Antagonism of Protein Kinase R by Large DNA Viruses

Abstract

Decades of research on vaccinia virus (VACV) have provided a wealth of insights and tools that have proven to be invaluable in a broad range of studies of molecular virology and pathogenesis. Among the challenges that viruses face are intrinsic host cellular defenses, such as the protein kinase R pathway, which shuts off protein synthesis in response to the dsRNA that accumulates during replication of many viruses. Activation of PKR results in phosphorylation of the α subunit of eukaryotic initiation factor 2 (eIF2α), inhibition of protein synthesis, and limited viral replication. VACV encodes two well-characterized antagonists, E3L and K3L, that can block the PKR pathway and thus enable the virus to replicate efficiently. The use of VACV with a deletion of the dominant factor, E3L, enabled the initial identification of PKR antagonists encoded by human cytomegalovirus (HCMV), a prevalent and medically important virus. Understanding the molecular mechanisms of E3L and K3L function facilitated the dissection of the domains, species-specificity, and evolutionary potential of PKR antagonists encoded by human and nonhuman CMVs. While remaining cognizant of the substantial differences in the molecular virology and replication strategies of VACV and CMVs, this review illustrates how VACV can provide a valuable guide for the study of other experimentally less tractable viruses.

Keywords: E3L; K3L; RNaseL; TRS1; cytomegalovirus; dsRNA; evolutionary arms race; gene amplification; protein kinase R; vaccinia virus.

Figure 1

Model of PKR antagonism by E3L and TRS1. When dsRNA is produced during viral infections, PKR undergoes a conformational change, leading to activation by homodimerization and autophosphorylation [12,40]. VACV E3L can bind to and sequester dsRNA, thereby preventing PKR activation [29,40]. Binding of E3L to dsRNA induces the formation of Z-RNA, which can trigger necroptosis. However, the N-terminal domain of E3L binds to this Z-RNA, blocking activation of the necroptosis signaling pathway. HCMV TRS1 binds to both dsRNA and PKR, possibly forming a heterodimer that prevents PKR activation [23,41]. The dsRBDs of each protein are shown in purple.

Figure 2

Genomic organization of CMV PKR antagonists. The genes encoding known PKR antagonists in the indicated CMVs, all of which are contained in the right end (~40 kb) of the genome, are depicted as red arrows (Accession numbers: HCMV, NC_006273.2; SqmCMV, NC_016448.1; RhCMV, NC_006150.1; MCMV, GU305914.1; GPCMV, NC_020231.1). HCMV IRS1 and TRS1 are partially encoded within repeats (blue boxes); thus, they have different C-termini, while the SqmCMV homologs are encoded entirely within the repeats and are identical. All CMVs have terminal repeats at each end of the genome (black rectangles), but the OWM and rodent CMVs do not have internal repeats. MCMV m142 and m143 are both required to antagonize PKR [57,58,59]. Deletion of GPCMV GP145 does not eliminate replication, suggesting that this virus may encode other not yet identified PKR antagonists [60,61].

Figure 3

HCMV replication depends on TRS1 binding to dsRNA and to PKR. HF or PKR-null HF were infected with recombinant HCMVs containing EGFP and wild-type TRS1 or triple residue mutations that eliminate binding to either dsRNA or PKR. Plaques were photographed 8 days after infection (bar = 400 μm, [11,41], and unpublished data).