Host-derived protease promotes aggregation of Staphylococcus aureus by cleaving the surface protein SasG

ABSTRACT Staphylococcus aureus is one of the leading causes of hospital-acquired infections, many of which begin following attachment and accumulation on indwelling medical devices or diseased tissue. These infections are often linked to the establishment of biofilms, but another often overlooked key characteristic allowing S. aureus to establish persistent infection is the formation of planktonic aggregates. Such aggregates are physiologically similar to biofilms and protect pathogens from innate immune clearance and increase antibiotic tolerance. The cell-wall-associated protein SasG has been implicated in biofilm formation via mechanisms of intercellular aggregation but the mechanism in the context of disease is largely unknown. We have previously shown that the expression of cell-wall-anchored proteins involved in biofilm formation is controlled by the ArlRS-MgrA regulatory cascade. In this work, we demonstrate that the ArlRS two-component system controls aggregation, by repressing the expression of sasG by activation of the global regulator MgrA. We also demonstrate that SasG must be proteolytically processed by a non-staphylococcal protease to induce aggregation and that strains expressing functional full-length sasG aggregate significantly upon proteolysis by a mucosal-derived host protease found in human saliva. We used fractionation and N-terminal sequencing to demonstrate that human trypsin within saliva cleaves within the A domain of SasG to expose the B domain and induce aggregation. Finally, we demonstrated that SasG is involved in virulence during mouse lung infection. Together, our data point to SasG, its processing by host proteases, and SasG-driven aggregation as important elements of S. aureus adaptation to the host environment. IMPORTANCE Here, we demonstrate that the Staphylococcus aureus surface protein SasG is important for cell-cell aggregation in the presence of host proteases. We show that the ArlRS two-component regulatory system controls SasG levels through the cytoplasmic regulator MgrA. We identified human trypsin as the dominant protease triggering SasG-dependent aggregation and demonstrated that SasG is important for S. aureus lung infection. The discovery that host proteases can induce S. aureus aggregation contributes to our understanding of how this pathogen establishes persistent infections. The observations in this study demonstrate the need to strengthen our knowledge of S. aureus surface adhesin function and processing, regulation of adhesin expression, and the mechanisms that promote biofilm formation to develop strategies for preventing chronic infections.

, with ~80% of infections caused by the patient's colonizing strain (8)(9)(10).S. aureus is one of the leading causes of healthcare-associated infections, such as surgical site infections and central line-associated bloodstream infections (11), imposing a substantial burden on the healthcare system.While these infections are often challenging to treat, the rise of methicillin-resistant S. aureus (MRSA), which causes over 119,000 of these infections annually in the United States, has further exacerbated treatment challenges and increased healthcare costs by nearly one billion dollars annually (12)(13)(14)(15).
S. aureus is one of the most prevalent pathogens in chronic wound infections (16)(17)(18) and is one of the first pathogens to colonize the cystic fibrosis (CF) lung (19).The occurrence of chronic and persistent S. aureus infections is in part due to aggregation mechanisms and the ability of this pathogen to adhere to indwelling medical devices as a biofilm (20,21).However, in the absence of an implanted medical device, S. aureus can form free-floating aggregates that are physiologically similar to biofilms and are likewise more antibiotic resistant (22,23).It has been suggested that bacterial aggre gates predominate in chronic infections such as osteomyelitis (24), chronic wounds (25), and the lungs of CF patients (26,27).Intensive efforts to clear MRSA lung infections in CF patients, sometimes using up to five different antibiotics, have shown some promise, although ~15% of patients still harbor MRSA at the end of the intervention period (28)(29)(30).A better understanding of S. aureus biofilm formation and aggregation may lead to alternative therapies for these difficult-to-treat infections.
MRSA aggregation observed in clinical infections has been described as groups of closely attached cells that are not surface attached, and similarly to mature biofilms, they protect from environmental stress and allow for persistence (22).Aggregates and biofilms are difficult to treat in part because they are up to 1,000-fold more resistant to antibiotics than planktonic cells (22,31,32).This increased tolerance is thought to be due to a combination of slowed diffusion of antibiotics through the extracellular matrix and slower growth of cells within the community of cells (33).In addition, aggregates are more resistant to clearance by the innate immune system, in part due to their large size, which impedes phagocytosis, and their ability to secrete and concentrate toxins that target leukocytes (34)(35)(36)(37).
One of the key drivers of biofilm formation and aggregation in S. aureus is the large, cell-wall-attached surface protein G (SasG) (38)(39)(40).SasG and its S. epidermidis homolog Aap consist of multiple domains with distinct functions (Fig. 1A).The A domain, which has 59% identity to Aap, is implicated in binding to corneocytes (41) and nasal epithelial cells (42), and has a short, variable repeat region and an L-type lectin subdomain.In full-length SasG, the B domain, which has 60%-67% identity to Aap depending on B-repeat number, consists of 2-17 repeats of alternating G5 subdomains and E spacers (38,43,44).These G5-E repeats can dimerize in a Zn-dependent manner to form a twisted cable structure that facilitates intercellular interactions (45).In S. epidermidis, the Aap A domain is removed by the metalloprotease SepA, allowing the exposed B domains to dimerize and promote biofilm accumulation (46).Exogenous addition of the host proteases trypsin and cathepsin G can also enhance S. epidermidis biofilm formation through the processing of Aap (43).Whether SasG also needs to be proteolytically processed is not known, although it appears that none of the known proteases secreted by S. aureus can specifically target SasG (38).
Expression of sasG is variable across S. aureus clinical isolates.SasG is constitutively expressed by some clinical isolates (39), and the presence of anti-SasG human antibodies demonstrates its expression during infection (47,48).While the majority of S. aureus strains encode sasG (49), some do not express it under laboratory conditions (39,50).Recently, it has become apparent that this lack of SasG expression might be due to its repression by an ArlRS-MgrA regulatory cascade under in vitro conditions (48,51).
In this project, we took advantage of the high level of SasG expression in certain S. aureus strains to investigate the role of SasG in aggregation and virulence.We identified that the presence of SasG increases S. aureus virulence during lung infection and that the cleavage of the N-terminal portion of the A domain of SasG is necessary for S. aureus to aggregate.Since S. aureus does not cleave SasG on its own (38), SasG cleavage during infection must be mediated by host proteases.Such cleavage leads to SasG-mediated aggregation of S. aureus, which is reflected in increased virulence of SasG-expressing strain during lung infection.Overall, the host-driven cleavage of SasG establishes an unusual and novel way of sensing and responding to the host environment.

SasG saliva interaction and expression levels across S. aureus strains
Aspiration of saliva is often a precursor to lung infections (52)(53)(54)(55), leading us to investigate how MRSA reacts to the presence of human saliva.We made a somewhat surprising observation that a USA400 MRSA ΔmgrA mutant strain aggregated to high levels when the cells were resuspended in human saliva, while the WT strain remained in suspension (Fig. 1B).Knowing there is differential surface protein expression in ΔmgrA mutants (48), we ran Coomassie protein gels (Fig. 1C) and observed dramatic salivary processing of a large protein that we reasoned might be surface protein G (SasG).Upon constructing a MRSA ΔmgrA ΔsasG double-mutant, the protein and aggregation phenotype both disappeared (Fig. 1B and C), demonstrating this phenotype is due to SasG.In addition, the aggregation could be complemented by providing mgrA on a plasmid (Fig. 1D).
We next investigated the generality of this phenotype in S. aureus.We compared sequenced S. aureus strains containing functional chromosomal copies of sasG including community-acquired MRSA (CA-MRSA) USA400 strain MW2, Newman, 502 a, and a CF clinical MSSA isolate AH4654.We also included strains that expressed a truncated form of SasG such as those of USA300 strain LAC and USA100 strain N315.Together, these strains represent different subtypes of SasG that have been recently described (49).Finally, we included strains lacking a copy of the sasG gene altogether, such as USA200 strains MN8 and MRSA252, as controls for comparison.Strains with a functional, full-length version of SasG protein exhibited high levels of saliva-induced aggregation in the absence of mgrA (Fig. 1E) and we observed abundant SasG in cell wall preparations (Fig. 1F).The CF clinical isolate AH4654 exhibited lower expression levels and intermediate aggregation (Fig. 1E), although the genetic composition is almost identical to MW2, the functionality of the ArlRS-MgrA system in relation to SasG is not clear in this strain.Unexpectedly, Newman exhibited no visible expression of SasG protein (in WT or ΔmgrA mutant) and little aggregation despite having a full-length version of SasG encoded in the genome (Fig. 1E and F).N315 expressed a protein of size similar to SasG but did not clump at all.These data were confirmed by qPCR quantifying sasG expression (Fig. 1G).In general, our observations indicate that S. aureus strains with a full-length SasG, under conditions that induce sasG gene expression, will aggregate in the presence of human saliva.

Molecular details of MgrA repression of sasG gene
To investigate transcriptional control of sasG in the (CA-MRSA) USA400 strain MW2, we constructed a P sasG -sGFP reporter plasmid (pHC127) with sasG promoter fused to a gene encoding sGFP.This plasmid was transformed into mutants of the ArlRS and MgrA regulatory systems, previously suspected to repress the expression of SasG, and the expression levels were monitored over 24 h (Fig. 2A).The highest expression was observed in the ΔmgrA mutant, followed by the ΔarlRS mutant, with minimal expression in WT.The high expression in ΔmgrA mutant was confirmed at the protein level (Fig. 2B).We analyzed the sasG promoter region by 5′RACE to identify a putative housekeeping promoter and transcriptional start site (Fig. 2C).Putative MgrA repressor binding sites are shown that overlap the promoter region (48).Overall, our findings confirm that the expression of SasG in laboratory growth media is repressed by the ArlRS-MgrA regulatory cascade.

SasG processing after A domain repeats promotes aggregation in human saliva
As noted in Fig. 1, a large protein consistent with the size of SasG was upregulated in the ΔmgrA mutant and processed to a smaller version after incubation with human saliva (Fig. 1C and 3A).These observations suggest that proteases present in saliva could process SasG to smaller sizes.A previous report suggested that SasG possessed self-processing capability and that this cleavage occurred at multiple sites within the B domain (38).While the self-processing might be occurring in other experimental conditions, we did not observe background processing in our experiments when bacteria were incubated in PBS (Fig. 1C).By contrast, our results indicate that SasG may be processed by a host protease(s), and there may be a single cleavage site near one end of the protein, similar to what is seen with Aap (46).
To determine the location of the cleavage site within SasG, we cloned and purified the extracellular portion of SasG.The LPXTG cell wall anchor was replaced with a hexahisti dine tag, and the protein was expressed in a S. aureus strain that lacks secreted proteases (56).Purified SasG was incubated with saliva and then re-purified before N-terminal sequencing to determine the cleavage site.The results revealed a cut site after Arg-144, which falls between the A repeats and lectin subdomain (Fig. 3A).This is similar to one of the two reported cleavage locations in Aap ( 46), but it is somewhat surprising because the removal of the entire A domain was thought to be required for both Aap and SasG B domain homodimerization and subsequent aggregation (43,46).The cleavage of SasG by saliva was found to be dose-dependent (Fig. 3B), suggesting the presence of specific cleaving protease(s) inside the saliva.Therefore, purified SasG was incubated with saliva and protease inhibitors to identify the responsible protease(s).Minimal inhibition was seen with EDTA or PMSF alone, but in combination, they almost completely inhibited cleavage of SasG (Fig. 3C).This result suggests that saliva contains at least two protea ses, a metalloprotease and a serine protease, that process SasG and promote bacterial aggregation.
To test whether this truncated form of SasG could promote aggregation, we cloned both full-length and truncated versions of sasG and expressed them in strain USA300 LAC, which does not express a functional SasG on its own due to a frameshift mutation in its sasG gene (57).While expression of full-length SasG had only minimal effect on aggregation in buffer and required saliva to facilitate a full-scale aggregation, the  truncated version of SasG facilitated aggregation in buffer alone (Fig. 3D).This confirmed that removal of the 94 N-terminal amino acids of the A repeat region is sufficient to allow SasG to dimerize and promote aggregation.

Fractionation to identify host proteases processing SasG
Clarified saliva was concentrated, filtered, and passed over multiple columns to separate the proteins into fractions.First, we used anion exchange chromatography followed by size exclusion chromatography.These fractions were then tested to see whether they could cleave purified SasG by running the reactions on SDS-PAGE gels and looking for a shift in SasG size (Fig. 4A).The level of SasG cleavage was highest in fractions 19-22 and these fractions were used going forward.In parallel, we tested the response of the isolated active saliva fraction with protease inhibitors to determine the exact class of the enzyme.The most inhibition was observed with AEBSF, antipain, and leupeptin, suggesting the enzyme present in the active fractions is a serine protease (Fig. 4B).After electrophoresis separation of fractions with the highest activity (Fig. 4A), individ ual bands were extracted from the gel and the protein(s) identified by MALDI mass spectrometry.Seven proteases were detected in these bands with significant peptide coverage, including trypsin-1, prostasin, serine protease 27, and various cathepsins (Table S1).Considering the protease inhibitor patterns (Fig. 4B), the best hit from the proteomics assessment was human trypsin.

Validation of identified proteases
We used commercially available human trypsin to test SasG processing and promotion of S. aureus aggregation.We performed aggregation assays of S. aureus strains MW2 and 502a with a range of recombinant human trypsin concentrations (0.02-200 μg/mL) (Fig. 5A and B).At the same concentrations, human trypsin was incubated with purified SasG, and dose-dependent SasG processing was observed via SDS-PAGE (Fig. 5C).Starting at 0.02 µg/mL, we observed cleavage of SasG, which correlated with an increase in aggregation.The levels of SasG cleavage and aggregation increased in a dose-depend ent manner, reaching a peak at 2 µg/mL and decreasing slightly up to 20 µg/mL (Fig. 5A and B).These findings demonstrated that recombinant human trypsin can recapitu late the phenotype of SasG processing and S. aureus aggregation seen with saliva.At 200 µg/mL human trypsin, aggregation began decreasing due to degradation of the SasG protein (Fig. 5B).

Role of SasG in pneumonia model
To examine the biological relevance of SasG in vivo, we intratracheally infected mice with MW2 ΔmgrA (thus, SasG-expressing) or with ΔmgrA ΔsasG double mutant (Fig. 6).No evidence of systemic dissemination was observed in this model (Fig. 6A).The mice that were infected with the double mutant lacking SasG showed decreased number of colonies in the lungs (Fig. 6B), compared to the ΔmgrA strain expressing SasG.At the same time, markers of inflammation and tissue damage, that is the number of leukocytes (Fig. 6C) and level of protein (Fig. 6D) in the bronchoalveolar lavage (BAL), remain similar irrespective of the injected strain.The same trend of decreased bacterial counts and not significantly affected leukocytes and protein levels was also observed when a lower dose of S. aureus was used for infection (Fig. S1A through C).To validate our in vivo findings, we repeated our in vitro aggregation and SasG cleavage experiments with commercially available mouse proteases, airway trypsin-like protease (TMPRSS11D), and Kallikrein 1 (KLKB1), which are serine proteases found in the airways and function similarly to human trypsin (58-61) (Fig. S2 and S3).
We performed aggregation assays and SasG cleavage experiments as in Fig. 5 using murine proteases at a concentration range of 0.002-2 µg/mL.We observed that both murine proteases TMPRSS11D and KLKB1 could cleave SasG and induce aggregation at a concentration of 2 µg/mL (Fig. S2 and S3).Notably, TMPRSS11D, which is an ortholog to the corresponding human trypsin-like protease, induced significant S. aureus aggrega tion similar to both saliva and what was observed at the same concentration with human trypsin (Fig. S2).Overall, this suggests that during lung infection, the presence of SasG on the S. aureus surface has no effect on host response or local damage, but it does the benefit survival of the pathogen when faced with host immune response.Overall, the mouse pneumonia data indicate that the presence of SasG contributes to S. aureus virulence in vivo.

DISCUSSION
Roughly one-third to half of healthy individuals are colonized by S. aureus in the nasal cavity and/or nasopharynx (62)(63)(64).While S. aureus colonization is benign in healthy adults, the presence of S. aureus in the respiratory tract is the major risk factor for developing pneumonia in the intensive care unit (65,66).Despite the high rate of S. aureus carriage in the oral cavity, only preliminary studies have been performed on S. aureus interactions with human saliva proteins (67,68).S. aureus predominantly binds human proteins using microbial surface components recognizing adhesive matrix molecules (MSCRAMMs) (69).We have previously shown that the ArlRS/MgrA regulatory cascade controls the expression of MSCRAMMs and other surface proteins that function in adhesion and immune evasion (48).Strains lacking either arlRS or mgrA overexpress these surface proteins, and in this work, we made the surprising discovery that a S. aureus mgrA mutant aggregates in the presence of human saliva.We found that intercellular aggregation is dependent on the expression of SasG but also requires host factors in saliva to process SasG.
In previous studies, we demonstrated that full-length SasG is sufficient to block clumping and adhesion of cells by physically interfering with other surface proteins' ability to bind to host matrix components (48,51,70,71).However, SasG expression is low in S. aureus laboratory strains under standard in vitro conditions, which masks these clumping interference and aggregation phenotypes.Through our mapping of the sasG promoter and transcriptional reporter assay, we show that sasG expression is repressed by ArlRS/MgrA, and we identify a potential MgrA-binding site that overlaps with the sasG promoter.Therefore, inactivation of the ArlRS-MgrA cascade allows for high expression levels of sasG(Fig.7A) .
We also found that there is significant variation in sasG expression and molecular characteristics among strains.It was recently reported that most S. aureus strains have a full-length copy of SasG (49), but not all express SasG at detectable levels (57) and some strains have truncated copies of SasG.USA400 MW2 and 502a encode full-length, surface-attached copies of SasG with 5 B-repeats which aggregate with high efficiency.Bioinformatic analysis of the CF isolate AH4654 genome revealed the sasG, mgrA, and arlRS genes, and their respective promoter regions, are all essentially identical to MW2.Interestingly, this CF isolate expresses SasG and aggregates natively (Fig. 1), similar to other S. aureus isolates that fall into the ST15/CC15 grouping (39).By contrast, strain Newman, despite encoding a full-length SasG, does not present it on its surface and does not aggregate with or without MgrA.The reason SasG is not functional in Newman is unclear at this time.Strains such as USA300 LAC and N315 have truncated copies of SasG due to frameshift mutations and therefore cannot aggregate (57).Other strains like MN8 and MRSA252 do not possess sasG and any observed aggregation was likely due to another surface protein.
SasG is one of the key drivers of biofilm formation in S. aureus (38)(39)(40)45).SasG and its S. epidermidis homolog Aap consist of multiple domains with distinct functions (Fig. 1A).In S. epidermidis, the Aap A domain is known to be removed by the secreted metallopro tease SepA to facilitate biofilm accumulation (46), but native S. aureus-secreted proteases have not been found to cleave SasG in the same manner (38).Previous studies on S. epidermidis Aap also showed that exogenously added host proteases, such as trypsin and cathepsin G, could cleave Aap and enhance biofilm formation through processing (43).Our studies have found a parallel role for host proteases in cleaving S. aureus SasG and triggering aggregation (Fig. 7).
During infection, S. aureus uses mechanisms of aggregation and biofilm formation as the survival strategy to protect itself long term in response to environmental stressors, such as antimicrobials or host immune factors.S. aureus is a frequent colonizer of the oral cavity and can be isolated from dental plaques, periodontal pockets, and saliva, which represents a significant reservoir for opportunistic respiratory infections like aspiration pneumonia (55).It has been reported that up to 24%-36% of healthy adults harbor S. aureus in the oral cavity, with a higher incidence in individuals who are immunocompro mised or undergoing antibiotic treatment (55,68,72).Oral S. aureus has been found to interact with and bind several salivary components, which is associated with inducing adhesion and aggregation mechanisms (67,68,73).Oral colonization and poor immune clearance of aggregated S. aureus present in saliva, particularly in immunocompromised individuals, are major risk factors for the development of lung infections following saliva aspiration (53,55,68,72,73).Our data demonstrate that the upregulation of sasG is associated with increased aggregation upon interaction with human saliva, which is known to contain numerous proteases (74).Considering that the aspiration of saliva secretions is a common precursor to lung infection (75), our findings indicate that salivary proteases are capable of cleaving SasG at a single site within the A domain.This processing removed the 94 amino acids that compose the A repeats, exposing the lectin and B domains to interact with neighboring cells and homodimerize (Fig. 7B).We fractionated the proteases to identify human trypsin and validated them with both commercially available human trypsin and murine salivary proteases.However, additional serine and metalloproteases may also contribute to the processing of SasG.From an adaptive standpoint, S. aureus may have evolved a surface protein like SasG that is proteolytically labile, which can sense environmental conditions and facilitate aggregation to protect S. aureus under stress.Despite significant biochemical and structural studies on SasG, accompanied by experiments in vitro, no prior studies have determined its contribution to virulence in animal models of infection.However, simultaneous deletion of SasG and Eap did reduce insect mortality in a silkworm infection model (76).In this work, we now provide evidence that SasG contributes to S. aureus in the establishment of a lung infection.We demonstrated that SasG is important for S. aureus to survive and proliferate at the infection site and that interaction with host saliva in conjunction with aspiration may promote S. aureus survival by inducing aggregation.However, the presence of SasG did not impact the host response or damage to the host, suggesting it is solely important for S. aureus survival in a stressful environment.
In summary, we have shown that the global regulator MgrA controls the expression of the surface protein SasG.There is variation in the type and amount of SasG expressed among S. aureus strains, but expression of full-length SasG is associated with increased aggregation which is dependent on the presence of host proteases.We identified the serine protease human trypsin as a component of saliva that can process the SasG A domain to trigger aggregation (Fig. 7).Finally, we showed that SasG is important for full virulence in a S. aureus lung infection.

Reagents and growth conditions
S. aureus strains and plasmids used in this work are listed in Table 1.AH4654 is one of 75 clinical isolates, isolated from 10 pediatric CF patients and kindly gifted by the Starner Lab, University of Iowa.S. aureus was cultured in tryptic soy broth (TSB) or brain heart infusion (BHI) broth, and E. coli was cultured in lysogeny broth (LB) at 37°C with shaking at 200 rpm.Antibiotics were added to the media at the following concentrations: chloramphenicol (Cam), 10 µg/mL; erythromycin (Erm), 5 µg/mL; and tetracycline (Tet), 1 µg/mL.E. coli strains with plasmids were maintained on media supplemented with ampicillin at 100 µg/mL; kanamycin at 50 µg/mL; or spectinomycin at 50 µg/mL.Porcine trypsin and the Protease Inhibitors Set (Roche) were purchased from Sigma.Recombi nant proteases (human trypsin, mouse kallikrein 1, and airway trypsin-like protease) were purchased from G-Biosciences, Sigma-Aldrich, and ThermoFisher Scientific (Biotechne).Stimulated saliva was collected over 10-30 min by chewing on paraffin wax.Particulate material was removed by centrifugation, and this clarified saliva was stored at 4°C for up to 2 days.

Recombinant DNA and genetic techniques
E. coli DH5α and DC10B were used as cloning hosts for plasmid construction.Restric tion enzymes, DNA ligase, and Phusion DNA polymerase were purchased from New England Biolabs.The plasmid mini-prep and gel extraction kits were purchased from Invitrogen. S. aureus genomic DNA was purified using the Puregene yeast/bacteria kit B (Qiagen).Lysostaphin, used for S. aureus DNA extractions, was purchased from Sigma.Plasmids were purified from S. aureus RN4220 or E. coli DC10B and electroporated into MRSA LAC strains as described previously (77,89).Bacteriophage transductions between S. aureus strains were performed with phage 11 as described previously (90).All oligonucleotides were ordered from IDT (Coralville, IA) and are listed in Table 2. Routine DNA sequencing was performed at the University of Iowa DNA Core Facility or the Molecular Biology Service Center at the University of Colorado Anschutz Medical Campus.Whole-genome sequencing was performed at the University of Iowa DNA Core Facility with the Illumina MiSeq platform followed by de novo contig generation with the SPAdes genome assembler (91), and quality assessed with QUAST (92).Assemblies were annotated with Prokka (93).

RNA purification and RT-qPCR
Bacterial cultures were grown overnight in TSB and then subcultured to an OD 600 of 1.5.Cells were then pelleted and washed with RNAprotect Bacterial Reagent (Qiagen).To extract RNA, cells were lysed with lysostaphin for 30 min at room temperature, and RNA was purified using the RNeasy Mini Kit (Qiagen).Following RNA purification, genomic DNA was then removed using the Turbo DNase Kit (Ambion).cDNA was generated from a DNase-treated RNA template using the iScript cDNA synthesis kit (Bio-Rad).To perform quantitative PCR (qPCR), Primers KK15 and KK16 were used for sasG, and KK23 and KK24 for DNA gyrase (gyrB), as described previously (48).qPCR was performed by amplifying cDNA in 20 µL reaction volumes with iTaq Universal SYBR Green Supermix (Bio-Rad) in the CFX96 Touch Real-Time PCR System (Bio-Rad) under the following conditions: 3 min at 95°C, 40 cycles of 10 s at 95°C, and 30 s at 59°C, followed by a dissociation curve.No template and no reverse transcription controls were performed in parallel.Experiments were performed in biological triplicate with two technical replicates, and expression was normalized to gyrB.

sasG promoter mapping and GFP fusion plasmid
The sasG promoter was mapped using rapid amplification of 5′ cDNA ends (5′ RACE) (94).Template RNA was purified from MW2 ΔmgrA using the RNeasy Mini Kit (Qiagen) as previously described (48).Primers used were the general 5′ RACE primers (94) HC608, HC609, and HC610, and the sasG-specific primers HC611, HC612, and HC613.To generate the P sasG -GFP fusion plasmid, the region upstream of sasG was amplified using primers HC598 and HC599.The fragment was digested using XbaI and KpnI before ligating into pCM29 (87).The resulting plasmid, pHC127, encodes the sasG promoter upstream of an optimized ribosome binding site and codon-optimized gene for superfolder GFP.To assess expression, overnight cultures were diluted 1:100 in TSB containing chloramphe nicol in a black 96-well plate.Plates were incubated at 37°C with shaking in a humidified microtiter plate shaker (Stuart).A Tecan Infinite M200 plate reader was used to periodi cally measure OD 600 and fluorescence intensity with excitation at 495 nm and emission at 515 nm.Values represent averages and standard deviations of triplicate wells.

S. aureus aggregation assay
S. aureus cultures (5 mL) were grown overnight in TSB with shaking at 37°C.One milliliter of culture was harvested by centrifugation and the media was discarded.Saliva was collected and clarified by centrifugation and filtration.Recombinant human trypsin was activated by incubating 0.4 mg/mL with 0.2 µg/mL porcine trypsin for 1 h.Following activation, the recombinant trypsin was diluted in phosphate-buffered saline to the indicated concentrations.The cells were resuspended in 1 mL of either phosphate-buf fered saline, clarified human saliva, or activated recombinant human trypsin solution.
Tubes were allowed to sit for 1 h at room temperature, and then aggregation was visually assessed.For quantification of aggregation, 100 µL of liquid was removed from the top of the tube at 0 h and 1 h, and the optical density at 600 nm was measured in a 96-well plate in a Tecan infinite M200 plate reader.Measurements represent averages and standard deviations of experiments performed on three separate days.
To assess S. aureus aggregation induced by the recombinant mouse kallikrein one and airway trypsin-like proteases, S. aureus cultures were prepared as described above.Recombinant proteases were prepared and/or activated according to the manufactur er's guidelines to a final concentration of 100 µg/mL.Proteases were then diluted in phosphate-buffered saline to the indicated reaction concentrations.Due to volume limitations of the recombinant proteases, cells were then resuspended in 500 µL of either clarified human saliva, phosphate-buffered saline, or recombinant protease solutions and incubated for 1 h at room temperature.Aggregation was quantified as described above.

Cell wall preparations
For the preparation of cell wall proteins after aggregation assays, the tubes were centrifuged, and the cells were washed twice with PBS.The cells were resuspended in 500 µL of protoplasting buffer (10 mM Tris pH 8, 10 mM MgSO 4 , 30% raffinose).Lysostaphin (25 µg) was added and the cells were incubated for 1 h at 37°C.The tubes were centrifuged for 3 min at max speed, and 500 µL of supernatant was transferred to a new tube.Proteins were precipitated by adding 125 µL of cold trichloroacetic acid and leaving it on ice for 2 h.Precipitated proteins were pelleted by centrifuging at max speed for 10 min.The pellet was washed twice with 500 µL of cold 100% ethanol and then inverted to dry.The pellets were resuspended in 36 µL of SDS-PAGE loading dye, heated to 85°C, and then 10 µL was loaded on a 7.5% acrylamide gel.

Purification of full-length SasG
The sasG gene from S. aureus MW2 was amplified using primers HC416 and HC418 (Table 2), which remove the last 33 amino acids of SasG, including the LPXTG cell wall anchor, and replace them with a glycine followed by six histidine residues.This C-terminally tagged, secreted version of sasG was cloned into pALC2073 under the control of an anhydrotetracycline-inducible promoter, generating pHC90.We decided to purify this version of SasG from S. aureus LAC, which does not have an intact copy of sasG on the chromosome.To avoid potential proteolysis, we used a previously developed strain of LAC lacking secreted proteases (AH1919).In addition, we modified AH1919 to be resistant to anhydrotetracycline by integrating the empty vector pLL29 (95) in the phage 11 attachment site, generating host strain AH4607.
For expression of SasG, pHC90 was moved into AH4607 and a 5 mL culture was grown overnight at 37°C in TSB with chloramphenicol.This overnight culture was used to inoculate 1 L of TSB supplemented with chloramphenicol and 0.15 µg/mL anhydrote tracycline.The culture was grown with shaking for ~6.5 h at 37°C.Cells were removed by centrifugation, and the culture supernatant was concentrated to ~30 mL using an Amicon stirring pressure concentrator with a 100 kDa cutoff filter.The supernatant was dialyzed twice against binding buffer (50 mM sodium phosphate, 300 mM NaCl, pH 8).SasG-His6 was then purified using a pre-packed 5 mL IMAC cartridge (Bio-Rad) on a Bio-Rad FPLC.SasG-His6 was eluted with a linear gradient up to 100% elute buffer (50 mM sodium phosphate, 300 mM NaCl, 250 mM imidazole, pH 8).The protein was then concentrated and dialyzed against the storage buffer (20 mM sodium phosphate, 150 mM NaCl, pH 7.5).Glycerol was added to 20% before flash freezing and storing at −80°C.

SasG processing assays
Purified, full-length SasG was diluted 10-fold in phosphate-buffered saline, and 2 µL of this dilution was combined with 2 µL of water and 16 µL of clarified saliva or saliva fraction.Dilutions of recombinant human trypsin and murine proteases were prepared as described above.2 µL of SasG was added to 18 µL of recombinant human trypsin solution to initiate the reaction.Reactions were incubated for 1-3 h at 37°C as indicated.Processing was then quenched by adding 10 µL of 2× Laemelli SDS-PAGE loading buffer (Bio-Rad) and heating to 65°C. 10 µL of this was loaded on a 7.5% or 10% gel, or a 4%-20% gradient gel.For calculating the percentage of SasG processed, Coomassie-stained gels were scanned and quantified using Image Studio Lite (LiCor).

Identifying the cleavage site within SasG
A large SasG cleavage reaction was set up using 100 µL of purified SasG-His6, 900 µL of PBS, and 4 mL of clarified filtered saliva.The reaction was allowed to incubate for 1.5 h at 37°C.The solution was then exchanged to a binding buffer (same as above) using a 100 kDa molecular weight cutoff filter (Amicon).SasG-His was re-purified using HIS-Select resin (Sigma) and eluted with binding buffer containing increasing concen trations of imidazole.Fractions containing SasG-His were pooled and concentrated to ~0.5 mL, and 2, 4, 6, and 8 µL aliquots were mixed with SDS-PAGE buffer, boiled, and run on a 4%-15% gradient gel.Proteins were then transferred to a PVDF membrane using a Trans-blot Turbo transfer system (Bio-Rad) and the membrane was stained with Coomassie.N-terminal sequencing of cleaved SasG was carried out by Edman degrada tion using a Shimadzu PPSQ-53A Gradient Protein Sequencer at the Protein Facility at Iowa State University.

Partial purification of proteases from human saliva
Stimulated saliva (~90 mL) was collected over 1 day and centrifuged at 30,000 × g to remove debris.Clarified saliva was filtered and then concentrated to ~3 mL using 30,000 MWCO centricon concentrators (Amicon) and dialyzed against buffer A (20 mM Tris pH 8, 2 mM NaCl).The sample was then separated by anion exchange chromatography using a HiScreen Capto Q column (GE Life Sciences), eluting with a linear gradient up to 100% buffer B (20 mM Tris pH 8, 1 M NaCl).Fractions were tested using the SasG processing assay described above, except that tubes were incubated for 2 h at 37°C before running on an SDS-PAGE gel.Active fractions were pooled, concentrated to ~350 µL, and loaded on an SEC70 size exclusion column (Bio-Rad).The running buffer consisted of 20 mM Tris pH 8 and 100 mM NaCl.0.5 mL fractions were collected and tested for their ability to cleave SasG as described above, and a couple of fractions (20 and 21) was selected for further analysis.Protease inhibitors (Sigma) were used according to the manufacturer's instructions.For protein identification, bands were excised from an SDS-PAGE gel and analyzed at the Proteomics Facility in the University of Iowa Carver College of Medicine.

Pneumonia model
All mouse experiments were conducted in accordance with National Institutes of Health guidelines and previously approved by the University of Colorado Institutional Animal Care and Use Committee.Wild-type (WT) female BALB/c, 6-8 weeks old, were purchased from Jackson Laboratories (Bar Harbor, ME).Mice were anesthetized with isoflurane inhalation and challenged with approximately 2 × 10 8 colony-forming units (CFU)/30 µL of either mutant (ΔmgrA or ΔmgrA ΔsasG) S. aureus MW2 strain intratracheally.A blunt-tipped, bent 18 g Hamilton syringe was used to administer 30 µL of S. aureus directly into the lungs.Mice were left to recover for 24 h after which were euthanized using a lethal dose of ketamine/xylazine.The trachea was cannulated and the right lobes were tied off allowing for unilateral bronchial alveolar lavage (BAL) fluid isolation from the left lung.The right lobes were weighed and then homogenized for CFU determina tion.As a measure of lung inflammation and injury, leukocytes and protein in BAL fluid were measured.

FIG 1 S
FIG 1 S. aureus aggregates in the presence of human saliva and high SasG levels.(A) Schematic of SasG domains.(B-D) Overnight cultures of the indicated MRSA MW2 strains were spun down and resuspended in either phosphate-buffered saline or clarified human saliva.(B) The photo shows an aggregation of the ΔmgrA mutant after 1 hour of incubation at room temperature.(C) Coomassie-stained SDS-PAGE gel shows cell wall preps from these same samples after 1-h incubation as described above.The experiment is representative of at least three replicates.(D) Quantification of aggregation of MW2 WT with the empty vector pCM28, or ΔmgrA mutant with either pCM28 or the complementation vector pCM28-mgrA (pHC66) in the presence of saliva.Data represent averages and standard deviations of three separate experiments.Statistical significance was calculated by One-way ANOVA.****, P ≤ 0.0001; ns, not significant.(E) Various S. aureus strains with full-length, truncated, or lacking sasG were incubated with human saliva and aggregation was measured following 2 h of incubation.(F) Cell wall proteins were precipitated from overnight cultures and run on SDS PAGE to observe relative SasG expression levels.(G) Quantification of sasG gene expression of various S. aureus mgrA mutant strains relative to the respective wild-type sasG expression (n = 3).Values are normalized to gyrB expression in each strain.

FIG 2
FIG 2 SasG expression is regulated by ArlRS and MgrA.(A) Expression of a P sasG -GFP transcriptional reporter in the wild-type strain MW2 and isogenic ΔmgrA and ΔarlRS mutants.(B) Coomassie-stained SDS-PAGE gel of shed surface proteins from MW2, as well as ΔmgrA and ΔmgrAΔsasG mutants.The SasG band is indicated.(C) The transcription start site (in bold) of sasG was determined using 5′RACE.The ATG start codon is boxed, and putative −35 and −10 elements are shaded in gray.A potential MgrA binding site is underlined.

FIG 3
FIG 3 Saliva cleaves SasG within the A domain.(A)The sasG gene from S. aureus MW2 was cloned with a C-terminal His 6 tag in place of the cell wall anchor, allowing it to be purified from S. aureus culture supernatants.This purified, full-length SasG was then incubated with human saliva for 1.5 h, resulting in SasG cleavage (shown in Coomassie-stained gel on the left).Cleaved SasG was re-purified and subjected to N-terminal sequencing, which showed the cleavage site to be N-terminal to the lectin domain.(B) Human saliva was concentrated ~5-fold before generating a 2-fold dilution series.Purified SasG was then added, and the reactions were incubated for 1 h at 37°C.(C) Saliva was pre-incubated with either 2.5 mM EDTA, 2.5 mM PMSF, or both, before adding purified SasG.Reactions were incubated for 2 h at 37°C before resolving on an SDS-PAGE gel.SasG bands were quantified, and the percentage processed to the shorter product was calculated.Results are averages of three experiments, with statistical significance calculated by ANOVA.***, P < 0.001.(D) Aggregation of LAC strain, lacking its own SasG, and expressing from a plasmid either a full-length SasG construct or SasG construct with truncated N-terminal domain which replicates the effect of saliva processing.Aggregation was measured on S. aureus from overnight cultures suspended in saliva or PBS buffer for 1 h.N = 7. Coomassie-stained SDS-PAGE gels showing expression and processing of SasG constructs in each strain were prepared from cell wall preparations of the above-mentioned samples after the incubation.

FIG 4
FIG 4 Partial purification of SasG processing enzyme from human saliva.(A) Pooled active fractions from passing saliva over an anion exchange column were then passed over a size exclusion column.Coomassie-stained gel shows SasG cleavage by selected fractions from the size exclusion purification.Fraction numbers are indicated above the gel, and bars show a percent SasG cleavage for each fraction.Molecular weight standards in kDa are indicated on the left.(B) Aliquots of fraction 20 were pre-incubated with the indicated protease inhibitors for 15 min before adding SasG.Cleavage of SasG was measured after 1.5 h at 37°C by separating on an SDS-PAGE gel and quantifying percent cleavage.

FIG 5
FIG 5 Trypsin can process SasG and promote S. aureus aggregation.S. aureus (A) 502a or (B) MW2 mgrA and sasG mutant strains were resuspended in either saliva or PBS supplemented with recombinant human trypsin and allowed to aggregate for 1 h.(C) Purified full-length SasG was incubated for 1 h with either human saliva or serial dilutions of trypsin before running on an SDS-PAGE gel and staining with Coomassie.Measurements are averages and standard deviations of three separate experiments.Significance was calculated by one-way ANOVA ****P < 0.0001.

FIG 6
FIG 6 SasG is involved in S. aureus virulence in lung infection.Mice were infected intratracheally by S. aureus MW2 ΔmgrA and by its congenic strain ΔmgrA ΔsasG lacking SasG, and severity of pneumonia was assessed by weight loss (A), counting the CFU burden in lung homogenates (B), lung leukocyte recruitment in bronchoalveolar lavage (C), and protein infiltration in lavage fluid (D) after 24 h.The results were presented as means ± SEM, with statistical significance calculated by the Mann-Whitney test.*, P < 0.05.

FIG 7
FIG 7 Model of SasG transcriptional and post-translational regulation.(A) The expression of sasG is repressed by the ArlRS-MgrA regulatory cascade.At the post-translational level, host proteases such as trypsin can remove the N-terminal end of the A domain.(B) Removal of the end of the A domain (dark blue) allows SasG to oligomerize with SasG molecules on neighboring cells, resulting in aggregation of S. aureus.

TABLE 1
Bacterial strains and plasmids