Differential Virulence Gene Expression of Group A Streptococcus Serotype M3 in Response to Co-Culture with Moraxella catarrhalis

Streptococcus pyogenes (group A Streptococcus, GAS) and Moraxella catarrhalis are important colonizers and (opportunistic) pathogens of the human respiratory tract. However, current knowledge regarding colonization and pathogenic potential of these two pathogens is based on work involving single bacterial species, even though the interplay between respiratory bacterial species is increasingly important in niche occupation and the development of disease. Therefore, to further define and understand polymicrobial species interactions, we investigated whether gene expression (and hence virulence potential) of GAS would be affected upon co-culture with M. catarrhalis. For co-culture experiments, GAS and M. catarrhalis were cultured in Todd-Hewitt broth supplemented with 0.2% yeast extract (THY) at 37°C with 5% CO2 aeration. Each strain was grown in triplicate so that triplicate experiments could be performed. Bacterial RNA was isolated, cDNA synthesized, and microarray transcriptome expression analysis performed. We observed significantly increased (≥4-fold) expression for genes playing a role in GAS virulence such as hyaluronan synthase (hasA), streptococcal mitogenic exotoxin Z (smeZ) and IgG endopeptidase (ideS). In contrast, significantly decreased (≥4-fold) expression was observed in genes involved in energy metabolism and in 12 conserved GAS two-component regulatory systems. This study provides the first evidence that M. catarrhalis increases GAS virulence gene expression during co-culture, and again shows the importance of polymicrobial infections in directing bacterial virulence.


Introduction
Streptococcus pyogenes (group A Streptococcus, GAS) is responsible for a wide range of infections including the relatively benign GAS pharyngitis but also life-threatening necrotizing fasciitis and toxic shock syndrome. GAS is also an important cause of post-infectious sequelae such as post-streptococcal glomerulonephritis and rheumatic fever [1,2,3,4]. In contrast, Moraxella catarrhalis is a rather exclusive colonizer and primarily opportunistic pathogen of the human respiratory tract, the bacterium being associated with both upper (e.g. otitis media (OM)) and lower (e.g. exacerbations of chronic obstructive pulmonary disease (COPD)) respiratory tract infections. However, both GAS and M. catarrhalis share the same respiratory biological niche, particularly with respect to upper respiratory tract infections [5]. For example, although Streptococcus pneumoniae, Haemophilus influenzae and M. catarrhalis are considered the predominant organisms causing OM, studies have shown that GAS may also be included as one of the more frequent causative agents of OM [5,6]. This includes the development of OM complications such as acute mastoiditis [7,8]. However, the role of GAS in OM is often underappreciated due to the effectiveness of b-lactam antibiotics in eliminating this organism [9,10].
Co-colonization by bacterial respiratory pathogens may lead to increased rates of colonization and infection. For example, it has been reported that M. catarrhalis and H. influenzae enhance the adherence of S. pyogenes to human epithelial cells [11,12]. Further,  showed that co-colonization of M. catarrhalis and H. influenzae could influence biofilm formation and antibiotic resistance, while at the same time, Verhaegh et al. (2010) showed that co-colonization of infants by M. catarrhalis and H. influenzae was significantly more likely than single species colonization [13,14].
Finally, polymicrobial infections may also promote the survival of bacterial species via the phenomenon of 'indirect pathogenicity'. For example, S. pneumoniae is protected from the action of certain b-lactam antibiotics in the presence of BRO b-lactamase positive M. catarrhalis [15,16].Most of the OM knowledge has been derived from work involving single bacterial species, though there is growing evidence for a role for polymicrobial infection. For example, bacteria and viruses are known to cooperate to cause respiratory diseases which are more severe than those caused by either pathogen alone [17].
Given that M. catarrhalis and GAS are common bacterial OM pathogens, we hypothesized that polymicrobial infections involving these 2 bacterial pathogens would significantly affect the transcription profile of GAS compared to growth of GAS in isolation.

Bacterial Strains and Growth Conditions
The isolates used in this study were a serotype M3 GAS strain MGAS16655 (cultured from a patient with pharyngitis and closely related to the reference strain MGAS315 [18]), and M. catarrhalis strain JMF150 (isolated from a child presenting with OM and provided by Texas Children's Hospital, Houston, United States). For co-culture experiments, both organisms were cultured in Todd-Hewitt broth supplemented with 0.2% yeast extract (THY) at 37uC with 5% CO 2 aeration. Further, each strain was grown in triplicate so that triplicate experiments could be performed. In order to quantify viable GAS and M. catarrhalis, both in monocultures and in co-culture, bacteria were grown in THY broth, and enumerated on the basis of colony forming units (CFU/ml) and OD 600 readings determined at serial time points. RNA was extracted from co-cultures of isolates in the exponential/early-stationary phase of growth after an incubation period of 2.5 hours, with control RNA being extracted from MGAS16655 grown under identical but mono-cultural conditions. Two milliliter of co-culture was added to 4 milliliter of RNA protect (Qiagen) and bacteria were incubated at room temperature for 5 minutes, then pelleted by centrifugation at 40006g at 4uC for 10 minutes, and stored at 280uC until ready for full RNA isolation.

RNA Isolation
Bacterial RNA was isolated from exponential/early-stationary phase bacterial cells using the QIAGEN RNeasy mini kit according to the manufacturer's instructions. RNA quality was analyzed using an Agilent 2100 bioanalyzer (Agilent Technologies Inc., Palo Alto, CA).

cDNA Synthesis, Fragmentation, and Labeling
The methods used for cDNA synthesis, fragmentation, and labeling have been described extensively elsewhere [19].

Microarray Transcriptome Expression Analysis
Microarray expression studies were performed in triplicate using a custom-made Affymetrix GeneChip that contained 100% of the ORFs of the GAS reference strain MGAS315 [20,21,22]. Briefly, samples used for microarray analysis were collected from cultures in THY at 2.5 hours after co-culturing. Controls comprised RNA extracted from GAS grown in pure culture under identical conditions. Estimates of gene expression were calculated using GCOS software (Affymetrix) and normalization of data carried out as previously described [23]. Differences in gene expression that were statistically significant (two-sample t-test; P,0.05) and had a $4-fold change in expression were included in the analysis. As a final step, the complete genome of M. catarrhalis strain RH4 [24] was compared to the genome of GAS strain MGAS315 for regions of sequence similarity. If a gene was found to be similar in both species using a cut off of .95% similarity, then this gene was removed from the analyses due to potential cross-hybridization of GAS and M. catarrhalis RNA transcripts.

Quantitative Real-Time PCR Analysis
Confirmatory quantitative gene transcript analysis was performed on 2 important GAS virulence associated genes, and the constitutively expressed control gene proS, using TaqMan quantitative real-time PCR (qPCR) and an ABI 7500 Fast Real-Time PCR System (Applied Biosystems) as previously described [20,25,26]. Transcripts were compared to the internal reference gene tufA as previously described [27]. Sequence data for the respective TaqMan primers and probes are listed in Table 1. All reactions were performed in quadruplicate using RNA purified from at least three biological replicates.

Results
Characterization of Serotype M3 GAS Strain MGAS16655 and M. catarrhalis JMF150 Growth as Mono-cultures and in Co-culture Prior to characterizing the transcriptome of GAS, we studied the growth of strains MGAS16655 and JMF150 in THY medium in co-culture and as pure cultures. Both strains grew rapidly in this medium, and the bacterial density reached ,10 8 CFU/ml in the exponential/early-stationary phase ( Figure 1). Further, these experiments showed that we were able to reproducibly correlate OD 600 readings to CFU concentrations for both MGAS16655 and JMF150.
Changes in the GAS Transcriptome Induced by Coculture with M. catarrhalis Microarray expression analysis was used to test the hypothesis that GAS gene transcript levels differed significantly during growth in co-culture with M. catarrhalis compared to GAS pure culture  Figure 2) and indicated that the two different growth conditions (co-culture versus mono-culture) generated distinct GAS gene transcription profiles, with multiple virulence and energy metabolism genes being differentially expressed upon co-culture compared to the constitutively expressed control gene proS ( Table 2). GAS transcription expression levels were affected by a factor of at least 4-fold or greater for a total of 207 genes (61 genes with increased expression and 147 genes with decreased expression), after growth in co-culture and relative to the growth of GAS alone. Further, BLAST searching of whole genome sequences revealed that only a small percentage of the M. catarrhalis RNA transcripts were predicted to cross-hybridize with the GAS-specific microarray. When these potentially cross-hybridizing genes (n = 15) were subtracted from the list of statistically significant, $4-fold, GAS expression transcripts, 192 genes remained. The final list comprised 52 genes with significantly increased expression and 140 genes with significantly decreased expression (Figure 3).
Significant down-regulation was observed for many genes of the functional subcategories as shown in Figure 4 (73% of the genes whose expression was modified were down-regulated in co-culture relative to mono-culture), for example genes with products involved in signal transduction tended to be down-regulated. In contrast, genes encoding proteins involved in purine/pyrimidine metabolism were generally up-regulated. A key finding was that a ,180-fold increase and ,80-fold decrease in the expression of hyaluronic synthase (hasA) and streptococcal pyrogenic exotoxin B (speB) was observed, along with a significant decrease in the expression of many genes involved in carbohydrate utilization. Further, gene expression in co-culture resulted in significant differences in several GAS two-component systems (TCS) and response regulatory systems (Table 3), compared to the results observed for GAS pure culture alone.

Validation of DNA Microarray Data
TaqMan quantitative real-time reverse transcription PCR analysis confirmed the validity of the DNA microarray results for 2 representative and significantly up-or down-regulated virulence genes, and the constitutively expressed control gene proS (Table 4). Genes showing significant changes in expression, using a 4-fold cut-off value, for GAS serotype M3 MGAS16655 when grown in co-culture with M. catarrhalis. Results were calculated relative to gene expression values obtained for GAS M3 MGAS16655 grown in pure culture alone. One hundred and ninety-two genes were either 4-fold increased (52 genes) or 4-fold decreased (140 genes) in expression, which corresponds to 10% of the MGAS315 genome. The hasA, ideS, smeZ and speB genes have been associated with GAS virulence. doi:10.1371/journal.pone.0062549.g003

Discussion
The primary goal of this study was to investigate the changes in GAS gene expression in co-culture with M. catarrhalis. We discovered 192 GAS genes to be significantly affected ($4-fold difference). Several of the most significantly up-regulated genes during co-culture were known virulence genes (hyaluronan synthase, streptococcal mitogenic exotoxin Z and IgG endopeptidase), as well as genes involved in purine and pyrimidine metabolism. However, we also observed significant down-regulation in genes involved in energy metabolism and regulatory twocomponent systems. The hyaluronic acid capsule (hyaluronan synthase gene hasA) contributes to resistance to phagocytosis, bacterial aggregation, biofilm maturation, resistance to host defense peptides and neutrophil extracellular killing [28]. GAS also produces a number of highly potent exotoxins that act as superantigens. The streptococcal mitogenic exotoxin Z (SmeZ) is the most potent bacterial superantigen so far discovered and enhances the cascade of pro-inflammatory events that follow invasive streptococcal infection [29,30]. GAS also secretes a highly effective IgG endopeptidase (IdeS) that inhibits phagocytic killing by cleavage of specific IgG [31,32,33]. Finally, though not listed in the top 15 up-and downregulated GAS serotype M3 MGAS16655 genes, the virulence factors streptolysin O (SLO) and streptolysin S (SLS) showed significant differences in expression during co-culture. However, the expression of SLO and SLS are both growth phase dependent; SLO being expressed during the exponential and early stationary-phase of GAS growth, with SLS expression occurring when cells are in the stationaryphase [34,35]. The fact that mRNA was extracted during the exponential/early-stationary phase of co-culture in this manuscript explains why in our study, SLO was upregulated 4.5-fold and the nine-gene locus (sagA-sagI) that is associated with SLS production was downregulated .3-fold. That the transcripts of 12 of the 13 TCS in our study were significantly affected suggests a common signal or significant cross-talk between GAS TCS. TCS are integral in the regulation of the expression of many key virulence factors and serve as a basic stimulus-response coupling mechanism to allow organisms to sense and respond to changes in many different environmental conditions. Examination of whole genome sequence data revealed 13 TCS [36]; however, only 3 GAS TCS have been studied in detail with respect to their involvement in host-pathogen interactions [23,36]. Our data showed down-regulation of the 12 conserved GAS TCS, a finding which when taken in context with the concomitant upregulation of virulence genes, shows that: i) TCS respond to changes in the GAS environment, ii) TCS may be involved in the regulation of GAS pathogenesis, iii) TCS may exert a growth phase-associated control over GAS virulence factors, and iv) GAS TCS may be susceptible to M. catarrhalis regulatory signals [21,37].
Our transcriptomics data are particularly relevant in the context of recent research detailing the effect of quorum sensing on the polymicrobial interactions of other respiratory bacterial pathogens. In particular, a recent study by  showed that co-colonization of M. catarrhalis and H. influenzae (bacterial pathogens sharing the same respiratory niche as GAS) results in increased M. catarrhalis biofilm formation and resistance to antibiotics. This process was achieved via an autoinducer-2 (AI-2, luxS) quorum sensing system [13]. At the same time, Verhaegh Table 3. Fold-change in TCS transcripts in MGAS16655 grown in co-culture with M. catarrhalis and relative to MGAS16655 grown as a pure culture.

Spy no. in strain MGAS315
Fold-change Two-component systems in GAS genomes   [14]. In fact, many species of bacteria use quorum sensing systems to coordinate their gene expression and biofilm formation, dependent on the local bacterial cell density. A homolog of luxS, the genetic determinant for AI-2 production, has been identified in GAS, though a luxS homolog in M. catarrhalis has not been found. However, evidence from whole genome sequencing does suggest that M. catarrhalis possesses an AI-1 luxR sensing system, but no quorum sensing system [38]. Chang et al. (2011) reported that the Rgg family of transcriptional regulators function as quorumsensing effector proteins and comprise the first functional quorumsensing pathway conserved in all group A streptococci [39]. Expression of SpeB is dependent upon this regulator, and is upregulated with increasing cell density and during infection, although Rgg does not appear to be mediator of growth phase regulated control [40]. Currently however, a role for quorum sensing systems in facilitating the observed GAS -M. catarrhalis polymicrobial interaction is still open to question, not least due to the fact that less 'targeted' mechanisms could be responsible for the change in GAS transcriptome profile during co-culture, including competition for nutrients. The 15 genes that showed the most significant decrease in expression during co-culture were mainly involved in energy metabolism, with an emphasis on galactose metabolism. Nutrient availability is a signal by which pathogens sense their external environment and to which they respond through regulated production of various virulence factors. Studies have shown that co-cultured bacteria compete for nutrients and that interactions between co-cultured strains also involve species-specific pH limits for growth and differential utilization of growth substrates [41,42]. The success of GAS, which thrives in diverse host niches, depends on the acquisition of nutrients from very different sources and GAS virulence gene expression is highly responsive to carbohydrate source and availability; however, the pathways linking metabolism and virulence remain poorly understood [20,43]. Our findings could also be associated with a GAS transcriptional response to depletion of carbohydrate (and possible other nutrients) within the polymicrobial environment, thereby helping GAS to conserve its 'energy reserves', and ultimately facilitate entry into a sessile 'state of rest' (associated with biofilm formation) [44]. However, it should be noted that M. catarrhalis is biochemically a-saccharolytic, meaning that it is unable to metabolize glucose or other carbohydrates [45]. Therefore, the true relationship between M. catarrhalis and GAS co-culture, carbohydrate depletion, and the down-regulation of genes involved in carbohydrate utilization remains to be further elucidated.
One gene that showed a significant decrease in expression during co-culture, and is not involved in energy metabolism, was salA (salivaricin A). Salivaricin A is an antimicrobial peptide which is a member of the class of antimicrobial peptides called lantibiotics. Lantibiotics are produced by a large number of Gram-positive bacteria such as Streptococcus and Streptomyces to attack other Gram-positive bacteria. A study by Upton et al. (2001) described intra-and interspecies signaling between Streptococcus salivarius and S. pyogenes, mediated by SalA [46]. A single bacteriocin (McbC) has been described in M. catarrhalis, though the effect of this bacteriocin on GAS growth and virulence is not yet known [47].
Currently, watchful waiting and/or the prescription of antibiotics are the most favored options for the treatment of organisms causing OM disease. Treatment with antibiotics seems effective, but is not favorable due to the development of antibiotic resistance. Targeting bacterial communication (quorum signaling) could be a novel treatment methodology, as there is growing evidence for a role for polymicrobial infection in de development of OM. For example, M. catarrhalis and GAS together could generate a more severe infection than M. catarrhalis or GAS alone. Further, M. catarrhalis outer membrane vesicles (OMVs) provide protection to S. pneumoniae and H. influenzae by carrying active blactamase [48]. Many bacteria rely on quorum signaling molecules to control the expression of virulence and regulatory genes, which ultimately influence the establishment and progress of disease. However, the interspecies targeting of quorum signaling molecules may result in undesirable complications, such as the promotion of virulence or virulence-related phenotypes in opportunistic pathogens that share the same biological niche as S. pyogenes [49].

Conclusion
The transcriptome profile of GAS is markedly altered in response to co-culture with M. catarrhalis, with genes involved in virulence (up-regulated) and carbohydrate utilization (downregulated) being especially affected. Though the study findings are limited due the use of one strain per species and the lack of protein verification assays, our findings are a first step in understanding and elucidating the mechanisms that help facilitate colonization and disease during polymicrobial infections with GAS and M. catarrhalis.