Identification, cloning and characterization of SpEX exotoxin produced by Staphylococcus pseudintermedius

Staphylococci have evolved numerous strategies to evade their hosts’ immune systems. Some staphylococcal toxins target essential components of host innate immunity, one of the two main branches of the immune system. Analysis of the Staphylococcus pseudintermedius secretome using liquid chromatography mass spectrometry guided by genomic data, was used to identify an S. pseudintermedius exotoxin provisionally named SpEX. This exoprotein has low overall amino acid identity with the Staphylococcus aureus group of proteins named staphylococcal superantigen like proteins (SSLs) and staphylococcal enterotoxin- like toxin X (SEIX), but predictive modeling showed that it shares similar folds and domain architecture to these important virulence factors. In this study, we found SpEX binds to complement component C5, prevents complement mediated lysis of sensitized bovine red blood cells, kills polymorphonuclear leukocytes and monocytes and inhibits neutrophil migration at sub-lethal concentrations. A mutant version of SpEX, produced through amino acid substitution at selected positions, had diminished cytotoxicity. Anti-SpEX produced in dogs reduced the inhibitory effect of native SpEX on canine neutrophil migration and protected immune cells from the toxic effects of the native recombinant protein. These results suggest that SpEX likely plays an important role in S. pseudintermedius virulence and that attenuated SpEX may be an important candidate for inclusion in a vaccine against S. pseudintermedius infections.


Introduction
Staphylococcus pseudintermedius is the main cause of canine dermatological disease and has been isolated from wound and surgical site infections, endocarditis and mastitis in dogs [1,2]. Human infections with this organism have been reported sporadically, most of which have developed in clinically healthy dogs and the neutralizing activity of those antibodies was evaluated.

Identification of putative immune modulator from S. pseudintermedius
Liquid chromatography-mass spectrometry (LC-MS/MS) was used to screen culture supernatants of three clinical strains of S. pseudintermedius (06-3228, 08-1661 and NA45), representing the major S. pseudintermedius genotypes occurring in the United States as determined by multilocus sequence typing (8). We identified over 500 secreted or cell wall associated proteins and others with unknown cellular locations (manuscript under review). The secretome proteins were compared among 06-3228, 08-1661 and NA45 strains using their respective genomes as references. SpEX, a 234-aa protein, was detected in the supernatants of the three isolates. The mean predicted molecular weight was 26.093 kDa with a mean isoelectric point of 6.30.

Bioinformatics analysis of S. pseudintermedius SpEX
A BLAST search of SpEX in the GenBank database revealed that it is conserved among S. pseudintermedius strains including 06-3221, 08-1661 and NA45 with amino acid identities over 88% between strains. SpEX shares sequence similarity with other pathogenic bacterial species within the Staphylococcus intermedius group. It shares amino acid identity of 71.8% with Staphylococcus intermedius, 73.5% with Staphylococcus delphini and 72.2% with Staphylococcus cornubiensis. No similar proteins were identified in S. aureus. SpEX has approximately 47% amino acid sequence identity with S. aureus SSL11 and less than 30% amino acid identity with other SSL members and SEIX. Analysis of SpEX secondary and tertiary domain structures using Geneious 11.0.3 software [24] supported by a SpEX model developed with the Phyre 2 web portal showed that S. pseudintermedius SpEX share fold and domain architecture with S. aureus SSLs and SEIX (Fig 1A and  1B). The SpEX protein sequence contains a signal peptide sequence from positions 1 through 35 detected using SignalP 4.1 [25]. SpEX has an N-terminal OB domain in residues 43-126 that folds into a five-stranded beta-barrel structure consisting of β-strands (β1-β5) and a Cterminal β grasp motif in residues 150-234 ( Fig 1A). This comprises a β sheet formed from β7, β6, β12, β9 and β10 with a characteristic central helix orientated diagonally in the center of the field from the top left to the bottom right leaving the N-terminal domain on the left and C-terminal domain on the right (Fig 1B).
S. aureus SSL4 (accession number: WP_000705627.1), SSL5 (accession number: WP_000784244.1), SSL11 (accession number: WP_000769163.1) and SEIX (accession number: AEI60186.1) proteins have been shown to bind sialated glycan through a conserved sialated glycan binding domain located at the C-terminal β grasp domain [19,20,26,27]. It has the capacity to influence the host's innate immunity defenses through targeting sialated glycoproteins at the surface of immune cells. Amino acid multiple sequence alignment of SpEX with previously characterized S. aureus SSL4, SSL5, SSL11 and SEIX proteins showed that S. pseudintermedius SpEX possesses the conserved sialated glycan binding site of the SSLs located at the C-terminal region and forming the V-shape depression in the 3D model of SpEX (Fig 2).
In S. pseudintermedius SpEX, we predict aspartic acid (D) 102 and threonine (T) 125 in the OB-fold domain are critical residues for the cytotoxic effect of the protein on immune cells. T 206, serine (S) 208, lysine (K) 211, leucine (L) 213, glutamine (Q) 214, arginine (R) 217 and isoleucine (I) 227 lining the sides of the glycan binding domain were predicted as essential for  binding to monocytes, neutrophils and complement. These residues are highly conserved among SSLs and SEIX proteins.

Design of SpEX with amino acid substitutions
An attenuated S. pseudintermedius SpEX (SpEX-M) was designed with (D102A and T125P) in the OB-fold domain and (T206P and R217A) in the sialated glycan binding domain (Fig 3).
To ensure that protein stability and folding were not altered, only amino acids highly conserved within functional domains of S. pseudintermedius SpEX and S. aureus SSL4, SSL5, SSL11 and SEIX were selected. Furthermore, the designed SpEX was first modelled in the structure to confirm that its conformation was maintained (Fig 3).

Cloning, expression, and purification of recombinant S. pseudintermedius SpEX
Recombinant native and mutant SpEXs with C-terminal 6x histidines (Fig 3) were generated in E.coli and purified under native conditions using HisPur Ni-NTA affinity chromatography. Both SpEX and SpEX-M expressed well from E. coli and remained soluble in solution, indicating protein stability with the amino acid substitutions. The molecular weights of SpEX and SpEX-M determined in western blots were of the expected sizes (27. 63 and 27.49 kDa, respectively) (Fig 4).

SpEX interferes with complement function
Binding of SpEX to human complement C5 was detected using ELISA, wherein complement component C5 was coated on ELISA plates and binding of recombinant native and attenuated his-tagged SpEX to C5 was measured using horseradish peroxidase (HRP)-conjugated anti-his tag monoclonal antibody. SpEX bound to human C5 significantly higher, in a dose dependent manner (0.5 with P = 0.0171, 1 μg /ml with P = 0.0027, 2 and 4 μg /ml with P <0.0001), than SpEX-M ( Fig 5A).
A hemolytic assay was used to evaluate SpEX mediated-inhibition of RBC lysis by complement. SpEX fixed complement and caused inhibition of hemolysis in a concentration dependent manner. At concentrations of 0.5, 1, 2 and 4 μg /ml, SpEX reduced the hemolysis of sensitized bovine erythrocytes compared to the positive control with P < 0.0001. SpEX at a concentration of 4 μg/ml showed no significant difference in hemolysis compared with the negative control ( Fig 5B).

SpEX inhibits neutrophil migration
A trans-well neutrophil migration assay was used to determine the inhibitory effect of recombinant SpEX on neutrophil migration in vitro. Recombinant SpEX at 0.2 μg/ml (at a concentration selected to assure cells remained viable) inhibited the migration of neutrophils induced by fetal bovine serum compared to SpEX-M at 0.2 μg /ml with P < 0.0001 (Fig 6). To validate the results, the viability of neutrophils during experiments was examined to ensure that killing did not occur at the concentration of SpEX used.

Attenuated SpEX induces a strong antibody response
Recombinant SpEX-M at 20 μg in 0.5 ml in phosphate buffered saline (PBS) (pH 7.2) was injected into three clinically normal dogs subcutaneously in the lateral thorax. Sera were collected from dogs on days -7, 8, 15 and 29 (relative to SpEX-M injections). IgG reactive with SpEX-M and SpEX was detected by ELISA on day 15 (P < 0.0001) and at a higher level on day 29 (P < 0.0001) compared to pre-injection control sera (Fig 7). Both SpEX and SpEX-M were recognized by dog anti-SpEX-M, confirming that there were no major antigenic differences between native and mutant SpEX.

Canine anti-SpEX-M reduced the effects of SpEX on PMNs and monocytes in vitro
Canine anti-SpEX-M collected after 3rd injection, at a dilution of 1:100 in PBS (pH 7.2), preincubated with recombinant SpEX at a concentration of 0.2 μg /ml, significantly reduced the inhibitory effect of rSpEX on leukocyte chemotaxis compared to serum collected from dogs before SpEX-M injection and control serum with P < 0.0001 (Fig 9A).

Discussion
Using mass spectrometry referenced to genomic data, it was possible to identify a new S. pseudintermedius exotoxin, SpEX, in S. pseudintermedius. SpEX secretion was confirmed in three S. pseudintermedius strains representing the three clonal complexes that predominate in the United States [5]. SpEX from the most commonly isolated clonal complex, CC68, was used in this study. However, SpEX is highly conserved among all three clonal complexes. Although its sequence is unique, this protein has a typical SSL tertiary structure, shared by S. aureus SSLs and SEIX, consisting of an N-terminal OB-fold domain that folds into a five-stranded β-barrel and a C-terminal β-grasp domain. However, SpEX, in addition to sharing the chemotaxis and complement inhibitory properties of SSLs, has a cytotoxic effect against monocytes and PMNs. These functions likely promote bacterial survival in their hosts and increase the likelihood of transmission. The attenuated SpEX remained soluble following expression from E. coli and both SpEX and SpEX-M were recognized by dog anti-SpEX-M. Thus, SpEX-M with amino acid substitutions in its predicted functional domains did not induce significant antigenic or conformational changes in the protein, however it showed diminished immune evasive properties compared to SpEX.
Antibody-mediated toxin neutralization may provide a strategy for immunotherapeutic treatment and/or prevention of current and recurrent infection [28,29]. Canine antibodies to SpEX neutralize and diminish its chemotactic inhibitory and cytotoxic effects on PMNs and   The increased prevalence of multidrug resistant S. pseudintermedius leaves few options for antimicrobial therapy [5]. Therefore, the development of novel strategies to treat this pathogen are a research priority [5,30]. One alternative approach is the development of a vaccine that can confer protection or provide effective therapy. However, vaccines in S. aureus clinical trials have thus far failed to show efficient protection [31]. Previous studies [30,[32][33][34] concluded that multicomponent vaccines containing a cocktail of staphylococcal antigens would likely work best in preventing infections caused by S. aureus. We propose that a potent and effective vaccine against S. pseudintermedius would involve immunogenic targets, secreted and/or accessible on the surface of the bacterium, conserved among prevalent strains and that play important roles in virulence, such as immune evasion. Understanding S. pseudintermedius protein function, surface accessibility and epitope conservation is crucial for vaccine development against infections caused by this pathogen.
In conclusion, we describe a new exotoxin produced by S. pseudintermedius that binds to C5, inhibits complement activation and permeabilizes leukocytes. A mutant version of the protein has reduced cytotoxic effects on dog PMNs and monocytes in vitro. Canine anti-SpEX produced against an attenuated form of the protein reduced its chemotaxis inhibition. Therefore, these mutant proteins may serve as important components of a multivalent vaccine for prophylaxis of S. pseudintermedius infections. By neutralizing extracellular toxins responsible for host tissue destruction and immunosuppression, such a vaccine may help the host immune system control infections.
Future studies might examine potential interactions with TLR as well as regulation of SpEX expression and the effects of quorum sensing and biofilm on production of the protein to better understand its role in pathogenesis. In addition, orthologs of SpEX produced by other members of the Staphylococcus intermedius group should be studied to determine their role in the virulence of these species. Pre-incubation of canine anti-SpEX diluted 1:100 with SpEX resulted in a significant reduction in MFI in permeability assays as compared with that of SpEX treatment alone. There was no significant difference between SpEX and SpEX treated with pre-serum. The MFI of SpEX pre-incubated with anti-SpEX 1:100 relative to SpEX was significantly lower in PMN (P < 0.0001 ���� ) and monocyte (P = 0.012 �� ) permeability assays. The values calculated were based on average values from three independent experiments ( � P < 0.05 was considered significant). https://doi.org/10.1371/journal.pone.0220301.g009

Ethics statement
Experimental protocols were reviewed and approved by the University of Tennessee Institutional Animal Care and Use Committee (IACUC) including 2474-0716 for samples of dog blood and 2572-1217 for producing antibodies in dogs with recombinant protein.

Bacterial strains and plasmids
Log phase bacterial cultures of S.pseudintermedius strains (shown with their sequence types) 06-3228 (ST68), 08-1661 (ST71) and NA45 (ST84) were analyzed by mass spectrometry. These strains are representative of the major S. pseudintermedius genotypes (clonal complexes) most commonly isolated from canine infections in the United States. A plasmid construct containing a mutated, synthetic S. pseudintermedius spEX (designed as described below) with BamHI/NotI cloning sites, was obtained commercially (Genscript Piscataway, NJ USA) ( Table 1).

Media and growth conditions
Bacterial colonies grown on blood agar plates were inoculated into 5ml of trypticase soy broth (TSB) (BD Biosciences, San Jose, CA) at 37˚C with shaking at 225 rpm. For log-phase bacterial cultures, bacteria were grown until an optical density (OD 600 ) of 0.4-0.6 was reached.

LC-MS/MS analysis of S. pseudintermedius supernatant
Log phase bacterial cultures of 06-3228, 08-1661 and NA45 were centrifuged at 10,000g for 30 minutes at 4˚C and the supernatants were collected and passed through 0.45μm filters (Whatman, GE Healthcare Lifesciences, Pittsburgh, PA). The filtrates were concentrated using Amicon Ultra -4 centrifugal filters (EMD Millipore Corp., Billerica, MA) and stored at -20˚C until further analysis. Samples interrogating S. pseudintermedius supernatant were prepared for shotgun LC-MS/MS analysis. Trypticase soy broth (media alone) and bacterial culture supernatant passed through control Sepahrose beads (without IgG) served as controls. Peptides were separated and analyzed with a 2-step MudPIT LC-MS/MS protocol (salt cuts of 50mM and 50mM ammonium acetate) over a 4-hr period then measured with a hybrid LTQ XL-Orbitrap (Thermo Scientific, Waltham, MA) mass spectrometer (MS). The percent coverage of detected proteins was calculated as previously described [35].
Recombinant S. pseudintermedius native and mutant SpEX expression were performed from E. coli pETBlue-2 constructs as previously described [4] using LB broth containing 50μg/ ml ampicillin and 20μg/ml chloramphenicol as protein expression media. Protein expression was induced by addition of 1 mM Isopropyl β-D-1-thiogalatopyranoside (IPTG). Protein extraction was performed using BugBuster reagent (Novagen) and 100X protease inhibitor (Cocktail Set III, EDTA-Free Calbiochem,). Recombinant proteins were purified using affinity purification (HisPur Ni-NTA Spin Purification Kit, Thermo Scientific). Protein concentrations were determined using a bicinchoninic acid (BCA) assay (Thermo Scientific).

SDS-PAGE and western blot
SDS-PAGE in 4-12% polyacrylamide gels (Invitrogen) was used to measure the molecular weight of expressed recombinant proteins. The resolved protein samples were blotted onto nitrocellulose membranes (Thermo Scientific). Then the blots were incubated in 5% (wt/vol) nonfat dried milk powder in 0.05% polyethylene glycol sorbitan monolaurate (Tween 20) containing phosphate buffered saline (PBS-T) overnight at 4˚C. Horseradish peroxidase (HRP)conjugated anti-6xhis tag monoclonal antibody (Thermo Scientific) diluted 1:2,000 in 0.05% PBS-T was added to the blocked membranes and incubated for 1 h with 225 rpm shaking at room temperature. After five washes with 0.05% PBS-T, bound antibodies were detected using 1-Stepchloronaphthol substrate solution (Thermo Scientific).

Preparation of canine anti-S. pseudintermedius SpEX
The endotoxin concentration in purified recombinant SpEX-M was measured using a Toxin-Sensor Chromogenic LAL Endotoxin Assay Kit (Genscript). Recombinant SpEX-M at 20 μg was diluted in 0.5 cc in phosphate buffered saline (PBS) (pH 7.2) and injected into three clinically normal dogs subcutaneously in the lateral thorax. Three injections were given 7 days apart with a control dog receiving PBS (pH 7.2) only. Blood was drawn on days -7, 8, 15 and 29 (relative to injections) and serum separated.

Enzyme-linked immunosorbent assay
Recombinant SpEX and SpEX-M were coated separately onto ELISA plates (Corning) at 2μg/ ml in PBS. The plates were washed with 0.05% PBS-T and incubated with two-fold serial dilutions of serum from dogs (injected with recombinant proteins) for 1 h at 37˚C, then bound IgG was detected using HRP-conjugated goat anti-dog IgG-heavy and light chain (Bethyl Laboratories, Inc.). ELISA assay plates were washed three times with PBS-T between all incubations, bound antibodies were detected using TMB substrate (Thermo Scientific), and reactions were stopped with 0.18 M sulphuric acid and optical density read at 450 nm on a plate reader (Bio TEK, EL800). The experiment was repeated a minimum of three times and a p-value of <0.05 was considered significant for all the experiments unless otherwise stated.

Trans-well assay for neutrophil transmigration
Canine blood was collected from healthy dogs using a sterile blood collection system with EDTA anticoagulant (BD Vacutainer). Canine neutrophils were isolated according to O'Donnell [42], washed several times, and re-suspended in pre-warmed 1ml RPMI medium.
A neutrophil transmigration assay protocol was used as previously described [43] with modifications. The assay was performed in 24-well plates with polycarbonate membrane (3.0 μM pore size and 6.5 mm well diameter) chambers (Corning Incorporated). Six hundred microliters of Dulbecco's Modified Eagle Medium (DMEM) with 10% of fetal bovine serum, as a PMN chemoattractant, was placed in each bottom well. Subsequently, SpEX and SpEX-M treated canine neutrophils (1×10 6 ), were added to the top chamber and incubated at room temperature for 6 h. DMEM medium was used as a negative control. Neutrophil viability and migration to the bottom chamber was quantified using a Countess II FL Automated Cell Counter (Thermo Scientific). In order to measure the protective effect of anti-SpEX on neutrophil transmigration, recombinant SpEX was incubated with canine anti-SpEX at a dilution of 1:100 for 30 min at 37˚C, then tested with the PMN transmigration assay.

Complement C5 binding assay
Human complement component C5 (Sigma-Aldrich) was coated onto ELISA plates (Corning) at 2μg/ml in PBS. Recombinant SpEX and SpEX-M were added at 0.5, 1, 2 and 4 μg/ml in PBS (pH 7.2) for 1 h at 37˚C, then bound recombinant proteins were detected using (HRP)-conjugated anti-6xhis tag monoclonal antibody (Thermo Scientific) at a dilution of 1:1000 in PBS-T (pH 7.2). Bound antibodies were detected as described above. The experiment was repeated a minimum of three times and a p-value of <0.05 was considered significant for all the experiments unless otherwise stated.

Complement mediated hemolysis assay
A hemolysis assay was performed to determine the ability of recombinant SpEX to inhibit complement function. Bovine erythrocytes, collected in EDTA from a clinically normal cattle, were washed in PBS and sensitized to complement by incubation with rabbit IgG fraction anti-bovine red blood cells (ICN Cappel) diluted 1:25, for 30 min at 37˚C with gentle mixing. Dog serum diluted 1:4 was pre-incubated with 0.5, 1, 2 and 4 μg/ml of recombinant native SpEX for 30 min at 37˚C with gentle shaking (100 rpm). One hundred microliters of sensitized bovine RBCs were added to all samples including positive and negative controls and incubated under the same conditions for an additional 30 min. After centrifugation at 4200 ×ɡ for 5 min, the absorbance of the supernatant was measured at 450 nm. Heat-inactivated serum mixed with PBS (pH 7.2) was used as a negative control and normal dog serum, diluted 1:4 was used as a positive control.

PMN and monocyte permeability assay
PMNs and monocytes were separated as previously described. PMNs and monocytes were incubated with two-fold serial dilutions of recombinant proteins (SpEX or SpEX-M) from concentrations of 50 μg to 1.6 μg in a volume of 500 μl of RPMI medium supplemented with 10% fetal bovine serum in an incubator with 5% CO 2 for 30 min. The supernatant of S. pseudintermedius strain 06-3228 was harvested at log phase to test the toxic effect of secreted, native SpEX. PMNs and monocytes were stained with 1 μl Sytox green (Life Technologies, Inc.) for 30 min, then washed twice with PBS (pH 7.2) and analyzed using a flow cytometer (Attune acoustic focusing cytometer) by gating separately on PMNs and monocytes.
To evaluate the protective effect of canine anti-SpEX-M on PMNs and monocytes, recombinant S. pseudintermedius SpEX, at concentration of 3.1 μg diluted in 1ml of PBS (pH 7.2) was incubated for 30 minutes at 37˚C with serum from SpEX-M injected dogs. The cell death cut-off used for flow cytometry analysis was established using leukocytes incubated without SpEX. Mean fluorescent intensity was measured from all gated cells.

Statistical analysis
Each experiment was repeated at least three times with a minimum of duplicate samples. Data analysis was conducted for each response variable using ANOVA methods, with treatment and dose as the independent variables. Diagnostic analysis was performed to check the model assumptions for normality and equal variance. Multiple comparisons were made with tukey's adjust. Significance was identified at p<0.05. All analysis was conducted in SAS 9.4 for Windows x64 from SAS Institute (Cary, NC) and some graphical outputs were generated by GraphPad Prism software (Version 7, GraphPad Software Inc.).