Journal list menu

Volume 584, Issue 11 p. 2351-2355
Short communication
Free Access

Nucleotide-dependent behavior of single molecules of cytoplasmic dynein on microtubules in vitro

Michi Miura

Michi Miura

Department of Life Sciences, Graduate School of Arts and Sciences, University of Tokyo, Komaba 3-8-1, Meguro-ku, Tokyo 153-8902, Japan

Present address: Laboratory of Virus Control, Institute for Virus Research, Kyoto University, Kyoto 606-8507, Japan. Search for more papers by this author
Aiko Matsubara

Aiko Matsubara

Department of Life Sciences, Graduate School of Arts and Sciences, University of Tokyo, Komaba 3-8-1, Meguro-ku, Tokyo 153-8902, Japan

Search for more papers by this author
Takuya Kobayashi

Takuya Kobayashi

Department of Life Sciences, Graduate School of Arts and Sciences, University of Tokyo, Komaba 3-8-1, Meguro-ku, Tokyo 153-8902, Japan

Search for more papers by this author
Masaki Edamatsu

Masaki Edamatsu

Department of Life Sciences, Graduate School of Arts and Sciences, University of Tokyo, Komaba 3-8-1, Meguro-ku, Tokyo 153-8902, Japan

Search for more papers by this author
Yoko Y. Toyoshima

Corresponding Author

Yoko Y. Toyoshima

Department of Life Sciences, Graduate School of Arts and Sciences, University of Tokyo, Komaba 3-8-1, Meguro-ku, Tokyo 153-8902, Japan

Corresponding author. Fax: +81 3 5454 6722.Search for more papers by this author
First published: 13 April 2010
Citations: 20

Abstract

We visualized the nucleotide-dependent behavior of single molecules of mammalian native cytoplasmic dynein using fragments of dynactin p150 with or without its N-terminal microtubule binding domain. The results indicate that the binding affinity of dynein for microtubules is high in AMP-PNP, middle in ADP or no nucleotide, and low in ADP·Pi conditions. It is also demonstrated that the microtubule binding domain of dynactin p150 maintains the association of dynein with microtubules without altering the motile property of dynein in the weak binding state. In addition, we observed bidirectional movement of dynein in the presence of ATP as well as in ADP/Vi condition, suggesting that the bidirectional movement is driven by diffusion rather than active transport.

1 Introduction

Cytoplasmic dynein is a minus-end directed microtubule motor, and functions as a large complex within the cell [1, 2]. The N-terminal region of the heavy chain forms the tail domain, which is responsible for dimerization and associating with intermediate and light intermediate chains [3]. Cytoplasmic dynein hydrolyzes ATP at the ring-shaped motor domain located in the C-terminal two-thirds of the heavy chain [4-6]. Cytoplasmic dynein binds to microtubules via the stalk head, the distal tip of the anti-parallel coiled-coil protruding from the ring structure [7].

Results of previous studies on recombinant cytoplasmic dynein suggest that the binding affinity for microtubules depends on the registry of the coiled-coil of the stalk and the nucleotide states at AAA (ATPases associated with diverse cellular activities) modules [8-10]. For native cytoplasmic dynein, the motile property of single dynein molecules associated with the dynactin complex was investigated in the presence of ATP [11], but their behavior on microtubules in various nucleotide conditions has not been explored sufficiently.

Here, we examined the behavior of single molecules of native cytoplasmic dynein on microtubules under different nucleotide conditions. We observed single molecules of native cytoplasmic dynein using the dynein binding protein, dynactin p150. Dynactin p150 is the major component of the dynactin complex, and is responsible for binding to dynein by the interaction with dynein intermediate chains via its coiled-coil region [12]. Its N-terminus contains a conserved CAP-Gly (cytoskeleton-associated protein, glycine-rich) motif, which binds to microtubules [13], and has been shown to promote the motor processivity of cytoplasmic dynein [14, 15]. We prepared two fragments of dynactin p150 fused to gelsolin. Molecules of the dynein/dynactin complex can be visualized as fluorescent actin spots by epifluorescence microscopy according to the method of Yajima et al. [16]. This method is suitable for long term observation, which is often prevented by fluorescence quenching in single molecule motility assays, because each short actin filament contains multiple fluorescent dyes.

By investigating the behavior of the dynein/dynactin complex on microtubules, we will discuss the nucleotide dependence of binding affinity for microtubules and the effect of dynactin p150 on the behavior of cytoplasmic dynein on microtubules.

2 Materials and methods

2.1 Preparation of proteins

Expression plasmids of gelsolin-fused proteins (C1G and NC1G) were prepared by inserting the DNA fragments coding human dynactin p150 into the 5′ terminus of human plasma gelsolin. Proteins were expressed bacterially and purified (for the details, see Supplementary data).

Cytoplasmic dynein was prepared from porcine brain [17]. To exclude the possible dynein fraction that binds to actin filaments used as a fluorescent probe, we removed this fraction by co-precipitating dynein with F-actin.

Tubulin was purified from porcine brain [18], and was labeled with BODIPY FL (Molecular Probes) [19]. Microtubules were polymerized [19] and stabilized with 40 μM taxol.

Actin was prepared from rabbit skeletal muscle [20], and labeled with phalloidin-tetramethylrhodamine B isothiocyanate conjugate (Fluka) [21].

2.2 Single molecule imaging of the gelsolin–actin complex

The procedures for preparing gelsolin-capped actin filaments and epifluorescence microscopy were as described previously [16]. In this study, dynein and gelsolin–actin mixture in motility buffer (10 mM Pipes-KOH, pH 6.7/4 mM MgSO4/1 mM DTT) containing 20 mM K-acetate, 0.6 mg ml−1 casein, 40 μM taxol, 0.1% Tween 20 (Nacalai Tesque) and the oxygen scavenger system [22] was loaded. This solution was supplemented with either of 1 mM AMP-PNP, 1 mM ADP/Vi, 1 mM ADP or no nucleotide. Each fluorescent actin filament was severed by gelsolin shortly enough to be visualized as a spot rather than a rod-like structure. Analysis of the behavior of fluorescent molecules is described in Supplementary data.

3 Results and discussion

3.1 Nucleotide dependence of C1G-labeled cytoplasmic dynein

Fig. 1 shows the fusion proteins used in this study and schematic diagrams of the complex of cytoplasmic dynein and these proteins on a microtubule. C1G contains a dynein binding region of 217–548 amino acids [12], and NC1G contains both dynein binding and microtubule binding regions.

figure image
Schematic diagrams of the fusion proteins used in this study and the complex of cytoplasmic dynein and these proteins. C1G has a dynein binding domain and NC1G contains both dynein binding and microtubule binding domains. Gelsolin caps rhodamine-labeled short actin filament (red) and, thus, native cytoplasmic dynein is visualized as a fluorescent actin spot by epifluorescence microscopy.

First, we labeled dynein with C1G/rhodamine–actin and observed its behavior on microtubules under four nucleotide conditions (AMP-PNP, ADP/Vi, ADP, or no nucleotide). The presence of vanadate inhibits ATPase activity of dynein and mimics the ADP·Pi state [23]. Each nucleotide was added at a final concentration of 1 mM, which was an excess amount compared with the trace amount of ATP (≈1 μM) from the final purification step of native cytoplasmic dynein.

We counted the numbers of dynein molecules per unit length of microtubules and tracked those individual molecules. As shown in Fig. 2 a, a relatively large amount of dynein molecules was associated with microtubules in the presence of AMP-PNP, and they remained on microtubules for tens of seconds. Under the condition of ADP or no nucleotide, the number of molecules on microtubules was 10–50% of that in AMP-PNP. In the presence of ADP/Vi, dynein molecules rarely associated with microtubules, resulting in <1% of dynein molecules on microtubules compared with AMP-PNP, and dynein molecules dissociated from the microtubules within a few seconds after binding to the microtubules. The histogram of the duration on microtubules in ADP/Vi was fitted to a single exponential decay, and its mean duration was calculated to be 1.34 s (Fig. 2b). Thus, the numbers of molecules on microtubules were correlated to the duration on microtubules. The trajectory data show that dynein did not move widely on microtubules under the condition of AMP-PNP. The center position of typical tracked molecules in the presence of AMP-PNP distributed within approximately 60 nm, which is consistent with previous results of kinesin bound to microtubules in the AMP-PNP condition [16]. Under the condition of ADP or no nucleotide, dynein fluctuated slightly. However, dynein fluctuated widely along the longitudinal axis of microtubules in the presence of ADP/Vi (Fig. 2c and also see Supplementary Movie S1 and S2).

figure image
The behavior of the dynein–C1G complex. (a) Relative numbers of dynein molecules per unit length of microtubules in each nucleotide condition standardized by the numbers in the AMP-PNP condition. (b) The histogram of duration of the dynein–C1G complex on microtubules in the presence of ADP/Vi. The graph was fitted to a single exponential decay. The mean duration was 1.34 s. (c) Displacement of the dynein–C1G complex along the longitudinal axis of microtubules in each nucleotide condition. Three typical examples are shown in three colors.

The dynein–C1G complex interacts with microtubules through the dynein stalk heads because C1G lacks the N-terminal microtubule binding domain (Fig. 1), and the numbers of bound molecules on microtubules are considered to correspond to the affinity for microtubules. Thus, these results are comparable to those of the previous study by Imamula et al. [9] on the recombinant single-headed cytoplasmic dynein. The dissociation constant was approximately 0.2 μM under the condition of AMP-PNP, ADP or no nucleotide as the strong binding state, whereas the constant was >10 μM in the presence of vanadate as the weak binding state. In our study, the number of dynein molecules associated with microtubules in the presence of AMP-PNP was distinct from the other three conditions. Among these three conditions, dynein frequently associated with microtubules and did not move widely under the condition of ADP or no nucleotide, whereas dynein rarely associated with microtubules and fluctuated largely once associated with microtubules in the presence of ADP/Vi, indicating that the microtubule binding property of dynein under these three conditions can be categorized further into two groups. Taken together, our results suggest that the affinity of native cytoplasmic dynein for microtubules is high in AMP-PNP, middle in ADP or no nucleotide, and low in ADP·Pi conditions. These results show that the binding affinity of dynein for microtubules under the condition of ADP or no nucleotide is distinguishable from that of AMP-PNP, whereas the binding affinity in these three conditions was similar in the previous study by Imamula et al. [9]. The difference might be attributed to the double-headed dynein in this study and the single-headed dynein in the study of Imamula et al. Namely, both heads of the double-headed molecule may bind to the microtubule in AMP-PNP, whereas only one head of the double-headed molecule may bind to the microtubule in ADP or no nucleotide as is the case for kinesin [24, 25], although the relation between the two heads of dynein has not been elucidated.

3.2 The behavior of cytoplasmic dynein associated with NC1G

Since NC1G contains the microtubule binding domain at its N-terminus, we examined whether NC1G/rhodamine–actin binds to microtubules by itself without dynein. NC1G alone bound to microtubules (Fig. 3 a), and fluctuated on the longitudinal axis of microtubules (Fig. 3b), as observed previously [26]. At a high molar ratio of NC1G/dynein in solution, unassociated free NC1G fragments may stay on microtubules together with the dynein–NC1G complexes, and it must be difficult to distinguish the dynein–NC1G complexes from free NC1G fragments. We determined that a NC1G/dynein molar ratio of 1/4 is appropriate, and confirmed that almost all the fluorescent actin spots represents NC1G-associated dynein, because all the molecules observed on microtubules did not fluctuate in the presence of AMP-PNP, which is not the property of NC1G alone, but the property of dynein, as was observed above.

figure image
The behavior of the dynein–NC1G complex. (a) The numbers of NClG molecules without dynein and the dynein–NC1G complex in each nucleotide condition per unit length of microtubules, standardized by the numbers in the AMP-PNP condition. (b) Displacement of NC1G along the longitudinal axis of microtubules. (c) Displacement of the dynein–NC1G complex along the longitudinal axis of microtubules in each nucleotide condition. Three typical examples are shown in three colors.

When cytoplasmic dynein was associated with NC1G, the numbers of molecules on microtubules in the four nucleotide conditions became similar (Fig. 3a). In particular, the numbers of molecules associated with microtubules in ADP/Vi were significantly increased, and its duration on microtubules became comparable to that of the other nucleotide conditions, but the dynein–NC1G complex still fluctuated widely on microtubules (Fig. 3c). These results are consistent with the idea that the microtubule binding domain of dynactin p150 contributes to an increase in the duration of dynein on microtubules by maintaining dynein association with microtubules [14, 15]. Moreover, this tethering force of dynactin p150 does not seem to alter the behavior that dynein fluctuates widely on microtubules in the weak binding state.

3.3 Motile property of cytoplasmic dynein and contribution of dynactin p150

Next, we observed the behavior of the dynein–C1G complex and the dynein–NC1G complex in the presence of 1 mM ATP. The dynein–C1G complex rarely associated with microtubules, while the dynein–NC1G complex associated with microtubules for a while and moved in both directions along the longitudinal axis of the microtubules (Fig. 4 and also see Supplementary Movie S3), demonstrating the effect of the microtubule binding region at the N-terminus of NC1G, as discussed above. The behavior of the dynein–NC1G complex in ATP is consistent to previous studies showing that the microtubule binding region of dynactin increases the duration of cytoplasmic dynein on microtubules [14, 15], and that the mammalian native dynein/dynactin complex moves on microtubules with a run length of hundreds of nanometers in both directions [11]. However, the behaviors of the dynein–NC1G complex in ADP/Vi (Fig. 3c, ADP/Vi) and in ATP (Fig. 4) were bidirectional and indistinguishable, although we have not determined the directionality of movement. We estimated the diffusion constant of the dynein–NC1G complex in the condition of ADP/Vi and ATP from the MSD (mean square displacement) plots, and found that the preliminary values for each condition were within a similar range (approximately 6–10 × 10−10 cm2 s−1). These results suggest that the dynein–microtubule interaction is weak and the contribution of diffusion is large in ATP, and that the observed movement in ATP may be attributed mostly to diffusion, but not to energy transduction from ATP hydrolysis. The motility of dynein molecules prepared in this study was confirmed to be active by the microtubule gliding assay (average velocity of 700 nm s−1), and the method of preparing dynein used in this study was identical to that in the single molecule bead assay [27]. Thus, the diffusive movement shown here is not because the dynein molecules are inactivated, but probably due to the nature of single molecules of mammalian cytoplasmic dynein in vitro without the reducing effect of diffusion by the bead or the laser trap. Ross et al. [11] suggested that the bidirectional motility is attributed to an active energy transduction property of dynein because the velocity in both direction is dependent on the ATP concentration. Based on our observation that the dynein–NC1G complex in ADP/Vi (ATP non-consuming state) moves largely enough, the movement by Ross et al. may be explained by the idea that the waiting time of ATP binding decreases as the ATP concentration increases, and, thus, the ratio of the ADP·Pi state (weak binding state) in the ATPase cycle increases, resulting in dynein being more susceptible to the driving force of diffusion.

figure image
Displacement of the dynein–NC1G complex along the longitudinal axis of microtubules in the presence of 1 mM ATP. The polarity of microtubules was not examined, and thus, the plus and the minus orientations in the vertical axis are arbitrary.

This behavior of mammalian native dynein seems to be different from that of Saccharomyces cerevisiae dynein, which moves on microtubules towards the minus-end of the microtubules with infrequent backward steps [28]. It is plausible that mammalian dynein associates with microtubules more weakly than yeast dynein, resulting in the frequent backward movement of the mammalian dynein/dynactin complex over a long distance as a diffusional movement. Recently, it was reported that the yeast dynactin complex increases the run length of native yeast dynein [29]. Notably, the microtubule binding region of Nip100 (yeast dynactin p150 homolog) was not essential, and the region between the CAP-Gly domain and dynein binding domain of Nip100 was sufficient for increasing the duration of dynein on microtubules [29]. These results suggest that Nip100 increases the duration not by binding to the microtubule, but by binding to dynein itself. Meanwhile, the effect of the microtubule binding region of mammalian dynactin p150 has been suggested to be significant for increasing the duration of dynein on microtubules in previous studies [14, 15] and this study. A possible explanation for this difference between the function of mammalian and yeast dynactin is that the effect of the microtubule binding domain of dynactin is more prominent in mammalian dynein than in yeast dynein because the binding affinity of mammalian dynein for microtubules is weak compared with that of yeast dynein, as discussed above.

Within a cell, the retrograde transport of cargoes occurs over several micrometers and continues for over several seconds [30, 31]. This movement of cargoes is thought to be driven by dynein. However, single molecules of mammalian dynein do not show such movement in vitro. It is plausible that dynein-bound cargoes are not much affected by the driving force of diffusional movement. Furthermore, such movement of dynein-driven cargoes is thought to be achieved by multiple molecules of dynein on the same cargoes. Another possibility is that there are the regulatory factors promoting the unidirectional movement of mammalian dynein. To further investigate dynein motility, the in vitro motility of multiple cytoplasmic dynein molecules should be addressed, and the roles and functions of dynein associated proteins should be explored.

Acknowledgements

The authors thank Dr. Ken'ya Furuta for developing the tracking program, MARK. This research was supported by grants from CREST of Japan Science and Technology Agency and Special Coordination Funds for Promoting Science and Technology of the Ministry of Education, Culture, Sports, Science and Technology, the Japanese Government to Y.Y.T.

    Appendix A A

    Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.febslet.2010.04.016.