GA98 and GB98 differ only in the identity of the amino acid at position 45, which is leucine in GA98 and tyrosine in GB98. The tyrosine seems to be critical in stabilizing a hairpin turn in the GB structure, possibly by interactions with an adjacent aromatic residue (F52 – see Fig. 6). In the structure of the GA95 protein (Fig. 5), L45 is partially solvent exposed, and therefore probably does not contribute substantially to stability. Of course, as a leucine is also hydrophobic, one might expect that it would also interact favorably with F52. This seems to be the case, as the GA98 protein binds to both HSA and IgG, suggesting that it transiently samples the GB structure.
As with studies I have talked about previously on lymphotactin, which has two structures in its native state, and the cro proteins, which can have different structures despite similar sequences, this research indicates that the relationship between sequence space and fold space may not be as straightforward as previously believed. We have known for some time that distance in sequence space does not imply distance in fold space — the convergent evolution of numerous proteins has shown us that, at least. These experiments inform us that distance in fold space need not imply distance in sequence space, that it is possible for a protein to undergo a structural saltation in which a single mutation gives rise to a dramatic difference in structure.
It remains to be seen whether this is true in larger proteins as it seems to be in small proteins and domains. If it is generally the case that proteins can drift towards new folds without losing their native function, then “jump” over to a completely different structure, without transiting the molten-globule state, then our understanding of evolution, particularly early evolution, may change significantly. For instance, the objection that proteins could not evolve new functions without wandering into the unfolded state loses its weight. Moreover, given that in two of these cases the proteins actually sample both structures (and hence, both functions), it is not even necessarily true that a protein must lose its original function before evolving a new one. Equilibrium switching between structures would allow a protein to fulfill both the old and new functions, albeit likely at reduced efficiency. As I’ve mentioned before, one answer to this lack of efficiency would be altering protein concentration by adjusting gene dosage, i.e. gene duplication. From that point, the generation of fold and functional diversity by evolution might be a straightforward process.
Alexander, P., He, Y., Chen, Y., Orban, J., & Bryan, P. (2009). From the Cover: A minimal sequence code for switching protein structure and function Proceedings of the National Academy of Sciences, 106 (50), 21149-21154 DOI: 10.1073/pnas.0906408106 OPEN ACCESS