Catalytic robustness and torque generation of the F1-ATPase
The F1-ATPase is the catalytic portion of the FoF1 ATP synthase and acts as a rotary molecular motor when it hydrolyzes ATP. Two decades have passed since the single-molecule rotation assay of F1-ATPase was established. Although several fundamental issues remain elusive, basic properties of F-type ATPases as motor proteins have been well characterized, and a large part of the reaction scheme has been revealed by the combination of extensive structural, biochemical, biophysical, and theoretical studies. This review is intended to provide a concise summary of the fundamental features of F1-ATPases, by use of the well-described model F1 from the thermophilic Bacillus PS3 (TF1). In the last part of this review, we focus on the robustness of the rotary catalysis of F1-ATPase to provide a perspective on the re-designing of novel molecular machines.
KeywordsF1-ATPase ATP synthase Single-molecule techniques Molecular motor
The FoF1 ATP synthase is a ubiquitous enzyme, spreading across all kingdoms of biology. It is found in the inner membrane of mitochondria, the thylakoid membrane of chloroplasts, and the plasma membranes of bacteria and some archaea. However, most archaea, and some eukaryotic bacteria, employ the evolutionary related V-type ATPase as an ATP synthase (Yoshida et al. 2001). FoF1 ATP synthases catalyze ATP synthesis using ADP and inorganic phosphate as substrates. ATP synthesis is a thermodynamically uphill reaction. To catalyze ATP synthesis, the FoF1 ATP synthase utilizes an electrochemical potential of protons or sodium ions across biological membranes (proton motive force, pmf, or sodium motive force, smf) as the driving force (Abrahams et al. 1994; Dimroth et al. 2006; Yoshida et al. 2001). The use of chemiosmotic catalysis by the ATP synthase is such a prevalent reaction in biology that the FoF1 is widely considered to be one of the most ancient enzymes. A scenario on the last universal common ancestor (LUCA) proposes that an ancestor molecular complex of ATP synthase prebiotically emerged to confer energy production for LUCA (Lane and Martin 2012), although there are arguments against this scenario (Jackson 2016).
Fo and F1
Upon ATP hydrolysis, the β subunit undergoes a large conformational change to induce the unidirectional rotation of the γ subunit, counterclockwise when viewed from the membrane side (Abrahams et al. 1994; Masaike et al. 2008; Noji et al. 1997). The ε subunit binds to the protruding domain of the γ to form the binding interface to Fo. The ε subunit is also known to be a catalytic regulator (Cingolani and Duncan 2011; Feniouk et al. 2010; Hara et al. 2001; Keis et al. 2006; Tsunoda et al. 2001). The δ subunit attaches to the bottom tip of the α3β3-ring (Dunn and Chandler 1998; Ogilvie et al. 1997).
Coupling between F1 and Fo
In the complete complex, Fo and F1 are connected via two stalks: the central and the peripheral. The central stalk is the rotary shaft of FoF1 ATP synthase, composed of the rotor parts of Fo and F1 (Stock et al. 1999; Wächter et al. 2011). Together, the γε complex and the c-ring form the combined entity of the rotary shaft. Fo and F1 form the second stalk, termed the peripheral stalk. This is comprised of the δ subunit and the b 2 dimer. The peripheral stalk is located on the lateral side of ATP synthase and firmly holds both of the stator parts of F1 and Fo from the outside, spanning from the bottom tip of F1 to the opposite side of the lipid bilayer (Rodgers and Capaldi 1998; Stock et al. 1999). In the ATP synthase, Fo and F1 exert rotational torque against each other (Junge et al. 1997; Noji and Yoshida 2001; Oster and Wang 2000). Under physiological conditions where pmf is sufficient and the rotational torque of Fo surpasses that of F1, Fo rotates the γ subunit in the reverse direction, a clockwise rotation viewed from the membrane side (from the top of Fig. 1).
The reverse rotation of F1 induces the reverse conformational change of the α3β3-ring, leading to the reverse reaction of ATP hydrolysis (ATP synthesis). When the free energy of ATP hydrolysis dominates, and the F1 generates a larger torque than Fo, F1 reverses the rotation of Fo to induce active proton pumping and build up pmf across the membrane. Thus, Fo and F1 interchange the energy between pmf and free energy of ATP hydrolysis via the mechanical rotation of the rotor complex. For efficient energy interconversion, the central stalk and the peripheral stalk have to be resistant against mechanical torsional stress (Wächter et al. 2011). The δb2 subcomplex is known to form a stable complex (McLachlin et al. 1998). Regarding the central stalk, although the c subunit oligomer ring has multiple symmetric binding sites for the γε complex (Müller et al. 2004; Schulenberg et al. 1999; Watts et al. 1995), there have been, to date, no reports of slippage at the interface of the γε subcomplex and the c-ring. On the other hand, some studies suggest that the stalks are twisted in the ATP synthase at work. Although the elasticity of the peripheral stalk has been reported previously (Junge et al. 2009), recent studies suggest that the γ subunit of the central stalk is twisted (Okazaki and Hummer 2015; Vahidi et al. 2016). It remains elusive how the elasticity of the stalks contribute to energy transduction within the ATP synthase.
All of F1, of which crystal structures are available, show a pseudo three-fold symmetry in the αβ stator ring (Abrahams et al. 1994; Böttcher and Gräber 2000; Cingolani and Duncan 2011; Stocker et al. 2005). Thus, the stoichiometry of F1 (m) should be 3 ATPs per turn. Actually, all of the F1 tested in the rotation assay show stepping rotation with 120° intervals (Bilyard et al. 2013; Konno et al. 2006; McMillan et al. 2016; Suzuki et al. 2014; Yasuda et al. 1998). Thus, it is well established that m = 3.
On the other hand, the stoichiometry of proton translocation per turn (n) is variable between species, yet is invariant within species. It is assumed that n is determined by the number of c subunits in the oligomeric c-ring rotor of Fo (Junge et al. 1997). While the bacterial Fo has generally 10–15 c subunits in the oligomer ring (Ballhausen et al. 2009; Matthies et al. 2009; Meier et al. 2005a; Pogoryelov et al. 2005, 2009), mammalian mitochondrial Fo has 8 (Watt et al. 2010) and, depending on the origin, chloroplasts have 13–15 subunits in the c-ring (Pogoryelov et al. 2007). Several groups reported that FoF1 from Escherichia coli makes the rotation with 10 steps, which is consistent with the number of c subunits being 10 (Ballhausen et al. 2009; Düser et al. 2009; Ishmukhametov et al. 2010). It should be noted that the apparent number of rotation steps of FoF1 depends on the experimental conditions, because 10-step rotation should be observed only when the proton translocation or conformational change of Fo determines the overall rate of rotation. Otherwise, step rotation of FoF1 should represent the structural symmetry of F1, as previously reported (Watanabe et al. 2013).
When we consider possible uncoupling rotation of Fo, i.e., slipped rotation without accompanying proton translocation, this value sets the upper limit of the H+/ATP ratio. Before structural information on the c-ring became available, the H+/ATP ratio was considered to be an integer, such as 3 or 4, and to be common among species. Recent studies show that this is not the case. The H+/ATP ratio estimated from the structure is 2.7 for mammalian mitochondrial F1, 3.3 for E. coli FoF1, and 4 or more for chloroplast FoF1 (Ballhausen et al. 2009; Pogoryelov et al. 2005; Watt et al. 2010). Biochemical analysis supports that FoF1 with a larger number of the c subunit shows a larger H+/ATP ratio (Turina et al. 2016). One example is a comparative study on the H+/ATP ratio of FoF1 from Saccharomyces cerevisiae or Spinacia oleracea (Petersen et al. 2012), each of which has 10 or 14 c subunits (Symersky et al. 2012; Vollmar et al. 2009). The experimentally determined H+/ATP ratio was 2.9 and 3.9, respectively. The reason for the lower values than expected is unclear, although it could be attributable to the slipped rotation or a biased experimental error.
A higher H+/ATP ratio means that the Fo must exert higher torque to induce ATP synthesis under a given pmf. The difference in the H+/ATP ratio would represent different physiological requirements. When pmf is sufficiently high and stable over the life cycle of a cell, ATP production requires less protons (such as in mammalian mitochondria). However, when pmf is small and/or environmental ΔpH is inverted, as it happens in alkaliphilic extremophiles (McMillan et al. 2009) or chloroplasts (Pogoryelov et al. 2005), the Fo requires more protons to ensure constant ATP production.
The modulation of c-rings between organisms must be subtle, yet it is clearly essential, since it has been demonstrated, using mutagenesis, that a minor modification at the c–c interface can modulate the stoichiometry of the c-ring (Pogoryelov et al. 2012). The rotation-mediated energy coupling mechanism of FoF1 ATP synthase would ease the modulation of the H+/ATP ratio. This mechanism may be an advantageous feature to allow the adaptation of the ATP synthase for a variety of environmental conditions.
Basic properties as a motor
Rotation assays of the F1-ATPase
Although viscous friction-limited conditions allow us to estimate the rotational torque individual molecules generate (see the section entitled “Torque”), detailed rotation dynamics such as stepping rotation and short-lived pausing states are frequently unclear due to the slow response of the probes used. For a detailed investigation of stepping rotation, nanoparticles of size several tens of nanometers, such as gold colloids or gold nanorods, are regarded as the most useful (Spetzler et al. 2006; Ueno et al. 2010; Yasuda et al. 2001). Nanoparticles do not behave like larger probes. Their response time is much faster, thereby enabling the observation of short-lived rotational pauses of F1. In the following section, the basic rotation properties of F1 revealed by rotation assay are discussed. Due to the breadth of knowledge available, we focus on the thermophilic Bacillus PS3 (TF1).
Step rotation and reaction scheme
Resolving mechanical motion into steps is a starting point to elucidate the chemomechanical coupling reactions scheme of molecular motors (Svoboda and Block 1994; Yildiz et al. 2003). Differing from architecturally simpler molecular motors like myosin and kinesin, F1 has three reaction centers. Naturally, this results in a comparatively more challenging puzzle to solve the reaction scheme. One may understand the difficulty when considering at which angle ATP is hydrolyzed after binding. It is important to note that the F1 has three catalytic angles in one turn and, thereby, there are, principally, three possible angles for each ATP to be hydrolyzed. A review paper that concisely introduces how this difficult puzzle has been addressed is available elsewhere (Okuno et al. 2011). Each reaction center conducts a single ATP hydrolysis event per 360° γ rotation, while the reaction phase of a catalytic site is always different from each other by 120°. Here, it should be stated that there is an unsettled argument on the timing of inorganic phosphate release. We propose that it occurs at 320°, i.e., during the third catalytic pause after binding (Watanabe et al. 2010), while Kinosita’s group support a 200° model (Adachi et al. 2007; Shimo-Kon et al. 2010). We are gaining strong supporting evidences for a 320° model from theoretical studies and structural analysis of MF1 (Rees et al. 2012); however, it remains to be resolved in TF1. We are certain further discussion and analysis will resolve this curious facet.
Interestingly, the rotation assay on mammalian F1 (human mitochondria F1; hMF1) showed that the 120° step is resolved into three substep positions at 65°, 25°, and 30° (Suzuki et al. 2014). Although the reaction scheme of MF1 apparently differs from that of TF1, it can be considered to be a variation of the reaction scheme of TF1, where the hydrolysis dwell and inorganic phosphate release dwell are split into different angles. Interestingly, a recent intensive analysis using data mining techniques revealed that TF1 makes a small rotational movement during the catalytic pause that had not been previously resolved using conventional data analysis (Li et al. 2015). Thus, the reaction scheme seems likely to be conserved among species, although there are some variations in substep size and sequence of reaction that might reflect physiological requirements. Regarding this point, a theoretical study suggests that the position of the substeps is not critical for energy coupling efficiency (Mukherjee and Warshel 2015).
The motion of micron-sized objects (Brownian particles) in fluid is overdamped. The acceleration and deceleration time for Brownian particles is extremely short, meaning that the Brownian motion can be treated as a Markov process. When the F1 rotates a probe in a buffered solution, the viscous friction against the rotating probe is always balanced with the torque that the F1 generates. Thus, the torque of the F1 can be estimated from the viscous friction determined from the probe size (i.e., the radius for a spherical probe) and observed angular velocity (Noji et al. 2001; Pänke et al. 2001; Yasuda et al. 1998). The torque of TF1 has been repeatedly reported to be around 40 pNnm (Hayashi et al. 2010; McMillan et al. 2016; Noji et al. 2001; Ueno et al. 2014). In addition, the angular velocity profile is almost constant against rotational angle. Thus, TF1 is generally considered to generate a constant 40 pNnm of torque, irrespective of rotational angle.
One possible concern about this torque estimation is that the torque could be underestimated. This is because the viscosity of fluid near the surface is known to be higher than that in the middle. On the other hand, the surface effect could be minor, because a probe is lifted by at least 20 nm perpendicular to the coverslip surface by the molecules involved in the rotation assay (F1, biotin, streptavidin, and polymer cushion on the probe surface). Furthermore, rotating particles have a slight elevation angle, being apart from the surface at the rotation edge of the probe where viscous friction is dominant. This concern was addressed, at least partially, by application of a statistical physical theory known as the fluctuation theorem. A variation of the fluctuation theorem formula adjusted to the rotation assay parameters allows for an estimation of entropy generation, i.e., the force generation only from the trajectory of probes without the consideration of physical parameters requisite for the previous method, such as fluid viscosity or size or shape of probes. A sole, but critical, requirement of this analytical method is the rapid and precise recording of the probe position. This method reconfirmed that TF1 generates a torque of 40 pNnm (Hayashi et al. 2010).
The torque of 40 pNnm has important implications for the energy coupling of TF1. Torque multiplied by angular displacement gives the work done by the molecular motor. In the case of the F1, the estimated work that the F1 does per 120° rotational movement (=single turnover of ATP hydrolysis) is approximately 80 pNnm (Yasuda et al. 1998). This is consistent with the free energy of ATP hydrolysis in a cell where ΔG’ ATP is 48 kJ/mol, which is equivalent to 80 pNnm/molecule. In addition, theoretical studies suggest that a constant torque over rotation angle is also reasonable for efficient energy coupling (Oster et al. 2000). However, recent reports challenge this view, which invite revision of the current understanding of the torque profile. Saita et al. (2015) reported that the torque profile showed three sawtooth-like peaks in a 120° step. Because the observed profile is not consistent with any reported substep positions, a reasonable explanation for this conflict with previous reports is yet to be described. Sielaff (2016) reported a different torque profile. They observed that the instantaneous angular velocity of rotation probed with a gold nanorod showed a distinct peak during a 120° step. Because the nanorod has a low viscous drag coefficient, it is not clear if the observed angular velocity represents the genuine torque profile of the F1 or it includes some effect from intermolecular friction between the rotor and the stator of the F1.
We should note here that the torque could be different among species. Another model F1 well studied with rotation assay is the F1 from Escherichia coli (EF1). Although several groups have reported different values, the reported torque of EF1 ranges from 30 to 63 pNnm (Bilyard et al. 2013; Hornung et al. 2008; Pänke et al. 2001). Recently, an F1 from a thermoalkaliphilic bacterium, Caldalkalibacillus thermarum, was reported to generate torque of over 50 pNnm (McMillan et al. 2016). The V1 motor of V-type ATPase from bacteria is also well studied using the rotation assay. There is a clear trend that V1-ATPases generate a lower torque than F1-ATPases: the V1 from Thermus thermophilus generates 30–35 pNnm and the Na+-transporting V-ATPase generates 23 pNnm of torque, respectively (Hayashi et al. 2010; Imamura et al. 2005; Ueno et al. 2014). Although more data are required, we can see a tendency that F1 or V1 dedicated only to ATP synthesis in a cell generates a larger torque than motors working as ATP-driven ion pumps (McMillan et al. 2016).
A milestone to meet in elucidating the mechanochemical coupling of molecular motor proteins is to identify which reaction step is principally responsible for force generation. ATP hydrolysis is resolved into at least four elementary steps: ATP binding, hydrolysis, ADP release, and inorganic phosphate release. This is a technically challenging question to address. Force measurement is a straightforward strategy. However, the catalytic reactions following ATP binding are often too fast to be resolved using force measurement as a sole strategy. Allostericity among multiple reaction sites on a motor protein also hampers the identification of the force-generation step. Another strategy that gives clues to this process is structural analysis. Interestingly, after nucleotide binding, large conformational rearrangements have been observed in the crystal structures of motor proteins.
The crystal structures of the F1 have revealed distinct conformational difference between the β subunits with and without bound nucleotide (Fig. 2) (Cingolani and Duncan 2011; Duncan et al. 1986; Ferguson et al. 2016; Menz et al. 2001). Several crystal structures of F1-ATPases prepared in different conditions and with different ligands have been reported so far (Abrahams et al. 1996; Braig et al. 2000; Cabezón et al. 2003; Kagawa et al. 2004; van Raaij et al. 1996). The β subunits are the principal torque generators, and assume two distinct conformations in most structures. The β subunit without bound nucleotide is in open conformation, while the β subunit with bound nucleotide takes closed conformation, in which the C-terminal domain swings toward the γ subunit (Fig. 2b) (Abrahams et al. 1994). From this observation, the ATP-binding step is proposed to be a major torque-generating step. This is quite consistent with the observation that the substep size induced by ATP binding is two times larger than that initiated after hydrolysis and inorganic phosphate release (Yasuda et al. 2001).
Thus, many lines of experimental results have shown that the ATP-binding process, or more precisely the affinity change to bound ATP, is a major torque-generation step in F1 catalysis, while the chemical cleavage of ATP has relatively minor contribution. It is less well understood how torque is generated upon ADP and inorganic phosphate release. Classic biochemical studies have shown that the inorganic phosphate uptake step is an energy-requiring step during ATP synthesis, suggesting that the inorganic phosphate-releasing step is another torque-generation step. This idea was supported by kinetic analysis on rotation of F1 in the presence of excess inorganic phosphate (Adachi et al. 2007).
Robustness of torque transmission
The stator–rotor interactions have been actively studied since 1994, when the crystal structure of F1 became available (Abrahams et al. 1994) and many residues that impair catalysis were identified. Subsequently, the rotation assay was established, and some of these mutants were closely examined. No significant impact on the torque generation was observed using these mutants; however, kinetic parameters were affected to some extent (Hara et al. 2000; Omote et al. 1999). The robustness of the torque transmission between stator and rotor was more clearly shown by Kinosita’s group. They showed that the F1 still rotates even after being devoid of the entire γ subunit coiled-coil axle (Fig. 5b) (Furuike et al. 2008). Although the rotation rate of 1 Hz was much slower than that of the wild-type F1 (∼200 Hz) and the torque was too low to measure, the rotation was apparently unidirectional. This finding came as a big surprise. This result means that, for the most part, the stator–rotor interaction is dispensable for unidirectional rotation. The remaining interaction in the axle-less F1 was the upper half of the orifice interaction.
Following this, the orifice interaction for torque transmission became the focus of the following study. The helix-turn-helix (HTH) structure of the β subunits forming the main interface of the stator orifice was investigated (Usukura et al. 2012). Figure 5a highlights the HTH structure in blue. The HTH structure is a part of the C-terminal domain of the β subunit, which makes a large swing motion upon nucleotide binding (Abrahams et al. 1994; Masaike et al. 2008). The HTH structure was shortened by the deletion of one or two turns (Fig. 5c). The resultant mutant F1 still showed clear rotation (Usukura et al. 2012). The orifice interaction was comprehensively shown to be dispensable. This point was confirmed in another mutagenesis study where all residues of the HTH in contact with the γ subunit were substituted with alanine or glycine residues (Fig. 5d) (Tanigawara et al. 2012). Surprisingly, the mutant F1 with Ala substitutions generated a torque comparable with the wild-type F1 values (97% of the wild-type), suggesting that specific interactions at the orifice interface are not entirely requisite for full torque transmission. The mutant F1 with Gly substitutions replacing 22 residues of the HTH still rotated; however, the torque was reduced to 60% of the wild-type. Molecular dynamics simulations suggested that the Gly mutant completely lost the rigidity of the HTH structure, while the Ala mutant still keeps the structure of the HTH (Tanigawara et al. 2012). These findings suggest that full torque transmission at the orifice interface does not require specific interaction, but the rigidity of the HTH structure. It is highly likely that the HTH structure works as a “push rod” for the β subunit to induce unidirectional rotation.
Thus, the series of extensive mutation studies revealed that all parts of the stator–rotor interface are dispensable, and that the torque transmission mechanism is far more robust than previously thought. The mutagenesis studies suggest that the orifice interaction is responsible for roughly half of the torque transmission. The sleeve interaction probably transmits the other half of the torque. The remaining issue to be explored is how the α3β3-ring generates torque at the sleeve interaction. From the crystal structures, it is not clear what type of conformational change is responsible for the torque transmission at the sleeve interaction.
Robustness of allostericity: from dictator model to professor model
The robustness of torque transmission also implies the robustness of the cooperativity among the three catalytic reaction centers. The rotation of axles-less F1 brought surprise not only in the context of the torque transmission, but also in that of the allostericity of F1. This is because, at the time, the so-called γ-dictator model was prevailing. This model assumes that the interaction with the γ subunit fully controls the timing of catalysis in the β subunits. Another underlying reason is a biochemical study which showed that the isolated α3β3-ring did not show obvious cooperativity among catalytic sites (Kaibara et al. 1996). The following experiments that reinforced the γ-dictator model were demonstrations that ATP synthesis is possible upon forced reverse rotation of F1 (Itoh et al. 2004; Rondelez et al. 2005). These experiments showed that the equilibrium constant of ATP hydrolysis is over 105 at room temperature in aqueous solution, and can be easily modulated to be less than 1 by controlling the angle orientation of the γ subunit.
This is the decisive evidence that the α3β3-ring has intrinsic cooperativity among the three catalytic sites, and that the γ subunit dictator model is likely incorrect for F1. However, the rate of the observed conformational transition, 1–4 Hz, was remarkably slower than with the γ subunit, being 100–200 Hz. In addition, the occasional back-step was also observed, suggesting that the cooperativity does not reach perfection without the γ subunit. Therefore, we propose that the “professor” model might be a better representation for the allostery of F1 than “intrinsic cooperativity”. Here, “professor” means an existence that is not indispensable but enhances the activity of a system or organization.1
Robustness of catalytic power
Robustness was also revealed in the catalytic power of the F1. Although the F1 is unique in its high reversibility and efficiency of the chemomechanical coupling, the F1 shares common structural features of the catalytic site with other nucleotide-driven molecular machines. One of the most well conserved structural features is the phosphate-binding motif, the so-called p-loop, or the Walker motif B that is found in most NTPases. The p-loop motif has a GXXXXGKT/S sequence (Walker et al. 1982). The seventh lysine, the p-loop lysine, is the most conserved, and is the gold standard target to knock out the catalytic activity of any described NTPase. In support of this, when substituted with another non-charged residue such as alanine, the catalytic activity reduces to undetectable levels in biochemical analysis (Senior and Al-Shawi 1992; Yagi et al. 2009). Substitution of the p-loop lysine in F1 is known to abolish all ATPase activity (Hanson and Whiteheart 2005).
The F1-ATPase has another well-conserved charged residue directly associating with bound ATP. Many of the NTPase proteins share a glutamic acid residue that seems to associate with bound nucleotide via a coordinated water molecule at the distal end of the γ-phosphate. This glutamic acid residue is critically important for the catalysis of F1 (Shimabukuro et al. 2003). When substituted with an aspartic acid, the rate of F1 catalysis is reduced by over 100-fold. When substituted with non-charged residue, the catalytic activity was undetectable in biochemical assays (Amano et al. 1994, 1999). Previously, it was proposed that the glutamic acid residue activates the coordinated water by removing a proton and induces the in-line attachment of the activated water to the γ-phosphate of ATP. Therefore, the glutamic acid residue was termed a “general base”. However, a recent theoretical study revised the working mechanism of the glutamic acid residue, showing that hydrolysis at the γ-phosphate spontaneously occurs before the protonation of the glutamic acid residue (Hayashi et al. 2012). The proposed role of this residue is to build a hydrogen network for the enhancement of proton transfer, the kinetically rate-determining step of ATP hydrolysis.
Another common feature of the catalytic site of F1 with other NTPase proteins is the catalytic arginine residue, the so-called arginine finger. The arginine finger was first found in a G-protein activating protein (GAP), and was proposed to be a catalytic switch to initiate the hydrolysis of GTP bound on G-protein (Ahmadian et al. 1997). The corresponding arginine residues were widely found in other NTPases, including RecA type proteins, of which F1 is a more distal family member, and to the AAA+ family proteins (Hanson and Whiteheart 2005). The arginine finger of F1 resides on the α subunit. Many lines of experimental research have shown that the arginine finger has a crucial role in the catalysis of F1 (Hatch et al. 1995; Komoriya et al. 2012).
These three charged residues of F1 were identified as catalytically critical residues in theoretical studies (Dittrich et al. 2004). Mutagenesis studies also showed that, when these residues are substituted with non-charged residues such as alanine, the catalytic power is abolished (Senior and Al-Shawi 1992; Yagi et al. 2009). Alanine mutants at these residues were re-investigated in rotation assay systems to confirm their critical role in catalysis (Watanabe et al. 2014). Against all expectations, all of the mutants showed unidirectional rotation. The rotational rate was significantly lower by a factor of 103 to 104. Such a low catalysis rate is not detectable in biochemical analysis. While the large impact of the mutation on catalysis is consistent with previous reports and theoretical analysis, the robustness of the catalytic power is unexpected. This finding reveals that the catalytic site architecture is designed to be quite robust against both point mutations and possibly more severe mutagenic perturbations.
Interestingly, all of the alanine mutants had a lower torque than the wild-type enzyme. The most remarkable impact was found in the alanine substitution at p-loop lysine, causing a 75% reduction of torque. Arginine finger and catalytic glutamic acid residue mutants also resulted in a torque reduction of 50%. This is in contrast to the impact of chemical modification of ATP on torque generation (Arai et al. 2014). The F1 has broad nucleotide specificity, hydrolyzing other nucleotides: GTP, CTP, and UTP, albeit with changes in the catalysis rate. Even when the base structure is removed, the F1 hydrolyzes base-free nucleotide, inducing γ subunit rotation. The impact of the modification or depletion of base in kinetics was remarkable; the binding constants of UTP and base-free nucleotide were 103 and 106 times slower than ATP. However, rotational torque was retained at a comparable level to that of ATP-driven rotation. This means that the interaction with the base structure of nucleotide is critical for the enhancement of binding, but not for torque generation. This is in contrast to the impact of mutation at the phosphate-binding residues, which largely impaired the kinetic power as well as torque generation (Arai et al. 2014; Watanabe et al. 2014).
During the past two decades, single-molecule rotation assays have uncovered many aspects of the F1-ATPase catalytic mechanism. Basic properties of the F1 as a rotary motor, such as step size, rotational velocity, torque, and kinetic parameters of rotation, were all clearly revealed. Most parts of the reaction scheme have also been elucidated in combination with knowledge from structural analysis and biochemical studies. Theoretical studies have also been exceedingly helpful in contributing to the elucidation of the molecular mechanism of the F1. The torque-generation mechanism is also partly revealed. However, one must note that analysis and interpretation do not guarantee our understanding. A practical test would be an engineering approach where we design, build, and test chimeras of molecular motors. A hybrid motor of F1 and FliJ is an initial step toward the engineering approach (Pogoryelov et al. 2012), but we are still clearly at the preliminary stages of this approach. Theoretical toolboxes for de novo designing of proteins are growing rapidly. By combining the rational designing methods with directed evolution technology, we will soon be able to build de novo molecular motors. The feedback loop between analytical approaches represented by single-molecule rotation assays, and structural analysis, and engineering approaches will further our understanding of mechanochemical coupling mechanisms and design principles in naturally occurring protein machinery.
Of course, this is a self-deprecating analogy. It could only be the case in Japan.
This work was supported in part by Core Research for Evolutional Science and Technology (CREST) from Japan Science and Technology Agency (JST), and Grant-in-Aids for Scientific Research (nos. 18074005 and 22247025 to H.N.) from the Ministry of Education, Culture, Sports, Science and Technology of Japan. We thank all members of our laboratory.
Compliance with ethical standards
Conflict of interest
Hiroyuki Noji declares that he has no conflicts of interest. Hiroshi Ueno declares that he has no conflicts of interest. Duncan G. G. McMillan declares that he has no conflicts of interest.
This article does not contain any studies with human participants or animals performed by any of the authors.
- Amano T, Matsui T, Muneyuki E, Noji H, Hara K, Yoshida M, Hisabori T (1999) α3β3γ complex of F1-ATPase from thermophilic Bacillus PS3 can maintain steady-state ATP hydrolysis activity depending on the number of non-catalytic sites. Biochem J 343(Pt 1):135–138PubMedPubMedCentralCrossRefGoogle Scholar
- Bilyard T, Nakanishi-Matsui M, Steel BC, Pilizota T, Nord AL, Hosokawa H, Futai M, Berry RM (2013) High-resolution single-molecule characterization of the enzymatic states in Escherichia coli F1-ATPase. Philos Trans R Soc Lond Ser B Biol Sci 368:20120023. doi: 10.1098/rstb.2012.0023 CrossRefGoogle Scholar
- Hayashi S, Ueno H, Shaikh AR, Umemura M, Kamiya M, Ito Y, Ikeguchi M, Komoriya Y, Iino R, Noji H (2012) Molecular mechanism of ATP hydrolysis in F1-ATPase revealed by molecular simulations and single-molecule observations. J Am Chem Soc 134:8447–8454. doi: 10.1021/ja211027m PubMedCrossRefGoogle Scholar
- Le Trong I, Freitag S, Klumb LA, Chu V, Stayton PS, Stenkamp RE (2003) Structural studies of hydrogen bonds in the high-affinity streptavidin–biotin complex: mutations of amino acids interacting with the ureido oxygen of biotin. Acta Crystallogr D Biol Crystallogr 59:1567–1573PubMedCrossRefGoogle Scholar
- Pogoryelov D, Klyszejko AL, Krasnoselska GO, Heller EM, Leone V, Langer JD, Vonck J, Müller DJ, Faraldo-Gómez JD, Meier T (2012) Engineering rotor ring stoichiometries in the ATP synthase. Proc Natl Acad Sci U S A 109:E1599–E1608. doi: 10.1073/pnas.1120027109 PubMedPubMedCentralCrossRefGoogle Scholar
- Schulenberg B, Aggeler R, Murray J, Capaldi RA (1999) The γε-c subunit interface in the ATP synthase of Escherichia coli. Cross-linking of the ε subunit to the c subunit ring does not impair enzyme function, that of γ to c subunits leads to uncoupling. J Biol Chem 274:34233–34237PubMedCrossRefGoogle Scholar
- Shimabukuro K, Yasuda R, Muneyuki E, Hara KY, Kinosita K Jr, Yoshida M (2003) Catalysis and rotation of F1 motor: cleavage of ATP at the catalytic site occurs in 1 ms before 40° substep rotation. Proc Natl Acad Sci U S A 100:14731–14736. doi: 10.1073/pnas.2434983100 PubMedPubMedCentralCrossRefGoogle Scholar
- Tsunoda SP, Rodgers AJ, Aggeler R, Wilce MC, Yoshida M, Capaldi RA (2001) Large conformational changes of the ε subunit in the bacterial F1F0 ATP synthase provide a ratchet action to regulate this rotary motor enzyme. Proc Natl Acad Sci U S A 98:6560–6564. doi: 10.1073/pnas.111128098 PubMedPubMedCentralCrossRefGoogle Scholar
- Usukura E, Suzuki T, Furuike S, Soga N, Saita EI, Hisabori T, Kinosita K, Yoshida M (2012) Torque generation and utilization in motor enzyme F0F1-ATP synthase: half-torque F1 with short-sized pushrod helix and reduced ATP synthesis by half-torque F0F1. J Biol Chem 287:1884–1891. doi: 10.1074/jbc.M111.305938 PubMedCrossRefGoogle Scholar
- Wächter A, Bi Y, Dunn SD, Cain BD, Sielaff H, Wintermann F, Engelbrecht S, Junge W (2011) Two rotary motors in F-ATP synthase are elastically coupled by a flexible rotor and a stiff stator stalk. Proc Natl Acad Sci U S A 108:3924–3929. doi: 10.1073/pnas.1011581108 PubMedPubMedCentralCrossRefGoogle Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.