ABSTRACT: In eukaryotes, regulation of mRNA translation enables a fast, localized and finely tuned expression of gene products. Within the translation process, the first stage of translation initiation is most rigorously modulated by the actions of eukaryotic initiation factors (eIFs) and their associated proteins. These 11 eIFs catalyze the joining of the tRNA, mRNA and rRNA into a functional translation complex. Their activity is influenced by a wide variety of extra- and intracellular signals, ranging from global, such as hormone signaling and unfolded proteins, to specific, such as single amino acid imbalance and iron deficiency. Their action is correspondingly comprehensive, in increasing or decreasing recruitment and translation of most cellular mRNAs, and specialized, in targeting translation of mRNAs with regulatory features such as a 5’ terminal oligopyrimidine tract (TOP), upstream open reading frames (uORFs), or an internal ribosomal entry site (IRES). In mammals, two major pathways are linked to targeted mRNA translation. The target of rapamycin (TOR) kinase induces translation of TOP and perhaps other subsets of mRNAs, whereas a family of eIF2 kinases does so with mRNAs containing uORFs or an IRES. TOR targets translation of mRNAs that code for proteins involved in translation, an action compatible with its widely accepted role in regulating cellular growth. The four members of the eIF2 kinase family increase translation of mRNAs coding for stress response proteins such as transcription factors and chaperones. Though all four kinases act on one main substrate, eIF2, published literature demonstrates both common and unique effects by each kinase in response to its specific activating stress. This suggests that the activated eIF2 kinases regulate the translation of both a global and a specific set of mRNAs. Up to now, few studies have attempted to test such a hypothesis; none has been done in mammals. We use array analysis to determine the global mRNA shift into polysomes following a stress response, and to compare the translational response following activation of GCN2 versus PERK, two of the four eIF2alpha kinases. Experiment Overall Design: Perk wild-type or knockout mouse liver were perfused without or with 2,4-di-tert-butylhydroquinone (tBuHQ) for RNA extraction and hybridization of Affymetrix microarrays. RNA was extracted from unfractionated liver samples and polysome fraction of samples separated on sucrose density gradient. To minimize biological variations, we pooled RNA from two perfused liver samples to use in each array analysis. The conditions were total and polysome fraction of Perk+/+, -tBuHQ or +tBuHQ; total and polysome fraction of Perk-/-, -tBuHQ or +tBuHQ. Each array analysis was done in duplicate.