KIF1A Alternately Uses Two Loops to Bind Microtubules

See allHide authors and affiliations

Science  30 Jul 2004:
Vol. 305, Issue 5684, pp. 678-683
DOI: 10.1126/science.1096621


The motor protein kinesin moves along microtubules, driven by adenosine triphosphate (ATP) hydrolysis. However, it remains unclear how kinesin converts the chemical energy into mechanical movement. We report crystal structures ofmonomeric kinesin KIF1A with three transition-state analogs: adenylyl imidodiphosphate (AMP-PNP), adenosine diphosphate (ADP)–vanadate, and ADP-AlFx (aluminofluoride complexes). These structures, together with known structures of the ADP-bound state and the adenylyl-(β,γ-methylene) diphosphate (AMP-PCP)–bound state, show that kinesin uses two microtubule-binding loops in an alternating manner to change its interaction with microtubules during the ATP hydrolysis cycle; loop L11 is extended in the AMP-PNP structure, whereas loop L12 is extended in the ADP structure. ADP-vanadate displays an intermediate structure in which a conformational change in two switch regions causes both loops to be raised from the microtubule, thus actively detaching kinesin.

To move along microtubules, kinesins (1) must alternate rapidly between tightly bound and detached states. How both dimeric (2, 3) and monomeric (4, 5) kinesins achieve this remains unclear. Because the binding energy in the strong-binding state [10 to 20 kBT (3, 4), where kB is the Boltzmann constant and T is absolute temperature] is too large for rapid spontaneous release, the energy for fast detachment of kinesin from the microtubule must come from a step of the ATP hydrolysis cycle. Large change(s) in free energy are expected to occur during four steps: ATP binding, hydrolysis, phosphate release, and ADP release. Both conventional kinesin and KIF1A bind tightly to microtubules in the nucleotide-free state and in the ATP-bound state. In the ADP-bound state, conventional kinesin is detached from microtubules, whereas KIF1A is partially detached and diffuses freely along the microtubule. This is because loose binding of ADP-bound KIF1A is supported by the KIF1 family–specific K-loop at loop L12. A mutant KIF1A that lacks the K-loop detaches from the microtubule in the ADP-bound state, and the dissociation constant markedly varies depending on the type of bound nucleotide, as is true for conventional kinesin (4). For historical reasons, the tightly bound state is called the strong-binding state, and the fully or partially detached state is called the weak-binding state. Recent work detected the phosphate release from a mutant kinesin, which stalls before the detached state (6, 7). This means that detachment occurs just at or after the phosphate release. Thus, the active process to detach kinesin from the microtubule should occur at the transition from the strong-binding state to the weak-binding state.

The active detachment process can be detected in KIF1A because of its property of binding to the microtubule during adenosine triphosphatase (ATPase) cycling. The apparent dissociation constant of KIF1A in the presence of ATP is the weighted average of the equilibrium dissociation constant of various intermediate states during the ATPase turnover. Because the dissociation constant is not significantly different between two major intermediate states, the AMP-PNP–bound and ADP-bound states (Table 1) (fig. S1) (8), the apparent dissociation constant during the ATPase turnover was not expected to be fundamentally different from these values. However, the apparent dissociation constant in the presence of 2 mM ATP was twice the expected value (Table 1) (fig. S1). Considering that the duration of the intermediate states other than the AMP-PNP–bound and ADP-bound states is much shorter, this large apparent dissociation constant implies that KIF1A has another intermediate state with a considerably lower affinity to the microtubule, which likely corresponds to the actively detaching state.

Table 1.

(Apparent) equilibrium dissociation constants (Kd) for microtubules. Kd values are reported as means ± SEM of at least three independent experiments. Conditions: 2 mM nucleotide or its analog, 50 mM imidazole, 5 mM Mg-acetate, 1 mM EGTA, and 50 mM K-acetate, pH 7.4 at 27°C (nd, not determined).

Nucleotide Kd (nM)
Wild type L12View inline L11View inline L8View inline
AMP-PNP 4.2 ± 1.3 6.0 ± 1.4 20.2 ± 4.0 25.0 ± 6.0
ADP 6.8 ± 2.5 23.5 ± 8.4 12.3 ± 4.0 26.5 ± 5.0
ATPView inline 10.8 ± 1.8 40.5 ± 11.8 nd nd
ADP-AlFx 5.9 ± 1.5 7.1 ± 1.7 nd nd
ADP-Vi 21.4 ± 4.3 167 ± 66 nd nd
  • View inline* ATP regeneration system was used to maintain ATP/ADP level.

  • View inline L12: CK1 (4).

  • View inline L11: K261A/R264A/K266A.

  • View inline§ L8: K161A/R167A/R169A/K183A.

  • Because this detached state must occur between the ATP- and ADP-bound states, we examined the KIF1A-microtubule interaction in the presence of two ADP-phosphate analogs, ADP-vanadate (ADP-Vi) and ADP-AlFx (9). Interestingly, the two analogs had different effects. ADP-Vi markedly lowered the affinity of KIF1A to the microtubule, whereas ADP-AlFx did not significantly change the affinity (Table 1) (fig. S1). Thus, ADP-Vi and ADP-AlFx stop the ATPase cycle of KIF1A at two different substates of the ADP-phosphate state. The ADP-Vi state likely corresponds to the actively detaching state (10). Extremely weak binding in the presence of ADP-Vi is not specific to KIF1A. For example, conventional kinesin and the Drosophila kinesin-like proteins ncd and nod also show extremely low affinity for microtubules in the presence of ADP-Vi, whereas ADP-AlFx increases affinity (11, 12). Because the affinity of conventional kinesin in the ADP-AlFx state is closer to the affinity of the ATP-bound state than to the ADP-Vi or the ADP states (13), ADP-AlFx likely stops the ATPase cycle of kinesin at a state between the ATP-bound and ADP-Vi states (Fig. 1).

    Fig. 1.

    Kinesin (KIF1A)–microtubule hydrolysis cycle. Abbreviations: K, kinesin; M, microtubule; T, ATP; D, ADP; P, phosphate. Asterisks refer to the second ATP state (K*T) or the second ADP-phosphate state (K*DP) of kinesin. The main reaction pathway is enclosed (45).

    Thus, biochemical analyses suggest that the affinity of kinesin-type motors to the microtubules undergoes a drastic change during the hydrolysis of ATP into ADP. In the ATP-bound state and in the early ADP-phosphate state (ADP-AlFx), kinesin has its highest affinity to the microtubules (the strong-binding state). This state is followed by a transient late ADP-phosphate state (ADP-Vi), at which point kinesin has its lowest affinity (the actively detaching state). After phosphate release, kinesin in the ADP state has a moderate affinity (the weak-binding state).

    To understand the atomic details of these mechanical states, we solved the crystal structures of KIF1A in the ADP-phosphate (ADP-AlFx and ADP-Vi) states and compared them with known structures of the ADP-bound state (1I5S) (14) and the AMP-PCP–bound state (1I6I) (14), as well as with an AMP-PNP–bound form solved here. As reported for myosin (15), AMP-PCP blocks the ATPase cycle before an isomerization step, which precedes ATP hydrolysis. Thus, the AMP-PCP–bound form would represent a collision complex or a preisomerization state, rather than a prehydrolysis state, as suggested from the lack of interactions between the γ-phosphate of AMP-PCP and the two highly conserved residues in the switch regions (Ser215 and Gly251) (see below) (fig. S2). In this study, we solved the structure in the AMP-PNP–bound state in order to examine the prehydrolysis state.

    The x-ray crystal structures of the motor domain of KIF1A with ADP-Vi, ADP-AlFx, and AMP-PNP were refined to values of 2.2 Å, 2.6 Å, and 1.9 Å, respectively (see table S1 for details of our crystallographic work). To identify the regions that change conformation during ATP hydrolysis, we superimposed these three structures and the previously solved ADP-bound structure with the use of the Cα atoms of the conserved nucleotide-binding loop (P-loop: amino acids 97 to 104), which remains unchanged throughout ATP hydrolysis (Fig. 2) (14). The overall architectures of the ADP-Vi, ADP-AlFx, and AMP-PNP forms are similar to the previously solved ADP and AMP-PCP forms. However, although the central β sheets and surrounding α helices hardly change their conformations [root mean square deviation (RMSD) < 0.6 Å], three loop regions do undergo marked changes (16). These are the switch I region (loop L9; amino acids 202 to 218), the switch II cluster (loop L11, helix α4, loop L12, helix α5, and loop L13; amino acids 248 to 324), and the neck-linker region (strand β9; amino acids 353 to 361), which are all thought to play a crucial role in the nucleotide-driven motility of kinesin-type motors.

    Fig. 2.

    Crystal structures of KIF1A. (A) The AMP-PNP form of KIF1A. The switch I, switch II, and neck-linker regions are highlighted in red. (B) Superposition of AMP-PNP, ADP-AlFx, ADP-Vi, and ADP structures. The AMP-PNP, ADP-AlFx, ADP-Vi, and ADP forms are shown in red, blue, green, and yellow, respectively.

    The switch I loop L9 changes its conformation, along with the movement and release of the γ-phosphate, during ATP hydrolysis. Before isomerization, loop L9 has a tight β-hairpin structure (AMP-PCP form, Fig. 3A). This structure melts so that, in the prehydrolysis state, the conserved serine residue (Ser215) at the C-terminal end of the loop can approach the γ-phosphate (AMP-PNP form) (Fig. 3B) (fig. S2). After hydrolysis, the γ-phosphate moves toward the gap between the switch I loop L9 and the switch II loop L11 (ADP-AlFx form) (Fig. 3C) (fig. S3) (17), and the loop-to-helix transition of loop L9 likely supports the release of the γ-phosphate out of the nucleotide-binding pocket (ADP-Vi form) (Fig. 3D) (fig. S4) (18, 19).

    Fig. 3.

    The conformational changes that occur in the two switch regions and the neck-linker region during ATP hydrolysis. (A to D) AMP-PCP (A), AMP-PNP (B), ADP-AlFx (C), and ADP-Vi (D) forms are shown in light brown, red, blue, and green, respectively. These panels are seen from the upper right in Fig. 2A (indicated by the black arrow). Nucleotides and the coordinating residues around them are shown as ball-and-stick models. Missing C-terminal residues and loops are shown by dashed lines. The structural details around the nucleotide-binding pocket are also shown in figs. S2 to S4 (46). (E and F) The conserved salt bridge between Glu170 and Arg316 (enclosed by a red circle), shown in ball-and-stick models in the AMP-PNP (E) and ADP-Vi (F) forms. Loop L8 and the switch II cluster are shown in dark blue and dark red, respectively. Nucleotides are shown as space-filling models.

    This dynamic structural change in the switch I loop L9 might explain the rapid hydrolysis and release of the γ-phosphate. The γ-phosphate is efficiently released through the “back door” formed by the highly conserved residues Arg216 and Glu253 (fig. S4). In myosin, the salt bridge between these two residues in the prehydrolysis structure closes this back door, which aligns the nucleophilic water with the γ-phosphate, and opening this back door generates a tunnel through which the γ-phosphate leaves the nucleotide-binding pocket (20, 21). KIF1A resembles myosin in the efficient release of the phosphate, which is generated by the swinging movement of the back door. The mutation of the back-door residue Arg216 significantly slows ATP hydrolysis, which also supports the functional importance of these back-door residues in fast ATP hydrolysis (2225).

    The conformational change of the back-door loop L9 not only contributes to the rapid ATPase turnover, but also triggers large conformational changes in the microtubule-binding loops in the switch II cluster. Before hydrolysis, the conserved Gly251 at the N-terminal end of the switch II cluster forms a hydrogen bond with the γ-phosphate (AMP-PNP form) (Fig. 3B) (fig. S2). The hydrolysis of ATP breaks this hydrogen bond and releases L11, the N-terminal loop of the switch II cluster (ADP-AlFx form) (Fig. 3C). The release of L11 triggers the rotation of the whole switch II cluster and the elongation of helix α4 from both ends by shortening and raising of both loops L11 and L12 during the phosphate release (Fig. 3, C and D). The raised conformation of loop L11 is stabilized by a salt bridge that forms between the conserved Glu267 in L11 and the conserved Arg216 in L9, and the hydrogen-bond network involving the γ-phosphate that connects switch I loop L9 and switch II loop L11 (fig. S4) (26, 27). Release of the phosphate breaks these bonds, which results in partial melting of the raised loops L11 and L12 in the ADP form. The rotation of the switch II cluster is supported by the salt bridge between the conserved Glu170 in loop L8 and the conserved Arg316 in helix α5 (Fig. 3, E and F) (28). This rotational movement also leads to the displacement of the neck, as we have proposed previously (14, 29).

    Thus, apparently, the rotational movement of the switch II cluster is supported by two conserved salt bridges: that between Arg216 and Glu253 (back door), and that between Glu170 and Arg316. The back door might serve as a “latch” to keep the switch II cluster to the strained “up” position. The release of the latch triggers the rotation of this helix, which is pulled to the relaxed “down” position by the salt bridge between Glu170 and Arg316. The interaction between loop L8 and the switch II cluster, mediated by the salt bridge between Glu170 and Arg316, might serve as the mechanochemistry coupler that couples the ATPase chemical reaction to the conformational change of the microtubule-binding surface. Consistent with this hypothesis, mutation of Glu170 results in the decoupling of the ATPase turnover from the detachment from the microtubule (6). Interestingly, a recent structural study with another kinesin KIF2C suggested that L8 might serve as the sensor for the microtubule, triggering the closure of the back door (30). The conserved salt bridge between L8 and α5 might also contribute to closing the back door, because the mutation of either Glu170 or Arg316 considerably lowers the microtubule-activated ATPase activity (31).

    Loop L11, helix α4, and loop L12 are the main microtubule-binding elements of kinesin (4, 3133); thus, their structural changes alter the affinity of kinesin to the microtubules. To examine the effect on microtubule binding, we docked the atomic structures of KIF1A and tubulin (34) to a KIF1A-microtubule complex whose structure was determined by cryogenic transmission electron microscopy (Fig. 4B) (14, 33). In the AMP-PNP form, loop L11 of KIF1A is extended down to the H11′ helix of tubulin, and helix α4 of KIF1A is inserted deep into the intradimer groove of tubulin. KIF1A is tilted toward L11 so that L12 is extended up and away from the microtubule. During the hydrolysis of ATP, both loops L11 and L12 are raised from the microtubule, and only helix α4 supports the interaction between KIF1A and the microtubule (ADP-Vi form). The rotation of the KIF1A core around helix α4 and the melting of loop L12 into the flexible conformation then allow loop L12 to interact with the C-terminal E-hook of tubulin in the ADP form while loop L11 is still raised.

    Fig. 4.

    Hydrolysis-induced conformational changes on the microtubule-binding surface of KIF1A. (A) Schematic illustration of the conformational changes in KIF1A. The switch I region (blue) and the switch II cluster (amino acids 251 to 269 in red, 270 to 288 in yellow, and 289 to 305 in green) are shown as ribbon models. Nucleotides are shown as space-filling models. The residues important for γ-phosphate release (Ser215, Arg216, and Glu253) are shown as ball-and-stick models. (B) The KIF1A-microtubule complex seen from the minus end of the microtubule. The microtubules (gray) are shown as space-filling models, except for helices H11 and H12 and the E-hook (blue). See movies S1 and S2 for details.

    These structural models suggest a third transient mechanical state, in which kinesin detaches itself from the microtubule, between the strong-binding and weak-binding states (Fig. 4B). In the ATP-bound state, kinesin is tightly fixed to the microtubule by interactions between loop L11 of kinesin and the H11′ helix of tubulin (the strong-binding state). ATP hydrolysis then generates the actively detaching state by raising the two main microtubule-binding loops L11 and L12. Thus, kinesin is only weakly supported by helix α4 and other minor binding elements. After phosphate release, the flexible loop L12 is extended down onto the tubulin C-terminal E-hook, and this flexible interaction allows one-dimensional diffusion in the weak-binding state (4, 5, 32). In short, three successive states of the kinesin-microtubule interaction exist during ATP hydrolysis, and kinesin uses the energy of ATP hydrolysis to detach loop L11 from the microtubule and to extend another loop (L12) to the microtubule.

    We tested this proposal of alternating use of the two microtubule-binding loops by introducing mutations to these loops. The positively charged residues Lys261, Arg264, and Lys266 in loop L11 were mutated to Ala, and this L11 mutant (K261A/R264A/K266A) was compared with an L12 mutant (CK1) (4) and an L8 mutant in which Lys161, Arg167, Arg169, and Lys183 were mutated to Ala (K161A/R167A/R169A/K183A). Loop L8 was selected for mutation because it corresponds to the third microtubule-binding region and does not change its conformation or position relative to the microtubule during ATPase turnover (14, 33). Relative to the wild-type protein, all of the mutants showed similar ATPase turnover rates but a lower affinity to the microtubule. As previously shown (4), the mutation in loop L12 selectively weakened affinity in the ADP state but not in the AMP-PNP state (Table 1). By contrast, the mutation in loop L11 had a stronger effect in the AMP-PNP state than in the ADP state. The mutation in loop L8 affected both forms to the same extent. Thus, the mutational studies support the proposal that KIF1A uses two loops, L11 and L12, in an alternating manner during ATPase turnover.

    This alternating use of microtubule-binding loops provides structural support for our model of KIF1A's motility (4, 5, 32, 35). The model assumes that KIF1A alternates between two different binding states: the rigor-binding state with the affinity biased toward the forward tubulin subunit, and the diffusive binding state that allows one-dimensional diffusion. The rigor-binding loop L11 is extended toward the forward tubulin subunit. This configuration of loop L11 gives directional bias, or asymmetric binding potential, to the rigor-binding state. The force and binding energy supported by this binding are more than 10 pN and 10 kBT (3, 4), respectively. This strong binding is broken by the energy of ATP hydrolysis. ATP hydrolysis causes conformational changes that first raise loop L11 from the microtubule and then bring loop L12 into contact with the microtubule. The flexible tether between loop L12 and the E-hook of tubulin allows diffusive binding in the ADP form. Thus, KIF1A kinesin takes three structures successively: L11 down, both L11 and L12 up, and L12 down. It is therefore possible to surmise that a fourth structure might exist: both L11 and L12 down (at the transition from the L12 down, ADP state, to the L11 down, ATP state). Bringing both L11 and L12 to the microtubule might open the nucleotide-binding pocket and accelerate ADP release. Further structural analysis of the nucleotide-free state or the ADP-release intermediate state of KIF1A will be required to examine this possibility.

    Finally, we note that the design principle of KIF1A is similar to that of heterotrimeric guanine nucleotide–binding proteins (G proteins) (36, 37) and other proteins such as protein kinases (38). The binding energy of KIF1A to the microtubule is used to achieve a mechanical step of KIF1A (5). With G proteins and protein kinases, the binding energy to an effector molecule or substrate is used to cause structural changes. Nucleophilic attack by water molecules or serine residues on the γ-phosphate catalyzes hydrolysis. Hydrolysis triggers a conformational change at the binding surface (the switch II helix) that detaches the enzyme from the target molecule. The binding between a G protein and its effector molecule, or between kinase and its substrate, is so strong that energy is required for dissociation. The energy of hydrolysis is used for this active detachment, which is necessary to cycle the reaction. This conserved strategy might reflect the evolutionary pathway of these classes of proteins, and also suggests a design principle for nanomachines.

    Supporting Online Material

    Materials and Methods

    Table S1

    Figs. S1 to S4

    Movies S1 and S2


    References and Notes

    Stay Connected to Science

    Navigate This Article