ABSTRACT
Cell division of the alfalfa symbiont, Sinorhizobium meliloti, is regulated by the CtrA signaling network. The gene expression of regulatory proteins in the network is affected by nutrient signaling. In this study, we found that NtrX, one of the regulators of nitrogen metabolic response, can directly regulate the expression of several regulatory genes from the CtrA signaling network. Three sets of S. meliloti ntrX mutants, including the plasmid insertion strain, the depletion strain and the substitution of the 53rd aspartate (ntrXD53E) from a plasmid in the wild-type strain (Sm1021), showed similar cell division defects, such as slow growth, abnormal morphology of partial cells and delayed DNA synthesis. Transcript quantitative evaluation indicated that the transcription of genes such as ctrA and gcrA was up-regulated, while the transcription of genes such as dnaA and ftsZl was down-regulated in the insertion mutant and the strain of Sm1021 expressing ntrXD53E. Correspondingly, 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, which were consistent with the transcription data. The transcriptional regulation of these genes requires phosphorylation of the conserved 53rd aspartate in the NtrX protein. The NtrX protein binds directly to the promoter regions of ctrA, gcrA, dnaA and ftsZ1 by recognizing the characteristic sequence CAAN2-5TTG Our findings reveal that NtrX is a novel transcriptional regulator of the CtrA signaling pathway genes, and positively affects bacterial cell division, associated with nitrogen metabolism.
IMPORTANCE Sinorhizobium meliloti infects the host alfalfa and induces formation of nitrogen-fixing nodules. Proliferation of rhizobia in plant tissues and cells is strictly controlled in the early stage of symbiotic interactions. However, the control mechanism is not very clear. Cell division of S. meliloti in the free-living state is regulated by the CtrA signaling network, but molecular mechanisms by which the CtrA system is associated with environmental nutrient signals (e.g., ammonia nitrogen) need to be further explored. This study demonstrates that NtrX, a regulator of nitrogen metabolism, required for symbiotic nodulation and nitrogen fixation by S. meliloti 1021, can act as a transcriptional regulator of the CtrA signaling system. It may link nitrogen signaling to cell cycle regulation in Rhizobium species.
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 (1–6). 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.
Sinorhizobium meliloti is a model strain of rhizobia that can infect the host plant and form nitrogen-fixing 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 (8–11), several legumes do not produce NCR peptides. Therefore, there may be other mechanisms that control cell division of symbiotic rhizobia in host plants. Since S. meliloti and C. crescentus belong to α-proteobacteria, based on the research results of C. crescentus, with the aid of DNA sequence homology analysis, many cell cycle regulatory genes such as ctrA, ccrM, cpdR1, divJ, divK, gcrA and pleC, have been identified in S. meliloti (12–16). In addition, the hybrid histidine kinase CbrA is linked to the CtrA regulatory system, which is an important regulator of cell division in S. meliloti (17, 18). However, it is still unclear whether these regulatory proteins conduct environmental nutrition signals (e.g., ammonia nitrogen) and whether they play a regulatory role in the symbiotic process.
The NtrY/NtrX two-component system, first discovered in Azorhizobium caulinodans,regulates nitrogen metabolism under free-living conditions and affects nodulation and nitrogen fixation in the host plant Sesbania rostrata (19). Subsequently, ntrY/ntrX homologous genes were found in Rhizobium tropici to regulate nitrogen metabolism and symbiotic nodulation (20). NtrY/NtrX homologs regulate nitrate uptake in Azospirillum brasilense and Herbaspirillum seropedicae (21, 22), and this regulatory system has been found to simultaneously control nitrogen metabolism and cellular redox homeostasis in Rhodobacter capsulatus (23). Moreover, NtrX is involved in the regulation of cell envelope formation in R. sphaeroides (24). In Brucella abortus, the histidine kinase NtrY participates in micro-oxygen signaling and nitrogen respiration (25), while the response regulator NtrX controls the expression of respiratory enzymes in Neisseria gonorrhoeae (26). Interestingly, the NtrY/NtrX system regulates cell proliferation, amino acid metabolism and CtrA degradation in Ehrlichia chaffeensis (27). Finally, NtrX is required for the survival of C. crescentus cells and its expression is induced by low pH (28). These findings show that NtrY/NtrX appears to be a regulatory system for nitrogen metabolism, which may be involved in the regulation of cell division.
NtrX is an NtrC family response regulator protein, which consists of a receiver (REC) domain and a DNA-binding domain (29, 30). X-ray crystal diffraction results indicate that the NtrX protein of B. abortus can form a dimer; the REC domain is composed of 5 α-helices and 5 β-sheets; the DNA-binding domain contains an HTH motif, which includes 4 α-helices. The three-dimensional structure of the C-terminus has not been resolved (30). In vitro experiments demonstrated that the NtrX protein of B. abortus can recognize and bind to the palindromic DNA sequence (CAAN3-5TTG) in the ntrY promoter region via the HTH motif to regulate gene transcription (29, 30). In S. meliloti 1021, our previous work found that NtrX protein can regulate bacterial growth and proliferation, flagellar synthesis and motility, succinoglycan production, and symbiotic nodulation and nitrogen fixation with the host plant alfalfa (31, 32). In the present study, we investigated the control mechanism by which NtrX regulates S. meliloti cell division at the transcriptional level.
RESULTS
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 wild-type Sm1021 (31). According to the determined growth curve, the doubling time of bacterial cell proliferation was calculated to be 180 mins for SmLL1 compared to 160 mins for the wild-type 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 (Fig. 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 180 mins, and then harvested the cells for flow cytometric analysis. The results showed that most of the Sm1021 cells were haploids, only a few diploids, whereas the most of SmLL cells were diploid (Fig. 1B), indicating a deceleration of replication and segregation of their genomic DNA as compared to the wild type. These observations indicate that the SmLL mutant has cell division defects.
Because the deletion of ntrX may be fatal, the deletion mutant in S. meliloti 1021was not yet successfully screened. Therefore, we constructed a depletion strain that the ntrX gene on the genome has been deleted, but carries an IPTG inducible-expression ntrX gene from a plasmid (ΔntrX/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 (Fig. 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 (Fig.1F). After the one-hour induction of IPTG, the peaks were similar to the wild-type cells (Fig. 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), the putative phosphorylation site is predicted as the conserved 53rd aspartate residue (D53) on the REC domain (Fig. 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 successfully to screen 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 NtrXD53A or NtrXD53N almost did not form visible colonies with IPTG induction; however, the strain expressing NtrXD53E or NtrX formed many colonies in the same condition (Fig. S1). GFP-labeled S. meliloti cells (35) cultured in LB/MC/IPTG broth up to the logarithmic phase were observed under a fluorescence microscope, and more than 20% of Sm1021/pntrXD53E cells had abnormal morphology, while Sm1021/pntrX cells were almost normal (Fig. 1G). The growth curve determination also showed that the growth of Sm1021/pntrXD53E in LB/MC/IPTG broth was apparently slower than that of Sm1021/pntrX (Fig. 1G). Synchronized S. meliloti cells were subcultured into LB/MC/IPTG broth and grown for 180 mins 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/pntrXD53E cells, including haploid and diploid (Fig. 1H). These results suggest that the D53 phosphorylation of the NtrX protein is required for the regulation of cell division of S. meliloti.
Transcription of cell cycle regulated genes under the regulation of NtrX in S. meliloti
Since NtrX is involved in controlling cell division of S. meliloti, does it regulate the transcription of cell cycle regulatory genes associated with CtrA system? To test this possibility, we performed a preliminary RNA-Seq analysis between Sm1021 and SmLL1 cells. The results indicated that many genes of cell cycle regulation such as chpT, sciP, dnaA, ftsZ1, ccrM, podJ1, cckA, cbrA, pleD, divK, cpdR1, mucR and clpP were differentially expressed in the mutant strain compared with the wild-type strain (Fig. S2 and Table S1), suggesting that transcription of many cell cycle regulatory genes is regulated by NtrX.
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 synchronized Sm1021 and SmLL1 cells were subcultured into LB/MC broth for shaking incubation, and then total RNA was extracted from cells grown for every half an hour. 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 the ntrY gene cyclical transcription trend was not obvious (Fig. 2A). Known cell cycle regulatory genes, such as ctrA, gcrA and dnaA, also exhibited a cyclical transcription trend (Fig. 2A and S3A). Compared to the wild-type cells, transcript levels of the ntrX gene were significantly low 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 down-regulation trend (Fig. 2A and S3A). Contrary to these results, ctrA was gradually up-regulated in the SmLL1 cells, and gcrA, ccrM and ntrY were significantly upregulated at the same time (Fig. 2A and S3A). These findings suggest that the NtrX protein may repress the transcription of genes such as ctrA and gcrA and activate the transcription of genes such as dnaA and ftsZ1.
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 (Fig. 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 (Fig. 2B and S3B). The transcript levels of cell cycle regulatory genes were significantly altered in depleted cells cultured in the broth with IPTG for 2 or 3 h compared to the cells cultured in the broth without IPTG: the transcription of ctrA, gcrA and ccrM was down-regulated; the transcription of dnaA, ftsZ1, pleC, chpT and cpdR1 was up-regulated (Fig. 2B and S3B). These results further confirm that the NtrX protein represses the transcription of genes such as ctrA and gcrA and simultaneously activates the transcription of genes such as dnaA and ftsZ1.
We analyzed the transcript levels of major cell cycle regulatory genes in Sm1021/pntrXD53E and Sm1021/pntrX cells to determine whether the conserved D53 residue on NtrX is essential for transcriptional regulation. The qRT-PCR results showed that transcripts of the ntrX gene were 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 were significantly reduced or tended to decrease, while the transcripts of dnaA and ftsZ1 genes were significantly increased (Fig. 2C). Contrary to the above results, as transcripts of the ntrXD53E gene increased significantly under IPTG induction, transcripts of ctrA, gcrA,and ntrY also increased significantly, while transcripts of dnaA and ftsZ1 genes decreased significantly (Fig. 2C). These results further confirm that the NtrX protein represses the transcription of genes such as ctrA and gcrA and activates the transcription of genes such as dnaA and ftsZ1, and that this regulation depends on the D53 residue on NtrX.
To determine whether the expression of the CtrA system genes 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 (Fig. S3C). In contrast, the activity of the dnaA promoter is apparently elevated in the cells co-expressing ntrX compared with those cells carrying pSRK-Gm (Fig. S3C). These data supported the conclusion that NtrX negatively controls transcription of ctrA and gcrA, but positively regulates transcription of dnaA.
Negative regulation of protein levels of CtrA and GcrA by NtrX in S. meliloti
To determine whether protein levels of key cell cycle regulators are affected by the ntrX mutation, we first expressed His-tagged NtrX, CtrA and GcrA proteins in E. coli. After purification by 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 (Fig. 3A). Unlike Sm1021, the total NtrX protein level in SmLL1 cells was apparently reduced and tended to increase gradually at different growth times (Fig. 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 were apparently increased in SmLL1 cells compared to Sm1021 cells, (Fig. 3A). These results indicate that SmLL1 is a down-regulated mutant of ntrX and that NtrX protein levels are negatively correlated with CtrA and GcrA proteins.
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 high-level expressed NtrX protein (Fig. 3B). CtrA and GcrA proteins were high-level expressed in cells cultured in LB/MC broth without IPTG, whereas their levels were apparently reduced in cells cultured in broth containing IPTG for 1 to 3 h (Fig. 3B). These results also prove that NtrX protein levels are negatively correlated with CtrA and GcrA proteins.
To determine whether the D53 residue on the NtrX protein affects the protein levels of CtrA and GcrA, we performed immunoblot assays of lysates from Sm1021/pntrXD53E and Sm1021/pntrX cells. The results showed that the protein levels of NtrX and NtrXD53E increased apparently when cultured in LB/MC broth containing IPTG for 23 h (Fig. 3C). Under the same culture conditions, the protein levels of CtrA and GcrA were reduced somewhat in Sm1021/pntrX cells, while they were elevated to some extent in Sm1021/pntrXD53E cells (Fig. 3C). These results reaffirm that NtrX protein negatively regulates the expression of CtrA and GcrA in the dependent manner of the D53 residue.
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 (Fig. 4C). After mutating the 53rd aspartate to glutamate, phosphorylated NtrX protein was not detected by treatment of acetyl phosphate (data not shown). To further determine whether NtrX is phosphorylated in vivo, western blotting assays were performed using anit-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 (Fig. 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/pntrXD53E 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/pntrXD53E cells tended to decrease (Fig. 4D). These results reveal that the D53 residue of NtrX is phosphorylated in S. meliloti cells.
Direct binding of phosphorylated NtrX protein to the promoter DNA of key 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 (Fig. 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 (Fig. 5A). After sequence analysis in detail, we found that the promoter DNA fragments of cell cycle regulatory genes such as ctrA, dnaA, mucR and cpdR1 were specifically enriched (Fig. 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 them in the precipitated DNA fragments, and found that nine of possible motifs are located in the promoter regions of cell cycle regulatory genes such as ctrA, danA, gcrA and ftsZ1(Fig. 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 different degrees (Fig. 5C), indicating that the NtrX protein in Sm1021 cells can interact directly with 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 (Fig. 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 (Fig. 6A). To determine whether the D53 phosphorylation of the NtrX protein is involved in the protein-DNA binding reaction, we replaced the phosphorylated NtrX protein with the NtrXD53E protein. The gel retardation results showed almost no protein-DNA complex formation (Fig. 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 biotin-labeled probe of dnaA promoter DNA (containing the CAAAACCCTTG motif) and found that they bound specifically to form a protein-DNA complex (Fig. 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 (Fig. 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 (Fig. 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.
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 can affect cell growth and division. However, the molecular mechanism by which combined nitrogen levels regulate 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 such as ctrA, gcrA, dnaA and ftsZ1 by specifically interacting with the promoter regions, to promote cell division (Fig. 7), which provides a preliminary answer to the above question.
NtrX is a bacterial cell cycle regulator. 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 (19–22, 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 for the first time. This conclusion is supported by multiple experimental evidences: 1) three sets of ntrX gene mutation materials are defective in bacterial growth, cell morphology and genomic DNA synthesis (Fig. 1); 2) the transcript levels of cell cycle regulatory genes such as ctrA, gcrA, dnaA and ftsZ1 are differentially altered in ntrX mutants (Fig. 2; 31; 3) the protein levels of CtrA and GcrA are correspondingly altered in the mutant and they were negatively correlated with the level of NtrX protein (Fig. 3); 4) the phosphorylated NtrX protein binds directly to the promoter regions of ctrA, gcrA, dnaA and ftsZ1 (Fig. 5–6).
Cell division defects showed a little difference for three sets of ntrX mutants. For example, the fewest cells of abnormal shapes were observed for SmLL1 cells, the most cells were found for the depletion cells (Fig. 1A-B, D, G). In fact, the depletion cells are easy to die in LB/MC medium without IPTG addition. Even in LB/MC broth containing different concentrations of IPTG, it still grew slower than the wild-type strain, Sm1021 (Fig. 1E). Irregular cell shapes of the ntrX mutant may be associated with genomic DNA content. We noticed that the flow cytometric peak of haploid cells is very sharp for Sm1021 or Sm1021/pntrX, but it was not for SmLL1, the depletion strain and Sm1021/pntrXD53E (Fig. 1C, F, I).
Differential expression of cell cycle regulatory genes exhibited some difference for three sets of ntrX mutation materials. First, qRT-PCR data showed larger expression differentials between SmLL1 and Sm1021 than those of the depletion strain (with and without IPTG induction) and the Sm1021/pntrXD53E strain (compared with Sm1021/pntrX, Fig. 2 and S3A-B). It may be result from the different cell cycle status of tested cells. Secondly, preliminary transcriptomic data showed more cell cycle regulatory genes differentially expressed between the synchronized cells of SmLL1 and Sm1021 than those cells subcultured in LB/MC broth (Fig. S2). Expression differentials of these genes from the cells subcultured in LB/MC broth were further determined by qRT-PCR (Fig. 2A and Fig. S3A), since this method is more sensitive for mRNA level analysis than RNA-seq from our experience.
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 (Fig. 2 and S2). This conclusion was supported by heterogeneous expression and western blotting results (Fig. S3C and 3). Moreover, the expression results coincide with data of interactions between of NtrX and promoter regions of ctrA, gcrA and dnaA (Fig. 5–6). 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 (37) (Fig. 6A, 6C). We noticed that the bands of the complex between NtrX and the gcrA probe were relatively weak, which may be associated with the selected probe (Fig. 6D).
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 (28–30). The same result was gained in S. meliloti (Fig. 4C and 6A), suggesting that NtrX phosphorylation is conserved in these species. Based on homology and conservativeness of NtrX proteins (Fig. 4A-B), the conserved 53rd aspartate was predicted as the phosphorylation residue. The mutation protein NtrXD53E was neither phosphorylated by acetyl phosphate in vitro (data not shown), nor by histidine kinase in S. meliloti cells (Fig. 4E), confirming that D53 is the real phosphorylation site. The NtrXD53E may mimic the phosphorylated NtrX protein to retain partial functions, which is completely different from NtrXD53A and NtrXD53N (Fig. 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 CAAN3-5TTG have been identified in the promoter region of ntrY by footprinting (30). In R. sphaeroides, the GCAN9TGC 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 CANxTG. We neither found GCAN9TGC motifs from the probes specifically binding to NtrX in S. meliloti, nor identified them by analyzing our ChIP-seq data (Fig. 5A, 6). Furthermore, at least one CAAN2-5TTG motif located in the promoter regions of cell cycle regulatory genes such as ctrA, gcrA, and dnaA were found (Fig. S5). These observations are not only consistent with our footprinting assays of the promoters of visN (36) and dnaA (data not shown), but also with our EMSA results (Fig. 6, 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 (37) (Fig. 2, 3, 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 CAAN2TTG of the promoters of visN and ctrA (36) (Fig. 6C, S5), but we still don’t 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) (Fig.4C). We noticed that both ORFs of ntrY and ntrX are overlapped and the repression of ntrY gene expression by NtrX in S. eliloti (31); Fig. 2). The phenotypes of the ntrY deletion mutant did not coincide with the ntrX mutant (16). These observations are consistent with the report that NtrY may be the phosphatase of NtrX in C. crescentus (38). 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 (39), 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.
MATERIALS AND METHODS
Strains and culture medium
Escherichia coli DH5α and BL21 were cultured in LB medium at 37 °C. S. meliloti (including Sm1021, SmLL1, ΔntrX/pntrX and derivatives) (31) were cultured in LB/MC medium at 28 °C. MOPS-GS broth was utilized for the cell synchronization of S. meliloti (33). The following antibiotics were added to the medium: kanamycin (Km), 50 μg/ml; gentamicin (Gm), 10 μg/ml; chloramphenicol (Cm), 30 μg/ml; neomycin (Nm), 200 μg/ml; streptomycin (Sm), 200 μg/ml.
Recombinant plasmid construction
Primers PntrX1 and PntrX2 bearing HindIII and XbaI digestion sites were used to amplify the S. meliloti ntrX gene (Table S3). The ntrX gene fragment was amplified using overlapping PCR primers NMF and NMR with the substitution of aspartate to glutamate, asparagine or alanine (Table S3). Overlapping PCR was performed as described by Wang (31) in 2013. The PCR products were digested with HindIII and XbaI (Thermo) and ligated with digested pSRK-Gm (34) to obtain the recombinant plasmids pntrX, pntrXD53E, pntrXD53A and pntrXD53N. Each plasmid was transferred into Sm1021 to gain merodiploids.
After introducing pntrX into SmLL1 by triparental mating with help of MT616, the cells were streaked on LB/MC/Sm agar plates containing 1 mM IPTG and 10% sucrose to screen the ntrX depleted cells. The depletion strain (ΔntrX/pntrX) was identified by PCR with the primers of PntrYk1 and PntrX2 (Table S3).
Primers PntrYk1/2, PntrXD53E1/2, PctrA1/2, and PgcrA1/2 were used to amplify the NtrY kinase domain, ntrXD53E, ctrA and gcrA, respectively (Table S3). The DNA fragments amplified by high fidelity PCR (Takara) were digested with the appropriate restriction enzymes, and ligated into pET28b (Sangon) to obtain pntrYk, pntrXD53E, pctrA and pgcrA, use for recombinant protein purification. The cloned genes on the plasmids were identified by DNA sequencing (Sangon).
Primers PctrAp1/2, PgcrAp1/2, PdnaAp1/2 were used for amplification the promoter regions of the ctrA, gcrA and dnaA, respectively (Table S3). The PCR fragments were digested by appropriate restriction enzymes and ligated into pRG960 (40) to gain the recombinant plasmids pPctrA, pPgcrA and pPdnaA.
Bacterial cell synchronization
De Nisco’s method was used for bacterial cell synchronization (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. 100 μl of the bacterial culture was transferred into 100 ml LB/MC broth and shaken cultured overnight until OD600 = 0.1-0.15. The cells were collected by centrifugation (6,500 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 (6,000 rpm, 5 min, 4°C), and washed twice with DEPC-treated water. RNA extraction was performed using 1 ml of Trizol (Life Technology). Total RNA was treated with genomic DNA Eraser (Takara) 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; ddH2O, 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-ΔΔCT 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 (4) using rabbit anti-NtrX polyclonal antibodies prepared by Wenyuange, Shanghai (36). In brief, Sm1021 cells (2ml, OD600 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™) on ice using 15 bursts of 30 sec (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 (1:1,000 dilution), incubated overnight at 4°C with 50 μl of protein-A agarose beads pre-saturated by BSA, 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 NaHCO3) supplemented with NaCl to a final concentration of 300 mM, and incubated for 6 h or overnight at 65°C to reverse the crosslinks. DNA was purified by a MinElute kit (QIAGEN) and resuspended in 40 μl of Elution Buffer. DNA sequencing was completed using Illumina HiSeq 2000 in BGI. 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 (6,000 rpm, 5 min, 4°C), washed twice with a 0.85% NaCl solution (stored at 4 °C). 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 hours. 1 μl of 10 μM SYTOX Green dye (Sigma) was added to each sample. Each sample was assessed using a MoFlo XDF (Beckman Coulter) flow cytometer, and the results were analyzed by Summit 5.1 software (Beckman Coulter).
EMSA (electrophoretic mobility shift assay)
EMSA was performed as described by Zeng (36). 30 μl of the purified NtrX protein solution (200 ng/μl) was incubated with 20 μl of 100 mM acetyl phosphate (Sigma) in 50 μl of 2X buffer (100 mM Tris-HCl pH 7.6, 100 mM KCl, 40 mM MgCl2) for 1 h at 28 °C. The remaining acetyl phosphate was removed by ultra-filtration (10 KD Amicon Ultra 0.5, Millipore). The protein-DNA binding reaction (20 μl) included 3, 6, 15 ng phosphorylated NtrX protein, 2 or 40 nM DNA probe, 1X binding buffer, 5 mM MgCl2, 50 ng/μl poly(dI·dC), 0.05% NP-40, 1% glycerol, and ddH2O (up to 20 μl). The mixture was incubated for 30 min at 28 °C, after which 1 μl of loading buffer was added for PAGE. The protein-DNA complexes were transferred onto a nylon membrane (Thermo) and irradiated with a 254 nm UV lamp for 10 min. The protein-DNA complexes were detected using a Light Shift Chemiluminescent EMSA Kit (Thermo). Probes of ntrY, ctrA, dnaA, gcrA and ftsZ1 promoter DNA labeled with biotin were synthesized in Invitrogen, Shanghai, and listed in Table S3.
NtrX phosphorylation assay and western blotting
The procedure of NtrX phosphorylation assays was modified from Pini (15). 1 mg His-NtrX fusion protein (NtrXr) and His-NtrY kinase domain fusion protein (NtrY-Kr) purified through a Ni2+column were used for in vitro phosphorylation assays. 2 mM acetyl phosphate (Sigma) was mixed with 300 μg NtrY-Kr in 1 ml of phosphorylation buffer (50 mM Tris-HCl pH 7.6, 50 mM KCl, 20 mM MgCl2) and then incubated for 1 h at room temperature. The remaining acetyl phosphate was removed using an ultra-filtration tube (10 KD Amicon Ultra 0.5, Millipore). 1, 3 and 10 μg phosphorylated NtrY-Kr protein were added to 200 μl of phosphorylation buffer containing 10 μg NtrXr, and incubated overnight at 28 °C. Samples were separated by a Phos-Tag™ Acrylamide SDS-PAGE gel (Mu Biotechnology, Guangzhou). The gel was prepared by mixing 50 μM Phostag™ acrylamide (29:1 acrylamide: N, N’’-methylene-bis-acrylamide) with 100 μM MnCl2.
Synchronized S. meliloti cells (Sm1021, SmLL1, Sm1021/pntrX and Sm1021/pntrXD53E) 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 as Tang (41), with rabbit anti-NtrX (1:10000) antibodies (36). Chemiluminescent detection was performed using an ECL fluorescence colorimetric kit (Tiangen) and fluorescent signals were visualized using a Bio-Rad Gel Doc XR. Band intensities were evaluated by Image J (42).
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, the synchronized cells of Sm1021/pntrX and Sm1021/pntrXD53E were subcultured in 100 ml of LB/MC broth containing 1 mM IPTG for one to three hours. ~ 108 cells were collected by centrifugation every half or one hour for each strain. 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 (OD600 = 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 (35) 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 (43). DNA library was prepared including DNA-end repair, 3’-dA overhang, ligation of methylated sequencing adaptor, PCR amplification and size selection (usually 100-300bp, including adaptor sequence). Bioinformatics analysis was performed as follows. The ratio of N was over 10% in whole read. Removed reads in which unknown bases are more than 10%. The ratio of 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 was then mapped to the reference genome by SOAP aligner/SOAP2 (Version: 2.21t). BWA (Burrows-Wheeler Aligner, Version: 0.7.10) is 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 Possion 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 < le-05. Peak Calling Parameter: macs14 −s 50 −g 6691694 −p 1e-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 (44). The 3D structures of NtrX were analyzed by the software Pymol (Delano Scientific).
AUTHOR CONTRIBUTIONS
L. L. designed research; S. X., L. Z., F. A., X. Y., L. H., S. Z., W. Z., and N. L. performed research; S. X., F. A., J. Y., L. Y. and L. L. analyzed data; L. L. and K. O. wrote the paper.
ACKNOWLEDGMENTS
This research was supported by the National Natural Science Foundation of China (31570241 to L.L). We thanks to Dr. Yiwen Wang (Eastern Normal University) for help of SEM.