The Subunit Composition of the Human NADH Dehydrogenase Obtained by Rapid One-step Immunopurification

Defects of the NADH dehydrogenase complex are predominantly manifested in mitochondrial diseases and are significantly associated with the development of many late onset neurological disorders such as Parkinson's disease. Here we describe an immunocapture procedure for isolating this multisubunit membrane-bound complex from human tissue. Using small amounts of immunoisolated protein, one-dimensional and two-dimensional gel electrophoresis, matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) peptide mass finger printing (PMF), and nanoflow liquid chromatography mass spectrometry/mass spectrometry (LC-MS/MS), we can resolve and identify the human homologues of 42 polypeptides detected so far in the more extensively studied beef heart complex I. These polypeptides include the GRIM-19 protein, which is claimed to be involved in apoptosis, a polypeptide first identified by gene screening as a neuronal protein, as well as a protein thought to be in differentiation linked processes. The concordance of data from human and bovine complex I isolated by different procedures adds to the certainty that these novel proteins of seemingly diverse function are a part of complex I. Defects of the NADH dehydrogenase complex are predominantly manifested in mitochondrial diseases and are significantly associated with the development of many late onset neurological disorders such as Parkinson's disease. Here we describe an immunocapture procedure for isolating this multisubunit membrane-bound complex from human tissue. Using small amounts of immunoisolated protein, one-dimensional and two-dimensional gel electrophoresis, matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) peptide mass finger printing (PMF), and nanoflow liquid chromatography mass spectrometry/mass spectrometry (LC-MS/MS), we can resolve and identify the human homologues of 42 polypeptides detected so far in the more extensively studied beef heart complex I. These polypeptides include the GRIM-19 protein, which is claimed to be involved in apoptosis, a polypeptide first identified by gene screening as a neuronal protein, as well as a protein thought to be in differentiation linked processes. The concordance of data from human and bovine complex I isolated by different procedures adds to the certainty that these novel proteins of seemingly diverse function are a part of complex I. Increasingly studies are highlighting the major extent to which defects of complex I, the NADH-ubiquinol reductase of the mitochondrial electron transport chain, contribute to human disease pathology. Genetic mutations of subunits in this complex are the leading cause of inherited mitochondrial diseases, which include Leigh's syndrome, Leber's hereditary optic neuropathy, and mitochondrial encephalopathy with lactic acidosis and stroke-like episodes (reviewed in Ref. 1Triepels R.H. Van Den Heuvel L.P. Trijbels J.M. Smeitink J.A. Am. J. Med. Genet. 2001; 106: 37-45Crossref PubMed Scopus (149) Google Scholar). In addition there is strong evidence that complex I inhibition, through accumulated damage caused by reactive oxygen and nitrogen species and by the binding of environmental toxins, has a role in the development of the more prevalent neurodegenerative disorders including Parkinson's disease (2Mizuno Y. Ohta S. Tanaka M. Takamiya S. Suzuki K. Sato T. Oya H. Ozawa T. Kagawa Y. Biochem. Biophys. Res. Commun. 1989; 163: 1450-1455Crossref PubMed Scopus (658) Google Scholar, 3Schapira A.H. Cooper J.M. Dexter D. Clark J.B. Jenner P. Marsden C.D. J. Neurochem. 1990; 54: 823-827Crossref PubMed Scopus (1681) Google Scholar, 4Betarbet R. Sherer T.B. MacKenzie G. Garcia-Osuna M. Panov A.V. Greenamyre J.T. Nat. Neurosci. 2000; 3: 1301-1306Crossref PubMed Scopus (2984) Google Scholar), Alzheimer's disease (5Kim S.H. Vlkolinsky R. Cairns N. Fountoulakis M. Lubec G. Life Sci. 2001; 68: 2741-2750Crossref PubMed Scopus (128) Google Scholar), Huntington's disease (6Arenas J. Campos Y. Ribacoba R. Martin M.A. Rubio J.C. Ablanedo P. Cabello A. Ann. Neurol. 1998; 43: 397-400Crossref PubMed Scopus (146) Google Scholar), amyotrophic lateral sclerosis (7Vielhaber S. Winkler K. Kirches E. Kunz D. Buchner M. Feistner H. Elger C.E. Ludolph A.C. Riepe M.W. Kunz W.S. J. Neurol. Sci. 1999; 169: 133-139Abstract Full Text Full Text PDF PubMed Scopus (88) Google Scholar, 8Vielhaber S. Schroder R. Winkler K. Weis S. Sailer M. Feistner H. Heinze H.J. Schroder J.M. Kunz W.S. J Neuropathol. Exp. Neurol. 2001; 60: 1032-1040Crossref PubMed Scopus (20) Google Scholar), Down's syndrome (5Kim S.H. Vlkolinsky R. Cairns N. Fountoulakis M. Lubec G. Life Sci. 2001; 68: 2741-2750Crossref PubMed Scopus (128) Google Scholar), schizophrenia (9Dror N. Klein E. Karry R. Sheinkman A. Kirsh Z. Mazor M. Tzukerman M. Ben-Shachar D. Mol. Psychiatry. 2002; 7: 995-1001Crossref PubMed Scopus (110) Google Scholar), and even the process of aging itself (10Lenaz G. D'Aurelio M. Merlo Pich M. Genova M.L. Ventura B. Bovina C. Formiggini G. Parenti Castelli G. Biochim. Biophys. Acta. 2000; 1459: 397-404Crossref PubMed Scopus (159) Google Scholar). Complex I is a large multimeric enzyme complex with an approximate molecular weight of 1 million. The bovine form isolated from beef heart has been analyzed extensively. In total, 45 different putative subunits have been identified, with 7 encoded by the mitochondrial DNA and the remainder of nuclear DNA origin (11Carrol J. Shannon R.J. Fearnley I.M. Walker J.E. Hirst J. J. Biol. Chem. 2002; 277: 50311-50317Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar). Purification protocols described thus far to obtain the enzyme for detailed characterization require large amounts of mitochondria as starting material, often in gram quantities that are not easily obtained from human tissue sources. Given the importance of the enzyme in human diseases, protocols that permit isolation of complex I from small amounts of biopsy tissue and even cell culture material would have great diagnostic utility. Moreover, a rapid procedure is important if multiple samples are being analyzed, for example in screening for protein markers of the above diseases. We have recently described a procedure to obtain complex I at high purity from human heart tissue by the simple method of solubilizing mitochondria in detergent and separating component complexes based on macromolecular size by sucrose gradient fractionation (12Hanson B.J. Schulenberg B. Patton W.F. Capaldi R.A. Electrophoresis. 2001; 22: 950-959Crossref PubMed Scopus (88) Google Scholar). The large size of complex I causes this enzyme to sediment in the densest fractions of a sucrose gradient and thus can be isolated in a convenient manner. However, 2.5 mg of mitochondrial protein are typically required, which in turn requires collection of 0.1–1 g of tissue depending on tissue type. More recently, we have described an expedient isolation protocol for both the ATP synthase and PDH complexes from small amounts of human tissue by immunocapture using monoclonal antibodies (mAbs) 1The abbreviations used are: mAbmonoclonal antibodyF1 α and βATP synthase α and β chainCHAPS3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acidMALDI-TOFmatrix-assisted laser desorption ionization time-of-flightLCliquid chromatographyMSmass spectrometry made against these respective antigens (13Aggeler R. Coons J. Taylor S.W. Ghosh S.S. Garcia J.J. Capaldi R.A. Marusich M.F. J. Biol. Chem. 2002; 277: 33906-33912Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar, 14Lib M. Marusich M.F. Rodriguez-Mari A. Capaldi R.A. Anal. Biochem. 2003; (in press)PubMed Google Scholar). Here we report the successful purification of complex I from human heart with mAbs made against the bovine form of the enzyme. We provide a proteomic analysis of the complex that identifies the polypeptide constituents and describe gel-based techniques, which allow separation of subunits for analyses of stoichiometry and the state of post-translational modification. monoclonal antibody ATP synthase α and β chain 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid matrix-assisted laser desorption ionization time-of-flight liquid chromatography mass spectrometry Human heart mitochondria were obtained from Analytical Biological Services by tissue homogenization and differential centrifugation essentially as described by Smith et al. (15Smith A. Methods Enzymol. 1967; 10: 81-86Crossref Scopus (472) Google Scholar). All mitochondrial samples were washed with 20 mm Tris-HCl, pH 7.5, 1 mm EDTA. The mitochondria were resuspended in the same buffer, supplemented with protease inhibitors (leupeptin at 0.5 mg/ml, pepstatin at 0.5 mg/ml, and 1 mm phenylmethylsulfonyl fluoride) at a concentration of 10 mg/ml. An equal volume of 2%n-dodecyl-β-d-maltoside (Calbiochem) was added to a final crucial concentration of 1% detergent at 5 mg/ml protein and incubated for 30 min on ice. Insoluble material was removed from the samples by centrifugation in a TLA100.2 (Beckman) at relative centrifugal force max 184,000 × g for 30 min at 4 °C. For immunoisolation, 80 μg of monoclonal antibody 17G3D9E12 was bound to 20 μg of swollen protein G-agarose beads (Sigma). The antibody was cross-linked to the beads with 25 mm dimethylpimelimidate (Sigma) for 30 min at room temperature in 0.2 m sodium borate, pH 9.0. Cross-linking was terminated with 0.2 methanolamine solution, pH 8.0, for 3 h at room temperature. Antibody cross-linked beads were collected by gentle centrifugation at 3000 rpm in a microfuge and resuspended in phosphate-buffered saline. This conjugate was incubated overnight at 4 °C with the supernatant from 10 mg of solubilized mitochondria. Beads were washed six times with phosphate-buffered saline supplemented with 0.05%n-dodecyl-β-d-maltoside. Immunocaptured NADH dehydrogenase was eluted twice with 60 μl of 0.1 mglycine, pH 2.5, supplemented with 0.05%n-dodecyl-β-d-maltoside. The sample was then dialyzed in phosphate-buffered saline, 0.05%n-dodecyl-β-d-maltoside to neutralize the pH of the solution. For one-dimensional electrophoresis samples separated on a 10–22% acrylamide gels containing 0.05% SDS, 0.375 m Tris-HCl pH 8.6. For two-dimensional electrophoresis, 100-μl samples were denatured in 350 μl of rehydration buffer (7 m urea, 2m thiourea, 65 mm dithiothreitol, 0.8% pH 3–10 carrier ampholyte (Fluka), 2% CHAPS (Sigma), 1% Zwittergen 310 (Sigma), 0.1% SDS) for 15 min at room temperature. Each sample was used to hydrate 18-cm immobiline pH gradient strips pH 3–10 (Amersham Biosciences) for 12 h. Then isoelectric focusing was performed in three stages of applied potential difference: 500 V for 1 h, 1000 V for 1 h, and 8000 V for up to 10 h, until 60,000 Vh were achieved. Focused strips were then soaked in SDS-PAGE buffer (50 mm Tris-HCl, pH 8.8, 6 murea, 30% glycerol, 2% SDS, 0.01% bromphenol blue, 100 mm dithiothreitol) for 15 min at room temperature. Strips were applied to 15% acrylamide gels for SDS-PAGE. Gels were stained with Coomassie Brilliant Blue (Sigma) or Sypro Ruby (Molecular Probes) gel stains. Silver staining was performed according to Shevchenko et al. (16Shevchenko A. Wilm M. Vorm O. Mann M. Anal. Chem. 1996; 68: 850-858Crossref PubMed Scopus (7831) Google Scholar). Sypro Ruby and silver-stained two-dimensional gel spots were excised using a ProteomeWorksTM Robotic Imager and Spot Cutter (Bio-Rad) and processed for mass spectrometric analysis as described previously (17Taylor S.W. Warnock D.E. Glenn G.M. Zhang B. Fahy E. Gaucher S.P. Capaldi R.A. Gibson B.W. Ghosh S.S. J. Proteome Res. 2002; 1: 451-458Crossref PubMed Scopus (74) Google Scholar). Silver-stained two-dimensional samples were manually destained in 5 mm potassium ferricyanide and 1 mm sodium thiosulfate, while Sypro Ruby-stained gel spots were destained in a ProGestTM digestion robot (Genomic Solutions Inc.). Reduction, alkylation, and digestion of both Sypro Ruby- and silver-stained two-dimensional gel spots were performed using the ProGest. Silver-stained one-dimensional gel bands (destained as described above) and Coomassie-stained one-dimensional gel bands were manually processed as described previously (13Aggeler R. Coons J. Taylor S.W. Ghosh S.S. Garcia J.J. Capaldi R.A. Marusich M.F. J. Biol. Chem. 2002; 277: 33906-33912Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). Half of the final volume of digests from silver- and Sypro Ruby-stained gels (5–8 μl) was further subjected to strong cation exchange ZipTip (Millipore) clean up and concentration. Peptides were eluted directly onto the MALDI targets, and spectra were automatically acquired on a Voyager DE-STRTM as reported previously (17Taylor S.W. Warnock D.E. Glenn G.M. Zhang B. Fahy E. Gaucher S.P. Capaldi R.A. Gibson B.W. Ghosh S.S. J. Proteome Res. 2002; 1: 451-458Crossref PubMed Scopus (74) Google Scholar). Peptide mass fingerprints from base-line-corrected, noise-filtered, and de-isotoped peaks were obtained and then analyzed using the program Protein Prospector, MS-Fit (18Clauser K.R. Baker P. Burlingame A.L. Anal. Chem. 1999; 71: 2871-2882Crossref PubMed Scopus (977) Google Scholar) with and without the Intellical algorithm from Applied Biosystems as described previously (17Taylor S.W. Warnock D.E. Glenn G.M. Zhang B. Fahy E. Gaucher S.P. Capaldi R.A. Gibson B.W. Ghosh S.S. J. Proteome Res. 2002; 1: 451-458Crossref PubMed Scopus (74) Google Scholar). Since many of the NADH dehydrogenase subunits are small, predicted peptide mass fingerprints often contained only a limited number of peptides to match to peaks in the MALDI spectra. In these instances we resorted to manual inspection of the MALDI data as outlined in our earlier study (13Aggeler R. Coons J. Taylor S.W. Ghosh S.S. Garcia J.J. Capaldi R.A. Marusich M.F. J. Biol. Chem. 2002; 277: 33906-33912Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). The remaining portion of selected digests was subject to automated LC-MS/MS analysis using a MicroTech Ultimate LC system coupled to a Finnigan LCQ™ DECA ion trap mass spectrometer equipped with a Finnigan dynamic nanospray source as described previously (13Aggeler R. Coons J. Taylor S.W. Ghosh S.S. Garcia J.J. Capaldi R.A. Marusich M.F. J. Biol. Chem. 2002; 277: 33906-33912Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). Each chromatogram was subsequently analyzed with the program SEQUEST (19Ducret A. Van Oostveen I. Eng J.K. Yates III, J.R. Aebersold R. Protein Sci. 1998; 7: 706-719Crossref PubMed Scopus (280) Google Scholar) as described previously (13Aggeler R. Coons J. Taylor S.W. Ghosh S.S. Garcia J.J. Capaldi R.A. Marusich M.F. J. Biol. Chem. 2002; 277: 33906-33912Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). The requirement for a protein assignment was at least one peptide for a particular protein having anXcorr of >1.7 for a +1 ion,Xcorr of >2.0 for a +2 ion, andXcorr of >3.0 for a +3 ion. In all cases Δcorr must be greater than 0.1. cDNAs were amplified from a human heart Quickclone library (Clontech) using standard PCR protocols. The following pairs of primers were used (i) GRIM-19: 5′-ATG GCG GCG TCA AAG GTG AAG CAG GAC AT, 3′-CTA CGT GTA CCA CAT GAA GCC GTG GC; (ii) NP17.3, 5′-ATG GCG GCT GGG CTG TTT GG, 3′-TCA CTC ATC CTC TGG CAG CTG G; (iii) NDUFA11, 5′-ATG GCG CCG AAG GTT TTT CG, 3′-TCA CAC CTT GGG TTT TGC AAA CAC CT. We have obtained several monoclonal antibodies that can immunocapture complex I from human tissue. For the experiments described here, we used the antibody 17G3D9E12 a mouse IgG1 monoclonal antibody specific for the NDUFA6 subunit and now available from Molecular Probes Inc. This mAb immunoprecipitated 1 μg of enzyme/250 μg of heart mitochondria, a yield of purified organelle membranes that could be obtained routinely from as little as 10 mg of heart tissue. Fig. 1 shows human complex I subunits separated by one-dimensional SDS-PAGE and visualized with silver stain. Under the gel electrophoresis conditions chosen, 29 protein bands were resolved. Individual subunits were identified by slicing out the protein bands, in gel trypsinization, followed by MALDI-TOF MS peptide mass fingerprinting and LC-MS/MS analysis. As shown in TableI, MALDI-TOF spectrometry identified 23 different complex I polypeptides while LC-MS/MS identified 40 subunits. The subunits A1 and B2 were identified by MALDI-TOF but went undetected by LC-MS/MS, and therefore combining the data from these methods results in identification 42 of the 45 putative complex I subunits. Several impurities of the preparation were detected. Two of the bands resolved in Fig. 1 contained IgG and the F1 α and β subunits, respectively. Other contaminants include ANT1 (ADP ATP carrier protein, human heart/skeletal muscle isoform T1), the complex II 70-kDa protein, the complex III core protein 2, and complex IV subunit 5a (see Fig.2). There is recent evidence for supermolecular complexes of the oxidative phosphorylation machineryin situ (20Schägger H. Pfeiffer K. EMBO J. 2000; 19: 1777-1783Crossref PubMed Scopus (1049) Google Scholar), and it appears that at the levels of detergent used (1% lauryl maltoside), which was chosen to minimally disrupt complex I, a small level of association of complexes remains. Nevertheless complex I subunits are greatly concentrated in the immunocapture, and the other oxidative phosphorylation complexes are overwhelmingly present in the supernatant based on Western blotting with specific mAbs to each of these complexes.Table IMALDI-TOF and LC-MS/MS detection of tryptic peptides from complex I subunitsBandMALDI-TOF identificationLC-MS/MS identificationLC-MS/MS peptidesPredicted mature subunit masskDa1(IgG)(IgG)2S1S13077.03(F1α/β)(F1α/β)(55.2/51.8)4V1V11048.65ND5567.06S2S21349.17A10437.18ND4151.69A9A91638.910(ANT1)(ANT1)/ND21(33.0)/39.011S3S3/ND110/326.4/35.712V2V2723.713B9/B10B9/B108/521.7/20.714S7/S8S7/S84/719.8/20.315NP17.31-aNeuronal protein (AAH07362).317.316B8/A8B8/A85/618.8/20.017B7B7816.318B6B6/S45/415.4/15.419B5B5417.02017.2 kDa1-b17.2-kDa protein related to the 13-kDa differentiation association protein (Q9UI09).17.2 kDa1-b17.2-kDa protein related to the 13-kDa differentiation association protein (Q9UI09).1817.121GRIM-191-cGRIM-19 (Q9P0J0).GRIM-191-cGRIM-19 (Q9P0J0).1016.622B4B4515.123A6/C2A6/C25/515.0/14.223A7/ND35/112.4/13.224A11/A5/S55/7/614.9/13.3/12.425S6/B34/410.7/11.326A2/B210.8/8.626A2/V34/110.8/8.427A3/A42/19.3/9.428A11-dThe A4 subunit was detected by MALDI-TOF MS after two-dimensional separation (see Fig. 2).8.029B1B1/AB11/37.0/10.21-a Neuronal protein (AAH07362).1-b 17.2-kDa protein related to the 13-kDa differentiation association protein (Q9UI09).1-c GRIM-19 (Q9P0J0).1-d The A4 subunit was detected by MALDI-TOF MS after two-dimensional separation (see Fig. 2). Open table in a new tab Figure 2Immunopurified human heart complex I separated by two-dimensional SDS-PAGE. 10 μg of complex I proteins were first resolved by a linear pH 3–10 immobilized pH gradient strip and then separated in a second dimension by 15% acrylamide SDS-PAGE. Two representative samples are shown for constancy of separation, followed by staining with Sypro Ruby (A) or silver nitrate (B). MALDI-TOF mass spectrometry identified spots are labeled by their SWISS-PROT gene name and predicted isoelectric point. The MS data for spots A9, A3, and S4 did not contain a sufficient number of tryptic peptide detections for automatic polypeptide assignment. Their positions were predicted by a preliminary two-dimensional analysis of bovine complex I (data not shown) and confirmed for A3 and S4 by manual inspection of the MALDI-TOF data as described by Aggeler et al. (13Aggeler R. Coons J. Taylor S.W. Ghosh S.S. Garcia J.J. Capaldi R.A. Marusich M.F. J. Biol. Chem. 2002; 277: 33906-33912Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar) and for A9 by Western blotting with an A9 specific mAb.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Fig. 2 shows the immunocaptured complex I separated by two-dimensinoal SDS-PAGE. Two different gels are presented that illustrate the reproducibility of the two-dimensional analysis, one stained with Sypro Ruby, the most sensitive protein detection method available, and the second with silver nitrate for comparison. In total, 21 complex I subunits could be resolved reproducibly when loading 10 μg of protein. While these include mostly the large hydrophilic subunits, good coverage of basic polypeptides is also obtained by this single pH 3–10 isoelectric focusing gel method. Hirst and colleagues (11Carrol J. Shannon R.J. Fearnley I.M. Walker J.E. Hirst J. J. Biol. Chem. 2002; 277: 50311-50317Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar) have been able to resolve more subunits of the biochemically purified bovine complex I by two-dimensional electrophoresis than the current study by isolating more complex I allowing analysis after subfractionation and also after heavy loading of pH 3–10 and 6–11 isoelectric focusing gel strips. In all, 27 subunits were fully resolved from each other or from impurities either as single bands in one-dimensional or as one or more unique spots on two-dimensional gels. These include all of the nuclear-encoded subunits in which mutations causing human complex I deficiencies have been identified to date. Also uniquely resolved are the two complex I polypeptides, NDUFA9 and NDUFS4, that are known to undergo post-translational modification (11Carrol J. Shannon R.J. Fearnley I.M. Walker J.E. Hirst J. J. Biol. Chem. 2002; 277: 50311-50317Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar). The pattern of spots seen in Fig. 2 has been reproduced more than 10 times with several different preparations of human heart. Interestingly the NDUFA9 protein and the NDUFB4 protein have isoelectric points significantly different from that calculated from published gene sequences (see Fig.2B). While the DNA sequences of almost all of the putative subunits of complex I have been described, this is the first analysis of the protein composition of the human enzyme. Three proteins were found in our complex I preparations that, until this point, have not been studied in the context of human complex I. These subunits are the GRIM-19 protein, the human neuronal protein NP17.3, and a hitherto unnamed hypothetical protein (GenBankTM accession number NP_783313). The homologous proteins B16.6, ESSS, and B14.7, respectively, have all been found in bovine complex I (11Carrol J. Shannon R.J. Fearnley I.M. Walker J.E. Hirst J. J. Biol. Chem. 2002; 277: 50311-50317Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar, 21Fearnley I.M. Carroll J. Shannon R.J. Runswick M.J. Walker J.E. Hirst J. J. Biol. Chem. 2001; 276: 38345-38348Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar). We have cloned and confirmed the sequence of the cDNAs encoding these proteins from a human heart library (data not shown). Based on the nomenclature developed by Robinson and colleagues (22Duncan A.M. Chow W. Robinson B.H. Cytogenet. Cell Genet. 1992; 60: 212-213Crossref PubMed Scopus (14) Google Scholar, 23Ali S.T. Duncan A.M. Schappert K. Heng H.H. Tsui L.C. Chow W. Robinson B.H. Genomics. 1993; 18: 435-439Crossref PubMed Scopus (30) Google Scholar), the unnamed subunit has been termed NDUFA11, since its bovine counterpart was shown to be a part of the 1α subcomplex. Furthermore, it has been postulated that this protein may be involved in protein import into the mitochondrion (11Carrol J. Shannon R.J. Fearnley I.M. Walker J.E. Hirst J. J. Biol. Chem. 2002; 277: 50311-50317Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar, 24Sasanov L.A. Peak-Chew S.Y. Fearnely I.M. Walker J.E. Biochemistry. 2000; 39: 7229-7235Crossref PubMed Scopus (157) Google Scholar). This subunit nomenclature has been adopted by the Human Genome Nomenclature Committee. There is a critical need for efficient isolation procedures of intact complex I from human tissue and cell culture material for the molecular diagnosis of genetically derived and late onset neurodegenerative diseases that are associated with deficiencies of this component of the electron transport chain. Because of constraints on the amount of tissue that can be harvested from patients, a rapid protocol, which works for small amounts of biopsy material, blood cells, or cell cultured material, is most desirable. In this report, we describe such a method that uses specific monoclonal antibodies to immunocapture complex I in amounts suitable for one-dimensional and two-dimensional electrophoresis followed by mass spectrometry for compositional analysis. We show that the human complex I purified by this immunocapture method contains the same polypeptides described in the studies of the bovine heart enzyme. A total of 42 different component polypeptides were identified by mass spectrometry after electrophoresis (see Table I). The ND6 protein was not identified in the current study but has been located in band 16 of the gel by a mAb specific for this polypeptide (see Fig. 1). This subunit has also has been detected in sucrose gradient purified complex I, when a larger amount of protein than used here was applied to the gel (25Taylor S.W. Fahy E. Zhang B. Glenn G.M. Warnock D.E. Wiley S. Murphy A.N. Gaucher S.P. Capaldi R.A. Gibson B.W. Ghosh S.S. Nat. Biotechnol. 2003; 21: 281-286Crossref PubMed Scopus (598) Google Scholar). The smallest putative subunit, the NDUFC1 protein, was also detected at a position just below band 29 in this other investigation (25Taylor S.W. Fahy E. Zhang B. Glenn G.M. Warnock D.E. Wiley S. Murphy A.N. Gaucher S.P. Capaldi R.A. Gibson B.W. Ghosh S.S. Nat. Biotechnol. 2003; 21: 281-286Crossref PubMed Scopus (598) Google Scholar). In contrast, ND3 was identified by a single cysteine-containing peptide in the current study but not in the other investigation where reduction and alkylation steps were omitted from gel processing procedures (25Taylor S.W. Fahy E. Zhang B. Glenn G.M. Warnock D.E. Wiley S. Murphy A.N. Gaucher S.P. Capaldi R.A. Gibson B.W. Ghosh S.S. Nat. Biotechnol. 2003; 21: 281-286Crossref PubMed Scopus (598) Google Scholar). The only subunit undetected by either study was ND4L, a hydrophobic mitochondrially encoded polypeptide with only a single large 23 residue trypsin fragment. Present in all preparations was the protein GRIM-19, which has been linked to apoptosis, a 17.2-kDa protein homologous to a cell differentiation associated protein, a so-called neuronal protein NP17.3, and a previously uncharacterized complex I subunit NDUFA11. The bovine homologues of these proteins were recently reported in beef heart complex I as our work was in progress and were termed B16.6, B17.2, ESSS, and B14.7, respectively (11Carrol J. Shannon R.J. Fearnley I.M. Walker J.E. Hirst J. J. Biol. Chem. 2002; 277: 50311-50317Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar, 21Fearnley I.M. Carroll J. Shannon R.J. Runswick M.J. Walker J.E. Hirst J. J. Biol. Chem. 2001; 276: 38345-38348Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar, 26Skehel J.M. Fearnley I.M. Walker J.E. FEBS Lett. 1998; 438: 301-305Crossref PubMed Scopus (59) Google Scholar). It is difficult to establish whether any subunit is abona fide subunit or an adventitiously associated impurity of a complex until the stoichiometry of that polypeptide has been determined and/or a key functional role has been identified. In the case of three of the polypeptides listed above, there are functions ascribed to each that are not related to complex I activity, and at least one, GRIM-19, has been found in both the nucleus and the cytosol (27Angell J.E. Lindner D.J. Shapiro P.S. Hofmann E.R. Kalvakolanu D.V. J. Biol. Chem. 2000; 275: 33416-33426Abstract Full Text Full Text PDF PubMed Scopus (196) Google Scholar). Thus our finding that they are all present in the human complex I isolated by a different procedure than that used for the bovine enzyme is significant and supports the notion that they are constituents of complex I. Nevertheless, additional work is necessary to ascertain the function of these ancillary proteins. In summary, we provide rapid purification of complex I from small amounts of human tissue, which will be useful in studying mitochondrial diseases involving complex I and invaluable in screening for post-translational modifications of the complex in neurodegenerative diseases. Importantly, 21 of the 45 subunits can be resolved to unique spots by two-dimensional gels using as little as 10 μg of the enzyme complex. These include all of the nuclear subunits currently found to contain mutations causing diseases and the two subunits with known post-translational modifications. We show that the normal human heart enzyme has the same composition as the bovine counterpart and must contain at least 45 different subunits together having a molecular weight close to 1 million.