Transcripts:
Three (3) transcripts have been identified based on alternate splicing of exon 1 with exon 2 in the 5'UTR. No change to the protein is observed with these transcripts (Kawakubo et al., 1999).
One (1) alternately spliced transcript (Tissue: Placenta) resulting in the loss of exon 12 (Gerhard et al., 2004).
One (1) alternately spliced transcript (Tissue: Brain/Lung) resulting in the loss of exon 11 (Gerhard et al., 2004).
One (1) transcript (Tissue: Brain) which results from an alternate splice acceptor site in exon 17 (Gerhard et al., 2004).

Activation
PKR activation was originally thought to occur only in the presence of double-stranded RNA (ex. viral infection). Over time increasing evidence has indicated that PKR activation is induced by cytotoxic cytokines (tumor necrosis factor (TNF)-α and IFNγ), growth factor deprivation, oxidative stress and DNA damage (Garcia et al., 2006).
PKR is potentially serine/threonine and tyrosine phosphorylated on 105 different sites (54 Ser, 33 Thr and 18 Tyr), including 15 suspected autophosphorylation sites. Of these, only 8 sites have so far been identified, and their significance to PKR activation determined. Phosphorylation of Thr451 in the catalytic domain of PKR is required for minimal kinase activity (Romano et al., 1998; Zhang et al., 2001). An additional phosphorylation of PKR on Thr446 serves to augment PKR activity (Romano et al., 1998; Alisi et al., 2005).
In addition to Thr446/451 phosphorylation, phosphorylation on three key tyrosine residues (Tyr101/162/293) is also required for maximal PKR activity (Su et al., 2006). In cell culture, PKR appears to be constitutively tyrosine phosphorylated, but the exact tyrosine sites that are phosphorylated have not been determined nor has the kinase(s) responsible for these phosphorylations. PKR kinase assays using wild-type eIF2α or mutants Ser51Thr or Ser51Tyr revealed that PKR could phosphorylate the residue at position 51 equally (Lu et al., 1999). One suggestion is that PKR possess tyrosine kinase ability and is able to autophosphorylate (Lu et al., 1999). This is supported by the finding that a catalytically-inactive mutant (K296R) of PKR is not tyrosine phosphorylated in vitro and in vivo (Su et al., 2006). More recent, findings indicate PKR is associated with JAK1 and TYK2 kinases in resting cells. Following interferon stimulation, exogenously expressed JAK1 and TYK2 were demonstrated to phosphorylate Tyr101 and Tyr293 (Su et al., 2007). Similarly the catalytic mutant of PKR was also tyrosine phosphorylated by the JAK kinases. As tyrosine phosphorylation of PKR in response to dsRNA is not affected in cells deficient in JAK kinases, other tyrosine kinases may potentially phosphorylate these sites in response to different stresses (Su et al., 2007). The role of PKR as a non-receptor tyrosine kinase remains controversial.

eIF2α
In order to properly initiate translation, the eIF2 complex must hydrolyze GTP to GDP in the presence of Met-tRNA and the 40S ribosomal subunit. Efficient recycling of the complex then involves the removal of GDP and the re-loading of GTP to the eIF2 complex; a process carried-out by the GTP-exchange factor, eIF2B (Kimball et al., 1998). Phosphorylation of the eIF2α subunit turns the eIF2 complex into a competitive inhibitor. Those eIF2 complexes containing phosphorylated eIF2α demonstrate increased affinity for eIF2B and associate, blocking the eIF2 complex in the GDP bound state (Krishnamoorthy et al., 2001). As the eIF2 complex is in excess of eIF2B, a small amount of phosphorylated eIF2α can result in a shut-off of general translation (Kimball et al., 1998; Sudhakar et al., 2000; Krishnamoorthy et al., 2001; Nika et al., 2001; Wek et al., 2006). The inhibition of general translation is mainly thought to be pro-apoptotic, but recent evidence has suggested that this may be a cellular defense mechanism against stresses (Wek et al., 2006).
Phosphorylation of eIF2α results in a shut-off of general translation but at the same time allows for efficient translation of uORFs in particular mRNAs, such as ATF4, due to their 5' structure; or through what is termed internal ribosome entry site (IRES)-mediated translation (Fernandez et al., 2002; Gerlitz et al., 2002; Yaman et al., 2003). Many of these mRNAs encode proteins involved in the stress response (Koschmieder et al., 2007; van den Beucken et al., 2007; Lee et al., 2009). Short-term inhibition of general translation through eIF2α phosphorylation may in fact be pro-survival by allowing for cellular repair following a particular stress (Donze et al., 2004).

p53
PKR was shown to phosphorylate cytoplasmic p53 on Ser392 enhancing p53 tetramer stability and transcriptional activation of p53 targeted genes (Sakaguchi et al., 1997; Cuddihy et al., 1999; Keller et al., 2001). Among these are p21^Cip1, BAX, PUMA and several pro-caspases. The implications of this phosphorylation are a PKR-mediated cell cycle arrest and induction of apoptosis. Inhibition of constitutive PKR activity in several acute leukemia cells lines with a small molecule inhibitor has been observed to lead to p53 degradation (Unpublished results). Although the exact mechanism for p53 degradation has not been determined, it likely involves the activation of AKT, whose phosphorylation and activity are observed to increase, and AKT effects upon MDM2 (Blalock et al., 2009). Additionally, the cellular PKR activator RAX/PACT was demonstrated to result in increased cellular levels of p53, p53 transcriptional activity and growth arrest in a PKR dependent manner (Bennett et al., 2012). Expression of a siRNA to RAX, which blocks the ability of most stresses to activate PKR, resulted in the decreased expression of several p53 regulated genes such as p21^Cip1 and PUMA and lower constitutive levels of p53. RAX resulted in the SUMOylation of p53 in a PKR independent manner, through direct interaction and activation of the E2 ligase Ubc9 (Bennett et al., 2012).

NF-κB
PKR association with inhibitor κB kinase (IKK) was demonstrated to induce NF-κB nuclear translocation and transcriptional activity (Gil et al., 2000; Zamanian-Daryoush, et al., 2000). While initially PKR kinase activity was implicated in the activation of NF-κB, PKR catalytic activity is not a requirement. Truncated forms of PKR consisting of the amino terminus were shown to associate with the IKK complex and stimulate IκBβ phosphorylation (Bonnet et al., 2000; Bonnet et al., 2006). Later, Donze et al. showed that PKR irregardless of catalytic activity could induce NF-κB activation and the synthesis of some NF-κB dependent transcripts, but NF-κB activity and transcription of other NF-κB dependent genes was greatly potentiated when PKR kinase activity remained intact (Donze et al., 2004). These data suggest that both PKR association with IKK and PKR catalytic activity are important for PKR mediated effects on NF-κB. To this end the current understanding is that PKR activity is required for the full effects of PKR on NF-κB, although whether PKR catalytic activity influences NF-κB activation at the point of IκB phosphorylation and release or at later points, has not been sorted-out.

STATs
PKR has also been demonstrated to affect the transactivation of STATs 1 and 3 (Karehed et al., 2007). STAT1 activity is enhanced by phosphorylation on Ser727. Phosphorylation of this site is defective in PKR-/- fibroblasts resulting in a decrease of STAT1 transactivation (Ramana et al., 2000). PKR kinase activity is not necessary for PKR effects on STAT1 (Wong et al., 1997); instead, PKR associates through its NH₂-terminus with STAT1, which apparently enhances mitogen activated protein kinase (MAPK)-mediated phosphorylation of STAT1 on Ser727 (Deb et al., 2001). Similar to STAT1, PKR has also been demonstrated to be required for proper phosphorylation and transactivation of STAT3. Like STAT1, PKR effects were mediated through MAPK-dependent phosphorylation of STAT3 (Deb et al., 2001). In the absence of PKR, activation of STAT3 by platelet derived growth factor (PDGF) is impaired (Deb et al., 2001).

PP2A
PKR was shown in a yeast-two hybrid system to associate with B56α in a manner dependent on PKR catalytic activity. PKR phosphorylated B56α at multiple sites in vitro (among these Ser28) leading to enhanced PP2A activity (Xu and Williams, 2000). The enhancement of PP2A activity via PKR phosphorylation of B56α resulted in decreased phosphorylation of eIF4E and a lower rate of translation. More recently additional effects of PKR on PP2A activity have been observed. The lymphocytic leukemia cell line REH contains both elevated levels of active PKR and a BCL2 targeted phosphatase activity. PKR was shown to phosphorylate B56α on Ser28 in REH cells which led to PP2A targeting to the mitochondria and dephosphorylation of BCL2 (Ruvolo et al., 2008). PKR activity was also shown to stabilize B56α, but this stabilization was not dependent on Ser28 phosphorylation but instead on eIF2α phosphorylation.

CDK1
Yoon et al. demonstrated that during genotoxic stress PKR is responsible for phosphorylating Cdc2 (CDK1) on Tyr4. Phosphorylation at this site was shown to result in ubiquitination and proteosomal degradation of Cdc2 thus resulting in a G2 arrest (Yoon et al., 2010).

IRS-1
PKR was found to link chronic inflammatory responses to metabolic signaling through the phosphorylation of the insulin response substrate (IRS)-1 on Ser312. Phosphorylation at this site inhibits the phosphorylation of key tyrosine residues required for insulin induced signaling (Nakamura et al., 2010; Yang et al., 2010a).

Single Nucleotide Polymorphisms
SNP analysis revealed V428E (T1840A; source unknown), I506V (A2073G; source unknown).
Additional polymorphisms (1084) identified in the genomic sequence in the locus of EIF2AK2 can be found at PheGenI.