Skip Navigation

eIF2α Phosphorylation

Substrate Recognition by the eIF2α Kinases

The translation initiation factor eIF2 is a heterotrimeric complex that is responsible for binding the initiator methionyl tRNA (Met-tRNAiMet) to the small ribosomal subunit.  The γ subunit of eIF2 contains a classic GTP-binding domain, and GTP-binding is essential for binding Met-tRNAiMet to form the ternary complex of eIF2, GTP and Met-tRNAiMet.  During the course of translation initiation the GTP bound by eIF2 is hydrolyzed to GDP and eIF2 is released from the ribosome in a binary complex with GDP.  As eIF2 has a much higher affinity for binding GDP than GTP, a guanine-nucleotide exchange factor (GEF) termed eIF2B is required to recycle eIF2•GDP to eIF2•GTP [Figure 1].

This recycling reaction is a key control point to regulate translation initiation and it is the target of the eIF2α protein kinases.  Four protein kinases have been identified that specifically phosphorylate the α subunit of eIF2 on the residue Ser51 [Figure 1] [Figure 2].  These kinases are activated under various cellular stress conditions included heme-deprivation (HRI), virus infection (PKR), ER stress (PERK), and amino acid starvation (GCN2).  Interestingly, phosphorylation of eIF2α on Ser51 inhibits eIF2B and thus impairs general translation, but it can also lead to translational activation of specific mRNAs including the yeast GCN4 mRNA and the mammalian ATF4 mRNA.

Figure 1
Figure 1
Figure 2
Figure 2

In collaboration with Frank Sicheri at the Samuel Lunenfeld Research Institute in Toronto, Canada we obtained the x-ray structure of eIF2α bound to the catalytic domain of PKR (Dar et al., 2005).  The PKR kinase domain resembles typical eukaryotic protein kinases with the active site located between a smaller N-terminal lobe and larger C-terminal lobe.  In the crystal structure the PKR kinase domain dimerizes in a back-to-back orientation mediated by conserved residues in the N-terminal lobe (PDB 2A19) [Figure 3].  One molecule of eIF2α is bound to each PKR kinase domain protomer with the eIF2α OB-fold domain interacting with helix αG in the kinase domain C-terminal lobe and the Ser51 phosphorylation site positioned near the PKR active site (Asp414) [Figure 4].  As the PKR residues mediating kinase domain dimerization and eIF2α recognition are shared among all four eIF2α kinases we propose that all four kinases dimerize and recognize eIF2α in similar manners.

Figure 3
Figure 3
Figure 4
Figure 4

The structure of the Ser51 region of eIF2α was not well resolved in the PKR–eIF2α structure. When the structure of free eIF2α (PDB 1Q46), in which Ser51 is resolved, was positioned on the PKR–eIF2α (PDB 2A1A) structure [Figure 4], Ser51 in eIF2α was ~20 Å from the kinase active site (Asp414). Current efforts are directed at identifying how Ser51 gains access to the kinase active site.

Viral inhibitors of eIF2α Phosphorylation

Figure 5
Figure 5

The PKR is a component in the mammalian anti-viral defense mechanism.  Viral replication and gene expression is thought to generate dsRNA molecules that can serve to activate PKR.  The PKR-dependent phosphorylation of eIF2α inhibits translation initiation, and thus blocks viral protein synthesis leading to impaired viral replication.  In order for a virus to replicate effectively it must prevent the PKR-mediated inhibition of protein synthesis.  Many viruses have developed mechanisms to inhibit PKR [Figure 5].  In addition to inhibiting PKR, it is likely that viruses will seek to inhibit PERK to prevent any potential translation inhibition resulting from ER stress associated with massive influx of viral capsid and envelop proteins.  Viral inhibitors of particular interest to us include the vaccinia virus K3L and E3L proteins and the baculovirus PK2 protein.

Pox Virus PKR Inhibitors

The vaccinia virus K3L protein and related proteins from other poxviruses (including the smallpox C3L protein and the swine pox virus C8L protein) resemble a truncated form of eIF2α.  The ~90 amino acid poxviral proteins show striking similarity to the N-terminal third of eIF2α; however, the viral proteins lack a phosphorylatable residue in the position analogous to Ser51 in eIF2α.  We, and others, have demonstrated that the poxviral proteins are pseudosubstrate inhibitors of PKR.  Expression of either the K3L or C8L protein reduced eIF2α phosphorylation and blocked the toxic effects associated with expression of PKR in yeast. We have used this system to identify the K3L and PKR determinants required for their physical and functional interaction (Seo et al. PNAS 2008).

Related phylogenetic analyses of the eIF2α kinases plus four unrelated protein kinases revealed fast evolution of the PKR kinase domain in vertebrates. These evolutionary studies also revealed evidence of positive diversifying selection at specific sites in the PKR kinase domain. Substitution of positively selected residues in human PKR with residues found in other species altered the sensitivity to PKR inhibitors from different poxviruses. Comparing the sensitivity of human and mouse PKR to poxviral pseudosubstrate inhibitors revealed differences that were traced to positively selected residues near the eIF2α-binding site. Interestingly, 10 of the 12 mutations identified in the genetic screen for PKR mutations conferring resistance to K3L inhibition occurred at sites that were under positive selection during evolution. Taken together, our results indicate how an antiviral protein (PKR) evolved to evade viral inhibition while maintaining its primary function (phosphorylation of eIF2α). Moreover, our identification of species-specific differences in PKR susceptibility to viral inhibitors has important implications for studying human infections in nonhuman model systems (Rothenburg et al. NSMB 2009).

The vaccinia and variola (smallpox) virus E3L protein is a PKR inhibitor consisting of an N-terminal Z-DNA binding domain linked to a C-terminal dsRNA-binding domain. We are currently using yeast based assays to dissect the mechanism of action of the E3L protein.

Baculovirus PK2 protein

The baculovirus PK2 protein resembles the C-terminal half of the eIF2α kinase domain present in PKR and the other kinases.  This protein lacks the signature motifs associated with ATP binding by protein kinases and thus is not expected to be a functional protein kinase.  Based on the common identification of eIF2α kinase inhibitors in viruses, we proposed that the PK2 protein might be an eIF2α kinase inhibitor.  Consistent with this hypothesis, we found that the PK2 protein inhibited the activity of PKR and GCN2 in yeast cells, that PK2 was required for maximal viral yields in infected insect host cells, and that PK2 was necessary for the inhibition of eIF2α phosphorylation in baculovirus-infected insect cells.  Finally, we demonstrated that PK2 directly interacts with PKR, and we propose that PK2 prevents kinase activation by forming inactive heterodimers with the eIF2α kinases (see Dever et al, 1998).