Peptidoglycan-linked protein A promotes T cell-dependent antibody expansion during Staphylococcus aureus infection

Significance

Staphylococcus aureus infection of humans is associated with immunoglobulin expansions but not with the development of immunity against recurrent disease. The mechanism whereby staphylococci manipulate the host immune system to block adaptive immunity is not known. We show here that S. aureus infection of mice is also associated with immunoglobulin expansions and that staphylococcal protein A (SpA), a molecule that cross-links B-cell receptors, is both necessary and sufficient for this immune evasion mechanism. Further, SpA requires covalent linkage to bacterial peptidoglycan and CD4 T-cell help to manipulate the immunoglobulin responses of infected hosts. Our results provide insights into microbial signaling mechanisms for B-cell activation and antibody development.

Abstract

A hallmark of Staphylococcus aureus disease in humans is persistent infections without development of protective immune responses. Infected patients generate V_H3 plasmablast expansions and increased V_H3 idiotype Ig; however, the mechanisms for staphylococcal modification of immune responses are not known. We report here that S. aureus-infected mice generate V_H3 antibody expansions via a mechanism requiring MHC-restricted antigen presentation to CD4⁺ T cells and staphylococcal protein A (SpA), a cell wall-anchored surface molecule that binds Fcγ and V_H3 variant heavy chains of Ig. V_H3 expansion occurred with peptidoglycan-linked SpA from the bacterial envelope but not with recombinant SpA, and optimally required five tandem repeats of its Ig-binding domains. Signaling via receptor-interacting serine/threonine protein kinase 2 (RIPK2) was essential for implementing peptidoglycan-linked SpA superantigen activity. V_H3 clan IgG from S. aureus-infected or SpA-treated animals was not pathogen-specific, suggesting that SpA cross-linking of V_H3 idiotype B-cell receptors and activation via attached peptidoglycan are the determinants of staphylococcal escape from adaptive immune responses.

The gram-positive pathogen Staphylococcus aureus evolved to colonize humans and their domesticated animals (1). The nares, skin, and gastrointestinal tract of approximately one-third of the human population are stably colonized, whereas another third is intermittently colonized (2). Colonization increases the risk of S. aureus infection, specifically for skin and soft tissue infection, invasive disease, as well as bacteremia (3, 4). S. aureus infection leads to the formation of abscess lesions supporting pathogen replication and dissemination (5). Even with antibiotic and surgical therapy (drainage of abscess lesions), S. aureus persists and causes recurrent disease, manifesting as relapse infections by strains responsible for the index case (6, 7). S. aureus persistence, recurrent disease, and the overall outcome of infections (morbidity and mortality) are impacted by the development of antibiotic-resistant strains, designated MRSA (methicillin-resistant S. aureus) (8). The emergence of MRSA strains has led to dramatic increases in the mortality of S. aureus infections, highlighting the need for development of new therapeutic and preventive strategies (9, 10).

S. aureus-infected individuals display expansions of blood V_H3 clonal plasmablasts and V_H3 clonal antibodies (11). Staphylococcal protein A (SpA), a cell wall-anchored molecule that is also released from the bacterial surface, associates with the variant heavy chains of V_H3 clonal IgG and IgM (B-cell receptors) (12, 13). Earlier work reported that SpA on bacterial surfaces but also SpA purified from S. aureus cell walls stimulate the proliferation of human peripheral blood B lymphocytes (14). In contrast, injection of recombinant SpA, purified from Escherichia coli, into mice induced trafficking and apoptosis of marginal zone B cells and innate-like B1 cells (15–17). Persistent depletion of functional B-cell repertoires suggested that animals treated with recombinant SpA may be more susceptible to infections with blood-borne pathogens (18). In agreement with these observations, injection of recombinant SpA into mice with systemic lupus erythematosus (SLE) depletes B1 and marginal zone B cells to alleviate antibody-induced nephritis and renal failure associated with SLE (19).

We sought to address the apparent paradox between V_H3 clonal B-cell expansion during S. aureus infection in humans and the collapse of V_H3 clonal B cells in mice treated with recombinant SpA. Earlier work demonstrated that S. aureus sortase A cleaves SpA precursors and covalently links the C-terminal end of polypeptides to peptidoglycan cross-bridges in the cell-wall envelope (20, 21). LytM hydrolase cuts the staphylococcal cell wall to release peptidoglycan-linked SpA into the extracellular medium (22). We show here that peptidoglycan-linked SpA, but not recombinant SpA, induces V_H3 clonal antibody expansions in mice.

Results and Discussion

S. aureus Infection of Mice Triggers V_H3 Antibody Expansions.

Intravenous inoculation of C57BL/6 mice with 1 × 10⁷ colony-forming units (CFUs) of S. aureus Newman causes bacteremia and persistent abscess formation in all tissues of infected animals (23). To determine whether infected mice expand V_H3 clonal antibodies, animals were bled 5, 12, 19, and 26 d following S. aureus inoculation and serum was analyzed for V_H3 clan IgM and IgG by ELISA with SpA_KK, a protein A variant that binds V_H3 variant heavy chains but not Ig Fcγ (24). Compared with mock-infected mice, the abundance of V_H3 clonal IgM was increased by 48-fold on day 5 following inoculation (wild-type S. aureus vs. mock; P < 0.0001) and thereafter declined as measured on days 12 and 19 (Fig. 1A). Further, S. aureus infection caused a 247-fold increase in the abundance of V_H3 clonal IgG on day 12 (P < 0.0001), which remained elevated for the remainder of the experiment (days 12, 19, and 26; Fig. 1C). As a control, serum IgM and IgG from S. aureus or mock-infected animals did not bind to SpA_KKAA, a protein A variant that binds neither Ig V_H3 variant heavy chains nor Fcγ (Fig. 1 B and D) (25). Antibody expansions were not observed when mice were injected with formalin-killed wild-type S. aureus, indicating that live bacteria are required for this phenotype (Fig. S1). Thus, similar to V_H3 clonal plasmablast and Ig expansions in humans with staphylococcal disease, S. aureus-infected mice also expand their V_H3 clonal IgM and IgG antibodies and fail to produce Ig that recognizes protein A as antigen.

Fig. 1.

V_H3 clonal antibody expansion in S. aureus-infected mice requires SpA binding to Ig variant heavy chains. Cohorts of C57BL/6 mice (n = 5) were infected by i.v. inoculation with a sublethal dose (1 × 10⁷ CFUs) of wild-type S. aureus, protein A variants (spa_KK, spa_AA, and spa_KKAA), or left uninfected (naive). On days 5, 12, 19, and 26 post infection, serum samples were collected from mice. IgM (A and B) and IgG (C and D) responses were analyzed against SpA_KK (A and C), a protein A variant that specifically recognizes V_H3 clonal Fab fragments, and SpA_KKAA (B and D), a protein A variant that fails to recognize Ig molecules, to determine the abundance of V_H3 clonal antibody and protein A-specific antibody, respectively. Statistical analysis was performed with two-way ANOVA (*P < 0.05; **P < 0.001; ***P < 0.0001). Data points represent the mean ± SEM. Results are representative of three independent analyses.

Fig. S1.

Formalin-fixed S. aureus does not induce V_H3 clonal antibody expansions. Cohorts of BALB/c mice (n = 5) were inoculated by i.v. injection with 1 × 10⁷ CFUs of formalin-fixed S. aureus Newman (NM). On days 5, 12, 19, and 26 post injection, serum samples were collected from mice. IgM (A) and IgG (B) responses were analyzed against SpA_KK, which specifically recognizes V_H3 clonal Fab fragments, to determine the abundance of V_H3 clonal antibody. Data points represent the mean ± SEM. Results are representative of two independent analyses.

Antibody Expansion Relies on Protein A Binding to V_H3 Heavy Chains.

The structural gene for protein A (spa) of S. aureus Newman encodes a precursor with an N-terminal cleavable signal peptide, five Ig-binding domains (IgBDs), the cell wall-spanning region (region X), and the C-terminal LPXTG sorting signal for cell-wall attachment of the mature product (26–29). Each IgBD is endowed with two conserved glutaminyl (Q_9,10) and aspartyl residues (D_36,37) that are critical for Fcγ and Fab binding, respectively (25). A variant, spa_KKAA, with substitutions of these glutaminyl (Q₉K, Q₁₀K) and aspartyl (D₃₆A, D₃₇A) residues in each of the five IgBDs, is unable to bind the V_H3 variant chains or Fcγ of immunoglobulins (24). In contrast, spa variants carrying only glutaminyl (Q₉K, Q₁₀K in SpA_KK) or aspartyl substitutions (D₃₆A, D₃₇A in SpA_AA) in their five IgBDs retain the ability to bind either V_H3 variant chains or Fcγ, respectively (24). When used in the i.v. challenge model, the spa_KKAA variant of S. aureus Newman did not cause expansions of V_H3 clonal IgM and IgG (Fig. 1 A and C). A similar result was obtained in mice infected with the spa_AA variant, whereas S. aureus expressing spa_KK promoted V_H3 clonal IgM and IgG expansions resembling those of animals challenged with wild-type staphylococci (Fig. 1 A and C). Similar to wild-type infected mice, serum IgM and IgG from spa_KK-, spa_AA-, and spa_KKAA-infected animals did not bind to SpA_KKAA (Fig. 1 B and D). These data indicate that protein A, specifically the D₃₆,₃₇ residues in each of the five IgBDs responsible for heavy-chain binding to B-cell receptors, is required for the expansion of V_H3 clonal antibodies during S. aureus infection.

V_H3 Clonal Antibody Expansion Is a Universal Attribute of MRSA Isolates.

The amino acid sequences of IgBDs encoded by the spa gene are conserved among clinical isolates of S. aureus (30, 31). The vast majority of clinical S. aureus isolates express a SpA molecule with five tandem IgBD repeats (28), whose coding sequence is located upstream of the highly variable region X in the spa gene, otherwise used for genotyping S. aureus (27, 32, 33) (Fig. 2A). We investigated whether infection with MRSA clinical isolates would trigger V_H3⁺ antibody expansions in mice. Immunoblot analysis revealed that three representative MRSA isolates, strains USA300 LAC, MW2, and N315, expressed similar amounts of SpA in the bacterial envelope (Fig. 2B). However, USA300 LAC and MW2 released increased amounts of SpA into the extracellular medium compared with N315 (Fig. 2B). Cohorts of naïve mice were infected with 5 × 10⁶ CFUs of MRSA isolates and monitored for 26 d, and SpA_KK⁺(V_H3⁺) serum IgM and IgG were quantified (Fig. 2 C and D). Mice infected with methicillin-sensitive S. aureus (MSSA) or MRSA isolates generated a 5- to 12-fold increase in SpA_KK⁺(V_H3) IgM antibody on day 5 post infection (Fig. 2C). As SpA_KK⁺(V_H3⁺) IgM decreased, V_H3 IgG antibody levels increased sharply in MRSA USA300- or MW2-infected animals, leading to >200-fold expansion of Ig (Fig. 2D). In contrast, mice infected with MRSA isolate N315 displayed a moderate 51-fold expansion of V_H3⁺ IgG on day 12, which subsequently waned (Fig. 2D). The spa gene of MRSA N315 encodes only four IgBDs, suggesting that the amplitude of V_H3 antibody expansions may be a function of protein A IgBD repeats (Fig. 2A).

Fig. 2.

Clinical MRSA isolates elicit nonspecific V_H3 clonal antibody responses in infected mice. (A) Schematic representation of mature protein A gene products from MSSA strain Newman and three MRSA strains, USA300, MW2, and N315. PG, peptidoglycan-linked protein A. (B) S. aureus Newman, USA300, MW2, and N315 cultures were centrifuged (8,000 × g for 10 minutes) and the supernatant was separated from the bacterial pellet. After treatment of the staphylococci with lysostaphin, proteins in supernatant and cell-wall fractions were analyzed by immunoblotting with HRP-conjugated human IgM. (C and D) Cohorts of BALB/c mice (n = 5) were infected by i.v. inoculation with 1 × 10⁷ CFUs of S. aureus Newman or 5 × 10⁶ CFUs of S. aureus USA300, MW2, or N315. On days 5, 12, 19, and 26 post infection, serum samples were collected from infected mice. IgM (C) and IgG (D) responses were analyzed against SpA_KK, a protein A variant that fails to recognize Ig molecules, to determine the abundance of V_H3 clonal antibody. Data points represent the mean ± SEM. Results in B–D are representative of two independent analyses.

IgBD Repeat Variation Affects Staphylococcal Binding to Ig Heavy Chains.

Analysis of 3,905 community- and 2,205 hospital-associated S. aureus isolates from Oxfordshire, United Kingdom, revealed that all strains carried intact spa ORFs (34). Rearrangements in the spa gene were observed for IgBD coding sequence in 1.8% of community and 0.6% of hospital isolates (34). Most of these gene rearrangements are in-frame deletions of either the D, B, or C domain coding sequence, resulting in protein A variants with only four IgBDs, similar to that observed for MRSA strain N315 (34). Gene rearrangements with deletions of coding sequence for two, three, or four IgBDs are much more rare (1/6,110) (34). These findings suggest that the IgBD tandem repeats in the spa gene are subject to variation via homologous recombination, as has been observed for other microbial virulence factors with tandem repeat sequences (35, 36). Because spa genes with five IgBDs are present in >98% of isolates, it appears that the number of SpA Ig domains displayed on the surface of S. aureus is subject to purifying selection.

We wondered whether variations in the number of IgBD repeats of spa genes impact the B-cell superantigen activity of S. aureus isolates. To analyze this, we created five variants of S. aureus Newman-expressing spa genes with one (spa_E), two (spa_ED), three (spa_EDA), four (spa_EDAB), or six (spa_EDABCD) IgBDs and compared their biological attributes with the wild-type parent (spa_EDABC) and the Δspa deletion mutant (Fig. 3A). Flow cytometry analysis of bacteria stained with α-SpA_KKAA antibody revealed that all spa variants expressed and anchored protein A on the staphylococcal surface (Fig. S2A). The capacity to bind human IgG varied among mutants. Whereas the Δspa mutant retained only small amounts of IgG, increasing amounts of Ig were captured on the surface of staphylococci expressing protein A with one, two, three, four, five, or six IgBDs (Fig. S2B). A similar result was obtained following incubation with fluorophore-conjugated human IgG-Fcγ fragments (Fig. S2C). In contrast, S. aureus variants expressing protein A with one (spa_E), two (spa_ED), or three IgBDs (spa_EDA) did not retain significant amounts of fluorophore-conjugated human IgG F(ab)₂ on the bacterial surface (Fig. S2D). On the other hand, strains with four (spa_EDAB), five (spa_EDABC), or six (spa_EDABCD) IgBDs bound increasing amounts of IgG heavy chains (Fig. S2D). Mice were infected with the various spa mutant strains, and blood was sampled on days 5 and 14 post infection and analyzed for V_H3⁺ IgM and IgG (Fig. 3 B–E). Compared with the Δspa mutant, strains expressing more than three IgBDs (spa_EDAB, spa_EDABC, and spa_EDABCD) caused a significant increase in V_H3⁺ IgM on days 5 and 14, whereas the spa_E, spa_ED, and spa_EDA variants did not (Fig. 3 B and D). Only the spa_EDAB, spa_EDABC, and spa_EDABCD mutant strains increased V_H3 clonal IgG during infection (Fig. 3 C and E). Although SpA_EDABCD displayed the highest capacity for binding to human F(ab)₂ (Fig. S2D), mice infected with the spa_EDABCD mutant showed diminished expansion of V_H3 clonal IgM and IgG (spa_EDABC; Fig. 3 B–E). Thus, S. aureus strains expressing protein A with five IgBD repeats display greater B-cell superantigen activity than isolates with fewer or more repeats, a mechanism that may support purifying selection during infection.

Fig. 3.

Four IgBDs are required to induce nonspecific V_H3 clonal antibody expansion in infected mice. (A) A schematic representation of protein A variants harboring one (spa_E), two (spa_ED), three (spa_EDA), four (spa_EDAB), five (spa_EDABC; wild-type Newman), and six (spa_EDABCD) Ig-binding domains in S. aureus Newman. (B–E) Cohorts of BALB/c mice (n = 8–10) were i.v. infected with a sublethal dose (1 × 10⁷ CFUs) of S. aureus Newman (spa_EDABC) or protein A variants (Δspa, spa_E, spa_ED, spa_EDA, spa_EDAB, and spa_EDABCD). At days 5 (B and C) and 14 (D and E) post infection, serum samples were collected from mice. IgM (B and D) and IgG (C and E) responses were analyzed against SpA_KK, which specifically recognizes V_H3 clonal Fab fragments, to determine the abundance of V_H3 clonal antibody. Statistical analysis was performed using one-way ANOVA. Error bars represent the mean ± SEM. Results in B–E are representative of two independent analyses.

Fig. S2.

S. aureus variants expressing variable numbers of protein A IgBDs. Overnight cultures of staphylococci grown in TSB were diluted 1:100 and grown at 37 °C with rotation to A₆₀₀ 0.6. Bacteria were centrifuged (8,000 × g for 10 minutes) and washed in PBS (n = 3). Before staining, staphylococci were fixed in 2.5% paraformaldehyde and 0.006% glutaraldehyde at room temperature for 20 min, washed, and blocked in 3% BSA/PBS. (A) To quantify protein A in the cell wall, staphylococci were incubated with 2 µg of affinity-purified rabbit α-SpA_KKAA antibody, followed by Alexa Fluor 647-conjugated α-rabbit IgG (1:2,000; Invitrogen). (B–D) To examine Ig binding to staphylococci, cells were incubated with Alexa Fluor 647-conjugated human IgG (B), FITC-conjugated Fc (C), or FITC-conjugated F(ab)₂ fragments of human IgG (D) (1:500; Jackson ImmunoResearch), washed in 3% BSA/PBS, and analyzed on an LSR II (BD Biosciences) (University of Chicago flow cytometry core facility). Data were analyzed by FlowJo software for mean fluorescence intensity (MFI). Statistical analysis was performed with one-way ANOVA. Data represent the mean ± SEM. Results are representative of two independent analyses.

V_H3⁺ Antibody Expansion Requires Peptidoglycan-Linked Protein A.

Mature, cell wall-anchored protein A is released from the staphylococcal surface during bacterial growth, and this molecule may modify the adaptive immune system of infected hosts by cross-linking B-cell receptors (22). Unlike recombinant protein A (rSpA), purified without the C-terminal sorting signal from E. coli, native SpA (nSpA) released from the S. aureus cell wall is modified with murein tetrapeptide-tetraglycyl [l-Ala-d-iGln-(SpA-Gly₅-)l-Lys-d-Ala-Gly₄] linked to its C-terminal threonyl residue (22) (Fig. 4A). We asked whether i.p. administration of purified rSpA and nSpA promotes V_H3 antibody expansions. Cohorts of naïve mice were treated with 0.1–100 µg of rSpA on days 0, 5, 12, and 19 and V_H3⁺ IgM and IgG antibodies were quantified (Fig. 4 B and C). As expected from earlier work studying the impact of rSpA on V_H3⁺ B cells in mice (15), repeated injection of mice with rSpA did not expand V_H3 clan IgM and IgG antibody levels (Fig. 4 B and C). In contrast, mice treated with 10 µg of nSpA mounted a 4.3 (±1.4)-fold increase in V_H3⁺ IgM and a 41.1 (±23.8)-fold increase in V_H3⁺ IgG antibodies on day 26 post injection (Fig. 4 D and E). These data indicate that B-cell receptor cross-linking by rSpA is not sufficient to induce proliferation of V_H3⁺ B lymphocytes and expand V_H3⁺ serum Ig. In contrast, native, peptidoglycan-linked protein A can promote B-cell proliferation and V_H3⁺ serum Ig expansions.

Fig. 4.

Peptidoglycan-linked protein A, but not the recombinant protein A, elicits nonspecific V_H3 clonal antibody expansion in mice. (A) A schematic representation of recombinant protein A purified from E. coli (rSpA) and native protein A linked to peptidoglycan, purified from S. aureus Newman culture supernatant (nSpA). (B–E) Cohorts of BALB/c mice (n = 3) were treated with various amounts of affinity-purified rSpA (B and C) or nSpA (D and E) on days 0, 5, 12, and 19 post injection. On days 5, 12, 19, and 26 following injection, serum samples were collected from mice. IgM (B and D) and IgG (C and E) responses were analyzed against SpA_KK to determine the abundance of V_H3 clonal antibody. (D, Inset) Coomassie-stained SDS/PAGE gel depicting affinity-purified nSpA. (F and G) Cohorts of C57BL/6 mice (n = 5) were i.v. infected with a sublethal dose (1 × 10⁷ CFUs) of S. aureus Newman (wild-type) or protein A null mutant (Δspa). On days 0, 5, 12, and 19 post infection, mice were treated with 10 µg of affinity-purified nSpA. On days 5, 12, 19, and 26 post infection, serum samples were collected from mice. IgM (F) and IgG (G) responses were analyzed against SpA_KK, which specifically recognizes V_H3 clonal Fab fragments, to determine the abundance of V_H3 clonal antibody. Data points represents the mean ± SEM. Results in B–F are representative of two independent analyses.

Of note, nSpA-mediated V_H3 IgM and IgG expansions were observed only after prolonged treatment of mice, whereas S. aureus infection promotes IgM expansions within 5 d. We wondered whether pathogen-associated molecular patterns (PAMPs) derived from staphylococci can rescue the phenotypic defect of nSpA. Mice were infected with 1 × 10⁷ CFUs of the Δspa mutant with or without 10-µg nSpA treatment on days 0, 5, 12, and 19 post infection. Compared with mice infected with wild-type S. aureus, challenge with the Δspa mutant did not elicit SpA_KK⁺(V_H3⁺) IgM and IgG antibodies (Fig. 4 F and G). However, treatment of Δspa-infected mice with nSpA caused enhanced SpA_KK⁺(V_H3⁺) IgM and IgG antibody production as early as day 12 post infection (Fig. 4 F and G). Thus, proliferation of V_H3⁺ B cells and expansion of V_H3 clonal Ig require peptidoglycan-linked protein A on the staphylococcal surface or released into the extracellular medium.

V_H3 Clonal Antibody Expansion Is MHC-Restricted and Requires Receptor-Interacting Serine/Threonine Protein Kinase 2 Signaling.

Antigen recognized by B-cell receptors is endocytosed, processed, and loaded onto MHC class II (H-2) compartments for presentation to CD4⁺ T cells (37). The quality and quantity of antibody responses to a given antigen are under genetic control of the H-2 locus (38). SpA-activated B cells induce proliferation of human T cells in an MHC class II-dependent manner (39). To determine whether a specific H-2 type favors V_H3 antibody expansion, multiple inbred mouse strains with well-defined MHC haplotypes were treated with nSpA, and blood samples were analyzed for V_H3⁺ IgM and IgG responses (Fig. 5 A and B). H-2^a (A/J) and H-2^q (FVB/N) animals developed fulminant antibody expansions. H-2^d (BALB/c and DBA/2) animals generated variable responses, whereas H-2^b and H-2^k (C3H/HeJ, C57BL/6, and CBA) mice responded in a moderate manner (Fig. 5 A and B). The abundance of preexisting V_H3 clonal antibodies in the blood of various mouse strains was not correlated with the magnitude of antibody expansions following nSpA treatment (Fig. 5C).

Fig. 5.

MHC-restricted V_H3 clonal antibody expansion is dependent on RIPK2 signaling. (A–C) Cohorts of A/J, BALB/c, C3H/HeJ, C57BL/6, CBA, DBA/2, and FVB/N mice (n = 3) were treated with 10 µg of affinity-purified nSpA on days 0, 5, 12, and 19 post injection. On days 5, 12, 19, and 26 post injection, serum samples were collected from mice. IgM (A) and IgG (B) responses were analyzed against SpA_KK to determine the abundance of V_H3 clonal antibody. (C) IgM response of naive animals before nSpA treatment was quantified against SpA_KK by ELISA. (D and E) Cohorts of C57BL/6 and TLR2-, NLRP3-, and RIPK2-deficient mice (n = 4–6) were treated with 10 µg of affinity-purified nSpA on days 0, 5, 12, and 19 post injection. On day 26 following injection, serum samples analyzed for IgM (D) and IgG (E) against SpA_KK by ELISA. Statistical analysis was performed with two-way ANOVA (*P < 0.01). Data represent the mean ± SEM.

Nucleotide oligomerization domain 1 (NOD1) and 2 (NOD2) are intracellular receptors that activate nuclear factor κB (NF-κB)-mediated inflammatory signaling in response to bacterial peptidoglycan (40). Receptor-interacting serine/threonine protein kinase 2 (RIPK2) acts as an adaptor, downstream of NOD1 and NOD2, and is essential for peptidoglycan-mediated NF-κB activation (41). Toll-like receptor 2 (TLR2) recognizes peptidoglycan and lipoteichoic acid on the cell surface (42). The NACHT, LRR and PYD domains-containing protein 3 (NLRP3) is activated by a variety of pathogen-associated molecular patterns (PAMPs) to up-regulate caspase 1-mediated release of interleukin 1β (43). To investigate the role of peptidoglycan-linked protein A in promoting B-cell superantigen activity, wild-type C57BL/6 and TLR2, NLRP3, or RIPK2 knockout mice were treated with nSpA (Fig. 4 B and C). Upon nSpA treatment, TLR2 and NLRP3 knockout mice responded with V_H3⁺ antibody expansions similar to wild-type mice (Fig. 5E). In contrast, RIPK2 knockout mice failed to expand V_H3 clonal IgM and IgG, suggesting that peptidoglycan-linked protein A stimulates the immune system via the NOD1 and NOD2 signaling pathways and their RIPK2 adaptor.

Requirement of CD4⁺ T Cells for V_H3 Antibody Expansions.

B-cell antibody class switching is modulated by CD4⁺ T helper cells (44). Human plasmablasts isolated from patients with S. aureus infection produced affinity-matured, class-switched V_H3 clonal antibodies (11). V_H3⁺ B cells activated by S. aureus infection of mice undergo class switching from IgM to IgG between 5 and 12 d post infection, leading to the corresponding changes in V_H3 clonal Ig. We therefore wondered whether CD4⁺ T helper cells are also required for V_H3 clonal Ig expansion. Mice were treated with multiple doses of rat monoclonal antibodies targeting CD4 or CD8α or mock-treated with PBS. Depletion of CD4 or CD8α T cells was verified by flow cytometry analysis of splenocytes (Fig. S3 A–C). Following i.v. infection with wild-type S. aureus, mice were monitored for morbidity and mortality. As expected from earlier work demonstrating the contribution of T cells toward innate immunity against S. aureus (45, 46), mice depleted of CD4- or CD8α-positive T cells were more susceptible to staphylococcal challenge and displayed increased mortality (Fig. S3D). Blood samples of infected mice were collected on days 5 and 12 post challenge and serum IgM and IgG levels were analyzed (Fig. S3 E and F). Animals treated with PBS or α-CD8α displayed similar expansions of V_H3 clonal IgM and IgG (Fig. S3 E and F). In contrast, CD4 T cell-depleted mice were unable to expand V_H3 clonal Ig during staphylococcal infection (Fig. S3 E and F). These data suggest that CD4⁺ T helper cells contribute to V_H3 clonal Ig expansions and class switching from IgM to IgG.

Fig. S3.

CD4⁺ T cells are necessary to induce nonspecific V_H3 clonal antibody expansion in infected mice. Cohorts of C57BL/6 mice (n = 3–5) received α-CD4– (GK1.5) or α-CD8α– (53-6.72) specific rat monoclonal antibodies to deplete CD4- or CD8-specific T cells. (A–C) Flow cytometry analysis was performed on single-cell suspensions of splenocytes prepared from mice after three injections of α-CD4 or α-CD8α antibodies. Lymphocyte population was gated and further analyzed using α-CD3ε (145-2C11), α-CD4 (RM4-5), α-CD8α (5H10), and α-CD19 (1D3). A statistical analysis was performed using one-way ANOVA (*P < 0.05; **P < 0.01). (D–F) On day 6 following initial antibody treatment, mice were infected by i.v. inoculation with 1 × 10⁷ CFUs of S. aureus Newman. (D) Kaplan–Meier curve depicting mortality over 12 d post infection. On days 5 and 12 post infection, serum samples were collected from infected mice. (E and F) IgM (E) and IgG (F) responses were analyzed against SpA_KK, a protein A variant that fails to recognize Ig molecules, to determine the abundance of V_H3 clonal antibody. Statistical analysis was performed with one-way ANOVA (*P < 0.05; **P < 0.01). Data points represent the mean ± SEM. Results are representative of two independent analyses.

V_H3 Antibodies of S. aureus-Infected Mice Are Not Pathogen-Specific.

To assess whether B cells in S. aureus-infected mice generate adaptive immune responses specific to staphylococcal antigens, cohorts of mice were infected with S. aureus Newman or its spa_KKAA variant for 14 d. Hyperimmune serum from animals in the same cohort was pooled and subjected to affinity chromatography (Fig. S4 A and B). ELISA analysis revealed that affinity-purified V_H3⁺ IgG from wild-type S. aureus-infected mice was predominantly of the IgG_2a and IgG_2b isotypes, whereas V_H3⁺ IgG from spa_KKAA variant-infected mice harbored similar amounts of all isotypes (Fig. S4C). Of 12 staphylococcal antigens (Coa, ClfA, ClfB, EsxA, EsxB, Hla, IsdA, IsdB, LukD, SdrC, SdrD, and SdrE) examined, only 4 antigens—Hla, IsdA, IsdB, and LukD—were recognized by SpA_KK⁺ or SpA_KK⁻ IgG generated during infection with wild-type S. aureus (Fig. S5). The total amount of SpA_KK⁺ IgG from mice infected with the spa_KKAA variant recognizing IsdA, IsdB, Hla, and LukD was increased (Fig. S5). Conversely, there was a significant reduction in SpA_KK⁻ IgG in mice infected with the spa_KKAA variant recognizing these four antigens. These data suggest that S. aureus-infected mice cannot generate pathogen-specific adaptive immune responses via their V_H3 antibody repertoire due to the B-cell superantigen attributes of protein A.

Fig. S4.

Purification of SpA_KK⁺ and SpA_KK⁻ IgG antibodies by stepwise affinity column purification. Cohorts of BALB/c mice (n = 50) were infected by i.v. injection with S. aureus Newman (wild-type) or its spa_KKAA variant and monitored for 14 d. (A) Hyperimmune mouse sera were collected from cohorts of infected animals, pooled, and subjected to affinity chromatography. First, immune sera were chromatographed on a SpA_KKAA column (GE Healthcare Life Sciences) to eliminate polyreactive antibodies and protein A-specific antibodies (α-SpA_KKAA). Flow-through fractions were pooled and subjected to SpA_AA affinity chromatography (GE Healthcare Life Sciences), and IgG fractions were eluted with elution buffer (0.1 M glycine⋅HCl, pH 3.0), followed by immediate neutralization with 1 M Tris⋅HCl (pH 8.5). Neutralized Ig was dialyzed against PBS. Last, pooled IgG fractions were loaded onto a SpA_KK-conjugated column (GE Healthcare Life Sciences) to separate SpA_KK⁺ (V_H3⁺) IgG from SpA_KK⁻ (V_H3⁻) IgG. (B) Coomassie-stained SDS/PAGE gel depicting the affinity of purified SpA_KK⁺ and SpA_KK⁻ IgG antibodies. (C) ELISA to determine the abundance of IgG₁, IgG_2a, and IgG_2b isotypes in affinity-purified SpA_KK⁺ antibody. All values are normalized against isotype control antibodies (IgG₁, IgG_2a, and IgG_2b). Statistical analysis was performed with two-way ANOVA (*P < 0.01). Data represent the mean ± SEM. Results are representative of two independent analyses.

Fig. S5.

V_H3 antibodies of S. aureus-infected mice fail to target staphylococcal antigens. Cohorts of BALB/c mice (n = 50) were infected by i.v. injection with 1 × 10⁷ CFUs of S. aureus Newman or spa_KKAA. On day 14 post infection, blood was collected by cardiac puncture and serum was separated by stepwise affinity purification of SpA_KK⁺ and SpA_KK⁻ IgG as shown in Fig. S3A. ELISA was performed to determine staphylococcal antigen-specific antibody responses against IsdA (A), IsdB (B), Hla (C), and LukD (D). Statistical analysis was performed with two-way ANOVA (*P < 0.01). Data represent the mean ± SEM. Results are representative of two independent analyses.

Studies in humans with S. aureus infection revealed expansions of V_H3 clonal plasmablasts and Ig (11). The expanded repertoire of V_H3⁺ antibodies bind via variant heavy chains to SpA or SpA_KK; however, all plasmablast-derived monoclonal antibodies examined failed to recognize pathogen-specific antigens (11). We demonstrate here that V_H3 clonal Ig expansion occurs in S. aureus-infected mice via a mechanism requiring the Fab binding sites of SpA IgBDs, namely aspartyl residues 36 and 37. Moreover, V_H3 clonal Ig expansion is dependent on peptidoglycan-linked SpA, which is released from the bacterial cell-wall envelope by staphylococcal murein hydrolases (22). In contrast, recombinant SpA (without attached peptidoglycan) does not trigger expansions of V_H3 clan Ig, likely because it fails to activate NOD1/NOD2 and RIPK2 signaling. Activation of V_H3 clonal B lymphocytes requires pathogen replication in infected host tissues and the activation of CD4⁺ T helper cells, confirming earlier observations of T cell-dependent Ig expansions in S. aureus-infected mice (45, 47). Antigen-engaged B cells migrate to follicular borders facing T cells. The resulting germinal centers support B-cell proliferation, diversification of Ig-variable sequences, and switching of antibody effector functions. V_H3 clonal B cells are similarly activated by protein A cross-linking of B-cell receptors, RIPK2 signaling, and MHC class II presentation to activate CD4⁺ T cells. Thus, S. aureus uses protein A to exploit mechanisms of B-cell development, thereby suppressing host-adaptive immune responses and enabling persistent colonization and recurrent infection.

Materials and Methods

Flow Cytometry.

For T-cell depletion, 6-wk-old C57BL/6 mice (n = 3) received CD4- or CD8α-specific monoclonal antibodies. After 5 d, animals were killed and their spleens were removed and homogenized. Red blood cells were lysed in ACK buffer (0.15 M NH₄Cl, 10 mM KHCO₃, 0.1 mM EDTA). White blood cells were washed and suspended in PBS and stained with allophycocyanin-conjugated α-CD3ε (145-2C11; eBioscience) and R-phycoerythrin-conjugated α-CD4 (RM4-5; Life Technologies), α-CD8α (5H10; Life Technologies), or CD19 (1D3; eBioscience) monoclonal antibodies. Cells were washed in 1% FBS/PBS and fixed with 4% (vol/vol) formalin and analyzed on a FACSCanto (University of Chicago flow cytometry core facility). Data were analyzed by FlowJo software.

In Vivo Mouse Experiments.

Overnight cultures of S. aureus strains were diluted 1:100 into fresh tryptic soy broth and grown for 2 h at 37 °C. Staphylococci were sedimented, washed, and suspended in PBS to the desired bacterial concentration. Inocula were quantified by spreading sample aliquots on TSA (tryptic soy agar) and enumerating CFUs. BALB/c or C57BL/6 mice (Charles River Laboratories and Jackson Laboratory) were anesthetized via i.p. injection with 65 mg·mL⁻¹ ketamine and 6 mg·mL⁻¹ xylazine per kg of body weight. Mice were infected by injection with 1 × 10⁷ CFUs of S. aureus Newman and its variants or 5 × 10⁶ CFUs of S. aureus USA300, MW2, or N315 into the periorbital venous sinus of the right eye. On days 5, 12, 14, 19, and 26 following infection, mice were anesthetized and bled either by heparin-coated capillary tubes from the periorbital venous sinus of the left eye or by cardiac puncture. For B-cell superantigen activity of rSpa or nSpA, A/J, BALB/c, C3H/HeJ, C57BL/6, CBA, DBA/2, or FVB/N mice (Jackson Laboratory) were treated by i.p. injection with 0.1–100 µg rSpA or nSpA on days 0, 5, 12, and 19. CD4- or CD8-specific T-cell depletion was achieved by i.p. injections of 0.5 mg purified rat monoclonal antibodies specific to mouse CD4 (GK1.5; Bio X Cell) or CD8α (53-6.72; Bio X Cell). Acute T-cell depletion was achieved by injecting monoclonal antibodies for 3 consecutive days, followed by repeated injections of the monoclonal antibodies at 3-d intervals. All mouse experiments were performed at least twice, and were conducted in accordance with the institutional guidelines following experimental protocol review and approval by the Institutional Biosafety Committee (IBC) and the Institutional Animal Care and Use Committee (IACUC) at the University of Chicago.

Protein A Expression in S. aureus.

Overnight cultures of staphylococci were diluted 1:100 and grown at 37 °C with shaking to A₆₀₀ 2. Fractionation of staphylococci into medium and cell-wall compartments followed a previously established procedure (29). Briefly, bacteria were centrifuged (8,000 × g for 10 minutes) and the extracellular medium in the supernatant was precipitated with 5% TCA. The pellet was suspended in TSM [50 mM Tris (pH 7.5), 500 mM sucrose, 10 mM MgCl₂ with 100 µg·mL⁻¹ lysostaphin] and incubated at 37 °C to solubilize the cell-wall envelope with lysostaphin. The resulting protoplasts were sedimented by centrifugation (8,000 × g for 10 minutes), and the supernatant was precipitated with TCA (cell-wall fraction). TCA-precipitated proteins were washed in acetone, dried, solubilized in sample buffer, and separated on SDS/PAGE. Proteins were electrotransferred to PVDF membrane and analyzed by immunoblotting using HRP-conjugated human IgM.

SI Materials and Methods

Bacterial Strains.

S. aureus strains Newman and its variants (spa_KK, spa_AA, spa_KKAA, spa_E, spa_ED, spa_EDA, spa_EDAB, spa_EDABCD, and Δspa), USA300 LAC, N315, and MW2 were grown in tryptic soy broth (TSB) or agar at 37 °C (24, 48). E. coli strains DH5α and BL21(DE3) were grown in Luria broth or agar at 37 °C. Ampicillin (100 µg·mL⁻¹ for E. coli) and spectinomycin (200 µg·mL⁻¹ for S. aureus) were used for plasmid selection (pET15b⁺) and mutant allele selection (Δspa).

S. aureus spa Mutants.

A detailed method of generating spa_KK, spa_AA, spa_KKAA, and Δspa mutant strains was described previously (24). Protein A mutant strains expressing different numbers of Ig-binding domains (IgBDs) were constructed by allelic recombination. To generate pKOR1 vectors for spa_E, spa_ED, spa_EDA, and spa_EDAB, three-step PCR was performed using S. aureus Newman genomic DNA. PCR products (PCR1) were amplified as DNA fragments with primers that bind upstream of the spa gene and 3′ of DNA sequences coding for IgBD E (spa_E), D (spa_ED), A (spa_EDA), or B (spa_EDAB) domains. A second PCR product (PCR2) was amplified with primers that anneal 5′ of region X coding sequence and 3′ of sequences downstream of the spa gene. The corresponding PCR products (PCR1 and PCR2) were annealed by PCR. To generate the pKOR1 vector for spa_EDABCD, a four-step PCR was performed using S. aureus Newman genomic DNA. The first PCR (PCR1) amplified a 1-kb DNA fragment with upstream spa sequences ending at the C IgBD. PCR2 amplified DNA sequences of IgBD D. PCR3 amplified sequences from region X and downstream of the spa gene. All three PCR products (PCR1–3) were annealed together by PCR. The final PCR product (PCR4) was cloned into the pKOR1 vector using the BP Clonase II Kit (Invitrogen). Primers used in this study are listed in Table S1.

Table S1.

List of primers used in this study

Gene	Reaction	Primer, 5′-3′
spaE	PCR1-F	GGGGACAAGTTTGTACAAAAAAGCAGGCCTTATTTATTTCATCAGCAAGAAAACACAC
	PCR1-R	CTTCCTCTTTTGGTGCTTGAGAGTCATTAAGTTTTTGAGCTTCACCTAAAACGTTAGC
	PCR2-F	GAAGCTCAAAAACTTAATGACTCTCAAGCACCAAAAGAGGAAGAC
	PCR2-R	GGGGACCACTTTGTACAAGAAAGCTGGGTAACGAACGCCTAAAGAAATTGTCTTTG
spaED	PCR1-F	GGGGACAAGTTTGTACAAAAAAGCAGGCCTTATTTATTTCATCAGCAAGAAAACACAC
	PCR1-R	CTTCCTCTTTTGGTGCTTGAGATTCGTTTAATTTTTTAGCTTCACCTAAAACGTTAGTG
	PCR2-F	GCTAAAAAATTAAACGAATCTCAAGCACCAAAAGAGGAAGAC
	PCR2-R	GGGGACCACTTTGTACAAGAAAGCTGGGTAACGAACGCCTAAAGAAATTGTCTTTG
spaEDA	PCR1-F	GGGGACAAGTTTGTACAAAAAAGCAGGCCTTATTTATTTCATCAGCAAGAAAACACAC
	PCR1-R	CTTCCTCTTTTGGTGCTTGAGATTCATTTAACTTTTTAGCTTCTGAC
	PCR2-F	CAGAAGCTAAAAAGTTAAATGAATCTCAAGCACCAAAAGAGGAAGAC
	PCR2-R	GGGGACCACTTTGTACAAGAAAGCTGGGTAACGAACGCCTAAAGAAATTGTCTTTG
spaEDAB	PCR1-F	GGGGACAAGTTTGTACAAAAAAGCAGGCCTTATTTATTTCATCAGCAAGAAAACACAC
	PCR1-R	CTTCCTCTTTTGGTGCTTGAGCATCATTTAGCTTTTTAGCTTCTGCTAAAAGGTTAGCG
	PCR2-F	AGCAGAAGCTAAAAAGCTAAATGATGCTCAAGCACCAAAAGAGGAAGAC
	PCR2-R	GGGGACCACTTTGTACAAGAAAGCTGGGTAACGAACGCCTAAAGAAATTGTCTTTG
spaEDABCD	PCR1-F	GGGGACAAGTTTGTACAAAAAAGCAGGCCTTATTTATTTCATCAGCAAGAAAACACAC
	PCR1-R	CATCCGCTTTTGGAGCTTGAGCATCGTTTAGCTTTTTAGCTTCTGCTAAAATTTC
	PCR2-F	GCTAAAAAGCTAAACGATGCTCAAGCTCCAAAAGCGGATGCGCAGCAAAAC
	PCR2-R	CTTCCTCTTTTGGTGCTTGCGATTCGTTAAGTTTTTTTGCTTCG
	PCR3-F	CAAAAAAACTTAACGAATCGCAAGCACCAAAAGAGGAAGAC
	PCR3-R	GGGGACCACTTTGTACAAGAAAGCTGGGTAACGAACGCCTAAAGAAATTGTCTTTG

F, forward; R, reverse.

Protein A Purification.

For recombinant protein A (rSpA) purification, E. coli BL21(DE3) harboring pET15b⁺ plasmids for the expression of His₆-tagged wild-type SpA was grown overnight, diluted 1:100 into fresh medium, and grown at 37 °C to A₆₀₀ 0.5. Cultures were induced with 1 mM isopropyl β-d-1-thiogalatopyranoside (IPTG) and grown for an additional 3 h. Bacterial cells were sedimented by centrifugation (8,000 × g for 10 minutes), suspended in column buffer (50 mM Tris⋅HCl, pH 7.5, 150 mM NaCl), and disrupted with a French pressure cell at 14,000 psi. Lysates were cleared of membrane and insoluble components by ultracentrifugation at 40,000 × g. Cleared lysates were subjected to nickel-nitrilotriacetic acid (Ni-NTA) affinity chromatography, and proteins were eluted in column buffer containing successively higher concentrations of imidazole (100–500 mM). For purification of native protein A (nSpA) released during bacterial replication, S. aureus sbi::erm was grown overnight, diluted 1:100 into fresh medium, and grown at 37 °C to A₆₀₀ 0.4. Culture supernatant devoid of bacterial cells was collected after centrifugation at 8,000 × g, followed by filtration through 0.22-µm membrane (Millipore). The culture supernatant was subjected to human IgG affinity chromatography (IgG Sepharose 6 Fast Flow; GE Healthcare Life Sciences) and proteins were eluted with elution buffer (0.1 M glycine⋅HCl, pH 3.0), followed by immediate neutralization with 1 M Tris⋅HCl (pH 8.5). Eluted antibody samples were dialyzed with PBS. Protein concentrations were determined by bicinchoninic acid (BCA) assay (Thermo Scientific). Purity was verified by Coomassie-stained SDS/PAGE.

Enzyme-Linked Immunosorbent Assay.

To determine antigen-specific SpA_KK- and SpA_KKAA-specific antibodies, affinity-purified recombinant SpA_KK and SpA_KKAA were used to coat ELISA plates (NUNC MaxiSorp) at 1 µg·mL⁻¹ in 0.1 M carbonate buffer (pH 9.5) at 4 °C overnight. The following day, plates were blocked and incubated with serially diluted sera. Plates were incubated with HRP-conjugated secondary antibody specific to mouse IgG, IgG₁, IgG_2a, IgG_2b, or IgM and developed using OptEIA reagent (BD Biosciences). Half-maximal titer was calculated using GraphPad Prism software. Fold change was calculated by normalizing half-maximal titer of the experimental sample by half-maximal titer of the control sample (naive mouse serum or isotype controls). To determine staphylococcal antigen-specific antibodies, 12 staphylococcal antigen matrix recombinant proteins (Coa, ClfA, ClfB, EsxA, EsxB, Hla, IsdA, IsdB, LukD, SdrC, SdrD, and SdrE) were used to coat ELISA plates (NUNC MaxiSorp) at 1 µg·mL⁻¹ in 0.1 M carbonate buffer (pH 9.5) at 4 °C overnight. The following day, plates were blocked and incubated with 10 µg of affinity-purified mouse IgG [SpA_KK⁺ (V_H3⁺) or SpA_KK⁻ (V_H3⁻)] generated from mice infected with S. aureus Newman or spa_KKAA. Plates were incubated with HRP-conjugated secondary antibody specific to mouse IgG and developed using OptEIA reagent.