Analysis of the Peptidoglycan Hydrolase Complement of Lactococcus lactis: Identification of a Third N-Acetylglucosaminidase, AcmC

Abstract

The peptidoglycan hydrolase (PGH) complement of Lactococcus lactis was identified by amino acid sequence similarity searching of the L. lactis IL-1403 complete genome sequence. Five PGHs that are not encoded by prophages were detected, including the previously characterized AcmA and AcmB proteins. Four of these PGHs, AcmA to AcmD, contain a catalytic domain homologous to that of enterococcal muramidase, but they have different domain structures. The fifth one (YjgB) has sequence similarity with the active-site domain of peptidoglycan-specific endopeptidases. The three new PGH-encoding genes identified in this study are all actively transcribed in L. lactis subsp. cremoris MG1363. The relative abundance of their transcripts varied during growth and was maximal during the early exponential growth phase. The three encoded proteins have peptidoglycan-hydrolyzing activities which are detected only at acidic pHs by zymography. Like AcmA and AcmB, AcmC has N-acetylglucosaminidase activity rather than the N-acetylmuramidase activity predicted by sequence similarity.

Bacteria produce enzymes that are capable of hydrolyzing bonds in their own protective cell wall peptidoglycan. These peptidoglycan hydrolases (PGHs) are involved in different cellular functions that require cell wall remodeling during growth and division. These include cell wall expansion, peptidoglycan turnover, recycling and maturation, and daughter cell separation (36, 38). PGHs are classified into the following four classes according to their hydrolytic bond specificities: N-acetylmuramidases, N-acetylglucosaminidases, N-acetylmuramyl-l-alanine amidases, and endopeptidases. Peptidoglycan is essential for cellular integrity, and thus its hydrolysis by some PGHs, called autolysins, causes cellular autolysis, leading to the release of intracellular contents. Autolysis can be observed under conditions which result in the cessation of peptidoglycan synthesis, such as stationary phase or exposure to antibiotics, possibly from an uncontrolled action of PGHs (33, 36).

Lactic acid bacteria are used as starters for cheese making, and their autolysis in cheese during ripening plays an important role in flavor development. Indeed, bacterial lysis leads to the release of the bacterial intracellular enzymatic pool into the cheese curd, which contributes to the conversion of milk components to aromatic compounds (9, 15). The model lactic acid bacterium, Lactococcus lactis, is widely used for dairy fermentation, and several studies have shown the positive impact of its cellular autolysis on cheese flavor. Specifically, it was shown that the release of intracellular peptidases into the cheese curd results in a higher rate of production of free amino acids and in the degradation of bitter hydrophobic peptides (10, 40). More recently, it was shown that, in addition, L. lactis lysis can stimulate amino acid catabolism, leading to the formation of aromatic compounds (6).

Two PGHs were previously identified and characterized at the molecular level for L. lactis. The major autolysin AcmA is required for proper cell separation after division (8) and is involved in cellular autolysis during stationary phase after growth on a synthetic culture medium (7). AcmA has a modular structure with two domains (8). The N-terminal catalytic domain exhibits sequence similarity with Enterococcus hirae muramidase Mur-2 and the C-terminal domain contains three amino acid repeats, named LysM (4), that are involved in cell wall binding (39). The second characterized PGH, AcmB, can also contribute to cellular autolysis, but to a lesser extent than AcmA (24). AcmB has a modular structure with three domains: they are a central catalytic domain homologous to E. hirae muramidase Mur-2; an N-terminal domain that is rich in Ser, Thr, Pro, and Asn residues, resembling a cell-wall-associated domain; and a C-terminal domain with unknown function. Both AcmA and AcmB have been shown to exhibit N-acetylglucosaminidase activity (24; A. Steen, G. Buist, G. Horsburgh, S. Foster, O. Kuipers, and J. Kok, Abstr. Eurolab Conf., abstr. p72, 2001) rather than the N-acetylmuramidase activity predicted by sequence similarity.

The availability of the complete genome sequence of L. lactis subsp. lactis IL-1403 (5) prompted us to search for its PGH complement by performing a sequence similarity search of previously characterized PGHs (38). We found that L. lactis contains a relatively simple PGH complement composed of five enzymes, including AcmA and AcmB. We show here that the three additional putative PGHs are functional and that one of them, named AcmC, is the third N-acetylglucosaminidase identified in L. lactis.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

L. lactis strains were grown at 30°C in M17 medium (Difco) or chemically defined medium (CDM) (37). The media were supplemented with 0.5% glucose (wt/vol). Escherichia coli strains were grown in Luria-Bertani broth at 37°C with shaking. The medium was supplemented, when appropriate, with ampicillin (100 μg ml⁻¹) or kanamycin (25 μg ml⁻¹). E. coli TG1 [supE hsdΔ5 thi Δ(lac proAB) F′(traD36 proAB⁺ lacI^q lacZΔM15)] (18) was used for plasmid construction and purification. The pGEMT easy vector (Promega Corporation) was used to clone PCR products into E. coli.

General recombinant DNA techniques.

DNA manipulations were carried out essentially as described previously (34). Restriction enzymes (Eurogentec or Roche Molecular Biochemicals), T4 DNA ligase (Epicentre), and Taq DNA polymerase (Appligene Oncor) were used as recommended by the manufacturers. Oligonucleotides were purchased from Invitrogen. Plasmids were extracted by using a QIAprep Spin miniprep kit (Qiagen). The L. lactis total DNA was isolated as described previously (32). PCRs were performed with a GeneAmp 2400 PCR system (Perkin-Elmer). DNA sequences were determined with an Applied Biosystems 370A automated DNA sequencer and with ABI PRISM dye terminator cycle sequencing and dye primer cycle sequencing kits (Perkin-Elmer). DNA and protein sequences were assembled and analyzed with the Genetic Computer Group sequence analysis package (GCG Inc., Madison, Wis.) and the tools available on the Expasy (Expert Protein Analysis System) Biology Server of the Swiss Institute of Bioinformatics (http://www.expasy.org).

Cloning of yjgB from L. lactis MG1363.

With the primer pair AU7-AU26 (Table 1), selected from the sequence surrounding the yjgB gene identified in the L. lactis IL-1403 sequence (5), a 1.4-kb DNA fragment was amplified by PCR from L. lactis subsp. cremoris MG1363 total DNA. The nucleotide sequence of the fragment was determined. Subsequently, a 0.6-kb fragment located upstream of the 1.4-kb amplified fragment was obtained with the primers AU31, selected from the IL-1403 sequence, and AU32 (Table 1), selected from the sequence of the 1.4-kb MG1363 DNA fragment. Finally, the sequence of a 1,935-bp MG1363 DNA fragment encompassing the yjgB gene was obtained.

TABLE 1.

Oligonucleotides used for this study

Primer	Sequence (5′-3′)a	Restriction site	Usage
AU7	GACGTAATGACAATGCCAGGb		Cloning (yjgB, MG1363)
AU26	TTGGCTTTGTTCCAATCTTGb		Cloning (yjgB, MG1363)
AU31	GCTACACTTCGTGCACTTAAb		Cloning (yjgB, MG1363)
AU32	TACCATCATCAGTCACGTCCc		Cloning (yjgB, MG1363)
AU51	CATGCCATGGGACAAAAGCCGTCTGAAGCAb	NcoI	Northern blot probe, expression (acmC)
AU52	GAAGATCTAATACTTGAACTTTTTATCATATb	BglII	Northern blot probe, expression (acmC)
AU47	CATGCCATGGCATCTGTTCAAGAAATTATTb	NcoI	Northern blot probe, expression (acmD)
AU48	GAAGATCTGATTCTAATTGTTTGTCCTGb	BglII	Northern blot probe, expression (acmD)
AU53	CATGCCATGGATTCAGTATTTGAAGACAATc	NcoI	Northern blot probe, expression (yjgB)
AU54	GAAGATCTGTATTCTAAGGCAAAATCAGc	BglII	Northern blot probe, expression (yjgB)

^aRestriction sites are underlined.

^bSequence from L. lactis IL1403 (5).

^cSequence from L. lactis MG1363 yjgB (this study).

Northern blotting.

Total RNA fractions were extracted at different stages of growth, corresponding to optical densities at 650 nm (OD₆₅₀) of 0.25, 0.5, 0.75, 1.0, and 1.25, from L. lactis MG1363 grown at 30°C in CDM as previously described (1). Northern blot hybridization was performed according to a standard protocol (34). RNA samples (20 μg) were transferred to a Hybond N+ nylon membrane (Amersham Pharmacia Biotech). The membrane was successively hybridized with probes corresponding to the different entire genes. The probes were obtained by PCR amplification of DNA fragments with the following primers and templates: acmC probe, AU51-AU52 primers (Table 1) and IL-1403 DNA template; acmD probe, AU47-AU48 primers (Table 1) and IL-1403 DNA template; and yjgB probe, AU53-AU54 primers (Table 1) and MG1363 DNA template. The acmB probe was prepared as described previously (24), with the MG1363 DNA template, and the acmA probe was prepared with the PALA-4 and PALA-14 primers (8) and the MG1363 DNA template. Lastly, the membrane was hybridized with a 16S RNA probe obtained by PCR amplification with previously described primers (13). The probes were labeled with [α-³²P]dCTP by use of a random primed DNA labeling kit (Roche Molecular Biochemicals). Hybridizations were performed under high-stringency conditions (50% formamide). After hybridization with each probe, the membrane was dehybridized by immersion in boiling 1× SSC (0.15 M NaCl plus 0.015 M sodium citrate)-0.2% sodium dodecyl sulfate (SDS) and a subsequent incubation for 1 h at room temperature. It was then rehybridized with another specific probe obtained as described above. Radioactivity was quantified with a Storm imaging system (Molecular Dynamics) using ImageQuaNT software. Relative amounts of transcripts were standardized by hybridization with an L. lactis 16S rRNA-specific probe.

Expression and purification of His-tagged proteins in E. coli.

The proteins were overexpressed in E. coli as C-terminal hexa-His-tagged proteins from the pQE60 expression vector (Qiagen). DNA fragments encoding polypeptides corresponding to the PGHs devoid of their putative signal peptide sequences were amplified by PCR. The acmC and acmD fragments were amplified from the L. lactis IL-1403 total DNA with primer pairs AU51-AU52 and AU47-AU48, respectively, and the yjgB fragment was amplified from the L. lactis MG1363 total DNA with primers AU53 and AU54. The forward and reverse primers contained NcoI and BglII restriction sites, respectively, at their 5′ ends (Table 1). The PCR fragments were digested with NcoI and BglII and cloned in frame upstream of the hexa-His box sequence in the pQE60 vector, which was precut by the same enzymes. E. coli M15(pREP4) (Qiagen) competent cells were transformed with the resulting plasmids.

E. coli recombinant strains were grown at 37°C in Luria-Bertani medium containing ampicillin and kanamycin. Protein expression was achieved by induction with 1 mM IPTG and a subsequent incubation of the culture for 4 h at 37°C. To avoid the formation of AcmD-His inclusion bodies, we transferred the bacterial culture to 28°C after the addition of IPTG. The cells were harvested by centrifugation. In the case of soluble His-tagged proteins (YjgB-His and AcmD-His), cells were broken by one passage at a pressure of 1.6 × 10⁵ kPa in a cell disruption system (Constant System Ltd., Warwickshire, United Kingdom). The soluble fraction was collected by centrifugation at 15,000 × g for 15 min at 4°C and the hexa-His-tagged protein was purified by affinity chromatography through Ni²⁺-nitrilotriacetic acid spin columns (Qiagen) according to the manufacturer's protocol for purification under native conditions. Because AcmC-His inclusion bodies were obtained even at 28°C, the protein was purified through Ni²⁺-nitrilotriacetic acid spin columns under denaturing conditions after solubilization in 8 M urea according to the manufacturer's instructions. The purity of the His-tagged proteins was confirmed by SDS-polyacrylamide gel electrophoresis (PAGE). Purified His-tagged proteins were dialyzed against citrate-phosphate buffer, pH 4.0, before further use.

SDS-PAGE and renaturing SDS-PAGE.

SDS-PAGE was performed with 15% (wt/vol) polyacrylamide separating gels. Renaturing SDS-PAGE was performed as previously described (24, 28), with 0.2% (wt/vol) autoclaved Micrococcus lysodeikticus cells or 0.4% (wt/vol) L. lactis cell walls included in the gels as an enzyme substrate. To test the effect of pH on activity detection, we used the following incubation buffers: 25 mM sodium citrate-50 mM sodium phosphate at pH 3.0, 4.0, or 5.0, 50 mM morpholineethanesulfonic acid (MES) at pH 6.0, and 50 mM Tris-HCl at pH 7.0 or 8.0.

Determination of hydrolytic bond specificity.

The hydrolytic bond specificity of AcmC was determined with a Bacillus subtilis HR vegetative cell wall substrate as described previously (24). Cell walls (5 mg) prepared as described previously (2) were incubated overnight at 37°C with 500 μg of purified hexa-His-tagged recombinant protein in a final volume of 500 μl of buffer (25 mM sodium citrate-50 mM sodium phosphate, pH 4.0). Samples were boiled for 3 min to stop the reaction. The insoluble material was removed by centrifugation at 14,000 × g. Half of the soluble muropeptide fraction was further digested with Cellosyl (250 μg/ml; a gift from Hoechst AG, Frankfurt, Germany). The soluble muropeptides obtained after digestion were reduced with sodium borohydride and separated by reversed-phase high-performance liquid chromatography (RP-HPLC) as described by Atrih et al. (3). Muropeptide analysis by matrix-assisted laser desorption ionization-time of flight mass spectrometry was performed as described previously (24).

Nucleotide sequence accession number.

The EMBL/GenBank accession number for the MG1363 yjgB gene reported in this paper is AJ414527.

RESULTS

Identification of putative PGHs in the L. lactis subsp. lactis IL-1403 genome sequence.

As well as the four previously identified lytic enzymes encoded by prophages (11), five definite or putative PGHs were identified in the L. lactis IL-1403 complete genome sequence (5) by amino acid sequence similarity searching. These five PGHs can be classified into two families (Fig. 1). Four of them, the previously characterized AcmA and AcmB proteins as well as the putative enzymes AcmC and AcmD, belong to the family of enterococcal muramidases (12). Conversely, YjgB has similarity with the active-site domain of peptidoglycan-specific endopeptidases (23). All of them possess a putative signal peptide sequence at the N terminus (31).

FIG. 1.

Schematic representation of the domain structure of the five PGHs of L. lactis IL-1403. The total number of amino acids (aa), the calculated molecular mass (M.M.), and the calculated isoelectric point (pI) of each protein, with or without a putative signal peptide (SP), are indicated.

The four proteins of the enterococcal muramidase family possess similar catalytic domains but different overall structures. Like Mur1 from Streptococcus thermophilus (25) and Mur from Leuconostoc citreum (14), AcmC contains a catalytic domain but lacks a specific cell wall binding domain. In contrast, like AcmA, AcmD has a modular structural organization with two domains, namely an N-terminal catalytic domain (residues 26 to 169) fused to a C-terminal cell wall binding domain. The latter contains three sequence repeats known as LysM domains (4), which were previously described as cell-wall-targeting and peptidoglycan-binding sequences in AcmA (39). The catalytically important amino acid residues identified for the enterococcal muramidase family (27) are conserved in L. lactis AcmC (Glu-104 and Asp-125) and AcmD (Glu-89 and Asp-110). In the previously characterized enzyme AcmB, the catalytic domain is surrounded by an N-terminal putative cell-wall-associated domain and a C-terminal domain of unknown function (24).

The last putative PGH, encoded by yjgB, contains only one domain corresponding to the catalytic domain. This domain is homologous to those of γ-d-Glu-(l)-meso-diaminopimelate-endopeptidase II from Bacillus sphaericus (23) and the endopeptidases LytE and LytF from B. subtilis (26, 29) (Fig. 2). An Asp-Cys-Ser-Gly motif, which is present in known endopeptidases belonging to the dl-endopeptidase II family (29), was found at positions 118 to 121.

FIG. 2.

Amino acid sequence alignment of the putative catalytic domains of YjgB of L. lactis IL-1403, dl-endopeptidase II of B. sphaericus (EPII), and the LytE and LytF endopeptidases of B. subtilis. The alignment was performed with the ClustalW 1.74 program. *, identical amino acids in all four sequences. The motif Asp-Cys-Ser-Gly, characteristic of the dl-endopeptidase II family and containing the putative Cys residue present in the catalytic site, is shaded. The numbers refer to amino acid sequence positions.

Since the IL-1403 genome contains several prophages whose induction can cause cellular lysis (11), we chose to study the PGHs in the prophage-cured strain L. lactis subsp. cremoris MG1363 (17). Using a PCR-based strategy with primers selected from the L. lactis subsp. lactis IL-1403 genome sequence, we cloned yjgB from MG1363. The MG1363 YjgB protein exhibits 94% sequence similarity with its IL-1403 counterpart. The molecular mass and pI are nearly the same for the MG1363 protein, with values of 21.0 kDa and 5.1, respectively, as for its IL-1403 counterpart. Although several pairs of primers located inside and outside of the open reading frames were tested, cloning of the MG1363 acmC and acmD genes by PCR failed.

Growth-phase-dependent expression of the five PGH-encoding genes.

The transcription of the different PGH genes in L. lactis MG1363 during cellular growth was analyzed by Northern blotting (Fig. 3). The total RNA from strain MG1363 was extracted at different stages of growth in CDM at 30°C. The membrane was hybridized successively with probes corresponding to each of the PGH genes. A single band was detected with each probe. The sizes of the detected bands (Fig. 3) correspond to monocistronic expression of each gene. These results indicate that all five genes are transcribed in L. lactis MG1363. This also demonstrates that the acmC and acmD genes are present and transcribed in the MG1363 genome, although they could not be cloned by PCR.

FIG. 3.

Growth phase-dependent expression of PGH genes. (A) Northern blot analysis of PGH gene expression. Total RNAs were prepared from L. lactis MG1363 cells grown in CDM to different OD₆₅₀ values and analyzed by Northern blotting with probes specific for each gene. The hybridization signals were imaged and quantified with a Storm system. The transcript amount was standardized by the amount of 16S rRNA in each sample. (B) Growth was monitored by OD₆₅₀ measurements.

A quantitative analysis of the transcripts revealed that their relative abundance varied during growth. For all five genes, the highest levels of transcript were found during the early exponential growth phase. Afterward, the relative variations were different between the different genes. The acmC and yjgB transcript levels declined rapidly during growth, and almost no transcript could be detected at the end of the exponential growth phase. The acmA and acmB transcript levels also declined, but at a lower rate, and transcripts were still present at the end of the exponential growth phase. Lastly, the acmD transcript level was the least variable during growth and remained relatively high during the late log phase.

AcmC, AcmD, and YjgB exhibit peptidoglycan-hydrolyzing activity.

The AcmC, AcmD, and YjgB proteins were overproduced in E. coli, with C-terminal six-His tags and with their putative signal peptides removed. Upon induction by IPTG, the proteins AcmC-His, AcmD-His, and YjgB-His were produced in the M15(pREP4) strain of E. coli. When induction was performed at 37°C, the AcmC-His and AcmD-His proteins were produced as insoluble inclusion bodies, whereas YjgB-His was found in the soluble fraction. When induction was performed at 28°C, the bulk of the AcmD-His protein was soluble, whereas AcmC-His remained insoluble. The recombinant proteins were purified by nickel-affinity chromatography performed under native conditions for AcmD-His and YjgB-His and under denaturing conditions with 8 M urea for AcmC-His. The purity of the recovered proteins was checked by PAGE (Fig. 4, lanes 1, 4, and 6). Their activities were examined by zymogram analysis, with autoclaved M. lysodeikticus cells used as a substrate. AcmC and AcmD gave a clear hydrolysis band by the classical renaturing SDS-PAGE technique (Fig. 4, lanes 2 and 5), whereas YjgB activity was detected only in a polyacrylamide gel run under native conditions (Fig. 4, lane 7). YjgB-His activity was not detected by renaturing SDS-PAGE, indicating an irreversible denaturation of the protein by SDS. AcmC-His activity was also detected with L. lactis cells (Fig. 4, lane 3) or cell walls, whereas the others were not detected under these conditions (data not shown).

FIG. 4.

PAGE and zymogram analysis of purified AcmC-His, AcmD-His, and YjgB-His. Electrophoresis was done in the presence (AcmC-His and AcmD-His) or absence (YjgB-His) of SDS. Lanes 1, 4, and 6, Coomassie blue staining; lanes 2, 5, and 7, gel containing 0.2% (wt/vol) M. lysodeikticus cells; lane 3, gel containing 0.4% (wt/vol) L. lactis cells. Renaturation and activity detection were performed at pH 4.0 (citrate phosphate buffer) or at pH 6.0 (50 mM MES buffer) with 1% (vol/vol) Triton X-100. The apparent molecular masses are indicated in kilodaltons.

Influence of pH on activity detection.

The influence of the incubation buffer pH on enzymatic activity detection with the M. lysodeikticus cell substrate for zymogram analysis was tested. Each of the three PGHs exhibited different patterns of activity as a function of pH. Activity for the purified AcmC-His protein was detected with the broadest pH range, from pHs 4.0 to 6.0, whereas AcmD-His activity was detected only at pH 4.0. YjgB-His activity was detected at pHs 3.0 and 4.0, but not at higher pH values and only when electrophoresis was run under native conditions.

From these results, it appears that the activities of the three PGHs are detected only at acidic pHs by zymography, as was also reported previously for AcmB (24). The recombinant proteins AcmC-His, AcmD-His, and YjgB-His have acidic or close to neutral pIs, with calculated values of 7.4, 4.5, and 6.1, respectively. Thus, it is remarkable that the activities of the three PGHs were detected only at pHs below their pIs, i.e., when the proteins would bear a positive global charge which could favor their interaction with the negatively charged cell walls.

AcmC, the third N-acetylglucosaminidase of L. lactis.

We examined the peptidoglycan hydrolytic bond specificities of the new enzymes on peptidoglycan extracted from B. subtilis vegetative cells. Each of the purified six-His-tagged proteins and a previously described purified recombinant N-terminally truncated AcmB protein (AcmB|[C+Z]-His) (24) were incubated overnight with the peptidoglycan substrate at 37°C in buffer at pH 4.0. The soluble fractions were recovered and analyzed by RP-HPLC. For AcmC, two major muropeptides (1 and 2) were detected (Fig. 5). These muropeptides had the same retention times as the two muropeptides obtained after the digestion of B. subtilis peptidoglycan by recombinant AcmB[C+Z]-His (data not shown). AcmB was previously shown to have N-acetylglucosaminidase activity, and the two muropeptides released from B. subtilis peptidoglycan were identified as disaccharide tripeptide and disaccharide tripeptide disaccharide tetrapeptide (24). Thus, muropeptides 1 and 2 produced by AcmC were attributed the structures shown in Fig. 5. A matrix-assisted laser desorption ionization-time of flight analysis of peak 1 gave a molecular ion with an m/z value of 892.12, which confirms its identification. In addition, the soluble muropeptides produced by AcmC could be digested by Cellosyl, which is a muramidase, confirming that, as expected, N-acetylglucosamine is present at the reducing end of the disaccharides of the muropeptides (data not shown). These results indicate that, similar to AcmA and AcmB, AcmC is an N-glucosaminidase.

FIG. 5.

RP-HPLC analysis of soluble muropeptides released by AcmC from B. subtilis vegetative peptidoglycan. The structures determined for the two main muropeptides (1 and 2) are indicated. MurNAc, N-acetylmuramic acid; GlcNAc, N-acetylglucosamine; mDAP, meso-diaminopimelic acid.

The cleavage sites of AcmD and YjgB could not be identified by the same approach, most probably because they have little activity towards B. subtilis peptidoglycan.

DISCUSSION

The PGH complement of L. lactis was determined on the basis of an amino acid sequence similarity search of the L. lactis genome. In addition to the four prophage-encoded endolysins (11), there are five enzymes, including the previously characterized AcmA (8) and AcmB proteins (24). In this study, we have shown that the three additional putative enzymes are functional. They are encoded by genes which are transcribed during growth in L. lactis MG1363 and they hydrolyze peptidoglycan.

The L. lactis PGH complement of five enzymes appears to be of low complexity compared to those of B. subtilis and E. coli, which contain 35 and 18 PGHs, respectively (20, 38). Also, the five L. lactis enzymes can be classified into only two families, as defined by sequence similarities at the level of their catalytic domains.

Four of the five described enzymes exhibit sequence similarity with the E. hirae Mur-2 catalytic domain (27). However, three of them, specifically AcmA (Steen et al., Abstr. Eurolab Conf.), AcmB (24), and AcmC (characterized in the present study), exhibit N-acetylglucosaminidase hydrolytic specificity (Fig. 6). A similar observation was also made previously for B. subtilis LytG (22). Since the specificity of the E. hirae enzyme was established by detailed biochemical studies (16), the enterococcal muramidase family defined by Smith et al. (38) on the basis of sequence similarity could contain glycosidases with the two types of specificity, either muramidase or glucosaminidase. The determination of the hydrolytic specificities of other enzymes belonging to the same family could help us to define the residues which determine hydrolytic specificity.

FIG. 6.

PGH complement of L. lactis. The schematic structure of the L. lactis peptidoglycan and the sites of cleavage by the different PGHs. MurNAc, N-acetylmuramic acid; GlcNAc, N-acetylglucosamine.

Regarding the other PGH types present in L. lactis, only one putative endopeptidase, YjgB, and no N-acetylmuramylamidase were identified (Fig. 6). The specificity of AcmD, whether it is a glucosaminidase or muramidase, was not determined. Since muramidase activity was previously detected in cell wall autolysis experiments (30), we hypothesize that AcmD is a muramidase. A further investigation of the YjgB and AcmD hydrolytic bond specificities with the L. lactis peptidoglycan substrate is required. Indeed, it is worth noting that the B. subtilis and L. lactis peptidoglycan structures have two main differences. Firstly, the third residue of peptidic side chains is l-Lys in L. lactis, whereas it is diaminopimelic acid in B. subtilis. Secondly, the peptidic side chains are directly cross-linked in B. subtilis, whereas they are connected by a d-Asp bridge in L. lactis (Fig. 6). These differences could account for the low levels of activity of AcmD and YjgB observed toward the B. subtilis peptidoglycan.

L. lactis contains different PGHs with the same specificity but different modular structures. In addition, they seem to differ in terms of their transcription patterns as well as the pH range of their activity. These results suggest that these enzymes could be involved in different cellular functions or could play a role at different stages during growth. AcmA was shown to be involved in cell separation after division (8). Recently, it was shown that the LysM domains of AcmA bind peptidoglycan and that the binding is hindered by other cell wall components, resulting in localized binding of AcmA (39). Other enzymes which lack a specific cell wall binding domain, such as AcmC and YjgB, could bind less specifically to the cell wall and be present on the whole cell surface. Therefore, they could be involved in other cellular functions, such as cell enlargement.

The need to control starter L. lactis lysis in cheese has stimulated numerous studies, especially the construction of inducible lysis systems consisting of a lysis gene under the control of an inducible promoter (35). In these systems, the use of acmA is somewhat restricted because of its high toxicity when overexpressed in L. lactis. The other L. lactis PGHs exhibit interesting characteristics as alternative candidates for expression under the control of an inducible promoter to induce L. lactis lysis. AcmC and the previously characterized enzyme AcmB are active in L. lactis cells but have apparent lower specific activities than AcmA, as observed by zymography. They are also active at acidic pHs, close to the pHs found in cheese (4.5 to 5.0). Thus, when overexpressed, they could be less toxic for the producing cells and stable and active in cheese to induce the lysis of other cells.

PGHs are considered to have essential activities for the growth that is required to make breaks in peptidoglycan for cell wall enlargement (20, 21, 36). However, this is not supported as yet by experimental evidence. Indeed, E. coli or B. subtilis mutants with deletions in several PGHs were obtained, but their growth was not impaired (19, 38). These results can be explained by the presence of multiple enzymes with redundant functions. In addition, the construction of a mutant completely devoid of PGH activity has been hampered by the large numbers of different PGHs present in these bacteria. Thus, the L. lactis autolytic system, comprising only five characteristic enzymes, appears to be a simpler model that will allow for the inactivation of all of the PGH genes and for investigations of the essential role of these enzymes for cell growth.