JBC

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M407821200 on July 21, 2004

J. Biol. Chem., Vol. 279, Issue 40, 41942-41949, October 1, 2004
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
279/40/41942    most recent
M407821200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wilkens, S.
Right arrow Articles by Forgac, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wilkens, S.
Right arrow Articles by Forgac, M.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Three-Dimensional Structure of the Vacuolar ATPase

LOCALIZATION OF SUBUNIT H BY DIFFERENCE IMAGING AND CHEMICAL CROSS-LINKING*

Stephan Wilkens{ddagger}§, Takao Inoue¶||, and Michael Forgac¶

From the {ddagger}Department of Biochemistry, University of California, Riverside, Riverside, California 92521, and Department of Physiology, Tufts University School of Medicine, Boston, Massachusetts 02111

Received for publication, July 12, 2004


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
The structure of the proton-pumping vacuolar ATPase (V-ATPase) from bovine brain clathrin coated vesicles was analyzed by electron microscopy and single molecule image analysis. A three-dimensional structural model of the complex was calculated by the angular reconstitution method at a resolution of 27 Å. Overall, the appearance of the V0 and V1 domains in the three-dimensional model of the intact bovine V-ATPase resembles the models of the isolated bovine V0 and yeast V1 domains determined previously (Wilkens, S., and Forgac, M. (2001) J. Biol. Chem. 276, 44064–44068; Zhang, Z., Charsky, C., Kane, P.M., and Wilkens, S. (2003) J. Biol. Chem. 278, 47299–47306). To determine the binding position of subunit H in the V-ATPase, electron microscopy and cysteine-mediated photochemical cross-linking were used. Difference maps calculated from projection images of intact bovine V-ATPase and a V-ATPase preparation in which the two H subunit isoforms were removed by treatment with cystine revealed less protein density at the bottom of the V1 in the subunit H-depleted enzyme, suggesting that subunit H isoforms bind at the interface of the V1 and V0 domains. A comparison of three-dimensional models calculated for intact and subunit H-depleted enzyme indicated that at least one of the subunit H isoforms, although poorly resolved in the three-dimensional electron density, binds near the putative N-terminal domain of the a subunit of the V0. For photochemical cross-linking, unique cysteine residues were introduced into the yeast V-ATPase B subunit at sites that were localized based on molecular modeling using the crystal structure of the mitochondrial F1 domain. Cross-linking was performed using the photoactivatable sulfhydryl reagent 4-(N-maleimido)benzophenone. Cross-linking to subunit H was observed from two sites on subunit B (E494 and T501) predicted to be located on the outer surface of the subunit closest to the membrane. Results from both electron microscopy and cross-linking analysis thus place subunit H near the interface of the V1 and V0 domains and suggest a close structural similarity between the V-ATPases of yeast and mammals.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
The vacuolar ATPase (V-ATPase, V1 V0-ATPase)1 is an essential component of all eukaryotic cells (1–4). The complex is found in both the endomembrane system of subcellular organelles and the plasma membrane of certain specialized cells. The function of the V-ATPase is to pump protons across the membrane bilayer, a process powered by hydrolysis of ATP. The V-ATPase is a large, multisubunit complex that can be divided into three domains: a water-soluble V1, which is responsible for binding and hydrolysis of ATP; a membrane embedded V0, which contains the proton translocation pore; and a stalk domain, which functions as a structural and functional connection between the V1 and V0. The subunit composition of the bovine V-ATPase has been determined by quantitative amino acid analysis to be A3B3CDEFG2H2 for the V1 and a(cc')4–5c''d for the V0 (5, 6). Overall, the structure and mechanism of the V-ATPase is similar to the structure and mechanism of the related F1F0-ATPsynthase (7, 8). According to the current mechanistic model of V-ATPase function, ATP hydrolysis taking place on the V1 A subunits drives rotation of a central domain composed of subunits D, F, d, and the ring of proteolipids against the remainder of the complex (9, 10). Proton translocation occurs at the interface of the membrane bound domain of the V0 a subunit and the lipid exposed glutamate sidechains from the c subunits. As in the F-ATPase motor, a peripheral stalk prevents rotation of the static domain during enzyme functioning. However, electron microscopy studies conducted with the bacterial (11), bovine (12), and plant (13) enzyme have suggested that there are at least two, possibly three peripheral stalks in the V-ATPase. Based on biochemical studies, it has been proposed that one of the stators contains subunits E and G (14, 15) and another one the N-terminal domain of subunit a, possibly together with subunit H (16).

Unlike in the F-ATPase, the in vivo proton pumping activity of the eukaryotic V-ATPase is regulated by a dissociation into V1 and V0 (17, 18). The mechanism of this regulation is well studied in the yeast and insect enzyme, and it has been shown that the dissociation is reversible and substrate-dependent and that the regulation is employed by the organism to preserve ATP in periods of limited nutrient availability. An important aspect of this regulation is that the dissociated components are enzymatically inactive in vivo. The isolated V0 is sealed to protons (19) and the isolated V1 is inhibited by Mg-ATP (20). This is again different from the F-ATPase, in which dissociation of the enzyme results in a highly active Mg-ATPase, F1, and a passive proton pore, F0. Despite virtually no sequence homology between the accessory V- and F-ATPase subunits, it has been shown that almost all of the V-ATPase polypeptides have functional homologues in the F1F0-ATPsynthase except for subunits d, H, C, and the large cytoplasmic domain of the V0 subunit a. It has been proposed that subunit C and the cytoplasmic domain of subunit a are involved in the mechanism of reversible dissociation (21, 22), whereas subunit d might have a structural role (23). Subunit H, on the other hand, is the only subunit that is not strictly required for assembly (24), and it has been shown that subunit H can be removed from the enzyme without subsequent dissociation of the complex into V1 and V0 (6, 25), suggesting a peripheral binding location for this subunit. It is interesting that although there have been reports that subunit H is required for Mg-ATPase activity of intact V-ATPase (6, 24, 25), the situation in the membrane detached V1 seems to be the opposite, in that the homologous subunit in yeast (Vma13p) is an inhibitor of Mg-ATPase activity in the V1-ATPase (26). Furthermore, subunit H has been identified to be responsible for the interaction of the V-ATPase with other cellular proteins such as HIV-Nef (27, 28), AP-2 (28) and ecto apyrase (29). Subunit H in higher eukaryotes is expressed as two isoforms as a result of gene splicing (30), and there is evidence that there is one copy of each isoform bound in the complex, both of which seem to be required for enzyme activity in the intact complex (6, 30).

The structure of the eukaryotic V-ATPase and its V1 and V0 domains has been studied by electron microscopy (12, 13, 31–33); based on these studies and the available biochemical data as well as the enzyme's similarity to the F-ATPase, a structural picture of the V-ATPase is now emerging. However, the arrangement of the stalk subunits is still poorly understood and remains a matter of ongoing controversy.

In the present study, we have used electron microscopy together with computer-assisted image analysis to determine a low resolution three-dimensional structural model of the bovine V-ATPase. Overall, the model of the intact bovine enzyme is similar to the previously determined models of the bovine V0 (32) and yeast V1 (33). Furthermore, the binding site(s) of subunit H in the bovine and yeast V-ATPase complex were determined by electron microscopy and photochemical cross-linking. The data indicate that subunit H is binding at the bottom periphery of the V1, in the interface between the V1 and V0 domains, near the cytoplasmic domain of the V0 a subunit.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Materials and Strains—All reagents were from Sigma Chemical unless noted otherwise. The V-ATPase from bovine brain clathrin-coated vesicles was purified as described previously (34). Subunit H was removed from intact, purified V-ATPase by extended dialysis in the presence of the oxidizing agent cystine. This procedure leads to the removal of the majority of both subunit H isoforms as well as a residual 50-kDa subunit of the AP-2 adapter complex (AP-2), which has a molecular mass similar to that of subunit H (6). Zymolase 100T was obtained from Seikagaku America Inc., and 4-(N-maleimido)benzophenone (MBP) was from Sigma. The monoclonal antibody 13 D11 against the yeast subunit B was purchased from Molecular Probes. The rabbit polyclonal antibody against subunit H was the kind gift of Dr. Tom Stevens.

Yeast strains disrupted in the endogenous gene encoding the V-ATPase B subunit (vma2{Delta}) expressing mutant forms of Vma2p in which the endogenous cysteine at position 188 was replaced with serine and a unique cysteine was introduced were constructed as described previously (14). The following single-cysteine containing mutants of subunit B were tested for photochemical cross-linking using MBP as described below: A15C, K45C, E106C, D199C, D341C, A424C, E494C, and T501C.

Electron Microscopy—For negative staining, protein at a concentration of between 20 and 50 µg/ml was applied to glow-discharged, carbon-coated copper grids. Grids were washed once with buffer or water, stained with 1% uranyl acetate, and air dried. Grids were examined in a Philips CM300 transmission electron microscopes operating at 100 kV. Images were recorded in low-dose mode on Kodak SO163 film that was developed in D19 developer for 12 min at room temperature. Images of stained samples were recorded with an underfocus of between 382 and 576 nm and an electron optical magnification of 47,000x, placing the first zero of the contrast transfer function at around 1/20 Å–1.

Analysis of the V1V0-H images—Electron micrographs recorded at 47,000x were digitized on an Optronics drum scanner with a sampling rate of 18.75 µm corresponding to 4 Å/pixel on the specimen level. 6950 individual molecular images were selected interactively from 45 micrographs. Images were analyzed with the IMAGIC 5 software package (35) essentially as described previously (12, 32, 33). The images were normalized and band pass filtered to remove low (<0.015 Å–1) and high (>0.1 Å–1) spatial frequencies. The data set was then analyzed by the "alignment by classification" (alignment by classification (36)) procedure leading to sets of initial references which were used in a first multi reference alignment step. The multireference alignment was iterated until no further improvement in the class sums was obtained. The analysis of two data sets of intact V-ATPase has been described previously (12).

Three-dimensional Reconstruction of Intact V-ATPase—One of the data sets described in Ref. 12 served as a starting point for the three-dimensional reconstruction. The data set was sorted into 80 classes, and initial angles were assigned assuming 3-fold symmetry along the long axis of the complex. All further refinement steps were calculated assuming no (C1) symmetry. After eight rounds of refinement, the model did not improve significantly anymore, but because virtually all projections were seen perpendicular to the long axis of the complex (the variation in the Euler angles {beta} at this stage was approximately ±15°), angle assignment was fluctuating for many of the projections. The number of images in the data set was therefore increased to ~17,200, including images from 12 micrographs recorded with the specimen tilted at an angle of 45° (total number of micrographs = 54). The model after the eighth refinement iteration was forward-projected along 120 directions uniformly distributed on the Euler sphere between {beta}-angles of 45° and 135°, and the resulting projections were used as references in a multireference alignment. The aligned data set was sorted into 100 classes, and angles were assigned to 80 of these using 144 forward projections of the model after the eighth iteration as an anchor set. The procedure was iterated with up to 176 forward projections as references in the final refinement step. The resolution of the final reconstruction was estimated by the Fourier shell correlation method. The resolution was 27 Å using the 0.5 criterion (37) and 18 Å using the 3 {sigma} measure (38). The final model was filtered to a resolution of 1/24 Å–1 using a Gaussian filter.

Three-dimensional Reconstruction of Subunit H-depleted V-ATPase— The final 3-D model of the intact V-ATPase was projected along 80 directions uniformly distributed between {beta}-angles of 80° and 100° on the Euler sphere. The resulting projections then served as references to align the data set of subunit H-depleted molecules. The aligned data set was divided into 80 classes based on the cross-correlation coefficient obtained during the multireference alignment with the 80 projections of the intact V-ATPase complex. The classes were averaged and used to calculate a three-dimensional model of the subunit H-depleted complex.

Difference Mapping—To localize the density corresponding to subunit H in the images of the intact V-ATPase, difference images from two-dimensional projections and the three-dimensional models were calculated. Before calculation of difference images, the projections were aligned to each other and normalized to give an average density of zero and a standard deviation of 10. In general, images of subunit H-depleted V-ATPase were subtracted from images of intact V-ATPase so that the position of additional protein mass in the intact complex was indicated by positive (white) density.

Photochemical Cross-linking of the V-ATPase Using MBP—Vacuolar membranes were prepared from the vma2{Delta} strain expressing each of the single cysteine-containing mutants of Vma2p (A15C, K45C, E106C, D199C, D341C, A424C, E494C, and T501C) as described previously (14). Covalent cross-linking using MBP was performed by a modification of the procedure previously employed (15). Vacuolar membrane vesicles (200 µg of protein) were washed three times in phosphate-buffered saline, pH 7.2, containing 2 mM EDTA, 1 mM PMSF, 2 µg/ml aprotinin, 0.7 µg/ml pepstatin, and 5 µg/ml leupeptin, and then resuspended in 100 µl of the same buffer. The samples were reacted with MBP at 50 µM for 30 min at 23 °C in the dark. Unreacted MBP was quenched by addition of 10 mM dithiothreitol, and vacuolar membranes were washed twice with the above buffer and resuspended in 100 µl of buffer. The samples were irradiated for 5 min at 4 °C with a long-wavelength ultraviolet light (365 nm) and then solubilized with 2% C12E9. The V-ATPase was immunoprecipitated using the mouse monoclonal antibody 13D11 against the yeast subunit B and protein A-Sepharose, and the immunoprecipitated proteins were separated by SDS-PAGE on 7.5% acrylamide gels as described previously (14). Gel electrophoresis was followed by electrophoretic transfer to nitrocellulose and Western blotting (14) using the mouse monoclonal antibody against subunit B and a rabbit polyclonal antibody against subunit H.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
Three-dimensional Reconstruction of the Intact V-ATPase— To start up the three-dimensional reconstruction of the intact V-ATPase, a three-fold symmetry was imposed in the first angle assignment step based on the pseudo three-fold symmetry of the catalytic A3B3 core. The model was subsequently refined assuming no symmetry at all. Images recorded at a tilt angle of 45° were included to improve the accuracy of the angle assignment by the common line method. Surface representations of the final 3-D model of the intact V1-ATPase are shown in Fig. 1B. A selection of final input projections (images 1, 3, 5, and 7) and the corresponding re-projections (images 2, 4, 6, and 8) is shown in Fig. 1A. Fig. 1C shows contoured cross-sections of the 3-D model from top to bottom as indicated by the white lines on the left side of image 1 in Fig. 1B.



View larger version (102K):
[in this window]
[in a new window]
  		FIG. 1.
Three-dimensional reconstruction of the bovine V-ATPase. A, input projections (images 1, 3, 5, and 7) and re-productions (images 2, 4, 6, and 8). The total number of input projections was 176. B, surface representation of the final 3-D model of the bovine V-ATPase. The orientations are the same as in the projection images shown in part A. The views in images 2, 3, and 4 are rotated counter-clockwise looking down on the membrane by 60°, 120°, and 180° from image 1, respectively. C, contoured cross-sections through the model at positions indicated by the white lines shown on the left side of image 1 (B). Bar, 5 nm. For details, see text.

 
Overall Features of the Three-dimensional Model of the Intact V-ATPase—Overall, the three-dimensional model of the bovine V-ATPase has a very similar appearance for the V1 and V0 portions compared with the three-dimensional models of the isolated V0 (32) and V1 (33) domains. Starting at the luminal side of the complex (Fig. 1C, image 16) is the well defined density corresponding to subunit Ac45 followed by the ring of proteolipids (Fig. 1C, images 15–12). As in the model for isolated bovine V0, the individual c subunits are not resolved, but the putative subunit a C-terminal domain can be seen at the periphery of the ring (see arrowheads in images 15–12). Fig. 1C, image 11, shows the cross-section through the density sitting in the cytoplasmic opening of the proteolipid ring and as in the isolated V0 (32), the density has a clear three domain appearance (see small arrows). In the previously determined three-dimensional model of the bovine V0, the putative N-terminal domain of subunit a is seen to interact with the protein density bound in the cytoplasmic opening of the proteolipid ring. This is different in the intact bovine V-ATPase, which shows the protein density emerging from the membrane surface in a straight fashion (Fig. 1C, images 11–9, arrowheads), stretching the entire length of the stalk region to interact at the bottom of the V1 domain. In the stalk region, six major protein densities can be identified (image 10), including a central, spherical mass (asterisk) which is surrounded by five densities. Two of the five densities are well defined and near the spherical protein in the center (image 10, long arrows) whereas the two densities at the periphery of the stalk are weaker (image 10, short arrows). The density between the two peripheral ones is the putative subunit a N-terminal domain (image 10, arrowhead). The two weak densities that are pointed out in images 9 and 10 by short arrows are visible in the three-dimensional model, but they seem not to be connected to the main structure by protein density at the contour level chosen (Fig. 1B, images 1 and 2, long arrows). The arrowheads in image 8 point to the bottom of three of the large subunits of the V1. The typical pseudo trigonal arrangement of the (AB)3 hexagon is seen most clearly in images 7–5 (see alternating arrows and arrowheads in image 6). A similar pseudo three-fold symmetry is seen in the yeast V1-ATPase model (Zhang et al., 2003) and the crystal structure of the F1-ATPase (39). The projection images and the surface representations seen in Fig. 1, A and B respectively, show the previously described "knob" and "elongated" densities at the top of the V1-domain (see arrowheads and short arrows in images A1, 3, and 5 and B1–3, respectively). From the cross-section shown in image 4 of Fig. 1C, the knob-like densities seem less well resolved in the model of the intact V-ATPase (see arrowheads in image 4) compared with the yeast V1 ATPase (33). This might be explained by the absence of pure top views in the set of input projections used for calculating the three-dimensional reconstruction of the bovine enzyme. Nevertheless, rotation of the three-dimensional volume in steps of 60° (see arrows in images B, 1–3 in Fig. 1) shows the alternating knob-like densities switching from one side to the other of the complex as seen in the projections described before (12, 33). Cross-sections of the elongated densities visible at the very top of the complex can be seen in image 3 (see arrowheads). Although it seems that two of these densities are more separate from their corresponding large subunits (the two top ones in image 3 of C), the significance of this difference is unclear given the resolution of the current model. At the very top of the complex, a strong three-fold symmetry component can be seen (images 2 and 1; see arrowheads in 1), again consistent with the structural model of the yeast V1-ATPase (33).

Fitting the Crystal Structure of F-ATPase {alpha}3{beta}3 into the V-ATPase 3-D Model—Fig. 2 shows fitting of the crystal structure of the nucleotide free {alpha}3{beta}3 domain of the F-ATPase from the thermophilic bacillus PS3 (Protein Data Bank code 1SKY [PDB] ). The fit with the highest cross-correlation value is shown. In this orientation, the {alpha}-subunits of the F1 are superimposed onto the V-ATPase subunits, which have the elongated densities bound at the top, and the F-ATPase {beta} subunits superimpose onto the V-ATPase subunits containing the knob-like densities. Based on the subunit homology between F- and V-ATPases, this would suggest that the B subunits have the elongated densities bound at the top and the A subunits contain the knob-like extensions, consistent with our earlier assignment of the catalytic subunits (12, 33). The second best fit with a correlation value of 0.98 compared with a relative value of 1 for the best fit was rotated by 60° around the axis of three-fold symmetry with respect to the best fit. The fitting also suggests that the model shown in Fig. 1 has the correct handedness, because using a mirrored version of the 3-D model led to a fit with a significantly lower correlation value (0.87 for the mirrored version compared with a relative value of 1 for the unmirrored version).



View larger version (80K):
[in this window]
[in a new window]
  		FIG. 2.
Fitting the F-ATPase {alpha}3{beta}3 atomic coordinates into the bovine V-ATPase 3-D model. Fitting was done in SITUS (situs.biomachina.org) with the crystal structure of the nucleotide free {alpha}3{beta}3 hexamer of the F1-ATPase from the thermophilic bacillus PS3 (Protein Data Bank 1SKY [PDB] ). Top, side view; bottom, top view. A slice of the V1 domain is shown as indicated on the left in the top image. The figure was generated with the molecular display software VMD (www.ks.uiuc.edu/Research/vmd/). For details, see text.

 
Subunit H-depleted V-ATPase—Genetic analysis in yeast has shown that subunit H (Vma13p) is the only subunit of the eukaryotic V-ATPase that is not strictly required for assembly of the complex in vivo (24). In the bovine V-ATPase, two isoforms of subunit H are found (H{alpha} and H{beta} (25)) and it has been shown that both isoforms can be removed from the complex without subsequent dissociation into V1 and V0 (6, 25), consistent with the genetic studies in yeast. Fig. 3A shows denaturing polyacrylamide gel electrophoresis of the intact and subunit H-depleted bovine V-ATPase preparations analyzed in this study. Only the lower molecular mass isoform of subunit H is resolved on the gel, whereas the heavier of the isoforms is co-migrating with subunit B. Based on antibody blots with an antibody that recognizes both isoforms of the subunit, treatment of bovine V-ATPase with cystine leads to almost complete dissociation of the lower molecular weight isoform, whereas the heavier isoform is only partly removed (6).



View larger version (78K):
[in this window]
[in a new window]
  		FIG. 3.
Gel electrophoresis and image analysis of intact and subunit H-depleted bovine V-ATPase. A, lanes 1 and 2, SDS-polyacrylamide gel electrophoresis of intact V-ATPase, 15% and 12% gel, respectively. Lane 3, 12% gel of V1V0 depleted of subunit H by treatment with cystine. The gels were stained with Coomassie blue. Lanes 1 and 2 are from Ref. 12. B1, total average of the aligned data set of 6950 subunit H-depleted V-ATPase molecules. B2, average of the same number of intact V-ATPase molecules. B3, difference map. The difference map was calculated by subtracting the V1V0-H average from the average of the V1V0 data set after symmetrizing both averages with respect to the long axis of the complex. B4, same image as B3, displayed with different gray levels. C, image analysis of the subunit H-depleted V-ATPase. The data-set of 6950 molecules was treated by multivariate statistical analysis/classification and sorted into 24 classes. Shown are four averages representing the most characteristic views. Averages were calculated from between 120 and 250 images, respectively. D, image analysis of the data set of intact V-ATPase molecules. The data set used for the three-dimensional reconstruction shown in Fig. 1 was treated by multivariate statistical analysis/classification after the final round of alignment and sorted into 150 classes. Averages of molecules having a similar orientation compared with the subunit H-depleted molecules are shown. E, difference maps calculated by subtracting the images shown in C from the images shown in D.

 
Difference Mapping of Two-dimensional Projections—To determine the location of this subunit in the V-ATPase, a data-set of 6950 subunit H-depleted V-ATPase molecules has been analyzed by single particle image analysis. The results are summarized in Fig. 3, B–E. Comparison of the total averages of the V1V0-H and V1V0 data sets shown in Fig. 3B, images 1 and 2, respectively, revealed that there is slightly less protein density directly underneath the V1 domain. However, inspection of the averages indicated also that the subunit H-depleted molecules adopted slightly different preferred orientations on the carbon support film compared with the intact V-ATPase enzyme complex. To minimize the effect of this overall difference in the orientations, the averages were symmetrized along the long axis of the molecule before calculating the difference map shown in Fig. 3B, image 3. Image B4 shows the same difference map displayed with a contour level so that only the strongest positive difference peaks are visible. Two areas of strong signal can be seen in the difference map obtained by subtracting the H-depleted molecule from the H-containing enzyme. A negative (black) density is surrounding the position of the V0 and a strong positive (white) density can be seen just below the V1 (see black arrows and arrowheads, respectively, in image B3). The strong negative density surrounding the V0 can be explained based on the fact that the subunit H-depleted enzyme has been prepared in presence of phospholipid as opposed to the intact enzyme, which has been purified without extra lipid. The addition of phospholipid during the preparation of the subunit H-free enzyme is necessary to stabilize the enzyme. The strong positive density underneath the V1, on the other hand, must be caused by the removal of subunit H; therefore, it indicates the binding site of this subunit in the complex. However, because of the imposed mirror symmetry of the averages before calculating the difference map, the difference signal is spread out underneath the V1. To determine the binding site(s) for the subunit H isoforms in more detail, difference maps from averages of both subunit H-depleted and -intact V-ATPase were calculated after sorting the data sets into classes of molecules having similar orientations on the carbon film. Fig. 3, C and D, shows multivariate statistical analysis/classification of the subunit H-depleted and -intact V-ATPase data sets, respectively. Averages of molecules having a similar orientation were chosen for both data sets. The difference maps are shown in Fig. 3E. Images E1 and E4 show a clear difference peak in the V1-V0 interface, whereas images 2 and 3, especially image 2, show diffuse difference maxima either on the left (image 2) or right (image 3) periphery of the complex. The difference peak in image 2 corresponds to the poorly resolved density seen in some projections of the intact enzyme (see also Fig. 5, image 4, in Ref. 12). A similar density is not seen in any of the averages of the subunit H-depleted enzyme, suggesting that at least one of the subunit H isoforms binds in that region of the complex.



View larger version (43K):
[in this window]
[in a new window]
  		FIG. 5.
Molecular model of subunit B of the yeast V-ATPase. A molecular model of the yeast A3B3 available hexamer was derived from the x-ray coordinates of the bovine mitochondrial F1 (39), sequence homology between the nucleotide binding subunits of the V and F-ATPases and energy minimization (53). Shown is the model of the B subunit oriented such that residues substituted with cysteine that are predicted to face the exterior of the complex (shown in green; Lys-45, Glu-106, Asp-199, Glu-494 and Thr-501) are on the right, whereas residues substituted with cysteine that are predicted to face the interior of the A3B3 hexamer (shown in red; Asp-341 and Ala-424) are on the left. The molecule is oriented with the region farthest from the membrane at the top. Not shown is Ala-15, which is expected to reside on the exterior surface near Lys-45. The figure was generated with Molscript (54) and Raster3d (44).

 
Position of Subunit H in the Three-dimensional Model of the V1V0—To localize subunit H in the three-dimensional model of the intact V-ATPase, 80 projections of the final model were used to align the data sets of both subunit H-containing and -depleted enzyme. The aligned data sets were sorted into 80 classes based on the cross-correlation coefficients obtained during the alignment, and three-dimensional models of both data sets were calculated from the averages of the 80 classes. Fig. 4 summarizes the results. As can be seen, the three-dimensional models of the intact complex (Fig. 4A) and the subunit H-depleted enzyme (Fig. 4B) are virtually identical, except for the elongated density shaded in red that is present in the intact complex but not the subunit H-depleted enzyme. The contour level for calculating the surface representations was chosen so that both models contained the same volume. Fig. 4C shows the stalk region as seen in images 2 of A and B enlarged by a factor of 2. The position of the additional density seen in the intact enzyme is also evident in a three-dimensional difference volume obtained by subtracting the volume of the subunit H-depleted enzyme from the model of the intact complex. The largest positive difference matches the position where the elongated density seen in the intact complex interacts with the putative a subunit N-terminal domain at the bottom of the V1. Lowering the contour level in the difference volume will reveal additional minor difference peaks consistent with the two-dimensional difference maps shown in Fig. 3E. It is possible that only the part of the H subunit that interacts with the V1 is sufficiently ordered to give a clear signal in the difference volume (see arrow in Fig. 4C, image 3). The absence of the elongated density in the model of the subunit H-depleted enzyme is significant because one might expect that alignment of the images of subunit H-depleted enzyme with images of the subunit H-containing complex would bias the outcome in such a way that the difference between images of the two preparations would be minimized. At the contour level chosen, no difference peak is visible for the second peripheral density in the stalk region, although the density seems somewhat weaker in the subunit H-depleted complex (see arrowheads in images A2 and B2).



View larger version (78K):
[in this window]
[in a new window]
  		FIG. 4.
Position of subunit H in the three-dimensional model of the V-ATPase. Data sets of intact and subunit H-depleted V-ATPase molecules were aligned to 80 forward projections of the final three-dimensional model of the intact V-ATPase. The data sets were sorted into classes based on the cross-correlation coefficient obtained during the alignment and three-dimensional models from the two data sets were calculated. A, images 1 and 2 show surface representations rotated by 120° with respect to each other. B, surface representations of the subunit H-depleted enzyme shown in the same orientations as the intact complex in A. The density missing in the subunit H-depleted enzyme is shaded in red in the images of the intact enzyme. C, enlarged view of the stalk region showing the additional density in the intact enzyme more clearly. Image 3 in C shows the difference volume calculated by subtracting the model of the subunit H-depleted enzyme from the model of the intact enzyme. Scale bars, 5 nm.

 
The image on the right in Fig. 4 shows a surface representation of the crystal structure of subunit H from yeast (40), filtered at a resolution of 20 Å. As can be seen, the overall shape of isolated subunit H is very similar to the shape of the extra density seen in the intact V-ATPase, suggesting that this density indicates the binding position of one of the subunit H isoforms in the intact bovine V-ATPase.

Position of Subunit H in the Yeast V-ATPase Based on Photochemical Cross-linking—As a further test of the location of subunit H within the V-ATPase complex, photochemical crosslinking was performed on the yeast V-ATPase using the photoactivatable sulfhydryl reagent MBP. Unique cysteine residues were introduced into a Cys-less form of the yeast V-ATPase subunit B (Vma2p) at sites localized on the basis of a molecular model of the V1 domain described previously (14, 15). Fig. 5 shows a ribbon diagram of the modeled B subunit with the site-directed cysteine residues indicated in space fill. The B subunit is oriented such that the residues predicted to be facing the outer surface of V1 are on the right and the residues closest to the membrane are at the bottom of the structure. The unique cysteine residues were used as sites of attachment of MBP and photoactivated cross-linking was then carried out on isolated vacuolar membranes followed by separation on SDS-PAGE and Western blot analysis using antibodies against both subunits B and H.

Of the single-cysteine mutants tested, only two gave bands that were recognized by antibodies against both subunits B and H, namely E494C and T501C (Fig. 6). Cross-linking from E494C and T501C resulted in a doublet and triplet of closely spaced bands, respectively. The cross-link products for both mutants were centered around an apparent molecular mass of ~125 kDa. This molecular mass is close to the sum of the masses of the yeast subunits B (57 kDa) and H (54 kDa), suggesting the formation of a heterodimer. The reason that multiple bands rather than a single band were observed is not certain, but the doublet and triplet bands may be a result of the cross-linking of a single site on subunit B to multiple sites on subunit H. The resultant denatured heterodimeric complexes would have different hydrodynamic properties that may result in slightly different mobility in a polyacrylamide gel matrix. A similar ladder of closely spaced products was previously observed for cross-linking between subunits B and D (14). The results obtained by cross-linking thus support the electron microscopy data described above placing subunit H on the outer surface of the complex near the interface of the V1 and V0 domains.



View larger version (32K):
[in this window]
[in a new window]
  		FIG. 6.
Cross-linking of subunits B and H of the yeast V-ATPase using the photoactivatable sulfhydryl reagent MBP. Shown are the only two B subunit cysteine-containing mutants (E494C and T501C) to show cross-linking of subunit B and H. The positions of the monomeric B and H subunits and the various cross-linked species are indicated with the solid arrowheads. The positions of the heavy and light chains of the immunoprecipitating antibodies are indicated with the open arrowheads. Note that for both B subunit mutants, a B-E heterodimeric product is observed, as reported previously (14, 15).

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
We are presenting the first three-dimensional reconstruction of the mammalian vacuolar ATPase, determined from electron microscopic images at a resolution of 27 Å. We have chosen the enzyme purified from bovine brain clathrin-coated vesicles because of its superior stability and structural integrity which might be because brain clathrin-coated vesicles serve an endocytic function and are therefore likely to be free of the proteases that may result in the lower stability of V-ATPases from digestive organelles, such as lysosomes. The model was reconstructed from images of the complex embedded in negative stain (uranyl acetate), and despite the potential artifacts caused by the stain embedding, such as flattening distortion and uneven stain distribution, it has recently been shown that medium to high resolution three-dimensional models can be obtained from images of negatively stained protein complexes (41, 42). Our attempts to obtain high quality images of the intact V-ATPase by cryoelectron microscopy have not been successful. Protein concentrations of several micrograms per milliliter seem to be required for obtaining cryo-images of membrane proteins in presence of detergent (see, for example, Ref. 43), conditions under which the V-ATPase tends to form aggregates.

To start up the three-dimensional reconstruction, three-fold symmetry along the long axis of the complex was assumed based on the pseudo-three-fold symmetry of the A3B3 catalytic core, which represents nearly 50% of the total mass of the complex. The actual influence of this three-fold symmetry component on the angle assignment is probably even greater because the detergent-covered proteolipid ring does not offer significant features with which angles could be assigned. All subsequent refinement was done without imposed symmetry; the final model shows little if any three-fold symmetry in the stalk region and projections of the final model match the final input projections and the projections obtained earlier (12) quite well, all of which suggests that the many refinement steps done without assuming any symmetry resulted in an overall correct model at the given resolution of 27 Å.

At the current resolution, the overall appearance of the V1 and V0 domains in the here presented model is very similar to the models of the isolated bovine V0 and yeast V1 determined earlier (32, 33). One of the major differences is the position of the putative N-terminal domain of subunit a. In our earlier model of the bovine V0 domain, two interacting protein densities could be seen on the cytoplasmic side of the proteolipid ring. The density that was seen sitting in the opening of the ring had a clear three domain structure, and we speculated that this density corresponded to subunit d, whereas the other density, which was sitting above the peripheral membrane bound density, was the 50 kDa N-terminal domain of subunit a. This assignment is now supported by two studies conducted with the bacterial V-type ATPase. First, it has been found that the bacterial C subunit, which is the homologue to the eukaryotic d subunit, remains bound to the proteolipid ring in absence of both the V1 and the bacterial a subunit homologue (23). Second, the recently published crystal structure of the bacterial C subunit shows a clear three-domain structure with a central depression (45), consistent with the size and three domain structure of the density seen in the cytoplasmic opening of the bovine V0, isolated or as part of the model of the intact enzyme presented here. The position of the putative a subunit N-terminal domain is different in the isolated V0 compared with the V0 domain in the intact enzyme. Although this density is seen to interact with the putative d subunit density in the isolated V0 (32), the density emerging from the putative a subunit C-terminal domain seems to undergo a large conformational change in the intact enzyme to span the entire length of the stalk domain and bind to the bottom of the V1. It has been shown that the a subunit N-terminal domain interacts with the E subunit of the V1 (6), and it is possible that the density going from the membrane surface to the bottom of the V1 contains both subunit E and the N-terminal domain of the a subunit. The interaction of the a subunit cytoplasmic domain with the d subunit as seen in the isolated V0 (32) might provide a way to prevent rotation of the proteolipid ring with respect to the a subunit C-terminal domain, which would explain the fact that the isolated V0 is impermeable to protons (19).

We had speculated previously based on our electron microscopic studies with the bovine V-ATPase (12) and yeast V1-ATPase (33) that at least part of the "knob"-like densities seen at the top periphery of the V1 belong to the A subunit nonhomologous regions, which are not present in the F-ATPase {beta} subunits. We have now confirmed that the knob-like density is part of the A subunit with the use of electron microscopy of yeast V1-ATPase decorated with Fab fragments derived from a monoclonal antibody which recognizes the N terminus of the A subunit.2 Based on this assignment, it seems that the interaction of the density involving the a subunit N-terminal domain occurs at the bottom of one of the A subunits. This interpretation is consistent with immunoprecipitation studies showing that the isolated N-terminal domain of Vph1p interacts with the A subunit in yeast V-ATPase (16). Whether the N-terminal domain of a stretches all the way to the top of the V1, as had been suggested for the yeast (16) and plant V-ATPase (13), is unclear at the present resolution.

From looking at the projections of the intact enzyme (see arrows in Fig. 1A), it seems that also the elongated densities at the top of the V1 are present in three copies. Questions about the identity of these elongated densities cannot be answered with certainty, but based on cross-linking studies performed with the yeast V-ATPase complex (see below), it is likely that at least one of these densities contains part of the E and G subunits. Although quantitative amino acid analysis indicated that there is one copy of E and two copies of G per complex (5, 6), the presence of up to three copies of each E and G per complex has been suggested based on an analysis of Coomassie staining in denaturing polyacrylamide gels (46–48).

The function of subunit H in the V-ATPase complex is poorly understood. Whereas subunit H is required for Mg-dependent ATPase activity in the intact enzyme (6, 25), the subunit seems to act as an ATPase inhibitor in the isolated V1 (26). The structural data presented here and previously (33) indicate that a large portion of subunit H is poorly ordered in the complex. This can be explained by the fact that subunit H can potentially interact with a variety of other cellular proteins such as ecto apyrase (29), AP-2, and HIV NEF (27, 28). The crystal structure of the yeast homologue of bovine subunit H, Vma13p, shows that the subunit is a three-domain protein of ~100 Å in length (see Fig. 4, right image, in Ref. 40). It has been shown that subunit H interacts with the N terminus of subunit E (49) and the N-terminal domain of subunit a (16), consistent with our electron microscopic data. Cross-linking experiments conducted with the bovine V-ATPase have detected cross-link products for subunit H that are observed in the isolated V1 domain but not the intact V1V0 (6). Furthermore, it has been shown that a large portion of the N terminus of subunit H can be deleted without impairing assembly or function of the V-ATPase (24). All these data are consistent with the idea that subunit H bound to the enzyme with one of its three domains, possibly the C-terminal domain, and that the majority of the subunit is flexible in absence of the other cellular proteins, which normally interact with the subunit. Comparing the three-dimensional models of the intact and subunit H-depleted complexes suggests that one of the two subunit H isoforms binds near the site where the putative a subunit N-terminal domain binds at the bottom of the V1. Despite the fact that the extra density seen in the intact enzyme has an elongated shape that is very similar to the shape of the yeast subunit H as seen in the crystal structure, the current resolution does not allow an unambiguous docking of the crystal structure into the electron microscopy-derived electron density. Higher resolution data, as can be obtained by cryoelectron microscopy, for example, will be necessary to fit the x-ray diffraction data into the three-dimensional model of the V-ATPase. A second peripheral density is present in the stalk region of the intact complex (see arrowheads in Fig. 4, images A2 and B2). This density, although somewhat weaker, is also seen in the subunit H-depleted enzyme; it is possible that this density corresponds to the second, heavier subunit H isoform that is not entirely removed by the cystine treatment of the intact enzyme, and it is again possible that only part of the subunit is resolved in the three-dimensional models.

Fig. 7 shows our current structural model of the V-ATPase complex based upon the data presented in the current manuscript together with data available in the published literature. Previous data from photoactivated cross-linking of unique cysteine residues on subunit B suggested that subunits E and G are part of the peripheral stalk connecting the V1 and V0 domains and that subunit D is located in the interior of the A3B3 hexameric head, thus making it the likely homolog to the {gamma} subunit in F1 (14, 15). The peripheral location of subunit E has been challenged by electron microscopy data obtained on the V1 domain of the Manduca sexta V-ATPase (50). One question in attempting to reconcile these two sets of data is the accuracy of the structural model of the V1 domain on which the interpretation of the cross-linking data is based. This model was shown to correctly predict the identity of residues at both the catalytic and non-catalytic nucleotide binding sites of the V-ATPase (51–53), but it was possible that the model would be less accurate at sites distant from the sites of nucleotide binding. The data in the present study indicates that there is good agreement in the location of the H subunit in the V-ATPase complex derived from electron microscopy analysis and photoactivated cross-linking, suggesting that the molecular model of the B subunit employed is able to accurately predict the location of residues even distant from the nucleotide binding sites.



View larger version (69K):
[in this window]
[in a new window]
  		FIG. 7.
Structural model of the V-ATPase. Subunits of the rotor domain and stalk domain are shown in green and red, respectively. The catalytic domain (A3B3) is shown in blue. Ac45 is present in the mammalian but not the yeast V-ATPase.

 
Further support for the model of the V-ATPase as shown in Fig. 7 is derived from the recent observations on rotation of the yeast V-ATPase complex, which demonstrate that subunit G is part of the peripheral stator (10). Cysteine residues predicted from the molecular model of subunit B to be facing the periphery of the complex have been shown to cross-link to both subunits E and G (14, 15), placing both subunits in the peripheral stalk. Finally, rotation of subunit D in the V1 domain of the Thermus thermophilus V-ATPase has been directly demonstrated (9), showing that subunit D is indeed the {gamma}-subunit homolog in the V-ATPases. Comparison of the sequences of the D and E subunits from yeast and Thermus thermophilus show that the D subunits are 27% identical, whereas the D subunit of Thermus thermophilus shares only 12% identity with the E subunit of yeast. It is possible that the central location of the E subunit observed in electron microscopy studies of the Manduca V1 domain may be related to the rearrangement of subunits observed to occur in comparing the isolated V1 with the intact V1V0 complex (6).


   FOOTNOTES
 
* This work was supported in part by National Institute of Health Grants GM58600 (to S. W.) and GM34478 (to M. F.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

|| Supported by a Postdoctoral Fellowship from the American Heart Association, Northeast Affiliate. Back

§ To whom correspondence should be addressed. Tel.: 909-787-3131; Fax: 909-787-4434; E-mail: stephan.wilkens{at}ucr.edu.

1 The abbreviations used are: V1V0, proton-pumping vacuolar ATPase; V1, water soluble domain of the vacuolar proton pumping ATPase; V0, membrane bound domain of the proton pumping vacuolar ATPase; 3-D, three dimensional; MBP, 4-(N-maleimido)benzophenone. Back

2 Z. Zhang, and S. Wilkens, manuscript in preparation. Back


   ACKNOWLEDGMENTS
 
Roderick Nakayama is gratefully acknowledged for excellent technical assistance. We thank Dr. Jim Baleja for help with the molecular modeling.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Nishi, T., and Forgac, M. (2002) Nat. Rev. Mol. Cell. Biol. 3, 94–103[CrossRef][Medline] [Order article via Infotrieve]
  2. Nelson, N., and Harvey, W. R. (1999) Physiol. Rev. 79, 361–385[Abstract/Free Full Text]
  3. Stevens, T. H., and Forgac, M. (1997) Annu. Rev. Cell Dev. Biol. 13, 779–808[CrossRef][Medline] [Order article via Infotrieve]
  4. Finbow, M. E., and Harrison, M. A. (1997) Biochem. J. 324, 697–712[Medline] [Order article via Infotrieve]
  5. Arai, H., Terres, G., Pink, S., and Forgac, M. (1988) J. Biol. Chem. 263, 8796–8802[Abstract/Free Full Text]
  6. Xu, T., Vasilyeva, E. and Forgac, M. (1999) J. Biol. Chem. 274, 28909–28915[Abstract/Free Full Text]
  7. Wilkens, S. (2001) Cell Biochem. Biophys. 34, 191–208[CrossRef][Medline] [Order article via Infotrieve]
  8. Grüber, G., Wieczorek, H., Harvey, W. R., and Müller, V. (2001) J Exp Biol. 204, 2597–2605[Abstract/Free Full Text]
  9. Imamura, H., Nakano, M., Noji, H., Muneyuki, E., Ohkuma, S., Yoshida, M., and Yokoyama, K. (2003) Proc. Natl. Acad. Sci. U. S. A. 100, 2313–2315
  10. Hirata, T., Iwamotot-Kihara, A., Sun-Wada, G.-H., Okajima, T., Wada, Y., and Futai, M. (2003). J. Biol. Chem. 277,
  11. Boekema, E. J., Ubbink-Kok, T., Lolkema, J. S., Brisson, A. and Konings, W. N. (1997) Proc. Natl Acad. Sci. U. S. A. 94, 14291–14293[Abstract/Free Full Text]
  12. Wilkens, S., Vasilyeva, E., and Forgac, M. (1999). J. Biol. Chem. 274, 31804–31810[Abstract/Free Full Text]
  13. Domgall, I., Venzke, D., Lüttge, U., Ratajczak, R., and Böttcher, B. (2002) J. Biol. Chem. 277, 13115–13121[Abstract/Free Full Text]
  14. Arata, Y., Baleja, J. D., and Forgac, M. (2002a). J. Biol. Chem. 277, 3357–3363[Abstract/Free Full Text]
  15. Arata, Y., Baleja, J. D. and Forgac, M. (2002b) Biochemistry 41, 11301–11307[CrossRef][Medline] [Order article via Infotrieve]
  16. Landolt-Marticorena, C., Williams, K. M., Correa, J., Chen, W., and Manolson, M. F. (2000) J. Biol. Chem. 275, 15449–15457[Abstract/Free Full Text]
  17. Kane, P. M. (1995) J. Biol. Chem. 270, 17025–17032[Abstract/Free Full Text]
  18. Sumner, J. P., Dow, J. A., Earley, F. G., Klein, U., Jäger, D., and Wieczorek, H. (1995) J. Biol. Chem. 270, 5649–5653[Abstract/Free Full Text]
  19. Zhang, J., Myers, M., and Forgac, M. (1992) J. Biol. Chem. 267, 9773–9778[Abstract/Free Full Text]
  20. Xie, X. S., and Stone, D. K. (1988) J. Biol. Chem. 263, 9859–9867[Abstract/Free Full Text]
  21. Kane, P. M., and Parra, K. J. (2000) J. Exp. Biol. 203, 81–87[Abstract]
  22. Kawasaki-Nishi, S., Bowers, K., Nishi, T., Forgac, M., and Stevens, T. H. (2001) J. Biol. Chem. 276, 47411–47420[Abstract/Free Full Text]
  23. Yokoyama, K., Nagata, K., Imamura, H., Ohkuma, S., Yoshida, M., and Tamakoshi, M. (2003) J. Biol. Chem. 278, 42686–42691[Abstract/Free Full Text]
  24. Ho, M. N., Hirata, R., Umemoto, N., Ohya, Y., Takatsuki, A., Stevens, T. H., and Anraku, Y. (1993) J. Biol. Chem. 268, 18286–18292[Abstract/Free Full Text]
  25. Xie, X. S., Crider, B. P., Ma, Y. M., and Stone, D. K. (1994) J. Biol. Chem. 269, 25809–25815[Abstract/Free Full Text]
  26. Parra, K. J., Keenan, K. L., and Kane, P. M. (2000). J. Biol. Chem. 275, 21761–21767[Abstract/Free Full Text]
  27. Lu, X., Yu, H., Liu, S. H., Brodsky, F. M. and Peterlin, B. M. (1998) Immunity 8, 647–656[CrossRef][Medline] [Order article via Infotrieve]
  28. Geyer, M., Yu, H., Mandic, R., Linnemann, T., Zheng, Y. H., Fackler, O. T., and Peterlin, B. M. (2002) J. Biol. Chem. 277, 28521–28529[Abstract/Free Full Text]
  29. Zhang, X., Malhotra, R., and Guidotti, G. (2000) J. Biol. Chem. 275, 35592–35599[Abstract/Free Full Text]
  30. Zhou, Z., Peng, S.-B., Crider, B. P., Andersen, P., Xie, X.-S., and Stone, D. K. (1999) J. Biol. Chem. 274, 15913–15919[Abstract/Free Full Text]
  31. Grüber, G., Radermacher, M., Ruiz, T., Godovac-Zimmermann, J., Canas, B., Kleine-Kohlbrecher, D., Huss, M., Harvey, W. R., and Wieczorek, H. (2000) Biochemistry 39, 8609–8616[CrossRef][Medline] [Order article via Infotrieve]
  32. Wilkens, S., and Forgac, M. (2001) J. Biol. Chem. 276, 44064–44068[Abstract/Free Full Text]
  33. Zhang, Z., Charsky, C., Kane, P. M., and Wilkens, S. (2003) J. Biol. Chem. 278, 47299–47306[Abstract/Free Full Text]
  34. Arai, H., Berne, M., Terres, G., Terres, H., Puopolo, K., and Forgac, M. (1987) Biochemistry 26, 6632–6638[CrossRef][Medline] [Order article via Infotrieve]
  35. van Heel, M., Harauz, G., Orlova, E. V., Schmidt, R., and Schatz, M. (1996) J. Struct. Biol. 116, 17–24[CrossRef][Medline] [Order article via Infotrieve]
  36. Dube, P., Tavares, P., Lurz, R., and van Heel, M. (1993) EMBO J. 12, 1303–1309[Medline] [Order article via Infotrieve]
  37. Böttcher, B., Wynne, S. A., and Crowther, R. A. (1997) Nature 386, 88–91[CrossRef][Medline] [Order article via Infotrieve]
  38. Schatz, M., Orlova, E. V., Dube, P., Jager, J., and van Heel, M. (1995) J. Struct. Biology 114, 28–40[CrossRef][Medline] [Order article via Infotrieve]
  39. Abrahams, J. P., Leslie, A. G., Lutter, R., and Walker, J. E. (1994) Structure at 2.8 Å resolution of F1-ATPase from bovine heart mitochondria. Nature 370, 621–628[CrossRef][Medline] [Order article via Infotrieve]
  40. Sagermann, M., Stevens, T. H., and Matthews, B. W. (2000) Crystal structure of the regulatory subunit H of the V-type ATPase of Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. U. S. A. 98, 7134–7139
  41. Golas, M. M., Sander, B., Will, C. L., Luhrmann, R., and Stark, H. (2003) Science 300, 980–984[Abstract/Free Full Text]
  42. Orlova, E. V., Papakosta, M., Booy, F. P., van Heel, M., and Dolly, J. O. (2003) J. Mol. Biol. 28, 1005–1012
  43. Rubinstein, J.l., Walker, J. E., and Henderson, R. (2003) EMBO J. 22, 6182–6192[CrossRef][Medline] [Order article via Infotrieve]
  44. Merritt, E. A., and Bacon, D. J. (1997) Methods Enzymol. 277, 505–524[Medline] [Order article via Infotrieve]
  45. Iwata, M., Imamura, H., Stambouli, E., Ikeda, C., Tamakoshi, M., Nagata, K., Makyio, H., Hankamer, B., Barber J, Yoshida, M., Yokoyama, K., and Iwata, S. (2004) Proc. Natl. Acad. Sci. U. S. A. 101, 59–64[Abstract/Free Full Text]
  46. Xie, X. S. (1996) J. Biol. Chem. 271, 30980–30985[Abstract/Free Full Text]
  47. Supekova, L., Sbia, M., Supek, F., Ma, Y., and Nelson, N. (1996) J. Exp. Biol. 199, 1147–1156[Abstract]
  48. Wieczorek, H., Grüber, G., Harvey, W. R., Huss, M., Merzendorfer, H., and Zeiske, W. (2000) J. Exp. Biol. 203, 127–135[Abstract]
  49. Lu, M., Vergara, S., Zhang, L., Holliday, L. S., Aris, J., and Gluck, S. L. (2002) J. Biol. Chem. 277, 38409–38415[Abstract/Free Full Text]
  50. Rizzo, V. F., Coskun, U., Radermacher, M., Ruiz, T., Armbrüster, A., and Grüber, G. (2003) J. Biol. Chem. 278, 270–275[Abstract/Free Full Text]
  51. MacLeod, K. J., Vasilyeva, E., Baleja, J. D., and Forgac, M. (1998) J. Biol. Chem. 273, 150–156[Abstract/Free Full Text]
  52. MacLeod, K. J., Vasilyeva, E., Merdek, K., Vogel, P. D., and Forgac, M. (1999) J. Biol. Chem. 274, 32869–32874[Abstract/Free Full Text]
  53. Vasilyeva, E., Liu, Q., MacLeod, K. J., Baleja, J. D., and Forgac, M. (2000) J. Biol. Chem. 275, 255–260[Abstract/Free Full Text]
  54. Kraulis, P. J. (1991) J. Appl. Cryst. 24, 946–950[CrossRef]

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea