Double Mutant Cycle Thermodynamic Analysis of the Hydrophobic Cdc42-ACK Protein-Protein Interaction † Andrea E. Elliot-Smith, Darerca Owen,* Helen R. Mott, and Peter N. Lowe Department of Biochemistry, UniVersity of Cambridge, 80 Tennis Court Road, Cambridge CB2 1GA, U.K. ReceiVed August 2, 2007; ReVised Manuscript ReceiVed October 1, 2007 ABSTRACT: Protein-protein interactions such as those between small G proteins and their effector proteins control most cell signaling pathways and thereby govern many cellular processes in both normal and disease states. Each small G protein interacts with several effectors, some shared between similar G proteins and others unique to a single GTPase. Although there is knowledge of the structural basis of these interactions, there is limited understanding of their thermodynamic basis. This is particularly significant because of the intrinsic conformational flexibility of the interacting partners. Here we have conducted a double mutant thermodynamic cycle for two key hydrophobic interactions in the Cdc42-ACK interface: Val42 Cdc42 -Ile463 ACK and Leu174 Cdc42 -Leu449 ACK . Val42 and Leu174 are known to be energetically important in this complex from previous thermodynamic studies, and their respective partners were predicted from the structure of the complex. Such a study has not been hitherto performed on any hydrophobic protein-protein interaction. The results confirm that a significant proportion of the overall interaction is dependent upon these residues, but in neither case is the direct interaction between the side chains the predominant energetic force. Indeed, the interaction of the side chains of Val42 and Ile463 appears to exert an energetic penalty. Rather, the stabilization of the complex, which requires the presence of these two pairs of residues, appears to be due to conformational changes, or interactions, that are not easily visualized in the structure of the complexes. In this respect, it is noteworthy that isolated Cdc42 shows regions of disorder and isolated ACK has no stable tertiary structure, whereas the Cdc42-ACK complex has a well-defined quaternary structure. Such changes may well be critical for the known selectivity of Cdc42 and related proteins such as Rho and Rac, for their wide range of effectors. Members of the Rho family of small GTP 1 binding proteins, including Rho, Rac, and Cdc42, play important roles in the control of cell growth, differentiation, and migration (1). They act as molecular switches, interconverting between GDP-bound (inactive) and GTP-bound (active) forms. In response to stimuli, G proteins are activated to their GTP- bound form, in which state they are conformationally competent to interact with a large variety of downstream effector proteins (2). Several targets of the Rho family proteins have been identified, as have multiple regulatory proteins. Biochemical and structural characterization of these interactions has improved our understanding of these func- tional relationships (3). An important question in small G protein biology is how these GTPases recognize and bind their effector proteins and which residues contribute to this specificity, particularly given the high degree of sequence similarity between Rho family proteins. For example, Cdc42 and Rac are 72% identical, and while some effectors such as the serine/ threonine kinase PAK (p21-activated kinase) are common to both, others interact specifically with only one of these GTPases. One effector that specifically interacts with Cdc42 but not Rac is ACK, a tyrosine kinase (4) implicated in integrin signaling (5) and clathrin-mediated endocytosis (6). Extensive mutagenesis studies on the interaction between Cdc42 and ACK (as well as two other CRIB effector proteins) have identified energetically key residues for these interactions (7, 8). The analysis of single-residue mutations generally, however, enables only the relative energies of having one residue versus another (usually alanine) at a specific position to be determined. It is therefore difficult to identify how much the loss of a specific interaction contrib- utes to the effect of a mutation in comparison with other effects on the protein (e.g., structural changes). Double mutant cycle analysis is an analytical technique that can, however, determine the interaction energy between two specific residues, eliminating any effects that may result from changes in solvation or reorganization energies as a result of the mutation (9). Double mutant cycle analysis was first used to analyze enzyme-substrate interactions in tyrosyl-tRNA synthetase (10). Since then, it has been used extensively in the study of protein folding mechanisms (in particular of barnase) (11- † This research was supported by a BBSRC Studentship to A.E.E.-S. and CR-UK project Grants C1465/A2590 and C11309/A5148. * To whom correspondence should be addressed. Telephone: +44- 1223-764824. Fax: +44-1223-766002. E-mail: do@bioc.cam.ac.uk. 1 Abbreviations: CRIB, Cdc42-Rac interactive binding; GDP, guanosine 5′-diphosphate; GTP, guanosine 5′-triphosphate; PAK, p21 activated kinase; ACK, activated Cdc42-associated kinase; WASP, Wiskott-Aldrich syndrome protein; GST, glutathione S-transferase; IPTG, isopropyl -D-thiogalactopyranoside; PEP, phosphoenolpyruvate; DTT, dithiothreitol; SPA, scintillation proximity assay; ITC, isothermal titration calorimetry; GMPPNP, guanylyl 5′-imidodiphosphate; wt, wild type. 14087 Biochemistry 2007, 46, 14087-14099 10.1021/bi701539x CCC: $37.00 © 2007 American Chemical Society Published on Web 11/14/2007