Following the αA helix, the N-terminal region preceding the αB helix has been most frequently identified as a region of dynamic allostery in the PDZ domain. Here, the researchers refer to this region as the αB lower-loop. A previous study has recognized dynamic allostery at the αB lower-loop through various techniques, including NMR, Monte-Carlo sampling, and PRS. Taken together, the various approaches have identified dynamic allostery in PTP-BL PDZ2 and PSD-95 PDZ3.
3.1. Agreement between Experimental and Computational Techniques: PTP-BL PDZ2
Taken together, experimental and computational techniques have pointed to the dynamic allostery at the αB lower-loop of PTP-BL PDZ2. In 2004, Fuentes et al. used NMR experiments to show that residues V61, V64, L66, A69, T81, and V85 of the αB lower-loop have increased flexibility upon ligand binding
[27]. In a continuation of their original work
[35], they performed point mutations to key residues in ligand binding to explore dynamic changes. NMR revealed that I20F and I35V mutants induce dynamic changes to V58, V61, V64, A69, A74, L78, and V85. In another experimental study, Gianni et al. performed double mutant cycles to reveal that R86 is coupled to the ligand
[57]. It is worth noting that Fuentes et al. performed these experiments with the RA-GEF2 ligand, which is 15 residues in length. While the residues at positions P
0 and P
−2 are primarily responsible for interactions with the PDZ domain, a previous study has revealed that positions P
−4, P
−5 and P
−6 can also directly interact with the PDZ domain at the region of the αB lower-loop
[62]. This points to the importance of exploring the dynamics at the αB lower-loop in PDZ complex systems with ligands of longer length, such as RA-GEF2. In addition to experimental techniques, various computational approaches have also identified dynamic allosterism at the αB lower-loop of PTP-BL PDZ2.
As described above, Dhulesia et al.
[29], Kong et al.
[46], and Morra et al.
[31] each performed equilibrium MD simulations to identify dynamic allosterism in PTP-BL PDZ2. With NMR restraints
[27][51][52], Dhulesia et al. observed the motion of the αB lower-loop (V61, V64, L66, A69, and T81) increasing and decoupling from the motion of the αB stand upon binding to RA-GEF2
[29]. While Kong et al. and Morra et al. did not restrain their simulations, they also identified significant modulations to PTP-BL PDZ2 upon ligand binding. Kong et al. noted a cluster of residues associated with ligand binding, including G68, A69, L78, T81, and G82 of the αB lower-loop
[46]. Morra et al. observed both significant changes in energetic modulations (V61, N62, G63, V64, T70, H71, L72, Q73, A74, V75, E76, T77, L78, and V85) and dynamic fluctuations (V58, L59, A60, V75, E76, T77, and L78) at this region
[31]. Cilia et al. performed Monte-Carlo sampling to echo these results
[48]. Their sampling identified V58, L59, V61, L66, A74, V75, T77, L78, T81, and V85 as a cluster of dynamically affected residues.
In addition to traditional simulations, a variety of other computational approaches have been used to explore dynamic allosterism at αB lower-loop in PTP-BL PDZ2. Gerek et al. applied PRS to show that applied forces to specific residues in PTP-BL PDZ2 resulted in a relative displacement of the αB lower-loop (V58, L59, A60, V61, L64, L66, A69, H71, A72, Q73, A74, V75, E76 T77, L78, R79, N80, T81, and V85)
[50]. They noted a pathway through S17 of the carboxylate binding loop by which the αB lower-loop was coupled to ligand binding. Combining PSN and ENM, Raimondi et al. identified communication pathways within PTP-BL PDZ2
[53]. They identified a cluster of residues (V61, N62, L66, A69, T70, H71, Q73, V75, L78, R79, T81, Q83, and V85) that are correlated to the RA-GEF2 ligand. Lastly, Kalescky et al. explored the role of each residue in the global dynamics of the protein using rigid-residue molecular dynamics and entropy analysis
[54]. They identified several residues including V61, L78, T81, and V85 that have a key role in the allosteric network. Each study described above is summarized in
Figure 4. The summarized works point to the importance of V61, V64, L66, A69, A74, L78, T81, and V85 in ligand-induced dynamic allosterism at the αB lower-loop.
Figure 4. Ligand-induced dynamic allostery at the αB lower-loop in the PTP-BL PDZ2 (PDB ID: 3PDZ) domain as observed in experimental and computational studies. (
a) NMR chemical shifts by Fuentes et al.
[27]; (
b) point mutations and NMR chemical shifts by Fuentes et al.
[35]; (
c) equilibrium MD with NMR restraints by Dhulesia et al.
[29]; (
d) equilibrium MD by Kong et al.
[46]; (
e) equilibrium MD by Morra et al.
[31]; (
f) Monte-Carlo sampling by Cilia et al.
[48]; (
g) perturbation response scanning (PRS) by Gerek et al.
[50]; (
h) protein structure network (PSN) and elastic network model (ENM) by Raimondi et al.
[53]; and (
i) rigid-residue MD by Kalescky et al.
[54]. The neighboring table displays the sequence fragment of the αB lower-loop in the PTP-BL PDZ2 domain with allosteric residues colored accordingly. Note that the colored residues in the table directly correspond to the colored residues in the structural representations shown above.
3.2. Computational Conclusions Lacking Experimental Support: PSD-95 PDZ3
While computational results pointing to ligand-induced dynamic allosterism in the αB lower-loop of PTP-BL PDZ2 have been validated by NMR experiments, computational results pointing to ligand-induced dynamic allosterism in the αB lower-loop of PSD-95 PDZ3 have not been so consistent. As described above, original experimental work on the PSD-95 PDZ3 domain specifically noted the lack of conformational changes to the protein backbone upon ligand binding
[17]. Since then, various computational studies have suggested that the αB lower loop may have a role in the propagation of dynamic allostery.
Normal mode analysis (NMA) of the PSD-95 PDZ3 domain showed a shift of the αB helix that more widely opened the binding pocket to permit ligand binding
[22]. Additionally, CMCA identified a correlated network including various residue at the αB lower-loop (I359, L360, V362, G364, D366, and N369)
[58]. PRS has revealed that applied forces on PSD-95 PDZ3 resulted in relative displacements of the αB lower loop and pointed to I359, S361, V362, L367, and H372 as key residues in the pathway of allosteric propagation
[50]. Taken together, complete single-mutation scan coupled with SCA identified residues I359, V362, L367, and H372 as being significantly correlated with ligand binding
[56]. As previously described, Morra et al. performed equilibrium MD simulations to reveal significant energetic modulations to PSD-95 PDZ3 at residues I359, L360, S361, V362, N363, and H72 upon ligand binding
[31]. Finally, Kalescky et al. performed non-equilibrium rigid-body MD simulations to consider the contribution of each residue to global protein dynamics
[47]. H372 was identified as one of five “wire” residues that are responsible for the propagation of allosteric signal. It should be noted that while the approaches described the above point to dynamic allosterism at the αB lower loop, other computational studies on PSD-95 PDZ3 failed to recognize this allosteric region
[21][49][57]. Each study described above is summarized in
Figure 5. The summarized studies point to the importance of I359, V362, L367, and H372 in ligand-induced dynamic allosterism at the αB lower loop.
Figure 5. Ligand-induced dynamic allostery at the αB lower-loop in the PSD-95 PDZ3 (PDB ID: 1BE9) domain as observed in experimental and computational studies. (
a) Conservative-mutation correlation analysis (CMCA) by Du et al.
[58]; (
b) perturbation response scanning (PRS) by Gerek et al.
[50]; (
c) statistical coupling analysis (SCA) coupled with a complete single-mutation scan by McLaughlin et al.
[56]; (
d) equilibrium MD by Morra et al.
[31]; and (
e) rigid body MD by Kalescky et al
[47]. The neighboring table displays the sequence fragment of the αB lower-loop in the PSD-95 PDZ3 domain with allosteric residues colored accordingly. Note that the colored residues in the table directly correspond to the colored residues in the structural representations on the left.