TEXT

Top Abstract Text References

Oral streptococci, particularly aciduric species such as Streptococcus mutans, contribute to dental caries by degrading dietary sugars and sugar alcohols to metabolic acid end products, resulting in the demineralization of tooth mineral (4). Caries formation in the presence of readily fermentable carbohydrates, such as sucrose, has led to the use of low-cariogenic sugar substitutes, such as sorbitol (glucitol), in sugar-free gums and lozenges (3). More recently, however, the frequent use of sorbitol-containing products has been shown to result in increased levels of sorbitol-utilizing bacteria due to adaptation to sorbitol. The major sugar transport process in S. mutans is via the phosphoenolpyruvate: sugar phosphotransferase system (PTS) (17, 28), a group translocation process utilizing phosphoenolpyruvate as a substrate in phosphoryl transfer involving the general, non-sugar-specific proteins enzyme I and HPr and ultimately the sugar-specific, membrane-bound enzyme II (EII) complex, resulting in the transport and phosphorylation of the specific sugar being transported. The EII complexes are normally comprised of three functional domains, fused either within a single protein or on separate proteins, with domains IIA (formerly enzyme III) and IIB possessing the first and second phosphorylation sites, respectively, while the IIC domain forms the transmembrane channel and the sugar-binding site (17).

Early work with S. mutans revealed that sorbitol transport by glucose-grown cells required the concomitant induction of the sorbitol-PTS and sorbitol-6-phosphate dehydrogenase (SDH), resulting in the formation of fructose-6-phosphate (8, 19). Sorbitol-PTS and SDH activities were repressed by low concentrations of glucose (8, 19) by a mechanism that was at least in part due to inducer exclusion, a mechanism not observed with glucose-PTS-negative mutants. Sorbitol transport by Streptococcus sanguis also occurs via an inducible sorbitol-PTS (10, 21); however, unlike S. mutans, S. sanguis is not subject to catabolite repression by glucose, being capable of growth on glucose and sorbitol concurrently, with sorbitol utilized at a slightly lower rate than glucose. The first sorbitol-PTS to be genetically characterized was from Escherichia coli L163sr (30, 31). Sequence and expression analysis revealed the presence of the genes gutA, gutB, gutD, gutM, and gutR, coding for EIIBC, EIIA, SDH, an activator, and a repressor, respectively. Subsequent reanalysis of the E. coli L163sr sequence has revealed that the EIIC domain is encoded by two distinct genes, one half by gutA and one half by gutE, which also encodes the EIIB domain (16). The genetic designations gut, for glucitol, and srl, for sorbitol, have both been used by different groups in designating the genes from characterized sorbitol operons.

There is currently a single report of a genetically defined sorbitol mutant of S. mutans (32). This strain failed to ferment sorbitol anaerobically but did so aerobically, and it was determined that the defect was due to a chromosomal deletion that included the pfl gene, coding for pyruvate formate-lyase. Consequently, the specific aim of the present study was to clone, sequence, and identify the genes involved in sorbitol metabolism by S. mutans. A mutant strain of S. mutans LT11 defective in sorbitol metabolism was generated via transposon mutagenesis, and this strain led to the recovery of the genes coding for the sorbitol-PTS, as well as the gene coding for SDH. In addition, the identification of two regulatory genes within the sorbitol operon has led to a better understanding of the mechanism involved in the regulation of sorbitol-related enzymes.

Bacterial strains, plasmids, and growth conditions. Table 1 lists the bacteria and plasmids used in this study. The plasmids p, p, and pIS were a kind gift of R. Lunsford. S. mutans strains were maintained anaerobically at 37°C on Todd-Hewitt (TH) plates (Difco or BBL) with antibiotics as appropriate and grown for DNA isolation in TH broth supplemented with 0.3% yeast extract. Where appropriate, antibiotics were used at the indicated concentrations: erythromycin at 500 µg/ml for E. coli and 10 µg/ml for S. mutans, kanamycin (KM) at 30 µg/ml for E. coli and 300 µg/ml for S. mutans, ampicillin at 100 µg/ml, and tetracycline at 10 µg/ml. Growth studies were carried out in tryptone (1%)-yeast extract (0.5%) broth (TYE) with the appropriate carbon source.

DNA methodology. S. mutans DNA isolation, plasmid isolation, agarose gel electrophoresis, Southern hybridizations, DNA ligations, and transformation of E. coli were performed as previously described (5, 6). Transformation of S. mutans was essentially done as described by Perry et al. (15). Sequencing was carried out manually using Sequenase version 2.0 (Amersham) with the modifications described by Mytelka and Chamberlin (14) or the CircumVent Thermal Cycle DNA sequencing kit (New England Biolabs), and automatic sequencing was carried out using fluorescent dye terminators by the University of Florida (Gainesville) DNA Sequencing Core Laboratory. Custom-made primers for manual sequencing or for PCR were synthesized by the University of Calgary, University Core DNA Services, or the University of Florida DNA Sequencing Core Laboratory. The DNA sequences to complete the sequence of the srlB gene and the additional sequence downstream were obtained by directly sequencing genomic DNA using primers designed based on previously sequenced DNA. Searches for homologous proteins were carried out against the GenBank database using the BLAST suite of programs (2) at the National Center for Biotechnology Information via their World Wide Web interface (http://www.nih.nlm.ncbi/BLAST). Multiple alignment of proteins was carried out using CLUSTAL W (25).

Isolation of S. mutans BH96SR. S. mutans LT11 was transformed with p (13), and dilutions of the culture were plated out on TH-KM plates to allow growth of individual colonies. Approximately 600 transformants were picked onto TYE-sorbitol indicator plates containing KM and incubated overnight. One non-acid-producing colony (BH96SR) was tested for its ability to ferment PTS sugars on TYE-sugar indicator plates and shown to be unable to metabolize sorbitol. A marker rescue strategy involved the use of plasmid pIS (Table 1) for Tn4001 junction rescue in BH96SR as follows. Plasmid pIS was transformed into S. mutans BH96SR, and transformants were selected on TH-erythromycin plates. Integration of pIS via Campbell-type recombination at the Tn4001 copy was confirmed for six transformants by Southern hybridization analysis. The genomic DNA from one isolate, BH96SRIS, was cut with SstI, ligated at a dilute DNA concentration, and used to transform E. coli to Em^r. Fifteen E. coli transformants were screened, and all appeared to carry a plasmid with ~1 kb of DNA flanking the transposon junction. One plasmid was selected for further analysis and was named pIS-SR. The transposon sequences were removed from pIS-SR by HindIII digestion-religation to form p-SR. An 0.66-kb NspV fragment from p-SR was cloned into the ClaI site of p in both orientations to give pSR(+) and pSR(). Rescue of DNA from the sorbitol locus of LT11 was essentially performed as described above, except that EcoRI was used to cut the genomic DNA prior to rescue in E. coli after the integration of pSR(+) and pSR() to give BH97SRT+ and BH97SRT, respectively. The plasmids recovered were named pSR-EcUP and pSR-EcDN.

Inactivation of the srlD gene. A 1.1-kb BamHI fragment containing an Em^r gene was isolated from pGh9:ISS1 (22) and cloned into the BamHI site of pSDH1.6 to give pSDH-Em (Fig. 1). Linearized pSDH-Em was used to transform LT11 to Em^r, and six colonies were analyzed by Southern hybridization for the presence of the disrupted srlD gene. All six contained the Em^r gene in srlD, and one was picked for further analysis and named BH98SDH.

View larger version (13K): [in this window] [in a new window]   FIG. 1.   Schematic representation of the genetic characterization of the sorbitol transport operon in S. mutans LT11 obtained by sequencing. Inserts of plasmids used for sequencing, marker rescue, and/or complementation are indicated under the restriction map. A double-headed arrow represents the point of insertion of Tn4001 in S. mutans BH96SR. A putative hairpin structure between gpi and srlD is indicated. Restriction enzyme sites are BamHI (B), BglII (Bg), EcoRI (E), HindIII (H), MunI (M), NspV (N), and SstI (S). Gene symbols: orf1, unknown; gpi, glucose-6-phosphate isomerase; srlD, sorbitol-6-phosphate dehydrogenase; srlR and srlM, regulatory proteins; srlA to srlE, EIIBC; and srlB, EIIA.

Characterization of S. mutans BH96SR and rescue of the sorbitol locus from S. mutans LT11. The Tn4001 system (13) was employed with S. mutans LT11 to successfully isolate a strain, BH96SR, that did not ferment sorbitol but did ferment all other PTS sugars tested, indicating that the transposon was not located in the ptsHI operon. We used a marker rescue strategy to isolate approximately 1 kb of genomic DNA flanking one end of the transposon insertion in BH96SR(p-SR) and determined its nucleotide sequence. Analysis of the sequence revealed that, in BH96SR, Tn4001 had inserted itself into an open reading frame (ORF) whose putative translation product showed homology to several transcriptional regulators from Bacillus subtilis, including the LicR protein from a -glucoside PTS operon that transports lichenan (-1,3-1,4-glucan) degradation products, including cellobiose (26). Thus, it appeared that the transposon insertion inactivated a putative regulator of the sorbitol operon of LT11, causing the sorbitol-negative phenotype of BH96SR. To further characterize the upstream and downstream regions, the marker rescue plasmids pSR(+) and pSR() were used to obtain two overlapping clones (pSR-EcUP and pSR-EcDN) whose inserts represent approximately 8.5 kb of contiguous sequence from the LT11 genome (Fig. 1).

DNA sequence analysis. The inserts from pSR-EcUP and pSR-EcDN were completely sequenced to give a total of 7,957 bp defining an EcoRI fragment from this region (data not shown). A further 495 bp of sequence was obtained by direct genomic sequencing, PCR product sequencing, and sequencing from an additional overlapping clone, pALH128 (Fig. 1). Analysis of the 8,452 bp of total sequence obtained revealed a total of eight ORFs, all in the same orientation, whose putative translation products were used as query sequences in BLAST searches of the GenBank database to identify the corresponding genes (Fig. 1). ORF1 (564 bp) begins with an ATG codon and encodes a putative protein of 187 amino acids that produced no significant matches with any entries in GenBank. After an intergenic region of 556 bp begins an ORF (1,350 bp, ATG start codon) whose product (449 amino acids) shares between 59 and 66% identity with the glucose-6-phosphate isomerase A and B isozymes, respectively, from Bacillus stearothermophilus and 62% identity to the enzyme from B. subtilis. We, therefore, designated this ORF as the gpi gene of S. mutans LT11. We have detected a region of dyad symmetry (G = 71 kJ) beginning 18 bp downstream of the gpi stop codon and ending 113 bp upstream of the srlD start codon. This may play a role in transcriptional termination of gpi and/or in regulation of the sorbitol operon expression. The remaining six ORFs appear to constitute the genetic components of the sorbitol-PTS operon of S. mutans. Separated from the gpi gene by 164 bp is the srlD gene (801 bp, ATG start codon) coding for a protein of 266 residues sharing 58% identity with the gutD gene coding for a SDH from the sorbitol-PTS operon of Clostridium beijerinckii (22) and 57% identity to the sorD gene encoded by the L-sorbose PTS operon of Klebsiella pneumoniae (29). This system encodes components to transport and phosphorylate sorbose to sorbose-1-phosphate and to reduce it to sorbitol-6-phosphate followed by conversion to fructose-6-phosphate by the SDH. Interestingly, the S. mutans SDH shares only 28% identity with the SDHs from the sorbitol-PTS operons of E. coli and Erwinia amylovora (1, 30). Twenty base pairs downstream of the stop codon of the srlD gene is a GTG codon, which we believe signifies the start of a gene that we have designated srlR, coding for a putative transcriptional regulator protein. There is a ribosomal binding site (RBS) motif, AGAGGG, located 7 bp upstream of the GTG codon, whereas there is no suitable RBS upstream of the first or second ATG codons of the ORF. The ORF (1,866 bp) codes for a protein of 621 residues (SrlR), which shares 22% identity with the LicR regulator (641 residues) from the -glucoside PTS operon from B. subtilis (26) and 19.5% identity with the MtlR regulator (697 residues) from the mannitol-PTS operon from B. stearothermophilus (11). As a consequence of the B. subtilis genome sequencing project, two other similar putative regulators have been identified, YjdC (648 residues), sharing 21.5% identity with SrlR, and YdaA (694 residues), sharing 18% identity with SrlR. All five proteins share between 18 and 27% identity with each other, except for the MltR and YdaA proteins, which are 40.5% identical, indicating a closer relationship (data not shown). Overlapping the stop codon of srlR by 1 nucleotide is srlM (489 bp, ATG start codon), coding for a homolog of the GutM activator protein from the E. coli sorbitol operon (31). There is another ATG codon 7 nucleotides downstream; however, the spacing between the putative RBS (AAGGAG) and the first ATG is more optimal (7 versus 16 bp). An alignment between SrlM (162 residues) and GutM (119 residues) reveals that they share 23.5% identity and that the additional amino acids in SrlM are located at the C terminus (data not shown). The next three ORFs code for the EII components of the system as identified by their similarity to their E. coli and C. beijerinckii homologs. Beginning 79 bp downstream of srlM is srlA (543 bp, ATG start codon) coding for one half of the EIIC domain (180 residues), which shares 58 and 52% identity with the EIICs from C. beijerinckii and E. coli, respectively. Beginning 90 bp after the end of srlA is srlE (1,011 bp, ATG start codon), coding for a fusion protein (336 residues) consisting of the other half of the EIIC domain fused to an EIIB domain and sharing 65 and 57% identity with the EIIBCs from C. beijerinckii and E. coli, respectively. Beginning 42 bp downstream of srlE is srlB (369 bp, ATG start codon) coding for the EIIA subunit (121 residues), which shares 41 and 33% identity with the EIIAs from C. beijerinckii and E. coli, respectively. We could detect no other ORFs downstream of srlB or a structure resembling a possible transcriptional terminator. However, we did find evidence of a nonfunctional transaldolase-like-protein-encoding gene similar to the one in the sorbitol operon of C. beijerinckii (22).

Growth characteristics of S. mutans LT11 and sorbitol mutants. When S. mutans LT11 was grown in medium with sorbitol as the sole carbon source, there was a long lag period (5 to 6 h), indicating that the proteins involved in sorbitol transport and metabolism are inducible (Fig. 2A). In medium containing glucose or glucose-sorbitol, a shorter lag period of 1.5 to 2 h was observed (Fig. 2A). Analysis of the glucose-sorbitol culture medium showed that glucose was utilized to near exhaustion before sorbitol was consumed (Fig. 2B). A very short lag period was observed during transition from growth on glucose to growth on sorbitol, indicating that the enzymes required for sorbitol utilization were induced prior to complete exhaustion of glucose. These observations indicate that sorbitol metabolism in S. mutans is inducible and subject to catabolite repression in the presence of glucose, confirming earlier research (8, 19). In similar experiments with S. mutans BH96SR, growth was unimpaired in medium containing glucose or glucose-sorbitol, and as expected, this S. mutans strain failed to grow when sorbitol was the sole carbon source (Fig. 3). Strains BH97SRT+ and BH97SRT (Table 1) failed to utilize sorbitol when grown on indicator plates, as did BH98SDH, containing an insertionally inactivated srlD gene. Analysis of the sequence would seem to indicate that, in both cases, it is likely that there is a polar effect by the insertions in these strains on the transcription of the downstream genes. Transcript analysis, however, would be needed to confirm this hypothesis. We transformed pSR-EcUP and pSDH1.6 into E. coli JWL163, which contains a mutation in its srlD gene (12). When plated on MacConkey-sorbitol plates, the plasmid-containing colonies were dark red, indicating sorbitol utilization. Thus, the expression of the srlD genes from these plasmids in E. coli is not dependent on the presence of the S. mutans SrlR or SrlM protein.

View larger version (14K): [in this window] [in a new window]   FIG. 2.   (A) Growth of the wild-type S. mutans LT11 on glucose (), sorbitol (), and glucose-sorbitol (). (B) Glucose () and sorbitol () utilization in the glucose-sorbitol culture.

View larger version (14K): [in this window] [in a new window]   FIG. 3.   (A) Growth of the mutant S. mutans BH96SR on glucose (), sorbitol (), and glucose-sorbitol (). (B) Glucose () and sorbitol () utilization in the glucose-sorbitol culture.

View larger version (9K): [in this window] [in a new window]   FIG. 4.   The structure of the sorbitol-PTS operons from S. mutans, C. beijerinckii, E. coli, and E. amylovora. The operons are shown arbritrarily aligned by their respective srlA and srlE genes.

View larger version (28K): [in this window] [in a new window]   FIG. 5.   Alignment of the PRD-I and -II domains from the YdaA, LicR, and YjdC proteins from B. subtilis (Bs); the MtlR protein from B. stearothermophilus (Bt); and the SrlR proteins from S. mutans LT11 (Sm). Identical residues present in all domains at positions 14 and 61 are boxed, positions with similar residues in at least five domains are indicated by a period, and positions with identical residues in at least five domains are indicated by an asterisk. Positions 7, 14, 61, and 69 of the alignment (see text) are indicated at the top.

Nucleotide sequence accession number. The sequence of the region depicted in Fig. 1 has been submitted to the GenBank database under accession no. AF132127.