Peptides encoded in the Streptococcus mutans RcrRPQ operon are essential for thermotolerance

Abstract

The MarR-like transcriptional regulator and two ABC transporters encoded by the rcrRPQ operon in the dental caries pathogen Streptococcus mutans have important regulatory roles related to oxidative stress tolerance, genetic competence and (p)ppGpp metabolism. A unique feature of the rcrRPQ operon, when compared to other bacteria, is the presence of two peptides, designated Pep1 and Pep2, encoded in alternative reading frames at the 3′ end of rcrQ. Here, we show that the rcrRPQ operon, including Pep1 and 2, is essential for S. mutans to survive and maintain viability at elevated temperatures. No major changes in the levels of the heat shock proteins DnaK or GroEL that could account for the thermosensitivity of rcrRPQ mutants were observed. By introducing a single amino acid substitution into the comX gene that deletes an internally encoded peptide, XrpA, we found that XrpA is a contributing factor to the thermosensitive phenotype of a ΔrcrR strain. Overexpression of XrpA on a plasmid also caused a significant growth defect at 42 °C. Interestingly, loss of the gene for the RelA/SpoT homologue (RSH) enzyme, relA, restored growth of the ΔrcrR strain at 42 °C. During heat stress and when a stringent response was induced, levels of (p)ppGpp were elevated in the ΔrcrR strain. Deletion of relA in the ΔrcrR strain lowered the basal levels of (p)ppGpp to those observed in wild-type S. mutans . Thus, (p)ppGpp pools are dysregulated in ΔrcrR, which likely leads to aberrant control of transcriptional/translational processes and the thermosensitive phenotype. In summary, the genes and peptides encoded in the rcrRPQ operon are critical for thermotolerance, and in some strains these phenotypes are related to altered (p)ppGpp metabolism and increased production of the XrpA peptide.

Impact statement

Streptococcus mutans is a human pathogen that drives the biofilm infection that leads to dental caries (tooth decay). It is important to identify biological processes that aid the survival of S. mutans in oral biofilm communities. Stress tolerance, including tolerance to fluctuating temperatures, is essential for S. mutans persistence and virulence. Here we show that gene products encoded by the rcrRPQ operon are essential for S. mutans to survive heat stress. These phenotypes are linked to changes in the expression/activities of a genetic competence-related peptide, XrpA, and the production of nucleotide signal molecules. These findings highlight novel aspects of S. mutans physiology that may be attractive therapeutic targets for control of dental caries.

Introduction

There is increasing evidence that small peptides have important functions in bacteria beyond functioning as intercellular communication molecules or bacteriocins [1], but progress in this area has been slowed, primarily by two challenges. First, traditional genome annotation methods are not optimized for the discovery of proteins smaller than 100 aa [2]. Second, forward genetic techniques that rely on mutagenesis events often do not identify small open reading frames (ORFs) because of the low frequency of insertions and/or because the peptide coding sequence may exist within an annotated ORF. Despite these challenges, peptidomics research has slowly gathered pace and several bacterial peptides are now known to have key regulatory functions in bacteria. For example, peptides have now been discovered that regulate cell division (MciZ, 40 aa) [3], proteolysis (MgtR, 30 aa) [4], carbohydrate metabolism (SgrT, 43 aa) [5] and multidrug efflux pumps (AcrZ, 49 aa) [6].

Two small peptides were previously discovered to be encoded in the 3′ end of the rcrRPQ operon of Streptococcus mutans , a bacterium that is well established as a primary contributor to the initiation and progression of dental caries [7]. Certain mutations in the rcrRPQ operon of S. mutans diminish oxidative and acid stress tolerance, dramatically affect genetic competence and lead to changes in (p)ppGpp accumulation [8]. RcrP and RcrQ are predicted to be ABC efflux pumps and RcrR is a MarR-like transcriptional regulator that represses the entire operon by binding near a single promoter located 5′ to rcrR (Fig. 1a) [9]. Although the operonic arrangement and primary sequence of RcrRPQ are conserved across the oral streptococci, there are peculiarities of the S. mutans operon that distinguish it from other oral streptococci. For instance, polar and non-polar deletion/replacement mutations of rcrR in S. mutans dominantly impact the development of genetic competence, whereas there are no discernible effects of similar mutations on competence in rcrR mutants of Streptococcus gordonii [10]. Another important distinction is that the two small peptides encoded in the very end of the rcrQ gene are conserved in the rcrQ genes of all isolates of S. mutans for which genome sequence is available (Fig. 1a). These peptides are not only absent in all other oral streptococci for which sequence information is available, but there are no similar peptides identifiable in any current databases [7]. The peptides themselves have been shown to strongly influence the phenotypic behaviours of the parental strain and of rcrR, rcrP and rcrQ mutants [7]. Because of the uniqueness of these peptides and how profoundly they influence important physiological properties, it has been proposed that they may be targeted to inhibit the establishment, persistence and virulence of S. mutans [7].

Fig. 1.

Deletion of rcrR, rcrP and rcrQ leads to decreased growth at elevated temperatures. (a) Organization of the rcrRPQ-pep12 operon with cluster of orthologous group (COG) identities shown. RcrR (depicted by grey circles binding to the promoter region) is a MarR family autogenous repressor of the rcrRPQ operon. RcrP and RcrQ share homology with multi-drug efflux ABC transporters, although their substrates are currently unknown. The 3’ end of rcrQ also includes two short ORFs, designated pep1 and pep2. Strains were grown in rich medium (BHI) with a sterile mineral overlay at 42 °C. Growth curves for the following strains as indicated above each panel: (b) UA159, ΔrcrR-NP, ΔrcrR-P; (c) UA159, ΔrcrP-NP, ΔrcrP-P; (d) UA159, ΔrcrQ-NP, ΔrcrQ-P; (e) UA159, ΔrcrPQ-NP and ΔrcrPQ-P. Graphs are representative of three biological replicates.

During an analysis of rcrR deletion mutants of S. gordonii , thermosensitivity was observed [10]. We subsequently determined that particular mutations in the rcrR gene of S. mutans caused sensitivity to elevated temperatures. While our laboratory has made inroads into understanding how rcrRPQ and the associated peptides affect genetic competence, the molecular mechanisms by which these factors are integrated into the stress response circuit remain somewhat enigmatic. The finding that these unique and unusual peptides alone are able to determine whether S. mutans can grow and tolerate elevated temperature provided a convenient and reproducible phenotype to probe the molecular mechanisms connecting RcrRPQ and Pep1 and 2 to the stress response circuitry. The results further highlight the critical roles of rcrRPQ and the rcr-encoded peptides in S. mutans and disclose important linkages to (p)ppGpp metabolism and how novel elements of the genetic competence circuit integrate complex cellular behaviours with the physiological status of the cells in a ubiquitous human pathogen.

Methods

Bacterial strains, media, and growth conditions

S. mutans strains (Table 1) were routinely cultured in brain heart infusion (BHI) broth (Difco Laboratories). Strains were grown in a 5 % CO₂ aerobic environment at 37 °C, unless stated otherwise. When necessary, antibiotics were added to media as follows: 1 mg ml⁻¹ kanamycin, 5 µg ml⁻¹ erythromycin and 1 mg ml⁻¹ spectinomycin.

Table 1.

Strains and plasmids used in this study

Strain or plasmid	Description	Reference
S. mutans
UA159	Wild-type	Burne Laboratory
ΔrcrR-NP	rcrR::NPkanR	[8]
ΔrcrR-P	rcrR::ΩkanR	[8]
ΔrcrP-P	rcrP::ΩkanR	[8]
ΔrcrP-NP	rcrP::NPkanR	[8]
ΔrcrQ-P*	rcrQ::ΩkanR	This study
ΔrcrQ-NP*	rcrQ::NPkanR	[8]
ΔrcrPQ-P*	rcrPQ::ΩkanR	This study
ΔrcrPQ-NP	rcrPQ::NPkanR	[8]
ΔrcrRPQ	rcrRPQ::NPkanR	This study
ΔrcrRPQ Δpep12	rcrRPQpep12::NPkanR	This study
Δpep1	rcrQpep1::2T>C	[7]
Δpep2	rcrQpep2::2T>C	[7]
Δpep12	rcrQpep12::2T>C	[7]
ΔcomX	comX::eryR	[46]
ΔrcrR-P ΔrelA	rcrR::Ωkm, relA::eryR	[38]
ΔrcrR-P ΔxrpA	rcrR::Ωkm, comX::T162 >C	[33]
ΔrelA	relA::eryR	[17]
pIB184	UA159, pIB184, eryR	[31]
pIB184-pep12	UA159, pIB184::P23-pep12, eryR	[7]
pIB184-comX	UA159, pIB184::P23-comX (comX::T162 >C), eryR	This study
pIB184-xrpA	UA159, pIB184::P23-xrpA, eryR	[33]
Plasmids
pIB184	Shuttle expression plasmid with constitutive P23 promoter, eryR	[31]

*Kanamycin cassette is inserted in such a way that pep2 is deleted and the pep1 sequence is intact.

NPkan^R, non-polar kanamycin resistance; Ωkan^R, polar kanamycin resistance; ery^R, erythromycin resistance.

Construction of mutant strains

Standard DNA manipulation techniques were used to engineer plasmids and strains [11]. A PCR ligation mutagenesis method was used to replace coding regions with polar or non-polar kanamycin markers [12]. Transformants were selected on BHI agar containing kanamycin. Double-crossover recombination, without the introduction of nearby secondary mutations, was confirmed by PCR and Sanger sequencing. In addition, previously published rcrRPQ strains were sequenced (primers described in Table 2) to ensure that the strains were as described. The strains used or constructed during this study are described in Table 1.

Table 2.

Oligonucleotide primers used in this study

Primer name	Nucleotide sequence (5’ to 3’)	Use
rcrRPQSeq1	ACAGCGAAAAAGCAATCGTT	Strain sequencing
rcrRPQSeq2	ACCAGCAAAATCTGCCAAAG	Strain sequencing
rcrRPQSeq3	CGCCATTAGGTCAGGAAAAA	Strain sequencing
rcrRPQSeq4	CTCCTTGTGAATGGCTGTCA	Strain sequencing
rcrRPQSeq5	TGGTCAAGGAACACACAAGG	Strain sequencing
rcrRPQSeq6	TTTGCCATGGTACTCCAAGA	Strain sequencing
Sm16SF	CCGGTGACGGCAAGCTAA	qRT-PCR
Sm16SR	TCATGGAGGCGAGTTGCA	qRT-PCR
comXF	TGCGTCAGCAAGAAAGTCAG	qRT-PCR
comXR	TACCGCCACTTGACAAACAG	qRT-PCR
dnaKF	CGTAACGCCATTGTCACTTG	qRT-PCR
dnaKR	AACGGCTGGTTGGTTATCTG	qRT-PCR
groELF	TGCGGATGATGTAGATGGTG	qRT-PCR
groELR	ACGATCACCAAAACCAGGAG	qRT-PCR
rcrQF	AGATGCACCGCTTCTGATTC	qRT-PCR
rcrQR	ACAGACGATGGGCAATAACG	qRT-PCR

Growth phenotypes

A Bioscreen C automated growth curve instrument (Oy Growth Curves AB) was used to monitor the effects of stressors on bacterial growth. Wild-type and mutant S. mutans strains were cultured overnight in BHI or BHI containing antibiotics, when noted. The following day, cells were diluted 1 : 50 into fresh media and grown to OD₆₀₀=0.5. Thereafter, cells were inoculated in triplicate into the wells of a Bioscreen C plate in 300 µL BHI at a dilution of 1 : 100. The cultures were incubated at 37 °C or 42 °C with optical density (OD₆₀₀) recorded every 30 min for 48 h. The Bioscreen C plate was shaken at medium intensity for 10 s prior to each OD₆₀₀ reading. Experiments were performed in triplicate.

Enumeration of viable cells

Cultures of S. mutans mutants were removed from Bioscreen C plates at 0, 2, 4, 6, 24 and 48 h to determine viable counts. Two hundred microlitres of culture were serially diluted from 10⁻¹ to 10⁻⁷ in sterile BHI. Next, each dilution was plated on BHI agar with antibiotics, if necessary. Plates were incubated for 48 h before colonies were counted; 48 h was selected rather than 24 h because certain S. mutans strains took longer to form colonies than the wild-type strain after being subjected to heat stress. The total number of colony-forming units (c.f.u.) was determined using the following formula: c.f.u. ml⁻¹=colonies counted×(1/dilution)×(1/volume).

LIVE/DEAD staining of microbial populations

For the microscopy of planktonic micro-organisms, 200 µL of cultures grown in the Bioscreen were harvested at 0, 2, 4, 6, 24 and 48 h. Samples were centrifuged at 3500 g for 10 min at 4 °C and then resuspended in phosphate-buffered saline (PBS) plus LIVE/DEAD BacLight bacterial viability stain (Thermo Fisher Scientific), prepared as recommended by the supplier. After incubation at room temperature in the dark for 15 min, cell suspensions were mounted on glass slides with glass coverslips. Microscopy was performed using a Leica DMIRB inverted fluorescence microscope equipped with a Photometrics cascade-cooled EMCCD camera. Syto9 fluorescence (live) was detected by excitation at 488 nm and emission was collected using a 525 nm (±25 nm) bandpass filter. Detection of propidium iodide fluorescence (dead) was performed using a 642 nm excitation laser and a 695 nm (±53 nm) bandpass filter. Images were acquired with a 63×/1.40 oil objective and processed using VoxCell (VisiTech International).

Preparation of protein from S. mutans .

Strains of interest were cultured overnight, diluted 1 : 50 into pre-warmed media, grown to the mid-exponential phase (OD₆₀₀=0.5) and then transferred to a 42 °C water bath. Control samples were incubated at 37 °C. After 30 min, cell cultures were centrifuged at 4000 g for 5 min at 4 °C. Cell pellets were frozen overnight at −20 °C. The following day, each pellet was resuspended in 750 µL of lysis buffer (60 mM Tris, pH 6.8, 2 % SDS) and placed in a screw-cap microcentrifuge tube with 0.25 ml of 0.1 mm glass beads. Samples were homogenized in a Bead Beater for 30 s, three times, with 5 min intervals on ice. Samples were then centrifuged at 8000g for 10 min at 4 °C. Cell lysates containing S. mutans proteins were carefully removed into a fresh 1.5 ml microcentrifuge tube. Protein concentration was determined using the bicinchoninic acid assay (BCA) following the supplier’s protocol (Pierce), with purified bovine serum albumin as the standard.

Western blotting

Clarified cell lysates (10 µg protein) from strains of interest were boiled in Laemmli’s sample buffer (Bio-Rad) for 10 min. Protein samples were then separated on 12 % Mini-PROTEAN TGX precast gels (Bio-Rad). Afterwards, proteins were transferred to PVDF membranes using a Trans-Blot Turbo transfer system (Bio-Rad). Membranes were treated with rabbit antisera elicited to purified, recombinant S. mutans DnaK (1 : 5000 dilution) [13] or recombinant Streptococcus pyogenes GroEL protein (1 : 5000 dilution) [14] and a secondary peroxidase-labelled, goat-anti-rabbit immunoglobulin G (IgG) antibody (1 : 5000 dilution; SeraCare Life Technologies). Western blot signals were detected using the SuperSignal West Pico Chemiluminescent Substrate kit (Thermo Fisher Scientific) and visualized with a FluorChem 8900 imaging system (Alpha Innotech).

Quantification of mRNA expression

Quantitative real-time reverse transcriptase PCR (qRT-PCR) was performed to measure the mRNA of S. mutans 16S rRNA, rcrQ, comX, dnaK and groEL genes (oligonucleotide sequences in Table 2). Cells were harvested in the mid-exponential phase (OD₆₀₀=0.5) and treated with RNAprotect (Qiagen), and total RNA was extracted using phenol (pH 4.3) and the RNeasy mini kit (Qiagen). RNA concentration was determined using a Nanodrop spectrophotometer (Thermo Fisher Scientific). Purified RNA (1 µg) was used to synthesize cDNA according to the Superscript III (Invitrogen) supplier’s protocol. Primers for each gene (Table 2) were designed using the qPCR settings in the Primer3Plus online application [15]. Standard curves for each gene were generated using 8 10-fold dilutions of PCR products, starting with 10⁸ copies µl⁻¹. Triplicates of standard curve DNAs, samples and cDNA controls were added to wells containing iQSYBR green supermix (Bio-Rad) with primers (0.4 µM). Thermocycling was carried out using an CFX96 Real Time PCR detection system (Bio-Rad) set to the following protocol: 40 cycles of 95 °C for 10 s and 60 °C for 45 s, with a starting cycle of 95 °C for 30 s. Analysis of the data was performed as described elsewhere [16].

Detection of (p)ppGpp

Detection of (p)ppGpp was performed as previously described [17]. Briefly, overnight cultures were diluted 1 : 50 in the defined medium known as “FMC” [18], formulated using a reduced amount of phosphate (8.6 mM) and containing 25 mM glucose. Cultures were grown at 37 °C to OD₆₀₀=0.2, and then a 200 µl aliquot of each culture was incubated with 30 µCi of [³²P]-orthophosphate for an additional hour, with or without 500 ng ml⁻¹ of mupirocin, which induces robust (p)ppGpp accumulation in S. mutans [19]. The labelled cells were collected by centrifugation at 10 000g for 5 min and resuspended in 10 µl of fresh FMC medium. Nucleotides were extracted by resuspending cells in 10 µl of ice-cold 13 M formic acid, followed by three freeze-thaw cycles using a dry ice/ethanol bath. Next, the c.p.m. µl⁻¹ was measured using a scintillation counter and 2.0×10⁵ counts per minute (c.p.m.) of each sample was spotted onto polyethyleneimine (PEI)–cellulose plates (EMD Millipore Corp.) for the separation of the phosphorylated nucleotides by thin-layer chromatography (TLC). The plates were chromatographed with 1.5 M KH₂PO₄ (pH 3.5), air-dried and exposed to BioMax MS film (Carestream Health, Inc.) at −80 °C.

Results and discussion

The rcrRPQ operon affects thermotolerance

We previously described the inability of a mutant of S. gordonii carrying a deletion of the rcrR gene (ΔrcrR) to grow at 43 °C, whereas the mutant grew as well as the parental strain at 37 °C [10]. We hypothesized that a similar defect may manifest in a ΔrcrR mutant derived from S. mutans UA159. The growth defect at 42 °C was substantial in an S. mutans ΔrcrR-P strain (polar mutant), but was not apparent in S. mutans ΔrcrR-NP (non-polar mutant). It is established that the rcrR polar mutant expresses wild-type levels of rcrPQ mRNA [8], whereas the rcrPQ mRNA levels in the non-polar mutant are more than 100-fold higher than in the wild-type or rcrR polar mutant. To explore the basis for these phenotypes in more detail, we examined the thermosensitivity of various strains in which rcrRPQ gene expression and/or the production of the two peptide effectors encoded within the rcrQ gene were disrupted. At 37 °C, S. mutans UA159 and rcr deletion strains displayed similar growth phenotypes (Fig. S1, available in the online version of this article). Differences between the wild-type and rcrR-NP mutant strain and the other mutants were striking when growth was monitored at 42 °C. S. mutans UA159 grew to a final OD₆₀₀ of 0.57 (±0.02), with a doubling time of 62.6 (±4.9) min during the exponential phase of growth (Fig. 1b), very similar to the S. mutans ΔrcrR-NP strain. In contrast, no growth was observed at 42 °C for the ΔrcrR-P, ΔrcrP-P, ΔrcrP-NP and ΔrcrQ-NP (Fig. 1b–e), even after prolonged incubation (>30 h). Growth at 42 °C was observed for a ΔrcrQ-P strain (Fig. 1d). Double deletion of both of the rcrPQ-encoded ABC transporters restored growth at 42 °C (Fig. 1e). A descriptive account of the various phenotypes of rcr and related strains at 42 °C is shown in Table S1 in the supplementary information.

We next determined whether the thermosensitive strains were killed at elevated temperature or were simply growth-arrested. Cells (1×10⁶ c.f.u. ml⁻¹) were inoculated into BHI at 42 °C and cell viability was monitored by plating 2, 4, 6, 24 and 48 h after inoculation. Compared to the parental stain, S. mutans UA159, viability was substantially reduced in ΔrcrR-P, ΔrcrP-P and ΔrcrP-NP (Fig. 2a). The mutants began to lose viability shortly after inoculation, with approximately 50 % cell viability being evident after 2 h of incubation at 42 °C, whereas the wild-type had begun to grow during the same period. After 48 h, four to five logs of killing of the strains were observed (Fig. 2a), compared to the viable counts of the parental strain at the same time points. Additionally, we explored if the thermosensitive mutants were undergoing any changes in cell morphology and if membranes were compromised at 42 °C. Wild-type and mutant strains were incubated at 42 °C in BHI for 0, 24 and 48 h, stained with the LIVE/DEAD BacLight bacterial viability kit, and observed by phase contrast and fluorescence microscopy. The wild-type strain retained normal morphology and most cells stained as live after 48 h at 42 °C (Fig. 2b). However, the mutant strains were uniformly stained with propidium iodide after 24 and 48 h. Rather than being organized primarily in individual chains like S. mutans UA159, the mutant strains aggregated. Notably, the thermosensitive mutant cells were surrounded by a diffuse PI-stained material. PI stains DNA, so in lieu of retention of the dye by some other substance(s), it is reasonable to suggest that the observed material was extracellular DNA (eDNA) released during cell lysis. Thus, the mutant strains had compromised cell membranes and underwent cell lysis when incubated at 42 °C for extended periods. In S. mutans , eDNA is released both by cell lysis [20] and via membrane-bound vesicles [21].

Fig. 2.

Elevated temperature leads to loss of cell viability for thermosensitive mutants. (a) Percentage of surviving S. mutans when grown in BHI at 42 °C. At each time point, strains were serially diluted and cultured on agar plates to enumerate colony forming units (c.f.u.). The number of c.f.u. was then compared to the starting inoculum for each strain to determine the percentage change in bacterial cell numbers. (b) At 0, 24 and 48 h the viability of S. mutans wild-type, ΔrcrR, and ΔrcrP strains was visualized with LIVE/DEAD BacLight bacterial viability stain and fluorescence microscopy. Live bacteria with intact cell membranes fluoresce green, while bacteria with compromised cell membranes, presumably dead, fluoresce red.

The results described above imply that the RcrP ABC transporter contributes to thermotolerance in S. mutans . The role of RcrQ is more ambiguous as a polar rcrQ mutant was able to grow at 42 °C. RcrR is a MarR family transcriptional repressor that binds to the rcr promoter, repressing rcrRPQ expression [9]. Under inducing conditions, a chemical or peptide is proposed to bind to RcrR and decrease the affinity of its interaction with the rcr promoter region [9], leading to increased transcription of the rcrRPQ operon. MarR family regulators can be allosterically regulated by numerous different chemical signals, including Zn(II), Cu(II), reactive oxygen species, urate and salicylate [22], but it has not been determined what factor(s) interacts with RcrR. It is noteworthy that RcrP (Smu.922) protein levels are higher during heat stress [23], although rcrRPQ mRNA levels are not elevated during heat shock [24, 25], suggesting that elevated temperature or a chemical produced during heat shock is not an activator of rcrRPQ transcription. The ΔrcrR strains also exhibit defective responses to quorum-sensing peptides and display aberrant growth when exposed to acid or oxidative stress [8], although none of these conditions lead to complete growth arrest or cell death, as was the case for heat stress. Thus, the findings that ΔrcrRPQ strains have a thermosensitive phenotype have provided new insights into the dysfunctional physiology of ΔrcrR strains.

Heat shock proteins (HSPs) in temperature-sensitive rcr mutants

Denaturation and aggregation of proteins are consequences of thermal stress. The major HSPs DnaK (Hsp70), along with GrpE and DnaJ, as well as GroEL (Hsp60), are upregulated during heat shock in S. mutans and have primary roles in repairing damaged proteins or presenting them to the proteolytic machinery for degradation [24, 26, 27]. To determine whether mutations in the rcrRPQ operon might exert their effects through DnaK or GroEL, the levels of these proteins were monitored by Western blotting in wild-type S. mutans and ΔrcrR-P, ΔrcrP-P and ΔrcrP-NP. The results (Fig. 3) revealed only minor differences in the levels of DnaK and GroEL between the strains at 37 °C, and after heat shock at 42 °C for 30 min (unmodified versions of the Western blots are shown in Fig. S2). In particular, all strains responded to heat shock by increasing the levels of DnaK protein afterward, compared to those found in cells cultured at 37‌°C (Fig. 3). The ΔrcrR-P mutant had higher basal levels of DnaK compared to the wild-type strain, consistent with the 1.7-fold increase in dnaK transcript levels in ΔrcrR-P measured by qRT-PCR (Fig. S2). Both of the ΔrcrP strains had wild-type levels of dnaK mRNA (Fig. S3). All strains had increased levels of GroEL in response to heat shock, except for ΔrcrR-P (Fig. 3). Strains with polar and non-polar insertions in rcrP produced more GroEL protein than the wild-type when heat shocked at 42 °C. Despite their importance, and being under the transcriptional regulation of HrcA, with groEL expression also being controlled by CtsR [28, 29], the upregulation of the operons encoding GrpE-DnaKJ and GroESL in response to heat shock was relatively weak, approximately 1.5-fold as measured by qRT-PCR [24]. Therefore, the relatively minor changes noted in protein levels after heat shock of wild-type S. mutans are consistent with the increases in transcription of the operons. Both DnaK and GroEL are essential and have functions beyond adaptation to heat stress, with properly regulated levels of these proteins being required for an optimal response to a variety of stressors in S. mutans [26, 30]. It is notable, however, that significant downregulation (~90 %) of either dnaK and groESL is required to elicit substantive changes in the ability of S. mutans to grow at 43 °C [26]. None of the rcr strains tested here exhibited major changes in the levels of DnaK or GroEL. Therefore, we propose that a different mechanism leads to the extreme thermosensitivity of rcrRPQ mutant strains.

Fig. 3.

Western blot analysis of DnaK and GroEL. Protein samples were taken at 37 °C during the mid-exponential growth phase and after heat shock at 42 °C for 30 min. (a) For the Western blots, 10 µg of proteins in clarified cell lysates were separated by electrophoresis in 12 % Mini-PROTEAN TGX precast gels and transferred to PVDF membranes, and primary antisera raised against the corresponding protein were used for detection. The abundance of (b) DnaK and (c) GroEL proteins was measured with densitometry using three independent Western blot replicates.

Effects of the loss of rcr peptides on thermosensitivity

The thermosensitive mutant strains ΔrcrR-P and ΔrcrP (polar and non-polar) have intact pep1 and pep2 sequences, and are expected to produce functional rcr-encoded peptides. However, we anticipate that the amount of Pep1 and Pep2 produced by the different rcr mutants differs as a function of the polarity of the markers inserted upstream of the coding sequences for the peptides. As an example, a 20-fold downregulation of rcrQ mRNA was measured in the ΔrcrP-P mutant (Fig. S4). To begin to explore how the levels of rcr peptides influence the thermotolerance of S. mutans , we observed how deletion of these peptides impacted on S. mutans growth at elevated temperatures. We used start codon mutants (ATG>ACG) of individual rcr-encoded peptides (Δpep1, Δpep2) or a strain where both start codons were mutated (Δpep12), in both cases keeping the coding sequences for RcrRPQ intact, and cultured these strains in rich medium at 42 °C. Strains lacking individual or both peptides were unable to grow at 42 °C (Fig. 4a). The number of cells of a Δpep12 mutant decreased during incubation at 42 °C (Fig. 4b), and, as was the case for the ΔrcrR-P and ΔrcrP strains, Δpep12 mutants demonstrated compromised cell membranes when cultured at 42 °C (Fig. 4c). We also anticipated that the levels of the rcr-enoded peptides would be elevated in some of the non-polar rcr mutants. Indeed, rcrQ expression was increased 2.1-fold in the ΔrcrP-NP strain (Fig. S3). To test how elevated levels of the rcr-encoded peptides might impact on thermotolerance, we incubated a pep12 overexpression strain (peptides under the control of the constitutive P23 promoter carried on pIB184 [7, 31]) in rich media at 42 °C. This strain grew more slowly than a strain carrying pIB184 without pep12, but reached comparable final optical densities (Fig. 4d). Our observations of the growth phenotypes of rcr peptide mutants, and the double-peptide overexpression strain, helped tease apart loss of, or changes in the expression levels of, the ABC transporters from aberrant peptide expression and how these changes influence thermosensitivity. Across the various rcr mutants, the inability to grow at 42 °C appears to be related to aberrant expression of ABC transporters and peptides and/or loss of ABC transporters or peptides. These results mirror the behaviours of various rcrRPQ/peptide mutant strains in terms of their impact on genetic competence of S. mutans . Genetic competence phenotypes are profoundly altered by the various components of the rcrRPQ operon [7, 8]. One proposal for how rcr-encoded peptides are thought to influence genetic competence behaviours is by increasing the rate at which the alternative sigma factor, ComX, is degraded [7]. ComX controls late competence gene expression and is required for DNA transformation. The amount of ComX protein is fine-tuned by the MecA-ClpCP proteolytic machinery [32]. How ComX stability would be related to thermotolerance in the rcr peptide mutant strains is unknown, although comX mRNA is upregulated in response to elevated temperatures [24]. We also discovered that a ΔcomX mutant strain is exquisitely sensitive to 42 °C, similar to the various rcr strains reported here (Fig. S5). Thus, the effects of rcr mutations on ComX, or genes under the control of ComX, may modulate the ability of S. mutans to cope with heat stress.

Fig. 4.

Deletion of rcr-encoded peptides confers a loss of thermotolerance. (a) UA159, Δpep1, Δpep2 and Δpep12 strains were grown in BHI with a sterile mineral overlay at 42 °C and growth was measured using a Bioscreen C automated growth curve instrument. (b) Percentage of surviving S. mutans strains on BHI agar after incubation for the indicated times at 42 °C. At each time point, the number of viable cells was calculated and compared back to the original inoculum to determine the percentage change in bacterial cell number. (c) At 0, 24 and 48 h, the viability of a S. mutans Δpep12 strain was visualized with LIVE/DEAD BacLight bacterial viability stain and fluorescence microscopy. Live bacteria with intact cell membranes fluoresce green, whereas bacteria with compromised cell membranes, presumably dead, fluoresce red. (d) UA159 carrying the pIB184 plasmid and a strain overexpressing pep12 (pIB184-pep12) were grown in BHI (without antibiotic pressure) with a sterile mineral overlay at 42 °C and growth was measured using a Bioscreen C automated growth curve instrument. Graphs are derived from three biological replicates.

Deletion of the XrpA peptide, encoded within comX, restores growth at elevated temperatures

Within the comX structural gene of S. mutans is an additional gene, in an alternative reading frame, designated xrpA (Fig. 5a) [33]. XrpA (comX regulatory peptide A) is a 69 aa protein that adversely affects the development of competence and is required for optimal oxidative stress tolerance by S. mutans . In ΔrcrR-NP, the full-length comX mRNA and ComX all but disappear, and a truncated mRNA with a 5′ terminus within the comX coding sequence and containing the xrpA ORF is transcribed. Mutation of the start codon of, or introduction of a premature stop codon into, xrpA in a ΔrcrR-NP background restores the transformability of the strain, with a concomitant restoration of full-length comX transcripts and ComX protein levels [33]. Given the strong effect of the rcrR-NP mutation on comX and xrpA, we reasoned that XrpA could be a contributing factor to the thermosensitive phenotype of ΔrcrR-P. Indeed, the introduction of a single nucleotide substitution that disrupts the start site of xrpA in the ΔrcrR-P mutant background restored growth at 42 °C (Fig. 5b). However, the thermosensitive ΔrcrR-P mutant already has a full-length comX transcript and produces elevated levels of comX mRNA (upregulated 42-fold, Fig. S6) and ComX protein [33]. The restoration of growth at 42 °C following the introduction of the ΔxrpA mutant into the ΔrcrR-P genetic background is therefore not due to the restoration of full-length comX mRNA and must therefore be caused by the loss of production of XrpA. To investigate this further, we cultured an xrpA overexpressing strain, with xrpA under the control of the constitutive P23 promoter carried on pIB184 [31, 33], in rich medium at 42 °C. This strain, pIB184-xrpA, grew more slowly and reached a lower final yield than UA159 (Fig. 5c) at 42 °C. A strain overexpressing comX on a plasmid (pIB184-comX), but with the start codon of xrpA mutated (base 162, T>C), grew better than the xrpA-overexpressing strain (Fig. 5c). This result suggests that increased production of XrpA, which we attribute to increased expression of comX in the ΔrcrR-P mutant, contributes to the thermosensitive phenotype of the strain. While it is known that the loss of XrpA restores the comX transcript in the ΔrcrR-NP mutant, XrpA was recently shown to interact directly with ComR and is thought to fine-tune ComX levels in S. mutans by inhibiting ComR-dependent activation of transcription of comX [34]. It is currently unknown if the XrpA peptide modulates the activity of other transcriptional regulators in S. mutans . It is notable, though, that overexpression of xrpA impacts negatively on the ability of S. mutans to survive oxidative stress [33]. Therefore, an effect(s) of XrpA on some circuit other than genetic competence is probably responsible for the effects that loss of XrpA has on the thermotolerance of the ΔrcrR-P mutant.

Fig. 5.

Thermotolerance is restored by elimination of XrpA. (a) The xrpA ORF is located within the comX gene, which encodes the alternative sigma factor that controls genetic competence. (b) Growth measurements of UA159, ΔrcrR-P, ΔxrpA and ΔrcrR-P/ΔxrpA in BHI at 42 °C using a Bioscreen C automated growth curve instrument. (c) Growth measurements of UA159, pIB184-comX (ΔxrpA) and pIB184-xrpA (overexpressing xrpA) in rich media at 42 °C using a Bioscreen C. Graphs are derived from three independent experiments.

(p)ppGpp levels influence the thermosensitive phenotype of ΔrcrR-P

The tetra- and pentaphosphorylated guanosine alarmones, collectively abbreviated as (p)ppGpp, are intracellular signals used by bacteria to respond to nutritional stress [35]. Proper regulation of the metabolism of (p)ppGpp is essential for optimal stress tolerance [36]. S. mutans contains three enzymes that contribute in different ways to (p)ppGpp pools: RelA, RelP and RelQ [17]. RelA (also known as Rel) is a member of the RSH superfamily (RelA/SpoT homologue) that can synthesize and hydrolyze (p)ppGpp. The RelP and RelQ enzymes are small alarmone synthetases (SAS), with RelP being the major source of (p)ppGpp during the exponential phase of growth. The conditions under which RelQ has a significant effect on (p)ppGpp pools are not yet defined, but there is a relationship between acetate metabolism and expression of the four-gene relQ operon, which encodes RelQ, a phosphotransacetylase, an RNA-modifying enzyme and an NADP kinase [37]. The products of the RcrRPQ operon have a major impact on (p)ppGpp accumulation in S. mutans [8]. Therefore, we reasoned that the thermotolerance defect observed in ΔrcrR-P could be related to perturbations in the regulation of (p)ppGpp pools, which in turn lead to growth arrest and death at elevated temperatures. Also of note, there is the observation that the introduction of a ΔrelA mutation into a ΔrcrR-NP strain restores the ability of ΔrcrR-NP to become genetically competent [38]. Interestingly, a ΔrcrR-P/ΔrelA double mutant grew similarly to the wild-type at 42 °C (Fig. 6). Having made this observation, we next sought to determine if (p)ppGpp accumulation was similar in the wild-type and ΔrcrR strains. Heat shock at 42 °C did not lead to robust accumulation of (p)ppGpp in the wild-type or any of the other strains tested (Fig. 7). However, the levels of (p)ppGpp were elevated in the ΔrcrR-P strain versus S. mutans wild-type at both temperatures. Elevated cellular concentrations of (p)ppGpp in response to heat shock have been observed in both the Gram-positive bacterium Enterococcus faecalis (heat shock at 60 °C) [39] and the Gram-negative bacterium Escherichia coli (heat shock at 40 °C, 42 and 50 °C) [40, 41], indicating some conservation in this response. Deletion of rcrRPQ in a way that retained the expression of pep1/2 yielded strains that no longer had elevated basal levels of (p)ppGpp. This is important, as deletion of rcrRPQ – with or without concomitant elimination of the peptides – did not impact on thermotolerance; the mutants grew as well as S. mutans UA159 at 42 °C (Fig. S7).

Fig. 6.

Reversal of the thermosensitive phenotype of ΔrcrR-P through loss of RelA. Growth of UA159, ΔrcrR-P and ΔrcrR-P/ΔrelA in rich media at 42 °C using a Bioscreen C automated growth curve instrument. The graph is representative of three independent experiments.

Fig. 7.

Accumulation of (p)ppGpp in S. mutans and rcrR operon mutants after heat shock. (p)ppGpp accumulation in S. mutans strains at 37 °C and after heat shock at 42 °C. Cells were cultured in chemically defined medium to an OD₆₀₀ of 0.2, labelled with ³²P-orthophosphate and incubated at either 37 °C or 42 °C. After an hour, cells were harvested. Nucleotides were extracted using ice-cold 13 M formic acid with three freeze-thaw cycles. Supernatants were spotted on PEI–cellulose plates for thin-layer chromatography in 1.5 M KH₂PO₄. The identity of the migrating nucleotides is shown on the left. Images are representative of three independent replicates.

In E. coli , the increase in (p)ppGpp pools in response to heat shock is short-lived [40], so it is possible that the timepoints at which we measured (p)ppGpp were not those at which peak levels occurred. To address the issue in more detail, we examined whether rcrR mutants were able to accumulate (p)ppGpp as effectively as the parental strain. We used mupriocin, an inhibitor of isoleucyl-tRNA synthetase that causes a robust accumulation of (p)ppGpp accumulation in S. mutans UA159. Consistent with previous results [17], treatment of S. mutans UA159 with mupriocin greatly increased the amount of (p)ppGpp in cells, compared to untreated controls (Fig. 8). Again, S. mutans ΔrcrR-P exhibited higher basal levels of (p)ppGpp than UA159. However, following treatment with mupirocin, the ΔrcrR-P strain did not accumulate (p)ppGpp at greater levels than untreated controls (Fig. 8). The ΔrcrR-P /ΔrelA double mutant strain also failed to respond to mupirocin, but had lower levels of (p)ppGpp than the ΔrcrR-P strain, similar to that of a ΔrelA single mutant. A ΔrcrRPQ strain with intact pep1/2 peptides accumulated (p)ppGpp in response to mupirocin almost identically to the wild-type.

Fig. 8.

Accumulation of (p)ppGpp in S. mutans and rcrR mutants during mupirocin treatment. Exponentially growing cells were simultaneously labelled with ³²P-orthophosphate and treated with 500 ng ml⁻¹ mupirocin. After 1 h, cells were harvested. Nucleotides were extracted using ice-cold 13 M formic acid with three freeze-thaw cycles. Acid extracts were spotted on PEI–cellulose plates for thin-layer chromatography in 1.5 M KH₂PO₄. The identity of the migrating nucleotides is shown on the left. Images are representative of three independent replicates.

Our data point to a mechanism whereby intermediate levels, as opposed to higher than normal levels or no (p)ppGpp, lead to the growth defect observed in ΔrcrR-P. More specifically, deletion of relA in the ΔrcrR-P background restores basal (p)ppGpp levels to those seen in the wild-type, presumably due to loss of the synthetase activity of RelA, with the ability to grow at 42 °C concurrently restored. ΔrcrR-P is also unable to respond to isoleucyl-tRNA synthetase inhibition, but that is also the case for a ΔrelA mutant, which is able to grow at 42 °C. Therefore, it is not low, but rather higher basal levels of (p)ppGpp that may inhibit growth of the ΔrcrR-P strain. Although RelA has both (p)ppGpp synthetase and hydrolase activity, it is the SAS enzymes RelP and RelQ that are the primary contributors to basal (p)ppGpp levels in exponentially growing cells [17, 19]. RelA has an important role in fine-tuning the levels of (p)ppGpp in S. mutans owing to the allosteric modulation of its enzymatic activities that degrade or synthesize (p)ppGpp to maintain optimal levels for particular physiological situations. In ΔrcrR-P, it appears that the RelA (p)ppGpp enzymatic activity is not properly balanced, resulting in higher basal levels of (p)ppGpp during exponential growth. The amount of (p)ppGpp required to induce a growth arrest response is higher than the levels seen in ΔrcrR-P. Importantly, we also demonstrated that ΔrcrR-P responds to elevated temperature with death, not persistence (Fig. 2). Instead, the thermosensitivity of ΔrcrR-P may be related to a lack of proper control of core cellular processes caused by aberrant (p)ppGpp levels. Also of relevance here, elevated levels of (p)ppGpp levels are associated with decreased guanosine triphosphate (GTP) pools [42]. High (p)ppGpp levels with concomitant low GTP levels direct cells to growth arrest [42]. While lower levels of GTP might influence our results by affecting the efficiency of translation of genes with GTG start codons, GTP is not an allosteric effector of the global regulator CodY in S. mutans [43, 44]. Thus, there is no reason to believe that the thermotolerance phenotypes directly involve CodY.

In Gram-positive organisms, (p)ppGpp regulates both transcription and translation by altering the activity of enzymes involved in both of these processes [42, 45]. Without optimal control of gene transcription and protein synthesis, S. mutans ΔrcrR-P is unable to cope with heat shock or grow at elevated temperatures. The mechanism by which RelA activity is altered in ΔrcrR-P is unknown, but the possibility exists that the modification could be through a transcriptional, translational or post-translational mechanism exerted by the changes in RcrPQ and Pep1 and/or 2 levels in ΔrcrR-P. One avenue that we are exploring is the potential allosteric regulation of RelA by Pep1 or Pep2.

Conclusion

In summary, we have demonstrated that certain mutations that ablate, or alter the levels of, RcrR, P or Q, and the two small peptides encoded at the end of relQ (Pep1 and Pep2), cause cell death at elevated temperatures. We also observed relationships between the phenotype exhibited by ΔrcrR-P, XrpA production and RelA. This work provides further demonstration of the substantial complexity of stress tolerance in S. mutans and provides additional support for a critical link between stress tolerance, intercellular signalling and (p)ppGpp homeostasis (see Fig. 9 for a simplified model). Additional peptidomics research – combining bioinformatics, proteomic and genetic techniques – will be needed to fully understand the regulatory functions that unannotated peptides have in S. mutans and other bacteria.

Fig. 9.

Working model integrating the rcrRPQ operon of S. mutans with stress tolerance, competence development and (p)ppGpp metabolism. In this report we have added to a growing body of evidence that shows that S. mutans physiology is significantly disrupted when rcrRPQ/pep expression levels are aberrant (via either deletion or polarity). Our data, and previous reports, are consistent with (p)ppGpp levels being a central component of the phenotypes caused by rcrR mutations. How (p)ppGpp metabolism becomes disrupted is unclear and it may be caused by changes in the efflux of an unknown compound, uncontrolled competence gene expression, possible allosteric regulation by the Rcr peptides, or the combination of several disrupted pathways. Although S. gordonii rcrR mutants are slightly attenuated at increased temperature, the phenotype is much more dramatic in S. mutans . We attribute this difference to the peptides (they are not present in S. gordonii ) and the different competence regulation systems that the organisms employ. In S. mutans , competence gene regulation and (p)ppGpp metabolism appear to be connected, as mutations in both systems can restore the thermotolerance defects we have observed. This entanglement of various components allows S. mutans to fine tune its response to physiological or environmental conditions and decide whether to commit to competence, growth, or persistence in response to these signals.