Network of coordinated substitutions in the Hepatitis B virus polymerase

We analyzed coordinated variation among amino acid sites in 370 HBV polymerase sequences using Bayesian networks. Among the HBV polymerase domains the spacer domain separating terminal protein from the reverse-transcriptase domain, showed the highest network centrality. Coordinated substitutions preserve the hydrophobicity and charge of Spacer. Maximum likelihood estimates of codon selection showed that Spacer contains the highest number of positively selected sites. Identification of 67% of the domain lacking an ordered structure suggests that Spacer belongs to the class of intrinsically unstructured domains and proteins whose crucial functional role in the regulation of transcription, translation and cellular signal transduction has only recently been recognized. Spacer plays a central role in the epistatic network associating substitutions across the HBV genome, including those conferring viral virulence, drug resistance and vaccine escape.