Proc. Nati. Acad. Sci. USA Vol. 85, pp. 7972-7976, November 1988 Biochemistry Bafilomycins: A class of inhibitors of membrane ATPases from microorganisms, animal cells, and plant cells (membrane ATPase/vacuolar ATPase/macrolide antibiotics) EMMA JEAN BOWMAN*, ANNETTE SIEBERSt, AND KARLHEINZ ALTENDORFt *Department of Biology, University of California, Santa Cruz, CA 95064; and tUniversitat Osnabrfick, Fachbereich Biologie/Chemie, Postfach 4469, D-4500 Osnabruck, Federal Republic of Germany Communicated by Peter Mitchell, August 1, 1988

ABSTRACT Various membrane ATPases have been tested inhibitors that have proved useful for these enzymes include for their sensitivity to baflomycin Al, a macrolide antibiotic. (i) N,N'-dicyclohexylcarbodiimide, which also inhibits F1F0 F1Fo ATPases from bacteria and mitochondria are not affected and E1E2 ATPases, (ii) N-ethylmaleimide, which inhibits by this antibiotic. In contrast, ElE2 ATPases-e.g., the K+- E1E2 enzymes (23) and at least one F1F0 ATPase (24), and (iii) dependent (Kdp) ATPase from Escherichia coli, the Na4+,K+- NO3 -, which is effective only at millimolar concentrations or ATPase from ox brain, and the Ca2+-ATPase from sarcoplas- greater. Thus, no potent specific inhibitor of vacuolar ATP- mic reticulum-are moderately sensitive to this inhibitor. ases has yet been identified. Finally, membrane ATPases from Neurospora vacuoles, chro- The bafilomycins A1, B1, C1, and D, macrolide antibiotics maffin granules, and plant vacuoles are extremely sensitive. with a 16-membered lactone ring, were isolated from Strep- From this we conclude that bafilomycin Al is a valuable tool for tomyces sp. (25). These compounds inhibited growth of distinguishing among the three different types of ATPases and Gram-positive bacteria and fungi in a disc diffusion assay. In represents the first relatively specific potent inhibitor of va- addition, it has been reported that bafilomycin C1 inhibits the cuolar ATPases. enzymatic activity of the Na4,K4-ATPase (26, 27). In this communication, we compare the effects ofbafilomycin A1 on In both prokaryotic and eukaryotic organisms most ion- representative enzymes of the three classes of ATPases. The pumping ATPases that have been characterized fit into one of results show that bafilomycin A1 is useful for distinguishing three structural types.t (i) The F1Fo ATPases (F-type) have among the different types of ATPases and that it is an been found in the inner mitochondrial membrane (2), the extremely potent inhibitor of the vacuolar ATPases. thylakoid membrane of chloroplasts (3), and the bacterial cytoplasmic membrane (4). (ii) The E1E2 ATPases (P-type) are present in the cell plasma membranes of fungi (5), plants MATERIALS AND METHODS (6), and animals [including the Na',K4-ATPase (7) and the Preparative Procedures. The preparation of the ATP syn- H + ,K + -ATPase (8)], as well as in the sarcoplasmic reticulum thase from Escherichia coli KY 7485 was as described (28, of muscle cells (Ca2+-ATPase) (9) and the bacterial cyto- 29). The K4-dependent (Kdp)-ATPase from E. coli strain plasmic membrane (K+-ATPase) (10, 11). (iii) A third class TK2240-40 was prepared as described (30). Preparation of ofATPases (V-type) has been identified and partially purified plasma membranes, mitochondria, vacuolar membranes (31), from membranes of fungal and plant vacuoles (ref. 12 and solubilized vacuolar ATPase (32), and purified plasma mem- refs. therein), of coated vesicles (13, 14), and of chromaffin brane ATPase (33) from Neurospora crassa was as de- granules (15, 16). As suggested by Mellman et al. (17), we use scribed. Purified Na4,K4-ATPase from ox brain (34) was a the term "vacuolar ATPases" to refer to this third group. gift from W. Schoner (University of Giessen, F.R.G.). The F1Fo ATPases typically use the electrochemical gra- Bovine chromaffin granule membranes were provided by dient of H+ (18) or occasionally a Na+ gradient (19) to D. K. Apps (University of Edinburgh, Scotland). Sarcopla- synthesize ATP. This type of enzyme also exhibits ATPase smic reticulum vesicles (Ca2+-ATPase) from rabbit muscle activity, in some cases only after activation with proteases were a gift from G. Blickenstaff (Tufts University, Boston). (20). The enzymatic activity of F1Fo ATPases is inhibited by Plant vacuolar membranes were provided by L. Taiz (Uni- azide and N,N'-dicyclohexylcarbodiimide; the mitochondrial versity of California, Santa Cruz, CA). ATPase is inhibited by oligomycin as well (21). Assays. The following procedures have been described: In the E1E2 ATPases the energy released by the hydrolysis assays for ATPase activity of the E. coli F1F0 and the of ATP is coupled to the translocation of cations across the Kdp-ATPase (30, 35), for the plasma membrane ATPase, the membrane. The enzyme cycles through conformational mitochondrial F1F0 ATPase, and the vacuolar ATPase of N. states including the formation of a phosphorylated interme- crassa (31), for the plant vacuolar ATPase (36), for the diate. The enzymatic activity is not affected by azide or Ca2+-ATPase (37), and for the Na4,K4-ATPase (ref. 38, oligomycin but is specifically inhibited by vanadate, in most using the assay buffer described in ref. 39). To assay the cases by N-ethylmaleimide and fluorescein isothiocyanate, chromaffin granule ATPase, we used the procedure for the N. and, in the case of the Na4 ,K4-ATPase, by ouabain (5-11). crassa vacuolar ATPase. The vacuolar ATPases appear to hydrolyze ATP, gener- Stock solutions of bafilomycins were prepared in dimethyl ating a proton gradient that is used for acidification of sulfoxide and stored at - 20'C. The actual concentrations of compartments within cells (12, 17, 22). This group of ATP- the stock solutions were determined spectrophotometrically ases has been distinguished from the other two by virtue of using the molar extinction coefficients published in ref. 25. its inhibitor specificity. The vacuolar ATPases are not inhib- The final concentration of sulfoxide in the ited by azide, oligomycin, vanadate, or ouabain. Instead, dimethyl assay Abbreviation: Kdp-ATPase, K+-dependent ATPase. The publication costs of this article were defrayed in part by page charge tWe follow the classical nomenclature; the more recently proposed payment. This article must therefore be hereby marked "advertisement" categorization by Pedersen and Carafoli (1) is indicated in paren- in accordance with 18 U.S.C. §1734 solely to indicate this fact. theses.

7972 Biochemistry: Bowman et al. Proc. Natl. Acad. Sci. USA 85 (1988) 7973 mixture did not exceed 1%, and control samples without bafilomycin contained dimethyl sulfoxide. Bafilomycin A1 was present in the assay mixtures at the concentrations given in the figures; the incubation volume was 0.5 ml throughout. 0 80 Protein was determined as described in ref. 40 or 41. z o H + translocation by the vacuolar ATPase ofN. crassa was LL 0 monitored by the fluorescence quenching of quinacrine in a 60- 2-ml volume containing 0.01 mM quinacrine, 1 mM MgSO4, 0.1 mM Na3VO4, 100 mM KCI, bafilomycin Al (as indicated in Fig. 4), vacuolar membrane vesicles (-30 gg of protein), ~40- and 25 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid NaOH buffer (pH 7.5). The reaction was initiated with the addition of 1 mM Na2ATP (pH was adjusted to 7.5 with 20- Tris base) and the reaction was reversed with 2.5 IxM nigericin. Quinacrine fluorescence was measured with exci- 0~~~~~~~~~~~~~~~~~ tation and emission wavelengths of 420 and 496 nm, respec- 0.001 0.01 0.1 1.0 10 100 1000 tively. CONCENTRATION OF BAFILOMYCIN A1 PM Materials. The bafilomycins were gifts from H. Zahner (University of Jubingen, F.R.G.) and Bayer (Wuppertal, FIG. 2. Effect of bafilomycin Al on the activity of F1F0, E1E2, F.R.G.). Quinacrine and nigericin were purchased from and vacuolar ATPases. Assays were performed with ATP synthase Sigma. from E. coli (m) (3.3 ug; 7.9 ,umollmg- '-min'), Kdp-ATPase from E. coli (0) (90 Mig; 0.3 ,umol mg- '-min), Na+ ,K+-ATPase from ox brain, (A) (4.4 jig; 8.6 Amol-mg - '-min -), vacuolar membranes from RESULTS N. crassa (e) (5 tug; 3.3 tmol mg 'lmin-') and bovine chromaffin granule membranes (A) (20 Mxg; 0.28 tmol-mg-'-min -). Numbers in When different types of membrane ATPases were treated parentheses refer to the amount of protein present in the assay with bafilomycin Al, whose structure is shown in Fig. 1, a volume of0.5 ml and to the 100%o values ofspecific ATPase activities. wide range of sensitivities was observed. The measurements revealed a strict dependence of 150 values on the ratio of total The 150 for the membrane-bound form ofthis enzyme was 100 amount of inhibitor to protein present in the assay mixtures Amol/mg; for the purified enzyme assayed in the presence of (for example, see Fig. 3). To standardize the results obtained added phospholipids, it was 910 Amol/mg; and for the purified by the different assay procedures we refer to I50 values as enzyme assayed in the absence ofadded phospholipids, it was ,umol of bafilomycin Al per mg of protein giving 50% 150 ,umol/mg (data not shown). Similarly, bafilomycin Al at inhibition of ATPase activity. concentrations up to 1 mM (30 Amol/mg) failed to inhibit the The ATP synthase (F1Fo type) from E. coli was not affected solubilized or reconstituted vanadate-sensitive ATPase of at concentrations of bafilomycin Al up to 1 mM (equivalent Streptococcusfaecalis (M. Solioz, personal communication). to 150 /Lmol of bafilomycin A1 per mg of protein) (Fig. 2). In contrast, the K +-ATPase of the Gram-positive thermoaci- Similar results were obtained with the ATP synthase from dophile Bacillus acidocaldarius is at least moderately sensitive mitochondria (N. crassa; Table 1). to bafilomycin Al. Although this enzyme exhibits a large By contrast, ATPases of the E1E2 type, the Na',K+- structural and functional similarity to the E. coli Kdp-ATPase ATPase from ox brain (Fig. 2), and the Ca2+-ATPase from (ref. 42; J. Hafer, A.S., and E. Bakker, unpublished results) sarcoplasmic reticulum (Table 1) showed intermediate sen- the determined I50 value of8.7 Amol/mg is considerably higher sitivities with 150 values between- 1 and 4 gmol/mg. The (J. Hafer, personal communication). Kdp-ATPase was somewhat unusual. Although the solubi- The most striking result, however, was found with the lized enzyme resembled the other E1E2 ATPases in its vacuolar ATPases (Fig. 2, Table 1): the membrane-bound sensitivity to bafilomycin (15 = 0.24 ,umol/mg), the mem- ATPases from Neurospora vacuoles, chromaffin granules, brane-bound form of the enzyme was >100 times less and corn vacuoles exhibited high sensitivity to this antibiotic sensitive. Interestingly, the Neurospora plasma membrane with Io values of <0.5 x 10' I3mol/mg. Essentially ATPase was considerably less sensitive to bafilomycin Al. identical values were obtained with the Neurospora vacuolar

OH 0 OCH3 CH3 CH3 NH - A H OH3 OH OH OH 0 0 5 7 13 1 6 H3CA ' OH3C3'. OH H31 OH 0 O-H3 H3C O 5 H3C C 4 H3C 0 7 '' '" OH3 HO CH3 CH3 OCH3 CH3 H3 HO CH3 CH3 OCH3 CH3 Bafilomycin A1 Bafilomycin B

0 HO - - OCHH3 CH3 CH3 OCH3CH3 CH3 O0 OH ' OH H °2 OH 0 OH3 OH 0 OH 0 CH3 13 aS 5 1 H3CH-SOI 23 17SJ~15 11'.'.9 H3C W1-'. OH3 HO OH3 CH3 OCH3 CH3 CH-3 CH OH CH3 OCH3 CH3 Bafilomycin C1 Bafilomycin D FIG. 1. Chemical structures of bafilomycins (25). 7974 Biochemistry: Bowman et al. Proc. Natl. Acad. Sci. USA 85 (1988) Table 1. Sensitivity of different ATPases to bafilomycin Al ATP Nigericin Type of ATPase 1 value, ,umol/mg F1FO Bacterial membrane ATPase (E. colt) NI Mitochondrial ATPase (N. crassa) NI E1E2 IAF10l% Kdp-ATPase (E. coli) 0.24 Na',K+-ATPase (ox brain) 3.9 Ca2+-ATPase (sarcoplasmic reticulum vesicles) 0.94 H +-ATPase (N. crassa plasma membranes) 100 FIG. 4. Inhibition of H + pumping by the vacuolar ATPase of N. Vacuolar crassa by bafilomycin A1. H + pumping activity was determined by Vacuolar membranes (N. crassa) 0.4 x 10 fluorescence quenching of quinacrine with vacuolar membrane Chromaffin granule membranes 0.05 x 10-3 vesicles (27 ,ug of protein; specific ATPase activity of 3.2 gmol- (bovine adrenal medulla) mg - min-'). Bafilomycin Al was added to assay mixes at the Vacuolar membranes (Z. mays) 0.004 x i0- concentrations (0-10 nM) indicated in the figure. NI, not inhibited at 1 mM (150 Amol per mg ofprotein), the highest concentration tested. H'-pumping activity of the vacuolar ATPase of N. crassa. As shown in Fig. 4, H + translocation was completely ATPase in its solubilized partially purified form (data not inhibited by 10 nM bafilomycin Al. Again, the I value shown). deduced from this experiment was very low, -0.74 x 1i0 Fig. 3 shows that the effect of bafilomycin on the activity Amol/mg. We could not perform a similar experiment in E. of the Neurospora vacuolar ATPase was strictly dependent coli, testing the effect of bafilomycin Al on K+ transport, on the amount of protein in the'assay. Gel scans of vacuolar because membrane vesicles have lost the capacity to trans- membranes (32) have allowed us to estimate that the 70-kDa port K+ via the Kdp-ATPase (for reasons not yet under- polypeptide ofthe vacuolar ATPase constitutes =2.5% ofthe stood), and we, have not yet been able to reconstitute the total membrane protein. Using this assumption, together with Kdp-ATPase into liposomes. the data in Fig. 3, we can calculate that the amount of Although we have no positive results that shed light on how bafilomycin required for half-maximal inhibition of the va- bafilomycin interacts specifically with the vacuolar ATPases, cuolar ATPase is ='wmolecule ofbafilomycin per molecule of we have obtained the following information about the Neu- 70-kDa polypeptide. Although a value for the precise stoi- rospora enzyme. The inhibitory effects seen with bafilomycin chiometry is clearly premature at this time, the data are Al are not common to all polyene antibiotics. For instance, sufficient to suggest strongly that bafilomycin interacts spe- nystatin (at concentrations up to 10 tLM) had no effect on any cifically with the vacuolar ATPase. Since for the chromaffin of the three ATPases from Neurospora. Inhibition by bafi- granules (15) and the Zea mays'vacuoles (36) the amount of lomycin, which we have been unable to reverse, was not ATPases within the membranes is difficult to estimate, a competitive with ATP. Finally, bafilomycin'was unable to similar calculation cannot be made and, therefore, an expla- prevent labeling of the 151kDa polypeptide of the vacuolar nation for the low 150 values (Table 1) has to await further ATPase by radiolabeled N,N'-dicyclohexylcarbodiimide. investigations. Comparable calculations for the E1E2 ATP- With the exception of the substitution at C-21, the struc- ases give values of100 (or more) molecules ofbafilomycin per tures of bafilomycins Al, B1, and C1 are identical (Fig. 1). molecule of ATPase. Compared with these three compounds, bafilomycin D has a The inhibitory effect of bafilomycin Al on vacuolar ATP- quite different side chain at position C-15 of the macrolide ases was further substantiated by testing its influence on ring. These variabilities in structure and hydrophobicity result in different inhibitory effects with the Kdp-ATPase 0 from E. coli (unpublished results). The three Neurospora Z 10 I ATPases, in their membrane-bound forms, proved about Z equally sensitive to bafilomycins B1, C1, or D as to bafilomy- cin A1. Whether the effect observed with the Kdp-ATPase is c 8 particular for this enzyme or dependent on the specific lipid 0 LO environment must be determined. .s 6- DISCUSSION Inhibitors have played an important role in studying ion- 4 translocating ATPases. Specific inhibitors, such as vanadate and ouabain for E1E2 ATPases and azide and oligomycin for F1F0 ATPases, have served both to distinguish among groups 2 of ATPases and to identify new ATPases, including the vacuolar group (12, 17, 22). Inhibitors have also been valu- E able in analyzing the structure and mechanism of ATPases. 0 3 , , *- 0 2 4 6 8 10 For example, vanadate, which is a transition state analog of phosphate, provided a means to demonstrate the formation of VM, pg Protein a phosphorylated intermediate of the enzyme reaction cycle FIG. 3. Dependence of bafilomycin Al inhibition on protein. ofE1E2 ATPases (43), while N,N'-dicyclohexylcarbodiimide Neurospora vacuolar membranes (VM) (1, 5, or 10 wg of protein) has been used in numerous studies to investigate proton were assayed for vacuolar ATPase activity with increasing amounts translocation by ATPases (e.g., see refs. 44-46). Compounds of bafilomycin Al as in Fig. 2. The 1 values were estimated from such as fluorosulfonylbenzoyl-5'-adenosine, 7-chloro-4- these plots. nitrobenzo-2-oxa-1,3-diazole, and 2-azido ATP have been Biochemistry: Bowman et al. Proc. Nadl. Acad. Sci. USA 85 (1988) 7975 used to label specific amino acid residues, helping to identify 1. Pedersen, P. L. & Carafoli, E. (1987) Trends Biol. Sci. 12, 146- critical functional regions within polypeptides of ATPases 150. (47, 48). Finally, inhibitors that include N,N'-dicyclohexyl- 2. Hatefi, Y. (1985) Annu. Rev. Biochem. 54, 1015-1069. carbodiimide (49), oligomycin (50), and hygromycin (51) have 3. Strotmann, H. & Bickel-Sandkotter, S. (1984) Annu. Rev. provided a means for obtaining mutants of membrane ATP- Plant. Physiol. 35, 97-120. 4. Futai, M. & Kanazawa, H. (1983) Microbiol. Rev. 47, 285-312. ases. 5. Goffeau, A. & Slayman, C. W. (1981) Biochim. Biophys. Acta In this paper, we identify the first inhibitor that may be able 639, 197-223. to fulfill these roles for the vacuolar ATPases. Vacuolar 6. Serrano, R. (1984) Curr. Top. Cell Reg. 23, 87-126. ATPases from a fungus (N. crassa), a plant (Z. mays), and an 7. Jorgensen, P. L. (1982) Biochim. Biophys. Acta 694, 27-68. animal (bovine adrenal medulla) proved exquisitely sensitive 8. Sachs, G., Wallmark, B., Saccomani, G., Rabon, E., Stewart, to bafilomycin A1 at concentrations <1O' ,umol/mg. Be- H, B., di Bona, D. R. & Berglindh, T. (1982) Curr. Top. cause we now know that vacuolar ATPases are homologous Membr. Transp. 16, 135-159. to F1F0 ATPases (52, 53), it is interesting that bafilomycin Al 9. Ikemoto, N. (1982) Annu. Rev. Physiol. 44, 297-317. is not effective against the E. coli membrane or Neurospora 10. Epstein, W. (1985) Curr. Top. Membr. Transp. 23, 153-175. 11. Hugentobler, G., Heid, I. & Solioz, M. (1983) J. Biol. Chem. mitochondrial ATPase. This observation further implies that 258, 7611-7617. the effect of bafilomycin Al is highly specific for the class of 12. Bowman, B. J. & Bowman, E. J. (1986) J. Membr. Biol. 94, 83- vacuolar enzymes. 97. Although several assumptions are involved in the calcula- 13. Xie, X.-S. & Stone, D. K. (1986)J. Biol. Chem. 261, 2492-2495. tions (see Results), it appears highly probable that bafilomycin 14. Arai, H., Berne, M., Terres, G., Terres, H., Puopolo, K. & Al interacts stoichiometrically at least with the Neurospora Forgac, M. (1987) Biochemistry 26, 6632-6638. vacuolar ATPase. Thus, ifa stable labeled form ofthe inhibitor 15. Percy, J. M., Pryde, J. G. & Apps, D. K. (1985) Biochem. J. can be produced and if it proves to bind covalently, a number 231, 557-564. ofexperiments on enzyme mechanism become feasible. Based 16. Moriyama, Y. & Nelson, N. (1987) J. Biol. Chem. 262, 9175- 9180. on the broad range of organisms affected by bafilomycin A1, 17. Mellman, I., Fuchs, S. R. & Helenius, A. (1986) Annu. Rev. it is likely to prove useful in the characterization and purifi- Biochem. 55, 663-700. cation of further vacuolar ATPases. In addition, bafilomycin 18. Greville, G. D. (1969) Curr. Top. Bioenerg. 3, 1-78. Al represents the first good opportunity for selecting mutants 19. Hilpert, W., Schink, B. & Dimroth, P. (1984) EMBO J. 3, 1665- of the vacuolar ATPase in N. crassa (and other organisms). 1670. We have obtained mutant strains of N. crassa, whose growth 20. Munoz, E. (1982) Biochim. Biophys. Acta 650, 233-265. is resistant to bafilomycin Al. 21. Linnett, P. E. & Beechey, R. B. (1979) Methods Enzymol. 55, The effect of bafilomycin on E1E2 ATPases is more 472-518. Al 22. Sze, H. (1985) Annu. Rev. Plant Physiol. 36, 175-208. difficult to evaluate. Our interest in the compounds began 23. Brooker, R. J. & Slayman, C. W. (1982) J. Biol. Chem. 257, with the finding that micromolar concentrations of bafilomy- 12051-12055. cins Al, B1, C1, and Dwere all good inhibitors of the Kdp- 24. Wang, Y.-Z. & Sze, H. (1985) J. Biol. Chem. 260, 10434-10443. ATPase of E. coli and that the bafilomycins affected enzyme 25. Werner, G., Hagenmaier, H., Drautz, H., Baumgartner, A. & activity differentially in accordance with variabilities in the Zahner, H. (1984) J. Antibiot. 37, 100-117. structure of the antibiotic. We also observed that bafilomycin 26. Hensens, 0. D., Monaghan, R. L., Huang, L. & Albers- A1 was a moderately good inhibitor (1-4 umol/mg) of two Schonberg, G. (1983) J. Am. Chem. Soc. 105, 3672-3679. other E1E2 ATPases, the Na',K+-ATPase and the Ca2+- 27. Huang, L., Albers-Schonberg, G., Monaghan, R. L., Jakubas, ATPase. K., Pong, S. S., Hensens, 0. D., Burg, R. W., Ostlind, D. A., Conroy, J. & Stapley, E. Q. (1984) J. Antibiot. 37, 970-975. The effect of bafilomycin on E1E2 ATPases is not univer- 28. Schneider, E. & Altendorf, K. (1986) Methods Enzymol. 126, sal, however. The plasma membrane ATPase of N. crassa, 569-578. which belongs to this family on the basis of its structure, 29. Friedl, P. & Schairer, H. U. (1986) Methods Enzymol. 126, vanadate sensitivity, and sequence homology, but differs 579-588. from the other members of this class in that it appears to 30. Siebers, A., Wieczorek, L. & Altendorf, K. (1988) Methods pump only protons (5), was considerably less sensitive (150, Enzymol. 157, 668-680. 100 ,umol/mg) to bafilomycin A1. Similarly, in S. faecalis, 31. Bowman, E. J. & Bowman, B. J. (1988) Methods Enzymol. Solioz and colleagues (11, 54) have identified and character- 157, 562-573. 32. Bowman, E. J., Mandala, S., Taiz, L. & Bowman, B. J. (1986) ized an ATPase consisting of a single polypeptide. This Proc. Natl. Acad. Sci. USA 83, 48-52. ATPase forms an aspartyl-phosphate intermediate and is 33. Bowman, B. J., Blasco, F. & Slayman, C. W. (1981) J. Biol. inhibited by micromolar concentrations of vanadate (55). Chem. 256, 12343-12349. Based on these properties this ATPase represents the sim- 34. J0rgensen, P. L. (1974) Biochim. Biophys. Acta 356, 36-52. plest E1E2 ATPase described to date, and it is insensitive to 35. Arnold, A., Wolf, H. U., Ackermann, B. P. & Bader, H. (1976) bafilomycin A1 at concentrations up to 1 mM (30 ,umol/mg). Anal. Biochem. 71, 209-213. We conclude that the bafilomycins show promise for 36. Mandala, S. & Taiz, L. (1985) Plant Physiol. 78, 327-333. probing the structure of the Kdp-ATPase and may prove 37. Bowman, B. J., Berenski, C. J. & Jung, C. Y. (1985) J. Biol. useful for other E1E2 ATPases as well. As a relatively specific Chem. 260, 8726-8730. 38. Chen, P. S., Toribara, T. Y. & Warner, H. (1956) Anal. Chem. inhibitor of high potency, bafilomycin Al and its relatives 28, 1756-1758. appear excellent candidates for probing the structure and 39. Schoner, W., von Ilberg, C., Kramer, R. & Saubert, W. (1967) function of the least understood category of membrane Eur. J. Biochem. 1, 334-343. ATPases, the vacuolar ATPases. 40. Dulley, J. R. & Grieve, P. A. (1975) Anal. Biochem. 64, 136- 141. We thank Prof. Dr. H. Zahner (UniversitAt Tubingen) for suggest- 41. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. ing the use of bafilomycins, Gabriele Dirkes for excellent technical (1951) J. Biol. Chem. 193, 265-275. assistance, and Johanna Petzold for typing the manuscript. This work 42. Bakker, E. P., Borchard, A., Michels, M., Altendorf, K. & was supported by Research Grant GM-28703 from the National Siebers, A. (1987) J. Bacteriol. 169, 4342-4348. Institutes of Health to B.J.B. and E.J.B. and by the Deutsche 43. Cantley, L. C., Jr., Cantley, L. G. & Josephson, L. (1978) J. Forschungsgemeinschaft (SFB 171), the Niedersachsische Minister Biol. Chem. 253, 7361-7368. fur Wissenschaft und Kunst, and the Fonds der Chemischen Indus- 44. Sussmann, M. R. & Slayman, C. W. (1983) J. Biol. Chem. 258, trie to K.A. 1839-1843. 7976 Biochemistry: Bowman et al. Proc. NatI. Acad. Sci. USA 85 (1988)

45. Fillingame, R. H. (1980) Annu. Rev. Biochem. 49, 1079- 50. Edwards, D. L. & Unger, B. W. (1978) J. Biol. Chem. 253, 1113. 4254-4258. 46. Sun, S.-Z., Xie, X.-S. & Stone, D. K. (1987) J. Biol. Chem. 262, 51. McCusker, J. H., Perlin, D. S. & Haber, J. E. (1987) Mol. Cell. 14790-14794. Biol. 7, 4082-4088. 47. Cross, R. L., Cunningham, D., Miller, C. G., Xue, Z., Zhou, 52. Zimniak, L., Dittrich, P., Gogarten, J. P., Kibak, H. & Taiz, L. J.-M. & Boyer, P. D. (1987) Proc. Nati. Acad. Sci. USA 84, (1988) J. Biol. Chem. 263, 9102-9112. 5715-5719. 53. Bowman, E. J., Tenney, K. & Bowman, B. J. (1988) J. Biol. 48. Vignais, P. V. & Lunardi, J. (1985) Annu. Rev. Biochem. 54, Chem., in press. 977-1014. 54. Furst, P. & Solioz, M. (1986) J. Biol. Chem. 261, 4302-4308. 49. Fillingame, R. H. (1975) J. Bacteriol. 124, 870-883. 55. Furst, P. & Solioz, M. (1985) J. Biol. Chem. 260, 50-52.