encompass a diverse family of animal viruses that possess a large linear double-stranded DNA genome (120 up to 230 kbp), share a common virion morphology, and retain the capacity to enter a latent phase of infection in particular host cells (24
). Historically herpesvirus species, including eight known to infect humans (HHV-1 to HHV-8), have been classified according to these biological characteristics into three major subfamilies: Alphaherpesvirinae
, and Gammaherpesvirinae
. Recent analysis of herpesviral gene content and sequence similarity substantiates this classification scheme and defines a subset of common genes (43 core sequences) thought to be involved in fundamental viral processes such as capsid assembly and egress from the nucleus, as well as DNA replication, processing, and packing into assembled virions (8
). Of the core herpesviral genes, only one (UL24
in HHV-1) remains unassigned to any broad functional category. Using Meta-BASIC (http://basic.bioinfo.pl
), a highly sensitive fold recognition method that applies a comparison of sequence profiles enriched by predicted secondary structure, we predict that the UL24
gene encodes a novel PD-(D/E)XK endonuclease belonging to a large superfamily of restriction endonuclease-like fold proteins.
A consensus UL24 family (PFAM [3
] accession no. PF01646) sequence maps with an above-threshold Meta-BASIC score (Z-score, 12.22) to a domain of unknown function, DUF91 (PFAM accession no. PF01939), recently identified as possessing a restriction endonuclease-like fold (17
). In addition, for the majority of UL24 family members Meta-BASIC provides reliable hits (Z-scores above 12) to several PD-(D/E)XK endonuclease families such as DUF91, DUF911 (PFAM accession no. PF06023) (17
), or SfsA (PFAM accession no. PF03749) (18
). Further analysis performed using a Meta Server (http://bioinfo.pl/meta
) that combines several top-of-the-line fold recognition methods revealed additional weak hits (for both family consensus and HHV-1 UL24 sequences) to several structures with a restriction endonuclease-like fold: Holliday junction-resolving enzymes (PDB
1hhl and PDB
1ipi), DNA mismatch repair protein (PDB
1azo), and DNA restriction endonuclease NaeI (PDB
1ev7). The correct but highly nontrivial fold assignment includes good mapping of predicted (mainly with PSI-PRED) (16
) and observed secondary structure elements, general conservation of hydrophobicity patterns, and absolute conservation of the signature PD-(D/E)XK endonuclease motifs critical for function (Fig. ).
FIG. 1. Multiple sequence alignment for selected UL24 and UL12 family representatives and Holliday junction-resolving structures. UL24 and UL12 family sequences are labeled according to NCBI gene identification (gi) number followed by an abbreviation of the species (more ...)
Existing classification of PD-(D/E)XK endonuclease structures groups a number of families, including several restriction endonucleases (EcoRI, EcoRII, BamHI, BglI, Cfr10I, NaeI, etc.), DNA repair enzymes (MutH and Vsr), Holliday junction resolvases (Hjc and Hje), and other nucleotide-cleaving enzymes, into a large and diverse superfamily of restriction endonuclease-like fold proteins (20
). These enzymes contribute to important biological functions such as protecting against foreign DNA (restriction endonucleases), repairing damaged DNA (MutH and Vsr), resolving Holliday junctions (endonuclease I, Hjc, Hje, and XPF/Rad1/Mus81-dependent nuclease), or recombining DNA (lambda exonuclease and TnsA) (4
). The common core structure (αβββαβ topology) includes a four-stranded, mixed β-sheet flanked by an α-helix on both sides. The PD-(D/E)XK signature (2
) provides active site residues responsible for cleaving a variety of nucleic acid substrates.
UL24 homologs encompass an N-terminal restriction endonuclease-like domain (ααβββαβββα topology, with the common core elements underlined) followed by a low-complexity, highly basic region that is probably unstructured. UL24 family proteins include all restriction endonuclease-like core elements (with a short α-helical insertion between the N-terminal core α-helix and first β-strand) in addition to three C-terminal elements (two β-strands and α-helix) that probably extend the core of the fold. UL24 sequences retain absolute conservation of the PD-(D/E)XK signature (motifs II and III, Fig. ). Several additional UL24 invariant residues (H and Y in motif I, Q in motif IV, and E and R preceding motif II; highlighted in blue in Fig. ) map near the presumed catalytic residues and are poised to supplement the active site architecture, perhaps contributing to the substrate specificity of the enzyme family.
Assignment of UL24 to the PD-(D/E)XK endonuclease superfamily provides new insight into its functional role in herpesviral replication. Initially, the UL24
gene was identified in the HSV-1 genome as an open reading frame overlapping the gene for thymidine kinase (tk
). Disruption of conserved UL24
sequence elements [including the first D of the signature PD-(D/E)XK motif] was shown to correlate with a small syncytial plaque phenotype observed in mutants initially designed to study tk
). Subsequent analysis of UL24
mutants with wild-type tk
activity substantiated the small-plaque phenotype and showed less efficient replication in the eye (10- to 30-fold reduction) as well as severe impairment of replication in ganglia (14
). These phenotypes support a general and nonessential role for UL24 activity in mediating viral replication and perhaps membrane fusion events.
Interestingly, another PD-(D/E)XK nuclease (alkaline nuclease) is encoded in the HHV-1 genome by the UL12
). Like the UL24
gene product, UL12
is not essential for herpesviral replication (26
), although its mutation decreases viral DNA synthesis and processing. UL12 exonuclease activity mediates strand exchange (in vitro) in DNA recombination events thought to be an integral part of herpesviral replication (25
). Although UL12 and UL24 do not share any significant sequence similarity [apart from the PD-(D/E)XK signature], these two enzymes might perform redundant nucleotide cleavage activities essential for low levels of herpesviral replication.
The universal presence of UL24 in completed avian-mammalian and reptilian herpesviruses (although not detected in amphibian-fish and invertebrate herpesviruses) indicates a fundamental role of this protein in the viral life cycle. Identification of UL24 as a potential PD-(D/E)XK endonuclease suggests that this role might involve cleaving nucleic acid substrate. Consistent with this hypothesized activity, UL24 protein localizes to the nuclei of infected cells (12
). Accordingly, UL24 could participate in homologous recombination of viral DNA, a process intimately associated with replication and thought to drive herpesviral evolution (25
). Alternatively, UL24 could mediate linear-to-circular genome transitions that emerge in herpesviral latency (13
) or might recognize and resolve specific genome structures that result from concurrent replication and recombination. Finally, UL24 could act on host genetic material, perhaps to trigger cellular DNA damage response machinery shown to accumulate at viral replication centers (19
). Ultimately, experimental investigations should address the predicted UL24 PD-(D/E)XK endonuclease activity and clarify potential substrates, allowing further insight into the fundamental role of this protein in herpesviral latency and replication.