Residue‐level determinants of angiopoietin‐2 interactions with its receptor Tie2

Abstract

We combined computational and experimental methods to interrogate the binding determinants of angiopoietin‐2 (Ang2) to its receptor tyrosine kinase (RTK) Tie2—a central signaling system in angiogenesis, inflammation, and tumorigenesis. We used physics‐based electrostatic and surface‐area calculations to identify the subset of interfacial Ang2 and Tie2 residues that can affect binding directly. Using random and site‐directed mutagenesis and yeast surface display (YSD), we validated these predictions and identified additional Ang2 positions that affected receptor binding. We then used burial‐based calculations to classify the larger set of Ang2 residues that are buried in the Ang2 core, whose mutations can perturb the Ang2 structure and thereby affect interactions with Tie2 indirectly. Our analysis showed that the Ang2‐Tie2 interface is dominated by nonpolar contributions, with only three Ang2 and two Tie2 residues that contribute electrostatically to intermolecular interactions. Individual interfacial residues contributed only moderately to binding, suggesting that engineering of this interface will require multiple mutations to reach major effects. Conversely, substitutions in substantially buried Ang2 residues were more prevalent in our experimental screen, reduced binding substantially, and are therefore more likely to have a deleterious effect that might contribute to oncogenesis. Computational analysis of additional RTK‐ligand complexes, c‐Kit‐SCF and M‐CSF‐c‐FMS, and comparison to previous YSD results, further show the utility of our combined methodology.

Structural design principles for specific ultra-high affinity interactions between colicins/pyocins and immunity proteins

“…Energy calculations to map residue-level specificity determinants. We followed the methodology described previously [55][56][57][58][59][60][61]68 to analyze the per-residue contributions of residues in colicins/pyocins and immunity proteins to interactions with their partners in the crystal structures mentioned above. The Finite Difference Poisson-Boltzmann (FDPB) method, as implemented in DelPhi 69 , was used to calculate the net electrostatic/ polar contributions (ΔΔG elec ) of each residue found within 15 Å of the dimer interface in each complex.…”

“…To map the individual residues that contribute to colicin/pyocin-immunity protein interactions, we analyzed the representative X-ray structures of colicins bound to their immunity proteins, using an energybased computational methodology developed previously by our lab [55][56][57][58][59][60][61] . As described in the "Materials and methods" sectio, we calculate the net electrostatic/polar contributions (ΔΔG elec ) of each residue ≤ 15 Å of the colicin-immunity protein interfaces.…”

Structural design principles that underlie the multi-specific interactions of Gαq with dissimilar partners

“…We followed the methodology described previously 45–47,53–55,64 to analyze the per-residue contributions of Gα q residues to interactions with their partners (PLC-β3, p63RhoGEF, RGS2/8, GRK2) in the crystal structures mentioned above. The Finite Difference Poisson–Boltzmann (FDPB) method, as implemented in DelPhi 65 , was used to calculate the net electrostatic and polar contributions (ΔΔG elec ) of each residue found within 15 Å of the dimer interface.…”

“…Non-polar energy contributions (ΔΔG np ) were calculated as a surface-area proportional term by multiplying the per-residue surface area buried upon complex formation, calculated using surfv 66 , by a surface tension constant of 0.05 kcal/mol/Å 2 56 . Residues contributing ΔΔG np ≥ 0.5 kcal/mol to the interactions (namely, those that bury more than 10 Å 2 of each protein surface upon complex formation) were defined as making substantial non-polar contributions 64 . To reduce false positives and negatives, we applied a consensus approach across comparable biological replicates in multiple PDB structures (3OHM, 4QJ3, 4QJ4, 4QJ5 – see Supplementary Fig.…”