Functional analysis of the group A streptococcal luxS/AI-2 system in metabolism, adaptation to stress and interaction with host cells

Background The luxS/AI-2 signaling pathway has been reported to interfere with important physiological and pathogenic functions in a variety of bacteria. In the present study, we investigated the functional role of the streptococcal luxS/AI-2 system in metabolism and diverse aspects of pathogenicity including the adaptation of the organism to stress conditions using two serotypes of Streptococcus pyogenes, M1 and M19. Results Exposing wild-type and isogenic luxS-deficient strains to sulfur-limited media suggested a limited role for luxS in streptococcal activated methyl cycle metabolism. Interestingly, loss of luxS led to an increased acid tolerance in both serotypes. Accordingly, luxS expression and AI-2 production were reduced at lower pH, thus linking the luxS/AI-2 system to stress adaptation in S. pyogenes. luxS expression and AI-2 production also decreased when cells were grown in RPMI medium supplemented with 10% serum, considered to be a host environment-mimicking medium. Furthermore, interaction analysis with epithelial cells and macrophages showed a clear advantage of the luxS-deficient mutants to be internalized and survive intracellularly in the host cells compared to the wild-type parents. In addition, our data revealed that luxS influences the expression of two virulence-associated factors, the fasX regulatory RNA and the virulence gene sibA (psp). Conclusion Here, we suggest that the group A streptococcal luxS/AI-2 system is not only involved in the regulation of virulence factor expression but in addition low level of luxS expression seems to provide an advantage for bacterial survival in conditions that can be encountered during infections.


Background
Bacterial cell density-dependent signaling, also termed "quorum sensing", is used by bacteria to collectively modulate gene expression in response to changes in the population density [1][2][3][4][5][6]. Although signaling can be achieved through a variety of regulatory mechanisms, all systems described to date involve the production, secretion and detection of extracellular low-molecular-weight signaling molecules called "autoinducers" [1,3,7,8]. Due to the specificity of their respective sensors, recognition of acylated homoserine lactones (AHLs) (gram-negative bacteria) and peptide autoinducers (gram-positive bacteria) is restricted for communication within the same species. More recently, a novel quorum sensing system involving a furanone-like signaling molecule termed "autoinducer 2" (AI-2) was described to regulate bioluminescence in Vibrio harveyi [9,10]. AI-2 is synthesized by the luxS gene product, which has been identified in the genome of over 55 gram-negative and gram-positive bacterial species [1,11,12]. Since culture supernatants of several bacterial species had a complementary effect on luxS-deficient V. harveyi, AI-2 has been proposed to function as a "universal" signaling molecule for interspecies communication [1,9,10]. In addition to controlling bioluminescence, recent studies show that the luxS/AI-2 signaling is involved in the regulation of pathogenicity in several organisms [5,12].
Apart from its role in quorum sensing, the enzyme LuxS is tightly coupled to the S-adenosylmethionine (SAM) utilization pathway [11][12][13]. SAM is an essential donor of methyl groups for DNA, RNA and other methylation reactions. Its utilization by methyl transferases yields S-adenosylhomocysteine (SAH), which is toxic for the cells and is eliminated through hydrolysis by the nucleosidase Pfs to produce S-ribosylhomocysteine (SRH) and adenine. Finally, LuxS cleaves SRH into homocysteine and 4,5dihydroxy-2,3-pentanedione (DPD). DPD spontaneously forms the cyclic pro-AI-2 molecule, which in V. harveyi reacts with borate to form a stable cyclic furanosyl borate diester [1,11,13].
Knowledge about quorum-sensing systems in the grampositive human pathogen Streptococcus pyogenes is rather limited. S. pyogenes (Group A Streptococcus, GAS) is responsible for a broad range of diseases including mild illnesses such as pharyngitis, impetigo and scarlet fever and more severe invasive infections such as necrotizing fasciitis, streptococcal toxic shock syndrome and postinfectious rheumatic fever [14,15]. In GAS, like in other bacteria, pathogenicity is multifactorial and requires the coordinated temporal regulation of virulence factor expression in response to changing environmental conditions and cell population density [14,16,17]. Studies have confirmed that expression of several GAS virulence genes is temporal and dependent on growth phase [16][17][18][19][20]. Furthermore, a number of specific regulators (e.g. two-component systems and response regulators) modulate virulence gene expression in a growth phase dependent manner [16,[21][22][23][24]. Based on similarities with previously described peptide signaling regulons of S. pneumoniae (com system) and S. aureus (agr system), two putative quorum-sensing systems, sil and fasBCA, have been reported in GAS [21,25]. Although no signaling peptide could be identified within the fasBCA locus, the putative pherom-one peptide SilCR from the sil operon was shown to regulate DNA uptake and the ability of GAS to cause invasive infection [21,25]. In addition, the observation of a luxS homologue in the GAS genome suggested that quorum sensing via the luxS/AI-2 signaling could have a relevant function in the pathogenesis of this organism [26,27]. Two recent studies report a role of luxS/AI-2 in the regulation of virulence gene expression in M3 and M6 serotypes [26,27]. However, the role of the luxS/AI-2 in the AMCrelated metabolism and adaptation to stress conditions in GAS remains unknown.
Here, we were interested in investigating further the function of the luxS/AI-2 system in GAS serotypes M1 and M19. Expression of luxS in connection with production of AI-2 like activity was analyzed. The metabolic role of luxS in the activated methyl cycle (AMC) in GAS was examined. Wild-type and isogenic luxS-deficient strains were compared in regard to their adaptation to diverse growth and stress conditions as well as diverse aspects of pathogenicity including interaction with epithelial cells and macrophages. In summary, our data suggest an important function of the luxS/AI-2 system in survival and growth of GAS under conditions that are relevant during infections.

Construction of luxS-deficient mutants
As reported previously, GAS possesses a luxS homolog (coding sequence: 483 bp), the predicted translational product of which shares 36% identity with the LuxS protein from V. harveyi [26]. Analysis of the luxS coding sequences in clinical isolates of two different serotypes, RDN29 (M1 type) and RDN02 (M19 type), revealed 100% amino acid sequence identity. In a previous study, a GAS strain of serotype M6 (JRS4) was shown to contain in its genome the insertion element, IS 1239, inserted 111 bp upstream of the luxS translational start codon [26]. This insertion was assumed to have an influence on luxS expression [26]. Further analysis of the sequence up to 300 bp upstream of the luxS coding sequence in RDN29 and RDN02 did not reveal the presence of IS 1239 (data not shown). To analyze the function of the luxS/AI-2 system in GAS, we deleted the luxS gene in RDN29 and RDN02, thus creating EC478 and RDN306, respectively (Table 1).

Temporal luxS expression and production of AI-2 like activity
First, we investigated luxS expression at the transcriptional level. Northern blot analysis of RNA extracts from wildtype strains revealed a monocistronic luxS-specific transcript of approximately 650 bases in size (Fig. 1A). In both strains, the luxS transcript levels were abundant from lag to early-logarithmic phase, after which gradual decrease of the signal intensity towards stationary phase was observed ( Fig. 1A). Primer extension analysis further identified the transcription initiation site (+1) as a guanine located 20 nucleotides upstream of the translation start codon (ATG) (Fig. 1B). Then, we investigated whether the growth phase-dependent expression of luxS transcripts correlated with production of AI-2 like activity. Analysis of conditioned media (CM) from the luxS-deficient mutants demonstrated a loss of AI-2 like activity production in a bioluminescence reporter assay described earlier [9]. In contrast, CM from the wild-type parental strains induced luminescence in the AI-1 sensor defective reporter V. harveyi strain BB170 with AI-2 like activity peaking at late-logarithmic phase ( Fig. 2A, B). CM prepared from the complemented luxS-deficient strains restored luminescence production in the reporter V. harveyi strain to a level comparable to that of wild-type CM ( Fig. 2A). Furthermore, we showed that the S. pyogenes luxS gene complemented successfully the lack of AI-2 production in E. coli DH5-α (data not shown), thus confirming that the luxS homologue in both M1 and M19 strains is required for production of AI-2 like activity. We then addressed the question whether CM with AI-2 like activity has an autoregulatory effect on luxS expression in S. pyogenes. CM from wild-type and luxS-deficient cultures grown to different phases was added to early-logarithmic wild-type cultures. No significant differences in the intensity of the luxS transcript signal were observed by Northern blot analysis when comparing induction with CM from both wild-type and mutant strains and control media (data not shown). Taken together, our data demonstrate that luxS transcription and production of AI-2 like activity are shifted out of phase. AI-2 like activity peaked at a time point -late-logarithmic phase -when luxS transcript levels rapidly decreased. Importantly, an auto-regulatory function of AI-2 like activity containing CM on luxS expression could not be observed in GAS.

Role of luxS in growth and metabolism
Growth rates and yields, colony and chain morphology of wild-type and luxS-deficient strains in complex THY medium, THY medium supplemented with 10% fetal bovine serum, C-medium and CDM were similar (data not shown). In addition, wild-type and luxS-deficient strains did not significantly differ in their ability to form primary adhesion to uncoated polystyrene surfaces (initial step in biofilm formation) when grown in THY or CDM (data not shown). To investigate whether inactivation of luxS and therefore disruption of the AMC would lead to a metabolic burden that might influence growth or fitness, we analyzed the growth rates of wild-type and luxS-deficient strains in sulfur-limited CDM (CDM-S1 and CDM-S2). Although overall growth rates and yields were significantly reduced in the restricted media compared to those in CDM, they were similar when wild-type strains Analysis of the luxS transcript in S. pyogenes  were compared to the luxS-deficient mutants (data not shown). These results indicate that luxS does not have an essential AMC-related metabolic role in GAS.
Role of luxS in adaptation to host-induced stress conditions. Further, we investigated whether luxS is involved in the adaptation of S. pyogenes to stress conditions that can be encountered during infection. Oxidative, acid and salt stresses, which have been shown to interfere with virulence factor expression in GAS [28][29][30], were studied. Challenging wild-type and luxS-deficient strains with hydrogen peroxide (4 mM) and high salt concentrations (150 and 650 mM NaCl) did not induce any significant change in survival rate (data not shown). However, exposure of bacterial cultures to acidic conditions (pH 4) for 2, 4 and 6 h revealed a profound increase in acid tolerance in the luxS-deficient strains (Fig. 3A). Complementing the luxS-deficient strains with luxS in trans resulted in restoration of acid sensitivity to a wild-type level (Fig. 3A, shown only for the 6 h time point), clearly linking luxS to the observed phenotype. Importantly, we could show that exposing S. pyogenes cultures to low pH led to reduced expression of luxS (Fig. 3B), thus suggesting a correlation between reduced luxS expression and survival under acidic conditions. Accordingly, CM from these cultures showed reduced AI-2 like activity (Fig. 3C).
Then, we analyzed the role of the luxS/AI-2 system in adaptation to an environment that resembles in vivo conditions by using RPMI medium (used for culturing eukaryotic cells) and 10% serum. We observed reduced luxS expression at the transcriptional level in wild-type cultures grown in RPMI medium compared to cultures grown in the nutrient rich THY medium (Fig. 4A). Furthermore, AI-2 activity in CM from wild-type cultures grown in RPMI was reduced to a level similar to CM from luxS-deficient cultures (Fig. 4B). During growth in RPMI, due to the presence of buffer salts and cultivation in the presence of CO 2 , no acidification occurs. Therefore, under this growth condition, regulation of luxS expression cannot be attributed to low pH. Since low concentration of iron in the RPMI medium could be responsible for the observed phenotype, we supplemented RPMI medium with iron and depleted THY medium from iron. In both cases, no change in luxS expression could be observed (data not shown). Our data thus indicate that under specific conditions encountered in the host (acidic conditions and in vivo-like environment), there is an advantage for GAS to lower the expression of luxS and the production of AI-2 like activity.
Role of luxS in internalization and survival in human epithelial cells and macrophages. These results prompted us to determine whether a lack of luxS expression gives a beneficial advantage for GAS to survive in an intracellular environment. Interactions of wild-type and luxS-deficient strains of both serotypes M1 and M19 with monolayers of human pharyngeal epithelial cells were analyzed. No differences in the number of bacteria adhering to the epithelial cells (30 min, 1 h and 2 h after infection) were observed when comparing wild-type and mutant strains (data not shown). Using an antibiotic protection assay, the internalization and survival rates of the luxS mutant strains of both M1 and M19 serotypes were significantly higher (about 5-fold) at 2 h and 5 h after the addition of antibiotics compared to those of wild-type strains (Fig.  5A). Analyzing the number of bacteria that survived intracellularly in macrophages led to the same results. The luxS deficient mutants showed significantly higher internalization and survival rates (about 2.5-fold) in macrophages at 0 h, 1 h and 2 h after the addition of antibiotics compared to the wild-type parents (Fig. 5B). With both serotypes and in both epithelial and macrophage cells, complementing the mutant strains with the luxS allele in trans restored the intracellular survival rates to almost wild-type levels. Since the role of luxS in the AMC-related metabolic pathway seems limited, we analyzed the possibility of a signalling function of the luxS/AI-2 system in GAS survival in host cells. Antibiotic protection assays were performed with mixed cultures of wild-type and luxS-deficient cells (ratio 1:1) comparing with non-mixed wild-type and luxSdeficient cultures. The survival rates of the mixed culture in both eukaryotic cell types were almost identical to those of the wild-type strain (data not shown), thus suggesting a potential complementation effect of the luxSdeficient cells by the wild-type parent cells possibly by AI-2 production and signaling. In summary, the data indicate that lowering the expression of luxS provides GAS an advantage to survive the harsh conditions encountered in host cells.

Virulence gene expression in luxS mutants
We also studied whether luxS had regulatory effects on pathogenicity functions other than those already described (e.g. effect on cysteine protease, SLS and hyaluronic acid production) [26,27]. Expression of a number of genes encoding virulence factors and their regulators [16,21,28,[31][32][33][34][35][36] was then studied by Northern blot analysis throughout the entire growth phase ( Table  2). A consistent decrease in the fasX (encoding the effector molecule of the fasBCA regulatory system [21]) transcript and a consistent enhancement of the sibA (psp) (encoding a secreted immunoglobulin binding protein [32]) transcript were detected in the luxS-deficient RDN02 (M19) strain compared to the wild-type parent, however not in the M1 background (Fig. 6A, B). This finding does not only add fasX and sibA to the list of genes being influenced by luxS but underlines the fact that regulatory pathways linking the luxS/AI-2 pathway to virulence-associated

Discussion
Among the quorum sensing systems described to date, the luxS/AI-2 pathway has been shown to be involved in the regulation of virulence in a number of gram-negative and gram-positive bacteria [1,5,11,12]. In this study, we analyzed the expression of the streptococcal luxS/AI-2 system, its possible role in the AMC metabolic pathway and its function in adaptation to diverse host-induced stress conditions in two GAS clinical isolates of serotypes M1 and M19.
Expression analysis of luxS indicated that the luxS transcript is monocistronic (644 bases) in GAS serotypes M1 and M19. In a previous report, which was based on sequence analysis, the luxS transcript in GAS (M6 strain) was predicted to be polycistronic as a distal member of the fatty acid metabolism operon [26]. In our study, Northern blot analysis of luxS expression did not reveal any additional transcript of higher size. Furthermore, we showed that the monocistronic transcript is expressed in a growth phase-dependent manner, a finding already reported for bacterial species like Streptococcus bovis, whereas transcription of luxS in species like Salmonella typhimurium and Vibrio fischeri has been shown to be constitutive [37][38][39]. Accordingly, production of AI-2 like activity occurred in a temporal fashion in the serotypes analyzed, however, shifted out of phase compared to luxS expression. AI-2 like activity peaked at late-logarithmic phase when the luxS transcript had already declined being just above the detectable expression level. We also investigated the possibility of an auto-regulation mechanism of luxS expression in GAS. Induction experiments by Northern blot analysis failed to show a regulation of luxS expression by AI-2 dependent CM. These findings are in accordance with previous work in S. typhimurium, where AI-2 production and luxS transcription were not in phase either, also suggesting that luxS expression was not regulated by AI-2 [38].
Despite intensive research on luxS and AI-2 during the last years, a clear separation of the possible metabolic function of AI-2 from its possible signaling activity could not be achieved. A recent comparative genomic and phylogenetic analysis of synthesis and signal transduction pathways showed that S. pyogenes contains the two-step enzymatic pathway catalyzed by the Pfs and LuxS enzymes to detoxify SAH leading to AI-2 production [13,40]. Interestingly, in S. pyogenes the AMC is incomplete as it lacks the homocysteine methyltransferases (MetE or MetH) that convert the SAH-derived metabolic product homocysteine to methionine. Nevertheless, S. pyogenes encodes the SAM synthetase ortholog (MetK) that permits conversion of methionine to SAM [40]. Therefore, either GAS cannot recycle homocysteine to methionine (as it has been described in Borrelia burgdorferi [41]) or recycling occurs through other types of methyltransferases. In the former case of a de novo methionine synthesis defect, bacterial growth would exclusively depend on sulfur provided from external sources. In our study, decreasing media concentrations of cysteine and methionine reduced to a similar extent bacterial growth and yields in both wild-type and luxS-deficient strains. These data indicate that GAS, as a poly-auxotrophic organism, is unable to recycle homocysteine to methionine. Moreover, we showed that inactivation of luxS does not lead to a metabolic burden that would influence growth or fitness. Although AI-2 can certainly be considered as a metabolite, the viability and lack of growth defects of the investigated luxS-deficient mutants in various media including minimal media depleted of sulfur sources argues against an essential role of luxS in AMC-related metabolism.
The previously described regulatory role of luxS in the production of virulence factors in GAS [26,27] prompted us to investigate the role of the luxS/AI-2 system in the adaptation to host-induced stress conditions. During an ongoing infection, bacteria often have to face challenging conditions in particular niches of the host, including changes in pH, and therefore are forced to develop quickly an adaptive response, which requires fine-tuning of pathogenicity gene expression. Although we could not identify a possible involvement of the GAS luxS system in adaptation to oxidative and salt stress, we showed an increased survival of S. pyogenes under acidic conditions when the luxS/AI-2 system was down-regulated. This is in accordance with luxS expression and AI-2 production being significantly lowered when GAS cells were grown in low pH conditions. Interestingly, previous studies in S. mutans showed that acid sensitivity was enhanced in luxS deficient mutants [42,43] and that luxS expression was increased at low pH [44,45]. These reverse effects of luxS on tolerance to acidic conditions in S. pyogenes and S. mutans need to be considered in regard to their different living habitats and pathogenesis. Thus, our data reveal a link between the S. pyogenes luxS/AI-2 system and pathways involved in adaptation of the organism to stress [46,47]. In addition, we also observed a reduction of luxS expression and AI-2 like activity in GAS cells grown in serum enriched RPMI, a medium with a composition similar to that of human plasma. In a previous study, Marouni et al. reported that in an M6 serotype, the survival of a luxS-deficient mutant in epithelial cells at 4 h after infection was higher compared to the wild-type parent [27]. Here we analyzed further the role of luxS in interaction of S. pyogenes with host cells. No differences in adhesion rates to human pharyngeal epithelial cells were observed when comparing wild-type and luxS-deficient (A) Acid tolerance assay Figure 3 (A) Acid tolerance assay. The ability of RDN02 (wild-type M19 strain), RDN306 (isogenic luxS-deficient mutant), EC480 (RDN02 (pEC82)), EC481 (RDN306 (pEC82)) and EC482 (RDN306 (pEC83)) to survive an acid challenge (pH 4) is shown as survival in percent (y axis) at the 6 h time point. The average % survival values and standard deviations of three independent experiments are shown. The asterisk indicates that the difference in survival rates between the luxS-deficient mutant and the wild-type strain was statistically significant (P < 0.05). Similar data were obtained with the RDN29 (M1 type) derivative strains. (B) Effect of low pH on luxS expression. Northern blot analysis was performed with total RNA isolated from RDN02 cultures grown to early-logarithmic phase and then incubated for 90 min with THY medium at pH 5.5, 6.0 and 7.5. Blots were hybridized with probes specific to luxS and rRNA16S (acting as loading control). The estimated sizes of transcripts are indicated. Pixel rates of signals (pixel counts obtained with the luxS probe versus pixel counts obtained with the rRNA16S probe) are shown. The blot shown is representative of three independent experiments. Similar data were obtained with the RDN29 strain. (C) Induction of luminescence in V. harveyi reporter strain BB170 by CM of RDN02. CM was prepared from RDN02 cultures grown to early-logarithmic phase and then incubated for 90 min with THY medium at pH 5.5 (RDN02-pH 5.5), 6.0 (RDN02-pH 6.0) and 7.5 (RDN02-pH 7.5) for 90 min. THY medium at pH 7.5, 6.0 and 5.5 and CM of DH5-α (AI-2 deficient) were used as negative controls; CM of BB170 was used as positive control. The luminescence values expressed in specific light units (SLU = RLU (relative light units)/OD 620 nm ) (y axis) are plotted versus time (h) (x axis). Data shown are representative of three independent experiments. Similar data were obtained with CM of the RDN29 derivative strains. * mutants in both M1 and M19 serotypes. However, the luxS-deficient mutants had a significant advantage to survive intracellularly in both epithelial cells (over a period of 7 h after infection) and macrophages (over a period of 3 h after infection). Taken together, we show that low level of luxS and AI-2 expression seems to provide a competitive advantage for GAS survival under specific conditions encountered during infection.
With luxS being linked to stress and adaptation to the host, we were interested in investigating additional effects of luxS on virulence factor expression in GAS. In an M6 serotype, luxS was shown to regulate streptolysin S (SLS) expression at the transcriptional level and SpeB cysteine protease activity [26]. In an M3 serotype, luxS was reported to regulate expression of SpeB and M protein at the transcriptional level and hyaluronic acid capsule at the post-transcriptional level [27]. In our report, we show that luxS had a positive effect on fasX expression and a negative effect on sibA (psp) expression at the transcriptional level.
fasX is a small RNA molecule, effector of the fasBCA operon, which has regulatory functions on virulence factor expression. fasX was also suggested to be involved in local tissue destruction and general bacterial aggressiveness towards host cells [21,48]. sibA (psp) is a virulence gene encoding a secreted immunoglobulin binding protein [32]. Remarkably, the above described regulatory effects were only observed in the M19 strain, thus suggesting a strain or serotype dependent effect of luxS on the expression of these two targets. The strain-or serotypedependent effect of luxS observed in this study emphasizes differences in regulatory pathways among different GAS isolates. Along these lines, previous reports demonstrated that mutations in GAS regulators might alter disease progression [49,50]. This is well illustrated with the two-

Conclusion
Our data together with previous reports suggest a complex role of the luxS/AI-2 system in S. pyogenes. Here, we showed that expression of both luxS and AI-2 occurs in a temporal fashion but AI-2 like activity does not seem to have a regulatory effect on luxS expression. Analysis of the role of luxS in metabolism demonstrated a limited role of luxS in the AMC-related metabolism. However, studying the possible function of the luxS/AI-2 system in adaptation to stress revealed that a down-regulation of luxS expression provides an advantage for S. pyogenes to tolerate acidic conditions and grow in a host environmentmimicking medium. Accordingly, there was an increased ability of the luxS-deficient mutants to survive intracellularly in both epithelial cells and macrophages. Altogether, our data suggest an important function for the luxS/AI-2 system in survival and growth of GAS under conditions that are relevant during infections. Based on the data outlined in this article, it is tempting to speculate that in GAS, luxS is not exclusively a key part of a detoxification pathway but rather modulation of luxS expression levels would allow adjusting bacterial fitness in response to changing host conditions. Finally, our study revealed two novel virulence-associated targets of luxS in S. pyogenes: the regulator fasX RNA and the virulence gene sibA.

Bacterial strains and growth conditions
Bacterial strains and plasmids used in this study are listed in Table 1 shaking or on agar plates. V. harveyi was cultured aerobically at 28°C in an Autoinducer Bioassay (AB) medium [9]. Transformation of E. coli and S. pyogenes was performed as previously described [59,60]. Whenever required, suitable antibiotics were added to the medium to the following final concentrations: erythromycin 300 μg/ml for E. coli and 3 μg/ml for S. pyogenes; kanamycin 25 μg/ml for E. coli and 300 μg/ml for S. pyogenes. Bacterial cell growth was monitored periodically by measuring the optical density of culture aliquots at 620 nm using a microplate reader (SLT Spectra Reader).

DNA manipulations
DNA manipulations including DNA preparation, amplification, digestion, ligation, purification, agarose gel electrophoresis and Southern blot analysis were performed according to standard techniques [60]. Synthetic oligonucleotides used as primers in this study (Table 2) were supplied by VBC-Biotech Services GmbH. Sequencing reactions were performed at VBC-Biotech Services GmbH.

Construction of the luxS-deficient mutants
Replacement of the luxS coding sequence with a kanamycin resistance cassette, aphIII [61], was performed selecting for a double cross-over event. For this purpose, a 1209 bp fragment upstream of luxS (luxS-up) and a 1107 bp fragment downstream of luxS (luxS-down) were amplified using wild-type genomic DNA as template and primers (containing flanking restriction sites) oliRN22/oliRN21 and oliRN20/oliRN19, respectively. The aphIII cassette was amplified using plasmid pAT21 as template and primers oliRN23/oliRN24. After digestion with the respective restriction enzymes, the three fragments were ligated and cloned into pUC19 (suicide vector for S. pyogenes). The resulting plasmid pEC131 was purified, linearized with the restriction enzyme ScaI (which cuts within the ampicillin resistance cassette) and used to transform electro-competent wild-type S. pyogenes. Kanamycin resistant clones were selected and the correct replacement event was checked by PCR analysis using combinations of primers oliRN267 and oliRN268 derived from flanking regions upstream and downstream of the luxS-up and luxS-down fragments and primers oliRN69 and oliRN70 specific to the aphIII cassette. Southern blot analysis was done to further confirm that the recombination events had not affected the DNA regions located upstream and downstream of luxS. The S. pyogenes mutants were grown in liquid medium without antibiotic unless otherwise specified.

Construction of plasmids for complementation studies
Plasmid pEC82 contains repDEG-pAMβ1 (the origin of replication of pAMβ1), ermAM (an erythromycin resistance gene with its own promoter and transcriptional terminator), ColE1 (a pUC19-based ColE1 origin of replication for E. coli) and an expanded MCS (multiple cloning site). To create the pEC83 luxS-complementation vector, a 1152 bp large DNA fragment (PluxS-luxS-TT) containing the luxS coding sequence, its putative promoter region and putative transcriptional terminator, was amplified from wild-type genomic DNA using primers (containing flanking restriction sites) oliRN205/ oliRN280 and cloned into pEC82. Plasmids pEC82 and pEC83 were introduced in competent E. coli and S. pyogenes strains selecting for erythromycin resistant clones.

RNA analysis
Total RNA was prepared from culture samples harvested at different time points during growth and further processed for normalization and Northern blot analysis as previously described [62] with minor modifications. Specific α-32 P-dATP labeled DNA probes corresponding to internal fragments of genes were created by amplification of wild-type genomic DNA using primers described in Table  2. Primer extension analysis of the luxS transcript and DNA sequencing reactions were carried out according to standard protocols [60].

Preparation of conditioned medium for induction assays
Conditioned medium (CM) was prepared from bacterial cultures grown in THY buffered with 0.1 M HEPES (Sigma) (pH 7.5) to early-, mid-and late-logarithmic phase. Culture supernatants were separated from bacterial pellets by centrifugation at 9,500 × g for 30 min at 4°C, filter sterilized (0.45 μm) and stored at -20°C. For induction assays, early-logarithmic wild-type cultures were induced for 60 and 90 min with THY adjusted to pH 7.5, 6.0 and 5.0 or with CM of wild-type and luxS-deficient cultures from early-, mid-and late-logarithmic growth phase and then subjected to Northern blot analysis.
Internalization assays of S. pyogenes in epithelial cells and macrophages Figure 5 Internalization assays of S. pyogenes in epithelial cells and macrophages. RDN29 (wild-type strain), EC478 (isogenic luxS-deficient mutant) and EC490 (RDN29 (pEC85)); RDN02 (wild-type strain), RDN306 (isogenic luxS-deficient mutant) and EC482 (RDN29 (pEC85)). Bacterial cultures were used to infect (A) HEp-2 human pharyngeal epithelial cells and (B) RAW 264.7 macrophages in an antibiotic protection assay. For this, bacterial cultures were grown to mid-logarithmic (RAW 264.7) and late-logarithmic (HEp-2) phases, growth time points showing optimal internalization rates in RAW 264.7 and HEp-2, respectively. The percent intracellular invasion represents the average CFU/ml number of viable intracellular bacteria recovered at different times following addition of antibiotics versus the average CFU/ml number of inoculated bacteria × 100. The average values and standard deviations of three independent assays are shown. The asterisks indicate that the differences in survival rates between the luxS-deficient mutants and the wild-type strains were statistically significant (P < 0.01 for Hep-2 epithelial cells, P < 0.05 for RAW 264.7 macrophages). Similar data were obtained for both M1 and M19 serotypes. Transcriptional expression of virulence-associated genes in RDN02 (wild-type M19 strain) and RDN306 (isogenic luxS-deficient mutant) Figure 6 Transcriptional expression of virulence-associated genes in RDN02 (wild-type M19 strain) and RDN306 (isogenic luxS-deficient mutant). Northern blot analysis was performed with total RNA isolated from strains grown to (1) lag,  , penicillin (100 μg/ml) and streptomycin (100 μg/ml), and grown at 37°C in an atmosphere containing 5% CO 2 . Prior the infection assay, the cells were cultured overnight to semi-confluency (1.5 × 10 5 cells per well for HEp-2 cells and 2 × 10 5 cells per well for RAW264.7 cells) in 24-well tissue culture plates containing medium without antibiotic. GAS strains were grown to the same OD 620 nm corresponding to mid-logarithmic (RAW 264.7) or late-logarithmic (HEp-2) phase, washed with PBS, suspended in fresh cell culture medium and added to the semi-confluent host cell monolayers at a multiplicity of infection (MOI) of 5:1 (HEp-2) or 100:1 (RAW 264.7) in triplicate. Times of incubation were 30 min, 1 h and 2 h (adhesion, HEp-2 cells), 2 h (internalization, HEp-2 cells) and 30 min (RAW 264.7 cells). For adhesion assays (HEp-2 cells), infected monolayers were washed extensively with PBS before lysing. For internalization assays (RAW 264.7 cells), extracellularly adhered bacteria were killed by incubation with fresh medium containing either 100 μg/ml gentamicin and 5 μg/ml penicillin (HEp-2 cells) or 60 μg/ml penicillin (RAW 264.7 cells). At indicated time points, the infected monolayers were washed extensively with PBS to remove the antibiotics. Cells were then lysed with chilled sterile distilled water. The number of viable intracellular bacteria released from the lysed cells was determined by plating appropriate dilutions of the lysates on TSA-blood plates in triplicates followed by 24 h incubation at 37°C. The number of bacteria that survived intracellularly was calculated using the following equation: average number of bacteria recovered (cfu/ml) per well (triplicate) at designated time point/inoculated number of bacteria (cfu/ml)) × 100. Average values were plotted. , isp2 (immunogenic secreted protein homologue), nga (NAD-glycohydrolase), scl (streptococcal collagen-like protein), sibA (secreted immunoglobulin binding protein), sic (streptococcal inhibitor of complement), ska (streptokinase), slo (streptolysin O), speC (streptococcal pyrogenic exotoxin C), speF (mf) (mitogenic factor).