Optimizing variant-specific therapeutic SARS-CoV-2 decoys using deep-learning-guided molecular dynamics simulations
Katharina Köchl, Tobias Schopper, Vedat Durmaz, Lena Parigger, Amit Singh, A. Krassnigg, Marco Cespugli, Wei Wu, Xiaoli Yang, Yanchong Zhang, Welson Wen-Shang Wang, Crystal Selluski, Tiehan Zhao, Xin Zhang, Caihong Bai, Leon C. W. Lin, Yuxiang Hu, Zhiwei Xie, Zaihui Zhang, Jun Yan, Kurt Zatloukal, Karl Gruber, Georg Steinkellner, Christian Gruber
Abstract
Abstract Treatment of COVID-19 with a soluble version of ACE2 that binds to SARS-CoV-2 virions before they enter host cells is a promising approach, however it needs to be optimized and adapted to emerging viral variants. The computational workflow presented here consists of molecular dynamics simulations for spike RBD-hACE2 binding affinity assessments of multiple spike RBD/hACE2 variants and a novel convolutional neural network architecture working on pairs of voxelized force-fields for efficient search-space reduction. We identified hACE2-Fc K31W and multi-mutation variants as high-affinity candidates, which we validated in vitro with virus neutralization assays. We evaluated binding affinities of these ACE2 variants with the RBDs of Omicron BA.3, Omicron BA.4/BA.5, and Omicron BA.2.75 in silico. In addition, candidates produced in Nicotiana benthamiana , an expression organism for potential large-scale production, showed a 4.6-fold reduction in half-maximal inhibitory concentration (IC 50 ) compared with the same variant produced in CHO cells and an almost six-fold IC 50 reduction compared with wild-type hACE2-Fc.