The molecular characterization of the first autolytic lysozyme of Streptococcus pneumoniae reveals evolutionary mobile domains


  • Pedro García,

    1. Departamento de Microbiología Molecular, Centro de Investigaciones Biológicas, Consejo Superior de Investigaciones Científicas, Velázquez 144, 28006 Madrid, Spain.
    Search for more papers by this author
  • María Paz González,

    1. Departamento de Microbiología Molecular, Centro de Investigaciones Biológicas, Consejo Superior de Investigaciones Científicas, Velázquez 144, 28006 Madrid, Spain.
    Search for more papers by this author
  • Ernesto García,

    1. Departamento de Microbiología Molecular, Centro de Investigaciones Biológicas, Consejo Superior de Investigaciones Científicas, Velázquez 144, 28006 Madrid, Spain.
    Search for more papers by this author
  • José Luis García,

    1. Departamento de Microbiología Molecular, Centro de Investigaciones Biológicas, Consejo Superior de Investigaciones Científicas, Velázquez 144, 28006 Madrid, Spain.
    Search for more papers by this author
  • Rubens López

    1. Departamento de Microbiología Molecular, Centro de Investigaciones Biológicas, Consejo Superior de Investigaciones Científicas, Velázquez 144, 28006 Madrid, Spain.
    Search for more papers by this author

Rubens López. E-mail; Tel. (+34) 91 561 1800; Fax (+34) 91 562 7518.


A biochemical approach to identify proteins with high affinity for choline-containing pneumococcal cell walls has allowed the localization, cloning and sequencing of a gene (lytC ) coding for a protein that degrades the cell walls of Streptococcus pneumoniae. The lytC gene is 1506 bp long and encodes a protein (LytC) of 501 amino acid residues with a predicted Mr of 58 682. LytC has a cleavable signal peptide, as demonstrated when the mature protein (about 55 kDa) was purified from S. pneumoniae. Biochemical analyses of the pure, mature protein proved that LytC is a lysozyme. Combined cell fractionation and Western blot analysis showed that the unprocessed, primary product of the lytC gene is located in the pneumococcal cytoplasm whereas the processed, active form of LytC is tightly bound to the cell envelope. In vivo experiments demonstrated that this lysozyme behaves as a pneumococcal autolytic enzyme at 30°C. The DNA region encoding the 253 C-terminal amino acid residues of LytC has been cloned and expressed in Escherichia coli. The truncated protein exhibits a low, but significant, choline-independent lysozyme activity, which suggests that this polypeptide adopts an active conformation. Self-alignment of the N-terminal part of the deduced amino acid sequence of LytC revealed the presence of 11 repeated motifs. These results strongly suggest that the lysozyme reported here has changed the general building plan characteristic of the choline-binding proteins of S. pneumoniae and its bacteriophages, i.e. the choline-binding domain and the catalytic domain are located, respectively, at the N-terminal and the C-terminal moieties of LytC. This work illustrates the natural versatility exhibited by the pneumococcal genes coding for choline-binding proteins to fuse separated catalytic and substrate-binding domains and create new and functional mature proteins.


Autolysins are enzymes that degrade the cell wall of microorganisms. Most bacteria contain several functional autolysins of different specificity, which indicates that these enzymes play fundamental biological roles. According to the chemical bond that these enzymes break down in the peptidoglycan substrate, they are classified as glycosidases (muramidases or lysozymes, glucosaminidases and transglycosylases), amidases and endopeptidases. The main lytic enzyme of Streptococcus pneumoniae has been identified as an N-acetylmuramoyl-l-alanine amidase (Mosser and Tomasz, 1970). Probably, this amidase (LytA) is the best characterized autolysin described so far, and its study represented the first molecular report on the cloning and expression of a bacterial autolysin gene (García et al., 1985). Interestingly, the pneumococcal amidase is also considered a virulence factor (Paton et al., 1997). Comparative and documented analysis of the pneumococcal amidase with the lytic enzymes coded by several pneumococcal phages led us to propose that the pneumococcal cell wall lytic enzymes could be the result of the fusion of two independent functional domains (García et al., 1988; 1990). Experimental support for this proposal required the construction of functional, chimeric phage–bacteria lytic enzymes (Díaz et al., 1990; 1991) to demonstrate that the carboxy-terminal domain (choline-binding domain; ChBD) is responsible for the specific recognition of choline-containing cell walls, whereas the catalytic centre of these enzymes is localized in the N-terminal part of the protein. The construction of chimeric lytic enzymes between phages and bacteria that belong to different Gram-positive families suggested the possibility of an evolutionary relationship between these enzymes (Croux et al., 1993; for a review, see López et al., 1997). Experimental evidence showing that this modular design is widespread among proteins has been reviewed recently (Doolittle, 1995).

In remarkable contrast to the LytA amidase, most of the bacterial peptidoglycan hydrolases are firmly bound to the cell envelope (Shockman and Höltje, 1994), which represents a limitation to the study of the physiological role(s) of these enzymes. The isolation of a mutant of pneumococcus (M31) deleted in the lytA gene (Sánchez-Puelles et al., 1986a) coding for the pneumococcal amidase allowed the detection of a new lytic activity that was characterized as an endo-β-N-acetylglucosaminidase (Sánchez-Puelles et al., 1986b; García et al., 1989). The precise biological role of the pneumococcal glucosaminidase remains to be determined, although we have provided experimental evidence suggesting that the bactericidal effect induced by β-lactam antibiotics on pneumococcus may be due to a cooperative effect of the amidase and the glucosaminidase (López et al., 1990). In a preliminary account, we have recently reported that the pneumococcal lytB gene codes for a murein hydrolase (LytB) that plays a fundamental role in pneumococcal cell separation (García et al., 1999). However, a detailed characterization of LytB remains elusive.

We have now cloned and characterized the lytC gene coding for the first pneumococcal lysozyme described (LytC). Remarkably, LytC (and LytB) exhibits a structural organization different from all the other choline-binding proteins described so far in the pneumococcal system. This finding provides insights into the plasticity of the genes coding for proteins with a modular design to switch and remodel DNA regions encompassing entire domains.


Identification of envelope proteins with lytic activity

Most bacterial species described so far contain several lytic enzymes. However, in spite of this fact, only a limited number of these enzymes have been purified because of their presence in small amounts and/or because of their high-affinity binding to the cell wall. Using a procedure previously published to characterize the pneumococcal glucosaminidase (García et al., 1989), we identified, by SDS–PAGE, four protein bands with apparently strong choline-binding affinity (Fig. 1). Residual murein hydrolase activity was detected when the major protein band (Mr of about 55 000) was excised and eluted from the gel, renatured and assayed on choline-labelled cell walls (data not shown). The N-terminal amino acid sequence of this protein was determined yielding Asn–Glu–Thr–Glu–Val–Ala–Lys–Thr–Ser–Gln. This sequence was compared with the translated version of the partial nucleotide sequence of the genome of a pneumococcal strain that has been recently released (see below), and a perfect match with an internal peptide of a gene product was found. This gene (hereafter designated lytC ) was located in the 3832 bp contig no. 4270. Analysis of this contig (Fig. 2A) revealed that the lytC gene is preceded by an open reading frame, whose product showed strong similarity (66% identity, 78% similarity) to the Tpi triosephosphate isomerase of Lactobacillus delbrueckii (accession no. AJ000339). Other genes further upstream of tpilytC were also preliminarily identified on the basis of sequence similarities: the first putative gene product of the contig, albeit incomplete, is 48% identical (66% similar) to the homoserine O-succinyltransferase (MetA) of Escherichia coli (accession no. U00006), whereas the gene immediately preceding tpi appears to code for DnaD (58% identity, 76% similarity), a protein involved in the initiation of DNA replication in Streptococcus mutans (accession no. D78182) (Fig. 2A).

Figure 1.

. Analysis by SDS–PAGE (10%) of choline-binding proteins prepared from the cell envelope of S. pneumoniae. Lane 1, crude extract obtained from M31; lane 2, supernatant obtained after centrifugation of crude M31 extract; lane 3, autolytic walls; lane 4, supernatant obtained after ultracentrifugation of autolysed walls; lane 5, purified choline-binding proteins. The molecular mass of the protein standards is indicated on the left. Arrows indicate the protein bands corresponding to LytB and LytC.

Figure 2.

. Genetic organization of contig no. 4270 from the S. pneumoniae chromosome containing the lytC gene (A) and amino acid sequence alignment of LytC regions (B and C). A. Arrows represent putative genes. The black and the white rectangles in lytC correspond to the DNA regions coding for the putative signal peptide and the choline-binding motifs respectively. bsl00048 represents a putative transcription terminator (ΔG = −30.6 kcal mol−1). The location of the oligonucleotides (LYTC-HIN, ▸; LYTC-HINR, ◃ ) used to amplify lytC is also shown. The location of the oligonucleotides used to construct pLCC14 is also indicated (LYTC-N2, bsl00049; LYTC-C, bsl00050). Abbreviations: B, BamHI; H, HindIII; E, Ecl 136II; N, NdeI. The nucleotide sequence of the lytC gene has been submitted to the GenBank/EBI DataBank with accession number AJ009639. B. Self-alignment of the N-terminal part of the deduced sequence of LytC. The consensus motif was deduced by aligning the ChBDs analysed by García et al. (1998) and those reported recently: CbpA (accession no. O30874); PbcA (O69188); PspC (G4097980); SigA (O33753), and LytB (E1374791). When at least 50% of the sequences coincide, the corresponding amino acids are inserted in a black box, whereas amino acid residues identical to those of the consensus sequence are shaded. C. C-terminal domain of LytC. Vertical arrows mark the catalytic amino acids of Cpl1 muramidase (accession no. P15057). The deduced amino acid sequence of the lysin of the Lactobacillus phage φg1e (O03950) is also compared. Asterisks and colons indicate identical amino acid residues and conservative substitutions respectively.

PCR amplification and sequencing of the lytC gene

A 1.6 kb DNA fragment embracing lytC was PCR amplified using oligonucleotides LYTC-HIN and LYTC-HINR and M31 DNA as template. Comparison of the 1626 bp nucleotide sequence from the PCR-amplified fragment with that reported for the contig no. 4270 revealed 16 differences, 15 of which corresponded to A/T or G/C transitions, whereas an additional nucleotide (G) should be introduced after position 1797 in the reported sequence of the contig 4270. This sequence correction allows the lytC gene to be translated properly from an ATG initiation codon (positions 27–29) instead of a GTG one, as originally predicted. We do not know whether these differences are due to gene polymorphism or sequencing errors. The lytC gene (1506 nucleotides) putatively encodes a protein of 501 amino acid residues with a predicted molecular mass of 58 682 Da. The first 33 amino acid residues appear to correspond to a signal peptide, in which the cleavage site is probably located in the sequence Val–Ala–Ala ↓ Asn–Glu–Thr according to criteria previously described (von Heijne, 1988). This assumption was in perfect agreement with the N-terminal amino acid sequence previously determined (see above), which demonstrates that LytC contains a signal peptide. It should be mentioned that a cleavable signal peptide has not been found in the main pneumococcal lytic enzyme LytA (Díaz et al., 1989). The mature form of LytC (468 amino acid residues) has a predicted Mr of 55 210. A schematic representation of the primary structure of LytC is shown in 2Fig. 2A. Self-alignment of the N-terminal part of the deduced sequence of LytC revealed the presence of 11 repeat sequences (motifs), showing a significant similarity with a consensus motif present in all pneumococcal proteins containing a ChBD (Fig. 2B). Interestingly, the ChBD of LytC was located close to the signal sequence, whereas its catalytic domain was located in the C-terminal moiety of the protein. The search for similarities of this C-terminal domain of LytC to those available in the sequence databases showed that LytC is similar to several lysins, in particular to the catalytic domain of the Cpl1 muramidase (García et al., 1988) and the N-terminal part of the Lactobacillus phage φg1e lysin (Oki et al., 1996) (Fig. 2C). It should be emphasized that Asp-10, which appears to be essential for the enzymatic activity of Cpl1 (Sanz et al., 1992a) and other muramidases of the Chalara family (Fastrez, 1996), is also conserved in LytC. On the contrary, Glu-37, which is less critical for enzymatic activity (Sanz et al., 1992a), is not conserved in the pneumococcal LytC lysozyme (Fig. 2C).

Cloning and expression of lytC and biochemical characterization of the gene product

Sonicated extracts of E. coli cells harbouring the recombinant plasmid pLCC14 and induced with IPTG were able to degrade choline-labelled pneumococcal cell walls (see below). Plasmid pLCC14 contains the part of the lytC gene that codes for the mature (processed) protein, under the control of the φ10 promoter of pT7-7 (see Experimental procedures ). On the contrary, repeated attempts to clone the complete lytC gene were unsuccessful, suggesting that the signal peptide might be processed in E. coli and the mature protein, then located at the outer side of the cytoplasmic membrane, might degrade the bacterial peptidoglycan and cause the lysis of the culture, as previously reported for the Ej1 amidase (Díaz et al., 1996) and the Cpl1 lysozyme (Martín et al., 1998) of the S. pneumoniae phages EJ-1 and Cp-1 respectively.

Induced extracts prepared from E. coli cells carrying pLCC14 showed an overproduced protein band of about 55 kDa when analysed by SDS–PAGE (Fig. 3). This protein was purified to electrophoretic homogeneity by affinity chromatography on DEAE cellulose (Sánchez-Puelles et al., 1992). N-terminal analysis of the purified protein (not shown) confirmed that it corresponded to the mature form of LytC. Analysis of the degradation products of pneumococcal cell walls on Sephadex G-75 (not shown) suggested that LytC was a glycosidase. To determine the nature of this glycosidase, samples of untreated and LytC-treated walls were reduced with NaB3H4 to label the free reducing groups present. The reduced samples were hydrolysed, and after chromatography on Dowex 50W-X4 columns, the radioactivity was mainly found as [3H]-muramitol (not shown), demonstrating that the LytC is a muramidase.

Figure 3.

. SDS–PAGE (12.5%) analysis of the purified LytC enzyme. Lane A, crude soluble extract prepared from E. coli BL21(DE3) (pT7-7); lane B, crude soluble extract from E. coli BL21(DE3) (pLCC14); lane C, purified LytC lysozyme obtained by affinity chromatography on DEAE–cellulose. The molecular mass of the markers, in kDa, is indicated on the left.

Table 1 shows some biochemical properties of the LytC lysozyme compared with those of the LytA amidase and the phage Cpl1 lysozyme. The activity of LytC was maximal at pH 6.0 in phosphate buffer. The highest reaction rate was achieved at 30°C, a property shared with the pneumococcal glucosaminidase previously described (García et al., 1989). The LytC enzyme is also choline-dependent for activity, as it was unable to degrade efficiently ethanolamine-containing cell walls (data not shown). The choline concentration needed to inhibit LytC is much higher than that of Cpl1 and similar to that of LytA.

Table 1. . Comparison of the biochemical properties of the purified LytC, LytA and Cpl1 enzymes. a. Molecular masses were calculated from the deduced amino acid sequences.b. Assays for enzymatic conversion (activation) by choline-containing cell walls of the low activity form of the lytic enzymes were performed as previously described (Tomasz and Westphal, 1971) using crude extracts prepared from E. coli transformants synthesizing the corresponding enzyme. One unit of enzymatic activity is defined as the amount of enzyme that catalyses the hydrolysis (solubilization) of 1 μg of cell wall material in 10 min (Mosser and Tomasz, 1970).c. Assays were carried out with purified enzymes. Values are the average of three independent experiments.d. The value reported in brackets corresponds to the molecular mass of the unprocessed form.Thumbnail image of

The sequence comparison discussed above (Fig. 2C) strongly suggested that the C-terminal moiety of LytC (C-LytC) is the catalytic domain. To test this assumption, E. coli BL21(DE3) cells harbouring pLCC20 that codes for the 253 C-terminal amino acid residues of the lysozyme (see Experimental procedures ) were incubated in Luria–Bertani medium and induced with IPTG, as described above for the preparation of LytC. SDS–PAGE analysis of crude extracts prepared from these E. coli cells showed a 30 kDa protein that corresponds to C-LytC, as demonstrated by Western blot analysis using anti-LytC serum. When these extracts were assayed for lytic activity using choline-, lysine- or ethanolamine-labelled pneumococcal cell walls, a low but significant lytic activity was detected (data not shown). Although the enzymatic activity of crude cell extracts containing C-LytC on choline-labelled cell walls was about 103-fold lower than those having the LytC lysozyme, it was significant enough to conclude that the truncated protein was an active enzyme. It is also important to point out that the high sensitivity of the radioactive assay used to determine the hydrolytic activity of the pneumococcal enzymes allowed the precise detection of low enzymatic activity. Moreover, analysis of the degradation products showed that the truncated protein retained enzyme specificity (data not shown).

On the other hand, the choline-dependent enzymes are unable to degrade efficiently pneumococcal cell walls where choline has been replaced by the structural analogue ethanolamine (Tomasz and Westphal, 1971). However, and in sharp contrast to the behaviour of LytC (Table 1), its C-terminal moiety was capable of degrading with similar efficiency both choline- and ethanolamine-labelled cell walls, indicating that the truncated protein had lost its characteristic choline dependence (not shown).

Subcellular localization of the LytC lysozyme

The precise subcellular distribution of the unprocessed and mature forms of the LytC lysozyme has been studied by Western blot analysis of different cellular fractions of the pneumococcal M31 strain using an antiserum raised against LytC. Necessarily, the murein hydrolases have to be transported across the cytoplasm membrane in order to reach their peptidoglycan substrate. Cell fractions prepared from M31 strain revealed the presence of the unprocessed form of LytC in the cytoplasm (Fig. 4B). Interestingly, the mature form of LytC was localized, exclusively, tightly bound to the cell envelope. These results are in agreement with the observation reported above showing that crude extracts prepared from the M31 envelope degrade pneumococcal cell walls.

Figure 4.

. Construction and biological behaviour of a lytC pneumococcal mutant. A. Schematic representation of the DNA region surrounding the lytC gene in the insertion duplication mutant strain M31C (lytC ) and its parental strain M31. A hatched arrow represents the lytC gene. The white rectangle in lytC represents the Ecl 136II DNA insert of pMPG421 used for insertion duplication mutagenesis. The thick line corresponds to the vector plasmid pUCE191 used to construct pMPG421. Abbreviations: E, EcoRI; bla and ermC, genes encoding ampicillin and erythromycin/lincomycin resistance respectively. B. Left, Southern blot hybridization using lyt C as probe. Lane 1, EcoRI-digested M31 DNA; lane 2, EcoRI-digested M31C DNA. The size (in bp) of the DNA fragments is shown on the right. Right, subcellular localization of LytC lysozyme by Western blot analysis using anti-LytC serum. Envelope protein (lanes 1 and 2) and cytoplasmic extracts (lanes 3 and 4) were prepared from M31 (lanes 1 and 3) or M31C (lanes 2 and 4). Numbers on the right indicate the molecular mass (in kDa) of the corresponding bands. C. Sensitization of M31C strain to exogenously added LytC lysozyme. Exponentially growing cultures of M31C received purified LytC (13 Units ml−1) at N = 200 (▴). The growth curves of untreated cultures of M31C (○) and M31 (●) are also shown. Cultures were incubated at 30°C and cellular density was followed by nephelometry (N). One N corresponds to 7.5–105 cfu ml−1.

Biological role of LytC

The growth characteristics exhibited by the S. pneumoniae lytA mutants demonstrated that the pneumococcal amidase is not required for growth (Sánchez-Puelles et al., 1986a; Berry et al., 1989). The LytA enzyme is, however, responsible for the lytic phenotype of pneumococcus either in the stationary phase of growth or upon addition of detergents or β-lactam antibiotics (Sánchez-Puelles et al., 1986a; Ronda et al., 1987).

To get insight into the biological role of the LytC lysozyme, we constructed two lytC pneumococcal strains by insertion duplication mutagenesis using plasmid pMPG421 (Fig. 4A). Lincomycin-resistant transformants of either R6 or M31 were used for further study. The accuracy of the constructions was tested by Southern and Western blot analysis, and 4Fig. 4B shows the results for the M31 and M31 lytC (M31C, hereafter) strains. Identical results were achieved when the wild-type strain R6 was used (data not shown). The high specificity of the anti-LytC serum was demonstrated by the observation that only LytC was recognized in extracts prepared from any of these strains (Fig. 4B; data not shown). The lytC mutants, either derived from R6 or M31, exhibited a normal growth rate and an average chain length similar to that of the parental strains (data not shown). Moreover, the double mutant strain M31C was normally transformable, which discards a role for these lytic enzymes in competence development as previously demonstrated for the pneumococcal LytA amidase alone (Sánchez-Puelles et al., 1986a). Most interestingly, the M31C strain, in contrast to the parental M31, did not autolyse upon incubation at 30°C (Fig. 4C), although a residual cell wall-degrading activity that might be ascribed to the glucosaminidase previously described (García et al., 1989) was still present (data not shown). Furthermore, addition of purified LytC lysozyme to a culture of M31C restored the autolytic capacity of the cells in the stationary phase when incubated at 30°C, as previously demonstrated for the LytA amidase (Tomasz and Waks, 1975). These findings demonstrated that LytC behaves as an autolytic enzyme at this temperature.


A two-domain structural organization of lytic enzymes in which the N-terminal portion corresponds to its enzymatic activity while the C-terminal region is responsible for its substrate-binding specificity has been proposed for pneumococcus and its bacteriophages (López et al., 1997), the clostridial Lyc muramidase (Croux et al., 1993), and the lysins of several phages, e.g. that of Tuc2009 of Lactococcus lactis (Arendt et al., 1994), φg1e of Lactobacillus (Oki et al., 1996) and, more recently, in phages infecting Streptococcus thermophilus (Sheehan et al., 1999). CwlB (= LytC) (Kuroda and Sekiguchi, 1991), LytE (Margot et al., 1998), and PaqQ (= CwlF) (Ishikawa et al., 1998) are lytic enzymes of Bacillus subtilis that exhibit repeats in their N-terminal region. CwlG (= LytD), the major glucosaminidase of B. subtilis, also shows two types of repeats in the N-terminal part of the protein (Rashid et al., 1995). A direct role of the repeats found in wall binding has been documented only for the CwlB amidase and the CwlG glucosaminidase. On the other hand, P60, a murein hydrolase of Listeria monocytogenes (Wuenscher et al., 1993), exhibits four tandemly arranged units that are supposed to be involved in wall binding because they have extended sequence homology with repeat motifs suggested as possible substrate binding sites for an autolysin of Streptococcus faecalis (Béliveau et al., 1991). In Listeria and S. faecalis, these binding regions are located in the C-terminal part of the lytic proteins.

A peculiar case of domainal organization is provided by a Staphylococcus aureus autolysin. The atl gene of S. aureus codes for a bifunctional protein that has an N-terminal amidase domain and a C-terminal endo-β-N-acetylglucosaminidase domain, which must undergo proteolytic processing to generate the two extracellular lytic enzymes. It was suggested that the three repeat sequences located between the two catalytic regions were the binding domain (Oshida et al., 1995). More recently, it has been shown that the binding domain of the primary product targets the pro-Atl to the cell surface before its proteolytic processing (Baba and Schneewind, 1998). Therewith, pro-Atl is cleaved and there is a repeat located on the N-terminal side of glucosaminidase and two repeats found on the C-terminal side of the amidase. All three species, pro-Atl, amidase and glucosaminidase, remain largely bound to the cell surface. The repeats were each sufficient to direct reporter proteins to the equatorial surface ring, indicating that they represent the binding domain. It has been pointed out that these three repeat sequences (each 150 amino acids long) behave in targeting like the six-repeat domain (six motifs of 20 amino acids each) at the C-terminal end of S. pneumoniae LytA (Baba and Schneewind, 1998). These long repeats appear to be capable of forming a structural block possessing an established folding, as suggested by the facility to prepare functional fused proteins (Baba and Schneewind, 1998).

The above examples taken together illustrate that either N-terminal or C-terminal cell wall binding domains are common among Gram-positive bacterial cell wall hydrolases, although, to our knowledge, these alternative situations have not been reported in the same microorganism. It is also noteworthy to recall that a domain has been defined as part of a protein that can fold up independently of neighbouring sequences. However, this term is sometimes used less precisely because sequence-oriented investigators often use the term ‘domain’ when they mean sequence motif or signature sequence, although such sequences are frequently characteristic of domains (Doolittle, 1995). Interestingly, Doolittle (1995) has proposed that a domain needs only to occur in two different settings in otherwise non-homologous proteins, or at clearly different locations in homologous proteins, to be considered as evolutionary mobile.

Construction of functional chimeric phage–bacteria lytic enzymes in pneumococcus has given strong support to our observations on the modular design of these enzymes (Díaz et al., 1990; 1991). These findings show that interchanges of full modules between host and phage genes by a simple recombinational event are quite plausible as a rapid mechanism to gain new biological properties. The preparation of functional chimeric enzymes by exchanging modules between the genes coding for lytic enzymes of pneumococcus (or its phages) and those from Clostridium or Lactococcus (or their phages) enabled us to postulate that modular interchange is also possible between different genera (Croux et al., 1993; Arendt et al., 1994). We have recently characterized the lytic enzyme (Pal) of the pneumococcal phage Dp-1, an amidase that appears to be a natural chimeric enzyme of intergeneric origin between pneumococcus and a lactococcal bacteriophage (Sheehan et al., 1997). The LytC lysozyme described in this work illustrates an example of formation of genes by combining the fusion of distinct genetic modules in a disposition different from that previously reported, indicating that, in the case of lytC, the pneumococcal system has changed the general building plan for the construction of their lytic enzymes (López et al., 1997). The organization of LytC indicates that the recruitment and genetic adaptation of the cassettes to achieve catalytic efficiency is extremely versatile in the pneumococcal system. It has been proposed that the bacterial chromosome must be a mosaic of phylogenetic histories even within genes, as in the case reported here. Units of modular exchange, smaller than the gene, in the evolution of the lytic enzymes of the pneumococcal system have been reported (López et al., 1997) and a very similar situation was recently documented for S. thermophilus phage genomes (Neve et al., 1998). These interchanges of pre-existing structural blocks by a simple recombinational mechanism implicate, as previously proposed, that the modules possess an established folding design to facilitate the formation of co-operative structures (Privalov, 1989). We have shown here that the C-terminal moiety of LytC behaves as a true domain because it was able to acquire an active conformation when synthesized in E. coli, confirming that this part of the molecule is the catalytic domain of the enzyme. Moreover, the truncated protein had lost the choline dependence characteristic of this class of lytic enzymes. This was not unexpected because murein hydrolases that do not contain the ChBD, or that lack at least four of the six repeat units of the ChBD, degrade either choline- or ethanolamine-containing pneumococcal cell walls with similar efficiency (Sanz et al., 1992b; García et al., 1994). It has been also demonstrated that the ChBD and the catalytic domain of other pneumococcal lytic enzymes can function independently of each other (Sánchez-Puelles et al., 1990; Sanz et al., 1992b; Usobiaga et al., 1996). Furthermore, the feasibility of the ChBD to become functional when it is located either at the N- or C-terminal ends of a protein had been anticipated through the preparation of fusion proteins containing this binding domain at the N-terminal position (Sánchez-Puelles et al., 1992).

Binding of many of the hydrolases of different bacterial species to peptidoglycan has been found to be quite strong (Shockman and Höltje, 1994). This characteristic has frequently represented a handicap to the study of the biochemical and biological properties of these enzymes and has been suggested as a mechanism for the regulation and control of the potentially suicidal activity of the peptidoglycan hydrolases. It has been pointed out that a clue to understanding these mechanisms might be gained by determining the precise subcellular distribution of the murein hydrolases (Shockman and Höltje, 1994), as carried out in this work for LytC. These enzymes have to be transported across the cytoplasmic membrane to reach the peptidoglycan substrate. In the LytC lysozyme, the leader peptide directs its translocation to the outer surface where it remains tightly bound in a mature, active form (Fig. 4B). The ChBD of LytC has more recognizing motifs (up to 11 units) than LytA and the pneumococcal phage-associated lysins previously characterized. As suggested for the pneumococcal PspA protein (Yother and White, 1994), the binding strength of the LytC lysozyme to the substrate could be directly related to the number of repeat motifs.

We have provided biochemical evidence on the presence of a second lytic enzyme, characterized as a glucosaminidase, in the amidase-deficient strain M31 (Sánchez-Puelles et al., 1986b; García et al., 1989), which complicates the precise determination of possible biological significance and function(s) of newly detected pneumococcal cell wall hydrolases. Actually, it is difficult to postulate a mechanism through which the peptidoglycan network could be enlarged and divided without the contribution of the lytic enzymes (Shockman and Höltje, 1994). On the other hand, these hydrolytic enzymes can kill the cell by autolysis and it has been well documented in the literature that the bacteriolytic action of penicillin and other cell wall inhibitors is due to the triggering of lytic enzymes (Tomasz, 1979). Theoretically, the consideration of the lytic enzymes as space-making enzymes should make these enzymes an ideal target for antibiotics. The multiplicity of lytic enzymes in bacteria makes it a difficult task to provide experimental evidence to support this hypothesis. However, the molecular approach to the study of all the peptidoglycan hydrolases present in pathogenic organisms such as pneumococcus should address these kind of biological problems. We have now found that insertional inactivation of the lytC gene in strain M31C makes the cells resistant to autolysis upon incubation at 30°C. Addition of purified LytC lysozyme to this deficient strain (phenotypic curing) showed that this enzyme was kept under regulatory control of the ‘cured’ cell until the culture reached the stationary phase of growth and then the culture started to lyse (Fig. 4C). These findings demonstrate that LytC can induce the lysis of pneumococcus when incubated at 30°C, whereas the role of the glucosaminidase previously identified biochemically (García et al., 1989) remains to be investigated by using a genetic and molecular approach similar to the one described here. S. pneumoniae is an exclusive human pathogen that is carried in the nasopharynx of up to one-third of the adult population (Salyers and Whitt, 1994). Transformation of chromosomal genes has been shown to occur in vivo between strains of S. pneumoniae and between S. pneumoniae and viridans streptococci (Ottolenghi-Nightingale, 1972). It is tempting to speculate that lytic enzymes, such as LytC, that behave as autolysins at 30°C might contribute to the release of transforming DNA in environments such as the upper respiratory tract that are highly ventilated.

Recently, we have found that LytB, which might be the pneumococcal glucosaminidase previously studied (García et al., 1989), plays a fundamental role in the terminal step that leads to daughter cell separation in S. pneumoniae (García et al., 1999). Interestingly, this murein hydrolase exhibits a structural organization similar to the LytC lysozyme described in this work. The data reported here together with previous findings (López et al., 1997) show that we are moving towards producing, in the near future, a comprehensive picture on the lytic enzymes of the pneumococcal system.

Experimental procedures

Bacterial strains, plasmids, and growth conditions

E. coli BL21(DE3) (Studier and Moffatt, 1986) served as the host for the expression of lyt C from pLCC14 (see below). E. coli DH5α (pGL100) (García et al., 1987) and E. coli DH5α (pCIP100) (Sanz and García, 1990) were used as sources, respectively, of LytA and Cpl1, a lysozyme encoded by the S. pneumoniae phage Cp-1. Plasmid pT7-7 has been described elsewhere (Tabor, 1990). The S. pneumoniae strains used were the wild-type strain R6 and the autolysin-deficient mutant M31 (Sánchez-Puelles et al., 1986a). E. coli was grown in Luria–Bertani medium. S. pneumoniae was grown in C medium (Lacks and Hotchkiss, 1960) supplemented with yeast extract (0.8 mg ml−1; Difco Laboratories) (C+Y medium). Transformation of S. pneumoniae (Sánchez-Puelles et al., 1986a) and E. coli (Sambrook et al., 1989) has been described previously.

Preparation of choline-binding proteins from autolytic walls

To purify the choline-binding proteins bound to the cell envelope, we followed the procedure described by García et al. (1989). In short, an exponential culture of the M31 strain was disrupted in a French Pressure Cell Press (American Inst.), centrifuged and the pellet washed twice in 50 mM potassium phosphate buffer (pH 6.5) containing 10 mM MgCl2 and 0.1% Brij-58. After centrifugation, the pellet was resuspended in the same buffer containing 0.1% choline chloride (‘autolytic walls’) and allowed to autolyse overnight during incubation at 30°C. Afterwards, the concentration of choline chloride was raised to 2%, and, after incubation at 0°C for 15 min, the mixture was ultracentrifuged (120 000g, 1 h, 4°C). The supernatant was dialysed against 50 mM potassium phosphate buffer (pH 6.5), applied to a DEAE–cellulose column and washed with 1.5 M NaCl. Choline-binding proteins were eluted from the column upon addition of 2% choline chloride (Sanz et al., 1988).

Cloning, nucleotide sequencing, computer analyses and plasmid construction

Routine DNA manipulations were performed essentially as described previously (Sambrook et al., 1989). DNA fragments were purified by using the Geneclean II kit (Bio 101). The relevant oligonucleotide primers used were: LYTC-N2 (1857) (5′-TCATGTCcaTatgAATGAAACTGAAGTAGCAAAAAC-3′), LYTC-C (3287/c) (5′-CcggatcctcaTTAATACCAAACGCTGCATC-3′), LYTC-HIN (1693) (5′-GAAGCTGAAAGCTTCTTGGCTTTGC-3′), and LYTC-HINR (3388/c) 5′-CTACTATA-TAAGCTTCATTTCCGAGC-3′). The numbers in parentheses indicate the positions of the first nucleotide of the primer in the sequence corresponding to contig no. 4270, and /c indicates that the sequence corresponds to the complementary strand of the same contig that is included in the TIGR S. pneumoniae genome database ( html). Lowercase letters indicate nucleotides introduced to construct the appropriate restriction sites, which are shown underlined. DNA sequence was determined by the dideoxy chain-termination method (Sanger et al., 1977) with an automated Abi Prism 377 DNA sequencer (Applied Biosystems). All primers for PCR amplification and nucleotide sequencing were synthesized on a Beckman model Oligo 1000M synthesizer.

Amino acid sequence was analysed with the protein analysis tool at the World Wide Web molecular biology server of the Geneva University Hospital and the University of Geneva. Protein sequence similarity searches were carried out with the blastp program via the National Institute for Biotechnology Information server. Pairwise and multiple protein sequence alignment were carried out with the align and clustalw programs, respectively, at the Baylor College of Medicine Human Genome Center server (

To construct pLCC14, we amplified, by PCR, an M31 DNA fragment using oligonucleotides LYTC-N2 and LYTC-C. The resulting fragment was digested with NdeI and BamHI, and ligated to pT7-7 previously treated with the same enzymes. The recombinant plasmid pLCC14 was selected among the ampicillin-resistant transformants of E. coli BL21(DE3). Plasmid pMPG421 was constructed by cloning a 421 bp Ecl 136II internal fragment of lytC into SmaI-digested pUCE191, a non-replicative plasmid in S. pneumoniae (Arrecubieta et al., 1995). Plasmid pLCC20 containing the gene region encoding the C-terminal part of LytC was constructed as follows: pLCC14 was digested with NdeI and NcoI, filled in with the Klenow fragment of the E. coli DNA polymerase, and self-ligated. The ligation mixture was used to transform E. coli BL21(DE3). The accuracy of the construction was checked by restriction analysis.

Purification of the mature LytC lysozyme

E. coli BL21(DE3) (pLCC14) strain was incubated in Luria–Bertani medium containing ampicillin (100 μg ml−1) up to an OD600 of 0.5. At this time, isopropyl-β-d-thio-galactopyranoside (40 nM) was added, and incubation continued for 3 h at 22°C to minimize the presence of inclusion bodies. The culture was centrifuged (10 000g, 5 min) and the pelleted bacteria were resuspended in 20 mM sodium phosphate buffer (pH 6.0) and disrupted in a French Pressure Cell Press. The insoluble fraction was separated by centrifugation (15 000 g, 15 min) and the supernatant was loaded into a DEAE–cellulose column to purify, in a single step, the LytC lysozyme following a procedure previously described (Sánchez-Puelles et al., 1992).

Miscellaneous methods

A polyclonal antiserum against LytC was raised in rabbits as described previously (García et al., 1982) and was used at a 1:1000 dilution for Western blot analyses according to a previously published procedure (Sambrook et al., 1989). SDS–PAGE was carried out with the buffer system described by Laemmli (1970) in 10% or 12.5% polyacrylamide gels, and protein bands were visualized by staining with Coomassie brilliant blue R250. DNA probes were labelled with the DIG Luminescent Detection Kit (Boehringer Mannheim). Southern blots and hybridizations were carried out according to the manufacturer's instructions. Pneumococcal cell walls were radioactively labelled with [methyl-3H]-choline, [3H]-lysine, or [14C]-ethanolamine, as described previously (Mosser and Tomasz, 1970). The characterization of the free reducing groups liberated by the lytic enzyme studied here was carried out by the procedure described by Ward (1973). Preparation of subcellular fractions was carried out as described by García et al. (1989). N-terminal amino acid sequence analyses were carried out according to a published procedure (Speicher, 1994).


We thank E. Cano and M. Carrasco for their technical assistance, and V. Muñoz and M. Fontenla for the artwork. This work was supported by grant PB96–0809 from the Dirección General de Investigación Científica y Técnica.