|Home | About | Journals | Submit | Contact Us | Français|
The c subunit of Streptococcus mutans ATP synthase (FoF1) is functionally exchangeable with that of Escherichia coli, since E. coli with a hybrid FoF1 is able to grow on minimum succinate medium through oxidative phosphorylation. E. coli F1 bound to the hybrid Fo with the S. mutans c subunit showed N,N′-dicyclohexylcarbodiimide-sensitive ATPase activity similar to that of E. coli FoF1. Thus, the S. mutans c subunit assembled into a functional Fo together with the E. coli a and b subunits, forming a normal F1 binding site. Although the H+ pathway should be functional, as was suggested by the growth on minimum succinate medium, ATP-driven H+ transport could not be detected with inverted membrane vesicles in vitro. This observation is partly explained by the presence of an acidic residue (Glu-20) in the first transmembrane helix of the S. mutans c subunit, since the site-directed mutant carrying Gln-20 partly recovered the ATP-driven H+ transport. Since S. mutans is recognized to be a primary etiological agent of human dental caries and is one cause of bacterial endocarditis, our system that expresses hybrid Fo with the S. mutans c subunit would be helpful to find antibiotics and chemicals specifically directed to S. mutans.
FoF1-ATP synthase catalyzes both the synthesis and hydrolysis of ATP coupled to H+ transport across membranes. This enzyme is widely distributed from prokaryotes to eukaryotes. It is localized in bacterial plasma membranes and mitochondrial inner and chloroplast thylakoid membranes. F1 has three catalytic sites, and Fo is an H+-transporting pathway. The synthesis and hydrolysis of ATP at catalytic sites are coupled to H+ transport through Fo. Bacterial FoF1 is composed of membrane-peripheral F1 (α3β3γδε) and membrane-integral Fo (ab2c10) (1). The c subunit of Fo forms a hairpin structure in the membrane (2). It is further suggested that the linker region of the c subunit joining two transmembrane helices constitutes a part of the F1-binding site together with other Fo subunits (3). The essential residue for H+ transport is believed to be a conserved acidic residue (Glu or Asp) in the second transmembrane helix of the c subunit (4).
Recent mechanistic studies demonstrated that the bacterial FoF1 is a molecular motor, which can be dissected into the rotor (γ, ε, and c) and the stator (α, β, δ, a, and b) (5). According to the rotation mechanism of ATP synthases, an H+ gradient drives rotation of the c subunit ring accompanying H+ transport from the periplasm to the cytoplasm through the H+ pathway composed of the a and c subunits. The γ subunit on the c subunit ring concomitantly rotates together with the ring, inducing changes in the three catalytic sites and ATP synthesis. On the contrary, ATP hydrolysis at the catalytic sites induces rotation of the γ subunit together with the c subunit ring, resulting in the transport of H+ from the cytoplasm to the periplasm through the H+ pathway.
S. mutans is implicated as the principal causative agent of human dental caries, which is one of the most common infectious diseases (6). This bacterium is able to survive in the low pH environment of dental caries, which is ascribed to the presence of FoF1 (7). FoF1 of S. mutans is suggested to pump H+ back out of the cell and thereby protect acid-sensitive glycolytic enzymes (8). Furthermore, S. mutans is recognized to be associated with infective endocarditis (9). It is reported that the binding site for diarylquinoline is the c subunit of Mycobacterium tuberculosis (10, 11). Although the conserved acidic residue in the transmembrane helix of the c subunit functions in H+ transport, the surrounding residues differ among species (4). Thus, the c subunit of M. tuberculosis is suggested to be a specific target of antituberculosis agents. These observations prompted us to express S. mutans FoF1 and find specific inhibitors of this bacterial FoF1.
In this study, we expressed the S. mutans c subunit in E. coli together with other E. coli FoF1 subunits and characterized the resulting hybrid FoF1.
S. mutans DNA (7) was extracted, and its entire ATP operon was cloned as described for Fig. S1 and S2 in the supplemental material. The resulting plasmid containing the cloned ATP operon, pP5-2-SM-BWU13 (see Fig. S2A), did not produce active ATP synthase in E. coli DK-8, which is a deletion strain of the ATP operon (12). Thus, we attempted to substitute the E. coli c subunit for the corresponding one of S. mutans, since this is the smallest FoF1 subunit and can be easily manipulated.
The c subunit gene of E. coli was deleted from pBWU13 (13), as indicated in Fig. S3A in the supplemental material. The small HindIII-SphI fragment was cloned into pUC18, and then SpeI and NgoMIV sites were introduced between the a/c and c/b intersubunit genes, respectively, using a PrimeSTAR mutagenesis basal kit (TaKaRa) with primers mSpeI-F/mSpeI-R and mNgoMIV-F/mNgoMIV-R, respectively (see Table S1). After SpeI and NgoMIV digestion, the large SpeI-NgoMIV fragment was treated with Klenow enzyme and self-ligated. The HindIII-SphI fragment, from which the SpeI-NgoMIV fragment carrying the c subunit gene had been deleted, then was reintroduced into pBWU13. The resulting plasmid was named pBWU13Δc (see Fig. S4C).
The expression plasmids for c subunits were constructed as follows (see Fig. S3B in the supplemental material). The S. mutans c subunit gene with a promoter was amplified from pP5-2-SM-BWU13 by means of PCR using primer pair P5-(c)SD-F and SMC-R (see Table S1). The PCR conditions comprised preheating (96°C, 1 min), followed by 35 cycles of denaturation (96°C, 10 s), annealing (55°C, 5 s), and extension (72°C, 30 s). The amplified fragment was cloned into the HincII site of pMW218 (GenBank accession number AB005477) (Nippon Gene), and the resulting plasmid was named pSmcEE (expression plasmid for S. mutans c subunit with glutamic acids at positions 20 and 53) (see Fig. S4B). The E. coli c subunit gene with the promoter was amplified from pBWU13 using primer pair mSpeI-R and ECC-R. The resulting expression plasmid was named pEcc (see Fig. S4A).
Mutations (E20Q, E20A, E53Q, and E53A) of the S. mutans c subunit gene were introduced by using a PrimeSTAR mutagenesis basal kit together with pSmcEE and one of the primer pairs (SMC-mE20Q-F/R, SMC-mE20A-F/R, SMC-mE53Q-F/R, or SMC-mE53A-F/R) listed in Table S1 in the supplemental material, and then the mutated genes were amplified with primer pair P5-(c)SD-F and SMC-R as described above. The expression plasmids were named pSmcQE, pSmcAE, pSmcEQ, and pSmcEA, respectively. Second mutations (E20Q+E53A and E20Q+E53Q) were introduced with pSmcQE and the primer pairs SMC-mE53A-F/R and SMC-mE53Q-F/R, respectively. The DNA sequence was confirmed by the dideoxy chain termination method (14). The molecular biological techniques used for DNA manipulations were based on standard procedures (15).
A rich medium (L broth) and minimum medium containing thymine (50 μg/ml), thiamine (2 μg/ml), isoleucine (50 μg/ml), valine (50 μg/ml), and a carbon source (0.5% glucose or 0.5% succinate) were used (16). For preparation of membrane vesicles, 0.5% glycerol was used as the sole carbon source. Ampicillin (100 μg/ml) (Wako) and kanamycin (10 μg/ml) (Wako) were added to L broth to select cells harboring expression plasmid pBWU13 (and its derivatives) and pMW218 (and its derivatives), respectively. Both antibiotics were added together to select cells with both plasmids.
Membrane vesicles were prepared after passing cells through a French press (FA-078; AMINCO) using an FA-031 cell at 1,250 kg/cm2 pressure, and the ATPase activity was determined by measuring the release of inorganic phosphate colorimetrically (17). N,N′-dicyclohexylcarbodiimide (DCCD; Wako) was added as an ethanol solution prepared as a 20 mM stock solution. Formation of an H+ gradient was assayed by measuring the fluorescence quenching of acridine orange (Sigma) (18). Protein concentrations were determined by means of the Bio-Rad protein assay (19) using bovine serum albumin (fraction V; Sigma) as a standard.
Protein samples were mixed with 5× sample buffer and then heat treated at 98°C for 5 min. They were separated by SDS-polyacrylamide gel electrophoresis (small gel [10 by 10 cm], 1 mm thick, consisting of a 3% [wt/vol] stacking gel and a 9% [wt/vol] separation gel ) and subjected to Western blotting (21). Proteins were electroblotted (25 V, 3 h; ATTO model AE-8130) onto a Hybond-P polyvinylidene difluoride (PVDF) membrane (GE Healthcare). Mouse monoclonal antibodies recognizing the E. coli F1 β subunit (Mito Science) were reacted for 1 h as the first antibodies at ×1,000 dilution in TBS-T (25 mM Tris-HCl [pH 7.6], 150 mM NaCl containing 0.1% [wt/wt] Tween-20). The membrane was washed in TBS-T for 5 min (3 times). Horseradish peroxidase-linked sheep anti-mouse Ig (×5,000 dilution) was used as the second antibody with ECL prime western blotting detection reagent (GE Healthcare). The β subunit was visualized with a LAS-3000 (Fuji Film).
Restriction enzymes and calf intestine phosphatase were purchased from NEB. The DNA ligation kit (version 1) and PrimeSTAR HS DNA polymerase were supplied by TaKaRa. The Klenow enzyme and agarose-LE classic type were provided by Toyobo and Nacalai, respectively. Oligonucleotides were purchased from Hokkaido System Science. All other chemicals used were of the highest grade commercially available.
We initially attempted to express S. mutans ATP synthase in DK-8, an E. coli strain lacking ATP operon production (12). However, the cloned ATP operon of S. mutans could not function in DK-8 (see Fig. S1 and S2 in the supplemental material). Thus, we introduced pSmcEE, carrying the S. mutans c subunit gene, into DK-8 harboring only pBWU13Δc, which could not grow on succinate plates but could grow on glucose plates. Interestingly, the transformant could grow on succinate plates as well as on glucose plates, similar to DK-8 harboring both pBWU13Δc and pEcc (Fig. 1A). These results suggest that the c subunit of S. mutans assembles together with the E. coli a and b subunits into Fo and complements the function of the E. coli c subunit of ATP synthase.
We next examined the growth behavior of E. coli with the hybrid FoF1 in liquid medium of neutral and acidic pH. Since the S. mutans c subunit has an additional acidic residue, Glu-20, together with H+-transporting Glu-53 in transmembrane helices (Fig. 1C), it seems likely that the growth of DK-8 with the hybrid FoF1 would respond to the medium pH. DK-8 expressing E. coli FoF1 grew well on glucose medium at pH 7.5 compared to growth at pH 5.5 (Fig. 2A and andC).C). Without the c subunit (Δc) or FoF1 (DK-8), E. coli could not grow on succinate medium (Fig. 2C), and the growth behavior on glucose medium was similar to that with E. coli FoF1 (Fig. 2A).
The growth of E. coli with the hybrid FoF1 on glucose medium was similar to that with E. coli FoF1 (Fig. 2A, closed and open circles, respectively). On succinate medium, cells grew better at pH 5.5 (Fig. 2C, closed and open circles, respectively). These results clearly indicate that E. coli with the hybrid FoF1 could grow well at acidic pH on succinate medium, although the growth curve at pH 5.5 was almost identical to that of cells with E. coli FoF1 at pH 7.5 (Fig. 2C).
The functional substitution of the c subunits of E. coli and S. mutans caused us to be interested in how chimeric Fo transports H+ and affects the catalytic activity of F1. We first determined F1 assembly on the membranes using an antibody for the β subunit. As shown in Fig. 3, the β subunit could be detected in the positive control (lane 2). Without expression of the c subunit, a significantly small amount of the β subunit was recovered from the membrane (lane 3). The negative control without FoF1 did not show any background signal (lane 1). Upon expression of the E. coli c subunit, as much as a 7-times higher level of the β subunit was detected on the membranes (lane 4) compared to that on the membranes without the c subunit (lane 3). Interestingly, much greater amounts of the β subunit were found with the S. mutans c subunit (lane 5). These results indicate that the F1 binding site could be similarly formed with the S. mutans c subunit together with the subunits a and b of the E. coli Fo.
We further assayed the ATPase activity of F1 to confirm the functional assembly of F1 on Fo. The membrane F1-ATPase activities of E. coli FoF1 and the hybrid FoF1 were ten and twenty times higher than those in negative controls without the c subunit or without FoF1, respectively (Table 1) The F1-ATPase activity expressed from the E. coli ATP operon of pBWU13 was much higher than that from the c subunit gene plus the rest of the subunit genes from pBWU13Δc. When the ATPase activities per amount of β subunit were compared (Table 1, right), the calculated values for E. coli FoF1 and the hybrid enzyme (2.24 and 1.45, respectively) were not significantly different, although the value for the latter is slightly lower.
DCCD binding to an H+-transporting acidic residue of the Fo c subunit affects the catalytic activity of F1 through a stalk region (22). As expected from the functional substitution of the S. mutans c subunit, the F1-ATPase activity of the hybrid FoF1 on membranes was DCCD sensitive, similar to that of E. coli FoF1 (Fig. 4). At a concentration of 20 μM DCCD, 50% of the activities were inhibited, whereas the lower level of ATPase activity without the c subunit was not sensitive to DCCD. These results suggest that F1 on membranes interacted normally with the hybrid Fo composed of the S. mutans c subunit and the E. coli a and b subunits.
It was interesting to determine whether or not the hybrid FoF1 could transport H+ coupled with ATP hydrolysis, since F1 interacts with Fo with the S. mutans c subunit normally. The ability of the hybrid FoF1 to catalyze in vivo oxidative phosphorylation in succinate medium strongly indicated that H+ transport through Fo from outside to inside the plasma membrane is coupled with ATP synthesis at the catalytic sites of F1. Although ATP-dependent H+ transport was detectable using membrane vesicles with E. coli FoF1, the activity could not be detected with the hybrid FoF1 (Fig. 5A). This negative result was not due to the H+-related leakiness of the membranes, since a respiratory chain-dependent H+ gradient was formed in the presence of d-lactate (Fig. 5B) and the formation of the gradient was only slightly increased in the presence of DCCD (Fig. 5). The H+ transport of E. coli FoF1 decreased with reduction of the pH of the assay buffer from 8.0 to 5.5. Furthermore, H+ transport of the hybrid FoF1 could not be detected in this pH range (data not shown).
The properties of the hybrid FoF1 could be due to the presence of an acidic residue in the first transmembrane helix (Fig. 1C). Thus, we mutated Glu-20 of the S. mutans c subunit to Gln-20 or Ala-20 (E20Q [SmcQE] and E20A [SmcAE] mutants, respectively) and expressed it instead of the wild-type S. mutans c subunit. Both mutants grew on succinate plates (Fig. 1B). The E20Q mutant grew as well on succinate plates as the wild-type S. mutans c subunit, unlike the E20A mutant. The difference in growth between neutral and acidic pH in the liquid succinate medium observed with the hybrid enzyme containing the wild-type S. mutans c subunit was almost completely abolished with the E20Q mutation (Fig. 2D, closed and open triangles, respectively). The growth behavior on glucose medium was similar for the hybrid FoF1 with the wild-type (SmcEE) and mutant (SmcQE) c subunits (Fig. 2B, closed and open circles and closed and open triangles, respectively). The membrane ATPase activity and DCCD sensitivity of the hybrid FoF1 with SmcEE and SmcQE were essentially the same (Table 1 and Fig. 4). Interestingly, ATP-dependent H+ transport into the membranes was detectable with the hybrid FoF1 with the E20Q mutation (Fig. 5A).
We further replaced the conserved amino acid residue (Glu-53) with Gln or Ala together with the Glu-20 residue (E53Q, E53A, E20Q+E53A, E20A+E53Q, E20Q+E53Q, and E20A+E53A). None of these mutant c subunits was active in in vivo oxidative phosphorylation, since E. coli cells expressing either of these c subunits could not grow on succinate plates or liquid medium (Fig. 1B and and2D).2D). Furthermore, the F1 did not assemble on the membranes of the cells with E53Q and E20Q+E53Q mutations (Fig. 3, lanes 7 and 8). Thus, the Glu-53 residue of the S. mutans c subunit is essential for the formation of functional FoF1 as well as H+ transport. Further, it is interesting to know whether the mutated c subunits are assembled in Fo or not.
The present results indicated that the c subunit of the S. mutans ATP synthase complex is functionally exchangeable with the c subunit of E. coli ATP synthase. We could not express functional S. mutans ATP synthase from the cloned ATP operon (see Fig. S1 and S2 in the supplemental material), possibly due to the presence of unusual initiation codons and different codon usages in S. mutans (NCBI reference sequence NC_004350) (7). The initiation codons for the a and α subunits of the S. mutans ATP operon are both TTG. Furthermore, the codon usages are also different between S. mutans and E. coli; the third letters of the S. mutans codons are rich in T and A nucleotide residues. It is also likely that the chaperone required for membrane insertion of Fo subunits (23) cannot function for the S. mutans a subunit in E. coli.
Subunits β and ε of F1 from Chlorobium limicola could complement E. coli cells defective in oxidative phosphorylation (24). The β subunit of Streptococcus sanguinis is assembled into E. coli F1, and F1-ATPase activity is reconstituted in vivo. However, this activity is less sensitive to DCCD than the E. coli FoF1 complex, suggesting that the hybrid F1 binds abnormally to E. coli Fo, which is consistent with the insufficient growth of E. coli cells expressing the hybrid F1 on minimum succinate medium (25, 26). In our case with the S. mutans c subunit, the hybrid Fo binds to E. coli F1 normally, since the F1-ATPase activities of E. coli F1 bound to E. coli Fo and the hybrid Fo showed similar DCCD sensitivities. Thus, the S. mutans c subunit assembled into functional Fo together with the E. coli a and b subunits, forming not only a normal F1 binding site but also an H+ pathway for ATP synthesis. Although the H+ pathway should be functional, as was suggested by the growth on minimum succinate medium through oxidative phosphorylation in vivo, ATP-driven H+ transport could not be detected with inverted membrane vesicles in vitro. This observation is partly explained by the presence of an acidic residue (Glu-20) in transmembrane helix 1 (TM1) of the S. mutans c subunit, since a site-directed mutant carrying Gln-20 partly recovered the ATP-driven H+ transport (Fig. 5A).
E. coli cells with the hybrid FoF1 grow well at acidic pH on succinate medium. This suggests that the wild-type S. mutans c subunit is more functional at pH 5.5 than pH 7.5 with the ATP synthesis mode of the hybrid FoF1. Such pH-dependent behavior could be ascribed to the presence of an acidic residue in the TM1 of the S. mutans c subunit, since the E20Q mutant did not show any pH dependency of growth.
Asp-61, which is essential for oxidative phosphorylation through E. coli Fo, could be moved to D24 (D24/G61) (27, 28) but not to D28 (D28/G61) or E28 (E28/G61) (29). Similar to the latter substitutions at position 28 of E. coli, the S. mutans c subunit with the E20/Q53 substitution was not functional in oxidative phosphorylation. It must be noted that position 28 of the E. coli c subunit corresponds to position 20 of the S. mutans c subunit (Fig. 1C).
It has been reported that the binding site for diarylquinoline is the c subunit of Mycobacterium tuberculosis (11). Since S. mutans is recognized to be a primary etiological agent of human dental caries and causative of bacterial endocarditis (6, 9), it must be stressed that our system expressing a hybrid Fo with the S. mutans c subunit would be helpful to find antibiotics and chemicals specifically directed to the S. mutans c subunit.
This research was supported in part by a grant from MEXT (Grant-in-Aid for Strategic Medical Science Research Centers, 2010 to 2014 [The MIAST Project]).
Published ahead of print 23 August 2013
Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.00542-13.