Integrated modeling of yeast kinetochore microtubule dynamics and the imaging thereof by fluorescence microscopy

We were interested in determining the mechanisms by which microtubules mediate the segregation of chromosomes during mitosis in S. cerevisiae. In metaphase each chromosome has a single kinetochore assembled in its kinetochore region and each yeast kinetochore is in turn associated with a single microtubule plus end. A yeast strain containing a GFP fusion to the kinetochore protein, Cse4p, was used to track kinetochore microtubule dynamics by fluorescence microscopy. However, the images were blurry as a result of diffraction, and so rather than deconvolve the experimental images, we instead convolved the model predictions with a model of the image formation process to generate simulated microscopic images. Using the latter approach, which we call "model-convolution," it was impossible to mistakenly converge to a false reconstructed image, as can happen with the former approach, and it was computationally faster. The simulated images were compared statistically to the experimental images to determine that a simple dynamic instability model was unacceptable. However, a stable spatial gradient of microtubule catastrophe rate model provided reasonable agreement. These results show that the behaviors of proteins confined to subcellular compartments can be quantitatively analyzed, provided that both the intrinsic dynamics and the imaging of those dynamics are modeled.