Anti-AdcA serum mediates Opsonophagocytic killing of different S. aureus strains

Given the crucial role of the surface-exposed lipoprotein AdcA in S. aureus (which shares with AdcA of E. faecium a sequence identity of 66%) and the important challenge of addressing vaccine development against MRSA32,33, we tested whether antibodies raised against AdcA from E. faecium were also able to mediate opsonophagocytosis in three different S. aureus strains including MW2, Reynolds and LAC. We chose MW2 and Reynolds strains because they express dominant polysaccharide capsule, CP5 and CP8 respectively, among clinical isolates. Also S. aureus USA300 (LAC), which lacks a capsule, is prevalent in the United States34. To obtain antibodies, New Zealand rabbits were previously immunised via intramuscular injection with recombinantly expressed AdcA from E. faecium and sera containing anti-AdcA antibodies were collected as described (Fig. 1A, B)18. In an opsonophagocytic killing assay (OPKA) we observed that anti-AdcA was, indeed, opsonic against all tested S. aureus strains, with percentages of opsonophagocytic killing mediated by the anti-AdcA serum in the range 40–60%, compared with 0–20% killing of the pre-immune sera (Fig. 1C–E). To assess whether antibodies that mediate the opsonic killing of S. aureus are specific towards AdcA, we incubated the anti-AdcA sera with different amounts of recombinant AdcA (opsonophagocytic killing inhibition assay, OPIA)(Fig. 1F, G). As shown in Fig. 1G, the opsonophagocytic killing of S. aureus MW2 mediated by the anti-AdcA serum was inhibited in a dose-dependent manner by the purified AdcA of E. faecium.

Fig. 1: Opsonophagocytic killing activity (OPKA) of anti-AdcA serum against different S. aureus strains.

A schematic representation of rabbit immunization (A) and mechanism of OPKA (B). Antigen-specific antibodies and complement proteins opsonise bacteria and facilitate the uptake of the antibody-bacteria complex by phagocytes (PMN). C OPKA against S. aureus strains MW2, (D) LAC, and (E) Reynolds were tested with sera raised against the recombinant AdcA from E. faecium, used at total IgG concentrations ranging from 1 to 0.5 mg/mL (grey bars). The effectiveness of opsonophagocytic killing by the anti-AdcA rabbit sera (MW2 - C, LAC – D, and Reynolds - E) was compared to that by the pre-immune rabbit sera (white bars). Statistical significance was tested by the unpaired two-tailed T test with a 95% confidence interval with a Bonferroni post hoc test using pre-immune and terminal immune sera at the same concentration. Bars and whiskers denote mean values ± standard errors of the mean of replicates within one assay. *P ≤ 0.025, **P ≤ 001, ***P ≤ 0.001. F Schematic representation of an OPIA assay. Antibodies, once engaged by their antigens (inhibitors) are unable to opsonise, leading to a decrease in bacterial killing. G OPIA of anti-AdcA rabbit serum performed by incubation of serum with recombinant AdcA protein and tested against S. aureus MW2. Antibodies raised against the AdcA at 0,5 mg/mL of total IgG concentration were pre-incubated with different amounts of AdcA (black bars) in a range from 0.14 to 3.56 µM. Pre-immune serum (P) at 0,5 mg/mL of total IgG concentration was used as a negative control, and anti-AdcA sera (C) without incubation with AdcA, at 0.5 mg/mL of total IgG concentration was used as a positive control. Statistical significance of inhibition was performed by the One-way analysis of variances test, followed by the Dunnett’s multiple comparison post hoc test. Bars and whiskers denote mean values ± standard errors of the mean of replicates within one assay. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

Structure-based identification of AdcA epitope

The observed cross-reactivity of E. faecium AdcA sera binding to both E. faecium and S. aureus, prompted us to investigate the molecular determinants responsible for antigenicity of this protein. As shown in Fig. 2A, three distinct regions can be identified in the AdcA (protein ID: WP_002297324.1) sequence: a signal peptide at the N-terminus, a periplasmic Zinc-uptake complex component A (ZnuA) and a lipocalin-like zinc-recruitment domain (ZinT), as detected by sequence analysis in the PFAM (Protein Families Database) database35 (Fig. 2A). As a first step for structure-based epitope identification approach, we homology modelled the structure of AdcA, using the programme MODELLER36 after consensus-based sequence alignment with HHPred37 (Fig. 2). In AdcA homology model, the N-terminal ZnuA domain and the C-terminal ZinT domain are connected by a long loop region, embedded between residues 314–330 (Fig. 2A, B). The ZnuA domain is formed by two lobes adopting (β/α)4 folds and connected by a long α−helix (residues 169–199), whereas the ZinT domain adopts an 8-stranded β-barrel structure terminating with a helical region of four short helices (Fig. 2B). Each domain presents a cluster of His-residues which is putatively involved in Zn binding (Fig. 2B). Consistently, they were both shown to bind Zn in the homologue of AdcA from S. pneumoniae; albeit being the ZnuA domain necessary and sufficient for Zn acquisition38.

Fig. 2: Structural information on AdcA from E. faecium and epitope identification.

A Domain composition of AdcA based on PFAM analysis. B Cartoon representation of the homology model of full-length AdcA, computed with MODELLER using the structure of AdcA from S. pneumoniae (PDB code: 7jj9, sequence identity 65%) as a template. The ZnuA domain, formed by Lobe 1 (light grey) and Lobe 2 (light blue) is connected to the ZinT domain (violet) through a loop, represented in green. The two zoom panels report putative Zn-binding histidines, represented in stick, in each domain. C Epitope prediction using BepiPred (top) and Discotope (bottom); the region with highest score is coloured blue. D Cartoon and stick representation of the most antigenic region, here named as EH-motif, located in Lobe 1 of the ZnuA domain.

Preliminary sequence-based B-cell epitope prediction, using Bepipred339, identified the strongest immunogenic regions on the ZnuA domain (Fig. 2C). Structure-based epitope prediction using Discotope-2.040 (Fig. 2C) and ElliPro41 confirmed this region as the most immunogenic of the AdcA structure. The best-predicted antigen is located on a large and flexible loop between α4 and β4 secondary structure elements of the ZnuA domain, between residues G125-E141 (Fig. 2D). Antigenicity, allergenicity, and toxicity, further analysed with VaxiJen v2.042, AllerTop 2.0/AllergenFP1.043,44, ToxinPred45, showed that the best epitope 125-GSEEEDHDHGEEDHHHE-141 is also predicted to be non-allergenic and non-toxic (Table 1). Due to the abundance of glutamic acid (E) and histidine (H) in this loop sequence, we denoted it as EH-motif (Fig. 2D).

Table 1 Prediction of the antigenicity, allergenicity, and toxicity of AcdA portionsThe EH-motif sequence reacts with anti-AdcA antibodies and partially inhibits the opsonic killing elicited by the anti-AdcA serum

Prompted by the bioinformatic prediction of the EH-motif as a dominant epitope, we tested whether antibodies raised against AdcA were able to bind this specific peptide sequence. To this end, we obtained two synthetic versions of the epitope, one with biotinylation in the N-terminal part of the peptide designated N-biot-EH (K(Biot)GSEEEDHDHGEEDHHHE) and one with biotinylation in the C-terminal part of the peptide designated C-biot-EH (GSEEEDHDHGEEDHHHEK(Biot)). Both peptides were applied to streptavidin-coated ELISA plates. As shown in Fig. 3A, both peptides were recognised by the antibodies raised against full-length recombinant AdcA, with C-biot-EH showing a higher affinity for the anti-AdcA serum, compared with the N-biot-EH (Fig. 3A). Therefore, we tested if these peptides were able to inhibit the opsonic killing mediated by anti-AdcA antibodies against both E. faecium and S. aureus. To assess selectivity of antibodies, we performed OPIA experiments, where we preincubated the anti-AdcA serum with either C-biot-EH and N-biot-EH peptides and with the two positive controls, the full length recombinant AdcA and the ZnuA domain. We observed that both peptides were able to inhibit the opsonic activity of the anti-AdcA antibodies against the S. aureus MW2 (Fig. 3B–E) and E. faecium 11236/1 (Fig. 3F–I) in a concentration dependent manner. As expected, the observed inhibition of opsonic killing due to EH-peptides was lower than that observed for the full-length AdcA protein, since the tested serum was originally raised against full-length AdcA.

Fig. 3: The EH-motif is recognised by antibodies raised against full-length recombinant AdcA and inhibit the opsonic killing mediated by anti-AdcA antibodies against both E. faecium and S. aureus.

A Immuno-reactivity towards the EH-peptides detected by enzyme-linked immunosorbent assays (ELISA). Either the N-biot-EH (black) or C-biot-EH (blue) were added to streptavidin coated plates and tested by ELISA with pre-AdcA or anti-AdcA sera. B–E Inhibition of opsonophagocytic killing activity of anti-AdcA 0.5 mg/mL of total IgG concentration rabbit serum against S. aureus MW2 and (F-I) E. faecium 11236/1 by addition of EH-peptides. Antibodies raised against the recombinant AdcA were pre-incubated with the different inhibitors (in the concentration range 0.14 to 3.56 µM), including recombinant AdcA (white), ZnuA domain (grey), C-biot-EH (light blue) or N-biot-EH (blue). As positive control (C) we used sera anti-AdcA 0.5 mg/mL of total IgG concentration with no inhibitor (black). Names of tested strains are visible below corresponding assays. Statistical significance was computed by the One-way analysis of variances test, followed by the Dunnett’s multiple comparison post hoc test. Bars and whiskers denote mean values ± standard errors of the mean of replicates within one assay. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

The best AdcA epitope, the EH-motif, is highly conserved among multiple species and suitable as a cross-reactive antigenic sequence

To design a broadly cross-reactive antigen, we performed a thorough sequence analysis to evaluate the level of antigen conservation among different strains of Enterococcus sp, and among different pathogenic bacteria from the WHO list. Sequence dataset collection and the following alignment were performed using BlastP46 and Clustal Omega EMBL-EBI47 software, respectively.

First, we searched homologues of the ZnuA domain of AdcA among multiple bacterial species. Using a sequence identity (seqid) threshold of 60%, we identified 157 sequences in E. faecium isolates (taxid: 1352), and further 318 sequences in Enterococcus sp. (taxid: 1350). Based on the promiscuity of ZnuA-like domains of Zn-acquiring proteins in Enterococcus sp., we decided to broaden our investigation to include closely related Streptococcus sp. (taxid: 1301), where we identified 549 homologous sequences. The same analysis conducted on ZnuA domains of S. aureus (taxid: 1280) showed that a lower albeit significant sequence identity with ZnuA of E. faecium AdcA. By lowering the seqid threshold to 30%, a value that still identifies similar structural folds, we could detect 627 sequences in S. aureus (Supplementary data 1–4).

Using the sequence sets obtained for each bacterial isolate, we then checked the intra-species conservation of the EH-motif. In E. faecium, the seqid of the EH-motif is 80% when computed on all 157 sequences, of which 89 sequences present an almost fully conserved EH-motif (seqid 99%, Fig. 4). Almost all ZnuA sequences of Enterococcus sp. (303 out of 318) present an EH-motif with small or no variations (sequence identity 87%, Fig. 4), indicating strong conservation of the EH-motif also in clinically relevant E. faecalis isolates. In Streptococcus sp., 548 out of 549 sequences had a highly conserved EH-motif (seqid 81%, Fig. 4). Interestingly, 404 out of 627 sequences with seqid 30% identified for S. aureus, embed a highly conserved EH-motif (seqid 97.5%, Fig. 4). These sequences contain the S. aureus strains MW2, LAC and Reynolds, for which we have experimentally proven an opsonic activity of anti-AdcA antibodies (Fig. 1).

Fig. 4: The EH-motif is conserved in a wide spectrum of Gram-positive bacteria.

Computational analysis of (A) EH-motif conservation in different bacterial species, and (B) pairwise alignment of consensus sequences of each dataset in (A) with the EH-motif of AdcA of E. faecium. Only regions with the highest Waterman-Eggert score are shown (computed with Lalign).

To evaluate the inter-species conservation of the EH-motif belonging to E. faecium AdcA, we computed the consensus sequence of each species, using the Jalview48, and then aligned this consensus with the sequence of AcdA EH-motif. As shown in Fig. 4B, the EH-motif sequence shares 100% seqid with the consensus computed for E. faecium, consistent with the high conservation of this motif in E. faecium family. Almost full conservation of the EH-motif is also observed with the consensus sequence of streptococci (seqid 93.3%), Fig. 4B. Lower sequence identities, still higher than 60%, are observed with S. aureus (Fig. 4B). Full datasets, a list of bacteria and alignments are available in supplementary data. These data predict a good efficacy of the EH-motif sequence as an antigen not only against multiple E. faecium strains, but against all the described bacterial species.

Design of a hyper-stable, multi-presenting, and cross-reactive scaffolded antigen

Once we had established that the EH-motif is strongly immunogenic and promiscuous among different pathogens, we focused on designing a hyper stable protein embedding multiple copies of this peptide antigen. As a vehiculating molecule, we chose the domain D1 (residues 1-210, PDB code 6gpc) of the highly stable arginine binding protein (ArgBP) from T. maritima (PDB code 4prs) The rationale for the choice of this domain as an accepting scaffold for the antigenic EH-motifs mainly relies on its extraordinary stability to temperature and chemical denaturants49, also when dissected from the entire protein50, a characteristic that allows structural manipulations albeit preserving the protein conformational stability. Also, the D1 domain of ArgBP presents more loops to host the antigenic motifs, thus potentially acting as a multi-antigen presenting molecule. Molecular modelling sessions were performed to identify the proper locations of the EH-motif to prevent the destabilisation of the protein structure. The D1 domain has a compact α/β fold, whose main body is formed by 5 α helices surrounding a central β sheet. The structure contains two β-hairpin motifs protruding from the main fold and embedding loop regions between Asp26 and Asn70 and between Ala95 and Glu102 (Supplementary Fig. 1). Due to their flexibility and outward protrusion, we identified these loops as more amenable for antigen insertion, without inducing protein distortions (Supplementary Fig. 1). Therefore, we designed a chimeric protein with the D1 domain as accepting scaffold (Sc) and the EH-motif between E27 and N28, G97 and G98 and at the N-terminal side of the protein (Fig. 5A, Supplementary Table 1), here designated as Sc(EH)3. The obtained model was used for molecular dynamics (MD)51. DiscoTope analysis52 on the minimised model shows that the sole predicted antigenic regions of Sc(EH)3 are the inserted EH-motifs (Fig. 5B, Supplementary Fig. 1).

Fig. 5: Design, validation, production and characterisation of a hyper stable multi-presenting antigen.

A Molecular model of Sc(EH)3 after energy minimisation. B DiscoTope predictions of Sc(EH)3, showing predicted antigenicity solely for the inserted EH-motifs. C Root-mean-square deviations (RMSD) computed on Cα atoms, reported for either the core or the EH-motifs residues for the three MD simulations at three increasing velocities in black, red, and green. D Root-mean-square fluctuations (RMSF) on Cα atoms showing high flexibility solely for the three EH-motifs. E Far-UV CD spectra were measured at 0.2 mg mL−1 in 20 mM sodium phosphate buffer (pH 7.4) after incubation for 1 h at 20 °C (green), 37 °C (blue) and 100 °C (black). F Thermal denaturation curves of Sc(EH)3 monitored at 222 nm in 20 mM sodium phosphate buffer (black) and in the presence of 3 M GuHCl denaturant (red). G Resistance to storage at 37 °C for 30 days of Sc(EH)3 compared to AdcA, as shown by SDS-gel electrophoresis. Degradation bands of AdcA are highlighted by arrows.

MD dynamics was used to validate the computationally designed model through the analysis of its structural and dynamic features. MD outcomes are generally important for model quality estimation, an important and propaedeutic step to experimental work53. We run three MD trajectories, each of 500 ns, with different starting velocities to improve the conformational ensemble search and extensively explore the dynamic features of the protein upon insertion of EH-motifs. The root-mean-square deviation (RMSD) profiles of Sc(EH)3, computed on Cα atoms with the respect to the starting model, suggests that the simulated system consistently reaches an equilibrium phase in about 100 ns in the three parallel runs, showing comparable mean and standard deviations of RMSD values (Supplementary Fig. 2). To dissect conformational changes of the accepting scaffold from those of the inserted EH-motifs, we computed RMSD values of these isolated protein core and inserted EH-motifs (Fig. 5C). It appears clear that Sc(EH)3 presents as a compact and stable core with low and stable RMSD values, whereas the EH-motifs show large conformational variations (Fig. 5C). The high dynamic nature of EH-motifs is also confirmed by the root mean squared fluctuations (RMSF) of the relative residues (Fig. 5D), which show values in the range from 5 to 15 Å, while the RMSF of all other residues remains below 5 Å. The MD analyses suggest that the EH-motifs are a region of peculiar flexibility, whereas the stability of the protein core region is well preserved, with no distortion in the connecting regions. The high flexibility in antigenic sites is a good premise, since flexible epitopes can elicit a diverse antibody response through their conformational adaptation during the binding event to multiple antibodies54. After validation of Sc(EH)3 model, we sub-cloned the synthetic gene encoding for its amino acid sequence into pETM-13, expressed and purified in high yields, as monitored by SDS-PAGE electrophoresis. Light Scattering studies confirmed that Sc(EH)3 is a monomer in solution, with a weight-average molar mass of 22.0 ± 0.1 kDa (Supplementary Fig. 3A). Of note for vaccine development, Sc(EH)3 is extremely thermo-resistant and long-term stable at 37 °C. Indeed, circular dichroism (CD) spectroscopy shows a fully conserved spectrum of the protein after 1 h incubation at 37 °C, compared to that at 20 °C (Fig. 5E). Also, Sc(EH)3 is still endowed with a secondary structure content after incubation at 100 °C (Fig. 5E). Consistently, thermal unfolding, analysed by recording the CD signal at 222 nm, shows that the denaturation transition is not complete at 100 °C (Fig. 5F). Full unfolding could be achieved in the temperature range 20–100 °C only in the presence of a chemical denaturant, specifically 3 M guanidine hydrochloride (GuHCl), with a melting temperature (Tm) of 52 °C (Fig. 5F). As shown in Fig. 5G, Sc(EH)3 is also highly resistant to proteolysis, as it can be stored for at least 30 days at 37 °C with no sign of degradation. In the same conditions, AdcA exhibits a visible degradation pattern, with the accumulation of two major degradation bands in the SDS-PAGE gels (Fig. 5G, Supplementary Fig. 3B, C).

Sc(EH)3 elicits antibodies that generate strong and specific opsonic killing against E. faecium, S. aureus and E. faecalis

The Sc(EH)3 protein was used to immunise two female New Zealand white rabbits, which were exsanguinated two weeks after the last injection (see “Methods”). To generate a serum that represents an average immune response, the terminal bleeds of both rabbits were mixed in equal volumes, generating the anti-Sc(EH)3 serum. The same procedure was applied to the pre-immune sera, used as a negative control. As a preliminary test, we evaluated the ability of the anti-Sc(EH)3 antibodies to specifically recognise C-biot-EH by ELISA. We found that the anti-Sc(EH)3 could specifically bind to the epitope up to a dilution of 1:4000 whereas the pre-Sc(EH)3 did not show any binding at dilutions as high as 1:250 (data not shown). This confirmed that the anti-Sc(EH)3 serum contained antibodies directed against the EH-motif. We then used OPKA to test if the anti-Sc(EH)3 antibodies can mediate opsonic killing of E. faecium 11236/1, S. aureus MW2 and E. faecalis T2. As shown in Fig. 6A, the anti-Sc(EH)3 serum antibodies, in a concentration of 0.4 mg/mL, were significantly better at inducing opsonophagocytic-dependent E. faecium 11236/1 killing, compared to anti-AdcA sera. We observed an average opsonic killing of 67% at 0.4 mg/mL IgG dilution, compared to 46% of sera raised against full-length AdcA. Importantly, Sc(EH)3 was also able to induce opsonic killing of S. aureus MW2 (Fig. 6B) and E. faecalis T2 (Fig. 6C). In the case of S. aureus MW2, significantly higher killing activity than anti-AdcA was observed in particular at lower concentrations of anti-Sc(EH)3 sera, with 48% killing of anti-Sc(EH)3 at 0.2 mg/mL compared to 32% in the case of anti-AdcA. Even at 0.05 mg/mL, anti-Sc(EH)3 sera were significantly more effective (21% killing), compared anti-AdcA sera (6% killing) (Fig. 6B). As for E. faecalis T2, sera raised by both AdcA and Sc(EH)3 were able to kill E. faecalis T2, with a maximum killing of about 70% at the maximum antibody concentration of 0.4 mg/mL (Fig. 6C). Higher killing was induced by anti-Sc(EH)3 compared to anti-AdcA at lower antibody concentration (44% compared to 30% at 0.2 mg/mL) (Fig. 6C). These data prove that both AdcA and Sc(EH)3 act as cross-reactive antigens. However, Sc(EH)3 elicits antibodies that generate the strongest opsonic killing against all tested Gram positive bacteria.

Fig. 6: OPKA and OPIA assays against E. faecium 11236/1, S.aureus MW2 and E. faecalis T2.

OPKA assays against (A) E. faecium 11236/1, (B) S. aureus MW2, and (C) E. faecalis T2. OPKA killing percentages mediated by anti-Sc(EH)3 sera are compared to those mediated by anti-AdcA (black bars). The statistical significance was tested by the unpaired two-tailed T test with a 95% confidence interval with a Bonferroni post hoc test using anti-AdcA and anti-Sc(EH)3 sera at the same concentration. Bars and whiskers denote mean values ± standard errors of the mean. *P ≤ 0.0125, **P ≤ 0.01, ***P ≤ 0.001. OPIA assays with 0.4 mg/mL anti-AdcA or anti-Sc(EH)3 sera against E. faecium 11236/1 (D–G) S. aureus MW2 (H–K) and E. faecalis T2 (L–O) by addition of different inhibitors. Anti-AdcA and anti-Sc(EH)3 were pre-incubated with different concentrations of inhibitors, including recombinant AdcA (light blue bars), ZnuA domain (light grey bars), and recombinant Sc(EH)3 (dark grey bars). Anti-AdcA sera incubated with AdcA were used as positive controls (D, H, L). As negative controls (0) for each plot, we used sera at 0.4 mg/mL anti-AdcA (dark blue bar) and anti-Sc(EH)3 (black bars) without inhibitors. Statistical significance of inhibition was performed by the One way analysis of variances test, followed by the Dunnett’s multiple comparison post hoc test. Bars and whiskers denote mean values ± standard errors of the mean of replicates within one assay. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

To verify the specificity of opsonising antibodies against the EH-motifs of Sc(EH)3, opsonophagocytic inhibition assays were carried out by pre-incubating the anti-Sc(EH)3 sera with different inhibitors at different concentrations, including Sc(EH)3, AdcA, or ZnuA. Anti-AdcA sera incubated with AdcA protein were used as a control (Fig. 6D–F). In all cases, we observed a concentration dependent reduction of killing induced by sera, indicating specificity of antibodies against each added antigen. Specifically, for E. faecium 11236/1 we observed a reduction in the killing of anti-Sc(EH)3 antibodies, at the maximum concentration used, by 54%, 41%, and 53% upon incubation with AdcA, ZnuA, and Sc(EH)3, respectively (Fig. 6D). In the case of S. aureus MW2, killing reduction at the highest inhibitor concentration was 49%, 51%, and 65% for AdcA, ZnuA and Sc(EH)3, respectively (Fig. 6E). Similar OPIA results were obtained for E. faecalis T2, showing that sera-mediated killing was significantly inhibited by all proteins in a concentration dependent manner (Fig. 6F), thus confirming the specific recognition of antibodies for the EH-motif, which is common to all inhibitors.

Estimation of the LD50 of the different used strain on Balb/C mice

To determine the lethal dose of each strain, different bacterial concentrations were injected into mice and the mortality rate was recorded for each group. We defined an approximate LD50 around 5 × 109 CFU for each bacterium since we recorded 50% of mortality between 3 × 109 and 6 × 109 CFU (Supplementary Table 2).

Anti-Sc(EH)3 and anti-AdcA antibodies are protective in mouse infection models

To determine whether anti-Sc (EH)3 and the anti-AdcA antibodies confer protection to mice against Gram-positive bacterial infections due to bacterial strains E. faecium 11236/1, faecalis T2, and S. aureus MW2, mice were passively immunised in a lethal bacteremia mouse model (Supplementary Fig. 4). As shown in Fig. 7A, after 96 h of bacterial challenge with 8 × 109 CFU of E. faecium 11236/1, a full protection in the group immunised with Sc (EH)3 antibodies (0% of mortality) and a significant decrease of mortality in the group immunised with anti-AdcA antibodies (30%) were observed compared to 90% of mortality in the untreated control group (p < 0.0001 and p < 0.05, respectively). For E. faecalis T2 strain, after 96 h of bacterial challenge with 6 × 109 CFUs, no mortality was observed in the mice immunised with either Sc (EH)3 or AdcA antibodies, compared to 60% of mortality in the control group (p < 0.001) (Fig. 7B). For S. aureus MW2, after 96 h of bacterial challenge with 9 × 109 CFU, a statistically significant decreased mortality was observed in the group immunised with Sc(EH)3 antibodies (20% of mortality) but not significant with anti-AdcA (70% of mortality) compared to 100% mortality in the control group (p = 0.001 and p = 0.06, respectively).

Fig. 7: In vivo passive protective efficacy of anti-Sc (EH)3 and anti-AdcA polyclonal antibodies against E. faecium 11236/1, E. faecalis T2 and S. aureus MW2.

A The percentage of survival mice during the protection assay against 8 × 109 CFU of E. faecium 11236/1. The discontinuous blue line represents the group treated with the anti-Sc (EH)3 antibody (n = 10), the discontinuous green line represents the group treated with the anti-AdcA antibody (n = 10) and the discontinuous black line represents the negative control group with mice receiving inactivated baby rabbit sera (BRS) (n = 10). B The percentage of survival mice during the protection assay against 6 × 109 CFU of E. faecalis T2. The discontinuous blue line represents the group treated with the anti-Sc (EH)3 antibody (n = 10), the discontinuous green line represents the group treated with the anti-AdcA antibody (n = 10) and the discontinuous black line represents the negative control group with mice receiving inactivated baby rabbit sera (BRS) (n = 10). C The percentage of survival mice during the protection assay against 9 × 109 CFU of S. aureus MW2. The discontinuous blue line represents the group treated with the anti-Sc (EH)3 (n = 10), the discontinuous green line represents the group treated with the anti-AdcA (n = 10) and the discontinuous black line represents the negative control group with mice receiving inactivated baby rabbit sera (BRS) (n = 10). Statistical differences between groups were assessed using Log Rank (Mantel-cox test). Difference between the treated groups with anti-Sc (EH)3 antibody and the control groups are marked by star symbol, ns: not significant; *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. Difference between the treated groups with anti-AdcA antibody and the control groups are marked by hash symbol, ns: not significant; #: p < 0.05, # #: p < 0.01, # # #: p < 0.001 and # # # #: p < 0.0001.

Active immunisation with Sc(EH)3 and AdcA antigens induces protective antibodies in mouse infection models

To demonstrate that Sc(EH)3 and AdcA antigens stimulate the immune system by inducing protective antibodies, we performed active immunisation with four doses of 10 µg of antigen, as detailed in Supplementary Fig. 5. We first performed a whole bacterial ELISA to quantify the levels of specific IgG antibodies against three Gram positive pathogens (E. faecium 11236/1, S. aureus MW2 and E. faecalis T2) in the sera of mice immunised with Sc(EH)3 and AdcA antigens, after 2 subcutaneous doses and after 2 intraperitoneal doses (Supplementary Fig. 5). Higher levels of antibodies were found both in the sera of mice immunised with Sc(EH)3 and AdcA antigens (Fig. 8A, B, respectively), compared to non-immunised groups, already after the 2 subcutaneous doses. Antibody levels increased after 4 weeks of immunisation, following the two further intraperitoneal injections (Fig. 8A, B). These results show that the both the immunisation with Sc(EH)3 and AdcA antigens induced a strong immune response and triggered the production of antibodies. Then, mice were challenged with intraperitoneal injections of lethal doses (Supplementary Fig. 5) of three tested Gram-positive bacteria.

Fig. 8: In vivo active protective efficacy of Sc(EH)3 and AdcA antigens against E. faecium, E. faecalis and S. aureus.

A IgG responses of mice immunised with Sc(EH)3 antigen in each group (S. aureus MW2, E. faecium 11236/1 and E. faecalis T2) (n = 10 per group), after two subcutaneous doses (after 2 weeks) and two intraperitoneal doses (after 4 weeks), compared to the control group receiving only buffer and IFA (n = 10). B As in panel (A), IgG responses of mice immunised with AdcA antigen+IFA (n = 10 per group) compared to the control group (n = 10). C Percentage of survived mice during the active protection assay against 9 × 109 CFU of E. faecium 11236/1, (D) against 7 × 109 CFU of S. aureus MW2 and (E) against 6.5 × 109 CFU of E. faecalis T2. The discontinuous blue line represents the group treated with Sc(EH)3 antigen (n = 10), the green line the group treated with AdcA antigen (n = 10) and the black line the untreated control group (n = 10). Statistical differences between groups were assessed using Log Rank (Mantel-cox test). Difference between the treated groups with Sc(EH)3 antigen and the control groups are marked by star symbol, ns: not significant; *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. Difference between the treated groups with AdcA antigen and the control groups are marked by hash symbol, ns: not significant; # : p < 0.05, # # : p < 0.01, # # # : p < 0.001 and # # # # : p < 0.0001.

As shown in Fig.