The Ingenious Structure of Central Rotor Apparatus in V_oV₁; Key for Both Complex Disassembly and Energy Coupling between V₁ and V_o

Abstract

Vacuolar type rotary H⁺-ATPases (V_oV₁) couple ATP synthesis/hydrolysis by V₁ with proton translocation by V_o via rotation of a central rotor apparatus composed of the V₁-DF rotor shaft, a socket-like V_o-C (eukaryotic V_o-d) and the hydrophobic rotor ring. Reconstitution experiments using subcomplexes revealed a weak binding affinity of V₁-DF to V_o-C despite the fact that torque needs to be transmitted between V₁-DF and V_o-C for the tight energy coupling between V₁ and V_o. Mutation of a short helix at the tip of V₁-DF caused intramolecular uncoupling of V_oV₁, suggesting that proper fitting of the short helix of V₁-D into the socket of V_o-C is required for tight energy coupling between V₁ and V_o. To account for the apparently contradictory properties of the interaction between V₁-DF and V_o-C (weak binding affinity but strict requirement for torque transmission), we propose a model in which the relationship between V₁-DF and V_o-C corresponds to that between a slotted screwdriver and a head of slotted screw. This model is consistent with our previous result in which the central rotor apparatus is not the major factor for the association of V₁ with V_o (Kishikawa and Yokoyama, J Biol Chem. 2012 24597-24603).

Introduction

The Vacuole-type ATPases (V_oV₁) are found in many organisms and are involved in a variety of physiological processes [1–3]. V-ATPases in eukaryotic cells (eukaryotic V_oV₁) translocate protons across the membrane consuming ATP. Most prokaryotic V_oV₁ (also referred to as A-ATPase or A_oA₁ [1, 4]) produce ATP using the energy stored in a transmembrane electrochemical proton gradient [3, 5], while the V_oV₁ of some anaerobic bacteria, such as Enterococcus hirae, function as a sodium pump [6].

The V_oV₁ and F_oF₁ ATPases/synthases are evolutionarily related and share a rotary mechanism to perform their specific functions [3, 7–9]. The basic structures of the ATPases/synthases are conserved among species [1–3]. The soluble, cytoplasmic portion of F_oF₁ and V_oV₁ (called F₁ and V₁, respectively), responsible for ATP hydrolysis/synthesis, is connected via the central rotor stalk and the peripheral stator stalk to the transmembrane portion (F_o and V_o) that houses the ion transporting pathway [1–3]. In Thermus thermophilus V_oV₁, the V₁ portion is composed of a hexameric A₃B₃ cylinder and a central shaft comprised of the D and F subunits [3, 10, 11]. The V_o portion is composed of 5 different subunits with a stoichiometry of C₁E₂G₂I₁L₁₂ (see Fig. 1A). In F_oF₁, a γ -subunit (equivalent to subunits D and F of V_oV₁) binds directly to the rotor ring [12]. In contrast, at the boundary surface of V_oV₁, V_o-C forms a socket-like structure which accommodates the V₁-DF central shaft [13], indicating that V₁-DF does not contact the rotor ring directly. Thus, the boundary surface of V_oV₁ is significantly different from that in F_oF₁. V_oV₁ also has a more complex peripheral stalk structure than F_oF₁. The stator structure of F_oF₁ consists of a single peripheral stalk, while V₁ is connected with V_o by two or three peripheral stalks [4, 14, 15].

Fig 1. V_oV₁ and the short helix of V₁-D subunit.

A, Schematic representation of T. thermophirus V_oV₁ [16]. Subunits in V_o and V₁ are shown in gray and white, respectively. B, The structure of the central rotor apparatus of V_oV₁ obtained by EM density map (PDBID; 3J0J) with the short helix in V₁-D subunit circled in red. The V₁-D, V₁-F and V_o-C subunits are represented in blue, yellow, and pink, respectively. C, Sequence alignment of V₁-D subunit of T. thermophilus (Tt), E. hirae (Eh) and H. sapiens (Hs). Identical amino acid residues are represented by asterisks. The sequences of the short helix of the V₁-D subunit are surrounded by the red box.

Recent reconstitution studies of T. thermophilus V_oV₁ have demonstrated that the A₃B₃ domain is tightly associated with the two EG peripheral stalks of V_o, even in the absence of the central shaft subunits [16]. In other words, the peripheral stalks are the major factor mediating association of V₁ and V_o, consistent with the unique boundary surface between V₁-DF and V_o-C in V_oV₁. This arrangement is highly relevant for the detachment of V₁-DF from V_o-C [13, 16]. However torque needs to be transmitted between V₁-DF and V_o-C for tight energy coupling between V₁ and V_o [16]. Thus a sticky interaction which also allows detachment of V₁-DF from V_o-C is required. How the protein maintains these two somewhat contradictory properties has yet to be investigated.

Lau et al. reported a sub-nanometer resolution structure of T. thermophilus V_oV₁ by single particle cryo-electron microscopy [11] showing that the rod like structure of V₁-DF is positioned in the cavity of V_o-C. This suggests that the rod like structure might play an important role in binding of V₁-DF and V_o-C. A crystal structure of V₁-DF isolated from E. hirae V_oV₁ suggested that the rod like structure might be a short helix at the tip of the V₁-DF (Fig. 1B, [17]).

In this study, reconstitution and fluorescence resonance energy transfer (FRET) analysis of V_oV₁ subcomplexes reveal that the binding affinity of V₁-DF with V_o-C subunit is weak. Further investigations indicated that the short helix of the V₁-DF subunit has important roles in both reconstitution of V_oV₁ and torque transmission. We propose a structural model accounting for both the detachable and sticky nature of the interaction between V₁-DF and V_o-C.

Materials and Methods

Proteins isolation of V_o and CL₁₂

Wild-type or mutant V_oV₁ (C-T105C/C-C268S/C-C323S) from T. thermophilus strains incorporating a His₃ tag on the C terminus of subunit L were generated by the integration vector system [18]. Culture of the modified T. thermophilus strains, membrane preparation, solubilization of His-tagged V_oV₁ and purification of V_oV₁, V_o and CL₁₂ were carried out as described previously [19]. The mutated V_o and CL₁₂ (C-T105C/C-C268S/C-C323S) were used for the FRET experiments.

Preparation of V₁ (A₃B₃DF) and V₁-DF

Escherichia coli strain BL21-CodonPlus-RP (Stratagene) was used for expression of V₁ (A₃B₃DF) and V₁-DF. These recombinant subcomplexes were isolated as described previously [20, 21]. The expression plasmids for V₁ containing DF from H. sapiens or E. hirae were constructed by the same method as described in ref. [21]. The genes encoding the D and F subunits were amplified from human cDNA and pCemtp18 [22], containing the complete E. hirae V_oV₁ (ntp) operon. The D and F genes of the T. thermophilus V₁ expression plasmid were replaced with the amplified genes. The expression plasmids for mutated V₁ lacking the short helix or containing the swapped short helix of E. hirae were constructed by PCR mutagenesis. To introduce the swapped short helix of E. hirae, complementary oligonucleotide primers containing the sequences encoding the 8 amino acids of this region were used to amplify the gene fragment. The fragment was then digested with appropriate restriction enzymes, and inserted into the corresponding region of the T. thermophilus V₁ expression plasmid [20]. The mutant V₁ (A-His₈/ΔCys, A-C255A/A-S232A/A-T235S, F-S54C) and mutant DF (F- His₆, S54C) were used for either reconstitution or FRET experiments [16].

Reconstitution of V_oV₁

The purity of each subcomplex was confirmed by SDS-PAGE. V₁ or V₁-DF (each 1 mg/ml) in MOPDM buffer (50 mM MOPS (pH 7.0), 150 mM NaCl, 0.03% n-dodecyl-β-D-maltoside) was mixed with 1 mg/ml V_o or CL₁₂ at an equal volume ratio. The mixtures were incubated for 1 h at 25°C and then applied onto the Superdex HR-200 column equilibrated with the same buffer. The reconstituted V_oV₁ were collected and used for further analysis immediately.

FRET analysis

FRET analysis was carried out as described previously [16]. The purified V₁ (A-His₈/ΔCys, A- S232A/A-T235S, F-S54C) or DF (F-His₆, S54C) was immediately labeled with an excess amount of Cy3-maleimide (GE healthcare, used as a donor molecule) in MOPDM buffer. Following a 60 min incubation at 25°C, proteins were separated from unbound reagent with a PD-10 column (GE Healthcare). The mutated V_o (C-T105C/C-C268S/C-C323S) was labeled with Cy5-maleimide (GE Healthcare, used as an acceptor molecule) by the same method described above. The specific labeling of subunit F in V₁ or DF, and subunit C in V_o or CL₁₂ was checked by measurement of subunit fluorescence. FRET, as a result of reconstitution of V_oV₁, was monitored with a fluorimeter using an excitation wavelength of 532 nm and an emission wavelength of 570 nm (FP-6200, JASCO). A cuvette was filled with 1.2 ml of MOPDM buffer containing 2 nM labeled V₁ or DF and incubated at 25°C until the fluorescence intensity reached a constant level. For measurement of binding kinetics, 8.0 μl of labeled V_o or CL₁₂ was added into the cuvette at a final concentration of 10 nM.

Mesurements of ATP synthesis of the V_oV₁

The reconstituted complexes were incorporated into liposomes using a freeze-thaw method [23]. Acidification of the proteoliposomes and measurement of ATP synthesis were carried out at 25°C. To acidify the interior of proteoliposomes, 30 μl of the proteoliposome solution was mixed with 15 μl of acidification buffer (300 mM MES pH4.7) and then incubated for 5 min at 25°C. The ATP was measured as the increase of intensity of luminescence in a Luminescencer-PSN (ATTO). The ATP synthesis reaction was initiated by injection of 30 μl of the acidified proteoliposomes into 0.5 ml of base buffer containing 100 mM Tricin-sodium (pH 8.5), 2.5 mM MgSO₄, 10 mM phosphate, 2.2 mg of luciferin/luciferase compound, 0.5 mM ADP, 36 nM valinomycin, and 100 mM KCl. For calibration, ATP was injected into the base buffer.

Measurements of proton channel and proton pump activity by the V_oV₁

Crude soybean L-α-Phosphatidylcholine (Sigma) was washed with 20 mM Tricine-sodium (pH 8.5), 2.5 mM MgSO₄ to remove K⁺ as described [24]. K⁺-loaded proteoliposomes containing enzyme were prepared as follows; aliquots containing reconstituted complex were diluted to 0.5 mg/ml in 20 mM Tricine, 2.5 mM MgCl₂ 20 mM MES pH8.0, 4% n-Octyl-β-D-glucoside (Sigma), the washed 20 mg/ml lipid and 150 mM KCl for the measurement of proton pump or 500mM KCl for the measurement of proton channel activity. Bio-beads SM-2 (Bio-Rad) were added to remove the detergent and incubated for 2 h at 25°C. Resultant proteoliposomes were centrifuged and subjected to the proton pump and proton channel analysis.

The acidification of liposomes was measured by the quenching of 9-amino-6-chloro-2-methoxyacridine, (ACMA) (Sigma) fluorescence [25]. Aliquots containing 5 μg of protein were suspended in 1.2 ml of 20 mM Tricine, 2.5 mM MgCl₂, 40 mM MES pH8.0, 500 mM NaCl for the measurement of proton channel or 110 mM NaCl /40 mM KCl for the measurement of proton pump activity, in the presence of 15 ng of ACMA. The time course of fluorescence quenching was monitored using a fluorimeter (FP-6200, JASCO).

Other assays

Protein concentrations of V₁ were determined from UV absorbance calibrated by quantitative amino acid analysis; 1 mg/ml gives an optical density of 0.88 at 280 nm. Protein concentrations of V_o and V_oV₁ were determined by BCA protein assay, with BSA used as the protein standard. ATPase activity was measured at 25°C with an enzyme-coupled ATP regenerating system [5]. Polyacrylamide gel electrophoresis in the presence of SDS or AES was carried out as described previously [5]. The proteins were stained with Coomassie Brilliant Blue.

Results

Weak interaction at the boundary surface between V₁-DF and CL₁₂

Reconstitution of A₃B₃ and V_o has indicated that the interaction between A₃B₃ and two EG peripheral stalks is rigid [16]. However the precise nature of the interaction between V₁-DF and V_o-C has not been experimentally characterized. In order to examine the interaction the isolated V₁-DF subunits or subcomplexes from V₁ with V_o or CL₁₂, were mixed with V_o or CL₁₂. The purity and subunit stoichiometry of all subcomplexes were confirmed by both SDS-PAGE and AES-PAGE (Fig. 2A and B). The band corresponding to V₁-DF was not detected on AES-PAGE gel. Reconstituted V_oV₁ was identified as a single band on AES-PAGE when V₁ was incubated with V_o at a molar ratio of 1:1 (Fig. 2B, lane 5). However attempts to reconstitute V₁ with CL₁₂ were unsuccessful as assessed by both AES-PAGE and gel permeation analysis (Fig. 2B, lane 6 and Fig. 2C, line 7). The same was also true for V₁-DF and CL₁₂ or V_o (Fig. 2B, lane 7 and Fig. 2C, line 6 and 8). Together, these results strongly suggest that the interaction of the boundary surface between V₁-DF and V_o-C in the complex is insufficient for reconstitution of a stable complex.

Fig 2. Analysis of reconstitution of complexes and subcomplexes.

A, 15% SDS-PAGE analysis. Lane 1, V_oV₁; lane 2, V_o; lane 3 V₁; lane 4, CL₁₂; lane 5, DF. Molecular weights of each subunit are indicated in parentheses. B, 6% AES-PAGE. The mixtures containing the subcomplexes were incubated for 1h at 25°C respectively, prior to analysis. Lane 1, V_oV₁; lane 2, V_o; lane 3, V₁; lane 4, CL₁₂; lane 5, V_o and V₁; lane 6, V₁ and CL₁₂; lane 7, DF and V_o, lane 8; DF. The band of DF complex was not detected on AES-PAGE gel. C, Gel permeation analysis of complexes and subcomplexes. The mixtures containing subcomplexes indicated by the scheme were incubated for 1h at 25°C respectively, followed by analysis. The molecular weights of each complex are, V_oV₁ (659kDa), V_o (268kDa), V₁ (391 kDa), CL₁₂ (131 kDa) and V₁-DF (37 kDa). D, FRET analysis of reconstituted complexes and subcomplexes. Fluorescence of 3 nM V_1-Cy3 (purple and green lines) or 3 nM V₁-DF_-Cy3 (blue and red lines) was recorded (excitation at 532nm) before and after addition of V_o-Cy5 or CL_12-Cy5. The V_o-Cy5 or CL_12-Cy5 was added at the time indicated by the arrow.

Further analysis of the interaction at the boundary surface between V₁-DF and V_o-C was carried out by fluorescent resonance of energy transfer (FRET), a powerful method for detecting protein-protein interaction/association [16]. For FRET analysis, each component was labeled with Cy3 or Cy5 (fluorescent dyes as described in ref. [16]). A mutated V₁-DF or V₁ incorporating a single cysteine residue (F/S54C) was labeled with Cy3 as a donor, while a mutated V_o or CL₁₂ (C/T105C) was labeled with Cy5 as an acceptor. Reconstitutions were carried out in a cuvette containing 1.2 ml of the 2 nM V₁ or V₁-DF labeled with Cy3. The reconstitution efficiency was evaluated by the decrease in donor emission (Fluorescence at 570 nm, ref. [16]). As shown in Fig. 2D, the fluorescence of V₁ decreased sharply upon addition of V_o into the cuvette, indicating reconstitution of V₁ and V_o complex (purple line). In contrast, addition of 8.0 μl of 1.5 mM of CL₁₂ or V_o into a cuvette containing V₁-DF exhibited no decrease in fluorescence at 570 nm (red or blue lines). Furthermore the addition of 8.0 μl of 1.5 mM of CL₁₂ into a cuvette containing V₁ showed no decrease in fluorescence (green line). These results clearly indicate low binding affinity between V₁-DF and V_o-C, consistent with the findings from the reconstitution experiments [16].

Effect of exogenous V₁-DF on reconstitution of V₁ and V_o

The amino acid sequence of V₁-DF is not highly conserved among species (Fig. 1C). To investigate the effect of theses differences on reconstitution of V₁ and V_o, expression constructs of V₁ containing V₁-DF from Homo sapiens (V_1-DF-H.s) or E. hirae (V_1-DF-E.h) were generated. These chimeric V₁ were purified and subjected to ATPase analysis as described in Materials and Methods. Subunit stoichiometry of each chimeric V₁ was confirmed by SDS-PAGE analysis (Fig. 3A). These chimeric V₁ were ATPase active (Table 1, S1 Fig.), indicating that the exogenous V₁-DF functions as a rotor in A₃B₃ of T. thermophilus. As shown in Table 1, the presence of the exogenous DF shaft from H. sapiens markedly enhanced the ATPase activity of V₁. The enhanced ATPase activity of chimeric V₁ is likely due to primary sequence difference between the DF of T. thermophilus and that of H. sapiens. We will discuss the effect of the H. sapiens DF on the activity of V₁ fully elsewhere.

Fig 3. Analysis of reconstituted V₁ with V_o.

Isolated V₁ and V_o were analyzed by SDS-PAGE (A). Lane 1, V_oV₁; lane 2, V_o; lane 3, V₁; lane 4, chimeric V₁ containing E. hirae V₁-DF (V_1-Eh-DF); lane 5, chimeric V₁ containing H. sapiens V₁-DF (V_1Hs-DF); lane 6, mutated V₁ containing the exogenous short helix of E. hirae (V_1SH-Eh); lane 7, the mutated V₁ lacking the short helix (V_1ΔSH). Analysis of reconstituted mutated V₁ and V_o by gel permeation (B). The mixtures containing the mutated V₁ and V_o were incubated for 1 h at 25°C, prior to analysis. Subunit stoichiometry of the fraction containing each mutated V_oV₁ was confirmed by SDS-PAGE analysis (C).

Table 1. Kinetics parameters of the mutated V₁ for ATPase activity.

	V1	V1-E.h-DF	V1-H.s-DF	V1-SH-E.h	V1ΔSH
K m [mM]	0.23 ± 0.05	0.76 ± 0.04	0.48 ± 0.04	0.22 ± 0.05	0.24 ± 0.06
V max [s-1]	30.6 ± 1.8	38.0 ± 0.80	132 ± 3.0	27.5 ± 1.5	21.8 ± 1.4

K _m and V _max values represent means ± SD (n = 3).

Each chimeric V₁ was mixed with an excess amount of V_o from T. thermophilus and incubated for 1 hour. Analysis by gel permeation analysis revealed peaks corresponding to reconstituted V_oV₁ containing the exogenous V₁-DF (Fig. 3B, lines 9 and 10). Subunit stoichiometry of the fraction containing each chimeric V_oV₁ was confirmed by SDS-PAGE analysis (Fig. 3C). These results indicate that the chimeric V₁ incorporating the exogenous V₁-DF can assemble with V_o.

Role of the short helix of V₁-D subunit in interaction of V₁ and V_o

The EM density structure of intact V_oV₁ of T. thermophilus and the crystal structure of V₁-DF suggested that an α-helix of V₁-D (a.a. 73–80) is key for interaction between V₁-DF and V_o-C ([11,17] and Fig. 1B and C). This helix is referred to as the short helix hereafter. To investigate the role of the short helix in reconstitution of V₁ and V_o, V_1-SH-E.h containing the exogenous short helix of E. hirae (A⁷⁴FIDELLA⁸¹, Fig. 1C), and V_1ΔSH lacking the short helix of the V₁-DF were constructed. The subunit stoichiometry and purity of the mutated V₁ constructs were confirmed by SDS-PAGE analysis (Fig. 3A). Reconstitution of V_o and V_1-SH-E.h or V_1ΔSH was confirmed by gel permeation and FRET analysis (Fig. 3B, line 11, 12). Subunit stoichiometry of the fraction containing each mutated V_oV₁ was confirmed by SDS-PAGE analysis (Fig. 3C). These findings indicate that the short helix is not essential for reconstitution of V₁ with V_o.

Role of the short helix of V₁-D subunit in energy coupling between V₁ and V_o

To further investigate the role of the short helix of the V₁-D subunit in energy coupling between V₁ and V_o, ATP synthesis activity of the reconstituted V_oV₁ were assessed. The reconstituted V_oV₁ constructs were purified by gel permeation chromatography and reconstituted into liposomes by freeze-thaw methods. Proton motive force was generated across the membranes of the reconstituted liposomes by acid-base transition (ref. [23], see inset in Fig. 4A). As shown in Fig. 4A, wild type reconstituted V_oV₁ showed continuous ATP synthesis. In contrast, the reconstituted V_oV_1ΔSH showed no ATP synthesis activity, indicating that lack of the short helix in V_oV₁ causes an intra molecular uncoupling between V₁ and V_o. In addition, the reconstituted V_oV₁ including the exogenous V₁-DF (V_oV_1-DF-H.s, V_oV_1-DF-E.h) or the short helix (V_oV_1-SH-E.h) showed no ATP synthesis activity.

Fig 4. Function of the short helix of V₁-D subunit for energy coupling.

A, Analysis of ATP synthesis by the reconstituted complexes. inset; schematic model of the experimental system [16]. The reconstituted V_oV₁ was incorporated into liposomes, then energized by an acid base transition procedure described in the Materials and Methods. Each line shows the raw data for ATP synthesis by each reconstituted complex at a Δ pH of 3.8. The reactions were initiated by addition of acidified proteoliposomes into the base buffer, as indicated by the arrow. Final concentrations were 2 μg/ml of V_oV₁ and 1 mM ADP. The synthesized ATP was monitored by the luciferin-luciferase assay [18]. Analysis of proton channel (B) and proton pump activity (C) coupled with ATP hydrolysis by proteoliposomes. K⁺-loaded proteoliposomes containing the reconstituted complexes were prepared as described in the Materials and Methods. Proton influx was initiated by addition of 20 ng of valinomycin at the time indicated by the black arrow (B) and proton pump was initiated by addition of ATP-Mg at finally 1mM followed by the addition of valinomycin at the time indicated by the red arrow (C). The ACMA fluorescence emission (480nm) was recorded at 25°C.

Next, proton channel and proton pump activity of the reconstituted V_oV₁ were measured to investigate the energy coupling efficiency between the mutated V₁ and V_o in the complexes. To facilitate ATP hydrolysis activity measurements, a mutated V₁ incorporating the TSSA substitutions (A-S232A/A-T235S) to overcome ADP inhibition was used [5]. The wild-type V_oV₁ did not show proton channel activity but did show proton pump activity (Fig. 4B, C). These results indicate that the wild-type V_oV₁ is tightly coupled. In contrast, mutated V_oV₁ including the exogenous short helix or lacking the short helix did not show proton pump activity (Fig. 4C). In contrast, the mutated V_oV₁s had proton pump activity almost identical to the proton channel activity of V_o (Fig. 4B), indicating that the mutated V_oV₁s were completely uncoupled. These results strongly suggest that a proper match between the short helix in V₁-D subunit and V_o-C subunit is essential for tight energy coupling between V₁ and V_o.

Discussion

In this study, we have directly demonstrated a low binding affinity between V₁-DF and V_o-C by both FRET and reconstitution experiments (Fig. 2). This is consistent with our previous result indicating that the two EG peripheral stalks are the major mediators of association of V₁ with V_o [16]. This low binding affinity between V₁-DF and V_o-C is relevant to reversible dissociation/association of V₁ from V_o in eukaryotic V_oV₁ [26, 27]. However such low binding affinity is unfavorable for energy coupling between V₁ and V_o; for ATP synthesis (ΔG = ~55 kJ/mol, [16]), the torque generated in the V_o rotor ring needs to be transmitted to V₁-DF via V_o-C. Thus, an ingenious structure, that is both detachable and sticky, is required at the boundary surface between V₁-DF and V_o-C. The EM density map of T. thermophilus V_oV₁ [11] provided a clue to unraveling the molecular basis of these seemingly contradictory properties. The short helix of V₁-DF apparently lies in the cavity of V_o-C in V_o, suggesting that it may play an important role in association of V₁-DF and V_o-C.

Here, we have investigated the role of the short helix in V₁-D subunit on V_oV₁ assembly and energy coupling between V₁ and V_o. Not only the chimeric V₁ containing the exogenous V₁-DF and the short helix, but also V₁ lacking the short helix could reconstitute the complex with V_o (Fig. 3), indicating that the short helix is not a key factor for V_oV₁ complex formation. However, these mutant V_oV₁s showed neither ATP synthesis or proton pump activities (Fig. 4). As shown in Fig. 1C, amino acid residues in the short helix of V₁-D are not highly conserved among species (T. thermophilus, E. hirae, H. sapiens). The different surface shape of the exogenous short helix compared to the endogenous one likely generates repulsion between V₁-DF and V_o-C. The intact fitting of the short helix into the recess of V_o-C appears to be required for tight energy coupling between V₁ and V_o.

Here we propose a model to account for a detachable V₁-DF/V_o-C boundary surface, which can transmit torque. In this model, the relationship between V₁-DF and V_o-C is analogous to that between a slotted screwdriver and a head of slotted screw (Fig. 5). In the V_oV₁, two EG peripheral stalks push the short helix of V₁-D into the socket of V_o-C. Thus, the short helix of V₁-D binds into the socket of V_o-C, forming a sufficiently close interaction for transmission of torque from the rotating V₁-DF to V_o-C. The rigid interaction between the V₁-DF and V_o-C would be abolished by a loss of interaction between the two EG peripheral stalks and A₃B₃, consistent with the results of reconstitution experiments (Fig. 2) and EM analysis of sequential disassembly of the V_oV₁ [28]. F_oF₁ does not contain a counterpart of V_o-C so that the F₁ shaft composed of the γ subunit directly attaches to the c_6–10 rotor ring [12]. Thus, the F₁-c_6–10 stator-less complex is easily isolated from F_oF₁.

Fig 5. A schematic diagram for a slotted screwdriver and a head of slotted screw in V_oV₁ [16].

Supporting Information

S1 Fig. [S]-V plot of ATP hydrolysis rate catalyzed by mutated V₁.

The solid lines show fit with the Michaelis-Menten equation.

(TIF)

Data Availability

All relevant data are within the paper and its Supporting Information files.

Funding Statement

This work was supported by Grants-in-Aid from the Ministry of Education, Science, Sports and Culture of Japan No. 24370059 and 26650039 to KY. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.