chains over a 275-amino-acid core stretch of similar sequence. Alignment studies revealed several regions which were highly homologous to &chains, including.
THEJOURNALOF BIOLOGICAL CHEMISTRY
Vol. 263, No. 19, Issue of July 5,pp. 9102-9112, 1988 Printed in U.5’.A.
0 1988 by The American Society for Biochemistry and Molecular Biology, Inc.
HOMOLOGY TO THE @-CHAIN OF FoF1-ATPases* (Received for publication, March 1, 1988)
Ludwika ZimniakS, Peter Dittrichs, JohannPeter Gogartenll, Henrik Kibak, andLincoln TaizII From the Biology Department, Thimann Laboratories, Universityof California, Santa Cruz, California 95064
VacuolarATPasesconstitute a novelclass of NRecently, a new class of proton-pumping ATPases has been ethylmaleimide- and nitrate-sensitive proton pumps described on the endomembranes of plant, fungal, and animal associated with the endomembrane system of eukary- cells (1-5). This proton pump is responsible for acidifying a otic cells. They resemble FoF1-ATPases in that they variety of intracellular compartments,including vacuoles, lyare large multimericproteins,400-500kDa, com- sosomes, endosomes, coated vesicles,Golgi cisternae,and posed of three to nine different subunits. Previous secretory vesicles (6-8). Termed the vacuolar ATPases, they studies have indicated that the active site is located on differ from the E1E2-typeATPases in being anion-stimulated the -70-kDa subunit. Using antibodies to the -70-kDa and insensitive to vanadate. Insensitivity to vanadate indisubunit of corn to screen a carrot root Xgtll cDNA cates the absence of a phosphorylated intermediate and library, we have isolated cDNA clones of the carrot suggests that vacuolar ATPasesare similar to FoF1-type 69-kDa subunit. The complete primary structure of ATPases. Recently, Uchida et al. (9) have provided evidence the 69-kDa subunit was then determined from the that ATP binding and hydrolysis by the vacuolar ATPase of nucleotide sequence of its cDNA. The 69-kDa subunit yeast resembles the “unisite reaction mechanism” of FoFlconsists of 623 amino acids (Mr 68,835), with no ob- ATPases. However, vacuolar ATPases are substantiallymore vious membrane-spanning regions. The carrot cDNA sensitive to nitrate and NEM’ than FOR-ATPases, while sequence was over 70% homologous with exons of a being insensitive to the FoF1-ATPase inhibitors, azide and Neurospora 69-kDa genomic clone. The protein se- oligomycin (10-13).Thus, vacuolar ATPases clearly represent quence of the carrot 69-kDa subunit also exhibited a distinct class of H+-ATPases. 34.3% identity to four representativeFoF1-ATPase BThere are structural resemblances between the vacuolar chains over a 275-amino-acid core stretch of similar and FoF1-ATPasesas well. The vacuolar H’-ATPase of plants sequence. Alignment studies revealed several regions and fungi is a large, multimeric protein, 400-500 kDa, conwhich were highly homologous to &chains, including sequences previously implicated in catalytic function. sisting of at least three major subunits: -70 kDa, -60 kDa, This provides definitive evidence that the vacuolar and -16 kDa (13-18). In animal cells, a number of additional ATPase is closely related to theFoF1-type ATPases.A potential subunitshave been identified; the ATPases of coated major functional difference between the 69-kDa and vesicles and chromaffin granules have been reported to consist &subunits is the location of 3 critical cysteine residues: of five to nine different polypeptides, including 116,100, -70, two in the putative catalytic region (Cys-248Cysand 58,40,38,34,33,19, and-17 kDa (19-21). These polypeptides 256) and one in theproposed Mg2+-binding site(Cys- were shown to copurify with reconstitutable ATP-driven proton-pumping activity (22-24). The source of the discrepancies 279). These cysteines (and two others) probably account for the sensitivity of the vacuolar H+-ATPase to in the apparent subunit composition of plant, fungal, and the sulfhydryl reagent, N-ethylmaleimide. It is pro- animal vacuolar ATPases remains to be determined. Despite the uncertainty regarding the number of subunits, posed that the two ATPases may have arisen from a common ancestor bythe insertion or deletion of a large all vacuolar ATPases examined thus far contain the70-, 60-, and 16-kDa polypeptides. The 16-kDa polypeptide has been stretch of nonhomologous sequence near the aminoshown to bind DCCD (13-18), and it is highly hydrophobic terminal endof the subunit. since it can be extracted with organic solvents (17, 23, 25). * This research was supported by Grant DMB-8517168 from the Recently its function as a proton channel has been demonNational Science Foundation and by Grant DE-FG03-84ER13245 strated in reconstitution studies (25). from the Department of Energy. The costs of publication of this There is increasing evidence that the 70-kDa subunit conarticle were defrayed in part by the payment of page charges. This tains thecatalytic site. The covalent inhibitor, NBD-C1, which article must therefore be hereby marked “aduertisement” in accordbinds to the @-subunit of FoF1-ATPases, also binds the 70ance with 18 U.S.C. Section 1734 solelyto indicate this fact. The nucleotide sequence($ reported in thispaperhas been submitted kDa subunit of the vacuolar ATPase in an ATP-protectable totheGenBankTM/EMBLDataBankwith accession number(s) manner (13, 14, 16, 21,26). The 70-kDa subunit is also J03769. preferentially labeled by NEM (22, 27). Moreover, antibody j Current address: Dept. of Biochemistry, University of Texas, to the70-kDa subunit strongly antagonized both ATP-hydro-
Houston. § Permanent address: Botanisches Institut, Ludwig-Maximilians Universitat Munchen, Menzingerstrasse 67,D-8000, Munchen 19, West Germany. 7 Recipient of a fellowship from the Deutsche Forschungsgemeinschaft. 1) To whom correspondence and reprint requests should be addressed.
The abbreviations used are: NEM, N-ethylmaleimide; BzATP, 3O-(4-benzoyl)benzoyladenosine 5“triphosphate; DCCD, N,N’-dicyclohexylcarbodiimide; NBD-Cl, 7-chloro-4-nitrobenzo-2-oza-1,3-diazole; IPTG, isopropyl-1-thio-0-D-galactopyranoside; kb, kilobase; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis.
9102
Homology between Vacuolar ATPases and FoFl-ATPases lytic activity (14) and proton-pumping activity (28). Although the 57-kDa subunit of beet was shown to bind 3-044-benzoy1)benzoyladenosine 5"triphosphate (BzATP),BzATPbinding was not competitive with MgATPase activity (15). These findings suggest that the 60- and 70-kDa subunits of the vacuolar H'-ATPase may be functionally equivalent to the a- and0-subunits, respectively, of the FoF,-ATPases. Until now, however, there has been no direct evidence that the vacuolar ATPases and FoF1-ATPases are related evolutionarily. We have now cloned the cDNA for the 70-kDa subunit of carrot,and from the DNA sequence we have derived its primary structure. A computer search of existing DNA and protein sequence libraries, statistical analyses, and alignment studies have now provided definitive evidence that the 70kDa subunit isclosely related to the 0-chain of FoF1-ATPases. Features were identified in the 70-kDa sequence which provided clues not only to theenzyme's function, but also to the possible mechanism of evolutionary divergence of the vacuolar and FoF1-type ATPases. MATERIALS ANDMETHODS
Isolation of cDNAClones-A carrot (Daucus carota) root cDNA library in Xgtll was obtained from Dr. Joseph Ecker, University of Pennsylvania. The library, divided into three size classes, 0.5-1.0, 12, and 2-6 kb, was screened according to the method of Mierendorf et al. (29) with rabbit polyclonal antibodies prepared against the purified 72-kDa subunit of the tonoplast H'-ATPase of corn (Zen mays L.) coleoptiles (14). Approximately 1 X 10' Xgtll clones with 0.5-1.0-kb inserts were analyzed, but no positive clones in this size class were identified. Upon screening approximately 6.6 X lo' clones in the 1-2-kb size class, seven positive plaques were detected and, after several rounds of rescreening, three positive clones were isolated. The insert sizes of these clones (numbers 4, 5, and 6), estimated by agarose electrophoresis after EcoRI digestion, were 1.2, 1.1, and 1.5 kb, respectively. All of them contained BamHI and HindIII sites a t identical locations, suggesting that they contained overlapping fragments of cDNA. To obtain a full length cDNA insert, the fraction of the library containing 2-6-kb inserts was screened. Of approximately 1 X 10' Xgtll clones, 11positive plaques were identified and, after subsequent rescreening, five were isolated. All of these clones reacted positively with the antibody to the denatured, purified 72-kDa subunit as well as with antibody raised against the native enzyme (14). EcoRI digestion of X phage DNA from three of these clones (numbers 71,72,and 74) produced two fragments: a 2-kb fragment, common to all three clones, and a second fragment of varying length. The length of the second fragment was approximately 0.45 kb for clone 71, 0.3 kb for clone 74, and 0.2 kb for clone 72. The restriction pattern of the 2-kb fragment was the same for all three clones (Fig. 3). From the restriction map it is clear that clones 4, 5, and 6 represent a part of the 2kb fragment. Thus, all of the clones selected by the antibody contained regions of the same sequence. Subcloning of Phage Inserts-Recombinant phage DNA was prepared according to Helms et al. (30). The inserted cDNA was separated from EcoRI-digested X DNA by electrophoresis on l%agarose gel, eluted from the gel by electroelution onto DEAE-cellulose membranes (Schleicher & Schuel NA-45) or the Gene Clean procedure (Bio 101, San Diego, CA), and ligated to theEcoRI-digested SK M13+ Bluescript vector (Stratagene, San Diego, CA).Competent E. coli J M 109 cells were transformed with ligated DNA and the plasmid DNA was prepared by a miniprep procedure from white colonies according to theboiling method described in Maniatis etal. (31). Restriction Maps and DNA Sequencing-The location of restriction endonuclease cleavage sites in the inserted DNA was determined by digesting either recombinant Xgtll DNA or recombinant plasmid DNA with several restriction enzymes under the conditions recommended by the manufacturer. The DNA sequencing analysis was performed by the dideoxynucleotide chain termination method (32) modified by use of [35S]deoxyadenosine5'-thio)triphosphate (33). Bluescript vector containing the appropriate inserts were used as templates. The primers used were either provided by Stratagene or were custom synthesized, and extension of primers was carried out with the Klenow fragment of DNA polymerase I underthe conditions
9103
recommended by Stratagene for double-stranded DNA sequencing, with the following modifications: the annealing and sequencing reaction buffer was 10 mM Tris, pH 7.5,5mM MgC12,5 mM dithiothreitol, and 1 mM EDTA; sequencing reactions were performed a t 50 "c. Preparation ofAntibodies against the Corn Tonoplast H+-ATPmeRabbit polyclonal antibodies against the purified, SDS-denatured 72kDa subunit and against the native, sucrose gradient-purified enzyme were prepared as previously described (14). Before use in screening the antibody was diluted 1:500 and preabsorbed with crude E. coli extract (Promega, Madison, WI). Antibody Selection by Plaque Affinity Purification-Nitrocellulose filters with expressed Xgtll proteins of positive clones 4, 5, and 6, and a negative clone as a control, were treated with antibody as in the screening procedures, and theunbound antibody was removedby washing. The positive control was a filter coated with a crude corn vacuolar ATPase preparation (dextran step gradient fraction) (34). Bound antibodies were eluted from the filters by treatment with 0.2 M glycine-HC1,pH 2.5, for 2 min (30, 35). Following neutralization with 1.0 M phosphate buffer, pH 9.0, and dilution to lower ionic strength, the eluted antibodies were used to probe Western blots of corn tonoplast ATPase solubilized by octylglucoside (34). Isolation of Fusion Protein-Lysogenic E. coli (Y1089) were prepared from Xgtll clone 74 according to the procedure given by Mierendorf et al. (29). Growth of the lysogenic bacteria, induction of fusion protein synthesis, and harvest of the bacteria were performed as described by Huynh et al. (35). Bacteria from 100 ml of culture were resuspended in 1 ml of TEP (100 mM Tris/HCl, pH 7.4,lO mM EDTA, 1mM phenylmethylsulfonyl fluoride) and lysed by four freezethaw cycles (liquid Nz/ice water). The viscosity of the lysate was reduced by sonication (Branson Model W185,10% duty cycle). Sample preparation for electrophoresis, SDS-PAGE, and western blots were performed as described by Mandala and Taiz (34), but with a separating gel of 8% polyacrylamide and 10 mM dithiothreitol in the sample buffer. Bound antibodies against the 70-kDa subunit were visualized by biotinylated secondary antibodies and avidin-coupled biotinylated peroxidase (Vectastain Kit, Vector Labs, Burlingame, CA). Preparation of Tryptic Fragments-Sodium deoxycholate-washed, tonoplast-enriched membranes from corn coleoptiles were isolated and their proteins separated by SDS-PAGE as described previously (34), except that 0.1 mM sodium thioglycolate was included in the cathode reservoir. Protein (125 pg) wasloaded into each of nine lanes. Electrophoresis was continued until the prestained myoglobin marker entered the anode reservoir. After staining with Coomassie Blue and destaining in 10% glacial acetic acid, the 70-kDa band was excised with a scalpel. Each gel slice was estimated to contain 15 pg of protein by comparison to bovine serum albumin standards. Control slices from the unstainedportion of the gelwere excised and treated identically in subsequent steps. The tryptic digest was carried out as follows. After a 30-min wash in distilled water, nine gel slices (-135 pg) were cut into approximately 1-mm cubes and homogenized in a small glass homogenizer with a Teflon pestle in 2 ml of 50 mM Tris/HCl (pH 8.0). Trypsin (1 pg) was added to thehomogenate and incubated while shaking at 250 rpm, 37 "C. After 6 h, the digest was supplemented with 0.5 pg of additional trypsin, and theincubation was allowedto continue for 12 more hours. The final digest was then centrifuged for 30 min at 12,000 X g in a Sorvall RC-2B centrifuge. The digest was complete when the supernatant turned blue from the released Coomassie dye. Thesupernatant containing the tryptic peptides was passed through a 0.45-pm Nalgene filter and lyophilized. The bluish residue was dissolved in 100 p1 of deionized water (HPLC grade) and the peptides were separated by reversed phase high pressure liquid chromatography in 0.1% trifluoroacetic acid using a gradient of 0-80% acetonitrile on a CIScolumn (Vydac 218TP). Absorbance a t 210 nm was monitored, and 0.5-ml fractions were collected. Selected fractions containing peaks present in the 70-kDa digest but absent from the control were taken to the Biomolecular Resource Center, University of California, San Francisco, for amino acid sequence analysis. RESULTS
Characterization of the Fusion Protein-The molecular mass of the IPTG-inducible fusion protein for a positive clone containing an insert of 2.1 kb wasdetermined to be -188 kDa (Fig. 1). After subtracting P-galactosidase from the total, the insert encoded a protein with a molecular mass of -70 kDa.
Homology between Vacuolar ATPasesand FoFl-ATPases
9104
I PTG
1
+ 200.
-188 kDa
72K-
2
3
4
5
-
116. 92 1 661
4 51
FIG.2. Western blot of solubilized corn tonoplast proteins separated by SDS-PAGE and probed with plaque affinitypurified antibodies to the 70-kDa subunit. Lane 1 (control), antibody eluted from crude ATPase preparation; lane 2, antibody eluted from positive clone 4; lane 3, antibody eluted from positive clone 5; lane 4, antibody eluted from positive clone 6; lane 5,antibody eluted from a negative clone.
clone, number 71, containing a 2.4-kb insert, and is presented in Fig. 4. The start sites of the shorter clones are indicated kDa by the arrows. Partial sequencing of the shorter clones indicated that they were identical, except for the difference in length of the 5’ region. FIG.1. Western blot of 10 pl of E. coli Y1089 lysate conThe sequence contains a 155-nucleotide 5”untranslated taining Xgtll clone 74 and separated by SDS-PAGE. +/- sequence, a 1869-nucleotide open reading frame, and 373 IPTG denotes parallel samples with or without induction of the fusion nucleotides of 3’-untranslated sequence, followedby the protein by addition of IPTG. Molecular mass values correspond to the locations of Coomassie-stained protein standards (Bio-Rad). The poly(A) tail. The5”untranslated region of the cDNA is arrow indicates the location of the antibody-binding, IPTG-inducible enriched in AT (-60%) and the translation initiation site occurs at thefirst AUG codon in the sequence, features which p-galactosidase fusion protein. are characteristic of plant mRNAs (36). The triplet, ATC, Fig. 2 shows the results of an antibody selection experiment, was found to be abundant in the 5’-untranslated region which tests the ability of the carrot fusion protein to bind (underlined), as well as one direct repeat (I, I ’) and two specifically the antibody which recognizes the purified 70- inverted repeatsequences (u,u’ and b,b’). These featuresmay to be important for the secondary structure of the 5’ region. kDa subunit on Western blots. Plaqueproteinsbound nitrocellulose filters were used to affinity purify the 70-kDa Two TATA boxes, normally the site of RNA polymerase antibodies, and theeluted antibodieswere then used to probe binding to the promoter region of the gene, were located 92 Western blots of total tonoplast membrane proteins. Lune 1, and 105 nucleotides upstream from the initiation codon. Both the positive control, shows that antibodies elutedfrom a crude TATA sequences were absent from the two shorter clones, 72 and 74. The 3”untranslated region also contained several vacuolar ATPase preparation of corn specifically recognized inverted repeats (underlined).The polyadenylation signal, Athe corn 70-kDa polypeptide on a Western blot. Lunes 2-4 A-T-A-A-A (37), was located 26 nucleotides upstream from were probed with antibodies elutedfrom filters containingthe the poly(A) tract. Thepoly(A) sequence was directly preceded plaque proteins of three of the positive clones, numbers 4, 5, by a T-rich region. and 6, respectively. In all three cases, the antibodies eluted Nucleotide Sequence Homology with Other ATPases-The from the positive plaque proteins specifically bound to the70- nucleotide sequence of the D. curotu 69-kDa cDNA clone was kDa band. In contrast, lune 5 shows that antibodies eluted compared with the recently sequenced genomic clone of the from one of the negative clones failed to bind the 70-kDa 70-kDa subunit of the Neurospora vacuolar ATPase.’ A Puspolypeptide. tell matrix analysis (38) of the two nucleotide sequences is Nucleotide Sequence of the 70-kDu cDNA-Fig. 3 shows the shown in Fig. 5. Stretches of 15-nucleotides were compared, restriction maps of the positive cDNA clones and thesequencB. J. Bowman and E. J. Bowman, personal communication. ing strategy. The sequence was determined from a single
Homology between Vacuolar ATPases and Fg1-ATPases
9105
.. .
c!
*#5
a 0 E
.-
X
FIG. 3. Restriction map and sequencing strategy for cDNA inserts of isolated clones. Restriction endonuclease cleavage sites are indicated by vertical lines and numbered starting from the 5’ end of the DNA. The 5’ ends of the six positive clones are indicated by the arrows above the map.
and sequences less than 70% identical were rejected. Remarkable sequence similarity between the cDNA of carrot and the genomic DNA of Neurospora was observed, in some regions as high as 85%. Several breaks in the diagonal were present corresponding to introns in the Neurospora genomic DNA. The close similarity between the independently cloned DNA sequences of the two highly divergent species is convincing evidence that they both encode the same polypeptide. We also searched for sequences similar to the70-kDa cDNA using the current complete EMBL DNA sequence library. The search was conducted for us by Guenter Stoesser at Heidelberg. Out of a total of 15,344 entries, the computer program (WORDSEARCH) selected the 50 most similar sequences. Of the 50 sequences selected, 46 could be identified definitely as ATPases or sequences containing ATPase subunit genes. Representative examples are shown in Table I. These findings confirm previous suggestions (39) that the ATPases represent a large family of related proteins. Predicted Amino Acid Sequence of the 70-kDa cDNA-The predicted amino acid sequence of the carrot 70-kDa cDNA is shown in Fig.4. The protein consists of 623 amino acid residues and hasa predicted molecular mass of 68,835 daltons. To confirm that the carrot cDNA sequence corresponds to the 70-kDa subunit previously characterized in corn, a 12amino-acid tryptic fragment of the corn 70-kDa subunit was sequenced and compared with the deduced amino acid sequence of the carrot cDNA. The tryptic fragment was identical to the carrotsequence from Leu-223 to Arg-234, except that tyrosine was substituted for Gly-232. This discrepancy may be an artifact, since Gly-232 is highly conserved. Based on computer analysis (CHARGPRO, PCGENE), the theoretical isoelectric point was determined to be 5.07. This is comparable to the isoelectric point of -5.1 determined on Western blotsof two-dimensional gels? Hydrophobicity plots (Fig. 6) were calculated with a program from PCGENE, which utilizes the method given by Kyte and Doolittle (40). The protein lacks a hydrophobic leader sequence and contains no membrane-associated a-helices, based on the method of Eisenberg et al. (41). The 69-kDa subunit, like the 8-chain, appears to be a peripheral membrane protein. Homology of the 69-kDa Subunit to the P-Chain-When we searched the National Biomedical Research Foundation protein library (PIR, release 14 from September 30, 1987) for similarities to the69-kDa subunit using the FastPalgorithm as described by Lipman and Pearson (42) and theparameter ktup = 1 (lookup table for single amino acids), the only proteins that gave alignment scores greater than 90 (initial and optimized) were the a- and 8-subunits of the various FoF1-ATPases. The mean score (&.D.) for the whole data W. Hurkman, F. M. Dupont, S. Mandala, and L. Taiz, unpublished data.
base compared to the 70-kDa subunit was 30.1 ( k 7.41). The sequence with the highest optimized alignment score was the 8-subunit of the CFoF1-ATPase from spinach (score = 303); however, if this sequence was randomized the score (&S.D., n = 20) dropped to 43.2 ( k 9.56). Thus, for this comparison the z-value is 27.2 (i.e.the distance between the score of the actual sequences and the mean score of the randomized versions expressed as multiples of the S.D.). According to Lipman and Pearson (42), z-values > 10, derived from the FastPalgorithm, can be regarded as significant similarities. In order to show that this similarity is not only due to the similarity of enzymatic function, i.e. ATP binding and hydrolysis, but reflects an evolutionary relationship, a more thorough comparison was carried out with otherATP-binding and hydrolyzing proteins employing the Needleman-Wunsch algorithm, which aligns the whole sequence (Table 11) (43). When 8-chains of FoFl-ATPases were compared with each other, the z-values ranged from 76 to 132. Since the comparison of the two 69kDa subunits from corn and Neurospora gave a z-value of 105, the 69-kDa subunit is as conserved as the 8-chain. The only other proteins showing significant overall peptide sequence similarity to the 69-kDa subunits were the FoF1ATPase a- and @-chains.Note that z-values > 3 are regarded as significant for procedures aligning whole sequences (44). The vacuolar 69-kDa subunit is more similar to the 8- than the a-subunit, especially if conservative amino acid substitutions are given positive scores (Table IIB). Alignment Studies-An alignment of the carrot 69-kDa subunitwith several 8-chains and an a-chain of FoF1ATPases is shown in Fig. 7. In general, the 69-kDa subunit showed significant homology to the @-chainsfrom approximately Thr-73 to Pro-128 andfrom Leu-229 to Ala-515. There was little similarity at the amino and carboxyl termini and none along the 98-amino-acid segment, Leu-130 to Pro-228. This long stretch of internal nonhomologous sequence is labeled “N” inFigs. 4,6, and7. The percent identical matches to the 8-chains along the 286-amino-acid core stretch of homology from Leu-229 to Ala-515, ranged from 31.8% for yeast mitochondria to 29.4% for E. coli. The percent identical homology to the a-chain over the same stretch was lower, about 19% in E. coli where the subunit shows some homology from approximately Thr-10 to Thr-150 and from Thr-231 to Gln-497. The chloroplast and mitochondrial a-chains typically show homology through Lys-470. In addition to the overall homology, a number of specific regions were highly conserved, especially in the case of the 8chain. Moreover, these regions of high homology, labeled AF i n Figs. 4 and 7, wereconserved with respect to organization along the sequence, and most were located in regions already identified in FoF1-ATPases as areas of probable importance for function or assembly. The most important of these are the
9106
Homology between Vacuolar ATPases and FoFl-ATPases
9107
the @-chain.Ser-373 of the a-chain,which has been shown to be essential for activity in E. coli (51), is conserved in the 69kDa subunit. The amino terminal end of the 69-kDa subunit was not nearly as well conserved with respect to a- and @-chainsas the centralregion, the firstsignificant homology beginning at P a z Thr-73 and defined as region A. When the first 145 amino n 0 .acids were used as a query sequence against the PIR protein sequence database using the FastPalgorithm, it still returned 1000a- and @-subunits,with a preferred slightly over @.Interest.? ! ingly, when the same search was conducted using the next stretch of 82 amino acids, Pro-146 to Pro-228, no ATPases 24 1500were identified. This finding, and thepresence of the homol9 0 ogous sequence at region A in both carrot and Neurospora, .r were the bases for introducing the gap in the alignment found w in the nonhomologous region labeled N in Figs. 6 and 7. The carboxyl terminus of the carrot 69-kDa subunit also appeared to lack homology to other ATPases, based on a FastP search I of the final 66-amino-acid residues. FIG. 5. Comparison of carrot cDNA and Neurospora geIt is interesting to compare the hydropathy plots of the 69nomic DNA nucleotide sequences for the 70-kDa subunit of kDa subunit and a @-chain in the homologous regions, with the vacuolar H+-ATPaseusing the matrixcomparison method the nonhomologous region removed (Fig. 6, B and C ) . The developed by Pustell (38).Letters indicate regions where homology is 270%. P I:71%; N 5 75%; M 5 77%; K I:81%;J 5 83%; Z 5 85%. putative active site (region B) spans both hydrophobic and The Neurospora genomic DNA sequence was from unpublished re- hydrophilic zones in the @-chain,while the homologous region sults of Drs. B. J. Bowman and E. J. Bowman, University of California of the 69-kDa subunit corresponds to a large peak of hydroSanta Cruz. phobic sequence. Regions C-F are correlated with zones of a hydrophobicity inbothsubunits.Bothsubunitscontain TABLEI mixture of hydrophobic and hydrophilic sequences in region Representative sequences selected in a search of the EMBL DNA A, although hydrophobic sequences predominate. In general, database far homokxv to the 70-kDa subunit the conserved regions tend to be hydrophobic. The average ATPase Oreanism twe Subunit index of hydrophobicity was -0.07 for the @-subunit,-0.21 for the 69-kDa subunit, and -0.21 for the 69-kDa subunit Yeast a, 8 Mito.,” F-type minus the nonhomologous region. Thus, the nonhomologous N . crassa 6 Mito., F-type Tobacco Chloro., F-type a,P, I, 111 region doesnot alter the overall hydrophobicity of the 69-kDa Yeast -100 kDa H+, P-type subunit. N. crasa -100 kDa H+,P-type Rat Na+/K+,P-type a,B DISCUSSION E. coli Kdp, P-type A, B,C -100 kDa Rabbit Ca2+,P-type The cDNA clone of the carrot vacuolar ATPase 70-kDa Mito., mitochondria; Chloro., chloroplast; F-type, FoFI; P-type, subunit which we sequenced in this study was obtained by antibody screening of a carrot root cDNA library cloned in &E,. Xgtll. Proof of the identity of the sequence was derived from putative catalytic site in region B, which is the site of NBD- several lines of evidence. First, all of the clones selected by C1 binding (45, 46), and the proposed Mg+-binding region as the antibody contained the same sequence. Second, the IPTGdemonstrated in thermophilic bacteria (47, 48), indicated in inducible fusion protein had the predicted molecular mass. region C. Lys-258 in region B, indicated by the double asterisk, Third, the fusion protein selectively bound antibody which is the residue which binds NBD-C1 in the @-chain.One highly recognized the -70-kDa subunit on Western blots. Fourth, significant difference between the 69-kDa subunit and the @- the amino acid sequence of a tryptic fragment of the purified chains is the presence of cysteine residues in regions B and 69-kDa subunit of corn closely matched the predicted amino C. Cys-256 is located within the putative active site defined acid sequence of the cloned carrot cDNA. However,the most in region B, and Cys-248 is 7 residues upstream. Cys-279 is convincing evidence was the homology between the carrot69located in region C,which is proposed as the Me-protectable kDa cDNA and the exons of the independently isolated geDCCD-binding region. This concentration of cysteines within nomic clone of the Neurospora vacuolar ATPase 70-kDa and near the predicted ATP-binding region is probably the subunit (made available to us by our colleagues at University key factor determining the enzyme’s sensitivity to sulfhydryl of California Santa Cruz, Drs. B. J. Bowman and E. J. Bowman). This resultconfirms previous immunological crossreagents such as NEM. The most conserved sequence appears to be the one defined reactivity studies (13,14,61)which indicated that the70-kDa as region D, with 10 out of 11 residues being identical to the subunits of plant, fungal, and animal vacuolar ATPases are corresponding amino acids in the mitochondrial @-chain se- highly conserved. Previous estimates of the molecular mass of the vacuolar quence of yeast. Region D and the ll residues downstream have also been proposed as a nucleotide-binding region based ATPase “70-kDa” subunit were based on SDS-PAGE and on phylogenetic studies (49, 50). A region of the 69-kDa ranged from 67 to 89 kDa, depending on the organism (8). subunit showing homology with the phosphorylation site of The molecular mass of the carrot70-kDa subunit determined E1E2-ATPaseis shown in region E, although it is missing the from its cDNA sequence was 68,885daltons. A more accurate aspartic acid residue (indicated by an asterisk) that becomes designation is therefore the “69-kDa subunit.” phosphorylated during catalysis. Region F of the 69-kDa Both structurally and functionally vacuolar ATPases resubunit showed greater homology with the a-chain thanwith semble FoF1-ATPases. Their molecular mass is around 400Carrot 69kDa Subunit cDNA Sequence
500
1000
1500
N
NN
E
NY
ZOO0
2410
9108
Homology between VacuolarATPases and F&'l-ATPases 38
FIG. 6. Comparison of the hydrophobicity plots of the complete carrot 69-kDasubunit ( A ) ,the spinach chloroplast &subunit (23) (63),and
I
-I
"I
38
b
the carrot 69-kDasubunit with the nonhomologous region (region N) removed (C). Stretches of 11 amino acids were analyzed according to Kyte and Doolittle (40) using the program SOAP (Intelligenetics) with hydrophobic regions given positive values. The dashed line marks the predicted division between interior (hydrophobic) and exterior (hydrophilic) portions of the protein.The bars correspond to the conserved regions in Fig. 4.
AMINO ACID RESIDUE 500 kDa; they consist of three to nine different subunits; Rea et al. (62) have recently shown that the 70- and 60-kDa subunits of the vacuolar ATPase of red beet can be specifically dissociated from the membrane as a water-soluble complex similar to the F1-complex using chaotropic anions; and the smallest subunit (15-17 kDa) is an integral membrane proteolipid which functions as a proton channel,similar to subunit cof the Fo-complex (reviewedin Refs. 1-4,8).Inhibitorbinding and immunological studies have suggested that the 69-kDa subunit contains the active site. The 69-kDa subunit binds the adenine analog, NBD-C1, in an ATP-protectable manner and it is immunologically highly conserved (13, 14, 16, 21, 22, 61). In addition, antibodies to the 69-kDa subunit specificallyblock H+-ATPase activity (14, 28). Thus,the vacuolar ATPase 69-kDa subunit and the B-chain of FoF1ATPases arefunctionally analogous. The present study establishes for the first time that the two proteinsare clearly homologous. Computer searches of the current PIR protein sequence database using the FastP algorithm indicated that FoFIATPase@-chainsare the closest relatives of the 69-kDa
subunit, followedby &-chains. EIEz-ATPases were not selected by this method. Two otherATP-bindingproteins, myosin kinase and adenylate kinase, appeared to be unrelated to the 69-kDa subunit using the Needleman-Wunsch algorithm. Surprisingly, at the nucleotide level a computer search identified both FoF1- and E1E2-typeATPases almost exclusively among the 50 most similar sequences. For reasonswhich remain to be clarified, the program is able to recognize the entire ATPasefamily of proteins betterat thenucleotide level than at theamino acid level. Our alignment of the 69-kDa subunit with 8- and a-chain sequences was aided by several computer alignment programs, including FastP, PEP ALIGN, GENALIGN, and MICROGENIE, but none of them identified all of the homologous regions. Problems appeared to arise from the differences in molecular mass of the two subunit types (69 kDa uersw 5055 kDa) and theinability to weight highly conserved regions. Hence, many of the alignments were obtained by eye. A guiding principle used throughout was to introduce as few gaps as possible within conserved areas. A number of highly conserved regions of the p- and a-
9109
Homology between VacuolarATPases and FoF1-ATPases TABLEI1 Needleman-Wunsch comparison of carrot 69-kDasubunit with other ATP-bindingproteins For each given z-value, the mean and standard deviationof at least 25 comparisons of the randomized sequences were calculated. The sequences used were 1) 69-kDa subunit of D. carota (D.c.);2) 70-kDa subunit of Neurospora crussa (N.c.)(E. J . Bowman and B. Bowman, personal communication); 3) (3-chain from the FoF1-ATPaseof E. coli (48); 4) @-chain from thechloroplast (chl.) ATPase of Spinacia oleracea (53); 5) (3-chain from bovine mitochondrion (mito.) (54); 6) a-chain from the FoF1-ATPaseof E. coli (55); 7) a-chain from the chloroplast ATPase of Nicotiana tabacum (56); 8) a-chain from the chloroplast ATPaseof the liverwort Marchantiapolymorpha (57); 9) rabbit adenylate kinase (58); 10) the first 700 amino acids of the nematode myosin heavy chain (59). For the results given under Part A, the unitary matrix (44) and a gap penalty of 3 was used for computation; under Part B, the structure genetic matrix (60) and a gap penalty of 6 was chosen, i.e. positive scores were ascribed to functionally similar amino acids. Vacuolar ATPases
Matrix
8-Subunits
3
2
1
a-Subunits
4
5
6
7
15.2 2.1
2.0
9.1 5.5
7.8
Other 8
9
10
A. Unitary matrix 1. 69 subunit, D.c. 2.N.c. 70 subunit,
13.8
11.3 105.0 9.6
9.7 105.0 9.4 14.1 13.7 10.0 10.4
13.6 3. 13.5 (3,E. coli 131.9 105.5 10.4 14.1 4. (3,spinach chl. 105.5 10.0 15.2 76.0 13.7 mito. 5. (3,bovine 1.8 131.9 9.6 1.5
169.7 69.1 2.6 169.7 -0.2
3.7
1.5
6. a, E. coli 72.2 69.1 9.4 15.4 14.9 7. chl. a, tobacco 10.1 13.6 7.8 9.7 8.72.2 a, Marchantia 19.9 chl. 12.8 11.3 13.8 9.1.8Adenylate 0.7kinase 2.0 10. Myosin, nematode 2.1
12.8
10.1
19.9
5.5 12.8
0.3
12.8 14.9
0.5
76.0
0.7 1.8
1.0
-0.2 0.8
2.1 1.0
15.4
13.5
9.1
3.7 0.3
1.o
0.8 0.6 1.0 0.5 2.1
1.8
2.6 0.5
0.6
B. Structure genetic matrix 7.71. 69 6.6 subunit,7.4 D.c. 12.7
15.1
14.1 106.4
chainshad homologues in the 69-kDa subunit, and their relative positions within the sequence were similar. Of particular importancewas the identification of the NBD-C1-binding site (the putative active site) in the 69-kDa sequence (region B). Theadenine analog, NBD-C1, is a potent inhibitorof both FoF1-ATPases (63) and vacuolar ATPases (9, 13, 14, 16, 26), specifically labeling the @-chain and69-kDa subunits, respectively, in a substrate-protectable manner. In beef heart mitochondria, NBD-Cl reacts very specifically and reversibly at Tyr-311 of the @-chain,before binding irreversibly to Lys-162 (63, 64). This observation has been interpreted as evidence that these 2 residues participate in ATP binding and catalysis. Lys-162 is, in fact, highly conserved in @-chainsof FoF1ATPasesas well as a diverse group of nucleotide-binding proteins. It is also conserved in the homologous sequence of the carrot vacuolar ATPase 69-kDa subunit (region B, Lys258). Site-directed mutagenesis of the equivalent residue in the E. coli @-chain(Lys-155) confirmed the essential nature of this residue for ATP synthesis andhydrolysis (45). Replacement of this residue with either glutamine or glutamate resulted in a greater than 5-fold reduction in activity. Chemical and kinetic studies of the inactivation of the vacuolar ATPase of yeast by NBD-C1 have provided evidence that thereis a single reactive tyrosine per molecule of 89-kDa subunit (9). Although the region surrounding Tyr-311 in the beef mitochondrial @-subunit appearsto be conserved in the 69-kDa subunit (region E , Fig. 7), there is no obvious candidate for homology to theNBD-C1-binding tyrosine, unless the hydroxyl of Ser-414 is sufficiently reactive to bind the inhibitor. A more likely explanation isthat NBD-Cl binding occurs by an alternative mechanism in thevacuolar ATPase, an idea already suggested by Moriyama and Nelson (26). Mutagenesis experimentsin which Tyr-297 of the E. coli @-chain was replaced by phenylalanine caused only a slight reduction in enzyme activity, but a marked increase in resistance to NBDC1 (451, demonstrating that this tyrosine is not essential for catalytic function. Interestingly, there is a tyrosine in the correct Val-Tyr-Val motif just upstream from glutamic acid
-0.1
-0.3
residues shown in the thermophilic bacterium, PS3, to be involved in the M$+-protectable labeling by DCCD,an inhibitor of both FO andF1 portions of ATP synthases (65). This stretch of sequence (region C, Fig. 7) is highly conserved and contains the Glu-181 from PS3 that binds DCCD. However, in E. coli and mitochondria the residue implicated in DCCD binding is Glu-192 (66), which falls outside the conserved region C and is replaced by Asp-292 in the carrot sequence. Glu-185 has been shown in E. coli (67) to be essential for structure and assembly of the beta subunit and is conserved in thecarrot 69-kDa subunit as well. The region defined by Thr-332 to Asp-342 (region D ) is exceptionally conserved in both a- and @-subunits of FoFIATPases. Since it ispresent on both a- and @-chains,and no direct evidence for nucleotide binding inthis region has emerged, it is tempting to speculate that it may be involved in some other function, such as assembly, energy coupling and/or transduction. The 69-kDa subunit contains 5 cysteine residues and is relatively rich in sulfhydryl groups compared to FoF1-ATPase P-chains, which contain no more than 1cysteine residue (4648). The location of the cysteine residues is even more significant. Three are located in the putative active site region, including two in the ATP-protectable, NBD-C1-binding region (region B)and one in the Mp-protectable, DCCDbinding region (region C). The presence of these strategically located cysteine residues probably accounts for the sensitivity of vacuolar ATPases to sulfhydryl group inhibitors, such as NEM. It has also been suggested that NBD-Cl might act as a sulfhydryl group inhibitor of vacuolar ATPases under certain conditions.' It has already been proposed that there is a residual sequence from the phosphorylation domain of EIEz-type ATPases present inthe a- and @-chains of FoF,-ATPases (39, 68). This region in the CY- and @-chains exhibitshomology to the carrot 69-kDa sequence (region E , Fig. 7). Like the FoF1N. Nelson, personal communication.
Homology between Vacuolar ATPases and Fsl-ATPases
9110
A
53 01 G a 77 @ V L @ V
A
113 138 IW ~ V ~ E ~ ~ 59 v 1 ~ ~ ~ L L ~ N ~ K ~ R A ~ A ~ ~ A ~ D ~ ~ ~ I ~ M K ~ I D ~ G A @ L I ~ ~120V ~ E A I ~ ~ ~ Q V L @ V O O O L @ G G I V ~ I I A ~ G ~ S D ~ I I G L D Q K D L E ~ ~ I E @ V V ~ C A I ~ ~ ~ M Y V L G 100 E @ V D M K G $8 I I 1 r a n L Y L E a D s v G A G v V@D@A E G M K@K I IQE@V V @ R ~ L v ~ v RY I L G A ~ D tI K G 119
V.AlVas. 6WDa S r M i t 5. tererastae mfl.Alvasc Beta S m i l OOI~PT mfl.AtPbse Beta Slrbnit m. potylorpna C f l - A l v a s ~seta $ b i t t . E L I EFl.1IPwSr Beta S r h n l l c ~ Efl-ArPose t srhmit D . carol.
”
c8
L ~ Q V L
L.
l I L E @ D @ . A I I 0 0 1 E VEY D V m R @ I I KEA 0 II L @ E Y I V I 1 I A D G @ E m V R G E KWDOG C @ l m V
~
I
O
~
L
~
E
~
~
~
I
I
I
A
M
~
G
VOW
~
E
~
V
L
~
I
~
Y
(
i
r
~
G
~
K
~
D
~
K 1 IA 0 E R c
R@L E@I
Zt(iJ0D
Y v@C
V a n E l a U @ l
G
A
~
I
~
I
COG
Y
0.c.t.9 S.c.B
K
OW.B
P I K I K O ~ A A I Y A E A ~
M.p.6
P V E V l l l f ~ l H R A A ~ f
P
S
C
~
~
~
I
~
~
G
V
S
~
~
L
~
K
D
~
L
~
E
~
~
~
~
K
I
~
E
~
D
L
L
I
G
G
D
L
T
A
I
~
..................................................... E ~ V E ~ ~ V .E .Q. E . .L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . l O L D I K L S. . l . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
E192 Y
~
~
..................................................... .....................................................
L.c.0
E l t E E E R Y A l H P A A Q S V E E L S N S 0 E L
E.c.1
P L O ~ ~ t f S I V E I I A ~ G V I E R O S W ~ O V
D.c.69
V L E L E f ~ G V K K O F I M L O l Y V V R l V U V V A S K L A A D l V ~ L ~ O R ~ L @ A ~ f ~ S V L ~ l C A ~ P Q ~. .A . fS a ~ lC ~ 2l MS O A
Y
E.c.1
D.c.69
S
S.C.I BW.I m.p.8
A ~ G G ~ ~ ~ ~ A @ V ~ I I Q ~ Y D L ~ ~ E ~ A I I G G ~ S ~ ~ A ~ V ~ I R Q G Y D L ~ ~ E ~~ I E. S L G VK I D A ~ A ~ G G Y S ~ ~ ~ ~ Y ~ I ~ Q G ~ D L ~ ~ E M K E S K V I
E.c.0
E H ~ t l S ~ f A ~ V ~ l R ~ G N D f l H E M T D S N V I D
E.C.A
O ~ G I ~ ~ I ~ A I ~ ~ K A ~ I I ~ Y ~ V I K L E E Y G A L A Y I I V ~ V A ~ A S E ~ A A L O ’ I L A I M
m.p.o E.C.I
Y
~
C D
OE OV
0 0 1 K ~
~
E
D
~
~
K
I
V
@
L
~
A
~
V
A
V
@
L
~
A
~
T
A
~
V
@
~
K
L
~
A
R
~
K
~
~
~
K K
I
R
l
V
~
C
~
t
G
L
l
G
L
~
K
I
F f
~ ~
C G
G
L
A A
F
~
~
~
~
M
A
E
~
L
~
O
f
VN N~ I Af Kl
V
~
E
A
V
O
~
V
K
A
V
~
S
~
D
G
l
~
l
G
~
~
E
~
V
G ~ I
~ .
I
O
R
..
E
L
~
l
~
D
~
D N
~
f
l
~
207 M E
~
l
M 115 E ~
l
N
N
~189L
I
~
E
V
V
A
~
I
O E ~ I C ~ I @ L ~ C O F A K ~ K R VY I ACI L 169 ~ ~ ~ A
C l
M
B
B
.................................... .................................... .................................... .................................... ....................................
S.c.0 00v.O
L
I& 132 1k6 126 145
P l K S K L R K e ) l H A D P ~ S f A E O S l S A E l
V
O
L
l
M
l
L
V
R
E ~ ~ ~ . . ~ E O ~
M
~
~ ~
R .
A
l L
l . T
~
O
~
A
Y
0 l
E G
~
~
.
~
l
~
.SM @ l~ W .
M
~ ~
~
R 197 L A
M
l
346 ~ A
O
~
A
l
l
E
A
S
. 2a1~ A ~ ~250 ~ G 265
F
G A
~
~ ~ . . ~ . V A Q ( ~ Y ~ ~ ~ ~ E ~ P G ~ ~
...........
V
S
Q
(
a
V
G
O
M
~
E
~
P
...........
G
Y
~
L
R
V
A
L
~
L
~
M
~ 237 K ~
.
A L M G ~ . I O255E D
FIG. 7. Amino acid sequence homology between part of the 69-kDa subunitof the carrot tonoplast H+-ATPase and thej3- and a-subunits of FoF1-ATPases. The less homologous a-subunit E . coli is included for comparison. Identical matches between the 69-kDa subunit residues of carrot and any of the other sequences are circled. S. cerevisiae mitochondrion @subunit (52); Bovine mitochondrion P-subunit (54); Marchntiupolymorpha chloroplast P-subunit (Ohyama, K. (1986), submitted to EMBL and GenBank in computer-readable form.); E. coli 0-subunit (48); E. coli a-subunit (55).
ATPases, vacuolar ATPases do not form phosphoenzyme intermediates during catalysis and, like the a- and p-subunits, the carrot subunit is missing the aspartic acid residue that becomes phosphorylated during catalysis in the EIEz-type enzymes (region E , asterisk). Instead of the glycine found in a-subunits and the threonine found in @-subunits,the 69-kDa subunit has glutamic a acid residue at theposition corresponding to the phosphorylation site of E,E2-type ATPases. The highly conserved nature of this region, the reversible inhibition caused by the transient binding of NBD-C1 to Tyr-311, and the possiblehomology to the active site of EIEz-type ATPases argue that this region may play an important role in catalysis. Although the function of the a-subunitremains a matterof debate, the uncA4Ol mutant of E. coli suggests that Ser-373 of the a-subunit plays an essential role in catalysis (51). It occurs in a region which is highly conserved among a-chains,
and which also shows homology to P-chains and the 69-kDa subunit (region F, Fig. 7). Within this region are 5 residues, DSTSR, found in this position in all P-chains sequenced to date except yeast (DSKSR) and liverwort (DSTST). Interestingly, this five-amino-acid sequence is also found in the carrot 69-kDa sequence at anentirely different location (Asp-352 to Arg-356), downstream from region D, in a moderately conserved region. Because the 69-kDa subunit is significantly larger than the a-and p-subunits(50-55 kDa), proper alignment required the introduction of numerous gaps in the latter. The 69-kDa sequence from Leu-229 and Ala-515showed good overall homology to the @-chainbetween Leu-132 and Ala-407 (bovine) and is approximately the same length. Hence, very few gaps were required in this region of the sequence. The introduction of an 88-amino-acid gap into the &subunits before Leu-133 (bovine) revealed an hitherto undetected homologous
E
~
R
D
Y
Homology between Vacuolar ATPases and FoFl-ATPases region between Leu-53 and Gln-139 of the carrot sequence (region A). The presence of this large gap, if real, would suggest either of two possible evolutionary origins: either the vacuolar type 69-kDa subunit arose from a common smaller ancestor by an insertion near Leu-133, or the FoF1-type pchain arose from a common larger ancestor by deletion of the region defined approximately by Pro-140 to Pro-228. It is intriguing that the only palindromic sequence in the 69-kDa cDNA occurred in this region: AACATCATGTT (c,c’ in Fig. 4). It may or may not be significant that a FastP search of the protein data base for homology to region N identified transposase from the Tn3transposon in E. coli and excisionase from bacteriophage X among the 10 most related proteins, none of which were ATPases. These preliminary comparisons wouldbe consistent with an insertion-based origin of the vacuolar type ATPase. Studies are underway to isolate and sequence genomic clones which may shed more light on the nature of the nonhomologous region. There is increasing evidence that thevacuolar H+-ATPase may be more closely related to archaebacterial H’-ATPases than to the FoF1-type ATPases of eubacteria, mitochondria, and chloroplasts (3, 69). For example, the ATP synthase/ hydrolase of Halobacterium hulobium has recently been shown to be anion-stimulated and inhibited by NEM and nitrate(70, 71). The catalytic complex consists of two major subunits, 86 and 64 kDa, with a totalmolecular mass of 300-320 kDa (71). Two DCCD-binding polypeptides have been identified, 78 and 12 kDa (72). The smaller subunit presumably represents the protonchannelsubunit, by analogy tothe vacuolar H+ATPases. Furthermore, antibodies tothe H+-ATPase of Halobacterium cross-react with the -70-kDa subunit of red beet (73). Comparisons of the amino acid sequences of the vacuolar- and archaebacterial-type H+-ATPaseshave the exciting potential to provide further insights into the origin of ATPases aswell as theevolution of the eukaryotic endomembrane system. Acknowledgments-We (L. Z., H. K., and L. T.)thank Dr. Virginia Walbot, Stanford University, for an enjoyable and enlightening summer in her laboratory during the early stages of this project. The isolation of Xgtll clones was carried out with the excellent technical assistance of Holly Roberts. We also thank our colleagues, Dr. Barry Bowman and Dr. Emma Jean Bowman, for countless stimulating and helpful discussions, for providing the genomic sequence of the Neurospora 70-kDa subunit, and for a critical review of the manuscript. REFERENCES 1. Al-Awqati, Q. (1986) Annu. Reu. Cell Biol. 2 , 179-199 2. Bowman, B. J., and Bowman, E. J . (1986) J. Membr. Bwl. 9 4 , 83-97 3. Nelson, N. (1988) Plant Physiol. 8 6 , 1-3 4. Pedersen, P. L., and Carafoli, E. (1987) Trends Biochem. Sci. 1 2 , 146-150 5 . Sze, H.(1985) Annu. Reu. Plant Physiol. 3 6 , 175-208 6. Boller, T., and Wiemken, A. (1986) Annu. Reu. Plant Physiol. 37,137-164 7. Mellman, I., Fuchs, R., and Helenius, A. (1986) Annu. Rev. Biochem. 55,663-700 8. Njus, D., Kelley, P. M., and Harnadek, G. J. (1986) Biochim. Biophys. Acta 853, 237-265 9. Uchida, E., Ohsumi, Y., and Anraku, Y. (1988) J. Biol. Chem. 263,45-51 10. Wang, Y., and Sze, H. (1985) J . Biol. Chem. 260, 10434-10443 11. Xie, X.-S., and Stone, D. K. (1986) J. Biol. Chem. 2 6 1 , 24922495 12. Cuppoletti, J., Aures-Fischer, D., and Sachs, G. (1987) Biochim. Biophys. Acta 899, 276-284 13. Bowman, E. J., Mandala, S., Taiz, L., and Bowman, B. J. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, 48-52 14. Mandala, S., and Taiz, L. (1986) J. Biol. Chem. 2 6 1 , 1285012855
9111
15. Manolson, M. F., Rea, P. A., and Poole, R. J. (1985) J . Biol. Chem. 260,12273-12279 16. Randall, S. K., and Sze, H. (1986) J. Biol. Chem. 261,1364-1371 17. Rea, P. A., Griffith, C. J., and Sanders, D. (1987) J. Bwl. Chem. 262,14745-14752 18. Uchida, E., Ohsumi, Y., and Anraku, Y. (1985) J. Biol. Chem. 2 6 0 , 1090-1095 19. Cidon, S., and Nelson, N. (1986) J . Biol. Chem. 261,9222-9227 20. Percy, J. M., Pryde, J. G., and Apps, D. K. (1985) Biochem. J. 23 1,557-564 21. Arai, H., Berne, M., Terres, G., Terres, H., Puopolo, K., and Forgac, M. (1987) Biochemistry 26, 6632-6638 22. Moriyama, Y.,and Nelson, N. (1987) J. Biol. Chem. 262, 91759180 23. Arai, H., Berne, M., and Forgac, M. (1987) J. Biol. Chem. 2 6 2 , 11006-11011 24. Brown, D., Gluck, S., and Hartwig, J . (1987) J. Cell Biol. 1 0 5 , 1637-1648 25. Sun, S.-Z., Xie, X.-S., and Stone, D. K. (1987) J. Biol. Chem. 262,14790-14794 26. Moriyama, Y., and Nelson, N. (1987) J. Biol. Chem. 262,1472314729 27. Percy, J. M., and Apps, D. K. (1986) Biochem. J. 239,77-81 28. Rausch, T., Butcher, D.N., and Taiz, L. (1987) Plant Physiol. 85,996-999 29. Mierendorf, R. C., Percy, C., and Young, R. A. (1987) Methods Enzymol. 152,458-469 30. Helms, C., Graham, M. Y., Dutchik, J. E., and Olson, M. V. (1985) DNA 4, 39-49 31. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning, A Laboratory Manual, Cold Spring HarborLaboratory, Cold Spring Harbor, New York 32. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U. S. A. 74,5463-5467 33. Biggin, M. D., Gibson, T. J., and Hong, G. F. (1983) Proc. Natl. Acad. Sci. U. S. A. 80,3963-3965 34. Mandala, S., and Taiz, L. (1985) Plant Physwl. 7 8 , 327-333 35. Huynh, T. V., Young, R. A., and Davis, R. W. (1984) in DNA Cloning Techniques: A Practical Approach. (Glover, D., ed) pp. 49-78, IRL Press, Oxford 36. Heidecker, G., and Messing, J. (1986) Annu. Reu. Plant Physiol. 37,439-466 37. Proudfoot, N. J., and Brownlee, G. G. (1976) Nature 2 6 3 , 211214 38. Pustell, J., and Kafatos, F. C. (1984) Nucleic Acids Res. 1 2 , 643656 39. Pedersen, P. L., and Carafoli, E. (1987) Trends Biochem. Sci. 12, 186-189 40. Kyte, J., and Doolittle, R. F. (1982) J. Mol. Biol. 1 5 7 , 105-132 41. Eisenberg, D., Schwarz, E., Komaromy, M., and Wull, R. (1984) J.Mol. BWl. 179,125-142 42. Lipman, D. J., and Pearson, W. R. (1985) Science 227, 14351441 43. Needleman, S. B., and Wunsch, C.D. (1970) J. Mol. Biol. 4 8 , 443-453 44. Doolittle, R. F. (1981) Science 2 1 4 , 149-159 45. Parsonage, D., Wilke-Mounts, S., and Senior, A. E. (1987) J. Bwl. Chem. 262,8022-8026 46. Andrews, W. W., Hill, F. C., and Allison, W. S. (1984) J. Biol. Chem. 259,14378-14382 47. Yoshida, M., Poser, J. W., Allison, W. S., and Esch, F. S. (1981) J.Bwl. Chem. 256,148-153 48. Saraste, M., Gay, N. J., Eberle, A., Runswick, M. J., and Walker, J. E. (1981) Nucleic Acids Res. 9 , 5287-5296 49. Fry, D. C., Kuby, S. A., and Mildvan, A. S. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,907-911 50. Walker, J. E., Saraste, M., Runswick, M. J., and Gay, N. J. (1982) EMBO J. 1,945-951 51. Noumi, T., Futai, M., and Kanazawa, H. (1984) J. Biol. Chem. 259, 10076-10079 52. Saltzgaber-Muller, J., Kanapuli, S. P., and Douglas, M. G. (1983) J. Biol. Chem. 258,11465-11470 53. Zurawski, G., Bottomley, W., and Whitfeld, P. R. (1982) Proc. Natl. Acad. Sci. U. S. A. 7 9 , 6260-6264 54. Runswick, M. J., and Walker, J. E. (1983) J. Biol. Chem. 2 5 8 , 3081-3089 55. Kanazawa, H., Kayano, T., Mabuchi, K., and Futai, M. (1981) Biochem. Biophys. Res. Commun. 103,604-612
9112
Homology between VacuolarATPases and FoFl-ATPases
56. Deno, H., Shinozaki, K., and Sugiura, M. (1983) Nucleic Acids Res. 11,2185-2191 57. Ohyama, K., Fukazawa, H., Kohchi, T., Shirai, H., Sano, T., Sano, S., Umesono, K., Shiki, Y., Takeuchi, M., Chang, Z., Aota, S., Inokuchi, H., and Ozeki, H. (1986) Nature 322,5725 74 Chem. 58. Kuby, S. A., Palmieri, R. H., Frischat, A., Fischer, A. H.,Wu, L. H., Maland, L., and Manship, M. (1984) Biochemistry 23, 2393-2399 59. Karn, J., Brenner, S., and Barnett, L. (1983) Proc. Natl. Acad. Sei. U.S. A. 80,4253-4257 60. Feng, D. F., Johnson, M. S., and Doolittle, R. F. (1985) J. Mol. E d . 2 1 , 112-125 70. 61. Manolsson, M. F., Percy, J. M., Apps, D. K., Xie, X., Stone, D. K., Harrison, M., and Clarke, D. J. (1987) in Proceedings of the Membrane Protein Symposium, August 3-6, 1986, San Diego, CA (Goheen, S. C., ed) pp. 427-434, Bio-Rad Laboratories, Richmond, CA 62. Rea, P. A., Griffith, C. J., Manolson, M. F., and Sanders, D. (1987) Biochim. Biophys. Acta.9 0 4 , l - 1 2 63. Ferguson, S. J., Lloyd, W. J., Lyons, M. H., and Radda, G. K. (1975) Eur. J. Biochem. 5 4 , 117-126
64. Andrews, W. W., Hill, F. C., and Allison, W. S. (1984) J. Biol. Chem. 259,8219-8225 65. Yoshida, M., Poser, J. W., Allison, W. S., and Esch, F. S. (1981) J. Biol.Chem. 256, 148-153 66. Yoshida, M., Allison, W. S., Esch, F. S., and Futai, M. (1982) J . Bwl. 257,10033-10037 67. Noumi, T., Azuma, M., Shimomura, S., Maeda, M., and Futai, M. (1987) J.Bwl. Chem. 262,14978-14982 Hundal, Norling, B., Sandri, G., Wojtczak, L., 68. Ernster, L.,T., Grinkevich, V. A., Modyanov, N. N., and Ovchinnikov, Y. A. (1986) Chem. Scr. 2 6 B , 273-279 69. Maloney, P. C., and Wilson, T. H. (1985) Bioscience 35,43-48 Nanba, T., and Mukohata, Y. (1987) J. Biochem. (Tokyo) 1 0 2 , 591-598 71. Mukohata, Y., and Yoshida, M. (1987) J. Biochem. (Tokyo) 102, 797-802 72, Mukohata, Y., Isoyama, M., Fuke, A., Sugiyama, Y., Ihara, K., Yoshida, M., and Nanba, T.(1987) in Perspectives of Biological Energy Transduction (Mukohata, Y., Morales, M. F., Fleischer, S., e&) Academic Press, Tokyo, in press 73. Mukohata, Y., Ihara, K., Yoshida, M., Konishi, J., Sugiyama, Y., and Yoshida, M. (1987) Arch. Bwchem. Bwphys. 259,650-654
