J. Biochem. 99, 1-8 (1986)

ATP Synthesis in Cell Envelope Vesicles of Halobacterium halobium Driven by Membrane Potential and/or Base-Acid Transition1

Yasuo MUKOHATA, Masaharu ISOYAMA , and Ayumi FUKE

Department of Biology, Faculty of Science , Osaka University, Toyonaka, Osaka 560

Received for publication, August 15, 1985

Cell envelope vesicles active in ATP synthesis were prepared from Halobacterium halobium cells, which genetically lack , by sonication in the pres ence of substrates. ATP was synthesized when vesicles were illuminated to build up membrane potential through the action of halorhodopsin. The threshold value of membrane potential for ATP synthesis was about - 100 mV relative to the ex ternal medium, i.e., inside-negative. ATP synthesis also occurred in the dark upon acidification of the external medium of a suspension of cell envelope vesicles. This base-acid transition ATP synthesis took place when the pH difference was greater than 1.6 units. The threshold pH difference was lowered when the base-acid transition was carried out under dim light which induced a membrane potential of about - 100 mV.

Regardless of the sort of driving force, ATP synthesis was optimum at the

intravesicular pH of around 6.5 and almost nil at 8, where ATP syntheses by F0F1

type in other organisms are most active. The synthesis could be inhibited

by N,N•Œ-dicyclohexylcarbodiimide (DCCD) with a half-maximum inhibition at

around 25 ƒÊm/2 mg protein/ml. These results strongly suggest that in halobacteria a DCCD-sensitive H+- translocating ATP synthase is in operation which is driven by membrane potential and/or pH gradient, and obeys chemiosmotic energetics. The results also suggest that the ATP synthase may not be identical to F0F1 type H+-translocating ATPases found in mitochondria, chloroplasts and eubacteria.

1This work was supported in part by Grant-in-Aid for special project research on Bioenergetics (#57122006 , 58114006, 59108006) from the Ministry of Education, Science and Culture of Japan.

Abbreviations: DCCD, N,N•Œ-dicyclohexylcarbodiimide; PIPES, piperazine-N,N•Œ-bis(2-ethanesulfonic acid); TPP+, tetraphenylphosphonium cation; TPT, triphenyltin chloride; pHex, the external pH of a vesicular suspension; pHin, the intravesicular pH given by the pH of the "stuffing" solution; •¢pH, the pH difference across the vesicle membrane, =pHin-pHex; •¢ƒµ, membrane potential, electrical potential difference across the vesicle membrane, relative to the external medium.

Vol. 99, No. 1, 1986 1 2 Y. MUKOHATA, M. ISOYAMA, and A. FUKE

Halorhodopsin (1, 2) was originally detected (3, ented subsequently. The second aim was, using 4) as a light energy transducer in ATP synthesis in this vesicle preparation, to examine and expose a red strain (R1mR) of Halobacterium halobium some features of the machinery/mechanism of which lacks bacteriorhodopsin. This pro ATP synthesis in this extremely halophilic archae tein (4, 5) is distributed widely among halobacterial bacterium. strains and builds up an inside-negative 4,p (6) by In this article, we describe a reproducible pro pumping in Cl- (7) in the light. cedure for preparing cell envelope vesicles which The light-driven ATP synthesis in bacterio can synthesize ATP whenever halorhodopsin is -depleted strain R1mR (3, 4) [and simi illuminated and/or •¢pH is applied in the dark. lar strains isolated later as halorhodopsin-contain ATP can be synthesized by 4 over - 100 mV in ing strains, such as ET-15 (8), L33 (9), Y-1 (10), the absence of •¢pH or by •¢pH over 1.6 pH units and YH-10 (11)] under anaerobic conditions should in the dark. The additivity of effects of •¢ƒµ and

be driven by this J V. Meanwhile, the possible •¢ pH on ATP synthesis was also demonstrated. presence of trace bacteriorhodopsin in R1mR was These results favor the hypothesis of chemiosmotic suggested (12). Stoeckenius and Bogomolni (13) coupling of ATP synthesis by H+-translocating proposed that our results (3, 4) might be due to ATPase. On the other hand, some indications this trace bacteriorhodopsin. However, their sug were obtained that the ATPase may not be the gestion is not valid, since the R1mR strain was F0F1 type which is the only ATP synthase so far analyzed to carry a genetic defect in bacteriorhod known in all living organisms. (14). Weber and Bogomolni (8) more recently MATERIALS AND METHODS claimed that halorhodopsin could not drive ATP synthesis in their ET-15 cells even when they Preparation of Cell Envelope Vesicles-Halo illuminated their P588 (5), which is considered to bacterium halobium R1mR cells were cultured as be halorhodopsin with contaminating sensory described earlier (3) for 5 days at 40•Ž, then rhodopsin (15, 16) that is inert in light energy harvested and washed twice with a saline buffer transduction. (4 M NaCl and 10 mm PIPES pH 6.8) by centrifu In whole cells of halobacteria, it was reported gation. The pellet (about 2 g protein from 8 liters (17) that no chemiosmotic potential gradient culture) was suspended in 100 ml of the "stuffing" seemed to be needed for ATP synthesis. Another solution (3 M NaCl, 0.9 M KCl, 50 mm MgCl2, 10 report (18) supported this and stated that adenylate mm PIPES pH 6.8, 20 mm phosphate buffer pH kinase takes a major role in cellular energetics. In 6.8, and 5 mm ADP, unless otherwise noted) then addition, after light-driven ATP synthesis was sonicated (50 ml each in a 100 ml beaker; Heat found in halobacteria with (19) or without (3) System-Ultrasonics, type W-225R, power 7, duty bacteriorhodopsin, the machinery of ATP synthesis cycle 50%, 1/2" tip) at 15•Ž for 2 min. The has been postulated to be F0Fl ATPase. Inhibi disrupted cells were centrifuged at 22,000 x g for tion of ATP synthesis by DCCD (4, 20) seemed to 30 min at 4•Ž and the pellet was suspended in the confirm this view. However, the machinery/mech basal "stuffing" solution (the "stuffing" solution anism of this ATP synthesis has not yet been without P, and ADP). The suspension was lay proven. ered on top of 15 % sodium tartrate in 4 M NaCl as Therefore, it would be interesting to examine described by Clark and MacDonald (21) , and cen the features of the machinery/mechanism of ATP trifuged at 16,000 x g for 45 min at 10•Ž. The synthesis in halobacteria in the absence of any cell envelope vesicles were collected from the layer

interference which would arise when whole cells just above the tartrate layer, then washed twice are used. The first aim of the present research with the basal "stuffing" solution by centrifugation was thus to prepare cell envelope vesicles which at 30,000 x g for 30 min each. The final vesicle can steadily synthesize ATP. ATP formation in suspension was stored overnight at 4•Ž in the dark halobacterial cell envelope vesicles by light has in order to reduce ATP which mainly originated been reported (21) but no detailed results were from the reagent ADP used. Proteins were deter- given, nor has confirmation by others been pres mined by the Lowry, method using bovine serum

J. Biochem. ATP SYNTHESIS IN HALOBACTERIAL VESICLES 3 albumin as a standard. (pHex), i.e., pHin-pHex. The outputs of these ATP Synthesis in Light-Vesicles (about 14 electrode signals were recorded except for the mg protein) were suspended in 7 ml of the basal TPP+-electrode signal in the base-acid transition "stuffing" solution in a glass ve ssel kept at 30•Ž experiments, because the electrode is sensitive to and illuminated by yellow light (> 520 nm , about pH. 105 lux from a 750-W projector lamp). Portions Chemicals-ADP was purchased from Yamasa (routinely 50 ƒÊl) of the vesicle suspension were Shoyu Co. [nominal contamination (ATP+AMP) sampled at given time intervals throughout the < 3 %], fire-fly tails from Sigma (or ATP biolumi experiment and mixed with 1.5 ml each of 40 ms nescence assay kit from Boehringer), and all other MgCl2 and 10 mm N-tris(hydroxymethyl)methyl-2- reagents from Yashima Chemicals (reagent grade), aminoethanesulfonic acid (TES, pH 7.4) at 98•Ž except that NaCl (common grade) from the Japan for 5 min to dissolve ATP. Each sample was Tobacco Industry and tap water were used for the cooled and then assayed for ATP by a luciferin culture media. luciferase method (22) using a photon-counting ATP photometer (Toyo Scientific Industry, type RESULTS AND DISCUSSION ATP-237). ATP Synthesis by Base-Acid Transition- Substrate "Stuffed" Vesicles-As shown in the Vesicles (about 14 mg protein) were suspended in following results, the cell envelope vesicles were 7 ml of the basal "stuffing" solution in which 10 active in ATP synthesis almost every time when mm PIPES had been replaced by 1 mm PIPES pH prepared by the present procedure, described in " MATERIALS AND METHODS 6.8, and incubated for 30 min at 30•Ž in the dark. ." Although the The acidifying buffer (for routine experiments, procedure was mostly based on that described by 0.5 M Na-citrate buffer pH 4.0, 140 ƒÊl) was then Clark and MacDonald (21), the method of sub

injected into the vesicle suspension to bring the strate incorporation into vesicles was totally dif

pHex from 6.8 to 4.0. When similar experiments ferent. They incubated vesicles for a long time were run for the (intravesicular) pH dependence, with substrates which might be incorporated spon cells in the "stuffing" solutions with buffers (Good taneously. This method may have led to the buffer at 10 mm) at the given pH's (pHin) were difficulty in confirming their results.

separately vesiculized and washed. The initial In the present study, we sonicated cells in the "stuffing" solution which contained substrates in pH,,, was kept at the pHin with the same buffer. The pH's of acidifying buffers (Good buffer or a volume about 20 times larger than the total citrate, at 0.5 M) were adjusted for individual vesi intracellular space. Therefore, the substrate con cle suspensions to give •¢pH=3.0. To find the centrations in the vesicles were expected to be threshold •¢pH, the pHin was kept at 6.8 and the almost equal to those in the "stuffing" solution. pH of the acidifying buffer was adjusted to obtain By the present procedure, the composition of various values of •¢pH. Portions (50 Id) of the the intravesicular solution can be changed rather suspension were sampled before and after the drastically by changing the composition of the "stuffing" solution base-acid transition and assayed for ATP as above. . Vesicles could be prepared in

Membrane Potential and pH Monitoring-The •¢ƒµ the "stuffing" solution in which 3 M NaCl was

, value was estimated with a TPP+-sensitive replaced by 3.1 M KCl (making 4 M KCl in total) electrode (23) from the external concentration and were active in ATP synthesis (data not shown).

(initially 2 ƒÊM) of TPP+ which had been partitioned The pHin value, buffer and/or substrate concen between the intravesicular space and the external trations as well as salt contents could also be medium. The Ap value was calculated with the readily modified (see below). The present "stuff- intravesicular volume of 2.7 µl/mg protein (24) ing" procedure should be useful in other studies and expressed relative to the external medium. of halobacterial physiology.

The external pH was monitored by a combination Light-Driven ATP Synthesis-The intravesic glass electrode. The •¢pH value was expressed as ular ATP increased when the cell envelope vesicles the difference between the given pH of the "stuff- were illuminated and decreased after the illumina ing" solution (pHin) and the external pH measured tion was removed (Fig. 1). In the presence of

Vol. 99, No. 1, 1986 4 Y. MUKOHATA, M. ISOYAMA, and A. FUKE

Fig. 2. Dependence on AP of ATP synthesis in halo- bacterial vesicles. The ATP data as in Fig. 1 were obtained I min after the start of actinic illumination at

pH;n=pHex=6.8 (TPT=2 ƒÊM). •, the data obtained in a series of experiments with one vesicle preparation at various values of •¢ƒµ produced by dimming the actinic

Fig. 1. Light-driven ATP synthesis in cell envelope illumination. O , the data obtained with different vesicles of Halobacterium halobiuin R1mR which lacks vesicle preparations under various conditions, e.g., over bacteriorhodopsin. R1mR vesicles (2.1 mg protein/ml) a wide range of actinic light intensity, in the presence of

were incubated at 30°C and illuminated with yellow light uncouplers and under normal routine conditions. (> 520 nm, 105 lux) in the presence (+) or absence (-) of 21:MTPT. The initial pH values of both intravesicu lar ("stuffing"; see " MATERIALS AND METHODS") be over-estimated because of adsorbed (not dis solution and the external medium were 6.8. Changes in tributed in the intravesicular space) TPP+ which

the TPP+-electrode output are also shown with the cali could not be determined accurately. Upon illu

brated JP scale. ON and OFF denote the start and mination, •¢ƒµ, increased to nearly - 150 mV (Fig.

the end of actinic illumination. 1). When the intensity of actinic illumination was

reduced by changing the lamp current or inserting

TPT, an OH-/Cl- exchanger (25), any pH differ neutral density filters, the amounts of synthesized

ence (if produced) across the vesicle membrane ATP as well as the size of .d decreased. The

would have been canceled (26). Therefore, ATP relationship between the amounts of synthesized

was synthesized in the absence of •¢pH, by the ATP and the size of d p (Fig. 2) clearly shows a

•¢ƒµ, formed by illuminated halorhodopsin. In the threshold •¢ƒµ of about -100 mV below which no

absence of TPT, incorporation of H+ [through ATP could be synthesized. Here, the amounts

DCCD-insensitive pathways (2)] following •¢ƒµ for of ATP determined 1 min after the start of illu

mation would have caused acidification of intra mination (when •¢ƒµ, gave an almost steady reading)

vesicular space as in intact cells (26), which would are plotted and regarded as proportional to the

depress the synthesis of ATP (Fig. 1). As the initial rates of ATP synthesis. Although the •¢ƒµ

vesicles had lost their respiratory activity, the dark value depends on the intravesicular space, a 20

(and also light) ATP level was not affected by change in this volume affects the calculated •¢ƒµ by anaerobiosis (under N2) or addition of 2 mm KCN less than •} 5 mV. The value may also be dis

(data not shown). cussed in relation to the phosphorylation potential

The level of .A in the dark was roughly esti {4 G P = d Ga + 2.3 RT log([ATP]/[ADP][Pi])} . How- mated to be - 50 mV, which was the lower limit ever, even though the concentrations of ADP,

of the present •¢ƒµ measuring system, and this may Pi and ATP in the "stuffing" solution were varied

J. Biochem. ATP SYNTHESIS IN HALOBACTERIAL VESICLES 5

drastically, no appreciable shift of the threshold The pH dependence of J p-driven ATP syn-

was obtained [data not discussed here, since this thesis in the absence of a pH gradient (i.e., pHin= is a kind of fundamental problem in energetics, pHex in the presence of 2 ƒÊm TPT) showed a unique

as in mitochondria (27), chloroplasts (28), and profile (Fig. 3). ATP can be synthesized between bacteria (29)]. Thus the obtained threshold J p pH 5.5 and 7.5 with an optimum at around 6.5. is thought to be characteristic of halobacteria. If Since the observed •¢ƒµ was sufficiently large (- 142

the H+/ATP stoichiometry is assumed to be 2 and •}5 mV) between pH 5.5 and 8.0, the profile is not

3 with AG, of 8 kcal/mol, the proton motive force, due to the lack of a driving force, i.e., •¢ƒµ, but is

•¢ p=•¢ƒµ-60•¢pH (mV), to produce ATP, should characteristic of the halobacterial machinery/mech be more than -174 and -116 mV, respectively. anism of ATP synthesis. For ordinary F0F1

If the stoichiometry is 3, the calculated value is ATPase, the pH profiles are much broader with

fairly close to, but still larger than the observed the pH optimum at 7-8, even in mitochondria

one. The observed threshold value would suggest functioning under an external electric field (30).

that the machinery/mechanism which synthesizes The light -•¢ƒµ-driven ATP synthesis was in

ATP in halobacteria may not be identical to F0F1- hibited by DCCD (Fig. 4) as expected from the

ATPase (further discussion, see below). earlier results with cells (4, 20) and vesicles (21). The half-maximum inhibition occurred at about

25 ƒÊM for a vesicle suspension of 2 mg protein/ml.

The ATP synthesis was also inhibited completely

Fig. 4. Effects of DCCD on the ATP synthesis in

halobacterial vesicles. DCCD (final concentration) was added to the vesicle suspension 1 h before exerting the Fig. 3. The pH dependence of d F-driven ATP synthe driving force, i.e., actinic illumination (-,]P= --145 mV) sis in halobacterial vesicles. The experimental condi or acidification (•¢pH=2.8). For the •¢ƒµ driven synthe tions were as in Fig. 1 in the presence of 2 ƒÊm TPT, sis (O) the experimental conditions were as in Fig. 1 except that the buffers used were 10 mm each of 2-(N- morpholino)ethanesulfonic acid (MES), PIPES and N- (TPT=2 ƒÊM) except that 1.95 mg protein/ml of vesicles was used. For the •¢pH-driven synthesis (• ; see below) tris(hydroxymethyl)methylglycine (Tricine), and the the experimental conditions were as in Fig. 5 (no TPT) pHex (=pHin) was adjusted by adding HCl or NaOH. except that 2.0 mg protein/ml of vesicles was used. The The same symbols indicate the data obtained in one amounts of ATP synthesized I min after applying the series of experiments with one vesicle preparation. The driving force were plotted relative to the control data were normalized at pH 6.8 where the amounts of (DCCD=O) which synthesized ATP in amounts of 220 synthesized ATP were between 180 and 225 pmol/mg and 460 pmol/mg protein under J IF and •¢pH, respec protein. The pH dependence of light-induced •¢ƒµ is tively. also shown.

Vol. 99, No. 1, 1936 6 Y. MUKOHATA, M. ISOYAMA, and A. FUKE

by an uncoupler, 3,5-di-tert-butyl-4-hydroxyben- (31). The smaller •¢pH may reflect some con zylidenemalononitrile (SF6847) at S ƒÊM (data not tribution of the resting •¢ƒµ (< -50 mV, see above), shown). or may result from the involvement of a different

Base-Acid Transition ATP Synthesis-The in machinery/mechanism of ATP synthesis.

travesicular ATP increased after the external me The base-acid transition ATP synthesis was

dium was rapidly acidified in the dark (Fig. 5). optimum at around pH 6.5 (inside the vesicle) and

In the presence of 2ƒÊM TPT, the pH difference almost nil at pH 8 (Fig. 7). It is of interest that

across the membrane collapsed and no ATP in- the pH profiles of ATP synthesis driven by J p crease was detected. This ATP synthesis is thus (Fig. 3) and •¢pH (Fig. 7) are significantly different. to be considered as a •¢pH-driven synthesis, and so At pH 7.5, the activity is half the maximum in the

suggests that the machinery is a H+-translocating •¢ pH-driven synthesis, while it is almost nil in the ATPase. •¢ƒµ-driven synthesis . In so far as the same ma

The base-acid transition ATP synthesis de chinery is driven by the different driving forces,

pended on the size of •¢pH (Figs. 5 and 6) and the pH profiles might be expected to be identical.

there was clearly a •¢pH threshold for the synthesis The observed difference can be understood in terms

of ATP. The threshold •¢pH was estimated to of the affinity of the H+-binding site(s) on the

be 1.6, which agrees with the threshold •¢ƒµ of -100 ATPase. The H+-translocating ATPase takes at

mV described above, if the resting •¢ƒµ, in the dark least three steps to translocate H+ coupled with

is neglected. This value of 1.6 pH units is again ATP synthesis, i.e., binding of H+ on the external

smaller than that required for ordinary F0F1- medium side (the F0 side in the case of an F,,F,- ATPase; in chloroplasts, the acid-base transition

ATP synthesis required at •¢pH of at least 2.2-2.9

Fig. 6. Dependence on •¢pH of ATP synthesis in

Fig. 5. The base-acid transition ATP synthesis in halobacterial vesicles. 0,9, the ATP data obtained halobacterial vesicles. The "stuffed" (see " MATERI with two different vesicle preparations in the dark as in ALS AND METHODS") vesicles at pHin=6.8 were Fig. 5 (TPT=0) at 1 min after the base-acid transition . suspended in the basal "stuffing" solution at pHex =6.8 0, •, the data obtained similarly with two vesicle prepa at 30•Ž, and then the pH of the suspension was brought rations in the presence of d _ -100 mV (for 2 min be - to a given pH (pHex) to give a transmembrane pH dif fore the transition) under controlled actinic illumination . ference, •¢pH=pHin -pHex. 0, data obtained in the The amounts of ATP synthesized under illumination absence of TPT. •, the data at •¢pH=2.9 in the were usually smaller than those in the dark . The initial presence of 2 ƒÊM TPT. pHin (and pHex) was 6.8.

J. Biochem . ATP SYNTHESIS IN HALOBACTERIAL VESICLES 7

ATPase which is composed of two pairs of 86K and 64K peptides and is active in a narrow pH range with an optimum at around 6 (Nanba , T. & Mukohata, Y., in preparation). We have not been able to detect peptides which could be as- signed to Fl-ATPase. These and other results described so far strongly suggest that there is an H+-translocating ATP synthase in Halobacteri aceae (and possibly in Archaebacteriaceae) which is not identical to the ordinary F0F1-type ATPase (further discussion will be published elsewhere). Additivity of the Driving Forces on A TP Syn-

thesis-When the base-acid transition was carried

out under dim light which gave the -lip of -100

mV (in the absence of TPT), ATP was synthesized

at lower •¢pH values. The threshold •¢pH shifted Fig. 7. The pH dependence of •¢pH-deriven ATP to 1 pH unit (Fig. 6). However, if the threshold synthesis in halobacterial vesicles. The vesicles were d?y of -100 mV under conditions of pHin=pHex prepared as in Fig. 3 (TPT=O) at various pHin values and the pHex after the transition was chosen so as to permits •¢ƒµ-driven synthesis, the expected •¢pH threshold at •¢ƒµ= -100 mV should be zero but give •¢pH=3 in each run of experiments. The relative amounts (normalized at pHin=6.5; ATP assayed was not 1 pH unit. Therefore, the result was contra 510 pmol/mg protein) of ATP found 1 min after the dictory. In this connection, it should be noted

transition were plotted. The pH dependence of .1 l- that during preincubation under dim light to give driven synthesis (Fig. 3) is also shown (broken line). •¢ƒµ = -100 mV (2 min before the transition) , H+ was taken up by the vesicles through a DCCD-

ATPase), transfer of bound H+ to the internal insensitive pathway (2) as indicated by a slight

medium side (F, , side) and release of H+ there. alkalization of the pHex signal (data not shown).

When pHin (the abscissas in Figs. 3 and 7) is 7.5, This suggests that the actual pH difference upon

in the •¢pH-driven synthesis the H+-binding site(s) base-acid transition was smaller than the nominal on the ATPase is exposed to pHex=4.5 because •¢ pH=pHin-pHex. Even when PIPES at 100 mm •¢ pH=3, while in the •¢ƒµ-driven synthesis it is instead of 10 mm in the "stuffing" solution was

exposed to pHex=pHin=7.5. If the H+-binding used at sonication, the results were almost the

site(s) has an affinity for H+ of the order of 10 same as in Fig. 7. Therefore, for energetic cal

nM, the d?p-driven synthesis at pH 7.5 would have culations and discussion, further investigation is

much lower activity because of the lack of H+ needed, especially of the actual intravesicular pH

which should be translocated. values. Here, we simply describe the existence of

The •¢pH-driven ATP synthesis is also in qualitative additivity in the driving forces.

hibited by DCCD with essentially the same profile We can now prepare cell envelope vesicles

as •¢•¢•¢ƒµ-driven synthesis, (Fig. 4). The DCCD- from halobacterial cells which contain halorhodop

binding proteins from halobacterial vesicles have sin but not bacteriorhodopsin. The synthesis of

been reported (32) to have molecular weights of ATP in these vesicles is driven by •¢ƒµ and/or •¢pH,

62-45K (or a little larger; Konishi, T., personal apparently obeying chemiosmotic energetics. The

communication) and 12-10.2K, and not to be ex vesicles should prove useful in studies to under-

tracted by a chloroform-methanol mixture. The stand the physiology and energetics of this ex

former peptide bound DCCD with an apparent tremely halophilic archaebacterium.

affinity constant of 29 ƒÊM (32) and the latter was We can readily control the size of •¢ƒµ by

assigned as an Na+/H+ antiporter. Our prelimi varying the intensity of actinic illumination on the

nary data show that a 78K dalton peptide binds integral protein, Cl--pump halorhodopsin. The

DCCD in a fashion paralleling the inhibition of present experimental system should therefore be

ATP synthesis. We isolated a membrane-bound useful in studying the mechanism of ATP synthesis,

Vol. 99, No. 1, 1986 8 Y. MUKOHATA, M. ISOYAMA, and A. FUKE

because it is much simpler than others, e.g., a K+/ 14. DasSarma, S., RajBhandary, V., & Khorana, H.D.

valinomycin/uncoupler system to produce/control •¢ƒµ (1983) Proc. Natl. Acad. Sci. U.S. 80, 2201-2205 15. Bogomolni, R.A. & Spudich, J.H. (1982) Proc. Natl. We suggest here the existence of a different Acad. Sci. U.S. 79, 6250-6254 16. Tsuda, M., Hazemoto, M., Kamo, N., Kobatake, type of H+-translocating ATP synthase in place of Y., & Terayama, Y. (1982) Biochem. Biophys. Res. F0F1-ATPase. Because F0F1 ATPase has been Commun. 108, 970-976 believed to be ubiquitous in all living organisms, 17. Michel, H. & Oesterhelt, D. (1980) Biochemistry 19, characterization of this ATP synthase should be 4615-4619 of crucial importance in understanding the position 18. Helgerson, S.L., Requadt, C., & Stoeckenius, W. of halobacteria (and possibly archaebacteria) in (1983) Biochemistry 22, 5746-5753 evolution. 19. Oesterhelt, D. (1974) in Membrane Proteins in Trans port and Phosphorylation (Azzone, G.F., Klingen- berg, M.E., Quagliariello, E., & Siliprandi, N., eds.) REFERENCES pp. 79-84, North-Holland Publishing Co., Amster 1. Mukohata, Y., Matsuno-Yagi, A., & Kaji, Y. dam (1980) in Saline Environment (Morishita, H. & 20. Danon, A. & Stoeckenius, W. (1974) Proc. Natl. Masui, M., eds.) pp. 31-37, Business Center for Acad. Sci. U.S. 71, 1234-1238 Academic Societies, Tokyo 21. Clark, R.D. & MacDonald, R.E. (1981) Biochem. 2. Mukohata, Y. & Kaji, Y. (1981) Arch. Biochem. Biophys. Res. Commun. 102, 544-550 Biophys. 206, 72-76 22. Yagi, A. (1978) Ph.D. Thesis, Osaka Univ. 3. Matsuno-Yagi, A. & Mukohata, Y. (1977) Biochem. 23. Kamo, N., Muratsugu, M., Hongoh, R., & Koba Biophys. Res. Commun. 78, 237-243 take, Y. (1979) J. Membr. Biol. 49, 105-121 4. Matsuno-Yagi, A. & Mukohata, Y. (1980) Arch. 24. Eisenbach, M., Cooper, S., Garty, H., Johnstone, Biochem. Biophys. 199, 297-303 R.M., Rottenberg, H., & Caplan, S.R. (1977) Bio 5. Lanyi, J.K. & Weber, H.J. (1980) J. Biol. Chem. 255, chim. Biophys. Acta 465, 599-613 243-250 25. Selwyn, M.J., Dawson, A.P., Stockdale, M., & 6. Lindley, E.V. & MacDonald, R.E. (1979) Biochem. Bains, N. (1970) Eur. J. Biochem. 14, 120-126 Biophys. Res. Commun. 88, 491-499 26. Mukohata, Y. & Kaji, Y. (1981) Arch. Biochem. 7. Schobert, B. & Lanyi, J.K. (1982) J. Biol. Chem. Biophys. 208, 615-617 257,10306-10313 27. Wilson, D.F. & Forman, N.G. (1982) Biochemistry 8. Weber, H.J. & Bogomolni, R.A. (1981) Photochem. 21,1438-1444 Photobiol. 33, 601-608 28. Giersch, C., Heber, U., Kobayashi, Y., Inoue, Y., 9. Lanyi, J.K. & Oesterhelt, D. (1982) J. Biol. Chem. Shibata, K., & Heldt, H.W. (1980) Biochim. Biophys. 257, 2670-2677 Acta 590, 59-73 10. Ogurusu, T., Maeda, A., Sasaki, N., & Yoshizawa, 29. Hatchens, G.D. & Kell, D.B. (1982) Biochem. J. T. (1981) J. Biochem. 90, 1267-1273 206,351-357 11. Hazemoto, N., Kamo, N., Terayama, Y., Kobatake, 30. Hamamoto, T., Ohno, K., & Kagawa, Y. (1982) Y., & Tsuda, M. (1983) Biophys. J. 44, 59-64 J. Biochem. 91, 1759-1766 12. Greene, R.V. & Lanyi, J.K. (1979) J. Biol. Chem. 31. Jagendorf, A.T. & Uribe, E. (1966) Proc. Natl. Acad. 254,10986-10994 Sci. U.S. 55, 170-177 13. Stoeckenius, W. & Bogomolni, R.A. (1982) Anna. 32. Konishi, T. & Murakami, N. (1984) FEBS Lett. Rev. Biochem. 52, 587-616 169,283-286

J. Biochem.