Abstract
Methicillin-resistant Staphylococcus aureus (MRSA) is a leading cause of nosocomial infections worldwide. MRSA resists nearly all β-lactam antibiotics that have a bactericidal activity and a signal inducer effect. However, studies have yet to clarify whether the inducer effect of empirically used β-lactams stimulates MRSA pathogenicity in vivo. Here, we showed that a new cluster of tandem lipoprotein genes (tlpps) was upregulated in MRSA in response to the subinhibitory concentrations of β-lactam induction. The increased Tlpps significantly altered immune responses by macrophages with high IL-6 and TNFα levels. The deletion of the tlpps mutant (N315Δtlpps) significantly decreased the proinflammatory cytokine levels in vitro and in vivo. The bacterial loads of N315Δtlpps in the mouse kidney were also reduced compared with those of the wild type N315. The β-lactam-treated MRSA exacerbated cutaneous infections with increased lesion size, extended illness, and flake-like abscess-formation compared with those of the nontreatment. The β-lactam antibiotics that promoted the MRSA pathogenicity were SarA dependent, and the increasing expression of tlpps after β-lactam treatment was directly controlled by the global regulator SarA. Overall, our findings suggested that β-lactams should be used carefully because it might lead to a worse outcome of MRSA infection than inaction in the treatment.
Author summary β-lactams are widely used in practice to treat infectious diseases, however, β-lactams worsening the outcome of a certain disease is poorly understood. In this study, we have identified a new cluster of tandem lipoprotein genes (tlpps) that is upregulated in the major clinically prevalent MRSA clones in response to the subinhibitory concentrations of β-lactams induction. The major highlight in this work is that β-lactams induce SarA expression, and then SarA directly binds to the tlpp cluster promoter region and upregulates the tlpp expression in MRSA. Moreover, the β-lactam stimulated Tlpps are important virulence factors that enhance MRSA pathogenicity. The deletion of the tlpps mutant significantly decreases the proinflammatory cytokine levels in vitro and in vivo. The β-lactam induced Tlpps enhance the host inflammatory responses by triggering the expression of IL-6 and TNFα, thereby promoting bacterial colonization and abscess formation. These data elucidate that β-lactams can worsen the outcome of MRSA infection through the induction of tlpps that are controlled by the global regulator SarA.
Introduction
Infectious diseases caused by bacteria are common and adversely affect human health worldwide. The discovery of antibiotics for antibacterial application was a remarkable achievement in the 20th century. In the therapeutic use of antibiotics in humans and animals, bacteria encounter wide gradients of antibiotic concentrations in host bodies [1]. Antibiotics can serve as signal inducers, in addition to their clinically important antibacterial activity, and influence the physiological characteristics of bacteria and trigger various cellular responses in bacterial species. Low levels of antibiotics can induce extracellular DNA release, virulence factor production, and biofilm formation [2,3], resulting in a worse outcome of bacterial infections. Therefore, the mechanisms underlying the stimulation of pathogenicity in a bacterial population at subinhibitory antibiotic concentrations should be understood.
Methicillin-resistant Staphylococcus aureus (MRSA) is a leading pathogen with notable pathogenic effects. MRSA causes a wide range of diseases, including acute skin and soft tissue infections, chronic and persistent endocarditis, osteomyelitis, and pneumonia [4,5]. MRSA infections cause greater morbidity and mortality than methicillin-susceptible S. aureus (MSSA) infections do [6,7]. However, the underlying mechanisms of these effects remain unclear. Several studies have suggested that as-yet-unidentified virulence factors or inappropriate treatments contribute to the poor outcome of MRSA infections [6,8,9]. It has been reported that between 30% and 80% of individuals infected with MRSA are inappropriately treated often with β-lactam antibiotics because of the failure to recognize MRSA infection initially [10,11]. Accumulated data have demonstrated that the subinhibitory concentrations of β-lactam antibiotics can promote S. aureus pathogenicity by increasing the expression of alpha-toxin [12], Panton-Valentine leukocidin (PVL) [2], enterotoxins [13], or staphylococcal protein A (SPA) in vitro [14]. Nevertheless, the contribution of the certain altered virulence factor to the MRSA pathogenesis in vivo has yet determined, and the molecular mechanisms underlying β-lactams modulating MRSA pathogenesis remain largely unknown.
Lipoproteins (Lpps) are an abundant family of proteins anchored in the bacterial membrane, and they account for at least 2% of a bacterial proteome [15,16]. S. aureus encodes 55-70 putative Lpps, and approximately 50% of these Lpps are annotated as chaperones or as transporters for amino acids, peptides, iron, and zinc [17]. Many of the proposed Lpps (more than 30%) in S. aureus are conserved hypothetical proteins of unknown functions [16]. Most virulent MRSA strains, such as USA300, carry a conserved genomic island termed vSaα (nonphage and nonstaphylococcal cassette chromosome genomic island) that encodes numerous homologous lpps arranged in tandem, which is referred to as “tandem lipoproteins” (tlpps) or “lipoprotein-like” (lpl) [15,18]. This tlpp cluster likely represents the paralogous genes that have diverged after a duplication event in S. aureus [17]. MRSA USA300 belonging to the clonal complex CC8 carries 15 (22%) hypothetical Tlpps. Of these Tlpps, 9 are specific to the vSaα genomic island [15]. By comparison, N315 belonging to the clonal complex CC5 carries 12 (21%) hypothetical Tlpps. Of these Tlpps, 9 Lpl proteins are specific to the vSaα genomic island (S1 Table).
Some staphylococcal Lpps can trigger host cell invasion, increase bacterial pathogenicity, and contribute to the epidemic of CC8 and CC5 strains [19,20]. However, the exact roles of the Tlpp proteins are unclear. Whether these Tlpp proteins can be induced by the subinhibitory concentrations of antibiotics and contribute to the pathogenesis of MRSA have yet to be determined. In this study, we demonstrated that a new tlpp cluster of MRSA was upregulated in response to the subinhibitory concentrations of β-lactam induction. The increased Tlpps in MRSA significantly altered immune responses by increasing the IL-6 and TNFα levels of the macrophages. Null Tlpp MRSA mutant infection decreased the IL-6 and TNFα levels in serum and the bacterial burdens in kidney of a mouse model. β-lactam-treated MRSA N315 exhibited an increased pathogenicity with severe cutaneous infection and abscess formation. Moreover, the β-lactam antibiotics that promoted pathogenicity in MRSA were SarA dependent, and the increasing Tlpp expression after β-lactam treatment was directly controlled by the global regulator SarA.
Results
β-lactam antibiotics stimulated tandem lipoprotein expression in MRSA
It is reported that the subinhibitory concentrations of β-lactam antibiotics can induce the production of some S. aureus toxins, such as alpha-toxin [12], PVL [2], and enterotoxins [13], or immune evasion molecules, such as SPA [14]. Here, S. aureus N315, which is a globally prevalent sequence type 5 (ST5) MRSA strain [21], was tested for its antibiotic response to identify new factors contributing to MRSA pathogenesis in antibiotic induction. The minimal inhibitory concentrations (MICs) of β-lactam antibiotics, including oxacillin (OXA), methicillin (MET), cefoxitin (FOX), imipenem (IMI), meropenem (MER), chloramphenicol (CHL), vancomycin (VAN), kanamycin (KAN), and erythromycin (ERY) against N315 were determined (S2 Table). The sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) revealed that a protein band (approximately 30 kDa) was upregulated in the subinhibitory concentrations of β-lactam-treated bacteria compared with the nontreatment control (Fig 1A). By contrast, CHL, VAN, KAN, and ERY did not have induction effects on this protein band. Further observations indicated that the subinhibitory concentrations of OXA exerted a broad-spectrum induction effect on other major clinically prevalent MRSA clones, including ST88, ST239, ST59, ST1, and ST398, which displayed the same upregulated protein band as that in the OXA-treated MRSA compared with the nontreatment one (Fig 1B).
The protein band was excised from the SDS-PAGE gel and analyzed through liquid chromatography tandem mass spectrometry (LC-MS/MS) to characterize the β-lactam-induced proteins in the MRSA strains. The detected peptides matched with 68 proteins in the N315 proteome (S3 Table). Most known metabolic enzymes were excluded, and three putative Tlpps, namely, Tlpp3 (SA2273), Tlpp2 (SA2274), and Tlpp1(SA2275), encoded by a consecutive gene cluster were selected on the basis of theoretical molecular weights for the analysis (Fig 2A and S1 Table). SA2273, SA2274, and SA2275 were annotated as hypothetical proteins in the N315 genome (GenBank accession no. BA000018.3). A typical Lpp precursor contained a signal peptide at the N-terminal and a characteristic conserved three-amino acid lipobox in front of the invariable cysteine [(LVI) (ASTG) (GA)↓C] [16,17]. Both Tlpp1 and Tlpp3 possessed signal peptides and “lipobox” sequences, whereas Tlpp2 comprised a transmembrane helix domain at the N-terminal (S1A Fig). These proteins were annotated as Tlpps belonging to a domain of unknown function (DUF) 576 protein family on the Pfam database [22]. Considering that Tlpps account for more than 62.6% of amino acid identity (S1B Fig), we prepared recombinant Tlpp1 proteins (Fig 1C) and the corresponding antibodies, and verified the authority of the β-lactam-stimulated proteins through Western blot analysis (Fig 1D). Although several MRSA clones, such as ST8 and ST239, contained only two of the three tlpps in their genomes, the tlpp cluster was widely distributed among the major prevalent MRSA clones (S4 Table). This observation was consistent with the finding that several major MRSA clonal strains upregulated Tlpps in response to the subinhibitory concentrations of β-lactam treatment (Fig 1B).
tlppl, tlpp2, and tlpp3 were co-transcribed, and their expression was induced by β-lactam antibiotics in a dose- and time-dependent manner
Reverse-transcription polymerase chain reaction (RT-PCR) was performed using RNA extracted from MRSA N315 by specific primers to test whether tlpp1, tlpp2, and tlpp3 were co-transcribed (Fig 2A and S5 Table). The comparison of the RT-PCR results with the template of the genomic DNA or RNA only revealed that tlpp1, tlpp2, and tlpp3 were co-transcribed from the tlpp1 promoter (Fig 2B). We further examined the influence of β-lactams on tlpp expression. Reverse-transcription quantitative PCR (RT-qPCR) showed that the mRNA levels of tlpp1, tlpp2, and tlpp3 were upregulated in N315 after treatment with the subinhibitory concentrations of OXA (Fig 2C). Western blot analysis demonstrated that the protein levels of Tlpps in both N315 total cell lysates (Fig 2D and 2E) and the culture supernatant (Fig 2F and 2G) increased in a dose-dependent manner after OXA treatment was administered. Tlpp expression was also upregulated by N315 in a dose-dependent manner after MET treatment was given (S2 Fig). Furthermore, the Tlpp expression by N315 increased in a time-dependent manner after OXA treatment was administered (Fig 2H and 2I). These results verified that MRSA Tlpps could be released from the bacteria and secreted to the culture, and their production was influenced by the subinhibitory concentrations of β-lactam antibiotics.
β-lactam-induced Tlpps triggered proinflammatory cytokine production by macrophages
In Gram-negative bacteria, the cell wall-associated lipopolysaccharides are the main molecules involved in activating the innate immune system of hosts via TLR4 interaction [23], while in Gram-positive bacteria, the releasable Lpps may be the main factor that exert a similar function by triggering the TLR2-MyD88-NF-κB signaling pathway, thereby inducing proinflammatory cytokines [17,19]. To determine whether the β-lactam-induced MRSA Tlpps involved in the innate immune activation, we cultured mouse RAW264.7 macrophages in media containing 5% culture supernatant of N315 post-treated with different concentrations of OXA for 6 h. We then determined the levels of proinflammatory cytokines, such as IL-6 and TNFα, in the cell-cultured media (DMEN). The results revealed that IL-6 and TNFα levels were gradually increased by macrophages stimulated by the culture supernatant of N315 treated with OXA in a dose-dependent manner (Fig 3A and 3B). Then, we wondered whether the MRSA Tlpps promoted the cytokine expression in macrophages, a tlpps deletion mutant (N315Δtlpps) with a pYT3-Δtlpps plasmid and a tlpps-overexpressing strain (N315Δtlpps/pLI-tlpps) with a pLI-tlpps plasmid (Fig 3C and S3 Fig) were constructed for macrophage infection. Results showed that the production of IL-6 and TNFα in macrophages significantly decreased after treated with N315Δtlpps at a multiplicity of infection (MOI) of 30 compared with those of the wild-type N315 administered. By contrast, higher levels of IL-6 and TNFα were detected in macrophages treated with N315Δtlpps/pLI-tlpps but not with N315Δtlpps, which carried an empty pLI50 vector (Fig 3D and 3E). Similar results were observed when the bacterial culture supernatant was used (Fig 3G and 3H) because of the loss of Tlpp expression in the supernatant of N315Δtlpps (Fig 3F), indicating that the increased levels of IL-6 and TNFα by macrophages depended on the expression of N315 Tlpps that responded to β-lactams in a dose-dependent manner. However, the recombinant Tlpp1 purified from Escherichia coli (Fig 1C) exhibited no effect on the levels of IL-6 and TNFα by macrophages (S4 Fig), suggesting that a correctly triacylated Tlpp or a long-chain N-acylated Lpp was needed for the recognition by TLR2-TLR1 receptors to trigger immune response by macrophages [24].
Tlpps contributed to the virulence of MRSA
Consistent with the observed immune response by macrophages, the levels of IL-6 and TNFα in mice 6 h post-challenge with 1 × 107 N315Δtlpps cells by tail vein injection were significantly lower than those of N315 was administered (Fig 4A and 4B). The complementary overexpression of tlpps (N315Δtlpps/pLI-tlpps) strain stimulated even higher levels of IL-6 and TNFα in mice, but not of the empty pLI50 plasmid-carrying strain (N315Δtlpps/pLI50). We also investigated whether the tlpp cluster of MRSA was associated with the bacterial burden in a mouse model. The mice were infected intravenously with pGFP plasmid-transformed N315 and N315Δtlpps for 5 days (S6 Table), and bacterial colonization was tracked through an animal imaging system. The fluorescence intensity of the GFP in the murine organs (i.e., heart, lung, liver, spleen, and kidney) was measured, and the results indicated that the radiant efficiency in the kidneys of the mice injected with N315 was significantly higher than those infected with N315Δtlpps (Fig 4C and 4D). Consistent with the radiant efficiency, the bacterial loads in the kidneys of the N315-infected mice were also significantly higher than that of the N315Δtlpps-infected ones (Fig 4E). Overall, these data suggested that the systemic inflammatory response in MRSA infection was associated with Tlpps, and MRSA Tlpps contributed to bacterial colonization and virulence.
β-lactam-antibiotic-stimulated Tlpps promoted the pathogenesis of MRSA
We determined whether β-lactam-stimulated Tlpps enhanced the pathogenesis of MRSA. A mouse subcutaneous infection model was used to evaluate the contribution of the subinhibitory concentrations of OXA or Tlpps to skin and soft tissue infections. The mice were subcutaneously injected in both flanks with 5 × 107 OXA-treated N315 and N315Δtlpps cells and intraperitoneally injected with 1 μg of OXA per gram weight twice a day for 14 days. The course of infection was monitored every day. The untreated N315- and N315Δtlpps-infected and PBS-injected mice served as the controls. The abscesses caused by the OXA-treated N315 was significantly larger than those caused by the OXA-treated N315Δtlpps, untreated N315, and untreated N315Δtlpps (Fig 5A), and this observation was further shown in the photographs of the skin lesions (Fig 5B and S5 Fig). Histological examinations indicated that the skin of the OXA-treated N315-challenged mice exhibited less extensive inflammation with leukocyte infiltration, destroyed skin structure, and more flake-like abscess-formation than those of the untreated N315-infected and PBS-injected mice (Fig 5C). By contrast, the skin of the OXA-treated N315Δtlpps-challenged mice displayed more leukocyte infiltration and sporadic abscess formation than that of the untreated N315Δtlpps-infected and PBS-injected mice. The corium layer of the N315-challenged and PBS-injected mice showed an extensive inflammation with leukocyte infiltration, although abscess formation was not observed compared with that of the N315Δtlpps-infected and PBS-injected mice. These pathological phenomena might be caused by the β-lactam-stimulated MRSA Tlpps, which stimulated the IL-6 and TNFα levels in mice (Fig 5D and 5E), thereby silencing the immune responses through granulocytic and monocyticmyeloid-derived suppressor cells induced by IL-6 [17], increasing immune cells death due to the tremendous TNFα release [24], and promoting bacterial colonization and abscess formation. Overall, these data confirmed that β-lactam-stimulated Tlpps worsened the MRSA infections.
β-lactam-stimulated Tlpp expression in MRSA was SarA dependent
β-lactams as antibiotics block the cell wall synthesis of bacteria to exert antimicrobial effects. By contrast, β-lactams as inductors may trigger global regulatory networks to modulate virulence in S. aureus [13]. RT-qPCR revealed that the expression of global regulators, including sarA, agrA, RNAIII, rot, ccpA, and saeR, increased in OXA-treated N315 compared with those in the untreated ones (Fig 6A). SarA and agrA were the most altered regulators, whereas sigB was unchanged, which was consistent with the Western blot results of SigB stimulated with different OXA concentrations (S6 Fig). AgrA is a downstream regulator of SarA [25]. In this study, we examined whether SarA was upregulated in N315 upon OXA treatment. Western blot analysis indicated that both SarA and Tlpps increased in a dose-dependent manner in response to the β-lactam antibiotic treatment (Fig 6B, 6C and S2 Fig). The deletion of sarA reduced the Tlpps levels in both N315 and its culture supernatant (Fig 6D and S7 Fig). The sarA overexpressing strain (N315ΔsarA/pLI-sarA) produced even more Tlpps compared with the wild type N315, whereas the empty pLI50 plasmid-carrying strain (N315ΔsarA/pLI50) did not. Consistent with the decreased Tlpps in N315ΔsarA, N315ΔsarA or its culture supernatant stimulated less IL-6 and TNFα expression in RAW264.7 cells than N315 or its culture supernatant did. By contrast, N315ΔsarA/pLI-sarA or its culture supernatant induced more IL-6 and TNFα expression than the wild-type N315 or N315ΔsarA/pLI50 did (S8 Fig). The IL-6 and TNFα levels in mice challenged with N315ΔsarA decreased compared with those in the mice challenged with N315. N315ΔsarA/pLI-sarA stimulated even higher cytokine levels in mice compared with those of N315 or N315ΔsarA induced (Fig 6E and 6F). Taken together, these data indicated that the β-lactam-stimulated Tlpp expression in MRSA was SarA dependent.
SarA directly controlled the β-lactam-stimulated Tlpp expression in MRSA
To investigate the effect of SarA on β-lactam-stimulated Tlpp expression, we constructed a reporter vector (pOS1-tlppsP) containing the tlpp promoter-controlled lacZ gene (S6 Table) and performed β-galactosidase assay by transforming pOS1-tlppsP into the MRSA strains N315 and N315ΔsarA. The results revealed that the β-galactosidase activity was significantly lower in the sarA mutant than that in N315. Moreover, the β-galactosidase activity presented no significant change in the sarA mutant after OXA treatment compared with the untreated N315ΔsarA (Fig 7A). However, OXA treatment significantly increased the β-galactosidase activity in N315, and this increase was consistent with the increasing Tlpp in MRSA strains (Fig 7A, Fig 1A and 1B). Consistent with the β-galactosidase assay results, Western blot analysis demonstrated that the Tlpp expression in the sarA mutant could not respond to β-lactam simulation (Fig 7B and 7C), suggesting that SarA controlled the Tlpp expression in response to the β-lactam antibiotic induction.
A global regulator can recognize specific motifs in the promoter regions of a certain gene, thereby controlling the gene expression [13,26,27]. We analyzed the binding motif of SarA [28,29] in the promoter regions of tlpps and found a typically predicted SarA box (Fig 7D). The recombinant His-tagged SarA was prepared and purified from E. coli (S7C Fig), and the electrophoretic mobility shift assay (EMSA) showed that the recombinant SarA proteins bound to the tlpp cluster promoter region that carried the putative SarA binding box (Fig 7E, 7F and 7H). No shifting band was observed when the AT-rich SarA box was mutated to become GC-rich (Fig 7E, 7G and 7H). These data indicated that S. aureus SarA could directly bind to the tlpp cluster promoter region, thereby upregulating the tlpp expression in the presence of β-lactams. The AT-rich motif (ATTTAAT) in the promoter regions of tlpps is essential for SarA binding and regulating.
Discussion
MRSA is distinct from MSSA in terms of the acquisition of a genetic element called staphylococcal cassette chromosome mec (SCCmec) in which mecA encodes an alternative penicillin-binding protein 2a (PBP2a) with a low affinity for β-lactams [21]. As such, MRSA strains are resistant to nearly all β-lactam antibiotics [6]. As antibiotics, β-lactams bind to penicillin-binding proteins (PBPs) and inhibit the transpeptidation and transglycosylation of the cell wall, resulting in a weakened cell wall and inducing cell lysis and death [30]. This type of antibiotics, particularly cephalosporins and β-lactam/β-lactamase inhibitor combinations, has been empirically used for clinical treatments of infectious diseases [31]. The subinhibitory concentrations of antistaphylococcal agents might arise because of antibiotic-resistant microorganisms or pharmacokinetics of antibiotics [13,31]. For MRSA infections, which are not initially recognized, β-lactams are not only ineffective in the treatment of infections but also likely contributing to poor outcomes by enhancing the pathogenicity of MRSA. Nonetheless, the underlying mechanisms remain obscure [9]. In this study, we showed that a three-gene constituent tlpp cluster upregulated in response to β-lactam induction in a dose- and time-dependent manner. This tlpp cluster, belonging to a DUF576 protein family of unknown function [19,22], was widely distributed among the major prevalent MRSA clones (Fig 1B and S4 Table). Tlpps could be upregulated after treatment with nearly all β-lactam antibiotics, such as penicillin (oxacillin and methicillin), cephalosporins (cefoxitin), and carbapenems (imipenem and meropenem), but not with vancomycin, kanamycin, and erythromycin (Fig 1A).
In addition to antimicrobial activity, signal induction may be implemented by the subinhibitory concentrations of β-lactams. They actively promote S. aureus biofilm formation [3], induce PBP2a to reduce peptidoglycan crosslinking in MRSA [6], and enhance virulence factors, such as alpha-toxins, PVL, SPA, and enterotoxins [2,32–34]. In contrast to β-lactam-stimulated SPA and PVL, which have a controversial pathogenic role in S. aureus [2], the Lpps of S. aureus are crucial players in alerting the host immune system by recognizing TLR2/TLR1 or TLR2/TLR6 receptors [24,35]. In general, the Lpps of Gram-positive bacteria are anchored in the outer leaflet of the cytoplasmic membrane [17]. We observed that the β-lactam-stimulated MRSA Tlpps could be released; Tlpps could be detected in both cell lysates and the culture supernatant through Western blot analysis (Fig 2D, 2F, 3C and 3F). MRSA Tlpps could induce the production of IL-6 and TNFα proinflammatory cytokines of macrophages which stimulated with the culture supernatant of N315 post-treated with different OXA concentrations (Fig 3A and 3B). Although the tlpp deletion mutant (N315Δtlpps) infection induced less IL-6 and TNFα production in mice compared with the wild-type N315 was administrated, OXA-treated N315Δtlpps-infected mice still produced higher IL-6 and TNFα cytokines compared with those of the untreated N315-challenged mice (Fig 5D and 5E), suggesting that other mechanisms might be involved in the immune system modulation by β-lactam-treated MRSA. For instance, β-lactam-promoted PBP2a induction can diminish the peptidoglycan crosslinking, thereby enhancing phagocytic degradation and detection and resulting in strong IL-1β production [6].
In addition to stimulating the immune system, increasing the pathogenicity of MRSA was attributed to β-lactam-stimulated Tlpps. In comparison with wild-type N315, the N315Δtlpps strain significantly decreased the bacterial burden in mouse kidneys (Fig 4C, 4D and 4E). β-lactam antibiotic treatment exacerbated MRSA infections in the mouse skin infection model, and the histological examinations of the OXA-treated N315-challenged mouse skin displayed less extensive infiltration with leukocytes, destroyed skin structure, and easily promoted abscess formation (Fig 5). A possible explanation is that the induced IL-6 and TNFα expression by β-lactam-stimulated MRSA Tlpps silenced the innate immune responses [24], thereby facilitating MRSA colonization and infection. Our findings might indicate that MRSA infections were associated with higher morbidity and mortality than MSSA [7,36].
β-lactams can induce PVL expression in S. aureus by interfering with PBP1 and triggering SarA and Rot global regulators [2]. Our results showed that SarA and AgrA were the most upregulated regulators in MRSA N315 after OXA treatment (Fig 6A). The deletion of SarA (N315ΔsarA) failed to upregulate tlpps even under OXA treatment (Fig 7A and 7B), indicating that the β-lactam-induced Tlpp expression in MRSA was SarA controlled. EMSA data revealed that the regulation of SarA on Tlpp expression was direct (Fig 7F). However, further investigations should be performed to clarify how β-lactams trigger the SarA expression.
In conclusion, this work focused on the function and regulation of a new tlpp cluster in response to the induction of subinhibitory concentrations of β-lactams. We demonstrated that the increased Tlpps in MRSA significantly enhanced inflammatory response by triggering IL-6 and TNFα levels in vitro and in vivo, thereby possibly contributing to bacterial pathogenicity and worsening the outcome of MRSA infection by reducing host immune responses and promoting bacterial colonization. β-lactam-stimulated MRSA Tlpps was SarA dependent, and the upregulation in Tlpp expression after β-lactam treatment was directly controlled by the global regulator SarA (Fig 8). Nonetheless the pathways leading to the SarA expression trigged by β-lactam antibiotics remain unknown. Our data support the recommendation to clinicians that the discreet usage of β-lactams which possibly worsened the clinical outcome of MRSA infections.
Materials and methods
Ethics statement
BALB/c mice were purchased from Laboratory Animal Center of Army Medical University (Third Military Medical University). All animal experiments were approved by the Army Medical University Institutional Animal Care and Use Committee (protocol #SYXK-PLA-20120031). All animal experimental procedures were performed in accordance with the Regulations for the Administration of Affairs Concerning Experimental Animals approved by the State Council of People`s Republic of China.
Bacterial strains and plasmids
Bacterial strains and plasmids used in this study were listed in S6 Table. E. coli strains DH5α and BL21 (DE3) were cultivated in Luria Broth (LB) medium (Oxoid). S. aureus strains grown in Brain Heart Infusion (BHI) or Tryptic Soy Broth (TSB) medium (Oxoid).
Antibiotic susceptibility tests
Antibiotic susceptibility was determined using broth microdilution methods according to the protocols recommended by the Clinical and Laboratory Standards Institute (CLSI, 2017) [37]. The antibiotic susceptibility results for all strains were listed in S2 Table.
Preparation of recombinant Tlpp1 and SarA
Recombinant Tlpp1 and SarA proteins were prepared in our laboratory. In brief, the gene encoding for Tlpp1 or SarA was amplified from the genomic DNA of N315 by PCR, and cloned into pET28a(+) vector to construct the pET28a-tlpp1 and pET28a-sarA plasmids. Then, the recombinant plasmids were transformed into E. coli BL21(DE3) for the expression of Tlpp1-6×His and SarA-6×His fusion proteins. Cells were grown in LB with 100 μg/ml of ampicillin at 37 °C. Isopropyl-D-thiogalactopyranoside (IPTG, 5 μM) was added once an OD600 of 0.6 was achieved. The cells were cultured at 22 °C for another 6 h. The cells were then centrifuged at 10,000 × g for 15min, washed once with PBS, resuspended in the lysis buffer (50 mM Tris·Cl, 0.15 M NaCl, 1 mM phenylmethanesulfonyl fluoride (PMSF), 0.5mg/ml lysozyme, pH 8.0) and lysed using ultrasonic method. The recombinant proteins in the supernatant were purified by Ni-NTA affinity chromatography (GE Healthcare) and identified by Western blot.
Preparation of protein-specific polyclonal antibodies
Female BALB/c mice (6-8 weeks) were immunized subcutaneously with 40 μg of recombinant proteins (Tlpp1 or SarA) emulsified with complete Freund’s adjuvant (Sigma, USA) for the first time, and boosted with recombinant proteins (Tlpp1 or SarA) in incomplete Freund’s adjuvant on day 14, 28, and 35, respectively. Seven days after the last immunization, blood samples were collected, the titer of antibodies against Tlpp1 or SarA was determined by ELISA.
Preparation of the total bacteria proteins and the culture supernatant proteins
Overnight MRSA culture was diluted 1:100 in BHI medium with or without the addition of β-lactam antibiotics, cultivated at 37 °C to an optical density (OD) at 600 nm of 2.0. Then, bacterial cells in 3 ml culture were harvested by centrifugation at 10,000 × g at 4 °C. The cell pellets were washed twice with PBS, resuspended in 1 ml of cold PBS supplemented with 1% β-mercaptoethanol (Sigma, USA) and 1 mM PMSF (Beyotime, China) on ice. Cells were broken by the addition of 0.1-mm diameter zirconia/silica beads, shaking on the Minibeadbeater 16 instrument (Biospec, USA). Cell debris was removed after centrifugation at 10,000 × g for 10 min at 4 °C, and the total cell proteins in 1 ml of the supernatant were precipitated with 7.5% trichloroacetic acid (TCA)/0.2 % deoxycholic acid solution.
The culture supernatant of MRSA stimulated with or without antibiotics was harvested by centrifugation at 10,000 × g for 10 min at 4 °C. The proteins in 1 ml culture supernatant were prepared by TCA precipitation, collected by centrifugation at 15,000 × g for 10 min, washed once with ice-cold acetone, dissolved in 60 μl of PBS for use. The protein concentration was determined using the Bradford Protein Assay Kit (Beyotime, China).
Proteomic analysis
LC-MS/MS was performed to identify proteins induced by β-lactam antibiotics as previously described [38]. MRSA strain N315 was cultured in TSB with or without 2 βg/ml OXA. The total bacteria proteins were separated through SDS-PAGE, and the antibiotic-induced protein band was excised and analyzed through LC-MS/MS by using UltiMate3000 RSLCnano liquid chromatography/Bruker maxis 4G Q-TOF. The resulting peptide mass fingerprints were compared against the ORFs of N315 by using Mascot and Mascot Daemon software (Matrix Science).
RT-PCR and RT-qPCR
Total MRSA N315 RNA was extracted as previously described [39]. Overnight N315 cultures were 1:100 diluted in TSB containing 2 μg/ml OXA, cultivated at 37 °C to the early exponential phase (OD600 = 1.0). Total RNA was isolated using the TriPure isolation reagent (Roche Applied Science, Germany) after the collected cells were firstly lysed using lysostaphin (Sigma, USA). cDNA was synthesised from 500 ng of total RNA using gene-specific primers and a RevertAid First Strand cDNA Synthesis Kit (Thermo, USA). RT-PCR was used to test whether tlpp1, tlpp2, and tlpp3 were cotranscribed. RT-qPCR was performed to detect the expression levels of the tlpp genes (tlpp1, tlpp2 and tlpp3) and the global regulators (sarA, agrA, RNAIII, rot, ccpA, saeR, sigB) using SsoAdvanced™ Universal SYBR® Green Supermix (Bio-Rad, USA). The relative expression level of all tested genes was normalized to that of the 16S rRNA. All primers used are listed in S5 Table.
Construction of gene mutant strains
The tlpp cluster marker-less deletion mutant was constructed using homologous recombinant strategy described previously [39]. Briefly, the tlpp cluster deletion plasmid pYT3-Δtlpps (S3A Fig) was constructed by amplifying the upstream and downstream regions of tlpp cluster with primer pairs, up-tlpps fwd/up-tlpps rev and down-tlpps fwd/down-tlpps rev, and subcloning these fragments into E. coli-S. aureus temperature-sensitive shuttle vector pYT3. The recombinant plasmid was identified by DNA sequencing and subsequently transformed into RN4220 by electroporation, then transformed into N315. The resulting N315Δtlpps strain was generated after inducing the integration of plasmid into chromosome at 42 °C and following inducing the plasmid losing at 25 °C. The deletion of tlpp cluster was confirmed by PCR and DNA sequencing.
To construct the tlpps complementary strain, the tlpp cluster containing its potential promoter region was amplified by pLI-tlpps fwd/pLI-tlpps rev primer pairs, cloned into the expression plasmid pLI50 [40]. Then, the correctly constructed plasmid pLI-tlpps was electroporated into RN4220 and then N315Δtlpps to generate N315Δtlpp/pLI-tlpps strain. Similar strategy was used to construct N315ΔsarA/pLI-sarA. The empty pLI50 plasmid transformed N315Δtlpps and N315ΔsarA strains served the controls. All primers used are listed in S5 Table.
Cytokine determination
RAW264.7 (TIB-71™, ATCC) cells were cultured in high glucose DMEM (Thermo Fisher Scientific, USA) added 10 % (v/v) fetal bovine serum (Thermo Fisher Scientific, USA) at 37 °C with 5 % CO2. For stimulation experiment, RAW264.7 cells (106/well) were either infected with MRSA strain (MOI of 30), or cultured in DMEM medium containing 5% (v/v) bacterial culture supernatant, in a 24-well microtiter plate for 6 h as described [41]. Then, the supernatant was collected, and the levels of IL-6 and TNFα were measured with an ELISA kit following the manufacturer’s instructions (R&D Systems, USA).
To detect the ability of Tlpp proteins in inducing cytokine secretion in vivo, female BALB/c mice were infected via tail vein injection with 1 × 107 CFU of N315, N315Δtlpps, N315Δtlpps/ pLI-tlpps, N315Δtlpps/pLI50, N315ΔsarA, N315ΔsarA/pLI-sarA, and N315ΔsarA/pLI50, respectively. Blood samples were collected 6 h post infection, and the IL-6 and TNFα levels in mouse sera were determined by using ELISA.
Animal experiments
BALB/c mice were randomly divided into two groups and infected via tail vein injection with 1 × 107 CFU of the GFP-expression plasmid (pGFP) transformed N315 or N315Δtlpps, sacrificed 5 days after infection. Mouse organs (i.e., heart, lung, liver, spleen, and kidney) were isolated and subjected to the determination of GFP fluorescence efficiency in the organs with IVIS® Lumina LT system and analyzed by Living Image 4.4 Software. The bacterial loads in the infected kidneys were also counted via plate dilution assay as described [42].
For skin abscess formation, BALB/c mice were fully anesthetized with 1 % pentobarbital sodium (50 mg/kg) and the back hair was depilated completely with 6 % sodium sulfide (w/v). Then, mice were subcutaneously inoculated with 5×107 CFU of MRSA N315 and N315Δtlpps in both flanks of the murine back as described [43], and then randomly divided into two groups. The mice of the treatment group were intraperitoneally injected with 1 μg of OXA per gram weight twice a day for 14 days. The PBS-injected mice served as the controls. The abscess area assessed by the maximal length × width of the developing ulcer was measured daily.
For histopathological examination, the skin lesions were fixed by 4 % paraformaldehyde, paraffin embed, and stained with hematoxylin & eosin (H&E).
β-galactosidase assay
A tlpp promoter-LacZ reporter plasmid (pOS1-tlppsP) was constructed by inserting the tlpp promoter region into pOS1 vector, and transformed into N315 and N315ΔsarA, respectively. Then, pOS1-tlppsP-carried N315 and N315ΔsarA were cultured overnight, diluted 1:100 in BHI, and cultivated at 37 °C to an OD600 of 0.6. Bacterial cells in 200 μl culture were collected by centrifugation, suspended in 100 μl of AB buffer (100 mM KH2PO4, 100 mM NaCl, pH 7.0), and treated with lysostaphin (20 μg/ml, Sigma) for 15 min at 37 C. Then, the suspension was added with another 900 μl of ABT solution (AB buffer containing 0.1 % TritonX-100) [44], 50 μl of the solution was mixed with 10 μl of MUG (4-methylumbelliferyl-β-D-galactoside, 4 mg/mL, Sigma) and incubated for 1 h at room temperature. Then, 20 μL of the sample was mixed with 180 μl of ABT solution, and the reaction was monitored at 445 nm with an excitation wavelength of 365 nm. All samples were tested in triplicate. The LacZ activity was normalized to the cell density of OD600, and the relative activity was calculated by setting the LacZ activity from the N315 to 100%. The assay was repeated at least three times.
Electrophoretic mobility shift assays (EMSA)
The predicted tlpp cluster promoter, an AT-rich motif fragment (56 bp), was synthesized using the primer pairs (EMSA-tlppsP fwd/EMSA-tlppsPM rev) as described [45]. The corresponding mutated GC-rich motif fragment was also synthesized by primer pairs EMSA-tlppsPM fwd/EMSA-tlppsPM rev and served as the control. Ten picomole of DNA fragment was incubated with a variable amount of recombinant SarA (0 to 240 pM) in a 20 μl reaction mixture containing 10 mM HEPES (pH 7.6), 1 mM EDTA, 2 mM dithiothreitol, 50 mM KCl, 0.05 % Triton X-100, and 5 % glycerol. Binding reactions were equilibrated for 20 min at room temperature before electrophoresis. Reaction mixtures were separated on 6 % native polyacrylamide gel electrophoresis in 0.5 × TBE (Tris/boric acid/EDTA) buffer at 90 V for 2 h at 4 °C. Gels were stained by GelRed dye (Biotium, USA) and observed under UV light. The primers used are listed in S5 Table.
Statistical analysis
Statistical analysis was carried out using GraphPad Prism 6.0. Unpaired two-tailed student’s t-test, analysis of variance (ANOVA) and Mann-Whitney test were used appropriately to compare the difference between groups. Each experiment was carried out at least three times. Results were presented as mean ± standard deviations (S.D.), and a P value less than 0.05 was considered statistically significant. * P < 0.05, ** P < 0.01, *** P < 0.001, and ns represented no significance.
Acknowledgments
We would like to thank the members of the Xiancai Rao and Xuhu Mao research groups for their critical reading of the manuscript. We thank Professor Baolin Sun (University of Science and Technology of China) providing plasmid pGFP and pLI50. We also thank Professor Lefu Lan (Shanghai Institute of Materia Medica, Chinese Academy of Sciences) for providing plasmid pOS1.