ABSTRACT
Neisseria gonorrhoeae acetylates its cell wall peptidoglycan (PG) at the C-6 position on N-acetylmuramic acid. To understand the effects of PG acetylation on PG metabolism and release of PG fragments, we have made mutations in the genes responsible for PG acetylation. An insertion mutation in a putative PG acetylase gene (designated pacA) resulted in loss of PG acetylation as detected by a high-performance liquid chromatography-based assay. Sequence analysis of a naturally occurring nonacetylating strain revealed the presence of a 26-bp deletion in pacA. Introduction of the deletion mutation into wild-type gonococci resulted in lack of acetylation, and the phenotype was complemented by the addition of a wild-type copy of pacA at a distant location on the chromosome. Mutations were also introduced into three genes downstream of pacA. The gene directly downstream of pacA was required for acetylation and was designated pacB, whereas the next two genes were not required. Sequences highly similar to pacA and pacB were also found in N. meningitidis and N. lactamica strains, and an insertion in the meningococcal pacA eliminated PG acetylation. Phenotypic analyses of an N. gonorrhoeae pacA mutant did not show any decrease in lysozyme resistance or serum resistance, and the release of PG fragments during growth was unchanged. However, purified PG from the wild-type strain was significantly more resistant to the action of human lysozyme than was PG purified from the pacA mutant. Interestingly, the pacA mutant was more sensitive to EDTA, a compound known to trigger autolysis.
Gonococci acetylate their peptidoglycan (PG) at the C-6 position on N-acetylmuramic acid. The extent of acetylation varies by strain, with most strains acetylating 40 to 60% of the PG (35). One strain, RD5, was reported to have little or no PG acetylation (3, 30). Studies performed with PG purified from RD5 and acetylated PG purified from strain FA19 showed that the FA19 peptidoglycan was more resistant to digestion by lysozyme. Also, the acetylated PG showed greater activity in stimulating inflammation when administered to rats (11).
Toxic PG fragments are released by N. gonorrhoeae during growth (23, 32), and PG acetylation could affect release of these PG fragments. During bacterial growth and division, PG strands are degraded and removed from the cell wall in order to expand it or to change its shape for forming the septum. In Escherichia coli, liberated PG fragments are efficiently taken up into the cytoplasm and recycled for new cell wall synthesis (25). However, in gonococci a significant portion of the fragments are not recycled but are instead released by the bacteria (29). In the fallopian tube organ culture model of gonococcal pelvic inflammatory disease, monomeric PG fragments and lipooligosaccharide were shown to be responsible for causing ciliated cell death (23). The same types of PG fragments have been shown to elicit an inflammatory immune response in several other infection models (8, 14, 18, 22). The major fragments released by N. gonorrhoeae are the 1,6 anhydro-PG monomers, produced by the action of lytic transglycosylases (32). These enzymes cleave the glycan strand of PG at the N-acetylmuramic acid-β-1,4-N-acetylglucosamine bond, the same bond as that acted upon by lysozyme. It has been postulated that the presence of acetate on C-6 of N-acetylmuramic acid will block the lytic transglycosylase reaction (2). Consistent with this hypothesis, nonacetylating strain RD5 was reported to have the highest rate of PG turnover of any gonococcal strain examined (29, 35).
In this study we examined the role of a putative polysaccharide O-acetylase in PG acetylation and examined the effects of acetylation on PG metabolism. A strain lacking PG acetylation was found to have a mutation in the putative PG acetylase gene (pacA). When the mutation was transferred to a wild-type strain, PG acetylation was lost. The mutants were not affected in serum resistance or lysozyme resistance, nor were the mutants affected in release of PG fragments during growth. However, mutation of pacA resulted in increased sensitivity to EDTA, suggesting that PG acetylation may act to decrease autolysis or to stabilize the outer membrane.
MATERIALS AND METHODS
Bacterial strains and growth conditions.Bacterial strains and plasmids are listed in Table 1. Gonococci were grown with aeration in gonococcal base liquid medium (GCBL) containing Kellogg's supplements (20) and 0.042% NaHCO3 (24) or on GCB agar plates (Difco) with Kellogg's supplements and 5% CO2. E. coli was grown in Luria broth or on Luria agar plates (31). Antibiotics were used at the following concentrations: for N. gonorrhoeae, erythromycin at 10 μg/ml, chloramphenicol at 10 μg/ml, and streptomycin at 100 μg/ml; for E. coli, erythromycin at 500 μg/ml and chloramphenicol at 30 μg/ml.
Bacterial strains and plasmids used in this study
Mutation construction and transformation.All transformations of N. gonorrhoeae were performed by spot transformation according to the method of Gunn and Stein (15). Template DNA for screening transformants by PCR was produced by lysing individual colonies in 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl, pH 8.5, as described by Wright et al. (39). Each gonococcal mutant was confirmed to have the expected construction by DNA sequencing.
A point mutation was created in pacA by overlapping PCR of pKH45 using primers ATTTCTACATATGGCAGGGCTACGGC and AGTGACTCATATGCCTGAGAGTACCAT followed by digestion with NdeI and religation. The plasmid was transformed into N. gonorrhoeae strain MS11, and insertion of the plasmid was selected using erythromycin. Transformants were screened for streptomycin sensitivity. Erythromycin-resistant, streptomycin-sensitive transformants were grown for 5 h without antibiotic and plated for streptomycin resistance to select for bacteria that had excised the plasmid by homologous recombination. Isolates carrying the pacA mutation were identified by PCR amplification of pacA and digestion with NdeI.
The 26-bp pacA deletion from strain RD5 was introduced into MS11 by transformation with NheI-digested pKH40. Potential transformants were screened for a smaller pacA PCR product.
The internal deletion in NGO0532 was created by digesting pKH43 with StyI, blunting with T4 DNA polymerase, then digesting with SacI and ligating the 670-bp fragment into similarly digested pKH44 to create pKH62. NheI-digested pKH62 was used to transform N. gonorrhoeae strain MS11, and potential transformants were screened for the mutation by PCR.
A translational stop was introduced into pacB at the third codon. Plasmid pKH42 was amplified with primers TTTGGTCATGATATAACTTTCTTTCCCTTTTCGCCT and GCGGGATCATGACCAAACCCTTAAAAATTGGC. The product was digested with BspHI and religated to produce pKH46. Colonies of MS11 transformed with pKH46 were screened by PCR and digestion of the PCR product with BspHI. A translational stop was introduced into NGO0532 at the third codon by PCR of pKH43 with primers CGGATGACTTAAGAACACTTCATCGCATTTTCCGCCC and GCCCTTAAGTCATGGCTGTGTACTTGATGGTTGC, followed by digestion with AflII and religation to create pKH47. Potential transformants of MS11 with pKH47 were screened by PCR and digestion with AflII. The same method was used to replace the NG0530 start codon with a stop codon, using primers AGTTCTTAAGACCGGACTTATGCCAATTTCTACGAAATGC and AGTCCGGTCTTAAGAACTTCCTTTATGGTTGC.
To create the pacA complementation plasmid pKH38, pacA was amplified by PCR using primers TGCAGGTACCAAGGATGGTTTATGCCGCTGCT and CAGGACTAGTCAGGGCAGACATCAGTATGG. The product was digested with SpeI and KpnI and ligated to similarly digested pKH35. Gonococci were transformed with PciI-digested pKH38, and transformants were selected with chloramphenicol. Transformants were screened by PCR for the presence of pacA at the complementation site between lctP and aspC and for maintenance of the original mutation.
PG isolation.Gonococci grown overnight on 40 GCB agar plates were swabbed into 20 ml of ice-cold 25 mM sodium phosphate buffer, pH 6 (phosphate buffer), and centrifuged at 3,800 × g for 10 min at 4°C. The pellet was washed with 10 ml phosphate buffer and centrifuged again. The pellet was then suspended in 10 ml phosphate buffer and added drop-wise to 10 ml of boiling 8% sodium dodecyl sulfate. After boiling for 1 h, the PG preparation was centrifuged for 30 min at 30,000 × g at 15°C. Boiling and centrifugation were repeated once. The pellet was washed five times by suspension in 10 ml phosphate buffer followed by centrifugation at 30,000 × g for 30 min at 25°C. This preparation was ultracentrifuged at 162,000 × g for 30 min at 25°C and suspended in 3 to 5 ml phosphate buffer. A portion of the sample was lyophilized and weighed to quantify the PG. Purified PG was stored at −20°C.
Acetate detection by high-performance liquid chromatography (HPLC).Acetate was released from macromolecular PG and detected by HPLC essentially as described by Payie et al. (26). A solution of macromolecular PG containing 15 mg PG was centrifuged at 51,500 × g for 30 min at room temperature, and the PG was suspended in 360 μl H2O. Beta elimination was carried out by addition of a one-fourth volume of 2 N NaOH (final concentration of 0.4 N), and the reaction was incubated at room temperature for 2 h. Samples were centrifuged at 51,500 × g for 30 min at room temperature. Supernatants were removed and stored at 4°C. Acetate was detected by running 100 μl of supernatant over a Bio-Rad HPX-87H organic acid column using a mobile phase of 10 mM sulfuric acid running at 0.6 ml/min at 35°C. Elution profiles were compared to those generated using an organic acid standard mix (Bio-Rad) or acetic acid.
PG analysis.Procedures for PG labeling and assaying PG turnover were performed as previously described (6). Analysis of released fragments was performed by size-exclusion chromatography and comparison with known standards as previously described (4). For gauging sensitivity to human lysozyme, PG was labeled with [6-3H]glucosamine during log-phase growth and purified as previously described (6). Approximately 370 μg of labeled PG was used in a reaction with 28 μg human neutrophil lysozyme (Sigma) in 1 ml 50 mM sodium phosphate, pH 6.4. The reaction was incubated at 37°C for 18 h. PG fragments were separated by size-exclusion chromatography and quantified by scintillation counting.
Bacterial cell death in the presence of lysozyme, serum, or EDTA.Gonococci grown overnight on GCB plates were transferred into 3 ml GCBL with supplements and grown for 3 h with aeration as described above. For serum resistance assays, the bacteria were centrifuged for 30 s at 15,000 × g, washed once, and then suspended in Dulbecco's modified Eagle's medium. Serial dilutions were made in this medium in a 96-well plate, and an equal volume of normal human serum was added. The plate was incubated at 37°C in 5% CO2 for 30 min. Total CFU were determined at 0 and 30 min by dilution in GCBL and overnight growth on GCB plates. Killing by lysozyme was assayed in a similar manner, except that dilutions in the 96-well plate were done using GCBL with supplements. Hen egg white lysozyme was added to a final concentration of 100 μg/ml. EDTA was used at 1 mM concentration in addition to the lysozyme or alone.
Southern blotting.Chromosomal DNA was prepared from Neisseria species as previously described (7). Southern blotting was performed by standard procedures (31) except that DNA was transferred to a Duralon filter by vacuum blotting and covalently attached by UV cross-linking (Stratagene). The gonococcal pacAB genes were excised from pKH53 with BamHI and NotI and labeled with [α-32P]dCTP by random-primed labeling. Hybridization was allowed to proceed for 18 h at 62°C, and two low-stringency washes were performed using 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate), 0.1% sodium dodecyl sulfate at room temperature for 15 min each.
RESULTS
Identification of pacA as a putative PG acetylase gene.A search of the gonococcal genome sequence of strain FA1090 (accession no. AE004969 ) revealed a putative O-acetyltransferase of the membrane-bound O-acyltransferase (MBOAT) family (19). This open reading frame (NGO0534) showed significant similarity with AlgI, an enzyme of Pseudomonas aeruginosa that acts to acetylate the secreted exopolysaccharide alginate (13). We have designated the gene pacA, for peptidoglycan acetylation gene A, since further work described below indicates that it is necessary for PG acetylation. PacA shows 33% identity, 52% similarity to AlgI over 414 amino acids. The predicted protein has several hydrophobic regions, suggesting that, like AlgI, it is likely a membrane protein. PacA has homologues in many other bacteria (12), including Neisseria meningitidis, a pathogenic species closely related to N. gonorrhoeae. N. meningitidis was recently shown to have acetylated PG (1).
Mutations in pacA result in loss of acetylation.To determine if pacA is necessary for PG acetylation, an insertion mutation was made using a nonreplicating plasmid carrying a 311-bp fragment of pacA coding sequence. Recombination of the plasmid into the gonococcal chromosome resulted in an insertion mutation, disrupting the coding sequence after amino acid 354 (Fig. 1). The plasmid, pTHH5, was transformed into FA19, a gonococcal strain that has been shown to have significant levels of PG acetylation (35). The pacA interruption mutant (THH51) was compared to the wild-type strain in an HPLC-based assay for acetylation. PG was purified from each strain and subjected to mild alkali treatment to release O-linked acetate. Acetate was measured by retention on an organic acid column with detection at 210 nm. FA19 produced significant acetate, exhibiting a strong peak at approximately 15.2 min (Fig. 2A). However, the pacA interruption mutant showed very little acetate that could be liberated from the PG (Fig. 2B).
Genetic map of the pacA region. Genetic organization information was derived from the N. gonorrhoeae FA1090 genome sequence. The locations of mutations used in this study are indicated below the map. Black flags indicate the location of gonococcal DNA uptake sequences. The large gray arrow indicates the position of a putative sigma 70-type promoter. Frameshifts in the NGO0535 sequence relative to its homologues are indicated by breaks in the NGO0535 arrow.
HPLC-based detection of O-linked acetate from the PG of wild-type and pacA mutant strains. Acetate was monitored as absorbance at 210 nm, eluting at approximately 15.2 min and expressed here in arbitrary units (AU). wt, wild type.
One gonococcal strain, RD5, has been reported not to acetylate its PG (3). To determine if the lack of acetylation might be explained by a defect in PacA, we sequenced the pacA gene and the surrounding chromosomal region from strain RD5. The sequence showed that indeed, pacA carried a mutation in strain RD5, a 26-bp deletion compared to the strain FA1090 sequence. This mutation results in a loss of coding sequence and a premature stop codon 86 bp after the mutation, thus reducing the open reading frame to less than 40% of the coding sequence (Fig. 1). Consistent with the results of Rosenthal et al. (30), very little acetate could be detected in the RD5 PG (Fig. 2C).
As RD5 is a nonpiliated, nonreverting strain, it is not transformable by standard procedures. Therefore we chose to characterize the effect of the RD5 deletion by transforming the mutation into N. gonorrhoeae strain MS11, a strain we have previously used for PG characterization studies (4-6). An 827-bp fragment was amplified from RD5 by PCR and cloned into pIDN1. The mutation was introduced into the MS11 chromosome by transformation. HPLC analysis showed that while MS11 produced significant amounts of acetate on the PG (Fig. 2D), little or no acetate could be liberated from the PG of the pacA deletion mutant KH518 (Fig. 2E).
We also tested the effect of a point mutation in pacA on PG acetylation. Members of the membrane-bound O-acyltransferase family of acyltransferases contain a common motif with an invariant histidine residue (19), and this residue has been shown to be necessary for acetylation of alginate by P. aeruginosa AlgI (12). A mutation changing the histidine at position 329 of the PacA predicted protein to glutamine similarly resulted in loss of PG acetylation (Fig. 2F).
Complementation analysis.Expression of pacA from a distant site on the chromosome restored acetylation. To complement the mutation, wild-type pacA was cloned from N. gonorrhoeae strain MS11 into complementation plasmid pKH35. The resulting plasmid, pKH38, was transformed into the pacA insertion mutant THH51 and also into the deletion mutant KH518. The HPLC assay showed that the complementation construct was able to restore PG acetylation to the MS11 pacA deletion mutant KH518 (Fig. 3A) but not to the FA19 pacA insertion mutant THH51 (Fig. 3B). This result demonstrated that the deletion in pacA was responsible for the lack of PG acetylation. However, one or more additional genes downstream of pacA must also be necessary for PG acetylation, since the polar insertion mutation in THH51 could not be complemented by pacA alone. To determine which additional genes were required for acetylation, we made mutations in the three genes downstream of pacA (Fig. 1). The regions of DNA surrounding the start of each coding sequence were amplified from wild-type strain MS11, and stop codons were introduced into the coding sequences at, or immediately after, the starts. These mutations were transformed into MS11 without selection, and transformants carrying each mutation were identified by PCR. Mutation of fadD homologue NGO0530 did not affect acetylation (Fig. 3C), nor did mutation of NGO0532 cause any reduction in acetate (Fig. 3D). This result was also confirmed by introducing a deletion into NGO0532 that removed an internal 946 bp (79%) of the coding region. The deletion mutant also maintained PG acetylation (Fig. 3E). However, mutation of NGO0533 eliminated detectable acetate on the PG, and we designated this gene pacB (Fig. 3F). The putative PacB protein is predicted to be a 36-kDa protein. Few regions of significant hydrophobicity and the presence of a predicted signal sequence suggest that PacB is a periplasmic protein. An extensive analysis of the sequence of meningococcal homologue of PacB (NMA1479) has been published (12).
pacA and pacB are necessary for PG acetylation. O-linked acetate was detected as in Fig. 2. Complemented strains are indicated as pacA+ and carry wild-type pacA on the gonococcal chromosome inserted with the pKH35 construct between lctP and aspC.
PG acetylation in other Neisseria species.A recent study of the composition of Neisseria meningitidis PG showed several PG fragments with acetate at the C-6 position of N-acetylmuramic acid (1). To determine if PG acetylation occurs in the same way in N. meningitidis as it does in N. gonorrhoeae, pTHH5 was used to make an insertion mutation in the meningococcal pacA homologue. The resulting mutant was tested for PG acetylation using the HPLC-based assay. Although the amount of acetate liberated from meningococcal PG was less than that from gonococcal strains, it was readily detectable (Fig. 4A). However, the meningococcal pacA insertion mutant JD1628 did not exhibit detectable acetate (Fig. 4B). This result suggests that, like gonococci, meningococci require pacAB for PG acetylation.
PG acetylation in meningococci and commensal Neisseria species. O-linked acetate was detected as in Fig. 2. Wild-type strains used were N. meningitidis ATCC 13102, N. lactamica ATCC 49142, and N. flavescens ATCC 13115.
Although they exhibit significant divergence, meningococci and gonococci share several similar virulence mechanisms. Only a subset of these virulence determinants is generally present in commensal Neisseria, species that do not normally cause disease in healthy individuals. We used PCR to search for pacA homologues in several commensal Neisseria species. No specific PCR products were detected in N. flavescens, N. subflava, N. cinerea, or N. mucosa. However, two strains of N. lactamica did produce specific products for pacA by PCR (data not shown). DNA sequence of the pacA gene from N. lactamica ATCC 49142 showed 94% identity with the gonococcal pacA. Consistent with this result, PG from N. lactamica ATCC 49142 was found to be acetylated (Fig. 4C). Examination of six Neisseria species by Southern blotting identified similar sequences in every species (Fig. 5). However, PG prepared from N. flavescens strain ATCC 13115 showed no acetate present (Fig. 4D), suggesting that the sequences similar to pacAB may not be functional PG acetylation genes in the commensal Neisseria species other than N. lactamica.
Southern blot to detect pacAB homologues in commensal Neisseria species. HindIII-digested chromosomal DNA from N. lactamica ATCC 49142, N. flavescens ATCC 13115, N. subflava ATCC 14799, N. mucosa ATCC 49233, N. cinerea NNRL9, and N. sicca ATCC 9989 was probed with pacAB from N. gonorrhoeae MS11. M, molecular mass in kilobases. wt, wild type.
PG fragment release in the pacA mutant.Our main interest in examining PG acetylation was to determine if it affected release of toxic PG fragments. Therefore, we analyzed PG fragment release by metabolic labeling of the PG using [6-3H]glucosamine and following the appearance of soluble PG fragments in the supernatant. Loss of PG from the cell fraction occurs due to PG turnover during growth. The data in Fig. 6 show that release of PG into the medium occurred at the same rate in the mutant and wild-type cultures during log-phase growth. During stationary phase (at 26 h) the pacA mutant appeared to have released slightly more PG than the wild-type strain. Although this difference was seen in each of three trials, the difference is not statistically significant.
Mutation of pacA does not affect the rate of PG turnover during growth. PG was metabolically labeled with [6-3H]glucosamine. PG turnover was measured as loss of radioactive material from the macromolecular fraction during growth in liquid culture. The generation time under these conditions is approximately 60 min during log-phase growth. The values shown are the averages of three separate experiments. wt, wild type.
To determine if the released PG fragments were of the same types, the PG was metabolically labeled with [6-3H]glucosamine and PG fragments present in log-phase culture supernatants were separated by size-exclusion chromatography. Gonococci are known to release three major types of PG fragments that can be distinguished by size: PG multimers, PG monomers, and free disaccharide (32). PG monomers are predicted to be released by the action of lytic transglycosylases, and we have previously shown that lytic transglycosylase LtgA is responsible for approximately half of the PG monomers released (4). If PG acetylation blocks PG fragment production, then the pacA mutant would be expected to show a greater release of PG monomers. However, the profile of released fragments from the pacA mutant was identical to that of the wild-type strain (Fig. 7). This result suggests that PG acetylation does not block the lytic transglycosylases involved in PG fragment release.
Profile of PG fragments released by wild-type and pacA mutant strains. PG was metabolically labeled with [6-3H]glucosamine. PG fragments released into the culture medium were separated by size-exclusion chromatography and detected by scintillation counting. wt, wild type. CPM, counts per minute.
Digestion with human lysozyme.We compared purified PG from wild-type strain MS11 to PG from the pacA mutant KH518 for susceptibility to digestion with human lysozyme. PG from the mutant was digested to three major peaks detectable by size-exclusion chromatography, representing monomers, dimers, and trimers (Fig. 8). Very little large, macromolecular PG remained. However, PG from the wild-type strain was much less sensitive to digestion, showing two additional peaks, representing PG fragments larger than those obtained from the pacA mutant PG. This result demonstrates that the pacAB gene products make gonococcal PG more resistant to lysozyme and suggests that acetylation may allow macromolecular PG from wild-type gonococcal strains to persist in the host.
Macromolecular PG from the pacA mutant is more sensitive to digestion with human neutrophil lysozyme. PG was metabolically labeled with [6-3H]glucosamine. Isolated PG was digested with human lysozyme, separated by size-exclusion chromatography, and detected by scintillation counting. wt, wild type.
Sensitivity to lysozyme, serum, and EDTA.Since the purified PG from the wild-type strain was more resistant to lysozyme than that of the pacA mutant, we hypothesized that gonococci that acetylate the cell wall PG should be more resistant to killing by lysozyme. In the body, lysozyme gains access to the cell wall of gram-negative bacteria through the pore formed by the membrane attack complex of complement (36) or through membrane disruptions caused by lactoferrin (9). Whether the presence of lysozyme increases cell killing in this context is controversial (36). We tested the effect of the pacA mutation on serum resistance. The mutation was introduced into the serum-sensitive strain F62, and the degree of killing by human serum was determined. No difference was found between the mutant and wild-type strains (data not shown). This result likely indicates that complement killed the cells even without breakdown of the cell wall.
To measure the effect of lysozyme in cell lysis more directly, the wild-type, pacA mutant KH518, and the complemented strain were incubated in the presence of 100 μg/ml lysozyme for 30 min. EDTA was added to the assay at 1 mM concentration to allow lysozyme to cross the outer membrane. The wild-type and complemented strains were greater than 10-fold more resistant to the lysozyme-EDTA treatment than the pacA mutant (data not shown). Surprisingly, the death seen in the pacA mutant culture could not be attributed to the lysozyme but was also present in a no-lysozyme control, where the only added material was the EDTA. The increased sensitivity of the pacA mutant to EDTA was exhibited over a wide range of cell densities (Fig. 9).
The pacA mutant shows increased sensitivity to EDTA. Gonococci from exponential-phase cultures were diluted in GCBL medium containing 1 mM EDTA and incubated at 37°C for 30 min. Input CFU per milliliter (CFU/ml) were determined from identical dilutions in GCBL without EDTA. Bacteria surviving the EDTA treatment (output CFU/ml) were enumerated following dilution in GCBL and overnight growth. The values shown are the means of three separate experiments. wt, wild type.
DISCUSSION
We have shown that N. gonorrhoeae requires pacA and pacB for PG acetylation. The predicted PacA protein is similar to AlgI, a cytoplasmic membrane protein of P. aeruginosa that functions to acetylate the exopolysaccharide alginate (13). PacB is similar to a class of predicted type II membrane proteins, encoded by genes located adjacent to the algI homologues (12). AlgJ is the P. aeruginosa type II membrane protein required for alginate acetylation. Type II membrane proteins are found in the periplasm but maintain their signal sequences uncleaved, anchoring them in the cytoplasmic membrane (28). Homologues of pacA and pacB are found in multiple bacterial species (12), and it is not known if these bacteria produce acetylated PG. pacA and pacB are found in N. meningitidis and N. lactamica, and we found that an insertion in N. meningitidis pacA disrupted PG acetylation (Fig. 4B). Sequences similar to pacAB were detected in five other commensal Neisseria species (in addition to N. lactamica) by low-stringency Southern hybridization (Fig. 5). However, these sequences were not identified on high-stringency Southern blots and could not be amplified by PCR with primers designed for the gonococcal pacAB genes. Since we were not able to detect acetate on the PG of N. flavescens and the sequences are not highly similar in the commensal Neisseria compared to the gonococcal pacAB genes, we predict that PG acetylation may be limited to the pathogenic species N. gonorrhoeae and N. meningitidis and to the sometimes, though rarely, pathogenic species N. lactamica.
Acetylated PG is significantly more resistant to digestion with human lysozyme than is nonacetylated PG. The sensitivity of the PG to human lysozyme was shown here for the pacA mutant and was previously shown by Striker et al. for RD5 PG (33). The resistance of the acetylated PG to lysozyme digestion likely contributes to pathogenesis of gonococcal and meningococcal infections. Intradermal injection of acetylated gonococcal PG caused systemic arthritis in rats (11). Nonacetylated PG was greatly reduced in causing this inflammatory response. Arthritis is a common symptom of disseminated gonococcal infection. Although it is not clear what advantage would be afforded the bacterium by causing arthritis, a second activity of macromolecular PG fragments is more easily seen as advantageous. Large soluble PG fragments were shown to deplete complement components from human serum, specifically C3 and C4 (27). Although gonococci and meningococci have several mechanisms for serum resistance, complement is important in clearing Neisseria infections. Individuals with complement deficiencies are subject to high rates of gonococcal or meningococcal infections (21). As gonococci are prone to autolysis (24), large PG fragments are likely released by dying bacteria during infection and could protect the remainder of the surviving population.
It appears that PG acetylation does not significantly affect PG fragment release during growth. The rate of PG turnover and the profile of released fragments did not differ between the mutant and wild-type strains (Fig. 6 and 7). If PG acetylation inhibits the action of lytic transglycosylases, it must not inhibit those that are involved in PG fragment release. We have previously shown that gonococcal lytic transglycosylase LtgA acts in the release of PG fragments (4). However, lytic transglycosylase LtgC does not significantly contribute to PG fragment release (5). The gonococcal genome sequence contains five lytic transglycosylase homologues (accession no. AE004969 ), and the gonococcal genetic island contains two more lytic transglycosylase homologues (16). The activities of some of these enzymes might be affected by PG acetylation but not in a way that reduces PG fragment release during the growth phase.
An unexpected outcome of these studies was the observation that the pacA mutant was more sensitive to EDTA than the wild type. It was previously shown that EDTA stimulates gonococci to lyse when suspended in buffer (10). Similarly, increased Mg2+ concentration reduced cell lysis. Thus, one interpretation of these results is that PG acetylation protects the cell wall from PG hydrolases and is thus a mechanism for controlling autolysis. However, Wegener et al. found that PG hydrolysis occurred at the same rate in the presence of Mg2+ as in its absence and concluded that Mg2+ decreases lysis by stabilizing the outer membrane (37). It should be noted that gonococci do not contain peptidoglycan-linked (Braun's) lipoprotein (38). Thus, the presence of divalent cations may be of increased importance in gonococci for maintaining outer membrane integrity. It is conceivable that Mg2+ or other divalent cations complex with the acetyl groups on the PG and with phospholipids to strengthen PG interaction with the outer membrane or that the acetylated PG stores more Mg2+ that may be used to stabilize lipooligosaccharide during membrane perturbations.
ACKNOWLEDGMENTS
We thank R.S. Rosenthal for advice and encouragement as well as the gift of PG standards. We thank Tajie H. Harris for technical assistance.
This work was supported by NIH grant AI47958 to J.P.D.
FOOTNOTES
- Received 18 March 2005.
- Returned for modification 20 April 2005.
- Accepted 29 April 2005.
- Copyright © 2005 American Society for Microbiology
