Impaired tissue growth is mediated by checkpoint kinase 1 (CHK1) in the integrated stress response

Abstract

The integrated stress response (ISR) protects cells from numerous forms of stress and is involved in the growth of solid tumours; however, it is unclear how the ISR acts on cellular proliferation. We have developed a model of ISR signalling with which to study its effects on tissue growth. Overexpression of the ISR kinase PERK resulted in a striking atrophic eye phenotype in Drosophila melanogaster that could be rescued by co-expressing the eIF2alpha phosphatase GADD34. A genetic screen of 3000 transposon insertions identified grapes, the gene that encodes the Drosophila orthologue of checkpoint kinase 1 (CHK1). Knockdown of grapes by RNAi rescued eye development despite ongoing PERK activation. In mammalian cells, CHK1 was activated by agents that induce ER stress, which resulted in a G2 cell cycle delay. PERK was both necessary and sufficient for CHK1 activation. These findings indicate that non-genotoxic misfolded protein stress accesses DNA-damage-induced cell cycle checkpoints to couple the ISR to cell cycle arrest.

Fig. 1.

Overexpression of PERK impairs eye development in Drosophila melanogaster. (A) Representative photomicrographs (top) and electron micrographs (bottom) of Drosophila eyes. Driver control (GMR-Gal4), kinase-dead PERK (GMR-Gal4>UAS-PERK-KR), PERK (GMR-Gal4>UAS-PERK-WT), GADD34 (GMR-Gal4>UAS-GADD34), PERK × GADD34 (GMR-Gal4>UAS-PERK; GADD34). (B) Immunoblot analysis of fly heads (three per lane) from indicated genotypes. Black arrowhead indicates upper hyperphosphorylated PERK band, open triangle indicates unphosphorylated inactive PERK. Blotting for actin provided the loading control. Note the increased level of PERK-WT immunoreactivity in GADD34-rescued flies. This reflects the preservation of PERK-expressing retinal tissue. (C) Fluorescence microscopy images of representative eye imaginal discs labelled for DNA synthesis with BrdU (red) and for PERK expression (green). All images aligned anterior (right) to posterior (left). Note the stripe of BrdU labelling indicative of DNA synthesis (S phase) immediately posterior to the morphogenetic furrow (arrowheads). GMR>PERK and GMR>PERK × GADD34 discs show no differences. (D) Fluorescence microscopy images of representative eye imaginal discs stained for phospho-histone H3 (phospho-H3, green) and ELAV expression (red). Note the stripe of phospho-H3 labelling, indicative of mitosis (M phase) posterior to the morphogenetic furrow (arrowheads). In the GMR>PERK eye imaginal discs, this stripe is markedly broadened. (E) Fluorescence microscopy images of representative eye imaginal discs in which PERK expression (blue) was driven in mosaic clones (GFP absent) by the tubulin promoter using the Gal4–Gal80 temperature-sensitive system. Top panel shows the kinase-dead PERK-expressing discs, whereas the bottom panel shows those expressing wild-type PERK. Phospho-histone H3 staining (red) was used to mark mitosis. Note the absent second mitotic wave in the clones lacking GFP expression (lower panel). Genotype: ey-flp; tubulin>Gal80^ts; tubulin>FRT, GFP, stop, FRT, Gal4/uas-PERK.

Fig. 2.

Rescue of eye development by transposon insertion in the grapes gene. (A) Representative photomicrographs and eyes from animals expressing PERK without (GMR-Gal4>UAS-PERK-WT) and with a transposon in the grapes gene (GMR-Gal4>UAS-PERK-WT; grp+/−) (top). Corresponding electron micrographs are shown below. (B) SDS-PAGE and western blot analysis to assess the expression of PERK in the eyes rescued by transposable element insertion within the grapes gene. Note the elevated levels of transgenic PERK in eyes rescued by the transposable elements and by GADD34. (C) Representative photomicrographs of GMR>PERK-WT, GMR>PERK-WT × grapes RNAi. The crosses were repeated on at least three independent occasions. A minimum of three independent repeats were performed for each cross. Images of typical progeny are shown.

Fig. 3.

CHK1 is phosphorylated during ER stress. (A) Immunoblot of whole cell lysate from HCT116 cells treated for the indicated times with thapsigargin (500 nM). UV indicates irradiation with 150 J/m² UV used as the positive control. (B) Graphical representation of three replicates of experiment A showing P317-CHK1 band densitometry normalised to UV control (mean ± s.e.m.). (C) Immunoblot of mouse embryonic fibroblasts (MEFs) untreated (UT) or treated with thapsigargin 500 nM for 1 hour (Tg), tunicamycin 2.5 μg/ml for 2 hours (Tm) or DTT 1 mM for 1 hour (DTT). The gels are representative of three repeats.

Fig. 4.

CHK1 activation during ER stress is mediated by PERK phosphorylation of eIF2α. (A) Immunoblot of Perk^+/+ and Perk^−/− MEFs treated with thapsigargin 500 nM for the indicated times. (B) Immunoblot of CHO cells stably expressing Fv2E-PERK treated for indicated times with 100 nM AP20187 dimerisation compound. Salt-extracted nuclear proteins (nuclear) are shown in panels 2 and 3, all other panels are of soluble post-nuclear supernatant proteins. (C) Immunoblot of eIF2α^SS and eIF2α^AA MEFs treated for the indicated times with thapsigargin 500 nM (Tg) or cycloheximide 50 μg/ml (Cyc). The samples were run on a single gel but have been separated for clarity. The gels shown are representative of three repeats.

Fig. 5.

Transient G2 cell cycle delay during ER stress is mediated by CHK1. (A) Asynchronous cultures of CHO cells stably expressing Fv2E-PERK were treated for the indicated times with tunicamycin (2.5 mg/ml), thapsigargin (500 nM) or AP20187 (100 nM). One million cells were fixed in ethanol and DNA content labelled with propidium iodide was determined by FACS analysis. The positions of cell populations with 2N and 4N DNA content are illustrated. (B) HCT116 cells stably transfected with a Tet-ON CHK1 siRNA or control scrambled sequence were treated for 48 hours with 1 mg/ml doxycycline. Proteins were subjected to immunoblot for CHK1 and actin. (C) Asynchronous cultures of HCT116 Tet-ON CHK1 siRNA or control cells treated with 500 nM thapsigargin and subjected to FACS analysis of the cell cycle. Representative results from four independent repeats are shown. (D) Pooled data from experiment in C are shown as mean ± s.e.m. (E) Data from C expressed as ratio of G2/G1 phase, mean ± s.e.m. (F) Data from C for subG1 phase, mean ± s.e.m. *P<0.05.

Fig. 6.

CHK1 mediates CDC25 loss during ER stress. (A) HCT116 cells stably transfected with Tet-ON CHK1 siRNA or scrambled sequence control treated for 48 hours with 1 μg/ml doxycycline before treatment for the indicated times with 500 nM thapsigargin. The samples were run on a single gel but have been separated for clarity. (B) Pooled data from A expressed as mean ± s.e.m. summarises CDC25A levels normalised to untreated cells (n=3). (C) Representative photomicrographs (upper panels) and electron micrographs (lower panels) of GMR>PERK eyes rescued by cyclin E expression. Cyclin E (GMR-Gal4>UAS-cyc E), PERK (GMR-Gal4>UAS-PERK-WT), PERK × cyclin E (GMR-Gal4>UAS-PERK-WT, cyc E). A minimum of three independent repeats were performed for each cross. Images of typical progeny are shown. (D) Schematic pathway to show PERK-mediated phosphorylation of eIF2α leads to translation attenuation during ER stress. Inhibition of protein translation induces CHK1 phosphorylation, which in turn impairs cell cycle progression in part by depletion of cdk2–cyclin-E activity.