Characterization of the Iron-Sulfur Cluster N7 (N1c) in the Subunit NuoG of the Proton-translocating NADH-quinone Oxidoreductase from Escherichia coli*

The proton-pumping NADH-quinone oxidoreductase from Escherichia coli houses nine iron-sulfur clusters, eight of which are found in its mitochondrial counterpart, complex I. The extra putative iron-sulfur cluster binding site with a C XX C XXX C X 27 C motif in the NuoG subunit has been assigned to ligate a [2Fe-2S] (N1c). However, we have shown previously that the Thermus thermophilus N1c fragment containing this motif ligates a [4Fe-4S] (Nakamaru-Ogiso, E., Yano, T., Ohnishi, T., and Yagi, T. (2002) J. Biol. Chem . 277, 1680–1688). In the current study, we individually inactivated four sets of the iron-sulfur binding motifs in the E. coli NuoG subunit by replacing all four ligands with Ala. Each mutant subunit, designated (cid:1) N1b, (cid:1) N1c, (cid:1) N4, and (cid:1) N5, was expressed as maltose-binding protein fusion proteins. After in vitro reconstitution, all mutant subunits were characterized MBP-NuoG m M In Vitro Reconstitution the Iron-Sulfur Clusters into the MBP-NuoG Subunits— anaerobic purified MBP-NuoG Han et al. (31). Other Analytical Procedures— UV-visible absorption spectra were recorded on a Beckman DU640 spectrophotometer at room temperature. Protein estimation was routinely done by the method of Lowry et al. (32) and Bradford (33). SDS-PAGE and blue native PAGE were carried out according to Laemmli (34) and Scha¨gger (35), respectively. Immunoblotting was conducted as described previously (36). Non-heme iron and acid-labile sulfide were determined according to Refs. 37 and 38, respectively. Any variations from the procedures and other details are described in the figure legends.

Interestingly, in some organisms such as E. coli and T. thermophilus, the NDH-1 contains an additional iron-sulfur cluster binding motif ( 230 CXXCXXXCX 27 C 265 , E. coli numbering) in the subunit NuoG for E. coli and Nqo3 for T. thermophilus (12,13,20). This motif is not found in either P. denitrificans NDH-1 or mitochondrial complex I (Fig. 1). It has been proposed that this additional unique cysteine sequence motif is responsible for the EPR signals from the binuclear cluster N1c of the E. coli NDH-1 (21)(22)(23). In our previous report, however, we have shown that the Thermus N1c fragment containing this motif ligates a [4Fe-4S] cluster, not a [2Fe-2S] cluster, by UV-visible and EPR spectroscopic analyses (12).
In the present study, to characterize the N1c cluster at the subunit level and distinguish it from three other iron-sulfur clusters (N1b, N4, and N5 accommodated in the NuoG subunit ( Fig. 1)), we have individually inactivated four iron-sulfur cluster binding sites by substituting Ala for all four conserved Cys (or His) residues. We overexpressed wild type and the individual iron-sulfur cluster mutant NuoG subunits, designated ⌬N1b, ⌬N1c, ⌬N4, and ⌬N5, as maltose-binding protein (MBP) fusion proteins in E. coli. We compared EPR spectra of the individual mutant NuoG subunits and identified the [4Fe-4S] EPR signals, which could not be assigned to either N4 or N5 cluster. Based on our present analyses and the fact that the motif CXXCXXXCX 27 C coordinates only [4Fe-4S] clusters in other known enzymes such as the periplasmic nitrate reductase in P. denitrificans (24), we concluded that the N1c binding motif undoubtedly ligates a [4Fe-4S]. Therefore, we propose to change its misleading name, N1c, to N7. In addition, we attempted to see the effects of each cluster mutation on the assembly of the peripheral subunits by complementing the nuoG knock-out E. coli mutant with mutated plasmids.

EXPERIMENTAL PROCEDURES
Materials-The pCRScript Cloning kit was from Stratagene (La Jolla, CA). The expression vector pMALc2g, amylose resin, and anti-MBP serum were from New England Biolabs (Beverly, MA). Materials for PCR product purification, gel extraction, and plasmid preparation were obtained from Qiagen (Valencia, CA). The BCA protein assay kit and SuperSignal West Pico chemiluminescent substrate were from Pierce. E. coli strain BL21(DE3) was purchased from Novagen. The gene replacement vector pKO3 was a generous gift from Dr. George M. Church (Harvard Medical School). The E. coli strain MC4100 was a gift from Dr. Mutsuo Yamaguchi (The Scripps Research Institute). All chemicals used were of the highest grade available from Sigma.
Expression and Purification of the NuoG Subunits-Expression of the NuoG subunits with pET20b/nuoG (with or without a His tag) and pMALc2g/nuoG vectors was conducted basically according to Nakamaru-Ogiso et al. (12). Briefly, competent E. coli strain BL21(DE3) cells were transformed with expression vectors, and the cells were grown overnight at 30 -33°C until A 600 nm reached ϳ2.0. After the cells were cooled down to 20°C followed by the addition of 0.4 mM isopropyl 1-thio-␤-D-galactopyranoside, the cells were further incubated for 3-5 h. The cells harvested were suspended to 7.5% (w/v) in 50 mM Tris-HCl buffer (pH 8.0) containing 1 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, and 2 mM dithiothreitol, which had been degassed and purged with oxygen-free argon on ice before use. The cells were broken up by two passages through a French pressure cell (Spectronic Instruments, Rochester, NY) at 15,000 p.s.i. After unbroken cells and inclusion bodies were removed by centrifugation, the supernatant was ultracentrifuged at 50,000 rpm for 60 min in a 70Ti rotor. The supernatant and pellet were analyzed for expression levels of each form of the NuoG subunit. The non-fused and His-tagged NuoG proteins were predominantly expressed as inclusion bodies with dark reddish brown color under all conditions we tested. In contrast, the MBP-fused proteins were successfully expressed in the soluble fraction. Therefore, the MBP-fused NuoG (MBP-NuoG) subunit was purified under anaerobic conditions (oxygen level Ͻ1 ppm) at room temperature basically according to the manufacturer's protocols. All solutions were degassed and purged with oxygen-free argon and equilibrated in the chamber overnight. The supernatant containing the MBP-NuoG protein was loaded onto an amylose column (1.5 ϫ 5.0 cm) equilibrated with the column buffer, 50 mM Tris-HCl buffer (pH 8.0) containing 10 mM ␤-mercaptoethanol, 200 mM NaCl, 0.5 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride. The column was washed with 60 ml of the same buffer, and then the MBP-NuoG protein was eluted with the same buffer containing 10 mM maltose and 0.2 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride.
In Vitro Reconstitution of the Iron-Sulfur Clusters into the MBP-NuoG Subunits-The reconstitution procedures were carried out in the anaerobic chamber by a modified method of Yano et al. (10,16). The purified MBP-NuoG subunit was diluted to 10 -20 M with the column buffer above containing 10% (v/v) glycerol. ␤-Mercaptoethanol was added to the protein solution at 1.0% (v/v), and the solution was gently mixed and left for 90 min. Fe(NH 4 ) 2 (SO 4 ) 2 and Na 2 S were added to the solution at a final concentration of ϳ500 M. The reconstitution proceeded for 2-3 h. Excess Fe 2ϩ and S 2Ϫ were removed by a desalting column (Bio-Rad, 10-DG). The reconstituted MBP-NuoG protein was applied onto a DEAE (Toyopearl 650 M) column (0.8 ϫ 1.0 cm) equilibrated with 50 mM Tris-HCl buffer (pH 8.0) containing 1 mM dithiothreitol and 10% (v/v) glycerol. The protein was eluted with 500 mM NaCl.
Cloning of the E. coli nuoG Gene and Construction of the nuoG Knock-out Strain [⌬(nuoG::Spc)]-The gene encoding the NuoG subunit, together with an ϳ210-bp DNA segment upstream and an ϳ220-bp DNA segment downstream, was cloned by PCR from genomic DNA-purified E. coli DH5␣. We used intrinsic sites SmaI for the sense primer, 5Ј-CTTAGGCCCGGGTAAAACTTTCTGTGCCCACGC-3Ј, and BglII or BamHI for the antisense primers, 5Ј-CGCCCAGATCTGTAC-CGTCAGCGTAATC-3Ј or 5Ј-CCGAGAATTTCGGGATCCAGTCTTCT-TTAAAGAAC-3Ј, respectively (the italicized bases represent the restriction site sequence). The spectinomycin (spc) gene cassette from the Staphylococcus aureus transposon Tn554 (25) was cloned by PCR. The DNA fragments were assembled together with the spc cassette by using the intrinsic BglII site in pCRScript and finally transferred to pKO3. The resulting construct was designated pKO3 (nuoG-spc). The E. coli strain MC4100 was transformed with the pKO3 (nuoG-spc) plasmid, and homologous recombination was carried out as described in Link et al. (26). The presence of the spc cassette and its location in the genomic nuoG gene were verified by PCR and DNA sequencing.
Membrane Preparation-The cell pellet was resuspended at 10% (w/v) in a buffer containing 10 mM Tris-HCl (pH 7.4), 1 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 15% (w/v) glycerol. The cell suspension was then passed once in a French press at 20,000 p.s.i. and centrifuged again in the GSA rotor at 12,000 rpm for 10 min. Cell debris was discarded, and the supernatant was then ultracentrifuged in a 70Ti rotor at 50,000 rpm for 30 min. The pellet was resuspended in the same buffer and was used immediately for various analyses.
EPR Spectroscopy-EPR samples were prepared in an anaerobic chamber, and the redox mediators, methyl viologen and benzyl viologen, were added at 5 M each. Anaerobically prepared neutralized sodium dithionite solution (0.5 M) was added at 10 mM of a final concentration. The protein solution was transferred to an EPR tube, and after 5-10 min of incubation at room temperature, the samples were quickly frozen in dry ice/ethanol mixture and then stored in liquid nitrogen until EPR analyses. EPR spectra were recorded by a Bruker ESP 300E spectrometer at X-band (9.4 GHz) using an Oxford Instrument ESR900 helium flow cryostat, a Hewlett Packard 5350B microwave frequency counter, and an ITC4 temperature controller to control sample temperatures. EPR spectra of the iron-sulfur clusters were simulated by SimFonia (Bruker). Spin quantitations were carried out under non-saturating conditions using 0.5 mM Cu(II)EDTA or 0.5 mM Cu(II) perchlorate as standards according to Aasa et al. (27). Power saturation data were analyzed by a computer fitting method (28,29).
Antibody Production-For raising antibodies specific toward E. coli subunits, we cloned the DNA fragments encoding the individual subunits by PCR and prepared inclusion bodies of each subunit as described previously (30). The antibodies were affinity-purified according to Han et al. (31).
Other Analytical Procedures-UV-visible absorption spectra were recorded on a Beckman DU640 spectrophotometer at room temperature. Protein estimation was routinely done by the method of Lowry et al. (32) and Bradford (33). SDS-PAGE and blue native PAGE were carried out according to Laemmli (34) and Schä gger (35), respectively. Immunoblotting was conducted as described previously (36). Non-heme iron and acid-labile sulfide were determined according to Refs. 37 and 38, respectively. Any variations from the procedures and other details are described in the figure legends.

Expression, Isolation, and Reconstitution of the MBP-NuoG
Subunit and Its Iron-Sulfur Cluster Mutants-The purified wild type, MBP-NuoG(WT), was reddish-brown and exhibited absorption spectra as shown in Fig. 2A, which were similar to those of its homologue, the MBP-fused T. thermophilus Nqo3 subunit described by Nakamaru-Ogiso et al. (12). The purified MBP-NuoG(WT) contained only 1.9 mol of non-heme Fe and 1.4 mol of S 2Ϫ /mol of protein, indicating that the iron-sulfur cluster contents were much lower than expected amounts (12 or 14 mol of each Fe and S 2Ϫ /mol of protein). However, the contents of non-heme iron and acid-labile sulfide in MBP-NuoG(WT) were significantly improved upon the in vitro reconstitution to 7.9 mol of Fe and 5.9 mol of S 2Ϫ /mol of protein on average. The absorption spectrum of the reconstituted MBP-NuoG(WT) exhibited a broad shoulder around 380 -420 nm with a tailing reaching 700 nm. The absorption spectrum did not show peaks at 320, 460, and 560 nm, which are prominent for [2Fe-2S] clusters, suggesting that the reconstitution of a binuclear cluster(s) in the NuoG subunit was not successful (Fig. 2B). UVvisible spectra of the reconstituted MBP-NuoG mutants showed slightly different spectral features (Fig. 2, B and C). A broad absorption peak around 420 nm, which is characteristic of [4Fe-4S] 2ϩ clusters, was lower than that of wild type. The absorbance was roughly proportional to the iron-sulfur cluster contents in the reconstituted subunits, most likely because of the inactivation of the putative iron-sulfur cluster binding motifs (Fig. 2, B and C).
EPR Characterization of the Iron-Sulfur Clusters in the MBP-NuoG Mutants-To characterize the bound iron-sulfur clusters in the reconstituted MBP-NuoG mutants, we performed EPR analyses. First, we examined EPR spectra of the dithionite-reduced samples at the g ϭ 2 region at 40 -60 K. We did not detect any EPR signals in all preparations including MBP-NuoG(WT), indicating that the reconstituted subunits did not contain the binuclear cluster N1b (results not shown). When EPR spectra were measured at 6 -30 K, EPR signals arising from [4Fe-4S] 1ϩ clusters were detected, and the individual samples exhibited different spectra. As shown in Fig. 3, MBP-NuoG(WT) displayed multiple EPR resonances with g values of 1.89, 1.91, 1.94, and 2.05 at 12 K and 5 mW. At 6 K, EPR signals at g ϭ 1.91 and g ϭ 2.05 became smaller and broader because of the saturation, whereas the signal at g ϭ 1.89 and a broad resonance at g ϭ 2.06 became more pronounced. Further inspection of the EPR spectra at 4 -50 K and 5 mW showed that the g ϭ 1.91 and g ϭ 2.05 resonances were best seen at 14 -16 K, whereas the EPR signals at g ϭ 1.89 and g ϭ 2.06 showed the optimal temperature at ϳ10 K. These observations indicate that MBP-NuoG(WT) contains at least two distinct [4Fe-4S] clusters. The mutants, MBP-NuoG⌬N5 (Fig. 3) and MBP-NuoG⌬N1b (data not shown), exhibited almost the same EPR spectra as those of MBP-NuoG(WT) under these conditions except that slight g value shifts were noticed. When MBP-NuoG⌬N5 was examined at 4 K and various microwave power levels, the g ϭ 1.91 resonance was readily saturated at low power levels (P1 ⁄2 ϳ0.02 mW at 4 K), whereas the signals at g ϭ 1.89 and g ϭ 2.06 were saturated at higher power levels (P1 ⁄2 ϳ1.0 mW at 4 K). In contrast, MBP-NuoG⌬N4 showed different EPR spectra. At 12 K, it showed the EPR signal with g x,y,z ϭ 1.91, 1.94, and 2.05. However, the EPR signals at g ϭ 1.89 and g ϭ 2.06 were absent. At 6 K, the EPR signal was partially saturated, and the line width became broader. The behaviors of this EPR signal were similar to the corresponding EPR signals observed in wild type, MBP-NuoG⌬N1b, and MBP-NuoG⌬N5. From these observations, it was concluded that the EPR signal with g x,y,z ϭ 1.89, 1.94, and 2.06 arose from the [4Fe-4S] cluster coordinated in the N4 binding site.
MBP-NuoG⌬N1c exhibited very small and broad EPR signals at both 6 and 12 K, which were completely different from the EPR signals arising from the two [4Fe-4S] clusters detected in MBP-NuoG(WT) (Fig. 3). The contents of non-heme iron and acid-labile sulfide in this mutant after in vitro reconstitution were the lowest among all of the preparations, indicating that the inactivation of the N1c cluster-binding motif severely affected in vitro incorporation of the iron-sulfur cluster into this protein. Because the EPR signal with g x,y,z ϭ 1.91, 1.94, and 2.05 was detected in wild type and all other mutant subunits To resolve the EPR spectra of these two [4Fe-4S] clusters, we attempted to simulate the EPR spectra of MBP-NuoG⌬N5 that contained only two [4Fe-4S], one each in the N4 and N1c binding sites. For the first approximation, we retrieved EPR parameters for the [4Fe-4S] cluster coordinated in the N1c binding site from the EPR spectrum of MBP-NuoG⌬N4 measured at 14 K and 5 mW. Under this condition, neither of the EPR signals of the two [4Fe-4S] clusters was saturated, and almost no contribution from other iron-sulfur clusters can be assumed. Then, the second [4Fe-4S] cluster EPR signal was simulated. Fig. 4A shows the EPR spectra of MBP-NuoG⌬N5 using the following parameters for the two [4Fe-4S] clusters: for cluster N1c, g x,y,z ϭ 1.91, 1.94, and 2.05, L x,y,z ϭ 21, 12, and 12.5 gauss; and for cluster N4, g x,y,z ϭ 1.89, 1.94, and 2.06, L x,y,z ϭ 55, 38, and 65 gauss. The ratio of the clusters was estimated to be N1c:N4 ϭ 3:1. Using these parameters, we obtained a reasonably good fit for the EPR spectrum of MBP-NuoG(WT) at 14 K and 5 mW (Fig. 4B).
The EPR Signal Arising from Cluster N5-We attempted to resolve the EPR spectra of cluster N5 in MBP-NuoG proteins. Cluster N5 has been detected in bovine complex I (4), yeast Yarrowia lipolytica complex I (8), and the overexpressed Nqo3 subunit from P. denitrificans (39); however, the EPR signal of cluster N5 has not been reported for the E. coli NDH-1 thus far. In general, the spin relaxation of cluster N5 is so fast that the EPR signal is only detected at 4 -6 K. We examined EPR spectra of MBP-NuoG⌬N4 at 4 K and at 1 and 10 mW. It is apparent that the g ϭ 1.91 and g ϭ 2.05 resonances arising from the [4Fe-4S] cluster in the N1c binding motif were saturated at 10 mW (Fig. 5). At the same time, rather broad resonance became noticeable at the g ϭ 1.85-1.88 region as well as at g ϭ ϳ2.06, which seemed likely to be derived from a distinct [4Fe-4S] cluster with much faster spin relaxation properties. In the case of the P. denitrificans Nqo3 subunit, a portion of the electron spin of cluster N5 has been shown to take a S ϭ 3/2 ground state, giving rise to EPR signals at g ϭ ϳ5 (39). We also examined a lower magnetic field to search for EPR signals; however, we could not detect any EPR signal at the g ϭ ϳ5 region in any of NuoG mutants examined. Despite several attempts, we were not able to show conclusively that the EPR signals were arising from cluster N5 in the expressed subunits mainly because of the low concentration of the [4Fe-4S] cluster. However, the broad EPR signals shown in Fig. 5 somehow resemble those of cluster N5 in the overexpressed P. denitrifi-

FIG. 4. EPR spectra of the two iron-sulfur clusters detected in the reconstituted MBP-NuoG⌬N5 (A) and MBP-NuoG(WT) subunits (B) at 14 K and at 5 mW.
Solid lines are experimental EPR data. Dashed lines are the simulated EPR spectra of cluster N7 (N1c) with the parameters g x,y,z ϭ 1.91, 1.94, and 2.05 and L x,y,z ϭ 21, 12, and 12.5 gauss. Dotted lines are the simulated EPR spectra of cluster N4 with the parameters g x,y,z ϭ 1.89, 1.94, and 2.06 and L x,y,z ϭ 55, 38, and 65 gauss. The remaining EPR spectra after the simulation were shown in the lower panels. Other EPR conditions are the same as in Fig. 3. cans Nqo3 subunit (39) and of complex I preparations (4,8). These results may suggest that the reconstituted MBP-NuoG subunits contain cluster N5. Fig. 6A, the anti-E. coli NuoG antibody reacted with a single band of ϳ95 kDa in the E. coli membranes of strains BL21(DE3) and MC4100, which is in good agreement with the molecular mass of 100.5 kDa deduced from its primary sequence. The antibody did not react with any proteins from other sources. The NuoG subunit was not detectable in the membranes of the nuoG knock-out mutant MC4100⌬nuoG (Fig. 6A). This strain did not exhibit any dNADH-ferricyanide reductase activity in its membrane preparations. We did not detect any peripheral subunits (NuoB, -CD, -E, or -F) in the cytoplasmic membrane from MC4100⌬nuoG except for the NuoCD subunit (Fig. 6B, lane 2). This subunit in the cytoplasmic membrane was present at about half the amount of that of wild type. We expressed MBP-NuoG(WT) in MC4100⌬nuoG strain in trans to see how the MBP-NuoG subunit could restore the assembly of the peripheral subunits. Despite the MBPtagged NuoG subunit, the partial assembly of the peripheral subunits was detected in the membranes from the trans-conjugant (Fig. 6B, lane 7). It should be noted that the expression of MBP-NuoG(WT) could not restore the NDH-1 activity. This finding prompted us to examine whether the inactivation of the putative iron-sulfur cluster binding sites in the NuoG subunit would affect the assembly of the peripheral subunits. We found that the expression of the MBP-NuoG mutants resulted in different subunit assembly (Fig. 6B, lanes 3-6). In the case of MBP-NuoG⌬N1b, the expression of the NuoB, NuoCD, and NuoF subunits was detected, whereas the NuoE subunit was missing. For the expression of MBP-NuoG⌬N1c and MBP-NuoG⌬N5, only NuoB and NuoCD were detected. Interestingly, in the case of MBP-NuoG⌬N4, the NuoCD subunit was absent. The amount of NuoB subunit was much less in the MBP-NuoG⌬N4 and MBP-NuoG⌬N5 subunits. These results suggest that the incorporation of the iron-sulfur clusters is important for the assembly of the peripheral subunits and that the individual iron-sulfur clusters are involved in the association with the other peripheral subunits in different manners. DISCUSSION In this study, we were able to detect at least two [4Fe-4S] clusters and identify their binding sites by EPR analyses. The EPR signal with g x,y,z ϭ 1.89, 1.94, and 2.06 was assigned to cluster N4 and was consistent with the results of the expressed P. denitrificans Nqo3 subunit (18). The EPR signal of cluster N4 was similar to that in the entire NDH-1 enzyme from E. coli with g values of g x,y,z ϭ 1.89, 1.93, and 2.09 (23) except for the position of the g z resonance. The second EPR signal with g x,y,z ϭ 1.  6. Effect of the individual MBP-NuoG iron-sulfur mutants on the peripheral subunit assembly of NDH-1 in the nuoG knock-out E. coli strain (MC4100⌬nuoG). A, Western blot analysis of membranes from different E. coli strains, inclusion body (Ib) of NuoG, and various organisms with anti-NuoG antibody. 20-g samples except for the NuoG inclusion body (2 g) were applied to each lane of a Laemmli SDS-12.5% polyacrylamide gel. Pd, Paracoccus denitrificans; Th, Thermus thermophilus; Hp, Helicobacter pylori; SMP, bovine submitochondrial particle. B, Western blot analyses of membranes from the MBP-NuoG mutants expressed in E. coli MC4100⌬nuoG strain with affinity-purified antibodies specific to the individual subunits of E. coli NDH-1: anti-NuoB, anti-NuoCD, anti-NuoE, and anti-NuoF antibodies. Lane 1, MC4100 wild type; lane 2, MC4100⌬nuoG; lanes 3-7, MC4100⌬nuoG strain carrying pMALc2g/nuoG⌬N1b, pMALc2g/nuoG-⌬N1c, pMALc2g/nuoG⌬N4, pMALc2g/nuoG⌬N5, and pMALc2g/nuoG, respectively. Membrane samples (20 g) were applied to each lane of a Laemmli SDS-10% or -6% (for anti-NuoCD) polyacrylamide gel.

Effects of Each Iron-Sulfur Cluster Mutant of the MBP-NuoG Subunit on Interaction with Other NDH-1 Subunits in Vivo-As shown in
The 75-kDa/Nqo3/NuoG subunit has been suggested to play a structurally important role in connecting the flavoprotein subcomplex, other amphipathic subunits, and hydrophobic subunits together in the NDH-1/complex I (42,43). Consistently, the knock-out of the nuoG gene resulted in the loss of the NDH-1 assembly in the cytoplasmic membranes. All peripheral subunits were absent except that rather large quantities of the NuoCD protein were found in the membranes of MC4100⌬nuoG. It seems highly unlikely that the membrane subunits NuoH, -J, -K, -L, -M, and -N are expressed and assembled in the membranes of MC4100⌬nuoG because of a polar effect of the inserted spectinomycin cassette (44). Because the NuoA subunit (the gene of which is located upstream of the nuoG gene) was not detected in the membranes, 3 it seems likely that the NuoCD subunit alone is able to directly associate with the cytoplasmic membrane. When the MBP-NuoG(WT) subunit was expressed in the MC4100⌬nuoG strain, the assembly of the NuoB, -CD, -E, and -F subunits was partially restored, indicating that the NuoG subunit plays an essential role in assembling the other peripheral subunits. This observation is in accordance with the previous cross-linking studies that the 75-kDa/Nqo3/NuoG subunit directly interacts with the PSST/Nqo6/NuoB, 30-kDa/Nqo5/NuoC, 24-kDa/Nqo2/ NuoE, and 51-kDa/Nqo1/NuoF subunits (42,45). When the MBP-NuoG mutant subunits were expressed, we observed significantly different effects on the subunit assembly. Similar results were reported in subunit NuoI (46). These results suggest that the individual iron-sulfur cluster binding domains seem to play roles in interactions with the neighboring subunits, which are essential for the intersubunit electron transfer pathway. Further study is in progress in our laboratories.