Comprehensive Analysis of Coupled Proline Cis-Trans States in Bradykinin Using ωBP-REMD Simulations.

It is well-known that proline (Pro) cis-trans isomerization plays a decisive role in the folding and stabilization of proteins. The conformational coupling between isomerization states of different Pro residues in proteins during conformational adaptation processes is not well understood. In the present work, we investigate the coupled cis-trans isomerization of three Pro residues using bradykinin (BK), a partially unstructured nonapeptide hormone, as a model system. We use a recently developed enhanced-sampling molecular dynamics method (ω-bias potential replica exchange molecular dynamics; ωBP-REMD) that allows us to exhaustively sample all combinations of Pro isomer states and obtain converged probability densities of all eight state combinations within 885 ns ωBP-REMD simulations. In agreement with experiment, the all-trans state is seen to be the preferred isomer of zwitterionic aqueous BK. In about a third of its structures, this state presents the characteristic C-terminal ß-turn conformation; however, other isomer combinations also contribute significantly to the structural ensemble. Unbiased probabilities can be projected onto the peptide bond dihedral angles of the three Pro residues. This unveils the interdependence of the individual Pro isomerization states, i.e., a possible coupling of the different Pro isomers. The cis/trans equilibrium of a Pro residue can change by up to 2.5 kcal·mol-1, depending on the isomerization state of other Pro residues. For example, for Pro7, the simulations indicate that its cis state becomes favored compared to its trans state when Pro2 is switched from the trans state to the cis state. Our findings demonstrate the efficiency of the ωBP-REMD methodology and suggest that the coupling of Pro isomerization states may play an even more decisive role in larger folded proteins subject to more conformational restraints.

Efficient and accurate calculation of proline cis/trans isomerization free energies from Hamiltonian replica exchange molecular dynamics simulations.

Proline cis/trans isomerization plays an important role in many biological processes but occurs on time scales not accessible to brute-force molecular dynamics (MD) simulations. We have designed a new Hamiltonian replica exchange scheme, ω-bias potential replica exchange molecular dynamics (ωBP-REMD), to efficiently and accurately calculate proline cis/trans isomerization free energies. ωBP-REMD is applied to various proline-containing tripeptides and a biologically important proline residue in the N2-domain of the gene-3-protein of phage fd in the wildtype and mutant variants of the protein. Excellent cis/trans transition rates are obtained. Reweighting of the sampled probability distribution along the peptide bond dihedral angle allows construction of the corresponding free-energy profile and calculation of the cis/trans isomerization free energy with high statistical precision. Very good agreement with experimental data is obtained. ωBP-REMD outperforms standard umbrella sampling in terms of convergence and agreement with experiment and strongly reduces perturbation of the local structure near the proline residue.

Ligand binding and global adaptation of the GlnPQ substrate binding domain 2 revealed by molecular dynamics simulations.

Substrate-binding domains (SBD) are important structural elements of substrate transporters mediating the transport of essential molecules across the cell membrane. The SBD2 domain of the glutamine (GLN) transporter from bacteria consists of two domains D1 and D2 that bind GLN in the space between the domains in a closed conformation. In the absence of ligand, SBD2 adopts an open conformation with increased domain distance. In molecular dynamics (MD) simulations in the absence of ligands, no closing of the open conformation was observed on the MD time scale. Addition of GLN resulted in several reversible binding and unbinding events of GLN at the binding site on the D1 domain but did not induce domain closing indicating that binding and global domain closing do not occur simultaneously. The SBD2 structure remained in a closed state when starting from the GLN-bound closed crystal structure and opened quickly to reach the open state upon removal of the GLN ligand. Free energy simulations to induce opening to closing indicated a barrier for closing in the absence and presence of ligand and a significant penalty for closing without GLN in the binding pocket. Simulations of a Leu480Ala mutation also indicate that an interaction of a C-terminal D1-tail471-484 with a D2-helix418-427 (not contacting the substrate-binding region) plays a decisive role for controlling the barrier of conformational switching in the SBD2 protein. The results allow us to derive a model of the molecular mechanism of substrate binding to SBD2 and associated conformational changes.