Sepehr Ehsani

8 December 2015


Beyond Order in Protein Structure: Disordered Domains as Generative Domains


The overwhelming majority of "tasks" in a cell are performed by proteins. Proteins perform these tasks in part, but not exclusively, based on their structure. Since the early 1950s, protein structures have been defined in terms of domains that may be of (roughly) three flavors: alpha-helical, beta-sheet and "disordered". Approximately a third of human proteins are thought to contain disordered segments, and these segments contribute significantly to protein-protein interactions and therefore to protein function. A persistent challenge, however, has been that we do not yet have an overarching explanatory theory of what a disordered domain really is. Our aim here is to begin to move in that direction, i.e., to create an informal and then a formal model of disorder in protein structure. The main finding thus far, using a theoretical analysis of the work to-date on disordered domains, is that disordered domains are not inherently "disordered" at all, and that they in fact assume structures that are "appropriate to, but not caused by" their immediate cellular environment. It is therefore more appropriate to refer to these domains as "generative" domains (GDs). We propose a tripartite "M.G.C." framework, composed of elements in a protein domain's Medium (sequence-independent), Geometry (sequence-dependent) and Chemistry (sequence-dependent), to explain the structural behavior of GDs under physiological conditions. Using this framework along with the disordered domain of α-Synuclein as a model sequence, we argue that disordered regions are geometrically ordered based on their medium, and that the key overlooked chemical interaction of GDs may be the non-electrostatic "H-H bonding" (as opposed to electrostatic H-bonding) interactions.



Back to Projects page (Link)