Phosphotransferase System Uptake and Metabolism of the β-Glucoside Salicin Impact Group A Streptococcal Bloodstream Survival and Soft Tissue Infection

Streptococcus pyogenes (group A Streptococcus [GAS]), a major human-specific pathogen, relies on efficient nutrient acquisition for successful infection within its host. The phosphotransferase system (PTS) couples the import of carbohydrates with their phosphorylation prior to metabolism and has been linked to GAS pathogenesis. In a screen of an insertional mutant library of all 14 annotated PTS permease (EIIC) genes in MGAS5005, the annotated β-glucoside PTS transporter (bglP) was found to be crucial for GAS growth and survival in human blood and was validated in another M1T1 GAS strain, 5448.

ABSTRACT

Streptococcus pyogenes (group A Streptococcus [GAS]), a major human-specific pathogen, relies on efficient nutrient acquisition for successful infection within its host. The phosphotransferase system (PTS) couples the import of carbohydrates with their phosphorylation prior to metabolism and has been linked to GAS pathogenesis. In a screen of an insertional mutant library of all 14 annotated PTS permease (EIIC) genes in MGAS5005, the annotated β-glucoside PTS transporter (bglP) was found to be crucial for GAS growth and survival in human blood and was validated in another M1T1 GAS strain, 5448. In 5448, bglP was shown to be in an operon with a putative phospho-β-glucosidase (bglB) downstream and a predicted antiterminator (licT) upstream. Using defined nonpolar mutants of the β-glucoside permease (bglP) and β-glucosidase enzyme (bglB) in 5448, we showed that bglB, not bglP, was important for growth in blood. Furthermore, transcription of the licT-blgPB operon was found to be repressed by glucose and induced by the β-glucoside salicin as the sole carbon source. Investigation of the individual bglP and bglB mutants determined that they influence in vitro growth in the β-glucoside salicin; however, only bglP was necessary for growth in other non-β-glucoside PTS sugars, such as fructose and mannose. Additionally, loss of BglP and BglB suggests that they are important for the regulation of virulence-related genes that control biofilm formation, streptolysin S (SLS)-mediated hemolysis, and localized ulcerative lesion progression during subcutaneous infections in mice. Thus, our results indicate that the β-glucoside PTS transports salicin and its metabolism can differentially influence GAS pathophysiology during soft tissue infection.

INTRODUCTION

In order for pathogens to successfully adhere, colonize, and infect, they must adapt and overcome various challenges within the host environment. In particular, being able to obtain carbon sources for growth is essential during early stages of infection. The phosphoenolpyruvate (PEP)-dependent phosphotransferase system (PTS) is a carbohydrate transport system found in most bacteria that couples the translocation of a sugar with its phosphorylation (1–4). It is composed of cytosolic proteins, enzyme I (EI or PtsI), and the histidine-containing phosphocarrier protein (HPr), as well as sugar-specific enzyme II (EII) complexes, which have both cytosolic and membrane subunits (EIIA, EIIB, EIIC, and EIID) (1–5). Losing the function of even one sugar-specific PTS enzyme complex in Gram-positive pathogens has been shown to affect the pathophysiology of these pathogens. For instance, a functional sucrose PTS in Streptococcus pneumoniae was essential for colonization in the nasopharynx (6). In Streptococcus mutans, the EII complex of the mannose PTS was found to play a role in biofilm development, wherein a mutant was impaired for biofilm formation in the presence of glucose (7). Additionally, in the attenuated Streptococcus iniae vaccine strain ISNO, the fructose-specific PTS component was found to be missing, in contrast to its virulent parental strain (8).

Survival and growth in vivo require pathogens to utilize, in the absence of their preferred substrate (e.g., glucose), any available secondary sugar found within their environments. This preferential metabolism of a particular carbohydrate is known as carbon catabolite repression (CCR), a necessary process for bacteria to conserve energy (ATP). Pathogens utilize CCR especially in response to differences in nutrient availability within various host tissues, eliciting a hierarchical preference for carbohydrate utilization and thus linking sugar uptake and metabolism with pathogenesis and virulence (9–15). In several Gram-positive bacteria, the highly conserved catabolite control protein A (CcpA) mediates CCR along with its coeffector, a serine-phosphorylated form of HPr (HPr-ser∼P). The CcpA-HPr complex binds to catabolite responsive elements (cre), regulating the transcription and expression of catabolite-related genes. In the presence of a preferred sugar, this complex inhibits expression of PTS permeases for alternative carbohydrates (9, 10, 16). However, there are CcpA-independent mechanisms of CCR as well (17). In S. pneumoniae, inactivation of ccpA attenuated virulence but did not abolish CCR of most metabolic pathways (18). In S. mutans, mutations affecting HPr-ser∼P inhibited expression of particular PTS permeases but carbohydrate utilization was still prioritized (19). Importantly, pathogens can adapt to carbon limitation through the upregulation of specific PTS operons, often induced by the presence of their substrates (3, 20, 21).

β-Glucosides, which include salicin, arbutin, esculin, and cellobiose, are an interesting group of alternative sugar sources for bacteria. Although commonly found in plants, β-glucosides are rare in mammals (22–24). However, pathogenic bacteria often possess transport systems for β-glucosides that are important for virulence. One theory is that since the mammalian extracellular matrix is rich in β-linked disaccharide units, such as the β-glucosides arbutin and esculin, bacterial pathogens may have adapted to utilize these during infection. In a rabbit endocarditis model of infection, Streptococcus gordonii appears to metabolize β-glucoside sugars from injured heart valves, while Listeria monocytogenes was found to induce β-glucoside uptake genes during in vivo growth (24, 25).

This adaptation to utilize available nutrients in host niches is especially true for Streptococcus pyogenes (group A Streptococcus [GAS]), a strict human pathogen that typically colonizes in the nasopharynx or skin (26, 27). These are environments with limited glucose availability but rich in other carbohydrate sources. GAS, like most bacteria, prefers glucose as the primary carbohydrate, but it is also able to utilize various other carbohydrates (3, 26, 28–30). GAS commonly causes self-limiting infections (e.g., pharyngitis and impetigo) yet is also able to disseminate into the bloodstream to cause more severe, life-threatening diseases (e.g., necrotizing fasciitis and streptococcal toxic shock syndrome) (26, 31–33). This ability of GAS to infect multiple host environments and utilize various nutrients for successful growth and infection has suggested that GAS carbohydrate uptake systems and the subsequent metabolism of these sugars are essential for infection (29, 30, 34–37).

Our lab previously showed that a PTS null mutant (ΔptsI) in M1T1 MGAS5005 exhibited early-onset hemolysis, correlating with deregulation of sagA (streptolysin S [SLS]) and more severe ulcerative lesions during murine subcutaneous infections (38). In a global analysis of all 14 annotated sugar-specific PTS EIIC loci in MGAS5005, we found that the mannose-specific EII had a significant contribution to onset of early hemolysis in GAS, although it did not fully recapitulate a ΔptsI hemolytic profile (29). In this study, we found that a mutant in the putative PTS β-glucoside transporter gene was diminished for growth and survival in whole human blood. Using defined nonpolar mutants of the β-glucoside permease (bglP) and phospho-β-glucosidase enzyme (bglB) in the M1T1 5448 background, we showed that bglB, not bglP, was important for growth in blood. Furthermore, the bglPB genes are in an operon repressed by glucose and induced by the β-glucoside salicin. Investigation of the mutants found a role for redundant sugar transport, with the bglP mutant influencing in vitro growth in multiple sugars. Furthermore, data indicate that losing BglP and BglB affect biofilm formation, hemolysis, and localized ulcerative lesions during soft tissue infection in mice. Thus, uptake and metabolism of the β-glucoside salicin are key for the pathophysiology of group A streptococcal infections.

RESULTS

Survival of GAS in nonimmune human blood requires a functional PTS β-glucoside EII locus.

Previously, we generated a mutant library of all 14 annotated PTS EIIC-encoding genes in M1T1 MGAS5005 using a polar insertional inactivation strategy to characterize the sugar-specific PTS transporters (29) and investigate their role in hemolytic phenotypes observed in a ΔptsI (PTS EI null) mutant GAS (38). In this study, the EIIC mutant library was subjected to a Lancefield bactericidal assay using nonimmune whole human blood (Fig. 1A). Wild-type (WT) M1T1 MGAS5005 and the EII mutants were inoculated at mid-exponential phase into fresh heparinized human blood and incubated with rotation for 3 h at 37°C. Only 2 of the 14 tested PTS EII mutants (β-glucoside ΔbglP and fructose ΔfruA) exhibited defects in blood comparable to complete loss of PTS in the ΔptsI mutant (Fig. 1A) (38). Our group previously demonstrated that loss of the fructose EII permease (ΔfruA) conferred decreased survival in human blood and neutrophils (36). However, this was a novel finding for the annotated β-glucoside EII transporter bglP (Fig. 1A). GAS sugar utilization and subsequent virulence phenotypes appear to have strain-specific differences between the clinically relevant strains M1T1 MGAS5005 (covS) and M1T1 5448 (29). As we have performed multiple GAS genomic studies with M1T1 5448 (35, 39, 40), we chose to focus our study on bglP in the M1T1 5448 background.

FIG 1.

The PTS β-glucoside EII is necessary for GAS survival in whole human blood. Lancefield bactericidal assay was performed on the EIIC polar mutant library (A). Data are the averages of three biological replicates. Red bars indicate growth of mutants significantly below that of M1T1 MGAS5005 (P < 0.0168). Gray bars indicate those EIIs that survived similarly to the WT or better. (B) The β-glucoside ΔbglP and ΔbglB nonpolar deletion mutants as well as their corresponding rescue strains were subjected to a Lancefield bactericidal assay in comparison to parental strain M1T1 GAS 5448. Unpaired Student’s t test was used to evaluate the significance of differences between groups. Error bars indicate standard deviations. **, P = 0.0012; ****, P < 0.0001. n.s., not significant.

The β-glucoside EII transporter bglP is located immediately upstream of a putative phospho-β-glucosidase gene (bglB) in GAS M1T1 genomes in a potential operon with an upstream putative regulator, licT (Fig. 2A). Since the bglP mutant in the MGAS5005 EII mutant library was insertional, polar effects on the transcription and expression of the downstream gene bglB are plausible. To interrogate the role of both genes in GAS survival during growth in human blood, we generated nonpolar deletion mutants of bglP and bglB in 5448 via allelic exchange with an in-frame promoterless aphA3 kanamycin resistance cassette (Table 1). Strains 5448.ΔbglP and 5448.ΔbglB were generated (Table 1), as well as their corresponding rescue strains as controls (Table 1). A Lancefield bactericidal assay of each individual deletion showed that only the ΔbglB strain had significantly decreased growth compared to WT 5448, whereas the ΔbglP strain actually exhibited approximately 8-fold-increased survival over that of the WT (Fig. 1B). The rescue strains, on the other hand, grew comparably to WT levels, although the bglP rescue appeared to only partially complement, as the rescue strain’s growth was still significantly different from that of the WT (Fig. 1B). If the genes are indeed in an operon, the polar Δβ-glucoside (bglP) mutant in MGAS5005 (Fig. 1A) was likely a double ΔbglPB mutant with its defect in whole human blood linked to loss of the bglB β-glucosidase gene.

FIG 2.

The β-glucoside locus in M1T1 GAS is an operon. (A) The genetic architecture of licT (green), bglP (blue), and bglB (orange) is shown. Solid, bold lines indicate cotranscription between genes as determined by reverse transcription-PCR (RT-PCR). The sequence in the region indicated in the gray box is magnified in panel B to show the end of the mRNA as detected by 5′ RACE and the −10 and −35 signal sequences subsequently identified. Location of the promoter (thin arrow) and the rho-independent terminator (lollipop) are indicated. The promoter region (P_licT) and flanking sequences (green bar) were fused to firefly luciferase (PlicT-luc) to assess promoter activity in GAS 5448 using luciferase assays, expressed as relative light units (RLU) in rich THY versus low-glucose C-medium (C) and in low-nutrient chemically defined media (CDM) supplemented with glucose versus a β-glucoside sugar, salicin, during early log, late log, and stationary growth (D). Unpaired Student’s t test was used to evaluate the significance of differences between groups. Error bars indicate standard deviations. **, P = 0.0053; ***, P < 0.0005; ****, P < 0.0001.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Relevant genotype description	Reference
E. coli
DH5α	hsdR17 recA1 gyrA endA1 relA1	74
S. pyogenes
5005	M1T1; covS	75
5448	M1T1	76
5448AP	M1T1; 5448 isogenic covS mutant	77
Δβ-glucosideEII	5005 with polar insertional inactivation of bglP; Spr	29
5448.ΔbglP	5448 with non-polar allelic exchange deletion of bglP, Kmr	This study
5448.ΔbglB	5448 with non-polar allelic exchange deletion of bglB, Kmr	This study
5448.bglPR	Rescue strain of 5448.ΔbglP; Kms Sps	This study
5448.bglBR	Rescue strain of 5448.ΔbglB; Kms Sps	This study
5448 (pKSM720)	5448 with replicating pKSM720 plasmid; Spr	This study
5448 (pKSM210)	5448 with replicating pKSM210 plasmid; Spr	This study
5448 (PlicT-luc)	5448 with replicating pPlicT-luc plasmid; Spr	This study
Plasmids
pCRS	Temperature-sensitive conditional vector; Spr	35
pAX0475K	pCRS-based plasmid for allelic exchange of the bglP ORF; Kmr Spr	This study
pAX0476K	pCRS-based plasmid for allelic exchange of the bglB ORF; Kmr Spr	This study
pKSM720	Replicating plasmid with promoterless firefly luciferase (luc) gene; Spr	13
pKSM210	Pemm driving the expression of luc in pKSM720 background; Spr	78
pPlicT-luc	pKSM720-based plasmid with PlicT driving the expression of luc; Spr	This study

The bglP EII transporter gene is in a catabolite-repressed and β-glucoside-induced operon.

In the M1T1 MGAS5005 and 5448 GAS genomes, bglP is found closely flanked on either side by licT and bglB (Fig. 2A). To assess whether licT and bglPB were cotranscribed in an operon, total RNA was isolated from WT 5448 cells grown to late exponential phase in THY-rich medium at 37°C and converted to cDNA (see Materials and Methods) (41). RT-PCR of the resulting cDNA using locus-specific primer pairs (Table 2) found transcriptional linkage between licT and bglP, bglP and bglB, and licT and bglB (Fig. 2A; see also Fig. S1 in the supplemental material), suggesting that licT-bglPB is expressed as a single transcript under the growth conditions used here. To define the promoter region of the operon, 5′ rapid amplification of cDNA ends (RACE) was performed on the cDNA isolated as described above using primers complementary to licT (Table 2) in order to identify the 5′ end of the licT-bglPB mRNA. A transcriptional start site (TSS) that possessed a weak −10 site (TGATAA) and a canonical −35 site (TTGACA) (Fig. 2B) was found 172 bp upstream of the licT start codon.

TABLE 2.

PCR primers

Purpose and target	Primer name	Sequencea (5′–3′)	Reference
For allelic exchange
Promoterless aphA3	oKmF	ATGGCTAAAATGAGAATATCACC	35
	oKmR	CTAAAACAATTCATCCAGTAAAATA	35
pCRS	M13F	GTTTTCCCAGTCACGACGTTGTA	35
	M13R	CAGGAAACAGCTATGACC	35
Within aphA3	oKmV1	TAGCAGGAGACATTCCTT CC	35
	oKmV2	TGGGGATCAAGCCTGATTGG	35
5′ of bglP	oAX0475.1	cccggatccTCAGAGAGAATTATCAATCTAGG	This study
	oAX0475.2	GGTGATATTCTCATTTTAGCCATCACGCCAATCACATTTGAAGAAATCGC	This study
3′ of bglP	oAX0475.3	TATTTTACTGGATGAATTGTTTTAGATTATTCAAGATGACAAAAAAATGGTGAC	This study
	oAX0475.4	cccggatccACGAGGATCAGAGGTCATTGGG	This study
5′ of bglP-aphA3 junction	oAX0475V1	AAAGTTGGAGATAGTATTGAGC	This study
3′ of bglP-aphA3 junction	oAX0475V2	TAGAAAATCAACCGTATGTTCC	This study
5′ of bglB	oAX0476.1	cccggatccCTGGCAGGTATCACGGAACC	This study
	oAX0476.2	GGTGATATTCTCATTTTAGCCATTCTCCCTCTCCTTTTAGGTTATAAAATGAC	This study
3′ of bglB	oAX0476.3	TATTTTACTGGATGAATTGTTTTAGATAGGTCTTGTTTAAGATGGTGACTC	This study
	oAX0476.4	cccggatccTCCATGATTTTATCTTTCATTTTAGC	This study
5′ of bglB-aphA3 junction	oAX0476V1	AGCAATGTTGGCTTCTAATTTGGC	This study
3′ of bglB-aphA3 junction	oAX0476V2	TATCCGTTCTTATGATTGCCG	This study
For cDNA and RT-PCR
5′ of licT ORF	licT-RevTrans-F	AAACCATAATGCGGCGATTTC	This study
3′ of licT ORF	licT-RevTrans-R	TAAGGTGGTAGTTGTACCGC	This study
5′ of bglP ORF	bglP-RevTrans-F	CGCTTTAAGGTGACACCTGT	This study
3′ of bglP ORF	bglP-RevTrans-R	CCAACCAGACCACCTGAAAT	This study
5′ of bglB ORF	bglB-RevTrans-F	CGCAAAAGGTTTGTCCGTAC	This study
3′ of bglB ORF	bglB-RevTrans-R	CCCTGACCAGATGTGTAAGC	This study
Before 3′ end of bglB	bglB-cDNA-R	AAATTATCCTCAAGGTAA	This study
For 5′ RACE
licT ORF	licT-GSP1	TTTAATATGGGTCATAAAG	This study
Upstream of licT-GSP1	licT-GSP2	GTCAGGCGATGTGGCTCCTT	This study
Upstream of licT-GSP2	licT-GSP3	CCCACCGTAGTGGGTTATGAATCAACA	This study
For PlicT-luc fusion
PlicT upstream	PlicT-luc-F	ccaggcct TTTAAGGGTAACTGATAGAC	This study
PlicT downstream	PlicT-luc-R	cccctcgag GTATCACAATAAAAAATGAC	This study

^aBold type denotes restriction sites. Underlined segments indicate overlap with aphA3 for PCR-SO Eing. Lowercase letters indicate nucleotides not complementary to target DNA.

To account for potential regulatory elements, the promoter region and flanking sequences (Fig. 2A, green box; Table 2) were then fused to firefly luciferase (luc) and transformed into GAS 5448 (Table 1). To characterize the induction and expression of the promoter, 5448 (P_licT-luc) was subjected to luciferase assays under different growth conditions (Fig. 2C and D). The promoterless luc vector pKSM720 was used as a negative control to account for background luminescence activity and pKSM210 with the strong P_emm promoter driving the luciferase expression (P_emm-luc) as a positive control (Table 1; Fig. S2). Strains were grown to early-logarithmic, late-logarithmic, and stationary growth in either rich, high-glucose medium (THY) and rich, low-glucose C-medium (Fig. 2C; Fig. S2A) or nutrient-limiting chemically defined medium (CDM) with a sole carbon source of either glucose or the β-glucoside sugar salicin (Fig. 2D; Fig. S2B). Samples were collected at each point in growth and normalized to predetermined final cell units (see Materials and Methods), and promoter activity was assessed based on luciferase activity. Results are graphed as relative light units (RLU) divided by the cell units. In comparison to the strong P_emm (Fig. S2) that showed RLU expression in the thousands, P_licT exhibited much lower induction of luc over background under all conditions tested (Fig. 2C and D), suggesting a tight regulation of the licT-bglPB operon. The expression of P_licT was significantly increased in low-glucose (0.05%) C-medium (Fig. 2C) compared to that in the high-glucose (0.5%) THY. Since the operon was repressed in the presence of high glucose and induced in low glucose, this indicates that it is under carbon catabolite repression (CCR). In addition, P_licT expression was significantly increased during growth in CDM supplemented with salicin as the sole carbon source compared to CDM with glucose as the sole sugar (Fig. 2D). Thus, P_licT is not only induced under conditions of low glucose but is also upregulated in the presence of salicin, a likely β-glucoside substrate.

Loss of bglP and bglB impacts in vitro growth of M1T1 5448 in multiple PTS sugars as the sole carbon source.

Given the catabolite repression of P_licT in THY and its induction in the presence of a likely substrate during growth in CDM supplemented with salicin, ΔbglP and ΔbglB mutants were monitored over 24 h for growth in these different media (Fig. 3). In THY, the ΔbglP mutant had a significant growth rate defect compared to WT 5448, and it also grew to a lower final yield (Fig. 3A). This defect was complemented back to WT levels with the bglP rescue (Fig. S3). On the other hand, although the ΔbglB mutant was not significantly different from the WT, it grew at a slightly higher rate and reached a comparable final yield (Fig. 3A). Additionally, in CDM supplemented with glucose (0.25%) as the sole carbon source, both the ΔbglP and ΔbglB mutants grew to slightly lower yields than the WT, yet neither mutant showed any major differences (Fig. 3B). This may be due to the fact that the operon was not induced in this high-glucose environment (Fig. 2D), suggesting that bglP and bglB are not necessary during growth under this condition. However, in CDM with salicin (1%) as the sole carbon source, losing either bglP or bglB was very detrimental to GAS growth (Fig. 3C), while their corresponding rescue strains grew similarly to the WT (Fig. 3D). These data suggest that both the bglP and bglB genes are essential for utilizing the β-glucoside salicin as a carbon source. Since bglP is annotated as a β-glucoside transporter, we tested additional β-glucoside sugars for growth of WT 5448 and the individual bglPB mutants. Interestingly, WT 5448 was not able to grow in CDM supplemented with either arbutin (1%) or esculin (1%) as the sole carbon source (Fig. S4), which is consistent with previous reports that GAS is unable to utilize and produce acids from arbutin or esculin (42, 43). We also were not able to grow WT 5448 in cellobiose (1%) as the sole carbon source (Fig. S4) as previously seen for MGAS5005 (29), despite the presence of 2 additional annotated PTS loci for cellobiose uptake in the GAS genome.

FIG 3.

Loss of the β-glucoside transporter confers in vitro growth defects. WT 5448 (black) and ΔbglP (blue) and ΔbglB (orange) deletion mutants were grown in rich THY (A) and CDM supplemented with either glucose (B) or salicin (C) as the sole carbohydrate source. (D) Rescue strains for bglP (blue, dashed) and bglB (orange, dashed) were also compared to WT 5448 (black) in CDM with salicin. Unpaired Student’s t test was used to evaluate the significance of differences between time points and groups. *, P < 0.04; **, P < 0.0058. Each point graphed corresponds to the mean of at least three biological replicates, with error bars indicating standard deviations.

Previously, the MGAS5005 Δβ-glucoside EII polar bglP mutant was found to have growth defects in mannose and fructose as well (29). To determine if this is also the case for individual deletion mutants of 5448, the ΔbglP and ΔbglB mutants were monitored for growth in CDM supplemented with either fructose (1%) (Fig. 4A) or mannose (1%) (Fig. 4B). Only the ΔbglP mutant had significant growth defects in both media compared to WT 5448 (Fig. 4), while the bglP rescue strain grew at rates comparable to that of the WT (Fig. S3). Although the ΔbglB mutant appeared to grow at a slightly higher rate than WT in both fructose (Fig. 4A) and mannose (Fig. 4B), and reached slightly lower yields, this difference was overall not statistically significant. These data suggest that the β-glucoside transporter may interact with other sugars or indirectly regulate their utilization.

FIG 4.

The β-glucoside permease affects GAS growth in non-β-glucoside sugars. ΔbglP (blue) and ΔbglB (orange) mutants were compared to WT 5448 (black) during growth in CDM supplemented with non-β-glucoside sugars fructose (A) and mannose (B) as the sole carbohydrate source. Unpaired Student’s t test was used to evaluate the significance of differences between time points and groups. *, P < 0.045. Each point graphed corresponds to the mean of at least three biological replicates, with error bars indicating standard deviations.

Mutations in bglP and bglB impact biofilm formation and β-hemolysis profiles in GAS.

Several studies have connected PTS sugar uptake and metabolism with adhesion and biofilm formation, given that most pathogenic colonization and subsequent infections in humans occur under conditions with limited to no glucose present, such as in the nasopharynx (7, 24, 44). To determine the effect of the loss of bglP and bglB on GAS abiotic biofilm formation, an in vitro biofilm assay was performed following established protocols for GAS (45, 46). Strains were grown static in brain heart infusion (BHI) at 37°C for 48 h in 24-well tissue culture plates, and crystal violet staining was used to visualize the biofilm formation (Fig. 5A); optical density at 600 nm (OD₆₀₀) was measured to quantify the biomass (Fig. 5B). M1T1 5448 showed poor in vitro biofilm production compared to an isogenic M1T1 covS strain 5448AP used as a positive control (Fig. 5). Interestingly, qualitative images of the plates showed darker crystal violet staining for both the 5448 ΔbglP and ΔbglB mutants similar to that for 5448AP (Fig. 5A), with their corresponding rescue strains having lighter staining comparable to that of WT 5448 (Fig. 5A). Quantification of each strain’s biofilm formation confirmed these findings (Fig. 5B). Of note, ΔbglP produced biofilm to the same level as 5448AP, yet the bglP rescue strain produced biofilm quantity intermediate to those of the WT and 5448AP, indicating only partial rescue of the phenotype (Fig. 5B). The ΔbglB mutant also produced more biofilm than the WT, although its biomass was intermediate between the WT and 5448AP levels. Genetic rescue of the bglB mutant exhibited a biofilm level comparable to that of WT 5448, indicating full complementation (Fig. 5B). Additionally, all tested strains were subjected to the same biofilm assay in THY medium and exhibited levels of biofilm formation similar to those in BHI (Fig. S5). Therefore, a normally poor biofilm-forming strain is induced to produce much more biofilm as a consequence of the loss of a functional β-glucoside transport and metabolism.

FIG 5.

Mutations in ΔbglP and ΔbglB induce increased in vitro biofilm formation. Biofilms were assessed using the crystal violet assay after static in vitro growth in BHI medium for 48 h. Results are shown as qualitative images of plates (A) or quantification by absorbance at 600 nm (B). Strains analyzed are WT 5448, 5448AP (covS), 5448 ΔbglP, rescued 5448 bglP, 5448 ΔbglB, and rescued 5448 bglB. Data are representative of at least 3 biological replicates for each strain. Unpaired Student’s t test was used to evaluate the significance of differences between groups. Error bars indicate standard deviations. ****, P < 0.0001.

Our lab previously found that a PTS-defective mutant of MGAS5005 (ΔptsI) produced increased streptolysin S (SLS)-mediated hemolytic activity earlier in growth compared to its parental strain (38). However, when the EII polar mutant library was assayed for early hemolytic activity, no single PTS transporter mutant completely recapitulated the phenotype of ΔptsI (ΔEI) (29). To determine the role of the β-glucoside locus in hemolytic activity, the ΔbglP and ΔbglB mutants, along with their corresponding rescues and WT 5448, were assayed for SLS-mediated hemolytic activity across growth using established protocols, as previously described (29, 38). Cell-free supernatants were collected from each strain throughout growth in THY for 8 h and then coincubated with sheep red blood cells (RBCs) at 37°C for 1 h. Hemolysis was assessed by measuring hemoglobin release (OD₅₄₁) of the supernatants after centrifugation, with data presented as growth (OD₆₀₀) on the x axis and hemolysis (OD₅₄₁) on the y axis. Because WT GAS produces high levels of secreted SLS activity via upregulating sagA only during transition from late log to stationary phase (OD₆₀₀ ≥ 1), detectable hemolysis before this transition was considered “early onset” (Fig. 6, dotted line). Both ΔbglP (Fig. 6B) and ΔbglB (Fig. 6C) exhibited early onset hemolysis compared to the WT (Fig. 6A), with ΔbglB showing the most dramatic profile. In fact, the ΔbglB mutant recapitulated the PTS-null (ΔptsI) hemolysis profile observed in our earlier work (38), exhibiting extremely early hemolysis throughout its growth, starting in early to mid-log phase (Fig. 6B). Rescue strains (dashed bars) for both ΔbglP (Fig. 6D) and ΔbglB (Fig. 6E) showed hemolysis profiles comparable to WT. The increase in biofilm formation and early onset of hemolysis throughout growth in the absence of the β-glucoside uptake and metabolism genes bglPB could indicate a deregulation of pathways controlling virulence genes such as the sag (SLS) operon.

FIG 6.

Loss of a functional bglP or bglB gene induces early hemolytic profiles. Hemolytic activity of culture lysates against sheep red blood cells was measured at different points in growth. Data are from three biological replicates comparing onset of SLS-mediated hemolysis (OD₅₄₁ > 0.2) between WT 5448 (A) and the ΔbglP (B) and ΔbglB (C) mutants as well as their corresponding rescue strains (D and E). Hemolysis onset only at or after the dotted line (OD₆₀₀ ≥ 1) was considered similar to the hemolysis pattern of WT 5448, while hemolysis before was considered early.

Loss of bglP and bglB causes more severe localized soft tissue infections.

To determine if the altered hemolytic profiles or biofilm phenotypes of the ΔbglP and ΔbglB mutants affect GAS virulence, monoculture infections with WT 5448 and both of the Δbgl mutants were performed using a murine model of soft tissue infection. Six- to 7-week-old CD-1 outbred mice were infected with 2 × 10⁸ CFU of each strain and then monitored for lesion size and severity (24 h and 48 h) as well as morbidity (7 days). Lesion sizes at 24 h trended larger in both ΔbglP and ΔbglB mutant-infected mice than for the WT, although only the ΔbglB lesions reached statistical significance (Fig. 7A). Severity of the lesions was quantified using darker red pixel intensity as an indicator of increased necrosis as previously described (29). At 24 h, mice infected with the ΔbglB mutant already had both significantly larger and more severe lesions than mice infected with the WT (Fig. 7B; Fig. S6). By 48 h, the ΔbglP mutant yielded significantly larger lesions (Fig. 7C; Fig. S6), yet these were still not more severe than WT lesions (Fig. 7D; Fig. S6). On the other hand, ΔbglB lesions remained both larger (Fig. 7C; Fig. S6) and more severe (Fig. 7D; Fig. S6) than the WT at 48 h, correlating with the early onset of SLS-mediated hemolysis observed as described above (Fig. 6B). Interestingly, neither the ΔbglP nor the ΔbglB mutant was significantly different in systemic lethality from the WT (Fig. 7E). These findings indicate that loss of a functional BglPB and subsequent metabolism of its substrate affects the severity of a localized infection but does not lead to increased dissemination or systemic lethality.

FIG 7.

Deletion of bglP and bglB contributes to localized tissue damage during murine soft tissue infection. Sizes (A and C) and severity (B and D) of lesions infected from WT 5448 and the ΔbglP and ΔbglB mutants are shown. Each were measured at 24 h (A and B) and 48 h (C and D) postinfection. (B and D) Red pixel intensity was used to determine lesion severity and graphed with an inverted y axis whereby lesions with higher severity (darker pixels) show lower red intensity. Unpaired Student’s t test was used to evaluate the significance of differences between groups. *, P < 0.05; **, P < 0.004; ***, P < 0.0009. Error bars indicate standard deviations. (E) Lethality of mice was monitored over 7 days. Significance was determined using Kaplan-Meier survival analysis and log-rank test.

DISCUSSION

Survival in host environments is imperative for colonization and subsequent infection. Consequently, pathogens have adapted to utilize secondary carbohydrates such as the β-glucoside salicin in the absence of their preferred sugar, glucose. The M1T1 GAS β-glucoside transporter is part of a 3-gene operon composed of licT (encoding a putative regulator), bglP (encoding a putative permease), and bglB (encoding a putative phospho-β-glucosidase) (Fig. 2A; Fig. S1). This study found that the licT-bglPB operon is under catabolite repression by glucose (Fig. 2C) and is induced by a β-glucoside substrate, salicin (Fig. 2D). Additionally, loss of bglP and bglB resulted in GAS growth defects in multiple sugars (Fig. 3 and 4), as well as a possible deregulation of virulence-related genes involved in blood dissemination (Fig. 1), biofilm formation (Fig. 5), SLS-mediated hemolysis (Fig. 6), and localized formation of ulcerative lesions (Fig. 7; Fig. S6).

The importance of the β-glucoside operon for GAS pathophysiology.

Our results indicate, for the first time, that a secondary β-glucoside carbohydrate is important for GAS growth and infection. The importance of this class of sugar for overall GAS pathophysiology is highlighted by the fact that there are multiple annotated β-glucoside transporters within the GAS genome (β-glucoside EII, cellobiose EII [1], and cellobiose EII [2]) even though primarily plant-derived β-glucoside compounds are rare in mammals (22, 24). One theory for β-glucoside-specific genes in human bacterial pathogens suggests that β-glucosides can be derived from human tissues, such as the extracellular matrix. Many bacteria, including streptococci, can catalyze the degradation of cellobiose. Degradation of glycosaminoglycans (GAGs) within host extracellular matrices produces structural analogues of cellobiose and other β-linked disaccharides, which are then transported by bacteria through the PTS (47–49). Bacterial phospho-β-glucosidases or phospho-β-galactosidases, such as that encoded by bglB, can then hydrolyze these phosphorylated disaccharides, resulting in glucose and one other component, which can be further metabolized in the glycolytic pathway (50, 51). As a β-linked disaccharide, salicin can be transported by the permease BglP and subsequently hydrolyzed by the phospho-β-glucosidase BglB into glucose and salicyl alcohol, specifically cleaving the β-linked glycoside bond between its glucose residues. Thus, pathogens can use β-glucosides to their advantage and utilize these sugars after tissue damage during the course of an infection (24, 52).

The β-glucoside permease BglP and its effects on multiple sugars.

Loss of the β-glucoside permease BglP prohibited growth in multiple sugars in addition to salicin (Fig. 3 and 4). Previously, we showed that GAS PTS permeases can be promiscuous for multiple sugars and affect growth in those sugars even when they are not the substrate predicted for that transporter by annotation (29). This may not be uncommon, as there have been reports of sugars that may access the same PTS permeases, such as glucose and mannose in S. mutans (53). Moreover, a recent study suggests that specific protein-protein interactions of PTS permeases in GAS help regulate the transport of carbohydrates, as well as lead to the hierarchy of preferred carbohydrates. The E. coli fructose PTS (FruA/FruB) was observed to either enhance or inhibit transport activities of other PTS permeases, including the mannose PTS, depending on whether fruA and fruB were overexpressed, deleted, or induced (54). In S. mutans, activity of the glucose PTS has been reported to be mediated by the mannose and cellobiose PTSs (55, 56). Our findings suggest a potential interaction between the β-glucoside PTS permease and the mannose and fructose PTS permeases, as deletion of the BglP permease negatively impacted successful growth in mannose and fructose (Fig. 4). Additionally, previous work has also shown that the mannose PTS impacts overall GAS growth in multiple carbohydrates (29, 38). Together, these findings may indicate that a functional β-glucoside permease is important for GAS pathogenesis not only for transporting a β-glucoside sugar but also indirectly through regulation of the transport and metabolism of other, more abundant PTS sugars. This potential protein-protein interaction between PTS permeases may also explain why our previous study on GAS fitness during in vitro growth in THY indicated that bglP was important within this environment (39). Loss of bglP resulted in a significant growth rate defect during growth in rich THY (Fig. 3A), despite the licT-bglPB operon likely being under catabolite repression within this environment (Fig. 2C). Although BglP should not be needed during growth in THY, the availability of other carbohydrates in this rich medium may require a functional BglP for regulating the hierarchy and function of other PTS permeases.

Although the β-glucoside locus appears to be important for GAS growth and infection, WT 5448 grew only in media supplemented with salicin as the sole carbon source and not in other β-glucosides, such as arbutin, esculin, and cellobiose (Fig. 3C and D; Fig. S4). We previously showed that another M1T1 GAS, MGAS5005, was unable to grow in cellobiose (29). Thus, only salicin is indicated as a potential substrate for BglP in M1T1 GAS. In fact, most GAS strains may be unable to efficiently utilize these specific β-glucosides, which is supported by findings that other GAS serotypes are unable to hydrolyze arbutin and, very rarely, esculin (42, 43). As a strict human pathogen, it is possible that GAS has evolved to utilize only specific β-glucosides prevalent within its host niches. In Enterococcus faecalis, the PTS genBA operon, previously annotated as a cellobiose uptake operon, was specifically found to be involved in the utilization of β-1,6-linked (e.g., gentiobiose) but not β-1,4-linked (e.g., cellobiose) glucosides (57). Niche-specific nutrient utilization is also seen with the MalE protein: in Escherichia coli it transports maltose and maltodextrin, whereas GAS uses MalE for transport of maltotriose and longer maltodextrins found in saliva (58, 59). As our growth assays were done with β-glucoside sugars in vitro, we are currently investigating if these β-glucoside sugars are transported by GAS in vivo.

Contributions of BglP and BglB to GAS pathophysiology.

Despite being cotranscribed in an operon (Fig. 2; Fig. S1), it is evident that bglP and bglB have roles that differentially impact GAS growth and infection. This may reflect their various roles during utilization of the β-glucoside salicin, as described above. The polar insertional mutant of bglP in MGAS5005 likely also inactivates the downstream bglB (ΔbglPB), as supported by different growth phenotypes between the polar mutant (Δβ-glucoside EII [Fig. 1A]) and the nonpolar ΔbglP and ΔbglB mutants (Fig. 1B) in whole human blood. Thus, the polar ΔbglPB mutant conferred decreased survival in human blood due to loss of bglB, where the ability to utilize the β-glucoside would be inhibited due to the absence of the enzyme needed to cleave it into a usable monosaccharide. Within human tissues, β-glucoside analogues are found in environments such as the respiratory tracts and colon, and their utilization requires a β-glucosidase enzyme (24, 50). For instance, more than half of Firmicutes found as human gut microbiota actively express β-glucosidase activity (50). In S. pneumoniae, deletion of a predicted β-glucosidase similarly impaired the mutant’s dissemination in blood (60). The importance of bglB in human blood may indicate the availability of β-glucosides from GAGs found within this environment (61). Additionally, BglB may also process closely related structures of other disaccharides, not just salicin.

The loss of bglP resulted in increased growth and survival in blood (Fig. 1B); however, this did not translate into increased bloodstream dissemination during infection (Fig. 7D). The BglP permease may affect cell surface morphology, possibly preventing effective phagocytosis during growth in whole human blood. Since GAS is a strict human pathogen whose virulence factors have evolved to specifically recognize human cells and proteins (62, 63), the phenotype discrepancy between human blood and the murine model may be a result of host specificity. Our recent study involving the ATPase of a putative ABC transporter, ScfE, also showed attenuated phenotypes in human blood that were not recapitulated in murine models (41). Additionally, GAS would need to circumvent epithelial barriers during murine subcutaneous infection in order to reach the bloodstream. Although the ΔbglP mutant had increased growth and survival in human blood, the mutant strain may not be able to easily disseminate in the murine model, impeding its ability to reach the bloodstream.

Regulatory pathways involving the β-glucoside operon.

In most cases, bacterial genes involved in the transport and metabolism of secondary sugars are expressed only when two conditions are met: (i) a preferred carbohydrate substrate is absent and (ii) the corresponding secondary sugar is present within the environment. Based on our promoter-luciferase fusions, the licT-bglPB operon is carbon catabolite repressed in the presence of glucose or a preferred sugar and also induced by a potential substrate, the β-glucoside salicin (Fig. 2C and D). These findings are supported by previous transcriptional studies that showed that the licT-bglPB operon is induced in low-glucose C-medium compared to THY (64) and is repressed by CcpA (13, 34). Additionally, the operon was induced in murine vaginal mucosa, which may indicate that its substrate is present within this environment (65). It is striking that luciferase expression driven by P_licT (Fig. 2C and D) is much lower than the (high) levels of luc expression driven by the constitutive P_emm (Fig. S2), suggesting that licT-bglPB is a highly regulated operon. In Bacillus subtilis, bglP and licT are known to have a feedback regulation of each other (66), where licT is an antiterminator that binds to RNA polymerase to avoid dissociation from the DNA template, allowing continuous transcription of the operon (66). In GAS, there is a predicted stem-loop structure associated with RNA polymerase termination in the 3′ end of licT prior to its stop codon (Arnold software [http://rssf.i2bc.paris-saclay.fr/toolbox/arnold/]). Although the role of licT was not explored in this study, we can predict based on homology that it is a regulator of the licT-bglPB operon by preventing early termination of transcription for bglP and bglB. Ongoing studies seek to determine the specific role of LicT in GAS regulation and pathogenesis.

Mutations of GAS genes involved in carbohydrate uptake and utilization significantly impact regulatory pathways involving cell growth and virulence (13, 29, 36, 38, 67). Loss of BglP (ΔbglP) and BglB (ΔbglB) increased biofilm production (Fig. 5). Previous studies have shown the importance of secondary PTS components during biofilm formation in Gram-positive pathogens (7, 24, 44). However, loss of bglP or bglB instead increased the biofilm of M1T1 5448, a normally poor biofilm-forming strain. The ability of GAS isolates to form biofilms may be a strain-specific attribute (68). Compared to 5448, the M1T1 5448AP, a covS mutant of 5448 obtained after animal passage, produces a more robust biofilm in vitro, and the increased biofilm formation with ΔbglP and ΔbglB in 5448 more resembled that of 5448AP (Fig. 5). In our study on GAS PTS EIIs, they also exhibited strain-specific phenotypes between two M1T1 strains, 5448 and MGAS5005 (covS) (29). Additionally, GAS biofilm production is known to differ between tissue communities and abiotic surfaces (69). Although there are now increasing data on the importance of biofilm in GAS infections, further work is needed to fully elucidate the roles of GAS biofilm and provide better treatment options.

Early onset of SLS-mediated hemolysis (Fig. 6) was observed in both the ΔbglP and ΔbglB mutants, with the ΔbglB hemolytic profile very similar to the one exhibited by a PTS (EI) null mutant (ΔptsI) (38). SLS, which mediates red blood cell lysis, is encoded and secreted by the sag operon, and its upregulation is normally associated with the start of late logarithmic growth (38). Extremely early onset of SLS-mediated hemolysis could be indicative of induction of the sag operon, which is known to be catabolite repressed. Both ΔbglP and ΔbglB reflected this upregulation of hemolysis over growth, as exhibited in the extensive localized tissue necrosis during subcutaneous infection; however, this was more prominent with ΔbglB (Fig. 7; Fig. S6). In a serotype M14 strain, HSC5, the catabolite control protein A (CcpA) was able to regulate the sag operon indirectly, although the specific mechanism was not elucidated (67). The β-glucoside PTS operon may also indirectly influence the sag operon through similar and interconnected pathways. Additionally, PTS EIIB components have been reported to interact with PTS-regulatory domains (PRDs) (70). It is possible that disrupting these regulatory interactions by deleting the β-glucoside PTS EII components could indirectly lead to deregulation of downstream pathways involving SLS-mediated hemolysis and biofilm formation.

The complex network of sensing and regulation of available carbohydrates involves carbon catabolite repression (CCR) and CcpA (10). CCR serves as a metabolic program, signaling pathogens within host tissues of diminishing or changing nutrient levels. CcpA, in turn, controls the expression of various genes, including PTS-related genes, based on these carbon cues. A recent study suggested that a CCR-based hierarchical model does not involve only availability of nutrients but that CCR and CcpA may also dictate a growth/damage balance in GAS (15). Thus, changes that may affect uptake, such as deletions within the PTS complex, may then result in an imbalanced deregulation of growth and damage within specific tissues, as seen with the extreme localized tissue necrosis from infection with ΔptsI (38) and ΔbglB (Fig. 7; Fig. S6) mutants.

MATERIALS AND METHODS

Bacterial strains and media.

The group A Streptococcus (GAS) strains 5448 and MGAS5005 (Table 1) were isolated from patients with invasive GAS infections. Strain 5448 is representative of the globally disseminated invasive GAS M1T1 serotype. The strain MGAS5005 (Table 1) was used as the reference genome in this study (MicrobesOnline). GAS cells were cultured either in rich medium, Todd-Hewitt medium supplemented with 0.2% yeast extract (THY), or in chemically defined medium (CDM; Alpha Biosciences). CDM was supplemented with 4.8 μM ferric nitrate, 35.8 μM ferrous sulfate, 29.5 μM manganese sulfate, and freshly prepared 59.51 μM sodium bicarbonate and 11.68 μM l-cysteine, along with either 0.25% glucose or 1% concentrations of other carbohydrate sources (e.g., salicin), unless otherwise noted. Escherichia coli strain DH5α (Table 1) was used as the host for plasmid constructions. All E. coli strains were grown in Luria-Bertani (LB) broth. Antibiotics were used at the following concentrations: spectinomycin (Sp) at 100 μg/ml (Sp100) for both E. coli and GAS and kanamycin (Km) at 50 μg/ml for E. coli and 300 μg/ml for GAS. The growth of GAS was assayed by measuring absorbance through a Klett-Summerson colorimeter (Klett units) or a spectrophotometer (OD₆₀₀).

Molecular genetics.

Integrated DNA Technologies, Inc., synthesized the oligonucleotides used in this study (Table 2). Plasmids used (Table 1) were isolated with either the Wizard Plus SV miniprep kit (Promega) or the Qiagen plasmid purification midikit (Qiagen). Genomic DNA was extracted from GAS using the Master-Pure complete DNA and RNA purification kit for Gram-positive bacteria (Epicentre). Restriction enzymes, Antarctic phosphatase, and T4 DNA ligase (New England BioLabs) were used according to the manufacturer’s instructions. PCR was performed using either Taq DNA polymerase (New England BioLabs) or high-fidelity AccuPrime Pfx DNA polymerase (Life Technologies) with ca. 1 μg of DNA template and 10 pmol of the appropriate primers (Table 2). When necessary, PCR products were purified using the PCR Clean-Up system (Promega). Transformations were performed with a Gene Pulser Xcell electroporator (Bio-Rad) as recommended by the manufacturer, using prepared electrocompetent cells of E. coli or GAS. Genomic DNA (gDNA) from GAS was purified using the MasterPure Complete DNA purification kit (Epicentre Biotechnologies). Sanger DNA sequencing was performed by Genewiz, Inc.

Insertional inactivation of EII transporters.

Polar insertional inactivation mutants in MGAS5005 were created using the pCRS integrative plasmid as previously described (29, 35). Briefly, nonpolar EII transporter libraries were generated by integrating a 500-bp region of internal homology within the target open reading frame (ORF) into pCRS. The resulting mutagenic plasmids were then transformed into MGAS5005 electrocompetent cells and incubated at 30°C with 5% CO₂ for 2 days. Transformed colonies were streaked onto THY plates with Sp100 and then passaged in THY broth with Sp100 at 37°C for integration of the plasmid. Resulting colonies were verified by PCR using plasmid-specific and gene-specific primers (Table 1).

Nonpolar deletion of bglP and bglB in GAS 5448.

Nonpolar deletion mutants were generated in GAS 5448 by replacing the ORF of the corresponding gene with promoterless aphA3 using allelic exchange as previously described (71). For bglP, DNA fragments flanking the gene were amplified with primers (Table 2) oAX0475.1 and oAX0475.2 (5′ end) and oAX0475.3 and oAX0475.4 (3′ end), joined in frame by overlap extension (SOE) to the aphA3 cassette, and ligated into BamHI-digested pCRS to produce the plasmid pAX0475K (Table 1). Allelic replacement of the bglP gene by the aphA3 cassette was conducted as previously described (71), creating strain 5448.ΔbglP (Table 1). For the ΔbglB mutant, flanking DNA fragments amplified with primer pairs (Table 2) oAX0476.1 and oAX0476.2 (5′ end) and oAX0476.3 and oAX0476.4 (3′ end) were used to produce pAX0476K and generate GAS strain 5448.ΔbglB (Table 1). Rescue strains were produced by excising the entire plasmids through selection of spectinomycin-susceptible (Sp^s) and Km^s clones (Table 1), thereby recovering WT genotypes (71). Mutants and rescues were verified by PCR amplification using appropriate verification primers upstream and downstream of the regions of interest (Table 2) as well as DNA sequenced.

Mapping of the licT-bglPB operon and transcriptional start site (TSS).

Four independent colonies of each strain were grown static overnight in 10 ml of THY with the appropriate antibiotic, if applicable, at 37°C with 5% CO₂. RNA isolation was carried out as previously described (41). Briefly, overnight cultures were inoculated into fresh media and grown to late logarithmic phase (ca. 100 Klett units). Cell pellets were frozen at –20°C until RNA extraction. Pellets were resuspended in TRIzol (Zymo Research), and acid-washed glass beads (Sigma-Aldrich) were added to lyse GAS cells by vortexing. After centrifugation to collect the beads, cell lysates were used for total RNA extraction using the Direct-zol RNA miniprep kit (Zymo Research) as recommended by the manufacturer. RNA samples were treated with Turbo DNase-free kit (Invitrogen) to eliminate DNA contamination.

First-strand cDNA synthesis was done using 500 to 800 ng of total RNA with Moloney Murine Leukemia Virus (M-MulV) reverse transcriptase (New England BioLabs) with primer bglB-cDNA-R (Table 2), following the manufacturer’s instructions. Subsequent PCR was carried out as previously described (41). Primers (Table 2) within the 5′ start and 3′ end of each licT-bglPB ORF were used as controls, as well as samples containing full-length primers, but without RNA or RT enzyme. Primers (Table 2) used during subsequent PCR to determine cotranscription were the following: licT to bglP, licT-RevTrans-F and bglP-RevTrans-R; bglP to bglB, bglP-RevTrans-F and bglB-RevTrans-R; and licT to bglB, licT-RevTrans-F and bglB-RevTrans-R.

5′ rapid amplification of cDNA ends (RACE; Invitrogen) was performed according to the manufacturer’s instructions, as previously described (41). Briefly, total RNA was used to generate cDNA with licT-GSP1 within licT (Table 2). cDNA was column purified, and then a homopolymeric tail was added to the 3′ end of the cDNA. PCR amplification was done with the nested gene-specific primer licT-GSP2 (Table 2) and a kit-supplied deoxyinosine-containing anchor primer to recognize the homopolymeric tail. A second PCR was performed using another nested gene-specific primer, licT-GSP3 (Table 2), with a kit-supplied forward primer to generate enough specific product and enrich the 5′ RACE product. The 5′ end of the transcript was determined by sequencing this PCR product.

Luciferase assays.

GAS 5448 was transformed with the desired luciferase plasmids (Table 1). Cells were grown static overnight in 10 ml of THY with appropriate antibiotic, if needed, at 37°C with 5% CO₂. Overnight cultures were then inoculated (1:20 dilution) into fresh THY. For assays in C-medium and CDM, overnight cultures were first washed with saline before inoculation into the corresponding medium. At Klett units corresponding to various phases of growth (early log, late log, and stationary) as determined by previously established growth curves, samples were removed and placed on ice. The volume of each sample needed was calculated to normalize each strain during each phase of growth to have the same predetermined cell units of either 15 (for P_licT samples) or 5 (for positive control pKSM210), using the following equation: volume needed (in milliliters) = (cell units × 2)/Klett value. Samples were then pelleted, the supernatant was discarded, and pellets were placed at –20°C until luciferase activity was measured. Using a luciferase assay system (Promega), pellets were resuspended in 500 μl of 1× lysis buffer. A total of 100 μl of each lysed sample was measured for luciferase activity using a Centro XS3 LB 960 luminometer (Berthold Technologies), into which 50 μl of Luciferin-D reagent was injected. Background luminescence from promoterless pKSM720 was deducted from all samples. Data were graphed and presented as relative light units divided by cell units (RLU/CU).

Lancefield bactericidal assay.

Blood donations were approved by the University of Maryland Institutional Review Board (IRB no. 10-0735), and written consent of donors was archived. As previously described, the ability of GAS strains to survive in whole human blood were tested (72). Overnight cultures were inoculated 1:20 in THY, grown to early mid-logarithmic phase (OD₆₀₀ = 0.1 to 0.15), and then serially diluted in saline. A total of 50 μl of the 10⁻⁴ dilution was used to inoculate 500 μl of fresh whole human blood, which was then incubated while rotating at 37°C. After 3 h of growth in whole human blood, cultures were serially diluted and plated on either THY agar (WT) or THY with antibiotic (mutants). The multiplication factor (MF) was determined by dividing the CFU obtained after growth in blood by the initial CFU inoculum. Data are presented as percent growth in blood using the following formula: percent growth = (MF of mutant/MF of WT) × 100.

Biofilm assay.

Biofilm assay was performed as previously described (45), with modifications. GAS cultures were grown overnight as described above. The next day, 1.9 ml of biofilm-producing brain heart infusion (BHI) medium (Difco) (46) was added to individual wells of a 24-well polystyrene plate (Costar). Overnight cultures were then vortexed and added to each well (1:20), with samples performed in triplicates. Plates were rocked back and forth to ensure even dispersion of the cells and incubated at 37°C with 5% CO₂ for 48 h. After incubation, culture medium was removed, taking care not to disturb the biofilm at the bottom of the well. Any planktonic, nonadherent cells were gently washed away with 2 ml of saline. After the saline was completely removed, plates were baked at 50°C for 30 min to dry the biofilm and improve adhesion. A total of 500 μl of crystal violet was added to each well and incubated at room temperature for 15 min, and then excess crystal violet was removed. To allow any unbound crystal violet to elute off the biofilm, each well was washed and incubated for 5 min with 2 ml of saline three times. To quantify biofilm formation, 100% ethanol was then added to each well, incubated for 10 min, diluted 1:10 with water, and measured for OD₆₀₀.

Hemolysis assay.

The hemolysis assay was carried as previously described (38), with minor modifications. Overnight cultures of GAS strains were used to inoculate THY as described above, supplemented with 10% heat-inactivated horse serum. Cultures were grown statically at 37°C. At 1 h and every 30 min thereafter for 8 h, 1-ml samples of the cultures were taken, the OD₆₀₀ reading was taken, and the sample was saved at –20°C. The following day, samples were pelleted and 50 μl of the supernatant was added to a sterile 1.5-ml microcentrifuge tube containing 450 μl of sterile 1× phosphate-buffered saline (PBS) and 500 μl of prepared defibrillated sheep RBCs. RBCs were separately prepared by washing 500 μl of sheep blood several times with 1 ml of PBS at 3,000 × g until the supernatant was relatively clear and then resuspending in 1 ml of PBS. Following addition of the culture to the washed RBCs, mixtures were incubated at 37°C for 1 h and then centrifuged at 3,000 × g for 10 min to clear intact RBCs. The OD₅₄₁ of the supernatant of each sample was then measured on a 96-well plate (Corning/Costar) to determine the release of hemoglobin upon RBC lysis. Sodium dodecyl sulfate (SDS; 0.1%) added to prepared RBCs was used as a positive control for maximum hemoglobin release, and PBS added to RBCs was used as a negative control. An OD₅₄₁ above 0.2 was considered to be positive for RBC lysis, and data shown are from at least three biological replicates.

Ethics statement.

All animal work was performed in an Association of Assessment and Accreditation of Laboratory Animal Care (AAALAC)-accredited animal biosafety level 2 (ABSL2) facility at the University of Maryland, College Park. Institutional Animal Care and Use Committee (IACUC)-approved protocols (R-MAR-19-17) for humane treatment of animal subjects in accordance with the Office of Laboratory Animal Welfare (OLAW) at NIH, Public Health Service, and the Guide for the Care and Use of Laboratory Animals (73) guidelines were used. Extreme care was taken to limit the pain and distress of the animals.

Soft tissue infections in mice.

Infections were performed using the mouse model of subcutaneous soft tissue infection as previously described (40). Log-phase cells were prepared from overnight cultures as described above, and then cell pellets were resuspended in saline. Fur was removed from the haunch of anesthetized mice, and they were injected subcutaneously with 100 μl of the cell suspension (ca. 2 × 10⁸ initial CFU) and monitored twice daily for up to 7 days. Upon signs of systemic morbidity (hunching, lethargy, and leg paralysis), mice were euthanized via CO₂ asphyxiation. Lethality was assessed by Kaplan-Meier survival analysis over the entire 7-day period. Lesions were imaged at 24 and 48 h postinfection for determination of lesion size and severity using ImageJ software (NIH). Each lesion per time point was digitally measured 5 times for size and mean red pixel intensity, and then the average was used as one data point. Red pixel intensity (severity) was graphed with an inverted y axis to correspond with the severity of the lesions, as darker lesions have lower red pixel intensity values.