Influence of Wall Teichoic Acid on Lysozyme Resistance in Staphylococcus aureus

Abstract

Staphylococcus aureus peptidoglycan (PG) is completely resistant to the hydrolytic activity of lysozyme. Here we show that modifications in PG by O acetylation, wall teichoic acid, and a high degree of cross-linking contribute to this resistance.

The human defense system uses a variety of factors to destroy bacteria that include reactive oxygen substances (7), bacteriolytic enzymes (lysozyme [13] and phospholipase A2 [17]), the complement system, and antimicrobial peptides (9). One important and widespread defense enzyme is lysozyme (5, 10), a component of both phagocytic and secretory granules of neutrophils, monocytes, macrophages, and epithelial cells (9, 13). Pathogenic bacteria, such as Staphylococcus aureus, express a wide variety of virulence factors that enable the organism to cause acute and chronic infections (1, 15). Host-colonizing bacteria must have developed mechanisms to overcome the adaptive and innate immune system. As recently shown, S. aureus is completely resistant to lysozyme, and the primary mechanism for this resistance is the modification of its peptidoglycan (PG) by O acetylation at the C-6 position of the N-acetyl muramic acid (NAM) (3). However, in S. aureus the C-6 position of NAM also carries phosphoester-linked wall teichoic acid (WTA) (19, 21) (Fig. 1), and the question is whether WTA also contributes to lysozyme resistance. Our results indicate that WTA and the degree of PG cross-linking indeed contribute to resistance against the hydrolytic activity of lysozyme.

FIG. 1.

Possible modifications at the C-6 position of the NAM in S. aureus PG. The C-6 position of NAM can be unmodified or O acetylated by OatA (every second residue [3]) or carries phosphoester-linked WTA in every ninth residue (14). WTA is composed of three glycerol phosphate and 40 ribitol phosphate units that are modified with d-alanine ester and N-acetylglucosaminyl residues. Lysozyme cleaves the β-1,4-glycosidic bond between NAM and N-acetylglucosamine (NAG).

Characterization of the double mutant SAΔoatA/tagO.

In order to analyze the role of WTA of S. aureus SA113 in lysozyme resistance, we tested a tagO deletion mutant (SAΔtagO::erm) that is completely devoid of WTA (21). Surprisingly, this mutant turned out to be as resistant to lysozyme as the wild type (WT) (Fig. 2). This was astonishing insofar as WTA is a much more bulky residue than the O-acetyl group and should therefore hinder the interaction of lysozyme with PG more efficiently. Therefore, we created a double-deletion mutant (SAΔoatA/tagO) that lacked both WTA and O acetylation. Gene replacement of oatA in the mutant SAΔtagO::erm was performed as described before (6). As a control, the oatA gene in the double mutant SAΔoatA/tagO was complemented with the plasmid pRBoatA (3); the complemented mutant restored the lysozyme-resistant phenotype. In liquid medium (B medium or tryptic soy broth), growth of the wild type and that of the tagO mutant were not inhibited at lysozyme concentrations of >50 mg per ml, growth of the oatA mutant was inhibited by 1 mg per ml, and the double mutant SAΔoatA/tagO showed the highest lysozyme sensitivity (<0.05 mg per ml); the MIC of this mutant was 20-fold lower than that of the mutant SAΔtagO. The increase in the lysozyme inhibition zone of SAΔoatA/tagO over that of SAΔoatA is illustrated in Fig. 2.

FIG. 2.

Lysozyme sensitivity by agar diffusion assay. Four milligrams (upper four wells) or only 1 mg (lower row) of lysozyme (L) was applied. Note that the tagO mutant is completely insensitive to lysozyme, whereas the double mutant SAΔoatA/tagO is much more sensitive than the oatA mutant. The presence of sublethal concentrations of penicillin (5 ng P) enhanced lysozyme sensitivity markedly.

In a cell lysis assay, lysozyme (300 μg/ml) was added to exponentially growing cultures. The optical density at 578 nm (OD₅₇₈) values of SA113 and SAΔtagO were unaffected in the presence of lysozyme, whereas the OD₅₇₈ of the mutant SAΔoatA continued for 1 h before it declined slowly during the next hours. In contrast to the case with SAΔtagO, the OD₅₇₈ of the double mutant SAΔoatA/tagO declined rapidly after addition of lysozyme. These results show that growth of SAΔoatA/tagO is much more sensitive to lysozyme than that of the single mutant SAΔoatA (Fig. 3A).

FIG. 3.

Lysozyme induced cell lysis during growth in the absence or presence of penicillin. One percent overnight cultures were inoculated into fresh B medium broth. Lysozyme (A, B) was added to the culture during the late exponential phase. Penicillin (B) was added immediately after inoculation. Changes in optical density were measured at hourly intervals. Symbols: SA113 wild type, □; SAΔoatA, ▪; SAΔoatA/tagO, ▴. Like the wild type, the mutant SAΔtagO was completely resistant (data not shown).

Susceptibility of isolated PG to the hydrolytic activity of lysozyme.

More than 80% of SAΔoatA/tagO-derived PG was digested rapidly within 1 h, whereas only 6 to 7% of PG from SAΔoatA was degraded within the same time span (Fig. 4). In contrast to the case with these mutants, the PG of SA113 and that of its WTA-less mutant (SAΔtagO) were not solubilized within 3 h; only after 24 h did the turbidity decline a little, probably due to constant release of the rather labile O-acetyl group. Indeed, when we removed the O-acetyl groups by alkaline treatment of the PG of SA113 and SAΔtagO (3 h in 80 mM NaOH at 37°C) (2), both PGs became susceptible to lysozyme. We also removed WTA by hydrofluoric acid treatment in wild-type PG and found that the WTA-less PG was completely resistant to the hydrolytic activity of lysozyme (data not shown). This result indicates that the presence or absence of WTA alone has little effect on lysozyme activity; only in the absence of both WTA and O acetylation does PG become highly susceptible to lysozyme hydrolysis.

FIG. 4.

Digestion of the purified peptidoglycan with lysozyme. PG was suspended in 80 mM sodium phosphate-buffered saline (0.5 mg/ml) and digested with 300 μg/ml lysozyme. Aliquots were taken at hourly intervals, and lysis of PG was measured as a decrease in absorbance at 578 nm. Symbols: SA113 wild type, □; SAΔtagO, ▵; SAΔoatA, ▪; SAΔoatA/tagO, ▴.

Comparative analysis of peptidoglycan O acetylation.

The lack of effect of lysozyme on the tagO mutant could be due to an increased level of O acetylation owing to a higher abundance of O-acetylation sites in the absence of WTA. Soluble monomeric muropeptides were generated by digestion of isolated PG from all strains with lysostaphin and cellosyl (3) and separated by high-performance liquid chromatography (HPLC). HPLC profiles of SA113 and the mutant SAΔtagO revealed two peaks: P1, comprised of monomeric de-O-acetylated muropeptides, and P2, comprised of monomeric O-acetylated muropeptides (3). The HPLC profile revealed no significant differences between the muropeptides of SA113 and those of the tagO mutant (Fig. 5A and C), and peak P2 is absent in the HPLC profiles of the mutants SAΔoatA and SAΔoatA/tagO, indicating that both mutants possess only de-O-acetylated PG (Fig. 5B and D). These data were further confirmed by determination of acetate (3) that was released by alkaline treatment of PG. This result shows that the insensitivity of SAΔtagO-derived PG cannot be explained by a higher abundance of O acetylation.

FIG. 5.

Comparative analysis of O acetylation in monomeric muropeptides isolated from various S. aureus strains. PG was isolated from SA113 (A), SAΔoatA (B), SAΔtagO (C), and SAΔoatA/tagO (D) and digested with lysostaphin and cellosyl. The muropeptides were reduced with NaBH₄ (pH 8.0) and analyzed on a C₁₈ column using a linear gradient of 10 mM sodium phosphate buffer to 30% methanol for 190 min. Muropeptides were detected at 210 nm. Peak P1 (retention time, 54 min) represents de-O-acetylated muropeptides and peak P2 (retention time, 67 min) O-acetylated muropeptides. Results illustrate that PG of SAΔoatA and SAΔoatA/tagO contained no O-acetylated muropeptides and that the ratios of de-O-acetylated (P1) to O-acetylated muropeptides (P2) were very similar for SA113 and SAΔtagO.

Content of WTA in the oatA mutant.

The PG of the oatA mutant showed a much lower susceptibility to the hydrolytic activity of lysozyme than that of the mutant SAΔoatA/tagO (Fig. 4). One possible reason could be that the content of WTA is increased in the PG of SAΔoatA. As shown in Fig. 1, both WTA and O-acetyl modification occur at the C6—OH positions of NAM. It might therefore be possible that in SAΔoatA-derived PG, some of the free C6—OH positions of NAM are occupied by WTA. However, we found no significant difference in the content of phosphate (354.4 nmol/mg cell wall for the wild type and 335.8 nmol/mg cell wall for SAΔoatA) or N-acetylglucosamine (394.4 nmol/mg cell wall for the wild type and 306.9 nmol/mg cell wall for SAΔoatA) in PG or WTA preparations (8, 16) of the wild type and the SAΔoatA mutant. Apparently the WTA biosynthesis machinery does not use the free C6—OH positions of NAM in SAΔoatA PG.

Activity of endogenous autolysins in presence of lysozyme.

To rule out the possibility that the increased sensitivity to lysozyme of SAΔoatA/tagO is a consequence of increased endogenous autolysis activity, we compared the autolysin patterns of the wild type and the mutants in a zymogram gel (4, 11); however, there was no difference. Increased autolysin activity very likely does not account for the high lysozyme susceptibility of SAΔoatA/tagO.

Effect of penicillin on lysozyme resistance.

The number of β-1,4-glycosidic linkages in PG depends on the degree of cross-linking of muropeptides (18). It is well known that penicillin decreases cross-linkages by inhibiting the transpeptidase that catalyzes the final step in cell wall biosynthesis (20). We incubated SA113 and mutants with lysozyme and a sublethal concentration of penicillin (5 ng/ml) (Fig. 3B). Indeed, in the presence of penicillin the sensitivity against lysozyme was 10-fold higher with SAΔoatA/tagO (MIC, 12 μg/ml) than with SAΔoatA (MIC, 100 μg/ml), while the wild type and the SAΔtagO mutant stayed completely lysozyme resistant. Lysozyme has two modes of action: hydrolysis of PG and antimicrobial peptide activity (12). We questioned to which activity the penicillin effect was due. To answer this question, penicillin was added to a freshly inoculated growth medium, and after 5 h of cultivation, lysozyme was added. While the growth of the wild type and that of SAΔtagO were unaffected, the OD₅₇₈ values of the SAΔoatA mutant decreased gradually and that of SAΔoatA/tagO rapidly after addition of lysozyme (Fig. 3B). In the absence of lysozyme, no cell lysis was observed with any of the strains (data not shown).

The contribution of WTA in lysozyme resistance is not easy to understand. We see no difference in lysozyme susceptibility whether WTA is present (wild type) or absent (ΔtagO mutant); neither growth nor hydrolysis of isolated PG is affected by lysozyme. The contribution of WTA in lysozyme resistance became visible only for the SAΔoatA/tagO double mutant, which is 10-fold more susceptible than SAΔoatA. We have ruled out that the relative insensitivity of SAΔoatA is due to an increased integration of WTA in PG. Apparently, the WTA biosynthesis machinery does not use the excess of free C6—OH positions in SAΔoatA PG to increase the content of WTA; conversely, the OatA enzyme does not incorporate more O-acetyl groups in SAΔtagO than in the wild type. We are still left with the question of why WTA shows its effect in lysozyme resistance only in the double mutant SAΔoatA/tagO and not in the WTA-less mutant. One explanation could be that in S. aureus only every ninth NAM residue contains phosphoester-linked WTA (14) while approximately every second is O acetylated (3). Even if WTA is missing, binding of lysozyme, which recognizes a hexameric glycan strand, is still blocked by the O-acetyl groups. If O-acetyl groups are missing, lysozyme still may bind to the glycan strand between the WTA residues, albeit interaction might still be hampered by the WTA residues. Optimal binding and highest hydrolytic activity can be displayed by lysozyme only if both modifications are missing. This might explain why the double mutant SAΔoatA/tagO and its isolated PG are more susceptible to lysozyme than SAΔoatA alone.