The Quorum Sensing-Controlled Competence Regulon Drives H₂O₂ Production in Streptococcus gordonii

Abstract

Streptococcus gordonii sp. firmicutes is an early colonizer of the oral microbiome and contributes positively to oral health. While this species has been found to produce hydrogen peroxide by spxB expression, the relationship of this expression to the competence regulon has not yet been explored. To this end, this study sought to investigate the connection of the S. gordonii competence regulon quorum sensing (QS) circuitry with downstream proliferative phenotypic expression as a result of competence stimulating peptide (CSP) exposure, with specific attention to peroxide formation. Following confirmation of the native CSP, RNA-seq was completed to gain insights into transcriptomic variations resulting from CSP incubation. Later, structure-activity relationship (SAR) analyses of the native CSP were completed. The results revealed residues integral to CSP::ComD binding and activation, while indicating which residues were considered dispensable to this process. Phenotypic assessment revealed that peroxide formation was modulated via the competence regulon. Finally, interspecies competition assays were carried out to understand the interactions between S. gordonii and S. mutans, with S. gordonii demonstrating the profound capability of antagonizing S. mutans growth and proliferation. Our results support that this antagonism is mainly attributed to hydrogen peroxide production by S. gordonii. This finding suggests that S. gordonii may be exploited for its beneficial proliferative phenotypes downstream of the competence regulon.

Graphical Abstract

Quorum sensing (QS) is a method of bacterial cell-cell communication that allows for the regulation of gene transcription and proliferative phenotypic expression.^1–2 Both Gram-positive and Gram-negative bacteria engage in QS through the production of signaling molecules, referred to as autoinducers. Following biosynthesis of these molecules, they are secreted into the extracellular environment and eventually reach a critical concentration threshold.^3–4 Once this is accomplished, the autoinducer goes on to bind to receptor proteins, which induces the production of more autoinducer molecules.² The establishment of this positive feedback loop and the subsequent accumulation of autoinducers in the extracellular environment helps bacterial populations gauge population density, thereby allowing for population-wide modulation of gene expression and group behavior profiles.⁵

Numerous studies have suggested that bacteria in the Streptococcus genus engage in QS.^6–9 Autoinducing peptides (AIPs) have been found to be the most common signaling molecules employed in Gram-positive QS systems.¹⁰ In streptococci, these AIPs are commonly post-translationally modified oligopeptides referred to as competence stimulating peptides (CSP) due to their role in inducing competence.¹¹ In addition to inducing competence, CSPs and the associated comABCDE regulon have been shown to influence other downstream proliferative phenotypic expression profiles. For example, QS has been linked to biofilm formation, competence induction, and virulence factor production, all of which can incite cumulative, systemic health effects.^{4, 7, 9, 12–13} Additionally, the competence regulon has been shown to influence the production of hydrogen peroxide in S. sanguinis, S. oligofermentans, and S. cristatus, an important colonization strategy for early inhabitants of the oral microbiome.^13–15

Streptococcus gordonii is one such early colonizer in the oral cavity that can be also found throughout the human body in locations such as the skin, upper respiratory tract, and intestines.^16–17 In the oral cavity, S. gordonii has demonstrated the ability to contribute to oral health through various antimicrobial mechanisms, such as modulating the biofilm formation of other species.¹⁸ While S. gordonii has been widely regarded as a commensal, it can also act as an opportunistic pathogen when introduced to the bloodstream. Associated diseases such as endocarditis, periodontitis, and cholecystitis have been reported on rare occasions.^{16, 19–20}

There is evidence that S. gordonii makes use of differential gene expression to engage in environmental sensing mechanisms. This affords S. gordonii the ability to adapt to continually changing environmental conditions and dynamic population compositions experienced during biofilm formation.²¹ Understanding which genes are upregulated during these interspecies interactions can reveal insights to help control biofilm development and prevent the progression of related human diseases, such as periodontitis and dental caries.²¹ Additionally, other studies have demonstrated that S. gordonii is capable of producing hydrogen peroxide from pyruvate via pyruvate oxidase (spxB).^22–23 Studies have shown that S. gordonii peroxide formation antagonizes S. mutans.^{18, 22} This phenotypic capability could provide a potential biotherapeutic method for treating S. mutans driven infections.

Overall, the role of QS in regulating S. gordonii phenotypes is not well known. Following the recent characterization of other Streptococci QS systems, we hypothesize that the CSP aids in the regulation of phenotypes that govern documented intraspecies and interspecies interactions. Prior work identified the S. gordonii sp. firmicutes CSP as the 19-mer DIRHRINNSIWRDIFLKRK, which was found to be responsible for regulating the competence regulon.²⁴ Like many other streptococcal species, the comABCDE QS cascade begins when ComC is processed and secreted by ComAB, a transmembrane ATP-binding cassette (ABC) transporter protein (Figure 1).²⁵ Once in the extracellular space, the now mature CSP can circulate throughout the microenvironment and is able to bind a histidine kinase receptor, ComD, of neighboring bacterial cells of the same or closely related species.^26–28 ComD phosphorylates an internal response regulator, ComE, which then upregulates the transcription and expression of various genes, including an alternate sigma factor, comX, known to be correlated with downstream phenotypic expression, as well as the competence regulon as a whole.^{26, 29}

Figure 1.

S. gordonii comABCDE regulon. The competence regulon relies on the production and reception of CSP in a concentration dependent manner. This peptide signal is encoded as a propeptide by comC, which is then processed and secreted by ComAB. Once the peptide has reached a critical concentration threshold, CSP binds ComD and in turn ComD phosphorylates ComE. ComE goes on to bind to conserved promoted sequences and upregulates the transcription of the entire competence regulon, including comX, an alternate sigma factor related to downstream phenotypes.

While this cycle has not been extensively reviewed in S. gordonii, previous literature suggests that properties exhibited by the S. gordonii QS cascade resemble that of S. pneumoniae.³⁰ This includes conserved chromosomal location in the QS cascade region, similar transformational frequencies based on CSP, and similar competence induction mechanisms.^30–31 It is important to note that the possible probiotic abilities of these two species vary drastically. While S. pneumoniae has been widely regarded as a notorious pathogen, S. gordonii has instead demonstrated the potential to promote oral health.³² Due to its probiotic potential, further investigation of the S. gordonii QS cascade is of significant importance.

In this work, we sought to gain a greater understanding of S. gordonii QS-related phenotypes with emphasis on elucidating potential beneficial probiotic effects. To begin this investigation, the naturally produced and secreted S. gordonii CSP, previously suggested to be DIRHRINNSIWRDIFLKRK, was isolated from cell-free supernatants and verified using high-resolution mass spectrometry and MS/MS analysis. The wild-type S. gordonii system was then incubated with the verified CSP sequence, after which transcriptomic analyses were completed to reveal genes that experienced either an increase or decrease in transcript abundance as a direct result of CSP exposure. Later, the CSP sequence underwent a comprehensive structure-activity relationship analysis to reveal critical structural motifs that facilitate CSP::ComD binding and activation. Following this, several phenotypic assays were utilized to further understand the role of the competence regulon in biofilm formation, peroxide formation, and interspecies interactions. Results from this study revealed key structural insights about the S. gordonii CSP, while also highlighting antagonistic interactions between S. gordonii and S. mutans. Overall, data from this study have great potential to aid in the future development of QS-based methods to combat pathogenic streptococcal species.

Results and Discussion

Isolation and Verification of S. gordonii Competence-Stimulating Peptide

This investigation began with the confirmation of the competence stimulating peptide sequence. Previous studies investigating S. gordonii have reported the CSP sequence as DIRHRINNSIWRDIFLKRK, however it has never been isolated from cell-free supernatants and analyzed by mass spectrometry.²⁴ Therefore, cultures of wild-type S. gordonii were grown to early log-phase growth (OD₆₀₀ of approximately 0.200) after which the culture was spiked with a peptide analog with a similar sequence to induce the upregulation of competence-regulon-associated genes, including comC. Prior research has indicated that the C-termini of CSPs are generally tolerant to modification.^33–34 Consequently, synthetic peptide CSP-desK19 was used to spike the culture in order to maintain similar ComD activation while being distinguishable from the native CSP by mass. This resulted in transcriptional upregulation of comC, allowing the processed and mature peptide to accumulate in the extracellular environment.

Crude peptide mixtures were precipitated from the cell-free supernatants of wild-type S. gordonii cultures, after which they were purified via reverse-phase high performance liquid chromatography (RP-HPLC). Individual fractions were analyzed using MALDI-TOF to ascertain the mass of the putative CSP sequence. Following identification of fractions containing the desired mass, the isolated secreted peptide was analyzed with high resolution mass spectrometry (HRMS) and tandem mass-spectrometry (MS/MS). The observed mass of the peptide was within 5 ppm of the predicted 19-mer CSP confirming that the previously determined sequence was in fact the mature CSP sequence.

To further validate the CSP sequence, we sought to confirm the connectivity of the peptide by reproducing a synthetic 19-mer and running parallel analyses of the endogenous CSP and synthetic mimic. To this end, the exact CSP sequence was generated using Fmoc solid-phase peptide synthesis (SPPS) on Wang resin and purified by reversed-phase high performance liquid chromatography (RP-HPLC). The synthetic peptide and the endogenously produced CSP were subjected to the same analytical RP-HPLC conditions (Figure 2). Resultant traces for each peptide exhibited the same retention times, supporting that both peptides were the same sequence. Furthermore, when combined into a single fraction, these samples experienced a single peak with the same retention time as the individual peptides, all of which further reinforced the CSP sequence and connectivity. Lastly, the secreted CSP was subjected to MS/MS analysis to cross-check that sequence of the naturally produced CSP corresponded with the confirmed sequence (Figures S-2 and S-3, Table S-3). These results again demonstrated that the S. gordonii sp. firmicutes CSP does not undergo additional post-translational processing and is categorically DIRHRINNSIWRDIFLKRK.

Figure 2. Comparison of purified synthetic CSP with extracted CSP.

A. The detected masses of the extracted and synthetic S. gordonii CSP were within 5 ppm of the literature mass value. B. Analytical HPLC traces of the purified extracted CSP, synthetic CSP, and extracted & synthetic CSPs demonstrated the same retention times.

Transcriptomic Analyses

RNA-sequencing was completed to understand the regulatory role of the S. gordonii CSP on the competence regulon and associated transcriptional effects. To this end, S. gordonii cultures were grown to early exponential phase and treated with exogenous CSP to induce competence regulon activation prior to the natural induction by endogenous CSP production. Preliminary qPCR data evaluating comX expression indicated enhanced expression of competence-related genes 10 minutes post-treatment (data not shown). Additionally, a previous study on S. mitis by Milly and coworkers demonstrated a similar peak expression of comE and comX at 10-minutes post-CSP incubation.³⁵ As such, cultures of wild-type S. gordonii were incubated with either 10,000 nM CSP or DMSO (no treatment) for 10 minutes to assess the effects of CSP exposure on resultant transcript levels. Following the 10-minute incubation period, total RNA was isolated and prepared for sequencing.

Differential gene expression analysis revealed numerous genes associated with early competence induction that experienced an increase in transcript abundance and thus significant transcriptional upregulation (Table 1). Genes belonging to the comABCDE regulon were upregulated, as well as additional early competence genes belonging to the ComG operon. ComG-associated genes were previously explored in Bacillus subtilis and have been found in other Gram-positive species.^36–37 They have demonstrated involvement in bacterial competence induction, specifically selective DNA binding and uptake mechanisms.³⁸ Several other genes demonstrated significant upregulation, however there was no clear correlation between these expression profiles (Table S-4, Figure S-4).

Table 1. RNA-Seq: Streptococcus gordonii CSP vs. control – Differentially expressed genes (upregulation).

RNA-seq results revealed numerous genes associated with competence that experienced an increase in overall transcript abundance, and thus an upregulation of transcription, following 10-minute CSP incubation.

Name	Fold Change	FDR p-value	Product
I6L84_RS00235	3.418036861	3.45176E-06	Competence stimulating peptide ComC
I6L84_RS00320	4.892582222	1.1159E-10	ComX1; sigma-70 family RNA polymerase sigma factor
I6L84_RS00640	4.127334592	2.05508E-21	DUF3021 family protein
I6L84_RS00645	4.738027184	2.93508E-22	LytTR family DNA-binding domain-containing protein
I6L84_RS00685	2.497157655	0.000221561	Competence/damage-inducible protein A
I6L84_RS06095	3.936115744	1.82444E-06	Competence protein CoiA
I6L84_RS06225	3.415598462	0.000427597	DNA internalization-related competence protein ComEC/Rec2
I6L84_RS06715	6.731564872	9.57752E-12	ComX2; sigma-70 family RNA polymerase sigma factor
I6L84_RS07985	3.625044004	1.84755E-05	ComF family protein
I6L84_RS07990	4.161890607	1.08847E-05	DEAD/DEAH box helicase
comA	5.503982371	7.98497E-09	Peptide cleavage/export ABC transporter ComA
comB	5.31885301	1.50338E-11	Competence pheromone export protein ComB
comD	2.937410526	2.56172E-05	Competence system sensor histidine kinase ComD
comE	2.983684565	1.16162E-05	Competence system response regulator transcription factor ComE
comGA	2.873095256	0.001225261	Competence type IV pilus ATPase ComGA
comGB	2.70402739	0.003247555	Competence type IV pilus assembly protein ComGB
comGC	2.797708353	0.000152485	Competence type IV pilus major pilin ComGC
comGD	2.587389067	0.002043326	Competence type IV pilus minor pilin ComGD
comGE	2.640801908	0.001510279	Competence type IV pilus minor pilin ComGE
comGF	2.340263649	0.004605839	Competence type IV pilus minor pilin ComGF
dprA	3.835018274	5.47242E-06	DNA-processing protein DprA
holA	2.338884743	0.000100239	DNA polymerase III subunit delta
radC	3.258528536	3.72474E-05	DNA repair protein RadC
recX	2.063843094	8.91693E-05	Recombination regulator RecX
topA	2.500811771	0.000427597	Type I DNA topoisomerase

Interestingly, S. gordonii did not demonstrate any significantly downregulated genes. It is possible that the 10-minute incubation time did not allow for extensive altered genetic expression, as much of the cellular effort was directed at upregulation of competence-related genes. If an increased incubation period is allotted, downregulation of various genes may be visualized and explored to a greater extent. Also noteworthy was that transcription of genes associated with proliferative phenotypic expression, such as biofilm formation, virulence factor formation, or peroxide production, remained unchanged. However, as some of these phenotypes take longer to develop, these results do not rule out transcriptional changes of genes associated with known QS-related phenotypes.

Development of the S. gordonii Luciferase-Based Reporter Strain

Next, we focused our efforts on developing a luciferase-based reporter system in order to aid in the visualization and quantification of CSP::ComD binding and activation. With the promoter for comX annealed just prior to the gene for luciferase, any subsequent activation of comX would result in the transcription and translation of luciferase. This enzyme cleaves luciferin, resulting in a detectable luminescent signal directly proportional to comX activation. The accurate construction of the reporter plasmid was verified via sequencing. Adequate and selective luminescence development was assessed by incubating either wild-type cultures or the designed reported strain with 15 μg/mL D-luciferin and either 10,000 nM CSP or DMSO (no treatment). Significant luminescence was observed when CSP was incubated with the reporter system, no luminescence was seen with wild-type (CSP or DMSO), and minimal luminescence was observed for reporter cultures treated with DMSO (representing endogenous CSP production). In combination, these results validated the design and construction of the reporter (Figure 3).

Figure 3. Time-resolved luminescence assays.

(A) A culture of wild-type S. gordonii incubated with 10,000 nM synthetic CSP exhibits no luminescence (red) as OD increases (black). (B) The constructed S. gordonii luminescent reporter incubated with DMSO. (C) The constructed S. gordonii luminescent reporter incubated with 10,000 nM synthetic CSP. (D) The constructed S. gordonii luminescent reporter incubated with 1,500 nM synthetic CSP. Overnight cultures were diluted 1:20 to achieve early logarithmic phase growth, after which cultures were diluted again 1:150 in fresh THY. See Supporting Information for full experimental details.

Temporal Identification of CSP Production

The previously constructed luminescence-based reporter system was initially utilized to investigate the endogenous mode of CSP production, as well as gain a deeper understanding of the lifecycle of S. gordonii. Reporter cultures were incubated with 10,000 nM CSP (complete inundation), 1,500 nM CSP (~5 × EC₅₀, as determined by using this reporter system to quantify the potency of the native CSP and presented in Table 2 below), or DMSO (no treatment) and monitored for luminescence production. Wild-type cultures experienced no instances of elevated luminescence, with or without the exogenous addition of synthetic CSP (Figure 3A). Contrarily, when reporter cultures were incubated with DMSO, a small but noticeable spike in luminescence was visualized, generating maximal expression throughout early log phase growth (Figure 3B). This signal was representative of endogenous CSP production, as the naturally produced peptide was able to induce the competence regulon, and the reporter exhibited a slight increase in luminescent signal as a result. Furthermore, when reporter cultures were incubated with either concentration of CSP, a similarly timed, but prominently increased luminescent signal was observed at an OD₆₀₀ of approximately 0.240. Maximal expression was generated throughout early log phase growth before resolving completely just prior to the onset of stationary phase (Figure 3C & D) due to competence shutoff, a known phenomenon in streptococci.^39–40 This luminescent output aligns closely with the present understanding of early competence induction. As QS-controlled competence induction is a population size-dependent process, adequate numbers of bacterial cells are reached in early log phase growth, thus allowing for the appropriate coordination of group behaviors.⁴¹ Overall, these results demonstrated that endogenous CSP production and secretion can be detected early in the S. gordonii life cycle, and further confirmed the effectiveness of luminescence as a quantifiable marker of an operable, responsive reporter system.

Table 2. Alanine scan of S. gordonii CSP.^a.

This scan demonstrates the systematic replacement of individual residues with alanine, thus revealing the importance of specific side-chain residues in the binding and/or activation of ComD.

#	Peptide Name	Sequence	EC50 or IC50* (nM)b	95% CIc
1	CSP	DIRHRINNSIWRDIFLKRK	282	249 – 319
2	D1A	AIRHRINNSIWRDIFLKRK	> 1,000*	-
3	I2A	DARHRINNSIWRDIFLKRK	-d	-
4	R3A	DIAHRINNSIWRDIFLKRK	-d	-
5	H4A	DIRARINNSIWRDIFLKRK	722	528 – 988
6	R5A	DIRHAINNSIWRDIFLKRK	> 1,000	-
7	I6A	DIRHRANNSIWRDIFLKRK	> 1,000	-
8	N7A	DIRHRIANSIWRDIFLKRK	203	144 – 285
9	N8A	DIRHRINASIWRDIFLKRK	692	591 – 950
10	S9A	DIRHRINNAIWRDIFLKRK	-d	-
11	I10A	DIRHRINNSAWRDIFLKRK	> 1,000	-
12	W11A	DIRHRINNSIARDIFLKRK	-d	-
13	R12A	DIRHRINNSIWADIFLKRK	127	87.8 – 183
14	D13A	DIRHRINNSIWRAIFLKRK	-d	-
15	I14A	DIRHRINNSIWRDAFLKRK	> 1,000	-
16	F15A	DIRHRINNSIWRDIALKRK	> 1,000	-
17	L16A	DIRHRINNSIWRDIFAKRK	127	110 – 146
18	K17A	DIRHRINNSIWRDIFLARK	426	399 – 454
19	R18A	DIRHRINNSIWRDIFLKAK	406	334 – 495
20	K19A	DIRHRINNSIWRDIFLKRA	272	260 – 285

^aSee the Materials and Methods section or the Supporting Information for experimental methods. See the Supporting Information for details of the reporter strain and plots of agonism or antagonism dose-response curves. All assays performed in triplicate.

^bEC₅₀ (half-maximal effective concentration) or IC₅₀ (half-maximal inhibitor concentration) values determined by testing peptides over a wide range of concentrations.

^c95% confidence interval.

^dEC₅₀ not determined due to the analogue’s low induction in primary agonism screening assays. See the Supporting Information for details.

Systematic Modification of the S. gordonii CSP

Generally, streptococcal CSP sequences can be broken down into three principal regions: the N-terminus, core region, and C-terminus. Several studies have indicated that there is an enhanced degree of sequence conservation at the N-terminus, as this portion of the CSP is critical for ComD binding and activation.^42–43 Contrarily, there is a great deal of evidence supporting enhanced tolerance to modification at the C-terminus.^{14, 35, 44} This region has largely been associated with hydrophilic, charged residues that aid in improved CSP solubility properties, rather than playing an active role in ComD binding and activation.^33–34

To determine if a similar pattern of CSP::ComD structure-activity relationships (SAR) exist in S. gordonii, peptide libraries were designed to interrogate the CSP sequence and reveal insights critical for ComD binding and activation. An alanine scan, D-amino acid scan, and truncation scan were synthesized. These aided in understanding the importance of specific side chain residues, the significance of side chain orientation, and the minimal scaffold needed for activity, respectively. All peptide analogs were synthesized by solid-phase peptide synthesis (SPPS), purified by RP-HPLC and analyzed by high-resolution MS (Table S-5 through Table S-7). After preparing these peptide libraries, SAR analyses were used to delineate the complex interactions that take place during CSP::ComD binding and activation. Initial activation screenings, or agonism and antagonism assays, were performed at a high peptide concentration (10,000 nM) for all the S. gordonii analogs (Figure S-5 through Figure S-10) to demarcate their function as potential ComD activators or inhibitors.

Structure Activity Relationship Analysis

Dose response curve results from the alanine scan revealed structural insights consistent with previous knowledge of other streptococcal CSPs (Table 2). The N-terminus was not tolerant to modification, with changes of the first three residues resulting in the complete decimation of activity. Studies of S. pneumoniae, S. mitis, S. oligofermentans, and others demonstrated that alteration to the negatively-charged residue at positions 1, Asp or Glu, resulted in analogs with inhibitory activity.^{7, 9, 14} Similarly, S. gordonii CSP-D1A produced an inhibitory analog, further supporting the highly conserved role and imperative presence of a negatively-charged residue at the N-terminus for ComD activation.

Modification to several other residues, including Arg5, Ile6, Ser9, Ile10, Trp11, Asp13, Ile14, and Phe15, resulted in the complete loss of activity, however none demonstrated ComD inhibition. Results suggest that these residues are relevant for ComD binding, but do not play a role in receptor activation. Analogs CSP-H4A, -N7A, -N8A, -K17A, -R18A, and -K19A demonstrated either no change or a roughly two-fold decrease in activity, suggesting these residues are permissible to change and do not play critical roles in ComD binding or activation.

Interestingly, CSP-R12A and CSP-L16A resulted in analogs with activity enhanced by roughly two-fold compared to the native CSP. This suggests that the native residues at their respective positions may not experience optimized binding with ComD and potentially prevent optimal activation. Removal of Arg12, a relatively bulky and basic side chain, as well as removal of Leu16, a nonpolar and hydrophobic residue, improved ComD activity. This suggests these positions could be further interrogated to potentially reveal additional analogs with enhanced activity compared to the native CSP. For example, modifying these positions to include residues of varying charges or sizes, or generating double-mutation analogs, could lead to CSP variants with more potent EC₅₀ values.

A D-amino acid scan was completed to study the effects of epimerization at individual side chain residues (Table 3). Similar to the alanine scan, the N-terminus was not tolerant to modification, with modification to several of the N-terminal residues resulting in complete loss of activity. In this region, two analogs, CSP-d1 and i2 resulted in analogs with inhibitory activity, albeit at high peptide concentration, again reinforcing their role in ComD activation. Additionally, modification to Ile6 resulted in an analog with three-fold improved activity, suggesting that epimerization at this position actually improves binding.

Table 3. D-amino acid scan of S. gordonii CSP.^a.

This scan demonstrates the systematic replacement of individual side chains with the dextrorotatory counterpart, thus revealing the importance of bond angle orientation of specific side-chain residues in the binding and/or activation of ComD.

#	Peptide Name	Sequence	EC50 or IC50* (nM)b	95% CIc
21	d1	dIRHRINNSIWRDIFLKRK	> 1,000*	-
22	i2	DiRHRINNSIWRDIFLKRK	> 1,000*	-
23	r3	DIrHRINNSIWRDIFLKRK	-d	-
24	h4	DIRhRINNSIWRDIFLKRK	-d	-
25	r5	DIRHrINNSIWRDIFLKRK	-d	-
26	i6	DIRHRiNNSIWRDIFLKRK	103	79.2 – 133
27	n7	DIRHRInNSIWRDIFLKRK	-d	-
28	n8	DIRHRINnSIWRDIFLKRK	-d	-
29	s9	DIRHRINNsIWRDIFLKRK	646	575 – 725
30	i10	DIRHRINNSiWRDIFLKRK	959	846 – 1,090
31	w11	DIRHRINNSIwRDIFLKRK	> 1,000	-
32	r12	DIRHRINNSIWrDIFLKRK	963	828 – 1,120
33	d13	DIRHRINNSIWRdIFLKRK	664	534 – 828
34	i14	DIRHRINNSIWRDiFLKRK	461	414 – 514
35	f15	DIRHRINNSIWRDIfLKRK	753	666 – 852
36	l16	DIRHRINNSIWRDIFlKRK	345	281 – 423
37	k17	DIRHRINNSIWRDIFLkRK	453	387 – 531
38	r18	DIRHRINNSIWRDIFLKrK	-d	-
39	k19	DIRHRINNSIWRDIFLKRk	386	278 – 536

^aSee the Materials and Methods section or the Supporting Information for experimental methods. See the Supporting Information for details of the reporter strain and plots of agonism or antagonism dose-response curves. All assays performed in triplicate.

^bEC₅₀ or IC₅₀ values determined by testing peptides over a wide range of concentrations.

^c95% confidence interval.

^dEC₅₀ not determined due to the analogue’s low induction in primary agonism screening assays. See the Supporting Information for details.

Epimerization of amino acids in the core region, including Asn7, Asn8, Ile10, Trp11, and Arg12, resulted in analogs with substantially diminished activity. This suggests that the naturally occurring L-orientation of these stereogenic centers is necessary for appropriate receptor binding. A similar trend was exhibited at C-terminal Arg18, which was peculiar, as the C-terminus has widely been regarded as dispensable to ComD activity. Instead, it is plausible that alteration to this residue disrupts secondary structure formation enough to impact binding, rather than playing a direct role in ComD interactions. Strikingly, epimerization of all the other residues (CSP-S9, -D13, -I14, -F15, -L16, -K17, -K19) demonstrated EC₅₀ values with similar or slightly diminished activities compared to that of the CSP, suggesting the stereogenic center of these residues are generally tolerant of change. As such, subsequent epimerization at these positions is inconsequential for appropriate ComD binding and activation.

Lastly, analysis of a truncation scan revealed the minimal peptide scaffold necessary for ComD activation (Table 4). Changes to the N-terminus yielded inhibitory peptides, again demonstrating the roles of the first three residues in receptor activation. Cleavage of the last C-terminal residue, Lys19, resulted in an analog with similar activity to the native peptide, suggesting that this residue is dispensable. Removal of the next residue, Arg18, yielded an analog capable of activating the competence regulon at the highest tested concentration (10,000 nM), however this change resulted in significantly diminished activating potency. Again, it is possible that removal of this residue impacts secondary structure and thus limits the effective concentration needed to inspire ComD activity. Lastly, CSP-desK17R18K19 experienced complete loss of activity, indicating that Lys17 is needed for appropriate ComD activity. Results from C-terminal modification in the alanine and D-amino acid scan do not directly support the involvement of the C-terminus in ComD binding, however the removal of the C-terminus results in a complete loss of activity. Combined, these results indicate that the C-terminal region affects characteristics of the peptide that are necessary for proper receptor binding.

Table 4. Truncation scan of S. gordonii CSP.

^a This scan demonstrates the systematic removal of amino acids from either termini of the native peptide, thus revealing the importance of terminal amino acids in the binding and/or activation of ComD.

#	Peptide Name	Sequence	EC50 or IC50* (nM)b	95% CIc
40	desD1	IRHRINNSIWRDIFLKRK	985*	889 – 1090
41	desD1I2	RHRINNSIWRDIFLKRK	> 1,000*	-
42	desD1I2R3	HRINNSIWRDIFLKRK	> 1,000*	-
43	desK19	DIRHRINNSIWRDIFLKR	309	293 – 326
44	desR18K19	DIRHRINNSIWRDIFLK	> 1,000	-
45	desK17R18K19	DIRHRINNSIWRDIFL	-d	-

^aSee the Materials and Methods section or the Supporting Information for experimental methods. See the Supporting Information for details of the reporter strain and plots of agonism or antagonism dose-response curves. All assays performed in triplicate.

^bEC₅₀ or IC₅₀ values determined by testing peptides over a wide range of concentrations.

^c95% confidence interval.

^dEC₅₀ not determined due to the analogue’s low induction in primary agonism screening assays. See the Supporting Information for details.

Secondary Structure Determination of Peptide Analogs

Several studies demonstrate that CSPs take on an alpha-helical conformations upon final cleavage and secretion into the extracellular environment.^45–46 Previous SAR analyses revealed information regarding key residues in the binding and activation of ComD, however the effect of these systematic modifications on the secondary structure of the peptide were unknown. To gain greater insight regarding CSP secondary structure, all peptide analogs were evaluated using circular dichroism (CD) spectroscopy (Figure S-11 through Figure S-16). Peptides were evaluated in an aqueous environment consisting of phosphate-buffered saline (PBS), as well as a membrane-mimicking milieu comprised of 20% trifluoroethanol (TFE) in PBS. It was postulated that, generally, peptides in PBS would demonstrate random coil conformations while those subjected to TFE conditions would take on an alpha-helical structure.

The S. gordonii native CSP remained unfolded in aqueous solution, whereas it took on a helical structure with enhanced helicity of approximately 22.3% under membrane-mimicking conditions. All alanine scan analogs exhibited similar trends compared to the native CSP, with those in PBS taking on random coil conformations while TFE samples experienced alpha-helical character ranging from approximately 15 to 32% helicity. Results did not support a direct correlation between helical content and the excitatory/inhibitory potency of alanine scan analogs.

The D-amino acid scan generally demonstrated analogs with random coil structures in PBS and alpha helices in TFE, however all D-amino acid analogs exhibited significantly reduced helicity, ranging from 5 to 15%, when compared to the native CSP. Interestingly, CSP-i10, -w11, and -r12 all exhibited random coils in TFE, suggesting alteration to these residues negates the formation of an alpha helix. Additionally, SAR analysis revealed that the excitatory potency of these analogs was significantly reduced, supporting the concept that loss of the alpha helical conformation greatly impacts receptor binding in this case. Peptides with N-terminal modifications, including CSP-d1 and -i2 had greatly reduced helicity, but demonstrated inhibitory activity. CSP-r18 also exhibited slightly reduced helicity, but resulted in an analog with complete loss of ComD activation. Overall, results did not directly indicate a correlation between alpha helical content and biological activity, however, it appears that general helical characteristics are required for ComD binding.

Lastly, secondary structure evaluation of the truncation scan revealed insights similar to the D-amino acid scan. Truncation analogs in PBS demonstrated random coils, while those in membrane-mimicking conditions took on alpha helical character. While most truncated CSPs experienced a significant decrease in helical content, with the exception of CSP-desK19, truncation at the N-terminus established inhibitors and truncation at the C-terminus resulted in peptides with loss of activity.

Phenotypic Evaluation

A great deal of research has demonstrated the involvement of the competence regulon in modulating downstream proliferative phenotypic expression, such as biofilm formation, hydrogen peroxide formation, and even interspecies communication mechanisms in different streptococci species.^{10, 47–48} Furthermore, it was of great interest to illuminate how the competence regulon modulates proliferative phenotypic expression, as well as correlate information gained from RNA-seq data with observable S. gordonii phenotypic expression profiles.

Competence is one of the most widely studied and well-understood phenotypes associated with the comABCDE regulon.^49–52 This phenotype has been investigated in numerous streptococcal species, including S. pneumoniae, S. mitis, S. sanguinis, S. mutans, and other related streptococcal species.^{15, 53–54} These studies have demonstrated that the competence regulon can be governed in part by the comABCDE QS circuitry. Time-course assessments of the closely related strain of S. gordonii, sp. challis (ACTC 35105) have indicated that competence can be induced with exogenous CSP and several genes, many of which are related to early and late competence development, experience transcriptional upregulation.³⁰ Our RNA-seq data exhibited similar upregulation of early competence genes following a 10-minute CSP incubation, suggesting there are transcriptional parallels and analogous downstream phenotypic expression in these two strains of S. gordonii. Additionally, when constructing the luminescence-based reporter system, the constructed reporter plasmid was successfully transformed into S. gordonii via competence induction with synthetic native CSP. Overall, due to the profound genetic overlap between S. gordonii sp. challis and S. gordonii sp. firmicutes, coupled with the parallels observed between prior studies’ transcriptional data and data we obtained via RNA-seq analysis, it was deduced that competence can be modulated via the S. gordonii sp. firmicutes comABCDE QS circuitry. However, CSP:ComD structure-activity differences likely exist given the distinct sequence variation between the firmicutes and challis CSPs (DIRHRINNSIWRDIFLKRK and DVRSNKIRLWWENIFFNKK, respectively) and merit further exploration.

Biofilm formation is another prominent phenotype commonly associated with the competence regulon. Several streptococcal species engage in biofilm formation to establish protective microenvironments to shield themselves from harsh environmental conditions, including anaerobic conditions, nutrient scarcity, and immune response, among other stressors.^55–56 We sought to gain more information about biofilm formation in S. gordonii, with close attention to the potential role of QS in modulating this phenotype. Previous RNA-seq results indicated no significant increase or decrease in the transcript abundance of genes related to biofilm formation, so we did not anticipate a correlation between biofilm formation and CSP exposure. In accordance with this hypothesis, a crystal violet biofilm quantification assay demonstrated no significant difference between the amount of biofilm generated following treatment with varying concentrations of exogenous native CSP (Figure S-17). It is worthwhile noting that biofilm formation is an incredibly complex phenotypic process that relies on the adequate coordination and expression of several gene products. While this assay did not indicate a correlation between the competence regulon and biofilm formation, it is possible that conditions did not appropriately account for the complex nature of this process. Additional studies would aid in further delineating this phenotype and understanding its relation to QS mechanisms.

Hydrogen peroxide formation is another phenotype that is commonly associated with S. gordonii and its role in the oral microbiome. There have been numerous studies investigating the antagonistic production of peroxide by species including S. sanguinis, S. oligofermentans, and S. gordonii to combat the growth and proliferation of pathogenic S. mutans.^{22, 57–61} It has been well established that S. mutans engages in lactic acid secretion, thereby acidifying its microenvironment and inhibiting the growth of nearby bacterial populations.⁶² In self-defense, S. Oligofermentans has developed methods to utilize this secreted environmental lactic acid, converting it to inhibitory hydrogen peroxide to antagonize S. mutans proliferation.^58–60 While hydrogen peroxide production has been characterized in S. gordonii, its connection to the competence regulon QS circuitry has not been thoroughly evaluated. Following a protocol similar to Tong and coworkers, the ability of S. gordonii to produce hydrogen peroxide following incubation with varying concentrations of native CSP was evaluated.⁵⁸ Peroxide concentration was determined by fitting collected data to a standard curve from fixed H₂O₂ concentrations (Figure S-1).

Results indicated that there was a significant difference in hydrogen peroxide formation following incubation with all tested CSP concentrations compared to the untreated cultures, suggesting that this phenotype is at least in part modulated via competence regulon QS mechanisms (Figure 4). Interestingly, results also revealed a basal level H₂O₂ production (DMSO), suggesting that peroxide production, while capable of being modulated via the competence regulon, may also be controlled via other pathways as well. Overall, results suggest that the competence regulon is at least partially responsible for hydrogen peroxide formation and maintains the ability to modulate this phenotype.

Figure 4. Hydrogen peroxide production.

The quantification of hydrogen peroxide formation of wild-type cultures of S. gordonii incubated with varying concentrations of native CSP or DMSO (no treatment). Statistical significance was determined using an unpaired t-test with Welch’s correction; * P ≤0.05; *** P ≤ 0.0005; **** P ≤ 0.0001. The experiment was repeated three times in triplicate for a total of nine experiments.

Since the RNA-Seq analysis did not reveal upregulation of genes (spxB) involved in hydrogen peroxide production, while the phenotypic analysis revealed a significant increase in hydrogen peroxide production, we set out to further investigate this discrepancy. Since the RNA-Seq analysis was conducted 10 minutes post CSP treatment, we reasoned that spxB may be upregulated at a later time. Thus, we conducted qPCR analysis of the spxB gene at different time points post CSP treatment. Indeed, the qPCR analysis revealed a peak of spxB expression 20 minutes post CSP treatment. This increase appears to be short-lived, as spxB levels return to basal levels at 40 minutes post CSP treatment (Figure 5). Overall, the qPCR analysis reconciled the discrepancies observed between the RNA-Seq and hydrogen peroxide production results, and further validated the role of the competence regulon in hydrogen peroxide production in S. gordonii.

Figure 5. Time-resolved spxB expression in response to CSP addition.

RT-qPCR was conducted on S. gordonii liquid culture samples taken at 0, 20, and 40 minutes post-CSP addition, to assess changes in spxB expression in response to CSP. The obtained Cq values for spxB and gap (GAPDH) were used to calculate the fold change in spxB expression, relative to gap expression, using the 2^−ΔΔCt method. Error bars represent the standard error of the mean of three biological replicates.

Interspecies Interactions:

Enhanced H₂O₂ formation as a concentration-dependent function of CSP exposure encouraged further exploration of this phenotype. We aimed to investigate the practical antagonistic ability of S. gordonii against S. mutans. Previous research suggested that S. gordonii would indeed be able to antagonize the growth of S. mutans, however we sought to correlate inhibition with CSP exposure.⁵⁷ To this end, S. gordonii cultures incubated with varying concentrations of CSP were exposed to two different S. mutans UA159 strains: one with a comC knockout incapable of producing the bacteriocin mutacin (SMCC3) and one with intact QS circuitry (SMCOM3). It was hypothesized that S. gordonii would be capable of inhibiting both strains through hydrogen peroxide production, resulting in prominent zones of inhibition (ZOI).

Several trials indicated that, following supplementation with lactic acid and varying concentrations of CSP, S. gordonii developed zones of inhibition (ZOI) when plated on lawns of either SMCOM3 or SMCC3 (Figure 6). The visible impeded proliferation of S. mutans SMCOM3 indicated that S. gordonii could antagonize the growth of this species despite intact mutacin production. To further validate the role of hydrogen peroxide production in S. mutans inhibition, the same experiments were repeated in the presence of catalase, an enzyme that degrades hydrogen peroxide. Indeed, following exposure to catalase, significantly reduced ZOIs were observed, supporting our hypothesis regarding the importance of hydrogen peroxide in S. gordonii-mediated killing of S. mutans. Additionally, in the case of SMCC3, but not SMCOM3, the CSP concentration seemed to impact the degree of S. mutans inhibition, as could be visualized through a slight increase in the ZOI at high CSP concentrations, compared with the low CSP concentrations or the negative DMSO control. However, since this effect was minimal, a clear correlation cannot be confidently concluded. Lastly, no treatment (CSP only, with no S. gordonii present) samples exhibited no ZOI, suggesting that the CSP itself is not toxic towards S. mutans. Additional trials without lactic acid supplementation exhibited similar results, suggesting that either the lactic acid produced by S. mutans is sufficient to allow effective hydrogen peroxide production by S. gordonii, or, more likely, that lactic acid is not required for hydrogen peroxide production by S. gordonii (Figure S-18). Overall, these results highlight the important role hydrogen peroxide plays in S. mutans inhibition. Addition of exogenous CSP to induce the competence regulon in S. gordonii could improve the effectiveness of S. mutans clearing. As these experiments intentionally expose S. gordonii to higher than native concentrations of CSP, the role of CSP-regulated H₂O₂ production for interspecies competition in natural environments may not be as pronounced.

Figure 6. Interspecies competition assay.

S. gordonii was treated with 10 mM lactic acid and varying concentrations of CSP (10,000 nM, 1,000 nM, 100 nM, 10 nM), with or without the presence of catalase, after which its ability to inhibit the growth of S. mutans SMCC3 (ΔcomC) or SMCOM3 was assessed. (A) An SMCC3 lawn was spotted with S. gordonii previously incubated with 10,000 nM CSP (far left) to 10 nM (second from right), including a no treatment control (DMSO, far right), without the addition of catalase. (B) An SMCC3 lawn was spotted with S. gordonii previously incubated with 10,000 nM CSP (far left) to 10 nM (second from right), including a no treatment control (DMSO, far right), with the addition of catalase. (C) An SMCOM3 lawn was spotted with S. gordonii previously incubated with 10,000 nM CSP (far left) to 10 nM (second from right), including a no treatment control (DMSO, far right), without the addition of catalase. (D) An SMCOM3 lawn was spotted with S. gordonii previously incubated with 10,000 nM CSP (far left) to 10 nM (second from right), including a no treatment control (DMSO, far right), with the addition of catalase. The experiment was repeated three times in triplicate for a total of nine experiments.

Summary and Conclusions

The competence regulon has been widely linked to a variety of proliferative phenotypic responses, including competence induction, biofilm formation, and modulation of interspecies communication mechanisms, all of which have the capacity to affect oral and gut microbiome composition and thus human health. This study sought to explore the potential connection between the S. gordonii competence regulon and downstream phenotypic expression. Results from this study indicated that the S. gordonii CSP maintains significant involvement in competence induction and hydrogen peroxide formation via modulation of the competence regulon.

Previous studies have investigated the ability of S. gordonii to produce inhibitory hydrogen peroxide, however this phenotype has not been thoroughly investigated in relation to the competence regulon. Following validation of the S. gordonii CSP sequence as DIRHRINNSIWRDIFLKRK, RNA-seq demonstrated a significant upregulation of genes related to early and late competence. However, there was no indication of increased transcription of genes related to other downstream phenotypes, such as biofilm formation, virulence factor production, or peroxide formation.

Following this, we sought to explore the role of CSP in ComD binding and activation. To do this, a luminescence reporter was constructed in which pcomX was inserted prior to the gene for luciferase, rendering a system that would produce a detectable and quantifiable luminescent signal proportional to comX activation. The luminescence response at natural CSP concentrations was low, but starkly differential from the wild-type, and increased proportionally with the addition of exogenous CSP. Following initial evaluation of the CSP, SAR analyses were completed to reveal amino acid residues critical for ComD binding and activation, as well as residues profoundly tolerant to modification. Our results revealed the critical nature of the N-terminus for receptor binding and activation, while the C-terminus was permissive to change, but not removal. Three analogs were discovered with roughly two-fold enhanced activity (CSP-R12A, -L16A, and -i6), as well as an analog capable of ComD inhibition (CSP-desD1).

Finally, detailed phenotypic analyses were completed and used to correlate RNA-seq data and SAR results with the actual expression profiles for specific phenotypes, including biofilm formation and hydrogen peroxide formation. While biofilm formation was unaffected by CSP exposure, peroxide formation appeared to be a CSP concentration-dependent process determined by supplementing S. gordonii with increasing concentrations of exogenous CSP. To resolve the potential discrepancy between the RNA-Seq data and the phenotypic observations, time-resolved qPCR analysis of the spxB gene was conducted and revealed a spike in spxB expression 20 minutes post CSP treatment, a later time point than the time point used for the RNA-Seq analysis. Lastly, interspecies interaction assays highlighted the antagonistic relationship between S. gordonii and S. mutans, with S. gordonii demonstrating the ability to inhibit the growth and proliferation of this pathogen. Overall, our results suggest that S. gordonii may be exploited for its beneficial phenotypes, serving as a potential biotherapeutic against pathogenic S. mutans and perhaps other infective streptococci.

Materials and Methods

Peptide synthesis and purification:

All peptide analogs were synthesized using standard Fmoc solid-phase peptide synthesis protocols on a Liberty Prime 2.0 Automatic Peptide Synthesizer or CEM Discover, followed by purification via RP-HPLC to ≥95% purity. Masses of purified peptides were confirmed by high-resolution ESI-TOF mass spectrometry. See Table S-5 through Table S-7 for peptide masses and purities. See the Supporting Information for full details.

Isolation of crude peptides from bacterial supernatants:

The native S. gordonii CSP was isolated from cell-free supernatants using previously described protocols with the exception that S. gordonii CSP-desK19 was used to spike the culture and induce native CSP production.⁶³ The peptide sequence was verified using MS/MS. See Supporting Information for more details.

Circular Dichroism (CD) Spectroscopy:

CD spectra were recorded using a Jasco CD Spectrophotometer (model J-1500–150). See Supporting Information for full experimental details.

Construction of S. gordonii luciferase-based reporter system:

The luminescence-based reporter system was constructed using previously described protocols with some modifications.⁶⁴ See Supporting Information for full details.

Biofilm formation assay:

The biofilm formation assay was conducted using previously described protocols with some modifications.¹⁴ See Supporting Information for full experimental details.

Hydrogen peroxide production assay:

See Supporting Information for full experimental details.

Interspecies competition assays:

See Supporting Information for full experimental details.

Supplementary Material

The Supporting Information is available free of charge at

Full experimental procedures, peptide characterization, additional RNA-Seq table, primary reporter assay data, dose−response curves for CSP analogues, CD spectra of synthetic CSP analogues, and supplementary figures.

Data Availability

All data are presented in the article or ESI. Raw RNA-sequencing reads data can be obtained from National Center for Biotechnology Information database (NCBI: BioProject ID PRJNA1222368).