LHON (Leber hereditary optic neuropathy) is a maternally inherited disease that leads to sudden loss of central vision at a young age. There are three common primary LHON mutations, occurring at positions 3460, 11778 and 14484 in the human mtDNA (mitochondrial DNA), leading to amino acid substitutions in mitochondrial complex I subunits ND1, ND4 and ND6 respectively. We have now examined the effects of ND6 mutations on the function of complex I using the homologous NuoJ subunit of Escherichia coli NDH-1 (NADH:quinone oxidoreductase) as a model system. The assembly level of the NDH-1 mutants was assessed using electron transfer from deamino-NADH to the ‘shortcut’ electron acceptor HAR (hexammine ruthenium), whereas ubiquinone reductase activity was determined using DB (decylubiquinone) as a substrate. Mutant growth in minimal medium with malate as the main carbon source was used for initial screening of the efficiency of energy conservation by NDH-1. The results indicated that NuoJ-M64V, the equivalent of the common LHON mutation in ND6, had a mild effect on E. coli NDH-1 activity, while nearby mutations, particularly NuoJ-Y59F, NuoJ-V65G and NuoJ-M72V, severely impaired the DB reduction rate and cell growth on malate. NuoJ-Met64 and NuoJ-Met72 position mutants lowered the affinity of NDH-1 for DB and explicit C-type inhibitors, whereas NuoJ-Y59C displayed substrate inhibition by oxidized DB. The results are compatible with the notion that the ND6 subunit delineates the binding cavity of ubiquinone substrate, but does not directly take part in the catalytic reaction. How these changes in the enzyme's catalytic properties contribute to LHON pathogenesis is discussed.

INTRODUCTION

LHON (Leber hereditary optic neuropathy) is a maternally inherited human disease causing acute to subacute bilateral painless loss of central vision in young adults, with male predominance. After the acute phase, optic nerve atrophy and centrocaecal absolute scotoma develop within weeks in one eye, followed by the other eye after a mean interval of 2 months [1]. The final visual outcome varies, and patients occasionally experience a marked visual recovery after the acute phase, both being dependent on the genotype [2]. The underlying genetic cause in LHON is a mutation in the mitochondrial genome. The first reported mutation linked to LHON was G11778A [3], and this and two other missense mutations, G3460A and T14484C, currently account for 96% of all patients diagnosed [4], so that they are classified as common primary mutations.

Mitochondrial NADH:ubiquinone oxidoreductase (complex I; EC 1.6.5.3) extracts electrons from NADH and channels them through FMN and a set of serially arranged iron–sulfur clusters to the reduction site of the lipophilic electron acceptor ubiquinone [5]. The energy liberated in this electron transfer is used to drive proton pumping across the mitochondrial inner membrane with a stoichiometry of 4H+/2e [6]. It is a multi-subunit enzyme with 45 dissimilar proteins [7], 14 of which constitute the evolutionally conserved catalytic core, which is also found in NDH-1 (bacterial NADH:quinone oxidoreductase), the bacterial counterpart of the mitochondrial complex I, see [8] for a review. In mitochondrial complex I, seven of these 14 core subunits are encoded by the mitochondrial genome and synthesized within the organelle [9]. All of these seven subunits, designated ND1–ND6 and ND4L, are integral membrane proteins having 3–14 transmembrane segments [10].

The common primary LHON mutations occur in three subunits of complex I, namely ND1, ND4 and ND6, all of which belong to the membrane domain of the enzyme. Several additional LHON-associated replacement mutations have been found in other mitochondrially encoded subunits of complexes I, III, IV and F1Fo-ATPase-synthase [11]. Taking all the known LHON mutations into account, it appears that the ND1 and ND6 genes are the main ‘hot spots’ for such mutations [12,13], since a total of 16 of their codons have been identified as being associated with LHON (Mitomap, May 2007; http://www.mitomap.org). Interestingly, many of the pathogenic ND6 subunit mutations are located in the relatively highly conserved N-terminal region within its third transmembrane helix (Figure 1).

Alignment of the region around the third putative transmembrane helix of the human ND6 with selected homologues

Figure 1
Alignment of the region around the third putative transmembrane helix of the human ND6 with selected homologues

The predicted transmembrane helix in the E. coli sequence is underlined [52]. Highly conserved amino acids are highlighted in grey. The symbol ‘▲’ denotes residues for which site-directed mutants were constructed in the present study. The uppermost line shows substitutions associated with diseases [12,53,54]. The GenBank® accession numbers for the sequences are (from top to bottom) ABA08023 (human), AAV88355 (bovine), NP_007573 (Alligator mississippiensis), NP_059342 (zebrafish), CAC28088 (Yarrowia lipolytica), YP_271941 (Montastraea franksi), CAA60351 (maize chloroplast), P29922 (P. denitrificans), AAA97947 (Thermus thermophilus) and NP_416783 (E. coli). The alignment was performed with the ClustalW program.

Figure 1
Alignment of the region around the third putative transmembrane helix of the human ND6 with selected homologues

The predicted transmembrane helix in the E. coli sequence is underlined [52]. Highly conserved amino acids are highlighted in grey. The symbol ‘▲’ denotes residues for which site-directed mutants were constructed in the present study. The uppermost line shows substitutions associated with diseases [12,53,54]. The GenBank® accession numbers for the sequences are (from top to bottom) ABA08023 (human), AAV88355 (bovine), NP_007573 (Alligator mississippiensis), NP_059342 (zebrafish), CAC28088 (Yarrowia lipolytica), YP_271941 (Montastraea franksi), CAA60351 (maize chloroplast), P29922 (P. denitrificans), AAA97947 (Thermus thermophilus) and NP_416783 (E. coli). The alignment was performed with the ClustalW program.

Several hypotheses have been proposed regarding the primary cause of the disease phenotype. These include mild impairment of mitochondrial respiration due to a complex I defect, as found in the ND1-A52T and ND4-R340H mutations [14] but not in ND6-M64V [15,16]. A defect in ATP production with unaltered ATP content has been found in trans-mitochondrial hybrid cell lines carrying any of the common LHON mutations [17,18]. Increased production of ROS (reactive oxygen species) or increased sensitivity to oxidative stress has been found in an ND1-A52T and ND4-R340H mutant cybrid neuronal cell model [19], whereas ROS production was not markedly increased in an ND6-M64V mutant cell line [20]. Cell death in LHON appears to occur by an apoptotic mechanism [21,22], but it remains obscure how the apoptotic wave of cell death is initiated exclusively in the retinal ganglion cells, especially in the case of the mild ND6-M64V mutation.

The mitochondrial genome accumulates mutations at a rate approximately one order of magnitude faster than the nuclear genome, and it is difficult to distinguish between rare haplogroup-specific variants and true pathogenic LHON mutations [23]. The pathogenic mutations are also frequently accompanied by other genetic alterations in mtDNA (mitochondrial DNA), and the contribution of individual mutations to the clinical phenotype and the biochemical defect may be difficult to establish. It is also very difficult to introduce site-specific mutations into the mitochondrial genome, so that we employed a bacterial homologue of complex I, the NDH-1 of Escherichia coli, as a model system for studying the effects of LHON and other mutations in the NuoJ subunit, the bacterial counterpart of ND6, on the function of the enzyme. The method allows us to examine the putatively pathogenic mutations against a truly neutral background and to explore their effects on the enzyme's catalytic turnover in detail.

A bacterial system has been employed on several occasions to model disease mutations. During the execution of the present study, which started during our endeavours with site-directed mutagenesis of the NuoK subunit [24], Kao et al. [25] reported on the modelling of mutations in the NuoJ subunit in E. coli, and our study, carried out using the methodology we had developed for NuoK [24], may also be partly regarded as an extension of the experiments of Kao et al. [25] on NuoJ. We have now analysed the E. coli equivalent of the common LHON mutation 14484/ND6-M64V as well as several other point mutations in its vicinity. This is important, because current knowledge of the pathophysiology of LHON is insufficient and the delineation between pathogenic mutations and polymorphism is difficult, so that the purported role of some mtDNA mutations in the development of LHON has recently been challenged [26].

EXPERIMENTAL

Materials

DB (decylubiquinone; 2,3-dimethoxy-5-methyl-6-decyl-1,4-benzoquinone), NADH, d-NADH [deamino-NADH; nicotinamide-hypoxanthine dinucleotide (reduced form)], DL-malate, piericidin A, soya bean phospholipids (Asolectin), VNA (N-vanillylnonanamide), stigmatellin and myxothiazol were from Sigma. HAR (hexammine ruthenium) was purchased from Aldrich, Hepes and Mes from AppliChem and potassium cyanide and PMSF from Merck. Oligonucleotides were purchased from Sigma-Genosys, the QuikChange® XL mutagenesis kit was from Stratagene and the annonin extract was a gift from Dr Y. K. Gupta (Department of Medicinal Chemistry, Banaras Hindu University, Varanasi, India).

Mutagenesis

A series of site-directed mutants was generated employing the same strategy as was applied earlier to the study of mutations in the NuoK subunit of E. coli NDH-1, an equivalent of the ND4L subunit of complex I [24]. The point mutations listed in Table 1 were introduced into pJK [24] using the QuikChange® XL mutagenesis kit (Stratagene) according to the manufacturer's instructions, followed by DNA sequencing of the construct. The oligonucleotides used in mutagenesis are listed in Supplementary Table S1 (http://www.BiochemJ.org/bj/409/bj4090129add.htm). Wild-type or mutant nuoJ in conjunction with non-mutant nuoK (with the c-terminal His tag) were transformed into GVnuoJK [24], a nuoJK-knockout strain, for in trans complementation. It should be noted that the parent strain used here carries only the bd type of ubiquinol oxidase in the aerobic respiratory chain [27].

Table 1
LHON mutations in the third transmembrane helix of the human ND6 subunit of complex I and mutations introduced into the NuoJ subunit of E. coli NDH-1

T14484C mutation is one of the three common primary LHON mutations.

Mitochondrial mutations in ND6   
Nucleotide change Residue change Resulting disease Mutations introduced into E. coli NuoJ 
T14498C Y59C LHON Y59C, Y59F 
T14484C M64V LHON M64V, M64C 
C14482G M64I LHON M64V, M64C 
C14482A M64I LHON M64V, M64C 
G14459A A72V *LDYT †M72V, M72C, M72A 
   ‡V65G 
   ‡Y109F 
Mitochondrial mutations in ND6   
Nucleotide change Residue change Resulting disease Mutations introduced into E. coli NuoJ 
T14498C Y59C LHON Y59C, Y59F 
T14484C M64V LHON M64V, M64C 
C14482G M64I LHON M64V, M64C 
C14482A M64I LHON M64V, M64C 
G14459A A72V *LDYT †M72V, M72C, M72A 
   ‡V65G 
   ‡Y109F 
*

LDYT, LHON and dystonia.

There is a methionine residue at position 72 in NuoJ of E. coli, whereas there is alanine in this position in the human ND6 (Figure 1).

Additional not-disease-related E. coli mutation introduced.

Bacterial culture and membrane preparation

For membrane preparation, a colony from a fresh transformation plate was grown overnight in LB (Luria–Bertani) medium in the presence of streptomycin, ampicillin and 1% arabinose and diluted 1:100 in fresh LB medium supplemented with streptomycin and ampicillin. After 6–8 h of cultivation, a fraction of this LB culture was diluted 1:100 in malate-YE medium [65 mM potassium phosphate buffer (pH 7.1) supplemented with 90 mM DL-malate as the main carbon source, 0.1% (w/v) yeast extract, 1 mM citrate, 75 mM ammonium sulfate, salts and ampicillin] [24] and grown at 37 °C under high aeration. The cells were collected (4200 g, 20 min and 4 °C) when the attenuance at 600 nm (D600) reached 0.1–0.4, cooled and washed with 50 mM Mes/KOH (pH 6.5), 2.5 mM EDTA and 0.2 mM PMSF. The washed cells were used for membrane preparation as described by Leif et al. [28], except that 50 mM Mes/KOH (pH 6.5), 0.2 mM PMSF, 5 mM MgSO4 and 10 μg/ml DNase I were used during the French press treatment. Unbroken cells and debris were removed by centrifugation (4200 g, 20 min), only the supernatant was used for ultracentrifugation (180000 g, 65 min) and the membranes were suspended in 50 mM Mes/KOH (pH 6.5) containing PMSF as above, to approx. 0.5 ml/g of initial cell weight. Membrane samples for the inhibitor titration experiments were prepared as described in [29]. The protein concentration in the membrane preparations was assessed by the method of Lowry et al. [30]. Growth on malate was determined as the final attenuance of the culture at 600 nm (D600) after 24 h of growth in malate-YE medium at 37 °C.

Activity measurements

E. coli possesses two NADH:quinone oxidoreductases, NDH-1 and NDH-2. The former is an equivalent of complex I, while the latter is a single-subunit enzyme that, although catalysing electron transfer from NADH to ubiquinone, does not couple the electron transfer activity with proton translocation across the membrane or to the generation of a membrane potential [31]. The two E. coli NADH oxidases can be distinguished from each other in activity assays by employing the non-physiological NADH analogue d-NADH as a substrate, since this serves as an equally good electron donor for NDH-1 as NADH, whereas NDH-2 is unable to oxidize it. Accordingly, d-NADH was used as a substrate to measure NDH-1-dependent activities, while NADH was employed to assay the combined NDH-1 and NDH-2 capacity, which corresponds better to the in vivo situation during bacterial growth.

The activity assays were performed on a dual-wavelength spectrophotometer as previously described [24], with the modifications that the d-NADH:DB oxidoreductase assays were performed with freshly prepared membrane samples (0.1 mg/ml membrane protein), 0.5 mg/ml soya bean phospholipid and 90 μM d-NADH. Six DB concentrations, ranging from 50 to 600 μM, were used and three parallel experiments were performed at each concentration. The rates of DB reduction in the concentration range between 0.5 and 45 μM were estimated from the progress curve throughout the course of a complete reaction (in triplicate), the initial DB concentration being 30–45 μM. Three to five colonies for each mutant or control were analysed independently. The VNA sensitivity at 100 μM DB was tested by adding VNA to a final concentration of 400 μM after recording the initial DB-reductase activity for 1–2 min. For the d-NADH:HAR activity, the previously used annonin was replaced by 400 μM VNA. In the sensitivity tests, the inhibitors dissolved in ethanol or DMSO to a final concentration of 30–400 μM VNA, 0.5–20 μg/ml annonin, 0.5–19 μM piericidin A, 5–100 μM stigmatellin or 5–100 μM myxothiazol (or an equal amount of ethanol or DMSO) were added before DB and the assays were conducted similarly to the d-NADH:DB oxido-reductase assays in all other aspects. The DB concentration was determined spectrophotometrically using a molar absorption coefficient (ϵ) of 14.0 mM−1·cm−1 at 278 nm.

Statistics

Student's t test for independent samples was used to compare two groups. For multiple comparisons, one-way ANOVA followed by the Dunnett's test was used. Curves were fitted to the kinetic equation with Sigmaplot 9.0 (Systat Software, Erkrath, Germany).

RESULTS

Growth and NADH oxidation capacity

On the basis of our previous experiments on NuoK mutagenesis in E. coli, the inability to grow on malate is a useful measure of the energy-conserving function of NDH-1 [24]. Therefore the effects of mutations on complex I activity were first evaluated by testing their effect on cell growth in a minimal medium where malate was the main carbon source. The results, depicted in Figure 2, show that there is a positive sigmoidal correlation between d-NADH oxidase activity and cell growth measured as attenuance at 600 nm (r=0.796, P<0.001). Interestingly, the growth of the double mutant M64V/M72A on malate as the main carbon source, and also the d-NADH:O2 activity, were lower than for either mutant alone (Table 2).

Correlation between growth in a malate-YE medium and d-NADH oxidation activity in E. coli mutants

Figure 2
Correlation between growth in a malate-YE medium and d-NADH oxidation activity in E. coli mutants

Cell density was measured as the attenuance of the culture at 600 nm. The curve represents a least-squares fit of the logistic dose–response curve y=(ad)/[1+(x/c)b]+d. (r=0.796, P<0.001, and the values are means±S.E.M.)

Figure 2
Correlation between growth in a malate-YE medium and d-NADH oxidation activity in E. coli mutants

Cell density was measured as the attenuance of the culture at 600 nm. The curve represents a least-squares fit of the logistic dose–response curve y=(ad)/[1+(x/c)b]+d. (r=0.796, P<0.001, and the values are means±S.E.M.)

Table 2
Respiratory activities in NuoJ mutants and the control, and their growth capability on malate

Five colonies for the control and three for each mutant were subjected to membrane isolation and activity analyses (run in triplicate). Units for NADH:O2, d-NADH:O2 and d-NADH:HAR activities are nmol·min−1·mg−1 of protein. For growth on malate, which was determined as final attenuance of cultures at 600 nm after 1 day of growth in malate-YE medium, 11 cultures were analysed for the control, nine for Y59F, eight for M64C, seven for M72V and six for the rest of the mutants. The values are means±S.E.M.

Mutant Growth on malate NADH:O2 d-NADH:O2 d-NADH:HAR 
Control 0.86±0.05 309±53 184±56 1140±304 
Y59C 0.69±0.01 261±39 156±22 1038±93 
Y59F 0.16±0.01 242±12 76±6 1056±40 
M64V 0.68±0.03 467±72 313±52 2498±278 
M64C 0.35±0.08 230±57 90±34 1198±482 
M72V 0.22±0.01 228±35 70±8 1141±56 
M72A 0.68±0.03 338±78 233±79 1945±517 
M72C 0.34±0.01 176±15 86±9 1052±47 
M64V/M72A 0.41±0.03 198±19 97±7 645±53 
V65G 0.06±0.05 180±17 4±1 939±390 
Y109F 0.75±0.01 308±26 207±13 1338±50 
Mutant Growth on malate NADH:O2 d-NADH:O2 d-NADH:HAR 
Control 0.86±0.05 309±53 184±56 1140±304 
Y59C 0.69±0.01 261±39 156±22 1038±93 
Y59F 0.16±0.01 242±12 76±6 1056±40 
M64V 0.68±0.03 467±72 313±52 2498±278 
M64C 0.35±0.08 230±57 90±34 1198±482 
M72V 0.22±0.01 228±35 70±8 1141±56 
M72A 0.68±0.03 338±78 233±79 1945±517 
M72C 0.34±0.01 176±15 86±9 1052±47 
M64V/M72A 0.41±0.03 198±19 97±7 645±53 
V65G 0.06±0.05 180±17 4±1 939±390 
Y109F 0.75±0.01 308±26 207±13 1338±50 

Among the set of NuoJ mutants produced, Y59F and V65G exhibited a knockout-like growth pattern, while the final D600 of the M72V mutant in the malate-YE medium was only slightly higher (Table 2). To test whether the poor growth capacity of these mutants was due to a deficiency in the respiratory chain downstream from the NDH-1 site or to NDH-1 itself, their NADH:O2 activities were measured (Table 2). Subtraction of the d-NADH:O2 activity from the NADH:O2 activity gives an estimate of the NDH-2-linked respiration bypassing the NDH-1 coupling site, and this turned out to be even higher in the Y59F, V65G and M72V mutants than in the controls (results not shown), demonstrating that, apart from NDH-1, the respiratory complexes were functioning normally in these mutants.

Expression and assembly

The HAR reductase activity (d-NADH:HAR) in the membranes (Table 2) was used to estimate the level of expressed and assembled NDH-1, since its loss through nuoJK knockout could be restored by complementation in trans [24]. Despite the impaired malate growth, mutants Y59F and M72V exhibited considerable d-NADH:HAR reductase activity, similar to the control or mutants such as Y59C that grew much better on malate. Even in the case of V65G, the poorest-growing mutant, the d-NADH:HAR reductase activity was higher than in the double mutant M64V/M72A, although the latter, but not the former, grew reasonably well on malate (Table 2).

The results also revealed great variability in expression and assembly levels between membrane batches of individual mutants. Due to this observation, the d-NADH:DB activity of each batch of membranes was normalized to the corresponding HAR reduction activity (by taking the activity ratio) in order to better reveal changes in enzyme catalysis.

Ubiquinone reduction and binding

The d-NADH:DB reductase activity and its dependence on the electron acceptor concentration were measured in isolated membranes from the various strains (Table 3). Complementation of the knockout with wild-type nuoJK construct (control in Table 3) resulted in 34% of the assembly normalized NDH-1 activity in the parent strain GV102 under same conditions [29]. The reason for this is currently unsolved, but it may partly be due to non-synchronized expression and assembly of NDH-1 subunits from the nuo-operon and the plasmid. Assembly-normalized NDH-1 activities were, however, constant in control complementations with the nuoJK expression plasmid introducing a C-terminal His6 tag into NuoK, present in all NDH-1 enzymes analysed in the present study or with a plasmid where the original stop codon was reintroduced into nuoK (M. Kervinen, unpublished work). Therefore we consider it justified to use this system to evaluate the effects of site-directed mutations on NDH-1, and the results obtained with different mutants were compared with the wild-type complementation throughout the present study.

Table 3
Kinetic properties of control and NuoJ mutants

Km and Ks′ values for DB and the Vmax of d-NADH:DB oxidoreductase activity were calculated by fitting the data to the equation v=Vmax·[S]/{Km+[S]+([S]2/Ks′)}, as described in the Results section. Vmax of d-NADH:DB activity was normalized to d-NADH:HAR activity to account for variations in expression levels between different membrane batches of individual mutants. Only the mean is shown for Ks′, the other values are means±S.E.M. measured from the same samples as described in Table 2. *P<0.05; †P<0.01 when compared with control. ‡No exact value could be calculated because of low affinity.

Mutant Vmax/HAR Km (μM) Ks′ (μM) 
Control 0.216±0.018 (100%) 23±2 4227 
Y59C 0.158±0.008 (73%) 22±2 594* 
Y59F 0.083±0.003† (38%) 29±3 1285 
M64V 0.180±0.016 (83%) 33±4* 3409 
M64C 0.120±0.011† (56%) 38±2† 4615 
M72V 0.114±0.003† (53%) 39±4† ∞‡ 
M72A 0.170±0.020 (79%) 46±2† 12430 
M72C 0.147±0.008* (68%) 42±1† 16939 
M64V/M72A 0.264±0.034 (122%) 38±2† 1792 
Y109F 0.241±0.016 (112%) 42±4† 5273 
Mutant Vmax/HAR Km (μM) Ks′ (μM) 
Control 0.216±0.018 (100%) 23±2 4227 
Y59C 0.158±0.008 (73%) 22±2 594* 
Y59F 0.083±0.003† (38%) 29±3 1285 
M64V 0.180±0.016 (83%) 33±4* 3409 
M64C 0.120±0.011† (56%) 38±2† 4615 
M72V 0.114±0.003† (53%) 39±4† ∞‡ 
M72A 0.170±0.020 (79%) 46±2† 12430 
M72C 0.147±0.008* (68%) 42±1† 16939 
M64V/M72A 0.264±0.034 (122%) 38±2† 1792 
Y109F 0.241±0.016 (112%) 42±4† 5273 

As compared with the method used by us previously [24], the changes to the membrane preparation procedure (omission of EDTA, addition of Mg2+ and avoidance of freezing of the sample before the assay) resulted in higher respiratory chain activities and made it possible to perform more detailed analyses of NDH-1 function. The results demonstrated that one mutant, V65G, was practically devoid of d-NADH:DB reductase activity (5.5±0.6 nmol·min−1·mg−1 of protein at 100 μM DB compared with 196±22 nmol·min−1·mg−1 of protein in the control), and that this was not stimulated by higher DB concentrations (results not shown). No accurate analysis of the DB reduction kinetics could be performed in this case, however. Mutant Y59F exhibited the lowest (corrected) Vmax value, while the Vmax/HAR ratio was also significantly decreased in mutants M64C, M72V and M72C (Table 3). The same pattern was observed in the assembly-corrected d-NADH oxidase activities in these mutants (with endogenous ubiquinone as a substrate), the ratios of d-NADH:O2 activity to the HAR reduction rate being 47, 38 and 51% of the control respectively. Unlike the other mutants and the control, Y59C and Y59F showed higher NDH-1 activities with the endogenous electron acceptor than with DB as compared with the wild-type control [93% of wild-type activity with endogenous ubiquinone compared with 73% (Vmax/HAR) with DB, and 45% with endogenous ubiquinone compared with 38% with DB, in Y59C and Y59F respectively].

An interesting finding was that some mutations led to a clear substrate inhibition by oxidized DB (Figure 3). This was analysed by fitting the data to the equation:

 
formula
(1)

where Ks′ is the equilibrium constant for binding of a second substrate molecule to the active site of the enzyme–substrate complex [32]. The results are listed in Table 3. The Km value for the wild-type control was in the range 16–25 μM. All the Met64, Met72 and Tyr109-position mutants and also the double mutant M64V/M72A exhibited a higher Km value for this ubiquinone analogue (Table 3), suggesting somewhat lower affinity for the physiological electron acceptor as well. Mutations at Tyr59 did not have a significant effect on the Km value, but, in contrast, showed the largest increase in sensitivity to substrate inhibition by DB, the Y59C mutant having the clearest effect (Table 3 and Figure 3). The double mutant M64V/M72A also showed some increase in sensitivity to substrate inhibition, higher than with either of the mutations alone, but the difference in Ks′ was not statistically significant as compared with the control. Additionally, all position Met72 substitutions, which led to an increased Km for DB, also showed a tendency for an increase in Ks′.

Substrate inhibition in NuoJ-Tyr59-position mutants

Figure 3
Substrate inhibition in NuoJ-Tyr59-position mutants

Representative DB titrations of d-NADH:DB reductase activity in cytoplasmic membranes from the control or mutants at a fixed d-NADH concentration. Activities are given as ratios to d-NADH:HAR reductase activity to eliminate the effect of variations in expression level. The solid curves represent a fit to the equation v=Vmax·[S]/{Km+[S]+([S]2/Ks′)}, as described in the Results section. ○, Control; ◇, Y59C; ▽, Y59F.

Figure 3
Substrate inhibition in NuoJ-Tyr59-position mutants

Representative DB titrations of d-NADH:DB reductase activity in cytoplasmic membranes from the control or mutants at a fixed d-NADH concentration. Activities are given as ratios to d-NADH:HAR reductase activity to eliminate the effect of variations in expression level. The solid curves represent a fit to the equation v=Vmax·[S]/{Km+[S]+([S]2/Ks′)}, as described in the Results section. ○, Control; ◇, Y59C; ▽, Y59F.

The d-NADH:HAR, d-NADH:O2, NADH:O2 and d-NADH:DB activities in the control and the mutants growing poorly on malate, namely Y59F, V65G and M72V, were also analysed in membranes from cells grown on LB medium and were very similar to the results for the respective membrane batches prepared from cells grown in the malate-YE medium, as listed in Tables 2 and 3.

Inhibitor-sensitivity

The sensitivity of the d-NADH:DB oxidoreductase activity to 400 μM VNA in the mutants did not differ much from the wild-type (percentage inhibition of the initial rate was 73–84%), except perhaps in Y59F (66–74%). Titration experiments in the presence of 100 μM DB and 90 μM d-NADH showed, however, that the IC50 for VNA in mutant Y59F was very similar to the control value, 56 and 62 μM respectively. We selected M72V, the counterpart of the disease-associated ND6-A72V mutant, for further inhibitor-sensitivity experiments because this mutant showed the largest decrease in DB affinity. Sensitivity was tested for class I/A-type and C-type inhibitors and for stigmatellin and myxothiazol, which interact at or near the ubiquinone-binding site(s). The results indicate a clear increase in the IC50 value of mutant M72V for both annonin and VNA, but not for piericidin A, stigmatellin or myxothiazol (Table 4). It should be noted here that an equal amount of membrane protein was used in the cuvette for each of the mutants or control and excess phospholipid was added in both the DB binding and inhibitor-sensitivity experiments in order to eliminate the effect of changes in lipid content between separate samples.

Table 4
Inhibitor-sensitivity in the NuoJ-M72V mutant and wild-type control

IC50, the inhibitor concentration that produces 50% inhibition of the d-NADH:DB activity, was measured at 100 μM DB. Myxothiazol at the concentrations used (5–100 μM) showed a maximal inhibition of d-NADH:DB activity of only 37%, and no difference was observed between the control and the NuoJ-M72V mutant (results not shown). Values are means±S.D., numbers in parentheses indicate the number of independent experiments. *P<0.05.

 IC50 
Inhibitor Control NuoJ-M72V 
Annonin (μg/ml) 0.50±0.06 (n=2) 0.96±0.20* (n=4) 
VNA (μM) 62±13 (n=5) 133±24* (n=2) 
Piericidin A (μM) 12 10 
Stigmatellin (μM) 22 29 
Myxothiazol (μM) >100 >100 
 IC50 
Inhibitor Control NuoJ-M72V 
Annonin (μg/ml) 0.50±0.06 (n=2) 0.96±0.20* (n=4) 
VNA (μM) 62±13 (n=5) 133±24* (n=2) 
Piericidin A (μM) 12 10 
Stigmatellin (μM) 22 29 
Myxothiazol (μM) >100 >100 

DISCUSSION

Despite an abundance of research into the pathophysiology of LHON, the underlying cause of the onset of neuronal cell death remains obscure and the biochemical defects at the level of complex I are controversial. We have developed a disease model in E. coli in which we can estimate and compare the effects of different LHON mutations (including the rare ones) on complex I activity against a genetically neutral background without the contribution of other mitochondrial or nuclear DNA mutations. A similar approach was used recently to investigate the effects of amino acid changes within the ND1 subunit causing MELAS (mitochondrial encephalomyopathy, lactic acidosis and stroke like episodes) syndrome [29]. The effect of mutated mtDNA against neutral nuclear background can be studied, even in mammalian cells, using cybrid techniques and haplotype-matched controls, although lack of the possibility of its site-directed mutagenesis remains a shortcoming. A few mtDNA mutations have been found in derivatives of murine cell lines after culturing under conditions selective for defects in cell respiration, and random mutagenesis can be used in combination with production of transmitochondrial cybrids [33,34]. Human colonocytes from aged individuals harbour mtDNA mutations, and can be studied experimentally in cybrid cells, but they are different from those occurring in proven inherited mitochondrial disease [35]. Therefore focusing on a spectrum of defined mtDNA mutations is difficult on a neutral genetic background when one uses in vitro eukaryotic models. Although restricted to the structural genes of the respiratory enzymes and mainly to the effects expressed at the enzymatic level, use of bacterial models makes it possible to study also more deleterious substitutions than those caused by the naturally occurring mtDNA mutations, and may therefore be a more useful model in elucidating the catalytic function of complex I.

In trans complementation of the nuoJK-knockout strain was used here to introduce several replacement mutations into NuoJ, the E. coli counterpart of human ND6, at sites corresponding to the LHON mutations, residues Tyr59, Met64 and Ala72, and in some nearby residues, in order to investigate the effects of these amino acid substitutions on NDH-1 activity. It may be noted here that Kao et al. [25] have recently inserted mutations directly into the nuoJ gene of the E. coli genome. Importantly, the DB reductase activities in the Y59C, Y59F, M64V and V65G generated by Kao et al. [25] and in the present study by means of in trans complementation were comparable, so that it may be concluded that these two methods are equally applicable. In order to understand the nature of the effect these mutations have on the enzyme, we performed a further analysis of the mutants compared with the previous work of Kao et al. [25] by determining the NDH-1-dependent growth phenotype and kinetic parameters for the mutant enzymes in question, and probed the selected mutant enzymes with a series of NDH-1-specific inhibitors.

Substitution of glycine for Val65 had deleterious effects on NDH-1, in agreement with previous findings (see above), and the NDH-1 activity in this mutant was not ameliorated by higher DB concentrations, which suggests that the defect brought about by the substitution is probably not due to a change in DB affinity, although no detailed analysis of DB binding could be performed. Many of the other mutations generated in the present study affected DB binding by NDH-1. The ubiquinone-binding cavity in complex I is assumed to be large and elongated, harbouring three overlapping inhibitor-binding sites [36,37]. It has been concluded on the basis of changes in inhibitor affinities in several mutants that the ubiquinone-binding site resides in the contact region between the 49 kDa and PSST (20 kDa subunit of mitochondrial complex I) subunits in the hydrophilic part, or at the interface between the hydrophilic part and the membrane arm, see [38] for a review. Although the location of the ND6 subunit with respect to the hydrophilic part is currently unknown, some evidence for it contributing to ubiquinone binding has been presented. In cybrid submitochondrial particles carrying an A72V mutation in ND6, substrate and product inhibition by DB and decylubiquinol respectively have been observed [39]. Contrary to this, substitutions at Met72 of E. coli NuoJ in the present study did not cause any substrate inhibition by DB, but instead induced a 1.7–2.0-fold increase in Km for DB and lowered the affinity for binding a second substrate molecule at the active site, as well as affecting inhibitor resistance (Tables 3 and 4). There are no clinical studies on ubiquinone binding in the 14484/ND6-M64V mutant, although a change in this was proposed on the basis of changes in inhibitor-sensitivity [40]. The Met64 residue has been subjected to mutagenesis in E. coli previously, and according to capsaicin inhibition data it is not involved in ubiquinone binding [25]. So whereas reports on the effects on the human enzyme and its bacterial counterpart are discrepant, the present results demonstrate a 1.4-fold increase in the Km for DB in response to the M64V replacement in E. coli (Table 3).

It is evident that this domain of NuoJ/ND6, which harbours many disease-associated mutations, has some relationship to ubiquinone-binding site(s). In order to understand this relationship better, the mutant with the largest change in ubiquinone affinity, namely NuoJ-M72V, was tested for changes in inhibitor-sensitivity with several classes of complex I inhibitors. Unlike the C-type inhibitor VNA, the inhibitory action of piericidin A, which belongs to class I/type-A, and stigmatellin and myxothiazol did not affect the mutants differently from the controls. Since class I/type-A inhibitors, which are quinone-like, and C-type inhibitors do not compete with each other for binding [36], it is plausible that the C-type inhibitors are bound more distantly in the ubiquinone cavity, at sites leading to the actual active site. Annonin VI is considered a class I/type-A inhibitor [41], but in contrast with piericidin A, it was found here to bind to the NuoJ-M72V mutant enzyme with lower affinity than the control (Table 4). The apparent discrepancy in binding between these two inhibitors can be explained by the rather long structure of the annonin VI molecule, with the possible protein-interacting atoms well spaced out. Therefore this inhibitor molecule could interact in both type-A and type-C inhibitor-binding domains. This is not in conflict with the previous work of Okun et al. [36], where displacement of the rather short type-A inhibitor 2-decyl-4-quinazolinyl amine was not observed with the type-C inhibitor CC 44. In conclusion, NuoJ-M72V changes the binding affinity of inhibitors that interact with the type-C binding domain. The stable binding of the ubiquinone head group in NuoJ/ND6 is unlikely to be due to the low magnitude of the DB affinity changes and the inhibitor-sensitivity pattern observed in the present study. More probably, NuoJ/ND6 might be lining the cavity leading to the actual binding site and providing only transitory interaction with the ubiquinone head group (Figure 4). It was proposed earlier by DeHaan et al. [42] that single amino acid substitutions in ND6 may perturb the secondary structure of the transmembrane helices, thereby causing long-ranging structural derangements. We consider this option unlikely, because none of the mutations introduced should result in steric congestion due to the smaller size of the amino acid introduced as compared with the wild-type, and because the observed changes in inhibitor binding can be assigned to a certain class of inhibitors, whereas larger changes in protein structure would be likely to affect the closely situated type-A inhibitor-binding site as well.

Proximity relationship between the ND6/NuoJ, ND1/NuoH, PSST/NuoB and 49 kDa/NuoD subunits in complex I and tentative location of the inhibitor-sensitive ubiquinone-binding site

Figure 4
Proximity relationship between the ND6/NuoJ, ND1/NuoH, PSST/NuoB and 49 kDa/NuoD subunits in complex I and tentative location of the inhibitor-sensitive ubiquinone-binding site

ND6 may not participate in ubiquinone binding but it does contribute to a channel through which the binding site becomes accessible. The binding sites for class A/I, B/II and C inhibitors are marked with ‘a’, ‘b’ and ‘c’ respectively. The general shape of the E. coli enzyme and the arrangement of the membrane-embedded subunits are from single particle analysis and fragmentation studies performed by Sazanov and co-workers [55].

Figure 4
Proximity relationship between the ND6/NuoJ, ND1/NuoH, PSST/NuoB and 49 kDa/NuoD subunits in complex I and tentative location of the inhibitor-sensitive ubiquinone-binding site

ND6 may not participate in ubiquinone binding but it does contribute to a channel through which the binding site becomes accessible. The binding sites for class A/I, B/II and C inhibitors are marked with ‘a’, ‘b’ and ‘c’ respectively. The general shape of the E. coli enzyme and the arrangement of the membrane-embedded subunits are from single particle analysis and fragmentation studies performed by Sazanov and co-workers [55].

One of the novel findings of the present study is the substrate inhibition by DB in the Y59C mutant. Substrate inhibition by exogenous short-chain ubiquinones [DB, Q1 (ubiquinone-1) and Q2] has been reported for the LHON-associated ND1-A52T mutation in mitochondrial membrane preparations from lymphoblasts [15], and in the presence of ND6-A72V substitution (see above). Substrate inhibition by ubiquinone could be involved in the pathogenesis of LHON, although Q10 did not behave in the same way as its analogues with shorter side chains that were used in enzyme assays in the presence of the 3460/ND1-A52T mutation [15]. Congruently with this, we found higher enzyme activity for the Y59C mutant when endogenous ubiquinone Q9 was used by the enzyme than with DB. Filling the binding cavity with the long isoprenoid tail of the Q9–10 molecule could disable additional binding of ubiquinone, while the shorter tail structures of DB, Q1 or Q2 might allow the binding of another molecule, and high secondary binding affinity in some mutants could lead to the observed substrate inhibition. Despite the lack of substrate inhibition with endogenous ubiquinone, the phenomenon observed here may provide crucial information on the defect in these mutant enzymes under normal respiration conditions, which ultimately presents itself as LHON in humans. If short-chained ubiquinones present higher secondary binding affinities for the substrate in the long binding cavity, then endogenous ubiquinone will also be subject to stronger intermediate binding interaction within the substrate-binding cavity. This could lead to improper channelling of the enzyme for ubiquinone and, in parallel to the mutants, a lowering of the ubiquinone binding affinity. Both may lead to lower occupancy of the actual ubiquinone reduction site, with potentially disastrous consequences for the retinal ganglion cells, by promoting the production of ROS with consequent protein, lipid and DNA modification, impairment of energy production, and as a result, a higher probability of apoptosis. This hypothesis requires further testing, however. It is uncertain whether the altered ubiquinone interaction with complex I alone is sufficient for development of the disease, but the incomplete penetrance of the disease, gender bias, sudden onset of the symptoms and restriction of the affected tissue to a certain cell type are suggestive of other, currently unknown, genetic and/or environmental factors that could also be involved in the presentation of the disease.

In evaluating the effects of a pathogenic mtDNA mutation at the enzymatic level in a bacterial model, one of the main concerns must be the relevance of the results to mitochondrial disease. We compared the effects of NuoJ mutations in E. coli with human disease phenotypes caused by equivalent mutations and two LHON mutations in other subunits (Table 5). The clinical phenotype variables evaluated were: first, the presence of other neurological manifestations in addition to LHON; secondly, disease penetrance and its dependence on gender in the affected families; and thirdly, the occurrence of visual recovery. This comparison shows that although LHON patients with M64V substitution in the ND6 subunit are less severely affected than patients with the Y59C or A72V mutations, no substantial difference in disease penetrance or its male predominance is detectable. The infrequent visual recovery and the occurrence of additional neurological symptoms in the Y59C and A72V mutations seem, however, to be linked to poorer complex I and/or NDH-1 activity. In addition, the DB-binding properties were more severely hampered in the E. coli homologues of the Y59C and A72V mutants than in M64V. The effects on the NDH-1 activity of E. coli mutations homologous with the ND6 substitutions M64V and A72V evidently coincide with to those occurring in the mitochondrion, and therefore it is reasonable to expect that this will also hold good for the Y59C mutation, for which no reports on complex I activity in the mitochondria of patients are currently available.

Table 5
Clinical characteristics and complex I/NDH-1 activity of the ND6 subunit-associated LHON mutations generated in the present study and common LHON mutations in the ND1 and ND4 subunits

Number in parentheses for visual recovery indicates percentage of patients experiencing visual recovery. Complex I (CI) activities are from cybrid cells; no data on complex I activity in patients with C14498T mutation are available. n.a., Not applicable.

         The present study 
 
   Penetrance (%)      
Subunit Mutation Amino acid change Phenotype Male Female Visual recovery CI activity (% of control) References NDH-1 activity‡ (% of control) DB binding 
ND6 C14498T Y59C LHON 40 25 Poor No data [44,4566 Substrate inhibition 
 T14484C M64V LHON 50 12 Good (37–64%) 90 [2,464873 Km 1.4-fold 
 G14459A A72V LDYT* 52 52 No/poor 70 [39,4956§ Km 1.7-fold§ 
   LHON only 29 10      
   Dystonia only 24 38      
   L+D† n.a.      
ND1 G3460A A52T LHON 46 19 Intermediate (15–20%) 30 [47,48,50  
ND4 G11778A R340H LHON 52 10 Intermediate (4–22%) 90–100 [48,51  
         The present study 
 
   Penetrance (%)      
Subunit Mutation Amino acid change Phenotype Male Female Visual recovery CI activity (% of control) References NDH-1 activity‡ (% of control) DB binding 
ND6 C14498T Y59C LHON 40 25 Poor No data [44,4566 Substrate inhibition 
 T14484C M64V LHON 50 12 Good (37–64%) 90 [2,464873 Km 1.4-fold 
 G14459A A72V LDYT* 52 52 No/poor 70 [39,4956§ Km 1.7-fold§ 
   LHON only 29 10      
   Dystonia only 24 38      
   L+D† n.a.      
ND1 G3460A A52T LHON 46 19 Intermediate (15–20%) 30 [47,48,50  
ND4 G11778A R340H LHON 52 10 Intermediate (4–22%) 90–100 [48,51  
*

LDYT, LHON and/or dystonia.

L+D, LHON and dystonia.

Ratio of d-NADH:DB to d-NADH:HAR activity at 100 μM DB concentration.

§

Results of NuoJ-M72V mutant.

Replacement of Tyr109 with phenylalanine was initially selected as a neutral control, but the surprising effects of this replacement on the ubiquinone affinity gave us other thoughts. In addition, the close proximity of Tyr109 to the 14340/ND6-V112M mutation site associated with aminoglycoside-induced or non-aminoglycoside-related sensorineural hearing impairment [43] suggests that this region is of functional significance. A double mutant M64V/M72A was generated to examine whether a methionine residue in either of these positions is a prerequisite for proper complex I function. The background for this assumption was the observation that, whereas human ND6 has a methionine residue in position 64 and alanine residue at position 72, the highly homologous NDH-1 from the soil bacterium Paracoccus denitrificans has alanine in position 64 and methionine at position 72 in its equivalent subunit Nqo10 (Figure 1), while in NuoJ of E. coli, the model system used here, there are methionine residues in both positions 64 and 72. The respiratory defects caused by the single M64V and M72A mutations in NuoJ were not cumulative as expected, and the Km for DB was in the same range as for M64V and M72A. In contrast with the single M64V and M72V mutations, the double mutant did, however, show some tendency for substrate inhibition (Table 3). Co-occurrence of the two mutations in the same protein domain seems to have more drastic effects on the enzyme than the occurrence of either mutant alone, as expected.

This bacterial model for LHON-associated mutations in the mtDNA gene of the ND6 subunit of complex I turned out to be a useful tool for analysing the effects of these mutations at the enzymatic level. This same approach is applicable to the study of other mtDNA mutations, including the rare ones, and may serve to reveal not only the mechanism underlying the pathogenesis of LHON but also the function of complex I. We hope that this model system will prove useful in the search of suitable therapeutic molecules or remedies for this disease.

This study was supported by grants from the Academy of Finland Council of Health, Sigrid Juselius Foundation (to I. E. H.), the Eye Foundation (Silmäsäätiö) and the Evald and Hilda Nissi Foundation (to M. K.). Aila Holappa and Raija Pietilä are gratefully acknowledged for skilful technical assistance.

Abbreviations

     
  • d-NADH

    deamino-NADH

  •  
  • DB

    decylubiquinone

  •  
  • HAR

    hexammine ruthenium

  •  
  • LB

    Luria–Bertani

  •  
  • LHON

    Leber hereditary optic neuropathy

  •  
  • mtDNA

    mitochondrial DNA

  •  
  • NDH-1

    bacterial NADH:quinone oxidoreductase

  •  
  • PSST

    20 kDa subunit of mitochondrial complex I

  •  
  • Q1 etc.

    ubiquinone-1 etc.

  •  
  • ROS

    reactive oxygen species

  •  
  • VNA

    N-vanillylnonanamide

References

References
1
Nikoskelainen
E.
Hoyt
W. F.
Nummelin
K.
Ophthalmoscopic findings in Leber's hereditary optic neuropathy. II. The fundus findings in the affected family members
Arch. Ophthalmol.
1983
, vol. 
101
 (pg. 
1059
-
1068
)
2
Johns
D. R.
Heher
K. L.
Miller
N. R.
Smith
K. H.
Leber's hereditary optic neuropathy. Clinical manifestations of the 14484 mutation
Arch. Ophthalmol.
1993
, vol. 
111
 (pg. 
495
-
498
)
3
Wallace
D. C.
Singh
G.
Lott
M. T.
Hodge
J. A.
Schurr
T. G.
Lezza
A. M.
Elsas
L. J.
Nikoskelainen
E. K.
Mitochondrial DNA mutation associated with Leber's hereditary optic neuropathy
Science
1988
, vol. 
242
 (pg. 
1427
-
1430
)
4
Man
P. Y.
Turnbull
D. M.
Chinnery
P. F.
Leber hereditary optic neuropathy
J. Med. Genet.
2002
, vol. 
39
 (pg. 
162
-
169
)
5
Hinchliffe
P.
Sazanov
L. A.
Organization of iron–sulfur clusters in respiratory complex I
Science
2005
, vol. 
309
 (pg. 
771
-
774
)
6
Bogachev
A. V.
Murtazina
R. A.
Skulachev
V. P.
H+/e− stoichiometry for NADH dehydrogenase I and dimethyl sulfoxide reductase in anaerobically grown Escherichia coli cells
J. Bacteriol.
1996
, vol. 
178
 (pg. 
6233
-
6237
)
7
Carroll
J.
Fearnley
I. M.
Shannon
R. J.
Hirst
J.
Walker
J. E.
Analysis of the subunit composition of complex I from bovine heart mitochondria
Mol. Cell. Proteomics
2003
, vol. 
2
 (pg. 
117
-
126
)
8
Friedrich
T.
Steinmuller
K.
Weiss
H.
The proton-pumping respiratory complex I of bacteria and mitochondria and its homologue in chloroplasts
FEBS Lett.
1995
, vol. 
367
 (pg. 
107
-
111
)
9
Anderson
S.
Bankier
A. T.
Barrell
B. G.
de Bruijn
M. H.
Coulson
A. R.
Drouin
J.
Eperon
I. C.
Nierlich
D. P.
Roe
B. A.
Sanger
F.
Schreier
P. H.
Smith
A. J.
Staden
R.
Young
I. G.
Sequence and organization of the human mitochondrial genome
Nature
1981
, vol. 
290
 (pg. 
457
-
465
)
10
DiBernardo
S. D.
Yano
T.
Yagi
T.
Exploring the membrane domain of the reduced nicotinamide adenine dinucleotide-quinone oxidoreductase of Paracoccus denitrificans: characterization of the NQO7 subunit
Biochemistry
2000
, vol. 
39
 (pg. 
9411
-
9418
)
11
Carelli
V.
Ross-Cisneros
F. N.
Sadun
A. A.
Mitochondrial dysfunction as a cause of optic neuropathies
Prog. Retin. Eye Res.
2004
, vol. 
23
 (pg. 
53
-
89
)
12
Chinnery
P. F.
Brown
D. T.
Andrews
R. M.
Singh-Kler
R.
Riordan-Eva
P.
Lindley
J.
Applegarth
D. A.
Turnbull
D. M.
Howell
N.
The mitochondrial ND6 gene is a hot spot for mutations that cause Leber's hereditary optic neuropathy
Brain
2001
, vol. 
124
 (pg. 
209
-
218
)
13
Valentino
M. L.
Barboni
P.
Ghelli
A.
Bucchi
L.
Rengo
C.
Achilli
A.
Torroni
A.
Lugaresi
A.
Lodi
R.
Barbiroli
B.
Dotti
M.
, et al. 
The ND1 gene of complex I is a mutational hot spot for Leber's hereditary optic neuropathy
Ann. Neurol.
2004
, vol. 
56
 (pg. 
631
-
641
)
14
Majander
A.
Huoponen
K.
Savontaus
M. L.
Nikoskelainen
E.
Wikström
M.
Electron transfer properties of NADH:ubiquinone reductase in the ND1/3460 and the ND4/11778 mutations of the Leber hereditary optic neuroretinopathy (LHON)
FEBS Lett.
1991
, vol. 
292
 (pg. 
289
-
292
)
15
Majander
A.
Finel
M.
Savontaus
M. L.
Nikoskelainen
E.
Wikström
M.
Catalytic activity of complex I in cell lines that possess replacement mutations in the ND genes in Leber's hereditary optic neuropathy
Eur. J. Biochem.
1996
, vol. 
239
 (pg. 
201
-
207
)
16
Smith
P. R.
Cooper
J. M.
Govan
G. G.
Harding
A. E.
Schapira
A. H.
Platelet mitochondrial function in Leber's hereditary optic neuropathy
J. Neurol. Sci.
1994
, vol. 
122
 (pg. 
80
-
83
)
17
Baracca
A.
Solaini
G.
Sgarbi
G.
Lenaz
G.
Baruzzi
A.
Schapira
A. H.
Martinuzzi
A.
Carelli
V.
Severe impairment of complex I-driven adenosine triphosphate synthesis in Leber hereditary optic neuropathy cybrids
Arch. Neurol.
2005
, vol. 
62
 (pg. 
730
-
736
)
18
Komaki
H.
Akanuma
J.
Iwata
H.
Takahashi
T.
Mashima
Y.
Nonaka
I.
Goto
Y.
A novel mtDNA C11777A mutation in Leigh syndrome
Mitochondrion
2003
, vol. 
2
 (pg. 
293
-
304
)
19
Wong
A.
Cavelier
L.
Collins-Schramm
H. E.
Seldin
M. F.
McGrogan
M.
Savontaus
M. L.
Cortopassi
G. A.
Differentiation-specific effects of LHON mutations introduced into neuronal NT2 cells
Hum. Mol. Genet.
2002
, vol. 
11
 (pg. 
431
-
438
)
20
Beretta
S.
Mattavelli
L.
Sala
G.
Tremolizzo
L.
Schapira
A. H.
Martinuzzi
A.
Carelli
V.
Ferrarese
C.
Leber hereditary optic neuropathy mtDNA mutations disrupt glutamate transport in cybrid cell lines
Brain.
2004
, vol. 
127
 (pg. 
2183
-
2192
)
21
Danielson
S. R.
Wong
A.
Carelli
V.
Martinuzzi
A.
Schapira
A. H.
Cortopassi
G. A.
Cells bearing mutations causing Leber's hereditary optic neuropathy are sensitized to Fas-induced apoptosis
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
5810
-
5815
)
22
Ghelli
A.
Zanna
C.
Porcelli
A. M.
Schapira
A. H.
Martinuzzi
A.
Carelli
V.
Rugolo
M.
Leber's hereditary optic neuropathy (LHON) pathogenic mutations induce mitochondrial-dependent apoptotic death in transmitochondrial cells incubated with galactose medium
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
4145
-
4150
)
23
Bandelt
H. J.
Yao
Y. G.
Salas
A.
Kivisild
T.
Bravi
C. M.
High penetrance of sequencing errors and interpretative shortcomings in mtDNA sequence analysis of LHON patients
Biochem. Biophys. Res. Commun.
2007
, vol. 
352
 (pg. 
283
-
291
)
24
Kervinen
M.
Pätsi
J.
Finel
M.
Hassinen
I. E.
A pair of membrane-embedded acidic residues in the NuoK subunit of Escherichia coli NDH-1, a counterpart of the ND4L subunit of the mitochondrial complex I, are required for high ubiquinone reductase activity
Biochemistry
2004
, vol. 
43
 (pg. 
773
-
781
)
25
Kao
M. C.
Di Bernardo
S.
Nakamaru-Ogiso
E.
Miyoshi
H.
Matsuno-Yagi
A.
Yagi
T.
Characterization of the membrane domain subunit NuoJ (ND6) of the NADH-quinone oxidoreductase from Escherichia coli by chromosomal DNA manipulation
Biochemistry
2005
, vol. 
44
 (pg. 
3562
-
3571
)
26
Mitchell
A. L.
Elson
J. L.
Howell
N.
Taylor
R. W.
Turnbull
D. M.
Sequence variation in mitochondrial complex I genes: mutation or polymorphism?
J. Med. Genet.
2006
, vol. 
43
 (pg. 
175
-
179
)
27
Oden
K. L.
DeVeaux
L. C.
Vibat
C. R.
Cronan
J. E.
Jr
Gennis
R. B.
Genomic replacement in Escherichia coli K-12 using covalently closed circular plasmid DNA
Gene
1990
, vol. 
96
 (pg. 
29
-
36
)
28
Leif
H.
Sled
V. D.
Ohnishi
T.
Weiss
H.
Friedrich
T.
Isolation and characterization of the proton-translocating NADH: ubiquinone oxidoreductase from Escherichia coli
Eur. J. Biochem.
1995
, vol. 
230
 (pg. 
538
-
548
)
29
Kervinen
M.
Hinttala
R.
Helander
H. M.
Kurki
S.
Uusimaa
J.
Finel
M.
Majamaa
K.
Hassinen
I. E.
The MELAS mutations 3946 and 3949 perturb the critical structure in a conserved loop of the ND1 subunit of mitochondrial complex I
Hum. Mol. Genet.
2006
, vol. 
15
 (pg. 
2543
-
2552
)
30
Lowry
O. H.
Rosebrough
N. J.
Farr
A. L.
Randall
R. J.
Protein measurement with the Folin phenol reagent
J. Biol. Chem.
1951
, vol. 
193
 (pg. 
265
-
275
)
31
Matsushita
K.
Ohnishi
T.
Kaback
H. R.
NADH-ubiquinone oxidoreductases of the Escherichia coli aerobic respiratory chain
Biochemistry
1987
, vol. 
26
 (pg. 
7732
-
7737
)
32
Dixon
M.
Webb
E. C.
Enzymes
1971
London
Longman
33
Bai
Y.
Attardi
G.
The mtDNA-encoded ND6 subunit of mitochondrial NADH dehydrogenase is essential for the assembly of the membrane arm and the respiratory function of the enzyme
EMBO J.
1998
, vol. 
17
 (pg. 
4848
-
4858
)
34
Moreno-Loshuertos
R.
Acin-Perez
R.
Fernandez-Silva
P.
Movilla
N.
Perez-Martos
A.
Rodriguez
d. C.
Gallardo
M. E.
Enriquez
J. A.
Differences in reactive oxygen species production explain the phenotypes associated with common mouse mitochondrial DNA variants
Nat. Genet.
2006
, vol. 
38
 (pg. 
1261
-
1268
)
35
Pye
D.
Kyriakouli
D. S.
Taylor
G. A.
Johnson
R.
Elstner
M.
Meunier
B.
Chrzanowska-Lightowlers
Z. M.
Taylor
R. W.
Turnbull
D. M.
Lightowlers
R. N.
Production of transmitochondrial cybrids containing naturally occurring pathogenic mtDNA variants
Nucleic Acids Res.
2006
, vol. 
34
 pg. 
e95
 
36
Okun
J. G.
Lummen
P.
Brandt
U.
Three classes of inhibitors share a common binding domain in mitochondrial complex I (NADH:ubiquinone oxidoreductase)
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
2625
-
2630
)
37
Zickermann
V.
Bostina
M.
Hunte
C.
Ruiz
T.
Radermacher
M.
Brandt
U.
Functional implications from an unexpected position of the 49-kDa subunit of NADH:ubiquinone oxidoreductase
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
29072
-
29078
)
38
Brandt
U.
Energy converting NADH:quinone oxidoreductase (complex I)
Annu. Rev. Biochem.
2006
, vol. 
75
 (pg. 
69
-
92
)
39
Jun
A. S.
Trounce
I. A.
Brown
M. D.
Shoffner
J. M.
Wallace
D. C.
Use of transmitochondrial cybrids to assign a complex I defect to the mitochondrial DNA-encoded NADH dehydrogenase subunit 6 gene mutation at nucleotide pair 14459 that causes Leber hereditary optic neuropathy and dystonia
Mol. Cell. Biol.
1996
, vol. 
16
 (pg. 
771
-
777
)
40
Carelli
V.
Ghelli
A.
Bucchi
L.
Montagna
P.
De Negri
A.
Leuzzi
V.
Carducci
C.
Lenaz
G.
Lugaresi
E.
Degli
E. M.
Biochemical features of mtDNA 14484 (ND6/M64V) point mutation associated with Leber's hereditary optic neuropathy
Ann. Neurol.
1999
, vol. 
45
 (pg. 
320
-
328
)
41
Friedrich
T.
van Heek
P.
Leif
H.
Ohnishi
T.
Forche
E.
Kunze
B.
Jansen
R.
Trowitzsch-Kienast
W.
Hofle
G.
Reichenbach
H.
Weiss
H.
Two binding sites of inhibitors in NADH: ubiquinone oxidoreductase (complex I). Relationship of one site with the ubiquinone-binding site of bacterial glucose:ubiquinone oxidoreductase
Eur. J. Biochem.
1994
, vol. 
219
 (pg. 
691
-
698
)
42
DeHaan
C.
Habibi-Nazhad
B.
Yan
E.
Salloum
N.
Parliament
M.
Allalunis-Turner
J.
Mutation in mitochondrial complex I ND6 subunit is associated with defective response to hypoxia in human glioma cells
Mol. Cancer
2004
, vol. 
3
 pg. 
19
 
43
Zhao
L.
Wang
Q.
Qian
Y.
Li
R.
Cao
J.
Hart
L. C.
Zhai
S.
Han
D.
Young
W. Y.
Guan
M. X.
Clinical evaluation and mitochondrial DNA sequence analysis in two Chinese families with aminoglycoside-induced and non-syndromic hearing loss
Biochem. Biophys. Res. Commun.
2005
, vol. 
336
 (pg. 
967
-
973
)
44
Wissinger
B.
Besch
D.
Baumann
B.
Fauser
S.
Christ-Adler
M.
Jurklies
B.
Zrenner
E.
Leo-Kottler
B.
Mutation analysis of the ND6 gene in patients with Lebers hereditary optic neuropathy
Biochem. Biophys. Res. Commun.
1997
, vol. 
234
 (pg. 
511
-
515
)
45
Leo-Kottler
B.
Christ-Adler
M.
Baumann
B.
Zrenner
E.
Wissinger
B.
Leber's hereditary optic neuropathy: clinical and molecular genetic results obtained in a family with a new point mutation at nucleotide position 14498 in the ND 6 gene
Ger. J. Ophthalmol.
1996
, vol. 
5
 (pg. 
233
-
240
)
46
Mackey
D.
Howell
N.
A variant of Leber hereditary optic neuropathy characterized by recovery of vision and by an unusual mitochondrial genetic etiology
Am. J. Hum. Genet.
1992
, vol. 
51
 (pg. 
1218
-
1228
)
47
Brown
M. D.
Trounce
I. A.
Jun
A. S.
Allen
J. C.
Wallace
D. C.
Functional analysis of lymphoblast and cybrid mitochondria containing the 3460, 11778, or 14484 Leber's hereditary optic neuropathy mitochondrial DNA mutation
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
39831
-
39836
)
48
Spruijt
L.
Kolbach
D. N.
de Coo
R. F.
Plomp
A. S.
Bauer
N. J.
Smeets
H. J.
Die-Smulders
C. E.
Influence of mutation type on clinical expression of Leber hereditary optic neuropathy
Am. J. Ophthalmol.
2006
, vol. 
141
 (pg. 
676
-
682
)
49
Novotny
E. J.
Jr
Singh
G.
Wallace
D. C.
Dorfman
L. J.
Louis
A.
Sogg
R. L.
Steinman
L.
Leber's disease and dystonia: a mitochondrial disease
Neurology
1986
, vol. 
36
 (pg. 
1053
-
1060
)
50
Johns
D. R.
Smith
K. H.
Miller
N. R.
Leber's hereditary optic neuropathy. Clinical manifestations of the 3460 mutation
Arch. Ophthalmol.
1992
, vol. 
110
 (pg. 
1577
-
1581
)
51
Newman
N. J.
Lott
M. T.
Wallace
D. C.
The clinical characteristics of pedigrees of Leber's hereditary optic neuropathy with the 11778 mutation
Am. J. Ophthalmol.
1991
, vol. 
111
 (pg. 
750
-
762
)
52
Kao
M. C.
Di Bernardo
S.
Matsuno-Yagi
A.
Yagi
T.
Characterization and topology of the membrane domain Nqo10 subunit of the proton-translocating NADH-quinone oxidoreductase of Paracoccus denitrificans
Biochemistry
2003
, vol. 
42
 (pg. 
4534
-
4543
)
53
Ugalde
C.
Triepels
R. H.
Coenen
M. J.
van den Heuvel
L. P.
Smeets
R.
Uusimaa
J.
Briones
P.
Campistol
J.
Majamaa
K.
Smeitink
J. A.
Nijtmans
L. G.
Impaired complex I assembly in a Leigh syndrome patient with a novel missense mutation in the ND6 gene
Ann. Neurol.
2003
, vol. 
54
 (pg. 
665
-
669
)
54
Ravn
K.
Wibrand
F.
Hansen
F. J.
Horn
N.
Rosenberg
T.
Schwartz
M.
An mtDNA mutation, 14453G→A, in the NADH dehydrogenase subunit 6 associated with severe MELAS syndrome
Eur. J. Hum. Genet.
2001
, vol. 
9
 (pg. 
805
-
809
)
55
Baranova
E. A.
Holt
P. J.
Sazanov
L. A.
Projection structure of the membrane domain of Escherichia coli respiratory complex I at 8 Å resolution
J. Mol. Biol.
2007
, vol. 
366
 (pg. 
140
-
154
)

Supplementary data