Fig. 2. Overview of the hybrid strategy: A combination of cryo-EM, computational, and NMR.
(A) Schematic representation of the unbiased CG ab initio approach. The internal elastic networks used for both AVP (left) and V2R (right) are shown. The full system (center) used for the CG-REMD simulations included two receptors and two peptides. (B) Ten most populated clusters (67.5% of the whole conformations) obtained for the AVP-V2R complex after three independent CG-REMD simulations. (C) Schematic representation of the successive steps using the CDMD method to fit the models resulting from the CG-REMD simulations into the L cryo-EM map. (D) Mapping AVP contact surface by experimental STD NMR. The STD effect profile (in percentage) is shown as a function of AVP protons [aliphatic, N (backbone), and aromatic]. (E) Cross-correlation between computational and STD NMR. Variability of the position of AVP was calculated as mean RMSD values (in angstroms) after cross-comparison of five models resulting from the fitting procedure of each of the 10 clusters in the L density map. Cluster 5 showed the smaller variability (2.2 Å). Experimental STD values were compared to the expected STD values from all-atom models issued from MDSs, and correlation coefficients were calculated for the whole peptide (R1–9). Cluster 5 appeared as the best cluster fitting to experimental STD values. (F) Building of the final L structure based on remapping cluster 5 into the L density map. (G) On the basis of the L structure, the T structure was built to match the T density map.