Fig 1.
3D representation, nomenclature and sequence of the model molecules used during the simulations.
Dimer and monomer 14-3-3ζ systems (on the right) were simulated with fragments of phosphopeptide 1 and 2 (sequence shown on the left). The phosphoserine in the sequence is highlighted in red. 14-3-3ζ monomers are represented as cartoons in green and cyan, phosphopeptides are shown in stick representations in orange and purple. The phosphopeptide under DF restraining is labelled by an upper-case letter in the abbreviations.
Fig 2.
Pathway visualization of the DF/HRE-MD simulation dim_p1hT.
A) The volume sampled by the p1t phosphopeptide during the simulation is shown by dots around the 14-3-3ζ protein coloured based on their position along the pathway. The most probable points in space to find the peptide in for each replica (probability density peaks) are represented by larger spheres. B) Density peaks and a few representative structures corresponding to the density peaks, coloured according to their position along the pathway. Replica density peaks are connected by lines to visualize the binding pathway. The 14-3-3ζ protein in panels A-B is shown as a surface representation, with the two monomers shown in light and dark grey. C) Average interaction energy between 14-3-3ζ and p1t. D) Interaction map between any atom of the pulled p1t peptide and the amino acids of the 14-3-3ζ protein, summarized for each replica (only amino acids with at least 0.1 hydrogen bond/salt bridge on average are shown). The scale indicates the average number of interactions.
Fig 3.
Comparison of binding pathways of different phosphopeptides.
A) Binding pathways from DF/HRE-MD simulations depicted as connected dots referring to the probability density peaks along the respective pathway (See Table 1 for more information). B) Electrostatic surface potential (ESP) of 14-3-3ζ in blue, white and red for positive, neutral, and negative surface patches, respectively. The positively charged main interaction site (IS1), and secondary interaction site (IS2) are connected by a positive surface along the binding pathways. A negative surface patch (NSP) involved in the binding process is also indicated. Serine58 (S58), a phosphorylation site is located near the binding pathway. For clarity, the 14-3-3ζ C-terminal tail is not shown.
Fig 4.
14-3-3ζ residues involved in the phosphopeptide interactions.
The table on the left side shows the 14-3-3ζ amino acids, which were found important based on the interaction map analysis (Fig 2, S1–S7 Figs) of the DF/HRE-MD simulations. Entries in the various simulations are marked as not applicable (n/a), not significant (-), significant (+), important (++), or major interaction partners (+++). The last three amino acids in the table were interaction partners from the other monomer. The right side shows the three dimensional structure of the 14-3-3ζ protein, where the monomers are depicted in cartoon and surface representations, respectively. The amino acids are depicted in stick representation, coloured differently for the two monomers. In the cartoon representation, peptide-interacting amino-acids are coloured according their position in the secondary structure, where helices 3, 5, 7 and 9 are shown in red, blue, green and brown, respectively, while the C-terminal stretch is depicted in light purple. For the surface representation, amino acids found significant only in the bound-state replicas (IS1) are shown as dark purple, amino acids along the binding pathway are marked in orange, light green and light blue, if they appeared prior, in, or after the secondary interaction site (IS2). See S8 Fig for details.
Fig 5.
Flexibility of the 14-3-3ζ C-terminal tail.
A) Representative structures of the conformational sub-states of the C-terminal stretch (marked in red) interacting with the inner or outer surface of 14-3-3ζ or being detached and exposed to the solvent. Populations of these conformational states as obtained from DF/HRE-MD (in red) and MD (in black) simulations are listed under the cartoon figures. B) 14-3-3 model starting conformation, coloured according to the backbone RSMF, where blue and red represent the least and most flexible amino-acids, respectively. C) The atom-positional root-mean-square fluctuations (RMSF) of the 14-3-3ζ backbone atoms.
Fig 6.
Changes in the 14-3-3ζ monomer groove width.
A) Models of 14-3-3ζ at different levels of opening (the green, blue and orange colours mark closed, open and wide-open states, respectively), the last 3 helices are marked with a darker colour. B) Representative replicas of five different stages along the binding pathway (with the replica ID number shown in brackets) from DF/HRE-MD simulation dim_p1hT. The panel shows groove width distributions whilst the phosphopeptide moves from the binding site (IS1) to the unbound state (Unb). C) Average groove width distributions of 14-3-3ζ monomers as obtained from DF/HRE-MD bound (IS1) and unbound (Unb) states, and unrestrained MD simulations in holo (bound) and apo (unbound) states.
Fig 7.
A) The top view of the protein dimer, with the twist angle between the monomers displayed in dark blue. B) Probability distribution of inter-monomers twist angles averaged over all DF/HRE-MD simulations in the IS1 (bound) and Unb (unbound) stages, and for the whole pathway.
Fig 8.
Phosphopeptide p1h secondary structure.
Changes in the backbone structure are shown during DF/HRE-MD simulation dim_p1Ht (panel A) and unrestrained MD (panels B and C) simulations, analysed by the DISICL algorithm. The change in the average secondary structure content during the DF/HRE-MD simulation is shown in the middle of panel A, whilst the most dominant conformations in bound and unbound states are tabulated on the left and right side, respectively. Representative conformations for the bound and unbound states are depicted on the left and right sides, respectively, where the residues are coloured according to secondary structure classification. Intra-molecular hydrogen bonds are depicted as dashed lines. The tables besides the depictions show the 5 most populated secondary structure classes for the corresponding MD simulation and the appropriate (bound(IS1)/unbound) stage of the DF/HRE-MD simulation, respectively. The most populated DISICL classes are depicted in the following colours: π-helix (PIH)–cyan, Extended β-strand (EBS)–red, normal β-strand (NBS)–orange, polyproline-like (PP)–brown, turn type VIII (TVIII)–indigo, Gamma turn (GXT)–maroon, β-cap (BC)–gold, helix-cap (HC)–blue, turn-cap (TC)–black. DISICL secondary structure elements are listed in S1 Table.
Fig 9.
Free energy profiles of the DF/HRE-MD simulations.
A) The raw free-energy profiles, derived from WHAM analysis, as function of simulation time per replica (every ns) for the simulation tmon_p2H. B) Convergence of the free energy difference between the unbound and bound state of all eight DF/HRE-MD simulations. C) Final free energy profiles (ΔAwham) for the eight 14-3-3ζ peptide-binding simulations.
Table 1.
Details of the DF/HRE-MD simulation of tmon_p2H.
The table contains quantities for every replica that are required for the free energy calculations. The columns show the state assignment, the replica number, the sampled volume (Vsampled), replica exchange probability (Pex), reaction coordinate value (λ), ideal distancefield distance (l0), accessible volume (V(l)) and raw free-energy according to the free-energy profile (Awham). V(l) and Awham were calculated based on the restraining distance assigned to the replica. The last row contains the unbound volume (Vunb), calculated from the volume sampled by the phosphopeptide in all Unb state replicas, and the volume of simulation box (Vbox).
Table 2.
Comparison of the experimental and calculated binding free energies.
The ΔGexp shows the experimental free energies calculated from dissociation constants [8–10]. ΔAbind(mon) shows the binding free energy calculated from the tmon DF/HRE-MD simulations (20ns/replica).
Fig 10.
Graphical representation of pathway mapping in simulations.
A) Distancefield grid, where grid points are coloured according to distancefield (DF) distance. B) Direct distance (as a black dashed line), and DF distance (along the coloured grid points) between the ligand and the protein binding side. Grid points located within the protein (which should be avoided) are shown in purple. C-E) Grid points sampled during a replica exchange simulation, coloured according to their position along the pathway. Panels C and D are showing grid points at different levels of relative probability per replica; with all visited grid points in Panel C, and only the often-visited grid points per replica in Panel D. Panel E shows the derived peptide-binding pathway, with one peak maximum per replica. The surface of the protein is shown in grey.