Cells rely on ‘molecular motors’ travelling along tracks called microtubules to move proteins and other cargoes between different parts of a cell. Dynein is a molecular motor that moves along the microtubules by taking “steps” towards the slowly growing end of these tracks.

The trajectories of dynein motors have been studied extensively using techniques that can follow their movements in two dimensions. However, some molecular motors can also rotate as they travel, creating a twisting force called a torque that causes the motor to spiral around the microtubule in a helix.

To assess the torque that dynein can generate and to better understand its movements in three dimensions, Can et al. used a length of microtubule to build a ‘bridge’ between two polystyrene beads. The dynein motors were made to carry a smaller polystyrene bead as cargo, and the movement of this smaller bead was tracked using a computer algorithm to interpret the motion recorded by a microscope.

Can et al. found that dynein moves in a helical trajectory around the microtubule, rather than travelling along it in a straight line. As it travels it can twist in one direction or the other, generating torque in either direction. This is unlike other types of molecular motor, which produce torque in just one direction. However, dynein prefers to rotate to the right, suggesting that with every step along a microtubule, it binds to the closest available binding site in the forward direction.

Why might it be useful for molecular motors to behave in this way? Can et al. propose that the ability to rotate in both directions may allow dynein to avoid roadblocks or other obstacles in the dense and busy cellular environment in which it has to operate.

Cytoskeletal motors transport a wide variety of intracellular cargos by processively moving along linear tracks. These motors do not always follow a linear trajectory. Rather, they generate torque perpendicular to their direction of motion, resulting in a helical movement relative to the filament. Such helical movement was first observed in a filament gliding assay in which surface-immobilized Tetrahymena axonemal dynein motors rotated MTs about their principal axes while translocating them (Vale and Toyoshima, 1988). Helical movement of cargoes has subsequently been demonstrated for several members of the myosin and kinesin superfamilies (Nishizaka et al., 1993; Nitzsche et al., 2008; Yajima et al., 2008; Bormuth et al., 2012; Brunnbauer et al., 2012).

Cytoplasmic dynein is the primary MT minus-end directed motor responsible for diverse cellular processes in cargo transport, nuclear positioning, and cell division (Roberts et al., 2013). Despite its importance, key aspects of dynein's mechanism remain unclear, including whether or not the motor can produce torque. In contrast to kinesin-1, which follows a single protofilament track (Ray et al., 1993; Yajima and Cross, 2005), dynein takes frequent sideways steps (Reck-Peterson et al., 2006). However, the full trajectory of dynein motors in three dimensions (3D) remains unknown, because the motors are sterically prevented from moving around the circumference of surface-immobilized MTs.

To track dynein motility in 3D, we constructed MT ‘bridges’ (Brunnbauer et al., 2012) (Figure 1A), which allow unconstrained motion between the protofilaments. The bridges were constructed by suspending a fluorescently labeled MT between two large (2 µm diameter) polystyrene beads, immobilized on the surface. These beads were densely coated with a chimeric protein containing the dynein MT binding domain (MTBD) that stably binds to MTs (Gibbons et al., 2005; Carter et al., 2008). Multiple dynein motors were linked to a 0.5-µm diameter bead (referred to as cargo) and brought in proximity of the MT bridge with an optical trap. When the motors bound to the MT, the cargo bead was released from the trap and its motion was recorded with bright-field microscopy (Figure 1B). The xy position of the cargo was determined by a two-dimensional Gaussian tracking algorithm. The z position was determined by changes in the intensity of the bead center (Figure 1C–D), allowing us to track bead movement in 3D.

Figure 1 with 1 supplement see all

Download asset Open asset

We initially tested the robustness of our experimental geometry using human kinesin-1 motors, which follow a single protofilament (Ray et al., 1993; Yajima and Cross, 2005). Because the handedness of the MTs varies based on the number (11–14) of protofilament tracks (Hyman et al., 1995), we polymerized tubulin with a non-hydrolyzable GTP analog (GMP-CPP). 96% of GMP-CPP MTs is made of 14 protofilaments and has a left-handed supertwist with 6400 ± 1000 nm pitch (Hyman et al., 1995). The average velocity of kinesin-driven beads was 541 ± 23 nm/s (mean ± SEM, N = 10), comparable to the speed of single motors (Yildiz et al., 2008). We observed that kinesin-1-coated beads travelled along GMP-CPP MTs with a left-handed helical motion with a pitch of 6500 ± 400 nm (mean ±SEM, N = 6) (Figure 1—figure supplement 1, Video 1), similar to the helicity of 14 protofilament MTs. The results are also consistent with the reported values of kinesin-1 movement on GMP-CPP MTs (5700–7900 nm pitch, left-handed) using alternative geometries (Nitzsche et al., 2008; Brunnbauer et al., 2012).

Video 1

Download asset

Helical motility of dynein dimers

We next investigated the helical motility of cytoplasmic dynein. We used a tail-truncated yeast dynein artificially dimerized with glutathione S-transferase (GST-Dyn_331kD), which has similar motile properties to full-length dynein (Reck-Peterson et al., 2006). The motors were fused to GFP at the N-terminus and attached to cargo beads coated with anti-GFP antibodies (see ‘Materials and methods’). At 1 mM ATP, dynein-coated beads moved in helical trajectories with a pitch of 591 ± 32 nm (mean ± SEM, 67 rotations, 15 beads, Figure 2A–C). This rotation corresponds to a sideways movement to a neighboring protofilament (6 nm) for every six tubulin dimers (48 nm) in the forward direction. Unlike axonemal dynein and kinesin motors, which primarily rotate along their tracks in only one direction (Vale and Toyoshima, 1988; Brunnbauer et al., 2012), beads coated with cytoplasmic dynein exhibited both left- and right-handed helical movement on MTs (N = 67 rotations, Figure 2A–B, Figure 2—figure supplements 1 and 2, Videos 2,3). The speeds of left- and right-handed movements along the helical path (79 ± 4.7 nm/s and 89 ± 17 nm/s, mean ± SEM, respectively) were statistically indistinguishable (t-test, p = 0.28). Bidirectional helical movement was also evident from traces of single beads, which occasionally switch direction during a run (4 out of 67 rotations, Figure 2D). In contrast, reversal of bead motility along the MT axis was never observed.

Figure 2 with 5 supplements see all

Download asset Open asset

Video 2

Download asset

Video 3

Download asset

We next investigated the helical motility of a full-length yeast cytoplasmic dynein (Reck-Peterson et al., 2006). The motors were fused to GFP at the N-terminus and attached to cargo beads coated with anti-GFP antibodies. Dynein-coated beads moved in helical trajectories with a pitch of 500 ± 36 nm (mean ± SEM, Figure 2—figure supplement 3, Video 4) and exhibited both left- and right-handed helical movement. The velocities of left- and right-handed movement along the helical path (42 ± 11 nm/s and 43 ± 10 nm/s, mean ± SEM, respectively) were statistically indistinguishable (t-test, p = 0.21). Similar to GST-Dyn_331kD, beads driven by full-length dynein occasionally switch direction during a run (8 out of 33 rotations, Figure 2—figure supplement 4). Unlike GST-Dyn_331kD, which has a net preference for left-handed helical movement (75%), full-length dynein has a net preference for right-handed helical movement (58%, t-test, p = 10⁻⁵). This difference may be related to GST dimerization, in which the heads may be oriented differently relative to the MT surface.

Video 4

Download asset

The pitch of dynein-driven rotation is much shorter than the supertwist of the GMP-CPP MTs (∼6400 nm), suggesting that helical motility of dynein is independent from the helicity of the MT track. To verify this, we repeated the assay with GST-Dyn_331kD on taxol-stabilized MTs, which contain a mixture of 12 (77%) and 13 (11%) and 14 (2%) protofilaments (Ray et al., 1993). Out of 34 rotations, 68% were left-handed and 32% were right-handed. On average, the pitch of helical movement was 607 ± 50 nm (mean ±SEM) (Figure 2—figure supplement 5), similar to that of GMP-CPP MTs (t-test, p=0.77). The results demonstrate that our findings are not an artifact of the MT polymerization method.

To test the possibility whether bidirectional helical motility is driven by a rotational tug-of-war between dynein motors (i.e., some motors strictly rotate in a right-handed helix and the others rotate in the opposite direction), we tracked individual GST-Dyn_331kD motors labeled with a quantum-dot on MT bridges (Figure 3A). The average run length of single motors on MT bridges was 1.5 ± 0.2 µm (mean ± SEM, N = 10), which is similar to that measured on surface-immobilized MTs (Reck-Peterson et al., 2006). The velocity of single dimers was 64 ± 8 nm/s (mean ± SEM, N = 10), similar (t-test, p = 0.12) to that of cargo beads carried by multiple dimers (80 ± 10 nm/s, N = 15). The traces of single motors show high variability along the perpendicular axis of MTs and switch directions in their sideways movement more frequently than the beads carried by multiple motors (Figure 3B). These results exclude rotational tug-of-war.

Figure 3

Download asset Open asset

Helical motility of dynein monomers

We next tested the possibility that the relative orientation of the two motor domains in a dimer can influence the handedness of helical motility. To eliminate the contribution of interhead orientation, we used a cargo bead coated with monomeric Dyn_331kD, which is not processive on its own (Reck-Peterson et al., 2006). We observed that multiple monomers were able to drive processive motility of cargo beads at 61 ± 6 nm/s (mean ± SEM), consistent with the ability of multiple non-processive kinesins or myosins to drive processive motility (Kamei et al., 2005; Furuta et al., 2013).

The trajectories of beads driven by dynein monomers (Figure 4, Video 5) also displayed a helical component. The average pitch (579 nm ± 38 nm, 57 rotations, 11 beads) was similar to that of GST-Dyn_331kD (t-test, p = 0.78) and higher than full-length dynein (t-test, p = 0.16). The majority (59%) of the rotations was right-handed, similar to full-length dimers (N-1 two proportion test, p = 0.47). The speeds of left- and right-handed movements along the helical path (66 ± 10 nm/s and 57 ± 6 nm/s, respectively) were statistically indistinguishable (t-test, p = 0.18). The results indicate that right-handed preference of full-length dynein for helical movement is not due to the head–head orientation of the dimer.

Figure 4

Download asset Open asset

Video 5

Download asset

In this study, we showed that cytoplasmic dynein moves along MTs in a helical trajectory with a pitch of ∼500 nm. Single dynein dimers take frequent sideways steps (Reck-Peterson et al., 2006), whereas multiple dynein dimers persistently move in a helical path with a net preference for a right-handed rotation. What is the molecular basis of this preference? Kinesin-1 monomers are believed to move to the closest possible site on the microtubule which stays to the left of the previous binding site (Yajima and Cross, 2005). However, kinesin-1 dimers only walk along a single protofilament because their short neck-linker prevents off-axis stepping (Brunnbauer et al., 2012). In dynein, monomers may prefer to attach to the nearest tubulin binding site towards the MT minus end, favoring a rightward step (Figure 5A). In the case of a dimer, the heads are attached to neighboring protofilaments and the leading head prefers to be on the right side at 30° relative to the trailing head (DeWitt et al., 2012; Qiu et al., 2012). This orientation would make a rightward step more favorable for the trailing head, resulting in a net rightward bias (Figure 5B).

Figure 5

Download asset Open asset

In contrast to other motors studied to date, dynein moves along both left- and right-handed helical paths. The molecular basis of the switches in helical directionality remains unclear. The irregular stepping pattern of individual motors rules out the possibility of a rotational tug-of-war. Instead, we propose that the helical pattern of bead movement may be determined by the conformations of dynein motors associated with the MT track. It is likely that dyneins that are bound in sufficiently close proximity to other dyneins experience steric exclusion effects. If a bead is carried by motors which are oriented such that one is bound immediately forward and to the right of the other, the tubulin binding sites ahead and to the right of the trailing head will be obstructed. Therefore, these motors may prefer to step to the left, and the entire cargo will eventually trace out a left-handed spiral (Figure 5C). The number and orientation of the motors that are simultaneously in contact with the MT may change over the course of a recording. Individual dynein motors have 1400 nm run length (Reck-Peterson et al., 2006), indicating that motors on the bead detach and reattach during a processive run of a cargo bead. These may alter the orientation of the MT bound motors, resulting in the reversal of helical directionality during a processive run (Figure 5D), as occasionally observed.

Inside cells, the MT surface is crowded with associated proteins, which act as roadblocks during the transport of intracellular cargos (Stamer et al., 2002). Furthermore, the intracellular space is crowded with large structures, such as vesicles and organelles. The ability of molecular motors to produce torque and axial force may allow the motors to switch protofilaments and avoid these obstacles. In cells, the same MT track is used for both plus- and minus-end-directed transport. Sideways movement may prevent traffic jams on MTs. In vitro assays have shown that dynein can bypass roadblocks whereas kinesin-1 stalls when it encounters an obstacle and eventually releases (Dixit et al., 2008). Bidirectional helical movement may provide additional flexibility to dynein to transport cargos in dense cellular environments.

1. 1. Aitken CE
   2. Marshall RA
   3. Puglisi JD

   (2008) An oxygen scavenging system for improvement of dye stability in single-molecule fluorescence experiments

   Biophysical Journal 94:1826–1835.

   https://doi.org/10.1529/biophysj.107.117689

   • Google Scholar
2. 1. Bormuth V
   2. Nitzsche B
   3. Ruhnow F
   4. Mitra A
   5. Storch M
   6. Rammner B
   7. Howard J
   8. Diez S

   (2012) The highly processive kinesin-8, Kip3, switches microtubule protofilaments with a bias toward the left

   Biophysical Journal 103:L4–L6.

   https://doi.org/10.1016/j.bpj.2012.05.024

   • Google Scholar
3. 1. Brunnbauer M
   2. Dombi R
   3. Ho TH
   4. Schliwa M
   5. Rief M
   6. Okten Z

   (2012) Torque generation of kinesin motors is governed by the stability of the neck domain

   Molecular Cell 46:147–158.

   https://doi.org/10.1016/j.molcel.2012.04.005

   • Google Scholar
4. 1. Carter AP
   2. Garbarino JE
   3. Wilson-Kubalek EM
   4. Shipley WE
   5. Cho C
   6. Milligan RA
   7. Vale RD
   8. Gibbons IR

   (2008) Structure and functional role of dynein's microtubule-binding domain

   Science 322:1691–1695.

   https://doi.org/10.1126/science.1164424

   • Google Scholar
5. 1. DeWitt MA
   2. Chang AY
   3. Combs PA
   4. Yildiz A

   (2012) Cytoplasmic dynein moves through uncoordinated stepping of the AAA+ ring domains

   Science 335:221–225.

   https://doi.org/10.1126/science.1215804

   • Google Scholar
6. 1. Dixit R
   2. Ross JL
   3. Goldman YE
   4. Holzbaur EL

   (2008) Differential regulation of dynein and kinesin motor proteins by tau

   Science 319:1086–1089.

   https://doi.org/10.1126/science.1152993

   • Google Scholar
7. 1. Furuta K
   2. Furuta A
   3. Toyoshima YY
   4. Amino M
   5. Oiwa K
   6. Kojima H

   (2013) Measuring collective transport by defined numbers of processive and nonprocessive kinesin motors

   Proceedings of the National Academy of Sciences of the USA 110:501–506.

   https://doi.org/10.1073/pnas.1201390110

   • Google Scholar
8. 1. Gibbons IR
   2. Garbarino JE
   3. Tan CE
   4. Reck-Peterson SL
   5. Vale RD
   6. Carter AP

   (2005) The affinity of the dynein microtubule-binding domain is modulated by the conformation of its coiled-coil stalk

   The Journal of Biological Chemistry 280:23960–23965.

   https://doi.org/10.1074/jbc.M501636200

   • Google Scholar
9. 1. Hyman AA
   2. Chretien D
   3. Arnal I
   4. Wade RH

   (1995) Structural changes accompanying GTP hydrolysis in microtubules: information from a slowly hydrolyzable analogue guanylyl-(alpha,beta)-methylene-diphosphonate

   The Journal of Cell Biology 128:117–125.

   https://doi.org/10.1083/jcb.128.1.117

   • Google Scholar
10. 1. Kamei T
    2. Kakuta S
    3. Higuchi H

    (2005) Biased binding of single molecules and continuous movement of multiple molecules of truncated single-headed kinesin

    Biophysical Journal 88:2068–2077.

    https://doi.org/10.1529/biophysj.104.049759

    • Google Scholar
11. 1. Nishizaka T
    2. Yagi T
    3. Tanaka Y
    4. Ishiwata S

    (1993) Right-handed rotation of an actin filament in an in vitro motile system

    Nature 361:269–271.

    https://doi.org/10.1038/361269a0

    • Google Scholar
12. 1. Nitzsche B
    2. Ruhnow F
    3. Diez S

    (2008) Quantum-dot-assisted characterization of microtubule rotations during cargo transport

    Nature Nanotechnology 3:552–556.

    https://doi.org/10.1038/nnano.2008.216

    • Google Scholar
13. 1. Qiu W
    2. Derr ND
    3. Goodman BS
    4. Villa E
    5. Wu D
    6. Shih W
    7. Reck-Peterson SL

    (2012) Dynein achieves processive motion using both stochastic and coordinated stepping

    Nature Structural & Molecular Biology 19:193–200.

    https://doi.org/10.1038/nsmb.2205

    • Google Scholar
14. 1. Ray S
    2. Meyhofer E
    3. Milligan RA
    4. Howard J

    (1993) Kinesin follows the microtubule's protofilament axis

    The Journal of Cell Biology 121:1083–1093.

    https://doi.org/10.1083/jcb.121.5.1083

    • Google Scholar
15. 1. Reck-Peterson SL
    2. Yildiz A
    3. Carter AP
    4. Gennerich A
    5. Zhang N
    6. Vale RD

    (2006) Single-molecule analysis of dynein processivity and stepping behavior

    Cell 126:335–348.

    https://doi.org/10.1016/j.cell.2006.05.046

    • Google Scholar
16. 1. Roberts AJ
    2. Kon T
    3. Knight PJ
    4. Sutoh K
    5. Burgess SA

    (2013) Functions and mechanics of dynein motor proteins

    Nature reviews. Molecular cell biology 14:713–726.

    https://doi.org/10.1038/nrm3667

    • Google Scholar
17. 1. Stamer K
    2. Vogel R
    3. Thies E
    4. Mandelkow E
    5. Mandelkow EM

    (2002) Tau blocks traffic of organelles, neurofilaments, and APP vesicles in neurons and enhances oxidative stress

    The Journal of Cell Biology 156:1051–1063.

    https://doi.org/10.1083/jcb.200108057

    • Google Scholar
18. 1. Vale RD
    2. Toyoshima YY

    (1988) Rotation and translocation of microtubules in vitro induced by dyneins from Tetrahymena cilia

    Cell 52:459–469.

    https://doi.org/10.1016/S0092-8674(88)80038-2

    • Google Scholar
19. 1. Yajima J
    2. Cross RA

    (2005) A torque component in the kinesin-1 power stroke

    Nature Chemical Biology 1:338–341.

    https://doi.org/10.1038/nchembio740

    • Google Scholar
20. 1. Yajima J
    2. Mizutani K
    3. Nishizaka T

    (2008) A torque component present in mitotic kinesin Eg5 revealed by three-dimensional tracking

    Nature Structural & Molecular Biology 15:1119–1121.

    https://doi.org/10.1038/nsmb.1491

    • Google Scholar
21. 1. Yildiz A
    2. Tomishige M
    3. Gennerich A
    4. Vale RD

    (2008) Intramolecular strain coordinates kinesin stepping behavior along microtubules

    Cell 134:1030–1041.

    https://doi.org/10.1016/j.cell.2008.07.018

    • Google Scholar

Thank you for sending your work entitled “Bidirectional helical motion of cytoplasmic dynein around microtubules” for consideration at eLife. Your article has been favorably evaluated by Tony Hunter (Senior editor) and 3 reviewers, one of whom is a member of our Board of Reviewing Editors, and one of whom, Andrew Carter, has agreed to reveal his identity.

The Reviewing editor and the other reviewers discussed their comments before we reached this decision, and the Reviewing editor has assembled the following comments to help you prepare a revised submission.

In this study, the authors investigate the 3 dimensional motion of cytoplasmic dynein as it walks along a microtubule. In the cell, the nature of this movement has important implications for how cargo is transported and how motors navigate the crowded cytoplasmic environment. Kinesin-1 precisely follows a single protofilament track on a microtubule, while dynein has been shown to take frequent sideways steps to neighboring protofilaments. However, since most work has been done on surface-immobilized MTs, which constrains movement to two dimensions, little is known about dynein's 3-dimensional movement and whether it can generate torque. The surprise from this work by Can et al. is that the dynein bead shows long distance helical motion in both directions.

We would like in principle to publish your paper in eLife. However there are two sets of concerns that we would like addressed.

First, we feel that there needs to be more discussion of your work. You should include a discussion on the importance of the ability of dynein to generate torque in the cell. You bring up the importance of torque in the Introduction, but do not mention anything about it in the Discussion. A detailed discussion on the molecular basis of bi-directionality is called for. Since bidirectional helical motion is the major finding of the paper, this warrants a more thorough explanation of the mechanism. How could orientation of the motors lead to changes in bead step? How could the number of motors affect this motion? What, specifically, could lead monomeric and dimeric dynein to behave differently? A model figure would be very useful to explain this.

Second, there are a number of points that need to be tightened up before publication, with respect to the presentation of the data, and some of its implications. This is a list of points from one of the reviewers, with which I agree. Please consider these when revising your manuscript.

1) Most of the experiments are done with yeast dynein artificial dimers, which are reported to take a large number of processive steps before detaching from a microtubule (Reck-Peterson 2006). The authors have data both on the processive movement of beads powered by multiple yeast dynein artificial dimers and on the processive movement of beads powered by single yeast dynein artificial dimers. Antibodies are used to capture this dynein on to the beads and this defines the bead attachment geometry. Clarifying this at the outset is really important because the authors show that the behavior of multi-dynein dimer beads differs from that of single dynein dimer Qdots (Figure 2–figure supplement 1). I tend to think that these data belongs in the main Figure 2. The heading 'Helical movement of dynein dimer' should at the very least be dimerS, since the authors describe helical movement of beads carrying multiple dynein artificial dimers.

2) Kinesin-1 is used as a positive control to demonstrate that tracking of the long-pitch helical path of protofilaments in 14PF GMPCPP MTs is accurately reported. It would be helpful to know if beads in such experiments always moved unidirectionally along MT bridges, or whether some MT bridges contained multiple MTs with some potentially pointing in the opposite direction to others. The fluorescence signal from the MTs should also help confirm that only single MTs are in the bridges. If the kinesin velocities showed a narrow Gaussian distribution that can also usefully inform the dynein experiments.

3) Assuming [2] above does indicate 1 MT per bridge, in experiments detecting bidirectional rotational motion it is important to compare the velocities of dynein-driven bead motility in the two directions, and not just the helical pitches. In Figure 2, “speeds of left handed and right handed movements were statistically indistinguishable”;does this refer to speeds in the X direction only or along the helical paths? The pitches of the helical paths are statistically different. It is also really important to clarify that in this 'bidirectional' motion, the direction of net progress along the MT axis is not changing: both left handed and right handed helical paths have the motor heading in the same direction along the MT axis.

4) Regarding the brightfield method for tracking the Z position of the beads, I have not seen brightfield used to do this before, and have slight concerns because the bright-dark appearance is highly nonlinear with Z (Figure 1D) so that the bead goes briefly invisible as it flips between black and white quite suddenly in the videos and then remains black or white until its next flip. This needs more careful explanation, considering it looks like the brightfield signal may be insensitive to Z in parts of the trajectories. Could the flat tops of the trajectories in Figure 2A & B reflect this? The side to side (Y) tracking should not have this problem. Please clarify that the side-to-side tracking is using centroids and that both black and white images are included. Figure 2 does not currently make this clear. It might be better to plot X versus time and Y versus time and Z versus time one underneath the other for each sequence. This would be especially useful for Figure 2D, where it is hard to see the change in direction / handedness. The bw signal from the bead could be coded into these plots.

5) Figure 2–figure supplement 1: 'The oscillation of the MT bridges': please clarify what this is. Is the MT bridge moving side to side as the Qdot squeezes underneath or is this just random positional noise of the MT bridge? “Expected amplitude of rotation is 60nm”: Please clarify, why? Is this single dynein dimer in fact rotating around the MT or is it just drifting randomly side to side? What is the observed run length distribution? In the example, a single motor moves more than 2 µm, implying >250 8 nm steps. The velocity of multi-dynein dimer driven beads is 61 nm sec-1 (clarify that this is along the MT axis), what is the velocity of these single dimer Qdots?

6) Finally there is a concern that cytoplasmic dynein in vivo is not thought to rotate its cargo around the MT axis. The experiment reported here clearly shows that dynein monomers have an intrinsic tendency to step off-axis, and that yeast artificial dynein dimers retain this tendency. I was left wondering whether authentic cytoplasmic dynein dimers might be better at tracking the MT axis, just as kinesin-1 dimers can over-ride the intrinsic tendency of kinesin monomers to drift off axis. The discussion concludes firmly that (yeast) dynein rotates its cargoes around the MT with a pitch of about 600nm. It is worth pointing out that blocking this rotation, as it is implied happens when MTs are bound to the coverslip surface, does not impede axial motion.

First, we feel that there needs to be more discussion of your work. You should include a discussion on the importance of the ability of dynein to generate torque in the cell. You bring up the importance of torque in the Introduction, but do not mention anything about it in the Discussion.

We added a new paragraph, starting “Inside cells, the MT surface is crowded with associated proteins, which act as roadblocks during the transport of intracellular cargos (Stamer et al., 2002)...”

A detailed discussion on the molecular basis of bi-directionality is called for. Since bidirectional helical motion is the major finding of the paper, this warrants a more thorough explanation of the mechanism. How could orientation of the motors lead to changes in bead step? How could the number of motors affect this motion?

We added a paragraph starting “In contrast to the other motors studied to date, dynein can move along both left- and right-handed helical paths. The molecular basis of the switches in helical directionality remains unclear…”

What, specifically, could lead monomeric and dimeric dynein to behave differently?

In this revised manuscript, we tested whether the differences between monomeric and dimeric dyneins could be due to the GST-dimerization, which may orient the heads differently. To test this idea, we repeated the MT bridge assay using full-length yeast cytoplasmic dynein. We observed that full-length dynein also shows bidirectional rotation and reversals in helical directionality during a processive run of a bead, similar to GST-truncated dynein. Unlike GST-dimerized truncated dynein, full-length dyneins prefer to walk along a right-handed helical path. As a result, we removed the statement that monomeric and dimeric dyneins behave differently. We also added the following paragraph to the Discussion, starting “In this study, we showed that cytoplasmic dynein moves along MTs in a helical trajectory with a pitch of ∼500 nm. Single dynein dimers move along microtubules by taking frequent sideways steps (Reck-Peterson et al., 2006), whereas multiple dynein dimers persistently move in a helical pattern with a net preference for a right-handed rotation…”

A model figure would be very useful to explain this.

We have added the paragraphs above to discussion and added a new model figure, Figure 5, to the revised manuscript.

Second, there are a number of points that need to be tightened up before publication, with respect to the presentation of the data, and some of its implications. This is a list of points from one of the reviewers, with which I agree. Please consider these when revising your manuscript.

1) Most of the experiments are done with yeast dynein artificial dimers, which are reported to take a large number of processive steps before detaching from a microtubule (Reck-Peterson 2006). The authors have data both on the processive movement of beads powered by multiple yeast dynein artificial dimers and on the processive movement of beads powered by single yeast dynein artificial dimers. Antibodies are used to capture this dynein on to the beads and this defines the bead attachment geometry. Clarifying this at the outset is really important…

We modified the corresponding section to:

“We next investigated the helical motion of a tail-truncated yeast cytoplasmic dynein artificially dimerized with glutathione S-transferase (GST-Dyn_331kD), which has similar motile properties to full-length dynein (Reck-Peterson et al., 2006). The motors were tagged with GFP at the N-terminus and attached to cargo beads coated with anti-GFP antibody (see Methods).”

The authors show that the behavior of multi-dynein dimer beads differs from that of single dynein dimer Qdots (Figure 2–figure supplement 1). I tend to think that these data belongs in the main Figure 2.

We were not able to fit these data to Figure 2 (all panels and the legend have to fit into a single page). Instead, we put it as new Figure 3.

The heading 'Helical movement of dynein dimer' should at the very least be dimerS, since the authors describe helical movement of beads carrying multiple dynein artificial dimers.

This is fixed.

2) Kinesin-1 is used as a positive control to demonstrate that tracking of the long-pitch helical path of protofilaments in 14PF GMPCPP MTs is accurately reported. It would be helpful to know if beads in such experiments always moved unidirectionally along MT bridges, or whether some MT bridges contained multiple MTs with some potentially pointing in the opposite direction to others. The fluorescence signal from the MTs should also help confirm that only single MTs are in the bridges.

Motor-coated beads always moved unidirectionally along the microtubules, without changing the direction. This excludes the possibility of bridges containing multiple MTs pointing in opposite directions. We kept the bead concentration high and the MT concentration low such that there are many beads sticking to the surface, but very few bridges forming on the entire coverslip surface. Therefore, it is highly unlikely that multiple MTs form single bridges. We also carefully examined each bridge by fluorescence in order to ensure that the bridges are straight, and the bead pair contains a single MT bridge in between them. The average distance between the beads on the sample is around 15 µm, and there are 4 beads on average observed in 20 µm x 20 µm area. At this density of beads, we observe less than 20 bridges in 10⁶ µm².

Finally, we use the same bridge for testing the movement of many beads. MT polarity is first tested by placing the bead on a MT at the center of the bridge. Once the directionality is determined, the bead is moved away from the MT and placed at the plus end tip of the bridge to explore the motility throughout the entire length of the bridge. ALL of the beads moved towards the same direction on a single bridge, without any reversals.

These statements are added to the Methods.

If the kinesin velocities showed a narrow Gaussian distribution that can also usefully inform the dynein experiments.

The average velocity of kinesin molecules is 541 ± 73 nm/s (mean ± SD, N=10). This comparable to the speed of single motors of the same truncated version of human kinesin-1 (hK560-GFP)(Yildiz et al., Cell 2008). This is added to the text.

3) Assuming [2] above does indicate 1 MT per bridge, in experiments detecting bidirectional rotational motion it is important to compare the velocities of dynein-driven bead motility in the two directions, and not just the helical pitches. In Figure 2, “speeds of left handed and right handed movements were statistically indistinguishable”;does this refer to speeds in the X direction only or along the helical paths? The pitches of the helical paths are statistically different.

The speeds of the beads are calculated along the helical pitch.

Pitch corrected velocities for dimers (mean ± SEM):

Left Handed Rotations: 79 ± 4.7 nm/s

Right Handed Rotations: 89 ± 17 nm/s

One tail t-test between left and right velocities p=0.28 (statistically indistinguishable)

Pitch corrected velocities for monomers (mean ± SEM):

Left Handed Rotations: 66 ± 10 nm/s

Right Handed Rotations: 57 ± 6 nm/s

One tail t-test between left and right velocities p=0.18 (statistically indistinguishable)

One tail t-test between dimer vs monomer p=0.004 (statistically distinguishable)

These numbers are corrected in text and the issue is now clarified.

It is also really important to clarify that in this 'bidirectional' motion, the direction of net progress along the MT axis is not changing: both left handed and right handed helical paths have the motor heading in the same direction along the MT axis.

Both left and right handed rotating dynein molecules moved in a unidirectional manner. While we observed change in the helicity of movement during the processive runs of the beads, we never observed a change in the directionality along the MT long axis. We clarified this in the text.

4) Regarding the brightfield method for tracking the Z position of the beads, I have not seen brightfield used to do this before…

Brightfield imaging has been used to determine the Z position of the bead in previous studies. Most recently, the Okten lab used this method to track kinesin-1-driven beads in three dimensions (Brunnbauer et al. 2012, cited in our manuscript).

I have slight concerns because the bright-dark appearance is highly nonlinear with Z (Figure 1D)…

The bright-dark appearance of the bead is nearly linear ± 150 nm and becomes highly nonlinear between 150 and 250 nm. We fit our calibration curve to a third order polynomial to convert bead signal to the z position ($R^2=0.9982$). The calibration curve (Figure 1D, updated) and description for how we extrapolated the z position from this curve are added to the manuscript.

The bead goes briefly invisible as it flips between black and white quite suddenly in the videos and then remains black or white until its next flip. This needs more careful explanation.

The bead never becomes fully invisible to the eye. At z = 0 position, the bead does not become invisible (see Figure 1C), instead it contains a dark ring near its center. In our imaging conditions, we cannot reliably detect the bead center when the z position is between -25 nm and +25 nm. This is now added to the methods. However, this does not affect our ability to observe rotation of the beads around MTs.

It looks like the brightfield signal may be insensitive to Z in parts of the trajectories. Could the flat tops of the trajectories in Figure 2A & B reflect this? The side to side (Y) tracking should not have this problem.

The reviewers are correct. As we explain in Figure 1, the brightfield signal is sensitive to z within ±250 nm of the focal plane. If the bead moves higher or lower than this limit, we accept these positions as +250 nm and -250 nm, respectively. Traces in Figure 2A&B appear flat at the top of the helical turn, because the bead is beyond the detectable range in the z-axis. The primary purpose of the z measurement is to determine the handedness of the helical trajectory, and this is still clear from the data presented in Figures 2.

Please clarify that the side-to-side tracking is using centroids and that both black and white images are included. Figure 2 does not currently make this clear.

Side to side tracking is done by using 2D Gaussian algorithm (plus a constant term for the background) and the both white and black images are included. After beads are tracked in 2 dimensions, the intensities of corresponding center positions are obtained from original recordings for Z tracking. We have also performed centroid calculations to verify that our conclusions are not an artefact of the Gaussian fitting algorithm. The figure below shows that centroid and Gaussian algorithms have slight disagreements in the xy position, but the rotational pitch of the bid does not change (Author response image 1).

Author response image 1

Download asset Open asset

It might be better to plot X versus time and Y versus time and Z versus time one underneath the other for each sequence. This would be especially useful for Figure 2D, where it is hard to see the change in direction / handedness. The bw signal from the bead could be coded into these plots.

In the revised version, we show example traces with x, y and z plotted separately as a function of time, and encoded the bead image in Figure 2–figure supplement 2.

5) Figure 2–figure supplement 1: 'The oscillation of the MT bridges': please clarify what this is. Is the MT bridge moving side to side as the Qdot squeezes underneath or is this just random positional noise of the MT bridge?

The persistence length of MT’s is 5.2 mm (Gittes et al., 1993) which is much longer than the length of our MT bridges. Therefore, MTs remain straight between the bridges and oscillate due to thermal noise. To measure amplitude of this random oscillation, MTs were sparsely labeled with a quantum dot 585 using bition-streptevidin linkage. The standard deviation of the quantum dot position was 35 nm in perpendicular direction and 17 nm in parallel direction to the long axis of the MTs. These errors are much higher than the localization error from the Gaussian tracking. Therefore, we interpreted that large fluctuations of fixed positions at MT bridges are due to the thermal fluctuations. We included a statement to this effect in the figure legend.

“Expected amplitude of rotation is 60nm”: Please clarify, why? Is this single dynein dimer in fact rotating around the MT or is it just drifting randomly side to side?

We modified this distance to 50 nm. The distance corresponds to the estimated amplitude of rotation for a quantum-dot labeled dynein dimer rotating around the MT. Given the MT radius (12 nm), the estimated distance between dynein MTBD and the GST (∼25 nm) and the radius of the quantum dots (∼12 nm), the expected radius of rotation is ∼50 nm. Compared to the beads carried by multiple dynein motors, single motors show higher variability and more frequent switches in sideways movement. So the single dynein dimer is not rotating as uniformly as cargo bead carried by multiple dyneins around the MT.

What is the observed run length distribution? In the example, a single motor moves more than 2 µm, implying >250 8 nm steps. The velocity of multi-dynein dimer driven beads is 61 nm sec-1 (clarify that this is along the MT axis), what is the velocity of these single dimer Qdots?

61 nm/s is the speed of the monomers. We fixed this error.

The run lengths of single dynein motors on MT bridges is 1.5 ± 0.7 µm (mean ± SD, N= 10), which is similar to that measured on surface-immobilized MTs (Reck-Peterson et al. Cell 2006). Velocity of single dynein dimers is 64 ± 8 nm/s (mean ± SEM , N = 10), similar (t-test p= 0.12) to that of cargo beads carried by multiple dynein motors (80 ± 13 nm/s, N = 15).

6) Finally there is a concern that cytoplasmic dynein in vivo is not thought to rotate its cargo around the MT axis. The experiment reported here clearly shows that dynein monomers have an intrinsic tendency to step off-axis, and that yeast artificial dynein dimers retain this tendency. I was left wondering whether authentic cytoplasmic dynein dimers might be better at tracking the MT axis, just as kinesin-1 dimers can over-ride the intrinsic tendency of kinesin monomers to drift off axis. The discussion concludes firmly that (yeast) dynein rotates its cargoes around the MT with a pitch of about 600nm. It is worth pointing out that blocking this rotation, as it is implied happens when MTs are bound to the coverslip surface, does not impede axial motion.

Although the reviewers did not strictly require this experiment, we tested the motility of cargo beads driven by full length yeast cytoplasmic dynein. We added the following paragraph, new Video 4 and two supplemental figures to the revised manuscript. The reason why we did not place this data in the main text is because we are not able to express the full-length protein at a high yield. When overexpressed, protein aggregation sharply reduces yield and protein quality. Under a native promoter, the protein behaves well, but the yield is comparatively poor, making it difficult to efficiently coat the cargo beads with dynein. As a result the 3D traces do not appear as high quality as that of GST-Dyn_331kDa, because the smoothness of rotations requires more than a few motors on a bead. If the reviewer panel feels that the full length data need to be moved to the main text, we would be happy to do so.

“We next investigated the helical motility of a full-length yeast cytoplasmic dynein (Reck-Peterson et al., 2006). The motors were fused to GFP at the N-terminus and attached to cargo beads coated with anti-GFP antibodies. Dynein-coated beads moved in helical trajectories with a pitch of 500 ± 36 nm (mean ± SEM, Figure 2–figure supplement 3, Video 4) and exhibited both left- and right-handed helical movement. The speeds of left- and right-handed movements along the helical path (42 ± 11 nm/s and 43 ± 10 nm/s, mean ± SEM, respectively) were statistically indistinguishable (t-test, p = 0.21). Similar to GST-Dyn_331kD, beads driven by full-length dynein occasionally switch direction during a run (8 out of 33 rotations, Figure 2–figure supplement 4). Unlike GST-Dyn_331kD, which has a net preference for left-handed helical movement (75%), full-length dynein has a net preference for right-handed helical movement (58%, t-test, p = 10^-5). This difference may be related to GST dimerization, in which the heads are oriented differently relative to the MT surface.”

Author details

1. Sinan Can

   Department of Physics, University of California, Berkeley, Berkeley, United States

   Contribution

   SC, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

2. Mark A Dewitt

   Biophysics Graduate Group, University of California, Berkeley, Berkeley, United States

   Contribution

   MAD, Conception and design, Analysis and interpretation of data

   Competing interests

   The authors declare that no competing interests exist.

3. Ahmet Yildiz

   1. Department of Physics, University of California, Berkeley, Berkeley, United States
   2. Biophysics Graduate Group, University of California, Berkeley, Berkeley, United States

   Contribution

   AY, Conception and design, Drafting or revising the article

   For correspondence

   yildiz@berkeley.edu

   Competing interests

   The authors declare that no competing interests exist.

• 2,940

  Page views

• 251

  Downloads

• 49

  Citations

Article citation count generated by polling the highest count across the following sources: Scopus, Crossref, PubMed Central.