Search

Doppel protein has been discovered in prnp knock-out mouse lines, with overproduction of this protein in the brain causing ataxia and neurodegeneration. We investigated whether Doppel expression (i) affected or was affected by the course of prion propagation in neuroblastoma cells, or (ii) modulated Creutzfeldt-Jakob disease pathogenesis. No change in Doppel production was detected in N2a cells, before or after infection. Transient murine Doppel gene expression had no effect on N2a viability or PrP(Sc) production. A sensitive immunometric assay revealed low levels of Doppel in human brain, reflecting weak transcription of the corresponding gene. No difference in brain Doppel levels was observed between Creutzfeldt-Jakob disease patients and controls, adding further evidence that Doppel is unlikely to be involved in prion disease pathogenesis.

Prion diseases are probably caused by an abnormal form of a cellular glycoprotein, the prion protein. Recent evidence suggests that the prion strain causing BSE has been transmitted to humans, thereby provoking a variant form of Creutzfeldt-Jacob disease. In this work, we analyzed the behavior of normal and malformed isoforms of the bovine PrP in transfected mammalian cell lines. Biochemical and immunocytochemical assays were complimented with imaging of live cells expressing fusion constructs between PrP and GFP. Bovine homologues of human E200K and D178N (129M) mutations were used as models of pathogenic isoforms. We show that the GFP does not impair the metabolism of native and mutant bPrPs and is thus a valid marker of PrP cellular distribution. We also show that each amino acid replacement provokes alterations in the cell sorting and processing of bPrP. These are different from those ascribed to both murine mutant homologues. However, human and bovine PrPs carrying the D178N genotype had similar cellular behavior., (Copyright 2001 Academic Press.)

On the basis of far-Western blot and plasmon resonance (BIAcore) experiments, we show here that recombinant bovine prion protein (bPrP) (25-242) strongly interacts with the catalytic alpha/alpha' subunits of protein kinase CK2 (also termed 'casein kinase 2'). This association leads to increased phosphotransferase activity of CK2alpha, tested on calmodulin or specific peptides as substrate. We also show that bPrP counteracts the inhibition of calmodulin phosphorylation promoted by the regulatory beta subunits of CK2. A truncated form of bPrP encompassing the C-terminal domain (residues 105-242) interacts with CK2 but does not affect its catalytic activity. The opposite is found with the N-terminal fragment of bPrP (residues 25-116), although the stimulation of catalysis is less efficient than with full-size bPrP. These results disclose the potential of the PrP to modulate the activity of CK2, a pleiotropic protein kinase that is particularly abundant in the brain.

Ten protein kinases have been assayed for their ability to phosphorylate in vitro the recombinant bovine PrP (25-242) (rbPrP). Substantial phosphorylation was observed with PKC, CK2, and two tyrosine kinases, Lyn and c-Fgr. With regard to CK2, phosphorylation occurs at Ser 154 with a stoichiometry of about 0.1 mol phosphate/mol rbPrP, which is doubled by mild heat treatment of rbPrP. Heat also reduces the overall protein ellipticity, suggesting that reversibly unfolded conformers are more susceptible to phosphorylation. Our data disclose the possibility that phosphorylation might modulate PrP biological activity., (Copyright 2000 Academic Press.)

According to the 'protein only' hypothesis, modification of the 3-dimensional fold of the constituent cellular protein, PrP(C), into the disease-associated isoform, PrP(Sc), is the cause of neurodegenerative diseases in animals and humans. Here we describe the high-level synthesis in Escherichia coli, and purification in the monomeric form, of a histidine-tagged full-length mature PrP (25-249) of bovine brain, termed His-PrP. Based on biochemical and spectroscopic data, His-PrP displays characteristics expected for the PrP(C) isoform. The reported expression system should allow the production of quantities of bovine PrP(C) sufficient to permit 3-dimensional structure determinations.

The inner membrane of yeast and mammalian mitochondria has been studied in situ with a patch clamp electrode. Anion channels were found in both cases, although their behavior and regulation are different. In mammalian mitochondria, the principal channel is of around 100 pS conductance and opens mainly under depolarized membrane potentials. As no physiological compound able to alter its peculiar voltage dependence has yet been found, it is proposed that this channel may serve as a safeguard mechanism for recharging the mitochondrial membrane potential. Two other anion channels, each with a distinct conductance (one of approx. 45 pS, the second of at least a tenfold higher value) and kinetics are harbored in the yeast inner membrane. Matrix ATP was found to interact with both, but with a different mechanism. It is proposed that the 45 pS channel may be involved in the homeostatic mechanism of mitochondrial volume.

The inner membrane of mitochondria from various strains of Saccharomyces cerevisiae has been analyzed with the patch clamp technique for comparison with the better known homologous membrane in mammals (Sorgato, M. C., and Moran, O. (1993) CRC Crit. Rev. Biochem. Mol. Biol. 18, 127-171). Differently than in mammals, the yeast inner membrane was found to harbor essentially two channels with similar anionic selectivity but otherwise different functional behavior. One had a conductance of around 45 picosiemens (in symmetrical 150 mM KCl) and an activity only marginally sensitive to voltage. The other channel was prominent for the higher outwardly rectifying current and for the dependence upon voltage of the open probability that induced rapid closure at physiological (negative) membrane potentials. Particularly interesting was the effect of ATP (Mg2+ free) added on the matrix side of the membrane. In the case of the lower conducting channel, the nucleotide caused an immediate block of activity (IC50, 0.240 mM), whereas it locked the larger conductance in the open state at both positive and negative potentials. In proteoliposomes containing both mitochondrial membranes, the small conductance was clearly evident, whereas a larger channel, cationic and without the voltage dependence typical of that in the native inner membrane, was found.

Detergent-free rat brain outer mitochondrial membranes were incorporated in planar lipid bilayers in the presence of an osmotic gradient, and studied at high (1 M KCl) and low (150 mM KCl) ionic strength solutions. By comparison, the main outer mitochondrial membrane protein, VDAC, extracted from rat liver with Triton X-100, was also studied in 150 mM KCl. In 1 M KCl, brain outer membranes gave rise to electrical patterns which resembled very closely those widely described for detergent-extracted VDAC, with transitions to several subconducting states upon increase of the potential difference, and sensitivity to polyanion. The potential dependence of the conductance of the outer membrane, however, was steeper and the extent of closure higher than that observed previously for rat brain VDAC. In 150 mM KCl, bilayers containing only one channel had a conductance of 700 +/- 23 pS for rat brain outer membranes, and 890 +/- 29 pS for rat liver VDAC. Use of a fast time resolution setup allowed demonstration of open-close transitions in the millisecond range, which were independent of the salt concentration and of the protein origin. We also found that a potential difference higher than approx. +/- 60 mV induced an almost irreversible decrease of the single channel conductance to few percentages of the full open state and a change in the ionic selectivity. These results show that the behavior of the outer mitochondrial membrane in planar bilayers is close to that detected with the patch clamp (Moran et al., 1992, Eur. Biophys. J. 20:311-319). The neurotoxicological action of aluminum was studied in single outer membrane channels from rat brain mitochondria. We found that microM concentrations of Al Cl3 and aluminum lactate decreased the conductance by about 50%, when the applied potential difference was positive relative to the side of the metal addition.

Rapid diffusion of hydrophilic molecules across the outer membrane of mitochondria has been related to the presence of a protein of 29 to 37 kDa, called voltage-dependent anion channel (VDAC), able to generate large aqueous pores when integrated in planar lipid bilayers. Functional properties of VDAC from different origins appear highly conserved in artificial membranes: at low transmembrane potentials, the channel is in a highly conducting state, but a raise of the potential (both positive and negative) reduces drastically the current and changes the ionic selectivity from slightly anionic to cationic. It has thus been suggested that VDAC is not a mere molecular sieve but that it may control mitochondrial physiology by restricting the access of metabolites of different valence in response to voltage and/or by interacting with a soluble protein of the intermembrane space. The latest application of the patch clamp and tip-dip techniques, however, has indicated both a different electric behavior of the outer membrane and that other proteins may play a role in the permeation of molecules. Biochemical studies, use of site-directed mutants, and electron microscopy of two-dimensional crystal arrays of VDAC have contributed to propose a monomeric beta barrel as the structural model of the channel. An important insight into the physiology of the inner membrane of mammalian mitochondria has come from the direct observation of the membrane with the patch clamp. A slightly anionic, voltage-dependent conductance of 107 pS and one of 9.7 pS, K(+)-selective and ATP-sensitive, are the best characterized at the single channel level. Under certain conditions, however, the inner membrane can also show unselective nS peak transitions, possibly arising from a cooperative assembly of multiple substrates.

The outer and inner membranes of mitochondria have recently been studied with the patch clamp technique. What has emerged is still an ill-defined picture for either membrane, primarily for the wide range of conductances found. Interestingly, however, a few conductances (in the range of 10-80 pS) seem to be ubiquitously distributed. Parallel studies in situ and in reconstituted systems have allowed the assignment to distinct membrane locations of some conductances, whose physiological role is, however, not yet elucidated.

The ionic permeability of the outer mitochondrial membrane (OMM) was studied with the patch clamp technique. Electrical recording of intact mitochondria (hence of the outer membrane (OM], derived from mouse liver, showed the presence of currents corresponding to low conductances (less than 50 pS), as well as of four distinct conductances of 99 pS, 152 pS, 220 pS and 307 pS (in 150 mM KCl). The latter were voltage gated, being open preferentially at positive (pipette) potentials. Very similar currents were found by patch clamping liposomes containing the isolated OM derived from rat brain mitochondria. Here a conductance of approximately 530 pS, resembling in its electrical characteristics a conductance already attributed to mitochondrial contact sites (Moran et al. 1990), was also detected. Immunoblot assays of mitochondria and of the isolated OM with antibodies against the outer membrane voltage-dependent anion channel (VDAC) (Colombini 1979), showed the presence of the anion channel in each case. However, the typical electrical behaviour displayed by such a channel in planar bilayers could not be detected under our experimental conditions. From this study, the permeability of the OMM appears different from what has been reported hitherto, yet is more in line with that multifarious and dynamic structure which apparently should belong to it, at least within the framework of mitochondrial biogenesis (Pfanner and Neupert 1990).

In this report data are presented which firmly establish that by treating isolated F0 with the thiol reagent diamide, two 25 kDa F0 subunits react to form a dimer of 45 kDa apparent molecular mass. This dimerising effect is correlated to the impairment of the binding of F1 to F0, both at microM and mM diamide concentrations. Under the latter condition, modification of other F0 subunits also occurs. Passive proton conductance through F0, as well as its sensitivity to N,N'-dicyclohexylcarbodiimide, are affected at low diamide concentration. Thus perturbation of the cysteine residue of the 25 kDa F0 subunit is sufficient for altering the ATP synthase proton channel.

From morphological and biochemical studies it has been recognized that the regions where the outer and inner membranes of mitochondria come in close contact (contact sites) can be the route mechanism through which mitochondria interact directly with the cytoplasm. We have studied these regions electrophysiologically with the patch clamp technique, with the aim of understanding if this direct interaction is mediated by high conductance ion channels similar to the channel already detected in the inner membrane of mitochondria (Sorgato M. C., Keller, B. U., and Stühmer, W. (1987) Nature 330, 498-500). Contact sites isolated from rat brain mitochondria were thus incorporated into liposomes subsequently enlarged sufficiently to be patch clamped. This study shows that these particular fractions contain ion channels with conductances ranging from approximately 5 picosiemens to 1 nanosiemens (in symmetrical 150 mM KCl). Most of these channels are not voltage-dependent and can be open at physiological potentials sustained by respiring mitochondria. The lack of voltage sensitivity seems not to be the outcome of methodological artifacts, as voltage-gated channels are detected in giant liposomes containing either the outer mitochondrial membrane or a partially purified fraction of the inner mitochondrial membrane. These data therefore indicate that channels present in mitochondrial contact sites have properties which render them amenable to perform several of the functions hypothesized for these regions, particularly that of translocating macromolecules from the cytoplasm to the matrix of mitochondria.

The prime function of mitochondria is to provide the cell with adenosine triphosphate (ATP). ATP synthesis is driven by the protonmotive force (delta p), which is generated and maintained across the inner mitochondrial membrane (IMM) by the activity of the respiratory chain. It is widely believed that the IMM is unlikely to contain ion channels like those present in the plasma membrane, because the high rates of ion transport characteristic of open channels would be expected to dissipate the delta p. Although the small size of the organelle has prevented the use of classical electrophysiological methods, the recent introduction of the patch-clamp technique, which allows currents to be recorded from very small cells, has enabled us to test this hypothesis. By patch-clamping the IMM, we have identified a slightly anion-selective channel, which is voltage-dependent and has a mean conductance of 107 pS in the presence of symmetrical 150 mM KCl.

A phosphorylation potential deltaGp, where deltaGp = deltaGo' + RT2.303 log ([ATP]/([ADP][Pi])), of approx. 44.3 kJ.mol-1 (10.6 kcal.mol-1) was generated by submitochondrial particles that were oxidizing either NADH or succinate. Addition of adenylyl imidodiphosphate, which should suppress adenosine triphosphatase activity of any uncoupled particles, did not raise the phosphorylation potential. Raising the Pi concentration slightly increased the magnitude of the value for [ATP]/[ADP], but this did not fully compensate for the increased Pi concentration, so that the phosphorylation potential decreased slightly as the Pi concentration was raised. The phosphorylation potential developed by submitochondrial particles is lower than that generated by phosphorylating membrane vesicles from some bacteria, and is also less than that developed externally by mitochondria, but is strikingly close to the phosphorylation potential that is generated internally by mitochondria.

The proton-motive forces generated in submitochondrial particles by both hydrolysis of ATP and oxidation of succinate have been measured by flow dialysis and compared with the ambient phosphorylation potentials. It is concluded that three H+ are translocated for each ATP molecule hydrolysed or synthesised. By utilising rat liver mitochondria respiring with beta-hydroxybutyrate as a new system for regeneration of ATP from ADP and Pi, phosphorylation potentials were clamped at a range of values by using mixtures of particles and mitochondria in various ratios. As the rate of ATP hydrolysis by the particles was lowered, the proton-motive force decreased only slightly except at the very lowest rates, these results paralleling earlier studies on the relation between rate of respiration-driven proton translocation and proton-motive force.

By use of the patch-clamp technique, the inner membrane of mouse liver and heart mitochondria is shown to contain a highly conductive (around 100 pS in symmetrical 150 mM KCl) and voltage-dependent ion channel. This channel closely resembles that previously found in cuprizone-treated mouse liver inner mitochondrial membrane. The paper discusses the electrical properties of the channel and its possible physiological function. The reconstitution in giant liposomes of a partially purified ox heart inner membrane fraction containing the channel and the use of various inhibitors are also presented.

Respiring submitochondrial particles from which the F1 sector of ATP-synthase was displaced generated a membrane potential in the range of 115-140 mV. Addition of oligomycin raised the membrane potential by approximately 40 mV. The lower membrane potential in particles with F1 displaced is attributed to partial dissipation of the proton electrochemical gradient as a consequence of proton flow through the open proton channels provided by the F0 sectors of the ATP-synthase. The characteristics of proton flow through the open F0 channels were studied by varying the rate of electron transport-driven proton translocation which permitted the establishment of a range of steady-state membrane potentials. Open F0 channels appeared to have a gated response to the membrane potential such that they were inoperative when the potential fell below approximately 110 mV. The membrane potential was measured as a function of respiratory rate in intact Mg-ATP submitochondrial particles that had been treated with low concentrations of the protonophore carbonylcyanide-p-trifluoromethoxyphenylhydrazone. In general a linear dependence of membrane potential upon respiratory rate was observed except at the lowest concentrations of protonophore and highest respiratory rates, presumably because the effect of the protonophore was then offset by an increased rate of proton translocation driven by the respiratory chain. The effect of increasing concentrations of carbonylcyanide-p-trifluoromethoxyphenylhydrazone on the membrane potential of respiring submitochondrial particles was studied. It was found that equal amounts of the protonophore lowered the membrane potential to a lesser extent at lower values of the membrane potential. Treatment of Mg-ATP submitochondrial particles with oligomycin slightly increased (by approximately 10 mV) the size of the respiration-dependent membrane potential, but did not alter the profile of membrane potential as a function of succinate oxidation rate. The latter was controlled by titration with malonate. This result indicates that the F0 sector of the ATP-synthase does not significantly contribute to leak pathways in intact submitochondrial particles.

1. A comparison was made of two methods for estimating the membrane potential in chromatophores from Rhodopseudomonas sphaeroides Ga. Illuminated chromatophores generated a potential that is apparently much larger when estimated on the basis of the red-band shift of carotenoids rather than from the extent of uptake of the permeant SCN- ion. 2. In contrast, when the chromatophores were oxidizing NADH or succinate the uptake of SCN- indicated a larger membrane potential than was estimated from the carotenoid band shift. 3. The extent of SCN- uptake and the carotenoid-band shift respond differently to changes in the ionic composition of the reaction medium. 4. The effects of antimycin on the carotenoid band shift and SCN- uptake are reported. 5. It is concluded that the carotenoid band shift and the uptake of SCN- are responding to different aspects of the energized state.

(1) The effects of membrane potential (delta psi) and nucleotides on the interaction between the F1-ATP synthase and its natural inhibitor protein (IF1) are studied in ox-heart submitochondrial vesicles. (2) Membrane potential causes displacement of IF1 from submitochondrial vesicles, as shown by measuring both delta psi-dependent stimulation of ATPase capacity and release of 125I-labelled IF1 from the vesicles. These effects are abolished if ATP is included in the incubation. (3) There is a linear increase in the steady-state ATPase capacity of oxidising vesicles as delta psi is increased from 100 mV to 135 mV. Increasing delta psi above 140 mV leads to no further change. (4) At a constant membrane potential, ATP suppresses the increase in ATPase capacity, with a concentration for half maximal effect of 140 microM. This value is close to the Km for ATP hydrolysis by membrane-bound F1. This suppression is related to ATP concentration rather than to delta Gp or ATP/ADP ratio. (5) The unidirectional on- and off-rates of IF1 were measured separately. The off-rate of IF1 is increased by membrane potential but unaffected by ATP. The on-rate, conversely, is increased by ATP. Thus, the suppression of the potential-dependent net release of IF1 from submitochondrial vesicles by ATP results from an increase of the IF1 on-rate above the off-rate.

ITP hydrolysis catalysed by the ATPase of submitochondrial particles from both bovine heart and rat liver is shown to be linked to the generation of a membrane potential, and therefore also to proton translocation. The magnitude of the membrane potential is similar to that observed during ATP hydrolysis at equivalent concentrations of phosphate and nucleoside tri- and diphosphates. An explanation is suggested for why in other reports ITP was found to be a poor substrate for supporting energy-linked reactions that are driven by the membrane potential.

(1) The kinetics of the release of the mitochondrial inhibitor protein (IF1) is studied in bovine heart submitochondrial vesicles supplemented with 125I-labelled IF1, using a method for rapidly 'freezing' the state of F1-IF1 interaction. It is shown that generation of a protonmotive force leads to release of IF1 from F1 into solution, following an exponential process. (2) In one set of experiments the rate of IF1 release, in IF1 supplemented vesicles generating a protonmotive force, is correlated with the induction of ATP hydrolytic capacity. It is found that, even under different metabolic states (phosphorylating and non-phosphorylating conditions), both processes follow the same time-course (half-time of around 40 s) and that there is a direct correlation between induced ATPase capacity and IF1 released. This finding rules out the possibility of a non-inhibitory binding site for IF1 on the membrane. (3) In a second set of experiments, also using IF1 supplemented vesicles, the induction of the ATP hydrolytic capacity after energisation is correlated with the induction of the ATP synthetic capacity. Initial rates of both processes are monitored using firefly luciferase, keeping the assay systems as similar as possible. It is shown that the induction of each capacity follows an exponential time-course, with a half-time of around 40 s. This is in good agreement with the half-times obtained for the induction of ATP hydrolytic capacity and the rate of IF1 release, using the quench-stop method. (4) If the induction of ATP hydrolytic and synthetic capacities is followed in untreated submitochondrial vesicles, i.e., vesicles not supplemented with IF1, the extent and time-course of the change in both hydrolytic and synthetic capacities remain correlated, but the half-time of the transient falls to around 10 s. It is suggested that the length of the transient, observed in IF1 supplemented vesicles, results from partial loss of coupling during repeated centrifugations. (5) These results demonstrate that energy-dependent release of IF1 from F1 into solution results in a concomitant increase in both ATP synthetic and hydrolytic capacities of the ATP synthase complex, and that the time-course of this process is sensitive to the degree of coupling of the vesicles.

The relationship between the rate of substrate oxidation and the protonmotive force (electrochemical proton gradient) generated by bovine heart submitochondrial particles has been examined. Unexpectedly, oxidation of succinate generated a higher protonmotive force than the oxidation of NADH, although the rate of proton translocation across the membrane was inferred to be considerably lower with succinate as substrate. The data suggest that the flow of electrons through site 1 of the respiratory chain may increase the conductance of the mitochondrial membrane for protons. Upon reduction of the rate of succinate oxidation by titration with malonate, the protonmotive force remained essentially constant until the extent of inhibition was greater than 75%. The general conclusion from this work is that a constant passive membrane conductance for protons cannot be assumed.

1. The magnitude of the protonmotive force in respiring bovine heart submitochondrial particles was estimated. The membrane-potential component was determined from the uptake of S14CN-ions, and the pH-gradient component from the uptake of [14C]methylamine. In each case a flow-dialysis technique was used to monitor uptake. 2. With NADH as substrate the membrane potential was approx. 145mV and the pH gradient was between 0 and 0.5 unit when the particles were suspended in a Pi/Tris reaction medium. The addition of the permeant NO3-ion decreased the membrane potential with a corresponding increase in the pH gradient. In a medium containing 200mM-sucrose, 50mM-KCl and Hepes as buffer, the total protonmotive force was 185mV, comprising a membrane potential of 90mV and a pH gradient of 1.6 units. Thus the protonmotive force was slightly larger in the high-osmolarity medium. 3. The phosphorylation potential (= deltaG0' + RT ln[ATP]/[ADP][Pi]) was approx. 43.1 kJ/mol (10.3kcal/mol) in all the reaction media tested. Comparison of this value with the protonmotive force indicates that more than 2 and up to 3 protons must be moved across the membrane for each molecule of ATP synthesized by a chemiosmotic mechanism. 4. Succinate generated both a protonmotive force and a phosphorylation potential that were of similar magnitude to those observed with NADH as substrate. 5. Although oxidation of NADH supports a rate of ATP synthesis that is approximately twice that observed with succinate, respiration with either of these substrates generated a very similar protonmotive force. Thus there seemed to be no strict relation between the size of the protonmotive force and the phosphorylation rate. 6. In the presence of antimycin and/or 2-n-heptyl-4-hydroxyquinoline N-oxide, ascorbate oxidation with either NNN'N'-tetramethyl-p-phenylenediamine or 2,3,5,6-tetramethyl-p-phenylenediamine as electron mediator generated a membrane potential of approx. 90mV, but no pH gradient was detected, even in the presence of NO3-. These data are discussed with reference to the proposal that cytochrome oxidase contains a proton pump.

The problem of whether the rate of ATP synthesis is proportional to the magnitude of the protonmotive force has been studied in submitochondrial particles. It was found that the rate of ATP synthesis can decrease at constant protonmotive force and is more closely related to the rate of substrate oxidation.

The magnitude of the protonmotive force, and its division between pH gradient and membrane potential components has been further characterised in submitochondrial particles. In a reaction medium containing sucrose for osmotic support and 4-(2-hydroxyethyl)-1-piperazineethanesulfonate (Hepes) as buffer, with succinate as substrate, the total protonmotive force reached a maximum value of 245 mV. The presence of Cl- enhanced the pH gradient with a partial but not fully compensating decrease in the membrane potential. When submitochondrial particles were suspended in a medium of low osmolarity consisting of phosphoric acid neutralised with Tris, again with succinate as substrate, the protonmotive force was lower and did not exceed 185 mV, and the pH gradient component was equivalent to 25 mV or less. The final phosphorylation potential, delta Gp, (formula: see text); maintained by the particles was higher in the phosphate/Tris medium (46--47.7 kJ mol-1) than in the sucrose/Hepes/KCl medium (43.7 kJ mol-1). Thus, comparison of the phosphorylation potential with the protonmotive force would suggest that the mechanistic stoichiometry H+/ATP (H+ translocated per molecule of ATP synthesied) for the ATPase enzyme is 3 in the former medium and 2 in the latter, which might be taken to indicate two different types of mechanism required for ATP synthesis. However it is questioned whether a comparison of the protonmotive force with delta Gp in terms of equilibrium thermodynamics ought not to be complemented by analysis in terms of linear non-equilibrium thermodynamics. The latter treatment shows that it is possible to estimate only a value for the product of a phenomenological stoichiometry and the degree of coupling, which can be variable, but not the mechanistic stoichiometry. This treatment can also rationalise the observation of the higher delta Gp in reaction conditions where the lower values for delta p are estimated. Irrespective of possible explanations, the data show how an unprejudiced choice of reaction conditions can lead to different conclusions about the relationship between the phosphorylation potential and the protonmotive force.

In order to assess the role of thiol groups in the Fo part of the ATP synthase in the coupling mechanism of ATP synthase, we have treated isolated Fo, extracted from beef heart Complex V with urea, with thiol reagents, primarily with diazenedicarboxylic acid bis-(dimethylamide) (diamide) but also with Cd2+ and N-ethylmaleimide. FoF1 ATP synthase was reconstituted by adding isolated F1 and the oligomycin-sensitivity-conferring-protein (OSCP) to Fo. The efficiency of reconstitution was assessed by determining the sensitivity to oligomycin of the ATP hydrolytic activity of the reconstituted enzyme. Contrary to Cd2+, incubation of diamide with Fo, before the addition of F1 and OSCP, induced a severe loss of oligomycin sensitivity, due to an inhibited binding of F1 to Fo. This effect was reversed by dithiothreitol. Conversely, if F1 and OSCP were added to Fo before diamide, no effect could be detected. These results show that F1 (and/or OSCP) protects Fo thiols from diamide and are substantiated by the finding that the oligomycin sensitivity of ATP hydrolysis activity of isolated Complex V was also unaltered by diamide. Gel electrophoresis of FoF1 ATP synthase, reconstituted with diamide-treated Fo, revealed that the loss of oligomycin sensitivity was directly correlated with diminution of band Fo 1 (or subunit b). Concomitantly a band appeared of approximately twice the molecular weight of subunit Fo 1. As this protein contains only 1 cysteine residue (Walker, J. E., Runswick, M. J., and Poulter, L. (1987) J. Mol. Biol. 197, 89-100), the effect of diamide is attributed to the formation of a disulfide bridge between two of these subunits. These results offer further evidence for the proposal, based on aminoacid sequence and structural analysis, that subunit Fo 1 of mammalian Fo is involved in the binding with F1 (Walker et al. (1987]. N-Ethylmaleimide affects oligomycin sensitivity to a lesser extent than diamide, suggesting that the mode of action of these reagents (and the structural changes induced in Fo) is different.