Introduction

Trueperella pyogenes is a prevalent opportunistic pathogen that normally inhabits host mucous membranes and causes suppurative lesions of economically important animals (Zhao et al. 2011; Abdulmawjood et al. 2016; Bicalho et al. 2016; Ribeiro et al. 2015). It has been reported that T. pyogenes expresses several known and putative virulence factors, which play important roles during infection (Jost and Billington 2005; Pietrocola et al. 2007; Zhao et al. 2013a, b). Particularly, pyolysin as the main virulence factor can be used for specific clinical diagnosis of T. pyogenes and DNA vaccine development (Zhao et al. 2013a; Zhang et al. 2013; Huang et al. 2016). Although there are an increasing number of studies concerning the isolation of T. pyogenes from various hosts, little is known about the pathogenic mechanism of this organism due to the limited knowledge about the virulence regulation at the genome level. Additionally, enhanced antibiotic resistance has significantly challenged veterinary practice and potentially threatens human health (Plamondon et al. 2007; Zhao et al. 2011; Zastempowska and Lassa 2012; Rzewuska et al. 2016).

Signaling transduction systems function through intracellular information-processing pathways that link external stimuli to specific adaptive responses and regulate several important behaviors, such as virulence and basic metabolic functions (Zschiedrich et al. 2016). In prokaryotes, these signaling systems are dominated by the phosphorylation of two-component proteins, which consist of a histidine kinase (HK) and a response regulator (RR). Phosphotransfer from HK to RR results in the activation of RR and triggers the output response of downstream signaling (Capra and Laub 2012; Zschiedrich et al. 2016). Our previous study identified a T. pyogenes isolate, TP7, with high pyolysin production and lethality rate as compared with other isolates (Zhao et al. 2013a). With evidence to suggest that the hemolysis and virulence of T. pyogenes were significantly decreased in plo mutant strain (Jost et al. 1999), T. pyogenes TP7 was defined as a highly virulent isolate, which might have a different regulatory pattern on plo expression in comparison to the moderately virulent isolate, T. pyogenes TP8 (Zhao et al. 2013a). In this study, we first performed a comparative transcriptome analysis between TP7 and TP8, and the result revealed a LuxR-type two-component regulatory system, named PloS/PloR, which might govern the virulence of T. pyogenes. Intriguingly, the quorum-sensing (QS) signal molecules of Gram-negative bacteria Pseudomonas aeruginosa and Escherichia coli could suppress the growth, biofilm production and hemolysis of T. pyogenes. And this process might be due to the binding of QS signals to PloS, the sensor HK of the LuxR-type two-component regulatory system.

Materials and methods

Isolates and culture conditions

The T. pyogenes strains, TP7 and TP8 were isolated from forest musk deer, an economically important ruminant and categorized as a first-class key species protected by Chinese legislation since 2002 (Guha et al. 2007; Zhao et al. 2011). Differently, T. pyogenes TP8 was obtained from the abscess of body surface, whereas T. pyogenes TP7 was from the suppurative lung of a dead individual. Both T. pyogenes strains were cultured in a brain–heart infusion broth supplemented with 5% fetal bovine serum (BHIF) at 38 °C with 5% CO2. P. aeruginosa PAO1 and E. coli (ATCC 25922) strains were lab-preserved and routinely grown in Luria–Bertani (LB) broth.

Transcriptome analysis

Total RNAs of T. pyogenes TP7 and T. pyogenes TP8 were extracted by TRIzol® reagent (Invitrogen) after overnight culture in BHIF medium. The RNA samples from three independent cultures were mixed to minimize the deviation of RNA-SEq. The complementary DNA libraries were constructed and sequenced using Illumina Hiseq2000 technology. Tophat2 (Kim et al. 2013) was conducted to perform mapping RNA-seq data to TP8 genome (GenBank accession no.: CP007003). The differential expression between the two isolates was calculated using Cufflinks (Mortazavi et al. 2008), and then the data were processed by R packages pheatmap software (Kolde 2011). The gene expression profiles of T. pyogenes TP7 and T. pyogenes TP8 had been deposited at DDBJ/EMBL/GenBank under the accessions, PRJNA360053.

Inhibitory effect of HSLs on T. pyogenes growth

Clinical evidence suggested that there was only T. pyogenes in the abscesses of body surface from sick forest musk deer. However, various Gram-negative bacteria, especially P. aeruginosa and E. coli, were isolated as the dominative population from the intracorporal abscesses of dead individuals (Zhao et al. 2011). Therefore, we then set out to test whether P. aeruginosa and E. coli could inhibit the growth of T. pyogenes. P. aeruginosa PAO1 and E. coli (ATCC 25922) were first grown in 2 ml BHIF broth at 37 °C for 10 h, respectively. Subsequently, the supernatants of both cultures were harvested and followed by filtering through a 0.45-μm filter (Millipore) to remove the residual cells. The filtrates were mixed (1:1) with fresh BHIF broth and inoculated with T. pyogenes TP8 (1.0 × 108 CFU). Because bacterial QS signal molecules were confirmed to play a key role during interspecies interaction (Short et al. 2014), the same amount of T. pyogenes TP8 was then cultured in 4 ml BHIF containing gradient concentrations (0, 10, 25, 50, 100, and 200 μM) of N-(3-oxododecanoyl)homoserine-l-lactone (C12HSL, QS signal of P. aeruginosa) or N-octanoyl-dl-homoserine lactone (C8HSL, QS signal of E. coli) and cultured at 38 °C for 24 h (Parsek and Greenberg 2000; Papenfort and Bassler 2016). The cell densities of each culture were detected by measuring optical density at 600 nm (OD600). All the experiments were repeated three times.

Crystal structures of two-component response regulators

The crystal structures of two-component RRs identified by gene annotation were predicted by the program SWISS-MODEL (https://swissmodel.expasy.org/) (Kiefer et al. 2009; Zhao et al. 2014). The interactions between the putative PloS and QS signal molecules were simulated by SwissDock (http://www.swissdock.ch/) (Grosdidier et al. 2011).

Quantitative PCR

Trueperella pyogenes TP8 were cultured in BHIF containing 100 μM of C12HSL or C8HSL at 38 °C for 24 h, respectively. Total RNA was extracted by TRIzol® reagent, and the expression levels of assayed genes were determined using QIAGEN OneStep RT-PCR Kit according to the manufacturer’s instructions. Specific primers for ploS and ploR were designed using the Primer 3.0 software (http://frodo.wi.mit.edu/primer3) based on the consensus of sequences deposited in GenBank. PCR primers used are as follows: (F, forward primer; R, reverse primer; all sequences are 5′–3′) ploS, F: GCGGAATGCTCACGGTCTTTA, R: TGAGCGAGATGACGGTGAGG; ploR, F: GACATGGTGCGAGCCGTTC, R: GTAGACCCGTGCGGCGATA. Primers for other virulence genes and reference gene ftsY were according to a previous study (Zhao et al. 2013a). All the experiments were performed in triplicate. Gene expression was calculated using the 2−ΔCT method (Pfaffl 2001).

Biofilm production

Production of biofilm was detected by crystal violet staining and then quantified at OD595 as previously described (O’Toole and Kolter 1998). Briefly, equal amounts (1.0 × 108 CFU) of T. pyogenes TP8 were cultured in BHIF broth supplied with gradient concentrations of C12HSL or C8HSL at 38 °C for 24 h, respectively. Then the unattached bacteria were gently removed and the tubes were washed three times with PBS. Finally, the biofilms were stained with 0.2% (wt/vol) crystal violet for 30 min and quantified at OD595 after dissolution by 95% ethanol. All the experiments were performed in triplicate.

Hemolysis determinations

Trueperella pyogenes TP8 was cultured in BHIF broth containing gradient concentrations of C12HSL or C8HSL at 38 °C for 24 h, respectively. The supernatant of each culture was filtered through a 0.45-µm filter, and then the hemolysis was determined as previously described (Plank et al. 1994). Briefly, the filtered supernatant of T. pyogenes was incubated with erythrocyte suspensions (5 × 106 cells/ml) at 37 °C for 30 min. After soft centrifugation (1000×g, 5 min) at 4 °C, the cell supernatants were assayed for absorbance at 540 nm. The absorbance of a cell supernatant obtained using 1% Triton X-100 was defined as 100% lysis. The hemolytic activity of each experiment was expressed as the percentage of absorbance compared with that observed after 100% lysis induced by Triton X-100. The supernatant of an untreated erythrocyte suspension in PBS was used as a blank.

Statistical analyses

All the data were presented and statistically analyzed by the software Graphpad Prism version 5.0 (San Diego, CA). Mean values were compared by using t test or one-way ANOVA and a subsequent Tukey–Kramer post hoc test using a 95% confidence interval.

Results

Comparative transcriptome analysis

The results of comparative transcriptome analysis (Fig. 1) showed that the global expression profiles of the highly virulent isolate, T. pyogenes TP7 were generally similar to that of the moderately virulent isolate, T. pyogenes TP8. Nevertheless, there were also 39 significantly up-regulated genes in TP7 and 19 in TP8 (P < 0.05) (Supplementary Table S1). Notably, the expression levels of plo and a DNA-binding RR (X956_RS10170, locus NZ_CP007003: 2232171–2240748) were significantly increased in TP7. Based on the genome annotation of T. pyogenes TP8 (Zhao et al. 2014), we identified a total of ten typical DNA-binding RRs, and the expressions of most regulators were not significantly different among TP7 and TP8 (Supplementary Table S2), except for X956_RS10170 (up-regulated in TP7) and X956_RS02475 (up-regulated in TP8).

Fig. 1
figure 1

Comparative transcriptome analyses of a highly virulent isolate T. pyogenes TP7 and a moderately virulent isolate T. pyogenes TP8. a Heat map of differentially expressed genes of T. pyogenes TP7 and T. pyogenes TP8. b Volcano plot of differentially expressed. FDR false discovery rate, FC fold-change

Full size image

Protein crystal structure analysis

Crystal structures of the ten identified DNA-binding RRs were predicted and are shown in Fig. 2a–j. A previous review described a LuxR-type two-component protein named ploR with a CheY-type receiver domain and a C-terminal LuxR-type helix-turn-helix (HTH) DNA-binding domain as the activator of pyolysin encoding gene plo (Jost and Billington 2005). And there was a sensor HK named PloS immediately upstream on the open reading frame of PloR (unpublished data). Unfortunately, the information provided in that review was limited and yet, no follow-up appeared. Among the screened DNA-binding RRs in this study, only X956_RS10170 had a CheY-type receiver domain and a C-terminal LuxR-type HTH DNA-binding domain (Fig. 2j), as well as an adjacent sensor HK (X956_RS10165) immediately upstream (Fig. 2k). Therefore, based on the similar structure description of previous review and our transcriptome data (Jost and Billington 2005), we temporarily named the DNA-binding RR (X956_RS10170) as PloR, and the cognate sensor HK was named as PloS. These results were then validated by quantitative PCR, indeed, the expression levels of ploR and plo genes were significantly increased in TP7 (Fig. 3).

Fig. 2
figure 2

Predicted crystal structures of DNA-binding response regulators in T. pyogenes. a X956_RS00305; b X956_RS01350; c X956_RS01065; d X956_RS05910; e X956_RS02715; f X956_RS02475; g X956_RS04350; h X956_RS07525; i X956_RS10135; j X956_RS10170; k X956_RS10165

Full size image
Fig. 3
figure 3

Expression of virulence genes in the highly virulent isolate T. pyogenes TP7 and the moderately virulent isolate T. pyogenes TP8. Data are means ± SEM, and representative of three experiments. *P < 0.05; **P < 0.01, t test

Full size image

HSLs can influence the growth and virulence of T. pyogenes

As shown in Fig. 4a, the supernatants of P. aeruginosa and E. coli could inhibit the growth of T. pyogenes TP8. The result of mathematical modeling showed that the putative PloS protein could almost fully dock C12HSL in the periplasmic sensing domain followed by C8HSL (Fig. 4b). To probe the effect of HSLs on T. pyogenes growth, TP8 was cultured in BHIF containing gradient concentrations of C12HSL or C8HSL, and the results showed that both molecules could significantly inhibit the growth of T. pyogenes (Fig. 5a). Additionally, the effect of C12HSL from P. aeruginosa was higher than C8HSL from E. coli, especially when the concentrations were higher than 100 µM. This result was also according to the finding that the supernatant of P. aeruginosa could significantly inhibit the growth of T. pyogenes (Fig. 4a).

Fig. 4
figure 4

Supernatant filtrates of P. aeruginosa and E. coli can inhibit the growth of T. pyogenes TP8. a Cell density of T. pyogenes TP8 when cultured in the filtrates of P. aeruginosa and E. coli. Data are means ± SEM, and representative of three experiments. *P < 0.05, one-way ANOVA (Tukey–Kramer post hoc). b Docking of PloS with HSLs

Full size image
Fig. 5
figure 5

HSLs of P. aeruginosa and E. coli can influence the growth (a) and expression of virulence genes (b) of T. pyogenes. Data are means ± SEM, and representative of three experiments. *P < 0.05; **P < 0.01, one-way ANOVA (Tukey–Kramer post hoc)

Full size image

To investigate the effect of HSLs on the expression of virulence genes in T. pyogenes, TP8 was cultured in BHIF containing 100 µM of C12HSL or C8HSL. The results of qPCR showed that the expressions of ploS, ploR and plo were significantly impaired, whereas the expression of fimbriae-encoding genes fimA and fimC were up-regulated (Fig. 5b). Our previous study confirmed that the expression of pyolysin and fimbriae were peaked at different time points in vivo (Zhao et al. 2013a), and these interesting results suggested that the putative ploR might be a bifunctional regulator and play an important role during the process of colonization. As shown in Fig. 6a, the addition of HSLs could significantly inhibit the biofilm production of TP8, and the effect of C12HSL was stronger than C8HSL. The hemolysis of T. pyogenes TP8 was significantly suppressed by HSLs (Fig. 6b). Collectively, the QS signals of P. aeruginosa and E. coli can simultaneously inhibit the growth and virulence of T. pyogenes, and this effect is probably due to the binding of HSLs to the putative sensor kinase, PloS.

Fig. 6
figure 6

HSLs of P. aeruginosa and E. coli can inhibit biofilm production (a) and hemolysis (b) of T. pyogenes. Data are means ± SEM, and representative of three experiments. *P < 0.05; **P < 0.01, ***P < 0.001, one-way ANOVA (Tukey–Kramer post hoc)

Full size image

Discussion

As a commensal species, the source of infecting T. pyogenes is considered as autogenous (Jost and Billington 2005). Although T. pyogenes can act as a primary pathogen and most suppurative lesions of body surface can be easily cured by artificially removing the pus, the undetectable intracorporal abscesses will compromise the immune system of the host, leading to higher susceptibility to subsequent polybacterial infections and finally kill the host (Zhao et al. 2011). Our study here provides an explanation for the substitution of dominant bacterial population during the development of forest musk deer abscess disease, and the QS signal molecules of P. aeruginosa and E. coli may contribute to the development of novel antimicrobials against T. pyogenes infection.

Bacterial cell-cell communication is an important component in pathogenesis and disease control. The two-component RRs serve as a signaling activator initiate important intracellular responses or extracellular responses, and structural analysis identified that almost 95% of the reported prokaryotic transcription factors use the helix-turn-helix motif to bind the corresponding target DNA sequences (West and Stock 2001; Capra and Laub 2012; Zschiedrich et al. 2016). Small signal molecules can enhance intra- or interspecific communication and facilitate the adaption of bacteria during infection by guiding the expression of regulatory proteins (Short et al. 2014; Papenfort and Bassler 2016). In this study, the results of transcriptome analysis suggested that only the expression of two-component RR X956_RS10170 (PloR) was significantly up-regulated in the highly virulent strain T. pyogenes TP7 (Supplementary Tables S1 and 2). Although we do not know whether the X956_RS10165 and X956_RS10170 genes are actually the same as the predicted PloS/PloR (Jost and Billington 2005), our mathematical modeling and experimental results suggested that the putative ploS (X956_RS10165) could dock the QS signals of P. aeruginosa and E. coli, and might therefore cause a series of decreases in cell growth, PLO expression, biofilm production and hemolysis of T. pyogenes by suppressing the function of ploR (X956_RS10170) (Figs. 4, 5, 6). Chu et al. (2013) also suggested that the signal molecule analogs produced by Staphylococcus delphini could inhibit the QS system of Gram-negative bacteria by binding the central LuxR-type RR of QS system, and this confirmed that bacteria may employ a different strategy to collaborate or compete with their neighbors for space and resources.

In summary, here we report the first transcriptome analysis of T. pyogenes and provide eagerly awaited information about the genetic features on its virulence regulation. Importantly, we find the QS signals of P. aeruginosa and E. coli can inhibit the growth and virulence of T. pyogenes, and therefore, provide a new strategy for the development of antibiotic substitutes to control the related diseases on the basis of cell-cell communication.