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

The proton-translocating NADH-quinone (Q) oxidoreductase (NDH-1) from Escherichia coli is composed of two segments: a peripheral arm and a membrane arm. The membrane arm contains 7 hydrophobic subunits. Of these subunits, NuoM, a homolog of the mitochondrial ND4 subunit, is proposed to be involved in proton translocation and Q-binding. Therefore, we conducted site-directed mutation of 15 amino acid residues of NuoM and investigated their properties. In all mutants, the assembly of the whole enzyme seemed intact. Mutation of highly conserved Glu144 and Lys234 leads to almost total elimination of energy-transducing NDH-1 activities as well as increased production of superoxide radicals. Their NADH dehydrogenase activities were almost normal. Because these two residues are predicted to be located in the transmembrane segments of NuoM, the results strongly suggest that they participate in proton translocation. Although it is hypothesized that His interacts with a Q head group, mutations at four His moderately inhibited NDH-1 activities and had almost no effect on the Km values for Q or IC50 values of capsaicin-40, a competitive inhibitor for the Q binding site. The data suggest that these His are not involved in the catalytic Q-binding. Functional roles of NuoM and advantages of NDH-1 research as a model for mitochondrial complex I study have been discussed.

structural roles of conserved amino acid residues of the NuoM subunit by utilizing chromosomal DNA mutation methods (19 -21).
In this article, we constructed mutants of 14 conserved and one non-conserved residues by chromosomal DNA manipulation and explored characteristics of these mutants. Of these mutants, E144A, E144Q, and K234A significantly inactivated the energy-transducing NDH-1 activities but scarcely affected NADH dehydrogenase activities, suggesting that Glu 144 and Lys 234 are essential for the coupling site 1.

EXPERIMENTAL PROCEDURES
Materials-The QuikChangeII XL site-directed mutagenesis kit and the Herculase-enhanced DNA polymerase were from Stratagene (Cedar Creek, TX). The endonucleases were from New England Biolabs (Beverly, MA). PCR product, gel extraction, and plasmid purification kits were from Qiagen (Valencia, CA). NADH, dNADH, antibiotics, acetylated cytochrome c, and bovine superoxide dismutase were from Sigma-Aldrich. p-Nitroblue tetrazolium was from EMD Biosciences (La Jolla, CA). pGEM-T Easy Vector was from Promega (Madison, WI). Goat anti-rabbit IgG horseradish peroxidase conjugate was from GE Healthcare (Piscataway, NJ). The gene replacement vector pKO3 was a kind gift from Dr. George M. Church (Harvard Medical School, Boston, MA). Capsaicin-40 (cap-40) was generously provided by Dr. Hideto Miyoshi (Kyoto University, Kyoto, Japan). The peptides were synthesized by Sigma Genosys (The Woodlands, TX). Oligonucleotides were synthesized by IDT (Coralville, IA).
Cloning and Mutagenesis of the E. coli nuoM Gene-The strategies used for cloning and mutagenesis of the E. coli nuoM gene were in principle similar to those we reported for the nuoA, nuoJ, and nuoK genes (19 -21) and illustrated in Fig. 1. Primer sequences are summarized in the supplemental Table  S1. The nuoM gene together with a 1-kb DNA segment upstream and 1-kb DNA segment downstream was cloned by PCR from E. coli DH5␣, generating the fragment nuoM-T. To generate the restriction sites SmaI and SalI, the DNA fragment was amplified from the nuoM-T by PCR using primers A and D. The nuoM-T fragment was cloned in pGEM-T Easy Vector and finally inserted in pKO3 (Fig. 1A). The 1-kb DNA segment upstream with the 5Ј-sequence of nuoM gene was cloned by PCR using primers A and B, generating the fragment nuoM-U, and the 1-kb DNA segment downstream with the 3Ј sequence of nuoM gene was cloned by PCR using primers C and D, generating the fragment nuoM-D. Both fragments were cloned in pGEM-T Easy Vector. The spectinomycin-encoding gene (spc) from Transposon Tn554 from Staphylococcus aureus (32) was cloned by PCR using the primers E2 and F2, both containing the HindIII restriction site. The spc fragment obtained was cloned in pGEM-T Easy Vector and finally inserted between the nuoM-U and nuoM-D fragments, generating the nuoM(spc) fragment and assembled in pKO3 generating the pKO3/nuoM-(spc) plasmid (Fig. 1C). To introduce point mutations in the nuoM gene primers J and M were used to amplify the nuoM gene by PCR. The DNA fragment obtained was cloned in pGEM-T Easy Vector and used as a template for the mutagenesis primers (Fig. 1B). The mutated fragments were inserted in pKO3/nuoM-T using the PflMI and BsmFI restriction sites, generating the pKO3/nuoM-T (mutants).
Preparation of nuoM Knock-out and Mutant Cells-E. coli MC4100 (F Ϫ , araD139, ⌬(arg F-lac)U169, ptsF25, relA1, flb5301, rpsL 150. Ϫ ) was used to generate knock-out and sitespecific mutations of nuoM employing the pKO3 system. The process was carried out according to the method described by Link et al. (33) with the modification described previously by Kao et al. (19). E. coli nuoM knock-out (NuoM-KO mutant) was constructed by replacement of the nuoM gene in the NDH-1 operon for the nuoM(spc) fragment inserted in the pKO3 plasmid. The NuoM-KO mutant obtained were verified by PCR sequencing using specific oligonucleotides and stored in glycerol stocks at Ϫ80°C. In subsequent steps the pKO3/nuoM-T(mutant) plasmids were used for construction of site-directed nuoM mutants. The correct introduction of point mutations in the chromosome was verified by direct DNA sequencing.
Antibody Production-On the basis of facility of peptide syntheses and antigenicity, we selected two regions (Phe 205 to Ser 216 and Lys 447 to Asp 464 ) to raise antibodies specific to E. coli NuoM. Two oligopeptides (H-FNYEELLNTPMSC-OH (ECONuoM-1) and H-CKAKSQIASQELPGMSLRD-OH (ECONuoM-7)) were synthesized and conjugated to the maleimide-activated bovine serum albumin according to the manufacturer's protocol. The anti-ECONuoM-1 and anti-ECONuoM-7 antibodies were affinity-purified by using the ECONuoM-1 and ECONuoM-7 peptides linked to the Sulfo-Link Coupling Gel (Pierce) according to the manufacturer's protocol. The antibodies specific to peripheral subunits NuoB, NuoCD, NuoE, NuoF, NuoG, and NuoI were available from our previous works (34).
E. coli Membrane Preparation-The E. coli membranes were prepared according to the method by Kao et al. (20). In brief, the cells were grown in 250 ml of Terrific Broth medium until A 600 of 2 and were then harvested at 5,800 ϫ g for 10 min. The cells were resuspended at 10% (w/v) in a buffer containing 10 mM Tris-HCl (pH 7.0), 1 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride and 15% (w/v) glycerol. Lysozyme was added to the cell suspension at a final concentration of 0.5 mg/ml and incubated on ice for 30 min. After that, the cell suspensions were sonicated twice during 15 s, were passed twice in a French press at 25,000 p.s.i. and centrifuged at 23,400 ϫ g for 10 min. The supernatant was ultracentrifuged at 256,600 ϫ g for 30 min. The pellet was resuspended in the same buffer as described above, frozen in liquid nitrogen and stored at Ϫ80°C until their use.
Gel Electrophoresis and Western Blotting Analysis-The expression of the NDH-1 subunits was determined by using the antibodies specific to ECONuoM-1, ECONuoM-7, NuoB, NuoCD, NuoE, NuoF, NuoG, and NuoI in Western blot experiments. To investigate the assembly of the NDH-1 subunits Blue-Native PAGE was performed according to Schägger and Von Jagow (35). In brief, E. coli membranes equivalent to a 800 g of protein were resuspended in 160 l of 750 mM aminocaproic acid, 50 mM BisTris-HCl (pH 7.0), 0.1 mg/ml DNase, and 0.8% (w/v) dodecyl-␤-maltoside. After the incubation on ice for 2 h, the samples were centrifuged at 149,000 ϫ g for 5 min. The supernatants were recovered, and 2% (w/v) of Coomassie Blue G in 1 M aminocaproic acid was added at a final concentration of Coomassie Blue of 0.08% (v/v). 10 l of the samples were loaded on a 7% separating gel with a 4% stacking gel, and the electrophoresis was performed in the cold room at 80 V until the penetration of the proteins in the separating gel. After that, the cathode buffer, containing 0.02% of Serva Blue G, was replaced by the same but containing 0.002% of Serva Blue G, and the electrophoresis was continued at 150 V until the dye went out. When the electrophoresis was finished the gel was washed several times in 2 mM Tris-HCl (pH 7.5) and subjected to immunoblotting analysis, using the anti-ECONuoM-1 antibody. For the NADH dehydrogenase activity staining of NDH-1 the gels were incubated in 2 mM Tris-HCl (pH 7.5) containing 2.5 mg/ml of p-nitroblue tetrazolium and 150 M NADH for 2 h at 37°C. The reaction was stopped with 7% acetic acid.
Activity Analyses-It is known that NDH-1 can use dNADH as well as NADH as a substrate, allowing it to be distinguished from the NDH-2 (36). On the basis of this, we have used dNADH as substrate in the enzymatic assays. All the measurements were conducted at 30°C using a SLM DW-2000 spectrophotometer. The dNADH oxidase assays were performed spectrophotometrically with 80 g of protein/ml of membrane samples in 10 mM potassium phosphate buffer (pH 7.0) containing 1 mM EDTA. The reaction was started by the addition of 0.15 mM dNADH, and the measurements were followed at a wavelength of 340 nm as described previously (37). 10 M cap-40 was added for the complete inhibition of the reaction (38). The dNADH-DB reductase assays were performed in the same conditions, except that the reaction buffer contained 10 mM KCN and 0.1 mM of DB. The dNADH-K 3 Fe(CN) 6 reductase assays were performed in the same conditions as the dNADH-DB reductase assays, but DB was replaced by 1 mM K 3 Fe(CN) 6 , and the measurements were monitored at 420 nm. 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 . The extent of membrane potentials were assayed with 0.33 mg of protein/ml of membrane samples in 50 mM MOPS (pH 7.3), 10 mM MgCl 2 , 50 mM KCl, and 2 M oxonol VI as described previously (20,39). The reaction was started by addition of 0.2 mM dNADH and the absorbance changes at 630 -603 nm were recorded. 2 M of the proton ionophore FCCP was used as the uncoupler. Proton pump activity was detected by ACMA fluorescence quenching as described by Amarneh and Vik (40). 0.2 mM dNADH was used as the respiratory substrate. The production of superoxide radicals were measured at 550 nm with membrane samples at 0.24 mg of protein/ml in 10 mM potassium phosphate buffer (pH 7) containing 5 M acetylated cytochrome c as a reporter (41). The reactions were started by addition of 0.15 mM dNADH. In parallel, reactions included 15 units of superoxide dismutase were performed as control. The extinction coefficient of 21 mM Ϫ1 cm Ϫ1 was used for reduction of acetylated cytochrome c.
Other Analytical Procedures-Protein concentrations were determined by BCA protein assay kit (Pierce) with bovine serum albumin as standard according to the manufacturer's instructions. Any variations from the procedures and details are described in the figure legends.

Sequence Analysis of the NuoM Subunit-The E. coli
NuoM subunit is a polypeptide of 509 amino acid residues. Fig. 2A shows the amino acid sequence alignment of this subunit in various organisms. Our attempt to determine the topology of the E. coli NuoM subunit using anti-ECONuoM-1 and anti-ECONuoM-7 antibodies was unsuccessful. Therefore, a topology of the E. coli NuoM subunit was proposed by most reasonable predictions offered by a composite of computer programs including TopPredII (42,43), PHDhtm (44), TMHMM (45), SOSUI (46 -48) Tmpred (49), and HMMTOP (50). The final model also accommodated experimentally determined topology of Rhodobacter capsulatus NuoL subunit which exhibits amino acid sequence homologies to the NuoM subunit (51). According to this proposed model, the E. coli NuoM subunits contains 14 transmembrane segments (designated TM1-14 from the N terminus), with the N-and C-terminal regions facing the periplasmic sides of the membrane (Fig. 2B). The number of TM of the E. coli NuoM subunit seems to be consistent with the results from EM analyses of the membrane domain of the E. coli NDH-1 at 8-Å resolution (9).
As described before, there are at least two functional roles assigned for the membrane domain. Those are proton translocation and Q reduction. It is recognized that conserved residues located in the TM and with charge are involved in proton translocation of the energy transducing enzyme complexes exemplified by ATP synthases (52). Therefore, we searched for qualified amino acid residues. From Fig. 2 homologues. It should be noted that these two residues are also conserved in the NuoL and NuoN subunits of NDH-1/complex I, the MrpA and MrpD subunits of multisubunit cation antiporter complexes and the EchA subunit of energy-converting NiFe hydrogenase (see supplemental Fig. S1).
On the basis of the three-dimensional structures surrounding the Q A and Q B sites of bacterial photoreaction center and Q i site of bc 1 complex, Fisher and Rich (53) proposed that the triad sequence motif (aliphatic-(X) 3 -H-(X) 2/3 -(L/T/S)) may form a part of the Q headgroup pocket. In accordance with this sequence motif of Q-binding, His 241 , His 322 , and His 348 might be considered as candidates. Near His 241 , there are Trp 243 and Pro 245 . Because these residues are also conserved, we mutated these residues. In addition, it has been reported that a photoaffinity Q derivative labeled the fragment between Val 184 and Asn 206 of E. coli NDH-1 NuoM (22). Because the fragment houses the Q-binding triad motif (Leu 192 -XXX-His 196 -XXX-Thr 200 ) (53), the authors suggest that subunit NuoM bears the FIGURE 2. A, similarity of the deduced amino acid sequences among the E. coli NuoM subunit and its homologues from various organisms. The alignment was carried out by the ClustalW program (71). Black boxes with white letters show perfectly conserved residues among the listed organisms. Dark gray boxes with white letters illustrate conserved residues among at least four organisms. Gray boxes with black letters indicate similar residues among at least four organisms. Amino acids mutated in this study are marked by arrows.  (22). Although it is recognized that His 196 is not a conserved residue, this His was also listed up as the candidate for Q-binding.
In a large number of LHON families, predisposition to the disease is tightly associated with the point mutation G11778A in human mtDNA (54). This mutation results in the replacement of Arg 340 by His (human ND4 numbering). Sequence comparison between the human ND4 subunit and homologues of many species has shown that the ND4 Arg 340 residue is conserved in many NDH-1/complex I and corresponded to the Arg 369 residue in E. coli NuoM (see Fig. 2). The function of Arg 369 has been also examined. In addition, highly conserved Arg 365 located nearby Arg 369 was explored.
Subunit Assembly of NDH-1 in NuoM Mutations-To investigate whether mutation of the NuoM subunit affects subunit assembly of NDH-1, we first analyzed mutant E. coli membranes on SDS-PAGE by immunoblotting using subunit-specific antibodies (Fig. 3). Antibodies used were raised against six peripheral subunits (NuoB, NuoCD, NuoE, NuoF, NuoG, and NuoI) and two oligopeptides of NuoM (ECONuoM-1 and ECONuoM-7). The NuoM-KO mutant, as expected, lacked NuoM. It also showed diminished assembly of peripheral subunits, possibly due to incomplete assembly of the whole NDH-1. In the revertant mutant (KO-Rev), all peripheral subunits and NuoM existed at the same levels as the wild-type. Similarly, in all constructed point mutants, no significant quantitative difference was observed for the peripheral subunits and furthermore the NuoM subunits of those mutants exhibited the same mobility as the wild-type. The data suggest that these mutations did not result in unstable or truncated NuoM subunit. In contrast, the D135A mutant appeared to show considerably lower amounts of NuoM in the membranes.
Another procedure to investigate effects of NuoM mutants on subunit assembly is to utilize BN-PAGE. In Fig. 4, the assem-bled NDH-1 on a BN-PAGE gel of our samples was visualized by using anti-ECONuoM-1 antibodies and by NADH dehydrogenase activity staining. As expected, the NuoM-KO mutant apparently lacked a fully assembled NDH-1. The D135A mutation partially decreased the amount of the assembled NDH-1 and NADH dehydrogenase activity staining. Other point mutants were akin to the wild-type. Once again, our results indicate that there is no drastic change that may alter the assembly status of NDH-1 in the point mutants.
Effects of NuoM Mutation on NDH-1 Activities-We measured electron transfer activities of NDH-1 by using membrane vesicles prepared from wild type and NuoM mutants (Table 1). In addition to NDH-1, E. coli also houses an alternative NADH-Q oxidoreductase (NDH-2) lacking an energy coupling site (19 -21). The E. coli NDH-2 is insensitive to specific inhibitors for NDH-1 such as capsaicin (38,55) and cannot use dNADH as a substrate (36). NADH-driven respiration of the wild-type membranes (MC4100 strain) was 65% sensitive to capsaicin, indicating that ϳ35% of the NADH oxidase activity was ascribed to NDH-2 (data not shown). To measure the activities derived solely from NDH-1, we used dNADH as the substrate. We conducted three types of assays (dNADH oxidase activity; dNADH-DB reductase activity; dNADH dehydrogenase activity) to assess the effects of the mutations.
The NuoM-KO mutant exhibited almost null activity of the energy-coupled NDH-1 activities (dNADH oxidase and dNADH-DB reductase) similar to the NuoA-KO (19), NuoJ-KO (20), and NuoK-KO (9) mutants. In contrast, the NuoM-KO mutant retained considerable activity of dNADH-K 3 Fe(CN) 6 reductase (ϳ35% of parent membranes), which were higher than the other hydrophobic subunits-KO mutants (ϳ10% of the parent membranes). As shown in Fig. 3, the decrease of the E and F subunits (flavoprotein subcomplex) in the membranes is in agreement with loss of the NADH dehydrogenase activity. Recently, it has been reported that NuoA, NuoJ, and NuoK may directly interact with the peripheral arm, whereas NuoM and NuoL are distantly located from the peripheral arm (9, 14, 56). Therefore, it seems reasonable that effects of NuoA-, NuoJ-,  The dodecyl-␤-maltoside concentration was 0.8%. The electrophoresis was performed as described under "Experimental Procedures." For NADH dehydrogenase activity staining, the gels were incubated with 2.5 mg/ml p-nitroblue tetrazolium and 150 M NADH for 2 h at 37°C. The reaction was terminated by 7% acetic acid. For immunoblotting, after BN-PAGE, the E. coli membrane proteins were electrotransferred to nitrocellulose membranes. Subsequently, the membranes were immunostained with the affinity-purified NuoM-1 antibody.
and NuoK-KO mutants are more drastic than those of NuoM-KO mutant.
As shown in Table 1, the KO-Rev mutant displayed properties comparable to the wild-type strain in all enzymatic activities tested. This proved that the procedure used to introduce point mutations in the chromosomal nuoM gene worked out as described in NuoJ and NuoK mutations (19,21).
The point mutants constructed displayed various degrees of inhibition of dNADH oxidase activity and dNADH-DB reductase activity. The two activities behaved in a similar manner among the mutants tested, implying that the observed effects are due to the NDH-1 mutations. Mutation of the highly conserved Glu 144 in the TM had the greatest impact, with almost total abolishment of activities with E144A and E144Q. Interestingly, E144D mutation scarcely affected the energy-transducing NDH-1 activity. In addition, mutations of the highly conserved Lys 234 and Lys 265 in the TM significantly reduced activities with K234A (ϳ90% inhibition) and K265A (ϳ70% inhibition). We also investigated the four histidine residues (His 196 , His 241 , His 322 , and His 348 ) proposed to be involved in the catalytic Q-binding. No mutants showed detrimental effects on the activities. In addition, the K m values for DB of these mutants were determined to be in the range of 2.2-2.7 M, which is almost the same as that for wild-type (2.5 M). Mutation of R369H (equivalent to 11778 mutation found in mtDNA of LHON patients) suppressed the activities to 70% of the wild-type. These values seem to be reasonably in good agreement with the results that NADH-linked respiration in mitochondria of cells carrying 11778 mutation was 38% lower than those of control cells (57). D135A mutation decreased activities to 40% of the wild-type. However, D135N mutation slightly suppressed the energy-coupled activities to ϳ80% of the wild type. Therefore, it is unlikely that Asp 135 is essential for activities. The data suggest that other highly conserved residues Asp 84 , Trp 243 , Pro 245 , Arg 273 , Arg 365 , or Tyr 435 are not essential for activity.
It is generally recognized that cap-40 acts as a competitive inhibitor for Q in NDH-1/complex I and suppresses only energy-coupled activities (38) and that the binding site of this inhibitor is located within the membrane domain of NDH-1/complex I. In our preparations from the wild-type strain, cap-40 inhibited the dNADH-dependent respiration by 97% (21). We then investigated the effects of NuoM mutation on the inhibitory characteristics of cap-40. As shown in Table 1, there was no significant difference in the IC 50 values of cap-40 between the wild-type and the mutants, suggesting that the cap-40 binding site is not modified by these point mutations.
Measurements of Electrochemical Potential of NuoM Mutants-Membrane potential of mutant membrane vesicles was monitored by following the absorbance change of oxonol VI. As shown in Fig. 5, addition of dNADH to the membrane vesicles from the wild-type strain increased the absorbance indicating generation of ⌬⌿ (inside positive) which was then dissipated by FCCP. The signal was enhanced in the presence of nigericin and totally abolished by the presence of both nigericin and valinomycin. Preincubation of the membrane vesicles with cap-40 also prevented the ⌬⌿ formation (data not shown). As expected, no ⌬⌿ was observed in the membranes of the KO mutant, and the membrane vesicles of KO-rev mutant generated a signal comparable to that of the wild-type strain. Mutation of highly conserved Glu 144 to Ala or Gln, resulted in a complete loss of the ability of the enzyme to generate ⌬⌿. In contrast, mutant E144D exhibited ⌬⌿ comparable to the wildtype. These results are consistent with the data of the energytransducing electron transfer activities of NDH-1. In the case of another highly conserved Lys 234 , mutation to Ala caused a significant reduction in ⌬⌿.
The formation of proton gradient was monitored by using ACMA (see Fig. 6). Membrane vesicles from wild-type and the KO-rev mutant exhibited an almost identical response. The signal was completely reversed by FCCP. No proton translocation was observed in NuoM-KO mutant. Mutant E144A or E144Q did not show any sign of proton translocation. As expected, Mutant E144D showed proton translocation comparable to the wild type. K234A mutant displayed significantly slow proton translocation and did not reach stationary phase in the range of experimental time.
Generation of Superoxide Radicals by NuoM Mutants- Fig. 7 illustrates production of superoxide radicals by the NuoM mutant membranes. It is obvious that E144A, E144Q, and K234A mutants produced significant amounts of superoxide radicals induced by dNADH. Cap-40 also stimulated the production of superoxide radicals. In contrast, only small amounts of superoxide radicals were detected in the wild-type, E144D, K256A, and R369H mutants. These results suggest that the mutants with severe inhibition of the NDH-1 activities (Ͼ90% inhibition) could generate superoxide radicals. Similar mechanism has been observed in overproduction of superoxide radicals by complex I inhibition in mammalian mitochondria (58,59). Mutant R369H corresponding to the human 11778 mtDNA mutation associated with LHON barely generated superoxide radicals.

DISCUSSION
Membrane-bound enzymes involved in proton or ion translocation tend to contain highly conserved, charged amino acids in the hydrophobic environment of the protein. The Asp (or Glu) residue in the DCCD-binding subunit of ATP synthase is probably one of the best known examples of such highly pre-served carboxyl group essential for proton translocation and also is well characterized mechanistically (60). Membrane domain subunits of NDH-1 have a number of conserved residues with charged groups in the predicted membrane spanning regions. Site-directed mutation studies in our group and other   laboratories revealed several candidate residues that may be part of the proton translocation mechanism. In these cases, mutating a residue resulted in a partial decrease of coupled activities to a varying extent. However, there was only one residue, Glu 36 in NuoK (ND4L), whose elimination of its carboxyl residue lead to a complete loss of coupled activities (21,61). In the present study, we have shown that Glu 144 in subunit NuoM (ND4) is also absolutely essential for the energy-coupled NDH-1 activities. This is only the second case of complete elimination of coupled activity by a single mutation. A sequence comparison (Supplemental Fig. S1) indicates that a residue equivalent to Glu 144 of NuoM is also conserved in NuoL and NuoN, which may not be surprising because the three subunits are proposed to diverge from a single common ancestor (51). Interestingly, the same Glu residue is also conserved in MrpA and MrpD of the multisubunit cation antiporters and EchA of the energy-converting NiFe hydrogenases. Given the fact that a common function among all these enzymes is proton translocation, it is highly likely that the carboxyl group of Glu 144 in NuoM is directly involved in the mechanism of proton pump of NDH-1. This hypothesis is also supported by the fact that NDH-1/complex I are inhibited by amiloride derivatives which act as specific inhibitors to Na ϩ /H ϩ antiporters (62,63). It is possible, however, that the exact role of Glu 144 in NuoM might not be the same as that of Glu 36 in NuoK. First, there is no counterpart of NuoK/ND4L in the multisubunit cation antiporters or energy-converting NiFe hydrogenases. Second, NuoK seems to be close to the proposed Q site of the peripheral segment, but the center of NuoM is believed to be about 60 Å away from it (9). The presence of a carboxyl residue participating in the proton translocation distantly located from the peripheral domain where the electron transfer takes place would be supportive evidence for indirect energy coupling mechanism in NDH-1/complex I.
In addition to carboxyl residues, positively charged residues in TM may also be involved in proton channel as has been reported for a Lys in certain respiratory enzyme complexes (64) and an Arg in the ATP synthase (65). We investigated six highly conserved, positively charged residues predicted to be in the TM regions of E. coli NuoM. Mutations, K234A and K265A, had the greatest impact with the former resulting in nearly total abolishment of activities with K234A (ϳ90% inhibition). Lys 234 is highly conserved in both NuoL and NuoN, whereas Lys 265 is conserved in NuoN but not in NuoL. In addition, Lys 234 , but not Lys 265 , is also conserved in MrpA, MrpD, and EchA subunits (see Supplemental Fig. S1). It seems likely that almost perfectly conserved Lys 234 residue in TM of NuoM participates in the proton translocation in NDH-1/complex I similar to Glu 144 . Recently, Vik's group (40) reported that in the E. coli NuoN subunit the E133A mutant (equivalent to Glu 144 in NuoM) showed only a 30% loss of dNADH oxidase activity and the K217C mutant (corresponding to Lys 234 in NuoM) showed almost null activity. The data together with those in this study suggest that the functional role of the three homologous subunits (NuoL, M, and N) might be different.
The presence of an essential histidine residue in the Q-binding site has been reported. For example, mutations of His 217 of cytochrome b in R. capsulatus bc 1 complex or His 98 of subunit I in E. coli bo 3 complex were shown to have major effects on the activities and properties of the respective enzyme (66,67). In the case of E. coli NuoM, photoaffinity labeling experiments suggested a location of the Q-binding site to be between Val 184 and Asn 206 (22). This sequence region contains one histidine, His 196 that is part of the proposed Q-binding motif (LX 3 HX 3 (T/ S)). Independently, Fisher and Rich (53) proposed that highly conserved His 319 of human ND4 (His 348 of E. coli NuoM) is the Q-binding site on the basis of Q-binding motif. Our present study indicates that the mutations of these two histidine residues (His 196 and His 348 ) as well as two highly conserved histidines (His 241 and His 322 ) did not give rise to any significant alterations in the characteristics of the enzyme including energy-transducing activities, K m values for Q and IC 50 for cap-40. Therefore, it is highly unlikely that these His residues are involved in the catalytic Q binding.
It is known that a homoplasmic mutation R340H in ND4 (the G to A substitution at position 11778 of human mtDNA) is associated with LHON. Hofhaus et al. (57) and Park et al. (68) reported that NADH dehydrogenase-dependent respiration, as measured in digitonin-permeabilized cells, was specifically decreased by ϳ40% in the R340H mutant cells constructed by using cybrid systems. As described above, The ND4 Arg 340 is conserved in the many organisms including E. coli (Arg 369 NuoM). To investigate whether the E. coli NDH-1 is a useful model of complex I, we constructed mutant R369H and examined its properties. The energy-coupled NDH-1 activities of mutant R369H membranes were suppressed to 60 -70%. In contrast, Lunardi et al. (54) constructed R368H (equivalent to human R340H) mutants of NuoM using R. capsulatus chromosomal gene manipulation technique and investigated the properties of this mutant. This mutant showed a clear impairment in oxidative phosphorylation capacity. However, no significant reduction of NADH-dependent respiration and its coupled proton uptake was observed in isolated membranes. This difference between R. capsulatus and E. coli membranes may be due to the complicated electron transfer system of R. capsulatus or the fact that only the NADH oxidase activity was measured in R. capsulatus. By using the E. coli system, it is possible to carry out comprehensive studies concerning effects by mutations inducing LHON.
It is recognized that all seven ND subunits are encoded by mtDNA. The only known exception is ND3 and ND4L subunits of Chlamydomonas reinhardtii (69). At present, systematic studies of mtDNA-encoded subunits are technically difficult although it might become possible in a future to carry out site-directed mutagenesis of mitochondrially encoded subunits (70). In fact, to date mutants of mtDNA which have been investigated were limited (e.g. mutated mtDNA from patients with mitochondrial diseases). We have shown that E. coli NDH-1 works as a useful model of complex I by using systematic mutants of membrane domain subunits NuoA (ND3), NuoJ (ND6), NuoK (ND4L), and NuoM (ND4). To solve energy-transducing mechanism of NDH-1/complex I, it is prerequisite to thoroughly study properties of hydrophobic domain subunits. For this purpose, E. coli NDH-1 looks a suitable system.