The pgdA Gene Encodes for a PeptidoglycanN-Acetylglucosamine Deacetylase in Streptococcus pneumoniae *

Analytical work on the fractionation of the glycan strands of Streptococcus pneumoniae cell wall has led to the observation that an unusually high proportion of hexosamine units (over 80% of the glucosamine and 10% of the muramic acid residues) was not N-acetylated, explaining the resistance of the peptidoglycan to the hydrolytic action of lysozyme, a muramidase that cleaves in the glycan backbone. A gene, pgdA, was identified as encoding for the peptidoglycanN-acetylglucosamine deacetylaseA with amino acid sequence similarity to fungal chitin deacetylases and rhizobial NodB chitooligosaccharide deacetylases. Pneumococci in which pgdA was inactivated by insertion duplication mutagenesis produced fully N-acetylated glycan and became hypersensitive to exogenous lysozyme in the stationary phase of growth. The pgdA gene may contribute to pneumococcal virulence by providing protection against host lysozyme, which is known to accumulate in high concentrations at infection sites.

The unusual complexity and diversity of the cell surface of Streptococcus pneumoniae is apparent both in the capsular structure and in the underlying cell wall structure of this bacterium, and this complexity and diversity may be related to the multiplicity of interactions of this microbe and its human host. S. pneumoniae is capable of producing at least 90 chemically distinct capsular polysaccharides (1). The cell wall of this microorganism contains a teichoic acid of unusually complex chemistry (2, 3) the components of which include ribitol phosphate, galactosamine, trideoxydiaminohexose, and covalently linked phosphocholine residues (4). The peptidoglycan of S. pneumoniae is also unusual because its stem peptides are cross-linked in both a direct and an indirect manner (5). Furthermore, the proportion of distinct linear and branched muropeptides in the peptidoglycan is clonally related (6). The pneumococcal cell wall is a potential target for components of the first line host defense such as lysozyme. In addition, the peptidoglycan and teichoic acid may represent bacterial ligands recognized by the innate immune system of the host.
The recent introduction of high resolution analytical techniques and genetic approaches began to shed light on the determinants and biological functions of the pneumococcal cell wall. In this study we describe the identification of a genetic determinant, pgdA, of the first bacterial peptidoglycan GlcNAc deacetylase. We show that the innate activity of this enzyme is responsible for the high proportion of non-acetylated hexosamine residues in the peptidoglycan that appears to play a role in the resistance of S. pneumoniae to the activity of exogenous lysozyme, an enzyme that is known to accumulate in high concentrations at infection sites.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Plasmids, and Growth Media-Cultures of S. pneumoniae R36A, a non-encapsulated laboratory strain from the Rockefeller University collection, were grown in a casein-based semi-synthetic medium (C ϩ Y) containing 1 mg/ml yeast extract (7) or in a chemically defined medium (Cden) (8) at 37°C without aeration. Plasmid pJDC9 (9) was used for insertion duplication mutagenesis. Escherichia coli DH5␣ was grown in Luria broth medium at 37°C with aeration. If necessary, erythromycin (Sigma) was added in the following concentrations: 1 mg/ml (E. coli) and 1 g/ml (S. pneumoniae).
Standard DNA Methods-Routine methods were used for the isolation and manipulation of DNA (10,11). Enzymes were purchased from New England Biolabs and were used as recommended by the manufacturer. Preliminary sequence data were obtained from the Institute for Genomic Research. Open reading frames were analyzed using DNAS-TAR software. Sequence comparisons were performed with the BLAST algorithm. CLUSTAL was used for multiple sequence alignments.
Sequencing of pgdA-A DNA fragment (1801 base pairs) including the pgdA gene was amplified by PCR 1 from chromosomal DNA isolated from S. pneumoniae R36A with the primers TCTACAGATACGGATGT-TGG and CTATCTTTGATTGCTTGACC using the GeneAmp PCR reagent kit (Perkin-Elmer). The following conditions were used for amplification: 94°C for 5 min, 30 cycles of 94°C for 30 s, 53°C for 30 s, 72°C for 3 min, and one final extension step of 72°C for 5 min. After purification (PCR purification kit, Promega) the DNA sequence was determined by primer walking at the Rockefeller University Protein/DNA Technology Center with the Taq fluorescent dye terminator sequencing method by using a PE/ABI 377 automated sequencer.
Inactivation of the pgdA Gene-The gene encoding the peptidoglycan GlcNAc deacetylase was inactivated by insertion duplication mutagenesis. An internal fragment of the gene was amplified from chromosomal DNA of strain R36A by PCR using the GeneAmp PCR reagent kit (Perkin-Elmer) with the following primers: 5Ј-GGTGAATTCG-GAGTCGTTAATCGTAATGTGACC-3Ј and 5Ј-GGCGGATCCCAACAA-CATGACCTTCAGATTTTATC-3Ј. With these primers, EcoRI and BamHI restriction sites were introduced. The PCR reaction was performed in a total volume of 100 l with 20 ng of template DNA, 40 pmol of primers, and 2.5 units of Taq polymerase using the following conditions: 94°C for 5 min, 30 cycles of 94°C for 30 s, 53°C for 30 s, 72°C for 3 min, and one final extension step at 72°C for 5 min. After purification (Promega PCR purification kit) the PCR product and plasmid pJDC9 were restricted with EcoRI and BamHI. After purification (Promega DNA purification kit) the internal gene fragment and the plasmid were ligated, and the resulting vector pPGDA was transformed into E. coli DH5␣ with selection for erythromycin resistance. Next, pPGDA was isolated (Promega plasmid miniprep) and used as donor DNA to trans-* Partial support for this study was provided by National Institutes of Health Grant AI37275 and by the Irene Diamond Foundation. Sequencing of S. pneumoniae was accomplished with support from The Institute for Genomic Research, the National Institute of Allergy and Infectious Diseases, and the Merck Genome Research Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AJ251472.
‡ To whom correspondence should be addressed. Tel.: 212-327-8278; Fax: 212-327-8688; E-mail: tomasz@rockvax.rockefeller.edu. form competent S. pneumoniae R36A. Competent bacteria were obtained by a published procedure (12) with addition of competencestimulating peptide (13). Transformation was performed by 30 min of incubation at 30°C followed by a phenotypic expression period of 2 h at 37°C and growth on tryptic soy agar containing 3% sheep blood and 1 g/ml erythromycin (12). One transformant was picked, and the correct insertion of the plasmid into the chromosome was verified by PCR analysis (data not shown). The mutant was able to grow in 1 g/ml erythromycin. However, because the growth rate was reduced, we included erythromycin only in the pre-cultures and not in the cultures used for the experiments in order to have growth conditions similar to that of the parental strain.
Biosynthetic Labeling of Pneumococcal Cell Walls with [ 3 H]GlcNAc-The labeling of cell walls of strain R36A was as described previously (14). The bacteria were first grown in Cden synthetic medium and then transferred into Cden medium with a reduced concentration of glucose (0.1 mg/ml) containing 1.2 Ci/ml [ 3 H]GlcNAc (Amersham Pharmacia Biotech). After three generations of growth the cells were harvested, and cell walls were isolated. The specific radioactivity was 3.4 ϫ 10 6 cpm/mg of cell wall with 70% of the label being present in cell wall teichoic acid and 30% in the peptidoglycan glycan strands (data not shown).
Isolation of Cell Wall and Peptidoglycan-Pneumococcal cell walls were prepared from cultures in exponential growth phase as described (15) with the modifications described in Ref. 16. Wall teichoic acid was removed by treatment with hydrofluoric acid (16) to obtain peptidoglycan.
N-Acetylation of Peptidoglycan-Cell walls were N-acetylated according to the method of Heymann et al. (17) with the following modifications. A suspension of 2 mg/ml cell walls in water was cooled with ice water, and subsequently 0.25 volume of saturated NaHCO 3 and 0.25 volume of freshly prepared 5% acetic anhydride were added. The mixture was stirred for 30 min at 0°C. After a second aliquot of 0.25 volume of 5% acetic anhydride was added, the mixture was stirred for 30 min at 0°C and for 1 h at 25°C. The peptidoglycan was recovered by centrifugation at 50,000 ϫ g for 30 min, washed three times with water, and resuspended in water.
Assay for Murein Hydrolase Activity-[ 3 H]GlcNAc-labeled peptidoglycan (7.5 g, 14,000 cpm), either non-modified or N-acetylated in vitro, was incubated with different murein hydrolases in a total volume of 100 l for 60 min at 37°C. The enzyme concentrations and buffers were as follows: 50 mM sodium phosphate (pH 7.0) for affinity-purified LytA (10 g/ml) (18), 25 mM sodium phosphate (pH 5.5) for chicken egg white lysozyme (Roche Molecular Biochemicals) (20 g/ml) and for mutanolysin from Streptomyces globisporus (Sigma) (40 g/ml). After addition of 100 l of 1% cetyltrimethylammonium bromide and an incubation on ice for 30 min, the sample was centrifuged (10,000 ϫ g, 20 min, 4°C), and the radioactivity in 100 l of supernatant was determined as described (19).
Fluorodinitrophenylation of Peptidoglycan and Analysis of the Amino Sugars-Free amino groups in the glycan strands of the peptidoglycan were detected by derivatization with 2,4-dinitrofluorobenzene according to a modified published method (20). Tritium-labeled peptidoglycan (see above) was incubated with 4 volumes of 10% sodium carbonate and 0.8 volume of 20% fluorodinitrobenzene (in acetone) for 3 h at 37°C. The yellowish peptidoglycan was recovered by centrifugation at 50,000 ϫ g for 30 min, washed three times with water, and resuspended in water. Concentrated hydrochloric acid was added to give a final concentration of 4 M, and after flushing with nitrogen, the tube was sealed, and the sample was hydrolyzed at 105°C for 14 h. The HCl was removed in an air flow, and the samples were dried in vacuo. To prepare a standard of the derivatized products, either [ 3 H]GlcNAc or 3 H-labeled peptidoglycan (label in GlcNAc and MurNAc) was first hydrolyzed and then derivatized as described above. The fluorodinitrophenyl derivatives of glucosamine and of muramic acid were separated by reversedphase HPLC on a 3-m ODS Hypersil column (Keystone Scientific) at 50°C at a flow rate of 0.5 ml/min for 70 min using a slightly convex gradient beginning with 100 mM sodium phosphate (pH 2.0) and ending with a mixture of 40% 10 mM sodium phosphate (pH 6.0) and 60% acetonitrile. The tritium-labeled amino sugar derivatives were detected using a flow-through scintillation counter (Packard Instrument Co.) with Ultima-Flo AP scintillator (Packard Instrument Co.) at a flow rate of 1.5 ml/min.
Isolation of Glycan Strands and Analysis of Their Lysozyme Digestion Products-Peptidoglycan (2 mg/ml) was digested with 80 g/ml affinity-purified pneumococcal amidase for 24 h at 37°C as described (16), and the glycan strands were separated from the peptides by size-exclusion chromatography using a Keystone GFS-150 column. The compounds were eluted with water at a flow rate of 0.8 ml/min. The glycan-containing fraction was concentrated in a SpeedVac and stored at 4°C. To test for deacetylated amino sugars, the glycan fraction was digested with 0.1 mg/ml lysozyme in 50 mM potassium phosphate (pH 6.3) for 90 min at 37°C. The samples were divided, and one part was N-acetylated with acetic anhydride as described above. Both aliquots were reduced with sodium borohydride and analyzed by reversed-phase HPLC on a 3-m ODS Hypersil column (Keystone) at 50°C. Elution was at a flow rate of 0.5 ml/min for 70 min by a slightly convex gradient beginning with a mixture of 97.5% 100 mM sodium phosphate (pH 2.0) and 2.5% methanol and ending with 82% 100 mM sodium phosphate (pH 2.0) and 18% methanol. The eluted compounds were detected by UV absorption at 206 nm.
Analysis of Pneumococcal Muropeptides-Peptidoglycan (2 mg/ml) was digested with 0.5 mg/ml mutanolysin (Sigma) in 25 mM sodium phosphate (pH 5.5). The resulting muropeptides were reduced with sodium borohydride and analyzed by reversed-phase HPLC according to Severin et al. 2 using similar conditions as in a published method (21).

RESULTS
Lysozyme Digestion of Pneumococcal Peptidoglycan-In the course of studies on the characterization of cell wall glycan strands, we made the unexpected observation that the native peptidoglycan of this bacterium was a poor substrate for lysozyme; only 11% of the peptidoglycan was solubilized even after extensive treatment with this enzyme (Fig. 1). Re-testing lysozyme sensitivity with chemically acetylated peptidoglycan resulted in rapid and complete hydrolysis of the glycan chains, suggesting that the pneumococcal glycan strains may not be fully acetylated (Fig. 1). The LytA amidase had higher activity against the non-modified peptidoglycan, and the M1 muramidase showed similar activities toward both substrates. The mechanisms of these effects are not understood. mococcal Peptidoglycan-To test whether GlcNAc, MurNAc, or both amino sugars of the pneumococcal peptidoglycan were non-acetylated, the free amino groups of [ 3 H]GlcNAc-labeled peptidoglycan were derivatized with 2,4-dinitrofluorobenzene. After total hydrolysis the fragments were separated by reversed-phase HPLC, and the radioactive amino sugars and their derivatives were detected with a flow-through scintillation counter. As shown in Fig. 2, lack of acetylation was mainly the property of glucosamine residues. Assuming both amino sugars were 3 H-labeled equally, it was estimated that 84% of GlcN and 10% of MurN residues were present in non-acetylated form in the pneumococcal peptidoglycan (Fig. 2).

Identification of N-Deacetylated Amino Sugars in the Pneu-
Identification of the pgdA Gene-An N-acetylglucosamine deacetylase activity in E. coli that also deacetylates N-acetylglucosamine-6-phosphate was described previously (22). More recently, the E. coli NagA was reported to be an N-acetylglucosamine-6-phosphate deacetylase involved in N-acetylglucosamine metabolism (23). Partly deacetylated hexosamine polymers or oligosaccharides such as the extracellular galactosamine polymer produced by Aspergillus parasiticus have also been described (24). For the identification of genes encoding putative peptidoglycan deacetylases we searched data bases for proteins with homology to known N-deacetylases of macromolecules with structures similar to peptidoglycan such as poly(␤-1,4)GlcNAc (chitin) or oligo(␤-1,4)GlcNAc (nodulation factors). In the unfinished nucleotide sequence of S. pneumoniae obtained from the Institute for Genomic Research we found only one open reading frame encoding a putative GlcNAc N-deacetylase. We named the gene pgdA for peptidoglycan N-acetylglucosamine deacetylase A. Fig. 3 shows the sequence alignment of the C-terminal region of the protein with two chitin deacetylases from Saccharomyces cerevisiae and with NodB from Sinorhizobium meliloti. The deduced sequence of PgdA consists of 463 amino acids. The protein has a predicted molecular weight of 52,699.74 Da. Only the C-terminal part of the protein (amino acids 252-422) shows homology to other known deacetylases. This region has 27% identity and 50% similarity to chitin deacetylase 1 of yeast (CDA1). Specific residues conserved in the known deacetylases CDA1 and CDA2 in yeast and NodB proteins in rhizobium species were also identified in the pneumococcal deacetylase (Fig. 3). The Nterminal part of PgdA (amino acids 1-251) does not have significant homology to any known protein in the data bases.
Putative Peptidoglycan Deacetylases in Other Bacteria-Non-acetylated hexosamine residues have been reported to be present in the peptidoglycans of many bacteria, especially in Bacillus species, although no genes encoding for such an enzyme were identified (25)(26)(27)(28)(29). In searching the data bases for hypothetical proteins similar in sequence to the pneumococcal PgdA protein we found other putative peptidoglycan deacetylases in bacterial species having partially deacetylated peptidoglycan: three hypothetical proteins in Bacillus subtilis (Gen-Bank TM accession numbers AAC46306, BAA23389, and CAA74511) and one NodB homolog protein in both Bacillus stearothermophilus (GenBank TM accession number B47692) and Bacillus cereus (GenBank TM accession number CAB40600). In addition, there are hypothetical proteins with similarities in sequence to PgdA in the data base of the unfinished microbial genomes (provided by the Institute for Genomic Research), including hypothetical proteins in Streptococcus pyogenes, Enterococcus faecalis, and Clostridium difficile. It remains to be verified experimentally whether these proteins deacetylate the bacterial peptidoglycan or other macromolecules like chitin.
Inactivation of the Pneumococcal pgdA Gene by Insertion Duplication Mutagenesis-A 601-base pair insert of the gene (from base 370 to 970) was amplified by PCR from chromosomal DNA of the laboratory strain R36A. The fragment was cloned into plasmid pJDC9, which carries an erythromycin resistance marker. Cloning was performed in E. coli. The resulting plasmid, pPGDA, was used to transform competent cells of S. pneumoniae R36A. Because the plasmid could not replicate in S. pneumoniae, resistance to erythromycin could only be acquired by a homologous recombination event yielding two truncated copies of the target gene on the chromosome, flanking the inserted plasmid with the erythromycin resistance In C, the radioactivities in the DNP-G and DNP-M peaks represent 42% and 5% of the total radioactivity, respectively. The signals between 8 and 11 min are the non-derivatized amino sugars, which cannot be separated by this chromatographic system. Cell Wall Structure of the pgdA Mutant-The cell walls of the mutant strain as well as those of the parental strain R36A were isolated, and the peptidoglycan parts were analyzed in two different ways. First, the peptidoglycan was digested with a muramidase (mutanolysin), yielding disaccharides substituted by the peptide side chains (muropeptides), which were separated by HPLC. As shown in Fig. 4, compounds previously shown to represent partly N-deacetylated muropeptides were not present in the peptidoglycan of the pgdA mutant (arrows in Fig. 4). The absence of N-deacetylated muropeptides resulted in an overall simplification of the muropeptide pattern of the mutant cell wall as compared with that of R36A. In the parental strain, the region of higher cross-linked muropeptides (dimers to tetramers) between 35 and 70 min of the elution profile contained a large number of non-resolvable peaks presumably representing variants of N-deacetylated muropeptides carrying the same peptide side chains. In contrast, the muropeptide profile of the pgdA mutant showed fewer peaks, which were better resolved in this region.
In a second analytical approach the purified glycan strands free of the stem peptide side chains were analyzed. Stem peptides were released from peptidoglycan by the pneumococcal amidase LytA. Parental and mutant peptidoglycans were digested by the amidase with similar rates, whereas amidase activity was considerably slower with the chemically acetylated peptidoglycan. Glycan strands were purified by size-exclusion chromatography and were treated with lysozyme. The glycan strands of the parental strain were found to be partly Ndeacetylated as indicated by the incompleteness of the lysozyme digestion (Fig. 5): in addition to the expected main products, the reduced form of GlcNAcMurNAc and (GlcNAc-MurNAc) 2 , additional peaks appeared in the elution profile between 30 and 50 min. After chemical N-acetylation in vitro these peaks became better resolved and shifted toward higher retention times, indicating that the peaks represented glycan fragments with different N-deacetylation patterns.
In contrast to the glycan strands of the parental strain, the glycan purified from the pgdA-inactivated mutant was quantitatively digested by lysozyme to the disaccharide and the tetrasaccharide, and the lysozyme products showed no change in retention times after a chemical N-acetylation. These findings allowed two conclusions. First, the data identified the mechanism of the resistance of pneumococcal peptidoglycan to lysozyme digestion (demonstrated in Fig. 1) as the poor hydrolytic activity of this enzyme against deacetylated glycan strands. Second, the results also suggest that the gene product of pgdA is the primary, if not the only, enzyme responsible for the deacetylation of the hexosamine residues of the peptidoglycan.
Lysozyme Sensitivity of the pgdA Mutant-R36A::pPGDA and the parental strain R36A were grown in semi-synthetic medium (C ϩ Y) at 37°C. Either in exponential growth phase or shortly after entering the stationary phase, the cultures were divided, and to one part 80 g/ml lysozyme was added (Fig. 6). Control cultures received no lysozyme. The parental strain R36A was not affected by the addition of lysozyme until about 2 h after the onset of stationary phase, when the cultures with lysozyme showed a slightly increased rate of lysis. In contrast, cultures of the mutant strain began to lyse rapidly upon entering the stationary phase. Neither the parental nor the mutant strain was affected by lysozyme during exponential growth. If lysozyme was added to the mutant culture at the onset of stationary phase, lysis started immediately. DISCUSSION During studies on the chemical structure of the pneumococcal peptidoglycan we noted the resistance of this macromolecule to lysozyme, suggesting the absence of N-acetyl substitu- After digestion with lysozyme the samples were divided, and one part was chemically N-acetylated (B and D). The samples were reduced with sodium borohydride and analyzed by HPLC. The bracket in A shows the region of the non-resolved, partially deacetylated glycan strands, which shifted to the retention times of fully acetylated glycan fragments (GlcNAcMurNAc) n (n ϭ 1 to 6) indicated by the asterisks (*) upon chemical acetylation (B). The N-deacetylated glycan fragments generally eluted with lower retention times than the corresponding fully acetylated glycan strands and had lower intensities because of the contribution of the carbonyl moiety of the acetyl groups to the UV signal at 206 nm. The HPLC pattern of the lysozyme-digested glycan fragments of the mutant strain did not change upon acetylation, indicating that the mutant glycans are fully acetylated (C and D).
tions of hexosamine residues. The presence of such nonacetylated hexosamines was already observed earlier (30). More recently, mass differences of Ϫ42 and Ϫ84 Da were detected among different pneumococcal muropeptides that otherwise had the same amino acid and amino sugar composition, indicating the loss of acetyl groups from one or more of the N-acetyl amino sugars 2 (21). It was speculated that this could be the result of an artificial deacetylation reaction during preparation of the peptidoglycan (21). Our observations do not confirm this speculation. Rather, our results indicate that the deacetylated hexosamines of the pneumococcal peptidoglycan are the products of an enzymatic reaction by a pneumococcal deacetylase encoded by the gene pgdA, which we describe in this study.
Two observations strongly suggest that pgdA is the structural gene for a deacetylase. Inactivation results in a virtually complete disappearance of deacetylated residues (see arrows in Fig. 4), and the relevant C-terminal domain shows 27% identity and 50% similarity to enzymes with similar catalytic activities.
The pgdA gene described here is the first bacterial determinant encoding an enzyme activity responsible for the loss of N-acetyl groups of the peptidoglycan hexosamine residues. The pgdA gene was identified through its sequence similarity with known deacetylases, the fungal chitin deacetylases and the rhizobial NodB proteins. The substrates of these enzymes are polymers or oligomers of ␤-1,4-linked GlcNAc residues, structures that resemble peptidoglycan glycan strands. Thus, the presence of conserved blocks in the sequence of the catalytic domains of these proteins is not surprising.
The pgdA gene of S. pneumoniae encodes for a putative secreted protein with an N-terminal signal peptide typical for Gram-positive bacteria (31). Structural features of PgdA imply that the substrate of this enzyme is the polymerized peptidoglycan at some stage of its assembly. There is no cleavage site for any known leader peptidase. Most likely the protein is translocated across the cytoplasmic membrane by components of the general secretory pathway and remains anchored to the cytoplasmic membrane by its N-terminal membrane domain. Thereby it would face to the outside and would be able to reach its substrate, the peptidoglycan. By this location of the enzyme, the deacetylation reaction would be a secondary modification, which is in accordance with the known pathway of peptidoglycan synthesis that leads to a fully acetylated glycan backbone (32).
The sequence similarity of PgdA with other deacetylases covers only the C-terminal half of the protein. The function of the N-terminal part showing no homology to known proteins remains to be elucidated. It may be involved in substrate recognition and/or specificity or in interactions with other proteins.
The presence of non-acetylated aminosugars in the peptidoglycan is not limited to pneumococci (25,26,28,29,33,34), and the increased activity of lysozyme after N-acetylation of partly deacetylated peptidoglycan was also reported in B. cereus (35,36). The data search suggests that, similar to pneumococci in bacilli and other bacterial species, the presence of non-acetylated amino sugars in the peptidoglycan is related to the activity of PgdA-like enzymes.
The virtually normal growth rate of the pgdA insertional mutant shows that this gene is not essential for in vitro growth of S. pneumoniae. The mutant showed no differences in morphology, and the growth rate in exponential phase and the cell density of stationary cultures were only slightly lower as compared with the parental strain. Apparently, under in vitro laboratory conditions, the reduced number of positively charged amino groups in the peptidoglycan of the mutant seems to have no drastic effect on cell wall-related processes like binding and uptake of ions and nutrition or non-covalent binding of proteins.
It is conceivable that the PgdA protein plays a role in controlling the activity of pneumococcal cell wall-hydrolyzing enzymes, which may have different activities toward substrates with different levels of deacetylation. The LytB endo-␤-Nacetylglucosaminidase is involved in daughter cell separation (37), whereas the LytC 1,4-␤-N-acetylmuramidase increases the rate of autolysis at 30°C (38). The deacetylation of the glycan strand GlcNAc residues could be a way of controlling the activities of these potentially autolytic enzymes.
The most striking effect of the inactivation of pgdA was the appearance of hypersensitivity of the mutant bacteria to exogenous lysozyme. Lysozyme is part of the first defense of host organisms against bacterial invasion, including response to pneumococcal infection (39,40). Large amounts of lysozyme accumulate both in the cerebrospinal fluid of the rabbit after inoculation with pneumococci in an animal model for bacterial meningitis (41) and in human meningeal disease (42). Our observations suggest that the pgdA gene may be part of the virulence mechanism of S. pneumoniae, providing increased resistance of the bacterium against the lysozyme of the human host.