Streptococcal Histone-Like Protein: Primary Structure of hlpA and Protein Binding to Lipoteichoic Acid and Epithelial Cells

Abstract

In addition to its role in the nucleoid, the histone-like protein (HlpA) of Streptococcus pyogenes is believed to act as a fortuitous virulence factor in delayed sequelae by binding to heparan sulfate-proteoglycans in the extracellular matrix of target organs and acting as a nidus for in situ immune complex formation. To further characterize this protein, the hlpA genes were cloned from S. pyogenes, S. gordonii, S. mutans, and S. sobrinus, using PCR amplification, and sequenced. The encoded HlpA protein of S. pyogenes has 91 amino acids, a predicted molecular mass of 9,647 Da, an isoelectric point of 9.81, and 90% to 95% sequence identity with HlpA of several oral streptococci. The consensus sequence of streptococcal HlpA has 69% identity with the consensus sequence of the histone-like HB protein of Bacillus species. Oral viridans group streptococci, growing in chemically defined medium at pH 6.8, released HlpA into the milieu during stationary phase as a result of limited cell lysis. HlpA was not released by these bacteria when grown at pH 6.0 or below. S. pyogenes did not release HlpA during growth in vitro; however, analyses of sera from 155 pharyngitis patients revealed a strong correlation (P < 0.0017) between the production of antibodies to HlpA and antibodies to streptolysin O, indicating that the histone-like protein is released by group A streptococci growing in vivo. Extracellular HlpA formed soluble complexes with lipoteichoic acid in vitro and bound readily to heparan sulfate on HEp-2 cell surfaces. These results support a potential role for HlpA in the pathogenesis of streptococcus-induced tissue inflammation.

Prokaryotes contain several small, basic, heat-stable proteins in association with the nucleoid. These proteins bind to single- and double-stranded DNA without obvious sequence specificity and are termed histone-like proteins; however, they do not have sequence homology with eukaryotic histones (for reviews, see references 13, 19, 33, and 37). The best-studied histone-like proteins are HU of Escherichia coli (4, 15, 29, 35, 38) and HB of Bacillus species (10, 23, 24, 31, 44). HU is a heterodimer of HU1 and HU2 proteins, which contain 90 amino acid residues each and have 70% sequence identity. HB is a protein highly homologous to HU but existing as a homodimer of a 92-amino-acid subunit (10, 23, 24, 31). Although the biological functions of histone-like proteins are not fully understood, they are known to wrap DNA and restrain negative supercoiling (4, 35). The resulting alterations in DNA structure and topology affect several cellular processes, including initiation of DNA replication (11, 51), DNA partitioning and cell division (12, 50), binding of repressors (3, 17, 30, 34), and transposition of bacteriophage Mu (43).

In addition to the physiological functions of bacterial histone-like proteins, HlpA (previously called GAG-BP and HBP) of Streptococcus species may contribute fortuitously to the virulence of these bacteria when the protein is released into the tissues during infection. Purified HlpA binds selectively in vitro to heparan sulfate in proteoglycans of heart and kidney basement membranes (1, 5, 6, 49). The accumulation of intravenously administered HlpA on renal basement membranes of mice and rabbits and the ensuing in situ immune complex formation (7, 20) indicate that it might be an important virulence factor in acute poststreptococcal glomerulonephritis and the glomerulonephritis that is often associated with streptococcal endocarditis in humans (21, 47). Tissue-bound HlpA may serve as a nidus for in situ immune complex formation leading to the inflammation and immunopathology that typify these diseases. The HlpAs of Streptococcus pyogenes, S. mutans, S. gordonii, and S. mitis are immunologically cross-reactive and exhibit identical binding activities for basement membranes in animal tissues (5, 6, 49).

This study was undertaken to clone and sequence hlpA from group A and viridans group streptococci, to compare the primary structure of HlpAs, and to evaluate the ability of these bacteria to release HlpA protein into the culture medium during growth. The hlpA genes of four Streptococcus species encode proteins of 91 amino acids that have at least 90% sequence identities. Members of the viridans group streptococci released more HlpA during stationary phase of growth than did the group A streptococci, and extracellular HlpA was complexed with soluble lipoteichoic acid (LTA). These antigen complexes bind to the surfaces of human epithelial cells in vitro and can lead to immune complex formation in situ.

MATERIALS AND METHODS

Bacteria and growth conditions.

S. gordonii G9B, S. mutans MT703, and S. sobrinus B13 were grown in chemically defined broth medium (CDM) (45). S. pyogenes M1 strain SF370, M6 strain D471 (obtained from V. Fischetti, Rockefeller University, New York, N.Y.), M12 (ATCC 11434), M24 (ATCC 10782), and M49 strain F301 (J. Zabriski, Rockefeller University) were grown in CDM supplemented with ultrafiltered yeast extract (48). All cultures were grown at 37°C, and where indicated, the pH of the medium was maintained within the designated range by periodic addition of NaOH.

Autolytic activity of streptococci was determined in some experiments by using a radioactivity release assay. Streptococci were grown for 16 h in CDM containing 2 μCi of [2,8-³H]adenine (ICN Biomedicals, Inc., Costa Mesa, Calif.) per ml. At maximum growth, 1,000-fold-excess unlabeled adenine was added to prevent further uptake of radiolabel. At later intervals, 100-μl aliquots of culture were removed and the bacteria were harvested by filtration (0.22-μm pore size; Millipore Corp.). The bacteria were washed with 5 ml of phosphate-buffered saline (PBS) at pH 7.2, and the incorporation of radioisotope was quantitated by scintillation spectrometry.

HlpA extraction and purification.

S. pyogenes D471 cells were harvested at the end of exponential growth (optical density at 600 nm of 0.85 to 0.90), washed twice with PBS, and suspended in 4 volumes of PBS at 0 to 4°C. The suspension was adjusted to pH 11.5 with NaOH, stirred for 18 h at 4°C, and then centrifuged at 14,000 × g for 30 min to remove whole cells and cell wall fragments. Supernatant fluids were collected, sterilized by filtration (0.22-μm-pore-size filter), adjusted to pH 7.0, dialyzed (3,500-molecular-weight exclusion) against water, and lyophilized. HlpA was purified to homogeneity by affinity chromatography on a column of heparin-agarose as previously described (5, 49). HlpA was also isolated from spent culture medium after the bacteria were removed by centrifugation and filtration. The spent culture medium was dialyzed against water and lyophilized.

ELISA.

An enzyme-linked immunosorbent assay (ELISA) was used to quantitate HlpA. Lyophilized streptococcal components, isolated from spent culture medium, were dissolved (100 μg of protein/ml) in PBS and used to coat the wells of 96-well vinyl assay plates (Costar, Cambridge, Mass.) at 4°C overnight. After being washed with 0.05% Tween 20 in PBS to remove unbound components, the wells were incubated with dilutions of murine monoclonal antibody (MAb) 3C4 to HlpA of S. pyogenes followed by goat anti-mouse polyvalent immunoglobulins conjugated to alkaline phosphatase (Sigma Chemical Co., St. Louis, Mo.). The wells were washed three times with PBS-Tween 20 and incubated with p-nitrophenylphosphate in 9.7% (vol/vol) diethylamine buffer at pH 9.8 for 15 min. The reaction was stopped with 3-N NaOH, and the A₄₀₅ in each well was determined with a microplate spectrophotometer (model EL310; Bio-Tek Instruments). Wells without spent culture medium components and wells containing material from uninoculated culture medium were used as reagent controls. Purified HlpA served as the quantitative standard.

Antibodies to HlpA in human sera were quantitated by ELISA. Purified HlpA (100 ng in 0.1 ml of PBS) was incubated in microtiter wells at 4°C for 18 h. The HlpA-coated wells were blocked for 1 h at room temperature with PBS–0.3% Triton X-100 (assay buffer) and washed three times with assay buffer. Human serum, diluted in assay buffer, was added to the wells and incubated for 1 h at room temperature. After the wells were washed five times with assay buffer, they were incubated with alkaline phosphatase-conjugated goat antibodies to human immunoglobulin G (IgG), IgA, and IgM (Sigma), and color was developed as described above. Antibodies to streptolysin O were quantified by using a hemolysin-neutralizing microassay (27).

Gel filtration chromatography.

The molecular weight of HlpA from spent culture medium was determined, under nondissociating conditions, by using a column (1.5 by 81 cm) of Bio-Gel P-200 (Bio-Rad, Richmond, Calif.) equilibrated with PBS. The flow rate was 4.8 ml/hr, and the fraction size was 4.8 ml. Molecular mass standards included blue dextran (2,000 kDa), aldolase (158 kDa), ovalbumin (45 kDa), chymotrypsin (25 kDa), and bromphenyl blue (670 Da).

Mobility shift assay.

The formation of LTA-HlpA complexes was determined by measuring shifts in the electrophoretic mobility of HlpA during polyacrylamide gel electrophoresis (PAGE) under nondenaturing conditions. Mixtures of purified LTA and HlpA, in 50 mM Na₂HPO₄-NaH₂PO₄ buffer at pH 7.2, were incubated at 22°C for 15 min and loaded onto a 1-mm-thick, 5% polyacrylamide gel. The separating and running buffer was also 50 mM phosphate buffer, pH 7.2. Electrophoresis was run for 1.5 h at 18 mA per gel, and HplA was subsequently stained with silver nitrate (32).

Cell culture and IIF assay.

Indirect immunofluorescence (IIF) assays were conducted on HlpA-treated monolayers of HEp-2 cells (ATCC CCL-23), a human epithelioid cell line. The HEp-2 cells were grown in Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum (GIBCO, Grand Island, N.Y.), penicillin (100 U/ml), streptomycin (100 μg/ml), and amphotericin B (0.25 μg/ml). Streptococcal components from spent culture medium or purified HlpA were added to a confluent monolayer in a chambered microscope slide (Nunc, Inc., Naperville, Ill.), and the slides were incubated for 30 min at 37°C. After unbound reactants were removed by washing in PBS, the monolayer was incubated with MAb 3C4 for 30 min, washed in PBS for 30 min, and incubated for 30 min with fluorescein isothiocyanate (FITC)-labeled rabbit antibodies to mouse IgA, IgG, and IgM (Sigma). The slides were viewed with a Leitz fluorescence microscope (Orthoplan 2; Leitz, Wetzlar, Germany).

Cloning strategy.

To amplify the hlpA gene from S. pyogenes D471, two degenerate PCR primers were designed based on the N-terminal amino acid sequence of HlpA protein (49) and on the highly conserved C-terminal consensus region of the histone-like gene sequences of Bacillus species (31) (Table 1). Both degenerate primers were biased to a S. pyogenes codon usage table generated from S. pyogenes sequences in the GenBank database.

TABLE 1.

Oligonucleotides used as PCR primers in the cloning of hlpA or as probes in Southern blot assays

Primer	Strand	Sequencea
hlp-forward	+	−17 TTTGGAGGATTTGTTAAATGGCTAA  8
hlp-reverse	−	262 ATATATGCAATTATTTAACAGCGTC  286
hlp inverse 1	−	110 CCGATCAATTGTACTTTTCACCTTCAGCA 140
hlp inverse 2	+	198 GGTCGTAACCAAACTGGTGCAGAAATT216
hlp NT-degenerate	+	40 GCWACWGARYTIACWAARAARGAYTCIGCWGC 71
hlp CT-degenerate	−	246 ACWGCRTCTTTWARWGCTTTACC 269
hlp Sg	+	96 TGTAACTGAATACCTTTCAAAA 117
hlp Sm	+	96 AGTATCATCTTACCTTGCAAAG 117
hlp Sp	+	96 AATCGAAGCTTTCCTTGCTGAA 117
hlp Ss	+	96 AATTGAAGGATTCCTTTCAAAA 117

^aNumbers coincide with nucleotide positions of the hlpA gene relative to the adenine in the ATG initiation codon; negative numbers denote positions 5′ of this codon. The mixed base site code for the degenerate oligonucleotides is as follows: R = A or G; W = A or T; Y = C or T; and I = inosine. 

One microgram of chromosomal DNA from S. pyogenes D471 was used as the template for PCR amplification (30 cycles of denaturation at 94°C for 1 min, annealing at 37°C for 2 min, and extension at 72°C for 3 min) of hlpA genes. The reactions were examined by agarose gel electrophoresis for the presence of an amplified DNA fragment of approximately 300 bp (predicted size of the hlpA gene). The product was then cloned directly from the PCR mixture into the T overhang vector pT7Blue (Novagen) and grown in E. coli DH5αF′. The nucleotide sequences of the cloned inserts were determined by double-stranded DNA sequencing. To confirm that hlpA genes had been cloned, the predicted translation products of the clones were compared to the amino acid sequences of other histone-like proteins.

The regions upstream and downstream of the hlpA gene were cloned by inverse PCR. Five micrograms of DNA from D471 was digested to completion with ApoI (New England Biolabs), which produced 0.5- to 3-kb fragments but did not cut within hlpA. After the fragments were religated to form circular molecules, they were precipitated with ethanol and used as templates for PCR amplification (30 cycles of denaturation at 94°C for 1 min, annealing at 60°C for 2 min, and extension at 72°C for 3 min) with primers generated to internal regions of the hlpA structural gene (Table 1). The amplification products of these reactions were cloned into pT7Blue.

Two additional primers (Hlp-forward and Hlp-reverse) were generated specifically to the N-terminal and C-terminal sequences of the hlpA gene of S. pyogenes D471 (Table 1), incorporating the first three codons of the 5′ end of the coding sequence and the last four codons of the 3′ end of the coding sequence, respectively. These primers were used to amplify and clone the hlpA genes of S. sobrinus B13, S. gordonii G9B, and S. mutans MT703, using amplification conditions as described above for S. pyogenes.

Radiolabeled oligonucleotides.

Oligonucleotides corresponding to bases 96 through 117 of the hlpA sequences of S. pyogenes, S. gordonii, S. mutans, and S. sobrinus were synthesized and purified by Integrated DNA Technologies (Coralville, Iowa). Each oligonucleotide was radiolabeled for 30 min at 37°C in a 20-μl reaction mixture containing 20 pmol of oligonucleotide, 62.5 μCi of [γ-³²P]ATP (3,000 Ci/mmol; Dupont-NEN), 10 U of T4 polynucleotide kinase (GIBCO BRL, Grand Island, N.Y.), 0.1 mM spermidine, and 1× Kinase Forward reaction buffer (GIBCO BRL). The kinase reaction was stopped by addition of 25 μl of 10 mM Tris-HCl–1 mM EDTA (pH 7.5) (TE buffer) and 5 μl of 200 mM EDTA (pH 7.0) followed by extraction with phenol-chloroform (1:1) and ethanol precipitation. After being dried, the radiolabeled oligonucleotides were resuspended in 50 μl of TE buffer.

Dot blot hybridization.

Genomic DNA was isolated and purified from S. gordonii, S. mutans, S. pyogenes, and S. sobrinus by the procedure of Sun et al. (41). Fifteen micrograms of the purified DNA was spotted onto nitrocellulose paper (MSI Separations Inc., Westboro, Mass.), washed twice with TE buffer, and air dried. The DNA on the paper was denatured with 0.5 M NaOH–1.5 M NaCl for 2 min, neutralized with 0.5 M HCl–1.5 M NaCl (pH 7.4) buffer for 5 min, and washed with 300 mM NaCl–20 mM NaH₂PO₄–2 mM EDTA (2× SSPE) buffer. The blot was air dried and fixed for 120 min at 80°C in vacuo (36). The nitrocellulose paper was cut into four segments, each containing genomic DNA from all four Streptococcus species, and hybridized for 3 h at 48°C with radiolabeled oligonucleotides (36). The prehybridization solution was 6× SSPE–0.25% nonfat milk–5 mM Na₄P₂O₇. After hybridization, the blots were washed sequentially at 48°C with 2× SSPE–0.1% sodium dodecyl sulfate–5 mM Na₄P₂O₇ and 0.5× SSPE–0.1% sodium dodecyl sulfate–5 mM Na₄P₂O₇. The blots were exposed for 18 h and developed with the Bio-Rad 505 Molecular Imager system.

Sequencing and analysis.

All DNA sequencing was done by using double-stranded plasmid DNA and a Sequenase II kit as recommended by the manufacturer (United States Biochemical, Cleveland, Ohio). Analysis of the sequences was done with the University of Wisconsin Genetics Computer Group software package (9). Sequence similarity searches were done by using the on-line BLAST server at the National Institutes of Health.

LTA.

LTA concentration was measured by a passive hemagglutination assay as previously described (25). One unit of cell sensitizing activity was defined as the minimum amount of LTA required to sensitize a standard suspension of sheep erythrocytes and yield a positive agglutination with a standard antiserum to LTA of S. mutans. One hemagglutinating unit was equal to 100 ng of LTA. Purified LTA of S. pyogenes and S. mutans was purchased from Sigma. Antibodies to LTA were purified from rabbit antiserum by affinity chromatography on a column of deacylated cardiolipin conjugated to divinyl sulfone-agarose as previously described (25).

Statistical analysis.

The relationship between the antibody titers to streptolysin O and HlpA (ASO and AHA titers, respectively) were analyzed with the Spearman rank correlation coefficient and t test and with the Wilcoxon rank sum test.

Nucleotide sequence accession number.

The nucleotide sequences of the hlpA genes have been deposited in GenBank with accession numbers L38946 (S. pyogenes D471), L40356 (S. gordonii G9B), L38959 (S. sobrinus B13), and L40355 (S. mutans MT703).

RESULTS

Cloning and analysis of the hlpA gene.

The HlpA of S. pyogenes has been purified, and the amino acid sequence in the amino-terminal one-third of the protein has been determined (49). Comparison of this sequence with those reported for other bacteria showed extensive homology with the HB histone-like protein of Bacillus species. This information was used in the synthesis of two oligonucleotides, Hlp NT-degenerate and Hlp CT-degenerate (Table 1), which were used to amplify the coding region of the hlpA gene of S. pyogenes D471 by PCR. An amplified DNA fragment of 230 bp was obtained and cloned directly into pT7Blue. The insert was verified by sequencing both strands and comparing the translated sequence with that of the N-terminal sequence of HlpA and with the sequences of other bacterial histone proteins. The promoter and termination regions of the hlpA gene were cloned by inverse PCR of ApoI-digested and religated chromosomal DNA from S. pyogenes D471, using primers internal to the hlpA structural gene (Table 1). The amplification product was cloned and verified as described above. Figure 1 shows that there is a putative ribosome-binding site (GGAGGA) and a potential −10 RNA polymerase-binding site (TAATTA) upstream of the ATG start codon. Complete nucleotide sequencing showed that identical hlpA genes were cloned from three independent chromosomal preparations of S. pyogenes D471. In addition, the nucleotide sequence of hlpA in strain D471 is 100% identical to the single-copy hlpA gene in S. pyogenes SF370 determined by genome sequencing by Ferretti et al. (16).

FIG. 1.

DNA sequence of the hlpA gene of S. pyogenes and the deduced amino acid sequence. The hlpA gene is 276 bp in length and encodes a protein of 91 amino acid residues with a predicted pI of 9.81. The numbers on the left coincide with nucleotide positions relative to the adenine in the ATG initiation codon of hlpA; negative numbers denote positions 5′ of the adenine. Amino acid residues for the HlpA (numbered to the right) are given below the DNA sequence in the single-letter designation. The positions of the putative Shine-Dalgarno (SD) ribosome-binding site and the −10 promoter sequence are underlined. *, stop codon.

To obtain the sequence of the hlpA gene from the three other species of streptococci, a pair of PCR primers was generated to the N terminus and C terminus of the S. pyogenes sequence (Table 1). The primers overlapped the first three codons and the last four codons, respectively. The amplified DNA fragments from S. gordonii, S. mutans, and S. sobrinus were cloned, and the sequences were compared with that of hlpA gene of S. pyogenes. All gene sequences were cloned and verified from at least two independent amplification reactions, and multiple clones from each reaction were sequenced in both directions. Although several silent changes were detected in the primary sequences of the hlpA genes from different streptococci (data not shown), the deduced amino acid sequences are very similar (Fig. 2). Two regions show heterogeneity; region 1 (amino acids 31 to 39) has numerous variations in amino acid charge and polarity, whereas region 2 (amino acids 69 to 74) has more limited substitutions. As expected, HlpA has a very high degree of similarity with the other bacterial histones, particularly that of Bacillus species (69% identity). Comparison of the Streptococcus and Bacillus consensus sequences (Fig. 2) shows that the C-terminal half of each protein is highly conserved whereas the N-terminal half is more divergent.

FIG. 2.

Comparison of the amino acid sequences of HlpA proteins of the genus Streptococcus with the HB proteins of Bacillus species. Amino acids are shown for each protein only when they differ from the consensus residue. The boxed sequences represent the conserved α-helices and the potential DNA-binding arm of HB protein (43).

To confirm that the PCR-amplified hlpA segments were derived from the indicated Streptococcus species, 22-mer oligonucleotides were synthesized by using the base sequences most unique to each gene; the hypervariable region of hlpA was comprised of bases 96 to 117 (Table 1). Each radiolabeled oligonucleotide was hybridized with dot blots of chromosomal DNA from each Streptococcus species (Fig. 3). In each case, the probe annealed with the DNA of the homologous species but weakly or not at all with DNA from the heterologous species.

FIG. 3.

Southern dot blot showing hybridization of radiolabeled, synthetic oligonucleotides with genomic DNA from S. gordonii (G), S. mutans (M), S. pyogenes (P), and S. sobrinus (S). The sequences of the Hlp oligonucleotides are shown in Table 1 and are designated by the initials of the genus and species from which they were derived.

Release of HlpA by streptococci.

We showed previously, using a mouse model, that intravenously administered HlpA is carried by the blood to the kidneys, where it binds to heparan sulfate-proteoglycans of the glomerular capillaries and acts as a nidus for the formation of in situ immune complexes (6, 7). Although immune complex deposition can lead to glomerulonephritis, the tendency of streptococci to release HlpA during localized infections has not been documented. Therefore, serum specimens were obtained from 155 pharyngitis patients at the Buffalo Children’s Hospital and assayed for antibodies to streptolysin O and to HlpA. Seventy-seven of the patients had significant ASO titers (≥200), indicating that they had experienced an infection by S. pyogenes within the previous 3 months (26, 27). Ninety-seven patients had AHA titers of between 1,600 and 25,600 (Fig. 4). There was a strong positive correlation between the history of streptococcal infection, as indicated by the ASO titer, and the AHA titers (P < 0.0001 [Spearman rank correlation coefficient and t test] or P = 0.0017 [Wilcoxon rank sum test]). This result indicates that immunogenic quantities of HlpA are released with streptolysin O by streptococci growing in vivo. None of the patients developed symptoms of delayed sequelae.

FIG. 4.

AHA and ASO titers in human sera. The endpoint data are clustered in relation to the serial dilutions (twofold) of the sera, beginning at 1:100. The average AHA titers are indicated by +.

Analyses of spent culture media of several Streptococcus species indicated that HlpA was released only during the stationary growth phase, as exemplified by S. mutans MT703 (Fig. 5). Maximal yields were 320 ng/ml for S. mutans, 294 ng/ml for S. gordonii CH1, and less than 20 ng/ml for S. pyogenes D471, 11434, and F301. The period of HlpA release by S. mutans, in culture medium maintained at pH 6.8, coincided with a 17% decrease in culture turbidity (Fig. 5) and with a release of 28% of previously assimilated [³H]adenine (data not shown), indicating that limited autolysis had taken place. S. mutans growing at pH 6.0 and below did not release either HlpA or radiolabeled nucleic acids, although the bacteria grew at the same rate and to similar turbidity values as streptococci grown at the higher pH (Fig. 5). S. pyogenes cultures showed neither turbidity decreases nor leakage of cytoplasmic radiolabel during stationary growth phase in any of these growth conditions (data not shown), which is consistent with the relatively low amount of HlpA in the spent culture medium.

FIG. 5.

Release of HlpA by S. mutans during growth in CDM. Symbols: ○, turbidity of a culture at pH 6.8; •, turbidity of a culture at pH 6; □, extracellular HlpA at pH 6.8 as determined by enzyme immunoassay; ▪, extracellular HlpA at pH 6.

Gel filtration chromatography of S. mutans components, from spent culture medium, indicated that HlpA was complexed with other cell constituents because it eluted predominately at the void volume of the column (Fig. 6), whereas chromatography of purified HlpA resulted in a single peak (indicated by bar) with an apparent molecular mass of 31 kDa (6). Addition of DNase to the crude preparation prior to chromatography did not change the elution profile of HlpA (data not shown). LTA, a streptococcal surface polyanion and virulence factor (8, 22), was also detected in the spent culture medium, and its elution with HlpA at the void volume during gel filtration chromatography (Fig. 6) indicated that it might complex ionically with the HplA after their release from the bacteria.

FIG. 6.

Gel filtration chromatography of S. mutans cell components in spent medium. Symbols: ○, absorbance at 280 nm; •, absorbance at 405 nm in ELISA for HlpA; □, LTA concentration determined by passive hemagglutination assay. The elution point of purified HlpA is indicated by the bar.

To determine if HlpA has binding affinity for LTA, purified protein from S. mutans was incubated with LTA (S. mutans), and the mixture was resolved by a PAGE mobility shift assay. Purified LTA caused increased anionic mobility of HlpA during electrophoresis in nondissociating conditions (Fig. 7). Similar results were obtained with reagents from S. pyogenes. Mixing of S. mutans LTA and HlpA at a molar ratio of 5:1 prior to gel filtration chromatography caused HlpA to elute at the void volume instead of its normal location (data not shown). Thus, the constituents of these naturally occurring complexes may exhibit altered tissue-binding properties for host cell surfaces.

FIG. 7.

Mobility shift assay of HlpA-LTA complexes in nondenaturing PAGE. Lane 1, 2 μg of HlpA only. For lanes 2 to 8, 2 μg of HlpA was mixed with 63 ng, 125 ng, 250 ng, 500 ng, 1 μg, 2 μg, and 4 μg of LTA, respectively. Lane 9 contained only 4 μg of LTA. The LTA and HlpA were obtained from S. mutans.

Binding to epithelial cells.

To evaluate the tissue binding activity of the HlpA-LTA complexes released by streptococci, naturally occurring complexes were obtained from spent culture medium of S. mutans by gel filtration column chromatography and incubated with growing monolayers of HEp-2 cells (Fig. 8). IIF assay using MAb 3C4 showed that HlpA bound to the cells in a granular pattern (Fig. 8A), which was identical to that observed with purified HlpA (Fig. 8B). Staining of the treated HEp-2 cells with affinity-purified antibodies to streptococcal LTA (22) produced more uniform and linear patterns (Fig. 8D) which are consistent with the binding pattern of purified LTA (Fig. 8E), also reported by others (8, 40). The binding of HlpA, but not LTA, to HEp-2 cells was inhibited by preincubation of the bacterial extract with 10 μl of heparin per ml, further indicating that these bacterial antigens were binding to different receptors on the epithelial cell surfaces. Similar binding properties and IIF staining patterns were observed with LTA and HlpA of S. pyogenes.

FIG. 8.

IIF assay showing HlpA and LTA binding to HEp-2 cell monolayers. (A) Cells treated with HlpA-LTA complexes (50 μg of protein) isolated from spent culture medium of S. mutans and stained for HlpA with MAb 3C4 and FITC-conjugated rabbit antibodies to mouse immunoglobulins; (B) cells treated with 20 μg of purified HlpA and stained as described for panel A; (C) cells treated with PBS (control) and stained as described for panel A; (D) cells treated with HlpA-LTA complexes isolated from spent culture medium and stained with affinity-purified rabbit antibodies to LTA and FITC-conjugated goat antibodies to rabbit IgG; (E) cells treated with 50 μg of pure LTA and stained as described for panel D; (F) control cells treated with only the primary and secondary antibodies. Magnification, ×328.

DISCUSSION

We have cloned, sequenced, and characterized the hlpA genes of S. pyogenes, S. gordonii, S. mutans, and S. sobrinus. Analysis of the nucleotide sequence predicts that HlpA is comprised of 91 amino acid residues with a molecular mass of 9,647 Da. Most of the variations (22 of 29) in the deduced amino acid sequences occur between residues 31 and 38, which corresponds to the second α-helical domain of the HB histone-like protein of Bacillus species (44). A few amino acid substitutions were also found within the antiparallel β-ribbon comprising the DNA-binding arm (residues 53 to 78), which showed 92 to 100% identity with this region of the Streptococcus consensus sequence. Specifically, S. pyogenes has a glutamic acid-for-lysine substitution at position 71; S. mutans has glutamic acid-for-alanine and lysine-for-alanine substitutions at positions 68 and 73, respectively; and S. gordonii has lysine-for-alanine and threonine-for-isoleucine substitutions at positions 68 and 70, respectively. Based on the crystallographic data for HB (44), these latter amino acid substitutions fall within the return strand of the arm (residues 65 to 78), whereas the amino acid sequences in the outgoing strands (residues 53 to 64) of the four streptococcal HlpAs are identical. Significantly, the outgoing strand contains all four arginines of the protein and is believed to constitute the DNA-binding site of the histone (31, 44). The high percentage of sequence identity between HlpA and HB proteins and gel filtration chromatography (6) also suggests that HlpA forms a homodimeric structure similar to that of HB (44). The amino acids responsible for the hydrophobic core of the dimer (Phe 30, Phe 48, Phe 51, Phe 80, Leu 7, Ilu/Val 33, Leu 37, and Leu 45) are highly conserved among both Streptococcus and Bacillus species. Consistent with histone-like proteins of other bacteria, HlpAs of streptococci are rich in lysine and arginine and devoid of cysteine and tryptophan. The HlpAs of S. mutans and S. gordonii each contain one tyrosine residue, whereas S. pyogenes and S. sobrinus contain none. The high degree of sequence homology among HlpAs is also consistent with their immunological cross-reactivity and similar binding affinities for heparin and heparan sulfate-proteoglycans of animal tissues (6, 49).

In addition to its essential role in the bacterial nucleoid, HlpA may be an important virulence factor in the streptococcal sequelae, acute poststreptococcal glomerulonephritis and rheumatic fever, when it is released by S. pyogenes during infection of the skin or nasopharynx. Previous studies have shown that purified HlpA in the bloodstream readily adsorbs to heparan sulfate-proteoglycans in the basement membranes of animal tissues (1, 5, 6, 49), where it can complex with circulating antibodies (7, 20). In the present study, the coincident release of [³H]adenine-labeled components and a decrease in culture turbidity indicated that the release of cytoplasmic HlpA to the culture medium by viridans group streptococci during stationary growth phase was most likely the result of limited bacterial lysis. HlpA release was maximal when the pH was maintained between 6.5 and 7 but minimal at more acidic pHs, which is consistent with the pH optima of autolysins of S. pneumoniae (14), S. mitis (42), and Enterococcus faecalis (39). Release of HlpA by these streptococci and by viridans group streptococci during infective endocarditis may lead to glomerulonephritis. S. pyogenes did not release significant amounts of HlpA in vitro and is not known to autolyse; however, the amount of HlpA released into tissues by group A streptococci during infection may be enhanced by the interaction of the bacteria with host defense mechanisms (e.g., defensins) that perturb the integrity of bacterial membranes (18). Indeed, the presence of anti-HlpA antibodies in human sera and their close correlation with antibodies to streptolysin O of group A streptococci can be interpreted to indicate that significant amounts of HlpA are released by these streptococci at infection sites. Alternatively, the immune response may result from the processing of cell-associated antigens, after phagocytosis of streptococci, by macrophages. The close correlation of antibody production to HlpA and streptolysin O makes it unlikely that the immune stimulus was HlpA from viridans group streptococci because they lack this hemolysin.

In mice, specific antibodies cross-link and stabilize the HlpA deposits in glomerular capillaries of kidneys (7), whereas in the absence of antibodies, HlpA is removed from the blood and excreted in urine without inducing adverse effects. Extracellular HlpA also binds to heparan sulfate in the extracellular matrix of human epithelial (HEp-2) cells and might contribute to the injury of mucus membranes in some infections. Although the IIF assay showed that HlpA and LTA had different patterns of binding on HEp-2 cells, it is not clear whether HlpA from spent culture medium was bound as free protein or as a complex with streptococcal LTA. Complexing of HlpA with the polyanionic LTA was not expected to preclude its binding to tissue components because previous studies have shown that HlpA has a 100-fold-higher affinity for heparan sulfate than it has for LTA (6). The different IIF staining patterns seen in Fig. 8 probably reflect the selective binding of HlpA to heparan sulfate-proteoglycans (1, 5, 49) and LTA to lipophilic receptors and to membrane lipids (8, 25). Also, the preparation of HlpA-LTA complexes may have contained free LTA; the molar ratio of LTA to HlpA was more than 30:1. Upon entering the blood, HlpA-LTA complexes, as well as free HlpA and LTA, can be carried to the kidneys, where they may bind directly to the glomerular capillary walls. Glomerulonephritis may arise through the activation of complement by in situ immune complexes and from localized production of interleukin-1β, interleukin-6, and tumor necrosis factor alpha by LTA-stimulated monocytes (2, 46).

The sequence data presented here are important to understanding the pathogenic properties of HlpA during streptococcal infections. Epitope mapping experiments may reveal protein domains that elicit protective antibodies, which will inhibit HlpA binding to heparan sulfate-proteoglycans in tissues, rather than antibodies that stabilize tissue deposits and exacerbate immunopathology.