Beyond Small Molecules: Computational and AI-Assisted Protein and Peptide Engineering Against the SARS-CoV-2 Nsp12–Nsp7–Nsp8 Replication Complex

Antiviral development against SARS-CoV-2 has focused largely on small-molecule inhibitors of the main protease and RNA-dependent RNA polymerase and on antibodies targeting spike. Antigenic evolution has reduced the activity of many spike-directed antibodies, while resistance-associated substitutions have been identified for enzyme inhibitors. The nsp12–nsp7–nsp8 replication complex offers a complementary target: its subunit interfaces support assembly and processive RNA synthesis and may be accessible to engineered proteins and peptides. This review examines the structural and functional evidence for targeting these interfaces, the reported peptide and protein inhibitors, and the relevance of deep-learning-based binder design. The literature reviewed here reveals a methodological gap. Replication-complex inhibitors have predominantly been developed through molecular dynamics, interface analysis, docking, and structure-guided engineering, whereas contemporary generative design methods have been validated on other targets, including spike. These precedents support investigation of generative binders against the replication complex but do not establish their antiviral efficacy or resistance barrier. A proposed development framework integrates structural ensembles, interface energetics, sequence variation, and mechanistic validation. Intracellular delivery, target accessibility, proteolytic stability, and resistance remain important barriers to translation.