Export of Rgg Quorum Sensing Peptides is Mediated by the PptAB ABC Transporter in Streptococcus Thermophilus Strain LMD-9

In streptococci, intracellular quorum sensing pathways are based on quorum-sensing systems that are responsible for peptide secretion, maturation, and reimport. These peptides then interact with Rgg or ComR transcriptional regulators in the Rap, Rgg, NprR, PlcR, and PrgX (RRNPP) family, whose members are found in Gram-positive bacteria. Short hydrophobic peptides (SHP) interact with Rgg whereas ComS peptides interact with ComR regulators. To date, in Streptococcus thermophilus, peptide secretion, maturation, and extracellular fate have received little attention, even though this species has several (at least five) genes encoding Rgg regulators and one encoding a ComR regulator. We studied pheromone export in this species, focusing our attention on PptAB, which is an exporter of signaling peptides previously identified in Enterococcus faecalis, pathogenic streptococci and Staphylococcus aureus. In the S. thermophilus strain LMD-9, we showed that PptAB controlled three regulation systems, two SHP/Rgg systems (SHP/Rgg1358 and SHP/Rgg1299), and the ComS/ComR system, while using transcriptional fusions and that PptAB helped to produce and export at least three different mature SHPs (SHP1358, SHP1299, and SHP279) peptides while using liquid chromatography-tandem mass spectrometry (LC-MS/MS). Using a deep sequencing approach (RNAseq), we showed that the exporter PptAB, the membrane protease Eep, and the oligopeptide importer Ami controlled the transcription of the genes that were located downstream from the five non-truncated rgg genes as well as few distal genes. This led us to propose that the five non-truncated shp/rgg loci were functional. Only three shp genes were expressed in our experimental condition. Thus, this transcriptome analysis also highlighted the complex interconnected network that exists between SHP/Rgg systems, where a few homologous signaling peptides likely interact with different regulators.

• PptAB exports signaling peptides from intracellular quorum sensing systems in Streptococcus thermophilus; • PptAB only exports mature signaling peptides; • in the S. thermophilus strain LMD-9, PptAB, Eep (a protease), and Ami (a peptide importer) are involved in a QS mechanism that controls the expression of genes located downstream all rgg genes but also of a few distal targets; and, secreted peptides [20], biofilms and lysozyme resistance [21][22][23], and mediate pathogenicity [24,25]. ComR regulators help to mediate competence and bacteriocin production [3,26]. Rgg regulators are associated with AIPs, called short hydrophobic peptides (SHPs); ComR regulators are associated with ComS peptides, called sigX-inducing peptides (XIPs) in their mature form. Eep causes the maturation of SHPs, as observed for one SHP in Streptococcus thermophilus [27]; for SHP2 and SHP3 in Streptococcus pyogenes [21]; and, for the RovS-associated SHP in Streptococcus agalactiae [24]. Eep also controls the maturation of ComS in S. thermophilus [28], but only plays a contributing role in S. mutans [29]. Finally, PptAB is involved in the export of SHP2 and SHP3 in S. pyogenes [30] and of the RovS-associated SHP in S. agalactiae [24]; however, it only contributes to the export of mature ComS in Streptococcus mutans [30]. S. thermophilus plays a major role in the food industry, because it is a starter employed in massive quantities to manufacture yoghurt, Swiss-type cheeses, and Italian-type cheeses. It is also a useful representative of the genus Streptococcus with which to study QS: S. thermophilus encodes many QS systems in its genome, including at least five different SHP/Rgg systems [27] and one ComS/ComR system [31]. We have deciphered two different SHP/Rgg systems in the S. thermophilus wild-type strain LMD-9. The first system, SHP/Rgg 1358 , controls the transcription of the shp 1358 gene and at least three genes that are involved in the production of a cyclic secreted peptide, called streptide, which are located downstream from the rgg 1358 gene [20,27,32]. The second system, SHP/Rgg 1299 , controls the transcription of the shp 1299 gene and a putative operon located downstream from the rgg 1299 gene whose function remains undetermined [33]. Because little is known about the process by which the AIPs of the two SHP/Rgg systems and the ComS/ComR system are exported, we characterized the role of the PptAB exporter in these three QS systems with transcriptional fusions and a LC-MS/MS approach for three SHP 1358 , SHP 1299 and SHP 279 . We used a transcriptomic approach to explore genome-level expression of the target genes in all functional RRNPP QS systems.

Bacterial Strains and Growth Conditions
Our study strain was the S. thermophilus wild-type strain LMD-9 (hereafter, the wild-type strain). It has two plasmids-plasmid 1 and plasmid 2 [34]. Table 1 lists the mutant strains that we constructed using this wild-type strain. S. thermophilus was grown at 37 • C or 42 • C in either M17 medium (Difco) supplemented with 10 g L −1 lactose (M17lac) or in a chemically defined medium (CDM) [35]. Escherichia coli was grown at 37 • C in Luria-Bertani broth with shaking. Agar (1.5%) was added to the media, as needed. When required, antibiotics were added to the media at the following final concentrations: erythromycin at 150 µg mL −1 for E. coli and at 5 µg mL −1 for S. thermophilus and kanamycin at 1 mg mL −1 for S. thermophilus.

DNA Manipulation and Sequencing
Restriction enzymes, T4 DNA ligase (New England Biolabs), and Phusion DNA polymerase (Finnzymes) were used in accordance with the manufacturers' instructions. Standard methods were used to carry out DNA purification, restriction digestion, PCR, ligation, and sequencing. Table S1 lists the oligonucleotides we used (Eurofins). The E. coli strain TG1 repA + was used as the host for the cloning experiments [36]. S. thermophilus was transformed while using natural competent cells [36] with the addition of synthetic competence peptide (ComS, LPYFAGCL) at a concentration of 1 µM when required or using electrocompetent cells [20] when the strains where not naturally competent. The plasmids used are listed in Table S2.

Construction of the Mutant Strains
The overlapping PCR method was used to delete the pptAB genes and replace them with an erythromycin cassette (erm) as follows. The erm cassette was amplified via PCR with the oligonucleotides Erm-F and Erm-R and using pG + host9 as a template [37]; it was fused to fragments located upstream and downstream from the pptAB genes, also using PCR. The upstream and downstream fragments were amplified with the oligonucleotides pptABe_up-F/pptABe_up-R and pptABe_down-F/pptABe_down-R, respectively, while using chromosomal DNA from the wild-type strain as a template. The resulting PCR fragment was used to transform the wild-type strain, leading to the construction of TIL1486 strain (pptAB::erm). The TIL1488 (blp::P comX -luxAB pptAB::erm), TIL1489 (blp::P shp1358 -luxAB pptAB::erm), and TIL1491 (blp::P shp1299 -luxAB aphA3 pptAB::erm) strains were constructed by transforming the CB001, TIL1165, and TIL1038 strains, respectively, with chromosomal DNA from the TIL1486 strain.
We also constructed a mutant, the TIL1558 strain (∆pptAB), in which we deleted an internal fragment of the pptAB genes via a double crossover event that involves the plasmid pG + host9. Thus, this mutant did not have an antibiotic resistance cassette. Briefly, the upstream and the downstream fragments of the pptAB genes were obtained via amplification using the primer pairs pptAB_up-F/pptAB_up-R and pptAB_down-F/pptAB_ down-R. The overlapping PCR method was used to fuse both fragments. The resulting fragment was double digested using the restriction enzymes EcoRI and KpnI and then ligated into pG + host9 between the EcoRI and KpnI restriction sites. The resulting plasmid, pG + host9::updown.pptAB, was used to transform competent cells of the wild-type strain. The integration and excision of this plasmid [37] then generated the TIL1558 strain. In this mutant, the promoter of the pptA gene (including its ATG) and the last 74 bp of the pptB gene were conserved.
The presence of the plasmid pBV5030::P 32 -shp 1358 allowed for the expression of the shp 1358 gene under the strong P 32 promoter. This plasmid was used to transform the wild-type strain and the TIL1558 strain, resulting in the TIL1559 and TIL1563 strains, respectively.
A variety of strains were created in which the comR gene was replaced with a kanamycin resistance cassette. First, the upstream and downstream regions of the comR gene were amplified using the primers comR_up-F/comR _up-R and comR _down-F/comR _down-R, respectively. The kanamycin cassette (aphA3) was amplified via PCR using the primers AphA3-F/AphA3-R and plasmid pKa as the template [38]. This cassette was inserted between the upstream and downstream fragments of the comR gene via overlapping PCR. The resulting fragment was double digested with the restriction enzymes EcoRI and XhoI and then ligated into pG + host9 between the related restriction sites, creating the plasmid pG + host9::updown.comRaphA3. The linearized plasmid that was obtained via digestion with the restriction enzyme NdeI was used to naturally transform the wild-type, TIL773 (∆eep), and TIL1558 (∆pptAB) strains in order to generate the TIL1560 (comR::aphA3), TIL1561 (∆eep comR::aphA3), and TIL1562 (∆pptAB comR::aphA3) strains, respectively. The electrotransformation of the TIL883 strain (∆amiCDE) with the plasmid pG + host9::updown.comRaphA3 and the subsequent processes of integration and excision generated the TIL1564 strain (∆amiCDE comR::aphA3).
In order to study the expression of the pptAB genes, a derivative of pGICB004a, pGICB004a::P pptAB , was constructed, as follows. The promoter of the pptAB genes was amplified using the oligonucleotides PpptAB-SpeI and PpptAB-EcoRI; double digested with the restriction enzymes SpeI and EcoRI; and finally, ligated into pGICB004a between the related restriction sites. ScaI-linearized pGICB004a::P pptAB was used to transform the wild-type strain, leading to the TIL1557 strain (blp::P pptAB -luxAB aphA3).
Finally, the plasmid pBV5030::P 32 was used to transform the TIL1558 strain, resulting in the TIL1565 strain.
All of the constructions were verified by PCR and sequenced when necessary.

Mass Spectrometry Analysis
Liquid chromatography-tandem mass spectrometry (LC-MS/MS) was performed by coupling an LTQ-Orbitrap (Thermo Fisher Scientific, Waltham, MA, USA) with an UltiMate™ 3000 RSLCnano System (Thermo Fisher Scientific); analyses were carried out at the PAPPSO platform (MICALIS, INRAE, Jouy-en-Josas, France). The strains were grown in CDM until they reached an OD 600 of 0.8, which corresponded to the end of the exponential growth phase. Three independent cultures (i.e., biological replicates) were grown for each strain studied. Supernatants were recovered via centrifugation (at 10,000 rpm and 4 • C for 10 min). Samples (4 µL) were loaded at 20 µL min −1 on a precolumn (µ-Precolumn, 300 µm × 5 mm, C18 PepMap100, 5 µm, 100 Å, Thermo Fisher) and then washed with loading buffer. After 3 min, the precolumn cartridge was connected to the separating column (Acclaim PepMap ® , 75 µm × 500 mm, C18, 3 µm, 100 Å, Thermo Fisher). The flow rate was 300 nL min −1 , and elution employed a linear gradient of acetonitrile (0.8% per min) in 0.1% formic acid. The eluted peptides were ionized with a spray voltage of 1.3 kV. A full MS scan (range: 400-2000 m/z) was performed while using the LTQ-Orbitrap. The full MS scans were automatically calibrated using the lock mass option and two dimethylcyclosiloxanes (391.284 and 536.165). The MS/MS analysis was carried out employing the collision-induced dissociation (CID) technique (40% collision energy), which was applied to three charge states (+1, +2, and +3) and the eight most abundant ions (dynamic exclusion of 30 s; minimal intensity threshold of 5 × 10 2 ). We manually extracted the ion current signals (XIC) with Xcalibur Qual Browser (v. 3.0.63). We analyzed the masses of mature SHP 1358 (EGIIVIVVG), 898.561 Da, SHP 1299 (DIIIFPPFG), 1018.561 Da, SHP 279 (EGIIVIGVG), 856.514 Da, and the streptide (AKGDGWKVM with a cyclization between the first K and the W), and 495.247 Da at a high level of resolution (10 ppm). The presence of these mature forms was confirmed when we detected at least three masses that corresponded to fragments that were produced by MS2 fragmentation (accuracy of 0.5 Da).

Luciferase Assays
For the pptAB genes expression study, TIL1557 was the strain used ( Figure 1, Table 1). For the results that are described in Figure 2, strains used were CB001, TIL1488, TIL1165, TIL1489, TIL1038, and TIL1491 ( Table 1). The luciferase assays were carried out, as follows. The cells were grown overnight at 37 • C in CDM or M17lac. The cultures were diluted to a final OD 600 of 0.05 in 1 mL of CDM or M17lac. As necessary, synthetic peptides, SHP 1358 (EGIIVIVVG), SHP 1299 (DIIIFPPFG), or ComS (LPYFAGCL), prepared in DMSO for the two SHPs or water for ComS were added at a final concentration of 1 µM. We transferred 250 µL of these diluted cultures to the wells of a covered sterile white microplate with a transparent bottom (Greiner). The cultures' OD 600 and luminescence (as expressed in relative light units [RLU]) values were assessed at 37 • C in an Infinite M200 spectroluminometer (Tecan), as previously described [31].

Total RNA Extraction
The strains were grown in CDM at 42 °C, and samples were taken from the cultures at the end of the exponential growth phase (OD600 = 1). TIL1560, TIL1561, TIL1562, and TIL1564 were the strains used ( Table 1). The cells were harvested via fast centrifugation (at 13,400 g for 15 s); the supernatants

Total RNA Extraction
The strains were grown in CDM at 42 °C, and samples were taken from the cultures at the end of the exponential growth phase (OD600 = 1). TIL1560, TIL1561, TIL1562, and TIL1564 were the strains used ( Table 1). The cells were harvested via fast centrifugation (at 13,400 g for 15 s); the supernatants

Total RNA Extraction
The strains were grown in CDM at 42 • C, and samples were taken from the cultures at the end of the exponential growth phase (OD 600 = 1). TIL1560, TIL1561, TIL1562, and TIL1564 were the strains used ( Table 1). The cells were harvested via fast centrifugation (at 13,400 g for 15 s); the supernatants were removed; and, the bacterial pellets were frozen in liquid nitrogen and stored at −80 • C. The cells were then lysed in phenol-chloroform 5:1 (v/v) with a FastPrep FT120 tissue grinder (Thermo Savant, Holbrook, NY, USA). Two cycles of cell disruption were performed (both at 6.5 m s −1 for 45 s). Cellular debris was pelleted by centrifugation (at 16,000 g and 4 • C for 10 min.) and the supernatants were collected. The total RNA extraction was then performed using the Direct-Zol RNA MiniPrep Kit (Zymo Research, Irvine, CA, USA) in accordance with the manufacturer's instructions. Genomic DNA was eliminated with a DNA-free DNA Removal Kit (Invitrogen, Waltham, MA, USA). Finally, RNA quantity and quality were assessed using a Qubit RNA HS Assay Kit (Thermo Fisher Scientific, USA) and a Bioanalyzer system (Agilent, Santa Clara, CA, USA), respectively. For each strain, RNA was extracted from the three biological replicates.

RNA Sequencing and Transcriptome Analysis
RNA sequencing was performed at I2BC (www.i2bc.paris-saclay.fr), a sequencing facility. The total transcript libraries were created using a Stranded mRNA Library Prep Kit (Illumina). We followed the manufacturer's instructions, except in the initial steps (i.e., up until RNA fragmentation), where polyA purification was replaced by rRNA depletion using a Ribo-Zero Kit (Illumina). Quality control was carried out on the libraries using a Bioanalyzer HS DNA Kit (Agilent) and a dsDNA HS Assay Kit (Thermo Fisher Scientific, USA). The libraries were then sequenced (single-end reading, read length of 75 bp) using an Illumina NextSeq 500 System. The raw reads were demultiplexed, and fastq files were created from the output with bcl2fastq2 (v. 2.18.12; Illumina). The adapters were subsequently trimmed out of the reads using Cutadapt (v. 1.15; [39], and read quality was then verified via FastQC (v. 0.11.5; http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). The clean reads in the total transcript libraries were then mapped against the S. thermophilus LMD-9 genome while using BWA-aln (v. 0.6.2-r126; [40]). The LMD-9 genome was processed beforehand using BactgeneSHOW to predict short CDSs [41,42]. We used the following option parameters: -m 4C_si -rbs m1 -duprev -cdst 0.01. The dataset included predicted CDSs (45-180 nt) that were not present in the GenBank annotation for LMD-9 (chromosome: CP000419.1; plasmid 1: CP000420.1; plasmid 2: CP000421.1) [34] (Supplementary  Table S3). The raw read counts were calculated using the featureCounts tool (option -s 1) in the SubRead package (v. 1.5.2; [43]). We employed DESeq2 (v. 1.18.1; [44]) to carry out normalization and differential analysis. First, the total read counts were normalized using the median of ratios method [45]. Subsequently, gene expression was compared between our reference strain, TIL1560 (comR::aphA3), and the TIL1561 (∆eep comR::aphA3), TIL1562 (∆pptAB comR::aphA3), and TIL1564 (∆amiCDE comR::aphA3) strains to examine the roles of the eep, pptAB, and amiCDE genes. The resulting p-values were adjusted using the false discovery rate method [46]. In our further analyses, we only focused on genes with adjusted p-values ≤ 0.01 and for which there was at least a 1.9-fold difference between the reference strain and the other strains. The RNA-seq raw data from this study have been deposited in the ArrayExpress database at EMBL-EBI under accession number E-MTAB-9568 (www.ebi.ac.uk/arrayexpress).

Bioinformatic Tools
We employed BlastP (v. 2. 9.0) with standard parameters and the PptAB proteins of S. mutans UA159 as baits to identify the PptAB homologous proteins of S. thermophilus LMD-9. PptA of S. thermophilus was found to be encoded by the Ster_RS07730 gene (previous nomenclature: Ster_1572). PptB of S. thermophilus was found to be encoded by the Ster_RS07725 gene (previous nomenclature: Ster_1571). We used the 65 complete genome sequences available in the NCBI database (https: //www.ncbi.nlm.nih.gov/genome) to compare the amino acid sequences of PptA and PptB inside the S. thermophilus species.

The PptAB Genes Are Expressed in S. thermophilus and Required for the Positive Control of Three Signaling Systems
PptAB is an ABC transporter that is made of a membrane-bound cytoplasmic ATP-binding protein, PptA, which provides the energy for substrate translocation, and a transmembrane protein, PptB, which forms the pore. It was first identified in E. faecalis [19]. However, shortly thereafter, orthologous genes encoding exporters with similar functions were found in S. pyogenes, S. mutans [30] and S. agalactiae [24]. We used the PptAB proteins in the S. mutans UA159 strain to identify any corresponding proteins in the S. thermophilus LMD-9 strain [47]. PptA of S. thermophilus shared 77% identity with PptA of S. mutans (88% sequence similarity) and PptB shared 51% identity with PptB of S. mutans (70% sequence similarity). The PptA and PptB proteins are highly conserved in the species S. thermophilus, estimates of shared identity range between 98 and 100 for PptA and between 97 and 100 for PptB.
First, we characterized the expression patterns of the pptAB genes in two media, CDM and M17lac, while using a P pptAB -luxAB transcriptional fusion in the TIL1557 strain. We found that gene expression was constitutive throughout the exponential growth phase and it was higher in CDM than in M17lac ( Figure 1). Second, we deleted the pptAB genes and characterized the phenotype of the resulting strains. To date, three RRNPP signaling systems have been studied in the LMD-9 strain of S. thermophilus. They involve three different AIPs that are potentially exported by PptAB, but only two systems have been associated with a phenotype.
The first is a SHP/Rgg system that is based on a nine-amino-acid-long hydrophobic AIP, SHP 1358 (EGIIVIVVG). This system helps to produce the streptide encoded by the Ster_RS010575 gene [27]. Therefore, we measured streptide production by the TIL1179 strain, which had a wild-type background and the TIL1565 strain, which bore the ∆pptAB mutation using LC-MS/MS; streptide levels were quantified by measuring the integrated area under the curve in extracted-ion chromatograms for the three biological replicates per strain. The TIL1179 strain had a mean (±SD) streptide level of 2.7 × 10 8 (±4 × 10 7 ). In contrast, the streptide level was < 7 × 10 4 in the TIL1565 strain. This result indicates that streptide was only produced at trace levels when the pptAB genes were inactivated.
The second system is based on the mature form of ComS (IAILPYFAGCL), which is also hydrophobic, but composed of eleven amino acids [28]. It controls the activity of ComR, a transcriptional regulator that is involved in mediating competence. Indeed, the ComS/ComR system positively controls the transcription of the comX gene. ComX is an alternative sigma factor that controls the transcription of genes encoding the transformasome, a protein complex essential for competence [31,48]. Therefore, we assessed the competence of the TIL1558 strain, a ∆pptAB mutant, when it was grown in CDM with pG + host9. No transformants were detected, which indicated that the mutant had completely lost the ability to transform naturally. However, the transformability of strain TIL1558 was restored with the addition of synthetic ComS in the CDM medium, as demonstrated with the construction of strains TIL1562, TIL1563, and TIL1565 ( Table 1). The third RRNPP system is the SHP/Rgg 1299 system. It is known to be functional, but its target function is unknown.
Thus, we wanted to demonstrate that streptide production and competence were specifically influenced by the elimination of expression of the RRNPP QS target genes in the ∆pptAB mutant. Therefore, we quantified the expression of the comX gene, a target of the ComS/ComR system, the shp 1358 gene, a target of the SHP/Rgg 1358 system, and also the shp 1299 gene, a target of the SHP/Rgg 1299 system. The mature form of its AIP, SHP 1299 (DIIIFPPFG), is similar to SHP 1358 because it is also nine amino acids long and mostly hydrophobic. To this end, we used transcriptional fusions between the promoter of these three genes and the luxAB genes encoding a luciferase and then explored gene expression in the wild-type versus the pptAB::erm background. In the three systems, the genes encoding the AIPs, shp genes, or comS gene, are system targets, creating a positive feedback loop and resulting in the marked expression of all the target genes [27,31,33]. Interestingly, although the three systems function similarly because they involve the export and reimport of AIPs, the three target genes were not expressed at the same moment during strain growth (Figure 2). The comX gene and the shp 1299 gene were expressed at the beginning and the middle of the exponential growth phase, respectively (Figure 2A,E). In contrast, the shp 1358 gene displayed a bimodal pattern during growth: there was a lower peak of expression early on and then a higher peak at the end ( Figure 2C). These differences in the timing of expression indicate that, although the mechanisms that underlie each system are similar, they are not identical. In the pptAB::erm background, the expression of the comX and the shp 1299 genes was completely absent (Figure 2A,E) and it dropped to trace levels for the shp 1358 gene. In conclusion, these results show that PptAB is necessary for the expression of the three target genes and, therefore, for the functionality of the three QS systems.
We added the cognate mature AIPs ComS, SHP 1358 , and SHP 1299 to cultures of the strains that contained the three transcriptional fusions in the pptAB::erm background in order to verify PptAB's predicted role in AIP export. The addition of the cognate AIPs immediately triggered the expression of the comX, or shp 1299 genes at levels that are similar than those obtained when the pptAB genes were functional ( Figure 2B,F) and at a level six times higher for the shp 1358 gene ( Figure 2D). Consequently, the functionality of the three QS systems was fully restored, confirming that PptAB plays a role in signal production, not signal detection, for these three systems. We deleted the amiCDE genes in strains TIL1488 (blp::comX-luxAB pptAB::erm), TIL1489 (blp::Pshp 1358 -luxAB pptAB::erm), and TIL1491 (blp::shp 1299 -luxAB pptAB::erm). Using luciferase assays, we checked that the expression of the comX, shp 1358 and shp 1299 genes remained absent in the presence of the cognate AIP (data not shown), which confirmed that Ami is the importer of the AIPs in strain TIL1488, TIL1489, and TIL1491.

PptAB is Involved in the Export of SHP 1358 and SHP 1299
To determine whether PptAB helps export the AIPs in S. thermophilus, we used LC-MS/MS to ascertain whether mature forms of SHP 1358 and SHP 1299 were present in the wild-type strain and absent in the ∆pptAB mutant (TIL1558). We did not look for ComS, which is not detectable via LC-MS/MS in the supernatant of the wild-type strain [28].
As expected based on prior research, we found mature forms of SHP 1358 and SHP 1299 in the wild-type strain [27,33] (Figure S1) but not in the ∆pptAB mutant. This result fits with our prior research showing that the shp 1358 and shp 1299 genes were not expressed in the pptAB mutant due to the disruption of the positive feedback loop.
To further characterize the key role played by PptAB as a peptide exporter, we delved deeper into how the mature form of SHP 1358 is produced. We uncoupled the expression of the shp 1358 gene from the presence of PptAB by introducing the plasmid pBV5030::P32-shp 1358 into the ∆pptAB mutant, creating the TIL1563 strain. Because of the plasmid, shp 1358 expression was under the control of a strong promoter and it was not reliant on the autoregulation loop generated by the SHP/Rgg 1358 system. As a result, there was constant independent synthesis of a SHP 1358 -like precursor, which, due to technical constraints, contained a glycine residue between the methionine at position 1 and the lysine at position 2 of the wild-type sequence [27]. This 24-amino-acid-long precursor (hereafter, SHP* 1358 ) can serve as the basis for the 9-amino-acid-long mature wild-type peptide, as seen below in the results for the TIL1213 strain. We did not detect the presence of mature SHP 1358 in the TIL1563 strain, which demonstrated that the PptAB transporter is required for the peptide's production.

PptAB Controls Transcription of the Genes Located Downstream from the Rgg Genes in S. thermophilus
Because PptAB helps export the AIPs that control the activity of transcriptional regulators, we used a deep sequencing approach (RNAseq) to identify all of the genes whose expression is controlled by PptAB. The genome of LMD-9 contains three complete shp/rgg loci in addition to the shp/rgg 1358 and shp/rgg 1299 loci, as well as a locus with truncated shp and rgg genes (sCDS_305/Ster_RS03240) ( Figure S2A). Thus, we hypothesized that other genes in the near or far vicinity of the rgg genes could also be controlled by this transporter. The ComR regulon, partially controlled by PptAB through the export of the mature form of its cognate AIP ComS, was previously identified in S. thermophilus [26,48]. Because the regulon encompasses more than 100 genes, we chose to work in a ∆comR background, so that we could identify PptAB regulon in the absence of the genes that are regulated by the ComS/ComR system. Therefore, we compared gene expression between the TIL1560 (comR::aphA3) and TIL1562 (comR::aphA3 ∆pptAB) strains. The shp 1299 and shp 1358 genes, the two distinct targets of the SHP/Rgg systems, were highly expressed at the end of the exponential growth phase (Figure 1), so samples for mRNA extraction were obtained at this stage (OD 600 = 1). Genes were upregulated in the TIL1520 strain compared to in the TIL1562 strain (Table 2; Figure 3). Except sCDS_305/Ster_RS03240 locus ( Figure S2A), the expression of all genes located downstream of the other five rgg genes and some, but not all of their cognate shp genes were controlled by PptAB. Hereafter, these genes will be referred as the proximal target genes of the Rgg regulators. We discovered two distal target genes (i.e., not located near a shp/rgg locus) that were regulated by PptAB: one short CDS (CDS_560) that was poorly expressed and one operon (Ster_RS04380-90). The short CDS has an unknown function, and the operon contains genes that are annotated as encoding enzymes that are involved in cysteine metabolism.   Only genes with an adjusted p-value ≤ 0.01 and an at least 1.9-fold change were examined; NR, not significantly regulated (adjusted p-value > 0.01); ND, not determined; * Average of normalized read counts from the three biological replicates.

Figure 3.
Organizational schematic of the genes controlled by PptAB. The genes whose expression was upregulated in the TIL1560 strain (comR::aphA3) compared to the TIL1562 strain (comR::aphA3 ΔpptAB) are shown in blue. This set of genes was controlled by PptAB but did not include genes regulated by ComR, which controls competence. ORF orientation within each cluster is indicated by the direction of the arrowhead (not to scale). Genes encoding SHPs have hatched shading. Genes that were not regulated by PptAB are in gray. Genes were named and described in accordance with their GenBank annotations, to which we added in-lab annotations for the short genes (see the Materials and methods section and Table S3).
We could distinguish three classes of genes located downstream from the rgg genes. First, there were two sets of highly expressed genes (based on the average of the normalized read counts) located downstream from the two rgg genes already studied, rgg1299 and rgg1358; they were strongly controlled by PptAB and displayed greater than 45-fold changes in expression ( Table 2). The genes downstream from the rgg1299 gene were all transcribed in the same direction, but, the greater their proximity, the higher their level of transcription and fold change. The operon's initial genes encode an enzyme (annotated oligopeptidase F), a transporter, and a peptide. The function of these genes is unknown. The operon's later genes encode two peptides, an ABC transporter, and a two-component system whose function seems to be linked to rhamnose-glucose polysaccharide synthesis induced by bacitracin stress [49]. The genes downstream from rgg1358 form two transcriptional units. The first unit was transcribed in the same direction as rgg1358 and it was strongly controlled by PptAB; it encodes proteins that are involved in the production of the streptide [20,32] and two additional proteins with unknown functions. The second unit was transcribed in the opposite direction and it was weakly controlled by PptAB; it encodes peptides or small proteins with unknown functions. The transcription of the two shp genes at these loci was also strongly controlled by PptAB, as had already been demonstrated via the transcriptional fusions with the luciferase reporter ( Figure 2).
The second class of genes involved two shp/rgg loci, sCDS_273/Ster_RS07530 and sCDS_279/Ster_RS09420. For this gene class, gene expression was relatively high in the TIL1560 strain, with fold changes ranging between 13 and 29. The shp gene at the sCDS_273/Ster_RS07530 locus was not expressed. The shp gene at the sCDS_279/Ster_RS09420 locus was weakly expressed and weakly controlled by PptAB (1.9-fold change). The sole proximal target gene in the sCDS_273/Ster_RS07530 system encodes a transporter with an unknown function. The proximal target genes in the sCDS_279/Ster_RS09420 system are similar to those at the streptide locus, with a gene encoding a SAM radical enzyme, a gene encoding a transporter, and two genes encoding peptides with unknown functions. The third class of genes was associated with the locus sCDS_265/Ster_RS04620. The genes were weakly expressed and controlled by PptAB (three-fold change). The shp gene at this locus was not expressed. Again, the proximal target genes were similar to those at the streptide locus, with a gene encoding a SAM radical enzyme, a gene encoding a transporter, two genes encoding peptides, a gene encoding an Fe-S oxido-reductase, and a gene encoding an arginase. Organizational schematic of the genes controlled by PptAB. The genes whose expression was upregulated in the TIL1560 strain (comR::aphA3) compared to the TIL1562 strain (comR::aphA3 ∆pptAB) are shown in blue. This set of genes was controlled by PptAB but did not include genes regulated by ComR, which controls competence. ORF orientation within each cluster is indicated by the direction of the arrowhead (not to scale). Genes encoding SHPs have hatched shading. Genes that were not regulated by PptAB are in gray. Genes were named and described in accordance with their GenBank annotations, to which we added in-lab annotations for the short genes (see the Materials and methods section and Table S3).
We could distinguish three classes of genes located downstream from the rgg genes. First, there were two sets of highly expressed genes (based on the average of the normalized read counts) located downstream from the two rgg genes already studied, rgg 1299 and rgg 1358 ; they were strongly controlled by PptAB and displayed greater than 45-fold changes in expression ( Table 2). The genes downstream from the rgg 1299 gene were all transcribed in the same direction, but, the greater their proximity, the higher their level of transcription and fold change. The operon's initial genes encode an enzyme (annotated oligopeptidase F), a transporter, and a peptide. The function of these genes is unknown. The operon's later genes encode two peptides, an ABC transporter, and a two-component system whose function seems to be linked to rhamnose-glucose polysaccharide synthesis induced by bacitracin stress [49]. The genes downstream from rgg 1358 form two transcriptional units. The first unit was transcribed in the same direction as rgg 1358 and it was strongly controlled by PptAB; it encodes proteins that are involved in the production of the streptide [20,32] and two additional proteins with unknown functions. The second unit was transcribed in the opposite direction and it was weakly controlled by PptAB; it encodes peptides or small proteins with unknown functions. The transcription of the two shp genes at these loci was also strongly controlled by PptAB, as had already been demonstrated via the transcriptional fusions with the luciferase reporter ( Figure 2).
The second class of genes involved two shp/rgg loci, sCDS_273/Ster_RS07530 and sCDS_279/Ster_RS09420. For this gene class, gene expression was relatively high in the TIL1560 strain, with fold changes ranging between 13 and 29. The shp gene at the sCDS_273/Ster_RS07530 locus was not expressed. The shp gene at the sCDS_279/Ster_RS09420 locus was weakly expressed and weakly controlled by PptAB (1.9-fold change). The sole proximal target gene in the sCDS_273/Ster_RS07530 system encodes a transporter with an unknown function. The proximal target genes in the sCDS_279/Ster_RS09420 system are similar to those at the streptide locus, with a gene encoding a SAM radical enzyme, a gene encoding a transporter, and two genes encoding peptides with unknown functions. The third class of genes was associated with the locus sCDS_265/Ster_RS04620. The genes were weakly expressed and controlled by PptAB (three-fold change). The shp gene at this locus was not expressed. Again, the proximal target genes were similar to those at the streptide locus, with a gene encoding a SAM radical enzyme, a gene encoding a transporter, two genes encoding peptides, a gene encoding an Fe-S oxido-reductase, and a gene encoding an arginase.
We found very similar expression levels and fold changes when we compared gene expression across the TIL1560 (comR::aphA3), TIL1561 (∆eep comR::aphA3), and TIL1564 (∆amiCDE comR::aphA3) strains (Tables S5 and S6). The differences mainly involved short CDSs encoding genes that were poorly expressed and for which regulation resulted in a three-fold change at most (sCDS_180, sCDS_560, sCDS_279, and sCDS_854). Only one gene (Ster_RS06940) was found to be downregulated in the eep mutant (Table S5); it encodes a peptide-binding protein of the Ami transporter and it was expressed at high levels in the TIL1560 strain. The expression of this gene was unaffected by the amiCDE and pptAB mutations. We also found that one gene (Ster_RS09420) was upregulated 1.9 fold in the amiCDE mutant, but it was poorly expressed in general (data not shown). Furthermore, it is important to note that none of the genes in plasmid 1 were expressed in the TIL1561 strain. Upon PCR verification, we discovered that this strain had lost this plasmid during the construction process (data not shown).
Finally, we observed that the shp 279 gene was expressed at very low levels. As a result, we questioned whether the expression levels were sufficient to allow for the production of mature SHP 279 in the wild-type strain. We found that the mature form of SHP 279 was indeed present in the wild-type strain ( Figure S3) as well as in the TIL883 strain (∆amiCDE) (data not shown). Interestingly, it was absent in the TIL773 (∆eep) and TIL1558 (∆pptAB) strains.

Discussion
PptAB has been identified as the exporter of the pheromones of PrgX in E. faecalis [19] and of Rgg and ComR in pathogenic streptococci S. pyogenes, S. mutans [30] and S. agalactiae, [24]. PptAB also exports peptides in S. aureus via a mechanism that is similar to the one that produces the pheromones in E. faecalis. However, the biological function of these peptides remains unclear [50,51]. The role of PptAB in the starter species S. thermophilus had previously remained unexplored, even though this bacterium has a comS/comR locus and it is one of the streptococcus species with S. pneumoniae with the greatest number of paralogous shp/rgg loci. In this study, we used genetic approaches and mass spectrometry to show that PptAB can export the mature AIPs of four different systems: ComS/ComR, SHP/Rgg 1299 , SHP/Rgg 1358 , and SHP/Rgg 9420 . Focusing on the SHP/Rgg 1358 system, we found direct evidence, namely the presence of mature SHP 1358 , that PptAB acts as a peptide exporter of this specific peptide.
Here, we studied the mature forms of AIPs in S. thermophilus. These peptides are hydrophobic and short in length, ranging in size from mime amino acids (the SHPs) to 11 amino acids (the mature ComS). These characteristics are shared by PptAB-transported AIPs in other streptococci. In the wild-type strain, we found no traces of the complementary part of mature SHP 1358 , a short hydrophobic peptide (14 amino acids long) that should be released after the cleavage of the precursor; we also did not observe any fragments of this complementary part. This result raised questions regarding the specificity of the transporter. One hypothesis is that PptAB only exports mature SHP 1358 because its activity focuses on mature AIPs. Another hypothesis is that there is interplay between PptAB and the protease Eep, which is known to be involved in the maturation of these peptides [21,24,27]. Eep is a transmembrane zinc-metalloprotease whose active cleavage site is thought to occur in a transmembrane domain [52]. It is possible that only mature SHPs are directly delivered to PptAB. If such is the case, the fate of the complementary part remains unknown.
Most commonly, AIPs of Gram-positive bacteria are exported either by the general Sec-dependent pathway or by dedicated peptidase-containing ABC transporters that can export and mature double glycine signaling peptides, which play a role in competence or provide competitive fitness advantages in complex environments [26,53,54]. The PptAB export mechanism represents an intermediate solution that is neither highly generalized nor highly specific. Further research should explore the potential evolutionary benefits of this solution and, more importantly, assess whether it is a point of control in QS systems. While the expression of the pptAB genes seems constitutive, more work is needed to determine whether PptAB activity is somehow regulated.
The RNAseq approach was used to identify the genes whose expression was controlled by PptAB, Eep, and AmiCDEF. All of the genes located downstream from the five non truncated rgg gene are part of the targets genes. Our findings suggest that non-truncated shp/rgg loci are functional and involved in the actions taken by PptAB, Eep, and AmiCDEF to regulate these target genes. Nevertheless, further work is needed in order to confirm that the expression of these target genes is indeed controlled by their cognate rgg gene. These target genes were generally expressed and regulated at similar levels by the three proteins PptAB, Eep, and AmiCDEF, indicating that these target genes might belong to a single transcriptional unit. There were a few exceptions: for example, when the proximal targets of SHP/Rgg 1299 were examined, the genes farthest from the rgg gene (Ster_RS06385-Ster_RS06360) were poorly expressed and regulated. However, this effect may have resulted from readthrough transcription dynamics, rather than from issues with QS regulation. Surprisingly, these genes are also part of the ComR regulon [26]. Although some of these loci had been previously studied, using the RNAseq approach, we were able to identify additional proximal targets, such as Ster_RS06680 and Ster_RS06670 for the shp/rgg 1358 locus, and Ster_sCDS172 for the shp/rgg 1299 locus. Ster_RS06680 and Ster_RS06670 have been annotated as SPASM-domain-containing protein and sodium transporters, respectively. Here, we discovered that they belong to the str operon, but further work is needed in order to clarify their role in streptide production. The function of Ster_sCDS172 is also unknown. We additionally found significantly expressed distal target genes with a three-gene operon linked to cysteine synthesis. Research is currently underway to identify the SHP/Rgg system that is associated with this operon and to confirm the operon's function. To our knowledge, it is one of the few distal target gene of a SHP/Rgg system identified to date, although we must still verify that transcriptional control is direct. Indeed, few RNAseq approaches have been developed to exhaustively identify SHP/Rgg system targets in streptococci. These experiments highlighted genes located near the rgg genes in Streptococcus mitis [55], Streptococcus pneumoniae [23], or S. pyogenes [21].
The identity and expression patterns of those genes regulated by PptAB, Eep, and Ami were strikingly similar. This finding indicates that an equivalent role is being played in export, maturation, and reimport. It is highly likely that the deletion of these genes would completely inactivate the mechanism in action, leading to similar results for target gene transcription. However, a closer look at the expression of the shp genes in the wild-type strain versus the ∆pptAB, ∆eep, and ∆amiCDE background revealed the existence of three different regulation scenarios.
First, the shp 1299 and shp 1358 genes were expressed at high levels, and their expression was clearly controlled by the three proteins. The result was a positive feedback loop and marked expression of the proximal target genes at the end of the exponential growth phase in the wild-type strain background. These results correspond to a scenario, in which SHP is exported by PptAB, matured by Eep, and reimported by Ami. Consequently, the activity of the associated Rgg regulator is controlled, and there is the transcription of the genes located downstream from the rgg gene.
Second, the shp 273 and shp 265 genes were not expressed. This result suggests that a SHP encoded by another locus controls the activity of Rgg 7530 and Rgg 4620 . We feel that there is more support for this hypothesis than for a hypothesis that involves the involvement of a non-cognate SHP/Rgg complex, because all of the rgg genes in the wild-type strain were expressed at similar levels indicating that the genes are likely functional also this deserves further validation. The proximal target gene in the SHP 273 /Rgg 7530 system was expressed at high levels and regulated by PptAB, Eep, and Ami, whereas the proximal target genes of SHP 265 /Rgg 4620 were expressed at lower levels and subject to less regulation by the three proteins. Interestingly, SHP 273 is nearly identical to SHP 1358 : there is a single amino acid difference, a spot occupied by a valine in SHP 1358 is occupied by a leucine in SHP 273 . SHP 265 and SHP 1358 differ by two amino acids: spots occupied by a glycine and a leucine in SHP 1358 are occupied by a serine and an alanine, respectively, in SHP 265 ( Figure S2B). This finding suggests that SHP 1358 is a good candidate for interacting with Rgg 7530 and Rgg 4620 , and it could more effectively cross-activate Rgg 7530 , but less effectively cross-activate Rgg 4620 , if we assume that the cognate SHP of a Rgg is likely to be the most efficient SHP. Further experimental work is needed in order to confirm that SHP 1358 can interact with different Rggs.
Third, the gene encoding SHP 279 was poorly expressed and controlled by PptAB, Eep, and Ami, but the mature form of SHP 279 was present in the wild-type strain background. In contrast, the system's proximal target genes were expressed at higher levels and more strongly regulated by these three proteins than the shp 279 gene. The Rgg regulator that was encoded by Ster_RS09420 plays a role in oxidative stress defense in the S. thermophilus CNRZ368 strain [56]. This system did not display a positive feedback loop related to shp 279 expression, providing an additional element of support for the hypothesis that another SHP might interact with the Rgg 9420 regulator. SHP 279 is nearly identical to SHP 1358 : a spot that is occupied by a valine in SHP 1358 is occupied by a glycine in SHP 279 . Consequently, SHP1358 is again a good candidate for Rgg 9420 activation.
In a previous study, we demonstrated the absence of cross-activation between the SHP 1299 /Rgg 1299 and SHP 1358 /Rgg 1358 systems [33]. This result was expected, since the amino acid sequences of the two SHPs are not similar. Our findings here suggest that some SHP/Rgg systems can cross-activate each other, likely via highly similar SHPs, although this hypothesis must be experimentally tested. However, the fact that some SHP/Rgg loci encoded shp genes that were not expressed and that appeared to be activated by SHPs from other systems was surprising. Therefore, we wonder if, in S. thermophilus, these accumulated SHP/Rgg systems have degenerated or if they may prove to be useful under certain natural conditions. We also cannot exclude the possibility that other growth conditions exist, in which these shp genes are efficiently expressed. Previous work has shown that intracellular cross-activation involving 2 SHP/Rgg systems exists in S. pneumoniae [25] and in S. pyogenes [21]. However, circumstances are slightly different in the latter case because the two SHPs can interact with two Rggs with relatively similar affinities [57], although one Rgg activates transcription whereas the other inhibits it [21]. Another level of complexity is present if we consider interspecies crosstalk. This cross-talk can occur between streptococci species via SHP/Rgg systems, as demonstrated between S. pneumoniae and S. mitis [55], but also between distinct species via another RRNPP mechanism between S. aureus and E. faecalis. S. aureus can produce at least four linear peptides during the maturation process of secretion signal peptides of lipoproteins; one of these peptides can mimic a pheromone of E. faecalis and cause aggregation, the first step in horizontal gene transfer in this species [50]. Interestingly, in S. aureus, the transporter EcsAB, encoded by paralogs of the genes that encode PptAB and Eep, are necessary for the production of these peptides [50,51].

Conclusions
In conclusion, in S. thermophilus, PptAB plays a central role in the export of the AIPs that are involved in intracellular QS mechanisms. A RNAseq approach designed to identify the genes regulated by the exporter PptAB, the protease Eep and the importer Ami highlighted cross-talk between SHP/Rgg systems. More work is needed now in order to dissect this crosstalk at the strain level, but also at intraan inter-species levels and to understand their various roles in natural and complex environments.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4425/11/9/1096/s1, Table S1: Oligonucleotides used in this study, Table S2: Plasmids used in this study, Table S3: List of CDSs annotated in GenBank and short putative CDSs predicted by BactgeneSHOW in the chromosome and the two plasmids of the S. thermophilus wild-type strain (LMD-9), Table S4: List of the SHP*1358 (MGKKQILLTLLLVVFEGIIVIVVG) fragments whose masses was looked for in the TIL1213 strain (∆amiCDE pBV5030::P 32 -shp 1358 ) via LC-MS/MS., Table S5: Genes significantly upregulated in the TIL1560 strain (comR::aphA3) versus in the TIL1561 strain (comR::aphA3 ∆eep) (positively controlled by Eep), Table S6: Genes significantly upregulated in the TIL1560 strain (comR::aphA3)versus in the TIL1564 strain (comR::aphA3 ∆amiCDE) (positively controlled by Ami), Figure S1: Detection of the mature forms of SHP 1299 and SHP 1358 in cultures of the S. thermophilus wild-type strain (LMD-9) via LC-MS/MS. Fragmentation spectra for SHP 1299 (A) and SHP 1358 (B); mass, retention time, and the integrated area under the curve obtained from the extracted-ion chromatograms for both peptides (C). The data shown are representative of three independent experiments, Figure S2: Information on the shp/rgg loci in the S. thermophilus wild-type strain (LMD-9). (A) Schematic representation of the six shp/rgg loci based on genome annotation for the LMD-9 strain, as described in Table S4. Annotations for the rgg genes come from GenBank (CP000419.1), and annotations for the shp genes were developed in our lab using BactogeneSHOW, as described in the Materials and methods section. The rgg genes and shp genes are represented by black and gray arrows, respectively. The size of the protein encoded by each gene is indicated underneath. (B) Alignment of the sequences for the SHP precursors. The predicted or experimentally validated sequences for the mature forms are in bold, Funding: This study was funded by the French National Institute for Agriculture, Food, and Environment (INRAE) and the French Ministry of Higher Education, Research, and Innovation (MESRI). Proteomics analyses were performed by the PAPPSO platform (http://pappso.inra.fr), which is supported by INRAE, the Regional Council of Ile-de-France, IBiSA, and CNRS.