The M Protein of Streptococcus pyogenes Strain AP53 Retains Cell Surface Functional Plasminogen Binding after Inactivation of the Sortase A Gene

Group A Streptococcus pyogenes (GAS) is a human-specific pathogen that produces many surface factors, including its signature M protein, that contribute to its pathogenicity. M proteins undergo specific membrane localization and anchoring to the cell wall via the transpeptidase sortase A. Herein, we explored the role of sortase A function on M protein localization, architecture, and function, employing, a skin-tropic GAS isolate, AP53, which expresses a human plasminogen (hPg)-binding M (PAM) Protein. We showed that PAM anchored in the cell membrane, due to the targeted inactivation of sortase A, was nonetheless exposed on the cell surface and functionally interacted with host hPg. We demonstrate that M proteins, and possibly other sortase A-processed proteins that are retained in the cell membrane, can still function to initiate pathogenic processes by this underappreciated mechanism.

ABSTRACT

Streptococcus pyogenes (Lancefield group A Streptococcus [GAS]) is a β-hemolytic human-selective pathogen that is responsible for a large number of morbid and mortal infections in humans. For efficient infection, GAS requires different types of surface proteins that provide various mechanisms for evading human innate immune responses, thus enhancing pathogenicity of the bacteria. Many such virulence-promoting proteins, including the major surface signature M protein, are translocated after biosynthesis through the cytoplasmic membrane and temporarily tethered to this membrane via a type 1 transmembrane domain (TMD) positioned near the COOH terminus. In these proteins, a sorting signal, LPXTG, is positioned immediately upstream of the TMD, which is cleaved by the membrane-associated transpeptidase, sortase A (SrtA), leading to the covalent anchoring of these proteins to newly emerging l-Ala–l-Ala cross-bridges of the growing peptidoglycan cell wall. Herein, we show that inactivation of the srtA gene in a skin-tropic pattern D GAS strain (AP53) results in retention of the M protein in the cell membrane. However, while the isogenic AP53 ΔsrtA strain is attenuated in overall pathogenic properties due to effects on the integrity of the cell membrane, our data show that the M protein nonetheless can extend from the cytoplasmic membrane through the cell wall and then to the surface of the bacteria and thereby retain its important properties of productively binding and activating fluid-phase host plasminogen (hPg). The studies presented herein demonstrate an underappreciated additional mechanism of cell surface display of bacterial virulence proteins via their retention in the cell membrane and extension to the GAS surface.

IMPORTANCE Group A Streptococcus pyogenes (GAS) is a human-specific pathogen that produces many surface factors, including its signature M protein, that contribute to its pathogenicity. M proteins undergo specific membrane localization and anchoring to the cell wall via the transpeptidase sortase A. Herein, we explored the role of sortase A function on M protein localization, architecture, and function, employing, a skin-tropic GAS isolate, AP53, which expresses a human plasminogen (hPg)-binding M (PAM) Protein. We showed that PAM anchored in the cell membrane, due to the targeted inactivation of sortase A, was nonetheless exposed on the cell surface and functionally interacted with host hPg. We demonstrate that M proteins, and possibly other sortase A-processed proteins that are retained in the cell membrane, can still function to initiate pathogenic processes by this underappreciated mechanism.

INTRODUCTION

Streptococcus pyogenes (Lancefield group A Streptococcus [GAS]) is a human-selective pathogen that is responsible for approximately 700 million infections/year of the throat and skin (1). These infections are usually self-limiting, and the bacteria are cleared by penicillin-type antibiotics. However, some strains of GAS are highly virulent and invade deep-tissue sites, causing diseases such as necrotizing fasciitis and toxic shock syndrome with multiple organ failure, as well as postinfection sequelae, e.g., glomerulonephritis and rheumatic fever. Approximately 40% of the severely infected patients expire after admission to an intensive care unit (ICU) (2).

At least 250 strains of GAS have been categorized through nucleotide sequencing of the hypervariable 5′ terminus of the signature emm gene (emm typing) that encodes the strain-specific surface-resident M protein (3). M proteins are genetically tuned to contain special sets of virulence determinants that assist the bacteria in evasion of the human innate immune response, often by conscripting host proteins for this purpose. For epidemiologic purposes, these distinct GAS strains are further subclassified into five pattern types, viz., nasopharynx-tropic patterns A to C, skin-tropic pattern D, and generalist pattern E strains (4). This subtyping is based on the presence and organization of a small subset of five closely linked emm paralog genes in a small region of the genome termed the multiple gene activator (mga) regulon, and all are positively regulated by the DNA binding protein, Mga (5).

In this communication, we focus on pattern D strains of GAS, the subclass containing the largest number of emm types (6). As an example, we employ the prototype GAS AP53, a well-studied pattern D GAS strain (7–10). M proteins from pattern D strains have the unique property of directly interacting with host human plasminogen (hPg), which is necessary for its pathogenicity, and are thus globally referred to as plasminogen-associated M (PAM) proteins. This PAM-bound hPg is then activated to the serine protease, plasmin (hPm), by a GAS-secreted streptokinase subtype (SK2b) that maximally activates hPg bound to PAM and not hPg in solution (11, 12). These critical events provide GAS with a proteolytic surface that is utilized by the bacterium for its dissemination through localized proteolysis of the extracellular matrix and epithelial/endothelial tight junctions (13–17). In this manner, pattern D strains of GAS can invade into deep-tissue sites and cause morbid and mortal diseases.

During translation, the nascent M protein is threaded through cell membrane secretory channels where it anchors within the cytoplasmic membrane via a single-pass hydrophobic transmembrane domain (TMD) positioned near the COOH terminus. Sortase A (SrtA), a transpeptidase cell membrane protein residing primarily in the division septum (18), functions to catalyze cleavage of a peptide bond between Thr and Gly of the LPXTG recognition sequence of its target protein that is located immediately upstream of the TMD in this class of proteins (19). Cleavage at the Thr-Gly bond allows the newly formed COOH terminus of Thr to form a peptide bond, via a required lipid II-linked intermediate (20), with the N terminus of the newly forming peptide cell wall cross-bridge, l-Ala–l-Ala (21). In this manner, the ∼600-Å-long M protein (22) is covalently anchored to the ∼70-Å-deep cell wall (23), thus allowing its NH₂-terminal region to easily extend through the wall into the extracellular space (24).

The full importance of SrtA processing for the surface display of PAM proteins is uncertain since the rod-like PAM (25), tethered to the mature cytoplasmic cell membrane via its TMD, can predictably span the cell wall and display a large region of its functional NH₂ terminus to the extracellular medium. While SrtA serves to allow covalent binding of M proteins to the peptidoglycan cell wall, we propose that M protein anchored in the cell membrane through its COOH-terminal hydrophobic domain would also be functional on the cell surface. This study presents a test of this important hypothesis.

RESULTS

GAS strains also contain two-component gene-regulatory systems, e.g., CovR (GenBank accession no. AMY96834.1) CovS mutant (GenBank accession no. AMY96835.1) involving an environmental sensor kinase (CovS) that communicates with and regulates an intracellular transcriptional regulator (CovR) (26). This system is used by the bacteria to adjust gene expression to changing niches in the host. In the original patient GAS isolate, strain AP53, the covS gene was mutated and CovS protein was not expressed (27). This allowed GAS AP53 to become highly virulent. We converted the covS mutant gene to wild-type (WT) covS⁺ in earlier work (27) and used the latter strain for these studies since our aim was to investigate the strain possessing a genome that was intact with regard to regulation of gene expression. The covS mutant isolate has a compromised genome and transcriptome compared to those of the covS⁺ strain (28).

Search of the AP53 genome for candidate SrtA substrates.

In order to identify all proteins in AP53 cells that would be affected by srtA gene inactivation, we conducted a bioinformatic search for potential SrtA (GenBank accession no. AMY97447.1) protein substrates in the genomic sequence (GenBank accession no. CP013672.1) of GAS AP53 (29), which is the parent strain used in this study. The search for candidate proteins included all of the following criteria: an NH₂-terminal signal sequence, a COOH-terminal SrtA consensus sequence (e.g., LPXTG), followed downstream by a short COOH-terminal hydrophobic transmembrane domain (TMD), and finally a small, positively charged COOH-terminal tail for cytoplasmic insertion. The search revealed at least 17 different proteins (Table 1) that are SrtA candidate substrates in GAS AP53, including the signature M protein, M53 (PAM), expressed by the emm53 gene. In examination of the resulting list in Table 1, the SrtA cleavage signal contained Leu¹ and Pro² residues as invariant, X³ as variable, and Thr as a highly preferred residue in the critical position 4 at the SrtA cleavage site. However, one other potential SrtA substrate was identified, viz., the chemokine-degrading protease ScpC (GenBank accession no. AMY96892.1), which contains Ala in place of Thr at position 4. Similarly, one protein fulfilling all other search criteria, the complement 5a-degrading protease, ScpA (GenBank accession no. AMY98238.1), contains Asn at sequence position 5 (LPTTN). Both proteins are known to be at the GAS surface (30–32). Consequently, both proteins are likely SrtA substrates.

TABLE 1.

Sortase A substrates in GAS strain AP53

Protein	Gene	GenBank no.	SrtA signal	SecA signal
Fibronectin binding protein	fbaA53	AMY98237	LPSTG	YSIRKLXXG
C5a peptidase	scpA53	AMY98238	LPTTN
Enn	enn53	AMY98239	LPSTG	YSLRKLXXG
M protein, PAM	emm53	AMY98240	LPSTG	YSLRKLXXG
IgG Fc receptora	fcR	AMY98241	LPSTG	YSLRKLXXG
Cell surface protein leucine-rich repeat		AMY97257	LPRTG
Protein G-related α2-macroglobulin binding protein	graB	AMY97671	LPTTG
Multifunctional 2′,3′-cyclic nucleotide, 2′-phosphodiesterase, 5′-nucleotidase, 3′-nucleotidase		AMY97282	LPITG
Collagen-like surface protein	sclB-1	AMY97374	LPATG
Collagen-like surface protein	sclB-2	AMY97371	LPATG
Collagen-like surface protein	sclA	AMY98216	LPATG
Fibronectin binding protein	prtF2, fbaB	AMY96676	LPATG
Serine endopeptidase ScpC/SpyCep	prtS	AMY96892	LPSAG
6-Aminohexanoate-cyclic-dimer hydrolase		AMY98224	LPQTG
Hypothetical protein		AMY97774	LPQTG
Endonuclease/exonuclease/phosphatase family protein		AMY97118	LPKTG
Pullulanase	pulA	AMY98209	LPKTG

^aEarly mutation at amino acid residue 85 to a stop codon in GAS-AP53 renders this gene nonfunctional.

Five of the candidate SrtA substrates that are listed in Table 1, viz., the gene products of fcr53 (an early termination pseudogene in GAS AP53), emm53 (PAM; GenBank accession no. AMY98240.1), enn53 (Enn; AMY98239.1), scpa53 (ScpA; AMY98238.1), and fbpA53 (FbpA; AMY98237.1), are involved in various aspects of the potential ability of GAS AP53 to evade the human innate immune response that would otherwise quickly eliminate the bacteria (33). The genes for these five bacterial proteins reside consecutively in the Mga regulon of the genome, and four of these also contain within their signal polypeptides the sequence YS(I/L)RKLXXG. This suggests that they are secreted through the Sec pathway and are targeted to the division septum of the bacterial cells in order to attach to the peptide cross-links in the newly forming cell wall (34, 35).

The domain structure of PAM that is the focus of this communication is illustrated in Fig. 1. The 427-amino-acid-residue protein contains, consecutively from its NH₂ terminus, a 41-residue signal peptide, followed by the hypervariable region (HVR), the A through D domains, the Gly/Pro (P/G) region, the SrtA cleavage site, the type 1 membrane insertion sequence (TMD), and, last, a short positively charged COOH-terminal cytoplasmic peptide region. Only the N-terminal A domain of mature PAM (residues 61 to 91) tightly binds host hPg, a major virulence factor for this strain of GAS (8, 36). We suggest that whether PAM is covalently attached to the cell wall via residue Thr³⁵⁴ or anchored in the cytoplasmic membrane by the downstream TMD, the N-terminal hPg binding region of PAM could be functionally present on the GAS surface, depending on the residence time of PAM in the cytoplasmic membrane. To address this issue, we inactivated the srtA gene in bacteria and examined the distribution and function of PAM in this altered strain.

FIG 1.

Schematic diagram of PAM protein. The full 427-amino-acid PAM sequence (molecular weight, 47,980) begins with a 41-amino-acid signal peptide containing a YSLRK motif for secretory channel selection. This is followed consecutively by the HVR, A (hPg binding), B, C, D, and proline/glycine (P/G) domains. Downstream of the P/G domain is the SrtA recognition region, L³⁵¹PXTG, followed by a hydrophobic transmembrane domain-spanning domain (TMD) and terminating with a short positively charged cytoplasmic COOH-terminal region. SrtA cleavage occurs between residues T³⁵⁴ and G³⁵⁵, resulting in the processed PAM being linked through T³⁵⁴ to a terminal Ala branch of the newly forming cell wall.

Generation of isogenic AP53 strains.

A targeting vector (TV) for deletion of the srtA gene by a double-crossover (DCO) strategy in AP53 covS⁺ cells (27) was designed. Here, the srtA gene was replaced by the chloramphenicol acetyltransferase (cat) gene, yielding the isogenic ΔsrtA strain (Fig. 2A), and then complemented into the srtA-deleted cells, yielding the complemented srtA-C strain, all in a targeted manner. These cells were used in this study, along with isogenic Δpam cells from previous work (27). Figure 2B shows that pam was absent in the genomic DNA (gDNA) of the Δpam cells (lane 2) but was present in the gDNAs of covS⁺ (lane 1), ΔsrtA (lane 3), and ΔsrtA-C (lane 4) cells. Further, the srtA gene was absent in the ΔsrtA cells (lane 8) but present in the covS⁺ (lane 6), Δpam (lane 7), and srtA-C (lane 9) cells. In addition, the cat gene was found in the Δpam (lane 12) and ΔsrtA (lane 13) cells, since this gene was inserted to replace pam and srtA, but was absent in the covS⁺ (lane 11) and srtA-C (lane 14) cells. These results show that the genomes of the gene-deleted and gene-complemented GAS cells were altered as expected.

FIG 2.

The construction and characterization of the isogenic mutated GAS cells. (A) Targeting vectors. (Step 1) A targeting vector (TV) was constructed for replacement of the srtA gene by the cat gene in GAS AP53 covS⁺ cells. In the TV, the cat gene is flanked by 340 bp of 5′ and 345 bp of 3′ srtA genomic fragments (striped boxes). This fragment was inserted into plasmid pHY304. (Step 2) After transfection and double crossover (DCO) into covS⁺ cells, the ΔsrtA cell line was generated. (Step 3) For targeted complementation of the WT srtA gene in the ΔsrtA cells, another TV was constructed, where 271-bp 5′ and 111-bp 3′ flanks (striped boxes) of the srtA gene were linked to the srtA cDNA in plasmid pHY304. (Step 4) After transfection and DCO with ΔsrtA cells, the complemented GAS line (srtA-C cells) was generated. (B) Genotyping. Confirmation of the mutations present in the GAS strains by PCR of gDNA. Four clones each were chosen for the following GAS cell lines, viz., covS⁺ (S⁺), ΔsrtA, srtA-C, and Δpam cells, and examined by PCR with the primers diagrammed in panel A. The amplicons with one of the clones are shown and are the same in all clones chosen. Lanes 5 and 10 represent a DNA ladder. The amplicon sizes in all cell lines were confirmed to correspond to their predicted sizes. (C) Growth curves of isogenic GAS strains. After 8 h of growth of single colonies of GAS cells in THY medium at 37°C in 5% CO₂, aliquots of the cells were diluted 1:10 with THY medium. After dilution, the OD₆₀₀ values were adjusted to 0.1, and the OD₆₀₀ was monitored with time as a measure of the growth rate of each isogenic strain. Each data point represents the average of OD₆₀₀ values from three different colonies. (D) Effect of growth phase on pam transcription in covS⁺ cells. Transcript levels of pam from covS⁺ cell lines at different growth stages, viz., early log (OD₆₀₀ of ∼0.3), mid-log (OD₆₀₀ of ∼0.55), and early stationary (OD₆₀₀ of ∼1.0) phases, were determined by qRT-PCR with specific primers for pam. The results are presented relative to those of the pam levels in covS⁺ cells at mid-log growth phase, which was set at 1. *, P < 0.0001 (for results compared to cells at mid-log growth phase with covS⁺ cells). (E) Transcript levels of pam and srtA from GAS cell lines at mid-log growth phase. The covS⁺, ΔsrtA, srtA-C, and Δpam cell lines were grown to mid-log phase, and transcript levels were determined by qRT-PCR with specific primers for each gene. The results are presented relative to pam and srtA levels in covS⁺ cells, which was set at 1. The GAS cell lines are labeled in the illustration. *, P < 0.001; **, P < 0.005; ns, not significant (for results compared to those with covS⁺ cells).

We first examined the growth rates of the isogenic mutant cells and found that these rates were very similar for covS⁺, srtA-C, and Δpam cells, while the ΔsrtA cells grew only slightly more slowly (Fig. 2C). The data show that cells with the SrtA mutation have a slightly lower growth rate, but the difference does not reach statistical significance. In order to evaluate the growth phase of the GAS cells maximally suited for subsequent experiments, we first estimated the expression levels of the pam transcript as a function of the growth phase of the cells (Fig. 2D). We found that the pam transcript is highly expressed at early (optical density at 600 nm [OD₆₀₀] of ∼0.3) and mid-log (OD₆₀₀ of ∼0.55) growth phases and is degraded, as expected, at the early stationary phase (OD₆₀₀ of ∼0.9) of growth, thus providing little new net pam expression at the stationary phase (37, 38). Thus, most subsequent experiments were conducted on cells at the mid-log phase of growth when expression of pam is optimal. At this growth stage, quantitative reverse transcription-PCR (qRT-PCR) results demonstrated that the pam mRNA level was elevated by ∼2.8-fold in the ΔsrtA cells and returned to near normal levels in the srtA-C cells, suggesting regulation of pam transcription by srtA in this GAS line. Additionally, transcription of srtA was only slightly upregulated in the Δpam cells (Fig. 2E).

Upon analysis of PAM protein levels in GAS cell supernatants using a PAM polyclonal antibody, we observed a small amount of PAM in covS⁺ cell supernatants at all three growth stages (Fig. 3A). However, PAM levels were highly enhanced in the supernatants of ΔsrtA cells, likely due to the elevated transcriptional rate of pam mRNA in ΔsrtA cells (Fig. 3A). Soluble PAM was reduced in srtA-C cell supernatants to levels seen in covS⁺ cells (Fig. 3A), confirming that a functional complementation of srtA had occurred. The gel inset of Fig. 3A shows that the intensity of the PAM bands in the supernatants generally reflected the enzyme-linked immunosorbent assay (ELISA) results in that ΔsrtA cells at mid-log growth phase produced far more PAM in the supernatants than the covS⁺ or srtA-C cells.

FIG 3.

PAM distribution in whole GAS cells. (A) Supernatants. The bacterial lines were grown to early log (OD₆₀₀ of ∼0.3), mid-log (OD₆₀₀ of ∼0.55), and early stationary (OD₆₀₀ of ∼0.9) phases, and ELISAs for PAM secretion were performed on the supernatants at various dilutions in 96-well microtiter plates after 7-min incubations. The inset represents Western blot analysis with anti-PAM of the cell supernatants at the mid-log phase of growth. (B and C) Cell walls and spheroplasts. Cells from covS⁺ (S⁺), ΔsrtA, and srtA-C cultures were grown to mid-log phase and then washed and treated with 100 μg of PlyC for 15 min. The cell wall fraction was separated from the spheroplast fraction and assayed for PAM content by Western blot analysis. The spheroplasts were disrupted prior to Western blotting by hypoosmotic lysis, and the lysates were applied to the gels. (D) The lysates were subjected to high-speed ultracentrifugation, and the membrane pellets were isolated, dispersed, and analyzed by Western blotting. In all cases of ELISA and Western analyses, the primary antibody was rabbit anti-PAM and the secondary antibody was goat anti-rabbit IgG conjugated with horseradish peroxidase (HRP). The wells were developed with 3,3′,5,5′-tetramethylbenzidine (TMB). Identical volumes of supernatants, cell walls, and spheroplasts were applied to each gel in each group of assays.

For approximate comparisons of band sizes, the recombinant PAM (rPAM) (Fig. 3A), which contains six His residues at its COOH terminus, has a calculated molecular weight of 40,950, which was confirmed earlier by mass spectroscopy (27). The molecular weights of PAM in the covS⁺ and srtA-C cells without the six His residues (calculated molecular weight, 40,130) were predictably slightly less than the weight of rPAM, suggesting that PAM in the supernatants of these cells represents a small level of PAM cleaved by SrtA, or a similar enzyme, but not bound to the cell walls. On the other hand, PAM in the ΔsrtA cell supernatants was larger than rPAM, and its size was consistent with the size of full-length mature PAM (calculated molecular weight, 43,640). This suggests that a portion of the full-length mature PAM in ΔsrtA cells may have emerged from the Sec channels into the AP53 culture medium, perhaps through an overload of this channel due to the large amounts of PAM in these cells.

PAM redistribution in AP53 ΔsrtA cells.

In order to further assess the degree of redistribution of PAM in cellular preparations of ΔsrtA cells, PAM levels were estimated in the cell walls (Fig. 3B) and spheroplasts (Fig. 3C) at the mid-log phase of cell growth. The data of Fig. 3B demonstrate that PAM is present in large amounts in the phage lysin C (PlyC)-digested (18) cell wall fractions of covS⁺ and srtA-C cells. The many PAM-positive bands present in these cell walls are due to numerous partially degraded cell wall fragments still attached to PAM, as would be expected after incomplete PlyC-catalyzed cleavage of the cell wall.

In cell wall preparations of the ΔsrtA cells (Fig. 3B), the molecular weight of the major PAM-positive band, albeit at a comparatively very low level, is approximately the same as that of rPAM. Since SrtA is not present and thus cannot be responsible for the PAM present in the cell wall, a candidate protease for such a cleavage is LPXTGase, an enzyme present in AP53 and other GAS cell lines, that catalyzes cleavage at the Gly⁵ residue of the SrtA consensus site (39). This would yield a protein similar in size to rPAM. These data suggest that LPXTGase might serve as a transpeptidase similar to SrtA, a consideration not necessarily surprising but also not heretofore known.

Figure 3C shows the presence of PAM in covS⁺-derived spheroplasts. The bands corresponding to PAM in covS⁺ and srtA-C membranes are shown to be larger than the band corresponding to rPAM. The larger size of PAM is expected if membrane-bound mature PAM retains its COOH-terminal domains downstream of the LPSTG domain (∼386 total residues; molecular weight, 43,640). These data also clearly reconfirm the effectiveness of the complementation of srtA in the srtA-C cells.

However, while PAM in ΔsrtA spheroplasts (Fig. 3C) retains this same higher-molecular-weight component, a subpopulation of PAM has a major band smaller than that of rPAM, in part due to the fact that rPAM has six additional His residues at the C terminus (molecular weight increase, 1,267) and also perhaps that a minor proteolytic event had occurred. Such proteolysis would represent a small fragment released from the HVR region of the NH₂ terminus (estimated as ∼40 residues based on its molecular size) with retention of the membrane-bound COOH terminus that still maintains PAM in the spheroplasts. Thus, this band is intact at its C-terminal TMD and possesses a short truncation at the N terminus.

In order to ensure that PAM was present in cell membranes, the spheroplasts were lysed and subjected to preparative high-speed ultracentrifugation for enrichment of the membrane pellet. After thorough resuspension of the pellets, the gel shown in Fig. 3D was obtained. This gel shows that cytoplasmic membranes of covS⁺ and srtA-C cells contain mature PAM that has not yet been cleaved by SrtA, whereas the membranes of ΔsrtA cells, similar to the spheroplasts, contain large amounts of PAM.

By size considerations alone, the PAM retained in the cytoplasmic membrane in the absence of SrtA processing should allow PAM to extend to the surface of the ΔsrtA GAS cells, projecting its NH₂ terminus into the GAS cell medium in a functional manner. As shown in previous work (40–43), peptides have been expressed in the absence of up to 55 NH₂-terminal residues without an effect on hPg binding and activation. This also seems to be the case here.

Function of PAM in ΔsrtA AP53 cells.

We evaluated the functional GAS surface exposure of PAM in the four different isogenic GAS strains. A very important property of GAS surface-localized PAM is that it serves as a singular tight-binding receptor for hPg via the NH₂-terminal A domain of PAM and the kringle 2 domain of hPg (K2_hPg) (8, 36). We have previously demonstrated that hPg binds avidly to PAM (dissociation constant [K_d] of ∼1 nM) through this binding mechanism (40), and we have also structurally modeled the X-ray crystal structure (44) and the nuclear magnetic resonance (NMR) solution structure and dynamics several of these complexes (43, 45–47). This work allowed us to propose a mechanism for this tight-binding interaction. The binding of hPg to each isogenic GAS strain was measured at 200 nM hPg, at which near saturation (80%) is achieved for each of the strains. The data of Fig. 4A show that hPg binds similarly to covS⁺ and srtA-C whole cells, showing that the levels of PAM on these cell surfaces are very similar. The data further show that isogenic strains in which the pam gene is inactivated (Δpam cells) interact with hPg to approximately 5% of the levels of the covS⁺ and SrtA-C cells, demonstrating that little binding of hPg occurs in the absence PAM and that PAM is the major tight-binding hPg protein on AP53 cells (10, 48, 49). In isogenic cells with a targeted inactivation of srtA, the PAM that is retained in the cytoplasmic membrane is nonetheless displayed on the GAS cell surface to ∼75% of the PAM level in covS⁺ cells (Fig. 4A). Very similar results were observed when the cells were titrated with a polyclonal antibody to PAM (data not shown).

FIG 4.

Binding and activation of hPg on GAS cells. (A) Binding by whole-cell ELISA. Each isogenic GAS cell line, viz., covS⁺ (S⁺), ΔsrtA, srtA-C, and Δpam cells, was grown to mid-log phase at 37°C in the presence of 5% CO₂ and then washed with PBS and resuspended in PBS–2% BSA for blocking. The cells were then incubated with hPg (200 nM), followed by incubation with the primary mouse anti-human Pg monoclonal IgG, after which the secondary antibody, rabbit anti-mouse IgG-HRP, was added. The color was developed with TMB. The reaction was then terminated by addition of 2 N H₂SO₄, and the A₄₅₀ and A₅₇₀ were determined. The reading at A₅₇₀ was subtracted from each of the readings obtained at A₄₅₀ in order to exclude the effect of optical aberrations due light scatter. *, P < 0.001 (for results compared to those with covS⁺ cells). (B) Activation rates of hPg bound to PAM on whole cells by SK2b. Mid-log-phase cells were washed with PBS, and the OD₆₀₀ was adjusted to 1.0. After cells were blocked with PBS-BSA, hPg (200 nM) was added to replicate wells of a 96-well microtiter plate. A solution containing SK2b (5 nM) and the chromogenic hPm substrate, S2251 (0.25 mM), was added next. The p-nitroanilide released from S2251 via hPm-catalyzed hydrolysis of S2251 was continuously measured by the A₄₀₅ at 25°C in a microtiter plate reader. The mA₄₀₅ was plotted against t², and mean slopes of the lines from duplicate measurements of four different clones of each isogenic strain were taken as the initial activation rate at 200 nM hPg. *, P < 0.005 (for results compared to those with covS⁺ cells). (C) Activation of hPg with PlyC-generated cell wall subfractions of GAS cells. Cell wall subfractions of mid-log-phase covS⁺, ΔsrtA, and srtA-C cells were assayed. Solutions of hPg (200 nM), S2251 (0.25 mM), and 20 μl of the cell wall fractions were added to 96-well microtiter plates, following which SK2b (5 nM) was added, and the rate of increase of the A₄₀₅ was measured. The mA₄₀₅ was plotted against t², and mean slopes of the lines were plotted for each cell wall preparation. In all cases the buffer was 10 mM HEPES–150 mM NaCl, pH 7.4, at 25°C. *, P < 0.0001 (for results compared to those with covS⁺ cells).

To examine whether the PAM presented to the extracellular environment in ΔsrtA cells possesses functional integrity, the activation capability of hPg in each of the isogenic cells was examined. In covS⁺ cells, hPg is properly aligned on PAM to undergo rapid activation by the coinherited SK subtype, SK2b (12). This PAM-mediated hPg binding (Fig. 4A) and its activation (Fig. 4B) are essential properties of pattern D GAS pathogenicity (50). In covS⁺ and srtA-C whole cells and in their isolated cell walls, SK2b effectively activates hPg (Fig. 4B), a feature that is greatly reduced in Δpam cells due to the lack of binding of hPg to PAM-deficient cells and the concomitant resistance of solution phase hPg toward SK2b activation (12). In ΔsrtA cells, hPg is effectively activated by SK2b to about 80% of the levels of covS⁺ and srtA-C cells (Fig. 4B) but not in the cell wall subfraction (Fig. 4C), demonstrating that functional PAM is present on the cell surface and originates from the spheroplast subfraction of the cells.

Properties of AP53 ΔsrtA cells that would influence their survival in the host.

We also examined a number of properties of ΔsrtA cells that would affect their survival in a hostile host environment. One such property is the ability to form biofilms, an important feature of bacterial survival (51). The data of Fig. 5A show that covS⁺ cells form adherent biofilms and that this property is highly attenuated after the deletion of the pam or srtA gene. Cells in which srtA was complemented into ΔsrtA cells regained the ability to form adherent biofilms. Thus, SrtA-processed, cell wall-anchored PAM plays a strong role in formation of adherent biofilms in AP53 cells, which is also reflected in the ΔsrtA-C cells.

FIG 5.

Functional properties of srt-deleted AP53 GAS cells. (A) Biofilm development. Four replicates of covS⁺ (S⁺), ΔsrtA, srtA-C, and Δpam cell lines were placed in 24-well microtiter plates and grown for 24 h in THY medium at 37°C in 5% CO₂. After the plates were washed, crystal violet was added to the wells for 15 min at room temperature. The solution was then removed, and the plate was washed, after which a 30% acetic acid solution was added. The A₅₅₀ was then recorded. *, P < 0.0001; ns, not significant (for results compared to those with covS⁺ cells). (B) Salt tolerance. Four replicates of the same four GAS strains were grown overnight and adjusted to the same OD₆₀₀ values (0.5). A 10× volume excess of THY–3% NaCl medium was added, and the growth rate was measured (OD₆₀₀) at 37°C in 5% CO₂ for 8 h. *, P < 0.006. (C) Gentamicin resistance. Overnight cultures of the GAS cell lines (3 replicates), adjusted to the same OD₆₀₀ (0.1) values, were placed in various concentrations of gentamicin in a 96-well microtiter plate and allowed to grow for 24 h at 37°C in 5% CO₂ in THY medium. The OD₆₀₀ values were then measured. The gentamicin concentration is indicated on the graph. *, P < 0.02; **, P < 0.0005 (for results compared to those with covS⁺ cells at the indicated gentamicin concentration). (D) Human keratinocyte cytotoxicity. Overnight cultures of the GAS cell lines were washed, resuspended in THY medium, and normalized to the same OD₆₀₀ values. The bacterial cells were added at an MOI of 10 to keratinocytes (HaCaT) that were 90% confluent. After incubation at 37°C in 5% CO₂ for 6 h, the bacterial suspensions were removed, and ethidium homodimer was added. The samples were left for 30 min in the dark. Fluorescence levels were then read at 617-nm emission and 528-nm excitation. A second set of readings was then taken after addition of saponin. The percent membrane permeabilization values were obtained by dividing the fluorescence level pre-saponin treatment by the level post-saponin treatment. V, vehicle control. *, P < 0.006; ns, not significant (for results compared to those with covS⁺ cells).

In addition, the cytoplasmic membrane is somewhat compromised in ΔsrtA cells, as shown by growth rate studies under high-salt conditions or in the presence of the antibiotic gentamicin. In both cases, the ΔsrtA cells showed decreased growth rates at 3% NaCl (Fig. 5B) and at concentrations of gentamicin of >3 μg/ml (Fig. 5C), compared to the levels in covS⁺, srtA-C, and Δpam cells, which were less affected under the conditions of the experiments. These data suggest that a deficiency of PAM alone does not compromise the integrity of the cell membrane, but an accumulation of SrtA-dependent proteins in the membrane as a result of inactivation of the srtA gene slightly increases membrane porosity to small molecules but not to an extent that influences cell growth in the absence of these challenges.

Last, ΔsrtA cells display a greatly reduced ability, compared to that of covS⁺ and srtA-C cells, to kill human keratinocytes, as shown by the keratinocyte membrane permeability studies of Fig. 5D. This property is not as significantly reduced in Δpam cells. Overall, the data suggest that AP53 ΔsrtA would not be virulent in a human host environment and would be eliminated by innate host response mechanisms, a conclusion that agrees with whole-animal findings. However, since the covS⁺ line is weakly virulent and since the ΔsrtA GAS cells are expected to be even less virulent, whole-animal studies were not performed, and our conclusions, which may also apply to other GAS lines, are based on in vitro data.

DISCUSSION

The cell walls of Gram-positive bacteria assist the organism in its survival. Not only does the cell wall provide a major passive function as a rigid exoskeleton in protection of the cytoplasmic membrane against hypotonic death of the bacteria, but it also serves actively in binding proteins, carbohydrates, and lipids that are displayed on the bacterial surface. In turn, these surface-exposed molecules are involved in growth and division of the bacterial cells, and, importantly, they also serve as adhesins and invasins for eukaryotic host cells. Additionally, proteins on the GAS surface assist GAS in evading innate immune systems of the host, e.g., modulating entrapment of the bacteria by neutrophil extracellular nets and/or fibrin, resistance to antimicrobial peptides, and complement-mediated opsonophagocytosis, among others. It is thought that in order for many of these bacterial surface proteins to reach the surface of Gram-positive cells, the unfolded cytoplasmic proteins must be sorted, cotranslationally or posttranslationally, by migration through the cytoplasmic membrane secretory channels to allow their proper folding and migration to the cell membrane in order to be transported to the outer bacterial surface.

While Gram-negative bacteria employ an outer membrane to stabilize proteins at the cell surface, this is not the case with Gram-positive bacteria that do not have an outer cell membrane but, instead, contain a larger, highly cross-linked peptidoglycan cell wall, and in some organisms, e.g., GAS, an outer polysaccharide capsule is present (27). Proteins at the Gram-positive surface exist in various conformations, from fibrillar elongated molecules to globular proteins. These surface proteins are linked by at least four types of binding mechanisms, as follows: (i) globular proteins that interact with the surface through noncovalent interactions (52, 53); (ii) lipoproteins with Cys side chains in their leader sequences that interact as thioethers with diacylglycerol of the extracellular surface of the cytoplasmic membrane (54); (iii) Srt-dependent proteins that covalently bind at their COOH termini to the peptide cross-bridge of the newly forming cell wall (55, 56); and (iv) other types of molecules, e.g., cell wall-bound teichoic acids linked to the cell wall by phosphodiester bonds (57) and cytoplasmic membrane-anchored lipoteichoic acids which can also serve to bind to proteins (57).

The cotranslational signal recognition pathway (SRP), which is present in all GAS types, is a mechanism for translocating presecretory proteins to the cell wall (58). An alternative posttranslational mechanism for this same purpose involves the chaperone SecB, which stabilizes the unfolded protein prior to transport. However, SecB is not present in GAS. Thus, the latter mechanism is not functional in GAS cells, and intact Srt-dependent proteins cannot be present in the cytoplasm.

Thus, the SRP pathway is the primary pathway for translocating M proteins (and other like proteins) across the cell membrane into the extracellular environment. To access this pathway, a signal peptide is needed, and, in the case of PAM, this signal peptide contains the YSIRK sequence, which targets PAM to the division septum of the developing cell wall, covalently attached to the l-Ala–l-Ala cross-link (34). In this case, as PAM traverses the cell membrane, it is halted in this membrane by its TMD, and SrtA then processes PAM prior to attachment to the cell wall. For the SRP to be fully functional, the fifty-four homolog protein (FfH; GenBank accession no. AMY97480.1) and filament temperature-sensitive Y (FtsY; GenBank accession no. AMY97024.1) must be present, and we find that both full-length genes encoding these proteins are present in the genome of AP53 cells (29). The SRP system then delivers the emerging protein to the Sec translocation system, viz., the membrane ExPortal, that contains these translocons (59).

In the absence of srtA, with the PAM delayed in the membrane, a protease, e.g., SpeB and or HrtA, might be more available to cleave the retained PAM. Interestingly, the membrane ExPortal in AP53 cells also contains the proteases SpeB (GenBank accession no. AMY98254.1) and HrtA (AMY98408.1), which may function in cleaving an NH₂-terminal peptide from the HVR of the PAM population that is delayed in the cell membrane in ΔsrtA cells and thus display the smaller spheroplast-residing PAM (Fig. 3D). Further, the PAM present in the cell supernatants is also of slightly higher molecular weight than rPAM and PAM in covS⁺ and srtA-C cells, which is consistent with a different proteolytic event that may have even occurred in the supernatant itself. These cleavages do not affect the hPg binding affinities of the resulting PAM.

According to the mechanism described, the PAM retained in the cell membrane would still extend to the GAS surface and possess its tight hPg binding NH₂-terminal A domain since only a maximum of 40 to 50 residues is lost from the NH₂ terminus (41, 43, 60). Thus, these minor truncations in this smaller PAM in the spheroplasts of ΔsrtA cells are not functionally important in terms of membrane retention and binding to hPg.

In summary, we conclude that in the absence of srtA, PAM is highly transcribed, is retained in the cytoplasmic membrane, is processed through the intact Sec system, spans the cell wall, and is exposed to the extracellular environment. SrtA potentially mediates cell wall anchoring of at least 17 proteins in GAS AP53 cells. In the absence of SrtA, many of these proteins are also likely retained in the cytoplasmic membrane by their TMDs, and some of these proteins might still be exposed on the cell surface. While other properties of GAS cells are shown to be affected by the srtA deletion, due to a more porous cell membrane likely caused by the retention of numerous missorted proteins, we nonetheless have identified another mechanism that could account for the presence of functional PAM on the cell surface. This mechanism could occur in part in regions of the cell that are functionally low in SrtA and/or accessory proteins.

MATERIALS AND METHODS

Generation of a targeted deletion of the srtA gene.

For construction of the targeting plasmid for the srtA deletion in covS⁺ cells, the cat gene (660 bp), flanked by 348 bp upstream of the ATG signal peptide for srtA and 435 bp downstream of the TAG stop codon for srtA, was constructed. During this process, the restriction endonuclease sites 5′ NotI and 3′ XhoI were also cloned into the two ends of the entire DNA segment by PCR primers and used for insertion into the same sites of the temperature-sensitive plasmid pHY304 (from M. J. Walker, Queensland, Australia). This plasmid also contained the erythromycin resistance (erm) gene downstream of the multiple-cloning site. The resulting targeting plasmid was then transformed into covS⁺ cells by electroporation. Chromosomal integration via allelic replacement was achieved by single crossover (SCO) at 30°C for 2 h for plasmid replication in Todd-Hewitt broth–1% yeast extract (THY)–0.25 M sucrose medium and then switched to 37°C overnight on blood agar plates for double crossover (DCO) to occur. These steps resulted in the replacement of the srtA gene and the targeting vector backbone by the cat gene, as we have described previously (27). The screened colonies from the blood agar plates that grew in THY medium, but not erythromycin-THY medium, due to the loss of the erm gene after DCO, were further selected by their resistance to chloramphenicol, indicating the presence of the cat gene. For confirmation, the DCO-positive colonies were checked by PCR with primers internal to the cat and srtA genes. The final isogenic GAS cells, with the srtA gene replaced by cat, are referred to as ΔsrtA cells.

Targeted complementation of the srtA gene.

The srtA gene was complemented into the ΔsrtA GAS cells through a similar approach by replacing the targeted cat gene. A DNA fragment comprising the entire srtA gene, along with 271 and 111 bp of its 5′ and 3′ genomic flanking regions, respectively, was isolated by PCR. BamHI and EcoRI restriction endonuclease sites were incorporated by PCR into the 5′ and 3′ ends, respectively, of the fragment for ligation into these same sites of plasmid PHY304. The resulting plasmid was then transformed into the ΔsrtA strain, and SCO/DCO was accomplished as described above. Proper genome allelic replacement of srtA in the ΔsrtA strain was confirmed by the presence of an 840-bp amplicon spanning 348 bp of 5′ flanking sequence upstream of the ATG of srtA to nucleotide 492 of srtA and, similarly, by a 952-bp amplicon spanning nucleotide 517 of srtA to 435 bp downstream of the TAG stop codon of srtA, as well as by PCR using internal primers that amplify only srtA. The correctly complemented DCO clones also lacked the cat and erm genes. Confirmation of the complementation was accomplished through sequencing of the srtA gene. The final isogenic GAS cells are referred to as srtA-C cells.

Targeted deletion of the emm (pam) gene.

The isogenic AP53 Δpam cells were constructed similarly to the AP53 ΔsrtA line via targeted insertion of the cat gene in place of the pam gene. The details have been described previously (27). These GAS cells are referred to as Δpam cells.

Genotyping.

In order to prepare gDNAs, single colonies of the GAS lines were selected from streaks on sheep blood agar and grown in THY medium overnight at 37°C in 5% CO₂. The cells were treated with lysozyme-proteinase K and suspended in 100 mM Tris–5 mM EDTA–0.2% SDS–200 mM NaCl, pH 8.5. Genomic DNA was extracted with phenol-chloroform-isoamyl alcohol (25:24:1), precipitated with isopropanol, and washed with 70% ethanol in nuclease-free H₂O.

PCR was employed with gDNA from the various strains to determine whether the desired gene alterations were present in the genomes. To detect the srtA gene in covS⁺ and srtA-C GAS lines, 24-bp internal primers that consisted of a forward primer (srtA-F) beginning at nucleotide 233 of srtA and a reverse primer (srtA-R) beginning at nucleotide 492 of srtA were used. PCR of the gDNA with these primers resulted in an amplicon of 260 bp when srtA was present. In the case of the srtA and pam gene deletions in both the ΔsrtA and Δpam strains, PCR was employed for analysis of the presence of the cat gene in each. The internal 20- to 21-bp primers used were cat-forward (cat-F) beginning at nucleotide 241 of the cat DNA and cat-reverse (cat-R) beginning at nucleotide 460 of the cat DNA. PCR of the gDNA provided an amplicon of 220 bp when the cat gene was present. Also, PCR of gDNA using an external 5′ primer upstream of the srtA or pam gene with cat-R and PCR of gDNA with an external 3′ primer downstream of the srtA or pam gene with cat-F provided correct amplicons in both cases, showing that cat was present and appropriately targeted in each of the strains.

For genotyping of pam, a 21- bp forward (pam-F) primer beginning at bp 1081 of the pam gene and a 23-bp reverse primer which began at bp 1387 of the pam gene yielded an amplicon of 308 bp when pam was present.

qRT-PCR.

Transcript levels of the genes of interest were measured by qRT-PCR. For these experiments, total RNA was extracted from GAS cells at their mid-log growth phases (OD₆₀₀ of ∼0.55 to 0.6). For this step, 3- ml aliquots of cell cultures were first centrifuged, and the cell pellet was resuspended in 0.2 ml of 20 mM Tris-HCl–10 mM MgCl₂–26% raffinose–100 μg/ml each of chloramphenicol and streptomycin, pH 6.8, with the addition of 25 U of mutanolysin or 0.1 mg of PlyC. After a 2-h incubation at 37°C, the cells were again pelleted and resuspended in RLT cell lysis buffer–1% 2-mercaptoethanol (Qiagen, Valencia, CA). The RNA isolation then continued as described for a Qiagen RNeasy minikit which also contained RNasin. Any remaining gDNA contamination was eliminated by two treatments of the sample with 20 Kunitz units of DNase I (Qiagen). For qRT-PCR, three independent extractions of total RNA from each of the isogenic GAS strains were used. qRT-PCR was performed as described previously (27).

Gene expression levels were assessed for srtA, pam, and gapdh (plr) RNAs using the internal gene-specific primers in covS⁺, ΔsrtA, srtA-C, and Δpam extracts. For srtA, the same primers were used in both genotyping and in transcriptional analysis. For pam mRNA analysis, we employed a different primer set than that used for gDNA analysis of pam. Here, message levels of pam were assayed by detection of a 385-bp amplicon with forward and reverse primers comprising bp 37 to 59 and bp 407 to 421, respectively, of pam. Similarly, for analysis of the levels of the reference housekeeping gene transcript, gapdh (plr), forward and reverse primers from bp 421 to 443 and bp 623 to 654, respectively, yielded a gapdh amplicon of 234 bp.

Three independent extractions of total RNA from each of the strains were used. Relative gene expression levels were analyzed by the 2^ΔΔCT (where C_T is threshold cycle) method (61). The statistical means of triplicate C_T values were calculated for the target and gapdh reference genes from all strains.

Growth curves of various isogenic GAS strains.

Single colonies of GAS cells were grown at 37°C in THY–5% CO₂ for 8 h to reach the stationary phase. A 5-ml aliquot was removed and inoculated into 45 ml of THY medium. After small adjustments to equalize the initial OD₆₀₀, growth rates were monitored by changes in the OD₆₀₀ of 1-ml samples measured at regular intervals.

PAM content of cell supernatants.

The supernatants were harvested by centrifugation and sterile filtered with 0.22-μm-pore-size polyvinylidene difluoride (PVDF) membrane vacuum filters. The resulting supernatants were concentrated 8-fold using 30-kDa-cutoff centrifuge concentrators. Serial dilutions (0- to 16-fold) of each fraction were prepared and added in replicates into wells of a high protein binding 96-well microtiter plate. The wells were then washed and blocked with 1% bovine serum albumin (BSA) at room temperature. Subsequently, rabbit anti-PAM polyclonal antibody was added, followed by anti-rabbit IgG conjugated with horseradish peroxidase (HRP; Bio-Rad). In order to assess the relative levels of PAM in the supernatants of the GAS strains, 100 μl of the HRP chromogenic substrate, viz., 3′,5,5′-tetramethylbenzidine (TMB; R&D Systems), was added and incubated for 7 min. At this time, the reaction was terminated by acidification of the medium using 50 μl of 2 N H₂SO₄, and the A₄₅₀ was determined.

For Western blot analysis, the samples (24 μl) were mixed with 6 μl of 5× SDS-PAGE loading buffer and boiled for 5 min, and then 10 μl of each sample, along with rPAM (100 ng), was added to individual lanes of a 10% SDS-PAGE gel. The gels were then used for Western blotting by addition of the in-house polyclonal rabbit anti-PAM as the primary antibody and goat anti-rabbit IgG–HRP as the secondary antibody. The gels were developed with luminol-peroxidase (Thermo-Fisher).

PAM content of isolated GAS cell walls and spheroplasts.

Cells were fractionated, and cell wall and spheroplast fractions were collected as previously described (62), with modifications. Specifically, overnight cultures of the GAS strains were grown to mid-log phase (OD₆₀₀ of ∼0.55 to 0.6) in THY medium at 37°C in 5% CO₂. The cultures (40 ml) were pelleted and washed twice with 5 ml of cold phosphate-buffered saline (PBS). The cells were then resuspended in 0.8 ml of PBS–30% raffinose–100 μg of the bacteriophage lysin PlyC (18) and incubated for 15 min at room temperature. The samples were centrifuged in a microcentrifuge, which provided cell wall samples in the supernatants and spheroplasts in the pellets. The supernatants were removed for Western blotting. The spheroplast pellets were lysed by resuspension in 1 ml of sterile H₂O for osmotic lysis, briefly vortexed, and used for Western blotting as described above.

To isolate the cell membrane in spheroplast lysates, 500 μl of each sample was digested with 20 μl of DNase I (2 units/μl) for 20 min at 37°C and then centrifuged at 165,000 × g for 2.5 h at 4°C. The soluble phase was removed, and the membrane pellet was resuspended in 520 μl of buffer.

Binding of hPg to isogenic GAS cells by whole-cell ELISA.

The binding of hPg to each of the GAS strains was measured by whole-cell ELISA. GAS cells grown to mid-log phase (OD₆₀₀ of ∼0.55 to 0.6) at 37°C in 5% CO₂ were washed with PBS and resuspended in PBS–2% BSA for blocking. Approximately 10⁸ bacterial cells of each strain were incubated with 200 nM hPg for 30 min at 25°C. The cells were then incubated with monoclonal mouse anti-hPg IgG, followed by washing and incubation with rabbit anti-mouse IgG–HRP for an additional 30 min at 25°C. All antibody incubations were conducted in microcentrifuge tubes in PBS–1% BSA, followed by centrifugation of the cells at 8,000 rpm for 5 min prior to the next addition. Finally, after adjusting the OD₆₀₀ of all the samples to ∼0.15, the antibody-treated and washed cells (40 μl) were transferred to individual wells of a 96-well microtiter plate, followed by addition of 100 μl of TMB. Color development progressed for 7 min at 25°C. The reaction was then terminated by addition of 2 N H₂SO₄, after which the A₄₅₀ values were determined. In addition, the OD₅₇₀ was determined, which was subtracted from the A₄₅₀ values in order to correct for any light scatter due to cells.

Activation of hPg in the presence of isogenic GAS strains.

Bacterial strains were grown to the mid-log phase (OD₆₀₀ of ∼0.55 to 0.60) and then washed with PBS, after which the OD₆₀₀ of the cell suspension was adjusted to 1.0. The cells (50 μl) were then used for the hPg activation assay. For this, an equal volume of 2% BSA was added to 50 μl of washed cells and incubated for 10 min at 25°C. These steps were followed by the addition of 200 nM hPg to replicate wells of a 96-well microtiter plate. Streptokinase subclass 2b (SK2b) and the chromogenic substrate d-Val-Leu-Lys-p-nitroanilide (S2251; Diapharma, West Chester, OH) were added together at final concentrations of 5 nM and 0.25 mM, respectively, to the reaction mixture. The release of p-nitroanilide from the substrate catalyzed by the hPm formed was continuously measured at 405 nm and 25°C in a microtiter plate reader. To obtain initial rates of activation under these conditions, the mA₄₀₅ was plotted against time (t; as minutes²). The slopes of the lines and linear regressions of the data were obtained using GraphPad Prism, version 8. The data were calculated as the means ± standard errors of the means (SEM) from four independent experiments.

Functional consequences of an srtA deletion.

(i) Biofilm formation. Biofilms were generated from GAS covS⁺, ΔsrtA, srtA-C, and Δpam lines and stained with crystal violet as described previously (63) with modifications. Overnight single cultures of GAS were diluted 1:100 into fresh medium, and 500 μl of diluted culture was added to four wells each of 24-well cell culture-treated plates. The plates were then incubated at 37°C in 5% CO₂ for 24 h. The supernatants were then removed, and the plate was washed twice via gentle submersion in, and subsequent shaking out of, 1× PBS. The wells were then stained with 500 μl of 0.1% crystal violet for 15 min at room temperature. A set of blank wells was also stained as background-staining blanks. The crystal violet was then removed, and the plate was washed twice and then left to dry upside down overnight. For quantification of staining, 500 μl of 30% acetic acid (HOAc) was added to each well and incubated at room temperature for 15 min. Afterwards, 200 μl from each well was transferred to a 96-well plate, and the A₅₅₀ was measured in a plate reader as a measure of attached biofilm generation.

(ii) Salt tolerance. Bacterial cultures containing covS⁺, ΔsrtA, srtA-C, and Δpam cells were inoculated from single colonies and incubated overnight. The cultures were then normalized to an OD₆₀₀ of 0.5, and 2 ml of this culture was added to 20 ml of fresh THY medium supplemented with 3% NaCl. The OD₆₀₀ was measured, and the cultures were then incubated at 37°C in 5% CO₂ for 8 h, and the OD₆₀₀ was measured every hour. The OD values were then plotted against time of growth.

(iii) Antibiotic resistance. Gentamicin (10 mg/ml in sterile H₂O; Sigma-Aldrich) was diluted in THY medium to 2× the final concentrations used, and 100 μl of each concentration was added to three wells for each strain in a 96-well plate. THY medium (100 μl) without gentamicin was added to four wells as a positive control for growth, and 200 μl of THY medium without bacteria was added to four wells as a medium control. Overnight bacterial cultures were normalized to an OD of 0.1, and 100 μl of each strain was added to the appropriate wells. The plate was then incubated at 37°C in 5% CO₂ for 24 h. After incubation, the wells were mixed by pipetting, after which the plate was placed in a SpectraMax Plus 384 microplate reader and shaken for 30 s, and the OD₆₀₀ was measured. The OD₆₀₀ readings were then adjusted using the medium blank and plotted.

(iv) Infectivity of human keratinocytes. Human keratinocytes (HaCaT) were grown to 90% confluence in Dulbecco’s modified Eagle’s medium (DMEM)–10% fetal bovine serum (FBS) using 24-well cell culture-treated plates at 37°C. The covS⁺, ΔsrtA, srtA-C, and Δpam cell lines were also grown in THY medium at 37°C overnight. Prior to infection, spent medium was removed from the HaCaT. Next, the cells were washed with PBS, and THY medium was added. Bacterial overnight cultures were centrifuged, and the pellets were resuspended in new THY medium. After normalization of the OD₆₀₀ values, the bacteria were added to the HaCaT at a multiplicity of infection (MOI) of 10. Infected HaCaT were incubated at 37°C in 5% CO₂ for 6 h. The supernatant was removed from the HaCaT monolayers at the end of the infection and then washed with PBS. After the washing step, 250 μl of ethidium homodimer solution (4 μM in PBS; Molecular Probes) was added to the wells, and the plates were incubated in the dark for 30 min. The fluorescence level was determined using a plate reader set to 528-nm excitation and 617-nm emission (cutoff, 590 nm). Saponin (0.1%, wt/vol) was then added to each well of the plates. The plates were rocked for 20 min in the dark, and then a second fluorescence reading was taken. The percent membrane permeabilization values were obtained by dividing the first fluorescence reading (posttreatment) by the second reading (post-saponin addition).

Statistical analyses.

Statistical analyses of three to four individual replicates were performed using Student's t test and calculated as the means ± standard errors of the means (SEM). A P value of <0.05 between groups using analysis of variance (ANOVA) was considered statistically significant.