- Email: [email protected]
Kinesin passing permanent blockages along its protofilament track
Biochemical and Biophysical Research Communications 395 (2010) 490–495
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Kinesin passing permanent blockages along its protofilament track Kerstin Dreblow, Nikolina Kalchishkova, Konrad J. Böhm * Leibniz Institute for Age Research – Fritz Lipmann Institute (FLI), Beutenbergstrasse 11, 07745 Jena, Germany
a r t i c l e
i n f o
Article history: Received 17 March 2010 Available online 23 April 2010 Keywords: KIF5A Eg5 Motility Bypassing Non-exchangeable blockages
a b s t r a c t During movement along microtubules, kinesin usually follows a track parallel to the axis of a single protofilament. The question arises what happens when kinesin encounters blockages. The present study describes the movement of kinesin labeled by 20-nm gold beads along immobilized microtubules artificially decorated with blocking proteins. To guarantee that exactly the kinesin-binding sites were occupied and to avoid steric effects exerted by large molecules, the KIF5A motor domain was used for blocking. After binding, the blockages were irreversibly cross-linked to the microtubules to make them non-exchangeable. Under such conditions, kinesin movement became a non-continuous one. As a rule, after temporary stopping the kinesin moved on without being released from the microtubule. The results strongly suggest a bypassing mechanism based on the postulation that kinesin changes to and continues movement along a neighbouring protofilament. Bypassing is considered to ensure an efficient long-distance transport of cellular cargoes by kinesins. Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction Kinesins belong to the large family of microtubule-associated motor proteins [1], which convert the chemical energy of adenosine triphosphate (ATP) into mechanical energy, used to maintain the various events of intracellular motility and cell movement. Among the members of the kinesin protein family, conventional kinesin (kinesin-1) is the best characterized one. Kinesin-1 molecules represent elongated homodimers with two distinct globular heads [2], each of them containing an ATP-hydrolyzing center and a microtubule-binding motif. Dependent on the ATP hydrolysis cycle, kinesin-1 moves along the microtubule surface towards the microtubule plus end by alternating binding of its heads from one tubulin dimer to the next free one. At any time, one of both heads is bound to the microtubule, ensuring processive movement over hundreds of steps [3]. The high level of processivity makes it superior to other cytoskeleton-associated motor proteins to transport cellular cargoes over long distances, which is required to fulfill physiological tasks of specialized cell types like the neurons with their extended axons. It is commonly accepted that kinesin follows a path parallel to the axis of the protofilaments [4,5]. However, in the viscous dense intracellular matrix, the kinesin-mediated transport is exposed to a great number of associated cell constituents, which might interact with the motor itself or with the microtubule rail. There seems to be a high probability that a kinesin molecule moving
* Corresponding author. Fax: +49 3641 656410. E-mail address: kboehm@fli-leibniz.de (K.J. Böhm). 0006-291X/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2010.04.035
along its protofilament track encounters other proteins, which hinder its binding to the microtubule surface and consequently inhibit its movement [6,7]. As a result, the kinesin-mediated cargo transport might come to a stop and provoke a situation comparable with a traffic jam [8,9]. Moreover, irregularities in microtubule structure, e.g., protofilament number variations [10], incorrect dimer positioning or dimer failure, partially denatured tubulin dimers with alterations in their kinesin recognition site, or any other discontinuities in the microtubule lattice might additionally limit the kinesin-mediated transport. In this context, the question arises what happens when a kinesin molecule encounters disturbed sites of kinesin binding during movement along its protofilament track. Concerning this subject, Seitz and Surrey [11] observed that moving kinesin, competing with a motility-deficient kinesin construct for binding sites on the microtubule surface, stopped at the site where this construct was located and that movement could be continued only after freeing the blocked site. In a recent paper, Telley et al. [12] reported that on the one hand kinesin has a low probability to wait and usually detaches from its protofilament track when encountering obstacles. On the other hand, the authors discussed that kinesins occasionally pass obstacles by changing protofilaments. Korten and Diez [13] decorated the surface of biotinylated microtubules with streptavidin and found that kinesin frequently stopped upon encounters with streptavidin molecules. They considered that passing would be possible only in cases when the streptavidin does not block the kinesin-binding site. Alternatively, they suggested that kinesin temporarily detaches, diffuses along the microtubule lattice, and rebinds.
K. Dreblow et al. / Biochemical and Biophysical Research Communications 395 (2010) 490–495
491
Summarizing available relevant reports dealing with the influence of blockages onto kinesin movement, a rather conflicting situation becomes evident (see also [14]). So far, the question what happens when exactly kinesin-binding sites along the protofilament track were irreversibly blocked has been unanswered. We expect the existence of more effective mechanisms of passing blocked kinesin-binding sites, which are not based on the principle of contingency like in cases of waiting for the release of the blockage or of temporary detachment of the transporter and its rebinding to the microtubule. To contribute to elucidation of the mechanism of overcoming blockages, we recently presented an experimental approach in which native kinesin labeled by 20-nm gold beads was allowed to move along microtubules whose surface was decorated by irreversibly bound, inactivated full-length kinesin or a C-terminally truncated kinesin construct (still containing about one-half of the stalk), acting as blocking protein [15]. This experimental approach ensured that kinesin-binding sites were specifically occupied, but had the disadvantage that blocking was superimposed by steric hindrance exerted by the about 80-nm long kinesin stalk and tail. To exclude steric effects, we present here a refined procedure which is based on the usage of the relatively small (8–9 nm in diameter [2]) kinesin motor domain as blocking protein, instead of the full-length kinesin or the truncated kinesin construct. Under such conditions, kinesin movement changed from a commonly continuous fashion to a conspicuously non-continuous one. But, after stopping the kinesin continued to move without being detached from the microtubule surface. Similar results were obtained when the slowly moving Eg5 motor was bound to the microtubules and kept in native state to permit its exchange by the transporter kinesin. To explain the non-continuous movement, we provide a theoretical bypassing concept based on the assumption of sideways stepping, described by Yildiz et al. [16]. 2. Materials and methods 2.1. Proteins The tubulin was purified from porcine brain homogenates by two cycles of temperature-dependent disassembly/reassembly [17] followed by phosphocellulose column chromatography [18]. Microtubules were formed by taxol-promoted self-assembly at 37 °C from tubulin (2 mg/ml) in assembly buffer (20 mM Pipes, 80 mM NaCl, 0.5 mM MgCl2, 1 mM EGTA, 0.2 mM GTP, pH 6.8). The microtubules prepared were free of associated proteins and mostly consisted of 12 protofilaments [19]. The full-length human neuron-specific kinesin KIF5A (KIF5Afl) [20] was expressed in Escherichia coli as recombinant protein and purified using the protocol recently described [21]. The pure kinesin, lacking any artificial sequences, was adjusted in motility buffer (50 mM imidazole, 0.5 mM MgCl2, 0.5 mM EGTA, 0.5 mM DTT, pH 6.8) and stored at 80 °C. Analogously, the KIF5A motor domain including amino acids 1–330 (KIF5A330) and the dimeric human Eg5 [22] including amino acids 1–513 were expressed and purified. The monomeric KIF5A330 had a microtubule-dependent ATPase with a kcat of 25 s 1. The Eg5513 had a kcat of 1 s 1 and moved gold beads at 40 nm/s (for detailed characterization see [21]). 2.2. Blocking of kinesin-binding sites along the microtubules To ensure that the blocking protein competes with the transporter protein for the same binding site on the microtubule, KIF5A was used for both blocking and transporting. Firstly, the KIF5A330 was irreversibly attached to the microtubules as follows: Taxolstabilized microtubules (20 lM tubulin) were mixed with the KIF5A330 in the presence of the non-hydrolyzable ATP analogue
Fig. 1. Discontinuous movement along microtubules with irreversibly bound KIF5A330. (A) Distance–time diagrams of individual transporter molecules. (a) Continuous movement in the absence of blockages. Within the time interval of this sequence, the mean velocity was 1056 nm/s. (b) Alternating phases of movement and resting observed in the presence of blockages. The mean velocity including the resting phases was 365 nm/s. The KIF5A330-to-tubulin ratio was 0.2. The data were obtained from digitized video sequences (30 frames/s) by measuring the X–Y shift of single transporter molecules. (B) AVEC-DIC image sequence of single transporter molecules moving along microtubules. (a) Continuous transport in the absence of blocking protein. (b) Blockage-induced discontinuous transport at a KIF5A330-totubulin ratio of 0.2. The moving objects were marked by arrows. The time scale indicates seconds and tenth of a second. Within the time intervals shown in (a) and (b), the kinesin moved over 4.4 and 0.9 lm, respectively.
492
K. Dreblow et al. / Biochemical and Biophysical Research Communications 395 (2010) 490–495
AMP-PNP (2 mM) or tripolyphosphate (pentasodium salt, Sigma– Aldrich, Germany) and transferred into a flow chamber formed between a glass slide (76 mm 26 mm) and a coverslip (18 mm 18 mm), pretreated with poly (diallyl dimethyl ammonium chloride) (PDDA, Sigma–Aldrich) and casein (5 mg/ml). The density of blocking molecules on the microtubule surface was varied by changing the KIF5A330-to-tubulin ratio at constant tubulin concentration. After binding to the glass surface, the KIF5A330–microtubule complexes were chemically fixed for 15 min by fresh 0.1% glutaraldehyde (grade I, Sigma–Aldrich) in motility buffer. The blocked microtubules were washed twice with phosphate-buffered saline (PBS). Residual active aldehyde groups were neutralized by 15min treatment with 0.1 mM glycine in PBS [23], followed by double washing with motility buffer. As a result, we obtained microtubules with a reproducible density of blocking molecules, which occupy kinesin-binding sites on the microtubule surface. For control experiments, microtubules without blocking proteins were bound to pretreated glass slides and chemically fixed. Chemical fixation of microtubules by low-concentrated glutaraldehyde was proved not to impair kinesin-mediated motility [23–25]. Analogously, Eg5513 was used as blocking protein. In this case, the microtubules were either treated by glutaraldehyde to get
non-exchangeable blockages like done with KIF5A330 or kept in native state to study the motility behaviour of the KIF5A transporter in the presence of exchangeable Eg5513. 2.3. Kinesin labeling by gold beads As transporter protein, KIF5Afl was used generally. KIF5Afl (90 ng in motility buffer) was 100-fold diluted in distilled water and allowed to bind to 9.1 109 beads (20-nm gold colloid, Sigma–Aldrich). After adding 10-fold concentrated motility buffer to restore normal motility buffer conditions, 10 ll of this bead suspension were mixed with casein (4.5 mg/ml final concentration; Sigma–Aldrich) and MgATP (5 mM final concentration; Sigma–Aldrich) and transferred onto the blocked microtubules immobilized onto the glass slides. 2.4. Microscopy and data analysis The movement of the gold bead-labeled kinesin along the microtubules was visualized by video-enhanced differential interference contrast microscopy (AVEC-DIC) using an Axiophot microscope (Zeiss, Germany) with an oil immersion objective (100/ 1.30), equipped with a Chalnicon video camera (Hamamatsu Pho-
Fig. 2. Effect of irreversibly bound KIF5A330 on motility parameters. (A, B) Velocity frequency histograms in the absence and in the presence of KIF5A330 blockages, respectively. The x-values indicate velocity intervals ±50 nm/s. n is the numbers of moving objects measured. (C) Dependence of the mean velocity on the KIF5A330-to-tubulin ratio. (D) Mean distance the transporter kinesin moved from the moment of association with a microtubule upon its final stopping or detachment as function of the KIF5A330to-tubulin ratio.
K. Dreblow et al. / Biochemical and Biophysical Research Communications 395 (2010) 490–495
493
tonics GmbH, Germany) and the image processing system Argus 20 (Hamamatsu). Video sequences, documenting movement on a digital hard-disk recorder, were digitized resulting in image sequences of 30 frames/s. The sequences were analyzed using a special Plug-In for the ImageJ software [26]. The mean velocity and distance of movement were calculated from trajectories gold bead-labeled kinesin molecules.
3. Results 3.1. Irreversible blocking of kinesin-binding sites along the microtubules To ensure a reproducible control of single motor behaviour, actuation, and functionalization [27], we performed a motility assay, which was based on the ability of motor proteins to move cargoes along proteinaceous filaments immobilized to a substrate. As transporter full-length neuron-specific kinesin KIF5A (KIF5Afl) was used throughout the present study. To study kinesin-dependent transport mechanisms under conditions of blocked kinesin-binding sites, the assay was modified as follows: before inducing the movement of the transporter kinesin, kinesin-binding sites along the microtubules were specifically occupied by molecules of the monomeric KIF5A motor domain (KIF5A330) in the presence of a non-hydrolyzable ATP analogue. The blocking KIF5A was irreversibly cross-linked to the microtubules by glutaraldehyde treatment to prepare stable microtubule rails. After neutralization of residual free aldehyde groups, the native transporter KIF5Afl, labeled by 20-nm gold beads, was added to the microtubules and allowed to move after ATP supply. This procedure guarantees that the blocking protein competes directly with the transporter for binding sites. As a result of glutaraldehyde fixation, KIF5A330 became irreversibly bound to kinesin-binding sites and acted as non-exchangeable blockage. Control experiments with microtubules lacking protein blockages proved that under the conditions of glutaraldehyde fixation the binding of the transporter kinesin and its movement along the microtubules was not affected (see [15]). 3.2. Discontinuous kinesin movement in the presence of blockages In control preparations without blocking protein, the kinesin moved continuously at a mean velocity of 783 ± 194 nm/s (arithmetic mean ± SD). In the presence of KIF5A330 blockages, phases of movement were observed to alternate with resting phases (Fig. 1). Accordingly, the mean velocity decreased (Fig. 4A–C). Between stop events, the velocity was found to be nearly the same as in control experiments. The increase of the KIF5A330-to-tubulin ratio to values above 0.1 did not cause stronger effects on the velocity of kinesin movement (Fig. 2C). But with increasing this ratio, we observed a lowered frequency of binding of transporter kinesin to the microtubules, obviously due to the reduced number of free binding sites that could be accessed. However, once a kinesin molecule was bound and started to move, it was usually (in more than 80% of cases observed) able to overcome blockages after pausing without being released from its microtubule rail. In the other cases, the transporter either detached from the microtubule or finally stalled. Unlike the mean velocity characteristics, the mean distances the transporter moved from the moment of binding to the microtubule upon its final stopping or detachment were widely unchanged (Fig. 2D).
Fig. 3. Kinesin movement along microtubules in the presence of the slow Eg5 motor. (A) Distance–time diagrams documenting alternating phases of movement and resting observed in the presence of exchangeable and non-exchangeable Eg5513 (mean velocities including resting phases 354 and 223 nm/s, respectively). Eg5-totubulin ratio 0.2. Further details see Fig. 1A. (B) Velocity frequency histogram in the presence of exchangeable Eg5513. The x-values indicate velocity intervals ±50 nm/s. n is the numbers of moving objects measured. Remarkably, the velocities significantly exceeded those typical for Eg5 (40 nm/s).
3.3. Kinesin movement in the presence of slowly moving Eg5 In addition to KIF5A, slowly moving dimeric human Eg5513 motor protein was used to block kinesin-binding sites along the microtubule surface. Control experiments in which this motor was labeled by 20-nm gold demonstrated that Eg5513 moved along immobilized microtubules at a mean velocity of 40 ± 20 nm/s. When microtubules were irreversibly blocked by Eg5, the transporter KIF5Afl moved in discontinuous fashion, which resulted in lowering the mean velocity (Fig. 3A). As expected, like in the case of the irreversibly bound, non-exchangeable KIF5A330 (see Fig. 1A) phases of movement alternated with resting ones, causing a decrease of the mean velocity. Surprisingly, principally the same characteristic was observed when the Eg5 was not chemically cross-linked to the microtubules to keep it in native, motility-competent state. However, the resting phases were usually shorter and the mean velocity was tendentiously higher than in the case of non-exchangeable Eg5 (Fig. 3A). It has to be emphasized that in the presence of the native, exchangeable Eg5, velocities in the range of 40 nm/s, which are
494
K. Dreblow et al. / Biochemical and Biophysical Research Communications 395 (2010) 490–495
4. Discussion
Fig. 4. Concept of kinesin bypassing: variations of overcoming blockages. (A) The transporter kinesin is a dimer with two heads (green) moving along a microtubule and encountering a blockage (red). Considering only forward steps, to overcome the blockage the trailing head can theoretically access four b-tubulin binding sites (orange dots) both at the right and at the left related to moving direction. (B) Bypassing using the adjacent protofilament. (a) The trailing head binds to the btubulin of the adjacent protofilament and becomes located side-by-side with the leading head. In the 2nd step, the former leading head is pushed forward and binds to the next free b-tubulin of the adjacent protofilament and becomes located sideby-side with the blockage. (b) The trailing head binds side-by-side with the blockage. In the 2nd step, the former leading head is pushed forward and binds to the next free b-tubulin along the new protofilament track. (C) Bypassing using the next but one protofilament. (a) The trailing head turns sideward and binds approximately perpendicular to the b-tubulin of the next but one protofilament. In the 2nd step, the former leading head is pushed forward and binds also to this protofilament. (b) The trailing head is pushed forward and binds to the b-tubulin of the next but one protofilament. In the 2nd step, the former leading head is pushed forward and binds to the next free b-tubulin along the new protofilament track.
typically measured for Eg5 in control experiments, were not found (see Fig. 3B).
The present study aims at elucidation of mechanisms enabling kinesin to pass non-exchangeable blockages along its protofilament track. With this intention, an experimental approach has been worked out in which kinesin-binding sites along the microtubule surface were specifically occupied by KIF5A heads. To avoid their release, the heads were chemically cross-linked to the microtubule surface to get stable, non-exchangeable blockages. Thereafter, native gold-labeled full-length KIF5A as transporter was allowed to bind and to move along the blocked microtubules. Unlike observed for control preparations lacking blockages, movement was observed to be a non-continuous one, whereby phases of active transport alternated with resting ones. As a result, the mean velocity was lowered. Regardless of the presence of blocking molecules, the transporter kinesin was found to move processively over relatively long distances which are in the same range like in control preparations without blockages. As in our motility experiments the blocking protein was irreversibly attached to the microtubule rail by cross-linking with a chemical fixative, the continuation of movement after stopping at a blocked site can be explained by postulating a bypassing mechanism (Fig. 4). Kinesin moves in an asymmetric hand-overhand like fashion [28,29], whereby the leading head remains bound until the trailing head attaches to the next tubulin dimer. In the case of an irreversible bound blockage, the trailing head is not able to bind to the next free tubulin dimer along its protofilament track and the leading head cannot detach what causes stalling of the cargo transporter. Recently, Yildiz et al. [16] reported that, even in the absence of a lateral force, there is a small portion of sideways steps which kinesin performs during stepping. Taken the molecular flexibility of kinesin into consideration, we hypothesize that the leading head remains bound to the original protofilament whereas the trailing one turns sideward and contacts a tubulin binding site on the adjacent or on the next but one protofilament (for different modes of binding see Fig. 4). We assume that both free protofilaments on the right and on the left, corresponding to movement direction, can be used. Finally, the leading head leaves the blocked protofilament and follows the trailing one along its new track. Our model of bypassing obstacles by making sidewards steps of various sizes is consistent with the theoretical torsional spring model of Bolterauer et al. [30]. Bypassing by taking sideways steps in the presence of blocked kinesin-binding sites is regarded to be a complex mechanism, including ‘‘recognition” of the blocked site and diffusional search of the trailing head until finding an accessible free binding site on a neighbouring protofilament, which finally cause temporary stopping events. Table 1 Theoretical step widths during transition of the trailing head to a free b-tubulin of a neighbouring protofilament. Calculations were performed using the following data: 5 nm lateral spacing between the protofilaments; 1 nm axial shift between adjacent protofilaments; 8 nm tubulin dimer length. Mode
Bypassing in moving direction
Transition to the adjacent protofilament (see Fig. 4B) a Right a Left b Right b Left Transition to the next but one protofilament (see Fig. 4C) a Right a Left b Right b Left
Step width (nm) 8.6 10.2 15.8 17.7 11.7 14.1 17.2 20.6
K. Dreblow et al. / Biochemical and Biophysical Research Communications 395 (2010) 490–495
For the single modes of bypassing described (Fig. 4B and C) theoretical step sizes between 8.6 and 20.6 nm were calculated for the first head (Table 1). These values seem to be realistic as single kinesin heads have been reported to take steps of 17.3 ± 3.3 nm [29]. In our model, we consider only forward steps, as in this case the ATPase cycle and the resulting coordinated alternating movement of both heads proceeds in regular fashion. In the case of transition to the adjacent protofilament (Fig. 4B), steps should occur during which either both heads of the kinesin transporter or one head of the transporter and a head of the blocking KIF5A are located side-by-side at about 5 nm distance. Chrétien et al. [10] demonstrated that under artificial conditions, suppressing ATPase activity, kinesin heads are able to bind in an equimolar ratio to microtubules. However, we consider that under conditions of active movement, the heads of the dimeric kinesin require a higher degree of freedom. Consequently, the probability that its free head, looking for an alternative binding site by diffusional search, binds directly side-by-side to another head seems to be rather low. Therefore, we favor the transition of the free head of the moving kinesin to the next but one protofilament (Fig. 4C) in cases of blocking the kinesin-binding site by an obstacle with dimensions in the range of a kinesin head. Transitions to the adjacent protofilament (Fig. 4B) might occur in cases of disturbed kinesin binding to the next dimer along the original protofilament track, which might be caused by e.g., structural defects of the microtubule lattice. Concerning the question whether kinesin encountering a blocked binding site along its protofilament track waits until its track becomes free, we performed additional experiments in which the transporter kinesin competed with native, exchangeable Eg5, known to move at velocities which are lower by a factor of about 25 compared to kinesin-1 [21]. We found that not only in the case of irreversibly bound Eg5 blockages, but also in the presence of native, exchangeable Eg5 the transporter kinesin moved in non-continuous fashion. If the fast kinesin would not be able to overtake the slowly moving motor it should move behind it at velocities of 40 nm/s. As such velocities were not measured (see Fig. 3B), we suggest that the transporter kinesin does not solely wait for the release of Eg5, but is assumed to bypass also the slow motor. 5. Conclusions Bypassing blockages can be regarded as a fundamental mechanism by which the track fidelity of a molecular motor can be improved. Elucidation of this mechanism should contribute to deeper understanding of the molecular mechanisms of cytoplasmic transport processes over long distances, required for e.g., motor neurons. Acknowledgments This study was supported by the European Commission within the Sixth EU Framework Programme, Contract No. NMP4-CT-2004516989. We thank Dr. Ronald D. Vale (University of California, San Francisco) for supplying the KIF5A cDNA, Janina Beeg and Ruben Serral Graciá (MPI of Colloids and Interfaces, Golm) for kindly providing the ImageJ Plug-In, and Marina Wollmann for her skilful technical assistance.
495
References [1] N. Hirokawa, Y. Noda, Y. Tanaka, S. Niwa, Kinesin superfamily motor proteins and intracellular transport, Nat. Rev. Mol. Cell Biol. 10 (2009) 682–696. [2] N. Hirokawa, K.K. Pfister, H. Yorifuji, M.C. Wagner, S.T. Brady, G.S. Bloom, Submolecular domains of bovine brain kinesin identified by electron microscopy and monoclonal antibody decoration, Cell 56 (1989) 867–878. [3] W.O. Hancock, J. Howard, Processivity of the motor protein kinesin requires two heads, J. Cell Biol. 140 (1998) 1395–1405. [4] S. Ray, E. Meyhöfer, R.A. Milligan, J. Howard, Kinesin follows the microtubule’s protofilament axis, J. Cell Biol. 121 (1993) 1083–1093. [5] S. Ray, S.G. Wolf, J. Howard, K.H. Downing, Kinesin does not support the motility of zinc-macrotubes, Cell Motil. Cytoskel. 30 (1995) 146–152. [6] H. Hagiwara, H. Yorifuji, R. Sato-Yoshitake, N. Hirokawa, Competition between motor molecules (kinesin and cytoplasmic dynein) and fibrous microtubuleassociated proteins in binding to microtubules, J. Biol. Chem. 269 (1994) 3581– 3589. [7] R. Dixit, J.L. Ross, Y.E. Goldman, E.L. Holzbaur, Differential regulation of dynein and kinesin motor proteins by tau, Science 319 (2008) 1086–1089. [8] M. Martin, S.J. Iyadurai, A. Gassman, J.G. Gindhart, T.S. Hays, W.M. Saxton, Cytoplasmic dynein, the dynactin complex, and kinesin are interdependent and essential for fast axonal transport, Mol. Biol. Cell 10 (1999) 3717–3728. [9] A. Seitz, H. Kojima, K. Oiwa, E.M. Mandelkow, Y.H. Song, E. Mandelkow, Singlemolecule investigation of the interference between kinesin, tau and MAP2c, EMBO J. 21 (2002) 4896–4905. [10] D. Chrétien, F. Metoz, F. Verde, E. Karsenti, R.H. Wade, Lattice defects in microtubules: protofilament numbers vary within individual microtubules, J. Cell Biol. 117 (1992) 1031–1040. [11] A. Seitz, T. Surrey, Processive movement of single kinesins on crowded microtubules visualized using quantum dots, EMBO J. 25 (2006) 267–277. [12] I.A. Telley, P. Bieling, T. Surrey, Obstacles on the microtubule reduce the processivity of kinesin-1 in a minimal in vitro system and in cell extract, Biophys. J. 96 (2009) 3341–3353. [13] T. Korten, S. Diez, Setting up roadblocks for kinesin-1: mechanism for the selective speed control of cargo carrying microtubules, Lab Chip 8 (2008) 1441–1447. [14] Y. Chai, R. Lipowsky, S. Klumpp, Transport by molecular motors in the presence of static defects, J. Stat. Phys. 135 (2009) 241–260. [15] K. Dreblow, N. Kalchishkova, K.J. Böhm, Kinesin bypassing blockages on microtubule rails, Biophys. Rev. Lett. 4 (2009) 139–151. [16] A. Yildiz, M. Tomishige, A. Gennerich, R.D. Vale, Intramolecular strain coordinates kinesin stepping behavior along microtubules, Cell 134 (2008) 1030–1041. [17] M.L. Shelanski, F. Gaskin, C.R. Cantor, Microtubule assembly in the absence of added nucleotides, Proc. Natl. Acad. Sci. USA 70 (1973) 765–768. [18] M.D. Weingarten, A.H. Lockwood, S.Y. Hwo, M.W. Kirschner, A protein factor essential for microtubule assembly, Proc. Natl. Acad. Sci. USA 72 (1975) 1858– 1862. [19] K.J. Böhm, W. Vater, H. Fenske, E. Unger, Effect of microtubule-associated proteins on the protofilament number of microtubules assembled in vitro, Biochim. Biophys. Acta 800 (1984) 119–126. [20] J. Niclas, F. Navone, N. Hom-Booher, R.D. Vale, Cloning and localization of a conventional kinesin motor expressed exclusively in neurons, Neuron 12 (1994) 1059–1072. [21] N. Kalchishkova, K.J. Böhm, The role of kinesin neck linker and neck in velocity regulation, J. Mol. Biol. 382 (2008) 127–135. [22] A. Blangy, H.A. Lane, P. d’Hérin, M. Harper, M. Kress, E.A. Nigg, Phosphorylation by p34cdc2 regulates spindle association of human Eg5, a kinesin-related motor essential for bipolar spindle formation in vivo, Cell 83 (1995) 1159–1169. [23] K.J. Böhm, R. Stracke, P. Mühlig, E. Unger, Motorprotein-driven unidirectional transport of micrometer-sized cargoes across isopolar microtubule arrays, Nanotechnology 12 (2001) 238–244. [24] D. Turner, C. Chang, K. Fang, P. Cuomo, D. Murphy, Kinesin movement on glutaraldehyde-fixed microtubules, Anal. Biochem. 242 (1996) 20–25. [25] A.K. Boal, H. Tellez, S.B. Rivera, N.E. Miller, G.D. Bachand, B.C. Bunker, The stability and functionality of chemically crosslinked microtubules, Small 2 (2006) 793–803. [26] J. Beeg, S. Klumpp, R. Dimova, R.S. Gracià, E. Unger, R. Lipowsky, Transport of beads by several kinesin motors, Biophys. J. 94 (2008) 532–541. [27] M.G.L. van den Heuvel, C. Dekker, Motor proteins at work for nanotechnology, Science 317 (2007) 333–336. [28] C.L. Asbury, A.N. Fehr, S.M. Block, Kinesin moves by an asymmetric hand-overhand mechanism, Science 302 (2003) 2130–2134. [29] A. Yildiz, M. Tomishige, R.D. Vale, P.R. Selvin, Kinesin walks hand-over-hand, Science 303 (2004) 676–678. [30] H. Bolterauer, J.A. Tuszynski, E. Unger, Directed binding: a novel physical mechanism that describes the directional motion of two-headed kinesin motor proteins, Cell Biochem. Biophys. 42 (1984) 95–119.
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Kinesin passing permanent blockages along its protofilament track Kerstin Dreblow, Nikolina Kalchishkova, Konrad J. Böhm * Leibniz Institute for Age Research – Fritz Lipmann Institute (FLI), Beutenbergstrasse 11, 07745 Jena, Germany
a r t i c l e
i n f o
Article history: Received 17 March 2010 Available online 23 April 2010 Keywords: KIF5A Eg5 Motility Bypassing Non-exchangeable blockages
a b s t r a c t During movement along microtubules, kinesin usually follows a track parallel to the axis of a single protofilament. The question arises what happens when kinesin encounters blockages. The present study describes the movement of kinesin labeled by 20-nm gold beads along immobilized microtubules artificially decorated with blocking proteins. To guarantee that exactly the kinesin-binding sites were occupied and to avoid steric effects exerted by large molecules, the KIF5A motor domain was used for blocking. After binding, the blockages were irreversibly cross-linked to the microtubules to make them non-exchangeable. Under such conditions, kinesin movement became a non-continuous one. As a rule, after temporary stopping the kinesin moved on without being released from the microtubule. The results strongly suggest a bypassing mechanism based on the postulation that kinesin changes to and continues movement along a neighbouring protofilament. Bypassing is considered to ensure an efficient long-distance transport of cellular cargoes by kinesins. Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction Kinesins belong to the large family of microtubule-associated motor proteins [1], which convert the chemical energy of adenosine triphosphate (ATP) into mechanical energy, used to maintain the various events of intracellular motility and cell movement. Among the members of the kinesin protein family, conventional kinesin (kinesin-1) is the best characterized one. Kinesin-1 molecules represent elongated homodimers with two distinct globular heads [2], each of them containing an ATP-hydrolyzing center and a microtubule-binding motif. Dependent on the ATP hydrolysis cycle, kinesin-1 moves along the microtubule surface towards the microtubule plus end by alternating binding of its heads from one tubulin dimer to the next free one. At any time, one of both heads is bound to the microtubule, ensuring processive movement over hundreds of steps [3]. The high level of processivity makes it superior to other cytoskeleton-associated motor proteins to transport cellular cargoes over long distances, which is required to fulfill physiological tasks of specialized cell types like the neurons with their extended axons. It is commonly accepted that kinesin follows a path parallel to the axis of the protofilaments [4,5]. However, in the viscous dense intracellular matrix, the kinesin-mediated transport is exposed to a great number of associated cell constituents, which might interact with the motor itself or with the microtubule rail. There seems to be a high probability that a kinesin molecule moving
* Corresponding author. Fax: +49 3641 656410. E-mail address: kboehm@fli-leibniz.de (K.J. Böhm). 0006-291X/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2010.04.035
along its protofilament track encounters other proteins, which hinder its binding to the microtubule surface and consequently inhibit its movement [6,7]. As a result, the kinesin-mediated cargo transport might come to a stop and provoke a situation comparable with a traffic jam [8,9]. Moreover, irregularities in microtubule structure, e.g., protofilament number variations [10], incorrect dimer positioning or dimer failure, partially denatured tubulin dimers with alterations in their kinesin recognition site, or any other discontinuities in the microtubule lattice might additionally limit the kinesin-mediated transport. In this context, the question arises what happens when a kinesin molecule encounters disturbed sites of kinesin binding during movement along its protofilament track. Concerning this subject, Seitz and Surrey [11] observed that moving kinesin, competing with a motility-deficient kinesin construct for binding sites on the microtubule surface, stopped at the site where this construct was located and that movement could be continued only after freeing the blocked site. In a recent paper, Telley et al. [12] reported that on the one hand kinesin has a low probability to wait and usually detaches from its protofilament track when encountering obstacles. On the other hand, the authors discussed that kinesins occasionally pass obstacles by changing protofilaments. Korten and Diez [13] decorated the surface of biotinylated microtubules with streptavidin and found that kinesin frequently stopped upon encounters with streptavidin molecules. They considered that passing would be possible only in cases when the streptavidin does not block the kinesin-binding site. Alternatively, they suggested that kinesin temporarily detaches, diffuses along the microtubule lattice, and rebinds.
K. Dreblow et al. / Biochemical and Biophysical Research Communications 395 (2010) 490–495
491
Summarizing available relevant reports dealing with the influence of blockages onto kinesin movement, a rather conflicting situation becomes evident (see also [14]). So far, the question what happens when exactly kinesin-binding sites along the protofilament track were irreversibly blocked has been unanswered. We expect the existence of more effective mechanisms of passing blocked kinesin-binding sites, which are not based on the principle of contingency like in cases of waiting for the release of the blockage or of temporary detachment of the transporter and its rebinding to the microtubule. To contribute to elucidation of the mechanism of overcoming blockages, we recently presented an experimental approach in which native kinesin labeled by 20-nm gold beads was allowed to move along microtubules whose surface was decorated by irreversibly bound, inactivated full-length kinesin or a C-terminally truncated kinesin construct (still containing about one-half of the stalk), acting as blocking protein [15]. This experimental approach ensured that kinesin-binding sites were specifically occupied, but had the disadvantage that blocking was superimposed by steric hindrance exerted by the about 80-nm long kinesin stalk and tail. To exclude steric effects, we present here a refined procedure which is based on the usage of the relatively small (8–9 nm in diameter [2]) kinesin motor domain as blocking protein, instead of the full-length kinesin or the truncated kinesin construct. Under such conditions, kinesin movement changed from a commonly continuous fashion to a conspicuously non-continuous one. But, after stopping the kinesin continued to move without being detached from the microtubule surface. Similar results were obtained when the slowly moving Eg5 motor was bound to the microtubules and kept in native state to permit its exchange by the transporter kinesin. To explain the non-continuous movement, we provide a theoretical bypassing concept based on the assumption of sideways stepping, described by Yildiz et al. [16]. 2. Materials and methods 2.1. Proteins The tubulin was purified from porcine brain homogenates by two cycles of temperature-dependent disassembly/reassembly [17] followed by phosphocellulose column chromatography [18]. Microtubules were formed by taxol-promoted self-assembly at 37 °C from tubulin (2 mg/ml) in assembly buffer (20 mM Pipes, 80 mM NaCl, 0.5 mM MgCl2, 1 mM EGTA, 0.2 mM GTP, pH 6.8). The microtubules prepared were free of associated proteins and mostly consisted of 12 protofilaments [19]. The full-length human neuron-specific kinesin KIF5A (KIF5Afl) [20] was expressed in Escherichia coli as recombinant protein and purified using the protocol recently described [21]. The pure kinesin, lacking any artificial sequences, was adjusted in motility buffer (50 mM imidazole, 0.5 mM MgCl2, 0.5 mM EGTA, 0.5 mM DTT, pH 6.8) and stored at 80 °C. Analogously, the KIF5A motor domain including amino acids 1–330 (KIF5A330) and the dimeric human Eg5 [22] including amino acids 1–513 were expressed and purified. The monomeric KIF5A330 had a microtubule-dependent ATPase with a kcat of 25 s 1. The Eg5513 had a kcat of 1 s 1 and moved gold beads at 40 nm/s (for detailed characterization see [21]). 2.2. Blocking of kinesin-binding sites along the microtubules To ensure that the blocking protein competes with the transporter protein for the same binding site on the microtubule, KIF5A was used for both blocking and transporting. Firstly, the KIF5A330 was irreversibly attached to the microtubules as follows: Taxolstabilized microtubules (20 lM tubulin) were mixed with the KIF5A330 in the presence of the non-hydrolyzable ATP analogue
Fig. 1. Discontinuous movement along microtubules with irreversibly bound KIF5A330. (A) Distance–time diagrams of individual transporter molecules. (a) Continuous movement in the absence of blockages. Within the time interval of this sequence, the mean velocity was 1056 nm/s. (b) Alternating phases of movement and resting observed in the presence of blockages. The mean velocity including the resting phases was 365 nm/s. The KIF5A330-to-tubulin ratio was 0.2. The data were obtained from digitized video sequences (30 frames/s) by measuring the X–Y shift of single transporter molecules. (B) AVEC-DIC image sequence of single transporter molecules moving along microtubules. (a) Continuous transport in the absence of blocking protein. (b) Blockage-induced discontinuous transport at a KIF5A330-totubulin ratio of 0.2. The moving objects were marked by arrows. The time scale indicates seconds and tenth of a second. Within the time intervals shown in (a) and (b), the kinesin moved over 4.4 and 0.9 lm, respectively.
492
K. Dreblow et al. / Biochemical and Biophysical Research Communications 395 (2010) 490–495
AMP-PNP (2 mM) or tripolyphosphate (pentasodium salt, Sigma– Aldrich, Germany) and transferred into a flow chamber formed between a glass slide (76 mm 26 mm) and a coverslip (18 mm 18 mm), pretreated with poly (diallyl dimethyl ammonium chloride) (PDDA, Sigma–Aldrich) and casein (5 mg/ml). The density of blocking molecules on the microtubule surface was varied by changing the KIF5A330-to-tubulin ratio at constant tubulin concentration. After binding to the glass surface, the KIF5A330–microtubule complexes were chemically fixed for 15 min by fresh 0.1% glutaraldehyde (grade I, Sigma–Aldrich) in motility buffer. The blocked microtubules were washed twice with phosphate-buffered saline (PBS). Residual active aldehyde groups were neutralized by 15min treatment with 0.1 mM glycine in PBS [23], followed by double washing with motility buffer. As a result, we obtained microtubules with a reproducible density of blocking molecules, which occupy kinesin-binding sites on the microtubule surface. For control experiments, microtubules without blocking proteins were bound to pretreated glass slides and chemically fixed. Chemical fixation of microtubules by low-concentrated glutaraldehyde was proved not to impair kinesin-mediated motility [23–25]. Analogously, Eg5513 was used as blocking protein. In this case, the microtubules were either treated by glutaraldehyde to get
non-exchangeable blockages like done with KIF5A330 or kept in native state to study the motility behaviour of the KIF5A transporter in the presence of exchangeable Eg5513. 2.3. Kinesin labeling by gold beads As transporter protein, KIF5Afl was used generally. KIF5Afl (90 ng in motility buffer) was 100-fold diluted in distilled water and allowed to bind to 9.1 109 beads (20-nm gold colloid, Sigma–Aldrich). After adding 10-fold concentrated motility buffer to restore normal motility buffer conditions, 10 ll of this bead suspension were mixed with casein (4.5 mg/ml final concentration; Sigma–Aldrich) and MgATP (5 mM final concentration; Sigma–Aldrich) and transferred onto the blocked microtubules immobilized onto the glass slides. 2.4. Microscopy and data analysis The movement of the gold bead-labeled kinesin along the microtubules was visualized by video-enhanced differential interference contrast microscopy (AVEC-DIC) using an Axiophot microscope (Zeiss, Germany) with an oil immersion objective (100/ 1.30), equipped with a Chalnicon video camera (Hamamatsu Pho-
Fig. 2. Effect of irreversibly bound KIF5A330 on motility parameters. (A, B) Velocity frequency histograms in the absence and in the presence of KIF5A330 blockages, respectively. The x-values indicate velocity intervals ±50 nm/s. n is the numbers of moving objects measured. (C) Dependence of the mean velocity on the KIF5A330-to-tubulin ratio. (D) Mean distance the transporter kinesin moved from the moment of association with a microtubule upon its final stopping or detachment as function of the KIF5A330to-tubulin ratio.
K. Dreblow et al. / Biochemical and Biophysical Research Communications 395 (2010) 490–495
493
tonics GmbH, Germany) and the image processing system Argus 20 (Hamamatsu). Video sequences, documenting movement on a digital hard-disk recorder, were digitized resulting in image sequences of 30 frames/s. The sequences were analyzed using a special Plug-In for the ImageJ software [26]. The mean velocity and distance of movement were calculated from trajectories gold bead-labeled kinesin molecules.
3. Results 3.1. Irreversible blocking of kinesin-binding sites along the microtubules To ensure a reproducible control of single motor behaviour, actuation, and functionalization [27], we performed a motility assay, which was based on the ability of motor proteins to move cargoes along proteinaceous filaments immobilized to a substrate. As transporter full-length neuron-specific kinesin KIF5A (KIF5Afl) was used throughout the present study. To study kinesin-dependent transport mechanisms under conditions of blocked kinesin-binding sites, the assay was modified as follows: before inducing the movement of the transporter kinesin, kinesin-binding sites along the microtubules were specifically occupied by molecules of the monomeric KIF5A motor domain (KIF5A330) in the presence of a non-hydrolyzable ATP analogue. The blocking KIF5A was irreversibly cross-linked to the microtubules by glutaraldehyde treatment to prepare stable microtubule rails. After neutralization of residual free aldehyde groups, the native transporter KIF5Afl, labeled by 20-nm gold beads, was added to the microtubules and allowed to move after ATP supply. This procedure guarantees that the blocking protein competes directly with the transporter for binding sites. As a result of glutaraldehyde fixation, KIF5A330 became irreversibly bound to kinesin-binding sites and acted as non-exchangeable blockage. Control experiments with microtubules lacking protein blockages proved that under the conditions of glutaraldehyde fixation the binding of the transporter kinesin and its movement along the microtubules was not affected (see [15]). 3.2. Discontinuous kinesin movement in the presence of blockages In control preparations without blocking protein, the kinesin moved continuously at a mean velocity of 783 ± 194 nm/s (arithmetic mean ± SD). In the presence of KIF5A330 blockages, phases of movement were observed to alternate with resting phases (Fig. 1). Accordingly, the mean velocity decreased (Fig. 4A–C). Between stop events, the velocity was found to be nearly the same as in control experiments. The increase of the KIF5A330-to-tubulin ratio to values above 0.1 did not cause stronger effects on the velocity of kinesin movement (Fig. 2C). But with increasing this ratio, we observed a lowered frequency of binding of transporter kinesin to the microtubules, obviously due to the reduced number of free binding sites that could be accessed. However, once a kinesin molecule was bound and started to move, it was usually (in more than 80% of cases observed) able to overcome blockages after pausing without being released from its microtubule rail. In the other cases, the transporter either detached from the microtubule or finally stalled. Unlike the mean velocity characteristics, the mean distances the transporter moved from the moment of binding to the microtubule upon its final stopping or detachment were widely unchanged (Fig. 2D).
Fig. 3. Kinesin movement along microtubules in the presence of the slow Eg5 motor. (A) Distance–time diagrams documenting alternating phases of movement and resting observed in the presence of exchangeable and non-exchangeable Eg5513 (mean velocities including resting phases 354 and 223 nm/s, respectively). Eg5-totubulin ratio 0.2. Further details see Fig. 1A. (B) Velocity frequency histogram in the presence of exchangeable Eg5513. The x-values indicate velocity intervals ±50 nm/s. n is the numbers of moving objects measured. Remarkably, the velocities significantly exceeded those typical for Eg5 (40 nm/s).
3.3. Kinesin movement in the presence of slowly moving Eg5 In addition to KIF5A, slowly moving dimeric human Eg5513 motor protein was used to block kinesin-binding sites along the microtubule surface. Control experiments in which this motor was labeled by 20-nm gold demonstrated that Eg5513 moved along immobilized microtubules at a mean velocity of 40 ± 20 nm/s. When microtubules were irreversibly blocked by Eg5, the transporter KIF5Afl moved in discontinuous fashion, which resulted in lowering the mean velocity (Fig. 3A). As expected, like in the case of the irreversibly bound, non-exchangeable KIF5A330 (see Fig. 1A) phases of movement alternated with resting ones, causing a decrease of the mean velocity. Surprisingly, principally the same characteristic was observed when the Eg5 was not chemically cross-linked to the microtubules to keep it in native, motility-competent state. However, the resting phases were usually shorter and the mean velocity was tendentiously higher than in the case of non-exchangeable Eg5 (Fig. 3A). It has to be emphasized that in the presence of the native, exchangeable Eg5, velocities in the range of 40 nm/s, which are
494
K. Dreblow et al. / Biochemical and Biophysical Research Communications 395 (2010) 490–495
4. Discussion
Fig. 4. Concept of kinesin bypassing: variations of overcoming blockages. (A) The transporter kinesin is a dimer with two heads (green) moving along a microtubule and encountering a blockage (red). Considering only forward steps, to overcome the blockage the trailing head can theoretically access four b-tubulin binding sites (orange dots) both at the right and at the left related to moving direction. (B) Bypassing using the adjacent protofilament. (a) The trailing head binds to the btubulin of the adjacent protofilament and becomes located side-by-side with the leading head. In the 2nd step, the former leading head is pushed forward and binds to the next free b-tubulin of the adjacent protofilament and becomes located sideby-side with the blockage. (b) The trailing head binds side-by-side with the blockage. In the 2nd step, the former leading head is pushed forward and binds to the next free b-tubulin along the new protofilament track. (C) Bypassing using the next but one protofilament. (a) The trailing head turns sideward and binds approximately perpendicular to the b-tubulin of the next but one protofilament. In the 2nd step, the former leading head is pushed forward and binds also to this protofilament. (b) The trailing head is pushed forward and binds to the b-tubulin of the next but one protofilament. In the 2nd step, the former leading head is pushed forward and binds to the next free b-tubulin along the new protofilament track.
typically measured for Eg5 in control experiments, were not found (see Fig. 3B).
The present study aims at elucidation of mechanisms enabling kinesin to pass non-exchangeable blockages along its protofilament track. With this intention, an experimental approach has been worked out in which kinesin-binding sites along the microtubule surface were specifically occupied by KIF5A heads. To avoid their release, the heads were chemically cross-linked to the microtubule surface to get stable, non-exchangeable blockages. Thereafter, native gold-labeled full-length KIF5A as transporter was allowed to bind and to move along the blocked microtubules. Unlike observed for control preparations lacking blockages, movement was observed to be a non-continuous one, whereby phases of active transport alternated with resting ones. As a result, the mean velocity was lowered. Regardless of the presence of blocking molecules, the transporter kinesin was found to move processively over relatively long distances which are in the same range like in control preparations without blockages. As in our motility experiments the blocking protein was irreversibly attached to the microtubule rail by cross-linking with a chemical fixative, the continuation of movement after stopping at a blocked site can be explained by postulating a bypassing mechanism (Fig. 4). Kinesin moves in an asymmetric hand-overhand like fashion [28,29], whereby the leading head remains bound until the trailing head attaches to the next tubulin dimer. In the case of an irreversible bound blockage, the trailing head is not able to bind to the next free tubulin dimer along its protofilament track and the leading head cannot detach what causes stalling of the cargo transporter. Recently, Yildiz et al. [16] reported that, even in the absence of a lateral force, there is a small portion of sideways steps which kinesin performs during stepping. Taken the molecular flexibility of kinesin into consideration, we hypothesize that the leading head remains bound to the original protofilament whereas the trailing one turns sideward and contacts a tubulin binding site on the adjacent or on the next but one protofilament (for different modes of binding see Fig. 4). We assume that both free protofilaments on the right and on the left, corresponding to movement direction, can be used. Finally, the leading head leaves the blocked protofilament and follows the trailing one along its new track. Our model of bypassing obstacles by making sidewards steps of various sizes is consistent with the theoretical torsional spring model of Bolterauer et al. [30]. Bypassing by taking sideways steps in the presence of blocked kinesin-binding sites is regarded to be a complex mechanism, including ‘‘recognition” of the blocked site and diffusional search of the trailing head until finding an accessible free binding site on a neighbouring protofilament, which finally cause temporary stopping events. Table 1 Theoretical step widths during transition of the trailing head to a free b-tubulin of a neighbouring protofilament. Calculations were performed using the following data: 5 nm lateral spacing between the protofilaments; 1 nm axial shift between adjacent protofilaments; 8 nm tubulin dimer length. Mode
Bypassing in moving direction
Transition to the adjacent protofilament (see Fig. 4B) a Right a Left b Right b Left Transition to the next but one protofilament (see Fig. 4C) a Right a Left b Right b Left
Step width (nm) 8.6 10.2 15.8 17.7 11.7 14.1 17.2 20.6
K. Dreblow et al. / Biochemical and Biophysical Research Communications 395 (2010) 490–495
For the single modes of bypassing described (Fig. 4B and C) theoretical step sizes between 8.6 and 20.6 nm were calculated for the first head (Table 1). These values seem to be realistic as single kinesin heads have been reported to take steps of 17.3 ± 3.3 nm [29]. In our model, we consider only forward steps, as in this case the ATPase cycle and the resulting coordinated alternating movement of both heads proceeds in regular fashion. In the case of transition to the adjacent protofilament (Fig. 4B), steps should occur during which either both heads of the kinesin transporter or one head of the transporter and a head of the blocking KIF5A are located side-by-side at about 5 nm distance. Chrétien et al. [10] demonstrated that under artificial conditions, suppressing ATPase activity, kinesin heads are able to bind in an equimolar ratio to microtubules. However, we consider that under conditions of active movement, the heads of the dimeric kinesin require a higher degree of freedom. Consequently, the probability that its free head, looking for an alternative binding site by diffusional search, binds directly side-by-side to another head seems to be rather low. Therefore, we favor the transition of the free head of the moving kinesin to the next but one protofilament (Fig. 4C) in cases of blocking the kinesin-binding site by an obstacle with dimensions in the range of a kinesin head. Transitions to the adjacent protofilament (Fig. 4B) might occur in cases of disturbed kinesin binding to the next dimer along the original protofilament track, which might be caused by e.g., structural defects of the microtubule lattice. Concerning the question whether kinesin encountering a blocked binding site along its protofilament track waits until its track becomes free, we performed additional experiments in which the transporter kinesin competed with native, exchangeable Eg5, known to move at velocities which are lower by a factor of about 25 compared to kinesin-1 [21]. We found that not only in the case of irreversibly bound Eg5 blockages, but also in the presence of native, exchangeable Eg5 the transporter kinesin moved in non-continuous fashion. If the fast kinesin would not be able to overtake the slowly moving motor it should move behind it at velocities of 40 nm/s. As such velocities were not measured (see Fig. 3B), we suggest that the transporter kinesin does not solely wait for the release of Eg5, but is assumed to bypass also the slow motor. 5. Conclusions Bypassing blockages can be regarded as a fundamental mechanism by which the track fidelity of a molecular motor can be improved. Elucidation of this mechanism should contribute to deeper understanding of the molecular mechanisms of cytoplasmic transport processes over long distances, required for e.g., motor neurons. Acknowledgments This study was supported by the European Commission within the Sixth EU Framework Programme, Contract No. NMP4-CT-2004516989. We thank Dr. Ronald D. Vale (University of California, San Francisco) for supplying the KIF5A cDNA, Janina Beeg and Ruben Serral Graciá (MPI of Colloids and Interfaces, Golm) for kindly providing the ImageJ Plug-In, and Marina Wollmann for her skilful technical assistance.
495
References [1] N. Hirokawa, Y. Noda, Y. Tanaka, S. Niwa, Kinesin superfamily motor proteins and intracellular transport, Nat. Rev. Mol. Cell Biol. 10 (2009) 682–696. [2] N. Hirokawa, K.K. Pfister, H. Yorifuji, M.C. Wagner, S.T. Brady, G.S. Bloom, Submolecular domains of bovine brain kinesin identified by electron microscopy and monoclonal antibody decoration, Cell 56 (1989) 867–878. [3] W.O. Hancock, J. Howard, Processivity of the motor protein kinesin requires two heads, J. Cell Biol. 140 (1998) 1395–1405. [4] S. Ray, E. Meyhöfer, R.A. Milligan, J. Howard, Kinesin follows the microtubule’s protofilament axis, J. Cell Biol. 121 (1993) 1083–1093. [5] S. Ray, S.G. Wolf, J. Howard, K.H. Downing, Kinesin does not support the motility of zinc-macrotubes, Cell Motil. Cytoskel. 30 (1995) 146–152. [6] H. Hagiwara, H. Yorifuji, R. Sato-Yoshitake, N. Hirokawa, Competition between motor molecules (kinesin and cytoplasmic dynein) and fibrous microtubuleassociated proteins in binding to microtubules, J. Biol. Chem. 269 (1994) 3581– 3589. [7] R. Dixit, J.L. Ross, Y.E. Goldman, E.L. Holzbaur, Differential regulation of dynein and kinesin motor proteins by tau, Science 319 (2008) 1086–1089. [8] M. Martin, S.J. Iyadurai, A. Gassman, J.G. Gindhart, T.S. Hays, W.M. Saxton, Cytoplasmic dynein, the dynactin complex, and kinesin are interdependent and essential for fast axonal transport, Mol. Biol. Cell 10 (1999) 3717–3728. [9] A. Seitz, H. Kojima, K. Oiwa, E.M. Mandelkow, Y.H. Song, E. Mandelkow, Singlemolecule investigation of the interference between kinesin, tau and MAP2c, EMBO J. 21 (2002) 4896–4905. [10] D. Chrétien, F. Metoz, F. Verde, E. Karsenti, R.H. Wade, Lattice defects in microtubules: protofilament numbers vary within individual microtubules, J. Cell Biol. 117 (1992) 1031–1040. [11] A. Seitz, T. Surrey, Processive movement of single kinesins on crowded microtubules visualized using quantum dots, EMBO J. 25 (2006) 267–277. [12] I.A. Telley, P. Bieling, T. Surrey, Obstacles on the microtubule reduce the processivity of kinesin-1 in a minimal in vitro system and in cell extract, Biophys. J. 96 (2009) 3341–3353. [13] T. Korten, S. Diez, Setting up roadblocks for kinesin-1: mechanism for the selective speed control of cargo carrying microtubules, Lab Chip 8 (2008) 1441–1447. [14] Y. Chai, R. Lipowsky, S. Klumpp, Transport by molecular motors in the presence of static defects, J. Stat. Phys. 135 (2009) 241–260. [15] K. Dreblow, N. Kalchishkova, K.J. Böhm, Kinesin bypassing blockages on microtubule rails, Biophys. Rev. Lett. 4 (2009) 139–151. [16] A. Yildiz, M. Tomishige, A. Gennerich, R.D. Vale, Intramolecular strain coordinates kinesin stepping behavior along microtubules, Cell 134 (2008) 1030–1041. [17] M.L. Shelanski, F. Gaskin, C.R. Cantor, Microtubule assembly in the absence of added nucleotides, Proc. Natl. Acad. Sci. USA 70 (1973) 765–768. [18] M.D. Weingarten, A.H. Lockwood, S.Y. Hwo, M.W. Kirschner, A protein factor essential for microtubule assembly, Proc. Natl. Acad. Sci. USA 72 (1975) 1858– 1862. [19] K.J. Böhm, W. Vater, H. Fenske, E. Unger, Effect of microtubule-associated proteins on the protofilament number of microtubules assembled in vitro, Biochim. Biophys. Acta 800 (1984) 119–126. [20] J. Niclas, F. Navone, N. Hom-Booher, R.D. Vale, Cloning and localization of a conventional kinesin motor expressed exclusively in neurons, Neuron 12 (1994) 1059–1072. [21] N. Kalchishkova, K.J. Böhm, The role of kinesin neck linker and neck in velocity regulation, J. Mol. Biol. 382 (2008) 127–135. [22] A. Blangy, H.A. Lane, P. d’Hérin, M. Harper, M. Kress, E.A. Nigg, Phosphorylation by p34cdc2 regulates spindle association of human Eg5, a kinesin-related motor essential for bipolar spindle formation in vivo, Cell 83 (1995) 1159–1169. [23] K.J. Böhm, R. Stracke, P. Mühlig, E. Unger, Motorprotein-driven unidirectional transport of micrometer-sized cargoes across isopolar microtubule arrays, Nanotechnology 12 (2001) 238–244. [24] D. Turner, C. Chang, K. Fang, P. Cuomo, D. Murphy, Kinesin movement on glutaraldehyde-fixed microtubules, Anal. Biochem. 242 (1996) 20–25. [25] A.K. Boal, H. Tellez, S.B. Rivera, N.E. Miller, G.D. Bachand, B.C. Bunker, The stability and functionality of chemically crosslinked microtubules, Small 2 (2006) 793–803. [26] J. Beeg, S. Klumpp, R. Dimova, R.S. Gracià, E. Unger, R. Lipowsky, Transport of beads by several kinesin motors, Biophys. J. 94 (2008) 532–541. [27] M.G.L. van den Heuvel, C. Dekker, Motor proteins at work for nanotechnology, Science 317 (2007) 333–336. [28] C.L. Asbury, A.N. Fehr, S.M. Block, Kinesin moves by an asymmetric hand-overhand mechanism, Science 302 (2003) 2130–2134. [29] A. Yildiz, M. Tomishige, R.D. Vale, P.R. Selvin, Kinesin walks hand-over-hand, Science 303 (2004) 676–678. [30] H. Bolterauer, J.A. Tuszynski, E. Unger, Directed binding: a novel physical mechanism that describes the directional motion of two-headed kinesin motor proteins, Cell Biochem. Biophys. 42 (1984) 95–119.