The final step of chromosome segregation is the division of the cell into two at cytokinesis. The position of the cytokinetic furrow is determined by the position of the mitotic spindle. The majority of cell divisions are symmetric: i.e. the two daughters have the same developmental fate. However, many divisions are asymmetric: the fate of the two daughter cells is different, Such asymmetric divisions are often accompanied by unequal division: the mother cell divides into cells of different sizes. Unequal divisions are determined by the unequal position of the mitotic spindle, which is eccentrically placed prior to cell division. The position of a mitotic spindle is thought to be determined by the interaction between microtubules and the cell cortex.
The establishment of an unequal division can be thought of as requiring three steps. First a polarity cue. Second the polarity machinery must respond to this cue to polarize the cell. Third, the polarity must communicate to the microtubules to eccentrically place the mitotic spindle. In C.elegans embryos, the polarity of the embryo is determined by the PAR proteins, which segregate into two domains.
In C.elegans, the key questions have been a) What is the polarity cue. b) How to the PAR proteins segregate into two domains. c) How do the PAR proteins position the mitotic spindle.
We have investigated the biophysical basis of asymmetric cell division. In the first cell C.elegans embryo, the spindle initially assembles in the middle of the cell.
At anaphase one of the spindle poles (the anterior) stays relatively still; the other (the posterior) moves towards the posterior cortex, generating an asymmetrically placed spindle. We have taken a biophysical approach to this problem, in collaboration with Ernst Stelzer at EMBL and Joe Howard in Dresden. Laser cutting of spindles at anaphase has demonstrated that pulling forces pull the spindle poles apart. Importantly, more force is exerted on the posterior aster than the anterior aster. Further experiments have shown that this difference in force is under control of the PAR proteins. Therefore the PAR machinery, by segregating into two domains, modulates the force generated on each aster, thus generating asymmetric cell division.
A question that arose from this first set of experiments was how is a greater net force exerted on the posterior pole? The pulling force on a spindle pole will probably come from the sum of the individual pulling forces of each of the cortex-attached astral microtubules. Therefore to distinguish different possible mechanisms for asymmetric movement of the spindle poles, we needed to measure the contribution of individual astral microtubules to the movement of the spindle poles. In developing such an assay we made the following assumptions: The cortex provides a pulling force on some of the plus ends of microtubules. However this force is limited by the attachment of the minus ends to the centrosome. Destruction of the connections that link microtubules to the spindle poles should reveal those microtubules subject to cortical force generation. We severed the connection of microtubules to the spindle pole, by disintegrating the pericentriolar material into a number of fragments using a pulsed UV laser. We term this technique Optically Induced Centrosome Disintegration or OICD. The analysis of aster fragment movement after OICD revealed striking differences in the distribution of fragment velocities between the anterior and posterior asters. In collaboration with Joe Howard at the CBG we have analyzed the statistics of fragment movement using a mean-variance analysis. Analysis of the mean-variance suggests that the force imbalance is due to a larger number of force generators pulling on astral microtubules of the posterior compared to the anterior aster.
An important function of microtubules is to move cellular structures such as chromosomes, mitotic spindles and other organelles around inside cells. Microtubules do this by attaching the ends of microtubules to various structures inside the cell. As the microtubules grow and shrink, these structures are pulled or pushed around inside the cell.
Microtubules (MTs) are polymers of the tubulin heterodimer involved in diverse functions such as the regulation of membrane traffic and chromosome separation during mitosis. They consist of protofilaments (PFs), head to tail alignments of the tubulin α/β heterodimer that run lengthwise along the MT axis and interact through lateral contacts to form the MT wall. They grow by an interesting behaviour known as dynamic instability. Here, microtubules interconvert between growing and shrinking phases. This interconversion is driven by GTP hydrolysis, and is stochastic. One can estimate the change of a transition over a certain time frame, but never predict exactly when this transition tales place.
The central question in understanding the function of microtubules is to understand how the dynamics ends of microtubules couple to cellular structures, and to understand how this coupling regulates the dynamics behaviour of the end.
We think of the microtubule end as having at least three different properties. The growing and shrinking ends have alternate structures. Secondly, the biochemistry of the end changes by GTP hydrolysis. Thirdly, the end forms a distinct target for distinct proteins.
Together with Joe Howard we study how the properties of microtubule ends correspond to the function of microtubules in spindle assembly and spindle position. For this purpose, we study the activity of microtubules and microtubule-associated proteins in vitro. We relate these activities to the dynamic behaviour of microtubules in vivo, and their functional properties in spindle assembly and position, primarily using C.elegans embryos.
We have worked on how GTP hydrolysis could destabilize MTs came using a non-hydrolyseable analogue of GTP called GMPCPP (ref). We used cryoelectron microscopy to analyze the structure of PFs derived from MTs polymerized with GMPCPP. These studies showed that the GTP protofilaments were straighter than protofialments derived from GDP protofilaments. Furthermore, the tubulin subunits in the lattice shortened after hydrolysis, changing the repeat length of the tubulin dimers, suggesting that a conformational change within tubulin subunits induced by GTP hydrolysis could induce PF curvature. Recent structural analysis of the two alternative structures, either straight or curved protofilaments at atomic resolution have shown that GTP hydrolysis produced a kink between the tubulin subunits of 12° (Nogales and Wang 2006). Such a kink would indeed increase the curvature of the PFs.
Microtubule dynamic can be easily followed under physiological conditions using Xenopus egg extracts. We have identified the key factors required to control the steady state length of microtubules in Xenopus extracts. These are the stabilizing protein XMAP 215 and the destabilizing protein MCAK, a kinesin. The results in extracts suggested that the differential activities of these two proteins could control the length of microtubules. Using these proteins we have successfully reconstituted the dynamic behaviour of microtubules under physiological conditions. Thus the dynamic behaviour of microtubules in Xenopus extracts appears to be determined by the relative activity of XMAP215 and MCAK.
The defining characteristic of the XMAP family of proteins is the presence of TOG domains, which are HEAT repeats. The function of these HEAT repeats and thus the biochemical activity of XMAP have been obscure. In collaboration with the laboratory of Steven Harrison, we have shown the first TOG domain of the Xenopus (unpublished) and the yeast protein binds tubulin very tightly. This suggests that a key role of XMAP215 could be to deliver tubulin to the growing end of the microtubule. Thus we now think of XMAP215 as a polymerase, in the same way that MCAK is a depolymerase. The relative activity of the polymerases and depolymerases will regulate microtubule length.
To understand the relative contributions of XMAP to the control of microtubule growth in C.elegans, we started with our genome-wide screen for cell division defects. We used a marker of growing ends, called EB1 to screen for defects specifically in microtubule behaviour. The surprising result of this screen is that the only two molecules that have a noticeable effect on microtubule growth in the cytoplasm are the Polymerase (Zyg-9) and its associated protein TACC. All the rest of the genes we studied are involved in regulation of the number of microtubules that grow from the centrosomes. Thus our global view of microtubule growth in C.elegans is as follows. Microtubules are nucleated at centrosomes. The growth of microtubules from the centrosome depends on the relative activity of centrosome-associated proteins. Once a microtubule grows from the centrosome, it continues at the astonishing speed of over 40 µM per minute, driven by the activity of polymerases until it reaches the cortex, where it pauses for a few seconds before depolymerizing.
As cells enter mitosis they build a microtubule-based structure called a mitotic spindle. Chromosomes attach to the spindle, and the interaction between the spindle and the microtubules drives chromosome segregation. The assembly of a mitotic spindle depends critically on the assembly of mitotic specific organelles, such as centrosomes of kinetochores. We are studying the assembly of a centrosome in C.elegans embryos, and how it contributes to the cell division of the C.elegans embryo. There are three interesting aspects to this problem. How does the centriole assemble duplicate once per cell cycle. Secondly, how does the centriole recruit pericentriolar material, and thirdly, how does the pericntriolar material mediate its function in assembling a mitotic spindle, and its signaling function in cytokinesis and cell polarity. Our lab is working on various aspects of this problem (see also Genomics).