Regeneration and Learning in a Single Cell with Wallace Marshall of UCSF

We generally think about cells as small, stupid building blocks, with the idea that to build anything complicated, or do any sort of complex behavior, one needs large numbers of cells. But in fact cells can have highly complex forms, both internal and external, and are capable of an amazing array of behaviors. It is not obvious how cells are able to develop complex internal anatomies, and indeed very little is known about the origins of cell geometry. Where does shape arise from collections of molecules? Perhaps even more fascinating than the complex shapes of cells is the fact that individual cells can often show a wide range of apparently purposeful behaviors. If you watch a cell going about its business it is hard not to start thinking of it as an animal or a robot. These behaviors require decision making and computation, and again it is not at all clear how these computational activities arise from the collection of molecules in the cell.

Understanding how cells generate form and carry out behaviors is likely to be valuable in practice. Metabolic engineering is currently focused entirely on altering enzyme expression levels, but unless the size and shape of the cell's internal reaction vessels (organelles) can be reprogrammed, the potential impact of changing enzyme levels is likely to be limited by the capacity of those vessels. In medicine, many diseases show alterations of cell shape but we have literally no understanding of why this happens. Learning how cells regenerate could suggest new ways to induce regeneration in damaged human cells, providing an alternative to stem cell replacement therapy. Finally, if it is really true that cells are computing devices, perhaps we can learn how to reprogram a cell's behavior, either to turn cells into micron-scale robots (for example by making white blood cells that can attack cancer targets) or to alter the behavior of cells throughout the body.

The giant single celled ciliate Stentor coeruleus has a size and structural complexity that rivals that of the animal embryo. Stentor can perform all of the key processes of development biology, including axiation, pattern formation, and induction, all within a single cell. Moreover, when a Stentor cell is cut into pieces, the pieces can regenerate intact cells with normal geometry. The ability of Stentor cells to heal wounds and regenerate allowed microsurgical experiments on Stentor development, which were carried out between 1890 and the mid 1980's. But Stentor was never developed as a molecular model system and despite the wealth of surgical results, virtually nothing is known about the molecular mechanisms by which Stentor is able to develop and regenerate. We have sequenced the Stentor genome and developed methods to perturb gene function using RNA interference. We are now conducting RNAi based screens to identify genes involved in regeneration. At the same time, the genome now allows us to use RNA sequencing to map the genomic programs that underlie specific regeneration events. Stentor is also notable for its rich behavioral repertoire. Each cell can choose between a range of different motile, contractile, and feeding behaviors, implying the existence of decision making circuitry. It has even been reported that Stentor cells can show simple forms of nonassociative learning, an idea that we are now exploring by combining automated testing systems with RNAi screening.