Allosteric Communication Occurs via Networks of Tertiary and Quaternary Motions in Proteins

Citation: Daily MD, Gray JJ ( Allosteric Communication Occurs via Networks of Tertiary and Quaternary Motions in Proteins Michael D. Daily 0 1 2 Jeffrey J. Gray 0 1 2 Matthew P. Jacobson, University of California San Francisco, United States of America 0 Current address: Department of Chemistry, University of Wisconsin Madison , Madison, Wisconsin , United States of America 1 Funding: JJG was supported by National Institutes of Health grant R01-GM078221 and a Beckman Young Investigator Award 2 1 Program in Molecular & Computational Biophysics, Johns Hopkins University , Baltimore , Maryland, United States of America, 2 Department of Chemical & Biomolecular Engineering, Johns Hopkins University , Baltimore, Maryland , United States of America Allosteric proteins bind an effector molecule at one site resulting in a functional change at a second site. We hypothesize that allosteric communication in proteins relies upon networks of quaternary (collective, rigid-body) and tertiary (residueresidue contact) motions. We argue that cyclic topology of these networks is necessary for allosteric communication. An automated algorithm identifies rigid bodies from the displacement between the inactive and the active structures and constructs ''quaternary networks'' from these rigid bodies and the substrate and effector ligands. We then integrate quaternary networks with a coarse-grained representation of contact rearrangements to form ''global communication networks'' (GCNs). The GCN reveals allosteric communication among all substrate and effector sites in 15 of 18 multidomain and multimeric proteins, while tertiary and quaternary networks exhibit such communication in only 4 and 3 of these proteins, respectively. Furthermore, in 7 of the 15 proteins connected by the GCN, 50% or more of the substrate-effector paths via the GCN are ''interdependent'' paths that do not exist via either the tertiary or the quaternary network. Substrateeffector ''pathways'' typically are not linear but rather consist of polycyclic networks of rigid bodies and clusters of rearranging residue contacts. These results argue for broad applicability of allosteric communication based on structural changes and demonstrate the utility of the GCN. Global communication networks may inform a variety of experiments on allosteric proteins as well as the design of allostery into non-allosteric proteins. - The modern concept of allostery began with the models of Monod et al. (MWC model) [1] and Koshland et al. (KNF model) [2], which sought to account for allostery based upon gross properties of the transition between two well-defined end-states. More recent thermodynamic models of allostery characterize population shifts in conformational ensembles in more detail [3 5], and there is experimental evidence that alternate allosteric states are simultaneously populated in solution [6,7]. Nonetheless, mechanical and chemical transitions in individual molecules underlie the thermodynamic properties of allosteric proteins. That is, in individual molecules, energetic pathways of spatially contiguous, physically coupled structural changes and/or dynamic fluctuations must link substrate and effector sites [810]. Crystal structures have revealed that most allosteric proteins are complex systems with both tertiary and quaternary structural changes [11]. Previously, we quantified allosteric communication through tertiary structure from graphs of residue-residue contacts that form, break, or rearrange in the transition between inactive and active state structures [12]. In such network representations of protein structure, putative paths between residues distant in threedimensional space can be readily identified. These tertiary networks or contact rearrangement networks (CRNs) identified substrate-effector paths in 6 of 15 proteins tested, which indicated that tertiary changes play a significant but incomplete role in allosteric communication. In this work, we broaden the CRN approach toward more completely quantifying allosteric coupling mechanisms from structure. Specifically, we develop a network representation of quaternary structural changes (collective / rigidbody motions) and integrate this representation with the CRN. We seek to infer information about the allosteric coupling mechanism from gross properties of the differences between inactive and active structures. In this, our work resembles the MWC [1] and KNF [2] approaches but differs from investigations of the kinetic mechanism, that is, the order of events in the transition between inactive and active structural regimes [1316]. Most current computational approaches to large-scale protein dynamics (e.g. normal mode analyses [1720], Go models [21], and all-atom simulations [22,23]) predict motions and/or associated energetics by applying to the structure(s) theoretical models like the elastic network [24] and potential functions. While these predictions address important problems, most of these approaches do not predict allosteric pathways. By contrast to these problems, we will argue that allosteric pathway identification is facilitated by a network representation of a protein structural transition. Network representations of protein structures have previously been used to illuminate dynamic and/or allosteric properties. For example, large-scale fluctuations predicted from normal mode Allosteric regulation is a major mechanism of control in many biological processes, including cell signaling, gene regulation, and metabolic regulation, and malfunctioning allosteric proteins are often involved in cancer and other diseases. In allostery, an effector-binding signal transmits over a long distance through the protein structure, resulting in a functional change at a second site. While many three-dimensional structures of allosteric proteins have been solved, the allosteric communication mechanism is usually not obvious from the motions between inactive and active state structures. In addition, allosteric structural transitions involve both small-scale motions at the level of amino acid residues and large-scale motions at the level of domains. Here, to address allosteric mechanisms, we transform the aforementioned protein motions into a multi-scale global communication network (GCN) representation from which substrate-effector pathways and other important allosteric communication properties can be identified. The GCN accounts for substrate-effector pathways in 15 of 18 proteins surveyed, and the GCN reveals that allostery often depends on linkage between the small- and the large-scale motions. This work will inform a wide variety of experiments investigating allostery, and it proposes concepts for engineering allostery into non-allosteric proteins. analysis of the elastic network correlate with known conformational changes [17,19,25]. In addition, rigid and flexible regions of protein structures have been predicted from the network of contact and hydrogen bond constr (...truncated)