Features of Subunit NuoM (ND4) in Escherichia coli NDH-1

The bacterial H+-pumping NADH-quinone oxidoreductase (NDH-1) is an L-shaped membrane-bound enzymatic complex. Escherichia coli NDH-1 is composed of 13 subunits (NuoA–N). NuoM (ND4) subunit is one of the hydrophobic subunits that constitute the membrane arm of NDH-1 and was predicted to bear 14 helices. We attempted to clarify the membrane topology of NuoM by the introduction of histidine tags into different positions by chromosomal site-directed mutagenesis. From the data, we propose a topology model containing 12 helices (helices I–IX and XII–XIV) located in transmembrane position and two (helices X and XI) present in the cytoplasm. We reported previously that residue Glu144 of NuoM was located in the membrane (helix V) and was essential for the energy-coupling activities of NDH-1 (Torres-Bacete, J., Nakamaru-Ogiso, E., Matsuno-Yagi, A., and Yagi, T. (2007) J. Biol. Chem. 282, 36914–36922). Using mutant E144A, we studied the effect of shifting the glutamate residue to all sites within helix V and three sites each in helix IV and VI on the function of NDH-1. Twenty double site-directed mutants including the mutation E144A were constructed and characterized. None of the mutants showed alteration in the detectable levels of expressed NuoM or on the NDH-1 assembly. In addition, most of the double mutants did not restore the energy transducing NDH-1 activities. Only two mutants E144A/F140E and E144A/L147E, one helix turn downstream and upstream restored the energy transducing activities of NDH-1. Based on these results, a role of Glu144 for proton translocation has been discussed.

The proton-translocating NADH-quinone oxidoreductase (designated complex I and NDH-1 for mitochondrial and bacterial enzymes, respectively) constitutes the first energy-coupling site in the respiratory chain in most eukaryotic and prokaryotic cells. It catalyzes the reduction of quinone using NADH as electron donor. This activity is coupled to the translocation of protons through the inner mitochondrial/cytoplasmic membranes (1)(2)(3)(4). Complex I is one of the largest and most intricate proteins of the respiratory chain, e.g. bovine complex I was reported to contain 45 different subunits (5,6). In contrast, NDH-1 has relatively simpler structure and seems suitable to study the coupling site 1. In case of Escherichia coli, NDH-1 is made up of only 13 subunits (designated NuoA-N), which are homologous to the 14 subunits that compose the central core of mitochondrial complex I (7,8). Electron microscopic images of both mitochondrial complex I and bacterial NDH-1 have revealed a characteristic L-shaped form with two distinct domains, a hydrophilic peripheral arm projected into the mitochondrial matrix (or bacterial cytoplasm) and a transmembrane hydrophobic arm (9). The hydrophilic domain catalyzes electron transfer from NADH to quinone and carries all of the cofactors (one FMN and eight or nine iron-sulfur clusters) (10 -13). In contrast, the membrane domain is considered to participate in proton translocation and quinone binding (14 -18).
The three-dimensional structure of the hydrophilic arm of NDH-1 from Thermus thermophilus has been determined, but there is little structural information available about the hydrophobic domain (19). In E. coli NDH-1 the transmembrane arm is composed of seven subunits (namely NuoA, -H, -J, -K, -L, -M, and -N), all of them homologous to the seven mitochondrialencoded subunits in eukaryotes (ND1-6 and ND4L) (20 -22). Preliminary structural studies have outlined the distribution of hydrophobic subunits in the membrane arm, in which NuoJ, -H, -A, -K, and -N are close to the hydrophilic arm, whereas NuoM and -L have a distal location (23)(24)(25). The last two subunits together with NuoN show a high degree of sequence homology. It was also described that NuoN, -M, and -L are well conserved with their counterpart subunits among the different complex 1/NDH-1 enzymes and with some Na ϩ /H ϩ membrane antiporters and energy-converting [NiFe]-hydrogenases (26 -28). This sequence similarity, mainly with Mrp-like Na ϩ /H ϩ antiporters, has led to relating the function of NuoL, -N, and -M subunits with proton translocation activity of NDH-1 (16,18,29,30).
Despite the absence of clear structural data on the membrane domain and the difficulty involved in working with hydrophobic polypeptides, molecular biology techniques were found to be very useful for the characterization of membrane domain subunits (31). The possibility of generating chromosomal deletions and the introduction of site-directed mutations in the nuo genes by homologous recombination has allowed functional studies on various hydrophobic subunits, such as NuoA (ND3), NuoJ (ND6), NuoK (ND4L), NuoN (ND2), NuoM (ND4), and NuoH (ND1) (32)(33)(34)(35)(36)(37)(38)(39). The NuoM subunit, counterpart of the eukaryotic ND4 subunit, is one of the largest polypeptides in the hydrophobic arm of NDH-1. This highly hydrophobic subunit was predicted to be composed of 14 helices (23,30). In a recent work, we introduced site-directed mutations in several conserved charged amino acids of the NuoM subunit (36). This study, lately corroborated by Euro et al. (40), showed the impor-tance of NuoM in the NDH-1 activity. Mutations of the conserved Glu 144 and Lys 234 completely abolished the energy-coupled activities of NDH-1 in the E144A mutant and led to a 90% activity suppression in the case of K234A mutant. Interestingly, in the case of the transmembrane conserved Glu 144 , its conservative mutation to Asp did not show any effect on NDH-1 activity. In this regard, highly conserved transmembrane carboxyl residues in NuoK (ND4L) and NuoA (ND3) (34,35,38) and in some other membrane proteins were reported to play important roles in proton translocation (41)(42)(43)(44)(45). Thus, it was suggested that the essential Glu 144 of the NuoM subunit could play a key role in the proton pump activity of NDH-1. As a first step to clarify the mechanism of H ϩ translocation in NDH-1, further characterization of NuoM was a prerequisite.
In this work, we have first attempted to experimentally clarify the topological structure of the E. coli NuoM subunit by introducing His 6 tag in the N-and C-terminal regions and the potential extramembrane loops of NuoM. Second, we have studied the importance of the location of essential carboxyl group of Glu 144 for the proton pump activity of NDH-1 by shifting its position in transmembrane helix V.

EXPERIMENTAL PROCEDURES
Materials-Chemicals, NDH-1 substrates and antibiotics were from Sigma-Aldrich. p-Nitroblue tetrazolium was from EMD Biosciences (La Jolla, CA). The endonucleases were from New England Biolabs (Beverly, MA). PCR product, gel extraction, and plasmid purification kits and anti-histidine antibodies (Penta⅐His antibody) were from Qiagen. All of the antibodies against E. coli NDH-1 subunits NuoM-1, NuoM-7, NuoAc, NuoB, NuoCD, NuoE, NuoF, and NuoG were obtained previously in our lab (10,35,36,46). The secondary antibodies, antimouse IgG (horseradish peroxidase conjugate) and anti-rabbit IgG (horseradish peroxidase conjugate) were from Calbiochem (La Jolla, CA) and GE Healthcare, respectively. The E. coli strains and vectors used in this work are listed in supplemental  Table S1. The primer sequences used in this study are summarized in supplemental Table S2. Capsaicin 40 was generously provided by Dr. Hideto Miyoshi (Kyoto University, Kyoto, Japan). The pKO3 vector was a generous gift from Dr. George M. Church (Harvard Medical School, Boston, MA).
Mutagenesis of E. coli nuoM Gene and Preparation of Mutant Cells-The strategy used to obtain the double site-directed NuoM E144A/XnE (EAE) 2 mutants was the same as we reported in the previous paper (36). In summary, the NuoM mutants were constructed using the QuikChangeII XL kit (Stratagene, Cedar Creek, TX). Fig. 1 shows the locations of insertion points of His 6 tag (pHis) on the NuoM polypeptide. To obtain the NuoM site-specific pHis mutants, we used the same strategy used for the EAE mutants, although the pHis were inserted in two consecutive mutagenic PCRs in which His 3 in the first round and remaining His 3 in the second PCR were inserted in the positions of interest. pGEM/nuoM was used as a template to obtain the pHis mutants and pGEM/ nuoM(E144A) for the NuoM EAE mutants. Afterward, the mutated nuoM fragments were cloned in pKO3/nuoM-T, generating the pKO3/nuoM-T(pHis mutants) or pKO3/nuoM-T(EAE mutants) as we described previously (36). The above plasmids were used to introduce the site-specific mutations in the nuoM gene from E. coli MC4100, by replacing the spc (spectinomycin) gene for the individual mutated nuoM gene in NuoM-KO mutant. The process was performed according to the method described by Link et al. (47) and modified by Kao et al. (35). The mutations were confirmed by DNA sequencing.
E. coli Membrane Preparation-E. coli mutants were grown in terrific broth medium at 37°C until A 600 of 2. Then the cells were harvested at 5800 ϫ g for 10 min and frozen at Ϫ80°C. Inside-out (ISO) membrane vesicles were prepared according to the method described previously (33)(34)(35)(36). Right-side-out (RSO) membrane vesicles were prepared following the method described by Kao et al. (48).
Immunochemical Analyses and Blue Native Gel Electrophoresis-The hydrophilic subunits of the membrane-associated NDH-1 were examined by Western blot of the E. coli membranes using antibodies specific to the NuoM-7, NuoB, NuoCD, NuoE, NuoF, and NuoG subunits. Blue native PAGEs were performed according to the method described previously (35). The assembly of NDH-1 was confirmed by immunoblotting using anti-NuoM-1 antibodies and NADH dehydrogenase activity staining.
The topological characterization of NuoM was performed by a dot-blot experiment. 50 l of 2.5 g of protein/ml of ISO and RSO membrane vesicles of the different NuoM pHis mutants were allowed to bind directly on nitrocellulose membrane in a dot-blot apparatus (Bio-Rad). Afterward, the wells were washed three times with Tris-buffered saline followed by blocking for 1 h with 2% bovine serum albumin in Tris-buffered saline buffer. Penta⅐His antibody was used as primary antibody to determine the orientation of N and C termini and different extramembrane loops in NuoM. Anti-mouse IgG was used as a secondary antibody.
Activity Analysis-The activity assays for NDH-1 mutants were performed according to the methods described previously (49). In brief, the correct assembly of the peripheral subunits of NDH-1 was determined by dNADH-K 3 Fe(CN) 6 reductase activity. The assay was performed at 30°C with 80 g of protein/ml of membrane samples in 10 mM potassium phosphate (pH 7.0) containing 1 mM EDTA, 10 mM KCN, and 1 mM K 3 Fe(CN) 6 . The samples were preincubated for 1 min, and the reaction was started by the addition of 150 M dNADH. All of the measurements were monitored at 420 nm. The dNADH-2,3-dimethoxy-5-methyl-6-decyl-1,4-benzoquinone (DB) reductase activity was performed in the same condition, but the K 3 Fe(CN) 6 was replaced by 50 M DB as electron acceptor. The measurement was followed at a wavelength of 340 nm. The reaction was completely inhibited when 10 M capsaicin-40 was added (50). The dNADH oxidase activity was performed in the same condition as that of the dNADH-DB reductase assay but neither KCN nor DB were added. The extinction coefficients used for activity calculations were ⑀ 340 ϭ 6.22 mM Ϫ1 cm Ϫ1 for dNADH and ⑀ 420 ϭ 1.00 mM Ϫ1 cm Ϫ1 for K 3 Fe(CN) 6 .
Membrane potential generated for the NDH-1 mutants were measured using oxonol VI as reported (33,51). The reaction mixture contained 0.33 mg of protein/ml of membrane vesicle samples in 50 mM MOPS (pH 7.3), 10 mM MgCl 2 , and 50 mM KCl buffer with 2 M oxonol VI. The reaction was started by the addition of 200 M dNADH, and the absorbance changes at 630 -603 nm were recorded. When necessary, 2 M of the proton ionophore FCCP was added as uncoupler. The NDH-1 proton pump activity was followed by ACMA fluorescence quenching (32). 200 M dNADH was used as substrate.
Other Analytical Procedures-Protein concentrations were determined by a BCA protein assay kit (Pierce) with bovine serum albumin as standard, according to the manufacture's instructions. Any variations from the procedures and details are described in the figure legends.

RESULTS
Topological Analysis of NuoM-The E. coli NuoM subunit is the second largest subunit of the hydrophobic arm of NDH-1. We proposed a probable topology for the NuoM subunit that includes 14 potential transmembrane helices (designated helices I-XIV) that derived from six computer programs ( Fig. 1) (36). However, experimental data confirming this conformation were lacking. In the present study, we have attempted to corroborate the previously predicted NuoM topology by detection of the cytoplasmic or periplasmic positions of the out-ofmembrane loops of NuoM. We introduced 18 pHis in loops 1-13 and in the N and C termini by site-directed mutagenesis of the nuoM gene in the chromosome. The effect of the insertion of pHis to the NuoM subunit on its content and other NDH-1 subunits was analyzed by immunoblot (supplemental Fig. S1). In all of the above cases the content of different NDH-1 peripheral subunits such as NuoE (24k), NuoF (51k), and NouCD (30k-49k) was unaffected, but in several NuoM pHis mutants the content of NuoM was found to be altered. In the case of pHis8 mutant, the quantity of NuoM was detected to be lower than the wild type, and NADH dehydrogenase staining of NDH-1 in blue native gel was also lower. However, dNADH-K 3 Fe(CN) 6 reductase activity was scarcely affected in this mutant (supplemental Table S3). In the case of mutant pHis7, the antibody against NuoM-7 reacted with the band of the expected molecular size, whereas the antibody specific to NuoM-1 did not show reactivity with this mutant. Because the His 6 tag was inserted in the middle of the epitope that is recognized by NuoM-1 antibody, this antibody did not work. In fact, as described below, the antibody to pHis recognized the His 6 tag of the pHis7 mutant. In contrast, mutants pHis5, -9, -14, -15, and -17 did not exhibit detectable levels of NuoM subunit nor NDH-1 assembly (supplemental Figs. S1 and S2). In addition, these five mutants almost completely lost dNADH oxidase and dNADH-DB reductase activities (supplemental Table S3). It should be noted that mutations of pHis5 and pHis17 are present in loop 4, mutations of pHis9 and pHis15 are located in loop 8, and a mutation of pHis14 is in loop 12 ( Fig. 1). It is conceivable that His 6 tag insertion in mutants pHis5, -9, -14, -15, and -17 stimulates degradation of NDH-1.
The topology of the loops and the N and C termini was examined by the detection of the pHis on membrane vesicles facing the cytoplasmic side (ISO) or the periplasmic side (RSO) in a dot-blot experiment using anti-penta⅐His antibodies. There was clearly a preferential reaction of the antibody when ISO and RSO were compared. The results of the dot-blot experiments ( Fig. 2A) showed that the antibody reactivity was stronger for the ISO membrane vesicles as opposed to the RSO membrane vesicles for the pHis inserted in the positions pHis4, -6, -8, -11, -12, -13, -16, and -18 (loops 3, 5, 7, 9, 10, 11, 13, and 1, respectively) (Fig. 1). In contrast, the His tag antibody reaction was relatively higher for the pHis inserted in the positions pHis1, -3, -7, -10, and -19 of the RSO membrane vesicles, which corresponds to the C and N termini and the loops 6 and 2 (Fig. 1). As expected, the attempts to detect the His tags inserted in the loops 4, 8, and 12 (NuoM pHis-5, -9, -14, -15, and -17 mutants) were unsuccessful. As stated, the above result showed that helices X and XI are probably located outside of the membrane and in the cytoplasm (Fig. 2B).
NDH-1 Subunit Expression and NDH-1 Assembly in E. coli NuoM EAE Mutants-Glu 144 in subunit NuoM is predicted to be present in the middle of transmembrane helix V (Fig. 2B). Previously we elucidated that Glu 144 in subunit NuoM is almost perfectly conserved, and the mutations E144A and E144Q completely suppressed the energy transducing NDH-1 activities, whereas these mutations did not affect scaler electron transfer such as dNADH-K 3 Fe(CN) 6 reductase activity (36). In addition, mutation E144D barely affected the characteristics of the wild type NDH-1. Therefore, we proposed that Glu 144 might be involved in H ϩ translocation. If that is the case, it is of interest to study how Glu located at position 144 in the NuoM subunit plays a role in H ϩ translocation. One of the approaches is to investigate the effects of repositioning Glu 144 on the properties of NDH-1. For this purpose, we systematically shifted Glu 144 from its original position within the helix V (from Phe 139 to Leu 153 ), and in three positions each in the adjacent helices IV (Val 127 , Ile 128 , and Gly 129 ) and VI (Ile 189 , Leu 190 , and Ala 191 ). We also constructed double site-directed mutants (EAE1 to EAE20) ( Table 1).
The site-directed NuoM EAE mutations were introduced directly in the E. coli chromosome by homologous recombination (36) by using, as a template, the NuoM E144A mutant obtained previously in our laboratory. To verify that the double site-specific mutation of chromosomal DNA does not affect NDH-1, we first analyzed immunochemically the subunit content of NuoM and several other hydrophilic subunits in the membranes from the NuoM EAE mutants (Fig. 3). There was no detectable change in the contents of the analyzed subunits (NuoM, -CD, -B, -E, -F, and -G) among the mutants except for NuoM-KO.
The correct assembly of the NuoM EAE mutants was detected through NADH dehydrogenase activity staining and by immunoblotting, using specific antibody against the NuoM subunit. As shown in Fig. 4, the NDH-1 assembly was not affected by the NuoM EAE mutations. These results, together with the immunoblotting analysis of the subunit content, suggest that none of the EAE mutations affected the expression levels and the assembly of E. coli NDH-1.
Another method employed to analyze whether the mutations introduced in this study altered the NDH-1 assembly includes the measurement of dNADH-K 3 Fe(CN) 6 reductase activity in membrane vesicles of the mutants. This activity is directly related to the dehydrogenase part (subunits NuoE and NuoF), and significant changes in the detected activity can be related to differences in the NDH-1 content and an incorrect assembly. As shown in Table 1, all of the site-directed EAE mutants constructed in this study have a dNADH-K 3 Fe(CN) 6 reductase activity similar to the wild type.
Measurements of NDH-1 Activities in NuoM EAE Mutants-The energy-coupled activities (NADH oxidase and NADH-DB reductase) of NDH-1 were measured in E. coli membrane vesicles using dNADH as substrate (52). As mentioned earlier, the mutation NuoM E144A led to a complete inhibition of the energy-coupled NDH-1 activities. As shown in Table 1, when the glutamic residue was repositioned within the transmem-  brane helix V (mutants EAE1, 3, 5, 6, and 8 -14) or helix IV (EAE15-17) or helix VI (EAE18 -20), none of the original energy-coupled activities of the NDH-1 were restored. In fact, these mutants showed completely abolished dNADH oxidase and dNADH-DB reductase activities similar to the NuoM E144A mutant. Interestingly, these activities were partiality restored when this Glu residue was relocated in positions either immediately preceding or immediately after helix turn. The double mutant E144A/F140E (EAE4) in which the glutamic residue was shifted from position 144 to position 140 (an helix turn upstream the original position) showed energy-coupled NDH-1 activities ϳ40 -60% of wild type. In the case of double mutant E144A/L147E (EAE7), the activities were ϳ30 -45% of the wild type. The mutant E144A/W143E (EAE2) restored moderate activity with only 15% of the wild type. The data suggest that Glu relocation to four residues upstream and three residues downstream from position 144 can significantly restore the energy transducing NDH-1 activities. In other words, Glu shifting to one helix turn upstream and downstream can catalyze the energy transducing NDH-1 activities. Effects of the inhibition by capsaicin-40 on the energy-coupled activities of NDH-1 were studied in the NuoM EAE4 and -7 mutants. Capsaicin-40 is a competitive inhibitor for the quinone in NDH-1 having its binding site located in the membrane domain (50). The IC 50 capsaicin-40 values (0.12 M) of the mutants were the same as the wild type. In complementation, the K m values for the DB were calculated for the above mutants.
Both mutants showed affinity for DB similar to the wild type (36). Taken together, these data suggest that Glu 144 is not involved in the catalytic quinone binding.

Measurement of Membrane Potential and Proton Translocation by
NuoM EAE Mutants-The analyses of membrane potential and proton translocation provide direct evidence that the electron transfer is coupled to energy transduction. We analyzed the generation of membrane potential and proton translocation by using, respectively, oxonol VI and ACMA as described in Ref. 36. As expected, the EAE mutants lacking NADH oxidase or NADH-DB reductase activity did not exhibit any membrane potential (Fig. 5) or proton translocation (Fig.  6). It was of particular interest to study membrane potential and proton translocation of mutants EAE2, EAE4, and EAE7 that could restore dNADH oxidase and dNADH-DB reductase activities. As shown in Fig. 5A, the extent of membrane potential corresponds to the levels of the dNADH oxidase activity in the wild type, EAE7, EAE4, and EAE2. In addition, the wild type, EAE7, EAE4, and EAE2 also showed membrane potential coupled to dNADH-DB reductase (Fig. 5B). Similar trends were observed in the extent of proton translocation (Fig. 6). Therefore, it is clear that restoration of electron transfer in EAE7, EAE4, and EAE2 is coupled to the proton translocation.

DISCUSSION
On the basis of analyses of the primary structures by several computer programs (TopPreIII, PHDhtm, TMHMM, SOSUI, Tmpred, and HMMTOP), the E. coli NuoM subunit was previously predicted to contain 14 transmembrane segments (designated helices I-XIV from the N terminus) (36). Our present topology experiments determined that (i) loops 1, 3, 5, 7, 9 10, 11, and 13 face the cytoplasmic phase and (ii) N and C termini and loops 2 and 6 are directed to periplasmic phase. These results indicate that helices X and XI are not membrane-spanning segments but are present in the cytoplasm. Determination of the sidedness of loops 4, 8, and 12 was not successful. As expected from the primary sequence of the stretch between helix IV, loop 4, and helix V, the peptide is more hydrophobic compared with the stretch between helix X, loop 10, and helix XI. It seems likely that helix V is a transmembrane segment. Taking these together, we propose the topology of E. coli NuoM as shown in Fig. 2B.
The secondary structure of NuoM is considered to be similar to that of NuoL except that the latter has two extra helices at the  C-terminal end. The topology of Rhodobacter capsulatus NuoL based on the alkaline phosphatase fusion method has been reported (30). It was suggested that loops 8 and 12 of NuoL (corresponding to loops 8 and 12 of NuoM) face the periplasmic phase. However, between the two loops in NuoL, when the fusion site was shifted by two residues, the results changed from high activity (suggesting the periplasmic side) to low activity (the cytoplasmic side or in the membrane). The proposed topology of NuoL was based on the latter, and the high activity result was regarded as a false positive. These ambiguous data may be related to known issues of reliability of the procedure involving alkaline phophatase in membrane topological studies (53,54). Determining the final topology of subunits NuoM and NuoL will need further scrutiny.
We constructed two mutants each for loops 4 and 8 and one mutant for loop 12 for topology research. It was interesting to observe that in none of these five His 6 insertion mutants could we detect NuoM in the membrane preparation or NDH-1 assembly in the native gels. These loops seem to be directed to the periplasmic phase. In addition, they are all short and hydrophobic (loop 4, M-FL-F; loop 8, F-SLPLFPNAS-A; loop 12, F-VGEFMILFGSF-Q) (a chargeable residue is underlined). In contrast, loops 8 (F-EYAPEAK-M) and 12 (V-GFAGYLSK D -AIIESAFAS-G) of R. capsulatus NuoL are more hydrophilic. It might be possible that loops 4, 8, and 12 in NuoM are located in the interface between membrane and periplasm, and insertions of His 6 tag into these loops unstabilize NuoM. In other words, structures of loops 4, 8, and 12 seem to be essential to maintain the intact structure of NuoM.
There is a hypothesis for the structural importance of hydrophilic residues in keeping the stability of transmembrane helices and in maintaining the transmembrane position or controlling the movements of the helix inside the membrane depending on their ionization state (55-57). However, mutants EAE1-20 and E144A scarcely affected NDH-1 assembly or the contents of the peripheral subunits and the dNADH dehydrogenase activity of the mutant membranes. In addition, mutant E144D illustrated dNADH dehydrogenase activities similar to the wild type membranes (36). Therefore, it is unlikely that Glu 144 significantly contributes to the structure of NDH-1.
It is recognized that proton-and ion-translocating enzymes in the membrane contain essential carboxyl residue(s) in the hydrophobic regions (e.g. the DCCD-binding subunit of ATP synthase (58) and Na ϩ /H ϩ antiporters (59 -61)). Two such carboxyl groups have been described for NDH-1. Our group and other laboratories showed that Glu 36 in NuoK (ND4L) and Glu 144 in NuoM (ND4) are absolutely essential for the energycoupled NDH-1 activities (34,36,38). Based on the facts that Glu 144 is conserved in energy-converting [NiFe]-hydrogenases  and Mrp (Na ϩ /H ϩ ) antiporters and that NDH-1/complex I are inhibited by amiloride derivatives that are specific inhibitors for Na ϩ /H ϩ antiporters (14,62), we postulated that Glu 144 in NuoM plays a role in the proton translocation of NDH-1. Recently, Kajiyama et al. (61) reported that subunit MrpD in Bacillus subtilis Mrp-like (Na ϩ /H ϩ ) antiporter (a homologue of NuoM) conserves Glu 144 of NuoM (Glu 137 MrpD numbering) so that Glu 137 in MrpD is critically involved in ⌬pH-dependent Na ϩ efflux on the basis of mutagenesis experiments. Glu 137 in MrpD is hypothesized to participate in proton and/or Na ϩ translocation (61), suggesting that NDH-1 and Mrp antiporters may share the same mechanism of energy transduction.
It has been reported that certain repositioning of essential Asp 61 in the DCCD-binding subunit retained the energy-coupled activities of ATP synthase. Mutants D61N/A24D (horizontal repositioning of Asp 61 by switching Asp in adjacent transmembrane helix) and D61N/M57D (repositioning Asp 61 by one helical turn toward the cytoplasmic side) restored the energy transducing activities (58,63). These results suggest that Asp 61 has site flexibility for the proton-translocating function of the carboxyl residue. As described in Ref. 4, NDH-1/complex I can catalyze reverse reaction similar to ATP synthase. We systematically constructed mutants with Glu shifting in helix V and characterized these mutants. The mutants in which Glu was positioned at 140, 144, or 147 can catalyze the energy transduction. Moving the Glu to other positions resulted in a loss of energy coupling. It should be noted that positions 140, 144, and 147 align vertically on helix V. The results are quite similar to what was reported for the Asp of the DCCD-binding protein. It will be interesting to study whether these properties of carboxyl residues are common in other proton (Na ϩ ) translocation systems in general. Based on our results and the proposed mechanism of Na ϩ /H ϩ -antiporter (64), we may speculate that a proton to be pumped by NuoM can penetrate from the cytoplasm to positions 140, 144, and 147 and ligate to the carboxyl residue. Further, the carboxyl residue could translocate the bound proton and release it to the periplasmic side. During one turnover of Na ϩ /H ϩ antiporter, two conformational change steps are considered to be triggered by proton transports. Whether these steps are related to the energy-coupled electron transfer of NDH-1 remains to be seen.