Structural Basis of a Thiol-Disulfide Oxidoreductase in the Hedgehog-Forming Actinobacterium Corynebacterium matruchotii

ABSTRACT

The actinobacterium Corynebacterium matruchotii has been implicated in nucleation of oral microbial consortia leading to biofilm formation. Due to the lack of genetic tools, little is known about basic cellular processes, including protein secretion and folding, in this organism. We report here a survey of the C. matruchotii genome, which encodes a large number of exported proteins containing paired cysteine residues, and identified an oxidoreductase that is highly homologous to the Corynebacterium diphtheriae thiol-disulfide oxidoreductase MdbA (MdbA_Cd). Crystallization studies uncovered that the 1.2-Å resolution structure of C. matruchotii MdbA (MdbA_Cm) possesses two conserved features found in actinobacterial MdbA enzymes, a thioredoxin-like fold and an extended α-helical domain. By reconstituting the disulfide bond-forming machine in vitro, we demonstrated that MdbA_Cm catalyzes disulfide bond formation within the actinobacterial pilin FimA. A new gene deletion method supported that mdbA is essential in C. matruchotii. Remarkably, heterologous expression of MdbA_Cm in the C. diphtheriae ΔmdbA mutant rescued its known defects in cell growth and morphology, toxin production, and pilus assembly, and this thiol-disulfide oxidoreductase activity required the catalytic motif CXXC. Altogether, the results suggest that MdbA_Cm is a major thiol-disulfide oxidoreductase, which likely mediates posttranslocational protein folding in C. matruchotii by a mechanism that is conserved in Actinobacteria.

IMPORTANCE The actinobacterium Corynebacterium matruchotii has been implicated in the development of oral biofilms or dental plaque; however, little is known about the basic cellular processes in this organism. We report here a high-resolution structure of a C. matruchotii oxidoreductase that is highly homologous to the Corynebacterium diphtheriae thiol-disulfide oxidoreductase MdbA. By biochemical analysis, we demonstrated that C. matruchotii MdbA catalyzes disulfide bond formation in vitro. Furthermore, a new gene deletion method revealed that deletion of mdbA is lethal in C. matruchotii. Remarkably, C. matruchotii MdbA can replace C. diphtheriae MdbA to maintain normal cell growth and morphology, toxin production, and pilus assembly. Overall, our studies support the hypothesis that C. matruchotii utilizes MdbA as a major oxidoreductase to catalyze oxidative protein folding.

INTRODUCTION

Previously classified as Bacterionema matruchotii (1), the Gram-positive actinobacterium Corynebacterium matruchotii is a member of the genus Corynebacterium that is microscopically seen as a filament joined to a bacillus-like body (2, 3). Early studies of dental plaque revealed that C. matruchotii is part of morphological units termed corncobs, comprised of filamentous organisms enclosed by adherent cocci (4, 5). Consistently, it has recently been shown in supragingival dental plaque that C. matruchotii is a predominant corynebacterial species, forming hedgehog-like structures—with corynebacterial filaments efferently arranged and streptococci at the boundary (6)—that may nucleate microbial consortia leading to biofilm formation (6). Early investigations of C. matruchotii were focused on its classification (7) and its ability to calcify (8, 9) and coaggregate with other oral bacteria (4, 5). Perhaps due to the lack of allelic exchange systems for genetic manipulation and biochemical characterization, although transposon mutagenesis has been developed for C. matruchotii (10), little is known about different cellular processes, including protein secretion and folding, in this organism.

Gram-negative and Gram-positive bacteria secrete a large number of unfolded proteins across the cytoplasmic membrane via the Sec apparatus (11, 12). In the Gram-positive actinobacteria, oxidative protein folding via disulfide bond formation appears to be the major pathway for posttranslocational folding of these unfolded proteins (13). Recent studies in the two actinobacteria Actinomyces oris and Corynebacterium diphtheriae revealed that a membrane-bound thiol-disulfide oxidoreductase named MdbA functions similarly to DsbA (14, 15), the archetypal thiol-disulfide oxidoreductase of Escherichia coli, which catalyzes disulfide bond formation of unfolded proteins translocated into the bacterial periplasm (16). In the two actinobacteria, it is proposed that MdbA catalyzes oxidative folding of unfolded proteins emerging from the Sec translocon (13). In E. coli, a membrane-anchored oxidoreductase named DsbB reoxidizes DsbA (17). In A. oris, reoxidation of MdbA requires a vitamin K epoxide reductase (VKOR)-like protein (14, 18), whereas in C. diphtheriae an equivalent reductase enzyme has yet to be identified. Similar to E. coli, A. oris utilizes the electron transport chain for reactivation of the MdbA/VKOR system, although menaquinone appears to be a key electron acceptor (19). Furthermore, DsbA-like proteins have also been identified in the Gram-positive Firmicutes, including Bacillus subtilis (20, 21), Staphylococcus aureus (22), and Streptococcus gordonii (23). However, while these firmicute oxidoreductases are involved in protein stability, oxidative stress resistance, bacteriocin production, and virulence, disulfide bond formation is not a major pathway for posttranslocational protein folding in these organisms (24).

Intriguingly, unlike the deletion of dsbA in E. coli (16), deletion of mdbA is lethal in A. oris; lethality is also observed in C. diphtheriae cells grown at elevated temperatures (14, 15). This supports the central role of oxidative protein folding in various cellular processes in Actinobacteria, which predictably encode more than 60% exported proteins with multiple cysteine residues (25). Consistent with these findings, a gene encoding a DsbA-like protein in the actinobacterium Mycobacterium tuberculosis (Mtb-DsbA) is essential for optimal growth (26, 27). Interestingly, a VKOR homolog has been identified in M. tuberculosis to be a DsbB functional homolog, which is capable of reoxidizing E. coli DsbA (28). Thus, it is likely that Actinobacteria may share a disulfide bond-forming pathway that mediates posttranslocational folding of exported proteins. However, it remains to be seen if this mechanism is conserved in other Actinobacteria such as C. matruchotii.

Here we report a gene deletion method for C. matruchotii and structural and biochemical characterizations of an MdbA-like protein in this organism. We showed by X-ray crystallography that C. matruchotii MdbA is structurally homologous to the C. diphtheriae MdbA. By reconstituting disulfide bond formation in vitro, we demonstrate that C. matruchotii MdbA exhibits disulfide bond-forming activities similar to those of the C. diphtheriae MdbA enzyme. Consistent with its role as a major disulfide bond-forming enzyme in C. matruchotii, deletion of mdbA is lethal to cells. Importantly, heterologous expression of C. matruchotii MdbA in the C. diphtheriae mdbA mutant rescues the defects in cell growth and morphology, pilus assembly, and toxin production of this mutant. Thus, this work not only supports the hypothesis that C. matruchotii MdbA is a major thiol-disulfide oxidoreductase in this organism but also provides genetic tools and an experimental system to study oxidative protein folding and other cellular processes in C. matruchotii.

RESULTS

Sensitivity of C. matruchotii to reducing conditions.

We analyzed the genome of C. matruchotii strain ATCC 14266 and found that proteins possessing 2 or more cysteine (Cys) residues made up 58.4% the C. matruchotii proteome (1,530 of 2,619 proteins) (https://biocyc.org); nearly half of those Cys-containing proteins harbored a signal peptide. This analysis suggests that disulfide bond formation may be an important trait of this anaerobe. To examine this possibility, bacterial cells were grown in liquid broth with dithiothreitol (DTT) to a final concentration of 0, 1, 5, or 10 mM (see Materials and Methods). Cell growth was monitored by optical density at 600 nm (OD₆₀₀), and the doubling time was determined. As shown in Fig. 1A, at DTT concentrations of 5 and 10 mM C. matruchotii doubling time increased more than 2.5- and 3-fold (from 196.85 ± 14.60 min to 558.78 ± 72.19 min and 883.44 ± 61.77 min), respectively. Consistent with the notion that disulfide bond formation is the major folding pathway in Actinobacteria, DTT treatment of C. diphtheriae also caused a 3-fold increase in doubling time at 5 mM (from 88.00 ± 21.50 min to 306.37 ± 48.59 min), and at 10 mM C. diphtheriae ceased to grow (Fig. 1B). On the other hand, no growth defects were observed with the firmicute Staphylococcus aureus in the presence of DTT (Fig. 1C), whereas treatment of E. coli cells with 10 mM DTT caused a cell growth defect as previously reported (25).

FIG 1.

Corynebacterial sensitivity to reducing agent DTT. (A to D) The doubling times of C. matruchotii, C. diphtheriae, S. aureus, and E. coli cells grown in liquid broth in the presence of 0, 1, 5, and 10 mM reducing reagent DTT were measured. The results are presented as averages from three independent experiments; *, P < 0.01; **, P < 0.001. (E to H) Aliquots of samples in the presence of 0 and 5 mM DTT were analyzed by electron microscopy, in which harvested cells were immobilized onto nickel grids, washed, and stained with 1% uranyl acetate prior to imaging by an electron microscope. Bars, 0.5 μm.

To determine if treatment of corynebacterial cells with DTT, which causes cell arrest as seen above, affects cell morphology, corynebacterial cells in the presence or absence of DTT were harvested for electron microscopy, whereby cells were immobilized on nickel grids, washed, and stained with 1% uranyl acetate prior to viewing with an electron microscope. Compared with untreated C. matruchotii cells, which displayed a slender, elongated shape (Fig. 1E), C. matruchotii cells in the presence of 5 mM DTT were shorter and heavily stained (Fig. 1F). Similarly, the phenotypes of C. diphtheriae cells in the presence or absence of DTT mirrored those of C. matruchotii samples (Fig. 1G and H).

Structural analysis of a thiol-disulfide oxidoreductase in C. matruchotii.

The results above suggest that C. matruchotii possesses a thiol-disulfide oxidoreductase that catalyzes disulfide bond formation critical for oxidative protein folding. Using the C. diphtheriae thiol-disulfide oxidoreductase MdbA (MdbA_Cd) sequence as the query, homology and BLAST searches revealed in the genome of C. matruchotii the hypothetical protein HMPREF0299_7193 as an MdbA homolog with 49% identity (E value of 1 × 10⁻⁶³), here called MdbA_Cm. Predicted to be a 26.8-kDa transmembrane protein, MdbA_Cm harbors a CHYC motif, in which the His residue appears to be characteristic of thiol-disulfide oxidoreductase enzymes in Actinobacteria (13–15).

To define the function of MdbA_Cm, a recombinant protein (residues 44 to 242) lacking a transmembrane domain was cloned in and purified from E. coli for crystallization. The X-ray crystal structure of C. matruchotii MdbA was determined by molecular replacement using the C. diphtheriae MdbA structure (PDB 5C00) (15) as the starting model. The MdbA_cm structure was refined to 1.2-Å resolution with R_work and R_free factors equal to 12.2 and 15.5%, respectively (Tables 1 and 2). The structure represents a reduced form of the protein with the catalytic residue C91 converted to S-hydroxy-l-cysteine (CSO91). The overall structure (Fig. 2A) harbors a typical DsbA/MdbA protein family fold (13, 29–31), which incorporates a thioredoxin-like domain and an α-helical domain. The thioredoxin-like domain (residues 60 to 126 and 196 to 242) consists of a 6-strand β-sheet in the order of β1↓-β2↑-β4↓-β3↓-β5↑-β6↓ and 4 flanking helices (two 3₁₀-helices and two α-helices, α3 and α9) (Fig. 2A; see also Fig. S1A in the supplemental material). The conserved catalytic CHYC motif (residues 91 to 94) forms the active site together with a conserved cis-Pro loop (residues S221 and P222) (Fig. 2C and D). The MdbA α-helical domain comprises 7 α-helices (Fig. 2A and S1A).

TABLE 1.

Crystal data collection statistics

Characteristic	Value(s)a
X-ray wavelength (Å)	0.9792
Space group	C 2221
Unit cell dimensions
a, b, c (Å)	35.6, 82.6, 123.5
α = β = γ (°)	90
Resolution (Å)	39.2–1.2 (1.22–1.20)
No. of unique reflections	55,421 (1,871)
Completeness	96.2% (65.6%)
Rmerge	0.054 (0.34)
CC1/2b (Å2)	—c (0.83)
I/σ	30.7 (2.01)
Redundancy	5.0 (2.2)
No. of molecules per asymmetric unit	1
No. of protein residues	200

^aNumbers in parentheses are for the highest-resolution shell.

^bCC1/2, Pearson's correlation coefficient for one-half of the data set.

^c—, not determined.

TABLE 2.

Structure refinement statistics

Characteristic	Value(s)
Resolution range (Å)	39.2–1.20 (1.23–1.199)
No. of reflections	55,381 (2,891)
σ cutoff	None
R (all) (%)	12.36
R (Rwork) (%)	12.21 (25.4)
Rfree (%)	15.45 (27.8)
RMS deviations from ideal geometry
Bond length (Å)	0.011
Angle (degrees)	1.46
Chiral (Å)	0.082
No. of atoms
Protein	1,690
CAPSO	13
Water	306
Mean B factor (Å2)
All atoms	20.5
Protein atoms	18.3
Protein main chain	15.5
Protein side chain	20.9
CAPSOa	32.1
Water	32.7
MolProbity Ramachandran plot statistics
Residues in favored regions (%)	97.3
Residues in disallowed region (%)	0.0

^aCAPSO, 3-(cyclohexylamino)-2-hydroxy-1-propanesulfonic acid.

FIG 2.

Crystal structure of C. matruchotii MdbA. (A) The C. matruchotii MdbA structure was determined to 1.2-Å resolution and is presented with rainbow colors from blue (Asn44) to red (Lys242). (B) Structure alignment of C. matruchotii MdbA (green) and C. diphtheriae MdbA (gray) (PDB 5C00) was performed using secondary structure matching (SSM) superposition in the COOT program (44). (C) Superposition of the active centers of C. matruchotii MdbA (rainbow colors) and C. diphtheriae MdbA (gray). (D) Electrostatic potential surface for C. matruchotii MdbA generated by Pymol with CSO91 replaced by cysteine; areas of positive charge are shown in blue and those of negative charge in red. The MdbA peptide-binding groove is marked by an oval dotted line.

The protein active site closely resembles active sites of other MdbA/DsbA enzymes. The superposition of C. matruchotii and C. diphtheriae MdbA active sites does not show notable changes of active-site arrangement (Fig. 2C). The CSO91 residue in the MdbA_Cm structure was modeled as sulfenic acid, which was detected in alternate conformations, based on electron density maps showing additional densities next to the sulfhydryl group of the cysteine side chain. Modeling of the residue as either sulfinic or sulfonic acid resulted in increased R factors in refinement as well as the occurrence of negative peaks in electron density F_o − F_c difference maps (where F_o is the structure factor observed and F_c is the structure factor calculated). Although we cannot exclude the possibility of the presence of a mixture of reduced and oxidized forms of cysteine in this position, the observed oxidation of residue C91 is most likely an artifact of crystallization. Therefore, the oxidation of C91 probably does not reflect protein activity regulation. The other cysteine residue, C94, showed no sign of oxidation most likely because its position within the protein body shields it from solvent.

The MdbA_Cm structure is very closely related to the C. diphtheriae MdbA structure (15) (47.6% amino acid sequence identity for protein constructs used for crystallization). According to DALI server analysis (32), alignment of those structures has a Z score and a root mean square deviation (RMSD) of superimposed atoms (in angstroms) value of 33.6 and 1.0, respectively, for 197 equivalent residues (Fig. 2B). The major difference between C. matruchotii and C. diphtheriae MdbAs is observed in the secondary structures of their N termini. The MdbA_Cm displays an α-helix (residues 45 to 54) that is broken at K50 due to an interaction with a series of hydrophobic residues (177 to 181) in α-helix 7. The equivalent region in MdbA_Cd has an unstructured or semihelical coil in this region, with 4 chains in the asymmetric unit of the PDB entry (15). The next closest structural homologs to MdbA_Cm are considerably more distant: Mycobacterium tuberculosis DsbA (PDB 4K6X) (33) and Actinomyces oris MdbA (PDB 4Z7X) (14) have Z scores of 19.5 and 18.2 and RMSD values of 2.2 and 2.5, respectively (Fig. S1B and C). Altogether, the results support the hypothesis that C. matruchotii encodes a thiol-disulfide oxidoreductase that is structurally homologous to the other actinobacterial MdbA enzymes.

The C. matruchotii MdbA catalyzes disulfide bond formation in vitro.

To demonstrate that MdbA_Cm directly catalyzes disulfide bond formation in proteins, we employed a previously reported disulfide bond-forming assay using the pilus protein FimA of A. oris as a substrate (14, 15). In this assay, recombinant proteins FimA, MdbA_Cd, and MdbA_Cm were expressed in and purified from E. coli (Fig. 3A), and the first (FimA) was reduced with DTT. Reduced FimA was incubated with MdbA_Cd or MdbA_Cm at a 1:8.75 molar ratio in redox buffer. At 5-min intervals, aliquots were removed and treated with methoxypolyethylene glycol-maleimide (Mal-PEG) to quench the reactions. Protein samples were then analyzed by SDS-PAGE with Coomassie blue staining (see Materials and Methods). As shown in Fig. 3B, reduced FimA was completely oxidized within 30 min by MdbA_Cm (compare lane 1 to lane 4), a reactivity that was similar to that of the MdbA_Cd enzyme (compare lane 5 to lane 8) as previously reported (14). This finding further supports the hypothesis that MdbA_Cm is a thiol-disulfide oxidoreductase.

FIG 3.

C. matruchotii MdbA catalyzes disulfide bond formation in A. oris pilin FimA. (A) Recombinant proteins FimA, MdbA_Cd, and MdbA_Cm were expressed in and purified from E. coli. Purified proteins were analyzed by SDS-PAGE and Coomassie blue staining. (B) Reduced recombinant FimA was left untreated (with buffer only) or treated with recombinant MdbA_Cd or MdbA_Cm proteins. At the indicated times, the reactions were quenched by Mal-PEG; protein samples were analyzed by SDS-PAGE and Coomassie blue staining. Oxidized and reduced forms of FimA are indicated by arrows.

Genetic and biochemical analyses of C. matruchotii MdbA.

To examine the role of MdbA_cm in cell physiology, we sought to generate a nonpolar, in-frame deletion mutant of C. matruchotii mdbA. Because an allelic exchange system for C. matruchotii was not available at the time of this investigation, we adapted a gene replacement method from C. diphtheriae for C. matruchotii, in which the conjugative vector pK19mobsacB, a nonreplicating plasmid in corynebacteria, was employed to deliver deletion constructs to corynebacterial cells (34). This vector also harbors sacB for counterselection (35). To generate an mdbA deletion mutant in C. matruchotii, a deletion cassette carrying 1-kb fragments flanking mdbA_Cm was cloned in pK19mobsacB, and the resulting plasmid was introduced into E. coli S17-1; the recombinant plasmid was then introduced into C. matruchotii ATCC 14266 by bacterial conjugation as previously reported (35). Integration of the plasmid into the C. matruchotii chromosome was selected by kanamycin, an antibiotic marker of pK19mobsacB in corynebacteria (34) (see Fig. S2A in the supplemental material; cointegrant). Of note, the cointegrant cells exhibited a cell morphological defect, short cell length, compared to the parental strain (Fig. S2B). Excision of the plasmid by homologous recombination was selected by bacterial growth on agar plates supplemented with sucrose, which kills cells harboring sacB on the integrative plasmid, leading to generation of C. matruchotii strains with wild-type or mutant mdbA alleles. Selection of mutant alleles was performed by PCR as previous described (35). Over 200 independent clones were screened; however, all were wild-type alleles, suggesting that mdbA is an essential gene in C. matruchotii, a genotype characteristic that has been reported for A. oris mdbA (14).

Since neither a conditional gene deletion method nor a complementing vector is currently available for this organism, we asked if C. diphtheriae could be used as a surrogate system for testing MdbA_Cm functionality in vivo. Recombinant plasmids expressing MdbA_Cm or its catalytically inactive variant (C91A/C94A) under the control of the C. diphtheriae mdbA_Cd promoter were cloned and electroporated into the C. diphtheriae ΔmdbA mutant (15). Expression of MdbA proteins was examined by Western blot assays, whereby cells of C. matruchotii ATCC 14266, C. diphtheriae NCTC 13129, or its isogenic ΔmdbA mutant harboring the aforementioned plasmids were subjected to cell fractionation, and protein samples from membrane fractions were immunoblotted with polyclonal antibodies against MdbA_Cm (anti-MdbA_Cm) or MdbA_Cd (anti-MdbA_Cd) (15). With anti-MdbA_Cm, a single band migrating above the 25-kDa marker was observed in C. matruchotii ATCC 14266 but was absent in the C. diphtheriae ΔmdbA mutant (Fig. 4A, lanes 1 and 3). A similar band was detected in this mutant expressing MdbA_Cm or MdbA_Cm-C91A/C94A, although an additional band was also seen (Fig. 4A, lanes 5 and 6, asterisks). The nature of this band is not clear. Interestingly, anti-MdbA_Cm also recognized C. diphtheriae MdbA in the wild-type and complementing strains, albeit with less intensity (Fig. 4A, lanes 2 and 4). Of note, the predicted molecular mass of MdbA_Cd is 27 kDa (15).

FIG 4.

Heterologous expression of C. matruchotii MdbA rescues the cell growth defect of the C. diphtheriae mdbA mutant. (A) Membrane fractions of C. matruchotii ATCC 14266 (Cm), C. diphtheriae NTCC 13129 (Cd), its isogenic mdbA mutant (ΔmdbA_Cd), or this mutant expressing C. diphtheriae MdbA (MdbA_Cd), C. matruchotii MdbA (MdbA_Cm), or C. matruchotii MdbA with C91A/C94A mutations were analyzed by immunoblotting with polyclonal antibodies against C. matruchotii (anti-MdbA_Cm) or C. diphtheriae MdbA (anti-MdbA_Cd) proteins. Molecular mass markers are shown. (B) Log-phase cells of C. diphtheriae strain grown at 30°C were used to inoculate duplicate cultures, one grown at 30°C and the other at 37°C for 3 h prior to imaging.

The same set of above-described samples was immunoblotted with anti-MdbA_Cd, which recognized both MdbA_Cm and MdbA_Cd, as was evident by the detection of equivalent bands in the C. matruchotii ATCC 14266 and C. diphtheriae NCTC 13129 strains (Fig. 4A, lanes 1 and 2), which were absent in the C. diphtheriae ΔmdbA mutant (Fig. 4A, lane 3). Expression of MdbA_Cd, MdbA_Cm, or C91A/C94A in this mutant resulted in detection of these proteins (Fig. 4A, lanes 4 to 6), although additional bands were similarly observed in MdbA_Cm and MdbA_Cm-C91A/C94A (Fig. 4A, asterisks). These results support the hypothesis that MdbA_Cd and MdbA_Cm are highly homologous, if not interchangeable.

Heterologous expression of the C. matruchotii MdbA rescues multiphenotypic defects of the C. diphtheriae ΔmdbA mutant.

Since MdbA_Cm is expressed in C. diphtheriae (Fig. 4A), we next examined if MdbA_Cm expression rescues the reported growth defect of the C. diphtheriae ΔmdbA mutant. Overnight cultures of corynebacterial strains grown at 30°C were used to inoculate duplicate fresh cultures, one maintained at 30°C and the other at 37°C. Growth was visually observed by turbidity. The C. diphtheriae wild-type strain grew normally at both temperatures, but the C. diphtheriae ΔmdbA mutant was able to grow only at 30°C (Fig. 4B, compare tubes 1 to tubes 2, top and bottom panels) as previously reported (15). This defect was rescued by expression of MdbA_Cd (Fig. 4B, tubes 3). Remarkably, expression of MdbA_Cm in this mutant also rescued the growth defect (Fig. 4B, compare tube 2 to tube 4, bottom panel), in contrast to expression of the catalytically inactive MdbA_Cm enzyme with C91A/C94A mutations (Fig. 4B, tubes 5).

Aliquots of the above cultures were also taken for electron microscopy, whereby corynebacterial cells were immobilized on carbon-coated nickel grids, washed, and stained with 1% uranyl acetate prior to viewing with an electron microscope. As previously reported (15), the C. diphtheriae ΔmdbA mutant exhibited a morphological defect at 37°C, with cells growing into chains, clumped and coccoid, unlike the wild-type (WT) cells (Fig. 5, compare panels F and G). Expression of MdbA_Cd or MdbA_Cm in this mutant restored the normal cell morphology (Fig. 5H and I). However, the C. diphtheriae ΔmdbA mutant expressing the MdbA_Cm-C91A/C94A protein remained defective at 37°C, although these mutant cells grown at 30°C were indistinguishable from the WT cells (Fig. 5J).

FIG 5.

Electron microscopy of corynebacterial cells. Aliquots of cell cultures grown at 30°C (A to E) and 37°C (F to J) (Fig. 3B) were analyzed by electron microscopy as described for Fig. 1; bars, 0.5 μm.

We next examined the corynebacterial ability to produce diphtheria toxin (DT) by collecting supernatants of C. diphtheriae cultures grown in the presence of the iron chelator EDDHA [ethylenediamine-N,N′-bis(2-hydroxyphenylacetic acid)], which induces toxin production (15), for immunoblotting with monoclonal antibodies against DT. As previously reported (15), DT was significantly degraded in the ΔmdbA mutant compared to the WT and complementing strains (Fig. 6A, lanes 1 to 3). Consistent with the above-described results, expression of MdbA_Cm in this mutant resulted in DT stability, in contrast to expression of the catalytically inactive enzyme (Fig. 6A, lanes 4 and 5).

FIG 6.

Heterologous expression of C. matruchotii MdbA rescues the defects in toxin production and pilus assembly of the C. diphtheriae mdbA mutant. (A) Cells of indicated C. diphtheriae strains were treated with 1 mM ethylenediamine-N,N′-bis(2-hydroxyphenylacetic acid) (EDDHA) iron chelator to induce production of diphtheria toxin (DT). Protein samples from the culture medium were immunoblotted with monoclonal antibodies against DT. The arrowhead indicates the intact toxin (A+B). Molecular mass markers are shown in kilodaltons. (B) Mid-log-phase cultures of indicated C. diphtheriae strains were subjected to cell fractionation. Protein samples from the culture medium (S) and cell wall (W) fractions were immunoblotted with polyclonal antibodies against the C. diphtheriae pilus shaft SpaA. SpaA polymers (P), monomers (M), and molecular mass markers are indicated.

Finally, we examined pilus assembly by subjecting the strains described above to cell fractionation, and protein samples isolated from culture supernatant (S) and cell wall (W) fractions were immunoblotted with antibodies against the pilus shaft SpaA (15, 36, 37). In the WT strain, the high-molecular-mass pilus polymers (P) were detected mainly in the cell wall fractions, whereas the ΔmdbA mutant failed to produce any (Fig. 6B, lanes 1 to 4). Complementing this mutant with the plasmids expressing MdbA_Cd or MdbA_Cm rescued the pilus assembly defect, in contrast to expression of the inactive C91A/C94A mutant, which resulted in secretion of low-molecular-mass products, the phenotype that was similar to the ΔmdbA mutant (Fig. 6B, lanes 5 to 10). Of note, MdbA_Cm appeared to function better than MdbA_Cd in the rescued strain (Fig. 6A and B); this could be due to the different expression levels of MdbA_Cm in C. diphtheriae and in MdbA_Cd.

Collectively, the results indicate that MdbA_Cm is a functional homolog of the C. diphtheriae thiol-disulfide oxidoreductase MdbA_Cd.

DISCUSSION

The sensitivity of C. matruchotii to reducing reagent DTT (Fig. 1) and the large number of proteins possessing even-numbered cysteine residues encoded by this actinobacterium are consistent with the capability of disulfide bond formation. While disulfide bonds can be formed spontaneously in vitro, eukaryotic and prokaryotic cells rely on oxidoreductase enzymes such as the mammalian protein disulfide isomerases (PDIs) (38) and the Gram-negative bacterium E. coli DsbA (16) to form these linkages. Therefore, it is logical to suggest that C. matruchotii also encodes a disulfide bond-forming enzyme. Indeed, in this report we have provided structural and biochemical evidence that the C. matruchotii hypothetical protein HMPREF0299_7193 is a thiol-disulfide oxidoreductase, which we have renamed MdbA.

The 1.2-Å crystal structure of C. matruchotii MdbA is very similar to the C. diphtheriae MdbA structure, evident by its high degrees of structural homology (Fig. 2) and RMSD of 1.0. The MdbA_Cm structure is also closely related to M. tuberculosis DsbA and A. oris MdbA structures (see Fig. S1 in the supplemental material). Like C. diphtheriae and A. oris MdbA enzymes, MdbA_Cm catalyzes disulfide bond formation in the common pilin substrate FimA in vitro (Fig. 3), suggesting that the C. matruchotii MdbA enzyme might function in a manner similar to those of the two actinobacterial counterparts in vivo. Certainly, when expressed in the C. diphtheriae ΔmdbA mutant (Fig. 4), MdbA_Cm was able to restore not only the normal cell growth and morphology at nonpermissive temperature (Fig. 5) but also toxin production and pilus assembly at the wild-type levels (Fig. 6). This heterologous complementation strongly supports the hypothesis that both MdbA_Cm and MdbA_Cd enzymes are structurally and functionally equivalent, given that the antibodies against each protein recognize one another (Fig. 4). Collectively, the results also support the hypothesis that the mechanism of oxidative protein folding mediated by MdbA is conserved in Actinobacteria.

The fact that deletion of mdbA is lethal in C. matruchotii, a phenotype characteristically observed in actinobacterial MdbA proteins studied to date (14, 15), suggests that MdbA_Cm is the major thiol-disulfide oxidoreductase in this microbe. It has been speculated that mdbA essentiality might be linked to some substrates of MdbA that are critically required for the biogenesis of peptidoglycan, which is the essential constituent of the bacterial cell envelope (13). Indeed, a comprehensive survey of the C. matruchotii genome reveals that many cell wall transpeptidases, i.e., penicillin-binding proteins, which are important components of the cell wall biosynthesis machine, contain multiple cysteine residues that may be subject to oxidation (Table 3). Future experiments might be focused on investigating if these penicillin-binding proteins require MdbA for posttranslocational protein folding in C. matruchotii. Thus, the presence of a genetic toolkit for this organism, as reported in this study, is highly important. On the other hand, the fact that the C. diphtheriae disulfide bond-forming machine MdbA is not essential at permissive temperature provides an excellent experimental system for these mechanistic investigations.

TABLE 3.

Cysteine-containing cell division proteins in C. matruchotii

Locus tag predicted to contain a signal peptide	Predicted function	No. of cysteine residues
HMPREF0299_6823	FtsW	2
HMPREF0299_6738	FtsY	2
HMPREF0299_7233	Penicillin-binding protein; transglycosylase	8
HMPREF0299_6684	Penicillin-binding protein	3
HMPREF0299_5161	Penicillin-binding protein	2
HMPREF0299_5029	Penicillin-binding protein; transglycosylase	2

MATERIALS AND METHODS

Bacterial strains and media.

Bacterial strains and plasmids used in this study are listed in Table S1 in the supplemental material. C. matruchotii was grown in brain heart infusion (BHI) broth (Becton Dickinson) or on BHI agar plates plus 0.5% (wt/vol) yeast extract at 30°C and 5% CO₂. E. coli DH5α, BL21, and Shuffle (NEB), used for cloning and protein purification, were grown in Luria-Bertani (LB) broth or on Luria agar at 37°C. Ampicillin (Amp) was added at 50 μg ml⁻¹ or 100 μg ml⁻¹. C. diphtheriae was grown in heart infusion (HI) broth (Becton Dickinson) or on HI agar plates at 30°C or 37°C. When necessary, kanamycin (Kan) was added at 25 μg ml⁻¹.

Recombinant plasmids. (i) pH 6-MdbA_Cm.

To generate a recombinant plasmid expressing a His-tagged MdbA protein of C. matruchotii, primers LIC-CmMdbA-F and LIC-CmMdbA-R (see Table S2 in the supplemental material) were used to amplify the extracellular coding region of C. matruchotii mdbA (corresponding to residues 46 to 242) from the genomic DNA of C. matruchotii strain ATCC 14266 as a template. The generated PCR product was cloned into pMCSG7 using ligation-independent cloning (39). The resulting plasmid was introduced into E. coli BL21(DE3) after verification by DNA sequencing.

(ii) pMdbA_Cm.

A recombinant plasmid expressing C. matruchotii MdbA under the control of the C. diphtheriae mdbA promoter was constructed as follows. The primer pair pCmMdbA-HindIII-A/pCmMdbA-B (Table S2) was used to amplify the C. diphtheriae mdbA promoter region from the genomic DNA of strain NCTC 13129, whereas the primer pair pCmMdbA-C/pCmMdbA-BamHI-D was used to amplify the C. matruchotii mdbA coding sequence from the genomic DNA of strain ATCC 14266. Overlapping PCR was employed to link the promoter region to the coding sequence using both PCR products as the templates and primers pCmMdbA-HindIII-A and pCmMdbA-BamHI-D according to a published protocol (18). The generated PCR product was gel purified and digested with HindIII-HF and BamHI-HF (New England BioLabs) and subsequently ligated into pCGL0243 (40). The resulting plasmid was introduced into E. coli DH5α, and plasmid DNA was isolated for sequencing to confirm the cloned sequence. Finally, the plasmid was electroporated into the C. diphtheriae ΔmdbA mutant accordingly (15).

(iii) pC91A/C94A.

Cys-to-Ala mutations within C. matruchotii MdbA were generated using a published PCR-based method (15). First, overlapping primers harboring the indicated mutations (Table S2) were 5′-phosphorylated and used to PCR amplify the pMdbA_Cm template with Phusion HF DNA polymerase. Purified products were cyclized by ligase (NEB) and then introduced into E. coli DH5α. Plasmid DNA was isolated for DNA sequencing to confirm the mutations prior to introduction into the C. diphtheriae ΔmdbA mutant.

Allelic exchange in C. matruchotii.

An allelic exchange system was developed for C. matruchotii, based on a published protocol for gene deletion in C. diphtheriae (35). Briefly, to generate a deletion construct of C. matruchotii mdbA (HMPREF0299_7193), 1-kb fragments upstream and downstream of mdbA were amplified by PCR using appropriate primers (Table S2) and cloned into pK19mobsacB at BamHI and PstI sites. The resulting plasmid was introduced into E. coli S17-1, and the inserted fragments were subsequently confirmed by DNA sequencing. E. coli S17-1 harboring the deletion plasmid was used for bacterial conjugation with C. matruchotii ATCC 14266. C. matruchotii cells with the plasmid integrated into the chromosome were selected by kanamycin and nalidixic acid, which selectively kills S17-1. Selected integrates were grown overnight at 30°C without antibiotics to induce double-crossover homologous recombination leading to plasmid excision and generating wild-type and mutant alleles. These alleles were screened by cell growth on agar plates containing 1% sucrose. Colony PCR was employed to screen for clones that harbor mutant alleles.

Protein purification.

His-tagged MdbA was purified according to a published procedure (14). Briefly, E. coli cells harboring pH6-MdbA_Cm or pH6-MdbA_Cd were grown in LB at 37°C until an OD₆₀₀ of ∼0.6 was reached. Cells were treated with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) for 3 h at 30°C to induce protein expression. Cells were harvested by centrifugation and resuspended in EQ buffer (50 mM Tris-HCl [pH 7.5], 100 mM NaCl) prior to cell lysis by using a French press cell. Clear lysates obtained by centrifugation were subject to nickel-nitrilotriacetic acid (Ni-NTA) affinity chromatography, and purified proteins were subjected to desalting and concentrating using Amicon Ultra centrifugal filters (Merck) at 4°C and stored at −20°C. One milligram of His-tagged MdbA_Cm was used for antibody production (Cocalico Biologicals, Inc., PA).

For crystallization studies, His-tagged MdbA was purified from lysates of E. coli cells harboring pH 6-MdbA_Cm as previously described (15). The purified protein was treated with tobacco etch virus (TEV) protease (concentration, 0.15 mg per 20 mg MdbA) for 16 h at 4°C and then passed through an Ni-NTA column to remove both the TEV protease and the cleaved N-terminal tags. The final step of purification was gel filtration on a HiLoad 16/60 Superdex 200pg column (GE Healthcare) in 10 mM HEPES buffer (pH 7.5), 200 mM NaCl, and 1 mM DTT. The protein was concentrated on Amicon Ultracel 10K centrifugal filters (Millipore) to approximately 25 mg/ml.

Protein crystallization, data collection, structure determination, and refinement.

Using the sitting-drop vapor-diffusion technique in 96-well plates with the MCSG crystallization suite (Microlytic) and Pi-minimal and Pi-PEG screens (Jena Bioscience) (41), the initial crystallization condition was determined with a sparse crystallization matrix at 4°C and 16°C. Only E9 conditions (31.4% PEG 8000, 150 mM citrate buffer [pH 5.5]) of the Pi-minimal screen yielded crystals at both temperatures after 10 days. Crystals were flash-cooled in liquid nitrogen without any cryoprotectant.

Single-wavelength X-ray diffraction data were collected at a temperature of 100 K at the 19-ID beamline of the Structural Biology Center at the Advanced Photon Source at Argonne National Laboratory as previously described (15). The structure was determined by molecular replacement using the HKL3000 suite (42) incorporating MOLREP programs (43). The coordinates of C. diphtheriae MdbA (PDB 5C00) (15) were used as the starting model. Several rounds of manual adjustments of structure models were performed using COOT (44) and refinements with the Refmac program (45) from the CCP4 suite (46). The Phenix suite (42) incorporating MolProbity (47) tools was used to validate the stereochemistry of the structure. Data collection and refinement statistics are presented in Tables 1 and 2.

Cell fractionation and Western blotting.

Cell fractionation and Western blotting were performed according to published procedures with some modifications (15, 48). Briefly, mid-log-phase cultures of C. diphtheriae strains grown at 30°C were normalized an OD₆₀₀ of 1.0 and subjected to cell fractionation. Protein samples obtained from culture medium (S), cell wall (W), and membrane (M) fractions were trichloroacetic acid (TCA) precipitated and acetone washed. For induction of DT, bacterial cultures grown to OD₆₀₀ of ∼0.3 were treated with 1 mM EDDHA (Sigma) for 3 h. Only culture medium fractions were harvested for protein precipitation.

Protein samples were resuspended in sample buffer containing SDS and heated at 60°C for 10 min prior to SDS-PAGE analysis using 3 to 12% or 3 to 20% Tris-glycine gradient gels. Detection of proteins was performed by immunoblotting with specific antibodies (1:10,000, anti-SpaA; 1:5,000, anti-MdbA_Cm; 1:3,000, anti-MdbA_Cd; and 1:1,000, anti-DT).

DTT sensitivity assay.

E. coli, A. oris, S. aureus, C. matruchotii, and C. diphtheriae were grown in appropriate media (E. coli in LB broth; A. oris, S. aureus, and C. diphtheriae in HI broth; and C. matruchotii in BHI broth) until an OD₆₀₀ of ∼1.0 was reached. Cells were harvested by centrifugation and washed with fresh medium before inoculating in fresh medium supplemented with 0, 1, 5, or 10 mM DTT, at the ratio of 1:50 (culture volume prior to washing). Cell growth (OD₆₀₀) was recorded every hour for the calculation of doubling times using GraphPad Prism 5.0. The results are presented as averages from three independent experiments.

Reconstitution of disulfide bond formation in vitro.

Recombinant FimA isolated from E. coli was purified and reduced with 100 mM DTT as previously described (14, 15). For disulfide bond formation, 5 μl of 13.75 μM reduced FimA was reacted with 10 μl of 60 μM MdbA_Cm or MdbA_Cd in 10 μl of redox buffer (100 mM Tris-HCl, pH 3.0, 2 mM EDTA, 0.2 mM glutathione disulfide [GSSG], and 1 mM glutathione [GSH]) at 37°C. Mock-treated samples were used as controls. At time intervals of 0, 5, 15, and 30 min, the reactions were stopped by treatment with 25 μl of Mal-PEG solution (1% SDS, 200 mM Tris-Cl [pH 8], 20% Mal-PEG) for 1 h at 37°C. Protein samples were then TCA precipitated and acetone washed. Protein pellets were resuspended into 20 μl of sample buffer containing 3 M urea and boiled for 10 min prior to SDS-PAGE using 4 to 20% Mini-Protean TGX precast protein gels (Bio-Rad). Proteins were visualized by Coomassie blue staining.

Electron microscopy.

Log-phase cells of various C. diphtheriae and C. matruchotii strains grown in the presence or absence of 5 mM DTT were harvested and subjected to electron microscopy as previously described (49, 50). Cells immobilized on grids were washed 5 times with distilled water and stained with 1% uranyl acetate for 1 min prior to viewing with a JEOL JEM-1400 electron microscope.

Statistical analysis.

Statistical analysis was performed using GraphPad Prism 5.0 (La Jolla, CA), with significant differences calculated using the unpaired t test with Welch's correction. Results are presented as averages from 3 independent experiments ± standard deviations (SD). A nonparametric, two-tailed value for P of ≤0.05, ≤0.01, or 0.001 was considered significant.

Accession number(s).

The atomic coordinates and structure factors of C. matruchotii MdbA were deposited into the Protein Data Bank under accession number 6BO0.