Inhibition of a constitutive translation initiation factor 2alpha phosphatase, CReP, promotes survival of stressed cells

Abstract

Phosphorylation of eukaryotic translation initiation factor 2alpha (eIF2alpha) on serine 51 is effected by specific stress-activated protein kinases. eIF2alpha phosphorylation inhibits translation initiation promoting a cytoprotective gene expression program known as the integrated stress response (ISR). Stress-induced activation of GADD34 feeds back negatively on this pathway by promoting eIF2alpha dephosphorylation, however, GADD34 mutant cells retain significant eIF2alpha-directed phosphatase activity. We used a somatic cell genetic approach to identify a gene encoding a novel regulatory subunit of a constitutively active holophosphatase complex that dephosphorylates eIF2alpha. RNAi of this gene, which we named constitutive repressor of eIF2alpha phosphorylation (CReP, or PPP1R15B), repressed the constitutive eIF2alpha-directed phosphatase activity and activated the ISR. CReP RNAi strongly protected mammalian cells against oxidative stress, peroxynitrite stress, and more modestly against accumulation of malfolded proteins in the endoplasmic reticulum. These findings suggest that therapeutic inhibition of eIF2alpha dephosphorylation by targeting the CReP-protein-phosphatase-1 complex may be used to access the salubrious qualities of the ISR.

Figure 1.

Reversal of eIF2α phosphorylation in the absence of GADD34. (A) Immunoblots of eIF2α phosphorylated on serine 51, detected with an epitope-specific primary antiserum, eIF2α (P), total eIF2α(T), PERK (which detects both the unphosphorylated, inactive form of the kinase PERK° and activated, phosphorylated form PERK(P)), and GADD34 on lysates prepared from mouse embryonic fibroblasts with the indicated GADD34 genotypes. The cells were treated for 30 min with 1 mM dithiothreitol (DTT) and placed in DTT-free media for the indicated period of time (wash). Cells were also treated for 4 h with the ER stress-inducing drug thapsigargin (400 nM) serving as positive control for GADD34 induction. (B) Similar experiment to A performed in mouse embryonic stem cells.

Figure 2.

Identification of a retroviral clone encoding CReP. (A) FACS^®-derived fluorescent profiles of untreated (UT), tunicamycin (Tm, 2 μg/ml, 8 h), and sodium arsenite (As, 50 μM, 8 h)–treated parental CHOP::GFP cells and CHOP::GFP cells transduced with the CD retrovirus. (B) Immunoblot of endogenous CHOP and total eIF2α from cells as in A. (C) Immunoblot of eIF2α phosphorylated on serine 51 and total eIF2α from cells as in A.

Figure 3.

CReP is a regulatory subunit of an eIF2α holophosphatase complex with similarity to GADD34 in its COOH-terminal, PP1c–binding region. (A) Alignment of the predicted amino acid sequence of the COOH termini of mouse (mu) and human (hu) CReP and GADD34. The asterisk indicates the phenylalanine residue conserved in all PP1 regulatory subunits that is required for interaction with PP1c. (B) Immunoblot of FLAG-epitope tagged CReP COOH-terminal fragment encoded by the CD retrovirus and full-length CReP immunoprecipitated from lysates of transfected 293T cells (top). Immunoblot of endogenous PP1c in the immunoprecipitates (bottom). (C) Autoradiogram of an in vitro dephosphorylation assay of ³²P-radiolabeled eIF2α on serine 51. The labeled protein was incubated with immune complexes purified from untreated or tunicamycin-treated (Tm) HT22 cells by means of preimmune sera (PI) or antisera directed to CReP or GADD34. Incubation times were 0, 10, and 20 min (lanes 1, 2, and 3, respectively). The bottom panels are immunoblots of the endogenous CReP and GADD34 from the immunoprecipitates used in the dephosphorylation assay. (D) Autoradiogram of an in vitro dephosphorylation assay of ³²P-radiolabeled eIF2α on serine 51. The labeled protein solution was incubated with buffer alone (No lysate) or crude lysate from parental CHO cells or CHO cells stably transduced with the CReP-expressing CD retrovirus described in Fig. 2 (CReP). Reaction times were 10 and 20 min. The asterisk indicates GST-PERK, which is autophosphorylated.

Figure 4.

CReP is a constitutively expressed short-lived protein. (A) Immunoblot of endogenous CReP, GADD34, and total eIF2α in untreated (UT), and tunicamycin (Tm, 2 μg/ml)- and arsenite-treated cells (As, 50 μM). (B) Autoradiogram of pulse chase–labeled endogenous CReP immunoprecipitated with an anti-CReP immune serum. The label in the CReP band is quantified with the signal at the end of the pulse set arbitrarily at 100 (Q). (C) Immunoblot of endogenous CReP, phosphorylated eIF2α and total eIF2α in cells treated with the protein synthesis inhibitor cycloheximide (CHX, 50 μg/ml).

Figure 5.

CReP RNAi activates CHOP::GFP, an ISR target gene. (A) FACS^®-derived fluorescent profiles of nontransfected, CReP RNAi and CD2 RNAi transfected parental CHOP::GFP cells or CHOP::GFP cells stably overexpressing GADD34. The profile of nontransfected cells is shown in black. The light and dark tracings were obtained 16 and 28 h after RNAi transfection. (B) Immunoblots of eIF2α phosphorylated on serine 51, detected with an epitope-specific primary antiserum, eIF2α (P), total eIF2α(T), and PERK, which detects both the unphosphorylated, inactive form of the kinase (PERK°) and activated, phosphorylated form PERK(P) on lysates prepared from mouse ES cells with a CReP knockdown elicited by stable transduction of a CReP shRNA expressing plasmid or control scrambled shRNA plasmid. The cells were treated for 30 min with 1 mM dithiothreitol (DTT) and placed in DTT-free media for the indicated period of time (wash). (C) Immunoblot of endogenous CReP and PP1c in lysates of parental ES cells, CReP shRNA cells, and cells stably transfected with the control scrambled shRNA plasmid. The abundantly expressed TLS protein, detected with a specific antiserum, served as a loading control in this gel.

Figure 6.

CReP RNAi promotes a stress-resistant state. (A) Immunoblot of endogenous CReP in lysates of nontransfected HT22 cells and cells transfected with a small interfering RNAi directed to CReP or to GFP. Lysates were harvested for immunoblot at the indicated time. (B) Survival, measured by MTT assay after exposure to glutamate (7 h at 5 mM, gray bar; or 10 mM, white bar), H₂O₂ (7 h at 1 mM, gray bar; 2 mM, white bar), SIN-1 (7 h at 1.5 mM, gray bar; 3 mM, white bar) and tunicamycin (12 h at 0.125 μg/ml, gray bar; 0.25 μg/ml, white bar followed by 24 h recovery). Cells were nontransfected, mock transfected with no siRNA, or transfected with siRNA to GFP (a control) or CReP. Shown are mean ± SEM of experiments performed in triplicates and reproduced three times. The MTT signal of the untreated, nontransfected cells is arbitrarily set at 100. (C) Dual channel FACscans of dichlorofluorescein fluorescence (DCF axis, reporting on endogenous peroxides) and propidium iodide fluorescence (P.I. axis, reporting on permeabilized, damaged cells) of untreated (boxes 1 and 4) and glutamate-treated (boxes 2, 3, 5, and 6) mock-transfected (boxes 1–3), and CReP RNAi-transfected (boxes 4–6) HT22 cells. The fraction of cells in each quadrant of the FACS^® is indicated in Table I.