We have developed a polypeptide lysostaphin FRET (fluorescence resonance energy transfer) substrate (MV11F) for the endopeptidase activity of lysostaphin. Site-directed mutants of lysostaphin that abolished the killing activity against Staphylococcus aureus also completely inhibited the endopeptidase activity against the MV11 FRET substrate. Lysostaphin-producing staphylococci are resistant to killing by lysostaphin through incorporation of serine residues at positions 3 and 5 of the pentaglycine cross-bridge in their cell walls. The MV11 FRET substrate was engineered to introduce a serine residue at each of four positions of the pentaglycine target site and it was found that only a serine residue at position 3 completely inhibited cleavage. The introduction of random, natural amino acid substitutions at position 3 of the pentaglycine target site demonstrated that only a glycine residue at this position was compatible with lysostaphin cleavage of the MV11 FRET substrate. A second series of polypeptide substrates (decoys) was developed with the GFP (green fluorescent protein) domain of MV11 replaced with that of the DNase domain of colicin E9. Using a competition FRET assay, the lysostaphin endopeptidase was shown to bind to a decoy peptide containing a GGSGG cleavage site. The MV11 substrate provides a valuable system to facilitate structure/function studies of the endopeptidase activity of lysostaphin and its orthologues.
- colicin E9
- fluorescence resonance energy transfer assay (FRET assay)
- Staphylococcus aureus
Lysostaphin is an antimicrobial protease that is often incorrectly termed a peptide; it is produced by Staphylococcus simulans biovar staphylolyticus and is active against Staphylococcus aureus strains . It is encoded by the pACK1 plasmid  and is synthesized as a 493-residue pre-proenzyme that consists of a 36-residue signal peptide and 15 tandem repeats of 13 amino acids, 14 of which have the identical sequence AEVETSKAPVENT. Removal of the signal peptide and the tandem repeats results in the mature enzyme of 247 amino acids  that consists of an N-terminal 132-residue catalytic domain, followed by a 13-residue linker that is attached to the 102-residue targeting domain. The signal peptide is cleaved during protein export from the bacterial cytoplasm, and the 15 tandem repeats are removed by a cysteine protease that is also secreted into the culture medium . It has been demonstrated that the target specificity of the mature lysostaphin is mediated by a 92-residue C-terminal targeting domain that is also required for rapid processing of the proenzyme form of lysostaphin to the mature form . Lysostaphin is a member of the M37/M23 zinc endopeptidase family that cleaves the pentaglycine cross-bridges  that are found in the cell wall of S. aureus, S. simulans and Staphylococcus carnosus . It has been suggested that lysostaphin cleaves specifically between the third and fourth glycine residues of the pentaglycine cross-bridge based on the release of staphylococcal surface proteins by lysostaphin . However, analysis of lysostaphin-cleaved muropeptides by MS revealed that cleavage occurs between glycine residues 2 and 3, 3 and 4 and 4 and 5, with a preference for the latter cleavage site .
The potential therapeutic use of lysostaphin has been investigated in several animal models [9,10]. Despite encouraging results, lysostaphin was not developed for clinical use due to (i) the difficulty in producing large amounts of pure preparations of lysostaphin; (b) the availability of other effective antibiotics; and (iii) possible concern over the potential immunogenicity of the protease . The continued rise in the problem of multiple antibiotic resistance in pathogenic bacteria, especially in MRSA (methicillin-resistant Staphylococcus aureus) isolates, has led to the search for alternative therapeutic agents and renewed interest in lysostaphin [11–14], especially as it disrupts biofilms of S. aureus  that are normally difficult to treat with antibiotics. Lysostaphin-coated catheters have also been shown to prevent catheter colonization by several strains of S. aureus, and activity was maintained for at least 4 days . Transgenic cows secreting lysostaphin in their milk were shown to be protected against S. aureus mastitis, which was responsible for up to 30% of mastitis infection in cattle at a cost of US$600 million annually in 2005 .
Despite almost 40 years of research on lysostaphin, there are still considerable gaps in our knowledge of how lysostaphin targets S. aureus cells and then cleaves the pentaglycine cross-bridge. Information on the nature of the lysostaphin receptor in S. aureus cells has only recently been revealed , and there are no reported mutants in lysostaphin that inactivate its biological activity. Other glycyl-glycine endopeptidase orthologues of lysostaphin have been reported, including ALE-1 from Staphylococcus capitis EPK1 that is 83% identical with the mature lysostaphin sequence  and LytM from S. aureus  where residues 185–315 are 49% identical with the endopeptidase domain of lysostaphin (residues 8–138). The existence of other members of this M37/M23 protease family such as zoocin A in Streptcoccus equi subsp. zooepidemicus 4881 , enterolysin A in Enterococcus faecalis DPC5280  and millericin B in Streptococcus milleri NMSCC 061 , produced by host organisms that do not contain pentaglycine cross-bridges in their cell walls, implies a wealth of structure/function information and novel antimicrobial agents that could be revealed by a comparative study of this group of endopeptidases.
The published assays of endopeptidase activity of lysostaphin are not sufficiently adaptable for structure/function studies of this family of endopeptidases [24–26]. We recently described the design and synthesis of three internally quenched substrates for lysostaphin based on the peptidoglycan cross-bridge of S. aureus and their use in FRET (fluorescence resonance energy transfer) assays . In the present paper, we describe the development of a novel polypeptide lysostaphin FRET substrate (MV11F) that facilitates the analysis of the endopeptidase activity of lysostaphin and LytM. We have validated the FRET assay by demonstrating that site-directed mutants that abolish the biological activity of lysostaphin similarly inhibit the endopeptidase activity against the FRET substrate. We have also identified the site of lysostaphin cleavage of the pentaglycine target sequence in the polypeptide FRET substrate and generated new information on the effect of serine residues in the pentaglycine target sequence on cleavage. Our long-term aim is to solve the structure of lysostaphin and its homologues bound to a non-cleavable peptide/peptidomimetic and to use the structural information to help design novel antibiotics with an extended spectrum of activity against Gram-positive pathogens.
MATERIALS AND METHODS
Escherichia coli DH5α (Novagen) was used as the host for cloning and mutagenesis. E. coli BL21 (DE3) and ER2566 (NEB) were used as host strains for the expression vector pET21a (Novagen). E. coli were grown at 37 °C in LB (Luria–Bertani) or 2YT medium [1.6% (w/v) tryptone, 1% (w/v) yeast extract and 0.5% (w/v) NaCl]  with ampicillin added (100 μg/ml) for selection of resistant cells. The plasmid pGFPuv (Clontech) was the source of the GFPuv sequence. Alexa Fluor® 546–C5-maleimide was obtained from Invitrogen.
Cloning of lysostaphin and its endopeptidase domain
Although lysostaphin has been cloned previously by several groups [2,29], we engineered a C-terminal hexahistidine tag in order to facilitate the purification and analysis of mutant proteins. Using PCR with S. simulans biovar staphylolyticus template DNA, we introduced a methionine codon immediately upstream of the first alanine codon of mature lysostaphin (primer RJ84), and an XhoI site, and thus two additional amino acids (LE), in place of the stop codon of the lysostaphin gene (primer DW41) . The resulting 747 bp PCR product was restricted with NdeI and XhoI and then ligated into the expression vector pET21a restricted with the same enzymes, resulting in the plasmid pEA3.
The cloning and expression of the isolated endopeptidase domain of lysostaphin have not been reported previously. The endopeptidase domain of lysostaphin was amplified by PCR from the plasmid pEA3 using the primer RJ84 and the primer PTB7 that inserts an XhoI site in place of the Y141 codon of the lysostaphin gene, and cloned into pET21a restricted with NdeI and XhoI, resulting in the plasmid pPTB201. The cloning of the endopeptidase domain of the lysostaphin orthologue zoocin A has previously been reported .
Cloning of S. aureus LytM and the active truncation LytM-(185–316)
S. aureus RN450 genomic DNA was the template in a PCR using primers RJ81 (inserts an NcoI site, and thus a methionine codon, in place of the Ser-25 codon of LytM) and PTB63 (inserts an XhoI site, and thus two additional amino acids (Leu-Glu), in place of the stop codon of the lytM gene). The resulting 882 bp PCR product was restricted with NcoI and XhoI and then ligated into similarly restricted pET21d, resulting in the plasmid pPTB105. The cloning and expression of an active isolated endopeptidase domain of LytM was reported previously . This endopeptidase was amplified by PCR from the plasmid pPTB105 using primers PTB66 (NcoI site in place of the His-184 codon of the lytM gene) and PTB63 (replaces the stop codon with an XhoI site as above). The resulting PCR product was restricted with NcoI and XhoI, ligated into similarly restricted pET21d, resulting in the plasmid pPTB107.
Construction of the FRET substrate
A DNA fragment encoding the N-terminal 57 amino acids of colicin E9 was amplified from the plasmid pNP69 , using primers T7 promoter and MV14 (introduced N44C and S50G mutations and an XmaI site at the 3′-end of the PCR product). This PCR product was then subcloned into the cloning vector pET21a already containing the GFPuv sequence followed by a C-terminal hexahistidine tag. The cysteine residue introduced by the N44C mutation allowed labelling of the encoded protein MV11 with Alexa Fluor® 546–C5-maleimide and thus resulted in a biological FRET substrate, MV11F.
Site-directed mutants of the mature lysostaphin gene were constructed from pEA3 using the ‘megaprimer’ PCR method .
Biological activity assays
Lysostaphin-producing E. coli ER2566 (DE3) cultures were stabbed into LB agar plates containing ampicillin (100 μg/ml) and were incubated at 37 °C overnight. The plates were exposed to chloroform vapour for 15 min to lyse the cells and were then overlaid with 5 ml of molten 0.7% (w/v) non-nutrient agar containing 50 μl of an overnight culture of the MRSA indicator strain. The plates were incubated at 37 °C for at least 6 h and then inspected for a clear zone of growth inhibition around the test culture. Each culture was tested independently in triplicate.
Protein expression was induced by IPTG (isopropyl β-D-thiogalactoside; 1 mM) in cultures of E. coli BL21 (DE3) carrying the relevant plasmid growing in 2YT–ampicillin (100 μg/ml) medium at 25 °C. Initial purifications of lysostaphin and its endopeptidase using nickel-affinity chromatography yielded proteins with reduced biological (and catalytic) activity when compared with zinc-purified proteins . For zinc-affinity chromatography, frozen cells were gently resuspended in buffer A [20 mM Tris/HCl (pH 7.0), 0.5 M NaCl and 5% (v/v) glycerol] with 1 mM PMSF, sonicated on ice and then subjected to centrifugation. The cleared cell lysate was applied to a 5 ml HiTRAP chelate column (Pharmacia Biotech) previously charged with zinc and equilibrated in buffer A. Bound proteins were eluted from the column using a linear gradient of increasing imidazole. Fractions containing the required protein, identified by SDS/PAGE, were pooled and dialysed overnight (4 °C) in the appropriate buffer for ion-exchange chromatography.
His-tagged lysostaphin(s) and the lysostaphin endopeptidase domain were diluted 10-fold into buffer cexA [10 mM Tris, pH 7.0, 50 mM NaCl, 5% glycerol and 0.1% (v/v) Triton X-100]. His-tagged biological substrates were diluted 3-fold in buffer aexA [20 mM Tris, pH 8.1, 25 mM NaCl, 5% glycerol and 1 mM DTT (dithiothreitol)]. His-tagged LytM-(185–316) was stored overnight (4 °C) without dialysis prior to cation-exchange chromatography.
Lysostaphin or endopeptidase domain in buffer cexA was loaded on to a 5 ml HiTRAP SP column equilibrated in buffer cexA and then eluted using a linear gradient of increasing sodium chloride. Fractions containing only the required protein were identified using SDS/PAGE and stored at −80 °C. His-tagged LytM-(185–316) was purified using a 5 ml HiTRAP DEAE FF column as previously described . The flow-through fraction contained purified LytM-(185–316) that was concentrated using a Microcon Centriplus YM3 centrifugal filter device with a 3 kDa cut-off and stored at −80 °C.
Anion-exchange chromatography of the biological substrates
His-tagged biological substrate was loaded on to a 1 ml MonoQ column (Pharmacia Biotech) equilibrated in buffer aexA and then eluted using a linear gradient of increasing sodium chloride. Fractions containing isolated biological substrate were identified using SDS/PAGE and stored at −80 °C.
To obtain purified decoy peptide from E. coli cell lysates that have expressed decoy–colicin E9 immunity complexes, we used zinc-affinity chromatography to capture the complex and a denaturing elution buffer to isolate the decoy peptide from the immobilized immunity protein. E. coli cell lysates were loaded on to a 5 ml HiTRAP chelate column as described for the MV11 biological substrates, washed in buffer A and then eluted in 20 mM Tris/HCl (pH 7.9), 6 M guanidine and 0.5 M NaCl. Decoy substrates were dialysed in 100 mM ammonium bicarbonate (pH 7.8) and freeze-dried to allow MS and enable resuspension in FRET buffer at high concentration.
Protein concentration determination
The concentrations of all purified proteins were determined by UV absorbance at 280 nm. The following molar absorption coefficients (ε) were determined online by the ExPASy (Expert Protein Analysis System) tool Protpram , using the amino acid sequence of each protein: lysostaphin and mutants, 67840 M−1·cm−1; the endopeptidase domain, 36900 M−1·cm−1; LytM, 63720 M−1·cm−1; LytM-(185–316), 32890 M−1·cm−1; decoy substrates, 31970 M−1·cm−1. The concentration of the biological substrates was determined using the absorbance of GFP (green fluorescent protein) at 397 nm and a molar absorption coefficient of 30000 M−1·cm−1 (Clontech).
A turbidity assay, similar to that previously described , was used to demonstrate lysostaphin activity, in which freeze-dried EMRSA-16 cells  were resuspended in a buffer (25 mM sodium phosphate, pH 7.0, and 171 mM NaCl) and the attenuance at 600 nm (D600) was monitored continuously using a Shimadzu 160A spectrophotometer or a Wallac Victor2 multilabel counter at ambient temperature. After ∼5 min, lysostaphin (5 μM final concentration) was added and changes in D600 were observed. To show that a lysostaphin mutant protein retained the ability to bind S. aureus, the assay was repeated with the mutant lysostaphin (10 μM final concentration) mixed with lysostaphin.
Polypeptide cleavage assay
To assess the cleavage of solvent-exposed ‘pentaglycine’ amino acid sequences by lysostaphin, each of the MV11 family of biological substrates (20 μM final concentration) was incubated with lysostaphin (2 μM final concentration) for 4 h at 37 °C in a buffer (25 mM sodium phosphate, pH 7.0, and 150 mM NaCl). Digests were stopped with the addition of an equal volume of SDS/PAGE loading buffer and cleavage was assessed by SDS/PAGE.
Fluorescence measurements were acquired at 37 °C using an LS55B luminescence spectrometer fitted with a thermostatically controlled cell holder (PerkinElmer) operating on FLWinlab® software.
Labelling of the polypeptide FRET substrate
Polypeptide FRET substrates were labelled with a 15-fold molar excess of Alexa Fluor® 546–C5-maleimide for 2 h at room temperature (22 °C). Excess DTT was added and free label was removed by extensive dialysis of the substrates against PBS (pH 7.4). Energy transfer diminishes rapidly with increasing distance between the donor and acceptor [FRET efficiency E=R06/(R06+r6), where r is the distance between the donor and acceptor and R0 is the distance at which transfer efficiency is 50% of the maximum] . For the FRET substrate MV11F, R0=51 Å (1 Å=0.1 nm) .
Polypeptide FRET endopeptidase assay
Cleavage of the MV11F substrates by lysostaphin was followed continuously over a 5 min period by measuring the fluorescence yield at 508 nm (emission slit width of 10 nm) that resulted from excitation at 475 nm (excitation slit width of 5 nm). The substrate in 150 μl of 25 mM sodium phosphate (pH 7.0) and 150 mM NaCl buffer was added into a 1 cm path length cuvette and equilibrated at 37 °C in the thermostatically controlled cell compartment of the fluorimeter prior to the addition of lysostaphin.
ESI–TOF (electrospray ionization–time-of-flight) MS
The masses of peptides were measured by ESI–TOF MS by using a Micromass LCT by operating the reflector in positive ion mode. The instrument was calibrated using equine heart myoglobin (Sigma). Peptides were dialysed extensively in 100 mM ammonium bicarbonate buffer (pH 7.8; 4 °C), freeze-dried and then dissolved in 50% (v/v) acetonitrile and ∼1% formic acid for application into the spectrometer. Peptide digestions were performed in 100 mM ammonium bicarbonate buffer (pH 7.8; 37 °C) using the minimum enzyme necessary and for up to 36 h. The multiply charged ion data were transformed into a singly charged ion indicative of the molecular mass using a custom software package.
Expression and purification of lysostaphin, endopeptidase and mutant proteins
Mature, recombinant lysostaphin, consisting of residues 247–493 , was purified from an IPTG-induced culture of E. coli BL21 (DE3) pEA3 by zinc metal-chelate chromatography, making use of the introduced C-terminal hexahistidine tag, followed by cation-exchange chromatography. Similarly, the isolated endopeptidase domain of lysostaphin, consisting of residues 1–140, was purified from an E. coli BL21 (DE3) pPTB201 culture. ESI–TOF MS analysis of the purified proteins gave a molecular mass of 28012.8 Da for lysostaphin and 16256.3 Da for the endopeptidase domain, compared with the predicted values of 28008.4 and 16254.1 Da respectively for the two proteins in which the N-terminal methionine residues are lacking.
In order to validate our lysostaphin assays, we constructed some endopeptidase mutants of lysostaphin. BLAST searches (http://www.ncbi.nlm.nih.gov/BLAST/) using the endopeptidase domain of lysostaphin reveal similar sequences in a large number of Gram-positive and Gram-negative bacteria, which on alignment revealed seven residues, other than glycine residues, that are absolutely conserved (Figure 1A). We used alanine scanning mutagenesis to investigate the effect of mutating all seven of these residues (His-33, Asp-37, Tyr-81, His-83, Thr-107, His-114 and His-116) of the mature lysostaphin sequence encoded by the plasmid pEA3 on the biological activity. The results clearly showed that mutagenesis of His-33, Asp-37, His-83, His-114 or His-116 to alanine abolished the biological activity of the mutant lysostaphin in the stab test, whereas mutations of Tyr-81 or Thr-107 to alanine resulted in a reduced killing zone size (Figure 1B). The seven mutant proteins were then assayed in a turbidity assay, in comparison with the mature lysostaphin (Figure 1C). To exclude the possibility that the five alanine mutations inactivated the endopeptidase activity by destabilizing the conformation of the mutant lysostaphin proteins, we demonstrated that the mutant proteins were able to protect S. aureus cells from killing by mature lysostaphin in a turbidity protection assay (results not shown), thus demonstrating that the mutant proteins retained normal S. aureus targeting activity and thus the C-terminal targeting domain at least must be normally folded.
A polypeptide FRET substrate for the glycyl-glycine endopeptidase activity
We recently described the design and synthesis of three internally quenched peptide FRET substrates for lysostaphin . The relatively low binding affinity of lysostaphin for the three FRET substrates (Km values of between 0.7 and 3 mM) required that they be used at relatively high substrate concentrations, with resulting solubility problems, and in the presence of high concentrations of enzyme. For this reason, we explored alternative polypeptide FRET substrates that might exhibit a higher sensitivity and would also be amenable to protein engineering and thus allow a range of mutant substrates to be rapidly produced. The polypeptide FRET substrate was based on the N-terminal region of the translocation domain of colicin E9 as this is known to be an NDR (natively disordered region) of the protein  and thus an introduced pentaglycine cleavage site might be readily accessible to cleavage by the endopeptidase. We engineered a chimaeric protein that consisted of the N-terminal 57 residues of colicin E9, a seven-residue linker and the GFPuv protein, with a C-terminal hexahistidine tag to facilitate protein purification. The amino acid sequence of the N-terminal 64 residues that are linked to GFPuv is shown in Figure 2. The mutation S50G created a pentaglycine cleavage site between residues 47 and 51, whereas the S52A and G53A mutations removed a possible extension of the endopeptidase cleavage site. The cysteine residue that was introduced by an N44C mutation allowed labelling of MV11 with Alexa Fluor® 546–C5-maleimide to produce the polypeptide FRET substrate, MV11F.
Cleavage of both the MV11 and MV11F substrates could be observed by SDS/PAGE and this demonstrated that labelling MV11 with Alexa Fluor® 546–C5-maleimide did not inhibit cleavage by lysostaphin or the endopeptidase domain (Figure 2). MS analysis of purified MV11 revealed a major species with a molecular mass of 33681.2 Da, which is consistent with the predicted molecular mass of MV11 without the N-terminal methionine residue of 33680.9 Da. After cleavage by the endopeptidase domain, the large C-terminal fragment of MV11 was determined to be 29445.5 Da (Table 1). Assuming cleavage between the second and third glycine residues of the pentaglycine target sequence of MV11, the predicted molecular mass of the large, C-terminal fragment that includes GFPuv is 29442.7 Da. If cleavage occurred between the third and fourth, or the fourth and fifth glycine residues, then the predicted molecular mass of the large fragment is 29386.1 and 29329 Da respectively. The observed cleavage site between glycine residues 2 and 3 of the pentaglycine target sequence of MV11 is consistent with one of the three proposed cleavage sites identified by MS of muropeptides after cleavage of S. aureus peptidoglycan with lysostaphin , and with the favoured cleavage site of lysostaphin in the synthetic FRET substrates .
Alexa Fluor® 546 has an excitation maximum at 554 nm and an emission maximum at 570 nm. With excitation at 475 nm, FRET will occur between the GFPuv and Alexa Fluor® 546 in MV11F such that the direct emission of GFPuv at 508 nm will decrease . This is shown in Figure 3(A). The FRET efficiency, E, can be calculated from the equation E=1−IDA/ID, where IDA is the intensity of the donor in the presence of the acceptor (MV11F) and ID is the intensity of the donor alone (MV11), and in this case E ∼96%. Incubation of the polypeptide FRET substrate MV11F with lysostaphin or the endopeptidase domain results in an increase in the emission at 508 nm, reflecting the cleavage of the pentaglycine target sequence and thus separation of the FRET donor and FRET acceptor, with the rate of increase being linear over a 5 min period (Figure 3B). Cleavage rates during the first 2 min for 200 nM enzyme and a range of MV11F concentrations (0–20 μM) were obtained (results not shown). The rates were converted from relative fluorescence units into concentrations using a calibration curve of the fluorescence of known concentrations of GFPuv (MV11) and corrected for quenching effects at higher substrate concentrations . Approximate kinetic constants were obtained by fitting the corrected rates to a Michaelis–Menten plot. The Km (app) was 65±27 μM for lysostaphin and 92±41 μM for the endopeptidase domain. The Vmax was 10.1±3.5 nM·s−1 for lysostaphin and 5.2±2.1 nM·s−1 for the endopeptidase domain. The narrow substrate concentration range that could be used for analysis, due to fluorescence quenching effects at higher substrate concentrations, resulted in rather large S.E.M. for the fitting. The increased sensitivity of this assay, reflected in shorter incubation times when compared with the chemical FRET assay, was consistent with the Km values obtained in the latter assay of between 70 and 200 μM for the three chemical substrates .
Cleavage of MV11 by other M37 endopeptidases
Zoocin A is produced by S. equi subsp. zooepidemicus 4881 and is reported to have inhibitory activity against Streptococcus pyogenes and Streptococcus mutans . The N-terminal endopeptidase domain of zoocin A shows considerable homology to the endopeptidase domain of lysostaphin (Figure 1A) and, surprisingly, has been reported to cleave a hexaglycine substrate in a colorimetric assay even though the peptidoglycan cross-link in the sensitive streptococcal strains is reported to consist of two or three L-alanine residues . We observed no cleavage of MV11 by high concentrations of purified zoocin A even though the protein showed the expected inhibitory activity against cultures of S. pyogenes (results not shown). LytM is an autolysin of S. aureus and is the first member of the group of lysostaphin metallopeptidases whose structure has been determined [26,41]. We engineered a truncated LytM from S. aureus, consisting of residues 185–316, and experimentally determined its mass as 15277.6 compared with the predicted value of 15274.0 for the protein lacking its N-terminal methionine. The truncated LytM protein cleaved MV11 with an experimentally determined mass for the large C-terminal cleavage product of 29327.1 (Table 1). This indicates that, in contrast with lysostaphin, cleavage of the MV11 substrate by LytM occurs between glycine residues 4 and 5 of the pentaglycine target sequence.
Effect of serine residues on the cleavage of MV11 by lysostaphin and LytM
Lysostaphin-producing S. simulans strains are protected from killing by the production of a Lif (lysostaphin immunity factor) that is homologous with the FemA and FemB proteins that are essential for the synthesis of pentaglycine cross-bridges in S. aureus . The presence of the Lif protein results in the incorporation of serine residues at positions 3 and 5 of the pentaglycine cross-bridge that prevent endopeptidase cleavage by lysostaphin [42–44], but it is not known whether the incorporation of one or both serine residues is required. Using site-directed mutagenesis, we engineered the pMV11 plasmid to produce four constructs that encode MV11 proteins containing a serine residue at positions 1, 2, 3 and 5 respectively of the pentaglycine cleavage site. We compared the cleavage of MV11 and the four serine-containing MV11 mutant proteins by lysostaphin using SDS/PAGE (Figure 4). The results demonstrated that a serine residue in the third position of the pentaglycine cleavage site (MV11-S3) abolished cleavage by lysostaphin or the endopeptidase, whereas the presence of a serine mutation at position 1 (MV11-S1), 2 (MV11-S2) or 5 (MV11-S5) was observed to decrease the rate of cleavage only after long incubation times (Figure 4). MS analysis of the cleavage products of these long incubations confirmed the absence of cleavage of the MV11-S3 protein and that cleavage of MV11-S1, MV11-S2 and MV11-S5 by lysostaphin occurs between glycine residues 3 and 4, 4 and 5 and 2 and 3 respectively (Table 1).
The truncated LytM polypeptide cleaved MV11, MV11-S1 and MV11-S2 less efficiently than lysostaphin, whereas cleavage of MV11-S5 was very efficient despite the preferential cleavage site being between glycine residues 4 and 5 (Table 1). MS analysis of the cleavage products of these long incubations confirmed the absence of cleavage of the MV11-S3 protein and that cleavage of MV11-S1 and MV11-S2 occurred between glycine residues 4 and 5. Cleavage of the MV11-S5 protein occurred between glycine residues 2 and 3 (Table 1). As expected, purified MV11-S3 labelled with Alexa Fluor® 546 (MV11-S3F) showed no cleavage when incubated with LytM (results not shown).
Other residues at position 3 inhibit cleavage of the MV11 pentaglycine substrate
In order to determine whether other amino acid residues at position 3 of the pentaglycine cleavage site in MV11 could inhibit cleavage by lysostaphin, we engineered a stop codon at this position in MV11, resulting in the loss of fluorescence from the expressed protein. Using a mutagenic primer that randomly mutated the stop codon to a codon for any amino acid, we screened for transformants in which GFP fluorescence was restored. This can only result from the replacement of the stop codon with an amino-acid-encoding codon, thus allowing expression of a full-size MV11 polypeptide. Rapid screening of cell lysates from these transformants, after incubating with the endopeptidase domain (20 μM) and then running the digests on SDS/16% PAGE, indicated that 8 of 25 transformants expressed an MV11 biological substrate that could be cleaved by the endopeptidase. DNA sequencing revealed that a glycine residue had been introduced in place of the stop codon in these eight clones in which the MV11 mutant protein could be cleaved. The other 17 clones had amino acids other than glycine present at this position including asparagine (in 3 clones), alanine (2), leucine (2), phenylalanine (2), valine, serine, threonine, cysteine and tryptophan. The MV11 substrate with an alanine residue at position 3 was purified and incubated with lysostaphin or LytM in buffer, rather than in the presence of LB medium in the cell digest incubations, before running the digest on SDS/PAGE. The results showed that both lysostaphin and LytM can cleave this substrate after long incubations, which suggests that the presence of an alanine residue at this position does not prevent binding of these enzymes to the modified substrate. However, purified MV11 substrates having valine or leucine in position 3 were not cleaved by lysostaphin or LytM in buffer. These experiments show the value of the MV11 substrate for studying the substrate preference of M37 endopeptidases such as lysostaphin and LytM.
A serine residue in the third position of the pentaglycine cleavage site does not prevent binding of the lysostaphin endopeptidase
One of our long-term aims is to solve the structure of lysostaphin bound to a non-cleavable peptide/peptidomimetic because this would enable rational mutagenesis of lysostaphin for the generation of new enzymes that cleave GGSGG sequences present in bacteria resistant to lysostaphin. To assess lysostaphin binding of MV11-S3, we performed a competition polypeptide FRET assay between fluorescently labelled MV11F and an unlabelled decoy peptide with a solvent-exposed GGSGG cleavage site. The decoy peptide was engineered by replacement of the GFPuv-encoding region of the MV11-3 substrate with the DNase-Im9 immunity protein-encoding region from colicin E9 . Further mutagenesis of the cleavage site resulted in two additional decoy peptides: a positive control peptide with a GGGGG site and a negative control peptide with an AASAA cleavage site. ESI–TOF MS analysis of the purified decoys gave a molecular mass of 20654.7 Da for the decoy with the GGSGG cleavage site, 20613.7 Da for the decoy with the GGGGG cleavage site and 20701.8 Da for the decoy with the AASAA cleavage site, compared with the predicted values of 20643.7, 20613.7 and 20699.8 Da respectively for the three proteins in which their N-terminal methionine residues are lacking. As expected, incubation of each of the decoy substrates alone with lysostaphin or its endopeptidase only led to cleavage of the decoy with the GGGGG cleavage site (results not shown).
The different masses of MV11 and the decoy substrate allowed observation on SDS/PAGE of MV11 cleavage by lysostaphin when co-incubated with each of the decoy substrates (results not shown). Cleavage of MV11F by the endopeptidase domain in the presence of excess of the GGSGG decoy was significantly reduced relative to cleavage of MV11F on its own, and similar to cleavage of MV11F in the presence of an equivalent excess of the decoy with the GGGGG cleavage site (Figure 5). Evidence for specific inhibition of endopeptidase cleavage by the above two decoy substrates was concluded from the finding that cleavage of MV11F in the presence of an equivalent excess of the AASAA decoy was not reduced. Inhibition of cleavage from the decoys with the GGGGG or GGSGG cleavage site was approx. 40-fold greater than product inhibition of the endopeptidase by triglycine.
We have developed an assay for the endopeptidase activity of mature lysostaphin, or of its endopeptidase domain, that uses an engineered polypeptide substrate, MV11. Cleavage of MV11 after long incubations with lysostaphin can be observed on SDS/PAGE gels, but labelling of MV11 with Alexa Fluor® 546 results in a sensitive FRET substrate (MV11F) in which cleavage by 200 nM lysostaphin can be observed during a 5 min assay period.
Clues as to the possible roles of the five essential residues that we have identified in our mutational studies in the endopeptidase activity of lysostaphin have come from the crystal structure of LytM [26,41], and from mutational analysis of the glycyl-glycine endopeptidase ALE-1 . The crystal structure shows that LytM is a two-domain protein in which the C-terminal domain is conserved in lysostaphin metallopeptidases. Three zinc ligands were identified in this domain (His-210, Asp-214 and His-293) that are conserved in lysostaphin homologues , and are equivalent to the His-33, Asp-37 and His-116 residues (Figure 1A) that we have shown to be essential for endopeptidase activity (Figure 1B). The fourth zinc ligand in LytM is Asn-117, which is located at the N-terminal domain of LytM and has no obvious equivalent in lysostaphin . In ALE-1 alanine mutations of residues His-150, Asp-154, His-200, His-231 and His-233 (that are equivalent to His-33, Asp-37, His-83, His-114 and His-116 in lysostaphin) resulted in almost complete loss of enzymatic activity, while only the alanine mutations at His-150 or His-233 resulted in loss of a zinc molecule .
One other residue was shown to be essential for the biological activity of LytM, His-291, which is equivalent to His-114 in lysostaphin and His-231 in ALE-1. This residue is at least 4 Å away from the Zn2+ in the LytM crystal structure and thus is not involved in metal chelation . The precise role of this histidine residue in LytM, ALE-1 or lysostaphin is currently unknown. One suggestion is that it could be acting as a general base to activate the water molecule by deprotonating it and then a general acid to protonate the nitrogen of the amine leaving group; however, it has been pointed out that residue His-260 in LytM could also play this role . It is significant that the His-83 residue that we have demonstrated is essential for the biological activity of lysostaphin is equivalent to His-260 residue of LytM. The crystal structure of the endopeptidase domain of lysostaphin, or LytM, in complex with a substrate analogue will be required to address this issue.
Resistance to lysostaphin is associated with the activity of the lif gene, which results in site-specific incorporation of serine residues at positions 3 and 5 of the pentaglycine cross-bridge of S. simulans . Our results on the effect of site-specific serine residues on the cleavage of MV11 by lysostaphin, or the endopeptidase domain, strongly suggest that the incorporation of a serine at position 3 in the pentaglycine will be a significant resistance factor in S. simulans if lysostaphin is used clinically. The first example of a plasmid-encoded lif gene that is not associated with a closely linked endopeptidase gene has been reported in a Staphylococcus species , which suggests that this resistance gene could spread rapidly by horizontal gene transfer if lysostaphin is used clinically and justifies research to develop second-generation lysostaphins that can cleave serine-containing modified cleavage sites.
This project was funded by the University of Nottingham and The Wellcome Trust [grant number 066850/C/02/A].
We thank Graham Coxhill (School of Chemistry, University of Nottingham) for assistance with MS.
Abbreviations: DTT, dithiothreitol; ESI–TOF, electrospray ionization–time-of-flight; FRET, fluorescence resonance energy transfer; GFP, green fluorescent protein; IPTG, isopropyl β-D-thiogalactoside; LB, Luria–Bertani; Lif, lysostaphin immunity factor; MRSA, methicillin-resistant Staphylococcus aureus; MV11, polypeptide lysostaphin substrate; MV11F, polypeptide lysostaphin FRET substrate
- © The Authors Journal compilation © 2009 Biochemical Society