Coevolving residues distant from the ligand binding site are involved in GAF domain function

Figures

• 

  Fig. 1. Coevolving residue positions and their role in GAF domain allostery. Schematic showing an approach to determine coevolving residues and delineating their role in the allosteric regulation of GAF domains. GAF domain sequences were aligned, and statistical coupling analysis (SCA) was performed to detect coevolving residue positions. The scoring matrix obtained from SCA was then used for clustering to determine the cluster of coevolving residues with highest scores, which were then mapped on to a GAF domain structure. MD simulations followed by biophysical assays with biosensors were then used to determine the role of individual coevolving residue positions in the allosteric regulation of GAF domains.

• 

  Fig. 2. SCA reveals a cluster of coevolving residues in GAF domains. A Color-coded heatmap showing pairwise scores of individual residue positions in GAF domains obtained from SCA. The positions of the nine coevolving residues are indicated. The area in the heatmap where the coevolving residue positions intersect is highlighted by a white box. B Cartoon representation of the GAFa domain of PDE5 (PDB: 2K3122) showing the coevolving cluster of residues (spheres in light blue) in the GAF domains. C Cartoon representation of the GAFa domain of PDE5 showing residues (side chains as sticks in pink) within a 5 Å distance from cGMP (left panel), coevolving residues (spheres in light pink; rest of the coevolving residues in light blue) within 5 Å distance from cGMP (middle panel) and coevolving residues L267 and F295 selected for further functional characterization (dark blue spheres, right panel).

• 

  Fig. 3. Mutation of distant, coevolving residue positions induces dynamic structural changes in the GAFa domain of PDE5. A Cartoon representation of GAFa domain (WT, left panel; L267A mutant, middle panel; and F295A mutant, right panel) dynamics across the 1000 ns-long MD simulation runs captured using a composite image of snapshots that are 200 ns apart. The ligand is shown as blue carbon sticks, and positions 267 and 295 are shown as green carbon sticks in each complex. B Graphs showing root mean squared deviation (RMSD) measurements of the apo and holo GAFa domains (WT and mutants) across the 1000 ns-long MD simulation runs. C Schematic (left panel) and graphs showing root mean square fluctuation (RMSF) measurements of the apo and holo GAFa domains (WT and mutants) across the 1000 ns-long MD simulation runs (right panel). D Graphs showing radius of gyration (Rg) values of the apo and holo GAFa domains (WT and mutants) across the 1000 ns-long MD simulation runs. E Schematic highlighting residues L267 and F295 (left panel) and graphs showing center-of-mass distance between amino acid positions 267 and 295 in the apo and holo GAFa domains (WT and mutants; right panel). Note how the…

• 

  Fig. 4. Mutation of the distant, coevolving residues alters cGMP-induced allosteric conformational change in the isolated GAFa domain of PDE5. A Cartoon representation of apo (PDB: 3MF016) and holo (PBD: 2K3122) PDE5 GAFa domain structures aligned using all Cα-atoms (left panel) and using Cα-atoms of α2 helix (right panel) highlighting structural changes (apo: gray, holo: black) induced upon cGMP binding. B Schematic showing the BRET2-based GAFa domain conformational biosensor that shows increased energy transfer upon cGMP binding likely due to the structural juxtaposition of helix α5 against helix α2, and thus, bringing the BRET donor and acceptor closer. C Cartoon representation of the PDE5 GAFa domain highlighting the distant, coevolving residues L267 and F295 (spheres in blue) that were mutated to A. The ligand, cGMP, is shown in the stick representation. D Western blot analysis of the cell lysates prepared from HEK293T cells transfected with either the WT or mutant GAFa domain biosensor plasmids and probed using an anti-GFP antibody showing the expression of biosensor constructs. Whole blot image used for generating the figure is shown in Supplementary Fig. 4. E Graphs showing…

• 

  Fig. 5. Mutation of distant, coevolving residues alters cGMP-induced allosteric regulation in the full-length PDE5. A Schematic showing the mechanism of the BRET2-based, full-length PDE5A2 conformational biosensor in the absence and presence of cGMP. B Western blot analysis of the cell lysates prepared from HEK293T cells transfected with either the WT or mutant full-length PDE5A2 biosensor plasmid constructs and probed using an anti-GFP antibody showing the expression of biosensor constructs. Whole blot image used for generating the figure is shown in Supplementary Fig. 6. C Graphs showing bioluminescence spectra obtained from lysates prepared from cells expressing the WT and L267A and F295A mutant PDE5 biosensor constructs. Note the reduction in the GFP2 emission peaks in the L267A and F295A mutant PDE5 biosensors. Data shown are mean ± S.D. from a representative experiment, with experiments performed three times. D Bar graph showing the basal BRET values of the WT, L267A, and F295A mutant PDE5 biosensors. Note the significant reduction in the basal BRET values of the L267A and F295A mutants compared to the WT GAFa domain. E Graphs showing percentage decrease in BRET values of the WT,…

• 

  Fig. 6. Mutating the distant coevolving residue positions affects the function of other ligand-binding GAF domains. A Cartoon representation of the miRFP670nano3 GAF domain (PDB: 7LSC) highlighting distant coevolving residues equivalent to PDE5 L267 and F295 (L100 and W125, respectively) and the ligand, biliverdin (BV). B Western blot analysis of the cell lysates prepared from HEK293T cells transfected with either the WT or mutant miRFP670nano3 plasmid constructs and probed using an anti-FLAG antibody showing the expression of biosensor constructs. Whole blot image used for generating the figure is shown in Supplementary Fig. 8. Graphs showing relative fluorescence (C) and bioluminescence (D) of the WT, L100A and W125A mutant miRFP670nano3. Data shown are mean ± S.D. obtained from five independent experiments. The p-values shown in (C) were obtained from Student’s t-test (two-tailed, unpaired, equal variance). For all tests, a p-value of <0.05 was considered statistically significant.

• 

  Fig. 7. Altered ligand binding site and allosteric activation upon mutation of distant, coevolving residues in the PDE5 GAFa domain. A Cartoon representation of the GAFa domain highlighting helix α4 and sheet β3 from the ligand binding site that “closes” upon cGMP binding (left panel) and N-terminal helix α2 and C-terminal helix α5 that move closer upon cGMP binding (right panel). B Graphs showing probability distribution (kernel density estimation) of distances between helix α4 and sheet β3 (x-axis) and N-terminal helix α2 and C-terminal helix α5 (y-axis) of the apo (upper panel) and holo (lower panel) GAFa domain (WT and mutants) determined from a cumulative of three independent, 1000 ns-long MD simulation trajectories. Color bars, probability densities.