Wall teichoic acid substitution with glucose governs phage susceptibility of Staphylococcus epidermidis

Abstract

The species- and clone-specific susceptibility of Staphylococcus cells for bacteriophages is governed by the structures and glycosylation patterns of wall teichoic acid (WTA) glycopolymers. The glycosylation-dependent phage-WTA interactions in the opportunistic pathogen Staphylococcus epidermidis and in other coagulase-negative staphylococci (CoNS) have remained unknown. We report a new S. epidermidis WTA glycosyltransferase TagE whose deletion confers resistance to siphoviruses such as ΦE72 but enables binding of otherwise unbound podoviruses. S. epidermidis glycerolphosphate WTA was found to be modified with glucose in a tagE-dependent manner. TagE is encoded together with the enzymes PgcA and GtaB providing uridine diphosphate-activated glucose. ΦE72 transduced several other CoNS species encoding TagE homologs, suggesting that WTA glycosylation via TagE is a frequent trait among CoNS that permits interspecies horizontal gene transfer. Our study unravels a crucial mechanism of phage-Staphylococcus interaction and horizontal gene transfer, and it will help in the design of anti-staphylococcal phage therapies.IMPORTANCEPhages are highly specific for certain bacterial hosts, and some can transduce DNA even across species boundaries. How phages recognize cognate host cells remains incompletely understood. Phages infecting members of the genus Staphylococcus bind to wall teichoic acid (WTA) glycopolymers with highly variable structures and glycosylation patterns. How WTA is glycosylated in the opportunistic pathogen Staphylococcus epidermidis and in other coagulase-negative staphylococci (CoNS) species has remained unknown. We describe that S. epidermidis glycosylates its WTA backbone with glucose, and we identify a cluster of three genes responsible for glucose activation and transfer to WTA. Their inactivation strongly alters phage susceptibility patterns, yielding resistance to siphoviruses but susceptibility to podoviruses. Many different CoNS species with related glycosylation genes can exchange DNA via siphovirus ΦE72, suggesting that glucose-modified WTA is crucial for interspecies horizontal gene transfer. Our finding will help to develop antibacterial phage therapies and unravel routes of genetic exchange.

Keywords: Staphylococcus epidermidis; bacteriophages; coagulase-negative staphylococci; glycosyltransferase; wall teichoic acid.

Fig 1

The tagE gene encodes a glycosyltransferase in S. epidermidis. (a) Genetic locus identified by transposon mutagenesis contains the S. epidermidis tagE, pgcA, and gtaB homologs. Transposon insertion sites are labeled in gold. (b) MUSCLE alignment of S. epidermidis TagE with B. subtilis TagE and S. aureus TarM protein sequences.

Fig 2

Efficiency of plating (EOP) (a and b) and binding (d and e) of ΦE72 is reduced on the tagE, pgcA, and gtaB deletion mutants, compared to the S. epidermidis wild type (WT). This defect can be restored by complementing the tagE mutant with the genetic locus containing tagE, pgcA, and gtaB on plasmid pRB473 (c and f). The data represent the mean ± SEM of at least three independent experiments. Ordinary one-way analysis of variance (ANOVA) was used to determine statistical significance vs S. epidermidis 1457 WT, followed by Dunnett’s multiple comparisons tests, indicated as: not significant (ns), ^**P < 0.01, ^***P < 0.001, and ^****P < 0.0001.

Fig 3

WTA analysis of the S. epidermidis mutants ∆tagE, ∆pgcA, ∆gtaB, and of ∆tagE containing the pRB473 plasmid carrying tagE, pgcA, and gtaB genes for complementation. (a) Ratio of glucose per phosphate content of WTA was measured enzymatically. (b) HPLC-MS: Extracted ion chromatograms of GroP-Gro [(M - H)^- = 245.0432] and GroP-GroP [(M - H)^- = 325.0095] with [GroP-Gro-Glc; (M - H)^- = 407.096] [GroP-GroP-Glc; (M - H)^- = 487.0623] or without glucose substitution. (c) ¹H NMR spectra reveal D-glucose on WTA of the S. epidermidis 1457 WT (at the C2-position of GroP), while the deletion of tagE results in the absence of glucose on WTA. For (a), data represent the mean ± SEM of at least three independent experiments. Ordinary one-way ANOVA was used to determine statistical significance, followed by Tukey’s multiple comparisons tests, indicated as: not significant (ns), ^**P < 0.01, and ^***P < 0.001.

Fig 4

Glycolipid detection by thin layer chromatography (TLC). (a) LTA glycolipid biosynthesis pathway as described for S. aureus and B. subtilis [adapted from reference (37)]. (b) Glycolipid detection on a TLC plate stained with α-naphthol/sulfuric acid. For positive control, 5 µg of digalactosyldiacylglycerol (DGDG) was used, while the solvent methanol/chloroform (1:1) was used as negative control (n.cont.). One representative experiment of three independent experiments is shown.

Fig 5

TagE-glycosylated WTA increases the binding of siphoviruses but reduces podovirus binding. WTA glycosylation-deficient mutants of S. epidermidis show decreased binding of ΦE72-related siphoviruses Φ459, Φ456, and Φ27 (a) but increased binding of the podoviruses ΦUKE3, ΦSpree, and ΦBE03 (c), while the GroP-GalNAc-specific siphovirus Φ187 still shows strong binding (a). WTA glycosylation-deficient mutants of S. epidermidis show less plaque formation by ΦE72-related siphovirus Φ459 (b), while plaque formation by the myoviruses ΦBE04 (d) and ΦBE06 (e) remains unchanged. (f) Lysis zones by siphoviruses decrease in the absence of tagE but increase for podoviruses. Myoviruses show the formation of lysis zones independent of the presence or absence of tagE [pRB = pRB473 (empty vector control); comp = complementation with tagE, gtaB, and pgcA genes]. The data represent the mean ± SEM of at least three independent experiments. Ordinary one-way ANOVA was used to determine statistical significance vs S. epidermidis 1457 WT, followed by Dunnett’s multiple comparisons tests, indicated as: not significant (ns), *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. P-values are shown if appropriate.

Fig 6

TagE-modified WTA determines horizontal gene transfer between CoNS. (a) The purified cell wall of S. epidermidis 1457 with glucose-modified GroP-WTA (black) prevents plaque formation by ΦE72 more efficiently than in the absence of glucose (gold). (b) Transduction of the plasmid pRB473 by ΦE72 is decreased if the receiving strain lacks WTA modification with glucose. (c) Transduction of pRB473 with Φ187 is increased in the absence of glucose on GroP-WTA. (d) ΦE72 binds to different strains of S. epidermidis, but binding is prevented by RboP expression encoded by the tarIJLM2 cluster. (e) ΦE72-mediated transduction of pBASE6 or pC183-S3-GFP to CoNS depends on high TagE homology. Strain names are given in brackets. The data represent the mean ± SEM of at least three independent experiments. For (a), two-way ANOVA was used to determine statistical significance comparing unsupplemented control vs WT cell wall supplementation (black) and supplementation of WT cell wall vs ΔtagE cell wall (gold), followed by Tukey’s multiple comparisons tests. For (b) and (c), unpaired t tests were used to determine statistical significance vs S. epidermidis 1457 WT, and for (d), ordinary one-way ANOVA was used to determine statistical significance vs S. epidermidis 1457 WT, followed by Tukey’s multiple comparisons tests. Statistical significance is indicated as: not significant (ns), *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.