Crystal Structure of A3B3 Complex of V-ATPase from ... - BioMedSearch

3 downloads 0 Views 411KB Size Report
Jul 16, 2009 - Crystal Structure of A3B3 Complex of V-ATPase from. Thermus thermophilus. Megan Maher, Satoru Akimoto, Momi Iwata, Koji Nagata, Yoshiko ...
The EMBO Journal Review Process File - EMBO-2009-71721

Manuscript EMBO-2009-71721

Crystal Structure of A3B3 Complex of V-ATPase from Thermus thermophilus Megan Maher, Satoru Akimoto, Momi Iwata, Koji Nagata, Yoshiko Hori, Masasuke Yoshida, Shigeyuki Yokoyama, So Iwata Corresponding author: Ken Yokoyama, Tokyo Institute of Technology

Review timeline:

Submission date: Editorial Decision: Revision received: Additional Correspondence: Accepted:

24 June 2009 16 July 2009 29 September 2009 30 September 2009 01 October 2009

Transaction Report: (Note: With the exception of the correction of typographical or spelling errors that could be a source of ambiguity, letters and reports are not edited. The original formatting of letters and referee reports may not be reflected in this compilation.)

1st Editorial Decision

16 July 2009

Thank you for submitting your manuscript to the EMBO Journal. Your study has now been seen by three experts in the field and their comments to the authors are provided below. As you can see the referees appreciate the structural analysis on the V-type ATPase and find that it is an important contribution to the field. While referees #2 and 3 raise relative minor concerns, referee #1 raises significant ones that have to be resolved before further consideration here. Should you be able to address the criticisms in full, we would consider a revised manuscript. I should remind you that it is EMBO Journal policy to allow a single round of revision only and that, therefore, acceptance or rejection of the manuscript will depend on the completeness of your responses included in the next, final version of the manuscript. Thank you for the opportunity to consider your work for publication. I look forward to your revision. Yours sincerely, Editor The EMBO Journal REFEREE REPORTS Referee #1 (Remarks to the Author): The manuscript describes the first reported crystal structure of a subcomplex of a bacterial V-type

© European Molecular Biology Organization

1

The EMBO Journal Review Process File - EMBO-2009-71721

ATPase, rather than structures of individual subunits. Given the many crucial cellular roles of Vtype ATPases (especially the eukaryotic enzymes), this is an area of significant interest. Unfortunately, the subcomplex described is lacking both bound nucleotides and the single copy subunits that, individually or in combination, produce the asymmetry that is an essential feature of their catalytic activity (by analogy with the closely related F-type ATPases). Nevertheless, the structure does show significant similarities with the F-type ATPases, especially in the region of the suggested catalytic site on the A subunits, and mutagenesis experiments are reported which support the predicted roles of a number of key residues in catalysis. Other aspects of the structural interpretation, concerning the likelihood of nucleotide binding to the B subunits, and the role of the Tyrosine/Serine pair located close to the catalytic site are necessarily more speculative, and the significance of the apparently swapped positions of these two residues is not clear. There are a significant number of issues that need to be addressed, which are listed explicitly below. In particular, the refinement statistics (table S1) must be presented in the main part of the paper and not relegated to the supplementary information (this could easily be merged with Table 1 on data collection statistics). It is quite unacceptable that the Rfree is not quoted anywhere in the main text. The discussion of the structural evidence that the B subunit is the non-catalytic subunit is not entirely convincing. In particular, the difference of 0.2Å in overall rmsd (1.7 vs 1.9Å) is not very significant. What is surely more important here is the arrangement of residues at the potential catalytic site (and the absence of bound nucleotide in the B subunit). The comparisons at the domain level are only quoted comparing B with alpha, not for B with beta. No mention is made of the sequence similarities. When comparing the P-loop regions, the equivalent sequence in the B subunit to the P-loop sequence should be given. A superposition of the two P-loops would make it easier to compare their conformations in Fig 3. The discussion of the difference in main chain conformation of P-loop region of subunit B compared to that of the F1 alpha subunit (with bound nucleotide), seems to assume that the B subunit P-loop could not change its conformation if nucleotide was bound. In spite of having a proline at one position, can the authors really rule out this possibility ? It should be noted that the main chain conformation of the P loop in the nucleotide-free beta subunit of F1 is different to that in the nucleotide binding subunits, so in F1 at least a conformational change does take place on nucleotide binding. In the absence of any knowledge of how nucleotide might bind to the B-subunit (which could be rather different to the way it binds to the A subunit or the alpha and beta subunits of F1) it is difficult to rationalise the results of the mutagenesis experiments on V1. It is possible that activity of V1 does not depend on nucleotide binding to the B subunits as seems to be assumed here. Indeed, it has been shown that E. coli F1 is still active in the absence of nucleotides binding to the noncatalytic alpha subunit (Weber et al., JBC 270, 21045-21049, 1995), and the reference quoted in the manuscript (Matsui et al., 1997) reports that the mutant that does not bind nucleotides at the noncatalytic subunits does indeed have catalytic activity, but this decays rapidly under high concentrations of ATP as the enzyme adopts the MgADP inhibited state. It does not seem that the statement that "it is very unlikely" that B-subunits bind nucleotide is fully justified based on the evidence presented. It was not clear that the discussion of the hydrophobic residues at the B-A interface reveals anything novel or incisive about the enzyme mechanism, particularly as the structure of the "closed" form of the A subunit is not known. The fact that a highly non-conservative mutation such as A/F230 to alanine significantly affects catalysis is not, of itself, surprising or particularly informative. It is clear that the enzyme undergoes significant conformational changes during catalysis, and that only a few of these states have been trapped in crystal structures, which makes it difficult to interpret the results of such mutagenesis experiments. The proposed role of a "hydrophobic cluster" as an important catalytic process for various ATPases seems highly speculative and is not justified on the basis of the results presented. The Rfree is rather high as mentioned in the text. Did the authors carry out any tests for twinning and are they confident about the space group assignment ? Are there any non-origin peaks in the native Patterson map, indicating translational non crystallographic symmetry (what is the relationship between the two heterodimers in the unit cell) ? Was there any evidence of disorder (streaking or unusual spots shapes) in the diffraction images. Was there significant radiation damage

© European Molecular Biology Organization

2

The EMBO Journal Review Process File - EMBO-2009-71721

in the data ? Minor comments 1. p4 line 4. It seems illogical to say that "a family of eukaryotic V-ATPases is found in prokaryotic cells", text requires modification. 2. p4 line 10. "the A subunit of the V-ATPase from T. thermophilus" 3. p4, 2nd para "In this model ...". The text is confusing because it could be assumed that the "model" refers to the binding change model (mentioned in the previous sentence) rather than to the crystal structure. 3 p4 last line. References should be provided for the EM studies. 4. p5 line 1. What are the sources of the V1 and F1 that have the quoted (~25%) sequence identity ? 5. p5 2nd para. Why is the A-B interface mainly composed of residues from the B subunit, it might have been expected to have approximately equal numbers of residues from both A and B ? 6. p6 The asymmetric unit surely contains two AB heterodimers, not two A3B3 dimers as stated. 7. p6 Average B-factors should only be stated to one decimal place. 8. p6 Fig 1. Captions for A and B need to be interchanged. The quality of the text in the figure itself is poor in my copy. 9. p7 middle. All B subunits (as well as A subunits) are described as being in an "open" conformation in A3B3, but in F1 all alpha (equivalent to B) subunits have a similar conformation which is closer to the "closed" B conformation than the "open" one, so this needs to be clarified. In addition, this comment is presumably not correct when considering the alpha3beta3 PS3 F1 structure. 10. p7. Fig 2 caption, again B and A need to be interchanged ! Fig S3 title is also incorrectly labelled S2 in the S.I. I doubt that Forgac, 2007 is the correct reference for identification of the B subunit as the noncatalytic subunit, the original paper should be referenced. 11 p9 top. Residues 73-112 in the A subunit have not been assigned to any domain, is this really the case ? The linker is quite invisible in Fig 2A, perhaps it could be shown in a different colour ? Even in Fig S3A it is not easy to follow. Sentence "In fact, the residues..." should read "In fact, some residues ..." as not all residues in the linker are conserved. 12. p10 end of 1st para: A comparison of the Pi binding site is made with the betaE conformation of the aluminium fluoride inhibited bovine F1. However, in this structure the BetaE subunit adopts a "half closed" conformation, not the "open" conformation reported for the A subunit, so it is not surprising that region of the Pi binding site is different in the two structures and it is not clear why this comparison is being made. 13 p.10, 2nd para. Fig 5 caption should state explicitly that A is V-type and B is F-type. The labelling of A for Y506 and F419 should be changed so that the two labels do not run into each other. The description of the colouring in B for Y311 and S344 is incorrect. 14 p11. top para. Because there is no nucleotide bound to the A subunit (making a direct structural comparison impossible), and only 3 mutants are considered, it would be more appropriate to say that the catalytic sites "are very similar" rather than "are almost identical" in V1 and F1. 15. p11, 2nd para. The reference given for the critical importance of F1 alpha Ser344 is based on modelling studies. It might be more appropriate to include experimental data on the significance of this residue.

© European Molecular Biology Organization

3

The EMBO Journal Review Process File - EMBO-2009-71721

16. p13 Experimental procedures. 6 lines up. The resolution is 2.8Å, not 2.5Å. What was the temperature of the data collection. If the crystals were cryocooled, how were they cryoprotected. Please give the wavelength used for data collection, the type of detector and the data processing programs used. What is the Wilson B-factor for the data ? 17. The X-ray source does not need to be specified in Table 1. The greek sigma symbols in the footnotes to table 1 have been changed to "S"'. Referee #2 (Remarks to the Author): This manuscript presents the high resolution crystal structure of the catalytic domain of the VATPase from Thermus thermophilus. The V-ATPase is a critical enzyme in the cell, which is used to maintain pH balance and acidification. While the V-ATPase has similarities with the F-type ATPases, there is much less known on the structure of the V-ATPases. This is the first high resolution crystal structure of the catalytic domain of the V-ATPase and it provides new insights and understanding into the structure function relationship of the V-ATPase. This manuscript is well written and systematically describes the structure function relationship of the A3 B3 subcomplex. This subcomplex contains the catalytic sites and there is a good discussion and comparison to that of the F1 ATPase. Importantly, site directed mutagenesis studies are included, which provide additional biochemical support for the conclusions made from the analysis of the structure. Overall, this is a new and important contribution to the understanding of the VATPase and its evolutionary relationship to the F1 ATPases. Referee #3 (Remarks to the Author): This manuscript describes the first high resolution structure for a major V-ATPase subdomain, the A3B3 subcomplex that encompasses the 3 catalytic and 3 "regulatory" subunits. Although several individual subunit or small subcomplex structures from T. thermophilus, archael ATPases (including the A and B subunits, whose structures were used for molecular replacement in this paper), and the yeast V-ATPase have been reported, the information from individual subunits is limited. The structure from the bacterial (T. thermophilus) V-ATPase reported here provides novel information about the two different interfaces between the A and B subunits, as well as defining the surfaces of the A3B3 hexamer exposed to the interior (presumably where the rotor binds) and exposed to the exterior (presumably where stator subunits bind). As expected, there is a strong similarity to the catalytic and regulatory subunits in the bovine and yeast F1 structures in the "open" βE conformation since the A3B3 has no nucleotide bound, but there are also enlightening differences. One of the most important differences is between the non-catalytic nucleotide binding sites in F1 and the comparable region of the A3B3 which does not appear to be able to accommodate nucleotide. This may resolve a number of longstanding questions about the existence and function of "non-catalytic" nucleotide binding sites in V-ATPases. Comparison of the putative catalytic sites in the V-ATPase suggests fundamental similarity to the F1 catalytic mechanism, although there are some variations on general themes, such as the arrangement and identity of hydrophobic residues that may help "seal" the catalytic pocket when nucleotide binds. Overall, this structure provides no huge surprises, but it is an important contribution to the field and platform for further work. I would recommend only the following relatively minor changes to the manuscript. 1) Inclusion of the E. coli F1β and F1α sequences in Fig. S1 would be helpful even though there is not yet a high resolution structure because so much of the mutagenesis information is from E. coli. Also, p. 4, 2nd paragraph, line 5 should read "The isolated E. coli F1 which has the subunit stoichiometry..." because the next sentence refers to the bovine enzyme, which has a somewhat different subunit composition. 2) One interesting conclusion from this structure is that the "non-homologous" region of the A subunit encompasses not only the obvious bulge in the A subunits that has been observed by EM, but also a linker that runs for much of the exterior face of the hexamer. This may be worth further comment. The text says that "the linker region is involved in assembly of the complex", but in the

© European Molecular Biology Organization

4

The EMBO Journal Review Process File - EMBO-2009-71721

Shao et al paper referred to in this section (p. 9, line 13), yeast mutants with substitutions in this region were shown to assemble, but to show poor coupling of ATP hydrolysis and proton transport. Also in Figure 2, the prolines from the linker region are highlighted in purple, but the legend does not explain what they are. 3) The authors relate the T. thermophilus structure reported here to existing information about A and B subunit mutations in the yeast enzyme, and this adds significance to the paper. However, there is some confusion in the references to the yeast A and B subunit mutations. On p. 10, last line, the A subunit mutation described was characterized in Liu and Kane (1996) Biochemistry 35: 1093810948. On the top of p. 11, a reference to mutation of the yeast B/R381 should be included (this ref. would be the Liu et al (1996) ref currently in the reference list). Yeast B/R381 is equivalent to T. thermophilus B/R373 and gives similar results when mutated. 4) Disulfide bond formation between two conserved cysteines in the A subunit has been proposed as a means of regulating V-ATPases by Feng and Forgac. Although these two cysteines are not present in this structure because they were mutated to serine to assist in the crystallization, it would be worth commenting on whether the structure suggests they could be in proximity in either a single A subunit or two different A subunits of the hexamer. 4) p. 9, line 8 from bottom: Because there are no bound nucleotides in this structure, it may be premature to say "A similar conformation change should be induced in the A subunit by nucleotide binding...". It might be more appropriate to say such a change "may" be induced.

1st Revision - authors' response

29 September 2009

Our response to reviewer’s comments: Reviewer #1 1-1. Reviewers comment: There are a significant number of issues that need to be addressed, which are listed explicitly below. In particular, the refinement statistics (table S1) must be presented in the main part of the paper and not relegated to the supplementary information (this could easily be merged with Table 1 on data collection statistics). It is quite unacceptable that the Rfree is not quoted anywhere in the main text. 1-1 Our response: In the revised manuscript, the refinement statistics have been incorporated into Table 1, which is in the main body of the paper. In addition, we have included measures of the quality of the final model, such as the residuals R and Rfree and average B factors, in the main body of the text (Experimental Procedures and Results and Discussion, respectively). 1-2. Reviewers comment: The discussion of the structural evidence that the B subunit is the non-catalytic subunit is not entirely convincing. In particular, the difference of 0.2A ; in overall rmsd (1.7 vs 1.9A) is not very significant. What is surely more important here is the arrangement of residues at the potential catalytic site (and the absence of bound nucleotide in the B subunit). The comparisons at the domain level are only quoted comparing B with alpha, not for B with beta. No mention is made of the sequence similarities. When comparing the P-loop regions, the equivalent sequence in the B subunit to the P-loop sequence should be given. A superposition of the two P-loops would make it easier to compare their conformations in Fig 3. 1-2 Our response: We agree that the comparison of the overall structure of the B subunit with either the α or β subunit of F1 alone, is not convincing to assess function of B subunit. Accordingly, we have omitted the followed sentence: P7; This comparison structurally supports the idea that the V-ATPase B subunit is non-catalytic and

© European Molecular Biology Organization

5

The EMBO Journal Review Process File - EMBO-2009-71721

equivalent to the F1-α subunit. In addition, we have elaborated on our analysis to compare the structures of the P-loop regions of the yeast F1, α subunit and the equivalent region of the Thermus V-ATPase B subunit, by superposition (Figure 3B). This includes a residue-by-residue comparison of sequence (which is reinforced by the sequence alignments in Figure S1), which shows that the P-loop sequence (GXXXXGKT/S, Walker motif A) is not conserved in the V-ATPase B subunit. As detailed on page 8, both this structural analysis, the lack of comparable nucleotide-binding residues at the A-B interface of V-ATPase, and the fact that site-directed mutagenesis of residues at the A-B interface neither effect the rate or profile of ATP hydrolysis by V1, support the idea that the A-B interface is non-catalytic. 1-3. Reviewers comment: The discussion of the difference in main chain conformation of P-loop region of subunit B compared to that of the F1 alpha subunit (with bound nucleotide) seems to assume that the B subunit P-loop could not change its conformation if nucleotide was bound. In spite of having a proline at one position, can the authors really rule out this possibility? It should be noted that the main chain conformation of the P loop in the nucleotide-free beta subunit of F1 is different to that in the nucleotide binding subunits, so in F1 at least a conformational change does take place on nucleotide binding. 1-3. Our response; As detailed above and on page 8 of the manuscript, the structure of loop region (H3) in the B subunit is strictly different to that of the α subunit. In addition, the loop region in B subunit is not hydrophilic. Thus, it is likely that structure of this loop region in the B subunit is unfavorable for nucleotide binding in a manner similar to that of the α subunit. Accordingly, we have amended the text as follow; P8 last paragraph; In conclusion, it is unlikely that the B subunit binds a nucleotide at the A-B interface in a similar manner to that of the α subunit. 1-4. Reviewers comment: In the absence of any knowledge of how nucleotide might bind to the B-subunit (which could be rather different to the way it binds to the A subunit or the alpha and beta subunits of F1) it is difficult to rationalise the results of the mutagenesis experiments on V1. It is possible that activity of V1 does not depend on nucleotide binding to the B subunits as seems to be assumed here. Indeed, it has been shown that E. coli F1 is still active in the absence of nucleotides binding to the noncatalytic alpha subunit (Weber et al., JBC 270, 21045-21049, 1995), and the reference quoted in the manuscript (Matsui et al., 1997) reports that the mutant that does not bind nucleotides at the noncatalytic subunits does indeed have catalytic activity, but this decays rapidly under high concentrations of ATP as the enzyme adopts the MgADP inhibited state. It does not seem that the statement that "it is very unlikely" that B-subunits bind nucleotide is fully justified based on the evidence presented. 1-4 Our response: As described in main text and preceding responses (1-2 and 1-3, above), our structure of the A3B3 complex suggests that the B subunit does not include a similar potential nucleotide binding site to that of α subunit in F-ATPase. This is supported by our previous analyses (Yokoyama et al, 1998 and Imamura et al, 2006). The work referred to by the referee (Mastui et al, 1998 describes a mutant F1 protein (alanine substitution at α /Q172, α K175 and α /T176), which shows initial burst of activity, which rapidly decays and is completely lost within a few seconds. This dramatic change is due to the loss of nucleotide binding activity at the subunits. In our results, the mutant V1 does not show a similar change to that of the F1 protein. Thus, it is safe to conclude that the B subunit cannot bind nucleotide in a similar manner to that of the F1-ATPase. To reinforce this analysis, we have amended the text (to include the facts that neither the rate or the profile of ATP hydrolysis by V1 are effected by mutagenesis) and added profiles of the ATP hydrolysis by both wild and mutant V1-ATPases in Figure S6. Taken together, several lines of evidence, including the structure, biochemical studies, and the mutagenesis studies, strongly suggest that the B subunit contains no nucleotide binding sites. 1-5. Reviewers comment:

© European Molecular Biology Organization

6

The EMBO Journal Review Process File - EMBO-2009-71721

It was not clear that the discussion of the hydrophobic residues at the B-A interface reveals anything novel or incisive about the enzyme mechanism, particularly as the structure of the "closed" form of the A subunit is not known. The fact that a highly non-conservative mutation such as A/F230 to alanine significantly affects catalysis is not, of itself, surprising or particularly informative. It is clear that the enzyme undergoes significant conformational changes during catalysis, and that only a few of these states have been trapped in crystal structures, which makes it difficult to interpret the results of such mutagenesis experiments. The proposed role of a "hydrophobic cluster" as an important catalytic process for various ATPases seems highly speculative and is not justified on the basis of the results presented. 1-5 Our response: As described in the main text, it is likely that the A subunit, in the presence of nucleotide, adopts a similar closed conformation to that observed for the β subunit. Based on the significant structural similarity between the A subunit and the β subunit, it follows that this conformation change leads to the clustering of several hydrophobic side chains (including A/F230, B/Y331, A/P381). Indeed, mutagenesis studies, including the substitution of A/230 to alanine, strongly suggest that this hydrophobic side chain plays an important role in the catalysis, even though A/F230 is not directly involved in hydrolysis/synthesis of ATP. In addition, mutagenesis studies for F1, such as the mutagenesis of β /Y311 to serine, support this idea. We are confident in this analysis as it applies to comparisons between the V-ATPase and F1 enzymes, however, we concede that the statement that "hydrophobic cluster formation seems an important catalytic process for various ATPases." is somewhat speculative. Accordingly we have amended the sentence as follows; p13, last sentence, Their proposal supports our idea that hydrophobic cluster formation is important for the ATPase reaction in rotary ATPases. 1-6. Reviewers comment: The Rfree is rather high as mentioned in the text. Did the authors carry out any tests for twinning and are they confident about the space group assignment ? Are there any non-origin peaks in the native Patterson map, indicating translational non crystallographic symmetry (what is the relationship between the two heterodimers in the unit cell)? Was there any evidence of disorder (streaking or unusual spots shapes) in the diffraction images. Was there significant radiation damage in the data ? 1-6 Our response: The apparent space group of the crystals is P321. The packing of the molecular replacement solution is reasonable and a non-origin peak, representing the translational symmetry was observed in the native Patterson (0.66, 0.33, 0.02). Prior to the original submission, we refined the structure as twinned P3 but this did not improve the refinement statistics. Thus we can confirm the space group of the crystal is, indeed, P321. After the referee's suggestion, we, however, tested other possibilities. Using the new version of the CCP4 refinement program Refmac5.5, which allows for the analsys and consideration of twinning during refinement, we found that a small fraction of the crystals (15%) is twinned with the operation (-h,-k,l) (Figure S3C). This improved the residuals of the final model, R and Rfree, to 25.0% (from 27.3%) and 28.1% (from 32.8%), respectively. Following this refinement, we have checked through the model to confirm that the original description, analysis and discussion of the structure still stands. We have amended the text to include: (1) Figure S3, which illustrates the packing of A3B3 molecules in the crystal including the contribution from the 15% twinned fraction; (2) updated statistics in Table 1, which give the residuals for the model after twin refinement; (3) analysis of the translational symmetry by native Patterson, details of the twinning analysis and refinement and details of crystal cryoprotection and data collection (the changes in text are detailed below). We would like to thank the referee for his constructive comment on this issue. p6, middle; The structure was refined using the data up to 2.8 Å resolution. The data collection and refinement statistics are summarized in Table I. A small fraction of the crystal (15%) is twinned with a twinning operator of (-h, -k, l) and this contribution was included during the final refinement of the

© European Molecular Biology Organization

7

The EMBO Journal Review Process File - EMBO-2009-71721

model. p14, line 8; By adding 20% glycerol, the reservoir solution was used to flash-freeze the crystals in liquid nitrogen. X-ray data were collected at 100 K using an ADSC Quantum-4 CCD detector with a wavelength of 0.933 Å and 0.5° oscillation steps over a range of 120° at the ESRF ID14-2 beamline (Grenoble, France). The data set was processed to 2.8 Å resolution using the programs Denzo and Scalepack (Otwinowski & Minor, 1997) and the CCP4 program suit (Collaborative Computational Project, Number 4, 1994.). The crystal belonged to the space group P321, with unit cell dimensions a = b = 199.4 Å, and c = 179.0 Å. Average B-factor estimated from the Wilson plot was 62.5 Å2. The 2.8 Å resolution structure of the A3B3 subcomplex was solved by molecular replacement with PHASER (McCoy et al, 2007), using partial structures of the A and B subunits from archea (PDB code 1VDZ and 2C61) as search models (Maegawa et al, 2006, Schafer et al, 2006). A single solution with two AB heterodimers located close to the crystallographic 3-fold axis was identified (Figure S3A, B). The two dimers were related by a translational symmetry, which is confirmed by peaks at (0.66,033,0.02) and equivalent positions in the native Patterson map. Preliminary rigid body and restrained refinement with REFMAC5 (Murshudov et al., 1997) gave residuals R=0.44 and Rfree=0.48 for this solution. Model bias was minimized by solvent flattening using RESOLVE (Terwilliger, 2000) using ‘prime-and-switch’ phasing and two-fold NCS averaging. The resulting electron density maps were readily interpretable and clearly showed areas with structural differences. Manual adjustments to the model were made using COOT (Emsley and Cowtan, 2004) and O (Jones et al., 1991). Model refinement was carried out using CNS (Brunger et al., 1998) and REFMAC5 with TLS (Painter J., and Merritt EA., 2006). We found that a small fraction of the crystal (15%) is twinned with a twinning operator (-h,-k,l) and this contribution was included in the final stage of the refinement using REFMAC5 (Figure S3C). This improved the residuals R and Rfree to 25.0% (from 27.3%) and 28.1% (from 32.8%), respectively. 1-7. Reviewers comment: p4 line 4. It seems illogical to say that "a family of eukaryotic V-ATPases is found in prokaryotic cells", text requires modification. And p4 line 10. "the A subunit of the V-ATPase from T. thermophilus" 1-7 Our response: We have modified the text as follow; P4 line 4; Some prokaryotes, such as a thermophilic eubacterium, Thermus thermophilus contains a member of the family (or homolog) of the V-ATPase in their membranes. 1-8. Reviewers comment: p4, 2nd para "In this model ...". The text is confusing because it could be assumed that the "model" refers to the binding change model (mentioned in the previous sentence) rather than to the crystal structure. 1-8. Our response: We have amended the text accordingly. 1-9. Reviewers comment: p5 line 1. What are the sources of the V1 and F1 that have the quoted (~25%) sequence identity ? 1-9. Our response: The sources are A subunit from T. thermophilus and FoF1 from Homo sapiens. We added this information in the text. 1-10. Reviewers comment: p5 2nd para. Why is the A-B interface mainly composed of residues from the B subunit, it might have been expected to have approximately equal numbers of residues from both A and B ? 1-10. Our response: As the reviewer pointed out, the section is confusing. We omitted the sentence "which is mainly composed of the residues from the B subunit" from the text.

© European Molecular Biology Organization

8

The EMBO Journal Review Process File - EMBO-2009-71721

1-11. Reviewers comment: p6 The asymmetric unit surely contains two AB heterodimers, not two A3B3 dimers as stated.. 1-11. Our response: The referee is correct. The asymmetric contains two AB heterodimers. We have amended the main text accordingly. 1-12. Reviewers comment: p6 Average B-factors should only be stated to one decimal place. 1-12. Our response: Agreed. We have amended the main text accordingly. 1-13. Reviewers comment: p6 Fig 1. Captions for A and B need to be interchanged. The quality of the text in the figure itself is poor in my copy. 1-13. Our response: Thank you for the indication. We have amended the caption and text in the figure accordingly. 1-14. Reviewers comment: p7 middle. All B subunits (as well as A subunits) are described as being in an "open" conformation in A3B3, but in F1 all alpha (equivalent to B) subunits have a similar conformation which is closer to the "closed" B conformation than the "open" one, so this needs to be clarified. In addition, this comment is presumably not correct when considering the alpha3beta3 PS3 F1 structure. 1-14. Our response: Agreed. We have amended the sentence as follows; P6, first paragraph, last sentence The diameter of the cavity is larger than those observed for the α3β3 sub-domain in F1 because all A subunits in the complex are present in an “open” conformation as described below. 1-15. Reviewers comment: p7 Fig 2. Captions for A and B need to be interchanged. Fig S3 title is also incorrectly labelled S2 in the S.I. I doubt that Forgac, 2007 is the correct reference for identification of the B subunit as the non-catalytic subunit, the original paper should be referenced. 1-15. Our response: We have amended the text accordingly. Please see the revised supplemental information of Figure S3. Also we cite two papers as references for identification of the B subunit as the non catalytic subunit (Manolson et al, 1985 and Vasilyeva et al, 1996). 1-16. Reviewers comment: p9 top. Residues 73-112 in the A subunit have not been assigned to any domain, is this really the case ? The linker is quite invisible in Fig 2A, perhaps it could be shown in a different colour ? Even in Fig S3A it is not easy to follow. Sentence "In fact, the residues..." should read "In fact, some residues ..." as not all residues in the linker are conserved. 1-16. Our response: Residues 73-112 are positioned between ñbarrel and bulge domains, and are distinguished from them. Thus we do not assign the region as part of any domain. We have improved the visibility of the loop in Figure 2 and Figure S4. We have amended the figure legends and text accordingly. 1-17. Reviewers comment: p10 end of 1st para: A comparison of the Pi binding site is made with the betaE conformation of the aluminium fluoride inhibited bovine F1. However, in this structure the BetaE subunit adopts a "half closed" conformation, not the "open" conformation reported for the A subunit, so it is not surprising that region of the Pi binding site is different in the two structures and it is not clear why this comparison is being made.

© European Molecular Biology Organization

9

The EMBO Journal Review Process File - EMBO-2009-71721

1-17. Our response: Agreed. We have omitted the description about the half closed bovine site. Instead, we have amended the text as follows; p10, first paragraph, last sentence, In the yeast F1 structure, the bound Pi was coordinated by residues β/K162 (equivalent to A/K234), β/R189 (A/R258), β/D256 (A/D325), β/R260 (A/R329), and α/R373 (B/R360), β/N257 (A/S326). In contrast, the A-subunit structure shows that, all arginine residues are orientated away from the putative Pi binding site, leaving the site much more open and structurally less favorable for Pi binding (Figure 4B). 1-18. Reviewers comment: p.10, 2nd para. Fig 5 caption should state explicitly that A is V-type and B is F-type. The labelling of A for Y506 and F419 should be changed so that the two labels do not run into each other. The description of the colouring in B for Y311 and S344 is incorrect. 1-18. Our response: We have modified the figure and have corrected the legend accordingly. 1-19. Reviewers comment: top para. Because there is no nucleotide bound to the A subunit (making a direct structural comparison impossible), and only 3 mutants are considered, it would be more appropriate to say that the catalytic sites "are very similar" rather than "are almost identical" in V1 and F1. 1-19. Our response: Agreed. We amended the text accordingly. 1-20. Reviewers comment: p11, 2nd para. The reference given for the critical importance of F1 alpha Ser344 is based on modelling studies. It might be more appropriate to include experimental data on the significance of this residue. 1-20. Our response: In addition to the modeling study, structural studies also suggested the importance of this serine residue for catalysis (Abraham et al, 1994, and Mentz et al, 2001). We added these references in the text. 1-21. Reviewers comment: The X-ray source does not need to be specified in Table 1. The greek sigma symbols in the footnotes to table 1 have been changed to "S"'. 1-21. Our response: The footnote was amended accordingly. Reviewer #3 3-1 reviewers comments Inclusion of the E. coli F1β and F1α sequences in Fig. S1 would be helpful even though there is not yet a high resolution structure because so much of the mutagenesis information is from E. coli. Also, p. 4, 2nd paragraph, line 5 should read "The isolated E. coli F1 which has the subunit stoichiometry..." because the next sentence refers to the bovine enzyme, which has a somewhat different subunit composition. 3-1 Our response We have included the sequences in Fig. S1 and amended the main text suggested accordingly. 3-2 reviewers comment One interesting conclusion from this structure is that the "non-homologous" region of the A subunit

© European Molecular Biology Organization

10

The EMBO Journal Review Process File - EMBO-2009-71721

encompasses not only the obvious bulge in the A subunits that has been observed by EM, but also a linker that runs for much of the exterior face of the hexamer. This may be worth further comment. The text says that "the linker region is involved in assembly of the complex", but in the Shao et al paper referred to in this section (p. 9, line 13), yeast mutants with substitutions in this region were shown to assemble, but to show poor coupling of ATP hydrolysis and proton transport. Also in Figure 2, the prolines from the linker region are highlighted in purple, but the legend does not explain what they are. 3-2 Our response We have added more detailed discussion about the linker region addressing the referee's comment as follows. P9 first paragraph, ‘The bulge is connected to the nucleotide binding domain through an extended linker (A185-204), which starts in the space between the N-terminal and the bulge domains and runs down the surface of the A3B3 complex (Figure2A, Figure S4A). Both the bulge and this extended linker are coded in so-called "non-homologous region" (Shao et al., 2003). Interestingly, the residues within this linker (A/P188, A/P194 and A/P204) are conserved amongst V-ATPases (Figure S1A). Substitution of the equivalent residues in yeast V-ATPase results in a change in coupling efficiency (Shao et al., 2003). These results suggest that the linker region is involved in the interaction between the peripheral stalk (composed of the E and G subunits) and the A3B3 complex.’ We have also clarified importance of the proline residues in the Figure 2 legend. 3-3 reviewers comment The authors relate the T. thermophilus structure reported here to existing information about A and B subunit mutations in the yeast enzyme, and this adds significance to the paper. However, there is some confusion in the references to the yeast A and B subunit mutations. On p. 10, last line, the A subunit mutation described was characterized in Liu and Kane (1996) Biochemistry 35: 1093810948. On the top of p. 11, a reference to mutation of the yeast B/R381 should be included (this ref. would be the Liu et al (1996) ref currently in the reference list). Yeast B/R381 is equivalent to T. thermophilus B/R373 and gives similar results when mutated. 3-3 our response Agreed. We have added the references accordingly. 3-4 reviewers comment Disulfide bond formation between two conserved cysteines in the A subunit has been proposed as a means of regulating V-ATPases by Feng and Forgac. Although these two cysteines are not present in this structure because they were mutated to serine to assist in the crystallization, it would be worth commenting on whether the structure suggests they could be in proximity in either a single A subunit or two different A subunits of the hexamer. 3-4 our response We have added comment about the issue in the text accordingly. p9 last paragraph; Regulatory disulfide bond formation between conserved cysteine residues in the A subunit (A/C261 and C539 in yeast V-ATPase) has been proposed (Feng and Forgac, 1994). The equivalent cysteine residues in the T.thermophilus A subunit are A/S232 and A/S539 (mutated from A/C539), respectively. In our open structure of A subunit, the distance between Oβ atom of A/S232 and the S β atom of A/C507 is 26 Å apart. It is therefore it is unlikely that a disufide bridge exists between them. However, it is possible that side chain of A/S232 might be proximate to that of A/C507 when the A subunit adopts the closed conformation. Disulfide bond formation between two different subunits is very unlikely because the relevant residues are more than 60 Å apart. 3-5 reviewers comment p. 9, line 8 from bottom: Because there are no bound nucleotides in this structure, it may be premature to say "A similar conformation change should be induced in the A subunit by nucleotide binding...". It might be more appropriate to say such a change "may" be induced.

© European Molecular Biology Organization

11

The EMBO Journal Review Process File - EMBO-2009-71721

3-5 our response: We have amended the text as suggested.

Additional Correspondence

30 September 2009

Thank you for submitting your revised manuscript to the EMBO Journal. I asked the original Referee #1 to review the revised version and I have now received the comments back from this referee. Referee #1 appreciates the significant efforts that you have undertaken to address the raised concerns and supports publication of your study here. Referee #1 has no further comments to authors. I am therefore very pleased to proceed with the acceptance of the study for publication here. Editor The EMBO Journal

© European Molecular Biology Organization

12