Can Simulations and Modeling Decipher NMR Data for Conformational Equilibria? Arginine–Vasopressin

Arginine vasopressin (AVP) has been suggested by molecular-dynamics (MD) simulations to exist as a mixture of conformations in solution. The <sup>1</sup>H and <sup>13</sup>C NMR chemical shifts of AVP in solution have been calculated for this conformational ensemble of ring conformations (identified from a 23 μs molecular-dynamics simulation). The relative free energies of these conformations were calculated using classical metadynamics simulations in explicit water. Chemical shifts for representative conformations were calculated using density-functional theory. Comparison with experiment and analysis of the results suggests that the <sup>1</sup>H chemical shifts are most useful for assigning equilibrium concentrations of the conformations in this case. <sup>13</sup>C chemical shifts distinguish less clearly between conformations, and the distances calculated from the nuclear Overhauser effect do not allow the conformations to be assigned clearly. The <sup>1</sup>H chemical shifts can be reproduced with a standard error of less than 0.24 ppm (<2.2 ppm for <sup>13</sup>C). The combined experimental and theoretical results suggest that AVP exists in an equilibrium of approximately 70% <i>saddlelike</i> and 30% <i>clinched open</i> conformations. Both newly introduced statistical metrics designed to judge the significance of the results and Smith and Goodman’s DP4 probabilities are presented.