ABSTRACT
Clostrididioides difficile causes severe antibiotic-associated diarrhea and colitis. C. difficile is an anaerobic, Gram-positive spore former that is highly resistant to β-lactams, the most commonly prescribed antibiotics. The resistance of C. difficile to β-lactam antibiotics allows the pathogen to replicate and cause disease in antibiotic-treated patients. However, the mechanisms of β-lactam resistance in C. difficile are not fully understood. Our data reinforce prior evidence that C. difficile produces a β-lactamase, which is a common β-lactam resistance mechanism found in other bacterial species. We identified an operon encoding a lipoprotein of unknown function and a β-lactamase that was greatly induced in response to several classes of β-lactam antibiotics. An in-frame deletion of the operon abolished β-lactamase activity in C. difficile strain 630Δerm and resulted in decreased resistance to the β-lactam ampicillin. We found that the activity of this β-lactamase, herein named BlaD, is dependent upon the redox state of the enzyme. In addition, we observed that transport of BlaD out of the cytosol and to the cell surface is facilitated by an N-terminal signal sequence. Our data demonstrate that a co-transcribed lipoprotein, BlaX, aids in BlaD activity. Further, we identified a conserved BlaRI regulatory system and demonstrated via insertional disruption that BlaRI controls transcription of the blaXD operon in response to β-lactams. These results provide support for the function of a β-lactamase in C. difficile antibiotic resistance, and reveal the unique roles of a co-regulated lipoprotein and reducing environment in β-lactamase activity.
IMPORTANCE Clostridioides difficile is an anaerobic, gastrointestinal human pathogen. One of the highest risk factors for contracting C. difficile infection is antibiotic treatment, which causes microbiome dysbiosis. C. difficile is resistant to β-lactam antibiotics, the most commonly prescribed class of antibiotics. C. difficile produces a recently discovered β-lactamase, which cleaves and inactivates numerous β-lactams. In this study, we report the contribution of this anaerobic β-lactamase to ampicillin resistance in C. difficile, as well as the transcriptional regulation of the gene, blaD, by a BlaRI system. In addition, our data demonstrate co-transcription of blaD with blaX, which encodes a membrane protein of previously unknown function. Furthermore, we provide evidence that BlaX enhances β-lactamase activity in a portion of C. difficile strains. This study demonstrates a novel interaction between a β-lactamase and a membrane protein in a Gram-positive pathogen, and due to the anaerobic nature of the β-lactamase activity, suggests that more β-lactamases are yet to be identified in other anaerobes.
INTRODUCTION
Clostridioides difficile, or C. difficile, is an anaerobic, Gram-positive, spore-forming bacterial pathogen that causes antibiotic-associated diarrhea (1-3). C. difficile infection, or CDI, can be severe, resulting in psuedomembranous colitis, intestinal rupture, and death. The Center for Disease Control (CDC) estimates that almost half a million people in the U.S. suffer from CDI per year, resulting in approximately 29,000 deaths per year (4). As a result, CDI cases add approximately $4.8 billion per year to U.S. healthcare costs (5). C. difficile was first linked to antibiotic-associated diarrhea in 1978, and antibiotic treatment is still one of the highest risk factors for CDI (2, 3). Antibiotic treatment results in gastrointestinal dysbiosis, eliminating important indigenous anaerobes, thereby allowing for C. difficile population expansion (6, 7). Antibiotic treatment of CDI is limited to the use of vancomycin, fidaxomicin, or metronidazole, due to the high resistance C. difficile exhibits for a wide array of antibiotics (8-10).
The most commonly prescribed class of antibiotics are the β-lactams, which comprise 62% of all prescribed antibiotics in the United States and are strongly associated with C. difficile infections (11-13). β-lactams are inhibitors of bacterial cell wall synthesis and are characterized by a four-membered core lactam ring (14). β-lactams are further classified into four groups based on adjoining structures: the penicillins, cephalosporins, monobactams, and carbapenems (15). All β-lactam antibiotics bind to, and thus disable, cell-wall synthesizers called penicillin-binding proteins (PBPs) of bacteria (16, 17). Since the introduction of β-lactams into modern medicine, multiple mechanisms of resistance to these antibiotics have been discovered in a variety of bacterial species. β-lactam resistance mechanisms include the production of β-lactamases, which hydrolyze the β-lactam ring and render the antibiotic ineffective, mutations acquired in PBPs that prevent binding of the β-lactams, reduced outer membrane permeability due to reduced porin expression, and efflux pumps, which prevent the antibiotic from reaching the cell wall (18-23).
The most common mechanism of β-lactam resistance occurs through the production of β-lactamase enzymes. Most of the characterized β-lactamases have been identified in Gram-negative species; in these bacteria, the β-lactamase is generally secreted into the periplasm, where the enzyme is concentrated, allowing for high levels of β-lactam resistance (24). Less common are the outer membrane-anchored β-lactamases, which may be further packaged into outer membrane vesicles, enabling the inactivation of nearby β-lactams (25-27). β-lactam resistance in Gram-positive bacteria, however, is more commonly conferred by the modification of the intended targets of the β-lactam, the penicillin-binding proteins (28). Still, β-lactamases do exist in Gram-positive bacteria (29-33). Although Gram-positive bacteria lack a periplasmic space, some species do produce membrane-bound β-lactamases (29, 34-37). A few of these enzymes are proteolytically cleaved, producing an exoenzyme that can be released from the membrane (31, 36, 38).
β-lactamase enzymes are classified into four classes: A, B, C, and D. Classes A, C, and D are serine hydrolases, while class B β-lactamases are metallohydrolases (18). Whereas β-lactamases of all classes have been discovered in Gram-negative bacteria, most Gram-positive β-lactamases belong to classes A or B (32). Class D β-lactamases were recently identified in Gram-positive bacteria, including one that is highly conserved among C. difficile isolates (33, 39). A recent study demonstrated that a β-lactamase in C. difficile confers resistance to the penicillin, cephalosporin, and monobactam class of β-lactams (39). According to the substrate profile of this enzyme, this β-lactamase belongs to the 2de functional group of β-lactamases (39, 40). The purpose of our study was to characterize the genetic organization, resistance contributions, biochemical activity, and regulation of the C. difficile β-lactamase. To accomplish this, we deleted the genes encoding the β-lactamase and the upstream predicted membrane protein in C. difficile, and examined the resulting resistance profiles and biochemical activity. Notably, we observed that the C. difficile β-lactamases are inactivated by oxygen, which has not been described for other class D β-lactamases. We also examined how this β-lactamase enzyme is transported, and detail its mechanism of regulation. We demonstrate that unlike other described β-lactamases, the C. difficile β-lactamase is co-transcribed with a membrane protein that facilitates β-lactamase processing and function. These results further our understanding of β-lactam resistance in C. difficile, which may expose approaches to prevent or treat β-lactam-associated CDI.
MATERIALS AND METHODS
Bacterial strains and growth conditions
Bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli was grown at 37°C in LB medium with 100 μg/mL ampicillin (Sigma-Aldrich) and 20 μg/mL chloramphenicol (Sigma-Aldrich) when necessary (41). C. difficile was grown anaerobically at 37°C as previously described (42) in brain heart infusion medium supplemented with 2% yeast extract (BHIS; Βecton Dickinson Company) or Mueller Hinton Broth (MHB; Difco) with 2 μg/mL thiamphenicol (Sigma-Aldrich), 3.125 – 60 μg/mL cefoperazone (Sigma-Aldrich), 0.25 – 2 μg/mL ampicillin, 0.125 – 1.5 μg/mL imipenem (US Pharmacopeia), 0.75 μg/mL vancomycin (Sigma-Aldrich), 75 μg/mL polymyxin B (Sigma-Aldrich), 1 mg/mL lysozyme (Fisher Scientific), 7.5 μg/mL nisin (MP Biomedicals), 2 μg/mL LL-37 (Anaspec), or 250 μg/mL kanamycin (Sigma-Aldrich) when specified.
Strain and plasmid construction
The oligonucleotide primers used in this study are listed in Table 2. Primer design and the template for PCR reactions were based on C. difficile strain 630 (GenBank accession NC_009089.1), except for pMC896, which was based on strain M120 (GenBank accession FN665653.1).
The blaX::erm and blaI::erm mutant strains were created by retargeting the Group II intron from pCE240 with the primers listed in Table 2, as previously described (43). To generate insertional disruptions, transconjugants were selected on 5 μg/mL erythromycin (Sigma-Aldrich), and 50 μg/mL kanamycin (Sigma-Aldrich) to select against E. coli.
The ΔblaXD mutant strain was created using a pseudo-suicide plasmid technique, as described previously, with slight variation (44). Briefly, 500 bp regions homologous to the 5’ and 3’ ends of the bla operon were amplified and Gibson assembled into the PmeI site of plasmid pMTLSC7215 to create plasmid pMC822. The plasmid was purified using a miniprep kit (Zymo Research), transformed into E. coli strain HB101 pRK24, and introduced into C. difficile by conjugation. C. difficile harboring the plasmid were selected on BHIS agar containing 15 μg/mL thiamphenicol, streaked onto BHIS agar, and subsequently on BHIS agar with 15 μg/mL thiamphenicol and 100 μg/mL kanamycin to force plasmid integration and counterselect against E. coli. A clone that screened positive for two crossover events was streaked to purity on BHIS agar for three more passages and the loss of plasmid was confirmed via sensitivity to 5 μg/mL thiamphenicol on BHIS agar.
Detailed construction of plasmids can be found in Figure S1. Plasmids were transferred to C. difficile as previously described, with slight variation (45, 46). Briefly, plasmids were chemically transformed into E. coli strain HB101 pRK24 and mated with C. difficile on agar plates for 48 h. Transconjugants were selected on BHIS agar containing 10 μg/mL thiamphenicol for plasmid selection and 100 μg/mL kanamycin to counterselect against E. coli.
Nitrocefin hydrolysis disk assays
β-lactamase activity was assessed by hydrolysis of nitrocefin, a chromogenic cephalosporin (Sigma-Aldrich). Briefly, C. difficile was grown overnight in BHIS to log phase, then diluted to an OD600 of 0.05 in BHIS medium with or without 2 μg/mL ampicillin. Cultures were grown to an OD600 of 0.45-0.55, and 1 mL of culture was collected and centrifuged for 3 minutes at 21,130 rcf. All but approximately 30 μL of the supernatant was decanted, the pellets were resuspended, and the cells were spotted onto a nitrocefin disk. The disks were incubated aerobically or anaerobically for 2 h at 37°C, as noted.
Nitrocefin liquid hydrolysis assays
β-lactamase activity was determined for complemented strains via anaerobic liquid nitrocefin assays, as previously reported, with some modifications (47). Briefly, C. difficile was grown overnight in BHIS with 2 μg/mL thiamphenicol to log phase, then diluted to an OD600 of 0.05 in BHIS medium with 2 μg/mL thiamphenicol and 2 μg/mL ampicillin. Cultures were grown to an OD600 of 0.45 – 0.55, 1 mL of culture was collected (in duplicate), and cells centrifuged for 3 min at 21,130 rcf. For whole cell reactions, supernatant was transferred to a fresh tube, and nitrocefin (BioVision) was added to supernatant or whole cell suspensions at a final concentration of 50 μM. For lysed cell reactions, pelleted cells were frozen at −20°C until use. Pellets were resuspended in lysis buffer (100 mM sodium phosphate + 50 mM sodium bicarbonate, pH 7.4), and DTT (Fisher Scientific) was added to each sample for a final concentration of 0.2 mM. Lysed samples were subjected to six freeze-thaw cycles (2 min in dry ice/Ethanol bath, 3 min at 37°C). 0.2 mL of the lysate was transferred to a fresh tube (designated ‘lysate’). The remaining volumes of samples were pelleted by centrifugation for 30 min at 21,130 rcf at 4°C, and then filtered via 0.22 μM syringe filters (BD Biosciences). 0.2 mL of this solution (designated ‘lysate filtrate’) was transferred to a fresh tube. Equal volumes of lysis buffer were added to each sample. Nitrocefin was added at a final concentration of 50 μM to bring the sample volume to 1 mL and samples were incubated anaerobically at 37°C for up to 7 minutes. Reactions were quenched by adding 100 μL of 1 M NaCl and immediately placed on ice. Samples were centrifuged for 3 min at 21,130 rcf to clear cell debris. The entire assay was performed anaerobically until this point. 300 μL of each supernatant was applied to a 96-well flat-bottom plate, and the OD490 was recorded with a BioTek microplate reader. β-lactamase units were calculated by the following equation: (OD490 * 1000) / (OD600 * time in min * vol of cells in mL), where OD600 is the value at the time of collection and the time is the number of minutes between the addition of nitrocefin and adding 1 M NaCl. Lysate results were normalized to the amount of lysate supernatant used. Time course experiments were run to confirm the linearity of the reaction. Results reported are the mean of at least three independent experiments.
Minimal Inhibitory Concentration determination (MIC)
β-lactam susceptibility of C. difficile was determined as described previously (48). Briefly, active C. difficile cultures were diluted in Mueller Hinton Broth (MHB; BD Difco) to an OD600 of 0.1, which were grown to an OD600 of 0.45, and further diluted 1:10 in MHB. 15 μL of this diluted culture (∼5×105 CFU/mL) was plated in a pre-reduced 96-well round bottom polystyrene plate that contained 135 μL of MHB with appropriate β-lactams in each well. The MIC was determined as the concentration at which there was no visible growth after 24 hours of anaerobic incubation at 37°C.
Alkaline phosphatase activity assays
Alkaline phosphatase activity assays in C. difficile were performed as described previously, with minor modifications to the original published assay (49, 50). Briefly, C. difficile cultures were grown anaerobically at 37 °C overnight in BHIS with thiamphenicol (2 μg/mL) to log phase, then diluted to an OD600 of 0.05 in 10 mL BHIS with thiamphenicol. 1 mL of cells was collected in duplicate when the OD600 reached 0.5. Cells were centrifuged at 21,130 rcf for 3 min and the pellets were stored in −20°C at least overnight. For the assay, cell pellets were thawed and resuspended in 500 μL of cold wash buffer (10 mM Tris pH 8.0 + 10 mM MgSO4) and pelleted for 3 min at 21,130 rcf. Alkaline phosphatase assays were performed as previously described (50) without the addition of chloroform (51). The OD550 (cell debris) and OD420 (pNP cleavage) were measured in a BioTek microplate reader. Values were averaged between the triplicate wells, and then between duplicate technical samples. AP units were calculated as ((OD420 – (1.75* OD550)) * 1000) / (OD600 * time), where OD600 is the value at the time of collection. Results reported are the average between three independent experiments.
Quantitative reverse transcription PCR analysis (qRT-PCR)
Actively growing C. difficile were diluted to an OD600 of 0.02 in 10 – 25 mL BHIS with appropriate antibiotic and grown to log phase. RNA was isolated as described previously (45, 52). Briefly, 3 mL samples were taken at an OD600 of 0.45 – 0.55, mixed with 3 mL ice-cold 1:1 acetone:ethanol, and stored immediately in −80°C. RNA was isolated (Qiagen RNeasy kit), treated for contaminating DNA (Invitrogen TURBO DNA-free kit), and RNA was reverse-transcribed into cDNA (Bioline Tetro cDNA synthesis kit). cDNA samples were used for qPCR (Bioline SensiFAST SYBR and Flourescein kit) in technical triplicates on a Roche Lightcycler 96 as described previously (53). Results are presented as the means and standard errors of the means for three biological replicates. Statistical significance was determined using a one-way ANOVA, followed by Dunnett’s multiple-comparison test (GraphPad Prism v6.0).
RESULTS
C. difficile produces an inducible, anaerobic β-lactamase
C. difficile was recently reported to produce a β-lactamase that can cleave β-lactam antibiotics (39). We further investigated the regulation and potential inducibility of C. difficile β-lactamase activity and examined the environmental conditions required for its function. Two diverse strains of C. difficile, 630Δerm (ribotype 012) and R20291 (ribotype 027), were grown in the presence or absence of cefoxitin, a cephalosporin, and applied to a membrane disk impregnated with nitrocefin, a chromogenic cephalosporin. As shown in Figure 1, both strains of C. difficile grown in the presence of cefoxitin caused a color change from yellow to red, indicating cleavage of nitrocefin. In the absence of cefoxitin, neither strain demonstrated observable nitrocefin cleavage. These results suggested that C. difficile produces a β-lactamase that is inducible by β-lactams and is present in diverse strains. During optimization of these assays, we observed markedly higher β-lactamase activity under anaerobic conditions, suggesting that this activity was impaired by oxygen. Indeed, when the nitrocefin assay was performed in the presence of oxygen, the disk did not change color, indicating a loss of β-lactamase activity. These results demonstrate that C. difficile strains produce an inducible β-lactamase, and that the activity of this enzyme is quenched by oxygen.
CD0458 encodes the putative class D β-lactamase, BlaD
Based on the observed induction of β-lactamase activity, we hypothesized that the expression of one or more putative β-lactamases would be induced upon exposure to β-lactams. To test this, C. difficile strain 630Δerm was grown in the presence of three classes of β-lactams: cefoperazone (a cephalosporin), ampicillin (a penicillin), and imipenem (a carbapenem). Using qRT-PCR, we measured the gene expression for 17 putative β-lactamases identified in the C. difficile genome (8, 54, 55). Figure S2 demonstrates that the expression of only one of these genes, CD0458, was significantly induced upon exposure to each of the three types of β-lactams. This result supports a previously reported hypothesis, as CD0458 was recently identified as a β-lactamase in C. difficile (39). This induction suggested that expression of CD0458 confers the β-lactamase activity that we previously observed. The expression of the homologous gene in C. difficile strain R20291 was also greatly induced by these three β-lactams (CDR20291_0399, 99% identity; Figure S2). CD0458 is analogous to two loci described recently by Toth et al. as cdd1 and cdd2 (39). However, based on the high similarity of the previously described Cdd1 and Cdd2 proteins, the existence of other genes already annotated as cdd, cdd2, and cdd3 in strain 630 (56, 57), and the sequence similarity of the CD0458/CDR20291_0399 proteins to class D β-lactamases, we renamed the locus blaD.
CD0457 encodes a putative membrane protein, BlaX, which is co-transcribed with blaD
Analysis of the region surrounding blaD revealed the presence of another gene, CD0457, which appeared to be part of an operon with blaD. Figure 2A illustrates the putative bla operon, in which CD0457 is located 27 nucleotides upstream of the start codon of CD0458. To determine if expression of CD0457 is similarly induced upon β-lactam exposure, we measured transcription of CD0457 in C. difficile strain 630Δerm upon exposure to cefoperazone, ampicillin, and imipenem. Figure 2B demonstrates that expression of CD0457 is comparably induced upon exposure to all three β-lactams. This co-regulation by β-lactams strongly suggested that CD0457 is co-transcribed with CD0458 and that the CD0457 predicted membrane protein product could play a role in the β-lactam resistance. The expression of the homologous gene in C. difficile strain R20291 was also comparably induced upon exposure to these β-lactams, indicating a similar organization in divergent strains (Figure S3).
To determine if the CD0457 and blaD genes are part of a single cistronic unit, we assessed the linkage of these transcripts by amplifying the region between CD0457 and blaD from cDNA generated after exposure of C. difficile strains 630Δerm and R20291 to ampicillin (Figure S4A). Figure S4B illustrates the results of the PCR from cDNA that generated a product of 1 kb, which matches the genomic DNA product from the same strain. These data demonstrate that the transcription of CD0457 and blaD are linked, indicating that they comprise a monocistronic unit. Since CD0457 and blaD form an operon and the function of CD0457 is unknown, we named the CD0457 gene blaX.
To further define the transcriptional organization of the bla operon, we examined promoter activity within the bla locus. Potential promoter activity was measured for putative promoter regions within the locus using phoZ reporter fusions, which produce alkaline phosphatase (50). As illustrated in Figure 3, regions of 300 nucleotides directly upstream of the start codons of blaX or blaD were fused to phoZ and expressed in C. difficile. The results of these reporter assays indicate that the region 300 nucleotides upstream of blaX, but not the region 300 nucleotides upstream of blaD, is able to promote transcription, resulting in measurable activity. To confirm the absence of a cryptic blaD promoter located within the blaX coding region, the entire region from the translational start of blaX to the start codon of blaD was also examined for possible promoter activity. However, no transcriptional activity was observed from this region (Figure 3). The only segment that produced significant and inducible activity contained the region upstream of the blaX coding sequence, strongly suggesting that soley this region drives blaX and blaD expression.
The bla operon contributes to ampicillin resistance in C. difficile
Notably, 36% of complete C. difficile genomes contain a homolog of blaX. Other sequenced genomes simply contain the same promoter and blaD region without the membrane protein. The membrane protein only shares approximately 23-40% amino acid identity to uncharacterized proteins found in a handful of other bacterial species. Thus, the function of this membrane protein cannot be inferred from other systems. To define the roles of BlaX and BlaD in β-lactam resistance and in β-lactamase activity, we created mutants of the 630Δerm strain with an insertional mutation in the blaX gene (MC905) or complete deletion of the blaX-blaD locus (MC1327). Compared to the parent strain, blaX::erm displayed decreased, but still inducible blaD expression (Figure S5). Although blaX transcription is measurable in the blaX::erm mutant, the product is presumably non-functional because of the insertional mutation. We confirmed that neither the blaX nor the blaD transcript was expressed in the ΔblaXD mutant (Figure S5).
Based on the induction of β-lactamase activity and the induction of the bla operon by β-lactams, we hypothesized that deletion of the operon would reduce C. difficile resistance to β-lactams. As shown in Figure 4, we performed growth curves with the ΔblaXD and blaX::erm strains in cefoperazone, ampicillin, and imipenem to measure the contribution of the bla operon to β-lactam resistance in C. difficile. While the deletion of blaX and blaD did not significantly affect growth in cefoperazone, ΔblaXD and blaX::erm growth was impaired in ampicillin compared to the parent strain. These data suggest that the bla operon contributes to ampicillin resistance in C. difficile. Interestingly, the deletion of blaX and blaD improved growth in imipenem, supporting the finding by Toth et al. that BlaD binds to, but does not hydrolyze imipenem (39).
Antibiotic resistance in clinically relevant bacteria is often characterized by minimum inhibitory concentrations (MIC) of antibiotics. To further define the contribution of blaX and blaD to β-lactam resistance in C. difficile, we measured the MIC of β-lactams in 630Δerm, ΔblaXD, and blaX::erm. Although the parent strain grew better in ampicillin, the MICs for both cefoperazone and ampicillin were similar in all three strains (Table S1), and higher for 630Δerm in imipenem, indicating a modest difference in resistance values.
The bla operon encodes the only functional β-lactamase in C. difficile
Although blaD was the only annotated β-lactamase induced by β-lactams (Figure 1), it was plausible that another β-lactamase existed in C. difficile. To determine if the bla operon encodes the only β-lactamase in C. difficile, we measured the β-lactamase activity of ΔblaXD in a nitrocefin hydrolysis assay. As shown in Figure 5A, no apparent β-lactamase activity was observed for the ΔblaXD strain. In comparison, the blaX::erm strain exhibits a slight change in color to a light pink, indicating that this mutant does not fully abolish production and activity of the β-lactamase, which is in agreement with the decrease in blaD gene expression observed for this strain (Figure S5). These results strongly suggest that blaD encodes the only functional β-lactamase in C. difficile.
The bla operon exhibits high level, dose-dependent expression in β-lactams
The induction of both blaX and blaD by β-lactams suggested that these genes are important for β-lactam resistance in C. difficile. To determine whether these genes could be induced by other cell wall targeting antimicrobials or if the induction is specific to β-lactam exposure, we measured the levels of gene expression for C. difficile strain 630Δerm in various cell wall targeting antibiotics (vancomycin, polymyxin B, and lysozyme) and cationic antimicrobial peptides (nisin and LL-37), as well as a ribosome-targeting antibiotic (kanamycin). Figure S6 shows that expression of blaX and blaD were induced in the presence of kanamycin and polymyxin B. However, these levels of expression are not statistically significant and were less than 3% of the levels seen for expression after β-lactam exposure, suggesting that the high levels of induction of blaX and blaD are specific to β-lactams.
Although the levels of blaX and blaD induction were high in all three β-lactams, expression varied greatly between each β-lactam. These results suggested that the level of induction of the bla operon is dependent upon the type of β-lactam C. difficile is exposed to and could be dose-dependent. To determine if the bla operon exhibits dose-dependent expression in β-lactams, we measured the relative expression of blaX and blaD in the 630Δerm strain in varying concentrations of cefoperazone, ampicillin, and imipenem. Figure S7 shows that the bla operon did indeed exhibit dose-dependent induction by β-lactams and that the response was different for the various classes of β-lactams. In increased concentrations of cefoperazone, induction of the bla operon trended downward, whereas expression trended upward in increased concentrations of ampicillin. Expression of the bla operon was high in all concentrations of imipenem, exhibiting only a modest increase in expression as the concentration of imipenem was increased. Furthermore, the level of induction of the bla operon was high even at concentrations of β-lactams far below the MIC (0.03125x MIC of cefoperazone, 0.125x of ampicillin, and 0.0625x MIC of imipenem). These results suggest that bla expression is controlled in a dose-dependent manner specific to the class of β-lactam administered.
BlaX is not necessary for BlaD activity
Of the 72 genomes retrieved from a blaD BLASTn search of C. difficile, 42 strains encode the upstream putative membrane protein, suggesting that the membrane protein BlaX may be important for β-lactamase activity in some strains, but not in others. To determine the importance of the membrane protein, we first assessed β-lactamase activity in C. difficile strain M120, which lacks a homolog of blaX. The BlaD enzyme from strains M120 and 630Δerm are highly similar, but the 4% variability clearly lies within the N-termini of these proteins (Figure S8A). As shown in Figure S8B, strain M120 does exhibit β-lactamase activity. The variability in the amino acid sequence of these two enzymes may be due to differences in signal sequence recognition, but a potential interaction with another protein cannot be ruled out.
As the function of BlaX was not immediately apparent, we examined whether BlaX is necessary to observe the β-lactamase activity of BlaD in strain 630Δerm. To test this, we complemented the ΔblaXD strain with blaX and/or blaD in trans. The nitrocefin disk assays in Figure 5B demonstrate that expression of blaD alone can restore β-lactamase activity in the ΔblaXD mutant, indicating that BlaD can act independently of BlaX, despite the co-transcription of these two genes. This result is further supported by the observation that the blaX::erm strain exhibits some β-lactamase activity (Figure 5A).
BlaD contains a predicted signal sequence and is associated with the cell membrane
A common characteristic of β-lactamases is an N-terminal signal sequence that directs the protein out of the cytoplasm. We hypothesized that the N-terminus of BlaD encodes a signal sequence based on the signal sequence prediction within the first 18 amino acid residues (58, 59). We generated a truncated version of BlaD missing these first 18 residues (BlaDΔ18; pblaDΔ18). As shown in Figure 5B, the expression of BlaDΔ18 is unable to complement the absence of β-lactamase activity in the ΔblaXD mutant in a whole cell assay. qRT-PCR results shown in Figure S9 confirm that blaX and/or blaD are expressed in the complemented strains, indicating that the absence of gene expression is not the cause of the lack of observable β-lactamase activity. This suggested that BlaDΔ18 is either not translated, is an unstable or inactive protein, or is active but trapped in the cytosol and unable to hydrolyze nitrocefin.
All of the characterized β-lactamases in Gram-positive bacteria are membrane-bound enzymes, although many of these proteins are cleaved, resulting in a smaller, soluble form that can be found in culture supernatants (29, 31, 34, 36). These findings are consistent with the lack of a periplasmic space for β-lactamases accumulation in Gram-positive bacteria. To determine if a soluble form of BlaD is secreted into the culture medium, we performed a nitrocefin hydrolysis assay using culture supernatants. As shown in Figure 6A and 6C, neither the supernatants of ΔblaXD cells harboring pblaD or pblaX-blaD, nor the wild-type strains 630Δerm or M120, react with nitrocefin, indicating that BlaD is not secreted into the medium. To confirm that BlaD is a membrane-associated enzyme, we lysed the cells and performed a nitrocefin hydrolysis assay using lysates containing cell debris (denoted as ‘lysates’) or the cleared cell lysates (denoted as ‘lysate filtrate’). Figures 6B and 6D show that when comparing the level of activity in the lysate to the lysate filtrate in strains containing a full-length blaD, 74-80% of the total β-lactamase activity is found in the cell debris, indicating that BlaD is associated with the cell surface. Furthermore, BlaDΔ18 activity is not associated with the cell surface, as demonstrated by the similar levels of activity in the lysate and the lysate filtrate (Figure 6B). This result indicates that BlaDΔ18 is an active, soluble form of BlaD that is trapped in the cytosol, and strongly suggests that the first 18 residues at the N-terminus of BlaD encode a signal sequence. Together, these results support the presence of a signal sequence that helps bring the protein to the cell surface.
BlaX aids in BlaD activity
Although BlaX is not necessary for BlaD activity (Figure 5A, B), blaX is conserved in many C. difficile strains. Thus, we examined whether BlaX enhances BlaD activity. The results shown in Figure 6A and 6B demonstrate that the presence of BlaX increases β-lactamase activity of the 630Δerm BlaD two to three-fold, suggesting that BlaX plays a role in the function of BlaD. To investigate the activity of a BlaD from a C. difficile genome that lacks BlaX, we also complemented the ΔblaXD strain with blaD cloned from the M120 genome, under the M120 native promoter. Figure 6A shows that in cell suspensions of ΔblaXD complemented strains, the M120 BlaD (pM120blaD) exhibits two-fold higher activity than the 630Δerm BlaD (pblaD). This result suggests that the M120 BlaD is superior to the 630Δerm BlaD at translocating to the cell surface when BlaX is not present. However, M120 BlaD is only two-thirds as active as the 630Δerm BlaXD complement (pblaXD). In lysed cells, the M120 BlaD β-lactamase activity levels are slightly higher than the 630Δerm BlaD (Figure 6B). Interestingly, the wild-type strains 630Δerm and M120 exhibit similar β-lactamase activity levels in both cell suspension and lysate samples, indicating that their overall efficacy is comparable (Figure 6C and D). Together, these results demonstrate that in 630Δerm, BlaX enhances BlaD activity, while in M120, β-lactamase activity is not dependent on BlaX. Finally, because the M120 BlaD does not fully complement the ΔblaXD strain, the N-terminal sequence variability of the BlaD proteins likely plays a role in strain-dependent translocation of BlaD to the cell surface.
The bla operon is regulated by BlaIR
Transcription of most β-lactamase genes in Gram-positive bacteria is regulated by the two-component BlaRI system (60-62). The C. difficile genome encodes several orthologs of the two genes that make up this system, blaI and blaR. In other bacteria, BlaR is a sensor that is activated upon β-lactam binding (63). Activated BlaR cleaves the BlaI repressor, which is bound as a dimer to the bla operon promoter in the absence of β-lactams (64-66). Once cleaved, BlaI can no longer bind to the bla promoter, thus allowing for active transcription. Two candidate orthologs CD0471 (blaI) and CD0470 (blaR) are located 11 kb downstream of the blaXD operon. To determine if these blaIR orthologs regulate the blaXD operon in C. difficile, we created an insertional disruption in blaI. Figure S10 shows that transcription of blaI and blaR are decreased in the blaI::erm mutant, confirming that blaI and blaR are organized in an operon, as is consistent with other bacteria. As shown in Figure 7, in the absence of β-lactams, blaX and blaD are transcribed at high levels in the blaI::erm mutant, as compared to the wild-type 630Δerm strain. These results confirm that BlaI acts as a repressor of the bla operon. Further, the induction of blaXD in β-lactams in the wild-type strain, but not in the mutant, strongly suggests that BlaI repression is relieved by the presence of β-lactams in wild-type strain. To verify that relief of BlaI repression results in β-lactamase production, we performed a nitrocefin hydrolysis assay on the blaI::erm mutant. Figure 5C confirms that the absence of BlaI results in active β-lactamase, independent of β-lactam presence. Together, these results show that C. difficile encodes a BlaRI system that represses bla transcription in the absence of β-lactams. Efforts to complement blaIR resulted in poor growth of E. coli mating strains, as well as C. difficile, and were not successful.
To further confirm that the BlaRI system regulates the bla operon and to define its contribution to ampicillin resistance, we examined the growth of the blaI::erm mutant in multiple β-lactams. Figure 8A illustrates that growth of the blaI mutant is not significantly different than the wild-type 630Δerm strain in the presence of cefoperazone. However, growth of the blaI mutant is significantly improved in the presence of ampicillin, as compared to 630Δerm (Figure 8B). Similarly, the blaI::erm mutant shows slightly impaired growth in imipenem, as compared to 630Δerm (Figure 8C). These results show that BlaIR contributes to ampicillin and impenem resistance in C. difficile through regulation of the bla operon.
DISCUSSION
This study provides evidence for β-lactam-dependent expression of the β-lactamase, BlaD, in two strains of C. difficile, 630Δerm and R20291, as well as activity of BlaD in both 630Δerm and M120. The blaD gene is located in an operon with blaX, which encodes a putative membrane protein (Figure S4). Our data indicate that the promoter for the blaXD operon is located within a 300 nucleotide region located directly upstream of the blaX start codon (Figure 3). The high level of blaD and blaX expression in response to β-lactams far below MICs (Figure S7), indicate that the promoter of the bla operon is quite strong, in contrast to a previous report in which part of the blaD locus was expressed in a heterologous host (39).
Our work has demonstrated that BlaD is a β-lactamase that is only active under anaerobic (reducing) conditions (Figure 1). To our knowledge, no other anaerobic β-lactamases have been reported, which is not surprising given that β-lactamase assays are generally performed in the presence of oxygen (67, 68). This, however, may be one reason that so few β-lactamases have been identified in anaerobic, Gram-positive bacteria (69-72). Indeed, the addition of 0.2 mM DTT to the nitrocefin hydrolysis assays, or steady-state enzyme kinetics (39), allowed for observation of BlaD activity (Figure 6) by maintaining reducing conditions. Assaying β-lactamases from other anaerobic, Gram-positive bacteria under reducing conditions may lead to the identification of more anaerobic β-lactamases in other species.
Our data indicate that BlaD acts at the cell membrane, in accordance with other β-lactamases from Gram-positive bacteria (Figure 6). We have shown that BlaD likely contains a signal sequence at the N-terminus, which facilitates translocation of BlaD to the membrane. BlaD is not secreted into the environment, but remains associated with the cell surface (Figure 6). While the exact function of BlaX is unknown, the data demonstrate that BlaD activity is enhanced by the presence of BlaX (Figure 6B). BlaX has five predicted transmembrane domains, with an approximate 125 residue-long extracellular loop (73). Because the activity of BlaD is membrane-associated across all samples except BlaDΔ18, and BlaD activity in cell lysates lacking BlaX is 60% less than when BlaX is present, it is possible that BlaX interacts with BlaD in a way that makes BlaD more accessible to substrates on the cell surface. Nitrocefin hydrolysis assays showed that in cell lysates, the activity of full length BlaD (pblaD) is 45% less than BlaDΔ18 (Figure 6B). This suggests that either BlaD is cleaved at the N-terminus after translocation to the cell membrane, or BlaX helps to relieve a steric hindrance caused by insertion into the cell membrane. The absence of β-lactamase activity in cell supernatants does not support cleavage of BlaD, unless BlaD remains anchored to the cell membrane after cleavage, which is unlikely due to the absence of a canonical lipobox immediately downstream of the signal peptide (74).
To date, only one other published β-lactamase is reported to be co-transcribed with a membrane protein (75). This membrane-bound β-lactamase, PenA, found in the Gram-negative Burkholderia psuedomallei, is encoded in an operon with nlpD1, a gene annotated as an outer membrane lipoprotein and thought to be involved in cell wall hydrolytic amidase activation (76). However, C. difficile does not contain an outer membrane, and nlpD1 is not homologous with blaX. Analysis of the blaD locus in the C. difficile strain M120, which does not contain a full blaX coding sequence, revealed regions of partial homology to the 5’ and 3’ ends of blaX, located between the promoter and the blaD start codon. This suggests that over the course of evolution of C. difficile, the majority of this gene was deleted. A search of the rest of the M120 genome revealed no other proteins similar to BlaX, further supporting the model that in many C. difficile strains, BlaX is not necessary for sufficient BlaD activity. However, the superior activity levels of M120 BlaD (Figures 6A and 6B), the 74% of cell surface-associated activity of M120 BlaD (Figure 6B), as well as the equal levels of β-lactamase activity of the 630Δerm strain compared to M120 (Figure 6D), suggest that M120 likely has a different mechanism of translocation.
We have shown that the bla operon confers resistance to ampicillin and is regulated by the BlaRI system in C. difficile (Figures 5, 8). Disruption of blaI resulted in constitutive expression of blaX and blaD (Figure 7), which resulted in improved growth in ampicillin (Figure 8), supporting the model that BlaI is a direct repressor of the bla operon. We identified a 52-nucleotide region of dyad symmetry in the promoter of the bla operon, which contains a canonical BlaI binding site, supporting the model of BlaI-PblaX binding, but does not rule out other binding partners. Our results align with previously reported data that BlaD confers resistance to penicillins (39). The discrepancy of the MIC values versus the growth curves can be attributed to the exact nature of a growth curve. Further investigation is needed to fully define the mechanisms of β-lactam resistance in C. difficile. Identification and characterization of the additional β-lactam resistance mechanisms may aid in preventing C. difficile infections and recurrence in the future.
ACKNOWLEDGEMENTS
We thank members of the McBride lab and the dissertation committee of B.K.S. for helpful suggestions and discussions throughout the course of this work. This research was supported by the U.S. National Institutes of Health (NIH) through research grants AI116933 and AI121684 to S.M.M., and training grant AI106699 to S.A. The content of this manuscript is solely the responsibility of the authors and does not necessarily reflect the official views of the National Institutes of Health.