Proteome analysis of Streptococcus mutans metabolic phenotype during acid tolerance

Alice C. L. Len, Derek W. S. Harty and Nicholas A. Jacques

Institute of Dental Research, Westmead Millennium Institute and Westmead Centre for Oral Health, PO Box 533, Wentworthville, NSW, Australia 2145

Correspondence
Nicholas A. Jacques
njacques{at}dental.wsahs.nsw.gov.au


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Two-dimensional gel electrophoretic analysis of the proteome of Streptococcus mutans grown at a steady state in a glucose-limited anaerobic continuous culture revealed a number of proteins that were differentially expressed when the growth pH was lowered from pH 7·0 to pH 5·0. Changes in the expression of metabolic proteins were generally limited to three biochemical pathways: glycolysis, alternative acid production and branched-chain amino acid biosynthesis. The relative level of expression of protein spots representing all of the enzymes associated with the Embden–Meyerhof–Parnas pathway, and all but one of the enzymes involved in the major alternative acid fermentation pathways of S. mutans, was identified and measured. Proteome data, in conjunction with end-product and cell-yield analyses, were consistent with a phenotypic change that allowed S. mutans to proliferate at low pH by expending energy to extrude excess H+ from the cell, while minimizing the detrimental effects that result from the uncoupling of carbon flux from catabolism and the consequent imbalance in NADH and pyruvate production. The changes in enzyme levels were consistent with a reduction in the formation of the strongest acid, formic acid, which was a consequence of the diversion of pyruvate to both lactate and branched-chain amino acid production when S. mutans was cultivated in an acidic environment.


Abbreviations: 2-DGE, two-dimensional gel electrophoresis; ASB-14, amidosulfobetaine-14; D, dilution rate; DE, differential expression (values); IPG, immobilized pH gradient; IPS, intracellular polysaccharide; MALDI-TOF, matrix-assisted laser desorption ionization time-of-flight; PEP-PTS, phosphoenolpyruvate : glucose phosphotransferase system; PMM, peptide mass mapping; YATP, ATP yield, dry weight cells per mol ATP; YGlc, cell yield, dry weight cells per mol glucose


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
The initiation of dental caries is clinically associated with acidogenic and aciduric bacteria that secrete and tolerate the organic acidic fermentation by-products of carbohydrate metabolism. By their demineralizing action upon tooth enamel, these organic acids both initiate and contribute to the aetiology of the disease. Over the past 50 years, studies of dental caries have focussed on the oral streptococci, particularly Streptococcus mutans, which is now accepted to be associated with the initiation of various forms of the disease (Hamada & Slade, 1980; Harper & Loesche, 1984; Loesche, 1986; van Houte, 1994; van Ruyven et al., 2000).

Much research has been devoted to studying streptococcal carbohydrate metabolism, particularly the biochemical and physiological adaptations that allow S. mutans to produce acids and to survive at low pH in the oral cavity. Historically, this research has employed a reductionist approach to study biochemical processes such as the allosteric regulation of enzymes and the alteration in fluxes of metabolites (Iwami & Yamada, 1980; Hamilton, 1984; Yamada, 1987; Carlsson & Hamilton, 1996; Quivey et al., 2001). A number of recent reports, however, have made use of the more holistic strategy enabled by proteome analysis to re-evaluate the mechanism of aciduricity in S. mutans. For example, an analysis by two-dimensional electrophoresis (2-DGE) of 14C-labelled cellular proteins revealed 64 proteins that were up-regulated, and 49 that were down-regulated, when cells were shocked from pH 7·5 to pH 5·0. The identities of these proteins remain to be determined (Svensäter et al., 2000). In a related study, 18 proteins were up-regulated and 12 down-regulated when a comparison was made between the 2-DGE proteomes of S. mutans grown in batch culture, without pH control, from starting pH values of 7·0 and 5·2. Of the 27 proteins identified by mass spectroscopic analysis, 13 were involved in sugar transport and central metabolism (Wilkins et al., 2002). In this latter report, S. mutans was harvested and analysed at mid-exponential phase, at which stage the pH values of the two cultures had fallen to 6·2 and 4·7, respectively. Significantly dissimilar generation times were obtained, 1·0 h and 6·6 h, at the two pHs (Wilkins et al., 2002). This variation in pH in batch culture makes an understanding of the observed changes in the proteome difficult to evaluate solely in relation to pH, since cells harvested from batch cultures have been continuously responding to an ever-changing external environment. In contrast, the continuous-culture method used in this study has the advantage of allowing cells to be sampled under strictly defined conditions at steady state, thus enabling true comparative proteome analysis when individual parameters are altered.

In the current study, the proteome of S. mutans, growing at steady state at pH 7·0 in a defined artificial salivary medium, has been compared with that of cells grown at pH 5·0. This approach has been chosen to mimic the conditions in the dental plaque of a healthy individual in comparison with those of a more susceptible host. While, at one level, this analysis has highlighted the extent of phenotypic adaptation during acid tolerance, at another it has shown that changes in the level of expression of proteins appear to be limited to key metabolic pathways; notably, glycolysis, acid production, and the synthesis of branched-chain amino acids. In contrast to previous studies, where none (Svensäter et al., 2000), or fewer than 18 % (Wilkins et al., 2002) of the proteins in these pathways were detected, this study identified 83 % of the enzymes in these three biochemical pathways. The data were consistent with the hypothesis that S. mutans adapts to an acidic environment by minimizing H+ production, using a series of different strategies. This was the case, despite the drain on ATP required for H+ extrusion and the resultant uncoupling of carbon utilization from anabolic processes.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Bacterial strain and growth conditions.
Continuous cultures of S. mutans LT11 (Tao et al., 1993) were grown under anaerobic conditions at a dilution rate (D) of 0·100±0·001 h–1, at pH 7·0±0·1 or at pH 5·0±0·1, with glucose as the limiting nutrient, as previously described (Jacques et al., 1979). DMM medium was used, without mucin, but modified to include adenine, guanine and uracil, at 20 µg ml–1, 15 mM KH2PO4 and 15 mM K2HPO4 (Sissons et al., 1991). Analyses showed that S. mutans was glucose limited at D=0·1 h–1, both at pH 7·0 and at pH 5·0, since 99·8 % and 99·9 %, respectively, of the glucose in the medium was utilized, while all of the added amino acids remained in excess (mean utilization, 75 % at pH 7·0 and 72 % at pH 5·0).

Preparation of cellular proteins.
When steady state had been achieved, the bacterial contents of the culture vessel were harvested, washed and lyophilized, before aliquots (10 mg dry wt) of cells were treated with mutanolysin (Len et al., 2003). Proteins that were to be separated on acidic immobilized pH gradient (IPG) strips (pH 4·0–6·7) were extracted as previously described (Len et al., 2003), except that 1 % (w/v) amidosulfobetaine-14 (ASB-14) and 65 mM DTT were added to produce a modified solubilizing solution for 2-DGE. While the addition of these reagents increased the total number of protein spots that could be readily discerned on 2-DGE gels, their inclusion selectively inhibited the extraction or subsequent separation of a small number of weakly expressed proteins that had been previously visualized and/or identified on 2-DGE gels (Len et al., 2003).

Proteins that were to be separated on basic IPG strips (pH 6–11) were obtained from the mutanolysin-treated cells by a two-fraction solubilization procedure. Following centrifugation of the cell lysate (12 000 g, 4 °C, 10 min), the cell pellet was stored at –20 °C, and proteins in the supernatant were precipitated overnight at –20 °C with 15 % (w/v) trichloroacetic acid. Following centrifugation (12 000 g, 4 °C, 10 min) and two washes in methanol, the precipitated proteins were solubilized in 300 µl of a 1 : 1 mixture of modified solubilization solution (without ASB-14) and Cellular and Organelle Membrane Solubilizing Reagent (Sigma-Aldrich), containing 1 % (v/v) Triton X-100 and 2 mM tributylphosphine. The frozen cell pellet was thawed and then resuspended by sonication (Branson Ultrasonics; 50 W, 10x10 s, with cooling on ice between each burst) in 700 µl of the same solubilization solution. Exonuclease III (150 U) was added, and the suspension incubated at room temperature (20–22 °C) for 15 min to degrade any DNA. The two cellular fractions were then combined and centrifuged at room temperature (12 000 g, 20–22 °C, 10 min). Prior to IEF, 100 µl iodoacetamide (500 mM) was added and the mixture incubated at room temperature (20–22 °C) for 2 h.

2-DGE.
In order to separate acidic proteins by 2-DGE, the equivalent of the amount of cellular proteins extracted from 0·75 mg dry cell weight was dissolved in 150 µl modified 2-DGE-solubilizing solution, prior to cup-loading onto narrow-range (pH 4·0–5·0, 4·5–5·5 and 5·5–6·7) IPG strips (Amersham Biosciences), pre-swollen with 350 µl of the same modified 2-DGE-solubilizing solution. For basic cellular proteins, IPG strips (pH 6–11; Amersham Biosciences) were first pre-swollen with modified 2-DGE-solubilizing solution, without ASB-14 but containing 10 % (v/v) 2-propanol. Cellular proteins extracted from the equivalent of 1 mg dry cell weight were then cup-loaded onto IPG strips. This modification resulted in better 2-DGE resolution of basic proteins.

Proteins were focused using a Multiphor II electrophoresis system (Amersham Biosciences) at 20 °C. Proteins were focussed on narrow-range IPG strips for a total of 141·5 kVh (100 V for 7 h, followed by 300 V for 6 h, 1000 V for 2 h, 2500 V for 2 h, 3500 V for 2 h and 5000 V for 25 h) and on broad-range IPG strips for a total of 79·8 kVh (100 V for 2 h, followed by 300 V for 2 h, 1000 V for 2 h, 2500 V for 2 h, 3500 V for 12 h and 5000 V for 6 h). Following separation in the second dimension on 10–18 % T gradient SDS-PAGE, the gels were stained with Sypro Ruby for image analysis, before being ‘double stained’ with Coomassie blue CBB G-250 for spot excision, tryptic digestion and mass spectroscopic analysis (Len et al., 2003). A high level of reproducibility of protein spots on Sypro Ruby-stained 2-DGE gels was observed between samples obtained from the the same growth pH (data not shown).

The density/volume, or differential expression (DE) value (arbitrary units), of each Sypro Ruby-stained protein spot was determined using the software package z3 (Compugen). In this study, the main source of error associated with this form of quantification was the reproducibility of the 2-DGE displays themselves, since biological variation was minimized through the use of a chemostat to culture the bacteria. That the 2-DGE displays were the main source of error was confirmed by comparing the DE values of 25 randomly chosen protein spots, selected from 2-DGE displays separated on broad-range IPG strips (pH 4·0–7·0). Equivalent protein spots from triplicate samples from each of three repeat continuous cultures were analysed. The data confirmed that 2-DGE was the main source of error in determining DE values. As a consequence, triplicate experimental samples from cells grown at each pH were used for all 2-DGE analyses, and the increase in the level of expression of a protein spot was based on the change in mean DE value.

Protein isoforms.
The term ‘isoform’ is used here to describe the multiple charged forms of a protein that exist on 2-DGE gels, where the mean observed Mr for each form calculated from the second (SDS-PAGE) dimension deviates by 5 % or less, and where there is no evidence from peptide-mass mapping of any form of truncation or degradation.

Mass spectrometry.
Tryptic in-gel digestion, concentration and desalting of low-copy-number proteins, and matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectroscopic analysis were carried out (as previously described) on each protein spot that was differentially or uniquely expressed at pH 7·0 or pH 5·0, or that formed part of the metabolic pathways of interest (Len et al., 2003). Mass spectroscopic analysis of protein spots from each of the replicate 2-DGE gels, at both growth pHs, was used to confirm the identity of the proteins that had been matched by the z3 software.

Protein identification.
Peptide mass mapping (PMM) analysis of proteins was undertaken as previously described, making use of the six contigs of the S. mutans UA159 genome downloaded on October 6, 2001 that were translated in all six reading frames (Len et al., 2003). The original six contigs were used in this manner as some genes identified in these contigs were not present in the final annotated version.

Parameters for protein identification included a mass tolerance of 150 p.p.m. and a maximum of one missed cleavage per peptide, while taking into consideration methionine sulfoxide and cysteine acrylamide modifications. Matches were defined on the basis of the number of matching peptide masses and the total percentage sequence covered by the peptides. As a general rule, a minimum total sequence coverage of 25 % was taken to match a given translated ORF of a high-Mr protein with confidence, though coverage as high as 80 % was observed with many low-Mr proteins. All translated ORFs that matched PMM data were then used to query the annotated S. mutans genome at the Oral Pathogen Sequence Databases (http://www.stdgen.lanl.gov/oragen), using the local BLAST search facility to determine the gene identification number. All gene names used are those associated with the S. mutans genome at the Oral Pathogen Sequence Database site. Theoretical Mr and pI were determined using MassLynx software version 3.4 (Micromass).

Amino acid analysis.
The concentrations of free amino acids in the spent medium were analysed by pre-column derivatization with 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate, using a Waters AccQFluor reagent kit (Cohen & Michaud, 1993; Cohen & DeAntonis, 1994). Amino acid derivatives were separated and quantified by reversed-phase (C18) HPLC, using a Waters AccQTag column (15 cmx3·9 mm i.d.), according to the manufacturer's protocol (Cohen, 2001). HPLC was carried out with a Waters Alliance 2695 Separation Module, 474 Fluorescence Detector and 2487 Dual l Absorbance Detector, in series. The control and analysis software was Waters Empower Pro Module. Tryptophan was analysed by UV detection, and all other amino acids by fluorescence detection.

Metabolite analyses.
The concentrations of glucose, acetate, ethanol, formate and lactate present in the medium at steady state were measured enzymically with the appropriate detection kit (Roche), according to the manufacturer's instructions (Harty & Handley, 1988). The amount of accumulated intracellular polysaccharide (IPS) was determined by the potassium iodide method of DiPersio et al. (1974).


   RESULTS AND DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
A combination of steady-state continuous-culture technology and narrow-range IPG strips allowed the detection, on Sypro Ruby-stained 2-DGE gels, of 155 differentially expressed cellular protein spots upregulated more than 1·5-fold during acid-tolerant growth of S. mutans at pH 5·0. Of these, 123 were identified by MALDI-TOF analysis and 53 found to be associated with regulatory and/or stress-responsive pathways (Len et al., 2004). The remaining 70 protein spots were associated with metabolism; the majority were associated with glycolysis, alternative acid production and branched-chain amino acid synthesis (Table 1).


View this table:
[in this window]
[in a new window]
 
Table 1. Differential expression of metabolic proteins by S. mutans grown at pH 7·0 or pH 5·0

Gene ID, gene names associated with the S. mutans genome at the Oral Pathogen Sequence Database site (http://www.stdgen.lanl.gov/oragen).

 
Glucose uptake and proton extrusion
At pH 7·0, S. mutans preferentially transports glucose by the Enzyme II complexes of high-affinity phosphoenolpyruvate : glucose phosphotransferase systems (PEP-PTS), capable of scavenging glucose molecules (Ellwood et al., 1979). Two such systems exist in S. mutans: EIIman, which has been well characterized, and EIIglc, the existence of which is primarily based on physiological observations (Vadeboncoeur, 1984; Néron & Vadeboncoeur, 1987). However, continuous-culture studies suggest that S. mutans is completely devoid of EIIman and EIIglc activity at pH 5·0 (Vadeboncoeur et al., 1991). At this acid pH, the bacterium preferentially utilizes a non-PTS glucose permease system (Hamilton & St Martin, 1982; Dashper & Reynolds, 1990; Buckley & Hamilton, 1994), while maintaining a more alkaline cytoplasm by increasing the level of its proton-translocating F1F0-ATPase (Hamilton & Buckley, 1991; Cvitkovitch et al., 1995). The recent annotation of the S. mutans genome, however, does not identify a gene encoding a specific glucose permease (Ajdic et al., 2002).

MALDI-TOF analysis failed to identify any intrinsic membrane proteins associated with the uptake of glucose in S. mutans. The failure to detect denatured intrinsic hydrophobic membrane proteins in Gram-positive bacteria is a common problem with 2-DGE proteome analysis (Len et al., 2003). Despite this limitation, two forms of the cytoplasmic component, EIIABman (ManL) of the EIIman PEP-PTS, were identified (Fig. 1, Table 1). Unfortunately, these two forms, of Mr 16 060±80 and 14 940±30, with pI 4·44±0·05 and 4·46±0·02, respectively (n=6), were C-terminally truncated. They represented only the EIIA portion of the protein, since peptide mass fingerprint coverage was limited to the EIIA region (data not shown). Although the EIIABman component of the EIIman PEP-PTS has been reported to be constitutively expressed in S. mutans (Vadeboncoeur, 1984), the truncated nature of the proteins observed in this study severely limited the interpretation of the observed 6·8-fold down-regulation of these proteins at pH 5·0. This limitation also applied to the question of whether the down-regulation reflected a reduction in the use of the glucose PEP-PTS under these acidic conditions (Fig. 1, Table 1).



View larger version (44K):
[in this window]
[in a new window]
 
Fig. 1. Comparative expression, in cells grown at pH 7·0 or 5·0, of S. mutans proteins involved in glucose transport and glycolysis. Columns in ovals represent the relative mean DE value (arbitrary units) for each protein spot identified on 2-DGE gels. DE values greater than 100 000 are shown with a height break. At pH 5·0, protein spots were up-regulated (green), down-regulated (red), non-differentially expressed (<1·5-fold difference, grey), or uniquely expressed (blue), relative to pH 7·0. Membrane proteins not identified by proteome analyses are shown in light blue rectangles. Numbers in parentheses correspond to protein numbers in Table 1; superscripts identify protein spots as N-terminal (a) or C-terminal (b) fragments.

 
In contrast, glucokinase (Glk), required for the ATP-dependent phosphorylation of any free glucose entering the cell by way of a permease, was up-regulated by 2·1-fold at pH 5·0. However, as glucokinase has a pH optimum of approximately 8·0 (Porter et al., 1982), its relative activity and stability are affected by acidification of the cytoplasm at a growth pH of 5·0, where the internal pH would be less than 6·5 (Dashper & Reynolds, 1990; Hamilton, 1990; Iwami et al., 1992). Since the optimal pH for glycolysis appears to be 7·0 (Dashper & Reynolds, 1992), the reduction in cytoplasmic pH most likely explains why glucokinase is a rate-limiting step in glycolysis when cells grown at pH 7·0 are resuspended at low pH in vitro (Iwami & Yamada, 1980). The observed increase in protein may therefore simply reflect a need to overcome this inhibition in cells growing at pH 5·0.

In studies of the aciduricity of S. mutans, the proton-translocating F1F0-ATPase that extrudes H+ has received considerable attention, since streptococcal plasma membranes are proton-permeable (Bender et al., 1986). A 7·5-fold increase in the {alpha} (AtpA) and 3·6-fold in the {varepsilon} (AtpC) subunits of the F1-component of this complex was observed at pH 5·0 (Fig. 1, Table 1). This observation was consistent with previous data showing that levels of the H+-ATPase increase in S. mutans in response to acidification of the environment in order to maintain a transmembrane pH gradient ({Delta}pH), with the interior of the cell more alkaline (Belli & Marquis, 1991; Hamilton & Buckley, 1991; Dashper & Reynolds, 1992; Quivey et al., 2001).

Glycolysis
Glucose entering the cell, by either the PEP-PTS or the permease/glucokinase pathway, enters the Embden–Meyerhof–Parnas glycolytic pathway as glucose 6-phosphate and is converted to pyruvate. Each of the enzymes in this pathway varied in the number of protein spots identified on 2-DGE gels (Fig. 1, Table 1).

Previous analysis of metabolic intermediates formed in the first half of the Embden–Meyerhof–Parnas pathway indicates that the level of glucose 6-phosphate is slightly elevated when S. mutans is grown in continuous culture at pH 5·5 and D=0·15 h–1, while the levels of all other intermediates are essentially the same, up to and including dihydroxyacetone phosphate and glyceraldehyde 3-phosphate (Iwami et al., 1992). On this basis, the doubling of glucose-6-phosphate isomerase (Gpi) and an overall 1·5-fold increase in the level of the three isoforms of fructose-1,6-bisphosphate aldolase (Fba), similar to that previously observed in batch culture (Wilkins et al., 2002; Fig. 1, Table 1), possibly reflected a need to overcome the effect of cytoplasmic acidification on enzyme activity at pH 5·0. Some care should be exercised in such a simplified interpretation, however, as both fructose 6-phosphate and glyceraldehyde 3-phosphate may be required as substrates for transketolase (Tkt), in order to increase ribonucleic acid synthesis at pH 5·0 (see below).

Not only is energy in the form of ATP generated during glycolysis, but also precursors for anabolic reactions. For instance, S. mutans possesses a specific NADP-dependent glyceraldehyde-3-phosphate dehydrogenase (GapN) that bypasses the first ATP-generating step in glycolysis in order to generate the NADPH required for reductive biosynthetic reactions (Brown & Wittenberger, 1971a; Crow & Wittenberger, 1979; Boyd et al., 1995; Fig. 1). This enzyme is essential, since S. mutans possesses neither the oxidative portion of the pentose phosphate cycle nor a transhydrogenase for the reduction of NADP by NADH (Brown & Wittenberger, 1971b). A 2·0- and 3·4-fold increase in the mean amount of the two isoforms of NADP-dependent glyceraldehyde-3-phosphate dehydrogenase at pH 5·0, along with alterations in the levels of the isoforms of the NAD-dependent glyceraldehyde-3-phosphate dehydrogenase (GapDH) and phosphoglycerate kinase (Pgk), highlighted the difficulty in determining the flux of carbon through the two branches of the pathway (Fig. 1, Table 1). However, since an overall 2·7-fold increase in NADP-dependent glyceraldehyde-3-phosphate dehydrogenase at pH 5·0 represented nearly twice the increase in the cumulative levels of each of the other two enzymes in the alternative oxidative branch of glycolysis, this suggested that NADP-dependent glyceraldehyde-3-phosphate dehydrogenase may indeed be of prime importance in determining this direction. This would be in keeping with the high level of anabolic activity required to maintain the structural integrity of molecules such as proteins and DNA in a non-optimal acidic cytoplasm (Crow & Wittenberger, 1979; Quivey et al., 2001; Len et al., 2004). This notion was supported by the detection of a high level of transketolase (Tkt) solely on 2-DGE gels at pH 5·0. Transketolase forms a junction in the glycolytic pathway that redirects glyceraldehyde 3-phosphate and fructose 6-phosphate into the pentose phosphate pathway, and ultimately to nucleotide synthesis (Fig. 1, Table 1).

Three enzymes are involved in the final conversion of 3-phosphoglycerate to pyruvate. The first of these, phosphoglyceromutase, is represented by six alleles in the S. mutans genome (Ajdic et al., 2002). Of the two protein products detected, one (PmgY) was up-regulated 2·7-fold at pH 5·0, while the other (Pgm) was absent at this pH (Fig. 1, Table 1). The four isoforms of the final glycolytic enzyme, pyruvate kinase (Pyk), were also up-regulated at pH 5·0 by 2·4-, 7·1-, 2·3- and 7·8-fold, respectively. Pyruvate kinase is believed to be the key rate-limiting step in the regulation of glycolysis in S. mutans, since it is activated by glucose 6-phosphate but inhibited by inorganic phosphate (Yamada & Carlsson, 1975b; Abbe & Yamada, 1982; Iwami & Yamada, 1980). However, the 93 % reduction of the dominant form of the preceding enzyme in the Embden–Meyerhof–Parnas glycolytic pathway, enolase (Eno), along with reductions in the other minor high-Mr and truncated forms, would be expected to have a significant effect on carbon flow at pH 5·0, especially as the pH optimum for enolase activity is 7·5 (Bunick & Kashet, 1981).

As S. mutans growing in continuous culture is devoid of measurable glucose PEP-PTS activity at pH 5·0 (Vadeboncoeur et al., 1991), phosphoenolpyruvate is not required to initiate the phosphorylation cascade for this glucose uptake system. The levels of the final three enzymes in glycolysis at pH 5·0 were consistent with a reduced pool of phosphoenolpyruvate, since the reduced levels of enolase, in conjunction with an overall 2·8-fold increase in the four isoforms of pyruvate kinase, would be expected to convert 2-phosphoglycerate to pyruvate without any build-up of phosphoenolpyruvate. This idea is supported by the measurement of the four intermediates, 3-phosphoglycerate, 2-phosphoglycerate, phosphoenolpyruvate and pyruvate, in S. mutans grown at pH 5·5 and D=0·15 h–1, where the levels of 3-phosphoglycerate and phosphoenolpyruvate decrease, that of 2-phosphoglycerate remains static and the amount of pyruvate increases (Iwami et al., 1992). This is further supported by another study, where a fall in intracellular pH was shown to significantly increase the pyruvate : phosphoenolpyruvate ratio (Iwami & Yamada, 1980). Phosphoenolpyruvate is also an inhibitor of L-lactate dehydrogenase (Yamada & Carlsson, 1975b; Yamada, 1987), so low levels of this metabolite would allow a higher concentration of lactate to accumulate at pH 5·0, as was indeed the case (Table 2; see below).


View this table:
[in this window]
[in a new window]
 
Table 2. Effect of growth pH on the yield and end-products of metabolism in S. mutans

Figures represent the mean of at least three determinations. YATP was determined according to the assumptions of Harty & Handley (1988).

 
Regeneration of NAD
In S. mutans, pyruvate can be converted into a variety of acidic end products (Fig. 2). In an aerobic environment, S. mutans utilizes pyruvate dehydrogenase to convert pyruvate to acetyl-CoA (Carlsson et al., 1985). Despite the fact that this enzyme should be inactive under the anaerobic conditions of growth utilized in this study, two of the three components of the multi-enzyme pyruvate dehydrogenase complex, the acetoin dehydrogenase E1 component, (AcoB), and a C-terminal fragment of dihydrolipoamide S-acetyltransferase (AdhC), were identified on 2-DGE gels (Fig. 2, Table 1).



View larger version (33K):
[in this window]
[in a new window]
 
Fig. 2. Comparative expression, in cells grown at pH 7·0 or 5·0, of S. mutans proteins involved in NADH oxidation and the alternative use of glucose 6-phosphate. Columns in ovals represent the relative mean DE value (arbitrary units) for each protein spot identified on 2-DGE gels. At pH 5·0, protein spots were up-regulated (green), down-regulated (red), non-differentially expressed (<1·5-fold difference, grey), or uniquely expressed (blue), relative to pH 7·0. Membrane proteins not identified by proteome analyses are shown in light blue rectangles, while those in tan ovals were identified previously, (Len et al., 2003). Numbers in parentheses correspond to protein numbers in Table 1; superscripts identify protein spots as N-terminal (a) or C-terminal (b) fragments.

 
Under anaerobic conditions, and in the presence of excess glucose, L-lactic acid is formed by L-lactate dehydrogenase (Ldh). In reducing pyruvate to L-lactate, NADH is oxidized to NAD and the oxidation–reduction balance of the cell is preserved. However, under anaerobic conditions, when glucose is limiting, the bacterium relies on the pyruvate formate-lyase pathway, which has two branches, leading either to the formation of formate and acetate or to the formation of ethanol (Carlsson & Griffith, 1974; Yamada & Carlsson, 1975b; Yamada, 1987). The ethanol branch involves two steps, in which 2 NADH are oxidized to 2 NAD. In S. mutans, these two steps, in which acetyl-CoA is first converted to acetaldehyde and then to ethanol, are catalysed by the bifunctional enzyme, alcohol-acetaldehyde dehydrogenase (AdhE). In the acetate branch of the pathway, an extra ATP is generated, as acetyl-CoA is first converted to acetyl phosphate and then to acetate, with the concomitant conversion of ADP to ATP. By appropriate use of the acetate and ethanol branches of the pyruvate-formate lyase pathway, S. mutans can readily adjust NADH oxidation to actual need. Without further anabolic diversion of the final products, the amount of formate produced by glycolysis is equivalent to the total amount of acetate and ethanol.

Proteome analysis showed that the two isoforms of L-lactate dehydrogenase detected on 2-DGE gels were up-regulated at pH 5·0 by 2·5- and 1·7-fold, compared with the levels at pH 7·0. These values are similar to reports of a total 1·7-fold increase observed on 2-DGE gels for the three isoforms of L-lactate dehydrogenase in S. mutans and the five isoforms in S. oralis when grown without pH control in batch culture (Wilkins et al., 2001, 2002). The increase in L-lactate dehydrogenase isoforms observed in our study is unlikely to account for the 127-fold increase in the amount of L-lactic acid produced under acidic conditions (Table 2), despite the fact that L-lactate dehydrogenase has a broad optimum for catalytic activity in the acidic pH range 5·0–6·2, similar to the pH range 6·0–6·5 of the cytoplasm of cells growing at pH 5·0 (Dashper & Reynolds, 1990; Hamilton, 1990; Iwami et al., 1992). The more likely explanation is that the enzyme is regulated by glycolytic intermediates at the lower growth pH (Yamada & Carlsson, 1975b; Takahashi et al., 1982; Yamada, 1987).

While no change in the level of expression of pyruvate formate-lyase was observed in the current study, since the enzyme was not detected on 2-DGE gels (Fig. 2, Table 1), the six isoforms of alcohol–acetaldehyde dehydrogenase were, on average, down-regulated 7·0-fold at pH 5·0 while the numerous C-terminal truncated forms were either down-regulated or missing at this pH. Acetate kinase (AckA), in the acetate branch, was also down-regulated by a factor of 2·1. In contrast, the level of phosphotransacetylase increased 1·7-fold at pH 5·0 (Pta; Fig. 2, Table 1). Despite the general overall reduction in protein expression observed on 2-DGE gels in these two branches of the pyruvate formate-lyase pathway, increased amounts of formate, acetate and ethanol were produced at pH 5·0, with the level of ethanol being 4·6-fold greater than at pH 7·0 (Table 2). This observation may be explained mainly by the increased glycolytic rate (Hamilton, 1984) needed to produce ATP for H+ extrusion at pH 5·0 (Bender et al., 1986; Belli & Marquis, 1991; Hamilton & Buckley, 1991; Miyagi et al., 1994; Smith et al., 1996). The fact that there was a futile use of carbon at pH 5·0 was reflected in the measured reduction in both ATP yield, (YATP; Harty & Handley, 1988) from 16·9 to 3·9 g (dry weight) cells per mol ATP, and the cell yield (YGlc), from 28·0 to 15·1 g (dry weight) cells per mol glucose, when the growth pH was reduced from 7·0 to 5·0 (Table 2). While clearly not preventing the production of alternative acids, the question remains whether the reduction in enzyme levels in the ethanol and acetate branches of the pyruvate formate-lyase pathway becomes limiting in cells growing at pH 5·0, and by so doing enhances the flow of pyruvate through the L-lactate dehydrogenase pathway.

Formation of branched-chain amino acids
A decrease in the formate : (acetate+ethanol) ratio, from the theoretical maximum glycolytic value of 1·0 to a value of 0·56, was observed at pH 5·0 (Table 2). Since the concentration of acetate was similar at both pHs, the relative reduction in formate at pH 5·0 suggested that S. mutans was further metabolizing this acid. The alternative scenario, of an increase in ethanol production by another catabolic process, could be ruled out, since ethanol can only be formed via the pyruvate formate-lyase pathway. Comparative proteome analysis suggested that the fate of some of the formate at pH 5·0 lay with the branched-chain amino acid biosynthesis pathway, since an increase was observed in the enzymes associated with this amino acid pathway alone at pH 5·0 (Fig. 3 and Table 1). The connection arises, since isoleucine biosynthesis requires 2-oxobutanoate, which is obtained by deamidation of serine by threonine ammonium lyase (IlvA, EC 4.3.1.19). Serine is formed from formate by way of the one-carbon donor cycle. The biosynthesis of both leucine and valine, on the other hand, relies directly on pyruvate (Fig. 3). Ultimately, however, branched-chain amino acid biosynthesis requires both pyruvate and NADPH for the synthesis of leucine, isoleucine and valine. De novo synthesis of these amino acids would reduce acid production in two ways, directly by removing reducing equivalents in the form of pyruvate and 2-oxobutanoate, and indirectly by the consumption of NADPH (Fig. 3). Furthermore, reducing the level of formate also reduces the concentration of H+, since formic acid is a stronger acid (pKa 3·75) than either lactic acid (pKa 3·86) or acetic acid (pKa 4·75).



View larger version (41K):
[in this window]
[in a new window]
 
Fig. 3. Comparative expression, in cells grown at pH 7·0 or 5·0, of S. mutans proteins involved in branched-chain amino acid biosynthesis. Columns in ovals represent the relative mean DE value (arbitrary units) for each protein spot identified on 2-DGE gels. Protein spots were up-regulated at pH 5·0 (green), and down-regulated at pH 7·0 (red). Proteins not identified by proteome analyses are shown in light blue rectangles. Numbers in parentheses correspond to protein numbers in Table 1.

 
At pH 5·0, the four isoforms of ketol-acid reductoisomerase (IlvC) were up-regulated 3·9-, 5·7-, 2·6- and 3·5-fold, while branched-chain amino acid aminotransferase (IlvE) and glutamine synthetase (GlnA) were up-regulated 2·7- and 8·7-fold, respectively (Fig. 3, Table 1). The product of branched-chain amino acid aminotransferase, 2-oxoglutaric acid ({alpha}-ketoglutaric acid), is also a weaker acid than glutamate, with pKa values of 2·8 and 5·0, compared with pKa values of 2·2 and 4·2 for the carboxyl groups of glutamic acid. Formation of branched-chain amino acids would have an added advantage, since the biosynthesis of all three amino acids involves the formation of NH3 by glutamine synthetase. This NH3, along with that produced from serine during the biosynthesis of isoleucine, would provide an alternative mechanism for buffering an acidic cytoplasm in S. mutans at a growth pH of 5·0, since the NH3 would react with H+ to form (Fig. 3). Interestingly, pyruvate kinase activity has an obligatory requirement for either K+ or , and a high concentration of does not seem to be detrimental to the growth of oral streptococci (Abbe & Yamada, 1982; Yamada, 1987; Pitty & Jacques, 1989).

Support for a role for branched-chain amino acid biosynthesis in the survival of streptococci at low pH comes from two recent observations. In Streptococcus thermophilus, the branched-chain amino acid biosynthesis pathway has been found to be necessary but not sufficient to ensure optimal growth, and a role has been postulated for the pathway in maintaining internal pH (Garault et al., 2000). Differential-display RT-PCR has also demonstrated that branched-chain amino acid aminotransferase is up-regulated by acidic conditions in S. mutans and that inactivation of ilvC leads to retarded growth at pH 5·5 (Chia et al., 2001). Further support for a role for branched-chain amino acid synthesis in maintaining internal pH comes from the fact that the 2-acetolactate that is synthesized from pyruvate by acetolactate synthase (AlsS, EC 2.2.1.6) occurs at the junction of the leucine/valine and acetoin biosynthesis pathways. The acetoin biosynthesis pathway leads to the formation of diacetyl (Fig. 2). While only the last enzyme in this pathway, acetoin dehydrogenase (ButA), was identified on 2-DGE gels, the amount of the enzyme at pH 5·0 was only 3 % of that at pH 7·0 (Table 1), suggesting that most of the 2-acetolactate had been directed to the leucine/valine biosynthesis pathway.

Although the proteome data, in concert with the acid and ethanol end-product analyses, were consistent with the formation of branched-chain amino acids, the concentration of these amino acids in the continuous-culture medium was only 2 % higher at pH 5·0 than at pH 7·0 (data not shown). Unlike other amino acid transport systems in S. mutans, branched-chain amino acid uptake has been well studied and has been shown to be driven by the proton-motive force, which results in high intracellular branched-chain amino acid levels (Dashper & Reynolds, 1990). Loss of these amino acids from the cell is due to passive efflux through the plasma membrane (Dashper & Reynolds, 1990). As the residual extracellular concentration of each of the branched-chain amino acids is approximately 0·3 mM at steady state in continuous culture, their cellular concentration would be as high as 10 mM (Dashper & Reynolds, 1990). As a consequence, an increase in the cellular biosynthesis of these amino acids at pH 5·0 would not necessarily have been reflected in an increase in extracellular levels. Also worthy of note is that analysis of codon usage shows that 24·4 % of the amino acids in the proteome of S. mutans are branched-chain amino acids, implying that there is a high biosynthetic requirement for these amino acids in protein synthesis.

Alterations in other biosynthetic pathways
Changes in the level of expression of individual proteins, representing incomplete metabolic pathways, were also identified by 2-DGE. The first of these was associated with the utilization of glucose 1-phosphate, following its conversion from glucose 6-phosphate by phosphoglucomutase (PgmA). A C-terminally truncated form of this enzyme was detected at pH 7·0 only (Fig. 2, Table 1). Glucose 1-phosphate, which is the initial substrate for intracellular polysaccharide synthesis, was synthesized to only a limited extent at either pH, consistent with glucose being limiting at D=0·1 h–1 (Hamilton, 1984; Table 2). Glucose 1-phosphate is also the initial substrate for four enzymes associated with the biosynthesis of rhamnose, a sugar component of S. mutans cell-wall polysaccharide (Schleifer & Kilpper-Bälz, 1987). While the first of these enzymes, glucose-1-phosphate thymidyltransferase (RmlA), existed as two isoforms at pH 7·0, only one was detected at pH 5·0, at a 2·4-fold lower level (Fig. 2, Table 1). The other three enzymes were not observed to be differentially expressed at the lower pH (Fig. 2, Table 1).

The second pathway was represented by only one protein: two isoforms of the NAD-dependent glycerol-3-phosphate dehydrogenase (GpdA), which forms the last step in glycerolipid degradation, directing carbon into the glycolytic pathway (Fig. 1). The levels of the two isoforms were reduced by factors of 2·7- and 1·9, respectively, at pH 5·0 (Table 1).

The third pathway was also represented by only one enzyme, lactoylglutathione lyase (GloA). This pathway constitutes a bypass in glycolysis, which converts the reactive dicarbonyl compound methylglyoxal into D-lactate. Methylglyoxal is toxic, as it can interact with and modify both nucleic acids and proteins (Thornalley, 1996). Methylglyoxal is formed from dihydroxyacetone and glyceraldehyde 3-phosphate, with or without catalysis by triose-phosphate isomerase. While the amount of D-lactate produced by S. mutans was constant, irrespective of the growth pH (Table 2), the level of lactoylglutathione lyase was increased 2·6-fold at pH 5·0 (Table 1). This level of up-regulation was far less than the 51-fold increase reported for S. mutans grown in batch culture without pH control (Wilkins et al., 2002).

Concluding remarks
Changes in the protein phenotype of S. mutans, leading to acid tolerance at pH 5·0, have been made possible by using a standardized, continuous-culture technique that enables a snapshot of the cellular proteome to be obtained during steady state. This differs from the situation in non-regulated batch cultures, where environmental changes are ongoing throughout growth and stationary phase. Continuous culture, in combination with enhanced 2-DGE proteome analysis with both mid-range and narrow-range IPG strips, has thus enabled us to measure the changes that take place, following a transition to an acidic environment, in the expression of metabolic proteins by S. mutans. While the cumulative changes in protein levels clearly impact on the flux of carbon through a given metabolic pathway, such changes cannot be viewed in isolation. Allosteric regulation, and also the stability and relative activity of a given enzyme at a given cytoplasmic pH, are important criteria in establishing the ultimate metabolic flux of carbon through a pathway (Yamada & Carlsson, 1975a; Yamada, 1987). Nevertheless, this study has highlighted the extent to which changes occur in the levels of enzymes involved in central metabolism.

Studies of changes in the levels of protein expression, along with measurement of the final by-products of carbon utilization, YGlc and YATP, have confirmed previous findings that streptococci tolerate growth at pH 5·0 by increasing their ability to extrude H+, while modulating the production of acid fermentation by-products to reduce the level of H+production (Hamilton, 1984; Iwami et al., 1992; Miyagi et al., 1994). The results further suggest that S. mutans utilizes branched-chain amino acid biosynthesis to lower the amount of reducing equivalents in the form of NAD(P)H that would otherwise be converted to acidic by-products, and by so doing helps buffer the cytoplasm by producing NH3. The exact role of branched-chain amino acid biosynthesis in the acid resistance of S. mutans, and the validity or otherwise of this hypothesis, is currently the subject of further investigation in our laboratory. The observed alterations in the levels of expression of proteins, however, clearly reflect the complexity required to maintain a phenotype commensurate with the survival of S. mutans in an acidic environment. In this regard, the nature of the multiple protein forms (particularly isoforms) remains to be elucidated, as do the mechanism(s) by which such an integrated change in the level of proteins is coordinated.


   ACKNOWLEDGEMENTS
 
This research was supported by Grant no. R01 DE 013234 from the Institute of Dental and Craniofacial Research, National Institutes of Health (NIH), USA, and was facilitated by access to the Australian Proteome Analysis Facility (APAF), established under the Australian government Major National Research Facility program. We wish to thank Dr S. J. Cordwell from APAF for his continued advice on all matters relating to 2-DGE proteomics, as well as Bernie McInerney and Liza Alfan for the amino acid analysis of culture fluids, and Dr K. Byth Wilson from Westmead Hospital for the statistical analyses. A. C. L. L. was the recipient of an Australian Postgraduate Award.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Abbe, K. & Yamada, T. (1982). Purification and properties of pyruvate kinase from Streptococcus mutans. J Bacteriol 149, 299–305.[Medline]

Ajdic, D., McShan, W. M., McLaughlin, R. E. & 16 other authors (2002). Genome sequence of Streptococcus mutans UA159, a cariogenic dental pathogen. Proc Natl Acad Sci U S A 99, 14434–14439.[Abstract/Free Full Text]

Belli, W. A. & Marquis, R. E. (1991). Adaptation of Streptococcus mutans and Enterococcus hirae to acid stress in continuous culture. Appl Environ Microbiol 57, 1134–1138.[Medline]

Bender, G. R., Sutton, S. V. W. & Marquis, R. E. (1986). Acid tolerance, proton permeabilities, and membrane ATPases of oral streptococci. Infect Immun 53, 331–338.[Medline]

Boyd, D. A., Cvitkovitch, D. G. & Hamilton, I. R. (1995). Sequence, expression, and function of the gene for the nonphosphorylating, NADP-dependent glyceraldehyde-3-phosphate dehydrogenase of Streptococcus mutans. J Bacteriol 177, 2622–2627.[Abstract]

Brown, A. T. & Wittenberger, C. L. (1971a). The occurrence of multiple glyceraldehyde-3-phosphate dehydrogenases in cariogenic streptococci. Biochem Biophys Res Commun 43, 217–224.[Medline]

Brown, A. T. & Wittenberger, C. L. (1971b). Mechanism for regulating the distribution of glucose between the Embden-Meyerhof and hexose-monophosphate pathways in Streptococcus faecalis. J Bacteriol 106, 456–467.[Medline]

Buckley, N. D. & Hamilton, I. R. (1994). Vesicles prepared from Streptococcus mutans demonstrate the presence of a second glucose transport system. Microbiology 140, 2639–2648.[Abstract]

Bunick, F. J. & Kashet, S. (1981). Enolases from fluoride-sensitive and fluoride-resistant streptococci. Infect Immun 34, 856–863.[Medline]

Carlsson, J. & Griffith, C. J. (1974). Fermentation products and bacterial yields in glucose-limited and nitrogen-limited cultures of streptococci. Arch Oral Biol 19, 1105–1109.[CrossRef][Medline]

Carlsson, J. & Hamilton, I. R. (1996). Differential toxic effects of lactate and acetate on the metabolism of Streptococcus mutans and Streptococcus sanguis. Oral Microbiol Immunol 11, 412–419.[Medline]

Carlsson, J., Kujala, U. & Edlund, M. B. (1985). Pyruvate dehydrogenase activity in Streptococcus mutans. Infect Immun 49, 674–678.[Medline]

Chia, J.-S., Lee, Y.-Y., Huang, P.-T. & Chen, J.-Y. (2001). Identification of stress-responsive genes in Streptococcus mutans by differential display reverse transcription-PCR. Infect Immun 69, 2493–2501.[Abstract/Free Full Text]

Cohen, S. A. (2001). Amino acid analysis using precolumn derivatisation with 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate. Methods Mol Biol 159, 39–47.

Cohen, S. A. & DeAntonis, K. M. (1994). Applications of amino acid analysis derivatisation with 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate: analysis of feed grains, intravenous solutions and glycoproteins. J Chromatogr 661, 25–34.[CrossRef]

Cohen, S. A. & Michaud, D. P. (1993). Synthesis of a fluorescent derivatizing reagent, 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate, and its application for the analysis of hydrolysate amino acids via high-performance liquid chromatography. Anal Biochem 211, 279–287.[CrossRef][Medline]

Crow, V. L. & Wittenberger, C. L. (1979). Separation and properties of NAD- and NADP-dependent glyceraldehyde-3-phosphate dehydrogenases from Streptococcus mutans. J Biol Chem 254, 1134–1142.[Medline]

Cvitkovitch, D. G., Boyd, D. A., Thevenot, T. & Hamilton, I. R. (1995). Glucose transport by a mutant of Streptococcus mutans unable to accumulate sugars via the phosphoenolpyruvate phosphotransferase system. J Bacteriol 177, 2251–2258.[Abstract]

Dashper, S. G. & Reynolds, E. C. (1990). Characterization of transmembrane movement of glucose and glucose analogs in Streptococcus mutans Ingbritt. J Bacteriol 172, 556–563.[Medline]

Dashper, S. G. & Reynolds, E. C. (1992). pH regulation by Streptococcus mutans. J Dent Res 71, 1159–1165.[Abstract]

Dashper, S. G. & Reynolds, E. C. (1993). Branched-chain amino acid transport in Streptococcus mutans Ingbritt. Oral Microbiol Immunol 8, 167–171.[Medline]

DiPersio, J. R., Mattingly, S. J., Higgins, M. L. & Shockman, G. D. (1974). Measurement of intracellular iodophilic polysaccharide in two cariogenic strains of Streptococcus mutans by cytochemical and chemical methods. Infect Immun 10, 597–604.[Medline]

Ellwood, D. C., Phipps, P. J. & Hamilton, I. R. (1979). Effect of growth rate and glucose concentration on the activity of the phosphoenolpyruvate phosphotransferase system in Streptococcus mutans Ingbritt grown in continuous culture. Infect Immun 23, 224–231.[Medline]

Garault, P., Letort, C., Juillard, V. & Monnet, V. (2000). Branched-chain amino acid biosynthesis is essential for optimal growth of Streptococcus thermophilus in milk. Appl Environ Microbiol 66, 5128–5133.[Abstract/Free Full Text]

Hamada, S. & Slade, H. D. (1980). Biology, immunology, and cariogenicity of Streptococcus mutans. Microbiol Rev 44, 331–384.[Medline]

Hamilton, I. R. (1984). Effects of changing environment on sugar transport and metabolism by oral bacteria. In Sugar Transport and Metabolism in Gram-Positive Bacteria, pp. 94–133. Edited by J. Reizer & A. Peterkofsky. Chichester: Ellis Harwood.

Hamilton, I. R. (1990). Maintenance of proton motive force by Streptococcus mutans and Streptococcus sobrinus during growth in continuous culture. Oral Microbiol Immuol 5, 280–287.

Hamilton, I. R. & Buckley, N. D. (1991). Adaptation by Streptococcus mutans to acid tolerance. Oral Microbiol Immun 6, 65–71.[Medline]

Hamilton, I. R. & St Martin, E. J. (1982). Evidence for the involvement of proton motive force in the transport of glucose by a mutant of Streptococcus mutans strain DR0001 defective in glucose-phosphoenolpyruvate : sugar phosphotransferase activity. Infect Immun 36, 567–575.[Medline]

Harper, D. S. & Loesche, W. J. (1984). Growth and acid tolerance of human dental plaque bacteria. Arch Oral Biol 10, 843–848.

Harty, D. W. S. & Handley, P. S. (1988). Fermentation products, amino acid utilization, maintenance energies and growth yields for the fibrillar Streptococcus salivarius HB and a non-fibrillar mutant HB-B grown in continuous culture under glucose limitation. J Appl Bacteriol 65, 143–152.[Medline]

Iwami, Y. & Yamada, T. (1980). Rate-limiting steps of the glycolytic pathway in the oral bacteria Streptococcus mutans and Streptococcus sanguis and the influence of acidic pH on the glucose metabolism. Arch Oral Biol 25, 163–169.[CrossRef][Medline]

Iwami, Y., Abbe, K., Takahashi-Abbe, S. & Yamada, T. (1992). Acid production by streptococci growing at low pH in a chemostat under anaerobic conditions. Oral Microbiol Immunol 7, 304–308.[Medline]

Jacques, N. A., Hardy, L., Knox, K. W. & Wicken, A. J. (1979). Effect of growth conditions on the formation of extracellular lipoteichoic acid by Streptococcus mutans BHT. Infect Immun 25, 75–84.[Medline]

Len, A. C. L., Cordwell, S. J., Harty, D. W. S. & Jacques, N. A. (2003). Cellular and extracellular proteome analysis of Streptococcus mutans grown in a chemostat. Proteomics 3, 627–646.[CrossRef][Medline]

Len, A. C. L., Harty, D. W. S. & Jacques, N. A. (2004). Stress-responsive proteins are upregulated in Streptococcus mutans during acid tolerance. Microbiology 150, 1339–1351.[Abstract/Free Full Text]

Loesche, W. J. (1986). Role of Streptococcus mutans in human dental decay. Microbiol Rev 50, 353–380.[Medline]

Miyagi, A., Ohta, H., Kodama, T., Fukui, K., Kato, K. & Shimono, T. (1994). Metabolic and energetic aspects of the growth response of Streptococcus rattus to environmental acidification in anaerobic continuous culture. Microbiology 140, 1945–1952.[Abstract]

Néron, S. & Vadeboncoeur, C. (1987). Two functionally different glucose phosphotransferase transport systems in Streptococcus mutans and Streptococcus sobrinus. Oral Microbiol Immunol 2, 171–177.[Medline]

Pitty, L. J. & Jacques, N. A. (1989). Dissimilar effects of Na+ and K+ on the promotion of glucosyltransferase secretion in Streptococcus salivarius. J Gen Microbiol 135, 1431–1439.[Medline]

Porter, E. V., Chassey, B. M. & Holmlund, C. E. (1982). Purification and kinetic characterization of a specific glucokinase from Streptococcus mutans OMZ70 cells. Biochim Biophys Acta 709, 178–186.[Medline]

Quivey, R. G., Jr, Kuhnert, W. L. & Hahn, K. (2001). Genetics of acid adaptation in oral streptococci. Crit Rev Oral Biol Med 12, 301–314.[Abstract]

Schleifer, K. H. & Kilpper-Bälz, R. (1987). Molecular and chemotaxonomic approaches to the classification of streptococci, enterococci and lactococci: a review. Syst Appl Microbiol 10, 1–19.

Sissons, C. H., Cutress, T. W., Hoffman, M. P. & Wakefield, J. S. (1991). A multi-station dental plaque microcosm (artificial mouth) for the study of plaque growth, metabolism, pH, and mineralization. J Dent Res 70, 1409–1416.[Abstract]

Smith, A. J., Quivey, R. G., Jr & Faustoferri, R. C. (1996). Cloning and nucleotide sequence analysis of the Streptococcus mutans membrane-bound, proton-translocating ATPase operon. Gene 183, 87–96.[CrossRef][Medline]

Svensäter, G., Sjögreen, B. & Hamilton, I. R. (2000). Multiple stress responses in Streptococcus mutans and the induction of general and stress-specific proteins. Microbiology 146, 107–117.[Abstract/Free Full Text]

Takahashi, S., Abbe, K. & Yamada, T. (1982). Purification of pyruvate formate-lyase from Streptococcus mutans and its regulatory properties. J Bacteriol 149, 1034–1040.[Medline]

Tao, L., MacAlister, T. J. & Tanzer, J. M. (1993). Transformation efficiency of EMS-induced mutants of Streptococcus mutans of altered cell shape. J Dent Res 72, 1032–1039.[Abstract]

Thornalley, P. J. (1996). Pharmacology of methylglyoxal: formation, modification of proteins and nucleic acids, and enzymatic detoxification – a role in pathogenesis and antiproliferative chemotherapy. Gen Pharmacol 27, 565–573.[CrossRef][Medline]

Vadeboncoeur, C. (1984). Structure and properties of the phosphoenolpyruvate : glucose phosphotransferase system of oral streptococci. Can J Microbiol 30, 495–502.[Medline]

Vadeboncoeur, C., St Martin, S., Brochu, D. & Hamilton, I. R. (1991). Effect of growth rate and pH on intracellular levels and activities of the components of the phosphoenolpyruvate : sugar phosphotransferase system in Streptococcus mutans Ingbritt. Infect Immun 59, 900–906.[Medline]

van Houte, J. (1994). Role of micro-organisms in caries etiology. J Dent Res 73, 672–681.[Abstract]

van Ruyven, F. O., Lingstrom, P., van Houte, J. & Kent, R. (2000). Relationship among mutans streptococci, "low-pH" bacteria, and iodophilic polysaccharide-producing bacteria in dental plaque and early enamel caries in humans. J Dent Res 79, 778–784.[Abstract]

Wilkins, J. C., Homer, K. & Beighton, D. (2001). Altered protein expression of Streptococcus oralis cultured at low pH revealed by two-dimensional gel electrophoresis. Appl Environ Microbiol 67, 3396–3405.[Abstract/Free Full Text]

Wilkins, J. C., Homer, K. A. & Beighton, D. (2002). Analysis of Streptococcus mutans proteins modulated by culture under acidic conditions. Appl Environ Microbiol 68, 2382–2390.[Abstract/Free Full Text]

Yamada, T. (1987). Regulation of glycolysis in streptococci. In Sugar Transport and Metabolism in Gram-Positive Bacteria, pp. 69–93. Edited by J. Reizer & A. Peterkofsky. Chichester: Ellis Harwood.

Yamada, T. & Carlsson, J. (1975a). Regulation of lactate dehydrogenase and change of fermentation products in streptococci. J Bacteriol 124, 55–61.[Medline]

Yamada, T. & Carlsson, J. (1975b). Glucose-6-phosphate-dependent pyruvate kinase in Streptococcus mutans. J Bacteriol 124, 562–563.[Medline]

Received 5 November 2003; revised 31 December 2003; accepted 12 January 2004.