Nuclear DNA origin of mitochondrial complex I deficiency in fatal infantile lactic acidosis evidenced by transnuclear complementation of cultured fibroblasts.

We have studied complex I (NADH-ubiquinone reductase) defects of the mitochondrial respiratory chain in 2 infants who died in the neonatal period from 2 different neurological forms of severe neonatal lactic acidosis. Specific and marked decrease in complex I activity was documented in muscle, liver, and cultured skin fibroblasts. Biochemical characterization and study of the genetic origin of this defect were performed using cultured fibroblasts. Immunodetection of 6 nuclear DNA-encoded (20, 23, 24, 30, 49, and 51 kDa) and 1 mitochondrial DNA-encoded (ND1) complex I subunits in fibroblast mitochondria revealed 2 distinct patterns. In 1 patient, complex I contained reduced amounts of the 24- and 51-kDa subunits and normal amounts of all the other investigated subunits. In the second patient, amounts of all the investigated subunits were severely decreased. The data suggest partial or extensive impairment of complex I assembly in both patients. Cell fusion experiments between 143B206 rho degrees cells, fully depleted of mitochondrial DNA, and fibroblasts from both patients led to phenotypic complementation of the complex I defects in mitochondria of the resulting cybrid cells. These results indicate that the complex I defects in the 2 reported cases are due to nuclear gene mutations.

The nuoM arg368his mutation in NADH:ubiquinone oxidoreductase from Rhodobacter capsulatus: a model for the human nd4-11778 mtDNA mutation associated with Leber's hereditary optic neuropathy.

Mutation at position 11778 in the nd4 gene of the human mitochondrial complex I is associated with Leber's hereditary optic neuropathy. Type I NADH:ubiquinone oxidoreductase of Rhodobacter capsulatus displays similar properties to complex I of the mitochondrial respiratory chain. The NUOM subunit of the bacterial enzyme is homologous to the ND4 subunit. Disruption of the nuoM gene led to a bacterial mutant exhibiting a defect in complex I activity and assembly. A nuoM-1103 point mutant reproducing the nd4-11778 mutation has been introduced in the R. capsulatus genome. This mutant showed a reduced ability to grow in a medium containing malate instead of lactate which indicated a clear impairment in oxidative phosphorylation capacity. NADH supported respiration of porous bacterial cells was significantly decreased in the nuoM-1103 mutant while no significant reduction could be observed in isolated bacterial membranes. As it has been observed in the case of the nd4-11778 mitochondrial mutation, proton-pump activity of the bacterial enzyme was not affected by the nuoM-1103 mutation. All these data which reproduce most of the biochemical features observed in patient mitochondria harboring the nd4-11778 mutation show that the R. capsulatus complex I might be used as a useful model to investigate mutations of the mitochondrial DNA which are associated with complex I deficiencies in human pathologies.

The 49-kDa subunit of NADH-ubiquinone oxidoreductase (Complex I) is involved in the binding of piericidin and rotenone, two quinone-related inhibitors.

Piericidin is a potent inhibitor of the mitochondrial and bacterial type I NADH-ubiquinone oxidoreductases (Complex I) and is considered to bind at or close to the ubiquinone binding site(s) of the enzyme. Piericidin-resistant mutants of the bacterium Rhodobacter capsulatus have been isolated and the present work demonstrates that a single missense mutation at the level of the gene encoding the peripheral 49-kDa/NUOD subunit of Complex I is definitely associated with this resistance. Based on this original observation, we propose a model locating the binding site for piericidin (and quinone) at the interface between the hydrophilic and hydrophobic domains of Complex I.

Genomic structure of the human NDUFS8 gene coding for the iron-sulfur TYKY subunit of the mitochondrial NADH:ubiquinone oxidoreductase.

The structural organization of the NDUFS8 gene coding for the TYKY subunit of the human mitochondrial NADH:ubiquinone oxidoreductase (Complex I) has been determined by sequencing of a genomic fragment cloned from a cosmid library. The NDUFS8 gene is located on chromosome 11q13 immediately downstream of the ALDH7 isoform gene. It spans about 6kb and contains seven exons ranging in size from 51 to 186bp. Three CCAAT box sequence motifs are present upstream of the transcription start. Sp1 and NRF1 binding site motifs are present in the first intron. Expression of the gene is ubiquitous but predominant in heart and skeletal muscle. Immunodetection of the TYKY subunit in placental mitochondria after two-dimensional gel electrophoresis revealed that the mature protein has a molecular mass of 22kDa and a pI in the range of 4.9-5.0.

The complex I from Rhodobacter capsulatus.

The NADH-ubiquinone oxidoreductase (type I NDH) of Rhodobacter capsulatus is a multisubunit enzyme encoded by the 14 genes of the nuo operon. This bacterial enzyme constitutes a valuable model for the characterization of the mitochondrial Complex I structure and enzymatic mechanism for the following reasons. (i) The mitochondria-encoded ND subunits are not readily accessible to genetic manipulation. In contrast, the equivalents of the mitochondrial ND1, ND2, ND4, ND4L, ND5 and ND6 genes can be easily mutated in R. capsulatus by homologous recombination. (ii) As illustrated in the case of ND1 gene, point mutations associated with human cytopathies can be reproduced and studied in this model system. (iii) The R. capsulatus model also allows the recombinant manipulations of iron-sulfur (Fe-S) subunits and the assignment of Fe-S clusters as illustrated in the case of the NUOI subunit (the equivalent of the mitochondrial TYKY subunit). (iv) Finally, like mitochondrial Complex I, the NADH-ubiquinone oxidoreductase of R. capsulatus is highly sensitive to the inhibitor piericidin-A which is considered to bind to or close to the quinone binding site(s) of Complex I. Therefore, isolation of R. capsulatus mutants resistant to piericidin-A represents a straightforward way to map the inhibitor binding sites and to try and define the location of quinone binding site(s) in the enzyme. These illustrations that describe the interest in the R. capsulatus NADH-ubiquinone oxidoreductase model for the general study of Complex I will be critically developed in the present review.

Immuno-purification of a dimeric subcomplex of the respiratory NADH-CoQ reductase of Rhodobacter capsulatus equivalent to the FP fraction of the mitochondrial complex I.

The Rhodobacter capsulatus genes encoding the NUOE and NUOF subunits, equivalent to the 24 kDa and 51 kDa subunits of the mammalian mitochondrial complex I, have been sequenced. According to the nucleotide sequence, the NUOE subunit is 389 amino acids long and has a molecular mass of 41.3 kDa. In comparison to the mitochondrial equivalent subunit, NUOE is extended at the C terminus by more than 150 amino acids. The NUOF subunit is 431 amino acids long and has a molecular mass of 47.1 kDa. A subcomplex containing both the NUOE and NUOF subunits was extracted by detergent treatment of R. capsulatus membranes and immuno-purified. This subcomplex is homologous to the mitochondrial FP fragment. Mass spectrometry after trypsin treatment of the NUOE subunit validates the atypical primary structure deduced from the sequence of the gene.

cDNA sequence and chromosomal localization of the NDUFS8 human gene coding for the 23 kDa subunit of the mitochondrial complex I.

We have sequenced the cDNA for the 23 kDa subunit of the human mitochondrial respiratory complex I. The deduced protein consists of 210 amino acids (Mr = 23705 Da) with a 34 amino acid N terminus presumably acting as a presequence for mitochondrial import. The predicted mature protein (Mr = 20290 Da) is 92% identical to the bovine mitochondrial subunit and 72% to the Rhodobacter capsulatus NUOI counterpart. Two clusters of four cysteine residues are conserved among these proteins. The gene (NDUFS8) coding for the human subunit has been mapped to chromosome 11q13.

The NuoI subunit of the Rhodobacter capsulatus respiratory Complex I (equivalent to the bovine TYKY subunit) is required for proper assembly of the membraneous and peripheral domains of the enzyme.

The nuoI gene that encodes a ferredoxin-like subunit of the Rhodobacter capsulatus Complex I (a subunit equivalent to the bovine TYKY subunit) was mutated by homologous recombination. Both a nuoI-deleted mutant (delta nuoI mutant) and a point mutant in which Cys74 was replaced by a serine (C74S mutant) proved to be completely deficient in Complex I activity. These strains were unable to grow under anaerobic photosynthetic conditions. Their cytoplasmic membranes were also characterized by the absence of specific EPR signals assigned to FeS clusters N1 and N2. Immunochemical analysis of the mutant membranes with subunit-specific antibodies showed that the peripheral subunits were not assembled. Trans-complementation of the mutant strains by a native nuoI gene restored the wild-type phenotypes. In the C74S mutant, a limited amount of NuoI subunit still bound to the membraneous domain of Complex I, which is an indication that NuoI directly interacts with this domain. All these results clearly show that NuoI plays a critical role in the connection between the membraneous domain and the peripheral domain of Complex I.

Identification of five Rhodobacter capsulatus genes encoding the equivalent of ND subunits of the mitochondrial NADH-ubiquinone oxidoreductase.

We previously reported the sequencing of two genes (ndhA and ndhI) encoding two of the subunits of the type-I NADH-ubiquinone oxidoreductase from Rhodobacter capsulatus (Rc). The present paper deals with the cloning and characterization of a chromosomal fragment clustering five new Rc genes which encode subunits of this enzyme. This gene cluster is located immediately downstream from ndhA and ndhI, and also contains two unidentified open reading frames (urf2, urf3). The five genes, nuoJ, nuoK, nuoL, nuoM and nuoN, encode proteins related, respectively, to mitochondrial (mt) subunits ND6, ND4L, ND5, ND4 and ND2. The overall organization of the nuo genes identified in Rc shows similarity to that of the Paracoccus denitrificans (Pd) nqo gene cluster.

A mutation located at the 5' splice junction sequence of intron 3 in the p67phox gene causes the lack of p67phox mRNA in a patient with chronic granulomatous disease.

Chronic granulomatous disease (CGD) is due to a functional defect of the O2(-)-generating NADPH oxidase of neutrophils. Mutations resulting in CGD have been shown to occur in only four genes, thus identifying the main components of the oxidase complex, namely the two subunits of a membrane-bound cytochrome b and two cytosolic factors of activation of 67 kD (p67phox) and 47 kD (p47phox). The present study deals with the biochemical and genetic analysis of the defect in a patient suffering from a p67phox-deficient form of CGD. The p67phox deficiency was ascertained by immunochemistry and the ability of recombinant p67phox to restore NADPH oxidase activity using a cell-free system of oxidase activation. The cellular extracts from the proband contained no p67phox protein and no p67phox mRNA when assayed by Western and Northern blot analysis. However, reverse transcription of mRNA and subsequent cDNA amplification by polymerase chain reaction using specific p67phox primers showed that trace amounts of a p67phox mRNA deleted for exon 3 were synthesized in the patient immortalized B lymphocytes. Sequence analysis of the genomic DNA showed a T-to-C transition at position +2 of intron 3. This point mutation in the consensus 5' splice site of the intron 3 was probably responsible for lack of accumulation of mRNA and also for the skipping of exon 3 detected in the few mRNA molecules that escaped cellular degradation.

Activation of Escherichia coli prohemolysin to the membrane-targetted toxin by HlyC-directed ACP-dependent fatty acylation.

Hemolysin (HlyA) and related toxins of Escherichia coli and other Gram-negative pathogenic bacteria form membrane pores in cells of the host immune system, causing cell dysfunction and death. An insight into the mechanism by which HlyA is targetted to mammalian cell membranes was achieved by establishing in vitro activation of the non-toxic precursor proHlyA. By this approach we have discovered that conversion of proHlyA to the post-translational active HlyA toxin is determined by fatty acylation of proHlyA in an apparently novel process directed by the HlyC homodimer activator protein, and dependent upon the cellular acyl carrier protein (ACP). By further exploiting the in vitro activation system it is now possible to obtain direct evidence that HlyC binds to an internal recognition sequence in the proHlyA precursor, in this way providing specificity for the transfer to proHlyA of a fatty acid moiety carried by the ACP. It is possible that the fatty acid modification determines directly the binding of HlyA to mammalian membrane lipids, thus initiating the toxin interaction with the target cells.

The ATP synthase (F0-F1) complex in oxidative phosphorylation.

The transmembrane electrochemical proton gradient generated by the redox systems of the respiratory chain in mitochondria and aerobic bacteria is utilized by proton translocating ATP synthases to catalyze the synthesis of ATP from ADP and P(i). The bacterial and mitochondrial H(+)-ATP synthases both consist of a membranous sector, F0, which forms a H(+)-channel, and an extramembranous sector, F1, which is responsible for catalysis. When detached from the membrane, the purified F1 sector functions mainly as an ATPase. In chloroplasts, the synthesis of ATP is also driven by a proton motive force, and the enzyme complex responsible for this synthesis is similar to the mitochondrial and bacterial ATP synthases. The synthesis of ATP by H(+)-ATP synthases proceeds without the formation of a phosphorylated enzyme intermediate, and involves co-operative interactions between the catalytic subunits.

Does pyrophosphate bind to the catalytic sites of mitochondrial F1-ATPase?

The interactions between the pyrophosphate (PPi) binding sites and the nucleotide binding sites on mitochondrial F1-ATPase have been investigated, using F1 preparations containing different numbers of catalytic and noncatalytic nucleotide-binding sites occupied by ligands. In all cases, the total number of moles of bound nucleotides and PPi per mole of F1 was less than or equal to six. F1 preparations containing either three or two filled noncatalytic sites and no filled catalytic sites (referred as F1[3,0] and F1[2,0]) were found to bind 3 mol of PPi/mol of F1. Tight binding of ADP-fluoroberyllate complexes to two of the catalytic sites of F1 converted the three heterogeneous PPi-binding sites into three homogeneous binding sites, each exhibiting the same affinity for PPi. The addition of PPi at saturating concentrations to F1 containing GDP bound to two catalytic sites (F1[2,2]) resulted in the release of 1 mol of GDP. Furthermore, the addition of PPi to F1 filled with ADP-fluoroberyllate at the catalytic sites resulted in the release of 1 mol of tightly bound ADP/mol of F1. Taken together, these results indicate that PPi binds to specific sites that interact with both the catalytic and the noncatalytic nucleotide-binding sites of F1.

In vitro activation of Escherichia coli prohaemolysin to the mature membrane-targeted toxin requires HlyC and a low molecular-weight cytosolic polypeptide.

The c. 110 kDa haemolysin toxin secreted by Escherichia coli and other pathogenic Gram-negative bacteria is synthesized as the non-toxic precursor, prohaemolysin (proHlyA), which is unable to target mammalian cell membranes until activated intracellularly by an unknown mechanism dependent upon the coexpressed c. 20 kDa protein, HlyC. We have established in vitro post-translational activation of proHlyA in membrane-depleted cell extract fractions from E. coli recombinant strains containing (separately) the proHlyA and HlyC proteins. In vitro activation was calcium-independent and effective over a pH range of 6 to 9 and at temperatures from 42 degrees C to 4 degrees C. HlyC cell extract was also able to activate proHlyA which had been secreted out of cells containing the export proteins HlyB and HlyD. Fractionation of HlyC cell extracts by sucrose gradient centrifugation and molecular weight chromatography revealed activating fractions as having a molecular mass of 40 kDa, suggesting that the HlyC activator is present physiologically in a multimeric form. Cell extracts containing activation-competent HlyC and proHlyA were inactive following dialysis, but activity was restored by complementation with a cell extract lacking both proteins. HlyC and proHlyA proteins which were overproduced separately from recombinant expression plasmids were inactive following purification, but activity could again be restored with a Hly-negative cell extract. These experiments demonstrated that HlyC is not sufficient for activation; an additional cellular factor is required. The cellular factor was found in enterobacteria but not other bacteria or eukaryotic cells. It was cytosolic, protease-sensitive, and behaved as a c. 10 kDa polypeptide in a number of assays including dialysis, sucrose gradient centrifugation, and gel filtration chromatography. Thus activation was possible in a defined in vitro reaction containing only purified proHlyA, HlyC, and the cellular factor. Kinetic studies in which the relative concentrations of the three components of proHlyA activation were varied suggested that neither HlyC nor the cellular factor acts as a conventional enzyme, with each participating in a finite number of activation events.

Activation of Escherichia coli prohaemolysin to the mature toxin by acyl carrier protein-dependent fatty acylation.

Haemolysin secreted by pathogenic Escherichia coli binds to mammalian cell membranes, disrupting cellular activities and lysing cells by pore-formation. It is synthesized as nontoxic prohaemolysin (proHlyA), which is activated intracellularly by a mechanism dependent on the cosynthesized HlyC. Haemolysin is one of a family of membrane-targeted toxins, including the leukotoxins of Pasteurella and Actinobacillus and the bifunctional adenylate cyclase haemolysin of Bordetella pertussis, which require this protoxin activation 1-5. HlyC alone cannot activate proHlyA, but requires a cytosolic activating factor6. Here we report the cytosolic activating factor is identical to the acyl carrier protein and that activation to mature toxin is achieved by the transfer of a fatty acyl group from acyl carrier protein to proHlyA. Only acyl carrier protein, not acyl-CoA, can promote HlyC-directed proHlyA acylation, but a range of acyl groups are effective.

Fluoroaluminum and fluoroberyllium nucleoside diphosphate complexes as probes of the enzymatic mechanism of the mitochondrial F1-ATPase.

The mechanism by which fluoride and aluminum or beryllium in combination with ADP inhibit beef heart mitochondrial F1-ATPase was investigated. The kinetics of inhibition depended on the nature of the anion present in the F1-ATPase assay medium. Inhibition required the presence of Mg2+ and developed more rapidly with sulfite and sulfate than with chloride, i.e., with anions which activate F1-ATPase activity. The ADP-fluorometal complexes were bound quasi-irreversibly to F1, and each mole of the inhibitory nucleotide-fluorometal complex was tightly associated with 1 mol of Mg2+. One mole of nucleotide-fluorometal complex was able to inhibit the activity of 1 mol of catalytic site in F1. Direct measurements of bound fluoride, aluminum, beryllium, and ADP indicated that the F1-bound ADP-fluorometal complexes are of the following types: ADP1A11F4, ADP1Be1F1, ADP1Be1F2, or ADP1Be1F3. Fluoroaluminates or fluoroberyllates are isomorphous to Pi, and the inhibitory nucleotide-fluorometal complexes mimicked transient intermediates of nucleotides that appeared in the course of ATP hydrolysis. On the other hand, each mole of fully inhibited F1, retained 2 mol of inhibitory complexes. The same stoichiometry was observed when ADP was replaced by GDP, a nucleotide which, unlike ADP, binds only to the catalytic sites of F1. These results are discussed in terms of a stochastic model in which the three cooperative catalytic sites of F1 function in interactive pairs.

Fluoride, beryllium and ADP combine as a ternary complex in aqueous solution as revealed by a multinuclear NMR study.

Different kinds of nucleotide binding enzymes are sensitive to fluoroberyllate complexes (BeFx) and fluoroaluminate complexes (AlFy). It has been hypothesized that the effects of these fluorometals are related to the generation at a nucleotide binding site of a pseudo nucleoside triphosphate, consisting of a fluorometal moiety bound to the beta phosphate group of a molecule of nucleoside diphosphate (Bigay et al. 1985; Lunardi et al. 1985). In order to establish whether ternary complexes comprising ADP, beryllium and fluoride can exist in slightly alkaline solution in the absence of enzyme, we have carried out a multinuclear (31P, 9Be and 19F) NMR study. In preliminary experiments, pyrophosphate (PPi) was substituted for ADP and taken as a simpler analog of nucleoside diphosphate. In the absence of fluoride, three types of PPi-Be complexes were generated: two of these were bidentate chelates with either one or two pyrophosphate molecules bound per beryllium; the third one was a monodentate complex. It is probable that the same types of combination exist between the polyphosphate chain of ADP and Be. In the presence of fluoride, both ADP and PPi combined with beryllium to form ternary complexes. These complexes consisted of monofluoroberyllate(-BeF) or difluoroberyllate(-BeF2) bound to the two phosphates of one molecule of ADP or PPi as a bidentate chelate. We failed to observe the formation of complexes between ADP and trifluoroberyllate(-BeF3). The relevance of this study to the biological effects of fluoride and beryllium on various enzymic reactions is discussed.

Synthesis and properties of azidonitrophenyl pyrophosphate, a photoaffinity probe of the nucleotide binding sites of mitochondrial F1-ATPase.

4-Azido-2-nitrophenyl pyrophosphate (azido-PPi) labeled with 32P in the alpha position was prepared and used to photolabel beef heart mitochondrial F1. Azido-PPi was hydrolyzed by yeast inorganic pyrophosphatase, but not by mitochondrial F1-ATPase. Incubation of F1 with [alpha-32P]azido-PPi in the dark under conditions of saturation resulted in the binding of the photoprobe to three sites, two of which exhibited a high affinity (Kd = 2 microM), the third one having a lower affinity (Kd = 300 microM). Mg2+ was required for binding. As with PPi [Issartel et al. (1987) J. Biol. Chem. 262, 13538-13544], the binding of 3 mol of azido-PPi/mol of F1 resulted in the release of one tightly bound nucleotide. ADP, AMP-PNP, and PPi competed with azido-PPi for binding to F1, but Pi and the phosphate analogue azidonitrophenyl phosphate did not. The binding of [32P]Pi to F1 was enhanced at low concentrations of azido-PPi, as it was in the presence of low concentrations of PPi. Sulfite, which is thought to bind to an anion-binding site on F1, inhibited competitively the binding of both ADP and azido-PPi, suggesting that the postulated anion-binding site of F1 is related to the exchangeable nucleotide-binding sites. Upon photoirradiation of F1 in the presence of [alpha-32P]azido-PPi, the photoprobe became covalently bound with concomitant inactivation of F1. The plots relating the inactivation of F1 to the covalent binding of the probe were rectilinear up to 50% inactivation.(ABSTRACT TRUNCATED AT 250 WORDS)

Photolabeling of the phosphate binding site of mitochondrial F1-ATPase by [32P]azidonitrophenyl phosphate. Identification of the photolabeled amino acid residues.

[32P]Azidonitrophenyl phosphate [( 32P]ANPP) is a photoactivatable analogue of Pi. It competes efficiently with Pi for binding to the F1 sector of beef heart mitochondrial ATPase and photolabels the Pi binding site located in the beta subunit of F1 [Lauquin, G. J. M., Pougeois, R., & Vignais, P. V. (1980) Biochemistry 19, 4620-4626]. By cleavage of the photolabeled beta subunit of F1 with cyanogen bromide, trypsin, and chymotrypsin, bound [32P]ANPP was localized in a fragment spanning Thr 299-Phe 326. By Edman degradation of the radiolabeled tryptic peptide spanning Ile 296-Arg 337, [32P]ANPP was found to be attached covalently by its photoreactive group to Ile 304, Gln 308, and Tyr 311. These results are discussed in terms of a model in which the phosphate group of [32P]ANPP interacts with a glycine-rich sequence of the beta subunit, spanning Gly 156-Lys 162, which is spatially close to the photolabeled Ile 304-Tyr 311 segment of the same subunit.