Cytokinesis is a broad topic. We are particularly interested in how the microtubules control the delivery of the “cytokinetic signal” which leads to activation of Rho and how the cell balances Rho activation and inactivation during cytokinesis. Each of these topics is considered in turn below.
Cytokinesis culminates the cell cycle by physically splitting the cell into two parts. In most animal cells, and many other eukaryotic cells, the splitting is powered by the so-called cytokinetic apparatus, which assembles at the equatorial plasma membrane (Cytokinesis figure 1). The cytokinetic apparatus is a complex structure that contains actin filaments and myosin-2. It is thought that the actin filaments and myosin-2 work together to contract, thereby driving inward closure of the cytokinetic apparatus and thus cytokinesis.
Cytokinesis figure 1—A schematic diagram representing the process of cytokinesis. The cytokinetic apparatus (green ring) contains actin filaments and myosin-2 and is thus able to contract. As it contracts, it pulls the plasma membrane (yellow oval) inward with it.
Two examples of the process of cytokinesis in living embryos of the purple sea urchin are shown in Cytokinesis movie 1 and Cytokinesis movie 2. These movies were made by George von Dassow (Oregon Institute of Marine Biology) using a probe developed by Brian Burkel, a former graduate student in the lab. The probe binds to actin filaments, and reveals their pattern of changing localization during cytokinesis. For more details see the paper published in 2007 in Cell Motility and the Cytoskeleton (http://onlinelibrary.wiley.com/doi/10.1002/cm.20226/abstract;jsessionid=…).
Cytokinesis movie 1—A fluorescence, time-lapse movie showing actin filaments in a dividing sea urchin embryo. The actin filaments accumulate in the cytokinetic apparatus at the cell equator and the apparatus then closes inward, dividing the cell.
Rho activity “zones” and cytokinesis
What triggers the formation of cytokinetic apparatus? One essential trigger is the small GTPase Rho (which is also involved in wound healing). Using a probe developed by Helene Benink (a former graduate student) to specifically detect the active (GTP-bound) form of Rho, we found that formation of the cytokinetic apparatus is initiated by the activation of Rho in a precisely defined “zone”, just as Rho is activated in a discrete zone during wound healing (see the wound healing section of the “research interests” section on this site). Examples of the cytokinetic Rho zone in sea urchin and Xenopus embryos are presented in Cytokinesis movie 3 and 4, respectively. See the 2005 paper in the Journal of Cell Biology for more details (http://jcb.rupress.org/content/170/1/91.long).
Cytokinesis movie 3—A fluorescence, time-lapse movie showing active Rho in a purple sea urchin embryo. Rho is activated in a precisely bounded “zone” at the cell equator; the cytokinetic apparatus subsequently develops in the same region and then closes inward.
Cytokinesis movie 4—A fluorescence, time-lapse movie showing active Rho in a Xenopus embryo. As in the sea urchin embryo, Rho is activated in a precisely bounded zone at the cell equator; the cytokinetic apparatus subsequently develops in the same region and then closes inward.
The cytokinetic Rho zones at least superficially resemble those discovered by Helene during single cell wound repair (see “Wound Healing”). It turns out that the resemblence gets even more stiking when one considers polar body emission. Polar body emission is an extremely asymmetric form of cytokinesis which takes place during oocyte meiosis. We first found that a Rho zone forms in concert with polar body emission, but it takes the form of a complete ring at the top (animal pole) of the oocyte (see http://jcb.rupress.org/content/170/1/91.long). Then, in a collaboration with Johne Liu (University of Ottawa) we found that a circular zone of Cdc42 activity forms inside the Rho zone during polar body emission (Cytokinesis figure 2; for details see the paper in Current Biology: http://www.sciencedirect.com/science/article/pii/S0960982205015174). This is fascinating in that it resembles the complementary Rho and Cdc42 zones observed during wound repair, but with the opposite organization (Rho zone around Cdc42 zone in polar bodyies; Cdc42 zone around Rho zone during wound healing). Johne’s lab, in collaboration with us, went on to show that each zone plays a different role in polar body emission, just as they do during wound healing. For details of the roles they play in polar body emission, see the paper published in Developmental Cell (http://www.ncbi.nlm.nih.gov/pubmed/18804436).
Cytokinesis figure 2—Rho and Cdc42 activity zones associated with polar body emission. The top row shows active Rho (red) and active Cdc42 (green) as seen when viewed from the outside of the cell looking down on the forming polar body. The Rho zone is ring-like and surrounds a circular zone of Cdc42. The bottom row shows the same process in a cross sectional (side) view. The Rho zone dives inward and pinches in, while the Cdc42 zone spreads over the surface of the forming polar body.
Microtubule-dependent control of the cytokinetic signal
How do cells control Rho activation at the equator? Because the mitotic spindle is needed to direct formation of the cytokinetic apparatus, and because the “astral” microtubules of the spindle extend to the cell surface, it has long been thought that signals are transported from the spindle to the cell surface along microtubules (Cytokinesis figure 3).
Cytokinesis figure 3—A schematic diagram representing control of cytokinetic apparatus formation by the mitotic spindle. The spindle is made of microtubules (white lines), chromosomes (purple rectangles) and centrosomes (red ovals) and is in the approximate middle of the cell. Signals are thought to travel from the middle of the spindle along astral microtubules that extend from the centrosomes to the cell periphery. Once at the periphery, the signals direct formation of the cytokinetic apparatus, apparently by promoting formation of a local “zone” of Rho activity.
Both positive and negative signals have been proposed to explain various features of cytokinesis, such that the positive signals are envisioned to undergo transport to the equator (to promote cytokinetic apparatus assembly there via Rho activation) and negative signals are envisioned to undergo transport to regions outside the equator (to prevent cytokinetic apparatus assembly there via inhibition of Rho (Cytokinesis figure 4).
Cytokinesis figure 4—A schematic diagram representing potential signals from the mitotic spindle that control of cytokinetic apparatus formation. Both positive signals that accumulate at the equator of the plasma membrane (green plus signs) and negative signals that accumulate in nonequatorial regions of the plasma membrane (red minus signs) are thought to travels along microtubules to the cell periphery.
Action at a distance during cytokinesis
Working with our collaborator, George von Dassow, we sought to monitor the distribution of microtubules with high spatial resolution in living embryos (see the 2009 paper in the Journal of Cell Biology; http://jcb.rupress.org/content/187/6/831.long). Surprisingly, when we eliminated microtubules that extend from the spindle to the cell periphery, cells were still able to generate a Rho zone and a cytokinetic apparatus and to divide. Moreover, cytokinesis was relatively accurate under these conditions, as judged by the fact that the cytokinetic apparatus closed inward between the separating chromosomes. The next series of movies show microtubules only, first in controls and then in experimentals; the movies reveal that cells can still divide when the microtubules that extend from the cell surface to the cell periphery have been eliminated.
Cytokinesis movie 5—A fluorescence, time-lapse movie showing microtubules in a normal (control) sea urchin embryo. The spindle is in the center of the cell and astral microtubules can clearly be observed to extend from the spindle to the cell periphery before and during cytokinesis.
Cytokinesis movie 6—A fluorescence, time-lapse movie showing microtubules in a normal (control) sand dollar embryo. The spindle is in the center of the cell and astral microtubules can clearly be observed to extend from the spindle to the cell periphery before and during cytokinesis.
Cytokinesis movie 7—A fluorescence, time-lapse movie showing microtubules in a sea urchin embryo during treatment with a drug that causes loss of astral microtubules extending from the spindle to the cell surface. The time of drug application is revealed by the rapid withdrawal of microtubules extending from the spindle to the cell surface. In spite of the fact that this pool of microtubules are eliminated, the cells finish mitosis and undergo cytokinesis.
Cytokinesis movie 8—A fluorescence, time-lapse movie showing microtubules in a sand dollar embryo during treatment with a drug that causes loss of astral microtubules. The time of drug application is revealed by the rapid withdrawal of microtubules extending from the spindle to the cell surface. In spite of the fact that this pool of microtubules are eliminated, the cell finishes mitosis and undergoes cytokinesis.
These experiments make it clear that cells can divide without microtuubles extending from the spindle to the cell surface. But is the division under these circumstances accurate? That is, does the cytokinetic apparatus find its way between the sets of chromosomes? Imaging of both microtubules and DNA provides the answer:
Cytokinesis movie 9—A fluorescence, time-lapse movie showing microtubules (green) and chromosomes (red) in a normal (control) sea urchin embryo. The spindle is in the center of the cell and astral microtubules can clearly be observed to extend from the spindle to the cell periphery before and during cytokinesis.
Cytokinesis movie 10—A fluorescence, time-lapse movie showing microtubules (green) and DNA (red) in a sea urchin embryo during treatment with a drug that causes loss of astral microtubules. Shortly after the drug is applied, the astral microtubules rapidly disappear. Nonetheless, the cell finishes mitosis and undergoes cytokinesis with accurate separation of the two masses of DNA into separate daughter cells.
These and other results (see http://jcb.rupress.org/content/187/6/831.long) led to the unexpected conclusion that contact or even close approach of the spindle microtubules to the plasma membrane and cell cortex are not strictly necessary for delivery of the cytokinetic signal and accurate cytokinesis although they are needed for proper shaping of the Rho zone, in that it is broader in the absence of cortical microtubules. How then is the signal delivered? We are currently trying to answer this question.
Balance of Rho activation and inactivation
Rho cycles between an active (GTP-bound) and inactive (GDP-bound) form. The ability to hydrolyze GTP to GDP is a basic feature of Rho and related G-proteins (e.g. Cdc42 and Rac) but they are not very good at it. That is, without help, a typical Rho class GTPase only hydrolyzes a single GTP every 5-50 minutes. This observation and several others led us (Bill, Ann Miller and George) to propose the “GTPase flux” model, which posits that the Rho and Cdc42 zones seen during cytokinesis and wound healing represent areas where Rho and Cdc42 are being rapidly driven through the GTPase cycle (for details, see the 2006 Bioessays paper: http://onlinelibrary.wiley.com/doi/10.1002/bies.20477/abstract). To put it another way, the Rho zone is not simply a place in the cell where Rho is being turned on (ie exchanging GDP for GTP) but also a place where it is being turned off (ie hydrolyzing GTP to GDP). This idea is sort of counterintuitive, but it has the virtue of explaining several observations, including the ability of Rho zones to form and focus rapidly, and to disappear within a few minutes after their formation.
Ann decided to test the GTPase flux idea by modifying one of the enzymes hypothesized to regulate Rho during cytokinesis–MgcRacGAP. This is a protein that is important for cytokinesis and which has GAP activity–the ability to promote GTP hydrolysis by Rho GTPases–although its precise targets remain controversial. In a study published in Nature Cell Biology (http://www.nature.com/ncb/journal/v11/n1/full/ncb1814.html), Ann used a “gene replacement” approach, meaning she simultaneously removed the endogenous Xenopus MgcRacGAP while replacing it with either a wild type Mgc (as a control) or one of two mutant forms–“R384A”, in which a single amino acid is changed to render the protein GAP inactive but otherwise keep it intact and “dGAP” (delta GAP) in which the GAP domain is entirely removed from the protein. The basics results of this procedure are shown in Cytokinesis figure 5.
Cytokinesis figure 5—Gene replacement of MgcRacGAP in Xenopus embryos. The top part of the figure shows an immunoblotting documenting the presence of MgcRacGAP in control (uninjected) embryos, the loss of MgcRacGAP following injection of two different morpholinos (Mgc MO.1 and Mgc MO.2) and another control, the replacement of the lost MgcRacGAP by reexpression of wild type MgcRacGAP. The bottom part of the figure shows loading controls in which the level of tubulin was monitored and the numbers to the left of each part of the figures refer to molecular mass (see Ann’s paper in Nature Cell Biology for details (http://www.nature.com/ncb/journal/v11/n1/full/ncb1814.html).
Having established the gene replacement approach, Ann next compared the ability of the different MgcRacGAP mutants to rescue cytokinesis following depletion of the endogenous MgcRacGAP, using wildtype Mgc as a positive control. As shown in Cytokinesis figure 6, while the wild type rescues the cytokinesis defect, the two mutant forms do not do so.
Cytokinesis figure 6–Analysis of cytokinesis following gene replacement of MgcRacGAP in Xenopus embryos. In this image, embryos are stained with a probe to actin filaments, which reveals the cell borders, and a probe for DNA which shows the nucleus. When cytokinesis is normal, each cell has one nucleus, and the cells are of a more-or-less uniform size, as shown in the uninjected control sample (Control). When cytokinesis fails, cells are many different sizes and contain more than one nucleus, or very large nuclei, as seen in the sample where MgcRacGAP has been depleted (Mgc KD). Reintroduction of wild type MgcRacGAP (Mgc KD + WT Mgc) largely rescues the cytokinetic defects while both of the GAP mutants (Mgc KD + R384A Mgc and Mgc KD + dGAP Mgc) are far less effective, indicating that the GAP activity of Mgc is important for cytokinesis.
Since the above results indicated that the GAP activity of MgcRacGAP is important for cytokinesis, Ann next analyzed cytokinesis in living cells while monitoring Rho activity. She found that while both the R384A and dGAP mutants were able to generate cytokinetic Rho zones, they differed both from the control Rho zones and from each other–both mutations resulted in broader-than-normal Rho zones, but zones in cells expressing the dGAP mutant often displayed rapid back and forth oscillations. The results are presented first in image form in Cytokinesis figures 7 and 8 and then as movies below those figures.
Cytokinesis figure 7–Effects of MgcRacGAP mutations on cytokinetic Rho zones. Each of the frames is from a time lapse movie of individual cells (circled in red in the first frame) undergoing cytokinesis and shows the distribution of active Rho (in white) in those cells. The cells were subjected to no special treatment (Control), expression of wild type MgcRacGAP (WT Mgc), MgcRacGAP with a single amino acid changed to inactivate the GAP activity (R384A Mgc) or MgcRacGAP lacking the entire GAP domain (dGAP Mgc). In the control and the cells expressin the wild type MgcRacGAP, the cytokinetic Rho zone is relatively narrow. In the point mutant (R384A) the zone is much broader, while the dGAP zone is broader but unstable, meaning it moves back and forth across the cell surface.
Cytokinesis movie 11—A fluorescence, time-lapse movie showing the cytokinetic Rho zones in a control sample.
Cytokinesis movie 12—A fluorescence, time-lapse movie showing cytokinetic Rho zones in cells expressing R384A MgcRacGAP–the zones are wide and poorly definited.
Cytokinesis movie 13—A fluorescence, time-lapse movie showing cytokinetic Rho zones in cells expressing dGAP MgcRacGAP–the zones are wide and poorly definited (although not as much as the R384A mutants) and undergo a dramatic side-to-side oscillation.
Cytokinesis figure 8–Effects of MgcRacGAP mutations on cytokinetic Rho zones viewed in cross section (side view). In these images, the cytokinetic Rho zones are seen from the side. Each time point is 20s apart; note that the zones of the two mutants are broader than the control. Note also that the dGAP mutant undergoes side-to-side oscillations. The arrow in the control image indicates the time at which the furrow starts to ingress into the cell; the asterisks in the dGAP images indicated the outer boundaries of the Rho zone.
The above results are consistent with the Rho GTPase flux hypothesis and raise many questions about the means by which Rho is regulated during cytokinesis. We are currently trying to answer some of these questions.