SIE-SCWRL, FoldX, Rosetta

AMBER, Rosetta

The atomic coordinates of the A26.8 VHH bound to the C-terminal portion of TcdA were taken from the structure of the A26.8H6-TcdA-A2 complex crystallized at pH 6.5 (PDB ID: 4NC0), and were used as a starting point for virtual affinity maturation. Two versions of the complex were prepared, differing by exclusion (preparation 1) or inclusion (preparation 2) of the C-terminal G262 of TcdA and N-terminal Q1 of the VHH, which are not visible in the crystal structure but may affect the calculated interactions in the complex. All TcdA amino-acid residues preceding A123, which are distant from the VHH, and the His-tag residue H125 at the C-terminus of the VHH, were deleted from the crystal structure. Hydrogen atoms were added to the resulting complex and adjusted for maximizing H-bonding interactions. Structural refinement of the complex was then carried out by energy-minimization using the AMBER force-field, with a distance-dependent dielectric and an infinite distance cutoff for non-bonded interactions. Non-hydrogen atoms were restrained at their crystallographic positions with harmonic force constants of 40 and 10 kcal/(mol.A2) for the backbone and side-chain atoms, respectively. The ADAPT platform was then used for affinity maturation. In the first round of affinity optimization, single-point scanning mutagenesis simulations were carried out at several positions within the CDRs of VHH A26.8. We used three protocols, SIE-SCWRL, FoldX, and Rosetta, for building the structures and evaluating the energies of single-point mutations to 17 other possible natural amino acids (Cys and Pro were excluded) at these positions of the parental sequence. A consensus approach over specific versions of these three protocols was applied for building and scoring the VHH mutants. Scoring of binding affinity was mainly based on the average Z-score and also on the average rank score over the scores calculated with the three component energy functions, SIE, FoldX-FOLDEF, and Rosetta-Interface. Further technical and implementation details of this consensus approach and its component methods can be found in J. Chem. Inf. Model. 56, 1292–1303 (2016). Prior to binding affinity predictions, the FoldX-FOLDEF energy function was used to estimate the effect of substitutions on the internal stability of the VHH structure. Thus, mutations predicted to be destabilizing by introducing folding free energy changes larger than 2.71 kcal/mol (i.e., 100-fold increase of unfolding equilibrium constant) relative to the parental molecule were discarded from further evaluation. In the second round of optimization, double- and triple-point VHH mutants were generated from combinations of the lead single-point mutations selected after experimental validation, and scored using the same computational protocol as for the single-point mutants.