UCSF

The ability to make sense of the complex sea of ever-changing molecules in the environment is one of the cell's most remarkable abilities. The molecular mechanisms by which these external cues are processed are well understood for many individual signaling pathways. However, fundamental questions for cell biology remain about how the collection of all such pathways within a cell—the signaling network—correctly responds to many inputs it receives. This dissertation focuses not only on determining the detailed biochemical mechanisms that resolve the complexities and ambiguities of individual cell signaling networks, but also on how these networks can change or grow to accommodate new pathways during evolution. To this end, reconstituted signaling networks are developed to probe the behavior of signaling systems under different environmental conditions or network compositions, as well as to rigorously compare evolutionarily related systems. Two different ambiguous network structures from model systems are explored from this perspective: overlapping signaling pathways arising from duplication/divergence events, derived from fungal ERK kinase signaling; and central signaling nodes that respond to multiple inputs and fan out to many possible outputs, derived from small Ras GTPase signaling. In the former case, allosteric dependencies on pathway-specific scaffold proteins were found to distinguish evolutionarily related molecules from one another and facilitate the use of homologous molecules in distinct signaling pathways. These allosteric dependencies appear to have evolved by exploitation of pre-existing differences in the conformational landscapes of otherwise-equivalent redundant signaling molecules. In the latter case, the behavior of Ras systems was explored in a multi-turnover in vitro setting for the first time. This revealed that Ras systems could transmit both sustained and transient signals and that the concentration and identity of signaling components strongly impacted the timing, duration, shape and amplitude of the output. Moreover, the extent to which oncogenic mutations in Ras distorted outputs was highly dependent on this underlying network configuration. Together, these studies demonstrate the utility of examining signaling systems not only from their current configuration, but the paths that led to that configuration during evolution, and the paths that might be taken through perturbation in the future.