The azeuz07-+arg Mutation in FIFo-ATP Synthase from Escherichia coli A MODEL FOR HUMAN MITOCHONDRIAL DISEASE*

The mitochondrial ATPase 6 gene encodes a subunit of FIFO adenosine triphosphate (ATP) synthase. A mutation in the ATPase 6 gene has been genetically linked to two maternally inherited genetic diseases: neurological muscle weakness, ataxia, and retinitis pigmentosa (NARP) and certain cases of subacute necrotizing encephalopathy (SNE). Although the sever- ity of both NARP and SNE disease were correlated with the quantity of the ATPase 61eu156+arg mutation in each patient, the mutation could not be shown to alter FIFo-ATP synthase activity. To investigate the biochemical effects of the ATPase 61eu156+arg muta- tion on FIF,-ATP synthase, the ~ l ~ ~ 2 0 7 + ~ ~ ~ mutation was constructed in the FIFO-ATP synthase from Escherichia coli to serve as a model for the disease mu- tation. Characterization of the model bacterial enzyme revealed that the mutation abolishes detectable ATP synthesis via oxidative phosphorylation. The ~ 1 ~ ~ 2 0 7 + ~ ~ ~ mutation results in a structural perturba- tion blocking proton translocation through FIFo-ATP synthase. The results suggest that a structural defect in human FIFo-ATP synthase is the biochemical basis for NARP and SNE. Oxidative aerobic triphosphate The electron transport


Printed in U.S.A.
complicated due to heteroplasmy arising from mitotic segregation of the mtDNA defect in somatic cells ( 5 ) .
Holt et al. (6) reported that a mutation in the mtDNA ATPase 6 gene encoding a subunit of FIFo-ATP synthase was linked to maternally inherited neurogenic muscle weakness, ataxia, and retinitis pigmentosa (NARP). A restriction fragment length polymorphism in mtDNA from the NARP family allowed a direct correlation between the severity of the disease and the percentage of mutant mtDNA genomes present in each patient. A single mutation was present at nucleotide 8993 in the mtDNA and resulted in a change in the ATPase 6 gene converting leucine 156 to arginine (ATPase 61eu156-rg). The ATPase 61eu156-wrg mutation has also been observed in patients from two families diagnosed with subacute necrotizing encephalopathy (SNE) or Leigh's disease (7,8). Like the NARP patients, the severity of disease in SNE patients was correlated to the percentage of mtDNA carrying the ATPase 61eu156+arg mutation. Neurological atrophy was the prevalent symptom in affected patients, suggesting that cellular energy thresholds were not maintained in neural tissues. The genetic evidence was consistent with the suggestion that the ATPase 61,u,56+arg mutation impaired FIFo-ATP synthase reducing oxidative phosphorylation activity in NARP and SNE patients (6)(7)(8). However, no reduction in FIFO-ATP synthase ATP hydrolysis activity was observed (7), and the ATPase 61eu156-rg mutation was not shown to alter FIFO-ATP synthase and oxidative phosphorylation.
FIFo-ATP synthases share a similar molecular architecture, which is conserved in virtually all organisms (9-12). The a subunit of the Escherichia coli FIFo-ATP synthase is homologous to the human ATPase 6 subunit, and a striking primary sequence homology exists between the "ATPase 6-like" subunits of FIFo-ATP synthases from mitochondria, chloroplasts and procaryotes (Fig. 1). Extensive site-directed mutagenesis studies of the a subunit in the E. coli FIFo-ATP synthase have demonstrated that the a subunit is essential for proton translocation (13-25). The location of the disease-linked ATPase 61eu156jarg mutation at a site displaying strong sequence conservation suggested that the biochemical basis for NARP and SNE might be a defect in proton translocation through the human FIFo-ATP synthase.
Leu-207 in the a subunit of E. coli FIFo-ATP synthase occupies the position comparable to Leu-156 in the human ATPase 6 subunit (Fig. 1). To determine if the ATPase 61eu156"targ mutation caused a defect in FIFo-ATP synthase accounting for the loss of oxidative phosphorylation in NARP and SNE patients, we have constructed the ~~l,,~207-,.~ mutation in the E. coli enzyme. Biochemical characterization of the recombinant FIFo-ATP synthase demonstrated that the aleu207jarg mutation was sufficient for a loss of oxidative phosphorylation. The present work clearly establishes the direct applicability for modeling human oxidative phosphorylation disease mutations in a bacterial system. coli strains RH305 and 1100ABC and control plasmids pPH12 ( a ' ) and pPHll ia-) have been described previously (25,26). Luria broth containing 0.2% (w/v) glucose (LBG) was used as the primary medium. Minimal medium employed for growth studies consisted of minimal A salts supplemented with succinate (0.2% w/v) (27). The concentration of chloramphenicol was 20 pgiml in liquid medium and 30 pg/ml in solid medium. Liquid cultures were aerated by continuous mixing on a n orbital shaker or in a roller drum, and incubations were performed at 37 "C. structed by cassette site-directed mutagenesis of plasmid pBDC26 (20).
Recombinant DNA Techniques-The ai,,r,YI,i ,~,~~ mutation was con-Plasmid DNA was prepared by a modification of the rapid screen DNA sequence of the recombinant plasmid pUNCB4.70. Standard con-method (28), and the Sequenase protocol (29) was used to determine the ditions were used for restriction endonuclease digestions, ligations, transformations, and agarose gel electrophoresis (30).
Cell Fractionation and Assays-E. coli were inoculated into LBG (500 ml) and grown to a cell density of approximately 150 Nett units. Preparation of subcellular fractions was as described previously (31). Protein concentrations were measured using the modified Lowry procedure of Markwell et al. (32). Membrane energization (750 pg of proteid3 ml of buffer; buffer: 50 mhf MOPS and 10 mbf MgCI,, pH 7.3) was assayed by fluorescence quenching of 9-amino-6-chloro-2-methoxyacridine (ACMA) (33). ATP hydrolysis was measured by a coupled assay to regenerate ATP in the system (25).

RESULTS
Effect of the Leu-207 + Arg Mutation on Oxidative Phosphorylation-Oxidative phosphorylation is required for utilization of tricarboxylic acid cycle intermediates as carbon sources for aerobic growth, so growth of mutant strains on succinate-based medium is a sensitive test for FIFo-ATP synthase function in uiuo. The recombinant plasmid pUNCB4.70 (aI<,u20i-,nrg) and the control plasmids pPH12 ( a + ) a n d p P H l l ( a -) were studied by complementation of the a subunit-defective E. coli strain ( a ' ) complemented strain RH305 for growth on succinate medium (Table I). In contrast, cells carrying either plasmid pPHll ( a -) or plasmid pUNCB4.70 (ab~u20i,arg) failed to grow, indicating that the a subunit Leu-207 --3 Arg mutation was sufficient for loss of biologically significant oxidative phosphorylation activity ( Table I).
Loss of FIFo-ATP Synthase Function-The activity of FIFO -RH305(a,,,tzsu,,ln. , 1 r t r~4 0 . + f r p . ,rp241"te,rd) (25). Plasmid pPH12 ATP synthase was studied in vitro in order to establish the biochemical basis for the loss of oxidative phosphorylation attributable to the Leu-207 + Arg mutation. Inverted membrane vesicles were prepared from the strains carrying both mutant and control plasmids, and the capacity of the mutant FIFo-ATP synthase to translocate protons was measured by studying ATP-dependent vesicle acidification. The acidification of the membrane vesicles was followed by the quenching of the fluorescent dye 9-amino-6-chloro-2-methoxyacridine (ACMA) (Fig.  2). The membranes derived from cells with plasmid pUNCB4 .70 (alc,uP07+rrrg) exhibited fluorescent quenching comparable to that from the negative control membranes. It was clear that the Leu-207 + Arg mutation resulted in a profound inhibition of proton translocation through FIFo-ATP synthase.
In support of this conclusion, the inability to translocate protons was further shown to result from a defect in the Fo portion of the enzyme complex by studying passive proton conductance. The F1 portion of ATP synthase was displaced from the membrane-associated Fo portion by a mild, low ionic strength wash procedure. Functional Fo remaining in the FIdepleted membrane vesicles derived from cells harboring plasmid pPH12 (a' ) acted a s a passive protonophore collapsing a n imposed proton gradient. Experiments with F1-depleted inverted vesicles from cells with plasmid pUNCB4.70 (a(euz07 displayed no apparent passive proton conductance (data not shown). Membrane vesicles and assays were performed as described previously (25).
Arrows mark the addition of ATP and nigericin. Traces: a + , strain RH305 carrying pPH12 (a-1; a-, strain RH305 carrying p P H l l la-); leu207+arg, strain RH305 carrying pUNCB4.70 Bacterial Model for Mitochondrial Disease cles when Fo was properly assembled. The membranes from cells carrying plasmid pUNCB4.70 (aleu207+arg) showed levels of membrane-associated ATPase activity equivalent to the negative control pPHll (a-1, indicating that the mutation affects either assembly or stability of Fo (Table I). DCCD inhibits FIFo-ATP synthase by covalently modifying the proteolipid subunit ( c subunit in E. coli) blocking proton translocation and inhibiting coupled ATP hydrolysis activity. The residual membraneassociated F1 activity from cells carrying plasmid pUNCB4.70 (aleu207+arg) was uncoupled since DCCD-sensitive ATPase activity was far below that of the wild type control (Table I). In summary, the Leu-207 + Arg mutation apparently caused a perturbation in the structure of Fo that disrupts proton translocation and its coupling to catalytic function in FIFo-ATP synthase. DISCUSSION The results indicate that the ~l~~2 0 7 *~~~ mutation is SUEcient to abolish detectable FIFO-ATP synthase function in E.
coli. The biochemical basis for the defect is a failure in Fomediated proton translocation resulting from a structural perturbation in the enzyme complex. The structural and functional similarities between E. coli and mitochondrial FIFo-ATP synthases suggests that both share a common molecular mechanism and that observations in the bacterial system are directly applicable to mutations occurring in the human enzyme. Therefore, the phenotype of the aleu207-targ mutation observed here in E. coli implies that the human ATPase 61eu156+arg mutation linked to NARP and SNE also results in a similar defect in FIFO-ATP synthase leading to a failure of oxidative phosphorylation.
Studies of the E. coli FIFo-ATP synthase a subunit indicated that many positions were sensitive to missense mutations (13-25). Mutations replacing the conserved amino acids in the a subunit with basic amino acids, such as occurs with the human ATPase 61eu156"larg mutation, frequently resulted in substantial losses of enzyme function (13, 20, 25). Similar mutations undoubtedly occur in the human mitochondrial ATPase 6 gene at positions other than Leu-156, implying that a large number of point mutations in the conserved region of ATPase 6 should be sufficient for significant losses in capacity for synthesis ofATP. Recently, mutations in mtDNA have also been linked to degenerative diseases associated with aging (2). Given that the mutation rate is 10-20-fold higher in mtDNA as compared t o the nuclear genome (31, missense mutations in the ATPase 6 gene are likely to be among the most common mutations affecting human FIFo-ATP synthase. Point mutations affecting the ATPase 6 gene are likely t o remain undiagnosed in the clinical setting, making it difficult to assess the contributions of these defects to degenerative disease. NARP patients do not display the gross mitochondrial abnormalities or the lactate imbalances characteristic of mitochondrial genetic diseases associated with electron transport deficiencies (6). Other ATPase 6 gene mutations, whether inherited or arising through somatic mutation, will probably exhibit a similar pathology. Additionally, most point mutations in the ATPase 6 gene cannot be expected to change a restriction endonuclease recognition sequence, so restriction fragment length polymorphism analysis will not suffice as a definitive diagnostic approach. In view of the large number of a subunit mutations known to impair FIFo-ATP synthase function in bacteria, it seems appropriate to consider a systematic survey of tissue samples from patients with degenerative disorders for missense mutations in the ATPase 6 gene.