Characterization of M-Type-Specific Pilus Expression in Group A Streptococcus

ABSTRACT

In addition to providing a typing mechanism for group A Streptococcus (GAS) isolates (T typing), cell surface pilus production impacts GAS virulence characteristics, including adherence and immune evasion. The pilus biosynthesis genes are located in the fibronectin- and collagen-binding T-antigen (FCT) region of the genome, and nine different FCT types, encoding more than 20 different T types, have been described. GAS isolates are not uniform in their degree or pattern of pilus expression, as highlighted by pilus production being thermoregulated in isolates that harbor the FCT-type FCT-3 (e.g., M-types M3 and M49) but not in isolates that harbor FCT-2 (e.g., M-type M1). Here, we investigated the molecular basis underlying our previous finding that M3 GAS isolates produce pili in lower abundance than M1 or M49 isolates do. We discovered that, at least in part, the low pilus expression observed for M3 isolates is a consequence of the repression of pilus gene expression by the CovR/CovS two-component regulatory system and of an M3-specific mutation in the nra gene, encoding a positive regulator of pilus gene expression. We also discovered that the orthologous transcriptional regulators RofA and Nra, whose encoding genes are located within FCT-2 and FCT-3, respectively, are not functionally identical. Finally, we sequenced the genome of an M3 isolate that had naturally undergone recombinational replacement of the FCT region, changing the FCT and T types of this strain from FCT-3/T3 to FCT-2/T1. Our study furthers the understanding of strain- and type-specific variation in virulence factor production by an important human pathogen.

IMPORTANCE Our ability to characterize how a pathogen infects and causes disease, and consequently our ability to devise approaches to prevent or attenuate such infections, is inhibited by the finding that isolates of a given pathogen often show phenotypic variability, for example, in their ability to adhere to host cells through modulation of cell surface adhesins. Such variability is observed between isolates of group A Streptococcus (GAS), and this study investigates the molecular basis for why some GAS isolates produce pili, cell wall-anchored adhesins, in lower abundance than other isolates do. Given that pili are being considered as potential antigens in formulations of future GAS vaccines, this study may inform vaccine design.

INTRODUCTION

The bacterial pathogen group A Streptococcus (GAS; Streptococcus pyogenes) significantly contributes to human morbidity and mortality worldwide, causing mild infections such as pharyngitis and impetigo, severe invasive infections such as necrotizing fasciitis and puerperal sepsis, and post-GAS infection sequelae such as acute rheumatic fever (1). The great disease burden caused by GAS warrants deconvolution of the mechanisms by which this pathogen circumvents the host immune response and causes disease (2). This is particularly true since, despite decades of efforts, there is currently no licensed GAS vaccine (3–5). Complicating GAS vaccine development are both strain- and type-specific variations in the presence, expression, and sequence of candidate vaccine antigens (6–8).

GAS isolates are divided into M types based upon the sequence of the 5′ end of the emm gene, a gene which encodes the prototypical GAS virulence factor, the M protein (9). To date, more than 220 M types have been characterized, and population-based epidemiological studies have uncovered M-type-specific differences in isolate disease potential; for example, M3 isolates are nonrandomly associated with causing particularly severe invasive infections (10–12), and M28 isolates are associated with cases of puerperal sepsis (13–16). Additional typing systems have also been developed for the analysis of GAS isolates, including T typing, which comprises more than 20 T types (17, 18). T typing is based upon the reactivity of antibodies to the Tee protein, which is the major (backbone) protein of the pilus, a cell surface protein that promotes the ability of GAS to adhere to host cells and tissues and to form biofilms, among other attributes (19–24). The tee gene is located within a region of the GAS genome known as the fibronectin- and collagen-binding T-antigen (FCT) region, of which nine different FCT types have been described based upon heterogeneity in gene content and sequence (18, 19, 25). Most isolates of a given GAS M type share common FCT and T types, such as FCT-2/T1 by M1 isolates and FCT-3/T3 by M3 isolates (Fig. 1A). However, large population-based studies have uncovered exceptions to this rule (6, 26, 27), although these variant strains have not been extensively characterized.

FIG 1.

The abundance of transcripts from pilus biosynthesis genes remains low in M-type M3 GAS isolates relative to M1 isolates regardless of growth temperature. (A) Schematic comparing the FCT-2 and FCT-3 regions. Genes are represented by arrows, with the direction of the arrows mirroring the direction of transcription. The colors of the arrows represent the functions of the encoded proteins (see the key). (B) Western blot analysis of cell surface pilus levels in a series of M-type M3 isolates grown at two temperatures. Cell wall proteins from exponential-phase THY broth cultures, grown at both 25°C and 37°C, were isolated and used with an anti-T3 antibody. Note the characteristic laddering pattern of the pili. Coomassie blue-stained SDS-PAGE gels of separated proteins were used as loading controls. (C) TaqMan-based absolute quantification was performed to enable comparison of the abundances of mRNAs encoding the major (tee) and minor (cpa) pilus proteins and pilus regulatory proteins (rofA or nra) in three M1 and three M3 clinical GAS isolates at both 25°C and 37°C. Each bar represents the mean (±standard deviation) for each strain, with the M1 isolates indicated in pink (25°C) and red (37°C) (in the following order from left to right, respectively: MGAS2221, MGAS9127, and MGAS3350) and the M3 isolates indicated in light (25°C) and dark (37°C) green (in the following order from left to right, respectively: MGAS10870, PD-824, and Di1253). The colored numbers are the mean value for each group of three strains. Asterisks indicate statistical significance (P < 0.0001) as determined by one-way analysis of variance (ANOVA), followed by Tukey’s multiple-comparison test.

GAS pilus gene expression is driven primarily by a positive-regulation transcription factor encoded within the FCT region (28–30). Two isoforms of the positive regulator exist, with some FCT types harboring a gene that encodes the protein RofA (e.g., FCT-2) (Fig. 1A) and other FCT types harboring a gene that encodes the protein Nra (e.g., FCT-3) (Fig. 1A). RofA and Nra share ~62% amino acid identity. Whether RofA and Nra are functionally interchangeable is unclear, since while replacement of nra by rofA in an M-type M53 strain led to enhanced pilus levels, a section of the promoter region was also replaced, and hence altered pilus production may have been due to differential regulation of rofA compared to nra or the alteration of DNA binding site locations, rather than different activities between RofA and Nra (31).

Previously, we determined that relative to M1 (FCT-2) and M49 (FCT-3) GAS isolates, M-type M3 (FCT-3) isolates produce pili in lower abundance (28). In large part, this is a consequence of lower nra expression in M3 isolates than in M49 isolates and also lower nra expression than of rofA expression in M1 isolates. Subsequently, an elegant study by Nakata et al. identified that pilus production by FCT-3 isolates, but not by isolates of other FCT types, is regulated in a thermosensitive manner, with higher production at 25°C than at 37°C (29). While the molecular mechanism controlling thermoregulation of pilus expression was not fully determined, modulation of nra mRNA translation was identified as key (29). Here, we investigate the molecular basis behind the low level of nra expression, and subsequently pilus gene expression, in M3 isolates. We also characterize, via whole-genome sequencing, an M3 isolate that harbors an FCT type and a T type that are atypical for isolates of this M type. Combined, our studies add to the growing literature describing the molecular mechanisms that drive strain heterogeneity in bacterial pathogens.

RESULTS

M3 isolates express pilus biosynthesis genes at a lower level than M1 isolates regardless of growth temperature.

We set out to verify the thermoregulation of pilus production by M3 isolates by comparing seven clinical M3 isolates and two modified derivatives in pilus Western blot analyses, following growth at both 25°C and 37°C. As expected, no or limited pili were detected on the surface of all seven of the tested clinical M3 isolates when grown at 37°C, while higher levels were seen following growth at 25°C for all isolates except MGAS1251 (Fig. 1B; see below for an analysis of the isolate MGAS1251). Expressing Nra from a multicopy plasmid in an M3 isolate background (strain M3 + pNra) resulted in extensive pilus expression regardless of growth temperature (Fig. 1B).

Next, we tested whether the low abundance of mRNAs from the FCT regions of M3 isolates relative to M1 isolates, as previously determined following growth at 37°C (28), is also observed at 25°C. To achieve this, we performed absolute quantification of rofA or nra, cpa, and tee1 or tee3 transcript levels by quantitative reverse transcription-PCR (qRT-PCR). We compared expression levels among three representative M1 and three representative M3 clinical isolates that had been grown at both temperatures. The data show that the low abundance of FCT transcripts in M3 isolates persists regardless of the growth temperature (Fig. 1C). In addition, the modestly higher abundance of transcripts at 25°C than at 37°C in the M3 background is consistent with the finding from Nakata et al. that the thermoregulation occurs primarily at the posttranscriptional level (29) and also with the same nra posttranscriptional thermoregulatory mechanism functioning in M3 isolates.

RofA and Nra are not interchangeable between M1 and M3 backgrounds.

To further investigate the regulation of pilus gene expression, and differences between harboring nra and rofA in particular, we set out to compare whether nra could replace rofA, or vice versa, in the M1 and M3 strain backgrounds, expanding upon data previously gained in an M53 background (31). We began by focusing on the M1 background, where an isogenic rofA mutant derivative, M1ΔrofA, was transformed with empty vector or plasmid constructs producing Nra or RofA. Western blot analysis following GAS growth at both 25°C and 37°C identified that only pRofA was able to restore pilus expression in M1ΔrofA (Fig. 2A), consistent with Nra not being functional in the M1 background. It was also noted that growth temperature had no impact on pilus production (Fig. 2A), consistent with a lack of thermoregulation by FCT-2 isolates (29). In an M3 strain background, we transformed the vector, pNra, and pRofA plasmids into an isogenic nra deletion mutant derivative and examined pilus expression by Western blotting. Plasmid pNra resulted in significant upregulation of pilus expression in the M3Δnra strain background, regardless of GAS culture temperature (Fig. 2B), although expression remained higher at 25°C than at 37°C. Plasmid pRofA had at best a minor enhancing effect on pilus expression in the M3Δnra background at both temperatures (Fig. 2B). Combined, the data are consistent with Nra and RofA not being functionally interchangeable in the tested M1 and M3 strain backgrounds.

FIG 2.

There is M-type specificity to the ability of RofA or Nra to promote pilus expression. (A and B) Western blot analyses of pilus expression in M1 (probed with anti-T1 antibody) (A) or M3 (probed with anti-T3 antibody) (B) GAS isolates. Coomassie blue-stained SDS-PAGE gels of separated proteins were used as loading controls.

Explanations for why Nra and RofA are not interchangeable include possible differences in DNA binding site sequences alongside the FCT-type-specific presence of those sequences. Consistent with this, the intergenic regions between rofA and cpa in the M1 background and between nra and cpa in the M3 background are highly divergent in sequence and length (Fig. 3). While the Nra binding site sequence has not been studied, RofA has three binding sites within the rofA-cpa intergenic region (Fig. 3, orange shading), none of which are conserved in the nra-cpa intergenic region (32). To facilitate testing the importance of the variable intergenic regions on the inability of RofA to induce pilus expression in an M3 GAS derivative, we created strain M3::M1IG, in which the nra-cpa intergenic region was precisely replaced with the equivalent rofA-cpa intergenic region from an M1 strain (Fig. 4A). In this strain background, neither pNra nor pRofA was able to induce pilus expression above the levels seen in the empty-vector-containing strain (Fig. 4B, right), despite these plasmids being able to induce pilus expression in the natural M3 background (Fig. 4B, left). Transcript analysis identified an ~6-fold-higher basal level of nra transcript abundance in M3::M1IG than in the parental M3 isolate, but this did not lead to any change in the basal level of cpa or tee mRNAs (Fig. 4C, green bars). The analysis also identified that RofA increases nra, cpa, and tee mRNA abundances in both of the tested backgrounds (Fig. 4C, blue and purple bars), but an associated increase in pilus protein levels was observed only in the parental strain background (Fig. 4B, M3 + pRofA).

FIG 3.

Comparison of the 430-bp nra-cpa intergenic region in M3 GAS with the 262-bp rofA-cpa region in M1 GAS. Conserved nucleotides are indicated by red asterisks. The nra and rofA start codons are shown in blue, while the cpa start codons are shown in green. Three previously identified RofA binding sites within the M1 intergenic region are highlighted in orange (32). Five previously identified putative CovR binding sites within the M3 intergenic region are highlighted in green (33).

FIG 4.

The inability of RofA to promote high-level pilus expression in M3 GAS is not a consequence of FCT-type-specific differences in nra-cpa or rofA-cpa promoter sequences. (A) Schematic of the rofA-cpa (M1, FCT-2) and nra-cpa (M3, FCT-3) regions in parental M1 isolate, parental M3 isolate, and our constructed chimeric strain M3::M1IG. (B) Western blot analysis of cell surface pilus expression from multiple derivatives of a parental M3 GAS isolate. The strains were grown to mid-exponential phase at 25°C and 37°C prior to the isolation of cell wall proteins. A Coomassie blue-stained SDS-PAGE gel of separated proteins was used as a loading control. (C) TaqMan-based quantitative RT-PCR analysis. Shown are the mean (±standard deviation) abundances of select mRNAs that were determined from triplicate exponential-phase GAS cultures of each strain, grown at 37°C and run in duplicate. Pairwise strain comparisons with the M3 + vector strain were investigated using Tukey’s multiple-comparison test (asterisks indicate significant values, P < 0.05).

The activity of the CovR repressor protein and a nonsynonymous mutation within nra combine to reduce pilus expression in M3 GAS.

We and others have previously identified that CovR, the response regulator component of the two-component regulatory system CovRS (also known as CsrRS), represses pilus production in M3 GAS but not M1 GAS backgrounds (28, 33). In addition, we showed that a nonsynonymous single nucleotide polymorphism (SNP) within nra that results in a D267E change in the protein (relative to most other Nra sequences, such as those seen in M49 isolates) also had a negative impact on pilus expression in M3 GAS (28). Here, we wished to test the cumulative consequences of CovR-mediated repression and the M3-specific nra allele on pilus expression in the M3 background. We did this by comparing M3 GAS derivative strains with covR deleted (M3ΔcovR), with an engineered nra allele that produces Nra lacking the D267E change (M3nra^fix) or containing both genetic alterations (strain M3nra^fixΔcovR). Both Western blot analysis using cell wall protein fractions of GAS strains grown at 25°C and 37°C (Fig. 5A) and qRT-PCR analysis from 37°C-grown cultures (Fig. 5B) identified a cumulative effect of CovR-mediated regulation and the nonsynonymous nra mutation on the inhibition of pilus expression. The temperature-mediated regulation of pilus production was maintained in strains M3ΔcovR and M3nra^fixΔcovR, even though there were significant increases in pilus levels at both temperatures, consistent with CovR not participating in the thermoregulation of pilus expression. Note that similar results were gained with strain MGAS1254 (Fig. 1B), which naturally harbors an inactivating mutation in covS, consistent with the CovR/CovS two-component regulatory system repressing pilus expression but not impacting thermoregulation.

FIG 5.

Low pilus expression by M3 GAS isolates is cumulatively impacted by the CovR-mediated repression of pilus gene expression and by a nonsynonymous SNP within nra. (A) Western blot analysis of pilus production in select M3 GAS isolates. A Coomassie blue-stained SDS-PAGE gel of separated proteins was used as a loading control. (B) TaqMan-based quantitative RT-PCR. Shown are the mean (±standard deviation) abundances, relative to that of the parental M3 isolate, of nra, cpa, and tee mRNAs that were determined from triplicate 37°C exponential-phase GAS cultures of each strain, run in duplicate. Pairwise strain comparisons were investigated by one-way ANOVA, followed by Tukey’s multiple-comparison test (asterisks indicate significant values relative to the parental strain, and hashtags indicate significance between the two covR mutant derivative strains; P < 0.05).

CovR binds with higher affinity to the nra-cpa intergenic region than to the rofA-cpa region.

There are five putative CovR-binding sites within the nra-cpa intergenic region (Fig. 3, green shading) (33), none of which are present in the rofA-cpa intergenic region, thus providing a potential explanation as to why CovR represses pilus expression in the M3 background but not in the M1 background. To assess whether CovR binds to the two respective intergenic regions, we performed an electrophoretic mobility shift assay (EMSA). Purified, native CovR was added to a 262-bp probe that spans the rofA-cpa intergenic region (M1) or to 430- or 429-bp probes that span the nra-cpa intergenic regions from M3 and M49 isolates, respectively. Note that in addition to the 1-bp difference between the M3 and M49 intergenic regions, there are also eight SNPs (see Fig. S1 in the supplemental material). Our data show that CovR fails to bind to the rofA-cpa intergenic region but binds to the nra-cpa intergenic regions (Fig. 6), with no appreciable difference in CovR binding between the M3 and M49 intergenic regions. Thus, the data are consistent with the production of pili being CovR regulated in FCT-3 isolates but not in FCT-2 isolates and that this is due to differences in the presence/absence of CovR-binding sites within the respective nra-cpa and rofA-cpa intergenic regions.

FIG 6.

CovR binds to the FCT-3 nra-cpa intergenic region at a higher affinity than to the FCT-2 rofA-cpa intergenic region. Shown is an EMSA in which PCR-generated probes spanning the rofA-cpa intergenic region (M1; 262 bp) or the nra-cpa intergenic regions (M3 and M49; 430 bp and 429 bp, respectively) were incubated with 0, 0.25, 0.5, and 1 μM recombinant CovR.

An M3 isolate from the 1920s harbors an FCT type similar to that of contemporary M1 isolates rather than that of contemporary M3 isolates.

As part of our investigation into pilus expression by clinical M3 isolates, we identified that surface proteins from MGAS1251 do not react with anti-T3 sera (Fig. 1B). This finding was expected, given that MGAS1251, also known as C203 and ATCC 12384, had previously been shown to react, for unspecified reasons, with anti-T1 sera and not anti-T3 sera as is typical for isolates of this M type (34). This led us to hypothesize that MGAS1251 has undergone recombinational replacement of the FCT region. To investigate this, we first performed Western blot analyses using anti-T1 antibodies to confirm that MGAS1251 produces a T1 pilus, which it does (Fig. 7A and data not shown). Next, we sequenced the MGAS1251 genome and compared it the published genome sequence of the M3 isolate MGAS315 (11). Sequencing reads failed to map, or mapped poorly, to six regions of the MGAS315 genome, indicating that these regions are either absent or variable in the MGAS1251 genome (Fig. 7B). Five of the six regions of variation overlapped with lysogenic bacteriophage within the MGAS315 genome, while the sixth overlapped the FCT region. Further analysis of the MGAS1251 sequencing reads identified that the FCT region of this strain is essentially identical to that seen in M-type M1 strains such as MGAS2221, which also produce the T1 pilus (Fig. 7C). Thus, MGAS1251 has a different FCT region (FCT-2), resulting in the expression of a different T type (T1) than that of most other tested M3 isolates. Note that multilocus sequence typing of MGAS1251 identified it as sequence type 15 (ST15), and of the 75 matches to ST15 within the MLST database (https://pubmlst.org), 74 are M3 strains and 1 is an M1 strain. This supports the idea that MGAS1251 arises from the lateral transfer and recombinational replacement of the FCT region from a donor strain (possibly an M1 isolate) into the M3 genetic background.

FIG 7.

MGAS1251, an M3 GAS isolate from the 1920s, harbors a different FCT type than contemporary M3 isolates. (A) Western blot analysis of cell surface T1 pilus expression for select M3 GAS isolates following growth at 25°C and 37°C. Coomassie blue-stained SDS-PAGE gels of separated proteins were used as loading controls. (B) MGAS1251 was whole-genome sequenced via Illumina-based next-generation sequencing, and an overview of how the MGAS1251 sequences mapped to the genome of the contemporary M3 isolate MGAS315 is shown. Note the gaps in the alignment located at the FCT region, Φ315.1, Φ315.2, Φ315.3, Φ315.4, and Φ315.6. (C) Schematic showing the nucleotide identities, rounded to the nearest whole number, for FCT region genes in MGAS1251 relative to those in the contemporary M1 isolate MGAS2221 and the contemporary M3 isolate MGAS315. Shading between genes is green (100% identity), blue (90 to 99% identity), or pink (<90% identity).

DISCUSSION

GAS isolates vary along strain- and/or type-specific lines in their regulatory pathways, gene content, and disease potential. Examples include the nonrandom association between M3 isolates and severe invasive infections (10–12), that M28 isolates represent one of the few M types that harbor the RD2 pathogenicity island (15, 16, 35, 36), and that isolates recovered from invasive infections often harbor mutations in the genes encoding the CovR/CovS two-component regulatory system (37, 38). Such variation represents a barrier to the characterization of GAS-host interactions. Thus, characterization of the mechanisms underlying phenotypic and regulatory differences between GAS isolates is of importance to both the clinical and basic science fields. Here, we investigated why M-type M3 isolates produce cell surface pili in lower abundance than M-type M1 and M49 isolates, finding that M-type-specific differences in the repression of the pilus biosynthesis operon by CovR, and of variation in the sequence of the Nra transcriptional activator, play contributing roles. The M3-specific adaptations reducing pilus expression likely contribute to the enhanced ability of M3 isolates to cause severe invasive infections, given that pili promote the killing of M3 GAS in human blood (28).

Despite that both RofA and Nra are positive regulators of pilus gene expression, these proteins are not functionally interchangeable. Plasmid pRofA complements the rofA mutation in strain M1ΔrofA but not the nra mutation in strain M3Δnra, while the opposite is true for plasmid pNra (Fig. 2). The data indicate that Nra and RofA have different binding site sequences. Consistent with this, swapping the nra-cpa intergenic region of an M3 isolate for the rofA-cpa intergenic region from an M1 isolate, strain M3::M1IG, abolished the ability of pNra to significantly enhance pilus protein (Fig. 4B) or pilus gene expression (Fig. 4C) levels, unlike what was observed in the parental M3 and M3Δnra strain backgrounds (Fig. 2B and 4B). Interestingly, while pRofA had no enhancing activity on pilus production in strain M3::M1IG, it did have enhancing activity in the parental (M3) strain background (Fig. 4B). This is despite the fact that RofA increased nra transcript abundance at a significantly greater level in strain M3::M1IG (~37-fold) than in the parental strain (~6-fold) and that in both backgrounds, the levels of cpa and tee transcripts increased equally (Fig. 4C). One possible explanation for these findings is that strain M3::M1IG pRofA harbors spurious mutations in one or more pilus assembly genes (e.g., srtC). However, neither Western blot data assaying for the accumulation of pilin subunits (data not shown) nor sequencing data from the FCT region (no mutations were observed in the sequenced region from hsp33 to fbaB) (Fig. 1A) supported this explanation. Rather, we believe that the data support (i) the inability of Nra to bind to the rofA-cpa intergenic region, (ii) the ability of RofA to bind, at least with low affinity, to the nra-cpa intergenic region of M3 isolates and induce nra expression, (iii) that the pilus expression seen in strain M3 + pRofA (Fig. 4B) is a consequence of the RofA-mediated production of Nra, rather than the direct promotion of pilus expression by RofA, as evident by the fact that strain M3Δnra + pRofA does not make appreciable amounts of pili (Fig. 2B), and (iv) that the inability of RofA to promote pilus expression in strain M3::M1IG may be due to some as-yet-uncharacterized regulation at the posttranscriptional level, given that cpa and tee transcripts are present at the same abundance in both M3 + pRofA and M3::M1IG pRofA (Fig. 4C) but pili are detectable only in strain M3 + pRofA (Fig. 4B).

Our previous work identified that despite that M3 and M49 isolates share the same FCT type (FCT-3), M3 isolates produce significantly lower levels of transcripts from the nra and pilus biosynthesis genes than do M49 isolates (28). This difference appears in part to be a consequence of the D267E Nra variant present in M3, but not in tested M49, isolates (Fig. 5). Additional factors in the differential level of pilus gene transcription between M3 and M49 strains could be due to sequence differences in their respective nra-cpa intergenic regions (see Fig. S1 in the supplemental material) and/or to differences in the repressing activity of CovR. Note that CovR differs by two amino acids in the M3 background compared to the M49 background, and the functional consequences of these amino acid changes have not been investigated.

While not common, M-type M3 isolates that T-type as T1 (corresponding to what is seen for most M1 isolates, rather than the standard T3) have been described, with 10 of 2,171 M3 isolates in one study having that designation (26). As these 10 isolates were not further investigated, there is the possibility that these isolates, and also MGAS1251, are not of the M3 background in which there has been recombinational replacement of the FCT region but rather are of M1 backgrounds in which there has been replacement of the emm gene (from emm1 to emm3). Our genetic analysis of MGAS1251 in this study (i.e., via genome sequencing and multilocus sequence typing [MLST] analysis) identified that it has an M3 genetic background, as most genes located outside of the FCT region in MGAS1251 are more closely related to genes from other M3 isolates than to those from M1 isolates (data not shown).

Since MGAS1251 is the oldest M3 isolate in our collection, having been isolated in the 1920s, it raises the possibility that ancestral M3 isolates all harbored FCT-2, with most post-1920s M3 isolates being descendant from an M3 strain that underwent recombinational replacement from FCT-2 to FCT-3. However, this can only be speculated about at this time, given that we have only a single isolate from the 1920s that can be analyzed. An alternative hypothesis is that most ancestral M3 strains harbored FCT-3 and that MGAS1251 underwent recombinational replacement from FCT-3 to FCT-2, possibly during cocolonization of an individual with both M1 and M3 isolates, with the M1 isolate serving as the donor of the FCT-2 region. At face value, the thermoregulation of pilus production is advantageous, for example, by expressing pili in high abundance at anatomic sites where expression could be of most value (e.g., to promote adherence to skin) and reducing it in environments where pili have been shown to be detrimental to GAS survival (e.g., in the case of M3 isolates, in human blood [28]). However, given that the thermoregulation of pilus expression is restricted to FCT-3 strains, this is likely an oversimplification of its importance and of the roles that pili play during infection.

In summary, this work furthers our understanding of how a human pathogen regulates, along strain- and/or type-specific lines, the production of a factor that influences virulence phenotypes such as adherence, biofilm formation, and survival in human blood. As pili are being considered as potential antigens in future GAS vaccine formulations (4, 8, 17), full characterization of the regulatory mechanisms controlling pilus production is warranted.

MATERIALS AND METHODS

Bacterial strains, growth conditions, and plasmids.

Multiple clinical and recombinant GAS strains (see Table S1 in the supplemental material) and plasmids (Table S2) were used in this study. GAS strains were grown in THY broth, with chloramphenicol (4 μg mL⁻¹), kanamycin (200 μg mL⁻¹), or spectinomycin (150 μg mL⁻¹) added when needed. For standard cloning, Escherichia coli DH5α cells were used. E. coli cells were grown in LB broth with agitation at 37°C, with chloramphenicol (20 μg mL⁻¹) added when needed.

Isolation of GAS cell wall protein fractions.

GAS cultures for all strains were grown to mid-exponential phase (optical density at 600 nm [OD₆₀₀] of 0.5) in Todd-Hewitt broth supplemented with yeast extract (THY broth). Forty-five-milliliter aliquots were recovered, and the bacteria were pelleted by centrifugation (5,000 × g for 11 min). The bacterial pellets were washed once in 10 mL of Tris-EDTA (TE) buffer and then resuspended in 450 μL of TE-sucrose buffer (50 mM Tris-HCl, pH 7.5, 550 mM sucrose, 1 mM EDTA, 250 μg mL⁻¹ hyaluronidase, 1 mg mL⁻¹ lysozyme, 250 μg mL⁻¹ mutanolysin). Samples were incubated at 37°C with end-over-end rotation for 2 h and then centrifuged at 15,000 × g for 5 min in order to pellet the protoplasts. The supernatants, containing cell wall proteins, were removed to clean 1.5-mL tubes, and the centrifugation step was repeated. The final supernatants were mixed 1:1 with 2× Laemmli buffer and stored at −20°C until used.

Western blot analysis of GAS cell surface pili.

Cell wall protein fractions were separated on two 7.5% SDS-polyacrylamide gels, one of which was stained with Coomassie blue (Imperial protein stain; Thermo Scientific) for use as a loading control, while the other was used in association with a Trans-Blot semidry transfer cell (Bio-Rad) to transfer the proteins to nitrocellulose membrane. The membrane was then subjected to overnight incubation with anti-T1 or anti-T3 monoclonal antibodies (1:1,000 dilution; TransEurope Chemicals). The next day, the blot was washed thoroughly and incubated with anti-rabbit IgG secondary antibody labeled with Alexa Fluor 680. A Li-Cor Odyssey system was subsequently used to visualize and capture the fluorescent signal.

Isolation of total GAS RNA samples.

Total RNA was isolated from tested GAS strains as previously described (38). Briefly, the strains of interest were grown to the exponential phase (OD₆₀₀ of 0.5) of growth in THY broth. Two volumes of RNAprotect bacterial reagent (Qiagen, Inc.) was added to 1 volume of GAS culture and incubated at room temperature for 5 min. Following centrifugation (5,000 × g for 10 min at 4°C), the supernatant was discarded, the cell pellets were snap-frozen in liquid nitrogen, and the frozen pellets were placed at −80°C until ready for processing. Cells were processed using a mechanical lysis method with lysing matrix B tubes in conjunction with a FastPrep24 homogenizer (MP Biomedicals). RNA was isolated using the miRNeasy kit (Qiagen) with contaminating DNA being removed with three treatments with a TURBO-DNA-free kit (Life Technologies). The quality and quantity of the purified RNA were determined using a Bioanalyzer system (Agilent Tech).

Relative and absolute quantification RT-PCR.

Total RNA samples from tested GAS strains were isolated as described above and converted into cDNA using the reverse transcriptase Superscript III (Invitrogen). The generated cDNA was analyzed via TaqMan-based quantitative RT-PCR analysis using a CFX Connect real-time system (Bio-Rad). TaqMan primers and probes for genes of interest are shown in Table S3. For relative quantification, transcript levels were determined from duplicate exponential-phase GAS cultures of each strain, run in duplicate, using the ΔΔCT method and the internal control gene proS. For absolute quantification, the expression of target transcripts was quantified from duplicate exponential-phase GAS cultures of each strain, run in duplicate, using a standard curve generated using 10-fold dilutions of MGAS2221 (M1) or MGAS10870 (M3) genomic DNA, normalized to the genome size of the respective strains. Statistical analyses were performed using GraphPad Prism 9.

Creation and verification of the RofA-expressing plasmid pRofA.

Plasmid pRofA was created by PCR amplifying the rofA gene from M1 GAS isolate MGAS2221 and cloning it into the shuttle vector pDCBB, which is a BamHI/BglII-digested (and religated) derivative of pDC123 (39). The correct plasmid construct was verified by sequencing the cloned insert. Note that plasmid pNra was constructed previously (28).

Creation and verification of GAS strain M3::M1IG.

“Allelic” replacement was used to replace the nra-cpa intergenic region in the M3 isolate MGAS10870 with the rofA-cpa intergenic region from the M1 isolate MGAS2221. The protocol made use of the suicide vector pBBL740 via standard techniques (40–42). Briefly, overlapping primers were designed to PCR amplify an ~1-kb fragment upstream of the M3 intergenic region, the M1 intergenic region, and an ~1-kb fragment downstream of the M3 intergenic region (Table S3). The PCR products were joined together, along with PCR-amplified pBBL740, via Gibson assembly (New England Biolabs), and the resultant plasmid was sequence verified. The plasmid was transformed into MGAS10870 competent cells, with selection for chloramphenicol resistance on THY agar plates. The passaging and patching protocol, to replace the M3 intergenic region with the M1 version, was performed as described previously (42). The generated strain was confirmed via PCR and targeted sequencing.

EMSAs.

The electrophoretic mobility shift assays (EMSAs) were essentially done as described previously (33). Briefly, the rofA-cpa intergenic region (from the M1 strain MGAS2221) and nra-cpa intergenic regions (from the M3 strain MGAS10870 and the M49 strain 591) were PCR amplified (primers are listed in Table S3). The resultant PCR products were incubated with increasing amounts of phosphorylated CovR (0.25 to 1 μM; overexpressed and purified from E. coli as we previously described [43]) at 37°C for 15 min in 10 mM Tris (pH 7.5), 50 mM KCl, 1 mM dithiothreitol (DTT), 32 mM acetyl phosphate, 0.05% NP-40, 5 mM MgCl₂, 2.5% glycerol, and 5 μg/mL salmon sperm DNA. Samples were then separated by native gel electrophoresis on a 6% Tris-borate-EDTA (TBE) gel, stained with ethidium bromide, washed, and imaged.

Whole-genome sequencing of GAS isolate MGAS1251.

Genomic DNA was isolated from MGAS1251, diluted to a concentration of 0.2 ng/μL, and used to prepare paired-end sequencing libraries according to the Illumina Nextera XT kit. Sequencing libraries were used in conjunction with an Illumina NextSeq 550 instrument (Nevada Genomics Center). Geneious Prime (Biomatters, Ltd.) software was used to map reads to the reference M3 genome of isolate MGAS315 (GenBank accession no. AE014074.1) and also to search for genetic variation (e.g., indels and SNPs) between the test and reference genomes. The generated raw sequence data are available for download from the NCBI Sequence Read Archive under BioProject accession no. PRJNA785991. Note that the American Type Culture Collection (ATCC) also recently sequenced the genome of this strain and has made it available for download from the ATCC website.