PerspectiveCell Biology

Microtubule Asymmetry

See allHide authors and affiliations

Science  27 Jun 2003:
Vol. 300, Issue 5628, pp. 2040-2041
DOI: 10.1126/science.1084938

Bacteria, yeast, and most mammalian cells exhibit polarity, that is, their proteins, lipids, and organelles are distributed asymmetrically between one end of the cell and the other. Establishing this asymmetry frequently depends on microtubules that respond to cues in the cell periphery. The prevailing search-and-capture model, elaborated by Kirschner and Mitchison 17 years ago (1), postulates that randomly assembled dynamic microtubules probe the cell cortex and are stabilized when they encounter localized cortical factors. However, a different picture is offered by three recent studies (24) in budding yeast that investigate the microtubule-dependent orientation of the spindle during the early stages of cell division (see the figure). These studies suggest that a cortical factor—previously thought to be positioned independently of microtubules—interacts with a specific microtubule organizing center before migrating to the tips of the associated microtubules (see the figure, steps 1 and 2). The newly decorated microtubule ends, rather than randomly probing the cortex, are selected when myosin V binds to the cortical factor and then “marches” the microtubule ends along polarized actin cables toward the cortex (see the figure, steps 3 and 4). This model might also be applicable to the microtubules of animal cells.

Guiding microtubules.

Polarization of selected microtubules ensures orientation of the spindle toward the yeast bud. Loading of the protein Kar9 onto the old SPB is mediated by Bim1 (step 1), whereas loading onto the new SPB is inhibited by Cdc28/Clb4. Kar9 moves out to the end of the microtubule (step 2) where it interacts with Myo2, a type of myosin V molecular motor (step 3). Myo2 then moves the Kar9-associated microtubule along polarized actin cables toward the bud tip where it may be anchored by additional factors (step 4). [Adapted from (24)]


In budding yeast, the axis of cell division is defined by the site of bud emergence, which is chosen early in the G1 phase of the cell cycle. Cells entering the cell cycle have a single microtubule organizing center called the spindle pole body (SPB), which is embedded in the nuclear envelope from which cytoplasmic microtubules emerge. The SPB duplicates itself and separates, giving rise to a second set of cytoplasmic microtubules and the intranuclear spindle. The initial step in spindle orientation occurs when microtubules from the “old” SPB are preferentially selected to become attached to the bud neck or bud tip (5). This process requires the cortical factor known as Kar9 (6). All three of the new studies use high-resolution imaging techniques combined with genetic analysis to address how this initial orientation is achieved.

The first step in this pathway is the asymmetric association of Kar9 with the old SPB in a manner that depends on binding to the Bim1 protein, but is independent of microtubules (2). Bim1, however, is found associated with both SPBs and thus cannot explain the asymmetric distribution of Kar9. Association of Kar9 with just one SPB requires activation of Cdc28, the only cyclin-dependent kinase in yeast, by a cyclin (Clb3 or 4), which itself is specifically associated with the newly assembled SPB (2). Cdc28-dependent phosphorylation of Kar9 (2, 3) inhibits its interaction with Bim1, which would explain why it only associates with the old SPB. In support of this proposal, when the Cdc28 phosphorylation sites of Kar9 are mutated, Kar9 is loaded onto both SPBs (2).

Remarkably, after loading on the SPB, Kar9 migrates down the associated microtubules to the microtubule tips (2, 3) (see the figure, step 2). Photobleaching experiments strongly support specific loading of Kar9 at the SPB, rather than further along the microtubules (2). Staining with Cdc28-green fluorescent protein revealed that some Cdc28 is also transported down the microtubules, and both transport processes seem to depend on Kip2, a plus-end microtubule motor (3). Interestingly, efficient transport may require the association of Cdc28 with a different cyclin, Clb5. The function of this microtubule-associated Cdc28 is not yet understood.

How are microtubules that become marked with Kar9 oriented toward the bud? Previous studies have implicated Myo2 (7, 8), a myosin V molecular motor that moves cellular cargoes toward the bud on polarized actin cables and whose tail binds to Kar9 (8). With the new evidence for SPB loading, two potential models emerge. Either microtubules tipped with Kar9 randomly probe the cortex to be captured by polarized Myo2 (7), or Myo2 actively transports microtubule ends to the bud neck and bud tip (8).

Barral and colleagues (2) distinguish between these models by watching the orientation of microtubules emerging from the SPB. Within half a minute, 93% of the microtubules emerging from the Kar9-loaded pole reorient themselves toward the bud, whereas only 9% of the microtubules from the other pole become reoriented, providing strong evidence for directed transport. This rapid reorientation was not seen in yeast either lacking Kar9 or expressing a Myo2 tail mutant that cannot bind to Kar9. The presence of some Myo2 at microtubule ends supports this model (4). Perhaps most persuasively, Hwang et al. (4) show that the rate of movement of microtubule ends is greatly reduced in a mutant where Myo2 translocates much more slowly. What happens when Kar9 with mutated phosphorylation sites associates with both SPBs and therefore with both sets of cytoplasmic microtubules? Both SPBs repeatedly attempt to align themselves along the mother-bud axis, with the net result that the spindle “dances” as each SPB tries to move toward the bud. Together, these data support the notion (8) of directed transport of microtubule ends by Myo2 along polarized actin cables (see the figure, steps 3 and 4).

So, it now appears that microtubules are led to the site of polarization like a “reluctant groom,” rather than finding the site without help like an enlightened one. Does this behavior apply to microtubules in animal cells, where distances and microtubule dynamics are greater and there is no single attachment point (like the yeast bud) for the actin cytoskeleton? Indeed, in fibroblasts, microtubules are targeted to focal adhesions probably by associated actin stress fibers (9). Also, there is a striking parallel between the behavior of Kar9 and that of the animal APC (adenomatous polyposis coli) protein. It has been suggested that APC is the functional equivalent of Kar9 based on its binding to EB1 (the mammalian Bim1) and on limited sequence homology. APC moves to the ends of microtubules in a kinesin-dependent manner (10) and may participate in reorientation of the microtubule organizing center (11).

The yeast studies raise a number of interesting questions. Although Cdc28 is involved in the asymmetric loading of Kar9, how do its associated cyclins become asymmetrically loaded? Also, how are the oriented microtubules ultimately anchored in the cell cortex? One possibility is the formin protein family. In yeast, formins participate in Kar9-mediated microtubule capture by nucleating the assembly of actin cables (12, 13). But in mammalian cells, formins regulate selective microtubule stabilization, binding to microtubules both in vivo and in vitro (14). Such cortical anchor proteins may also regulate the activity of captured microtubules to give controlled shrinkage in yeast or capping of stabilized microtubules in mammalian cells.

The new studies shift our view of how microtubule dynamics contribute to generating cellular asymmetries. They point to the guiding of microtubules by actin microfilaments through factors that initially interact separately with the two elements. The dynamics of microtubules are still likely to be important in this new view, especially because the myosin “target” is one that moves along actin cables. It will be interesting to test this model in other systems, such as migrating cells, neuronal growth cones, or polarized epithelia, where actin microfilaments and microtubules collaborate to effect cellular asymmetries.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.

Navigate This Article