Site-Specific Serine Incorporation by Lif and Epr into Positions 3 and 5 of the Staphylococcal Peptidoglycan Interpeptide Bridge

Abstract

The FemAB-like factors Lif and Epr confer resistance to glycylglycine endopeptidases lysostaphin and Ale-1, respectively, by incorporating serine residues into the staphylococcal peptidoglycan interpeptide bridges specifically at positions 3 and 5. This required the presence of FemA and/or FemB, in contrast to earlier postulations.

Glycylglycine endopeptidases are staphylolytic enzymes that cleave the pentaglycine interpeptide bridges of the staphylococcal peptidoglycan. Staphylococcus simulans biovar staphylolyticus, which produces lysostaphin, and Staphylococcus capitis, which produces Ale-1, protect themselves from their endopeptidases by the corresponding lysostaphin immunity factor Lif (1, 10) or the endopeptidase resistance factor Epr, respectively (9). Resistance is due to the integration of serine in place of glycine residues in the peptidoglycan pentaglycine interpeptide bridge (1, 9, 10). Pentaglycine interpeptide bridge formation in Staphylococcus aureus depends on at least three factors, FmhB (7), FemA (5, 8), and FemB (4), which are needed for the addition of the first, the second and the third, and the fourth and the fifth glycines, respectively. Lif and Epr show up to 41% identity to FemA and FemB, suggesting that they may be catalyzing serine incorporation, and Lif was suggested to complement FemB (11). Here, we determined the positions of serine residues within the interpeptide bridge in a wild-type strain and different femAB mutants complemented with lif or epr and found a high specificity of Lif and Epr for serine incorporation at positions 3 and 5. Neither Lif nor Epr alone was able to extend the shortened cross bridges in the femAB mutants by the addition of serine residues, suggesting that both proteins depend on FemA and FemB for activity.

Muropeptide profile of wild-type and femAB mutants expressing lif.

Upon expression of lif from plasmid pCXlif (10), the amino acid fraction of the peptidoglycan in parent strain BB270 (NCTC 8325, mec) and in the corresponding femB mutant BB815 (mec, Ω2006femB::Tn551) (4) showed an increased serine content, whereas that of the femA mutant UK17 (mec, ochre mutation in femA) (3) and the femAB null mutant AS145 (mec, ΔfemAB::tetK) (8) was not altered (11). The resulting muropeptide patterns of BB270/pCXlif and BB815/pCXlif (11) showed additional peaks in the monomeric and dimeric fractions (Fig. 1a and b) when compared to the muropeptide profiles of the parent strains BB270 and BB815 (formerly UT43-2) determined earlier (5, 8). The main monomeric peaks of the two strains were collected and desalted as described (7), and their amino acid sequence was analyzed by automated Edman degradation (2). Since the major monomeric peak M4 could not be separated from the novel peak S1 (BB270/pCXlif) (Fig. 1a), the amino acid sequence of the isolated muropeptides revealed a mixture of the normal peptide with five glycine residues (M4; ∼90%) and the peptide Gly-Gly-Ser-Gly-Ser (S1; ∼10%). Furthermore, minor amounts of the sequence Gly-Gly-Ser-Gly-Gly could be detected. Consistent with the shortened interpeptide bridge in the femB mutant, the major monomeric peak M3 present in strain BB815/pCXlif (Fig. 1b) contained ∼40% Gly-Gly-Gly (M3) and ∼60% Gly-Gly-Ser (S2) sequences. The retention times of muropeptides containing three glycine residues and muropeptides containing one serine and two glycine residues were nearly identical under the high-performance liquid chromatography conditions used. Since the peaks designated S in Fig. 1a and b were absent in the corresponding isogenic parent strains, it is likely that these muropeptides were also modified by serine-containing interpeptide bridges. Their amino acid composition could not be determined due to their low abundance. In contrast to the wild-type and the femB mutant described above, expression of lif did not alter the muropeptide pattern of strains AS145 and UK17, and the major monomeric peaks revealed only monoglycine side chains as described earlier for femAB mutants (data not shown). No serine residues were found in these two strains by amino acid analysis.

FIG. 1.

Muropeptide profile of the monomeric and dimeric fractions of S. aureus BB270/pCXlif (a) and S. aureus BB815/pCXlif (b), BB270/pTSF6 (c), and BB815/pTSF6 (d). Muropeptides were prepared as previously described (8). Peaks were identified by comparison with the known profile of the parent strains (4, 8). Novel peaks, not found in the isogenic parental strains, are boxed. S, peaks that were not analyzed; S1 to S6, peaks that could be isolated and were analyzed by Edman degradation. M, P, and Pn stand for monomer disaccharide, stem pentapeptide, and stem pentapeptide in which Gln was replaced with Glu, respectively. M1, MP; M2, MP-Gly; M3, MP-Gly₃; M4, MP-Gly₅; M8, MPn-Gly₃; M9, mixture of Mpn-Gly₅ and MP-Ala; and M10, MP-(Gly₄Ala). Peptides marked by asterisks are probably degradation products produced by endopeptidases and include the following: M2*, MT-(Gly₂), a total of two glycines substituted at two possible sites, either Lys in position 3 or Ala in position 4 of the stem tetrapeptide; M4*, MT-(Gly₄); and M6*, MT-(Gly₆). Dimer muropeptides include the following: D3, dimer of M4, and D4, dimer from M3 and M4, and D7, dimer of M3. D2* is a degradation product containing two M4* units. The structure of the S peptides, containing serine residues, is described in the text.

Muropeptide profile of wild-type and femAB mutants expressing epr.

Similar to Lif, Epr was shown to increase the serine content of the peptidoglycan (9). When plasmid pTSF6 containing epr under the control of its own promoter (9) was introduced into BB270, the serine content increased compared to strain BB270 carrying the control plasmid pGC2 (Table 1). Likewise, the amount of serine increased in the femB mutant BB815/pTSF6 expressing epr. However, only negligible amounts of serine could be detected in the femAB double mutant AS145 carrying pTSF6 (Table 1).

TABLE 1.

Effects of epr on cell wall composition of the wild type and different femAB mutants^a

Strain	Molar mass ratio
	Ser	Gly	Ala	Lys	Ser + Gly
BB270/pTSF6	0.62	3.20	1.83	0.90	3.84
BB270/pGC2*	0.02	4.10	2.00	0.95	4.12
BB815/pTSF6	0.39	1.90	1.57	0.86	2.29
BB815/pGC2	0.14	2.20	1.57	0.85	2.34
AS145/pTSF6	0.09	0.96	2.06	0.94	1.05
AS145/pGC2*	0.02	0.95	1.98	0.93	0.97

^aCells were grown at 37°C in brain heart infusion medium with 20 μg of chloramphenicol per ml prior to isolation of peptidoglycan, which was then subjected to amino acid analysis as described earlier (8). The mean value of three or two (*) independent experiments is shown. BB270, parent strain; BB815, femB mutant (4); AS145, ΔfemAB mutant (8); pTSF6, plasmid carrying epr (9); pGC2, control cloning vector. Molar mass ratio related to Glu = 1.0. 

The muropeptide profiles of strain BB270/pTSF6 and femB mutant BB815/pTSF6 expressing epr showed additional peaks (Fig. 1c and d) compared to the known muropeptide patterns of BB270 and BB815 (5, 7, 8). In addition to M4, a second major peak (S1) with a slightly increased retention time was found in BB270/pTSF6. The composition of the interpeptide side chain of peak S1 revealed by Edman degradation the following sequences: Gly-Gly-Ser-Gly-Gly (∼55%) and Gly-Gly-Ser-Gly-Ser (∼45%). Instead of the one major peak of control strain BB270 (8), the dimeric fraction of strain BB270/pTSF6 contained three major peaks (Fig. 1c). Analysis of the amino acid sequence of the interpeptide side chains of all three peaks revealed the same amino acid sequences as in the monomeric peak S1 with serine incorporated in position three or in positions three and five of the cross bridge (Table 2). Overlapping of the three dimeric peaks S3, S4, and S5 prevented isolation of pure peaks. We found two muropeptide species in each of the peaks, namely Gly-Gly-Ser-Gly-Gly and Gly-Gly-Ser-Gly-Ser (Table 2). Considering the cross contamination, the different relative amounts suggested that S3 contains as main species dimers of Gly-Gly-Ser-Gly-Gly, whereas S4 contains both Gly-Gly-Ser-Gly-Gly and Gly-Gly-Ser-Gly-Ser, and S5, though contaminated with S4, mainly represents dimers with Gly-Gly-Ser-Gly-Ser bridges.

TABLE 2.

Amino acid composition of the interpeptide side chains of the main dimeric muropeptides S3, S4, and S5 found in BB270/pTSF6^a

Peak/muropeptide	Sequence	Relative amount (%)
S3	Gly-Gly-Ser-Gly-Gly	78
	Gly-Gly-Ser-Gly-Ser	22
S4	Gly-Gly-Ser-Gly-Gly	60
	Gly-Gly-Ser-Gly-Ser	40
S5	Gly-Gly-Ser-Gly-Gly	50
	Gly-Gly-Ser-Gly-Ser	50

^aMuropeptides were collected, desalted, and analyzed by automated Edman degradation (2). 

The monomeric muropeptide pattern of femB mutant BB815 complemented with epr revealed one major peak containing 58% triglycine cross bridges (M3), typical for a femB mutant (4), and, similar to BB815/pCXlif, cross bridges with the sequence Gly-Gly-Ser (42%; S2). The two main dimeric peaks (Fig. 1d) consisted of triglycine cross bridges (D7) and Gly-Gly-Ser cross bridges (S6).

In accordance with the only slightly increased amount of serine found in the peptidoglycan isolated from the femA mutant strain AS145/pTSF6 (Table 1) or UK17/pTSF6 (data not shown), only one minor novel peak was found in the monomeric muropeptide fraction of AS145/pTSF6, which could not be isolated in sufficient amounts for Edman degradation but which is likely to be substituted by a serine residue (data not shown).

Conclusions.

We could show that expression of Lif or Epr in S. aureus led to the incorporation of serine residues specifically in positions 3 and 5 of the interpeptide side chain. The resulting alternating sequence of glycine and serine residues protects the peptidoglycan from glycylglycine endopeptidases, which are unable to hydrolyze glycylserine and serylglycine peptide bonds (6). It is important to note that the detailed muropeptide analysis revealed that neither Lif nor Epr extended the interpeptide bridge. This is in contrast to the earlier postulation, based solely on the total amino acid composition of the peptidoglycan, that Lif complements FemB (11). The observation that the overall cross-linkage degree was not changed in the femAB mutant strains compared to their isogenic parental strains upon expression of lif or epr (data not shown) made it additionally unlikely that an increased amount of extended interpeptide bridges was present in the highly cross-linked peptidoglycan fraction. Thus, it seems that Lif requires the function of FemA and/or FemB for the incorporation of serine to the existing interpeptide cross bridge and that this function cannot be catalyzed by Lif or Epr alone. This is in striking contrast to the glycyl-glycine elongation of the side chain by FemA and FemB in S. aureus, despite the high sequence identities of FemA and FemB with Lif and Epr. The fact that the serine incorporation was dependent on the presence of either FemA and/or FemB suggests possible protein interactions between FemA, FemB, and Lif or Epr. There appear to be qualitative differences between Lif and Epr, suggesting a more efficient Ser incorporation by Epr than by Lif. However, we must take into account that Epr was expressed here from a high-copy-number plasmid and from its own promoter, whereas Lif was under the control of a xylose-inducible promoter in a vector with a lower copy number. Compared to FemA and FemB, the function of Lif and Epr seems to be different with regard to (i) serine instead of glycine incorporation, (ii) the site specificity of the positions of the incorporated serine residues, and (iii) the inability to attach additional amino acids to the peptide cross bridge on their own.