Intracellular signaling through the comRS system in Streptococcus mutans genetic competence

Entry into genetic competence in streptococci is controlled by ComX, an alternative sigma factor for genes that enable the import of exogenous DNA. In Streptococcus mutans, the immediate activator of comX is the ComRS signaling system, which consists of the cytosolic receptor ComR and the 7-residue signal peptide XIP, which is derived from ComS. Extracellular XIP imported by an oligopeptide permease interacts with ComR to form a transcriptional activator for both comX and comS. Therefore, extracellular XIP can function as an exogenous signal to trigger S. mutans competence. However, the mechanisms that process ComS and export it as XIP are not fully known in S. mutans. The observation that comX is expressed bimodally under some environmental conditions suggests that ComR may also interact with endogenously produced XIP or ComS, creating an intracellular positive feedback loop in comS transcription. Here we use single cell and microfluidic methods to compare the effects of the native comS gene and extracellular XIP on comX expression. We find that deletion of comS reduces the response of comX to extracellular XIP. We also find that comS-overexpressing cells autoactivate their comX even when their growth medium is rapidly exchanged, although this autoactivation requires an intact copy of comS under control of its own promoter. However comS-overexpressing cells do not activate comS-deficient mutants growing in coculture. These data show that individual cells can activate comX without exporting or importing the XIP or ComS signal, and that endogenously and exogenously produced ComS/XIP have inequivalent effects on comX behavior. These data are fully consistent with a model in which intracellular positive feedback in comS transcription plays a role in ComRS signaling, and is responsible for the bimodal expression of comX. Author Summary Heterogeneous gene expression in genetically identical populations plays an important role in bacterial persistence and survival under changing environmental conditions. In the oral pathogen Streptococcus mutans, the physiological state of genetic competence can exhibit bimodality, with only some cells becoming competent. S. mutans controls its entry into competence by using the ComRS signaling system to activate comX, a gene encoding the master competence regulator ComX. The ComRS system is understood as a quorum sensing system, in which the extracellular accumulation of the small signal peptide XIP, derived from ComS, induces comX expression. We coupled observation of bacteria that fluoresce when comX is active with mathematical analysis and chemical binding assays to show that activation of comX does not necessarily require extracellular XIP or ComS, and that comX-active cells do not necessarily export XIP. Our experiments and mathematical modeling indicate that a positive feedback loop in comS transcription allows a cell to activate comX in response to its own XIP or ComS in the absence of extracellular XIP, or to amplify its comX response to extracellular XIP if present. Such positive feedback loops are often the cause of bimodal gene expression like that seen in S. mutans competence.


Introduction
cells that carry opp. Activation by XIP does not require comS [20]. Therefore, although 113 exogenous XIP can compensate for a comS deletion and activate comX, the bimodal 114 comX response still requires an intact comS gene. 115 The observation that competence in several streptococcal species is directly 116 stimulated by an extracellular ComS-derived peptide suggests that ComRS constitutes 117 a novel type of Gram positive quorum signaling system, in which the ComS-derived 118 signal XIP is processed and secreted, accumulates in the extracellular medium, and is 119 then reimported. This interpretation in S. mutans is supported by several experimental 120 observations. First, cells that carry opp take up exogenous XIP (in defined medium) and  suggests that XIP in S. mutans lacks its own exporter and is released from the cells 142 primarily through lysis. 143 The import of XIP presents an additional puzzle for ComRS quorum signaling, as 144 the permease OppA is not required for exogenous CSP to activate comX, but is 145 required for XIP to activate comX. It has been suggested that bacteriocin production 146 induced by CSP may create another entry route for extracellular XIP by increasing 147 membrane permeability [12]. However, such a model implies, contrary to data [11], that 148 CSP should also induce comX in defined growth media. In addition this permeability  Therefore the deletion of comS both elevated the threshold for a response to 180 extracellular XIP and reduced the overall response at saturation.

181
The deletion of comS also affected cell-to-cell variability (noise) in comX 182 expression. Figs. 1C and 1D show that the histograms of reporter fluorescence differ in 183 wild-type and ΔcomS cells. Wild type showed a generally broader (noisier) comX 184 response than did ΔcomS. We quantified this difference by fitting the histograms to a 185 gamma distribution Γ(n | a,b), a two-parameter probability distribution that can be used  were supplied with fresh complex medium flowing at rates between 0.02 ml h -1 and 1 ml 209 h -1 . These flow rates were sufficient to completely replace the growth medium within 210 each chamber on time intervals ranging from 6 seconds to 10 minutes. We also studied 211 (1) a PcomX-rfp reporter in a UA159 background (negative control), and (2) a 184comS 212 overproducing strain that was lacking a start codon (ATG point-mutated to AAG) on its 213 chromosomal comS (184comS PcomX-rfp ΔcomS).    Growth phase-dependent release of XIP 289 We previously showed that intercellular signaling by S. mutans ComRS is 290 impeded by deletion of the atlA gene, encoding a major autolysin. Loss of AtlA inhibits 291 cell lysis, which appears to occur primarily in stationary phase [26]. We therefore tested 292 whether signaling from a sender (comS overexpressing) strain to a receiver (ΔcomS) 293 strain is enhanced in the latter phases of growth. We prepared sender/receiver co-294 cultures in different ratios in defined medium. Every 2 h the pH of the cultures was 295 adjusted to 7.0 by addition of 2 N NaOH, the OD600 was recorded, and an aliquot of the 296 culture was collected for fluorescence imaging of the comX promoter activity. Low pH 297 suppresses the comX response, so the pH adjustment ensures that cells are able to 298 respond to comX-activating signals when present [9, 13]. The GFP fluorescence 299 histograms of Fig. 5A show that comX expression in the ΔcomS strain is slightly higher 300 at 12 h than at 2 h or 8 h. This increase was more pronounced at higher ratios of sender 301 to receiver cells, consistent with increased release of XIP from the senders in late      Our observations suggest more strongly that export or import of ComS or XIP is 458 not essential to ComRS control of comX in S. mutans, under the conditions examined.

459
The key role of native control of comS in our data argues that the more essential      Table S1 (see supporting information).

608
Flow rate dependence experiment 609 In order to measure the flow rate dependence of XIP signaling we loaded cells 40x into fresh FMC containing erythromycin (10 µg ml -1 ) and spectinomycin (1 mg ml -1 ).

646
Once grown to OD600 0.05, these were mixed in ratios varying from to 0% comS 647 overexpressers to 100% overexpressers, defined by volume of comS overproducers 648 added divided by the volume of the ΔcomS culture added. Low initial cell densities were 649 used to ensure that early, mid and late growth phases were probed for XIP release.

650
Every two hours the OD600 of the culture and its pH were measured. The pH was 651 corrected back to 7.0 using 2N sodium hydroxide if it had deviated below 6.5, in order to 652 measure reaction to any XIP released at late times into the culture. RFP and GFP   6A) and competition ( Figure 6B) experiments for a given peptide P. 730 In general the FP data are compatible with a range of parameter values. If n is 731 constrained to be less than 2.5 then optimal values are in the range k1 ∼ 1-6 μM and k2 732 ∼