Molecular Basis of ADP Inhibition of Vacuolar (V)-type ATPase/Synthase*

Background: ADP inhibition of rotary ATPases is a common mechanism to avoid wasteful ATP hydrolysis. Results: Domain swap approaches in V1 showed that domain interaction plays a key role in sensitivity of ADP inhibition. Conclusion: Increasing the affinity of V1 for phosphate correlates with reducing sensitivity to ADP inhibition. Significance: The molecular basis of ADP inhibition of V0V1 is clarified. Reduction of ATP hydrolysis activity of vacuolar-type ATPase/synthase (V0V1) as a result of ADP inhibition occurs as part of the normal mechanism of V0V1 of Thermus thermophilus but not V0V1 of Enterococcus hirae or eukaryotes. To investigate the molecular basis for this difference, domain-swapped chimeric V1 consisting of both T. thermophilus and E. hirae enzymes were generated, and their function was analyzed. The data showed that the interaction between the nucleotide binding and C-terminal domains of the catalytic A subunit from E. hirae V1 is central to increasing binding affinity of the chimeric V1 for phosphate, resulting in reduction of the ADP inhibition. These findings together with a comparison of the crystal structures of T. thermophilus V1 with E. hirae V1 strongly suggest that the A subunit adopts a conformation in T. thermophilus V1 different from that in E. hirae V1. This key difference results in ADP inhibition of T. thermophilus V1 by abolishing the binding affinity for phosphate during ATP hydrolysis.

Vacuolar-type ATPase/synthase (V 0 V 1 ) 3 functions as an ATP hydrolysis-driven proton pump that carries out acidification of cellular compartments in eukaryotes (1). A family of V 0 V 1 , sometimes referred to as the A-type ATPases or A 0 A 1 (2), is also found in archaea and some eubacteria (the prokaryotic V 0 V 1 family) (3). In most prokaryotes, such as Thermus thermophilus, the V 0 V 1 functions as an ATP synthase. However, it can also act as a primary ion pump as seen in Enterococcus hirae (4).
Similar to F-type ATP synthase (F 0 F 1 ), V 0 V 1 consists of two distinct motor subcomplexes: hydrophilic V 1 , which catalyzes either ATP hydrolysis or synthesis, and hydrophobic V 0 , which is responsible for proton translocation across membranes (see Fig. 1a). The two motors are coupled through a central rotor complex that rotates against a surrounding stator apparatus (3,(5)(6)(7).
ADP inhibition caused by entrapment of ADP at a catalytic site is believed to be a regulatory mechanism of F 0 F 1 to prevent wasteful ATP consumption when proton motive force is lost (8,9). V 0 V 1 from T. thermophilus (TthV 0 V 1 ) also exhibits sensitivity to ADP inhibition, resulting in rapid decay of the ATPase activity of the V 1 subcomplex (10,11). ADP inhibition has been investigated in F 0 F 1 , but the precise molecular mechanism remains poorly understood (8).
The prokaryotic V 1 is composed of four different subunits with a stoichiometry of A 3 B 3 DF (A 3 B 3 DG in E. hirae) (3). A 3 B 3 forms a hexameric ring in which three A and B subunits are alternately arranged (12). The D and F (G in E. hirae) subunits form the rotary shaft; the D subunit penetrates the central cavity of the A 3 B 3 ring, and the F subunit binds to the protruding part of subunit D (13)(14)(15). The catalytic reaction center resides at the A-B interface mainly on the A subunit. The A subunit consists of three subdomains given the abbreviations NT (N-terminal), NB (nucleotide-binding), and CT (C-terminal) based on their individual structures and functions (see Figs. 1b and 2) (12)(13)(14). The NT domain forms a ␤-barrel structure with adjacent subunits and is thought to be important for the formation of the A 3 B 3 hexamer (12). The NB domain includes the Walker A and B motifs that form a nucleotide-binding site (Fig.  2). Thus, the NB domain functions as a catalytic center. The CT domain interacts with the rotary shaft during rotary catalysis (13,14).
The primary sequences of the V 1 from T. thermophilus (TthV 1 ) and E. hirae (EhiV 1 ) are highly conserved (for example, the two A subunits share 57.8% sequence identity and 79.5% sequence similarity, respectively; see Fig. 2). Furthermore, the overall structures of the two V 1 subcomplexes are almost identical (13,14). However, there is a significant functional difference between the two V 1 subcomplexes. The TthV 1 rapidly lapses into an ADP inhibition state during ATP hydrolysis and ultimately loses all ATPase activity (10). This ADP inhibition mechanism of TthV 0 V 1 is advantageous as it prevents consumption of ATP when proton motive force is lost. In contrast, because the E. hirae and eukaryotic V 0 V 1 enzymes function as ion pumps coupled with continuous ATP hydrolysis, they do not exhibit sensitivity to ADP inhibition. Indeed, a rapid decrease of ATPase activity of the E. hirae and eukaryotic V 0 V 1 has not been reported (1,16).
In this study, domain swapping was applied to the V 1 -A subunit to investigate the factors that define sensitivity to ADP inhibition using V 1 from T. thermophilus and E. hirae. The results indicate that sensitivity to ADP inhibition is not due to differences in the NB domains of the two V 1 -ATPases. We have experimentally demonstrated the relationship between the affinity for P i and sensitivity to ADP inhibition of V 1 . With recent structural studies of both V 1 -ATPases (13,14), we discuss the molecular basis of the ADP inhibition mechanism of V 0 V 1 .

EXPERIMENTAL PROCEDURES
Preparation of Chimeric V 1 -The expression plasmids for chimeric V 1 were constructed by an overlap PCR-based method. DNA fragments for A subunit gene for chimeric V 1 were amplified from TthV 1 expression plasmid (17) and pCemtp18 (18), which contains EhiV 0 V 1 (ntp) operon, by PCR using oligonucleotide primers, respectively. These PCR fragments contain sequences complementary to each other. They were mixed and submitted to a further round of amplification using the 5Јand 3Ј-terminal primers used in the first PCR to produce a single fragment containing the chimeric gene sequence. The chimeric gene sequences were then digested with appropriate restriction enzymes and inserted into the corresponding region of the TthV 1 expression plasmid (17) (see Fig. 3). Escherichia coli strain BL21-CodonPlus-RP (Stratagene) was used for expression of chimeric V 1 . The chimeric V 1 constructs were isolated as described previously (17). Cells containing the expressed proteins were suspended in equilibration buffer (20 mM imidazole/HCl (pH 8.0) containing 300 mM NaCl and disrupted by sonication. After removal of debris by centrifugation at 4,500 ϫ g for 30 min, the supernatant was applied to a Ni 2ϩ affinity column (Qiagen; 3 ϫ 5 cm), the column was washed thoroughly with equilibration buffer, and bound protein was eluted with 200 mM imidazole/HCl (pH 8.0) containing 300 mM NaCl. The V 1 was exchanged into 20 mM Tris/HCl (pH 8.0), 1 mM EDTA by ultrafiltration (Millipore, Amicon Ultra), and the dialysis solution was applied to a UNOQ column (Bio-Rad) for further purification. The fractions containing chimeric  Single Molecule Observation for ATP Hydrolysis-For single molecular observation experiments, cysteine residues were introduced into the D subunit in the V 1-A011 . Then the modified V 1-A011 was biotinylated as described previously (5,11). The single molecule observation system using magnetic beads was described previously (11). Briefly, the biotinylated chimeric V 1 (ϳ1 nM) was bound to Ni 2ϩ -nitrilotriacetic acid-coated coverslips via their His tag. Then streptavidin-coated magnetic beads (nominal diameter, ϳ200 nm; Thermo Fisher Scientific) were bound to the chimeric V 1 , and finally, observation of the rotation was initiated after infusion of buffer containing Mg-ATP.
To observe the rotation under a low viscous load, instead of the magnetic beads, we used gold beads (nominal diameter, 80 nm; BBInternational, Cardiff, UK) that functionalized with Neutravidin (Pierce) and polyethylene glycol (molecular weight, 1214; Quanta BioDesign, Powell, OH) (19). The rotation was observed on an inverted microscope (IX71, Olympus, Tokyo, Japan) using a 100ϫ objective lens (UPlanApo; numerical aperture, 1.35; Olympus) and a condenser unit (U-UCD8, U-TLD, Olympus) with a stable microscope stage (KS-O, ChuukoushaSeisakujo, Tokyo, Japan) that reduces thermal drift. The bead images were captured with an sCMOS camera (Neo, Andor, Tokyo, Japan) at 1000 frames s Ϫ1 and analyzed by NIH ImageJ with home-made plug-ins (created by Dr. K. Adachi) (20). The centroids of the bead images were calculated as described previously (21).
Other Assays-Protein concentrations of the purified V 1 constructs were determined from UV absorbance calibrated by quantitative amino acid analysis; 1 mg/ml gives 0.88 OD at 280 nm. ATPase activity was measured at 25°C with an enzymecoupled ATP-regenerating system. The ATPase assay solution contained 50 mM Tris-HCl (pH 8.0), 100 mM KCl, 6 mM MgCl 2 , 2 mM phosphoenolpyruvate, 100 g/ml lactate dehydrogenase, 100 g/ml pyruvate kinase, 0.2 mM NADH, and a range of concentrations of Mg-ATP. The reactions were initiated by the addition of enzymes. ADP removal treatments for chimeric V 1 were carried out by methods described previously (11). Polyacrylamide gel electrophoresis in the presence of SDS or alkyl ether sulfate (AES) was carried out as described previously (22). The TthV 1 and TthV 0 were isolated from the membranes from the mutated T. thermophilus strain incorporating a His 3 tag on the C terminus of subunit L as described previously (23). The reconstituted chimeric V 0 V 1 was then incorporated into liposomes by a freeze-thaw method (24). The liposomes were used for ATP synthesis activities driven by acid-base transition as described previously (11). The condition used is as follows: pH in ϭ 4.9, pH out ϭ 8.5, [K ϩ ] in ϭ 1 mM, [K ϩ ] out ϭ 100 mM, 25°C. Figures of structural models were prepared using the UCSF Chimera package (25).

RESULTS
ATP Hydrolysis of Chimeric V 1 -All chimeric V 1 constructs were generated using the basic TthV 1 (Figs. 1, b and c, and 3). The subunit stoichiometry and complex formation of all the purified chimeric V 1 constructs were confirmed by both SDSand native PAGE (Fig. 1, d and e). The NB domain is thought to be critical for entrapment of inhibitory ADP at a catalytic site. ADP inhibition of TthV 1 can be overcome by introduction of the TSSA mutation in the P-loop region in the NB domain of the A subunit (S232A/T235S; see Refs. 5 and 11). In this case, entrapped inhibitory ADP in the catalytic site of TSSA mutated TthV 1 was easily detached from the enzyme during turnover because of its lower affinity for nucleotide (ADP and/or ATP) than that of wild-type TthV 1 . Assuming that the insensitivity of EhiV 1 to ADP inhibition is due to a difference in the molecular characteristics of the NB domain of the two enzymes, chimeric V 1-A010 containing the NB domain (residue numbers 194 -434; Figs. 1, b and c, 2, and 3) of the A subunit of EhiV 1 should show insensitivity to ADP inhibition. However, the ATP hydrolysis activity of isolated V 1-A010 was very low due to the presence of inhibitory ADP bound to the enzyme (Fig. 4). This inhibitory ADP in V 1-A010 was successfully removed by phosphate/EDTA treatment as described previously (17). After the treatment, V 1-A010 showed an apparent ATP hydrolysis activity of 34.4 Ϯ 0.6 s Ϫ1 , which is almost equivalent to that of TthV 1 (39.9 Ϯ 0.3 s Ϫ1 ) (11). However, the ATPase activity of V 1-A010 rapidly decayed due to transition into the ADP inhibition state. The inhibition rate of the V 1-A010 was 1.4 ϫ 10 Ϫ2 s Ϫ1 , which is higher than that of TthV 1 (7.9 ϫ 10 Ϫ3 s Ϫ1 ; see Ref. 11) (Fig. 4). These results clearly indicate that the different sensitivities of the two V 1 enzymes to ADP inhibition are not solely dependent on the properties of the NB domain.
A CT domain chimera, V 1-A001 (containing residue numbers 435-593 of the A subunit of EhiV 1 ; Figs. 1c and 3), was constructed and purified (Fig. 1, c and d). The isolated V 1-A001 exhibited extremely low ATP hydrolysis activity likely as the result of bound inhibitory ADP (Fig. 4), indicating that the V 1-A001 is sensitive to ADP inhibition. Following ADP removal, V 1-A001 exhibited ATPase activity of 5.0 Ϯ 0.2 s Ϫ1 , which is significantly lower than that of V 1-A010 or TthV 1 . The ATPase activity of V 1-A001 also decayed. The results indicate that sub-  The time course of ATP hydrolysis by chimeric V 1 at 2 mM ATP is shown. ATP hydrolysis activity was measured by monitoring the absorbance decrease at 340 nm as described under "Experimental Procedures." The reactions were initiated by addition of 20 l of 1 M enzymes without heat treatment to 2 ml of assay mixture (black lines). In addition, V 1-A010 and V 1-A001 were subjected to nucleotide removal treatment and then applied to the ATPase assay (red lines). mAU, milliarbitrary units. stitution of the CT domain of TthV 1 to that of EhiV 1 is not sufficient to suppress ADP inhibition of V 1 .
In contrast, the chimeric V 1-A011 containing both the NB and CT domains of A subunit of EhiV 1 maintained nearly continuous ATP hydrolysis activity of 57.7 Ϯ 0.8 s Ϫ1 without ADP removal treatment (Fig. 4). These results from three chimeric V 1 indicate that substitution with both the NB and CT domains of EhiV 1 -A subunit is required to suppress the ADP inhibition of the TthV 1 .
The first and second helices of the CT domain lie in close proximity to the lower side of the NB domain in the crystal structure of TthV 1 (Fig. 1b), suggesting that these helices play an important role in the interaction between the two domains. To investigate the inherent function of the CT domain helices in ADP inhibition, chimeric construct V 1-A010.1 , which contains both the NB domain and the first and second CT helices (residue numbers 435-481; Figs. 1c and 3) of the A subunit of EhiV 1 , was constructed and purified (Fig. 1, c and d). V 1-A010.1 showed continuous ATP hydrolysis activity without ADP removal treatment (Fig. 4), indicating that both of these regions of the A subunit of EhiV 1 are required to overcome ADP inhibition of TthV 1 . In other words, interdomain interaction between the NB domain and the first and second helices of the CT domain is likely to be important for suppression of ADP inhibition.
Kinetic Parameters of Chimeric V 1 for ATP Hydrolysis-Affinity of V 1 for nucleotides seems to be key for sensitivity of ADP inhibition because the ADP inhibition is due to entrapment of ADP in a catalytic site of the V 1 . In fact, TSSA mutated TthV 1 exhibited insensitivity to ADP inhibition because of its lower affinity for nucleotide (ADP and/or ATP) than that of wild-type TthV 1 (5,11). Thus, we examined kinetic parameters of V 1-A011 , which exhibits insensitivity to ADP inhibition, and other chimeric V 1 by both bulk phase assay and single molecule observation. The ATP hydrolysis rates of V 1-A010 and V 1-A011 obeyed simple Michaelis-Menten kinetics (Fig. 5a). The V max values of V 1-A010 and V 1-A011 were calculated to be 34.4 Ϯ 0.6 and 57.7 Ϯ 0.8 s Ϫ1 (mean Ϯ S.E.), respectively (Table  1), which are similar to that of TthV 1 . The K m for ATP (K m(ATP) ) of V 1-A011 , 21.5 Ϯ 1.8 M (mean Ϯ S.E.), was 10-fold lower than that of TthV 1 (205 Ϯ 7 M (mean Ϯ S.E.)) and 6-fold lower than that of V 1-A010 (132 Ϯ 11 M (mean Ϯ S.E.)) ( Table 1). The V max /K m value of V 1-A011 was ϳ14-fold higher than that of TthV 1 despite that V max of V 1-A011 was 1.5-fold higher than that of TthV 1 . These results suggest that V 1-A011 has a higher affinity for ATP compared with TthV 1 and V 1-A010 .
To obtain the kinetic parameters of chemical reactions in the ATP hydrolysis, we observed the rotations of single molecule V 1-A011 (Fig. 5b) under an optical microscope. A bead (nominal

Molecular Basis of ADP Inhibition of V-ATPase
JANUARY 3, 2014 • VOLUME 289 • NUMBER 1

JOURNAL OF BIOLOGICAL CHEMISTRY 407
diameter, ϳ200 nm) attached to the D subunit of V 1-A011 rotated stepwise at 2 and 0.5 M ATP concentration, pausing every 120°, as observed for TthV 1 (11) (Fig. 5b). The dwell time between successive 120°steps in V 1-A011 at low ATP concentrations is a result of the enzyme waiting to bind ATP because binding of ATP is the rate-limiting step under these conditions (11). Based on dwell time analysis (Fig. 5, c and d), the apparent binding rate, k on , for ATP of V 1-A011 was estimated to be (3.5 Ϯ 0.1) ϫ 10 6 and (5.1 Ϯ 0.2) ϫ 10 6 M Ϫ1 s Ϫ1 (mean Ϯ S.E.) at 2 and 0.5 M ATP, respectively. These values are 3-5-fold higher than that reported previously for TthV 1 (11). These results indicate a higher affinity of V 1-A011 for ATP, consistent with the result obtained for the bulk ATP hydrolysis assay ( Fig. 5a and Table 1). ATP Synthesis by Chimeric Enzymes-For the ATP synthesis assay, V 1-A010 , V 1-A011 , and V 1-A001 were individually reconstituted with TthV 0 , respectively. Reconstitution of each chimeric V 1 and TthV 0 was confirmed by alkyl ether sulfate-PAGE (Fig.  1e). Each reconstituted chimeric V 0 V 1 complex was individually incorporated into liposomes, and the ATP synthesis activities were measured. All the chimeric V 0 V 1 complexes exhibited ATP synthesis activity ( Fig. 6 and Table 2). The ATP synthesis rates were plotted against ADP concentration ([ADP]). The plots obeyed roughly Michaelis-Menten behavior (Fig. 6, a-d, and Table 2). The apparent K m values for ADP (K m(ADP) ) of V 0 V 1-A010 (14.0 Ϯ 3.0 M) and V 0 V 1-A001 (13.5 Ϯ 2.3 M) were slightly higher than that of was lower than those obtained for TthV 0 V 1 and the other chimeric V 0 V 1 complexes. These results indicate that V 0 V 1-A011 has a high affinity for ADP, suggesting that there is no direct relationship between sensitivity to ADP inhibition of V 1 and the affinity for nucleotides.
The rates of ATP synthesis of the chimeric V 0 V 1 were also measured as a function of the phosphate (P i ) concentration (Fig. 6, e-h, and Table 2). Interestingly, the apparent K m value for P i (K m(Pi) ) for V 0 V 1-A011 (10.6 Ϯ 3.5 M) was markedly lower than those of TthV 0 V 1 and the other two chimeric V 0 V 1 complexes, indicating that V 1-A011 has a much higher affinity for P i than the other constructs. P i Effect on ATP Hydrolysis of Chimeric V 1 -To confirm the change in affinity for P i of V 1-A011 , the effects of P i on ATP hydrolysis were examined by both bulk phase and single molecule observation. The ATPase activities of V 1-A011 were measured in the presence of various P i concentrations. The ATPase activities of TthV 1 , V 1-A010 , and V 1-A001 were not affected at low P i concentrations (1-30 mM). In marked contrast, the ATPase activity of V 1-A011 was strongly inhibited at low P i concentrations with an IC 50 of ϳ10 mM (Fig. 7). This result indicates that V 1-A011 has higher affinity for P i than the other constructs, consistent with the results of ATP synthesis experiments. V 1-A011 was subjected to single molecule measurement using an 80-nm gold bead for which viscous drag is very low (19,26). ATP dependence of the time-averaged rotation rate of V 1-A011 follows simple Michaelis-Menten kinetics (Fig. 8a). The V max value was 28.3 Ϯ 0.5 revolutions/s, which is comparable with a turnover rate of almost 90 s Ϫ1 (3 ϫ V max ). The K m(ATP) value of 28.7 Ϯ 2.8 M is almost identical to the K m(ATP) value of 21.5 Ϯ 1.8 M estimated from the bulk ATP hydrolysis assay (Table 1). At 2 mM ATP, the rotation rate (28.0 Ϯ 1.7 revolutions/s) decreased to 14.8 Ϯ 0.7 revolutions/s upon addition of 20 mM P i  Table 2. Error bars represent S.D. into the rotation assay buffer (Fig. 8, a and b). This indicates that addition of P i prolongs the dwell time probably due to rebinding of P i to the enzyme. At saturating ATP concentrations, V 1-A011 showed 120°steps, but no additional substep was observed as for the TthV 1 (26) both in the absence and presence of P i (Fig. 8, c and d). The results suggest that release of P i from the V 1 occurs at the same angles as that of ATP waiting to bind.

DISCUSSION
In this study, we investigated the molecular mechanism of ADP inhibition of V 1 by a domain swap approach using combinations of subunit A from ADP inhibition-sensitive TthV 1 and from ADP inhibition-insensitive EhiV 1 . The chimeric V 1-A010 containing the NB domain of EhiV 1 was highly sensitive to ADP inhibition, indicating that sensitivity of V 1 to ADP inhibition is not defined solely by the NB domain. In contrast, the chimeric V 1-A011 consisting of both NB and CT domains of EhiV 1 exhibited insensitivity to ADP inhibition without decreasing the binding affinity for nucleotides ( Fig. 4 and Tables 1 and 2). Thus, the V 1-A011 has a mechanism of reducing sensitivity to ADP inhibition that is different from that of the ADP inhibition-insensitive mutant TSSA V 1 , which was previously shown to have decreased binding affinity for nucleotides as a result of two point mutations in the P-loop region of the NB domain of the A subunit (5, 11).  For the ATP synthesis reaction, the affinity of V 0 V 1-A011 for P i was over an order of magnitude higher than that of V 0 V 1-A010 even though both contain the NB domain from EhiV 1 . The high affinity for P i of V 1-A011 was confirmed using the ATP hydrolysis assay (Fig. 7). In addition, single molecular analysis of V 1-A011 using a low viscous drag probe clearly indicates that addition of P i prolongs the dwell time at every 120°dwell position (Fig. 8). These results strongly suggest a tight correlation between the insensitivity to ADP inhibition of V 1-A011 and an increased affinity for P i . Feniouk et al. (8) proposed a mechanism for ADP inhibition of F 1 : when P i release from a catalytic site happens before ADP release during turnover, the ADP is stochastically and tightly entrapped in the catalytic site. As a result, the F 1 lapses into the ADP-inhibited state (8). Based on this assumption and our results, we think it is safe to conclude that the delayed release of P i from a catalytic site of V 1-A011 as indicated by the increased dwell time decreases the probability of lapsing into the ADPinhibited state.
In this study, we demonstrated that the chimeric TthV 1 incorporating both the NB and CT domains of EhiV 1 A subunit showed insensitivity to ADP inhibition as a result of an increased affinity for P i without an associated decrease in binding affinity for nucleotide. Here we propose the molecular basis of ADP inhibition of V 0 V 1 by comparing the recently reported crystal structures of EhiV 1 and TthV 1 (13,14).
In the crystal structure of TthV 1 , the three A subunits of TthV 1 adopt three different conformations, namely "TthA T " containing ADP, "TthA D " containing ADP, and nucleotide-free "TthA O " (referred to as A N , A NЈ , and A W , respectively in Ref. 13) (Fig. 9a and Table 3). In contrast, there are only two structurally distinct A subunit conformations in EhiV 1 : two A subunits including bound nucleotide (referred as A C and A CR in Ref. 14) and one A subunit without nucleotide (referred to as A O ) (Fig. 9b). The structure of A C is almost identical to that of A CR (Table 4). In this study, we call the two A subunits with bound nucleotide in EhiV 1 (A C and A CR ) EhiA T1 and EhiA T2 , respectively, and the one A subunit with no bound nucleotide EhiA O hereafter.
The structure of TthA O is almost identical to that of EhiA O (Table 5). In addition, the overall structure of TthA T is highly similar to that of EhiA T1 and EhiA T2 (Table 5). In contrast, the catalytic site of TthA D is significantly different from that of the TthA T and the EhiA T1 and EhiA T2 structures. The P-loop residues of the TthA D , TthA T , and EhiA T were superimposed, and the positions of the catalytic residues were compared (Fig. 9, c  and d). The amino group of Lys 234 is 2.9 Å from the ␤-P i group of bound nucleotide in TthA D but 3.8 Å distant in the TthA T (Fig. 9, e and f). The narrow space between the amino group of Lys 234 and the ␤-P i group blocks further binding of P i in the catalytic site of the TthA D . Together with the data from our biochemical analysis of the chimeric V 1 complexes, these findings indicate that the higher affinity of V 1-A011 for P i is due to the A subunit adopting the EhiA T form with ADP (but no P i ), which is capable of binding P i . In contrast, TthV 1 or V 1-A010 can lapse into the ADP-inhibited form when one A subunit in V 1 stochastically adopts the TthA D form with bound ADP, which is then incapable of binding P i . This leads to inhibition of catalysis.  The r.m.s.d. (Å) of C␣ between each A subunit, which is superimposed at the N-terminal ␤-barrel, was calculated by UCSF Chimera.

Molecular Basis of ADP Inhibition of V-ATPase
The overall structure of TthA T is also significantly different from that of TthA D . This difference is caused by movement of the CT domain relative to the NB domain in the A subunit. Some hydrogen bonds (Val 214 -Arg 451 , Leu 215 -Thr 432 , and Asn 425 -Gln 459 ) observed between NB and CT domains in TthA D are absent in TthA T by the movement (Fig. 10). This strongly suggests that the interaction between the NB and CT domains plays an important role in the conformation of the A subunit in V 1 . In the crystal structure of TthV 1 , the first and second helices of domain CT lie in close proximity to the lower part of NB domain (Fig. 10), suggesting that these two helices are important mediators of the interaction between the two domains. The chimeric V 1-A010.1 containing both the NB domain and the first and second helices of the CT domain from EhiV 1 exhibits insensitivity to ADP inhibition. However, the V 1-A010 shows sensitivity to ADP inhibition despite containing the NB domain of EhiV 1 . Taken together, these findings strongly indicate that the interaction between the NB and CT domains of the A subunit that prevents the A subunit from adopting the A D form during turnover is critical for insensitivity to ADP inhibition in V 0 V 1 . This ADP inhibition mechanism of TthV 0 V 1 is advantageous as it prevents consumption of ATP when proton motive force is lost. In contrast, because the E. hirae and eukaryotic V 0 V 1 enzymes function as ion pumps coupled with continuous ATP hydrolysis, they do not exhibit sensitivity to ADP inhibition.
Eukaryotic V 0 V 1 functions as an ion (proton) pump like EhiV 0 V 1 . Thus, we speculate that eukaryotic V 0 V 1 is also insensitive to ADP inhibition to sustain continuous ATP hydrolysis. A similar interaction between the NB and CT domains of the A subunit as seen in the EhiV 0 V 1 is also likely to exist in the eukaryotic enzymes.