Biological function, such as DNA replication and chromosome segregation, depend on the complex interactions among many different proteins. As a biological function changes during evolution, the network of proteins that supports it must change accordingly. But how does the network alter in response to change in a biological function? Do most of changes occur in the network’s most critical protein or do they occur in other proteins whose collective interactions produce the biological function? To address these questions, Phoebe Hsieh, a graduate student in Andrew Murray’s lab, focused on the protein rings, called cohesin, that hold sister chromosomes together, from the time of DNA replication to cell division. One component of the ring, kleisin, has two different forms (encoded by two different genes), one specialized for mitosis and the other specialized for meiosis. Phoebe investigated which mutations allow the meiotic kleisin to support mitotic chromosome segregation, in budding yeast. The experiment involves three steps: engineering cells to express the meiotic instead of the mitotic kleisin, evolving them to reproduce faster, and tracking down the mutations that allowed cells to adapt to expressing the “wrong’ protein. Most of the work was done in Cambridge, but one important set of experiments was done by Adele Marston and her postdoc Vasso Makrantoni at the University of Edinburgh.
Phoebe found that replacing the mitotic kleisin, Scc1, with its meiotic counterpart, Rec8, reduces the linkage between sister chromosomes at mitosis, advances the timing of genome replication, and reduces reproductive fitness by 45%. By evolving fifteen parallel, Rec8-expressing populations in the lab for 1750 generations, Phoebe asked which mutations, either in Rec8 or elsewhere in the genome, allow budding yeast cells to improve their fitness and sister chromosome linkage. Across the parallel populations, she found mutations altered three functional modules: 1) other components of cohesin and the enzyme that cleaves cohesin at the end of mitosis, 2) the transcriptional mediator complex, and 3) proteins that stimulate the transition from G1 to S phase. Both individually and in combination, these mutations improve sister chromosome cohesion, delay genome replication, and improve mitotic fitness. Only proteins in the first class have previously been implicated in controlling the linkage between sister chromosomes, demonstrating that experimental evolution can reveal previously undiscovered interactions that contribute to and regulate biological functions. In contrast to the frequent occurrence of the other mutations, only one population had mutated Rec8.
Why are mutations in Rec8, the protein that was swapped between meiosis and mitosis, so rare? Phoebe argues that answer lies in the target size for different classes of mutations: the number of base pairs that can mutate to produce particular class of mutation. Mutations at roughly 20% of the base pairs inactivate a typical protein, whereas there are very few base pairs where a mutation improves or qualitatively alters a protein’s function. Many of the the mutations in transcriptional mediator and the G1 to S phase regulators completely inactivate protein by creating a stop codon in the middle of a protein, explaining why they occur frequently. In contrast, we expect that there are very few changes that could improve Rec8’s function, making the rate at which those mutations occur very low, and explaining why we see them rarely, if at all. Because Rec8 and Scc1 arose from the duplication of an ancestral kleisin gene, the attempt to get Rec8 to take over Scc1’s function offers insight into what happens after genes duplicate, often repeatedly, and then each evolve to perform a different function. Phoebe’s work implies that this process of duplication and divergence is as likely to involve changes in the proteins that directly or indirectly interact with the duplicated proteins as it is to involve changes in the duplicated proteins themselves.