Quantitative Analysis of Mitotic Chromosome Motion & Mechanics

Wallace Marshall, Sedat Lab UCSF

Chromosome Mechanics

Mitosis in an inherently mechanical process. Chromosomes are captured by microtubules, aligned onto the metaphase plate, and then pulled to the spindle poles. Understanding mitosis, therefore, means understanding the forces that act on chromosomes, and how these forces lead to the observed behavior, which in turn means understanding the mechanical properties of chromosomes and the surrounding nucleoplasm. The goal of this project is to develop a way to measure, quantitatively and noninvasively, the mechanical properties of chromosomes and the forces that act on them, inside a living cell. Specifically, we are using measurements of the three-dimensional movements of chromosomes during mitosis, to deduce the flexibility of the chromosomes, the viscosity of the nucleoplasm, and the magnitude of the forces acting on the kinetochore and the chromosome arms (the polar ejection force). Nucleoplasm viscosity is important becase it determines the velocity of motion produced by a given force acting on a chromosomes, and also may shed light on the physical structure of the nucleus. Chromosome flexibility provides information about chromosome organization and also determines the degree to which forces acting on the chromosome arms will be transmitted to the kinetochore. Quantitative measurements of the kinetochore and ejection forces is important in thinking about possible mechanisms for generating these forces.

This whole approach requires a way to track the movements of chromosomes in vivo. By movements, we mean not just changes in overall position of the chromosome, but changes in chromosome shape and curvature (this provides the basis for measuring flexibility). The first step in measuring chromosome motion is acquiring high-resolution three-dimensional images of chromosomes in living cells.

Visualization of Chromosomes in 3D in living Drosophila embryos

Visualization is done by injection of Cy5-labelled histone protein or the DNA vital dye Oli-Green into early embryos of Drosophila melanogaster. We then use time-lapse 3D microscopy to obtain a series of 3D images of the chromosomes with a temporal resolution of one 3D image every 10-20 seconds. Wide-Field deconvolution microscopy is used to avoid cooking (and thus killing) the embryos in the high intensity laser beams typically employed in confocal microscopy, while still obtaining high quality 3D images. Here is an example of such an image sequence which shows a nucleus from a living Drosophila embryo injected with Cy5-histone where one 3D image was collected every 25s, and then a projection was made at each time point. The movie runs from prometaphase to anaphase and then backwards as a loop.

Motion Estimation

Once we have imaged the chromosomes, we need a way to measure and track the motion of the arms. We have developed a model-based motion estimation algorithm that can track nonrigid motion in 3D under conditions of limited temporal resolution (the nonrigidity and limited time resolution are challenging problems as is the rather uniform appearance of chromosomes which leads to an insoluble aperture problem for most optical-flow methods).

The key to the model based approach is to first construct a set of wire-frame models of the chromosome arms, and then track the deformations of these wireframe models. here is an example of a series of such wireframe models for one time-series of one nucleus. This is a stereo pair so you have to cross your eyes to see it in 3D.


This motion analysis algorithm is not limited to chromosomes but can in principle be used to track the dynamics of other cellular structures, such as mitochondria.

Inference of Mechanical Properties

Once we obtain a motion estimate, we have developed quantitative measures of parameters such as mobility and flexibility, in order to better understand chromosome mechanics. We are currently working on developing ways to relate these measures to the physical properties of the chromosomes.

Visualization of Motion

In general, time-lapse 3D movies are very confusing, creating a need for tools to visualize 3D motion. The motion estimate provides a way to do this. Rather than directly view the motion vector field, which is itself even more complex than the original images, we use the motion information to generate the trajectories of selected sites on a chromosome. These trajectories can be plotted in a single 3D image, making the motion much easier to visualize and compare.

Publications

so far we have only published the algorithms that we have developed. The results are on the way - stay tuned!!
Marshall, W.F., Agard, D.A., and Sedat, J.W. 1995 Quantitative Analysis of Chromosome Motion in Drosophila melanogaster. Proc. SPIE 2412,33-42.

Marshall, W.F., Agard, D.A., and Sedat, J.W. 1994 Motion Estimation and Visualization for four-dimensional optical microscopy. Proc. SPIE 2184,149-158.

Links to other lab projects involving motion of cellular structures

Analysis of interphase chromatin motion This is one of my other main projects, in which I have measured the diffusional motion of interphase chromatin using a GFP system to track specific loci. Follow the link to learn more about this.
Visualization of 4D structural dynamics of yeast mitochondria Mitochondria turn out to have very interesting dynamics, and by tracking yeast mitochondria in time-lapse 3D datasets we were able to show that mitochondria undergo constant fusion and fission in vivo, which leads to a network-like steady-state morphology that has implications for mitochondrial inheritance and biogenesis.

About the Movie

The movie on this page was generated from a time-lapse 3D movie of an embryo injected with Cy5-labelled histone. One image was acquired every 25 seconds.