INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Coaggregation of human oral bacteria is thought to occur by specific interactions between complementary surface molecules on the partner cells (64). Cell-surface adhesins on one cell type may recognize and bind to complementary receptors on the partner cell type. Most of the interactions are among members of different genera, for example Streptococcus spp. and Actinomyces spp., and are termed intergeneric coaggregations. Streptococci also participate in intrageneric coaggregations, which occur among the human oral viridans streptococci, and these coaggregations are galactoside inhibitable (38). A 100-kDa putative adhesin on the surface of Streptococcus gordonii DL1 was proposed to mediate specifically these galactoside-inhibitable intrageneric coaggregations, because streptococcal insertional mutants as well as spontaneous coaggregation-defective mutants lost specifically the intrageneric coaggregation capability and lack this protein (9, 10, 63).

A potential role of lipoteichoic acid (LTA) in coaggregation is the proper presentation of adhesins and receptors on partner cells. LTAs are macroamphiphiles that contain alditolphosphates as integral parts of the hydrophilic chain (17). Most glycerol LTAs are substituted with D-alanyl ester residues (18). Because LTA is polyanionic, it binds Ca⁺⁺ ions (53) and may contribute to the proper environment for coaggregation, which requires divalent cations, especially Ca⁺⁺ (38, 46). Out of a total of 86 strains examined, LTA was found in all viridans streptococci except Streptococcus mitis and Streptococcus oralis (28). Since LTA inhibits the attachment of S. gordonii to substratum-located glucan polymer, the adhesion is thought to be mediated by LTA (60). LTA is reported to mediate binding of group A streptococci to fibronectin receptors on pharyngeal epithelial cells (4), and it inhibits binding of human oral viridans streptococci to fibronectin-coated spheroidal hydroxyapatite beads (27). LTA is thought to be important in the first of two steps of adhesion to human cells; the second step is postulated to occur by a specific adhesin(s) that determines tissue tropism (24). LTA exhibits properties of an enterococcal binding substance that is recognized by a mating cell-expressing aggregation substance (16), and the resulting union of the two cell types is part of the well-studied pheromone-inducible conjugation system in Enterococcus faecalis (12, 14).

The dlt operon involved in the synthesis of the D-alanyl esters of LTA was discovered in Lactobacillus rhamnosus (25, 26, 47) and subsequently identified in Bacillus subtilis (21, 50). The operons encode, respectively, four and five genes in these organisms. The first gene, dltA, encodes the D-alanine-D-alanyl carrier protein ligase (Dcl), which catalyzes the D-alanylation of the D-alanyl carrier protein (Dcp) encoded by dltC. Dcp in turn transfers the D-alanine to a membrane acceptor for the D-alanylation of LTA. There is a coding sequence overlap between dltA and -B and between dltC and -D with putative ribosome binding sites preceding dltA and dltC (47). On the basis of the hydropathy profile of the putative DltB, it is hypothesized that DltB is located in the cytoplasmic membrane and displays 12 membrane-spanning domains. DltB is hypothesized to be involved in the efflux of activated D-alanine to the site of LTA acylation (47).

The function of the D-alanine esters has been a point of recent investigation for several genera of gram-positive bacteria. In Lactobacillus rhamnosus ATCC 7469 they appear to play a role in determining cell shape and cell septation (48), whereas in B. subtilis, the absence of D-alanine esters has no effect on ultrastructure or cell septation but does enhance autolytic and beta-lactam-induced cell lysis (62). D-Alanine esters provide limited protection to B. subtilis JH642 against methicillin but do not protect against phagocytosis and degradation of the bacterium in macrophages (61). In Staphylococcus aureus mutant strains defective in formation of D-alanine esters, the cells exhibited reduced autolysis and enhanced expression of methicillin resistance (49). Increased resistance to vancomycin in Enterococcus faecium D366 was accompanied by a doubling of the D-alanyl ester content of LTA (22). It was proposed that this event would reduce the ability of autolysins to bind to the heavily D-alanylated LTA, which may affect a later step in the pathway that triggers autolytic and beta-lactam-induced cell lysis. Insertion of ISS1 into dltD resulted in Lactococcus lactis MG1363 becoming UV-sensitive suggesting that cell envelope integrity and the ability to repair DNA are related (15). The L. lactis dltD mutant also grew more slowly and formed longer chains than the parent strain. Thus, the D-alanyl esters of LTA would appear to play a variety of roles in gram-positive organisms.

Here we report that inhibition of D-alanyl-LTA biosynthesis by specific insertional mutagenesis in S. gordonii DL1 resulted in undetectable levels of a 100-kDa putative adhesin on the cell surface, prevented intrageneric-galactoside-inhibitable coaggregation, contributed to altered cellular division and morphology, and reduced DNA transformation frequency. In contrast, intergeneric coaggregation with actinomyces was unchanged. We propose that the absence of D-alanyl esters disrupts one or more of the normal scaffolding functions of LTA in adhesin presentation on the streptococcal cell surface and in a variety of cellular processes including transformation, adherence, and cell division.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cultivation of bacteria. All bacterial strains used in this study are listed in Table 1. Streptococci and actinomyces were cultured in CAMG medium (9) at 37°C under anaerobic conditions with the GasPak system (BBL Microbiology Systems, Cockeysville, Md.) (2). S. gordonii strains containing transposon Tn916 were grown on CAMG medium supplemented with 10 µg of tetracycline (Sigma Chemical Co., St. Louis, Mo.) per ml of medium. All incubations for the generation of competent S. gordonii DL1 were carried out at 37°C. Transformants in the dlt operon were obtained by the method previously described (63) by using the transformation medium of LeBlanc and Hassel (39). The resultant dlt insertion mutants were tested for competence and transformation frequency by the method of Lunsford (42). All Escherichia coli strains were cultured aerobically at 37°C in Luria-Bertani (LB) broth or on LB agar (Gibco-BRL). LB medium supplemented with ampicillin (100 µg per ml), tetracycline (4 µg per ml), erythromycin (300 µg per ml), or chloramphenicol (20 µg per ml) was used to select for E. coli strains containing various plasmids (Table 2). E. coli CG120 was cultured in medium containing 100 µg of ampicillin and 4 µg of tetracycline per ml.

Isolation of Tn916-insertion mutants. The procedure of Behnke (5) was followed for the transformation of S. gordonii DL1 cultures with plasmid DNA as described previously (63). Tetracycline-resistant transformants were picked onto fresh CAMG agar medium containing tetracycline and incubated at 37°C under anaerobic conditions. The colonies were screened (63) for the absence of coaggregation with Streptococcus oralis 34 by the microtiter plate assay described previously by Kolenbrander (37). Cells of coaggregation-defective cultures were frozen in a dense suspension in CAMG broth and stored at 40°C.

Recombinant DNA methods. The procedures for preparing plasmid DNA from E. coli have been described previously (63), and genomic DNA from S. gordonii was prepared by the procedure described by Andersen et al. (2). To clone the region adjacent to the Tn916 transposon, we used the single HindIII site located 6.5 kb from the left end (19). The 7.1-kb fragment from HindIII-digested (New England BioLabs, Beverly, Mass.) genomic DNA from the Cog Tn916 insertion mutant G7 was cloned into the HindIII site of pUC19 (Gibco-BRL, Gaithersburg, Md.) (63). The recombinant plasmid containing the 7.1-kb HindIII fragment was called pDC3 (Table 2), and the E. coli DH5 strain containing pDC3 was designated PK3322.

Insertional inactivation of the S. gordonii DL1 dltA gene with ermAM. Restriction sites for KpnI and BamHI were incorporated into the DNA primers with the respective 5' and 3' sequences of the dltA gene and used to amplify a 1.4-kb fragment of the dltA gene from pDC5 as template. The two primer sequences were as follows: PCR1, 5'-CCGGATCCTGACCTCGCTGATTAAGCCC-3'; PCR2, 5'-GGGGGTACCTCTCCTGTCGTGGTCTATGGTGGGC-3'. PCR of the S. gordonii DL1 dltA homolog was performed with the GeneAmp PCR core reagents (Perkin-Elmer Cetus, Norwalk, Conn.). pDC5 DNA (5 ng) was mixed with 1 µM each primer, 2 mM MgCl₂, and 2.5 U of AmpliTaq DNA polymerase (Perkin-Elmer Cetus) in a final reaction volume of 50 µl. The thermocycler program (Perkin-Elmer Cetus) was 4 min at 95°C; 30 cycles of 1 min at 95°C, 1 min at 55°C, and 1 min at 72°C; and a final elongation of 5 min at 72°C. The 1.4-kb dltA fragment contained a single HincII site near the center and was used to clone a blunt-ended 922-bp ermAM cassette from pKSerm2 (43). The recombinant plasmid was linearized with BamHI and used to transform S. gordonii DL1 (63) to disrupt the dltA gene with ermAM. Two transformants, PK3241 and PK3242, were chosen for study. Southern analyses of AccI and ScaI restriction endonuclease-digested genomic DNAs from strains DL1, PK3241, and PK3242 were performed according to the procedure of Whittaker et al. (63) and were conducted in order to confirm insertion of ermAM into the S. gordonii dltA homolog. Both the 0.9-kb ermAM and the 1.4-kb dltA probes were end labeled with ³²P by using T4-polynucleotide kinase (Lofstrand Labs Limited, Gaithersburg, Md.).

Computer analyses. Sequence assembly and analyses were performed with Assembly-LIGN and MacVector (International Biochemicals, Inc., New Haven, Conn.). Database searches were conducted with the BLAST algorithm (1) by using the current version of the nonredundant protein database at the National Center for Biotechnology Information (NCBI-NLM, Bethesda, Md.). Protein alignments were performed with PILEUP, GAP, and PRETTY programs (Genetics Computer Group, version 9.1, for VMS-VAX).

Immunoblot analyses. The proteins from streptococcal surface-sonicate extracts were separated on sodium dodecyl sulfate-6% polyacrylamide gel electrophoresis gels (Novex, San Diego, Calif.) as described previously (63). Detection of the 100-kDa putative adhesin was done by previously published methods with anti-DL1 polyclonal serum that had been adsorbed with spontaneously occurring coaggregation-defective mutant strain PK1897 (9, 63).

Streptococcal surface protein preparations. Streptococcal cells were washed three times with distilled water by centrifugation (10,000 × g for 10 min at 4°C) and resuspended to a concentration of 0.5 g (wet weight) cells per ml of distilled water up to a maximum of 1 ml and sonicated on ice for 1 min at maximum power (50 W) with a Micro Ultrasonic cell disrupter (Kontes, Vineland, N.J.). The sonicated suspension was centrifuged at 20,000 × g for 15 min at 4°C, and the supernatant fluid was stored frozen until used. Total protein was determined with the Bio-Rad protein assay (Bio-Rad, Richmond, Calif.).

Electron microscopy. Exponential-phase cells were fixed overnight in 0.1 M cacodylate buffer (pH 7.4) containing 2.5% (wt/vol) glutaraldehyde at 4°C. For scanning microscopy fixed samples were washed four times with 0.1 M cacodylate buffer. A drop of each sample on a glass coverslip was dehydrated in a graded series of ethanol concentrations, critical point dried, and sputter coated with gold-palladium. Specimens were examined with a JEOL JSM-1 or an ultra-high-resolution Hitachi S-4500 microscope. For transmission microscopy the fixed cells were washed four times with 0.1 M cacodylate buffer and postfixed for 1 h at 25°C with 0.5% (wt/vol) osmium tetroxide-0.8% (wt/vol) potassium ferricyanide in 0.1 M cacodylate (pH 7.4). After washing the samples three times with water, they were stained with 0.5% (wt/vol) uranyl acetate for 1 h at 25°C and washed three more times. The samples were pelleted in 2% (wt/vol) agar, cut into 1-mm cubes, and dehydrated in increasing concentrations of ethanol (34). The samples were infiltrated with Spurr's resin (56), embedded in BEEM capsules, and sectioned (80 nm, silver) on a Reichert Ultracut-E. The sections were poststained with 1% (wt/vol) uranyl acetate for 20 min and examined with a JEOL JEM-100CX 11 microscope at 80 kV.

D-Alanine incorporation assay. Incorporation of D-[¹⁴C]alanine into toluene-treated cells was performed in a reaction mixture (20 µl) which contained 30 mM bis-Tris (pH 6.5), 10 mM ATP, 10 mM MgCl₂, 1 mM dithiothreitol, 0.11 mM D-[¹⁴C]alanine (43 mCi/mmol), and 4 µl of permeabilized cell suspension. For the preparation of permeabilized S. gordonii, a modification of the procedure of St. Martin and Wittenberger (57) was used. The cells (0.65 gm/ml), suspended in 100 mM phosphate buffer (pH 7.0), were treated with 0.15 volumes of toluene-acetone (1:9 [vol/vol]) for 1 min of vigorous agitation on a Vortex mixer. The agitation was repeated three more times with intermittent cooling in an ice bath. The reaction mixture was incubated at 37°C, and the reaction was terminated by the addition of 0.9 ml of 30 mM bis-Tris (pH 6.5). The labeled cells were collected on a GN-6 Metricel filter and washed with three 1 ml and one 10 ml portion of 30 mM bis-Tris (pH 6.5). The filters were dissolved in ethyl acetate and assayed for radiolabel.

Nucleotide sequence accession number. The assigned GenBank accession number is AF059609.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of S. gordonii DL1 Cog mutant. Transposon mutagenesis was conducted (63) to isolate Cog mutants that failed to coaggregate with streptococcal partners of the wild-type strain DL1. Transposon Tn916 was inserted into the chromosome of S. gordonii DL1 by transformation of strain DL1 with the donor plasmid pAM120 (19, 63). This process yielded mutant strain G7, which lost specifically the ability to coaggregate with streptococcal partners while exhibiting wild-type levels of coaggregation with an actinomyces partner (Table 3). Intergeneric coaggregations of the mutant or wild type with Actinomyces naeslundii ATCC 51655 are not inhibited by lactose, whereas the intrageneric coaggregations between the wild-type DL1 with S. oralis 34, S. oralis C104, and Streptococcus sp. strain SM PK509 are inhibited by lactose (38). In these galactoside-inhibitable coaggregations, S. gordonii DL1 was sensitive to heat and protease treatment and was presumed to contain a putative adhesin which recognizes a carbohydrate-containing receptor on the surface of its partner strains (38). On the basis of the current work with Cog mutant G7, it was discovered that Tn916 inserted into a new coaggregation-relevant locus, which is different from the locus in the Cog insertion mutant described previously by Whittaker et al. (63).

Cloning and sequencing of the streptococcal DNA flanking the Tn916 insert. To obtain the nucleotide sequence of the coaggregation-relevant locus, the HindIII junction fragment (7.1 kb), which included the left end of Tn916 and flanking S. gordonii DNA, was isolated and cloned into the HindIII site of plasmid pUC19 to yield pDC3. The 626-bp sequence of the streptococcal DNA flanking the left end of the HindIII fragment in strain G7 was obtained and is presented as a partial restriction map (Fig. 1A). Translation of the two partial open reading frames (ORFs) showed significant sequence identity to DltA (D-alanine-D-alanyl carrier protein ligase [Dcl]) and DltB (13, 25, 26). The two partial ORFs were 519 bp (ORF 1, dltA homolog) and 111 bp (ORF 2, dltB homolog) in length with the 3' end of ORF 1 sharing four nucleotides with the 5' end of ORF 2. A similar overlap of dltA and dltB in the dlt operon from L. rhamnosus was reported by Heaton and Neuhaus (25). Both ORFs appear to be transcribed towards the Tn916 insertion with dltB being directly interrupted by the transposon insert. To examine the possibility that there was more than one streptococcal locus homologous to dltA from L. rhamnosus, a 60-bp probe (Fig. 1A) was used for Southern blot analysis of BamHI-, ClaI- and HindIII-digested genomic DNA from S. gordonii. A single hybridizing band was observed in genomic digests with each of the three restriction enzymes (data not shown), suggesting a single locus for the dlt operon. The 60-bp probe (G7) hybridized with L. rhamnosus dltA cloned into pDAE1. These data confirmed the presence of a single homologous sequence in S. gordonii.

View larger version (50K): [in this window] [in a new window]   FIG. 1.   (A) Partial restriction map of the streptococcal flanking region and the left end of Tn916 from mutant G7. The location of the 60-bp G7 probe is indicated. Putative transcription orientation is from left to right. Abbreviations of restriction enzyme sites: C, ClaI; H, HindIII. (B) Comparison of deduced amino acid sequence of the 626-bp fragment from S. gordonii (panel A) with truncated sequences of DltA and DltB homologs from L. rhamnosus, S. aureus, B. subtilis, S. mutans, and S. pyogenes. Upper case reverse font indicates amino acid sequence identity in five or more proteins; lower case reverse font indicates identity among all three streptococcal strains. Regions II and III are indicated and are discussed in the text. The 60-bp G7 probe includes the sequence encoding the first 19 amino acids of region II. S. pyogenes, contig 320 (51a); S. mutans, accession no. AF049357 (55a) accession no. AF051356 (5a); B. subtilis, accession no. X73124 (21); Staphylococcus aureus, accession no. D86240 (46a) (nucleotide 3939 was changed from A to C to remove false stop codon); L. rhamnosus, accession no. U43894 (25). (C) Strategy for obtaining the ermAM insertion in the dltA gene of S. gordonii DL1. The 1,385-bp region was obtained by PCR with primers PCR1 and PCR2 (Materials and Methods). The location of the KpnI (K) and BamHI (B) restriction sites are shown. A portion of panel A is shown for orientation. The HincII site (Hc) of insertion of the 922-bp ermAM cassette is indicated. The transcriptional orientation of the cassette and the ScaI site (S) are indicated. The dltA fragment was cut with KpnI and BamHI and cloned into KpnI- and BamHI-digested pBluescript IIKS(+) (Stratagene).

Cloning and expression of the S. gordonii DL1 dltA. To determine if a functional dltA gene product could be expressed from a cloned streptococcal dltA, we used the 60-bp G7 probe to locate the putative dltA locus in the fosmid-constructed genomic library of wild-type S. gordonii DL1 (Challis). Three fosmids hybridized with the 60-bp G7 probe; one was selected for further study and designated pDC5. Recombinant E. coli PK3324 containing the fosmid pDC5 exhibited Dcl activity (6.2 U/mg of protein) which was comparable to that of the cloned L. rhamnosus dltA in E. coli (11.2 U/mg of protein). These data indicated that the cloned streptococcal dltA in pDC5 encodes Dcl.

Insertional inactivation of the S. gordonii DL1 dltA with ermAM. To prepare plasmid for insertional inactivation, it was necessary to sequence dltA. Fosmid pDC5 was chosen as the template for sequencing the 5'-region flanking the 626-bp fragment described in Fig. 1A. From the sequence two PCR primers were constructed to amplify a fragment containing partial dltA (Fig. 1C). This product, which hybridized with the 60-bp G7 probe, was cloned into pBluescript to form pDC8. To insertionally inactivate dltA, the 0.9-kb streptococcal erythromycin resistance determinant (ermAM) was ligated into the unique HincII site (Fig. 1C, Hc) within the 1.4-kb fragment of cloned DNA in pDC8 (43, 45), and the resulting plasmid was designated pDC11. By digesting this plasmid with ScaI, which cuts the ermAM cassette asymmetrically (Fig. 1C), it was shown that the ermAM determinant was cloned in the opposite transcriptional orientation to the dltA. Linearized plasmid pDC11 DNA was transformed into S. gordonii DL1. Two Em^r isolates, PK3241 and PK3242 (Table 3), unable to coaggregate with streptococcal partners but still able to coaggregate with A. naeslundii ATCC 51655, were identified and used for further study.

Characterization of Em^r S. gordonii DL1 dltA insertion mutants PK3241 and PK3242. The insertion mutants exhibited an inability to participate in galactoside-inhibitable coaggregation with other streptococci while maintaining strong noninhibitable coaggregations with actinomyces, for example A. naeslundii ATCC 51655 (Table 3). The wild-type strain DL1 coaggregated strongly with the streptococcal partners, and these coaggregations were reversed by adding lactose. Also, the growth of the mutants was slower than that of the wild type, with doubling times at 37°C for S. gordonii DL1, PK3241, and PK3242 of 51, 96, and 117 min, respectively. Thus, insertional inactivation of dltA results both in a loss of galactoside-inhibitable coaggregation and in a slowing of the growth rate compared with that of the parent strain DL1.

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Incorporation of D-[14C]alanine into the parent and mutant strains. The D-alanine incorporation assay (described in Materials and Methods) was used with permeabilized cells of the indicated strains.

View larger version (48K): [in this window] [in a new window]   FIG. 3.   D-Alanine ester content of S. gordonii wild-type DL1 and mutant strains PK3241 and PK3242. As described in the text, a Western blot analysis using serum with antibodies to D-alanyl esters (40) was used to detect the presence of D-alanyl esters. (A) Photograph of plate; (B) colony immunoblot.

View larger version (143K): [in this window] [in a new window]   FIG. 4.   Scanning electron micrographs of DL1 (A and D) and PK3241 (B and C). The specimens shown in panels A to C were examined with a JEOL JSM-1 microscope. Bar = 1.0 µm. The specimen shown in panel D was examined with a Hitachi S-4500 microscope after fixing and coating with gold-palladium (Materials and Methods). Bar = 0.5 µm. The arrow in panel B indicates nonlinear fission and the arrows in panel C indicate multiseptate pleomorphs.

View larger version (133K): [in this window] [in a new window]   FIG. 5.   Aberrant morphology of PK3241. Exponential-phase cells of DL1 (A) and PK3241 (B to D) were fixed, embedded, sectioned, and examined with a JEOL-100CX 11 microscope as described in Materials and Methods. Bars = 0.5 µm.

View larger version (27K): [in this window] [in a new window]   FIG. 6.   Immunoblot analysis of sonic surface extracts of PK3241 and PK3242, ermAM insertion mutants in dltA, compared with wild-type S. gordonii DL1. The position of the putative adhesin (arrow) was determined by using prestained molecular weight standards (Bio-Rad). The standards of 112 and 84 kDa are indicated by dashes.

3

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This is the first report of a mutation in the dlt operon resulting in altered adherence properties. The dltA mutants, PK3241 and PK3242, of S. gordonii DL1 specifically lost the ability to participate in intrageneric coaggregations while maintaining intergeneric coaggregations. The concomitant loss of the 100-kDa putative adhesin suggests that this protein binds to D-alanyl LTA. LTA has been implicated in the adhesion of a number of gram-positive bacteria to a variety of target cells and surfaces (6, 24, 58). In these examples, the participation of LTA as a direct mediator of adhesion has been postulated. In contrast, the present report describes the function of a putative adhesin which binds to D-alanyl-LTA and is presented to its partner cell. Thus, the D-alanyl esters would appear to play a role in determining the binding and presentation of this adhesin.

Some cell surface proteins of S. gordonii may bind to D-alanyl LTA in a way analogous to the binding of choline-binding proteins of S. pneumoniae (54, 65). Among the family of surface-located choline-binding proteins on S. pneumoniae is CbpA, a 75-kDa adhesin and virulence determinant. The family of proteins is noncovalently bound to the phosphoryl-choline of the wall teichoic acid. By analogy, the 100-kDa putative adhesin that is lacking in the S. gordonii DL1 mutants studied here may normally bind to the D-alanine-substituted LTA. In support of a weakly bound protein is the observation that mild sonication of parent DL1 cells is sufficient to remove the 100-kDa protein and render them unable to participate in galactoside-inhibitable coaggregation with streptococci but still capable of other galactoside-noninhibitable coaggregations with actinomyces (data not shown). In this regard, the D-alanyl LTA may act as a scaffolding for presenting the bound 100-kDa protein on S. gordonii surface.

Another important class of protein ligands which bind to LTA are autolysins. These proteins are cell wall hydrolases, e.g. MurNAc L-ala amidase, which the bacterium must regulate for growth (11, 29). Gel-permeation chromatography demonstrated the binding of the amidase to LTA (30). In B. subtilis, insertional inactivation of the genes in the dlt operon results in an increased rate of autolysis (61, 62). The increased negative charge of LTA in the mutants resulting from a decrease in D-alanylation appeared to increase the amount of autolysin (s) bound. While the mechanism of adhesin binding would appear to be different from that of autolysin binding, the conclusion is made that D-alanylation of LTA provides a feature for regulating the ability of LTA to bind selected protein ligands.

Interestingly, many S. gordonii and S. sanguis strains in biovars that are positive for LTA (the Lancefield group H antigen [52]) also are positive for galactoside-sensitive adhesins detected by intrageneric coaggregation (32, 38). The group H antigen occurs in most strains of S. gordonii and S. sanguis and not at all in strains of S. oralis and S. mitis (35). These latter species include strains with GalNAc-containing cell wall polysaccharides, which are the receptors for the galactoside-sensitive adhesins, including the putative 100-kDa adhesin on S. gordonii DL1 (32, 38). Significantly, all streptococcal strains that are positive for the group H antigen are negative for GalNAc-containing cell wall polysaccharides and visa versa (7, 32). These findings implicate an association between the galactoside-sensitive 100-kDa putative adhesin and the group H antigen on the streptococcal cell surface. Moreover, the anti-DL1 polyclonal serum absorbed with spontaneous Cog mutant PK1897 and used to detect the 100-kDa putative adhesin in DL1 (see Fig. 6) also identifies a 100-kDa protein in other streptococcal strains that also possess the group H antigen (9, 63). This absorbed antiserum does not react with S. oralis strains (9). The correlation of the reactivity of both anti-adhesin and anti-LTA sera with the same strains strongly supports the association of 100-kDa putative adhesin with LTA.

Previous results with D-alanine ester-deficient mutants of L. rhamnosus showed aberrant morphology and defective cell separation (48). However, these mutants may have resulted from multiple mutations or single mutations with pleiotropic effects. The results with the S. gordonii mutants show a correlation between D-alanine ester deficiency resulting from insertional inactivation of dltA and aberrant cell morphology, slower growth rate, and defective cell separation. Inactivation of the dltD of L. lactis also results in a mutant that grows more slowly and forms longer chains than the wild-type strain (15). In contrast to these results, mutations in dltA-dltD of B. subtilis did not result in changes in cellular morphology, cell growth, basic metabolism, and formation of flagella (61, 62). The only changes correlated with D-alanine ester deficiency in this organism were an enhanced rate of autolysis and a higher susceptibility to methicillin (61). Thus, at this time it cannot be concluded that D-alanine ester deficiency will have the same effect in each gram-positive organism.

A wide range of species exhibited dltA-reactive fragments in Southern blots (data not shown). Out of 23 strains tested, all but one strain of S. sanguis and one strain of S. oralis were positive for dltA. The surprising finding was that S. pneumoniae, S. oralis, and S. mitis possess dltA but do not have D-alanine esters of LTA. In fact, S. pneumoniae has the entire dlt operon (3). The expression of the dlt operon may be silent or may occur under environmental conditions that have not been tested in the laboratory. S. oralis and S. mitis do not express detectable amounts of glycerol LTA (28, 35), but S. oralis can incorporate radiolabelled choline into membrane components, suggesting that it may produce a choline-containing macroamphiphile similar to that of S. pneumoniae (31). Alternatively, these three species may use the dlt operon for D-alanylation of a different molecule or macroamphiphile than LTA (8). One possibility has been suggested for the DltA-DltD homologs of S. mutans (accession no. AF049357): mutants in genes encoding these proteins accumulate elevated levels of intracellular polysaccharide in the presence of fructose or sucrose (55). When grown in continuous culture, S. mutans Ingbritt produced two to six times as much LTA during growth on fructose compared to glucose, depending on the generation times of the cells (33). And, the LTA content of strain Ingbritt increased four- to fivefold when the pH of the culture medium was raised from 5.0 to 7.5, irrespective of the carbon source (33). Thus, environmental factors can greatly influence the amounts of LTA in streptococcal cells. A recent entry into GenBank, accession no. AF051356, reports a potential relationship of defects in the dlt operon to acid sensitivity in S. mutans (59). Taken collectively, the properties of adherence, intracellular polysaccharide accumulation, and acid sensitivity all being affected by mutations in the dlt operon suggest that the LTA macroamphiphile may act as a supporting matrix or scaffolding for binding of a family of proteins with specific functions for sensing the streptococcal environment.