luxS contributes to intramacrophage survival of Streptococcus agalactiae by positively affecting the expression of fruRKI operon

Abstract

The LuxS quorum sensing system is a widespread system employed by many bacteria for cell-to-cell communication. The luxS gene has been demonstrated to play a crucial role in intramacrophage survival of piscine Streptococcus agalactiae, but the underlying mechanism remains largely unknown. In this study, transcriptome analysis, followed by the luxS gene deletion and subsequent functional studies, confirmed that impaired bacterial survival inside macrophages due to the inactivation of luxS was associated with reduced transcription of the fruRKI operon, encoding the fructose-specific phosphotransferase system. Further, luxS was determined not to enhance the transcription of fruRKI operon by binding its promoter, but to upregulate the expression of this operon via affecting the binding ability of catabolite control protein A (CcpA) to the catabolite responsive element (cre) in the promoter of fruRKI. Collectively, our study identifies a novel and previously unappreciated role for luxS in bacterial intracellular survival, which may give a more thorough understanding of the immune evasion mechanism in S. agalactiae.

Introduction

Streptococcus agalactiae, also known as group B streptococcus (GBS), is an important pathogen of both humans and animals [1]. This bacterium was first isolated from cow milk with mastitis, and subsequently, it was found to be a major etiologic agent of neonatal sepsis and meningitis [2]. Although several mortality events associated with bacterial infection in fish were reported in the 1980s, not much attention was paid at the outset because of the limited epidemic area. However, since 2009, large-scale streptococcal outbreaks caused by this bacterium continuously occurred in tilapia farms with high mortality and brought a deleterious impact on aquaculture industry worldwide [3]. Pathological phenotypes of S. agalactiae infection are mainly septicemia and meningoencephalitis in farmed tilapia [4]. Most notably, episodes of bacteremia and meningitis in humans were recently reported to be associated with consumption of raw fish infected with GBS sequence type (ST) 283 in Singapore [5,6,7] and Hong Kong (official figures).

Pathogenic mechanisms of meningoencephalitis caused by bacteria have been extensively researched in recent years. The mechanism by which pathogens target the brain and cross the blood–brain barrier (BBB) in the early phase of infection is dependent on successful evasion of the host innate immune system. Phagocytes play central roles in the innate immune response, and bacterial survival within phagocytes may contribute to dissemination of the pathogen within its host. Pathogens have developed diverse strategies to survive within phagocytes and even to take advantage of the intracellular environment. For example, Staphylococcus aureus is able to perturb the acquisition of lysosomal hydrolases, e.g., cathepsin D and β-glucuronidase, in macrophages, thereby preventing its degradation in the phagolysosome [8]. In Streptococcus pneumoniae, pneumolysin (Ply), a member of the thiol-activated cytolysin family of toxins, could inhibit the initial macrophage inflammatory response and improve bacterial immune evasion [9]. In Streptococcus pyogenes, a carbohydrate metabolism-related operon called fruRBA was found to be critical for the survival of this bacterium in neutrophils [10]. S. agalactiae has been known to utilize multiple virulence factors to survive inside host phagocytes. To defend itself against oxidative stress and reduce ROS production inside macrophages, this bacterium expresses an NADH-dependent peroxidase [11]. Also, hyaluronidase has a positive influence on the intracellular survival of S. agalactiae by inhibiting the secretion of proinflammatory cytokines [12]. Our previous study indicated that S-ribosylhomocysteine (SRH) lyase (LuxS) contributes to the intracellular survival of S. agalactiae within macrophages [13]. But paradoxically, the luxS inactivation has been reported to enhance the intracellular survival of S. pyogenes [14] or S. aureus [15].

As a homodimer, LuxS can formally catalyze the non-redox cleavage of bonded sulfides in S-rybosylhomocysteine (SRH) to produce L-homocysteine and 4,5-dihydroxy-2,3-pentanedione (DPD), which spontaneously cyclizes to active the autoinducer-2 molecule (AI-2) [16]. It is well-known that AI-2 serves as a universal signaling molecule in quorum sensing (QS) which mediates both intra- and interspecific communication [17]. Many studies have demonstrated that LuxS can regulate bacterial physiological processes through AI-2, such as biofilm formation [18], swimming motility [19] and antibiotic resistance [20]. However, our study has demonstrated that LuxS contributes to intracellular survival of S. agalactiae independent of the effect of AI-2 [13]. Knowledge about the contribution of LuxS/AI-2 to bacterial intracellular survival is rather lacking.

In this study, we explored the mechanism by which LuxS is responsible for the intracellular survival of S. agalactiae GD201008-001. We found that inactivation of luxS caused highly downregulated expression of the fruRKI genes as an operon. Interestingly, this regulation effect of luxS is associated with the impact on the binding ability of catabolite control protein A (CcpA) to the fruRKI promoter. The novel function of luxS identified in this study will broaden our understanding of the pathogenesis of S. agalactiae.

Materials and methods

Bacterial strains, cell lines and culture conditions

Streptococcus agalactiae GD201008-001 was isolated in 2010 from tilapia with meningoencephalitis in Guangdong Province, China [21]. GD201008-001 wild-type strain (WT) and its derived luxS mutant strain (ΔluxS) and the luxS complemented strain (CΔluxS) [13] were maintained in Todd-Hewitt broth (THB) or in chemically defined medium (CDM) [22]. Escherichia coli was cultured in Luria–Bertani (LB) medium. For plasmids screening required, media were supplemented with antibiotics using the concentration below: 100 μg/mL spectinomycin (Sp, Sigma, St. Louis, MO, USA), 10 μg/mL erythromycin (Em, Sigma), 100 μg/mL kanamycin (Km, Sigma) or 100 μg/mL ampicillin (Ap, Sigma). The details of bacterial strains and plasmids are listed in Additional file 1. Macrophage cell line RAW 264.7 were maintained in DMEM (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS, Gibco).

DNA methods and construction of mutant strains

PCR for cloning and generating fragment fusions was performed using 2 × Phanta Max Master Mix (Vazyme, Nanjing, China) and diagnostic assays were performed using Green Taq Mix (Vazyme) following the manufacturer’s protocol. DNA fragments and plasmids digested with enzymes were gel purified from agarose using the Gel Extraction Kit (Omega, Beijing, China). All DNA sequencing was done by Genewiz, Inc (Suzhou, China).

The whole fruRKI operon was knocked out using the suicide plasmid pSET4S [23] and performed according to the method described previously with some modulations [24]. All primers used in this study are listed in Additional file 2. For construction of the ΔfruRKI mutant, DNA fragments flanking the fruRKI operon were amplified using the primer pairs fruRKI 1/2 (before the start codon of fruR) and fruRKI 3/4 (after the stop codon of fruI), fused together using the primer pair fruRKI 1/4, and then cloned into the pSET4S. The recombinant plasmid was transformed into the chemically competent E. coli DH5α. After sequencing, the recombinant plasmid was transformed into the S. agalactiae GD201008-001 competent cells by electroporation. After electroporation, the strain containing plasmid was sub-cultured twice daily at 28 °C for five days and then THB medium with Sp was used to check for plasmid loss after the successful double-crossover recombination between plasmid and genome. The ΔfruRKI mutant by double-crossover recombination was confirmed by amplifying the fruRKI locus using primers fruRKI-F and fruRKI-R followed by DNA sequencing.

To construct the fruRKI complementary stain (CΔfruRKI), the whole operon with the flanking fragments was amplified and cloned into the pSET4S. The recombinant plasmid was transformed into the ΔfruRKI strain. The resultant strain was cultured on Sp-containing THB agar medium, and positive clones were verified by PCR followed by DNA sequencing. The cre site mutation strains were constructed following a similar procedure as above, except that the fragment containing the point mutation was synthesized and cloned into the pSET4S. The mutation sites of the cre were chosen based on the conserved critical nucleotides identified by sequence alignment. In this study, five strains with the cre site mutation were generated, i.e., WT/ΔluxS-G1 (the conserved nucleotides of the cre region were replaced in the fruRKI promoter of the WT or the ΔluxS strain), and WT/ΔluxS-G2 (control strains, non-conserved nucleotides of the cre were replaced in the fruRKI promoter).

The shuttle vector pSET2 [25] that drives fruI, fruK or fruR, respectively, were constructed to complement the loss of each gene function of fruRKI operon in the ΔluxS mutant strain. For ΔluxS::fruI strain, the promoter region of fruRKI (amplified by the primer pair CfruI1/2) and the fruI coding sequence (amplified by the primer pair CfruI3/4) was fused together using the primer pair CfruI1/4 and then cloned into pSET2. The recombinant plasmid was transformed into the ΔluxS competent cells by electroporation and the ΔluxS::fruI strain was selected by THB medium with Sp.

To construct a promoter reporter strain, the 129 bp-DNA fragment containing the promoter region of fruRKI operon was amplified using the primers PfruRKI-lacZ-F/PfruRKI-lacZ-R and cloned in front of a promoterless β-galactosidase gene in a shuttle vector, pTCV-lac [26]. The resultant strain was named PfruRKI-lacZ. For point mutations on pTCV-lac plasmids, two DNA fragments with 5’ homology arms containing mutation sites were amplified and fused, and cloned into pTCV-lac. All resulting plasmids were sequenced and then transformed into the wild-type S. agalactiae GD201008-001 and its derivative mutants by electroporation and named PMfruRKI-lacZ, T1PfruRKI-lacZ and T2PfruRKI-lacZ.

Bacterial growth in carbon defined medium

For the carbohydrate metabolic assay, S. agalactiae strains to be tested were cultured in THB medium at 37 °C with 180 rpm until to reach the stationary phase and then adjusted to the optical density at 600 nm (OD₆₀₀) of 0.8. The adjusted suspensions were inoculated into 100 mL of CDM supplemented with 0.1% (w/v) of either glucose or fructose (Sigma) as a sole carbon source at a ratio of 1:100 and incubated under the same conditions for 1 day. The cell densities at OD₆₀₀ were measured every 2 h. The experiment was repeated three times independently.

Transcriptome analysis

Bacteria were cultured in THB medium at 37 °C. After reaching an OD₆₀₀ of 0.8, bacterial RNA was extracted by an RNAqueous kit (Thermo Fisher Scientific, San Jose, CA, USA). Before the RNA library assembly, ribosomal RNA was removed using Ribo-Zero Magnetic Kit (Illumina, San Diego, CA, USA). Libraries construction and transcriptome sequencing were conducted by OE Biotech (Shanghai, China). A GO enrichment analysis was conducted by the OECloud tools based on GO Database. To perform RNA-Seq, the S. agalactiae strains were purified and divided into three for triplicate repeats. After that, total RNA of each repeat was extracted and sent for RNA-Seq. All samples were analyzed in one sequencing run.

Real-time quantitative PCR (RT-qPCR)

The RT-qPCR was carried out to measure the transcription levels of target genes using the primers listed in Additional file 3. The total RNA was extracted using the Total RNA kit (Omega, Norcross, GA, USA) and then reverse transcribed into cDNA by Hiscript II QRT Supermix (Vazyme). The mRNA levels of target genes were measured by RT-qPCR according to the protocol of One Step RT-qPCR SYBR Green kit (Vazyme). The recA gene was used as an internal control. The fold-changes of mRNA levels were calculated using the comparative cycle threshold (2^−ΔΔCT) method [27]. The experiment was repeated three times independently.

S. agalactiae intracellular survival and phagocytosis assay

RAW264.7 cells were cultured in 24-well plates at a density of 2 × 10⁵ cells/well with 10% FBS added DMEM at 37 °C with 5% CO₂ for 20 h, until to 85–90% confluence. S. agalactiae strains were cultured in THB medium overnight at 37 °C. Then bacteria were washed three times in PBS and adjusted to 4 × 10⁶ bacterial cells/mL using DMEM. The cells were washed and then inoculated with S. agalactiae at a multiplicity of infection (MOI) of 1:1 for 1 h. To eliminate residual adherent bacteria, the cells were washed five times and then added 10% FBS-DMEM containing 1% penicillin G and incubated for 1 h.

To measure the phagocytotic rate, infected cell samples were taken 1 h after antibiotic treatment and subjected to lysis. Ten-fold serial dilution of lysates were made with PBS and then cultured on THB agar to give a bacterial count of colony forming unit (CFU). The percentage of phagocytosis was calculated based on the CFU of intracellular bacteria relative to the total CFU of bacteria added in the cell monolayers. To measure the survival rate of intracellular bacteria, the cell sampling period was started after 1 h antibiotic treatment (time point 0) and samples were taken every 4 h during 12 h period. The final timepoint of sampling was 24 h after time point 0. Infected cells were treated as above to measure the CFU of intracellular bacteria. The relative survival rate was calculated as follows: (CFU at a specific time point/CFU at time point 0) × 100. The assays described above were repeated three times with three independent replicates.

Electrophoretic mobility shift assay (EMSA)

The LuxS and CcpA proteins were expressed and purified as previously described [13]. The ribosylhomocysteinase activity of LuxS has been confirmed by the ability to synthesize AI-2 in vitro [13]. The primers used to amplify the ccpA gene are listed in Additional file 2. The DNA fragments used in EMSAs were amplified by primer pairs PMfruRKI-F/PMfruRKI-R and PMluxS-F/PMluxS-R. A 70 bp DNA fragment served as the negative control. The purified protein (0.2–1.0 μM) was incubated with the DNA fragment (25 nM) in binding buffer [20 mM Tris–HCl (pH = 7.5), 30 mM KCl, 1 mM DTT, 1 mM EDTA (pH = 7.5), 10% (v/v) glycerol)] in a final volume of 20 μL for 30 min at 37 °C. Samples were loaded on a 10% polyacrylamide gel and electrophoresed in 0.5 × TBE (44.5 mM Tris, 44.5 mM boric acid, 1 mM EDTA, pH = 8.0) under 120 V with the ice bath for 1 h. The gel was stained in Gold nucleic acid staining solution for 10 min, and then watched and recorded under the UV Trans illumination by Gel Doc XR (Bio-Rad, CA, USA). The experiment was repeated three times independently.

β-galactosidase assay

Β-galactosidase assay was performed as previously described [26]. S. agalactiae strains were cultured in THB medium for 12 h and then cultured in fresh THB to an OD₆₀₀ of 0.8. Bacteria were incubated on ice for 20 min and washed three times in β-mercaptoethanol (BME) free Z buffer and adjusted to an OD₆₀₀ of 1.0. The diluted cells were permeabilized by treatment with 0.05 M BME, 0.5% toluene and 4.5% ethanol for 5 min at 30 °C. The substrate o-nitrophenyl-β-D-galactoside (ONPG, 4 mg/mL) was added to start reaction until sufficient yellow color has developed. The reaction time was recorded. The activity of β-galactosidase was calculated by the formula: (10³) × (OD₄₂₀−1.75 × OD₅₅₀)/(reaction time × volume of culture × OD₆₀₀). The experiment was repeated three times independently.

Statistical analyses

Data were presented as the mean ± standard deviations (SD). GraphPad Prism version 8.0.1 was used for the statistical analysis and graph preparation. All statistical analyses were performed using unpaired two-tailed Student’s t test. Comparisons with P ≤ 0.05 were accepted as statistically significant.

Results

A number of differentially expressed genes (DEGs) were identified in the luxS deficiency mutant of S. agalactiae

To determine the role of luxS in intracellular survival of S. agalactiae, we performed RNA transcriptome sequencing (RNA-Seq) for WT and ΔluxS mutant strains. Genes with over twofold change ((log₂ ≤  −1.0 or log₂ ≥ 1.0) were considered differentially expressed. A total of 264 genes were identified in the ΔluxS strain, including 155 upregulated (red) and 109 downregulated (green) genes (Fig. 1A; Additional file 4). The hierarchical clustering clearly illustrated the magnitude difference of the differentially expressed genes between WT and ΔluxS (Fig. 1B). To validate the reliability of our RNA-Seq results, RT-qPCR was performed to measure the expression levels of the DFGs. The expression patterns of 20 genes randomly selected (10 up-regulated and 10 down-regulated genes) were consistent with the transcriptomic data (Fig. 1C). Further, we classified the DFGs according to the Gene Ontology (GO) descriptions. The DEGs were primarily classified into biological process, cellular component, and molecular function (Fig. 1D). Specifically, the TOP 30 enriched GO terms (down-regulated in ΔluxS) included carbohydrate metabolic process, membrane component and ion transport activity. Notably, the putative fructose metabolic operon (fruRKI) was significantly down-regulated in ΔluxS, which has been reported to be involved in the intracellular survival of S. pyogenes [10]. The enrichment scores, p-value and the down-regulated genes contained in the GO terms are listed in Additional file 5.

Fig. 1

Comparative transcriptomics analysis. A Volcano plot of the differentially expressed genes. The x-axis represents fold change (log₂) and y-axis represents p-value (log₁₀). Red dots represent upregulated genes and green dots indicate downregulated genes with a significant difference. B Heat map of gene expression. A total of 264 genes were differentially expressed in the ΔluxS strain compared to the wild-type (WT) strain. Red and blue fonts represent up- and down-regulated genes, respectively in the WT or ΔluxS strains. C Relative mRNA levels of 20 differently expressed genes determined by RT-qPCR. Data are expressed as n-fold change normalized to mRNA level of WT. D Gene Ontology (GO) classification of downregulated genes in the ΔluxS strain.

luxS deficiency decreases carbohydrate metabolism of S. agalactiae

Considering the repressed expression of carbohydrate metabolic process-related genes in the ΔluxS strain, we sought to ascertain whether luxS is involved in the utilization of carbon sources in S. agalactiae. As shown in Additional file 6, there was no difference in growth kinetics between the WT and ΔluxS strains in THB. When we used CDM supplemented with glucose or fructose as the sole carbon source, the ΔluxS strain showed significantly decelerated growth. In CDM supplemented with 0.1% (w/v) (Fig. 2A) or 1% (w/v) (Fig. 2B) fructose, ΔluxS showed a significantly lower bacterial cell density as indicated by OD₆₀₀ when compared to the WT or CΔluxS strains at all sampling time points. A similar alteration was observed with supplementation of 0.1% (Fig. 2C) or 1% (Fig. 2D) of glucose.

Fig. 2

Influence of luxS deficiency on carbon source availability of S. agalactiae strains. All strains were cultivated in a chemically-defined medium (CDM) supplemented fructose (A, B) or glucose (C, D) at a range of concentrations from 0.1% to 1% (w/v) as the sole carbon source, separately. Data are presented as the mean ± SD for three independent experiments. *P < 0.05, **P < 0.01, or ***P < 0.001, indicates a significant difference between the indicated strain and the WT strain.

luxS contributes to intracellular survival of S. agalactiae via upregulating the transcription of fruRKI operon

Based on the transcriptomic data, we identified the putative fructose operon fruRKI was downregulated in ΔluxS mutant strain. In the genome of S. agalactiae GD201008-001, fruR (encoding the DeoR/GlpR transcriptional regulator), fruK (encoding 1-phosphofructokinase) and fruI (encoding PTS fructose-specific EIIC) are three contiguous genes with a 3-bp overlap, indicating that they might represent an operon (Fig. 3A). To further support the notion, we isolated total RNA from WT and reverse-transcribed into cDNA as a template and successfully amplified a 3.1 kb transcript between fruR and fruI (Additional file 7A), indicating that fruI, fruK and fruR genes comprised an operon in S. agalactiae. Further, we want to investigate whether AI-2 molecules can negate the significantly down-regulated expression of fruRKI caused by luxS deficiency. As a result, the addition of AI-2 could not effectively improve the mRNA levels of fruRKI in the ΔluxS strain (Fig. 3B).

Fig. 3

Growth characters and intramacrophage survival capabilities of S. agalactiae strains. A Genetic structure of the fruRKI operon in S. agalactiae. B Relative mRNA levels of the fruRKI genes. C Growth curves of the WT, ΔluxS, ΔfruRKI, and CΔfruRKI strains cultivated in THB. D Growth curves of the WT, ΔluxS, ΔfruRKI and CΔfruRKI strains cultivated in CDM supplemented 1% fructose. E Growth curves of the WT, ΔluxS, ΔfruRKI and CΔfruRKI strains cultivated in CDM supplemented 1% glucose. F Intracellular survival of the WT, ΔluxS, ΔfruRKI and CΔfruRKI strains in macrophages. G Intracellular survival of the WT, ΔluxS, ΔluxS:: fruI, ΔluxS:: fruK and ΔluxS:: fruR strains in macrophages. Data are presented as the mean ± SD of three independent experiments. *P < 0.05 or ***P < 0.001, indicates a significant difference between the indicated strain and the WT strain. ^# < 0.05, ^## < 0.01, or ^### < 0.001, indicates a significant difference between the indicated strain and the ΔluxS strain.

To determine whether decreased bacterial intracellular survival caused by luxS deficiency was due to the down-regulation of fruRKI operon, we constructed the whole fruRKI operon knock-out mutant strain ΔfruRKI and its complementary strain CΔfruRKI. Before proceeding intracellular assay, we compared the growth characteristics of the WT and ΔfruRKI mutant strains in THB or CDM supplemented with 1% fructose (fructose-CDM) or 1% glucose (glucose-CDM). As shown in Fig. 3C, the growth kinetics of ΔfruRKI in the THB medium was similar to that of the WT. However, ΔfruRKI failed to grow in fructose (Fig. 3D) or glucose (Fig. 3E) as the sole carbon source. The intracellular survival assay was performed by infecting RAW264.7 macrophage cell line, which has been validated as an in vitro platform to evaluate piscine streptococcus-macrophage interactions [12, 13, 24]. As shown in Fig. 3F, at each time point, the number of intracellular viable bacteria in both ΔluxS and ΔfruRKI mutant strains was significantly decreased compared to that of the WT and CΔfruRKI strains. To further confirm whether the down-regulation of the fruRKI operon could be accountable for the decreased intracellular survival of ΔluxS mutant strain, we have tried to overexpress the whole fruRKI operon in the ΔluxS strain, however, the operon is too large to be cloned into the pSET2 vector. Thus, we overexpressed the individual genes in the fruRKI operon in the ΔluxS strain, and obtained the strains ΔluxS:: fruI, ΔluxS:: fruK and ΔluxS:: fruR. The measurement for the number of surviving bacteria showed that overexpression of any gene located in the fruRKI operon significantly restored the viability of ΔluxS mutant strain in RAW264.7, although the single restoration could not make the number of viable bacteria reach the level of the WT strain (Fig. 3G).

luxS indirectly regulates the promoter activity of the fruRKI operon

Considering that inactivation of luxS downregulated the expression of all three genes located in the fruRKI operon as evidenced by the transcriptomic data, we make a conjecture that luxS may have affected the transcription of fruRKI via modulating the promoter. To determine the transcription level of fruRKI, we examined the promoter activity by utilizing a transcriptional fusion of lacZ reporter gene with the promoter of fruRKI in the WT and ΔluxS strains grown in THB or fructose-CDM. At any of the stages of bacterial growth in THB, the fruRKI promoter showed decreased activity in the ΔluxS strain compared with WT, while the introduction of luxS in the ΔluxS strain almost completely restored the activity of fruRKI promoter (Fig. 4A). A similar result was also observed in the WT and ΔluxS strains grown in fructose-CDM (Fig. 4B). Further, we performed the EMSA and confirmed that the LuxS protein could not bind to the promoter of fruRKI operon (Additional file 7B).

Fig. 4

The promoter activity of fruRKI in S. agalactiae strains grown in rich medium or chemically defined medium (CDM). A The fruRKI promoter activity in the WT, ΔluxS and CΔluxS strains cultivated in THB (rich medium). B The fruRKI promoter activity in the WT, ΔluxS and CΔluxS strains cultivated in CDM supplemented 1% fructose. Data are presented as the mean ± SD of three independent experiments. ***P < 0.001.

A catabolite responsive element (cre) is shared by the promoters of luxS and fruRKI operon

The CcpA has previously been reported to be the major regulator involved in controlling carbon-metabolism by binding to the cis-regulatory element cre to activate or repress the transcription of target genes in Gram-positive bacteria [28]. In this study, a cis-acting sequence was identified in the fruRKI promoter (Fig. 5A). Interestingly, a putative cre was also found in the promoter region of luxS (Fig. 5B). This shared feature by the two promoters led us to hypothesize that CcpA may play a certain role in the relationship between luxS and fruRKI. To illustrate this point, we first performed the EMSA using purified CcpA protein with the DNA fragments containing the promoter of fruRKI or luxS. As shown in Fig. 5C, CcpA could directly bind to the promoter regions of both fruRKI and luxS. This result indicated that both luxS and fruRKI transcription were directly regulated by CcpA. Further, we have tried to delete ccpA from S. agalactiae using homologous recombination, but unfortunately, all our attempts failed, suggesting that this gene might indeed be essential for bacterial survival. Then we performed RT-qPCR to detect the transcription level of ccpA in the WT and ΔluxS strains. No significant difference was observed in the ccpA transcription between the two bacterial strains (Additional file 8). In addition, our transcriptomic data also confirmed that the deletion of luxS does not result in altered ccpA expression. These findings exclude the possibility that loss of luxS results in the upregulation of CcpA, thereby repressing the expression of the fruRKI operon.

Fig. 5

The in vitro binding of CcpA to the fruRKI promoter or luxS promoter. A Diagrams depicting the noncoding region of the fruRKI operon. The magnified region indicates the promoter sequences, including the putative cre for CcpA binding (underlined), and the −10 and −35 positions are highlighted and labeled. B Diagrams depicting the noncoding region of the luxS gene. The putative cre is underlined, and the −10 and −35 positions are highlighted and labeled. C An electrophoretic mobility shift assay (EMSA) showing the binding ability of CcpA protein to the fruRKI or luxS promoters (25 nM). The concentrations of CcpA ranged from 0 to 1.0 μM.

Deletion of luxS promotes the binding of CcpA to the fruRKI promoter

To determine whether the reduced transcription level of the fruRKI gene in ΔluxS is due to the increased binding ability between CcpA and the fruRKI promoter, we mutated the conserved nucleotides of cre in the luxS and fruRKI promoter regions, respectively, and tested their binding capacity to CcpA protein by EMSA. According to the RegPrecise Database, the cre regions from eight species of Gram-positive bacteria were aligned and a total of eight nucleotides were identified to be most conserved (Fig. 6A). To confirm mutation of these conserved nucleotides on cre abolishes the binding of CcpA, we amplified the luxS and fruRKI promoter regions containing the cre and cre mutant (denoted as luxS mutant promoter and fruRKI mutant promoter), respectively. The EMSA was performed with four groups of different concentrations of CcpA protein (Fig. 6B). The EMSA showed that the mutation of cre in either luxS promoter or fruRKI promoter reduced detectable binding of respective promoters to CcpA (Fig. 6B and C). Results of β-galactosidase activity assay revealed that the PfruRKI activity was remarkably decreased in ΔluxS compared to the WT strain (Fig. 6D), but no significant difference was observed in promoter activity between ΔluxS and WT after mutating the cis-acting sequences of fruRKI promoter (PMfruRKI). Moreover, the activities of T1PfruRKI and T2PfruRKI were not affected due to mutation in non-cis region of the fruRKI promoter. To determine the importance of cre in the regulation of fruRKI transcription by luxS, we mutated the cre region of fruRKI promoter in the WT and ΔluxS strains. The resulting mutant strains were named WT-G1 and ΔluxS-G1, respectively. As expected, transcription of fruRKI in WT-G1 was significantly enhanced compared with the WT strain while no significant difference was seen between the WT-G1 and ΔluxS-G1 strains (Fig. 6E). Overall, our results indicate that deletion of luxS leads to increased binding of CcpA to the fruRKI promoter.

Fig. 6

luxS indirectly regulates the expression of fruRKI operon through the cre region in the fruRKI promoter. A The sequence alignment of cre from eight gram-positive bacteria. The conserved regions of cre are highlighted. B Competitive binding of the luxS promoter and the fruRKI promoter to CcpA by EMSA analysis. The 129-bp fragment of luxS promoter was tested for its ability to bind with CcpA (0.4, 0.6 and 0.8 μM) in presence of 188-bp fragment of wild type fruRKI promoter (lane 4, 7 and 10) or fruRKI promoter with mutations on cre (lane 6, 9 and 12); the 188-bp fragment of fruRKI promoter was tested for its ability to bind with CcpA (0.4, 0.6 and 0.8 μM) in presence of 129-bp fragments of wild type luxS promoter (lane 4, 7 and 10) or luxS promoter with mutations on cre (lane 5, 8 and 11); Lane 1 to 3. Different components of internal controls. C The DNA-binding capacity of CcpA was measured with grayscale analysis of the blots. D The fruRKI promoter activity in the WT and ΔluxS strains containing the PMfruRKI-lacZ reporter plasmid. The WT and ΔluxS strains containing T1PfruRKI-lacZ or T2PfruRKI-lacZ reporter plasmids serve as control groups. The β-galactosidase activity was expressed as relative miller units. E Relative mRNA levels of the fruRKI genes determined by real-time PCR. WT-G1: eight-base substitution in the cre conserve region of fruRKI promoter in WT; ΔluxS-G1: eight-base substitution in the cre conserve region of fruRKI promoter in ΔluxS; WT-G2: eight-base substitution in the cre non-conserve region of fruRKI promoter in WT; ΔluxS-G2: eight-base substitution in the cre non-conserve region of fruRKI promoter in ΔluxS. F Intracellular survival rates of the WT, ΔluxS, WT-G1, ΔluxS-G1, WT-G2 and ΔluxS-G2 strains in macrophages. Data are presented as the mean ± SD of three independent experiments. *P < 0.05, **P < 0.01, or ***P < 0.001.

In addition, to determine whether reduced intracellular survival ability exhibited by the ΔluxS strain can be attributed to cre-mediated downregulation of the fruRKI operon, we infected the RAW264.7 macrophages with WT, ΔluxS, WT-G1 and ΔluxS-G1 strains. The WT-G1 incubated with RAW264.7 showed a comparable ability in intracellular survival to that of the WT, whereas the ΔluxS mutant exhibited a significant decrease during the entire experiment (Fig. 6F). Importantly, the survival ability of ΔluxS-G1 was completely restored to the WT level (Fig. 6F).

Discussion

Our understanding of luxS has been predominantly focused on its role in quorum sensing. Here, we demonstrated that luxS also affected the ability of S. agalactiae to survive within macrophages by regulating the transcription of the carbohydrate utilization operon fruRKI. Using site-directed mutagenesis, we demonstrated that the cre located in the promoter of fruRKI operon is a critical locus for luxS to play an indirect regulatory role. Our study reveals a new mechanism by which S. agalactiae can adapt their metabolism to the available nutrients in macrophages and thus survive efficiently.

S. agalactiae could withstand the extreme environment in macrophages and persist inside the fully mature phagolysosome for a relatively long period [8]. We have previously found that luxS deficiency decreased the intracellular survival ability of S. agalactiae but this weakening effect was not mediated by the signaling molecule AI-2 [13]. LuxS has been reported to be involved in many cellular processes, but little is known about its function in intracellular survival. LuxS is required for AI-2 biosynthesis, but it also plays a crucial role in the activated methyl cycle (AMC), which is involved in the utilization of S-adenosylmethionine (SAM) [29,30,31]. In Streptococcus sanguinis, luxS deletion has resulted in a large number of gene expression changes due to the accumulation of intermediates of SAM metabolism [32]. Therefore, we speculate that reduced intracellular survival ability in the ΔluxS might be linked to the altered gene expression pattern. To verify this, we performed a comparative transcriptomics analysis of the WT and ΔluxS strains and identified 264 differentially expressed genes. Interestingly, the genes involving carbohydrate metabolisms, such as sugar-specific phosphotransferase system (PTS) and carbohydrate ABC transporter, account for a large proportion of downregulated transcripts.

One previous study has revealed that S. agalactiae genomes lack genes involved in the biosynthesis of the tricarboxylic acid cycle (TCA) [33]. However interestingly, Patron et al. [34] reported that this bacterium has a broad ability to import carbohydrate sources to adapt to the host environment. Therefore, we speculated that the utilization of carbohydrates may be essential for S. agalactiae survival within phagocytes. In this study, the downregulation of the fruRKI operon in the luxS mutant of S. agalactiae attracted our attention. The fruRKI operon has been demonstrated in S. mutans to play important roles in various biological processes including sugar metabolism and biofilm formation [35]. Additionally, the operon fruRBA, similar to fruRKI, has been shown to be required for S. pyogenes growth in fructose and for resistance to neutrophil killing in human blood [10]. Our present study showed that the growth of ΔluxS exhibited significant weakening in a chemically defined medium with a low concentration of fructose or glucose, and the transcription of fruRKI was also significantly lower than that of the WT strain. Furthermore, we also demonstrated that the fruRKI operon was important for S. agalactiae intracellular survival. Accordingly, it is safe to reach the conclusion that downregulation of the fruRKI operon in the ΔluxS strain results in deficiencies in carbohydrate metabolism, thereby reducing the survival of this bacterium within nutrient-poor macrophages [36]. Notably, downregulation in the mRNA level of fruRKI could not be eliminated by the addition of exogenous AI-2, which further gives support to our previous finding that LuxS contributes to intracellular survival of S. agalactiae independent of the effect of AI-2 [13].

It is widely known that pathogens utilize carbon catabolite repression (CCR) to effectively assimilate a preferred carbon in response to local differences in nutrient availability within the host [28]. CcpA is the master transcriptional regulator of CCR, which can repress or activate gene transcription by binding cis-acting DNA known as the cre [37, 38]. In this study, we identified a putative cre in the promoters of fruRKI, which has been demonstrated to be directly bound by CcpA, thereby inhibiting fruRKI transcription in Lactococcus lactis [39]. Consequently, an intriguing question arose: can CcpA transcription be regulated by LuxS? If loss of luxS results in an increase in CcpA transcript levels, then it can explain our observations of repression of fruRKI operon. However, our study provides evidence that ccpA transcription has not been altered due to the inactivation of luxS, and therefore is not responsible for the downregulated expression of fruRKI in the ΔluxS strain.

We have attempted to verify the bridge role of CcpA between luxS and fruRKI by deleting the ccpA gene from S. agalactiae, but regretfully, we failed to obtain the ccpA deletion mutant. This gene has been reported to be essential for S. agalactiae survival [40]. However, Roux et al. [41] deleted the ccpA gene in S. agalactiae strain A909 successfully. No more than 95% query cover exist between the genomes of the piscine GD201008-001 and human-derived isolate A909, which might explain the contradiction between different researches. Interestingly, knockout of luxS caused a significantly reduced activity of the fruRKI promoter. And the cre in the promoter of fruRKI operon was determined to be an action target region of luxS since its nucleotide substitutions eliminated the effect of luxS on fruRKI operon and restored the intracellular survival ability of the ΔluxS strain. All the above findings suggest that the altered binding capability of CcpA to the cre in the ΔluxS strain accounts for decreased fruRKI expression. Based on the evidence presented here, we propose a model for the regulation between luxS-fruRKI-CcpA (Additional file 9). In short, the deletion of luxS enhances the binding of CcpA to the cre of the fruRKI promoter, resulting in reduced fruRKI transcription. Whilst our study has highlighted the important role that the recognition of cre by CcpA plays in affecting the transcription of fruRKI operon by luxS, it remains unclear how luxS deficiency exerts an effect on the binding capability of CcpA to the cre in the luxS promoter. In addition, we also cannot rule out the possibility that proteins other than CcpA that have the ability to bind this cre responsive element may contribute to the changes seen in this study, although only CcpA has so far been reported to preferentially bind to the cre. Future work will clarify these issues.

Notably, according to the RegPrecise database, the cre regions were found in the promoters of 23 genes related to carbohydrate metabolism (Additional file 10), representing 74.19% of the downregulated genes. To determine whether our findings relating to CcpA are specific to the fruRKI promoter, we randomly selected another two promoters (PptsG and PrbsR) containing the cre to perform competitive EMSA. Unsurprisingly, they could also bind to CcpA competitively with luxS (Additional file 11). This finding suggests that the regulatory role of luxS in carbohydrate metabolism may be broad by affecting the binding of cre to CcpA. But it is important to realize that not all carbohydrate metabolism-related genes with the cre element could be regulated by luxS. For example, a conserved PTS manLMN operon was not transcriptionally affected by luxS deficiency based on our RNA-seq data and subsequent RT-qPCR validation (data not shown). We speculate that there may be unknown regulators involved in manLMN transcription. In S. mutans, the manLMN operon is regulated not only by CcpA but also by FruR and EIIMan [34].

In conclusion, our findings establish, for the first time, a link between luxS and intracellular survival of S. agalactiae, which advances our understanding of the luxS function in mediating bacterial pathogenesis. Further studies providing insight into the precise mechanism of luxS effect may uncover novel therapeutic avenues.