NADH:ubiquinone oxidoreductase (complex I) is an entry point for electrons into the respiratory chain in many eukaryotes. It couples NADH oxidation and ubiquinone reduction to proton translocation across the mitochondrial inner membrane. Because complex I deficiencies occur in a wide range of neuromuscular diseases, including Parkinson's disease, there is a clear need for model eukaryotic systems to facilitate structural, functional and mutational studies. In the present study, we describe the purification and characterization of the complexes I from two yeast species, Pichia pastoris and Pichia angusta. They are obligate aerobes which grow to very high cell densities on simple medium, as yeast-like, spheroidal cells. Both Pichia enzymes catalyse inhibitor-sensitive NADH:ubiquinone oxidoreduction, display EPR spectra which match closely to those from other eukaryotic complexes I, and show patterns characteristic of complex I in SDS/PAGE analysis. Mass spectrometry was used to identify several canonical complex I subunits. Purified P. pastoris complex I has a particularly high specific activity, and incorporating it into liposomes demonstrates that NADH oxidation is coupled to the generation of a protonmotive force. Interestingly, the rate of NADH-induced superoxide production by the Pichia enzymes is more than twice as high as that of the Bos taurus enzyme. Our results both resolve previous disagreement about whether Pichia species encode complex I, furthering understanding of the evolution of complex I within dikarya, and they provide two new, robust and highly active model systems for study of the structure and catalytic mechanism of eukaryotic complexes I.

INTRODUCTION

In mitochondria, NADH:ubiquinone oxidoreductase (complex I) catalyses the oxidation of NADH and the reduction of ubiquinone, coupled to the translocation of four protons across the mitochondrial inner membrane, contributing to the protonmotive force [1,2]. Fourteen ‘core subunits’ are conserved throughout the complex I family. They comprise seven membrane-bound subunits, encoded by the mitochondrial genome in eukaryotes, and seven hydrophilic subunits, encoded in the nucleus, which form a domain extending into the mitochondrial matrix. The hydrophilic subunits bind a FMN (flavin mononucleotide), at the active site for NADH oxidation, and eight FeS (iron sulfur) clusters. The structure of the hydrophilic domain from Thermus thermophilus complex I shows how the FeS clusters enable electron transfer from the flavin towards the quinone-binding site [3]. In addition, eukaryotic complexes I contain a variable number of ‘supernumerary’ subunits. Complex I from Bos taurus heart mitochondria has been characterized extensively and contains a total of 45 different subunits [4]. In two fungal model systems, 40 different subunits have been identified in complex I from Yarrowia lipolytica [5], and 39 in the enzyme from Neurospora crassa [6]. Gabaldón et al. [7] have used genomic data to propose how the subunit composition of eukaryotic complex I has evolved; mammalian and fungal complexes I have 21 supernumerary subunits in common, whereas mammals have ten additional subunits and fungi have three. Mutations in both the nuclearly and mitochondrially encoded subunits contribute to the increasing number of complex I deficiencies which have been identified in a wide range of neuromuscular diseases, including Leigh syndrome, leukodystrophy and Parkinson's disease [8,9].

Studies of eukaryotic complex I rely on model systems to provide sufficient quantities of the enzyme for investigation of its structure and mechanism, and to provide insights into the dysfunction of the human enzyme. The B. taurus enzyme is highly homologous with the human enzyme [4], and has proved invaluable in the elucidation of its protein composition, as well as for spectroscopic and functional studies. However, lack of control over the source material, and the impracticality of genetic manipulation, are significant limitations. The intensively studied yeast, Saccharomyces cerevisiae, does not encode complex I, but N. crassa [10], and more recently Y. lipolytica [11], have provided two genetically manipulable fungal enzymes for study.

Figure 1 shows a phylogenetic tree for the dikarya (constructed using data from the OrthoDB database [12]), showing that N. crassa belongs to the pezizomycotina, and Y. lipolytica to the saccharomycetes. Y. lipolytica diverged relatively early in evolution from the saccharomycetaceae, which include Saccharomyces, Kluveromyces, Candida and Pichia. Some of the saccharomycetaceae have lost the complex I genes during evolution, and, in fact, no complex I has been isolated from any of them. Thus although the Pichia genus shows many promising characteristics for the development of a complex I model system (see below), it has not been clear, until now, whether Pichia species even express complex I. Sequences for some of the subunits of complex I which are encoded in the mitochondrial genome (known as the ND subunits) have been detected in several species, including Pichia pastoris [13], but attempts to identify an NADH dehydrogenase from P. pastoris which is inhibited by a canonical complex I inhibitor, rotenone, were unsuccessful [14].

Phylogenetic tree showing the distribution of complex I within the subkingdom of dikarya

Figure 1
Phylogenetic tree showing the distribution of complex I within the subkingdom of dikarya

Species in bold have complex I subunits encoded in their nuclear and mitochondrial genomes, and they encode an alternative oxidase and at least one alternative, single subunit NADH:ubiquinone oxidoreductase. Species in grey are able to anaerobically ferment, and they only encode alternative NADH:ubiquinone oxidoreductases. * denotes the presence of the genes for the three subunit mitochondrial transhydrogenase, and the NURM gene, which has been observed previously only in N. crassa complex I. The phylogenetic tree was created using the OrthoDB database [12].

Figure 1
Phylogenetic tree showing the distribution of complex I within the subkingdom of dikarya

Species in bold have complex I subunits encoded in their nuclear and mitochondrial genomes, and they encode an alternative oxidase and at least one alternative, single subunit NADH:ubiquinone oxidoreductase. Species in grey are able to anaerobically ferment, and they only encode alternative NADH:ubiquinone oxidoreductases. * denotes the presence of the genes for the three subunit mitochondrial transhydrogenase, and the NURM gene, which has been observed previously only in N. crassa complex I. The phylogenetic tree was created using the OrthoDB database [12].

The two species of yeast described in the present study are P. pastoris and P. angusta (previously named Hansenula polymorpha). They are both methylotrophic ascomycetes, which are able to catabolize methanol as their sole carbon source. They are obligate aerobes, and grow on inexpensive, simple medium to very high cell densities, particularly in batch-fed conditions. Notably, Pichia species grow almost exclusively as yeast-like spheroidal cells, whereas N. crassa grows as filaments, and Y. lipolytica switches between yeast and filamentous growth in response to various environmental stresses [15], a transition known to affect cellular metabolism and protein expression [16]. Both species can be used for overexpression, commonly controlled by the promoter for alcohol oxidase, and induced with methanol [17]. In addition, P. angusta can grow at temperatures as high as ∼50 °C [18], and therefore may express more stable proteins as part of its heat tolerance. In the present study, we describe the purification and characterization of the complexes I from P. pastoris and P. angusta, discuss their suitability as eukaryotic model systems, and compare their properties with those of other members of the complex I family.

EXPERIMENTAL

All chemicals were from Sigma–Aldrich or BDH (unless otherwise stated), and chromatographic media were from GE Healthcare. Buffer pH values are reported at room temperature (20 °C).

Yeast strains and growth

Haploid P. pastoris strain X33 (Invitrogen) was grown at 30 °C in 1% yeast extract (ForMedium), 2% peptone (ForMedium) and 2% glycerol (VWR International), initially at pH 5. A 60 litre volume of medium, in a 70 litre Applikon ADI 1075 fermentor stirred at 500 rev·min−1, maintained at 80% pO2, was inoculated with 2 litres of culture, and cells were grown until early stationary phase (approx. 24 h). P. angusta strain A-16 (NCYC 2310), a haploid leucine auxotroph of CBS4732, was from The National Collection of Yeast Cultures (U.K.). Cells were grown at 40 °C in 800 ml aliquots in 2 litre flasks (shaken at 250 rev·min−1) until mid-stationary phase (approx. 48 h) in 1% yeast extract, 2% peptone and 4% glucose, initially at pH 5. Y. lipolytica strain GB10 was provided by Professor U. Brandt (Universität Frankfurt, Fachbereich Medizin, Zentrum der Biologischen Chemie, Frankfurt am Main, Germany). The strain was grown at 27 °C in 2% yeast extract, 4% peptone and 4% glucose, initially at pH 5.5, in the 70 litre fermentor stirred at 500 rev·min−1, maintained at 2% pO2 (once this value was reached), until late-exponential phase (approx. 20 h). All cells were harvested by centrifugation (5000 g for 15 min), twice resuspended in Milli-Q water and recentrifuged (5000 g for 15 min). If necessary, cell pellets were frozen at −80 °C. The yields were approx. 41, 27 and 29 g of wet cells per litre of medium for P. pastoris, P. angusta and Y. lipolytica, respectively. All subsequent steps were performed at 4 °C.

Isolation of mitochondrial membranes

Methods for cell disruption and membrane isolation were based on those described for Y. lipolytica [19,20]. Cells were resuspended at 500 g·l−1, in 20 mM Mops (pH 7.2), 400 mM sorbitol (NBS Biologicals), 0.2% BSA, 5 mM EDTA, 1 mM benzamidine and 1 mM ε-caproic acid. Then, 1 mM PMSF was added immediately prior to cell disruption in a Dyno®-Mill bead mill (Willy A. Bachofen U.K.) maintained below 10 °C (two passes at 10 ml·min−1). Cell debris was removed by centrifugation (5000 g for 15 min), and then the membranes were collected by centrifugation [1 h at 40000 rev·min−1 in a Beckman Coulter Ti45 rotor (for P. pastoris and Y. lipolytica) or 20000 rev·min−1 in a SS34 Sorvall rotor (for P. angusta)], rehomogenized in 20 mM Mops (pH 7.2), 5 mM EDTA, 1 mM benzamidine and 1 mM ε-caproic acid, and recollected by recentrifugation. If necessary, membrane pellets were frozen at −80 °C.

Complex I purification

Mitochondrial membranes were homogenized in 20 mM sodium citrate (pH 6.5), to 20 mg of protein·ml−1, and solubilized for 45 min by the addition of DDM (dodecyl-β-D-maltoside; Glycon GmbH) to 2.3% (P. pastoris) or 2.9% (P. angusta). Following centrifugation [42000 rev·min−1 (Beckman Coulter Ti45 rotor) for 45 min], the supernatant was passed through a MacroCap SP cation-exchange column, and then dialysed against 20 mM Tris/HCl (pH 7.5) for at least 3 h. Then, it was loaded on to a Q Sepharose high-performance anion-exchange column, pre-equilibrated in 20 mM Tris/HCl (pH 7.5) and 0.15% DDM, and eluted at 2 ml·min−1. First, 0.27 M NaCl (P. pastoris) or 0.22 M NaCl (P. angusta) was applied until the 280 and 420 nm absorbancies became negligible, then complex I was eluted in a sharp gradient to 0.33 M NaCl (P. pastoris) or 0.27 M NaCl (P. angusta), and confirmed using its dNADH (deamino-NADH):HAR [hexaammineruthenium (III)] oxidoreductase activity. Pooled fractions were concentrated (100 kDa cut-off Vivaspin centrifugal concentrators; Sartorius AG), and then injected on to a Superose 6 size-exclusion column [0.24 ml·min−1, in 20 mM Mops (pH 7.5), 0.05% DDM and 150 mM NaCl]. Monomeric complex I eluted from 11 to 14 ml. If necessary, these complex-I-containing fractions were pooled, concentrated and reapplied to the column, providing a single, sharp peak. Complex I from Y. lipolytica was isolated as described previously, by affinity purification on a Ni-NTA (Ni2+-nitrilotriacetate) column using a hexahistidine tag on the C-terminus of the NUGM subunit (throughout the paper subunits are named according to their SwissProt codes, and the names for the B. taurus homologues are presented alongside where appropriate) [20,21], followed by size-exclusion chromatography on a Superose 6 column. B. taurus complex I was prepared as described previously [22].

Kinetic measurements

Kinetic measurements were carried out at 32 °C, either in 1 ml cuvettes (Ocean Optics diode array spectrometer), or in 200 μl wells (Molecular Devices microtitre plate reader). dNADH and NADH were added to 100 μM for ferricyanide {FeCN [hexacyanoferrate (III)]}, HAR and DQ (decylubiquinone) measurements, and HAR, FeCN and DQ to 3.5 mM, 2 mM and 100 μM respectively. For measurements using DQ, soya bean asolectin was added to 0.35 mg·ml−1 from a 10 mg·ml−1 stock solution in 20 mM Tris/HCl (pH 7.5) and 1% (w/v) CHAPS (Anatrace). Rotenone or piericidin A were added to 5 or 2 μM respectively, from ethanolic stock solutions. Initial rates of NADH oxidation (ε340–380nm=4.81 mM−1·cm−1) were obtained using linear regression. H2O2 and superoxide production were measured at pH 7.5, using 30 μM NADH and either 10 μM Amplex Red and 2 units·ml−1 horseradish peroxidase (Invitrogen) or 50 μM acetylated cytochrome c, as described previously [23].

To produce proteoliposomes containing complex I, 440 μl of 10 mg·ml−1 soya bean asolectin (Avanti Polar Lipids) in 20 mM K-Mops (potassium Mops) at pH 7.5 was homogenized with 60 μl of 10% octyl glucoside. Then, 50 μl of 2 mg·ml−1 complex I was added, and the detergent was removed by three successive additions of 50 mg of Amberlite XAD 2 biobeads (Supelpak) at 1 h intervals. The proteoliposomes were separated from the biobeads using a syringe fitted with a filter, collected by centrifugation (1 h at 45000 rev·min−1 in a Beckman-Coulter MLA 130 rotor), and then resuspended in 100 μl of K-Mops. RCRs (respiratory control ratios) were calculated from the rate of NADH:DQ oxidoreduction, measured (as described above) in the presence and absence of 7 μg·ml−1 gramicidin (RCR=uncoupled rate/coupled rate). The quenching of the ACMA (9-amino-6-chloro-2-methoxyacridine) fluorescence at 475 nm was measured using an excitation wavelength of 430 nm, in the presence of 500 nM ACMA and 5 μM valinomycin. Typically, 20 μl of resuspended proteoliposomes were added to 2 ml of assay buffer. When required, FCCP (carbonyl cyanide p-trifluoromethoxyphenylhydrazone; 8 μM) or piericidin A (2 μM) were added.

EPR spectroscopy

EPR samples (∼15 mg·ml−1) were reduced anaerobically with 5 mM NADH, and frozen immediately. Spectra were recorded on a Bruker EMX X-band spectrometer using an ER 4119HS cavity, maintained at low temperature by an ESR900 continuous-flow liquid helium cryostat (Oxford Instruments); the sample temperature was measured with a calibrated Cernox resistor (Lake Shore Cryotronics).

Analytical methods

Protein concentrations were measured using the Pierce BCA (bicinchoninic acid) assay. FMN concentrations were analysed fluorometrically using the method of Burch [24], modified for the plate reader. Iron concentrations were determined using ferene {5,5′[3-(2-pyridyl)-1,2,4-triazine-5,6 diyl]-bis-2-furansulfonate}, by the method of Pieroni et al. [25] (except the protein was denatured in 1% HCl and 1.5% trichloroacetic acid, the reducing agent was 0.5% hydroxylamine, and 5 mM thiosemicarbazide was added to bind any copper present).

Protein analyses

SDS/PAGE was carried out on an 18–22% acrylamide gradient, and gels were stained with 0.2% Coomassie Blue R250. Peptide mass fingerprinting and tandem MS were carried out using an Applied Biosystems/MDS SCIEX model 4800 Plus MALDI–TOF (matrix-assisted laser-desorption ionization time-of-flight)/TOF mass spectrometer. Database matches to peptide mass fingerprinting and tandem MS data were identified using Mascot (Matrix Science) [26]. Sequences for proposed subunits of P. angusta complex I were identified by performing tBLASTn searches to compare known Y. lipolytica and N. crassa sequences with the genome survey sequence information from P. angusta which is available on the NCBI website.

RESULTS

The presence of both complex I and alternative NADH dehydrogenases in Pichia species

Alternative NADH dehydrogenases are single subunit, non-energy transducing NADH:ubiquinone oxidoreductases which are found in many aerobically respiring organisms, either alongside complex I, or, in cases such as S. cerevisiae, as a replacement for it (see Figure 1) [27,28]. Alternative NADH dehydrogenases complicate the identification, analysis and purification of complex I, because the catalytic properties of the two enzymes overlap. Table 1 presents the catalytic properties of mitochondrial membranes prepared from P. pastoris and P. angusta; in both cases approx. 90% of the NADH-binding sites were accessible to the NADH, FeCN and HAR reactants. Three different electron acceptors were used: FeCN and HAR are considered to react at, or close to, the flavin site in complex I, so, unlike DQ, an analogue of the physiological ubiquinone, they do not rely on the functional integrity of the quinone-binding site or the membrane domain. First, based on a study of the NADH:ubiquinone oxidoreductases in Escherichia coli, dNADH is assumed to react only with complex I [29], so comparing the rates of NADH and dNADH:DQ oxidoreduction by the P. pastoris membranes suggests that ∼80% of the NADH:DQ oxidoreductase activity is due to complex I. In contrast, for P. angusta the rate with dNADH is only ∼10% of that with NADH, indicating that one or more alternative NADH dehydrogenases contribute significantly. There is little difference, in either case, between the rates of dNADH and NADH:HAR oxidoreduction, consistent with previous suggestions that HAR does not react with the alternative enzymes [11]. dNADH:FeCN oxidoreduction is significantly faster than dNADH:HAR oxidoreduction in both species, suggesting that FeCN reduction by the Pichia complexes I is much faster than HAR reduction. Furthermore, additional FeCN reductases may be present, as suggested by comparison of the dNADH and NADH:FeCN oxidoreductase activities. Note that the concentrations of FeCN and HAR used were chosen because they give close to maximal rates for complex I from B. taurus; an extensive investigation of the kinetics of the two Pichia enzymes has not been undertaken so far. Secondly, NADH:DQ and dNADH:DQ oxidoreduction by the P. pastoris membranes is 85 and 90% sensitive to the complex I inhibitor piericidin A, consistent with the results described above; the weaker inhibition from rotenone suggests that it is a relatively poor inhibitor in this case. NADH:DQ oxidoreduction by P. angusta is not inhibited strongly by either compound, although, as expected, the dNADH:DQ activity is more strongly inhibited. Taken together, the results in Table 1 indicate that only 10–20% of the NADH:DQ oxidoreductase activity from the P. angusta membranes is due to complex I. Interestingly, dNADH:HAR oxidoreduction by the P. angusta membranes was optimized by harvesting the cells during stationary phase, and by growing them under high aeration. Similar observations were made previously for P. jadinii (also known as Torulopsis utilis or Candida utilis) [30,31] and N. crassa [32]. In particular, mitochondria from P. jadinii cells harvested in stationary phase displayed piericidin-A-sensitive NADH oxidation and the complex I EPR signals N1b and N2, but mitochondria from cells harvested in exponential phase did not.

Table 1
Specific activities of the mitochondrial membranes from P. pastoris and P. angusta

Values are means±S.D. Conditions: 20 mM Tris/HCl (pH 7.5), 100 μM NADH or dNADH, 100 μM DQ, 3.5 mM HAR, 2 mM FeCN, 5 μM rotenone and 2 μM piericidin A, 32 °C. The accessibility of the NADH-binding sites in the two preparations was determined from the average increase in the rates of dNADH:HAR, dNADH:FeCN and NADH:FeCN oxidoreduction, following solubilization of the membranes in 3% DDM.

 Catalytic activity (μmol of NADH·min−1·mg−1
(a)     
 Catalytic activity (μmol of NADH·min−1·mg−1
Activity P. pastoris P. angusta 
NADH:FeCN oxidoreduction 9.39±0.56 9.65±1.02 
dNADH:FeCN oxidoreduction 6.85±0.18 6.88±0.88 
NADH:HAR oxidoreduction 1.00±0.07 1.41±0.18 
dNADH:HAR oxidoreduction 1.14±0.07 1.34±0.14 
NADH:DQ oxidoreduction 1.27±0.07 2.16±0.40 
dNADH:DQ oxidoreduction 1.00±0.07 0.27±0.14 
NADH:DQ oxidoreduction+piericidin A 0.19±0.06 (85%) 1.63±0.31 (25%) 
dNADH:DQ oxidoreduction+piericidin A 0.10±0.04 (90%) 0.10±0.01 (63%) 
NADH:DQ oxidoreduction+rotenone 0.95±0.52 (26%) 1.71±0.40 (21%) 
dNADH:DQ oxidoreduction+rotenone 0.28±0.03 (72%) 0.12±0.02 (56%) 
Accessibility of NADH-binding sites 86.2±1.8% 87.1±8.3% 
 Catalytic activity (μmol of NADH·min−1·mg−1
(a)     
 Catalytic activity (μmol of NADH·min−1·mg−1
Activity P. pastoris P. angusta 
NADH:FeCN oxidoreduction 9.39±0.56 9.65±1.02 
dNADH:FeCN oxidoreduction 6.85±0.18 6.88±0.88 
NADH:HAR oxidoreduction 1.00±0.07 1.41±0.18 
dNADH:HAR oxidoreduction 1.14±0.07 1.34±0.14 
NADH:DQ oxidoreduction 1.27±0.07 2.16±0.40 
dNADH:DQ oxidoreduction 1.00±0.07 0.27±0.14 
NADH:DQ oxidoreduction+piericidin A 0.19±0.06 (85%) 1.63±0.31 (25%) 
dNADH:DQ oxidoreduction+piericidin A 0.10±0.04 (90%) 0.10±0.01 (63%) 
NADH:DQ oxidoreduction+rotenone 0.95±0.52 (26%) 1.71±0.40 (21%) 
dNADH:DQ oxidoreduction+rotenone 0.28±0.03 (72%) 0.12±0.02 (56%) 
Accessibility of NADH-binding sites 86.2±1.8% 87.1±8.3% 

Purification and confirmation of the complexes I from P. pastoris and P. angusta

dNADH:HAR oxidoreductase activities were relied upon to guide the development of purification protocols for the two complexes I (see Figure 2). Methods for isolating the complexes I were based on the established protocol for B. taurus complex I: solubilization with DDM, anion-exchange chromatography and size-exclusion chromatography [22]. Passing the solubilized membranes through a MacroCap cation-exchange column, prior to application to the anion-exchange column, improved the efficiency of the anion-exchange step, particularly for P. angusta. For P. pastoris, dNADH:HAR oxidoreduction was observed in only one of the major peaks eluting from the anion-exchange column (270–300 mM NaCl). For P. angusta, dNADH:HAR oxidoreductase activity eluted in two major peaks, at 50–120 mM NaCl, and 220–270 mM NaCl. Only the latter peak catalysed dNADH:DQ oxidoreduction, and the results below confirm it was complex I. Typically, 1 g of wet cells yielded 12.7 μg of P. pastoris complex I, or 7.6 μg of P. angusta complex I. In comparison, Y. lipolytica has been reported to yield ∼30 μg per g of wet cells by conventional methods [33] or 50–100 μg by affinity chromatography [20].

Variation of the specific dNADH:HAR oxidoreductase activity during preparation of the complexes I from P. pastoris and P. angusta

Figure 2
Variation of the specific dNADH:HAR oxidoreductase activity during preparation of the complexes I from P. pastoris and P. angusta

Memb., mitochondrial membranes; Sol. Memb., membranes following solubilization in DDM; QS, pooled fractions from the Q-sepharose ion-exchange chromatography step; SE2, pooled fractions from the second size-exclusion chromatography step.

Figure 2
Variation of the specific dNADH:HAR oxidoreductase activity during preparation of the complexes I from P. pastoris and P. angusta

Memb., mitochondrial membranes; Sol. Memb., membranes following solubilization in DDM; QS, pooled fractions from the Q-sepharose ion-exchange chromatography step; SE2, pooled fractions from the second size-exclusion chromatography step.

During analytical size-exclusion chromatography both Pichia complexes I eluted as single, symmetrical peaks at 11.9±0.2 ml (P. pastoris) or 12.1±0.1 ml (P. angusta). The well-characterized complexes from B. taurus and Y. lipolytica eluted at 11.8 and 11.9 ml respectively. Their protein molecular masses of ∼980 kDa [34] and ∼940 kDa [5] respectively, were included in a molecular mass calibration to estimate masses for the P. pastoris and P. angusta enzymes of 920 and 890 kDa respectively. The FMN and iron contents of the purified enzymes were 1.0±0.1 FMN and 33.2±3.6 Fe per P. pastoris complex I, and 1.1± 0.2 FMN and 33.4±4.5 Fe per P. angusta complex I, using these estimated molecular masses.

The EPR spectra of both the P. pastoris and P. angusta enzymes (see Figure 3), are typical complex I spectra, and confirm, unambiguously, the identity of the purified enzyme [35]. Comparison with the spectra from B. taurus complex I [35,36] is straightforward, and four signals (N1b, N2, N3 and N4) are immediately identified by their similar g-values. The N5 signal, observed in complex I from B. taurus and Y. lipolytica [33], and the N1a signal, observed in complex I from E. coli [37], are not observed. The slowly relaxing N2 spectrum, assigned to the [4Fe–4S] cluster in NUKM (PSST), is evident at approx. 12 K, with gz-values matching those from the B. taurus and Y. lipolytica enzymes (gz=2.06 for both P. pastoris and P. angusta). At 40 K the N1b spectrum, from the [2Fe–2S] cluster in NUAM (75 kDa), is observed, again with gz-values matching those from B. taurus and Y. lipolytica complex I (gz=2.03 for both P. pastoris and P angusta). At lower temperatures (most clearly 7 and 9 K) two further [4Fe–4S] spectra are observed. The gx value for N3, assigned to the [4Fe–4S] cluster in NUBM (51 kDa), is 1.86 for both P. pastoris and P. angusta (the gz signal cannot be distinguished). The origin of N4 is controversial [38,39] but the gz and gx signals of 2.11 and 1.89 (in both cases) match closely to those from Y. lipolytica and B. taurus complex I.

EPR spectra showing signals from the FeS clusters in the complexes I from P. pastoris and P. angusta reduced by NADH

Figure 3
EPR spectra showing signals from the FeS clusters in the complexes I from P. pastoris and P. angusta reduced by NADH

Spectra are presented at different temperatures because each cluster signal is only observed over a limited range. The signals have been identified and labelled by matching their g-values and temperature-dependence to those from the well-characterized complexes I from B. taurus and Y. lipolytica. Conditions: ∼15 mg·ml−1 complex I, 5 mM NADH, 20 mM Mops (pH 7.5), 150 mM NaCl and 0.02% DDM; microwave power 1 mW, conversion time 81.92 ms, time constant 20.48 ms, modulation amplitude 10 G, microwave frequency ∼9.38 MHz.

Figure 3
EPR spectra showing signals from the FeS clusters in the complexes I from P. pastoris and P. angusta reduced by NADH

Spectra are presented at different temperatures because each cluster signal is only observed over a limited range. The signals have been identified and labelled by matching their g-values and temperature-dependence to those from the well-characterized complexes I from B. taurus and Y. lipolytica. Conditions: ∼15 mg·ml−1 complex I, 5 mM NADH, 20 mM Mops (pH 7.5), 150 mM NaCl and 0.02% DDM; microwave power 1 mW, conversion time 81.92 ms, time constant 20.48 ms, modulation amplitude 10 G, microwave frequency ∼9.38 MHz.

Catalytic activities of the purified enzymes

Table 2 summarizes the catalytic properties of the complexes I from P. pastoris and P. angusta, and compares them with those from Y. lipolytica and B. taurus (measured in the present study). All the rates of NADH:HAR oxidoreduction are similar, but the rates of NADH:FeCN oxidoreduction are significantly higher for the two Pichia enzymes. Note that our reported values refer to only a single condition, so they are only an indication of the relative catalytic activities of the enzymes, and do not allow the cause of the variation to be identified. Using the structure of the hydrophilic domain of T. thermophilus complex I as a template, sequence comparisons for the protein around the NADH-binding site in B. taurus, Y. lipolytica and P. angusta have not provided any insight. The rates of NADH:DQ oxidoreduction are similar in all cases, except for the P. pastoris enzyme, which displays a markedly higher rate. Note that the highest published values for Y. lipolytica (6–7 μmol·min−1·mg−1 [40]) exceed those achieved here, but are comparable with our values from P. pastoris. Interestingly, all four enzymes are close to completely inhibited by 2 μM piericidin A, but the three yeast enzymes are much less effectively inhibited by rotenone than the B. taurus enzyme (see below).

Table 2
Specific activities of the complexes I purified from P. pastoris and P. angusta, and comparison with the enzymes from Y. lipolytica and B. taurus

Values are means±S.D. Conditions: 20 mM Tris/HCl (pH 7.5), 100 μM NADH, 100 μM DQ, 3.5 mM HAR, 2 mM FeCN, 5 μM rotenone and 2 μM piericidin A, 32 °C. H2O2 measurements: 20 mM Tris/HCl (pH 7.5), 30 μM NADH, 10 μM Amplex Red and 2 units·ml−1 horseradish peroxidase. Superoxide measurements: 20 mM Tris/HCl (pH 7.5), 30 μM NADH and 50 μM partially acetylated cytochrome c. Values for Y. lipolytica and B. taurus were measured alongside those for P. pastoris and P. angusta for direct comparison. (a)

 Catalytic activity (μmol of NADH·min−1·mg−1
Activity P. pastoris P. angusta Y. lipolytica B. taurus 
NADH:FeCN oxidoreduction 302.6±22.9 248.9±22.9 140.8±3.1 111.1±8.3 
NADH:HAR oxidoreduction 70.5±9.4 63.4±3.2 68.8±13 72.5±6.8 
dNADH:HAR oxidoreduction 68.6±7.4 69.0±5.3 
NADH:DQ oxidoreduction 6.27±0.38 4.00±0.04 3.80±0.55 3.82±0.17 
dNADH:DQ oxidoreduction 7.00±0.86 3.27±0.26 3.55±0.40 3.62±0.50 
dNADH:DQ oxidoreduction+piericidin A 0.37±0.30 (95%) 0.06±0.01 (98%) 0.28±0.02 (92%) 0.24±0.11 (93%) 
(b)     
 H2O2 or superoxide production (nmol·min−1·mg−1
Activity P. pastoris P. angusta Y. lipolytica B. taurus 
H2O2 63.3±4.9 76.3±14.7 42.5±4.4 23.6±5.6 
Superoxide 105.0±11.6 127.2±7.5 90.3±2.8 41.3±1.8 
NADH:O2 oxidoreduction 55.1±5.8 68.6±15.6 41.3±1.8 20.1±1.7 
 Catalytic activity (μmol of NADH·min−1·mg−1
Activity P. pastoris P. angusta Y. lipolytica B. taurus 
NADH:FeCN oxidoreduction 302.6±22.9 248.9±22.9 140.8±3.1 111.1±8.3 
NADH:HAR oxidoreduction 70.5±9.4 63.4±3.2 68.8±13 72.5±6.8 
dNADH:HAR oxidoreduction 68.6±7.4 69.0±5.3 
NADH:DQ oxidoreduction 6.27±0.38 4.00±0.04 3.80±0.55 3.82±0.17 
dNADH:DQ oxidoreduction 7.00±0.86 3.27±0.26 3.55±0.40 3.62±0.50 
dNADH:DQ oxidoreduction+piericidin A 0.37±0.30 (95%) 0.06±0.01 (98%) 0.28±0.02 (92%) 0.24±0.11 (93%) 
(b)     
 H2O2 or superoxide production (nmol·min−1·mg−1
Activity P. pastoris P. angusta Y. lipolytica B. taurus 
H2O2 63.3±4.9 76.3±14.7 42.5±4.4 23.6±5.6 
Superoxide 105.0±11.6 127.2±7.5 90.3±2.8 41.3±1.8 
NADH:O2 oxidoreduction 55.1±5.8 68.6±15.6 41.3±1.8 20.1±1.7 

The markedly more active P. pastoris complex I could be reconstituted into liposomes, using methods based on those for Y. lipolytica [21] (see the Experimental section), to provide well-coupled proteoliposomes, typically with RCRs (the ratio of the rates of NADH oxidation in the presence and absence of an uncoupler) of ∼2. Proteoliposomes formed from the less active P. angusta enzyme did not show any coupling (RCR=1). Figure 4 shows typical traces monitoring the fluorescence of ACMA, a measure of ΔpH across the liposomal membrane [41], for reconstituted P. pastoris complex I. In the presence of valinomycin (so that potassium ion transport collapses Δψ and facilitates ΔpH formation) the ACMA fluorescence is rapidly quenched when catalysis is initiated. It is re-established upon the addition of a proton ionophore, FCCP, to dissipate the protonmotive force, or an inhibitor (piericidin A) to prevent further proton translocation. The results shown in Figure 4 are strong evidence for proton translocation by complex I from P. pastoris.

Proton pumping by complex I from P. pastoris reconstituted into liposomes

Figure 4
Proton pumping by complex I from P. pastoris reconstituted into liposomes

The formation of ΔpH was monitored by the fluorescence of ACMA, which is quenched upon accumulation into the lumen of the liposomes. A small decrease in fluorescence, not related to the formation of ΔpH, is observed upon the addition of DQ. Significant quenching (ΔpH formation) is observed in the presence of both NADH and DQ, and is dissipated by the presence of either the proton ionophore FCCP, or the complex I inhibitor piericidin A. The potassium ionophore, valinomycin, is present to collapse Δψ and allow ΔpH formation. The traces all begin at zero fluorescence intensity and have been offset from one another along the y-axis. Conditions: 100 μM NADH, 100 μM DQ, 500 nM ACMA, 5 μM valinomycin, 20 mM Mops (pH 7.5) and 50 mM KCl, excitation 430 nm, emission 475 nm.

Figure 4
Proton pumping by complex I from P. pastoris reconstituted into liposomes

The formation of ΔpH was monitored by the fluorescence of ACMA, which is quenched upon accumulation into the lumen of the liposomes. A small decrease in fluorescence, not related to the formation of ΔpH, is observed upon the addition of DQ. Significant quenching (ΔpH formation) is observed in the presence of both NADH and DQ, and is dissipated by the presence of either the proton ionophore FCCP, or the complex I inhibitor piericidin A. The potassium ionophore, valinomycin, is present to collapse Δψ and allow ΔpH formation. The traces all begin at zero fluorescence intensity and have been offset from one another along the y-axis. Conditions: 100 μM NADH, 100 μM DQ, 500 nM ACMA, 5 μM valinomycin, 20 mM Mops (pH 7.5) and 50 mM KCl, excitation 430 nm, emission 475 nm.

Rates of ROS (reactive oxygen species) generation by the complexes I from both Pichia species were measured as H2O2 (detected using Amplex Red) or superoxide (detected using acetylated cytochrome c), and verified by comparison with independent measurements of the rates of NADH oxidation (see Table 2). Within experimental error, the rates of NADH oxidation matched the rates of H2O2 production in a 1:1 ratio, validating both measurements. In both cases, the rates of superoxide formation (one electron) are close to twice the rates of H2O2 formation (two electrons), indicating that the enzymes from P. pastoris and P. angusta, like the enzyme from B. taurus [23], produce ROS predominantly as superoxide. Table 2 shows that complex I from Y. lipolytica falls into the same class, although the rate of superoxide production measured in the present study (90 nmol·min−1·mg−1, validated using the rate of NADH oxidation) is significantly higher than reported previously (18 nmol·min−1·mg−1; [42]). The cause of this discrepancy is not clear. Table 2 shows that the three yeast complexes I produce significantly more ROS than the B. taurus enzyme.

The protein composition of the complexes I from P. pastoris and P. angusta

Figure 5 shows SDS/PAGE analysis of the protein composition of the complexes I from B. taurus, Y. lipolytica, P. pastoris and P. angusta. The overall similarity of the pattern, particularly for the high-molecular-mass subunits, is striking, although differences in mass and/or migration become apparent lower down the gel. Bands corresponding to the NUAM (75 kDa), NUBM (51 kDa), NUCM (49 kDa), NUDM/NUEM (42/39 kDa) and NUGM (30 kDa) subunits are clearly visible in all cases (NUBM and NUCM co-migrate in Y. lipolytica), but one of the bands present in the mammalian enzyme, probably the 42 kDa subunit, does not appear in the lanes from the three yeast species.

Analysis of the subunit composition of the complexes I from P. pastoris and P. angusta, and comparison with the enzymes from B. taurus and Y. lipolytica

Figure 5
Analysis of the subunit composition of the complexes I from P. pastoris and P. angusta, and comparison with the enzymes from B. taurus and Y. lipolytica

SDS/PAGE analysis was performed on a 12–22% gradient gel, and the bands excised and analysed by peptide mass fingerprinting and tandem MS. Peptides were identified using publicly available genome sequences (see Supplementary Table S1 at http://www.BiochemJ.org/bj/422/bj4220151add.htm).

Figure 5
Analysis of the subunit composition of the complexes I from P. pastoris and P. angusta, and comparison with the enzymes from B. taurus and Y. lipolytica

SDS/PAGE analysis was performed on a 12–22% gradient gel, and the bands excised and analysed by peptide mass fingerprinting and tandem MS. Peptides were identified using publicly available genome sequences (see Supplementary Table S1 at http://www.BiochemJ.org/bj/422/bj4220151add.htm).

The complexes I from B. taurus and Y. lipolytica have already been characterized extensively, using publically available sequence data, but there is little available for P. pastoris and P. angusta. In order to identify as many subunits as possible in the two Pichia complexes, all of the visible protein bands were excised from the gels, digested with trypsin, and subjected to MALDI–TOF and tandem MS analysis. For both P. pastoris and P. angusta the observed peptide masses (from MALDI–TOF MS) were searched (for matches within 70 p.p.m.) against the NCBI non-redundant sequence database, and against an in-house database of expected complex I subunits from P. angusta (derived from the NCBI genome survey sequences by comparison with known sequences from other species). Then, the closest matches to the tandem MS data (within 120 p.p.m.) were re-evaluated using error tolerance searches, allowing for modifications and amino acid substitutions between species. A summary of the data assigned is presented in Supplementary Table S1 (at http://www.BiochemJ.org/bj/422/bj4220151add.htm). All seven hydrophilic core subunits, and two supernumerary subunits, were identified in the enzyme from P. pastoris. Six of the hydrophilic core subunits [the exception being NUKM (PSST)], and nine supernumerary subunits were identified in the enzyme from P. angusta. The subunits of P. pastoris complex I were identified using the similarity of observed peptide masses to predicted peptide masses from other fungal species, including P. stipitis [NUAM (75 kDa), NUBM (51 kDa), NUCM (49 kDa), NUEM (39 kDa), NUGM (30 kDa), NUHM (24 kDa), NUYM (18 kDa), NUIM (TYKY) and NUKM (PSST)], Lodderomyces elongisporus (NUAM), Debaryomyces hansenii (NUBM, NUCM and NUHM), Y. lipolytica (NUBM, NUCM and NUIM), P. guilleromondii (NUEM), N. crassa (NUKM), Aspergillus fumigatus (NUKM) and Candida albicans (NUYM). These matches confirm that P. pastoris complex I contains a number of canonical complex I subunits, but the lack of complete sequences for them precludes further discussion about its subunit composition and homology with other species. In the same way, masses from five subunits in P. angusta complex I were found to be similar to those in other species: P. stipitis [NUAM (75 kDa) and NUCM (49 kDa)], N. crassa [NUBM (51 kDa)], D. hansenii (NUCM) Dekkera bruxellensis [NUHM (24 kDa)], Magnaporthe grisea (NUHM) and A. niger [NB6M (B16.6)]. For P. angusta there are several full or partial sequences available publically as genome survey sequences, and they allowed the identification of eleven complex I subunit sequences [NUBM (51 kDa), NUGM (30 kDa), NUIM (TYKY), NUZM, NIAM (ASHI), N7BM (B17.2), NUKM (B14.7), NUFM (B13), NUMM (13 kDa), NIMM (MWFE) and NIDM (PDSW)]. A comparison of observed peptide masses from P. pastoris against these P. angusta sequences only provided a match to one complex I subunit [NUIM (TYKY)].

DISCUSSION

The presence of complex I in Pichia species

The results of the present study show that both P. pastoris and P. angusta assemble a piericidin-A-sensitive, multisubunit NADH:ubiquinone oxidoreductase in their mitochondria, with closely similar catalytic and spectroscopic properties to the enzymes from B. taurus and Y. lipolytica. A number of the core subunits of complex I have been identified in both Pichia species, along with several supernumerary subunits, which vary between kingdoms, subkingdoms and between genera and species [7]. Our data and preliminary sequence comparisons suggest that, within the species included in Figure 1, complex I from P. pastoris is most closely related to the enzyme from P. stipitis, whereas P. angusta complex I is most closely related to the enzymes from Y. lipolytica, L. elongisporus, P. stipitis, P. guilleromondii and D. hansenii (but not to any one of them in particular).

The presence of the complex I ND genes in the mitochondrial genomes of many fungi (including Pichia species) [13,28], suggests that most (if not all) obligate aerobic fungi are able to express complex I. Those species capable of facultative anaerobic fermentation encode only the single subunit enzyme (see Figure 1). Without a larger set of sequences for comparison, it is difficult to speculate further on the evolutionary development of complex I in some species, and its loss in others. However, intriguing correlations can be identified within the species in Figure 1. Only species that encode complex I encode the alternative, single subunit, non-proton pumping ubiquinol oxidase [28]: if complexes III and IV are substituted by the alternative oxidase, complex I is retained in order to maintain the protonmotive force (or vice versa). Only species in the subphylum pezizomycota encode the complex I supernumerary subunit NURM [28], and they are the only species that encode the three subunit, proton-pumping transhydrogenase also. In N. crassa NURM is a 17.8 kDa protein containing one transmembrane helix, with no known homology with any other protein.

The catalytic activities of the complexes I from Pichia species

The preparation of a highly pure eukaryotic complex I with native NADH:ubiquinone oxidoreductase activity has proved challenging, particularly for the mammalian complexes. Previously, by including phospholipids in the preparation, highly pure complex I has been prepared from B. taurus mitochondria with NADH:DQ oxidoreductase activity up to 4 μmol of NADH·min−1·mg−1 [22]. For Y. lipolytica, preparations with an activity of ∼2.8 μmol·min−1·mg−1 were obtained by conventional purification [33], but a histidine tag on the NUGM (30 kDa) subunit increased the activity to 6–7 μmol· min−1·mg−1 [40]. In the present study, conventional methods were used to purify the two Pichia complexes I. Initially, it was hoped that the P. angusta enzyme would be more stable because of the organism's thermotolerance [18], but the P. pastoris enzyme gave a significantly higher activity than that of P. angusta (6.3 and 4.0 μmol of NADH·min−1·mg−1 respectively). It is possible that phospholipid content is key to understanding the relative activities. The B. taurus enzyme is readily depleted of phospholipids during preparation [22]. The P. angusta enzyme requires more exogenous phospholipids in activity assays than the P. pastoris enzyme, to attain its highest activity, and it cannot be incorporated properly into liposomes. It is possible that this stems from its lower relative abundance in the membranes – it must be washed more extensively on the anion-exchange column, a procedure which is known to ‘strip’ the B. taurus enzyme [22]. Finally, variations in the DQ reductase activity are not reflected in any other activity measurements (see Table 2); in particular, NADH:HAR oxidoreduction is essentially the same in all four cases. Previously, a study of the complexes I from B. taurus and P. denitrificans also found that, whereas FeCN was reduced relatively slowly by the prokaryotic enzyme, the HAR reduction rates were similar [43].

‘Tests’ which distinguish complex I from other NADH dehydrogenases, and inhibitor sensitivity

It is widely assumed that dNADH reacts only with complex I, not with alternative NADH dehydrogenases, although this fact has only been demonstrated in E. coli [29]. All the complexes I tested in the present study (Table 2) react at comparable rates with dNADH and NADH. However, if the dNADH:HAR and dNADH:DQ oxidoreductase activities in the membranes are only from complex I, then the ratio between them should be conserved throughout the preparation. In fact, the ratio (HAR/DQ) increases from 1.1 (P. pastoris) and 5.0 (P. angusta) in the membranes to 9.8 and 21.1 respectively. Because the specific dNADH:HAR oxidoreductase activity is unlikely to vary significantly, the results indicate at least one additional enzyme in the membrane preparations catalyses dNADH:DQ oxidoreduction also (or a significant portion of the dNADH:DQ activity has been lost). Consequently, dNADH may differentiate less clearly between complex I and alternative NADH dehydrogenases than commonly supposed. It is widely assumed that only complex I catalyses NADH:HAR oxidoreduction also (see, for example [11]) but we have been unable to find any published data to support this fact. During the preparation of P. angusta complex I, two fractions with NADH:HAR oxidoreductase activity eluted from the anion-exchange column, but only the second catalysed dNADH:DQ oxidoreduction. Although the protein composition of the first peak was complex, SDS/PAGE, MS and Western blot analyses detected no evidence for complex I or any portion it, suggesting that this second ‘test’ for complex I is also not definitive.

Table 2 shows that all the NADH:DQ oxidoreductase activities tested are strongly inhibited by piericidin A, but rotenone does not inhibit the three yeast enzymes as strongly as the B. taurus enzyme. This observation may explain why rotenone-sensitive NADH oxidation was not observed previously in P. pastoris mitochondria [14]. Studies of the inhibition of the activity of P. pastoris complex I over a range of rotenone concentrations showed that almost complete inhibition can be observed, but only at much higher concentrations (∼96% in 50 μM rotenone). Relatively weak inhibition of Y. lipolytica complex I was also observed previously [42]. Thus rotenone is regarded as an archetypal complex I inhibitor simply because of the predominance of the B. taurus enzyme in early studies.

The production of ROS

Previously, we described a mechanism for ROS production by the reduced flavin in complex I from B. taurus, and demonstrated the same mechanism in complex I from E. coli [23,44]. In the present study, we compared ROS production by the complexes I from P. pastoris, P. angusta and Y. lipolytica with that by the B. taurus enzyme. Values for all three yeast species were significantly higher than those from B. taurus, and those for the two Pichia species were the highest. This is particularly interesting because B. taurus and E. coli complexes I exhibit similar rates [44]. When the two Pichia species are grown on methanol the induction of alcohol oxidases results in significant H2O2 production, which is metabolized by catalase and cytochrome c peroxidase [17,45]. The additional contribution from complex I may not be significant in this context. Furthermore, all four eukaryotic enzymes produce predominantly superoxide, not H2O2. This observation maintains the correlation between the superoxide to H2O2 ratio and the reduction of cluster N1a described previously [44]: the N1a EPR signal from the [2Fe–2S] cluster in NUHM (24 kDa) is not observed in any of the eukaryotic enzymes described in the present study (see Figure 3). Therefore the cluster is predominantly oxidized in the presence of NADH, and is, in principle, available to transiently accept the second electron from the fully reduced flavin, promoting the escape of nascent superoxide from the active site.

Abbreviations

     
  • ACMA

    9-amino-6-chloro-2-methoxyacridine

  •  
  • DDM

    dodecyl-β-D-maltoside

  •  
  • dNADH

    deamino-NADH

  •  
  • DQ

    decylubiquinone

  •  
  • FCCP

    carbonyl cyanide p-trifluoromethoxyphenylhydrazone

  •  
  • FeCN

    hexacyanoferrate (III), FeS, iron sulfur

  •  
  • FMN

    flavin mononucleotide

  •  
  • HAR

    hexaammineruthenium (III)

  •  
  • MALDI–TOF

    matrix-assisted laser-desorption ionization–time-of-flight

  •  
  • RCR

    respiratory control ratio

  •  
  • ROS

    reactive oxygen species

AUTHOR CONTRIBUTION

Hannah Bridges carried out all experimental work (unless otherwise stated) and analysed the data. Ljuban Grgic carried out the first preparations of complex I from P. pastoris and supervised the reconstitution experiments. Michael Harbour carried out MS experiments and analysis (with Hannah Bridges). Judy Hirst directed the research. Judy Hirst and Hannah Bridges wrote the paper.

We thank Professor Ulrich Brandt (Frankfurt, Germany) for providing the strain of Y. lipolytica, Dr Ian M. Fearnley (Medical Research Council, Cambridge, U.K.) for advice on MS, and Martin King (Medical Research Council, Cambridge, U.K.) for providing B. taurus complex I.

FUNDING

This work was supported by The Medical Research Council.

References

References
1
Brandt
U.
Energy converting NADH:quinone oxidoreductase (complex I)
Annu. Rev. Biochem.
2006
, vol. 
75
 (pg. 
69
-
92
)
2
Hirst
J.
Energy transduction by respiratory complex I: an evaluation of current knowledge
Biochem. Soc. Trans.
2005
, vol. 
33
 (pg. 
525
-
529
)
3
Sazanov
L. A.
Hinchliffe
P.
Structure of the hydrophilic domain of respiratory complex I from Thermus thermophilus
Science
2006
, vol. 
311
 (pg. 
1430
-
1436
)
4
Hirst
J.
Carroll
J.
Fearnley
I. M.
Shannon
R. J.
Walker
J. E.
The nuclear encoded subunits of complex I from bovine heart mitochondria
Biochim. Biophys. Acta
2003
, vol. 
1604
 (pg. 
135
-
150
)
5
Morgner
N.
Zickermann
V.
Kerscher
S.
Wittig
I.
Abdrakhmanova
A.
Barth
H.-D.
Brutschy
B.
Brandt
U.
Subunit mass fingerprinting of mitochondrial complex I
Biochim. Biophys. Acta
2008
, vol. 
1777
 (pg. 
1384
-
1391
)
6
Marques
I.
Duarte
M.
Assunção
J.
Ushakova
A. V.
Videira
A.
Composition of complex I from Neurospora crassa and disruption of two accessory subunits
Biochim. Biophys. Acta
2005
, vol. 
1707
 (pg. 
211
-
220
)
7
Gabaldón
T.
Rainey
D.
Huynen
M. A.
Tracing the evolution of a large protein complex in the eukaryotes, NADH:ubiquinone oxidoreductase (complex I)
J. Mol. Biol.
2005
, vol. 
348
 (pg. 
857
-
870
)
8
DiMauro
S.
Schon
E. A.
Mitochondrial respiratory-chain diseases
New Engl. J. Med.
2003
, vol. 
348
 (pg. 
2656
-
2668
)
9
Greenamyre
J. T.
Hastings
T. G.
Parkinson's: divergent causes, convergent mechanisms
Science
2004
, vol. 
304
 (pg. 
1120
-
1122
)
10
Videira
A.
Complex I from the fungus Neurospora crassa
Biochim. Biophys. Acta
1998
, vol. 
1364
 (pg. 
89
-
100
)
11
Kerscher
S.
Dröse
S.
Zwicker
K.
Zickermann
V.
Brandt
U.
Yarrowia lipolytica, a yeast genetic system to study mitochondrial complex I
Biochim. Biophys. Acta
2002
, vol. 
1555
 (pg. 
83
-
91
)
12
Kriventseva
E. V.
Rahman
N.
Espinosa
O.
Zdobnov
E. M.
OrthoDB: the hierarchical catalog of eukaryotic orthologs
Nucleic Acids Res.
2008
, vol. 
36
 (pg. 
D271
-
D275
)
13
Nosek
J.
Fukuhara
H.
NADH dehydrogenase subunit genes in the mitochondrial DNA of yeasts
J. Bacteriol.
1994
, vol. 
176
 (pg. 
5622
-
5630
)
14
González-Barroso
M. M.
Ledesma
A.
Lepper
S.
Pérez-Magán
E.
Zaragoza
P.
Rial
E.
Isolation and bioenergetic characterization of mitochondria from Pichia pastoris
Yeast
2006
, vol. 
23
 (pg. 
307
-
313
)
15
Ruiz-Herrera
J.
Sentandreu
R.
Different effectors of dimorphism in Yarrowia lipolytica
Arch. Microbiol.
2002
, vol. 
178
 (pg. 
477
-
483
)
16
Morín
M.
Monteoliva
L.
Insenser
M.
Gil
C.
Domínguez
A.
Proteomic analysis reveals metabolic changes during yeast to hypha transition in Yarrowia lipolytica
J. Mass Spectrom.
2007
, vol. 
42
 (pg. 
1453
-
1462
)
17
Gellissen
G.
Kunze
G.
Gaillardin
C.
Cregg
J. M.
Berardi
E.
Veenhuis
M.
van der Klei
I.
New yeast expression platforms based on methylotrophic Hansenula polymorpha and Pichia pastoris and on dimorphic Arxula adeninivorans and Yarrowia lipolytica – a comparison
FEMS Yeast Res.
2005
, vol. 
5
 (pg. 
1079
-
1096
)
18
van Uden
N.
Abranches
P.
Cabeça-Silva
C.
Temperature functions of thermal death in yeasts and their relation to the maximum temperature for growth
Arch. Mikrobiol.
1968
, vol. 
61
 (pg. 
381
-
393
)
19
Kerscher
S. J.
Okun
J. G.
Brandt
U.
A single external enzyme confers alternative NADH:ubiqinone oxidoreductase activity in Yarrowia lipolytica
J. Cell Sci.
1999
, vol. 
112
 (pg. 
2347
-
2354
)
20
Kashani-Poor
N.
Kerscher
S.
Zickermann
V.
Brandt
U.
Efficient large scale purification of his-tagged proton translocating NADH:ubiquinone oxidoreductase (complex I) from the strictly aerobic yeast Yarrowia lipolytica
Biochim. Biophys. Acta
2001
, vol. 
1504
 (pg. 
363
-
370
)
21
Dröse
S.
Galkin
A.
Brandt
U.
Proton pumping by complex I from Yarrowia lipolytica reconstituted into proteoliposomes
Biochim. Biophys. Acta
2005
, vol. 
1710
 (pg. 
87
-
95
)
22
Sharpley
M. S.
Shannon
R. J.
Draghi
F.
Hirst
J.
Interactions between phospholipids and NADH:ubiquinone oxidoreductase (complex I) from bovine mitochondria
Biochemistry
2006
, vol. 
45
 (pg. 
241
-
248
)
23
Kussmaul
L.
Hirst
J.
The mechanism of superoxide production by NADH:ubiquinone oxidoreductase (complex I) from bovine heart mitochondria
Proc. Natl. Acad. Sci. U.S.A.
2006
, vol. 
103
 (pg. 
7607
-
7612
)
24
Burch
H. B.
Fluorimetric assay of FAD, FMN and riboflavin
Methods Enzymol.
1957
, vol. 
3
 (pg. 
960
-
962
)
25
Pieroni
L.
Khalil
L.
Charlotte
F.
Poynard
T.
Piton
A.
Hainque
B.
Imbert-Bismut
F.
Comparison of bathophenanthroline sulfonate and ferene as chromogens in colorimetric measurement of low hepatic iron concentration
Clin. Chem.
2001
, vol. 
47
 (pg. 
2059
-
2061
)
26
Perkins
D. N.
Pappin
D. J. C.
Creasy
D. M.
Cottrell
J. S.
Probability-based protein identification by searching sequence databases using mass spectrometry data
Electrophoresis
1999
, vol. 
20
 (pg. 
3551
-
3567
)
27
Melo
A. M. P.
Bandeiras
T. M.
Teixeira
M.
New insights into type II NAD(P)H: quinone oxidoreductases
Microbiol. Mol. Biol. Rev.
2004
, vol. 
68
 (pg. 
603
-
616
)
28
Lavín
J. L.
Oguiza
J. A.
Ramírez
L.
Pisabarro
A. G.
Comparative genomics of the oxidative phosphorylation system in fungi
Fungal Genet. Biol.
2008
, vol. 
45
 (pg. 
1248
-
1256
)
29
Matsushita
K.
Ohnishi
T.
Kaback
H. R.
NADH-ubiquinone oxidoreductases of the Escherichia coli aerobic respiratory chain
Biochemistry
1987
, vol. 
26
 (pg. 
7732
-
7737
)
30
Katz
R.
Kilpatrick
L.
Chance
B.
Aquisition and loss of rotenone sensitivity in Torulopsis utilis
Eur. J. Biochem.
1971
, vol. 
21
 (pg. 
301
-
307
)
31
Grossman
S.
Cobley
J. G.
Singer
T. P.
Beinert
H.
Reduced nicotinamide adenine dinucleotide dehydrogenase, piericidin sensitivity, and site 1 phosphorylation in different growth phases of Candida utilis
J. Biol. Chem.
1974
, vol. 
249
 (pg. 
3819
-
3826
)
32
Schwitzguébel
J.-P.
Palmer
J. M.
Properties of mitochondria as a function of the growth stages of Neurospora crassa
J. Bacteriol.
1982
, vol. 
149
 (pg. 
612
-
619
)
33
Djafarzadeh
R.
Kerscher
S.
Zwicker
K.
Radermacher
M.
Lindahl
M.
Schägger
H.
Brandt
U.
Biophysical and structural characterization of proton-translocating NADH-dehydrogenase (complex I) from the strictly aerobic yeast Yarrowia lipolytica
Biochim. Biophys. Acta
2000
, vol. 
1459
 (pg. 
230
-
238
)
34
Carroll
J.
Fearnley
I. M.
Skehel
J. M.
Shannon
R. J.
Hirst
J.
Walker
J. E.
Bovine complex I is a complex of forty-five different subunits
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
32724
-
32727
)
35
Ohnishi
T.
Iron-sulphur clusters/semiquinones in complex I
Biochim. Biophys. Acta
1998
, vol. 
1364
 (pg. 
186
-
206
)
36
Reda
T.
Barker
C. D.
Hirst
J.
Reduction of the iron-sulfur clusters in mitochondrial NADH:ubiquinone oxidoreductase (complex I) by EuII-DTPA, a very low potential reductant
Biochemistry
2008
, vol. 
47
 (pg. 
8885
-
8893
)
37
Uhlmann
M.
Friedrich
T.
EPR signals assigned to Fe/S cluster N1c of the Escherichia coli NADH:ubiquinone oxidoreductase derive from cluster N1a
Biochemistry
2005
, vol. 
44
 (pg. 
1653
-
1658
)
38
Yakovlev
G.
Reda
T.
Hirst
J.
Reevaluating the relationship between EPR spectra and enzyme structure for the iron-sulfur clusters in NADH:quinone oxidoreductase
Proc. Natl. Acad. Sci. U.S.A.
2007
, vol. 
104
 (pg. 
12720
-
12725
)
39
Ohnishi
T.
Nakamaru-Ogiso
E.
Were there any misassignments among iron-sulfur clusters N4, N5 and N6b in NADH-quinone oxidoreductase (complex I)
Biochim. Biophys. Acta
2008
, vol. 
1777
 (pg. 
703
-
710
)
40
Dröse
S.
Zwicker
K.
Brandt
U.
Full recovery of the NADH:ubiquinone activity of complex I (NADH:ubiquinone oxidoreductase) from Yarrowia lipolytica by the addition of phospholipids
Biochim. Biophys. Acta
2002
, vol. 
1556
 (pg. 
65
-
72
)
41
Dufour
J.-P.
Goffeau
A.
Tsong
T. Y.
Active proton uptake in lipid vesicles reconstituted with the purified yeast plasma membrane ATPase
J. Biol. Chem.
1982
, vol. 
257
 (pg. 
9365
-
9371
)
42
Galkin
A.
Brandt
U.
Superoxide radical formation by pure complex I (NADH:ubiquinone oxidoreductase) from Yarrowia lipolytica
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
30129
-
30135
)
43
Zickermann
V.
Kurki
S.
Kervinen
M.
Hassinen
I.
Finel
M.
The NADH oxidation domain of complex I: do bacterial and mitochondrial enzymes catalyze ferricyanide reduction similarly?
Biochim. Biophys. Acta
2000
, vol. 
1459
 (pg. 
61
-
68
)
44
Esterházy
D.
King
M. S.
Yakovlev
G.
Hirst
J.
Production of reactive oxygen species by complex I (NADH:ubiquinone oxidoreductase) from Escherichia coli and comparison to the enzyme from mitochondria
Biochemistry
2008
, vol. 
47
 (pg. 
3964
-
3971
)
45
Duff
S. J. B.
Murray
W. D.
Metabolism of hydrogen peroxide by Pichia pastoris. Agric
Biol. Chem.
1990
, vol. 
54
 (pg. 
1967
-
1973
)

Supplementary data