C1 metabolism inParacoccus denitrificans: Genetics ofParacoccus denitrificans

Paracoccus denitrificans is able to grow on the C₁ compounds methanol and methylamine. These compounds are oxidized to formaldehyde which is subsequently oxidized via formate to carbon dioxide. Biomass is produced by carbon dioxide fixation via the ribulose biphosphate pathway. The first oxidation reaction is catalyzed by the enzymes methanol dehydrogenase and methylamine dehydrogenase, respectively. Both enzymes contain two different subunits in an ₂₂ configuration. The genes encoding the subunits of methanol dehydrogenase (moxF andmoxI) have been isolated and sequenced. They are located in one operon together with two other genes (moxJ andmoxG) in the gene ordermoxFJGI. The function of themoxJ gene product is not yet known.MoxG codes for a cytochromec _551i , which functions as the electron acceptor of methanol dehydrogenase. Both methanol dehydrogenase and methylamine dehydrogenase contain PQQ as a cofactor. These so-called quinoproteins are able to catalyze redox reactions by one-electron steps. The reaction mechanism of this oxidation will be described. Electrons from the oxidation reaction are donated to the electron transport chain at the level of cytochromec. P. denitrificans is able to synthesize at least 10 differentc-type cytochromes. Five could be detected in the periplasm and five have been found in the cytoplasmic membrane. The membrane-bound cytochromec ₁ and cytochromec ₅₅₂ and the periplasmic-located cytochromec ₅₅₀ are present under all tested growth conditions. The cytochromesc _551i andc _553i , present in the periplasm, are only induced in cells grown on methanol, methylamine, or choline. The otherc-type cytochromes are mainly detected either under oxygen limited conditions or under anaerobic conditions with nitrate as electron acceptor or under both conditions. An overview including the induction pattern of allP. denitrificans c-type cytochromes will be given. The genes encoding cytochromec ₁, cytochromec ₅₅₀, cytochromec _551i , and cytochromec _553i have been isolated and sequenced. By using site-directed mutagenesis these genes were mutated in the genome. The mutants thus obtained were used to study electron transport during growth on C₁ compounds. This electron transport has also been studied by determining electron transfer rates inin vitro experiments. The exact pathways, however, are not yet fully understood. Electrons from methanol dehydrogenase are donated to cytochromec _551i . Further electron transport is either via cytochromec ₅₅₀ or cytochromec _553i to cytochromeaa ₃. However, direct electron transport from cytochromec _551i to the terminal oxidase might be possible as well. Electrons from methylamine dehydrogenase are donated to amicyanin and then via cytochromec ₅₅₀ to cytochromeaa ₃, but other routes are used also.P. denitrificans is studied by several groups by using a genetic approach. Several genes have already been cloned and sequenced and a lot of mutants have been isolated. The development of a host/vector system and several techniques for mutation induction that are used inP. denitrificans genetics will be described.     

The cytochromec reductase/oxidase respiratory pathway ofParacoccus denitrificans: Genetic and functional studies

Data are presented on three components of the quinol oxidation branch of theParacoccus respiratory chain: cytochromec reductase, cytochromec ₅₅₂, and thea-type terminal oxidase. Deletion mutants in thebc ₁ and theaa ₃ complex give insight into electron pathways, assembly processes, and stability of both redox complexes, and, moreover, are an important prerequisite for future site-directed mutagenesis experiments. In addition, evidence for a role of cytochromec ₅₅₂ in electron transport between complex III and IV is presented.     

The reactions of the oxidase and reductases ofParacoccus denitrificans with cytochromesc

Electron transport in theParacoccus denitrificans respiratory chain system is considerably more rapid when it includes the membrane-bound cytochromec ₅₅₂ than with either solubleParacoccus c ₅₅₀ or bovine cytochromec; a pool function for cytochromec is not necessary. Low concentrations ofParacoccus or bovine cytochromec stimulate the oxidase activity. This observation could explain the multiphasic Scatchard plots which are obtained. A negatively charged area on the back side ofParacoccus c which is not present in mitochondrialc could be a control mechanism forParacoccus reactions.Paracoccus oxidase and reductase reactions with bovinec show the same properties as mammalian systems; and this is true ofParacoccus oxidase reactions with its own soluble cytochromec if added polycation masks the negatively charged area. Evidence for different oxidase and reductase reaction sites on cytochromec include: (1) stimulation of the oxidase but not reductase by a polycation; (2) differences in the inhibition of the oxidase and reductases by monoclonal antibodies toParacoccus cytochromec; and (3) reaction of another bacterial cytochromec withParacoccus reductases but not oxidase. Rapid electron transport occurs in cytochromec-less mutants ofParacoccus, suggesting that the reactions result from collision of diffusing complexes.     

Metabolic pathways inParacoccus denitrificans and closely related bacteria in relation to the phylogeny of prokaryotes

Denitrification and methylotrophy inParacoccus denitrificans are discussed. The properties of the enzymes of denitrification: the nitrate-nitrite antiporter, nitrate reductase, nitrite reductase, nitric oxide reductase and nitrous oxide reductase are described. The genes for none of these proteins have yet been cloned and sequenced fromP. denitrificans. A number of sequences are available for enzymes fromEscherichia coli, Pseudomonas stutzeri andPseudomonas aeruginosa. It is concluded that pathway specificc-type cytochromes are involved in denitrification. At least 40 genes are involved in denitrification.In methanol oxidation at least 20 genes are involved. In this case too pathway specificc-type cytochromes are involved. The sequence homology between the quinoproteins methanol dehydrogenase, alcoholdehydrogenase and glucose dehydrogenase is discussed. This superfamily of proteins is believed to be derived from a common ancestor. ThemoxFJGI operon determines the structural components of methanol dehydrogenase and the associatedc-type cytochrome. Upstream of this operon 3 regulatory proteins were found. The mox Y protein shows the general features of a sensor protein and the moxX protein those of a regulatory protein. Thus a two component regulatory system is involved in both denitrification and methylotrophy.The phylogeny of prokaryotes based on 16S rRNA sequence is discussed. It is remarkable that the 16S rRNA ofThiosphaera pantotropha is identical to that ofP. denitrificans. Still these bacteria show a number of differences.T. pantotropha is able to denitrify under aerobic circumstances and it shows heterotrophic nitrification. Nitrification and heterotrophic nitrification are found in species belonging to the -and -subdivisions of purple non-sulfur bacteria. Thus the occurrence of heterotrophic nitrification inT. pantotropha which belongs to the -subdivision of purple non-sulfur bacteria is a remarkable property. FurthermoreT. pantotropha contains two nitrate reductases of which the periplasmic one is supposed to be involved in aerobic denitrification. The nitrite reductase is of the Cu-type and not of the cytochromecd ₁ type as inP. denitrificans. Also the cytochromeb of theQbc complex ofT. pantotropha is highly similar to its counterpart inP. denitrificans. It is hypothesized that the differences between these two organisms which both contain large megaplasmids is due to a combination of loss of genetic information and plasmid-coded properties. The distribution of a number of complex metabolic systems in eubacteria and in a number of species belonging to the -group of purple non sulphur bacteria is reviewed. Two possibilities to explain this haphazard distribution are considered: 1. Lateral gene transfer between distantly related micro organisms occurs frequently. 2. The eubacterial ancestors must have possessed already these properties. The distribution of these properties is due to sporadic loss during evolutionary divergence.With respect to the occurrence and frequency of lateral gene transfer two opposing views exist. According to molecular biologists lateral gene transfer occurs frequently and is very easy. Bacteria are supposed to form one large gene pool. On the other hand population geneticists have provided evidence that strong systems operate that establish reproductive isolation between diverged species and even between closely related cell lines.Data on amino acid sequences of nitrogenase proteins, cytochromesc, cytochrome oxidases, -subunits of ATP synthase and tryptophan biosynthetic enzymes of various micro organisms were reviewed. In all these cases phylogenetic trees could be constructed based on the amino acid sequence data. In all cases this phylogenetic tree was similar to the one based on 16S rRNA homology. Only in one case evidence for the occurrence of lateral gene transfer was obtained. Therefore it is concluded that lateral gene transfer played a minor role in the distribution of complex metabolic systems among prokaryotes. It must be stressed that this does not exclude the possibility that lateral gene transfer occurred frequently in the initial stage of bacterial evolution. It is hypothesized that the appearance of nitrogen fixation, denitrification and cytochrome oxidase formation were early events in the evolution of micro organisms. Both systems are supposed to have evolved only once. Subsequently the capacity to fix nitrogen or to denitrifymust have been lost many times, just as photosynthetic capacity is supposed to have been lost many times. During evolution many systems have been lost leading to a haphazard distribution of metabolic characters among bacteria. As an example it is suggested that organisms with a respiratory chain similar to that ofEscherichia coli arose by loss of the capacity to form the Qbc complex andc-type cytochromes. The remaining systems could be controlled much better however than in the ancestral organisms.     

Regulation of methanol dehydrogenase synthesis inParacoccus denitrificans

The region downstream from the methanol dehydrogase (MDH) structural gene has been cloned and sequenced. MDH promoter activity have been studied by using a broad-host-range promoter probe vector.     

Electron transfer between cytochromec and cytochromec peroxidase

The reaction between cytochromec (CC) and cytochromec peroxidase (CcP) is a very attractive system for investigating the fundamental mechanism of biological electron transfer. The resting ferric state of CcP is oxidized by hydrogen peroxide to compound I (CMPI) containing an oxyferryl heme and an indolyl radical cation on Trp-191. CMPI is sequentially reduced to CMPII and then to the resting state CcP by two molecules of CC. In this review we discuss the use of a new ruthenium photoreduction technique and other rapid kinetic techniques to address the following important questions: (1) What is the initial electron acceptor in CMPI? (2) What are the true rates of electron transfer from CC to the radical cation and to the oxyferryl heme? (3) What are the binding domains and pathways for electron transfer from CC to the radical cation and the oxyferryl heme? (4) What is the mechanism for the complete reaction under physiological conditions?     

Regulation of methylammonium transport inParacoccus denitrificans

Depending on the growth conditionsParacoccus denitrificans synthesizes two different carriers mediating uptake of methylamine. When used as a nitrogen source, methylamine is transported via a NH ₄ ⁺ carrier, and its transport is inhibited by NH ₄ ⁺ but not by ethylamine. When used as a carbon source, methylamine is transported by a specific alkylamine carrier, and its transport is inhibited by ethylamine but not by NH ₄ ⁺ . The NH ₄ ⁺ carrier is under nitrogen control, the alkylamine carrier under carbon control.Abbreviations MA Methylamine - FCCP p-trifluormethoxycarbonylcyanide-phenylhydrazone     

The cytochromec oxidase from the yeastCandida parapsilosis

In the yeastCandida parapsilosis, the proteins encoded by mitochondrial DNA are different in number and size from those ofSaccharomyces cerevisiae. Nevertheless, the purified cytochromec oxidase fromCandida parapsilosis shows kinetic properties similar to those ofSaccharomyces cerevisiae.     

Isolation,sequencing and mutational analysis of a gene cluster involved in nitrite reduction inParacoccus denitrificans

By using the gene encoding the C-terminal part of thecd ₁-type nitrite reductase ofPseudomonas stutzeri JM300 as a heterologous probe, the corresponding gene fromParacoccus denitrificans was isolated. This gene,nirS, codes for a mature protein of 63144 Da having high homology withcd ₁-type nitrite reductases from other bacteria. Directly downstream fromnirS, three othernir genes were found in the ordernirECF. The organization of thenir gene cluster inPa. denitrificans is different from the organization ofnir clusters in some Pseudomonads.nirE has high homology with a S-adenosyl-L-methionine:uroporphyrinogen III methyltransferase (uro'gen III methylase). This methylase is most likely involved in the hemed ₁ biosynthesis inPa. denitrificans. The third gene,nirC, codes for a small cytochromec of 9.3 kDa having high homology with cytochromec _55X ofPs. stutzeri ZoBell. The 4th gene,nirF, has no homology with other genes in the sequence databases and has no relevant motifs. Inactivation of either of these 4 genes resulted in the loss of nitrite and nitric oxide reductase activities but not of nitrous oxide reductase activity.nirS mutants lack thecd ₁-type nitrite reductase whilenirE, nirC andnirF mutants produce a small amount ofcd ₁-type nitrite reductase, inactive due to the absence of hemed ₁. Upstream from thenirS gene the start of a gene was identified which has limited homology withnosR, a putative regulatory gene involved in nitrous oxide reduction. A potential FNR box was identified between this gene andnirS.Abbreviations SDS sodium dodecyl sulfate - NBT nitroblue tetrazolium - PAGE polyacrylamide gel electrophoresis     

Activation studies by phospholipids on the purified cytochromec 4:o oxidase ofAzotobacter vinelandii

A modified procedure is described that was used to solubilize and purify the TMPD-dependent cytochromec ₄:o oxidase fromAzotobacter vinelandii. Two functional components (Fractions I and V) were obtained after DEAE-cellulose chromatography. Fraction V contained both cytochromec ₄ (3.6 nmol/mg protein) and cytochromeo (1.6 nmol/mg protein). This cytochrome oxidase complex oxidized TMPD at moderate rates. Fraction I, a clear greenish-yellow fraction, contained primarily phosphatidylethanolamine with some phosphatidylglycerol. Fraction I itself could not oxidize TMPD, but when it was preincubated with Fraction V, a 2–4-fold stimulation in TMPD oxidase activity occurred. Other authentic micellar phospholipids also readily activited TMPD oxidase activity in Fraction V. Themaximum activation effect obtained with Fraction I was in essence duplicated with purified phosphatidylethanolamine.Dedicated to the memory of David E. Green, a fine gentleman, an excellent scientist, and a true scholar. He will be missed by many of his former colleagues.     

Phylogenetic analysis of the perissodactylan family Tapiridae using mitochondrial cytochromec oxidase (COII) sequences

The four extant members of the family Tapiridae have a disjunct, relictual distribution, with three species being Neotropical (Tapirus bairdii, T. terrestris, andT. pinchaque) and one found in Southeast Asia (T. indicus). Little recent work on tapir systematics have appeared, and no molecular studies of this group have been published. A phylogenetic analysis was undertaken using sequences of the mitochondrial cytochromec oxidase subunit II gene (COII) from representatives of the four species of tapirs, as well as a representative outgroup,Equus caballus. Analyses of the COII sequences indicate a close relationship between the two South American species of tapirs,T. terrestris andT. pinchaque, and estimates of divergence dates using rates of COII evolution are compatible with migration of a single tapir lineage into South America following the emergence of the isthmus of Panama, about 3 million years bp. Various methods of analysis, including maximum parsimony, maximum likelihood, and neighbor-joining, provided poorer resolution of other tapir relationship. The COII data suggest that three distinct tapir mitochondrial lineages, a South American (represented byT. terrestris andT. pinchaque), a Central American (represented byT. bairdii), and an Asian (represented byT. indicus) diverged relatively rapidly, 20–30 million years bp. Another goal of this study was to calibrate the rate of COII evolution in a eutherian mammal group which has a good fossil record, such as perissodactyls, to estimate accurately the rate of COII evolution in a nonprimate mammalian group. The rate of COII evolution in equids and tapirs has been relatively constant and, using corrected distances, calibrated to be approximately 0.22% lineage/million years. This rate is three-to fourfold lower than that of hominoid primates.     

Functional binding of cardiolipin to cytochromec oxidase

Bovine cytochromec oxidase usually contains 3–4 mol of tightly bound cardiolipin per cytochromeaa ₃ complex. At least two of these cardiolipins are required for full electron transport activity. Without the tightly bound cardiolipin, cytochromec oxidase has only 40–50% of its original activity when assayed in detergents that support activity, e.g., dodecyl maltoside. By measuring the restoration of electron transport activity, functional binding constants for cardiolipin and a number of cardiolipin analogues have been evaluated (K _d,app=1 µM for cardiolipin). These binding constants agree reasonably well with direct measurement of the binding using [¹⁴C]-acetyl-cardiolipin (K _d <0.1 µM) when the enzyme is solubilized with Triton X-100. These data are discussed in relationship to the wealth of data that is known about the association of cardiolipin with cytochromec oxidase and the other mitochrondrial electron transport complexes and transporters.     

Metabolic regulation including anaerobic metabolism inParacoccus denitrificans

Under anaerobic circumstances in the presence of nitrateParacoccus denitrificans is able to denitrify. The properties of the reductases involved in nitrate reductase, nitrite reductase, nitric oxide reductase, and nitrous oxide reductase are described. For that purpose not only the properties of the enzymes ofP. denitrificans are considered but also those fromEscherichia coli, Pseudomonas aeruginosa, andPseudomonas stutzeri. Nitrate reductase consists of three subunits: the subunit contains the molybdenum cofactor, the subunit contains the iron sulfur clusters, and the subunit is a special cytochromeb. Nitrate is reduced at the cytoplasmic side of the membrane and evidence for the presence of a nitrate-nitrite antiporter is presented. Electron flow is from ubiquinol via the specific cytochromeb to the nitrate reductase. Nitrite reductase (which is identical to cytochromecd ₁) and nitrous oxide reductase are periplasmic proteins. Nitric oxide reductase is a membrane-bound enzyme. Thebc ₁ complex is involved in electron flow to these reductases and the whole reaction takes place at the periplasmic side of the membrane. It is now firmly established that NO is an obligatory intermediate between nitrite and nitrous oxide. Nitrous oxide reductase is a multi-copper protein. A large number of genes is involved in the acquisition of molybdenum and copper, the formation of the molybdenum cofactor, and the insertion of the metals. It is estimated that at least 40 genes are involved in the process of denitrification. The control of the expression of these genes inP. denitrificans is totally unknown. As an example of such complex regulatory systems the function of thefnr, narX, andnarL gene products in the expression of nitrate reductase inE. coli is described. The control of the effects of oxygen on the reduction of nitrate, nitrite, and nitrous oxide are discussed. Oxygen inhibits reduction of nitrate by prevention of nitrate uptake in the cell. In the case of nitrite and nitrous oxide a competition between reductases and oxidases for a limited supply of electrons from primary dehydrogenases seems to play an important role. Under some circumstances NO formed from nitrite may inhibit oxidases, resulting in a redistribution of electron flow from oxygen to nitrite.P. denitrificans contains three main oxidases: cytochromeaa ₃, cytochromeo, and cytochromeco. Cytochromeo is proton translocating and receives its electrons from ubiquinol. Some properties of cytochromeco, which receives its electrons from cytochromec, are reported. The control of the formation of these various oxidases is unknown, as well as the control of electron flow in the branched respiratory chain. Schemes for aerobic and anaerobic electron transport are given. Proton translocation and charge separation during electron transport from various electron donors and by various electron transfer pathways to oxygen and nitrogenous oxide are given. The extent of energy conservation during denitrification is about 70% of that during aerobic respiration. In sulfate-limited cultures (in which proton translocation in the NADH-ubiquinone segment of the respiratory chain is lost) the extent of energy conservation is about 60% of that under substrate-limited conditions. These conclusions are in accordance with measurements of molar growth yields.     

Cytochromec oxidase fromParacoccus denitrificans in Triton X-100: Aggregation state and kinetics

Cytochromec oxidase fromParacoccus denitrificans was homogenously dispersed in Triton X-100. Using gel exclusion chromatography and sucrose gradient centrifugation analysis a molecular weight of the detergent-protein complex of 155,000 was determined. After subtraction of the bound detergent (111 mol/mol hemeaa ₃) a molecular weight of 85,000 resulted, which agreed well with the model of a monomer containing two subunits. This monomer showed high cytochromec oxidase activity when measured spectrophotometrically in the presence of Triton X-100 (V _max=85 s^–1). The molecular activity, plotted according to Eadie-Hofstee, was monophasic as a function of the cytochromec concentration. AK _m of 3.6×10^–6 M was evaluated, similar to theK _m observed in the presence of dodecyl maltoside [Naeczet al. (1985).Biochim. Biophys. Acta 808, 259–272].     

Cytochromec oxidase inParacoccus denitrificans. Protein,chemical, structural,and evolutionary aspects

Preparations and protein chemical characterizations performed with cytochromec oxidase (E.C. 1.9.3.1) from the purple bacteriumParacoccus denitrificans are reviewed. The simplest catalytically competent complex of the enzyme consists of two subunits of 62012 and 27999 Da. The theoretical hemea/protein ratio of the purified enzyme is 22.0 nmol/mg. The amino acid sequences of both proteins are compared with examples of subunits I and II of mitochondrial terminal oxidases from the main kingdoms of eukaryotes. The significance of the emerging conserved features such as membrane penetration patterns, invariant residues, stoichiometry, and sites of prosthetic groups are discussed. TheParacoccus enzyme represents the only prokaryotic oxidase detailed so far, which is directly related to the mitochondrial oxidases by common ancestry in the growing O₂ atmosphere.     

Inhibition of the phosphate-stimulated cytochromec oxidase activity by thiophosphate

Yeast and mammalian cytochromec oxidase activity is inhibited by thiophosphate. This inhibition was observed when using either whole mitochondria or the isolated or reconstituted enzyme. The kinetics of the reduction reaction enabled us to demonstrate that thiophosphate acted on th electrons transfer between hemesa anda ₃. With whole mitochondria, phosphate alone stimulated respiration. The inhibition induced by thiophosphate was suppressed by phosphate only in mitochondria, but not when the isolated enzyme was used. The possibility of a kinetic regulation is discussed.Abbreviations CCCP p-carbonylcyanidem-chlorophenylhydrazone - TMPD N,N,N,N-tetramethylp-phenylenediamine - SPi thiophosphate     

Evolutionary aspects of cytochromec oxidase

The presence of additional subunits in cytochrome oxidase distinguish the multicellular eukaryotic enzyme from that of a simple unicellular bacterial enzyme. The number of these additional subunits increases with increasing evolutionary stage of the organism. Subunits I–III of the eukaryotic enzyme are related to the three bacterial subunits, and they are encoded on mito-chondrial DNA. The additional subunits are nuclear encoded. Experimental evidences are presented here to indicate that the lower enzymatic activity of the mammalian enzyme is due to the presence of nuclear-coded subunits. Dissociation of some of the nuclear-coded subunits (e.g., VIa) by laurylmaltoside and anions increased the activity of the rat liver enzyme to a value similar to that of the bacterial enzyme. Further, it is shown that the intraliposomal nucleotides influence the kinetics of ferrocytochromec oxidation by the reconstituted enzyme from bovine heart but not fromP. denitrificans. The regulatory function attributed to the nuclear-coded subunits of mammalian cytochromec oxidase is also demonstrated by the tissue-specific response of the reconstituted enzyme from bovine heart but not from bovine liver to intraliposomal ADP. These enzymes from bovine heart and liver differ in the amino acid sequences of subunits VIa, VIIa, and VIII. The results presented here are taken to indicate a regulation of cytochromec oxidase activity by nuclear-coded subunits which act like receptors for allosteric effectors and influence the catalytic activity of the core enzyme via conformational changes.     

Ferricytochromec induces monophasic kinetics of ferrocytochromec oxidation in cytochromec oxidase

The kinetics of ferrocytochromec oxidation by reconstituted cytochromec oxidase (COX) from bovine heart was followed by a spectrophotometric method, using on-line data collection and subsequent calculation of reaction rates from a function fitted to the progress curve. When reaction rates were calculated at increasing reaction times, the multiphasic kinetics of ferrocytochromec oxidation gradually changed into monophasic Michaelis-Menten kinetics. The same phenomenon was observed when ferrocytochromec oxidation was followed in the presence of increasing amounts of ferricytochromec. From these results we conclude that ferricytochromec shifts the multiphasic kinetics of ferrocytochromec oxidation by COX into monophasic kinetics, comparable to high ionic strength conditions. Furthermore, we show that ferricytochromec inhibits the high affinity phase of ferrocytochromec oxidation in an apparently competitive way, while inhibition of the low affinity phase is noncompetitive. These findings are consistent with a regulatory site model where both the catalytic and the regulatory site bind ferro- as well as ferricytochromec.     

Isolation and characterization of human heart cytochromec oxidase

Cytochromec oxidase was isolated from human hearts and separated by SDS gel electrophoresis. The identity of polypeptide bands with known subunits was demonstrated by immunoblotting with monospecific antisera to rat liver cytochromec oxidase subunits. The polarographically determined kinetics of cytochromec oxidation were similar to those reported for the bovine heart enzyme.     

On the role of subunit III in proton translocation in cytochromec oxidase

Mammalian mitochondrial cytochromec oxidase catalyzes the transfer of electrons from ferrocytochromec to molecular oxygen in the respiratory chain, while conserving the energy released during its electron transfer reactions by the vectorial movement of protons across the inner membrane of the mitochondrion. The protein domain that translocates the protons across the membrane is currently unknown. Recent research efforts have investigated the role of one of the transmembrane subunits of the enzyme (III,M _r 29,884) in the vectorial proton translocation reaction. The data that favor subunit III as integral in vectorial proton translocation as well as the data that support a more peripheral role for subunit III in proton translocation are reviewed. Possible experimental approaches to clarify this issue are presented and a general model discussed.