Novel staphylococcal glycosyltransferases SdgA and SdgB mediate immunogenicity and protection of virulence-associated cell wall proteins

Abstract

Infection of host tissues by Staphylococcus aureus and S. epidermidis requires an unusual family of staphylococcal adhesive proteins that contain long stretches of serine-aspartate dipeptide-repeats (SDR). The prototype member of this family is clumping factor A (ClfA), a key virulence factor that mediates adhesion to host tissues by binding to extracellular matrix proteins such as fibrinogen. However, the biological siginificance of the SDR-domain and its implication for pathogenesis remain poorly understood. Here, we identified two novel bacterial glycosyltransferases, SdgA and SdgB, which modify all SDR-proteins in these two bacterial species. Genetic and biochemical data demonstrated that these two glycosyltransferases directly bind and covalently link N-acetylglucosamine (GlcNAc) moieties to the SDR-domain in a step-wise manner, with SdgB appending the sugar residues proximal to the target Ser-Asp repeats, followed by additional modification by SdgA. GlcNAc-modification of SDR-proteins by SdgB creates an immunodominant epitope for highly opsonic human antibodies, which represent up to 1% of total human IgG. Deletion of these glycosyltransferases renders SDR-proteins vulnerable to proteolysis by human neutrophil-derived cathepsin G. Thus, SdgA and SdgB glycosylate staphylococcal SDR-proteins, which protects them against host proteolytic activity, and yet generates major eptopes for the human anti-staphylococcal antibody response, which may represent an ongoing competition between host and pathogen.

Figure 1. mAb rF1 exhibits robust binding to and killing of S. aureus bacteria.

(A-C) Bacteria were preopsonized with huIgG1 mAbs rF1 (squares), 4675 anti-ClfA (triangles), or anti-herpes virus gD (circles). (A) Binding of mAbs to WT (USA300-Δspa) bacteria was assessed by flow cytometry, and expressed as mean fluorescent intensity (MFI). (B) CFSE-labeled, preopsonized WT (USA300-Δspa) bacteria were incubated with human PMN. Bacterial uptake was expressed as % of CFSE-positive PMN, after gating for CD11b-positive cells by flow cytometry. (C) Preopsonized WT (USA300-Δspa) bacteria were incubated with PMN to assess bacterial killing. Numbers of viable CFU per mL are representative of at least three experiments. (D) Flow cytometry analysis of binding of rF1 to S. aureus from various infected tissues. Homogenized tissues were double stained with mAb rF1 (X-axis), and with anti-peptidoglycan mAb 702 to distinguish bacteria from tissue debris (Y-axis) (left panel; gate indicated by arrow), followed by gating of bacteria to generate histogram figures. (E) Binding of rF1 to various staphylococcal and non-staphylococcal Gram-positive bacterial species by flow cytometry. Red lines, rF1; blue lines, isotype control mAb anti-gD; green lines, control without mAb. (See also Figure S1).

Figure 2. mAb rF1 binds to a family of serine-aspartate-repeat (SDR)-proteins.

(A) rF1-reactivity with USA300 CWP is sensitive to proteinase-K (PK) treatment. Lysostapahin-derived CWP from WT (USA300-Δspa) bacteria was left untreated (lane 1) or treated with 10 µg/mL PK for 1 hour (lane 2), and immunoblotted with rF1. (B) rF1-reacitivty is dependent on the presence of SDR-proteins. CWPs from WT, indicated deletion strains of various combinations of SDR-family proteins , , and a Δspa strain as control for non-specific binding, were immunoblotted with rF1. The lower molecular weight bands (∼50 kDa) were due to non-specific IgG binding to protein A. (C) rF1 also binds to additional SDR-proteins from S. epidermidis. Cell lysates from S. epidemidis were immunoprecipitated with rF1 (lane 1) or an isotype-control mAb (lane 2) and immunoblotted with rF1 mAb. Identities of rF1-reactive bands were revealed by mass-spectrometry of the same lysates (see also Figure S2). (D) Alignment of SDR-proteins revealed by mass-spectrometry from S. aureus and S. epidermidis. SDR-regions are indicated by red hatches. Three truncation mutants of clumping factor A (ClfA) that were fused with maltose-binding protein (MBP) are also shown. (E) SDR-region is sufficient for rF1 reactivity. CWPs from S. aureus expressing truncated recombinant constructs were immunoblotted with anti-MBP mAb or rF1 mAb.

Figure 3. mAb rF1 recognizes post-translational sugar modifications on SDR-proteins.

(A) rF1-reactivity is a pathogen-specific modification. His-tagged MBP-SDR constructs were over-expressed in E. coli, B. subtilis, or S. aureus, and whole cell lysates were immunoblotted with rF1 or anti-His mAb. (B) E. coli derived SDR-proteins can be post-translationally modified by S. aureus lysates to confer rF1-reactivity. His-tagged E. coli-expressed SDR-proteins were bound to nickel beads, incubated with or without ΔpanSDR mutant S. aureus whole cell lysate and immunoblotted with rF1 or anti-His mAbs. (C) Fractionation of S. aureus cellular components conferring rF1-reactivity. His-tagged, E. coli-expressed ClfA protein was bound to nickel beads and incubated with fractions of ΔpanSDR mutant S. aureus whole cell lysates that were prepared by size exclusion chromatography (upper panel), and further fractionated by ion-exchange chromatography (bottom panel). Individual fractions were spiked with B. subtilis lysate to provide necessary building blocks for the reactions. Beads were washed, eluted and immunoblotted with rF1. (See also Figure S3). (D) Genomic organization of the SDR-CDE locus. The glycosyltransferase genes sdgA (SAUSA300_0549) and sdgB (SAUSA300_0550) are found adjacent to the sdr-genes.

Figure 4. SdgB is the key rF1 epitope-modifying enzyme.

(A) SdgB is necessary for rF1 reactivity. Cell wall lysates from WT and various putative glycosyltransferase mutants were immunoblotted with mAbs rF1, anti-ClfA (9E10), anti-SdrD (17H4) or anti-panSDR (9G4 α-SDR; recognizes the unmodified SDR-domain. (B) Complementation of ΔsdgB with exogenous SdgB confers rF1 reactivity. Cell wall lysates from WT, glycosyltransferase mutants, and the SdgB-complemented strain were immunoblotted with rF1, anti-ClfA, and anti-SDR mAbs as in (A). (C) Binding of rF1 to whole USA300 bacteria requires SdgB. Binding of mAbs to ΔsdgB USA300 was assessed by flow cytometry as described in Figure 1A. (D) rF1-mediated killing of USA300 activity requires SdgB. Wild-type USA300 bacteria preopsonized with rF1 (closed square) or anti-gD (closed circle), and ΔsdgB preopsonized with rF1 (closed triangle) or anti-gD (open circle), were incubated with PMN, and bacterial killing was determined as in Figure 1C. (E) MBP-SDR-His construct was expressed in WT, ΔsdgA, ΔsdgB, or ΔsgdAΔsdgB S. aureus, and whole cell lysates were immunoblotted with rF1, anti-His and anti-SDR. (F) Preliminary model for step-wise glycosylation of SDR-proteins by SdgB and SdgA. SDR-domains are first glycosylated by SdgB, which appends sugar modifications creating the epitope of mAb rF1. SdgA further modifies these epitopes with additional sugar moieties (left panel). The ΔsdgA S. aureus mutant shows that SdgA-mediated modifications do not influence rF1-binding activity (middle panel). In ΔsdgB or ΔsgdAΔsdgB S. aureus, the unmodified SDR-region is now recognized by the anti-pan-SDR mAb (9G4).

Figure 5. SdgB and SdgA sequentially modify the SDR-domain with GlcNAc moieties.

(A) SdgB generates rF1 epitopes on SDR protein. A combination of MBP-SDR-His and SdgA or SdgB was co-expressed in E. coli, and cell lysates were immunoblotted with mAb rF1, or with mAb against unmodified SDR (9G4) or anti-His. (B) Cell-free system to reconstitute SDR glycosylation using purified components. Recombinant MBP-SDR-His was incubated with purified SdgA or SdgB, and in the presence or absence of UDP-GlcNAc; rF1 reactivity was induced only in the presence of SdgB and UDP-GlcNAc. (C) Final model for step-wise glycosylation of SDR proteins by SdgA and SdgB. First, SdgB appends GlcNAc moieties onto the SD-region on SDR proteins, followed by additional GlcNAc modification by SdgA. The epitope for mAb rF1 includes the SdgB-dependent GlcNAc moieties. (D) Mass spectrometry analysis to identify the SDR-sugar moieties using purified MBP-SDR-His expressed in E. coli. (Upper panel) Deconvoluted mass spectrum of purified MBP-SDR-His protein, showing the expected intact mass of 58719 Da. (Middle panel) MBP-SDR-His protein was treated with purified SdgB enzyme in the presence of UDP-GlcNAc for 2 h at 37°C. After incubation, the mass of the MBP-SDR-His protein showed several peaks, each peak being separated from the others by the mass of additional GlcNAc residues. (Bottom panel) The above-mentioned reaction mixture of MBP-SDR-His and SdgB (middle panel) was additionally treated with purified SdgA enzyme. After further incubation for 2 hrs at 37°C, up to an additional 47 GlcNAc groups were found to be added. Thus, most of the serines in the DSD motifs in MBP-SD can be modified with these disaccharide sugar moieties.

Figure 6. Recognition of SdgB-dependent epitope by human antibodies.

(A) Four different human IgG preparations were reacted with plate-bound CWP from WT or ΔsdgB USA300 by ELISA. To calculate the specific anti-staphylococcal IgG content, data were normalized using a calibration curve with known IgG concentrations of a mAb against peptidoglycan, which has the same reactivity with both USA300 strains by ELISA. Data are expressed as µg/mL of anti-staphylococcal IgG in the serum. The reduction in reactivity observed for CWP from ΔsdgB (red bars) as compared to wild-type CWP (black bars) reflects IgG specific for SdgB-dependent epitopes. Asterisks indicate significant differences (p < 0.05) from WT CWP. (B) CWP from WT, ΔsdgA, or ΔsdgB, ΔsdgAΔsdgB USA300 were immunoblotted with rF1 and three additional human mAbs (SD2, SD3, and SD4) from different patients. All four mAbs showed similar epitope specificity.

Figure 7. SdgB glycosylation protects SDR proteins from cleavage by human neutrophil-derived cathepsin G.

(A) Live, in tact WT or ΔsdgB USA300 bacteria were incubated in the presence or absence of human neutrophil lysosomal extracts (NLE). Culture supernatants were immunoblotted with a mAb against the A-domain of ClfA (9E10) to detect cleaved ClfA fragments released from the bacteria. (B) Live, in tact WT or ΔsdgB cells were incubated in the presence or absence of lysosomal extracts from human THP1 cells or mouse RAW cells and culture supernatants were immunoblotted with anti-ClfA. (C) Live, intact WT or ΔsdgB cells were incubated with a panel of purified human neutrophil serine proteases, ie. neutrophil elastase (NE), cathepsin G (CatG), proteinase-3 (P3), and neutrophil serine protease-4 (NSP4). (D) Δ sdgB cells were treated with human neutrophil lysosomal extract in the presence or absence of a biochemical inhibitor of cathepsin G. (E) WT or various Sdg-mutant strains were treated with purified human cathepsin G. (B-E) Culture supernatants were analyzed by immunoblotting as in (A) to detect released ClfA fragments. (F) Live bacteria of WT, Δ sdgB, or ΔsdgB complemented with exogenous SdgB (psdgB) were treated with purified human cathepsin G. Culture supernatants (Sup) or cell wall preparations (CWP) were immunoblotted with mAb against the A-domain of ClfA (S4675), SdrD (17H4), or IsdA (2D3). In addition to S4675, another mAb against the A-domain of ClfA (9E10) showed similar results (not shown). (G) Human cathepsin G inhibits adherence of glycosylation-deficient S. aureus to human fibrinogen. Live WT or ΔsdgB USA300 bacteria were pre-incubated with cathepsin G, and allowed to adhere to fibrinogen-precoated plates. Bacterial adhesion was quantified by measuring the amount of bacterial ATP associated with the plates.