Staphopains Modulate Staphylococcus aureus Biofilm Integrity

Abstract

Staphylococcus aureus is a known cause of chronic biofilm infections that can reside on medical implants or host tissue. Recent studies have demonstrated an important role for proteinaceous material in the biofilm structure. The S. aureus genome encodes many secreted proteases, and there is growing evidence that these enzymes have self-cleavage properties that alter biofilm integrity. However, the specific contribution of each protease and mechanism of biofilm modulation is not clear. To address this issue, we utilized a sigma factor B (ΔsigB) mutant where protease activity results in a biofilm-negative phenotype, thereby creating a condition where the protease(s) responsible for the phenotype could be identified. Using a plasma-coated microtiter assay, biofilm formation was restored to the ΔsigB mutant through the addition of the cysteine protease inhibitor E-64 or by using Staphostatin inhibitors that specifically target the extracellular cysteine proteases SspB and ScpA (called Staphopains). Through construction of gene deletion mutants, we determined that an sspB scpA double mutant restored ΔsigB biofilm formation, and this recovery could be replicated in plasma-coated flow cell biofilms. Staphopain levels were also found to be decreased under biofilm-forming conditions, possibly allowing biofilm establishment. The treatment of S. aureus biofilms with purified SspB or ScpA enzyme inhibited their formation, and ScpA was also able to disperse an established biofilm. The antibiofilm properties of ScpA were conserved across S. aureus strain lineages. These findings suggest an underappreciated role of the SspB and ScpA cysteine proteases in modulating S. aureus biofilm architecture.

INTRODUCTION

Biofilm formation is an important contributing factor for the establishment of chronic infection by the opportunistic pathogen Staphylococcus aureus (1). S. aureus readily forms biofilms on host surfaces such as bone (2), cartilage, and heart valves (3), as well as on surfaces coated in host-derived proteins, including catheters and orthopedic devices (4). The mature biofilm is composed of a community of cells encased in an extracellular matrix that provides inherent resistance to the innate immune system and other antimicrobials (5, 6). Disassembly of the biofilm restores bacterial susceptibility to active chemotherapies and is an active area of research interest (7, 8). Understanding the molecular mechanisms of S. aureus biofilm formation and dispersal may enhance therapeutic strategies to treat chronic infections.

At the heart of S. aureus biofilm development is the complex extracellular matrix composed of proteinaceous material, polysaccharides, extracellular DNA (eDNA), and other cellular components. Key roles for each component have been demonstrated through the use of exogenously added enzymes that target matrix constituents (1). There are growing reports that S. aureus modulates this complex biofilm matrix through self-targeting by its own secreted enzymes (8–12). We have demonstrated that upregulation of the chromosomally encoded nuclease prevents S. aureus biofilm formation (11) and that agr-dependent control of proteases can disperse biofilms (7, 8). The role of the proteases is particularly complex since there are over 10 proteases secreted by most S. aureus strains (13). The protease genes are organized into four distinct operons (see Fig. 1A), encoding seven serine proteases (SspA and SplA-F), two cysteine proteases (ScpA and SspB), and a single metalloprotease (Aur). While the six Spl enzymes are active upon secretion, the Aur, SspA, SspB, and ScpA proteases are produced as zymogens (Fig. 1B). The Aur and ScpA zymogens autoactivate outside the cell (14, 15), and SspA and SspB activation relies on a proteolytic cascade in which Aur processes SspA (16) and SspA subsequently processes SspB (17). There is preliminary evidence that the SspA (V8) serine protease might be important in biofilm remodeling (9, 10, 12), but the contribution of the other proteases is less clear.

Fig 1.

Schematic of S. aureus protease genes and activation cascades. (A) Genomic organization of secreted protease operons based on the USA300 genome. Colors are keyed as follows: red, metalloprotease; blue, serine proteases; green, cysteine proteases (staphopains); orange, cysteine protease intracellular inhibitors (Staphostatins). (B) S. aureus proteolytic cascade of activation. The Aur, SspA, SspB, and ScpA proteases are secreted as inactive zymogens (Pro-) that must undergo processing for full activity. Pro-Aur autoactivates, allowing processing of SspA. Upon SspA activation, SspA is able to process Pro-SspB. Like Pro-Aur, Pro-ScpA is able to autoactivate.

The underlying mechanism through which S. aureus proteases self-prevent or disassemble established biofilms is complicated not only by the number of proteases involved but also by the myriad targets. Many surface proteins have been identified with S. aureus biofilm roles, such as SasC (18), SasG (19, 20), FnbpAB (10), protein A (21), ClfB (22), and Bap (23), and there are additional secreted proteins with known biofilm roles, such as beta-toxin (24). How proteases specifically target and destroy these proteinaceous matrix components and which protease and/or component remains most important in the biofilm formation/dispersal mechanisms remain unclear. Adding to this complexity, global changes in S. aureus gene expression control biofilm remodeling. These changes are coordinated by a suite of regulators (25–27), and the accessory gene regulator (agr) quorum-sensing system and stress response sigma factor B (sigB) are in part key players in this process. The agr system is known to regulate biofilm dispersal across the staphylococci (8, 28), and loss-of-function mutations in sigB upregulate the agr system and result in the inability of the strain to form a biofilm (29). While the molecular and biochemical mechanisms behind these phenotypes are not completely understood, it is known that SigB is a negative regulator of secreted virulence factors (30, 31), which is largely due to the repressive effect exerted on the agr system (29).

Here, we took advantage of a sigB mutation to uncover the most important proteases involved in controlling S. aureus biofilm phenotypes. All our biochemical and genetic studies led to the somewhat surprising observation that the cysteine proteases, also called Staphopains, are the key modulators of S. aureus biofilm development. Additional exogenous experiments confirmed these findings and demonstrated the important role of Staphopains in controlling S. aureus biofilm integrity.

MATERIALS AND METHODS

Strains and growth conditions.

The bacterial strains and plasmids used in this study are described in Table 1. Escherichia coli cultures were grown in Luria Bertani (LB) broth or on LB agar plates supplemented with 100 μg/ml ampicillin (Amp) or 50 μg/ml kanamycin (Kan) as required for maintenance of plasmids. S. aureus cultures were grown in tryptic soy broth (TSB) or tryptic soy agar (TSA). S. aureus chromosomal markers or plasmids were selected for, or maintained in, 10 μg/ml of chloramphenicol (Cam), erythromycin (Erm), or tetracycline (Tet) and 50 μg/ml of kanamycin (Kan). Unless otherwise stated, all broth cultures were grown at 37°C with shaking at 200 rpm.

Table 1.

Strain and plasmid list

Strain or plasmid	Description	Source or reference
Strains
Escherichia coli BL21(DE3)	Cloning strain	New England Biolabs
Escherichia coli BW25141	Cloning strain	62
Escherichia coli ER2566	Cloning strain	New England Biolabs
Staphylococcus aureus KB600	Δspl::erm	38
Staphylococcus aureus RN4220	Restriction modification-deficient cloning host	63
Staphylococcus aureus AH843	USA400 clinical isolate MW2	64
Staphylococcus aureus AH1263	LAC* USA300 CA-MRSA Erms	65
Staphylococcus aureus AH1292	AH1263 Δagr::tetM	11
Staphylococcus aureus AH1483	AH1263 ΔsigB	11
Staphylococcus aureus AH1825	AH1263 ΔscpA	This work
Staphylococcus aureus AH1866	AH1263 ΔsigB Δaur	This work
Staphylococcus aureus AH1868	AH1263 ΔsigB Δspl::erm	This work
Staphylococcus aureus AH1887	USA600 clinical isolate	NARSA
Staphylococcus aureus AH1921	AH1263 ΔsigB nuc::ltrB	11
Staphylococcus aureus AH1922	AH1263 ΔsigB ΔsspAB	This work
Staphylococcus aureus AH1980	AH1263 ΔsigB ΔscpA	This work
Staphylococcus aureus AH1982	AH1263 ΔsigB ΔsspAB ΔscpA	This work
Staphylococcus aureus AH1985	AH1263 ΔsigB Δaur ΔsspAB ΔscpA Δspl::erm	This work
Staphylococcus aureus AH2032	AH1263 ΔsigB Δagr::tetM	This work
Staphylococcus aureus AH2383	AH1263 ΔsigB Δaur ΔsspAB Δspl::erm	This work
Staphylococcus aureus AH2413	USA200 clinical isolate MN8	66
Staphylococcus aureus AH2596	AH1263 ΔsigB ΔsspB::pSMUT	This work
Staphylococcus aureus AH2597	AH1263 ΔsigB ΔsspB::pSMUT ΔscpA	This work
Staphylococcus aureus AH2644	USA100 clinical isolate	NARSA
Staphylococcus aureus AH2650	USA500 clinical isolate	NARSA
Staphylococcus aureus AH2652	USA700 clinical isolate	NARSA
Staphylococcus aureus AH2654	USA800 clinical isolate	NARSA
Plasmids
pALC2073	Tet-inducible expression plasmid	67
pALC2109	sigB gene cloned into pALC2073	67
pCM11	sGFP reporter plasmid	29
pCM13	pCM11 with aur promoter	This work
pCM15	pCM11 with spl promoter	This work
pCM16	pCM11 with ssp promoter	This work
pCM34	sspAB knockout vector	37
pCM35	pCM11 with scp promoter	This work
pCM39	scpA knockout vector	37
pCM48	pET28a with full-length 6×His ScpA C237A	This work
pCM49	pET28a with full-length 6×His SspB C243A	This work
pCM50	pET28a with mature 6×His SspB C243A	This work
pCM51	pET28a with mature 6×His ScpA C237A	This work
pET28a	Protein expression vector	Novagen
pJM03	scpB gene cloned into pMAL-c2X	This work
pJM04	sspC gene cloned into pMAL-c2X	This work
pKOR1-Δaur	aur knockout vector	36
pKOR1-ΔsigB	sigB knockout vector	29
pMAL-c2X	E. coli MBP expression plasmid	New England Biolabs
pSMUT	Derivative of pMUTIN	34
sspB::pSMUT	sspB gene cloned into pSMUT	This work

Recombinant DNA and genetic techniques.

S. aureus chromosomal DNA was prepared using a Puregene DNA purification system from Gentra Systems (Minneapolis, MN). Oligonucleotides were synthesized by Integrated DNA Technologies (Coralville, IA) and are listed in Table S1 in the supplemental material. Restriction modification enzymes were purchased from New England BioLabs and used according to the manufacturer's instructions. Plasmid construction was carried out in E. coli BW25141 or ER2566. Plasmids were first electroporated into S. aureus RN4220 as previously described (32) and subsequently transformed into select strains. As needed, chromosomal mutations were transduced into selected strains using bacteriophage 80α (33). Nonradioactive sequencing was performed at the University of Iowa DNA Facility.

Plasmid construction. (i) Protease promoter fusions.

The protease promoters were PCR amplified from AH1263 genomic DNA and cloned into the HindIII and KpnI sites of pCM11 (7) to generate reporters expressing synthetic green fluorescent protein (sGFP). Oligonucleotides CLM369 and CLM370 (see Table S1 in the supplemental material) were used to generate the aur promoter, resulting in plasmid pCM13. Oligonucleotides CLM359 and CLM360 were used to generate the spl promoter, resulting in pCM15. Oligonucleotides CLM362 and CLM375 were used to generate the ssp promoter, resulting in pCM16. Oligonucleotides CLM453 and CLM454 were used to generate the scp promoter, resulting in pCM35. Plasmid construction was confirmed by PCR and enzymatic digestion.

(ii) MBP-Staphostatin fusions.

The Staphostatin-encoding genes (scpB and sspC) were cloned into the EcoRI and PstI sites of pMAL-c2X (New England BioLabs) to generate maltose binding protein (MBP)-Staphostatin protein fusions. pMAL-scpB was constructed by PCR amplification of scpB using oligonucleotides JMM007 and JMM008. To clone sspC into pMAL-c2X, JMM009 and JMM010 oligonucleotides were used. Plasmid construction was confirmed by PCR and enzymatic digestion.

(iii) sspB::pSMUT insertion mutation plasmid.

For sspB mutation, a 446-bp internal fragment of sspB was PCR generated using primers OL-5505 and OL-5506. This fragment was cloned into pSMUT (gift from Simon Foster, University of Sheffield), a derivative of pMUTIN (34) where the lacZ gene was removed and the multiple cloning site improved.

Strain construction.

To build Δaur, ΔsspAB, and ΔscpA mutations, we used the pKOR1 knockout protocol (35) to integrate pKOR1-Δaur (36), pCM34 (37), or pCM39 (37) into AH1263. Double and triple mutations were built as needed and are listed in Table 1. For the ΔsspB mutant, sspB::pSMUT was used to transform strain RN4220 with selection for Erm resistance. RN4220 sspB null isolates were confirmed by Southern blotting, and the mutation was moved by phage transduction mutation into other strains as needed. To inactivate the spl operon, the Δspl::erm mutation (38) was moved by phage transduction. In order to create mutant combinations in a ΔsigB background, we transformed pKOR1-ΔsigB (29) into select protease mutant strains and carried out the knockout protocol. pKOR1-ΔsigB was also transformed into AH1292 (11) to create a ΔsigB Δagr::tetM mutation (designated AH2032). All mutations were confirmed using PCR. The sspAB, scpA, and sspB::pSMUT mutants were confirmed by immunoblot analysis (information provided below). The lack of SspA and SspB was also confirmed by reduced protease activity using a Fluorescence Resonance Energy Transfer (FRET) assay and Bz-Pro-Phe-Arg-pNA substrates and the lack of ScpA using elastin agar (substrate and agar information provided below).

Hemolysis assay.

As a measure of alpha-toxin production, qualitative hemolysis was monitored using rabbit blood agar. Rabbit blood plates contained 5% (vol/vol) rabbit blood in TSA. Defibrinated rabbit blood was purchased from Hemostat Laboratories (Dixon, CA).

Protease reporter assay.

S. aureus reporter strains were grown overnight in TSB supplemented with Erm and subcultured in 25 ml to an optical density at 600 nm (OD₆₀₀) of 0.025. Samples (200 μl) were collected for 48 h throughout growth and measured for fluorescence (excitation, 490 nm; emission, 520 nm) in a microtiter plate (Corning) using a Tecan Infinity 200 M plate reader.

Protease activity assays.

Protease activity was monitored using milk agar, elastin agar, and two unique peptide substrates. To prepare samples for measurement of protease activity, cultures were grown overnight in TSB at 37°C. For milk plate assays, 2 μl of overnight culture was placed onto dried-milk plates. Milk plates consisted of 5% nonfat dry milk and 3% Bacto agar. Elastin plate assays were conducted as previously described (39). Elastin from bovine neck ligament was purchased from Elastin Products Co., Inc. (Owensville, MO).

A FRET assay was developed to examine protease activity. The FRET substrate [5-carboxyfluorescein (FAM)-Lys-Lys-Ala-Ala-Glu-Ala-Ser-Lys-(QXL520)-OH; AnaSpec, Fremont, CA] was based on a known SspA peptide substrate (40). The substrate was resuspended to 50 μM using 20 mM Tris (pH 7.4). For measurement of protease activity, overnight cultures were subcultured to an OD₆₀₀ of 0.1 and allowed to grow for an additional 24 h, and spent media were filtered through 0.22-μm-pore-size Costar Spin-X centrifuge tube filters (Corning, NY). For the FRET assay, filtered medium samples were buffered in Tris (pH 7.4) (25 μl of 20 mM Tris [pH 7.4] was added to 175 μl of filtered medium). To start the reaction, 175 μl of buffered media was mixed with 25 μl of FRET substrate in a microtiter plate, and fluorescence measurements (excitation, 490 nm; emission, 520 nm) were obtained at 37°C in a Tecan Infinity 200 M plate reader. To test the specificity of the FRET substrate, we used a combination of purified enzymes and protease mutational analysis. Purified aureolysin was purchased from BioCentrum (Krakow, Poland), and purified SspA was purchased from Worthington Biochemical Corporation (Lakewood, NJ). Our findings suggest that the FRET substrate is cleaved primarily by SspA, and also by Aur and ScpA to a lesser extent, but not by SspB or the Spl proteases (data not shown).

SspB protease activity was measured using the synthetic chromogenic substrate Bz-Pro-Phe-Arg-pNA (Bachem, Torrance, CA). A 10 mM stock solution of Bz-Pro-Phe-Arg-pNA was prepared in 100 mM HEPES (pH 6.4) containing 25% methanol. Activity of spent media was assayed in a manner similar to that described by Massimi et al. (17). Briefly, reaction mixtures consisted of 2 mM Bz-Pro-Phe-Arg-pNA–20 mM EDTA–10 mM CaCl₂–100 mM HEPES (pH 6.4). The mixture was preincubated 15 min at 37°C in a microtiter plate and then mixed in a 1:1 volume with spent media. Absorbance at 405 nm was measured at regular time intervals in a Tecan Infinity 200 M plate reader.

Construction of mature-length, inactive sspB and scpA expression plasmids.

To generate proteins for antibody production, SspB and ScpA were mutated to change the active-site cysteine to an alanine (C237A in ScpA and C243A in SspB) with overlap extension PCR. The oligonucleotides for scpA were CLM551 and CLM550 as well as CLM549 and CLM552, and genomic DNA from strain LAC (AH1263) was used as the template. The two PCR volumes were mixed and used as the template for a second PCR performed with CLM551 and CLM552. This product was digested by NheI and EcoRI and ligated into pET28a digested by the same enzymes, and the resulting plasmid, pCM48, encoded a full-length, 6×His ScpA C237A protein. To construct inactive SspB protein, oligonucleotides CLM547 and CLM544 as well as CLM545 and CLM546 were used pairwise for PCR performed using genomic DNA from strain LAC. The purified PCR products were mixed and used as the template for a second round of PCR with CLM546 and CLM547. The resulting PCR volume was digested by SpeI and EcoRI and cloned into pET28a digested by NheI and EcoRI. The cloned full-length, 6×His SspB C243A-expressing plasmid is designated pCM49.

Next, the mature-length (processed-form) plasmids were generated. To create a mature-length ScpA C237A protein, oligonucleotides CLM559 and CLM552 were used for PCR with pCM48 as the template. The resulting PCR was digested with NheI and EcoRI, ligated to pET28a digested by the same enzymes, and designated pCM51. The mature, SspB C243A protein was cloned using oligonucleotides CLM548 and CLM547 for PCR and pCM49 as the template. After digestion by SpeI and EcoRI, the PCR product was ligated to pET28a digested by NheI and EcoRI and designated pCM50. Expression of pCM51 and pCM50 was tested in the E. coli strain ER2566 grown in LB Kan at 37°C. Cells were grown to an OD₆₀₀ of 0.6, induced with isopropyl-β-d-thiogalactopyranoside (IPTG) to 1 mM, and harvested after 2.5 h of incubation. The ScpA protein migrates at about 21 kDa and the SspB protein at about 22 kDa. However, the ScpA protein was not soluble in our tests, in similarity to what has been previously reported (41).

Purification and refolding of inactive Staphopain A.

The scpA expression plasmid was electroporated into E. coli BL21(DE3) cells. Cultures (900 ml) were grown with Kan at 37°C to an OD₆₀₀ about 0.6, induced with 1 mM IPTG, grown another 4 h, and harvested. Pellets were lysed with BugBuster (Novagen) diluted to 1× in 100 mM sodium phosphate (pH 8). DNA was sheared by passing the lysate through a 26-gauge needle seven times. Centrifugation at 20,000 × g for 30 min at room temperature gave a pellet of inclusion bodies. This pellet was resuspended in 6 M guanidine hydrochloride–100 mM sodium phosphate (pH 8) (buffer A) by stirring overnight at room temperature. After centrifugation, the supernatant was mixed with 5 ml of prewashed His-Select HF Nickel Affinity (Sigma) resin and washed in batches. Washes were done with buffer A until the OD₂₆₀ of the absorbing material was negligible and with buffer B (6 M guanidine hydrochloride, 100 mM sodium phosphate [pH 6.2]). Elution was performed using buffer C (6 M guanidine hydrochloride, 100 mM sodium phosphate [pH 4.5]). Fractions containing Staphopain A were pooled and concentrated with an Amicon Ultra filter unit (Millipore) (10,000 molecular weight cutoff [10K MWCO]). To refold, 5 ml of concentrated eluant was added dropwise to 500 ml of cold refolding buffer {50 mM sodium phosphate, 150 mM sodium chloride, 1 mM EDTA, 0.5 M arginine (pH 6.0), 30% glycerol, 1 mM 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS)} and stirred at 4°C for 64 h. The refolded Staphopain A solution was centrifuged at 18,000 rpm for 15 min at 4°C. The supernatant was concentrated to 22 ml using an Amicon cell with a YM 10 membrane (10K MWCO). The concentrated Staphopain A was dialyzed overnight in phosphate-buffered saline (PBS). Centrifugation (10,000 × g, 30 min, 4°C) removed any precipitate. The protein was further concentrated to about 0.1 mg/ml and sent to Pacific Immunology for the production of chicken antibodies to ScpA.

Purification of inactive Staphopain B protein.

Two liters of E. coli ER2566 containing pCM50 was grown with LB Kan at 37°C with shaking to an OD₆₀₀ of about 0.6. Cultures were induced with IPTG at 1 mM, grown another 6 h, and centrifuged. Pellets were lysed with BugBuster diluted in equilibration buffer (50 mM sodium phosphate, 0.3 M sodium chloride, 10 mM imidazole [pH 8.0]). Lysates were passed through a 26-gauge needle to shear DNA and centrifuged at 16,000 × g for 20 min at room temperature. The cleared lysate was loaded onto a preequilibrated 5-ml His-Select HF Nickel Affinity resin column. The column was washed with equilibration buffer until the protein concentration in the effluent was negligible. Elution buffer (50 mM sodium phosphate, 0.3 M sodium chloride, 250 mM imidazole [pH 8.0]) was applied to the column, and fractions containing the 22-kDa SspB protein were pooled and dialyzed versus PBS at 4°C. Pure SspB protein at 2.4 mg/ml (as determined by a Bio-Rad protein assay) was sent to Pacific Immunology for production of chicken antibodies.

Purification of antibodies.

Purification of Staphopain B antibodies from chicken egg yolks was performed using a Pierce Chicken IgY purification kit (Thermo Scientific) according to the manufacturer's instructions. ScpA antibodies were purified from chicken egg yolks according to Ko et al. (42). The yolk was passed through cheesecloth to remove membrane, 10 volumes of water was added, and the pH was adjusted to 5.0. The mixture was stirred for 2 h at 4°C and allowed to sit at 4°C overnight. Centrifugation at 10,000 × g for 30 min at 4°C removed most of the lipids. The pH of the supernatant was adjusted to 4 with hydrochloric acid, and 0.01% activated charcoal was added to remove the last of the lipids. Charcoal was removed by filtration through Whatman no. 1 filter paper. The pH was adjusted to 7 by the addition of sodium hydroxide, and ammonium sulfate was slowly added to reach a concentration of 50% at 4°C. After 1 h, the mixture was centrifuged at 10,000 × g for 30 min at 4°C. The pellet was resuspended and dialyzed in PBS. A subsequent ammonium sulfate precipitation (50%) yielded cleaner antibodies.

Western blotting.

Purified protein or trichloroacetic acid (TCA)-precipitated spent supernatant samples were mixed with SDS-PAGE loading buffer, and 5 μl of each sample was subjected to electrophoresis on a 12% polyacrylamide gel. The proteins were transferred to Immobilon-P polyvinylidene difluoride (PVDF) membranes (Millipore) using a Protean II device (Bio-Rad Laboratories, Hercules, CA). Membranes were blocked overnight at 4°C with 5% milk–Tris-buffered saline (20 mM Tris-HCl [pH 7.0], with 137 mM NaCl) containing 0.1% Tween 20 (TBST). Primary antibodies were diluted in 5% milk–TBST and incubated with the membranes at room temperature for 2 h. Membranes were rinsed 3 times in TBST and washed with agitation for 15 min followed by two more washes for 5 min. Secondary horseradish peroxidase (HRP)-conjugated goat anti-chicken IgG was diluted 1:20,000 in 5% milk–TBST and incubated at room temperature for 1 h. The membrane was rinsed and washed again as described above. SuperSignal West Pico chemiluminescent substrate was added for 5 min at room temperature followed by exposure to X-ray film.

Biofilm assays. (i) Microtiter plate biofilms.

Biofilm formation was assessed using a plasma-coated static microtiter plate assay as previously described (43). Our protocol was slightly modified from the published version in that we eluted the crystal violet immediately following the second set of triplicate washes. Biofilm biomass was completely resuspended by scraping attached biomass off the bottom of wells followed by thorough pipetting. Absorbance was measured in a Tecan Infinity 200 M plate reader at 595 nm. If the absorbance value fell out of the linear range, samples were instead measured at 600 or 605 nm, thereby allowing all samples to be read.

All biofilm additives were added at time zero at concentrations that did not inhibit growth (data not shown). Antibiotics were used to maintain plasmids as necessary. pALC2073 and pALC2109 were induced with 50 ng/ml anhydrotetracycline. For the experiment involving inhibition of wild-type (WT) biofilm formation by strain ΔsigB spent supernatant, an overnight culture of strain ΔsigB grown in TSB was filtered through a 0.22-μm-pore-size Costar Spin-X centrifuge tube filter (Corning) to collect spent supernatant. 1,10-Phenathroline monohydrate (1,10-PA), 3,4-dichloro-isocoumarin (DIC), and E-64 were purchased from Sigma (St. Louis, MO), and each was used at a concentration of 10 μM. Purified MBP, MBP-ScpB, and MBP-SspC were used at a final concentration of 400 nM. Pure ScpA and SspB (BioCentrum) were used at a final concentration of 50 or 250 nM.

(ii) Flow cell biofilms.

Flow cell biofilm formation was assessed as previously described (11) with one exception. To mimic the plasma-coated surfaces used in the microtiter assays, we coated plastic coverslips in human plasma. Rinzle Platic coverslips from Electron Microscopy Sciences (Hatfield, PA) were UV sterilized for 4 h and glued to flow cell chambers. Once they had hardened, 20% human plasma was injected into each flow cell chamber and moved to 4°C for 24 h, allowing plasma protein attachment. Following attachment, flow cell chambers were washed 3 times with 1× PBS before fresh culture was used to inoculate the flow chambers.

Purification and testing of MBP fusion proteins.

Purification of MBP fusion proteins from pMAL-scpB and pMAL-sspC on amylose resin was performed according to the manufacturer's instructions (New England BioLabs). For lysis of cells containing induced proteins, we incubated harvested cells in 1× Novagen Bug Buster (EMD Chemicals, San Diego, CA) for 2 h at room temperature. Eluted fractions were analyzed via SDS-PAGE and concentrated with an Amicon Ultra-15 filter unit with a 10K MWCO (Millipore). The protein concentration was measured using the Bio-Rad protein assay (Hercules, CA). Activity of MBP-ScpB was measured by inhibition of protease activity of strain AH2383 (ΔsigB Δaur ΔsspAB Δspl::erm) against the FRET substrate. Activity of MBP-SspC was measured by inhibition of protease activity of the ΔsigB mutant against Bz-Pro-Phe-Arg-pNA.

RESULTS

SigB and agr-dependent regulation of protease expression.

We previously reported that sigB mutants were defective in biofilm formation in multiple different strains, and the agr system and extracellular proteases were linked to this phenotype (29). To investigate this interconnection in more detail, we evaluated sigB, agr, and protease regulation using a community-associated methicillin-resistant S. aureus (CA-MRSA) isolate of the USA300 group (strain LAC; see Table 1), here called LAC-WT. We developed a series of reporter constructs that fused the promoter for each protease transcript (aur, sspABC, scpAB, and splABCDEF) to a gene encoding a modified green fluorescent protein (sGFP) (44). The reporter plasmids were transformed into LAC-WT and isogenic ΔsigB and Δagr mutants, and changes in protease gene expression were monitored throughout a time course. Our results show that expression from each of the four protease operon promoters is markedly upregulated in a ΔsigB knockout and downregulated in the Δagr mutant compared to its isogenic parent (see Fig. S1A in the supplemental material). The changes in gene expression coincide with those seen in overall protease activity (see Fig. S1B in the supplemental material), as measured using a FRET-based substrate assay (see Materials and Methods). The Δagr mutation was also epistatic to ΔsigB in the double mutant (see Fig. S1B), demonstrating that agr acts downstream of SigB and supporting previous observations (29). The high protease level in the ΔsigB mutant inhibits the establishment of a biofilm, presumably by cleaving surface and secreted matrix proteins. While the molecular details of some of these cascade steps are not fully established, we can take advantage of the ΔsigB mutation to gain insight into the protease(s) responsible for the biofilm phenotype.

The ΔsigB mutant is deficient in biofilm formation on human plasma due to a secreted factor.

Throughout this work, we used developed biofilm assays with human plasma-coated surfaces to investigate the biofilm roles of S. aureus proteases. The use of conditioned surfaces has gained popularity due to the observations that implant materials are coated with host matrix proteins (43, 45–47) and that S. aureus attaches to these proteins in vivo rather than directly to abiotic materials (48, 49). Notably, the ΔsigB mutant is defective in biofilm formation on the plasma-coated surface (Fig. 2A), and this phenotype can be complemented. As with the protease phenotype, introduction of the Δagr mutation repaired the biofilm deficiency, confirming that the agr system acts downstream of SigB in this biofilm development pathway.

Fig 2.

S. aureus sigB mutants form a protein-dependent biofilm on plasma-coated surfaces. (A) Biofilm formation by strain LAC (WT) and regulatory mutants in an assay using plasma-coated biofilm. Bacterial cultures were grown overnight and inoculated 1:200 in biofilm media over microtiter wells precoated in 20% human plasma. Static biofilm cultures were grown for 24 h. Biomass was calculated by staining washed biofilms with crystal violet and measuring absorbance of solubilized cultures. (B) WT biofilm levels upon addition of exogenous ΔsigB spent media. Increasing amounts of cell-free spent media were added at time zero to WT microtiter plate cultures. (C) ΔsigB biofilm formation in the presence of protease inhibitors at subinhibitory concentrations [10 μM] of the metalloprotease inhibitor 1,10-phenanthroline (1,10-PA), serine protease inhibitor 3,4-dichloroisocoumarin (DIC), or cysteine protease inhibitor E-64. **, P < 0.01; ***, P < 0.001 (relative to ΔsigB strain as determined by paired t test).

Several factors, including proteases and nuclease, could act downstream of SigB to inhibit biofilm formation. To determine which factors are biofilm inhibitory, we initially examined whether secreted factors were responsible for the phenotype. We collected overnight spent media from a ΔsigB mutant and added various amounts to LAC-WT biofilms at time zero. As shown in Fig. 2B, low levels of ΔsigB mutant spent media can inhibit biofilm formation in a concentration-dependent manner. Growth of the organism is not negatively affected across the concentrations tested (data not shown). These data suggested that something is released into the extracellular environment by the ΔsigB mutant that causes biofilm disruption. Knowing that S. aureus-secreted nuclease could inhibit biofilm formation on an abiotic substratum (11), we tested whether this was also the case in the plasma-coated assay by inactivating the nuc gene in the ΔsigB mutant background. The ΔsigB nuc double mutant displayed a level of biofilm capacity similar to that seen with the ΔsigB mutant on a plasma-coated surface, suggesting that additional factors were involved (data not shown).

Cysteine protease inhibition restores biofilm capacity to a ΔsigB mutant.

Considering that the agr mutation was epistatic to SigB (Fig. 2A) and that the secreted nuclease (Nuc) was not involved in these phenotypes, we focused attention on the secreted proteases. Initially, we attempted to restore ΔsigB biofilm formation by the exogenous addition of chemical inhibitors. Three inhibitors were tested: 1,10-phenanthroline (1,10-PA), a metal ion chelator that blocks Aur metalloprotease activity (36); 3,4-dichloroisocoumarin (DIC), a serine protease irreversible inhibitor that blocks SspA (V8) activity and potentially that of the six Spl proteases A to F (data not shown); and E-64, a cysteine protease inhibitor that blocks activity of both Staphopain proteases (50). Using plasma-coated biofilm assays with the ΔsigB mutant (Fig. 2C), 1,10-PA had no effect and DIC had an intermediate effect. In contrast, E-64 completely restored biofilm formation, suggesting that the cysteine proteases Staphopain A (ScpA) and Staphopain B (SspB) are contributors to the ΔsigB biofilm-negative phenotype.

Staphopain inhibition restores biofilm capacity to a ΔsigB mutant.

The challenge of using protease inhibitors such as E-64 is that they can alter the function of other fundamental cellular processes in S. aureus, such as the activity of the critical enzyme sortase (51), which is required for proper presentation of matrix binding proteins and biofilm formation (10). Downstream of each Staphopain gene on the chromosome is a cotranscribed gene encoding a cytoplasmic inhibitor (Staphostatin) with specificity to its cognate Staphopain (Fig. 1A). The Staphostatins are thought to protect the cell against premature activation of Staphopain precursors during protein export by forming a 1:1 protein complex (52). Due to their unique target specificity, the Staphostatins can also be used as tools to selectively inactivate the Staphopains in a complex mixture, unlike the nonspecific effects that are associated with E-64.

We affinity purified the Staphostatins ScpB and SspC (called Staphostatin A and Staphostatin B, respectively, to indicate the cysteine protease target). To test the functionality and specificity of the purified statins, we measured the inhibition of ScpA and SspB protease activity using spent media from S. aureus strains producing the enzymes (see Materials and Methods). As shown in Fig. 3A, only Staphostatin A could inhibit ScpA protease activity in a complex mixture at concentrations as low as 200 nM and, similarly, only Staphostatin B could inhibit Staphopain B at low concentrations (Fig. 3B). Therefore, each Staphostatin protein inhibitor was functional and acted in a selective manner against its respective Staphopain. When tested in a biofilm assay, 400 nM Staphostatin A and Staphostatin B were individually able to partially restore biofilm formation of the ΔsigB mutant (Fig. 3C). When they were added together, ΔsigB biofilm formation was completely restored, reaching the same level as seen with E-64 inhibition and that of a double Staphopain knockout in the ΔsigB mutant (see below). These data further suggest that the Staphopains (ScpA and SspB) are contributing to the ΔsigB mutant biofilm phenotype.

Fig 3.

Staphostatin specificity and restoration of ΔsigB biofilm phenotype. (A) Specificity of ScpB Staphostatin against ScpA activity. Overnight ΔsigB Δaur ΔsspAB Δspl cultures were incubated with the FRET substrate for 60 min. Increasing concentrations of MBP-ScpB or MBP-SspC were incubated with culture supernatant and the FRET substrate at time zero. Percent ΔsigB activity was measured by comparing the rate of fluorescent change at each statin concentration to the rate of fluorescent change of the ΔsigB culture using the equation [(slope of ΔsigB Δaur ΔsspAB Δspl + MBP-statin culture)/(slope of ΔsigB Δaur ΔsspAB Δspl culture)] × 100. (B) Specificity of SspC Staphostatin against SspB activity. The activities of MBP-ScpB and MBP-SspC against SspB were measured using the SspB-specific substrate Bz-Pro-Phe-Arg-pNA. Overnight cultures of strain ΔsigB were incubated with this substrate for 60 min. Percent ΔsigB activity was measured as described above using the ΔsigB strain. (C) Strain ΔsigB biofilm formation in the presence of MBP-Staphostatin fusion proteins. Strain ΔsigB biofilm formation was assessed following treatment of cultures with MBP [250 nM], MBP-Staphostatins [250 nM], or E-64 [10 μM] at time zero. **, P < 0.01; ***, P < 0.001 (relative to ΔsigB strain as determined by paired t test).

Mutation of sspB and scpA restored ΔsigB biofilm formation.

To follow up on the inhibitor observations, we took a genetic approach and constructed protease mutations in the ΔsigB background. First, we made single deletion mutations in the aur, sspAB, scpA, and spl operons. Next, we made combinations of protease mutations, including a complete extracellular protease knockout, here referred to as Δprotease. All mutations were confirmed by PCR, and wherever possible, immunoblot and protease activity assays and/or antibiotic resistance testing was also utilized for confirmation (see Materials and Methods). To test for the occurrence of spontaneous agr mutations, hemolysis on blood agar was visualized for each mutant, and levels were similar to those of the LAC-WT or ΔsigB strain for each constructed strain (data not shown).

The complete extracellular protease knockout in a sigB mutant (ΔsigB Δprotease strain) was able to form a biofilm with capacity similar to that of the wild type, confirming that extracellular proteases have an inhibitory effect on biofilm formation (Fig. 4A). Taking our studies further, we tested each protease knockout individually in the sigB mutant background. While introduction of the Δspl operon mutation was unable to restore biofilm formation, deletion mutations in Δaur, ΔsspB, and ΔscpA showed partial restoration. In support of the inhibitor observations (Fig. 3), introduction of both Staphopain single mutations into the ΔsigB background completely restored biofilm formation to wild-type levels (Fig. 4A). Although SspA (V8) has a demonstrated role in biofilm remodeling (10), our findings indicate that SspA was not a significant factor in this assay.

Fig 4.

Deletion of the staphopain genes restores biofilm formation to a ΔsigB mutant. (A) Biofilm-forming capacity of protease mutants in ΔsigB background. The sigB deletion plasmid was transformed into single, double, triple, and quadruple protease mutants, and the resulting strains were tested for biofilm formation in the static plasma-coated microtiter plate assay. (B) Biofilm-forming capacity of staphopain mutants on an uncoated surface. Mutant combinations were tested for biofilm formation using experimental conditions identical to those described for panel A except that microtiter wells were not precoated in human plasma. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (relative to ΔsigB strain for panels A and B as determined by paired t test). (C to F) Plasma-coated coverslip biofilms grown in flow cells. As an additional assay, flow cell biofilm formation was assessed using plastic coverslips that were UV sterilized and treated overnight with 20% human plasma. Flow cell biofilms contained no cells (C) or the WT (D), ΔsigB (E), or ΔsigB ΔsspB ΔscpA (F) strain. Biofilms were poststained with SYTO-9 and imaged with confocal microscopy. Representative top-down images of flow cell biofilm are shown. Experiments were performed in triplicate.

Staphopains have been implicated in cleavage of fibrinogen, a human plasma protein capable of binding S. aureus adhesins (53). It is possible that the Staphopains cleave plasma proteins and that this is the explanation for the biofilm phenotypes in our assays. To assess this issue, we repeated our microtiter plate experiments using uncoated microtiter wells. As shown in Fig. 4B, the overall quantity of biofilm biomass is reduced in uncoated wells compared to that seen with plasma coating, which supports previous observations (43). Importantly, while the ΔsigB mutant had significantly less biomass than the LAC-WT strain, inactivation of Staphopain A or B in the ΔsigB background partially restored biofilm formation, and removal of both Staphopains restored biofilm capacity to WT levels (Fig. 4B). Taken together, and in conjunction with the inhibitor data, our findings suggest that SspB and ScpA are important determinants of biofilm formation.

To confirm these observations, we repeated experiments using flow cell biofilm assays. Unlike typical flow cell experiments, the substratum was coated with human plasma protein to maintain consistency with other assays in this study. The flow biofilms were poststained with SYTO-9 to detect biomass, and images were obtained with confocal microscopy. A no-bacteria control experiment revealed that the hydrophobic nature of the dye has some, albeit limited, background binding to the plasma proteins on conditioned surfaces (Fig. 4C). LAC-WT formed a robust biofilm in this assay (Fig. 4D), and the ΔsigB mutant was defective (Fig. 4E). Some small clumps of ΔsigB mutant cells did attach to the plasma proteins, giving a punctate staining appearance, but the level of biofilm accumulation observed with the LAC-WT strain was not achieved with the ΔsigB mutant. Finally, introduction of the double cysteine mutation (ΔsspB and ΔscpA) into the ΔsigB mutant background restored biofilm capacity to the WT level, supporting the observations made with the microtiter assay (Fig. 4A).

Staphopains are repressed under biofilm-forming conditions.

Knowing that Staphopain inhibition restores biofilm formation to a ΔsigB mutant, we reasoned that S. aureus maintains low Staphopain levels in order to facilitate establishment of the biofilm. To address this issue, LAC-WT and strains with either a ΔsigB single mutation or a ΔsigB ΔsspB ΔscpA triple mutation were grown under three different conditions: (i) in TSB, to mimic planktonic growth; (ii) in biofilm media (TSB with 3% NaCl and 0.5% glucose [43]), to simulate regulatory conditions during biofilm initiation; and (iii) under biofilm-forming conditions, where biomass was prepared using the plasma-coated microtiter assay. Cells under each condition were grown for 24 h, and the amounts of cells assayed were normalized to an OD₆₀₀ of 1.5. Staphopain levels were assessed using an immunoblot approach with anti-ScpA and anti-SspB antibodies prepared in this work.

For ScpA, the active, processed form of the protease was not detectable in LAC-WT cultures grown under any of the conditions tested, while this form was detectable in a ΔsigB mutant (Fig. 5A). Levels of ScpA (various processed forms) in a ΔsigB mutant were markedly higher in the TSB-grown cultures compared to those grown in biofilm media or the biofilm biomass. In our experience, the series of processed bands in immunoblots is typical, as the zymogen is cleaved at various positions to generate the active form of the Staphopain. As a control, a ΔsigB ΔsspB ΔscpA triple mutant was tested and ScpA was not detectable under any of the tested conditions. In LAC-WT, the fact that ScpA was not detectable at 24 h even under the TSB-grown conditions matched previous reports (54). To confirm that the anti-ScpA antibodies were functioning properly, an immunoblot time course experiment was performed (see Fig. S2A and B in the supplemental material), and ScpA levels were indeed high in late-logarithmic growth and early stationary phase, with the ScpA degrading by 24 h as previously observed (54).

Fig 5.

Staphopain levels are repressed during biofilm formation. The WT, ΔsigB, and ΔsigB ΔscpA ΔsspB strains were grown overnight and subcultured in TSB or biofilm media or under static biofilm conditions. Cultures were grown for 24 h, and cell-free spent media were collected and normalized to a cell suspension at an OD₆₀₀ of 1.5. Proteins in spent media were precipitated with TCA and prepared for electrophoresis. (A) Immunoblot detection of ScpA. (B) Immunoblot detection of SspB.

The SspB enzyme is more stable and facilitated a better assessment of Staphopain regulation within a biofilm (Fig. 5B). Processed SspB levels were high in a ΔsigB mutant, especially under the TSB-grown conditions, and all of these levels were substantially elevated over the LAC-WT levels. As seen with the ScpA immunoblot, processed SspB in the biofilm or the biofilm growth media was not observed in LAC-WT, except for trace levels of the SspB zymogen under each growth condition. However, LAC-WT levels of SspB were elevated in the TSB-grown culture, a striking contrast from the biofilm-grown conditions. To confirm functionality of the SspB antibody, a time course experiment was performed and the SspB zymogen was detected at all time points (see Fig. S2C in the supplemental material), except in sspB mutant strains. Processed SspB was detectable only in LAC-WT by 24 h, albeit at low levels. In all the SspB immunoblots, the sspB::pSMUT mutation resulted in an antibody-detectable, truncated form of the protein. The truncation was absent from TSB-grown cultures, possibly suggesting that it is degraded under conditions of high protease levels, and importantly, SspB protease assays confirmed the absence of activity (data not shown) in sspB::pSMUT mutant strains. Altogether, these experiments demonstrated that Staphopain accumulation is repressed during biofilm growth and that these proteases are overproduced in SigB-defective strains.

Exogenous addition of Staphopains inhibits biofilm formation.

Based on our observations that Staphopain overproduction is biofilm inhibitory, we tested whether the exogenous addition of purified Staphopains could inhibit LAC-WT biofilm formation. Since Staphopain A is produced in the 50 to 200 nM range by LAC-WT S. aureus when grown in TSB (54) and since the ΔsigB mutant produces even more proteases, we used similar levels in our experiments. As shown in Fig. 6A, pure ScpA was able to significantly inhibit biofilm formation at low enzyme concentrations (50 nM) and more dramatically at higher concentrations (250 nM). Addition of SspB was not as effective, with only higher concentrations of the enzyme (250 nM) inhibiting biofilm formation. Supporting these observations, when SspB and ScpA were simultaneously added, biofilm formation was completely prevented. A strain lacking all the proteases (the ΔsigB Δprotease mutant) was also tested, and addition of both SspB and ScpA also prevented biofilm formation. Taken together, these findings support our observations that the Staphopains have a biofilm-inhibitory role.

Fig 6.

Staphopain addition inhibits biofilm formation and disperses established biofilms. (A) Staphopain inhibitory capacity of LAC (WT) biofilm formation. Purified Staphopains [nM] were added at time zero to WT or ΔsigB Δprotease strains, and biofilm formation was assessed. (B) Staphopain disassembly of established WT biofilms. Purified Staphopains were added to WT biofilm cultures after 12 h, and biomass was assessed 12 h later (24 h from time zero). A concentration of 250 nM ScpA, 250 nM SspB, or 250 nM each in combination was used. *, P < 0.05; ***, P < 0.001 (relative to untreated control as determined by paired t test).

Due to our previous reports on protease-mediated biofilm dispersal (7, 8), we tested the ability of the Staphopains to disassemble established biofilms. We allowed a LAC-WT biofilm to form and achieve full biomass, and at this time point (12 h), the biofilm was subjected to ScpA or to SspB or to both enzymes at a concentration of 250 nM. ScpA protease was able to disperse the established biofilm (Fig. 6B), but SspB was less effective. The addition of SspB and ScpA together also dispersed the biofilm, but the inhibition level was similar to that seen with ScpA alone. These findings suggest that ScpA is the more proficient at dispersing an established biofilm.

Staphopain A prevents biofilm formation across S. aureus lineages.

The S. aureus biofilm matrix is a complicated meshwork of cellular components, and the importance of each component likely depends on both the strain examined and the environment in which the biofilm is growing. In this work, we demonstrated the role of Staphopain-susceptible protein components in biofilm formation by the USA300 isolate LAC-WT. To assess the generality of the observations, we examined the susceptibility of biofilms to Staphopains in a variety of strain backgrounds. Since Staphopain A was able to both prevent the formation of and disassemble LAC-WT biofilms at low concentrations of enzyme, we focused on this enzyme in the strain assessment. A single clinical isolate from each pulsed-field gel electrophoresis (PFGE) (type USA100 to USA800) was selected and assessed for biofilm formation in the plasma-coated microtiter assay. Each of USA100, USA200 (MN8), USA300 (LAC), USA400 (MW2), USA500, USA600, USA700, and USA800 was able to form a biofilm in this assay. As before, we used 250 nM ScpA for the biofilm treatments and this prevented biofilm formation in all strain types tested (Fig. 7). These findings demonstrate that ScpA enzyme has a conserved ability to inhibit biofilm formation across a broad range of S. aureus strains.

Fig 7.

ScpA inhibition of biofilm formation is conserved across S. aureus strains. Strains representing PFGE types USA100 to USA800 were selected to examine the capacity to form a biofilm in the presence of 250 nM ScpA. White bars, untreated; black bars, ScpA treated. All ScpA treatment results were significant relative to untreated control results (P < 0.001) as determined by paired t test.

DISCUSSION

The S. aureus biofilm matrix is complex and has been the focus of numerous studies in recent years (1). Of interest for this work, there are growing reports that proteins within the biofilm matrix are critical for S. aureus biofilm structure, with primary evidence coming from the ability of exogenously added proteases to inhibit biofilms (55–57) and findings showing that secreted proteases of S. aureus also self-inhibit biofilms (7–10). To date, studies have pointed at a key role for SspA (V8) protease in cleaving important surface adhesins, such as FnbpAB (10, 12), and preventing biofilm formation, thus making V8 the only known protease of S. aureus with biofilm-inhibitory functions. S. epidermidis Esp protease is a V8 homologue, and there is a recent report that it can destroy S. aureus biofilms (58), further supporting the antibiofilm activity of V8-like enzymes. In this study, we investigated the contribution of proteases to S. aureus biofilm phenotypes and all of our findings pointed to the Staphopain cysteine proteases being the most important modulators of biofilm integrity.

Considering published findings, why did our studies identify Staphopains and not V8 protease? Some of the properties of V8 make it an unusual choice to be the dominant enzyme involved in biofilm remodeling. V8 has narrow substrate specificity, cutting specifically after glutamate residues (13), and the zymogen also has to be activated by Aur in order to be functional (16), both properties that limit the utility of V8 as a biofilm modulator. In contrast, the other major proteases (Aur, ScpA, SspB) have broader target specificity and both Aur and ScpA self-activate (14, 15, 59). ScpA in particular seems like an obvious candidate to control biofilm matrix structure. The enzyme is one of the first proteases produced and activated in the extracellular environment and reaches high local concentrations (54), all properties that would be advantageous for controlling biofilm integrity. The fact that cysteine proteases would be inhibitory toward a bacterial biofilm is also not unique. In group A Streptococcus (GAS), the prominent SpeB protease (Streptopain) has been linked to biofilm formation (60). When SpeB levels are high, such as in srv mutants, GAS biofilm formation is eliminated unless the cysteine protease is inhibited genetically or biochemically, all findings that share striking parallels to our observations with S. aureus.

What are the Staphopains cutting that negatively impact S. aureus biofilms? The targets of the proteases are important to understanding how the biofilm matrix retains structure. Presumably, FnbpA and -B are potential targets, since they are produced by most S. aureus strains, have a key role in biofilm formation, and are also known to be targeted by proteases (10, 61). It is also possible that the Staphopains are targeting surface adhesins that bind human matrix proteins and, potentially, even the matrix proteins themselves. Fibrinogen is likely an important handle for binding and initiation of S. aureus biofilm formation, and the Staphopains are known to cleave fibrinogen (53); more specifically, SspB cleaves the Aα-chain under physiologically relevant conditions. However, our results indicate that the Staphopains can inhibit biofilm formation on an uncoated surface in a manner independent of fibrinogen (Fig. 4B). This finding suggests that Staphopains are capable of cleaving S. aureus biofilm matrix proteins, which results in a biofilm phenotype in overexpressed conditions. Thus, for the plasma-coated surfaces, the significance of fibrinogen cleavage in the context of a S. aureus biofilm remains unclear. It should also be noted that the biofilm-dispersing functions of the Staphopains differ from the biofilm inhibition functions (Fig. 6), suggesting that some of the targets might differ depending on the developmental stage of the biofilm. Taken together, the limited information on targets and the interconnection between the proteases are areas in need of further investigation.

Altogether, our studies have uncovered an unexpected new role for the S. aureus Staphopain proteases in controlling biofilm development. Under conditions where the Staphopains are upregulated, such as a sigB mutation, the proteases reach levels that prevent biofilm formation, presumably due to the cleavage of surface structures or secreted proteins in the biofilm matrix. ScpA exogenous-addition experiments, and other overall properties of this enzyme, suggest that it might be the more significant player in biofilm remodeling, although SspB clearly has an important role. Our observations were consistent across strains, suggesting that this is a general trend among S. aureus biofilms. Uncovering new strategies to upregulate the Staphopains could be an innovative approach to treating biofilm infections.