis, which would subsequently suppress the growth and proliferation of bacterial cells. This is one explanation to the effect of ntrX mutations on bacterial cell division. In E. chaffeensis cells, NtrX affected the stability of the CtrA protein through a post-translational mechanism [27], indicating that NtrX may act directly on the cell cycle regulatory system to regulate cell division. This work was carried out using S. meliloti as the study material and reveals the transcriptional control mechanism of the cell cycle regulatory genes mediated by the NtrX protein.  5 and 6).
Transcriptional regulation of ctrA, gcrA, and dnaA mediated by NtrX is confident, though the expression differentials are varied from different materials or detected by different methods (Figures 2 and S2). This conclusion was supported by heterogeneous expression and Western blotting results ( Figures S3C and 3). Moreover, the expression results coincide with data of interactions between of NtrX and promoter regions of ctrA, gcrA, and dnaA ( Figures 5 and 6). An NtrX homolog may bind to the promoter region of dnaA in R. sphaeroides based on the published ChIP-seq data [24]. NtrX may bind to the recognition sites that contain a transcription start site to prevent transcriptional initiation of ctrA and ntrY [42] (Figure 6A,C).
NtrX phosphorylation has been reported in B. abortus and C. crescentus, and it is required for the formation of ntrY promoter DNA-NtrX complex in B. abortus [28][29][30]. The same result was obtained in S. meliloti ( Figures 4C and 6A), suggesting that NtrX phosphorylation is conserved in these species. Based on homology and conservativeness of NtrX proteins ( Figure 4A,B), the conserved 53rd aspartate was predicted as the phosphorylation residue. The mutation protein NtrX D53E was not by histidine kinase in S. meliloti cells ( Figure 4E), confirming that D53 is the real phosphorylation site. The NtrX D53E may mimic the phosphorylated NtrX protein to retain partial functions, which is completely different from NtrX D53A and NtrX D53N ( Figure S1).
Phosphorylated NtrX can recognize cis elements on the promoters of downstream regulated genes in bacterial species [24,30]. These cis elements are not completely consistent from different literatures. In B. abortus, the NtrX binding sites CAAN 3-5 TTG have been identified in the promoter region of ntrY by footprinting [30]. In R. sphaeroides, the GCAN 9 TGC motifs have been suggested to be NtrX recognition sites by analyzing ChIP-seq data [24]. These NtrX recognition sites from above two species share the palindromic sequence CAN x TG. We neither found GCAN 9 TGC motifs from the probes specifically binding to NtrX in S. meliloti, nor identified them by analyzing our ChIP-seq data ( Figures 5A and 6). Furthermore, at least one CAAN 2-5 TTG motif located in the promoter regions of cell cycle regulatory genes such as ctrA, gcrA, and dnaA were found ( Figure S5). These observations are not only consistent with our footprinting assays of the promoters of visN [36], but also with our EMSA results (Figures 6 and S5). Interestingly, when the NtrX binding site is overlapped with one of transcriptional start sites of ctrA and ntrY, the expression of these genes is downregulated by NtrX [42] (Figures 2, 3, and 6). The possible explanation is that NtrX binding to the sites prevents transcription initiation of these genes. Although we identified that NtrX binds to the motifs of CAAN 2 TTG of the promoters of visN and ctrA [36] (Figures 6C and S5), but we still do not know why NtrX recognition sites contain the length-varied palindromic sequences.
The upstream kinase of NtrX may be not the cognate kinase NtrY, though the NtrY recombinant kinase from S. meliloti and C. crescentus phosphorylated NtrX protein in vitro [25] ( Figure 4C). We noticed that both ORFs of ntrY and ntrX are overlapped and the repression of ntrY gene expression by NtrX in S. meliloti [31] (Figure 2). The phenotypes of the ntrY deletion mutant did not coincide with the ntrX mutant [31]. These observations are consistent with the report that NtrY may be the phosphatase of NtrX in C. crescentus [43]. Moreover, the expression of ntrY gene (not ntrX) is induced by micro-oxygen in B. abortus [25]. The primary function of NtrX in bacteria was considered to regulate nitrogen metabolism. The nitrogen limitation signal transduction in bacteria is mainly mediated by the NtrB/NtrC system [44], so it cannot be ruled out that the NtrB/NtrC system can regulate the expression of ntrX under nitrogen lacking conditions. Under nitrogen rich conditions, the activity of NtrX may be regulated by an unknown kinase, which may be able to sense the level of combined nitrogen.
In the future, specific NtrX recognition sites on the promoters of other cell cycle regulatory genes (such as divL, cbrA, cckA, and cpdR1) in S. meliloti will be mapped by footprinting, and the details of the sites will be evaluated to help reconstruction of the regulatory network of NtrX. The upstream proteins in response to the nitrogen sources will be identified by immunoprecipitation and high-resolution mass spectrography.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/genes13061066/s1, Figure S1: Colony formation of Sm1021 expressing NtrX containing a D53 substitution on LB/MC/IPTG (1 mM) agar plates. Figure S2: Heat map showing the expression differentials of cell cycle regulatory genes from one RNA-seq assay. Figure S3: Expression differentials of some cell cycle regulatory genes in ntrX mutants. Figure S4: The NtrX protein detected by Western blotting using rabbit anti-NtrX polyclonal antibodies. Figure S5: The possible recognition sites of NtrX in promoter regions of cell cycle regulatory genes. Table S1: Expression differentials of cell cycle regulatory genes evaluated by a RNA-seq experiment. Table S2: DNA fragments from Sm1021 cells precipitated by anti-NtrX antibodies in a ChIP-Seq experiment. Table S3: DNA oligonucleotides used in this study.