Genomewide function conservation and phylogeny in the Herpesviridae

Abstract

The Herpesviridae are a large group of well-characterized double-stranded DNA viruses for which many complete genome sequences have been determined. We have extracted protein sequences from all predicted open reading frames of 19 herpesvirus genomes. Sequence comparison and protein sequence clustering methods have been used to construct herpesvirus protein homologous families. This resulted in 1692 proteins being clustered into 243 multiprotein families and 196 singleton proteins. Predicted functions were assigned to each homologous family based on genome annotation and published data and each family classified into seven broad functional groups. Phylogenetic profiles were constructed for each herpesvirus from the homologous protein families and used to determine conserved functions and genomewide phylogenetic trees. These trees agreed with molecular-sequence-derived trees and allowed greater insight into the phylogeny of ungulate and murine gammaherpesviruses.

Figure 1

Schematic representation of a homologous protein family (HPF). Identically shaded boxes represent identified regions of each protein that have sequence homology. HPFs are constructed computationally by identifying one (or more) region(s) of sequence homology (i.e., unfilled oval) that builds the largest group of sequences. The HPF-conserved sequence region can be found in all proteins in the HPF.

Figure 2

Phylogenetic profile of homologous proteins families (HPF) known to be involved in structural functions (capsid, tegument, virus assembly). The presence of the family in any genome is indicated by 1 and the absence by 0. Alpha, Alphaherpesviruses; Beta, Betaherpesviruses; and Gamma, Gammaherpesviruses. The HPF numbers are indicated and relate directly to accompanying data available at http://www.biochem.ucl.ac.uk/bsm/virus_database. HPF 1^* is indicated twice because it represents a shared domain present in both the major and minor capsid proteins of all herpesviruses.

Figure 3

Distribution of functional classes of homologous families across the 19 herpesviruses considered. The number of herpesvirus genomes that contain the functional class are shown on the X-axis (lowest graph) and the number of homologous families in the functional class are shown on each Y-axis.

Figure 4

Phylogenetic trees based on protein family phylogenetic profiles (A,B,C) or on sequence comparison of herpesvirus-conserved domains (D). The initial data or alignments were bootstrapped 100 times. Neighbor-joining trees were constructed using the fraction of nonshared homologous families (A) or the fraction of dissimilar homologous families (B), as described in the text. Maximum parsimony cladogram built from the phylogenetic profiles (C). Neighbor-joining tree constructed using conserved regions in 26 herpesvirus open reading frames (D). α, Alphaherpesviruses; β, Betaherpesviruses; and γ, Gammaherpesviruses. The rhadinovirus subgroup of the Gammaherpesviruses consists of the viruses MHV-68, AHV-1, HVA-3, HVS, RRV, HHV-8, and EHV-2.