GENERALIZED CORRELATION OF PAN-CORONAVIRUS 2′-O- METHYLTRANSFERASES

The coronavirus nsp16 protein encodes the 2’-O-ribose methyltransferase, an

enzyme whose activity relies on the formation of the nsp16/nsp10 heterodimer. The

nsp16 transfers a methyl group to the 2’-O position of the first nucleotide of viral RNA

at the 5’ end, generating the cap-1 structure. By capping the viral RNA, the complex

masks the genetic material, facilitating immune evasion and supporting viral

replication.

In this study, we analyzed dynamic correlations between nsp10 and nsp16

proteins from SARS-CoV (PDBid 3R24), MERS-CoV (PDBid 5YNM), and SARS-CoV-

2 (PDBid 6WKS). Missing residues were reconstructed with Modeller, and specific

deprotonation mutations in cysteines coordinating Zn²⁺ ions were applied with psfgen.

Molecular dynamics simulations were performed in NAMD3 with the CHARMM36

force field. Production runs consisted of 50 replicas of 100 ns each. Dynamic network

analyses were conducted with DyNetAn to identify residues that move cooperatively,

thereby revealing potential allosteric regulation sites. To assess whether residues

dynamics differed substantially between species, PCA and MDS of the correlation

matrices were applied using scikit-learn.

Results revealed distinct correlation patterns across species. MERS-CoV

exhibited stronger internal correlations (0.50–0.75) but weaker interface interactions

(~0.30), suggesting robust intraprotein communication but limited interprotein

dynamics. SARS-CoV showed higher internal correlations (0.75–0.80) and moderately

stronger interface interactions (~0.40), indicating greater flexibility. SARS-CoV-2

displayed a profile more similar to MERS-CoV, with dominant internal correlations and

limited interface communication, though with a slightly broader range of internal

correlation ranges.