Figure 1: An array of mitotic spindles obtained in vitro with Xenopus laevis egg extracts (Dinarina et al., Cell, 2009).
Figure 2: The metaphase spindle, a dynamic bipolar structure of filaments called microtubules (white) that are connected by molecular motors (orange). This simulation elucidates how a spindle can remain stable for hours, even though it is made of filaments that individually exist for less than a minute (Loughlin et al. 2010).
The Nédélec group combines experimental methods and physical theory to study how the interior of cells is spatially organised and how multiple cells are arranged in space.
Previous and current research
Modern microscopy has demonstrated the dynamic nature of biological organisation. The mitotic spindle, for example, is a stable and solid cellular structure: in a given cell type, it has a precise symmetry and very reproducible dimensions. Yet, except for the chromosomes, all the components of a spindle – polar filaments called microtubules and associated proteins – are in rapid turnover. Microtubules grow, shrink and disappear in less than a minute and their associated proteins continuously and stochastically bind and unbind even faster. The resulting assembly, although highly dynamic, is remarkably precise: it can remain steady for hours waiting for the right signal, to eventually apply the balanced forces necessary to position and segregate the chromosomes exactly.
The spindle is thus a fascinating structure that illustrates a central question in biology: how can the uncoordinated and inevitably imperfect actions of proteins and other molecules collectively fulfil the biological needs with the required accuracy? Today, understanding biological phenomena from their multiple biological components seems within our reach, as testified by the rise of systems biology. Yet, collective behaviours in biology require more than statistical averages. Understanding such complex collective behaviours is challenging for many reasons: 1) the diversity of molecular players is enormous; 2) their interactions are often dynamic and out-of-equilibrium; and 3) the properties of the constituents have been selected by natural evolution.
We approach this topic in practical terms by developing in vitro experiments and modelling tools, allowing us to reduce the number of components in the system: we can either remove specific proteins, or start from scratch by mixing purified components. Modelling allows us to recapitulate the process of protein organisation in a framework in which all the interactions are known exactly and can even be specified at will. We have developed an advanced simulation engine – called Cytosim – to simulate ensembles of multiple polar fibres and associated proteins, which can simulate problems involving microtubules, actin filaments or both. Simulations are often used to validate or refute existing ideas, but we also try to use them in a more creative way: one can generate systematically various properties for the molecules and automatically test their ability to form stable structures. The analysis of successful scenarios leads to the formulation of new hypotheses.
Future projects and goals
We will study systems in which experiments and theory can be synergistically combined. We currently focus on Xenopus egg extracts, an experimental system in which many aspects of mitosis can be recapitulated. We are also generally interested in modelling cellular processes in which the cytoskeleton is a major player, such as the different stages of mitosis, the generation of cell shape in S. pombe, or the generation of asymmetry during cell division.