Identification of the dicyclohexylcarbodiimide-binding subunit of NADH-ubiquinone oxidoreductase (Complex I).

The mitochondrial NADH:ubiquinone oxidoreductase complex (Complex I) is inhibited by N,N'-dicyclohexylcarbodiimide (DCCD), and this inhibition correlates with incorporation of radioactivity from [14C]DCCD into a Complex I subunit of Mr 29,000 (Yagi, T. (1987) Biochemistry 26, 2822-2828). Resolution of [14C]DCCD-labeled Complex I in the presence of NaClO4 showed that the labeled Mr 29,000 subunit was in the hydrophobic fraction of the enzyme. This fraction, which contains greater than 17 unlike polypeptides, was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and the Mr 29,000 subunit, containing bound [14C]DCCD, was isolated and purified. The amino acid composition and partial sequence of this subunit corresponded to those predicted from the mitochondrial DNA for the product of the mtDNA gene designated ND-1. The identity of the Mr 29,000 subunit with the ND-1 gene product was further confirmed by immunoblotting and immunoprecipitation experiments, using the hydrophobic fraction of [14C]DCCD-labeled Complex I and antiserum to a C-terminal undecapeptide synthesized on the basis of the human mitochondrial ND-1 nucleotide sequence. Thus, it appears that the DCCD-binding subunits of the respiratory chain Complexes I, III, and IV and in certain organisms the DCCD-binding subunit of the ATP synthase complex (Complex V) are all mtDNA products.


Takao Yagi$ and Youssef Hatefi
From the Division of Biochemistrv. Deuartment of Basic and Clinical Research, Research Institute of Scripps Clinic, plexes I11 and IV is also inhibited by DCCD (8)(9)(10). More recently, it was shown by Yagi (11) that bovine Complex I (NADH:ubiquinone oxidoreductase), the NADH-Q reductase isolated from Paracoccus denitrificans, and the NADH-Q reductase activity of Escherichia coli and Thermus thermophilus HB-8 membranes were all inhibited by DCCD. These and other results indicated further that DCCD inhibition of NADH-Q reductase activity in different species is correlated with the presence of an energy-coupling site in this segment of their electron transfer system. Where NADH-Q reductase activity was inhibited by DCCD, this span of the electron transfer chain was capable of proton translocation coupled to electron transfer, and where NADH-Q reductase activity was insensitive to inhibition by DCCD, reduction of Q by NADH was not coupled to proton translocation (11).
In addition to the above, it was demonstrated by Yagi (11) that treatment of bovine Complex I with [14C]DCCD resulted in incorporation of radioactivity into two Complex I polypeptides with M, values of 49,000 and 29,000. The time course of [14C]DCCD labeling of the M, 29,000 polypeptide, but not that of the M, 49,000 polypeptide, paralleled the time course of inhibition of NADH-Q reductase activity. Since bovine Complex I is composed of >25 unlike polypeptides ( 4 ) and contains several polypeptides in the M, 30,000 range, it was of interest to isolate and characterize the DCCD-binding polypeptide of this enzyme complex, especially since this polypeptide may be concerned with proton translocation at the level of Complex I, just as the DCCD-binding subunits of Complexes 111, IV, and V (ATP synthase) are considered to be so involved.

EXPERIMENTAL PROCEDURES
Enzyme Preparation-Bovine Complex I (12) and the iron-sulfur flavoprotein (FP), the iron-sulfur protein (IP), and the hydrophobic protein (HP) fractions of Complex I (13) were prepared according to the references given.
Purification of Polypeptides from HP-Eighty p g of HP were loaded in each well (1-cm width) of a sodium dodecyl sulfate slab gel (15 X 14 X 0.15 cm) composed of 10% acrylamide as described by Laemmli (14) except that the running buffer contained 0.1 mM thioglycolic acid. The gel was electrophoresed for 3 h at 30 mA and stained for 30 min at room temperature in a solution containing 0.05% Coomassie Brilliant Blue R-250, 25% 2-propanol, and 10% acetic acid. Destaining was carried out for 2-3 h at 4 "C in 5% acetic acid containing 16.5% methanol. The stained bands were sliced from the gel and washed several times with water. The polypeptides were then electroeluted as described by Hunkapiller et al. (15). For amino acid analysis and sequence determination, the isolated polypeptides were re-electrophoresed and re-isolated.
Removal of N-Formyl Groups-The amino terminal blocking group of the isolated polypeptide (assumed to be N-formyl) was removed by the modified method of Sheehan and Yang (16) and Fearnley and Walker (17). The purified polypeptide (68 Gg) was lyophilized, SUSpended in 1 ml of 1.5 N HC1 in methanol, and incubated for 4 h at room temperature. The solution was neutralized by addition of 1.5 ml of 1 M NaHCOs, and then concentrated by Centricon-10. The concentrate was washed with 3 ml of 10 mM NaHC03 containing 0.02% SDS, lyophilized, and subjected to sequence analysis by a gas phase sequencer (Applied Biosystem).
Other Analytical Procedures-Protein was estimated by the methods of Lowry et al. (18) or by biuret in the presence of 1 mg of sodium deoxycholate/ml (19). Enzymatic assays were carried out essentially according to Yagi (11,20). The gels were autoradiographed with Kodak X-AR5. Immunoblotting experiments were carried out using skim milk and Tween-20 as blocker (20). Any variations from these procedures and other details are described in the figure legends.
Materials-The sources of the chemicals used were as follows: NADH and dithiothreitol were from Behring Diagnostics Inc., Somerville, NJ; DCCD was from Aldrich; [''CIDCCD was from Research Products; acrylamide, SDS, and Coomassie Brilliant Blue R-250 were from Bio-Rad; rotenone was from S. B. Penick; NaClO. was from G. Frederick Smith Chemical Co.; and protein A-Sepharose CL-4B was from Pharmacia LKB Biotechnology Inc. Ubiquinone-1 was a generous gift from Eisai Chemical (Tokyo, Japan). The antiserum to human ND-1 C-terminal undecapeptide was kindly provided by Dr. R. F. Doolittle (University of California, San Diego) and Dr. G. Attardi (California Institute of Technology, Pasadena, CA). Other chemicals were reagent grade or of the highest quality available.

Labeling Pattern of Complex
I Polypeptides with r4C] DCCD-Treatment of Complex I with chaotropic salts results in the resolution of the enzyme complex into a water-soluble and a water-insoluble fraction. The former contains a threesubunit iron-sulfur flavoprotein (FP or the primary NADH dehydrogenase) and a six-subunit iron-sulfur protein (IP). The latter fraction (HP) contains the remainder of the polypeptides of Complex I, phospholipids, and one or two ironsulfur clusters (4). When this resolution procedure was applied to [14C]DCCD-treated Complex I, protein-bound radioactivity from [I4C]DCCD was found exclusively in the HP fraction. These results are shown in Fig. 1. Complex I at 10 mg/ml was incubated with [14C]DCCD until NADH-QI reductase activity was 70% inhibited. The labeled enzyme was precipitated with ammonium sulfate as described in the legend + b a n d l C band 2 to Fig. 1, resuspended in buffer at 10 mg/ml, and resolved in the presence of 0.5 M NaClO.,. HP, IP, and FP were separated as described and subjected to SDS-gel electrophoresis. A sample of Complex I not treated with DCCD was similarly manipulated and resolved. Fig. 1 shows on the left a proteinstained SDS gel of HP, IP, and FP from [I4C]DCCD-treated and control Complex I run side by side. This gel shows that treatment of Complex I with [14C]DCCD does not seem to have affected the resolution of Complex I into HP, IP, and FP, altered the mobility of the polypeptides in these fractions, or created new bands as a result of cross-linking. The right side of Fig. 1 shows an autoradiogram of the gel on the left. It is seen that radioactivity from [I4C]DCCD was incorporated only in the polypeptides of the H P fraction. Two polypeptides with M , values of 49,000 (band 1) and 29,000 (band 2) and several with M , values <20,000 were labeled. Among these, only the labeling of the M , 29,000 polypeptide correlated with inhibition of NADH-QI reductase activity, as described previously (11). The M, 49,000 band was labeled too rapidly, and the extent of labeling of bands with M , <20,000 was found to be related neither to the duration of incubation of Complex I with [I4C]DCCD nor to the degree of inhibition of NADH-QI reductase activity. Indeed, as will be seen below, these latter bands were not labeled when Complex I at concentrations lower than 10 mg/ml was incubated with [14C]DCCD. This extraneous labeling of low M , polypeptides may be related to the multiple labeling pattern reported by Voukila and Hassinen (21).
Isolation of the DCCD-binding Subunit-In the remainder of this report, the M , 29,000 polypeptide described above will be referred to as the DCCD-binding subunit, because the extent of labeling of only this polypeptide by [ 14C]DCCD was found to correlate with inhibition of the NADH-QI reductase activity of Complex I (11). Although distinct in the autoradiogram of Fig. 1, the DCCD-binding subunit was found in a region of the SDS gel of HP which contained three closely packed protein bands. These bands are shown in Fig. 2 and will be referred to as the upper, the middle, and the lower bands. Since in this situation autoradiography could not allow one to decide which band was the DCCD-binding subunit, each band was carefully excised from a number of comparable gels. The corresponding slices were combined, protein was DCCD (1.0 X 10" cpm/mol) in 1 ml of a solution containing 0.25 M sucrose and 50 mM Tris-HCI, pH 7.5, for 2 h at 20 "C. The mixture was diluted with 4 ml of the same solution, and Complex I was precipitated by addition of 2.8 ml of saturated ammonium sulfate solution. The mixture was allowed to stand on ice for 10 min and then was centrifuged for 20 min a t 42,500 rpm in a 50 T i rotor of Beckman ultracentrifuge. The supernatant containing unbound ["C] DCCD was discarded, and the pellet was suspended in 0.8 ml of 50 mM Tris-HCI, pH 8.0, containing 5 mM dithiothreitol. NaC104 was added from an 8 M solution to a final concentration of 0.5 M, and the labeled Complex I was resolved and fractionated into HP, IP, and FP according to Galante and Hatefi (13). Protein samples were denatured in the Laemmli sample buffer, containing 80 mM Tris-HCI, pH 6.8, 6% SDS, 0.005% bromphenol blue, and 20% glycerol and were applied to a Laemmli-type 10% acrylamide mini-slab (55 X 95 X 0.75 mm) SDS gel. The amounts of protein applied to the gel were 4 pg each of FP and IP and 12 pg of HP. The gel was electrophoresed for 1 h a t 200 V and then stained, destained, and autoradiographed as described previously (11). removed by electroelution and subjected to SDS gel electrophoresis. Results are shown in Fig. 3. The left side of Fig. 3 shows an SDS gel of the isolated upper, middle, and lower polypeptides stained with Coomassie Blue, and the right side shows an autoradiogram of the same gel. It is clear that radioactivity from [I4C]DCCD was associated mainly with the lower band. The small amount of radioactivity seen in the middle band is probably due to contamination with the lower band and related to the difficulty in cleanly separating these bands by excision from SDS gels of HP.
Relationship of the DCCD-binding Subunit to the Mitochondrial Gene Products in Complex I-The DCCD-binding subunits of Complex I11 (cytochrome b ) and Complex IV (subunit 111) are mitochondrial gene products (8,10,22). Also, in certain species, e.g. Saccharomyces cereuisiae and possibly maize, the DCCD-binding subunit of the ATP synthase complex is encoded and synthesized within the mitochondria (23). Therefore, it was of interest to see whether the DCCD-binding subunit of Complex I is related to any one of the seven Complex I polypeptides which are the products of mitochondrial genes, designated ND-1, -2, -3, -4L, -4, -5, and -6 (24,25). Among these, the ND-1 gene product was the most likely possibility, because ( a ) its molecular weight as predicted from the gene (36,000) was closest to that of the DCCD-binding subunit and ( b ) the human ND-1 gene product with the same predicted molecular weight exhibited an M , of 24,000 on SDSurea gels (26). Thus, the relationship between the DCCDbinding subunit and the ND-1 gene product was investigated chemically as well as immunochemically. For this purpose, the DCCD-binding subunit (lower band of Figs. 2 and 3) was further purified by repeated SDS-gel electrophoresis and electroelution until no hint of contamination could be detected on stained gels. This material was then subjected to amino acid analysis and N-terminal sequencing. Table I shows the amino acid composition of the DCCD-binding subunit as well as that of the bovine ND-1 gene product as predicted from the gene sequence (22). The similarity of amino acid composition is obvious. The results of the N-terminal sequencing are given in Fig. 4. It is seen that the 15 N-terminal amino acids of the DCCD-binding subunit are identical to those of the bovine ND-1 gene product as predicted from the ND-1 gene sequence.
The immunochemical experiment was carried out as follows. Because antiserum to the bovine ND-1 gene product was not available, an antiserum to a synthetic peptide representing the C-terminal undecapeptide of the human ND-1 gene product was employed. This antipeptide antiserum has  The purified DCCD-binding subunit (10 pg been shown to cross-react with the bovine ND-1 gene product, because the C terminus of the latter is highly similar to that of its human counterpart (27). As seen in the immunoblot of Fig. 5, the isolated lower band and the same polypeptide in H P cross-reacted with this antipeptide antibody, but the isolated middle and upper bands did not (see Figs. 2 and 3). Other experiments indicated that none of the FP or I P polypeptides cross-reacted with the above antipeptide antiserum either (results not shown). Thus, the analytical and the immunoblotting results indicated strongly that the bovine ND-1 gene product is identical to the polypeptide that was identified in Fig. 3 as the DCCD-binding subunit.
The experiments of Figs. 4 and 5 and of Table I were carried out with the isolated lower band of Fig. 2 or HP, neither of which had been treated with DCCD. It was, therefore, necessary to establish that the DCCD-labeled material in the lower band was indeed the same as the polypeptide identified above as the ND-1 gene product. This experiment was performed by labeling Complex I with [14C]DCCD as described in Fig. 1, isolating the labeled HP, and immunoprecipitating from it the DCCD-binding subunit with the use of the antiserum to the synthetic C-terminal undecapeptide of the human ND-1  Fig. 2. The three protein bands were isolated from SDS gels of H P a s described under "Experimental Procedures." The isolated proteins (1 LIE each) were lyophilized, dissolved in the Laemmli sample buffer as in Fig. l, and applied to a Laemmli-type mini-slab SDS gel. Other details were as in Fig. 1. gene -product. As seen in Fig. 6 Fig. 3) with antiserum to a synthetic C-terminal undecapeptide of the human ND-1 gene product. Proteins were electrically transferred from SDS-polyacrylamide gels to nitrocellulose membranes (Schleicher and Schuell, 0.22-pm pore size) as described by Towbin et al. (31). The nitrocellulose membranes were incubated with 1% skim milk and 0.05% Tween-20 in a buffer containing 50 mM sodium phosphate, pH 7.5, and 150 mM NaCl for 1 h a t 37 "C. After washing with the same buffer, the nitrocellulose membranes were incubated with antiserum (1:500 dilution) to the C-terminal undecapeptide of the human ND-1 gene product for 105 min a t 37 "C and then with anti-rabbit antibody conjugated to horseradish peroxidase (1:lOOO dilution) for 45 min a t 37 "C. The nitrocellulose membranes were transferred into substrate solution which contained in the above phosphate-NaCI buffer 0.36 mg/ml of 4-chloronaphthol and 0.012% H202. After incubation for 30 min, the nitrocellulose membranes were washed with H 2 0 and dried.

SDS-Gel Autoradiogram
FIG. 6. Immunoprecipitation of the DCCD-binding subunit with antiserum to the C-terminal undecapeptide of the human ND-1 gene product (Anti ND-IC). Immunoprecipitation experiments with the antipeptide antiserum were carried out according to Anderson and Blobel (32). ["CIDCCD-labeled H P (1.93 mg/ml) was incubated for 1 h at 30 "C in a buffer containing 0.25 M sucrose, 50 mM Tris-acetate, pH 7.5, and 4.5% SDS. Fifty pl of the H P solution was diluted with 200 pl of a buffer containing 190 mM NaCI, 60 mM Tris-HCI, pH 7.6, 6 mM EDTA, and 1.25% Triton X-100 and then incubated overnight at 4 "C with 10 pl of the antipeptide antiserum or control serum. The suspensions were centrifuged for 2 min in an Eppendorf centrifuge. The Supernatants were transferred to new Eppendorf tubes, 60 pl of 1:l suspension of protein A-Sepharose CL-4B in the dilution buffer were added to each supernatant, and the mixtures were incubated with end-over-end mixing for 2 h a t room temperature. The resins were washed three times with 1 ml of a buffer containing 145 mM NaCI, 50 mM Tris-HCI, pH 7.5, 5 mM EDTA, 0.1% Triton X-100, and 0.02% SDS and then washed once with the same buffer lacking detergent. Forty pl of Laemmli's sample buffer (see Fig. 1) was added to each tube and incubated for 2 h at 30 'C. After centrifugation for 30 s, the supernatants (8 and 15 pl) were subjected to SDS-gel electrophoresis and autoradiography as described in Fig. 1. The heavily stained bands near the top of the gels are IgG. Relationship between the DCCD-binding Subunit and the Site of Rotenone Inhibition-Complex I is inhibited by rotenone, piericidin A, Demerol, and barbiturates (4). Phenomenologically, all of these reagents interrupt electron transfer from FMN and all the EPR-visible iron-sulfur centers of Complex I to ubiquinone. There are also data suggesting that these reagents bind to a common or overlapping site (28). Recent studies of Earley et al. (29) have suggested that the rotenone-binding subunit of Complex I is the ND-1 gene product. Thus, it was of interest to see whether rotenone and DCCD bind to a common site. Fig. 7 shows the effect of rotenone on the labeling of Complex I by [14C]DCCD. On the 'eft is depicted an SDS slab gel of Complex I treated in lanes J-7 with increasing concentrations of rotenone, which resulted in 73-100% inhibition of NADH-QI reductase activity. The preparations were then incubated with [14C]DCCD and subjected to SDS-gel electrophoresis. The right-hand side of Fig. 7 shows an autoradiogram of the gel on the left. It is seen that complete inhibition of NADH-Ql reductase activity by rotenone had no apparent effect on labeling of Complex I by [I4C]DCCD. Assuming that rotenone binds to the ND-1 gene product as suggested by Earley et al. (29), the data of Fig. 7 would suggest that the binding sites of rotenone and DCCD on this subunit are different.

DISCUSSION
Two polypeptides of apparent M , 49,000 and 29,000 are labeled when bovine Complex I is treated with [14C]DCCD. The former is labeled rapidly and saturated when inhibition of NADH-Ql reductase activity of Complex I reaches about 50%. However, the labeling of the M , 29,000 polypeptide parallels the inhibition of NADH-Ql reductase activity by [14C]DCCD and reaches saturation only when this activity is completely inhibited (11). These results indicated, therefore, that the M , 29,000 polypeptide was involved in the NADHubiquinone reductase activity of Complex I and contained an essential carboxyl group whose modification by DCCD resulted in inhibition of ubiquinone reduction (11,30). Voukila and Hassinen (21) have reported that treatment of Complex I at 37 "C with [14C]DCCD results in the labeling of six subunits with molecular masses of 13.7, 16.1, 21.5, 39,43, and 53 kDa. First of all, this extensive labeling may be related to the incubation temperature used, because incubation of Complex I at 37 "C results in structural alterations and loss of rotenone-sensitive Q reductase activity (11,20). Second, as it was indicated above, we also observed the labeling of additional polypeptides when Complex I at high concentration (e.g. 10 mg/ml) was incubated with [14C]DCCD. However, the extent of labeling of these additional polypeptides appeared to be unrelated to the duration of incubation of Complex I with [14C]DCCD and to the degree of inhibition of NADH-Q1 reductase activity. Considering the fact that Complex I contains a large number of hydrophobic polypeptides (4), such spurious labeling by [14C]DCCD, especially under adverse conditions, is not surprising. Voukila and Hassinen (21) do not report on the correlation between inhibition of NADHubiquinone reductase activity and the extent of labeling by [14C]DCCD of the several polypeptides indicated above. However, in our experiments, such a correlation was observed with respect to labeling of only the M, 29,000 polypeptide. Thus, it was considered justified to designate this polypeptide as the DCCD-binding subunit of Complex I.
Resolution of [14C]DCCD-treated Complex I by chaotropic salts and separation of the water-soluble FP and IP fragments from the water-insoluble HP fraction showed that the DCCDbinding subunit was in HP, not in FP and IP which contain most of the electron carriers of Complex I. The DCCD-binding subunit was excised from SDS gels of HP, purified by repeated gel electrophoresis and electroelution, and subjected to amino acid analysis and N-terminal sequencing down to the 15th residue. These data showed excellent agreement with the predicted amino acid composition and N-terminal sequence of the ND-1 gene product of mitochondrial DNA. Furthermore, the antiserum to a synthetic peptide corresponding to the C-terminal undecapeptide of the human ND-1 gene product cross-reacted with the isolated DCCD-binding subunit of Complex I and immunoprecipitated this subunit labeled with [14C]DCCD from the HP fraction of [14C]DCCD-treated Complex I. The similarity of the C termini of the human and bovine ND-1 gene products is known (22), and the above antipeptide antiserum had been shown earlier to cross-react with the corresponding mtDNA gene product of bovine mitochondria (27,29). Thus, the results reported here establish that the DCCD-binding subunit of Complex I is the product of the bovine mtDNA gene designated as ND-1.
The finding that the DCCD-binding subunit of Complex I is a mitochondrial gene product is interesting in view of the fact that the DCCD-binding subunits of Complex I11 (cytochrome b ) and Complex IV (subunit 111) are also mtDNA products. In addition, the DCCD-binding subunit of the ATP synthase complex in S. cereuisiae and possibly maize is mtDNA-encoded as well. It might also be added that in Neurospora the gene for the DCCD-binding subunit of the ATP synthase complex is present in both the mitochondrial and the nuclear DNA, but only the latter is expressed (23).
In terms of the effect of DCCD, the above energy-transducing systems may be divided into two groups. In Complexes I11 and IV, DCCD modification appears to disengage the scalar and the vectorial reactions, because DCCD inhibits proton translocation much more than electron transfer (8,lO). In the ATP synthase complex, this disengagement does not occur. AS a result, ATP hydrolysis is inhibited when proton translocation through the Fo sector of the enzyme complex is blocked by DCCD. In Complex I, electron transfer and proton translocation are also blocked in parallel by DCCD (11). This could mean that the primary effect of DCCD is on electron transfer, which necessarily inhibits proton translocation as well. Should this be the case, then the DCCD-binding subunit of Complex I may be involved in electron transfer. This is entirely possible, because iron-sulfur center N-2 of Complex I, which has a relatively high reduction potential and is considered to be the immediate electron donor to ubiquinone, appears to fractionate into HP. The polypeptide bearing ironsulfur center N-2 would be expected to be hydrophobic, because it would have to interact with the water-insoluble ubiquinone-10. It could also be the site of rotenone binding, because this reagent inhibits electron transfer immediately on the substrate side of ubiquinone.
Another interpretation of concomitant inhibition by DCCD of electron transfer and proton translocation by Complex I would be that DCCD acts on Complex I as it does on the ATP synthase complex. It blocks proton translocation by Complex I without disengaging the proton channel from the subunits involved in electron transfer. Consequently, electron transfer becomes inhibited to the same extent that proton translocation is blocked by DCCD. Should this interpretation be correct, then another analogy between Complex I and the ATP synthase complex may be considered. In the ATP synthase complex, the catalytic sector (Fl) is water-soluble, whereas the sector involved in proton translocation through the membrane (Fo) is composed of hydrophobic polypeptides. In Complex I, also, the FP (which is concerned with NADH oxidation) and the IP are water-soluble and contain FMN and six of the eight iron-sulfur clusters of Complex I, whereas HP (which most likely contains the polypeptides that form the proton channel of Complex I) is composed of hydrophobic polypeptides and contains the DCCD-binding subunit of the enzyme complex. The fact that several energy-transducing NADH-quinone reductases that have been examined from prokaryotic and eukaryotic sources are all inhibited by DCCD, whereas those not containing an energy-coupling site are not ( l l ) , could be a hint in favor of the latter possibility.