Vanderbilt Institute of Chemical Biology



Discovery at the VICB







Exploring the Mechanism of Microtubule-Directed Motion



By: Carol A. Rouzer, VICB Communications
Published:  July 15, 2015



The kinesin-8 Kif18B exhibits both directed motion and diffusion in its trip along the microtubule, a pattern of motility with implications for its role in the mitotic spindle.


Kinesins and dyneins are motor proteins that convert the energy from ATP hydrolysis to mechanical energy for the transport of molecules along microtubule scaffolds. Kinesins comprise fourteen structurally distinct families, all of which are distinguished by the presence of a kinesin motor domain. In the majority of families, this domain is present at the N-terminus of the protein, and all of these kinesins move along the microtubule in the direction of the rapidly growing plus-end. Kinesins in which the motor domain is C-terminal move in the opposite direction, and a limited number of kinesins, which contain the motor domain in their center, are involved in regulation of microtubule dynamics. The kinesin-8 family of proteins comprises homodimers of subunits characterized by an N-terminal motor domain connected to a C-terminal specialized domain via a coiled coil protein interaction domain (Figure 1). Depending on the species, these proteins regulate a number of important cellular processes, including spindle positioning, chromosome alignment, and the length of cilia. The most studied of the kinesin-8 family are the yeast Kip3 and human Kif18A motors. Both motors regulate microtubule plus-end dynamics, but do so in different ways: Kip3 causes subunit removal (i.e., disassembly), whereas Kif18A prevents both growth and depolymerization. Kif18B, a less well-studied kinesin-8, is also noted for its ability to limit microtubule length, acting particularly on astral microtubules of the mitotic spindle. Previous work has suggested that the ability of Kif18A and Kip3 to regulate microtubule length is due, at least in part, to their ultraprocessivity - that is, their ability to travel long distances along the microtubule in order to reach the ends. In contrast, Kif18B’s localization at the ends of astral microtubules requires the plus-end tracking protein EB1. Now, Vanderbilt Institute of Chemical Biology member Ryoma “Puck” Ohi and his collaborator Matthew Lang (Department of Chemical and Biomolecular Engineering) show that Kif18B uses a distinctly different motility mechanism from those of Kif18A and Kip3, demonstrating that not all kinesin-8 family members act identically [Y. Shin, et al. (2015) Proc. Natl. Acad. Sci. U.S.A., published online July 6, DOI:10.1073/pnas.1500272112].



Figure 1.
Diagrammatic representation of the structure of kinesin-8 proteins.The proteins are homodimers of subunits comprising an N-terminal motor domain (light green) joined by a coiled coil (dark green) to a specialized C-terminal domain (blue). Figure reproduced by permission from Macmillan Publishers, Ltd. from K. J. Verhey & J. W. Hammond, (2001) Nat. Rev. Mol. Cell Biol., 10, 765. Copyright 2009.


To study the mechanism of Kif18B’s motility along the microtubule, the Ohi and Lang laboratories expressed and purified green fluorescent protein-tagged full-length Kif18B (GFP-Kif18B-FL). As expected, the protein was a homodimer in solution and acted as a plus-end-directed microtubule motor in a standard microtubule gliding assay (Figure 2). Having confirmed that GFP-Kif18B-FL retained the expected structural characteristics and activity of Kif18B, the investigators went on to assess its microtubule-dependent motor activity. They used total internal reflection fluorescence microscopy (TIRF) to observe the movement of GFP-Kif18B-FL along microtubules that had been fixed to the surface of a cover slip (Figure 3). The results showed that GFP-Kif18B-FL is only modestly processive, exhibiting a 10-fold reduction in processivity when compared to Kip3 and Kif18A. Consequently, relatively few of the GFP-Kif18B-FL molecules reached the plus-ends of the microtubules, and those that did remained associated with the microtubules for only a brief time, exhibiting a >50-fold reduction in dwell time as compared to Kip3 or Kif18A. These latter findings are consistent with prior evidence that Kif18B requires EB1 to target microtubule plus-ends in the mitotic spindle.

Figure 2. 
Microtubule-dependent motor assay. GFP-Kif18B-FL was fixed to the walls of a flow cell. Then fluorescent microtubules, labeled so that the minus-end is brighter than the plus-end, were added. Fluorescence microscopy captured the movement of the microtubules in the direction of the minus-end, as would be expected with a plus-end-directed motor. The asterisk indicates the location of a fixed point on the flow cell. Figure reproduced by permission Y. Shin, et al. (2015) Proc. Natl. Acad. Sci. U.S.A., published online July 6, DOI:10.1073/pnas.1500272112. Copyright 2015, Y. Shin, et al.

Figure 3.  (A) Assay for KIF18B movement along a microtubule. Microtubules were fixed to a coverslip via a biotin/streptavidin linker, and the GFP-Kif18B was added with or without ATP. (B) The movement of a GFP-Kif18B (green) along a microtubule (red) in the presence of ATP as monitored by TIRF microscopy. The numbers indicate elapsed time in seconds. Figure reproduced by permission Y. Shin, et al. (2015) Proc. Natl. Acad. Sci. U.S.A., published online July 6, DOI:10.1073/pnas.1500272112. Copyright 2015, Y. Shin, et al.



The investigators were intrigued by the observation that GFP-Kif18B-FL molecules that reached the microtubule plus-ends exhibited brief back and forth movements before dissociating. As the fluorescence-based assay provided limited resolution for observing these movements, the researchers devised an assay that used GFP-Kif18B-FL attached to a 550 nm polystyrene bead. This approach enabled them to image the protein’s motion with 6 nm resolution using differential interference contrast (DIC) microscopy (Figure 4). They observed that the overall motion of GFP-Kif18B-FL in the presence of ATP occurred in the direction of the microtubule plus-end, but it was punctuated by frequent backward motions, suggestive of diffusion. A mean-squared displacement analysis showed the data fit a model that incorporated both directed motion and diffusion, whereas the motion of the motor in the presence of ADP instead of ATP was consistent with diffusion alone. A mathematical analysis of the motility as monitored by high resolution video tracking indicated that the protein spends about 72% of its time in diffusion, with a diffusion coefficient of 0.016 μm2/s, and the remaining time in directed motion with a velocity of 0.183 μm/s.


Figure 4.
  High resolution assay for monitoring the motion of GFP-Kif18B along a microtubule using DIC microscopy. Figure reproduced by permission Y. Shin, et al. (2015) Proc. Natl. Acad. Sci. U.S.A., published online July 6, DOI:10.1073/pnas.1500272112. Copyright 2015, Y. Shin, et al.



The high processivity of Kip3 and Kif18A has been attributed to their C-terminal tail, which provides a secondary microtubule binding site. Kif18B also has a microtubule binding site in its C-terminal domain, leading the investigators to hypothesize that it may play a role in the diffusion behavior of the protein. They considered two possible mechanisms by which this could occur. The first proposed that all motility is directed by the kinesin motor domain in the protein’s head, but an interaction between the head and tail of the protein results in a switch between directed motion and diffusion. The second suggested that interaction of the protein head with the microtubule results in directed motion, while the interaction of the protein tail leads to diffusion (Figure 5). To explore these possibilities, they expressed a GFP-linked truncated protein (GFP-Kif18B-TL) lacking a tail, and a tail-only domain linked to the fluorescent mCherry protein (mCh-Kif18B-T). Using these proteins in their assays revealed that GFP-Kif18B-TL exhibits only unidirectional motion with a velocity of 0.166 μm/s in the presence of ATP. In contrast, mCh-Kif18B-T exhibits only diffusion, with a much higher diffusion coefficient (0.676 μm2/s) than is exhibited by either GFP-Kif18B-FL (0.015 μm2/s) or GFP-Kif18B-TL (0.014 μm2/s) in the presence of ADP. These results suggest that the tail is required to observe diffusion in the presence of ATP, but that the head of the motor limits the rate of diffusion. More detailed kinetics studies revealed that the tail reduces the off rate of Kif18B by approximately 5-fold and increases its processivity by 1.7-fold, confirming its importance in regulating the interaction of Kif18B with the microtubule. However, the tail-bound state is present for <5% of the time. This led the investigators to conclude that the head of the protein is responsible for both diffusion and directed motion while the role of the tail is to regulate the motility mode of the head.




Figure 5.  Possible mechanisms by which the C-terminal tail of Kif18B can regulate diffusion (Left) An interaction between the head and the tail converts the motility of the head from directed motion to diffusion. (Right) The head of the protein is responsible for directed motion, while the tail is responsible for diffusion. Figure reproduced by permission Y. Shin, et al. (2015) Proc. Natl. Acad. Sci. U.S.A., published online July 6, DOI:10.1073/pnas.1500272112. Copyright 2015, Y. Shin, et al.



From their findings, the investigators concluded that the motility of Kif18B is distinctly different from that of Kif18A and other previously characterized kinesin-8 family members. It is interesting to note that Kif18A concentrates at the kinetochores of mitotic spindles, a function consistent with its high processivity that allows it to reach the ends of long microtubules. In contrast, Kif18B is found mostly at the ends of spindle microtubules overlapping with EB1. To more thoroughly evaluate the interaction of Kif18B with EB1, the investigators carried out high resolution structured illumination microscopy of the localization of both proteins in the spindles of mitotic HeLa cells. The results (Figure 6) revealed that, as expected, most of the Kif18B protein colocalizes with EB1. However, with close inspection, the investigators also discovered a second class of Kif18B molecules located at the very tip of the microtubules, displaced from EB1 by approximately 24 microtubule subunits. These findings are intriguing in light of Kif18B’s purported role in preventing astral overgrowth. However, the investigators note that, although the data clearly indicate that Kif18B is present at the extreme end of spindle microtubules, it is not clear how it gets there. The overall velocity of Kif18B along the microtubule is much slower than the rate of microtubule polymerization. This, coupled with its poor processivity, suggests that Kif18B does not reach the plus-end of the microtubule by solely moving along its length. Alternatively, Kif18B’s ability to diffuse along the microtubule may allow it to capture the plus-end of a growing microtubule. Further work is obviously needed to explore these various possibilities and the role of EB1. However, these findings represent an important step forward in our understanding of the fundamental process of intracellular motility.


Figure 6.  High resolution structural illumination microscopy reveals the location of EB1 and Kif18B in the mitotic spindles of HeLa cells. Merging of the images reveals two classes of Kif18B proteins. Class I proteins overlap with EB1. Class II proteins are displaced from EB1, localizing to the very tip of the spindle fiber. Figure reproduced by permission Y. Shin, et al. (2015) Proc. Natl. Acad. Sci. U.S.A., published online July 6, DOI:10.1073/pnas.1500272112. Copyright 2015, Y. Shin, et al.









The Vanderbilt Institute of Chemical Biology, 896 Preston Building, Nashville, TN 37232-6304, phone 866.303 VICB (8422), fax 615 936 3884
Vanderbilt University is committed to principles of equal opportunity and affirmative action. Copyright © 2014 by Vanderbilt University Medical Center