Basic Properties of Rotary Dynamics of the Molecular Motor Enterococcus hirae V₁-ATPase

Background: The chemomechanical coupling scheme of the rotary motor V₁-ATPase is incompletely understood.

Results: Enterococcus hirae V₁-ATPase (EhV₁) showed 120° steps of rotation without substeps, as commonly seen with F₁-ATPase.

Conclusion: The basic properties of rotary dynamics of EhV₁ are similar to those of Thermus thermophilus V₁-ATPase.

Significance: This study revealed the common properties of V₁-ATPases as rotary molecular motors, distinct from those of F₁-ATPases.

Abstract

V-ATPases are rotary molecular motors that generally function as proton pumps. We recently solved the crystal structures of the V₁ moiety of Enterococcus hirae V-ATPase (EhV₁) and proposed a model for its rotation mechanism. Here, we characterized the rotary dynamics of EhV₁ using single-molecule analysis employing a load-free probe. EhV₁ rotated in a counterclockwise direction, exhibiting two distinct rotational states, namely clear and unclear, suggesting unstable interactions between the rotor and stator. The clear state was analyzed in detail to obtain kinetic parameters. The rotation rates obeyed Michaelis-Menten kinetics with a maximal rotation rate (V_max) of 107 revolutions/s and a Michaelis constant (K_m) of 154 μm at 26 °C. At all ATP concentrations tested, EhV₁ showed only three pauses separated by 120°/turn, and no substeps were resolved, as was the case with Thermus thermophilus V₁-ATPase (TtV₁). At 10 μm ATP (⪡K_m), the distribution of the durations of the ATP-waiting pause fit well with a single-exponential decay function. The second-order binding rate constant for ATP was 2.3 × 10⁶ m⁻¹ s⁻¹. At 40 mm ATP (⪢K_m), the distribution of the durations of the catalytic pause was reproduced by a consecutive reaction with two time constants of 2.6 and 0.5 ms. These kinetic parameters were similar to those of TtV₁. Our results identify the common properties of rotary catalysis of V₁-ATPases that are distinct from those of F₁-ATPases and will further our understanding of the general mechanisms of rotary molecular motors.

Introduction

V-ATPase is a rotary molecular motor that couples ion transport to ATP hydrolysis and synthesis. The main function of V-ATPase in eukaryotes is to transport protons across a membrane by using the energy derived from ATP hydrolysis (1–3). V-ATPase also catalyzes ATP synthesis, harnessing the energy of proton flow in certain eubacteria such as Thermus thermophilus. V-ATPases are composed of V₁-ATPase (V₁),⁶ a water-soluble moiety that hydrolyzes and synthesizes ATP, and a membrane-embedded moiety (V₀) that translocates ions. The V₁ and V₀ domains are connected by a rotary shaft and peripheral stalks (1–3). The V₁ complex is composed of A, B, D, and F subunits, in which the three A and three B subunits are alternately arranged, forming a hexameric stator A₃B₃ ring (4–7). ATP hydrolysis and synthesis occur on the catalytic sites that are located at the interfaces of the A and B subunits, with the majority of the catalytic residues residing in the A subunits. The rotary shaft is composed of D and F subunits penetrating into the central cavity of the A₃B₃ ring (6, 7).

The rotation of V₁ has been visualized using optical microscopy by attachment of a probe to the rotary shaft (8–11). V₁ of T. thermophilus (TtV₁), which functions as an ATP synthase, rotates stepwise in a counterclockwise direction (8). The basic step size is 120°, and similar to F₁-ATPase (F₁), the water-soluble moiety of F₀F₁-ATP synthase (12), each step is coupled to the consumption of a single ATP molecule (10). Although no substeps have yet been resolved in the rotation of TtV₁ (10, 11), the 120° steps of F₁ from the thermophilic Bacillus PS3 (TF₁) and Escherichia coli (EF₁) have been shown to be further divided into 80° and 40° substeps and into 85° and 35° substeps, respectively (13–15). The 80° and 85° substeps are triggered by ATP binding and ADP release, whereas the 40° and 35° substeps are known to occur after ATP cleavage and release of inorganic phosphate. Accordingly, the pauses before the 80° and 85° substeps are referred to as ATP-binding (ATP-waiting) pauses, and those prior to the 40° and 35° substeps are known as catalytic pauses. As described above, the chemomechanical coupling scheme of TtV₁ appears to be distinct from that of F₁. However, to date, the stepping rotations of V₁ complexes other than TtV₁ have not been described, and the chemomechanical coupling scheme of V₁ remains unclear (9).

Enterococcus hirae V-ATPase functions as a primary ion pump, similar in nature to eukaryotic V-ATPases (16, 17). We recently solved the crystal structures of the V₁ component of E. hirae V-ATPase (EhV₁) and proposed a model of its rotation mechanism (6). In this study, to characterize the stepping rotation of EhV₁, we analyzed and compared the basic properties of EhV₁ rotary dynamics with those of TtV₁, TF₁, and EF₁. As was the case with TtV₁, no substeps were resolved in the rotation of EhV₁, suggesting that 120° stepping rotation without substeps is a common property of V₁ complexes.

EXPERIMENTAL PROCEDURES

Preparation of Recombinant EhV₁ and AviTag-tagged EhV₁ Expressed in E. coli

The EhV₁ holocomplex (A₃B₃DF) was expressed in E. coli using the expression plasmid pTR19-FABD. We synthesized a DNA fragment containing the ntpF, ntpA, ntpB, and ntpD genes (in this order) and optimized its codon usage for E. coli expression. This fragment was then cloned into plasmid pTR19, the expression vector for the F₀F₁-ATP synthase of thermophilic Bacillus PS3 (18), after which a His₆ tag was introduced at the N terminus of the A subunit by PCR to obtain plasmid pTR19-FABD. For the rotation assay, we used the pTR19-FABD-Avi, in which the AviTag biotinylation sequence (GLNDIFEAQKIEWHE) (19) was inserted between Gly-121 and Tyr-122 of the D subunit by PCR-based mutagenesis. E. coli BL21(DE3) cells were transformed with pTR19-FABD or pTR19-FABD-Avi and cultured in Super Broth (32 g/liter Tryptone, 20 g/liter yeast extract, and 5 g/liter sodium chloride) containing 100 μg/ml ampicillin and 2 mm isopropyl β-d-thiogalactopyranoside at 37 °C for 20 h. Cells were suspended in buffer A (20 mm potassium P_i (pH 7.0), 230 mm NaCl, and 20 mm imidazole) and disrupted by sonication. After removal of the cell debris by centrifugation at 81,000 × g for 20 min at 4 °C, the solution was applied to a nickel-nitrilotriacetic acid column (Ni-NTA Superflow, Qiagen). After washing with 10 column volumes of buffer A, recombinant EhV₁ or AviTag-tagged EhV₁ (AviTag-EhV₁) was eluted with buffer B (20 mm potassium P_i (pH 7.0), 50 mm NaCl, and 200 mm imidazole). The eluted fractions were concentrated with an Amicon Ultra-10K unit (Merck Millipore) and then passed through a Superdex 200 gel-filtration column (GE Healthcare) equilibrated with buffer C (20 mm MES-NaOH (pH 6.5), 100 mm NaCl, and 10% glycerol). The purified proteins were flash-frozen in liquid nitrogen and stored at −80 °C until used.

Preparation of the A₃B₃ Subcomplex

The A₃B₃ subcomplex was expressed in E. coli BL21(DE3) cells harboring the expression plasmid pTR19-AB. Plasmid pTR19-AB was constructed by removing the ntpF and ntpD genes from pTR19-FABD by PCR. Expression and purification of the A₃B₃ subcomplex were performed using the same procedure as described for the EhV₁ holocomplex, and purified proteins were stored at −80 °C until used.

Preparation of the DF Subcomplex

An E. coli cell-free protein expression system was employed to synthesize the DF subcomplex using a mixture of plasmids containing the corresponding genes. The expressed DF subcomplex was purified as described previously (6). The homogeneity of each purified subcomplex was judged by SDS-PAGE analysis. After purification, cysteine residues in the DF subcomplex were biotinylated with a 3 m excess of biotinylation reagent (biotin-PEAC₅-maleimide, Dojindo) in 20 mm MOPS-KOH (pH 7.0) and 150 mm NaCl at room temperature for 20 min. The reaction was quenched using 10 mm DTT. We used a mutant DF subcomplex containing two engineered cysteine residues in its D subunit (T60C/R131C), substituted using a QuikChange site-directed mutagenesis kit (Agilent Technologies), and a single endogenous cysteine residue in both D (Cys-153) and F (Cys-54) subunits. A maximum of three of these residues can be expected to react with the biotinylation reagent. Specific biotinylation of the D subunit was confirmed by Western blotting with streptavidin-alkaline phosphatase conjugate (see Fig. 1).

FIGURE 1.

Gel electrophoresis. Lanes 1–3, SDS-PAGE of reconstituted EhV₁, recombinant EhV₁, and AviTag-EhV₁. A 16% gel was used; 12 pmol of protein was loaded in each lane. The molecular masses of the A, B, D, and F subunits are 65, 51, 24, and 11 kDa, respectively. Proteins were stained with Coomassie Brilliant Blue. Lanes 4–6, immunoblots stained by streptavidin-alkaline phosphatase conjugates, showing biotin labeling of the D subunit. Lane 4, reconstituted EhV₁ containing the biotinylated D subunit. Lane 5, non-biotinylated recombinant EhV₁. Lane 6, AviTag-EhV₁ containing the biotinylated D subunit.

Preparation of Reconstituted EhV₁

Reconstituted EhV₁ was prepared as follows. Purified A₃B₃ and biotinylated DF were mixed at a 1:2 molar ratio and incubated at room temperature for 2 h. Reconstituted EhV₁ was purified using a Superdex 200 gel-filtration column equilibrated with buffer C, flash-frozen in liquid nitrogen, and stored at −80 °C until used.

Biochemical Assay

The protein concentration of EhV₁ was determined based on UV absorbance using a molar extinction coefficient of 310,910 m⁻¹ cm⁻¹ calculated from its amino acid sequence (ProtParam tool, ExPASy). The ATP hydrolysis rate of EhV₁ was measured using an ATP-regenerating system. The reaction mixture contained 50 mm MES-KOH (pH 6.5), 50 mm KCl, 5 mm MgCl₂, 2.5 mm phosphoenolpyruvate, 0.2 mg/ml NADH, 0.1 mg/ml pyruvate kinase, and 0.1 mg/ml lactate dehydrogenase in addition to various concentrations of ATP. The rate of ATP hydrolysis was monitored as the rate of NADH oxidation at times ranging from 0 to 10 s after the addition of V₁, which was measured as a decrease in the absorbance at 340 nm. All measurements were carried out at 25 ± 1 °C.

Rotation Assay

To observe ATP hydrolysis-driven rotation of EhV₁, the A₃B₃ stator ring was immobilized on a Ni-NTA-coated glass surface via a His₆ tag introduced at the N terminus of the A subunit. Streptavidin-coated 40-nm gold colloid was then attached to the biotinylated cysteine or AviTag in the rotor DF subunit as a probe. The gold colloids were prepared as described previously (20). The rotations of the gold colloid were observed at 26 ± 1 °C by an objective-type total internal reflection dark-field microscope constructed on an Olympus IX-71 inverted microscope (20). The images were recorded with a high-speed CMOS camera (FASTCAM 1024PCI, Photron) at 1000–10,000 frames/s (fps). The flow cell was assembled from a Ni-NTA-coated glass (24 × 36 mm²) and an uncoated cover glass (18 × 18 mm²) separated by two spacers of ∼50 μm thickness. First, buffer D (5 mg/ml BSA, 20 mm potassium P_i (pH 7.0), 230 mm NaCl, and 20 mm imidazole) was infused into the flow cell to prevent nonspecific binding of EhV₁ and gold colloid. After incubation for 10 min, EhV₁ (0.5–2 nm in buffer D) was infused into the flow cell. After incubation for 5 min, unbound EhV₁ was washed out with buffer D, after which gold colloid suspended in buffer D was infused. After 10 min, unbound gold colloid was washed out. Observation of rotation was initiated after infusion of buffer E (50 mm MES-KOH (pH 6.5), 50 mm KCl, and 5 mm MgCl₂) containing ATP (10 μm to 3 mm) or MgATP (4 and 40 mm) and an ATP-regenerating system.

RESULTS

Recombinant EhV₁

We first tried to carry out a rotation assay using recombinant EhV₁ (A₃B₃DF complex) expressed in E. coli. To observe the rotation, the rotation probe must be attached to the rotor DF subunits through biotin-streptavidin linkage. Because three endogenous cysteine residues in the stator A subunit (Cys-28, Cys-174, and Cys-259) may react with the biotinylation reagent, we substituted these residues with serine or alanine (C28A/C174S/C259S) and measured the activity of this mutant. The ATP hydrolysis rate of the EhV₁ mutant, measured by a biochemical assay, was <10% of that of wild-type EhV₁ (data not shown), indicating that the effect of substitution was significant.

Reconstituted EhV₁

Because the rotor DF subunit can be biotinylated separately from the stator A₃B₃ ring prior to reconstitution, we next used reconstituted EhV₁ in the rotation assay. Reconstituted EhV₁ has also been used recently for crystal structural analysis requiring a pure homogeneous sample (6). No differences were observed in the subunit composition for reconstituted EhV₁ and recombinant EhV₁ (Fig. 1), indicating a high reconstitution efficiency. Furthermore, reconstituted EhV₁ had an ATP hydrolysis rate comparable to that of recombinant EhV₁ at all ATP concentrations used, ranging from 10 μm to 4 mm (Fig. 2). The expected rotation rates of reconstituted EhV₁ and recombinant EhV₁, calculated as one-third of the ATP hydrolysis rates, followed Michaelis-Menten kinetics. The maximal rotation rate (V_max) and Michaelis constant (K_m) were 73 ± 2 revolutions/s (rps) and 221 ± 17 μm, respectively, for reconstituted EhV₁ and 72 ± 1 rps and 246 ± 16 μm (mean ± S.E. of fitting), respectively, for recombinant EhV₁ (Fig. 2B, upper panel, red and blue dashed lines). These results indicate that the kinetic parameters of reconstituted EhV₁ are almost identical to those of recombinant EhV₁, suggesting that the catalytic properties of EhV₁ are not affected by the reconstitution process. Therefore, we decided to use reconstituted EhV₁ for the rotation assay.

FIGURE 2.

ATP dependence of ATP hydrolysis rate and rotation rate. A, time course of ATP hydrolysis by reconstituted EhV₁ (solid lines), recombinant EhV₁ (dashed lines), and AviTag-EhV₁ (dotted lines) at 25 ± 1 °C with 10 μm, 50 μm, 100 μm, 300 μm, 1 mm, and 4 mm ATP. ATP hydrolysis was monitored as NADH oxidation in the ATP-regenerating system. The reaction was initiated by the addition of EhV₁ (final concentration of 10 nm) at 0 s. The ATP hydrolysis rate of EhV₁ was estimated from the slope of 0–10 s. B, upper panel, red dots indicate rotation rates determined by single-molecule rotation assay of reconstituted EhV₁. Open red circles indicate average rotation rates (n ≥ 3). Error bars represent S.D. Open red squares, open blue triangles, and green crosses indicate the average of one-third of the ATP hydrolysis rates determined by the biochemical assay of reconstituted EhV₁, recombinant EhV₁, and AviTag-EhV₁, respectively (n ≥ 3). In biochemical assay, S.D. is smaller than the size of the symbols. The solid and dashed lines indicate fits with the Michaelis-Menten equation: V = V_max × [ATP]/(K_m + [ATP]). V_max = 107 ± 5 rps and K_m = 154 ± 33 μm (mean ± S.E. of fitting) were obtained by the single-molecule rotation assay for reconstituted EhV₁. V_max = 73 ± 2 rps and K_m = 221 ± 17 μm were obtained by the biochemical assay for reconstituted EhV₁, V_max = 72 ± 1 rps and K_m = 246 ± 16 μm for recombinant EhV₁, and V_max = 76 ± 1 rps and K_m = 300 ± 19 μm (mean ± S.E.) for AviTag-EhV₁. The apparent binding constant for ATP (3 × V_max/K_m) was estimated as (2.2 ± 0.4) × 10⁶ m⁻¹ s⁻¹ from the rotation assay. Lower panel, ratio of the rotation rate determined by the rotation assay to one-third of the ATP hydrolysis rate determined by the biochemical assay of reconstituted EhV₁.

EhV₁ Has Two Distinct Rotational States

The rotary motion of reconstituted EhV₁ was observed in a single-molecule assay using streptavidin-coated 40-nm gold colloid as a load-free probe at a rate of 1000–10,000 fps (Fig. 3). We found that the reconstituted EhV₁ complexes exhibited two distinct reversible states of rotation, namely clear and unclear (Fig. 4, A–C). In the clear rotational state, the majority of the centroids of gold colloid in each frame were distributed in three positions separated by 120° and were remote from the rotation center. Moreover, the time course showed clear unidirectional rotation in a counterclockwise direction (Fig. 4A). In contrast, in the unclear state, the centroids showed wide fluctuations toward the rotation center (Fig. 4B). Because the rotation rate in the unclear state seems to be similar to that in the clear state, the complex appears to rotate unidirectionally, although the measured rotation rate in the unclear state may not be accurate.

FIGURE 3.

Schematic image of the rotation assay. Shown is the experimental setup for single-molecule rotation assay of reconstituted EhV₁. The stator A₃B₃ ring of EhV₁ was fixed on the glass surface with His₆ tag at the N termini of the A subunits. Streptavidin-coated 40-nm gold colloid was attached to biotinylated cysteine residues in the rotor DF subunit.

FIGURE 4.

Two distinct states in the rotation of EhV₁. A and D, time courses of rotation, including two reversible states of single molecule-reconstituted EhV₁ (3 mm ATP) and AviTag-EhV₁ (2 mm ATP). The rotations in the clear and unclear states are highlighted in red and blue, respectively. B and E, The x-y trajectories of the centroid of a rotating gold colloid shown in A and D. C and F, distributions of rotary angles shown in A and D. The numbers in A–C and D–F indicate the corresponding parts. G and H, distributions of duration times of clear and unclear states of reconstituted EhV₁. The data at various ATP concentrations (from 10 μm to 40 mm) were analyzed collectively. The bin width was 0.1 s. The solid curves show the fit with single-exponential decay: constant × (exp(−t/τ)), where τ = 0.13 ± 0.003 s (mean ± S.E. of fitting, 238 events from 58 molecules) and 0.33 ± 0.009 s (mean ± S.E., 199 events from 58 molecules) for the clear (G) and unclear (H) states, respectively.

To assess whether the two rotational states are observed only in reconstituted EhV₁, the D subunit of recombinant EhV₁ was fused to AviTag, a 15-amino acid sequence that is subject to biotinylation by biotin ligase in E. coli (19). No differences were observed in the subunit compositions of AviTag-EhV₁ and reconstituted or recombinant EhV₁ (Fig. 1). Moreover, the ATP hydrolysis rate of AviTag-EhV₁ was comparable to that of reconstituted and recombinant EhV₁ in the biochemical assay (Fig. 2). Importantly, as was the case in the rotation assay with reconstituted EhV₁, AviTag-EhV₁ also exhibited two reversible rotational states (Fig. 4, D–F). This result clearly shows that the two rotational states are not an artifact of damage or inactivity caused by the reconstitution procedure but instead represent an intrinsic property of EhV₁. Given that the efficiency of biotinylation (Fig. 1) and the frequency of rotating probes for AviTag-EhV₁ were significantly lower than those for reconstituted EhV₁, we used reconstituted EhV₁ in the subsequent experiments.

Next, we analyzed the duration of the clear and unclear states for 10 μm to 40 mm ATP. Because we did not find a clear dependence of the duration on ATP concentration, we analyzed the data at various ATP concentrations collectively. The distributions of the duration times fit well to a single-exponential decay function, suggesting a single rate-limiting step in the transition between the clear and unclear states. The time constants were 0.13 ± 0.003 s (mean ± S.E. of fitting, 238 events from 58 molecules) for the clear state and 0.33 ± 0.009 s (mean ± S.E. of fitting, 199 events from 58 molecules) for the unclear state (Fig. 4, G and H). The ratio of the clear state to the total observation time was ∼0.3 (=0.13/(0.13 + 0.33)).

To date, it has been unclear why transitions occur between the clear and unclear states. A study reported that although unusual fluctuations have been reported in the rotation of TtV₁ (11), the behaviors are not entirely identical; in that study (11), the authors attributed the fluctuating state to the probe adopting two orientations relative to the D subunit and excluded these data from the analysis. Alternatively, the unclear state may be due to less stable interactions between the rotor and stator in V₁ compared with those in F₁. Because V₀ and V₁ are connected not only by the rotor but also by the two peripheral stalks composed of the E and G subunits, this unstable interaction would be anticipated to occur only in the isolated V₁ complex and not in the physiological V₀V₁ complex. We nevertheless concluded that the tight chemomechanical coupling of EhV₁ is achieved at least in the clear state and, accordingly, restricted our analysis in the remainder of the study to that state.

ATP Dependence of Rotation

The rotation rates of reconstituted EhV₁ were measured at various concentrations of ATP, ranging from 10 μm to 40 mm (Fig. 2B, upper panel, red dots and open circles). Below 100 μm ATP, the rotation rates were almost proportional to the ATP concentration, indicating that ATP binding is rate-limiting in this range. Above 1 mm ATP, the rotation rate was essentially constant. The rotation rates followed Michaelis-Menten kinetics with a V_max of 107 ± 5 rps and a K_m of 154 ± 33 μm (mean ± S.E. of fitting) (Fig. 2B, upper panel, solid red line). The second-order binding rate constant for ATP (k_on(ATP)) determined from 3 × V_max/K_m was (2.2 ± 0.4) × 10⁶ m⁻¹ s⁻¹ under the assumption that three ATP molecules were hydrolyzed per turn.

The value of K_m determined by the single-molecule rotation assay was slightly lower than that determined by the biochemical assay. Furthermore, the value of V_max estimated by the single-molecule rotation assay was ∼50% greater than that deduced from the biochemical assay. This ratio was essentially constant at each ATP concentration (Fig. 2B, lower panel). This result may arise from inaccuracy in protein concentration determination based on the molar absorbance coefficient calculated from the amino acid sequence and/or imply that the ATP hydrolysis rate during the unclear state is slightly lower than that during the clear state. Furthermore, in the biochemical assay, the ATP hydrolysis rate gradually decreased during measurement (Fig. 2A). It is well known that the ATP hydrolysis of TtV₁, which functions as ATP synthase, is strongly regulated by MgADP inhibition to prevent wasteful ATP consumption (21). The ATP hydrolysis activity of TtV₁ is inhibited rapidly and irreversibly in the presence of ATP. Because the decrease in the ATP hydrolysis rate of EhV₁ was much slower than that of TtV₁ (Fig. 2A) and some EhV₁ molecules showed reversible and irreversible long pauses in the rotation assay, unknown inhibited states of EhV₁ other than MgADP inhibition may exist.

Typical examples of rotation at 40 mm, 100 μm, and 10 μm ATP are shown in Fig. 5. At all ATP concentrations, EhV₁ exhibited stepwise rotation with three intervening pauses separated by 120°, as shown in the x-y trajectories and the distribution of the rotary angle (Fig. 5, insets). Because the 120° steps were completed within 0.2 ms (one to two frames captured at 10,000 fps), the rotation rate was determined primarily by the duration of the intervening pause. At 40 mm ATP, a concentration considerably higher than the K_m, the intervening pauses would represent the catalytic pauses (Fig. 5A) because the expected binding time constant for ATP (0.011 ms, estimated from 1/(40 × 10⁻³ m × 2.2 × 10⁶ m⁻¹ s⁻¹)) was 100-fold smaller than the duration times (3.1 ms) of pauses expected from the V_max (1/(107 × 3)). At 10 μm ATP, a concentration substantially lower than the K_m, the intervening pauses would correspond to the ATP-waiting pauses because ATP binding is rate-limiting under these conditions (Fig. 5C).

FIGURE 5.

Steps and pauses in the rotation of reconstituted EhV₁. Shown are typical time courses of rotation of reconstituted EhV₁ at various ATP concentrations. A, rotation at 40 mm ATP, captured at 10,000 fps. B, rotation at 100 μm ATP, captured at 10,000 fps. C, rotation at 10 μm ATP, captured at 5000 fps. The upper left insets show the x-y trajectories of the centroid of a rotating gold colloid. The lower right insets show the distributions of the rotary angle.

Even at 100 μm ATP, a concentration near the K_m and at which the duration of the ATP-waiting pause approaches that of the catalytic pause, EhV₁ exhibited only three pauses separated by 120° (Fig. 5B). This finding suggests that there are no substeps in the rotation of EhV₁, as is the case for TtV₁ (10, 11).

Buffer-exchange Experiment

To further confirm that the angles of the ATP-waiting pauses observed at low ATP concentrations correspond to those of the catalytic pauses observed at saturating ATP concentrations, we next conducted a buffer-exchange experiment. After recording the stepwise rotation of EhV₁ pausing every 120° at 10 μm ATP, we increased the ATP concentration to 3 mm by infusing buffer containing 3 mm ATP into the flow cell. After buffer exchange, although the rotation rate increased significantly, EhV₁ continued to rotate with discrete 120° steps, pausing at almost the same angles (Fig. 6, A and B). The distributions of the rotary angles indicated three peaks corresponding to three pauses in the rotation (Fig. 6B). To assess the difference in the angular position (Δθ) between 10 μm and 3 mm ATP, the positions of the peaks were determined by fitting the histogram with Gaussian functions and then comparing them with the nearest ones (Fig. 6B). The distribution of Δθ showed a single peak around 0°, with a mean value of 3.2 ± 12° (mean ± S.D., 21 events from seven molecules) (Fig. 6C). On the basis of these data, we confirmed again that the rotation of EhV₁ contains no substeps.

FIGURE 6.

Buffer-exchange experiments. A, time courses of rotation of the same molecule at 10 μm ATP (red) and 3 mm ATP (blue). The ATP concentration was increased from 10 μm to 3 mm. B, distributions of the rotary angle at 10 μm ATP (upper) and 3 mm ATP (lower) shown in A. Δθ represents the angular differences between the pause angles (peaks) before and after buffer exchange. C, distribution of Δθ. The mean value was 3.2 ± 12° (mean ± S.D., 21 events from seven molecules).

Dwell Time Analysis

To obtain the time constants and kinetic parameters for elementary reaction steps such as ATP binding, ATP cleavage, and product release, we analyzed the duration of the pauses. On analyzing four molecules at saturating ATP concentrations (40 mm), all distributions of the durations of the catalytic pauses showed a convex shape (Fig. 7, A–D). At 40 mm ATP, the expected time constant for ATP binding (0.011 ms) was too short to be resolved. Therefore, three elementary reaction steps could occur during the catalytic pauses, namely ATP cleavage, ADP release, and phosphate release. We first attempted to fit the distributions with a model of three consecutive reactions. This failed to improve the fits compared with a model of two consecutive reactions with two time constants. The average time constants from four molecules were 2.4 ± 1.1 and 0.6 ± 0.2 ms (mean ± S.D.) (Fig. 7F, open red circle and square). These values are consistent with the time constants of 2.6 ± 0.1 and 0.5 ± 0.02 ms (mean ± S.E. of fitting, 2097 events) determined by reproducing the accumulated distribution of the pause duration time from all four molecules (Fig. 7E). In addition, at 4 mm ATP (the expected time constant for ATP binding is 0.11 ms), similar accumulated distribution and time constants of 2.4 ± 0.05 and 0.5 ± 0.02 ms (mean ± S.E. of fitting, 4303 events from nine molecules) were obtained (Fig. 7, G and H), consistent with the saturation of the rotation rate at this concentration (Fig. 2B).

FIGURE 7.

Distributions of duration times of the pauses. A–D, distributions of duration times of four single molecules at 40 mm ATP with a 0.2-ms bin width captured at 10,000 fps. The solid curves show fits with a model of two consecutive reactions: constant × (exp(t/τ₁) − exp(t/τ₂)), where τ₁ = 3.0 ± 0.2 ms and τ₂ = 0.5 ± 0.1 ms (mean ± S.E. of fitting, 398 events) (A), 3.5 ± 0.1 ms and 0.6 ± 0.04 ms (863 events) (B), 1.2 ± 0.2 ms and 0.4 ± 0.1 ms (228 events) (C), and 1.8 ± 0.3 ms and 0.8 ± 0.2 ms (608 events) (D). E, accumulated distribution of duration times of all four molecules at 40 mm. The solid curve shows a fit with a model of two consecutive reactions with time constants of 2.6 ± 0.1 and 0.5 ± 0.02 ms (2097 events). F, red dots and square dots indicate τ₁ and τ₂, respectively, shown in A–D. The open red circle and square indicate average τ₁ (2.4 ± 1.1 ms) and τ₂ (0.6 ± 0.2 ms) (mean ± S.D.) from four molecules. Error bars represent S.D. The open blue circle and square indicate τ₁ and τ₂, respectively, shown in E. G, accumulated distribution of duration times (nine molecules) at 4 mm ATP with a 0.2-ms bin width captured at 10,000 fps. The solid curve shows a fit with a model of two consecutive reactions: constant × (exp(t/τ₁) − exp(t/τ₂)), where τ₁ = 2.4 ± 0.05 ms and τ₂ = 0.5 ± 0.02 ms (mean ± S.E. of fitting, 4303 events). H, red circles and square dots indicate τ₁ and τ₂ determined by fitting the individual distributions of duration times of nine single molecules with a model of two consecutive reactions, respectively (each distribution is not shown). The open red circle and square indicate average τ₁ (2.3 ± 0.5 ms) and τ₂ (0.4 ± 0.2 ms) (mean ± S.D.) from nine molecules. Error bars represent S.D. The open blue circle and square indicate τ₁ and τ₂, respectively, shown in G. I–L, distributions of duration times of four single molecules at 10 μm ATP with a 10-ms bin width captured at 5000 or 2000 fps. The solid curves show fits with single-exponential decay: constant × (exp(−t/τ)), where τ = 33 ± 1 ms (mean ± S.E. of fitting, 581 events) (I), 47 ± 6 ms (67 events) (J), 54 ± 1 ms (432 events) (K), and 46 ± 1 ms (978 events) (L). M, accumulated distribution of duration times of all four molecules at 10 μm. The solid curve shows a fit with single-exponential decay with a time constant of 43 ± 1 ms (2058 events). N, the red dots indicate τ shown in I–L. The open red circle indicates average τ (45 ± 9 ms) (mean ± S.D.). Error bars represent S.D. The open blue circle indicates τ shown in M.

These time constants would correspond to (i) ATP cleavage and (ii) ADP and/or phosphate release, although it is currently unclear which time constant corresponds to which elementary reaction step. At a low ATP concentration (10 μm), as discussed above, the duration of the pauses corresponded to that of ATP waiting. Analysis of the four molecules showed that the distributions of the ATP-waiting duration time followed single-exponential decay (Fig. 7, I–L), indicating that EhV₁ consumed one ATP molecule/120° step. The average time constant from four molecules was 45 ± 9 ms (mean ± S.D.) (Fig. 7N, open red circle), which is consistent with a time constant of 43 ± 1 ms (mean ± S.E. of fitting, 2058 events) determined by reproducing the accumulated distribution of the duration time from all four molecules (Fig. 7M). This value corresponds to a k_on(ATP) of (2.3 ± 0.03) × 10⁶ m⁻¹ s⁻¹, which is consistent with that determined by 3 × V_max/K_m ((2.2 ± 0.4) × 10⁶ m⁻¹ s⁻¹), shown in Fig. 2B.

DISCUSSION

In this study, using a single-molecule assay, we have shown that EhV₁ is a rotary molecular motor. To our knowledge, this is the first report showing that a eubacterial V₁ functions as an ATP-driven ion pump under physiological conditions. EhV₁ exhibited two rotational states, namely clear and unclear (Fig. 4). Assuming that the clear rotational state represents the tight chemomechanical coupling of EhV₁, we analyzed this state to elucidate the basic rotational properties of EhV₁. Our hypothesis that the unclear state is caused by unstable interactions between the rotor and stator of EhV₁ must be examined by rotation assay of the entire E. hirae V-ATPase complex, in which the interactions between the rotor and stator are stabilized by two peripheral stalks. To perform this study, we are currently designing an E. coli expression system in which an appropriately tagged recombinant V-ATPase complex can be produced for a rotation assay.

In the clear rotational state, at all ATP concentrations ranging from below to above the K_m, EhV₁ rotated unidirectionally in a counterclockwise direction, exhibiting three pauses separated by 120° (Fig. 5). No substeps were resolved, as has been reported for TtV₁ (10, 11). In contrast, in the region of their respective K_m values, TF₁ and EF₁ have been reported to rotate with six pauses/turn (13–15). Recently, the overall crystal structures of TtV₁ and EhV₁ were shown to be similar (4–7), especially with respect to the interaction sites between the rotor and stator. These structures are distinct from the structure of F₁ (22), although many amino acid residues associated with catalysis in the binding pocket are conserved between V₁ and F₁. These results imply that the degree of similarity in the interactions between the rotor and stator determines the presence or absence of substeps in the rotation.

Table 1 contains a comparison of the kinetic parameters determined by the single-molecule assay for EhV₁, showing values for TtV₁, TF₁, and EF₁. Despite the difference in physiological function between EhV₁ and TtV₁ and notwithstanding the large difference (>30 °C) in the optimal growth temperatures between E. hirae and T. thermophilus, the values for EhV₁ are closer to those for TtV₁ than for TF₁ and EF₁. This result implies that the basic properties of rotary dynamics are determined by their overall structures and that the difference in the physiological function derives from regulatory mechanisms such as MgADP inhibition.

TABLE 1.

Kinetic parameters of V₁-ATPases and F₁-ATPases from different sources

Protein/measurement temperature	kon(ATP)a	kon(ATP) (3 × Vmax/Km)a	τ1/τ2a	Ref.
	m−1 s−1	m−1 s−1	ms
EhV1d
26 ± 1 °C	(2.3 ± 0.03) × 106	(2.2 ± 0.4) × 106	2.6 ± 0.1/0.5 ± 0.02e	This work
TtV1
23 °C	1.5 × 106	0.84 × 106	2.8/2.8	11
24 ± 1 °C		1.39 × 106		21
TF1
23 °C	3.0 × 107	2.6 × 107	1.6/0.71	13
25 ± 1 °C			1.34/0.29	20
25 ± 1 °C		2.2 × 107		27
EF1
23 °C	4.7 × 107	6.4 × 107	0.41/0.29	15

^a The second-order binding rate constant for ATP (k_on(ATP)) was determined from the distribution of the duration of the ATP-waiting pause.

^b The second-order binding rate constant for ATP (k_on(ATP)) was determined from 3 × V_max/K_m.

^c Time constants were determined from the distribution of the duration of the catalytic pause, which corresponds to ATP cleavage and either ADP or phosphate release (or both).

^d The values are the mean ± S.E. of fitting.

^e The values were obtained at 40 mm ATP.

During the unclear rotational state, the centroids of the gold colloid showed wide fluctuations toward the rotation center. It should be noted that EhV₁ nevertheless rotated unidirectionally, implying that even if the interactions between the rotor and stator are not perfect, EhV₁ maintains unidirectional and cooperative rotary catalysis. Recently, rotary catalysis of the rotor-less stator α₃β₃ ring of TF₁ was demonstrated by high-speed atomic force microscopy (23), and we speculate that the stator A₃B₃ ring also likely exhibits rotary catalysis in the absence of the rotor DF subunits.

The chemomechanical coupling scheme of TF₁ has been extensively studied by advanced single-molecule techniques such as a rotation assay of hybrid molecules and single-molecule manipulation with magnetic tweezers (24–26). For a single catalytic site of TF₁, after ATP binding at 0°, ATP cleavage, ADP release, and phosphate release occur at 200°, 240°, and 320°, respectively (26). Further studies on EhV₁ using advanced single-molecule techniques and high-resolution structural analysis will provide details on its chemomechanical coupling scheme. Moreover, comparison of the schemes of V₁ and F₁ from various species will shed light on the general mechanism of rotary molecular motors.