Characterization of cis-Acting Sites Controlling Arginine Deiminase Gene Expression in Streptococcus gordonii
http://www.100md.com
《细菌学杂志》
Department of Oral Biology, University of Florida, Gainesville, Florida
ABSTRACT
The arginine deiminase system (ADS) is responsible for the production of ornithine, CO2, ammonia, and ATP from arginine. The ADS of the oral bacterium Streptococcus gordonii plays major roles in physiologic homeostasis, acid tolerance, and oral biofilm ecology. To further our understanding of the transcriptional regulation of the ADS (arc) operon, the binding of the ArcR transcriptional activator, which governs expression of the ADS in response to arginine, was investigated by DNase I protection and gel mobility shift assays. An ArcR binding sequence was found that was 27 bp in length and had little sequence similarity to binding sites of other arginine metabolism regulators. The presence of arginine at physiologically relevant concentrations enhanced the binding of ArcR to its target. Using cat fusions, various deletion and substitution mutations within the putative ArcR footprint were shown to cause dramatic reductions in expression from the arcA promoter in vivo, confirming that the 27-bp sequence is required for optimal expression and induction of the ADS by arginine. Mutation of two putative catabolite response elements (CREs) within the arc promoter region showed that both CREs contribute to catabolite repression. A thorough understanding of the regulation of the ADS in S. gordonii and related organisms is needed to develop ways to exploit arginine catabolism for the control of oral diseases. Identification of the ArcR and CcpA binding sites lays the foundation for a more complete understanding of the complex interactions of multiple regulatory proteins with elements in the arc promoter region.
INTRODUCTION
Bacteria in the oral cavity are subjected to constantly changing environments, with major fluctuations in nutrient availability and pH occurring with cycles of feeding and fasting by the host. Following ingestion of carbohydrates, the pH of oral biofilms can drop within minutes from above 7 to below 4 as a result of glycolysis by acidogenic bacteria, resulting in demineralization of teeth (23, 26). Oral bacteria have developed a variety of strategies to cope with intermittent and sustained acidification through complex and coordinated cellular responses that enhance acid tolerance, including conversion of salivary and dietary substrates to basic compounds (2, 5). Ammonia generation from hydrolysis of urea by urease and arginine catabolism by the arginine deiminase system (ADS) are believed to be crucial for oral biofilm pH homeostasis and are thought to be major factors promoting dental health (2).
The ADS is a three-enzyme system, consisting of arginine deiminase, catabolic ornithine carbamyltransferase, and carbamate kinase enzymes, which together convert arginine to ornithine, ammonia, and CO2, with concurrent production of ATP (6). There is considerable conservation in the gene organization and primary structure of the enzymes of the ADS, but there is also diversity in the regulation and physiological role of the pathway (8). The most thoroughly studied ADS, that of the gram-negative bacterium Pseudomonas aeruginosa, is only expressed under anaerobic conditions, and the presence of arginine enhances the expression of ADS genes (12, 18). Two regulators, ANR and ArgR, are required for responding to low oxygen tension and sensing exogenous arginine, respectively. Separate binding sites for ANR and ArgR have been identified in the promoter region of the P. aeruginosa arc operon (17). Similarly, the ADS in Bacillus licheniformis is induced by anaerobiosis, but only when arginine is present (20, 21). In both of these organisms, the ADS appears to function primarily in ATP generation when aerobic respiration cannot be used for growth. In contrast, the ADS of oral streptococci, including S. gordonii, Streptococcus sanguis, and Streptococcus rattus, appears to be a primary defense against killing by acidification (2). In these organisms, which lack a respiratory chain, the ADS is subject to carbon catabolite repression (CCR) and arginine acts as an induction signal to further increase expression (8-10, 14).
In a previous study from our laboratory, the ADS of S. gordonii was found to be induced by arginine and subject to CCR, with very low levels of expression observed in the presence of the repressing sugar glucose (8, 9). Deletion of the ADS regulatory gene arcR resulted in greatly reduced levels of arginine deiminase activity and arc transcription, while disruption of the catabolite control protein, CcpA, greatly reduced CCR of the ADS. The ADS was also found to be further induced by anaerobiosis through an Fnr-like protein (Flp) (8, 9). Significant levels of homology were found between ArcR of S. gordonii and other transcriptional regulators of arginine metabolism, including ArgR of Escherichia coli and B. licheniformis as well as AhrC of B. subtilis (8).
S. gordonii is one of relatively few ADS-positive oral bacteria, but it colonizes the mouth very early in life and remains abundant in healthy oral biofilms (4). S. gordonii has been suggested to play important roles in maintaining pH homeostasis in oral biofilms, and the ADS has been proposed to have significant potential to be exploited to prevent dental caries in humans (2). In this report, we expressed a recombinant ArcR protein and studied its interaction with the arcA promoter region to gain insight into the molecular basis for differential expression of the ADS in response to arginine and other environmental influences.
MATERIALS AND METHODS
Bacterial strains, growth conditions, and reagents. E. coli strains were grown in Luria-Bertani (LB) medium supplemented with antibiotics at the following concentrations: ampicillin (Ap; 100 μg ml–1), kanamycin (Km; 40 μg ml–1), erythromycin (Em; 300 μg ml–1), spectinomycin (Sp; 50 μg ml–1), and chloramphenicol (Cm; 20 μg ml–1). S. gordonii DL1 and its derivatives were maintained in brain heart infusion broth (BHI; Difco Laboratories, Detroit, Mich.) at 37°C in 5% CO2 and 95% air, with antibiotics added at the following concentrations when necessary: Km (250 μg ml–1), Em (5 μg ml–1), and Sp (250 μg ml–1). S. gordonii cells intended for chloramphenicol acetyltransferase (CAT) assays were cultured in tryptone yeast extract (TY) base medium supplemented with 10 mM glucose or galactose and 50 mM arginine, as specified. All chemical reagents and antibiotics were obtained from Sigma Chemical Co. (St. Louis, Mo).
DNA manipulations. To construct a ParcA::cat reporter gene fusion, a 337-bp fragment upstream of the arcA start codon (ParcA) was amplified by primers ParcA5' and ParcA3' (Table 1) and fused with a promoterless chloramphenicol acetyltransferase (cat) gene. After confirming by sequencing, the ParcA::cat construct was introduced into the S. gordonii genome by integration into the gtfG locus (9). Other ParcA::cat mutants were engineered by site-directed mutagenesis. Briefly, pairs of primers with designed deletions or mutations (Table 1) were synthesized and used in recombinant PCR with the original primers (ParcA5' and ParcA3') to generate new fragments with the desired deletions or mutations. These fragments were then fused with cat and integrated into the S. gordonii genome in the same way. All deletions and mutations were confirmed by sequencing. All primers were synthesized by Integrated DNA Technologies (IDT), Inc. (Coralville, IA).
Expression and purification of recombinant ArcR in E. coli. The arcR gene wasamplified by PCR using primers arcRBamHIPUR and arcRPstIPUR (Table 1) with BamHI and PstI sites added to the 5' and 3' ends, respectively. The amplified arcR gene was digested with BamHI and PstI and ligated into the expression vector pQE30 (QIAGEN, Inc., Valencia, CA), which allowed in-frame fusion of six histidine residues to the N terminus of the recombinant ArcR protein. The resulting plasmid, pQE30-ArcR, was introduced into E. coli M15. Expression of recombinant ArcR was induced with isopropyl--D-thiogalactopyranoside (IPTG) at a final concentration of 1 mM. After 3 h of IPTG induction, protein lysates were prepared by homogenization and the recombinant protein was purified by nickel affinity chromatography as recommended by the supplier. Purified ArcR was dialyzed against 1 liter of phosphate-buffered saline buffer (pH 7.4) and then stored at 4°C for up to 1 month.
Electrophoretic mobility shift assay. A 193-bp fragment upstream of the arcA start codon was amplified by PCR using primers arcRFT-5' and arcRFT-3' (Table 1). The PCR product was cloned into the TA-cloning vector pGEM-T Easy (Promega, Madison, WI), resulting in plasmid pLZ0402, and then released as a 213-bp fragment by EcoRI digestion. The EcoRI fragment was end labeled with [-32P]ATP using T4 polynucleotide kinase and was used as a probe in the mobility shift assays. The reaction mixture (10 μl) contained 50 fmol labeled DNA probe, 5 mM MgCl2, 50 mM KCl, 1 mM EDTA, 0.2 μg μl–1 poly(dI-dC), 10 mM HEPES (pH 7.9), 5 mM dithiothreitol, and 10% glycerol. Purified recombinant ArcR, L-arginine, or L-lysine was added to the mixture at the specified concentrations. The mixture was incubated at room temperature for 20 min, followed by electrophoretic separation through a 4% low-ionic-strength polyacrylamide gel. The gel was dried on Whatman 3 MM paper and exposed to X-ray film.
DNase I protection (footprinting) assay. The 213-bp DNA probe used in the DNase I protection assay was released from plasmid pLZ0402 with EcoRI. It was then purified and treated with shrimp alkaline phosphatase before labeling with [-32P]ATP using T4 polynucleotide kinase, followed by another digestion with either HindIII or EcoRV, as these sites were engineered into the primers arcRFT-5' and arcRFT-3', respectively (Table 1). The DNA probes with only one radiolabeled end were then purified using the QIAquick gel extraction kit. DNase I digestion in the presence of recombinant ArcR and subsequent analysis by polyacrylamide gel electrophoresis were carried out according to a standard protocol with minor modifications (11). Briefly, 20 fmol of the DNA probe was added into a 25-μl system containing 0 or 1 pmol of purified ArcR, 40 ng μl–1 poly(dI-dC), and 50 mM Tris-HCl, pH 8.0, and incubated at room temperature for 20 min. Fifty microliters of a 10 mM MgCl2, 5 mM CaCl2 solution was then added, followed by 1 μl of DNase I at 0, 10–3, 10–2, or 10–1 dilutions from the original concentration (114 U μl–1; Invitrogen, Carlsbad, CA). After 2 min of incubation at room temperature, 75 μl of stop solution (20 mM EDTA, 1% sodium dodecyl sulfate [SDS], 0.2 M NaCl, 125 μg ml–1 yeast tRNA) was added, and the DNA was precipitated after addition of 375 μl of ethanol. Digested DNA samples were then analyzed along side of a sequencing ladder that was generated using the Sequenase sequencing kit (USB, Cleveland, Ohio) and pLZ0402 as a template.
Chloramphenicol acetyltransferase (CAT) assay. Bacterial cells with various cat fusions were grown under different conditions to mid-exponential phase (optical density at 600 nm [OD600] 0.6 to 0.7) and washed twice with 10 mM Tris buffer (pH 7.8). Cells were then disrupted in a Bead Beater for 20 s, performed twice with a 2-min interval on ice. The cell lysate was centrifuged at 18,000 x g for 10 min, and the supernatant was recovered and used for measuring CAT activity using the method of Shaw (24). The concentration of the protein in the lysate was determined using the Bradford protein assay (1).
RESULTS AND DISCUSSION
Purified ArcR binding to the promoter region of arcA in vitro. To study the interaction between ArcR and its binding elements in vitro, the entire arcR coding sequence was cloned into plasmid pQE30 to create a translational fusion adding six histidine residues to the N terminus of ArcR, and His-ArcR was purified by nickel affinity chromatography. The purified protein displayed the expected apparent molecular mass of 18 kDa on SDS-polyacrylamide gel electrophoresis (Fig. 1A). To prove that purified ArcR had the ability to bind to the arcA promoter region, electrophoretic mobility shift assays were performed with purified ArcR and a 213-bp -32P-labeled DNA fragment, designated USR, containing about 0.2 kbp upstream of the arcA transcription initiation site (TIS) (Fig. 2). Without His-ArcR, the probe migrated to the bottom of the gel. When recombinant ArcR was added, the movement of USR was retarded (Fig. 1B). Increasing concentrations of His-ArcR caused more USR signal to shift upward and to shift to higher molecular weight levels, creating multiple slower-migrating bands on the gel. These results indicated that the recombinant ArcR was able to bind to the arcA promoter region and suggested that more than one molecule of His-ArcR may bind to the same DNA probe. ArgR of E. coli, one of the proteins with considerable similarity to ArcR of S. gordonii, has been crystallized, and an analysis of its structure revealed a symmetric hexamer formed by six identical ArgR molecules that was stabilized by the presence of arginine (27).
As an additional control, an irrelevant His-tagged protein, an autolysin from S. mutans, was used in a mobility shift assay in place of His-ArcR (data not shown). No mobility shift of USR was detected when this protein was used at concentrations similar to those for His-ArcR, indicating that the His tag did not play a role in the binding of recombinant His-ArcR to its target DNA.
Arginine enhances binding of ArcR to the arcA promoter region in vitro. ArcR shares considerable similarity with regulatory proteins that can interact with arginine in a way that alters protein conformation and binding affinity for DNA and ultimately changes in transcription of the target promoter (27). S. gordonii ArcR contains several amino acid residues that are conserved with arginine binding sites defined by studies of the E. coli arginine repressor ArgR; namely, glutamine at position 104 and aspartic acid residues at positions 125 and 126 (8). As Dong et al. have previously reported, expression of the arc operon of S. gordonii DL-1 is substantially enhanced by the addition of arginine to the growth medium in an ArgR-dependent manner (8). To determine whether arginine could influence the binding of ArcR to its target, arginine (arginine-HCl; Sigma) was added into the gel mobility shift assay. As shown in Fig. 3, addition of arginine clearly improved the interaction between ArcR and USR, with more USR being shifted when higher concentrations of arginine were present in the assay. Ina control experiment, lysine instead of arginine was added to the mixture of ArcR and USR before electrophoresis. No apparent effect on the migration of USR was observed in this experiment, suggesting that the enhancing effect of arginine on ArcR-USR binding was not due simply to the presence of a basic amino acid. Notably, arginine concentrations as low as 10 μM were sufficient to enhance binding, a concentration lower than that reported for free arginine in the human mouth (50 μM) (28) and well within the range of known concentrations of intracellular pools of amino acids in bacteria (15).
ArcR binds to a 27-bp region upstream of the arcA promoter. The experimental results from the gel shift assays revealed that the recombinant ArcR protein was active with regard to interaction with its target DNA and arginine. Two 213-bp DNA probes, one end labeled with -32P, were prepared from the same sequence that was used in the gel shift assays. DNA was subjected to DNase I digestion in the presence and absence of ArcR protein. The digested probe was then analyzed by electrophoresis on a sequencing gel. Results from the two strands (Fig. 4A) showed that a 27-bp region (ATATAAAATATGCAAAGAAAACGCTTC), spanning from position –122 to –96 upstream of the arcA transcription initiation site, was protected by ArcR. Three hypersensitive sites divided the footprint into two parts (Fig. 4A). The 27-bp sequence was predicted to have a very limited tendency for formation of secondary structure, and little sequence similarity was detected with known binding sequences of other ArcR-like proteins. The protected sequence also partly overlapped a potential catabolite response element (CRE).
To confirm the results of the footprinting assay, the ability of various DNA oligonucleotides to compete with the -32P-labeled USR probe for binding to ArcR was tested in electrophoretic mobility shift assays. As depicted in Fig. 4B, 1224-5x3, the full-length 27-bp, double-stranded DNA in the ArcR footprint, along with three shorter DNA fragments lacking the left half (1224-#1), the right half (1224-#2), or 6 bp from both ends (1224-d), were used to compete with labeled USR. Each of four DNA fragments was generated from a pair of synthesized single-stranded oligonucleotides by heating to 94°C and cooling slowly to room temperature. As shown in Fig. 4C, 1224-5x3 was able to compete for ArcR binding with radioactive USR, although higher concentrations were required compared to that for cold USR. At the same time, when single-stranded oligonucleotides (1224-5 and 1224-3) were used individually, no competition was detectable at similar concentrations. Furthermore, when the three shorter, double-stranded DNA fragments (1224-#1, 1224-#2, and 1224-d) were individually added to the assay, no competition for binding to ArcR was detected (Fig. 4D). These results indicate that the entire 27-bp sequence is required for efficient binding to ArcR.
Due to the range limitation of the sequencing gel, it was possible that the DNase I protection assay missed other ArcR binding elements within the 213-bp probe. To test for the existence of other ArcR binding elements within the upstream region of the arcA promoter, a probe that contained all of the sequence of the 213-bp probe except the 27-bp target region (USRFT) was engineered and prepared for gel shift assays. When USRFT was used to compete with the radioactive USR in binding to ArcR, no competition activity was detected (Fig. 4E), suggesting that the 27-bp DNA sequence is the only major ArcR binding site in the upstream region of the arcA promoter. However, due to the limited sensitivity of the competition assay, the existence of a low-affinity binding site(s) for ArcR in the arcA upstream region remained a possibility. When radiolabeled USRFT was used to bind to the recombinant His-ArcR in a mobility shift assay, a low residual level of binding activity was detected in this fragment (data not shown), suggesting the presence of a secondary binding site(s) for ArcR in the arcA promoter region. The presence of both high- and low-affinity ArcR binding sites in the arcA promoter region could allow for fine tuning the regulation of the ADS through cooperative binding, perhaps in response to various concentrations of arginine in the environment. In this report, we will only focus on the analysis of the 27-bp high-affinity binding site of ArcR in the arcA promoter region.
Research on the transcriptional regulators of other known bacterial arginine metabolism systems from a variety of genera has revealed binding specificities within the promoter regions of their target genes that share sequence similarity. Among these transcriptional regulators, the binding characteristics of the arginine repressor of E. coli (ArgR) are the best characterized. It has been determined that E. coli ArgR oligomerizes into a hexamer in the presence of arginine and binds to a consensus Arg box of A/TNTGAATAATTATTCANA/T (N represents any base) found in the genes of the arginine regulon (3, 19, 27). ArgR of B. licheniformis has a very similar 18-nucleotide binding site, with 15 nucleotides conserved with the E. coli Arg box, although the purified ArgR protein protected a significantly longer sequence from DNase I digestion (21). Similar Arg boxes have also been found in the arginine biosynthetic operon of B. subtilis (25). Another well-studied arginine deiminase transcriptional regulator, ArgR of P. aeruginosa, binds to two direct-repeat sequences upstream of the arc promoter, matching the Arg box (TGTCGCN8AA) commonly found in other ArgR binding sites of P. aeruginosa (17).
One common feature of the aforementioned Arg box sites is their sequence symmetry indicated by the presence of inverted or direct repeats. Unlike those cis elements regulating arginine metabolism, sequences with a strong tendency to form regions of dyad symmetry or direct repeats were not observed within the S. gordonii ArcR binding sequence. Half of the upstream portion of the sequence did show some symmetry, but this very A:T-rich sequence was not likely to form a very stable structure. Moreover, little sequence similarity was detected between the S. gordonii ArcR binding site and the Arg boxes of E. coli or P. aeruginosa. Our experiments also revealed that instead of having two symmetrical binding sites, as does ArgR in P. aeruginosa, ArcR probably has one high-affinity binding site and an additional low-affinity binding site(s) in the arc promoter region. Given the complexity of the regulation of the S. gordonii ADS and modulation of expression by growth phase, pH (unpublished data), CCR, and oxygen tension, it would not be surprising that sophisticated regulation by arginine involving cooperative binding of ArcR to its target cis elements has been evolved in this bacterium. Additionally, other regulatory proteins, including CcpA and Flp (9), may also interact near to, or overlap with, the ArcR binding sites.
Functional analysis of the ArcR binding site. Based on the knowledge of the interaction of ArcR with the upstream region of the arcA promoter in vitro, we engineered a series of mutants into ParcA::cat, a transcriptional fusion of the arcA promoter to a promoterless chloramphenicol acetyltransferase (cat) gene and monitored expression in vivo in response to environmental stimuli. All strains containing the fusion constructs were grown in BHI broth with appropriate antibiotics overnight and then transferred into fresh TY medium containing 10 mM glucose or galactose as the carbohydrate source, with or without addition of 50 mM arginine. Mid-exponential-phase cultures (OD600 0.6 to 0.7) were harvested and CAT assays were carried out immediately to measure the expression levels of the promoter fusions. In agreement with previous reports, arc operon expression in the wild-type strain was subject to CCR, as indicated by low CAT activity when glucose was used as the main carbohydrate source and increased CAT activity when glucose was replaced with galactose. Likewise, addition of arginine increased CAT activity (Table 2).
Based on the data obtained in the in vitro binding assays, three deletion mutants were engineered (Fig. 5A) with the left half (FTL), the right half (CRE, overlapping with the second CRE), or the entire 27-bp sequence (FT) deleted from the ParcA::cat fusion. Additionally, two substitution mutants were generated with changes in the left half (FTLM) or the right half of the ArcR binding site (FTRM) being mutated in a way that the weak dyadic sequence was disrupted. All constructs were then introduced in single copy into the gtfG gene of the wild-type strain DL-1 and an otherwise-isogenic ccpA mutant in which an erythromycin resistance cassette was inserted to disrupt the ccpA gene. Strains harboring these mutant fusions were cultured to mid-exponential phase in TY medium containing 10 mM glucose or galactose, with 0 or 50 mM added arginine (Table 2).
Strains carrying all three deletion mutants (FTL, CRE, and FT) produced similarly lower levels of CAT activities compared to the wild-type ParcA::cat fusion in either the DL-1 or the ccpA mutant backgrounds, with no apparent induction seen with the addition of arginine. This result is consistent with our in vitro observation that the entire 27-bp sequence was necessary for binding to ArcR. The requirement for the entire sequence is especially evident in the ccpA mutant background, since as suggested in previous studies, loss of the CcpA protein almost completely abolishes CCR of the ADS (9). More specifically, elimination of CCR by deletion of the ccpA gene allows for easier detection of the effects of deletions in the ArcR binding site, which partly overlaps with CRE-W. Accordingly, in the ccpA mutant background, compared to that of the wild-type ParcA::cat construct, all deletion mutants had much lower CAT activities when either glucose or galactose was added as the carbohydrate source or when arginine was present or absent. Moreover, at least for FTL and CRE, there was no statistically significant difference in CAT activity between samples with and without supplementation of arginine.
When substitution mutants were used in these tests, results were slightly different. As shown in Table 2, mutation of the left half of the ArcR binding site (FTLM) produced CAT activities comparable to those of the other three deletion mutants. This finding again supports the data obtained by footprinting with ArcR. Notably, statistical analysis indicated a significant increase in CAT activity by this mutant in the presence of arginine (P = 0.034 for DL-1 and P = 0.023 in the ccpA mutant background) when glucose was used. This effect will be discussed later in this section. In contrast, as shown in Table 2, mutations in the right half of the ArcR binding site (in FTRM), which also disrupted CRE-W, failed to bring down the CAT activity to the level of the deletion mutants, while the CCR level (indicated by the ratio of ParcA-cat expression in galactose versus that in glucose) was slightly decreased. It has been reported previously that protection from Arg repressors seen on DNase I digestion can extend beyond the limits of the putative Arg box that was defined by other analyses (7, 16, 21). Considering that in FTRM four out of five mutated nucleotides were on the outer edges of the putative ArcR footprint (Fig. 5A), it is possible that the region with which the ArcR protein needs to interact to effect activation of the operon is slightly shorter on the 3' end than the sequence protected by DNase I. The significance of these data in understanding the function of the second CRE is discussed below.
These results also emphasized the critical role of arginine-mediated activation in regulation of arc operon expression by ArcR, since there was more than a 95% reduction in CAT activity in the mutants when cells were grown under optimal induction conditions, i.e., with galactose and arginine. Thus, arginine induction of ADS expression through ArcR can be considered the primary activation mechanism for ADS expression in S. gordonii.
Lastly, seemingly contradictory to our model of ArcR regulating Parc expression in response to arginine, it has been observed during our study that arginine appears to be able to advance ParcA-cat expression in an arcR mutant background. Also notable is the similar trend with some of the ArcR-binding-site mutants in the DL-1 wild-type and ccpA mutant background. However, these minor increases of Parc expression in response to arginine are seen in the context of an overall dramatic reduction in ParcA-cat expression caused by either arcR deletion or disruption of the ArcR binding site in the arcA promoter region. One plausible explanation for this variation is the change in growth conditions brought on by inclusion of arginine in the growth medium. Since the ADS provides the organism with ATP, addition of arginine in the medium could significantly alter the growth rate or affect sugar metabolism by altering intracellular ATP pools, both of which could result in fluctuation of ParcA expression. Another possibility is the impact of arginine on pH of the growth medium through ammonia production. We have recently observed that pH is another factor regulating ADS expression in S. gordonii (unpublished data).
Involvement of two putative CREs in CCR. As previously reported, two potential catabolite response elements (CREs) were found near the arc promoter (Fig. 2). CRE-S (TGTAAGTGTTTTCA) spans from –35 to –22 upstream of the arc TIS, partly overlapping with the probable –35 region of the arc promoter and differing by 1 base from the consensus CRE sequence (TGWNANCGNTNWCA) (W is for A or T) found in gram-positive bacteria. CRE-W (AGAAAACGCTTCAA) spans from –107 to –94 and differs by 2 bases from the consensus sequence. It has also been reported that after an apparent ccpA homologue was disrupted in the genome of S. gordonii DL-1, CCR of arc expression was alleviated (9), suggesting that CCR was mediated by CcpA binding to one or both of the CREs in the arc promoter region. To determine the involvement of these two CREs in CCR, mutations were engineered by site-directed mutagenesis into one (FTRM and CRE-S') or both CREs (CRE-W'S') (Fig. 5B) of the ParcA::cat reporter construct to disrupt the dyad symmetry of the sequence, which has been proven to be critical for the functionality of CREs in other bacterial systems (29). Wild-type S. gordonii DL-1 was used as the host to express these fusion constructs, and CAT assays were carried out in bacteria grown to mid-exponential phase (OD600 of around 0.6 to 0.7) in TY medium with glucose or galactose as the carbohydrate source.
As shown in Table 2, mutation in either of the putative CREs alleviated CCR of arc expression, and mutation in both CREs further reduced the level of CCR, producing higher levels of CAT activities in glucose-based TY medium than the intact promoter fusion. This experiment indicates that both CREs participate in repression. It is not rare for gram-positive bacteria to have multiple CREs in one promoter region, most of which exist within or downstream of the promoter. For example, two CREs in the gnt operon of B. subtilis were suggested to function independently, and both were required for exertion of CCR by CcpA. Also, CcpA interacted only with one CRE when P-Ser-HPr was added in vitro and interacted with both when Glc-6-P was added (22). In another study, multiple CREs of the xyl operon from Bacillus megaterium were shown to be bound by CcpA either cooperatively or noncooperatively, depending on the presence of specific trigger molecules (P-Ser-HPr or Glc-6-P) and environmental pH (13). According to a mechanism suggested by this study, these two CREs within the arc promoter region could be working in a coordinated fashion by serving as binding sites for CcpA-cofactor complexes, allowing modulation of ADS activity in response to multiple signals, or the two CREs may simply be required for cooperative binding of CcpA to optimize repression.
As mentioned above, CRE-W (–107 to –94) overlaps with the ArcR binding site by 12 bases, which made it difficult to disassociate effects of activation by ArcR from CCR, particularly since S. gordonii cannot grow without carbohydrates. Initial results of the CAT assays in the DL-1 background suggested that this region might not be playing a role in CCR of arc, since all four mutants, two of which had an intact CRE-W (FTL and FTLM) and two of which had mutations in this element (CRE and FT), gave similar ratios of expression in the presence of glucose or galactose. However, the presence of wild-type ArcR could be a complicating factor in these experiments, since its binding capacity to this region, which could also be affected by these mutations, greatly affects cat expression. To rule out the involvement of ArcR, both ParcA::cat and ParcACRE::cat fusions were introduced into an otherwise isogenic arcR mutant strain, where the arcR gene had been disrupted by insertion of a kanamycin resistance marker. CAT assays (Table 2) in this background showed that deletion of CRE-W reduced the CCR ratio from 13.4 to 4. This result is consistent with our study of the substitution mutants in the two CREs (Fig. 5B, Table 2). By having ArcR and CcpA share an overlapping binding sequence, CcpA will not only block the RNA polymerase machinery by binding to CRE-S but also further reduce the transcriptional activity of the arc promoter by competing for the binding of the primary activator of the operon, ArcR.
Current experimental data are not sufficient to allow a conclusion to be reached regarding the hierarchy of the two CREs in binding to CcpA. CRE-S resides in close proximity to the –35 sequence, and CRE-W overlaps partly with the ArcR binding site; therefore, sequence alterations in either CRE could profoundly affect the transcriptional activity of the arc operon. A similar challenge was noted previously in our laboratory, with transcriptional analysis of two putative CREs in the promoter region of the fruA gene of S. mutans, when mutation of a CRE that overlapped with the extended –10 promoter greatly reduced fruA promoter activity (30). Notably, differences between changes in ADS expression in response to CCR in these two mutants were small and possibly not biologically significant, at least under the conditions tested. In vitro binding experiments using purified CcpA and upstream regions of the arcA promoter may provide additional insights.
ACKNOWLEDGMENTS
REFERENCES
Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248-254.
Burne, R. A., and R. E. Marquis. 2000. Alkali production by oral bacteria and protection against dental caries. FEMS Microbiol. Lett. 193:1-6.
Charlier, D., M. Roovers, F. Van Vliet, A. Boyen, R. Cunin, Y. Nakamura, N. Glansdorff, and A. Pierard. 1992. Arginine regulon of Escherichia coli K-12. A study of repressor-operator interactions and of in vitro binding affinities versus in vivo repression. J. Mol. Biol. 226:367-386.
Cook, G. S., J. W. Costerton, and R. J. Lamont. 1998. Biofilm formation by Porphyromonas gingivalis and Streptococcus gordonii. J. Periodont. Res. 33:323-327.
Cotter, P. D., and C. Hill. 2003. Surviving the acid test: responses of gram-positive bacteria to low pH. Microbiol. Mol. Biol. Rev. 67:429-453.
Cunin, R., N. Glansdorff, A. Pierard, and V. Stalon. 1986. Biosynthesis and metabolism of arginine in bacteria. Microbiol. Rev. 50:314-352.
Dion, M., D. Charlier, H. Wang, D. Gigot, A. Savchenko, J. N. Hallet, N. Glansdorff, and V. Sakanyan. 1997. The highly thermostable arginine repressor of Bacillus stearothermophilus: gene cloning and repressor-operator interactions. Mol. Microbiol. 25:385-398.
Dong, Y., Y. Y. Chen, J. A. Snyder, and R. A. Burne. 2002. Isolation and molecular analysis of the gene cluster for the arginine deiminase system from Streptococcus gordonii DL1. Appl. Environ. Microbiol. 68:5549-5553.
Dong, Y., Y.-Y. M. Chen, and R. A. Burne. 2004. Control of expression of the arginine deiminase operon of Streptococcus gordonii by CcpA and Flp. J. Bacteriol. 186:2511-2514.
Ferro, K. J., G. R. Bender, and R. E. Marquis. 1983. Coordinately repressible arginine deiminase system in Streptococcus sanguis. Curr. Microbiol. 9:145-150.
Galas, D. J., and A. Schmitz. 1978. DNAse footprinting: a simple method for the detection of protein-DNA binding specificity. Nucleic Acids Res. 5:3157-3170.
Gamper, M., A. Zimmermann, and D. Haas. 1991. Anaerobic regulation of transcription initiation in the arcDABC operon of Pseudomonas aeruginosa. J. Bacteriol. 173:4742-4750.
Gosseringer, R., E. Kuster, A. Galinier, J. Deutscher, and W. Hillen. 1997. Cooperative and non-cooperative DNA binding modes of catabolite control protein CcpA from Bacillus megaterium result from sensing two different signals. J. Mol. Biol. 266:665-676.
Griswold, A., Y. Y. Chen, J. A. Snyder, and R. A. Burne. 2004. Characterization of the arginine deiminase operon of Streptococcus rattus FA-1. Appl. Environ. Microbiol. 70:1321-1327.
Holden, J. T. 1962. The composition of microbial amino acid pools, p. 73-108. In J. T. Holden (ed.), Amino acid pools. Elsevier, New York, N.Y.
Klingel, U., C. M. Miller, A. K. North, P. G. Stockley, and S. Baumberg. 1995. A binding site for activation by the Bacillus subtilis AhrC protein, a repressor/activator of arginine metabolism. Mol. Gen. Genet. 248:329-340.
Lu, C. D., H. Winteler, A. Abdelal, and D. Haas. 1999. The ArgR regulatory protein, a helper to the anaerobic regulator ANR during transcriptional activation of the arcD promoter in Pseudomonas aeruginosa. J. Bacteriol. 181:2459-2464.
Luthi, E., A. Mercenier, and D. Haas. 1986. The arcABC operon required for fermentative growth of Pseudomonas aeruginosa on arginine: Tn5-751-assisted cloning and localization of structural genes. J. Gen. Microbiol. 132(Part 10):2667-2675.
Maas, W. K. 1994. The arginine repressor of Escherichia coli. Microbiol. Rev. 58:631-640.
Maghnouj, A., A. A. Abu-Bakr, S. Baumberg, V. Stalon, and C. Vander Wauven. 2000. Regulation of anaerobic arginine catabolism in Bacillus licheniformis by a protein of the Crp/Fnr family. FEMS Microbiol. Lett. 191:227-234.
Maghnouj, A., T. F. de Sousa Cabral, V. Stalon, and C. Vander Wauven. 1998. The arcABDC gene cluster, encoding the arginine deiminase pathway of Bacillus licheniformis, and its activation by the arginine repressor. argR. J. Bacteriol. 180:6468-6475.
Miwa, Y., K. Nagura, S. Eguchi, H. Fukuda, J. Deutscher, and Y. Fujita. 1997. Catabolite repression of the Bacillus subtilis gnt operon exerted by two catabolite-responsive elements. Mol. Microbiol. 23:1203-1213.
Quivey, R. G., W. L. Kuhnert, and K. Hahn. 2001. Genetics of acid adaptation in oral streptococci. Crit. Rev. Oral Biol. Med. 12:301-314.
Shaw, W. V. 1979. Chloramphenicol acetyltransferase activity from chloramphenicol-resistant bacteria. Methods Enzymol. 43:737-755.
Smith, M. C., L. Czaplewski, A. K. North, S. Baumberg, and P. G. Stockley. 1989. Sequences required for regulation of arginine biosynthesis promoters are conserved between Bacillus subtilis and Escherichia coli. Mol. Microbiol. 3:23-28.
Stephan, R. M. 1940. Changes in hydrogen-ion concentration on tooth surfaces and in carious lesions. Am. J. Dent. 27:718-723.
Van Duyne, G. D., G. Ghosh, W. K. Maas, and P. B. Sigler. 1996. Structure of the oligomerization and L-arginine binding domain of the arginine repressor of Escherichia coli. J. Mol. Biol. 256:377-391.
Van Wuyckhuyse, B. C., H. E. Perinpanayagam, D. Bevacqua, R. F. Raubertas, R. J. Billings, W. H. Bowen, and L. A. Tabak. 1995. Association of free arginine and lysine concentrations in human parotid saliva with caries experience. J. Dent. Res. 74:686-690.
Weickert, M. J., and G. H. Chambliss. 1990. Site-directed mutagenesis of a catabolite repression operator sequence in Bacillus subtilis. Proc. Natl. Acad. Sci. USA 87:6238-6242.
Wen, Z. T., and R. A. Burne. 2002. Analysis of cis- and trans-acting factors involved in regulation of the Streptococcus mutans fructanase gene (fruA). J. Bacteriol. 184:126-133.(Lin Zeng, Yiqian Dong, an)
ABSTRACT
The arginine deiminase system (ADS) is responsible for the production of ornithine, CO2, ammonia, and ATP from arginine. The ADS of the oral bacterium Streptococcus gordonii plays major roles in physiologic homeostasis, acid tolerance, and oral biofilm ecology. To further our understanding of the transcriptional regulation of the ADS (arc) operon, the binding of the ArcR transcriptional activator, which governs expression of the ADS in response to arginine, was investigated by DNase I protection and gel mobility shift assays. An ArcR binding sequence was found that was 27 bp in length and had little sequence similarity to binding sites of other arginine metabolism regulators. The presence of arginine at physiologically relevant concentrations enhanced the binding of ArcR to its target. Using cat fusions, various deletion and substitution mutations within the putative ArcR footprint were shown to cause dramatic reductions in expression from the arcA promoter in vivo, confirming that the 27-bp sequence is required for optimal expression and induction of the ADS by arginine. Mutation of two putative catabolite response elements (CREs) within the arc promoter region showed that both CREs contribute to catabolite repression. A thorough understanding of the regulation of the ADS in S. gordonii and related organisms is needed to develop ways to exploit arginine catabolism for the control of oral diseases. Identification of the ArcR and CcpA binding sites lays the foundation for a more complete understanding of the complex interactions of multiple regulatory proteins with elements in the arc promoter region.
INTRODUCTION
Bacteria in the oral cavity are subjected to constantly changing environments, with major fluctuations in nutrient availability and pH occurring with cycles of feeding and fasting by the host. Following ingestion of carbohydrates, the pH of oral biofilms can drop within minutes from above 7 to below 4 as a result of glycolysis by acidogenic bacteria, resulting in demineralization of teeth (23, 26). Oral bacteria have developed a variety of strategies to cope with intermittent and sustained acidification through complex and coordinated cellular responses that enhance acid tolerance, including conversion of salivary and dietary substrates to basic compounds (2, 5). Ammonia generation from hydrolysis of urea by urease and arginine catabolism by the arginine deiminase system (ADS) are believed to be crucial for oral biofilm pH homeostasis and are thought to be major factors promoting dental health (2).
The ADS is a three-enzyme system, consisting of arginine deiminase, catabolic ornithine carbamyltransferase, and carbamate kinase enzymes, which together convert arginine to ornithine, ammonia, and CO2, with concurrent production of ATP (6). There is considerable conservation in the gene organization and primary structure of the enzymes of the ADS, but there is also diversity in the regulation and physiological role of the pathway (8). The most thoroughly studied ADS, that of the gram-negative bacterium Pseudomonas aeruginosa, is only expressed under anaerobic conditions, and the presence of arginine enhances the expression of ADS genes (12, 18). Two regulators, ANR and ArgR, are required for responding to low oxygen tension and sensing exogenous arginine, respectively. Separate binding sites for ANR and ArgR have been identified in the promoter region of the P. aeruginosa arc operon (17). Similarly, the ADS in Bacillus licheniformis is induced by anaerobiosis, but only when arginine is present (20, 21). In both of these organisms, the ADS appears to function primarily in ATP generation when aerobic respiration cannot be used for growth. In contrast, the ADS of oral streptococci, including S. gordonii, Streptococcus sanguis, and Streptococcus rattus, appears to be a primary defense against killing by acidification (2). In these organisms, which lack a respiratory chain, the ADS is subject to carbon catabolite repression (CCR) and arginine acts as an induction signal to further increase expression (8-10, 14).
In a previous study from our laboratory, the ADS of S. gordonii was found to be induced by arginine and subject to CCR, with very low levels of expression observed in the presence of the repressing sugar glucose (8, 9). Deletion of the ADS regulatory gene arcR resulted in greatly reduced levels of arginine deiminase activity and arc transcription, while disruption of the catabolite control protein, CcpA, greatly reduced CCR of the ADS. The ADS was also found to be further induced by anaerobiosis through an Fnr-like protein (Flp) (8, 9). Significant levels of homology were found between ArcR of S. gordonii and other transcriptional regulators of arginine metabolism, including ArgR of Escherichia coli and B. licheniformis as well as AhrC of B. subtilis (8).
S. gordonii is one of relatively few ADS-positive oral bacteria, but it colonizes the mouth very early in life and remains abundant in healthy oral biofilms (4). S. gordonii has been suggested to play important roles in maintaining pH homeostasis in oral biofilms, and the ADS has been proposed to have significant potential to be exploited to prevent dental caries in humans (2). In this report, we expressed a recombinant ArcR protein and studied its interaction with the arcA promoter region to gain insight into the molecular basis for differential expression of the ADS in response to arginine and other environmental influences.
MATERIALS AND METHODS
Bacterial strains, growth conditions, and reagents. E. coli strains were grown in Luria-Bertani (LB) medium supplemented with antibiotics at the following concentrations: ampicillin (Ap; 100 μg ml–1), kanamycin (Km; 40 μg ml–1), erythromycin (Em; 300 μg ml–1), spectinomycin (Sp; 50 μg ml–1), and chloramphenicol (Cm; 20 μg ml–1). S. gordonii DL1 and its derivatives were maintained in brain heart infusion broth (BHI; Difco Laboratories, Detroit, Mich.) at 37°C in 5% CO2 and 95% air, with antibiotics added at the following concentrations when necessary: Km (250 μg ml–1), Em (5 μg ml–1), and Sp (250 μg ml–1). S. gordonii cells intended for chloramphenicol acetyltransferase (CAT) assays were cultured in tryptone yeast extract (TY) base medium supplemented with 10 mM glucose or galactose and 50 mM arginine, as specified. All chemical reagents and antibiotics were obtained from Sigma Chemical Co. (St. Louis, Mo).
DNA manipulations. To construct a ParcA::cat reporter gene fusion, a 337-bp fragment upstream of the arcA start codon (ParcA) was amplified by primers ParcA5' and ParcA3' (Table 1) and fused with a promoterless chloramphenicol acetyltransferase (cat) gene. After confirming by sequencing, the ParcA::cat construct was introduced into the S. gordonii genome by integration into the gtfG locus (9). Other ParcA::cat mutants were engineered by site-directed mutagenesis. Briefly, pairs of primers with designed deletions or mutations (Table 1) were synthesized and used in recombinant PCR with the original primers (ParcA5' and ParcA3') to generate new fragments with the desired deletions or mutations. These fragments were then fused with cat and integrated into the S. gordonii genome in the same way. All deletions and mutations were confirmed by sequencing. All primers were synthesized by Integrated DNA Technologies (IDT), Inc. (Coralville, IA).
Expression and purification of recombinant ArcR in E. coli. The arcR gene wasamplified by PCR using primers arcRBamHIPUR and arcRPstIPUR (Table 1) with BamHI and PstI sites added to the 5' and 3' ends, respectively. The amplified arcR gene was digested with BamHI and PstI and ligated into the expression vector pQE30 (QIAGEN, Inc., Valencia, CA), which allowed in-frame fusion of six histidine residues to the N terminus of the recombinant ArcR protein. The resulting plasmid, pQE30-ArcR, was introduced into E. coli M15. Expression of recombinant ArcR was induced with isopropyl--D-thiogalactopyranoside (IPTG) at a final concentration of 1 mM. After 3 h of IPTG induction, protein lysates were prepared by homogenization and the recombinant protein was purified by nickel affinity chromatography as recommended by the supplier. Purified ArcR was dialyzed against 1 liter of phosphate-buffered saline buffer (pH 7.4) and then stored at 4°C for up to 1 month.
Electrophoretic mobility shift assay. A 193-bp fragment upstream of the arcA start codon was amplified by PCR using primers arcRFT-5' and arcRFT-3' (Table 1). The PCR product was cloned into the TA-cloning vector pGEM-T Easy (Promega, Madison, WI), resulting in plasmid pLZ0402, and then released as a 213-bp fragment by EcoRI digestion. The EcoRI fragment was end labeled with [-32P]ATP using T4 polynucleotide kinase and was used as a probe in the mobility shift assays. The reaction mixture (10 μl) contained 50 fmol labeled DNA probe, 5 mM MgCl2, 50 mM KCl, 1 mM EDTA, 0.2 μg μl–1 poly(dI-dC), 10 mM HEPES (pH 7.9), 5 mM dithiothreitol, and 10% glycerol. Purified recombinant ArcR, L-arginine, or L-lysine was added to the mixture at the specified concentrations. The mixture was incubated at room temperature for 20 min, followed by electrophoretic separation through a 4% low-ionic-strength polyacrylamide gel. The gel was dried on Whatman 3 MM paper and exposed to X-ray film.
DNase I protection (footprinting) assay. The 213-bp DNA probe used in the DNase I protection assay was released from plasmid pLZ0402 with EcoRI. It was then purified and treated with shrimp alkaline phosphatase before labeling with [-32P]ATP using T4 polynucleotide kinase, followed by another digestion with either HindIII or EcoRV, as these sites were engineered into the primers arcRFT-5' and arcRFT-3', respectively (Table 1). The DNA probes with only one radiolabeled end were then purified using the QIAquick gel extraction kit. DNase I digestion in the presence of recombinant ArcR and subsequent analysis by polyacrylamide gel electrophoresis were carried out according to a standard protocol with minor modifications (11). Briefly, 20 fmol of the DNA probe was added into a 25-μl system containing 0 or 1 pmol of purified ArcR, 40 ng μl–1 poly(dI-dC), and 50 mM Tris-HCl, pH 8.0, and incubated at room temperature for 20 min. Fifty microliters of a 10 mM MgCl2, 5 mM CaCl2 solution was then added, followed by 1 μl of DNase I at 0, 10–3, 10–2, or 10–1 dilutions from the original concentration (114 U μl–1; Invitrogen, Carlsbad, CA). After 2 min of incubation at room temperature, 75 μl of stop solution (20 mM EDTA, 1% sodium dodecyl sulfate [SDS], 0.2 M NaCl, 125 μg ml–1 yeast tRNA) was added, and the DNA was precipitated after addition of 375 μl of ethanol. Digested DNA samples were then analyzed along side of a sequencing ladder that was generated using the Sequenase sequencing kit (USB, Cleveland, Ohio) and pLZ0402 as a template.
Chloramphenicol acetyltransferase (CAT) assay. Bacterial cells with various cat fusions were grown under different conditions to mid-exponential phase (optical density at 600 nm [OD600] 0.6 to 0.7) and washed twice with 10 mM Tris buffer (pH 7.8). Cells were then disrupted in a Bead Beater for 20 s, performed twice with a 2-min interval on ice. The cell lysate was centrifuged at 18,000 x g for 10 min, and the supernatant was recovered and used for measuring CAT activity using the method of Shaw (24). The concentration of the protein in the lysate was determined using the Bradford protein assay (1).
RESULTS AND DISCUSSION
Purified ArcR binding to the promoter region of arcA in vitro. To study the interaction between ArcR and its binding elements in vitro, the entire arcR coding sequence was cloned into plasmid pQE30 to create a translational fusion adding six histidine residues to the N terminus of ArcR, and His-ArcR was purified by nickel affinity chromatography. The purified protein displayed the expected apparent molecular mass of 18 kDa on SDS-polyacrylamide gel electrophoresis (Fig. 1A). To prove that purified ArcR had the ability to bind to the arcA promoter region, electrophoretic mobility shift assays were performed with purified ArcR and a 213-bp -32P-labeled DNA fragment, designated USR, containing about 0.2 kbp upstream of the arcA transcription initiation site (TIS) (Fig. 2). Without His-ArcR, the probe migrated to the bottom of the gel. When recombinant ArcR was added, the movement of USR was retarded (Fig. 1B). Increasing concentrations of His-ArcR caused more USR signal to shift upward and to shift to higher molecular weight levels, creating multiple slower-migrating bands on the gel. These results indicated that the recombinant ArcR was able to bind to the arcA promoter region and suggested that more than one molecule of His-ArcR may bind to the same DNA probe. ArgR of E. coli, one of the proteins with considerable similarity to ArcR of S. gordonii, has been crystallized, and an analysis of its structure revealed a symmetric hexamer formed by six identical ArgR molecules that was stabilized by the presence of arginine (27).
As an additional control, an irrelevant His-tagged protein, an autolysin from S. mutans, was used in a mobility shift assay in place of His-ArcR (data not shown). No mobility shift of USR was detected when this protein was used at concentrations similar to those for His-ArcR, indicating that the His tag did not play a role in the binding of recombinant His-ArcR to its target DNA.
Arginine enhances binding of ArcR to the arcA promoter region in vitro. ArcR shares considerable similarity with regulatory proteins that can interact with arginine in a way that alters protein conformation and binding affinity for DNA and ultimately changes in transcription of the target promoter (27). S. gordonii ArcR contains several amino acid residues that are conserved with arginine binding sites defined by studies of the E. coli arginine repressor ArgR; namely, glutamine at position 104 and aspartic acid residues at positions 125 and 126 (8). As Dong et al. have previously reported, expression of the arc operon of S. gordonii DL-1 is substantially enhanced by the addition of arginine to the growth medium in an ArgR-dependent manner (8). To determine whether arginine could influence the binding of ArcR to its target, arginine (arginine-HCl; Sigma) was added into the gel mobility shift assay. As shown in Fig. 3, addition of arginine clearly improved the interaction between ArcR and USR, with more USR being shifted when higher concentrations of arginine were present in the assay. Ina control experiment, lysine instead of arginine was added to the mixture of ArcR and USR before electrophoresis. No apparent effect on the migration of USR was observed in this experiment, suggesting that the enhancing effect of arginine on ArcR-USR binding was not due simply to the presence of a basic amino acid. Notably, arginine concentrations as low as 10 μM were sufficient to enhance binding, a concentration lower than that reported for free arginine in the human mouth (50 μM) (28) and well within the range of known concentrations of intracellular pools of amino acids in bacteria (15).
ArcR binds to a 27-bp region upstream of the arcA promoter. The experimental results from the gel shift assays revealed that the recombinant ArcR protein was active with regard to interaction with its target DNA and arginine. Two 213-bp DNA probes, one end labeled with -32P, were prepared from the same sequence that was used in the gel shift assays. DNA was subjected to DNase I digestion in the presence and absence of ArcR protein. The digested probe was then analyzed by electrophoresis on a sequencing gel. Results from the two strands (Fig. 4A) showed that a 27-bp region (ATATAAAATATGCAAAGAAAACGCTTC), spanning from position –122 to –96 upstream of the arcA transcription initiation site, was protected by ArcR. Three hypersensitive sites divided the footprint into two parts (Fig. 4A). The 27-bp sequence was predicted to have a very limited tendency for formation of secondary structure, and little sequence similarity was detected with known binding sequences of other ArcR-like proteins. The protected sequence also partly overlapped a potential catabolite response element (CRE).
To confirm the results of the footprinting assay, the ability of various DNA oligonucleotides to compete with the -32P-labeled USR probe for binding to ArcR was tested in electrophoretic mobility shift assays. As depicted in Fig. 4B, 1224-5x3, the full-length 27-bp, double-stranded DNA in the ArcR footprint, along with three shorter DNA fragments lacking the left half (1224-#1), the right half (1224-#2), or 6 bp from both ends (1224-d), were used to compete with labeled USR. Each of four DNA fragments was generated from a pair of synthesized single-stranded oligonucleotides by heating to 94°C and cooling slowly to room temperature. As shown in Fig. 4C, 1224-5x3 was able to compete for ArcR binding with radioactive USR, although higher concentrations were required compared to that for cold USR. At the same time, when single-stranded oligonucleotides (1224-5 and 1224-3) were used individually, no competition was detectable at similar concentrations. Furthermore, when the three shorter, double-stranded DNA fragments (1224-#1, 1224-#2, and 1224-d) were individually added to the assay, no competition for binding to ArcR was detected (Fig. 4D). These results indicate that the entire 27-bp sequence is required for efficient binding to ArcR.
Due to the range limitation of the sequencing gel, it was possible that the DNase I protection assay missed other ArcR binding elements within the 213-bp probe. To test for the existence of other ArcR binding elements within the upstream region of the arcA promoter, a probe that contained all of the sequence of the 213-bp probe except the 27-bp target region (USRFT) was engineered and prepared for gel shift assays. When USRFT was used to compete with the radioactive USR in binding to ArcR, no competition activity was detected (Fig. 4E), suggesting that the 27-bp DNA sequence is the only major ArcR binding site in the upstream region of the arcA promoter. However, due to the limited sensitivity of the competition assay, the existence of a low-affinity binding site(s) for ArcR in the arcA upstream region remained a possibility. When radiolabeled USRFT was used to bind to the recombinant His-ArcR in a mobility shift assay, a low residual level of binding activity was detected in this fragment (data not shown), suggesting the presence of a secondary binding site(s) for ArcR in the arcA promoter region. The presence of both high- and low-affinity ArcR binding sites in the arcA promoter region could allow for fine tuning the regulation of the ADS through cooperative binding, perhaps in response to various concentrations of arginine in the environment. In this report, we will only focus on the analysis of the 27-bp high-affinity binding site of ArcR in the arcA promoter region.
Research on the transcriptional regulators of other known bacterial arginine metabolism systems from a variety of genera has revealed binding specificities within the promoter regions of their target genes that share sequence similarity. Among these transcriptional regulators, the binding characteristics of the arginine repressor of E. coli (ArgR) are the best characterized. It has been determined that E. coli ArgR oligomerizes into a hexamer in the presence of arginine and binds to a consensus Arg box of A/TNTGAATAATTATTCANA/T (N represents any base) found in the genes of the arginine regulon (3, 19, 27). ArgR of B. licheniformis has a very similar 18-nucleotide binding site, with 15 nucleotides conserved with the E. coli Arg box, although the purified ArgR protein protected a significantly longer sequence from DNase I digestion (21). Similar Arg boxes have also been found in the arginine biosynthetic operon of B. subtilis (25). Another well-studied arginine deiminase transcriptional regulator, ArgR of P. aeruginosa, binds to two direct-repeat sequences upstream of the arc promoter, matching the Arg box (TGTCGCN8AA) commonly found in other ArgR binding sites of P. aeruginosa (17).
One common feature of the aforementioned Arg box sites is their sequence symmetry indicated by the presence of inverted or direct repeats. Unlike those cis elements regulating arginine metabolism, sequences with a strong tendency to form regions of dyad symmetry or direct repeats were not observed within the S. gordonii ArcR binding sequence. Half of the upstream portion of the sequence did show some symmetry, but this very A:T-rich sequence was not likely to form a very stable structure. Moreover, little sequence similarity was detected between the S. gordonii ArcR binding site and the Arg boxes of E. coli or P. aeruginosa. Our experiments also revealed that instead of having two symmetrical binding sites, as does ArgR in P. aeruginosa, ArcR probably has one high-affinity binding site and an additional low-affinity binding site(s) in the arc promoter region. Given the complexity of the regulation of the S. gordonii ADS and modulation of expression by growth phase, pH (unpublished data), CCR, and oxygen tension, it would not be surprising that sophisticated regulation by arginine involving cooperative binding of ArcR to its target cis elements has been evolved in this bacterium. Additionally, other regulatory proteins, including CcpA and Flp (9), may also interact near to, or overlap with, the ArcR binding sites.
Functional analysis of the ArcR binding site. Based on the knowledge of the interaction of ArcR with the upstream region of the arcA promoter in vitro, we engineered a series of mutants into ParcA::cat, a transcriptional fusion of the arcA promoter to a promoterless chloramphenicol acetyltransferase (cat) gene and monitored expression in vivo in response to environmental stimuli. All strains containing the fusion constructs were grown in BHI broth with appropriate antibiotics overnight and then transferred into fresh TY medium containing 10 mM glucose or galactose as the carbohydrate source, with or without addition of 50 mM arginine. Mid-exponential-phase cultures (OD600 0.6 to 0.7) were harvested and CAT assays were carried out immediately to measure the expression levels of the promoter fusions. In agreement with previous reports, arc operon expression in the wild-type strain was subject to CCR, as indicated by low CAT activity when glucose was used as the main carbohydrate source and increased CAT activity when glucose was replaced with galactose. Likewise, addition of arginine increased CAT activity (Table 2).
Based on the data obtained in the in vitro binding assays, three deletion mutants were engineered (Fig. 5A) with the left half (FTL), the right half (CRE, overlapping with the second CRE), or the entire 27-bp sequence (FT) deleted from the ParcA::cat fusion. Additionally, two substitution mutants were generated with changes in the left half (FTLM) or the right half of the ArcR binding site (FTRM) being mutated in a way that the weak dyadic sequence was disrupted. All constructs were then introduced in single copy into the gtfG gene of the wild-type strain DL-1 and an otherwise-isogenic ccpA mutant in which an erythromycin resistance cassette was inserted to disrupt the ccpA gene. Strains harboring these mutant fusions were cultured to mid-exponential phase in TY medium containing 10 mM glucose or galactose, with 0 or 50 mM added arginine (Table 2).
Strains carrying all three deletion mutants (FTL, CRE, and FT) produced similarly lower levels of CAT activities compared to the wild-type ParcA::cat fusion in either the DL-1 or the ccpA mutant backgrounds, with no apparent induction seen with the addition of arginine. This result is consistent with our in vitro observation that the entire 27-bp sequence was necessary for binding to ArcR. The requirement for the entire sequence is especially evident in the ccpA mutant background, since as suggested in previous studies, loss of the CcpA protein almost completely abolishes CCR of the ADS (9). More specifically, elimination of CCR by deletion of the ccpA gene allows for easier detection of the effects of deletions in the ArcR binding site, which partly overlaps with CRE-W. Accordingly, in the ccpA mutant background, compared to that of the wild-type ParcA::cat construct, all deletion mutants had much lower CAT activities when either glucose or galactose was added as the carbohydrate source or when arginine was present or absent. Moreover, at least for FTL and CRE, there was no statistically significant difference in CAT activity between samples with and without supplementation of arginine.
When substitution mutants were used in these tests, results were slightly different. As shown in Table 2, mutation of the left half of the ArcR binding site (FTLM) produced CAT activities comparable to those of the other three deletion mutants. This finding again supports the data obtained by footprinting with ArcR. Notably, statistical analysis indicated a significant increase in CAT activity by this mutant in the presence of arginine (P = 0.034 for DL-1 and P = 0.023 in the ccpA mutant background) when glucose was used. This effect will be discussed later in this section. In contrast, as shown in Table 2, mutations in the right half of the ArcR binding site (in FTRM), which also disrupted CRE-W, failed to bring down the CAT activity to the level of the deletion mutants, while the CCR level (indicated by the ratio of ParcA-cat expression in galactose versus that in glucose) was slightly decreased. It has been reported previously that protection from Arg repressors seen on DNase I digestion can extend beyond the limits of the putative Arg box that was defined by other analyses (7, 16, 21). Considering that in FTRM four out of five mutated nucleotides were on the outer edges of the putative ArcR footprint (Fig. 5A), it is possible that the region with which the ArcR protein needs to interact to effect activation of the operon is slightly shorter on the 3' end than the sequence protected by DNase I. The significance of these data in understanding the function of the second CRE is discussed below.
These results also emphasized the critical role of arginine-mediated activation in regulation of arc operon expression by ArcR, since there was more than a 95% reduction in CAT activity in the mutants when cells were grown under optimal induction conditions, i.e., with galactose and arginine. Thus, arginine induction of ADS expression through ArcR can be considered the primary activation mechanism for ADS expression in S. gordonii.
Lastly, seemingly contradictory to our model of ArcR regulating Parc expression in response to arginine, it has been observed during our study that arginine appears to be able to advance ParcA-cat expression in an arcR mutant background. Also notable is the similar trend with some of the ArcR-binding-site mutants in the DL-1 wild-type and ccpA mutant background. However, these minor increases of Parc expression in response to arginine are seen in the context of an overall dramatic reduction in ParcA-cat expression caused by either arcR deletion or disruption of the ArcR binding site in the arcA promoter region. One plausible explanation for this variation is the change in growth conditions brought on by inclusion of arginine in the growth medium. Since the ADS provides the organism with ATP, addition of arginine in the medium could significantly alter the growth rate or affect sugar metabolism by altering intracellular ATP pools, both of which could result in fluctuation of ParcA expression. Another possibility is the impact of arginine on pH of the growth medium through ammonia production. We have recently observed that pH is another factor regulating ADS expression in S. gordonii (unpublished data).
Involvement of two putative CREs in CCR. As previously reported, two potential catabolite response elements (CREs) were found near the arc promoter (Fig. 2). CRE-S (TGTAAGTGTTTTCA) spans from –35 to –22 upstream of the arc TIS, partly overlapping with the probable –35 region of the arc promoter and differing by 1 base from the consensus CRE sequence (TGWNANCGNTNWCA) (W is for A or T) found in gram-positive bacteria. CRE-W (AGAAAACGCTTCAA) spans from –107 to –94 and differs by 2 bases from the consensus sequence. It has also been reported that after an apparent ccpA homologue was disrupted in the genome of S. gordonii DL-1, CCR of arc expression was alleviated (9), suggesting that CCR was mediated by CcpA binding to one or both of the CREs in the arc promoter region. To determine the involvement of these two CREs in CCR, mutations were engineered by site-directed mutagenesis into one (FTRM and CRE-S') or both CREs (CRE-W'S') (Fig. 5B) of the ParcA::cat reporter construct to disrupt the dyad symmetry of the sequence, which has been proven to be critical for the functionality of CREs in other bacterial systems (29). Wild-type S. gordonii DL-1 was used as the host to express these fusion constructs, and CAT assays were carried out in bacteria grown to mid-exponential phase (OD600 of around 0.6 to 0.7) in TY medium with glucose or galactose as the carbohydrate source.
As shown in Table 2, mutation in either of the putative CREs alleviated CCR of arc expression, and mutation in both CREs further reduced the level of CCR, producing higher levels of CAT activities in glucose-based TY medium than the intact promoter fusion. This experiment indicates that both CREs participate in repression. It is not rare for gram-positive bacteria to have multiple CREs in one promoter region, most of which exist within or downstream of the promoter. For example, two CREs in the gnt operon of B. subtilis were suggested to function independently, and both were required for exertion of CCR by CcpA. Also, CcpA interacted only with one CRE when P-Ser-HPr was added in vitro and interacted with both when Glc-6-P was added (22). In another study, multiple CREs of the xyl operon from Bacillus megaterium were shown to be bound by CcpA either cooperatively or noncooperatively, depending on the presence of specific trigger molecules (P-Ser-HPr or Glc-6-P) and environmental pH (13). According to a mechanism suggested by this study, these two CREs within the arc promoter region could be working in a coordinated fashion by serving as binding sites for CcpA-cofactor complexes, allowing modulation of ADS activity in response to multiple signals, or the two CREs may simply be required for cooperative binding of CcpA to optimize repression.
As mentioned above, CRE-W (–107 to –94) overlaps with the ArcR binding site by 12 bases, which made it difficult to disassociate effects of activation by ArcR from CCR, particularly since S. gordonii cannot grow without carbohydrates. Initial results of the CAT assays in the DL-1 background suggested that this region might not be playing a role in CCR of arc, since all four mutants, two of which had an intact CRE-W (FTL and FTLM) and two of which had mutations in this element (CRE and FT), gave similar ratios of expression in the presence of glucose or galactose. However, the presence of wild-type ArcR could be a complicating factor in these experiments, since its binding capacity to this region, which could also be affected by these mutations, greatly affects cat expression. To rule out the involvement of ArcR, both ParcA::cat and ParcACRE::cat fusions were introduced into an otherwise isogenic arcR mutant strain, where the arcR gene had been disrupted by insertion of a kanamycin resistance marker. CAT assays (Table 2) in this background showed that deletion of CRE-W reduced the CCR ratio from 13.4 to 4. This result is consistent with our study of the substitution mutants in the two CREs (Fig. 5B, Table 2). By having ArcR and CcpA share an overlapping binding sequence, CcpA will not only block the RNA polymerase machinery by binding to CRE-S but also further reduce the transcriptional activity of the arc promoter by competing for the binding of the primary activator of the operon, ArcR.
Current experimental data are not sufficient to allow a conclusion to be reached regarding the hierarchy of the two CREs in binding to CcpA. CRE-S resides in close proximity to the –35 sequence, and CRE-W overlaps partly with the ArcR binding site; therefore, sequence alterations in either CRE could profoundly affect the transcriptional activity of the arc operon. A similar challenge was noted previously in our laboratory, with transcriptional analysis of two putative CREs in the promoter region of the fruA gene of S. mutans, when mutation of a CRE that overlapped with the extended –10 promoter greatly reduced fruA promoter activity (30). Notably, differences between changes in ADS expression in response to CCR in these two mutants were small and possibly not biologically significant, at least under the conditions tested. In vitro binding experiments using purified CcpA and upstream regions of the arcA promoter may provide additional insights.
ACKNOWLEDGMENTS
REFERENCES
Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248-254.
Burne, R. A., and R. E. Marquis. 2000. Alkali production by oral bacteria and protection against dental caries. FEMS Microbiol. Lett. 193:1-6.
Charlier, D., M. Roovers, F. Van Vliet, A. Boyen, R. Cunin, Y. Nakamura, N. Glansdorff, and A. Pierard. 1992. Arginine regulon of Escherichia coli K-12. A study of repressor-operator interactions and of in vitro binding affinities versus in vivo repression. J. Mol. Biol. 226:367-386.
Cook, G. S., J. W. Costerton, and R. J. Lamont. 1998. Biofilm formation by Porphyromonas gingivalis and Streptococcus gordonii. J. Periodont. Res. 33:323-327.
Cotter, P. D., and C. Hill. 2003. Surviving the acid test: responses of gram-positive bacteria to low pH. Microbiol. Mol. Biol. Rev. 67:429-453.
Cunin, R., N. Glansdorff, A. Pierard, and V. Stalon. 1986. Biosynthesis and metabolism of arginine in bacteria. Microbiol. Rev. 50:314-352.
Dion, M., D. Charlier, H. Wang, D. Gigot, A. Savchenko, J. N. Hallet, N. Glansdorff, and V. Sakanyan. 1997. The highly thermostable arginine repressor of Bacillus stearothermophilus: gene cloning and repressor-operator interactions. Mol. Microbiol. 25:385-398.
Dong, Y., Y. Y. Chen, J. A. Snyder, and R. A. Burne. 2002. Isolation and molecular analysis of the gene cluster for the arginine deiminase system from Streptococcus gordonii DL1. Appl. Environ. Microbiol. 68:5549-5553.
Dong, Y., Y.-Y. M. Chen, and R. A. Burne. 2004. Control of expression of the arginine deiminase operon of Streptococcus gordonii by CcpA and Flp. J. Bacteriol. 186:2511-2514.
Ferro, K. J., G. R. Bender, and R. E. Marquis. 1983. Coordinately repressible arginine deiminase system in Streptococcus sanguis. Curr. Microbiol. 9:145-150.
Galas, D. J., and A. Schmitz. 1978. DNAse footprinting: a simple method for the detection of protein-DNA binding specificity. Nucleic Acids Res. 5:3157-3170.
Gamper, M., A. Zimmermann, and D. Haas. 1991. Anaerobic regulation of transcription initiation in the arcDABC operon of Pseudomonas aeruginosa. J. Bacteriol. 173:4742-4750.
Gosseringer, R., E. Kuster, A. Galinier, J. Deutscher, and W. Hillen. 1997. Cooperative and non-cooperative DNA binding modes of catabolite control protein CcpA from Bacillus megaterium result from sensing two different signals. J. Mol. Biol. 266:665-676.
Griswold, A., Y. Y. Chen, J. A. Snyder, and R. A. Burne. 2004. Characterization of the arginine deiminase operon of Streptococcus rattus FA-1. Appl. Environ. Microbiol. 70:1321-1327.
Holden, J. T. 1962. The composition of microbial amino acid pools, p. 73-108. In J. T. Holden (ed.), Amino acid pools. Elsevier, New York, N.Y.
Klingel, U., C. M. Miller, A. K. North, P. G. Stockley, and S. Baumberg. 1995. A binding site for activation by the Bacillus subtilis AhrC protein, a repressor/activator of arginine metabolism. Mol. Gen. Genet. 248:329-340.
Lu, C. D., H. Winteler, A. Abdelal, and D. Haas. 1999. The ArgR regulatory protein, a helper to the anaerobic regulator ANR during transcriptional activation of the arcD promoter in Pseudomonas aeruginosa. J. Bacteriol. 181:2459-2464.
Luthi, E., A. Mercenier, and D. Haas. 1986. The arcABC operon required for fermentative growth of Pseudomonas aeruginosa on arginine: Tn5-751-assisted cloning and localization of structural genes. J. Gen. Microbiol. 132(Part 10):2667-2675.
Maas, W. K. 1994. The arginine repressor of Escherichia coli. Microbiol. Rev. 58:631-640.
Maghnouj, A., A. A. Abu-Bakr, S. Baumberg, V. Stalon, and C. Vander Wauven. 2000. Regulation of anaerobic arginine catabolism in Bacillus licheniformis by a protein of the Crp/Fnr family. FEMS Microbiol. Lett. 191:227-234.
Maghnouj, A., T. F. de Sousa Cabral, V. Stalon, and C. Vander Wauven. 1998. The arcABDC gene cluster, encoding the arginine deiminase pathway of Bacillus licheniformis, and its activation by the arginine repressor. argR. J. Bacteriol. 180:6468-6475.
Miwa, Y., K. Nagura, S. Eguchi, H. Fukuda, J. Deutscher, and Y. Fujita. 1997. Catabolite repression of the Bacillus subtilis gnt operon exerted by two catabolite-responsive elements. Mol. Microbiol. 23:1203-1213.
Quivey, R. G., W. L. Kuhnert, and K. Hahn. 2001. Genetics of acid adaptation in oral streptococci. Crit. Rev. Oral Biol. Med. 12:301-314.
Shaw, W. V. 1979. Chloramphenicol acetyltransferase activity from chloramphenicol-resistant bacteria. Methods Enzymol. 43:737-755.
Smith, M. C., L. Czaplewski, A. K. North, S. Baumberg, and P. G. Stockley. 1989. Sequences required for regulation of arginine biosynthesis promoters are conserved between Bacillus subtilis and Escherichia coli. Mol. Microbiol. 3:23-28.
Stephan, R. M. 1940. Changes in hydrogen-ion concentration on tooth surfaces and in carious lesions. Am. J. Dent. 27:718-723.
Van Duyne, G. D., G. Ghosh, W. K. Maas, and P. B. Sigler. 1996. Structure of the oligomerization and L-arginine binding domain of the arginine repressor of Escherichia coli. J. Mol. Biol. 256:377-391.
Van Wuyckhuyse, B. C., H. E. Perinpanayagam, D. Bevacqua, R. F. Raubertas, R. J. Billings, W. H. Bowen, and L. A. Tabak. 1995. Association of free arginine and lysine concentrations in human parotid saliva with caries experience. J. Dent. Res. 74:686-690.
Weickert, M. J., and G. H. Chambliss. 1990. Site-directed mutagenesis of a catabolite repression operator sequence in Bacillus subtilis. Proc. Natl. Acad. Sci. USA 87:6238-6242.
Wen, Z. T., and R. A. Burne. 2002. Analysis of cis- and trans-acting factors involved in regulation of the Streptococcus mutans fructanase gene (fruA). J. Bacteriol. 184:126-133.(Lin Zeng, Yiqian Dong, an)