Genetic removal of eIF2α kinase PERK in mice enables hippocampal L-LTP independent of mTORC1 activity

Abstract

Characterization of the molecular signaling pathways underlying protein synthesis-dependent forms of synaptic plasticity, such as late long-term potentiation (L-LTP), can provide insights not only into memory expression/maintenance under physiological conditions but also potential mechanisms associated with the pathogenesis of memory disorders. Here, we report in mice that L-LTP failure induced by the mammalian (mechanistic) target of rapamycin complex 1 (mTORC1) inhibitor rapamycin is reversed by brain-specific genetic deletion of PKR-like ER kinase, PERK (PERK KO), a kinase for eukaryotic initiation factor 2α (eIF2α). In contrast, genetic removal of general control non-derepressible-2, GCN2 (GCN2 KO), another eIF2α kinase, or treatment of hippocampal slices with the PERK inhibitor GSK2606414, does not rescue rapamycin-induced L-LTP failure, suggesting mechanisms independent of eIF2α phosphorylation. Moreover, we demonstrate that phosphorylation of eukaryotic elongation factor 2 (eEF2) is significantly decreased in PERK KO mice but unaltered in GCN2 KO mice or slices treated with the PERK inhibitor. Reduction in eEF2 phosphorylation results in increased general protein synthesis, and thus could contribute to the mTORC1-independent L-LTP in PERK KO mice. We further performed experiments on mutant mice with genetic removal of eEF2K (eEF2K KO), the only known kinase for eEF2, and found that L-LTP in eEF2K KO mice is insensitive to rapamycin. These data, for the first time, connect reduction in PERK activity with the regulation of translation elongation in enabling L-LTP independent of mTORC1. Thus, our findings indicate previously unrecognized levels of complexity in the regulation of protein synthesis-dependent synaptic plasticity. Read the Editorial Highlight for this article on page 119. Cover Image for this issue: doi: 10.1111/jnc.14185.

Keywords: LTP; PERK; Rapamycin; eEF2; eIF2α; mTORC1.

Fig. 1

L-LTP failure induced by the mTORC1 inhibitor rapamycin is reversed by genetic removal of eIF2α kinase PERK. (a) HFS-induced L-LTP was blocked in hippocampal slices from WT mice treated with the mTORC1 inhibitor rapamycin (1 µM, closed squares, n=6) compared with slices treated with vehicle (open squares, n=6). In contrast, in slices from PERK KO mice, L-LTP sustained in the presence of rapamycin (filled circles, n=8). Hippocampal L-LTP in PERK KO mice was not altered (open circles, n=7). Arrow denotes HFS. (b) Representative traces of fEPSPs before and after HFS for LTP experiments shown in a. (c) Cumulative data showing mean fEPSP slopes 90 minutes after delivery of HFS based on LTP experiments in a. Unpaired student t-test, **p=0.0013. (d) Treatment of WT slices with selective PERK inhibitor GSK2606414 (1 µM) failed to rescue rapamycin-induced LTP failure (filled circles, n=7). PERK inhibitor did not affect HFS-induced LTP in WT slices (open circles, n=10). (e) Western blot experiments showed that levels of eIF2α phosphorylation were reduced in hippocampi of PERK KO mice (left panel), and WT slices treated with PERK inhibitor GSK2606414 (1 µM, right panel). Unpaired student’s t-test, **p=0.003 *p=0.0188. n=4–6 mice for each group with up to five technical replicates. Cumulative data were shown in bar graphs.

Fig. 2

L-LTP in PERK KO mice is blocked by the general mRNA translation inhibitor anisomycin. (a) HFS-induced L-LTP was blocked in WT hippocampal slices treated with anisomycin (40 µM, filled squares, n=6) compared to normal L-LTP in vehicle-treated WT slices (open squares, n=5). Treatment of slices with anisomycin (40 µM) also blocked L-LTP in slices from PERK KO mice (filled circles, n=8). Arrow denotes HFS. (b) Cumulative data showing mean fEPSP slopes 90 minutes after delivery of HFS based on LTP experiments in a. Unpaired student t-test, *p=0.0179, **p=0.0004.

Fig. 3

Genetic removal of eIF2α kinase GCN2 fails to rescue rapamycin induced L-LTP failure. (a) L-LTP was blocked in hippocampal slices from WT mice treated with rapamycin (1 µM, closed squares, n=6) compared with slices treated with vehicle (open squares, n=8). Treatment of slices from GCN2 KO mice with rapamycin resulted in L-LTP failure (filled circles, n=8). Hippocampal L-LTP in GCN2 KO mice was not altered (open circles, n=7). Arrow denotes HFS. (b) Representative traces of fEPSPs before and after HFS for LTP experiments shown in a. (c) Cumulative data showing mean fEPSP slopes 90 minutes after delivery of HFS based on LTP experiments in a. Unpaired student t-test, *p=0.0443, **p=0.0012. (d) Western blot experiments showed that levels of eIF2α phosphorylation were reduced in hippocampi of GCN2 KO mice. Unpaired student’s t-test, *p=0.0327. n=4–6 mice for each group with up to 5 technical replicates. Cumulative data shown in bar graphs.

Fig. 4

Repression of either PERK or GCN2 activity does not affect the mTORC1 signaling cascade in hippocampus. (a–b) Western blot experiments on hippocampal slices from PERK KO or GCN2 KO mice showed no change in mTOR phosphorylation at either Ser2448 or Ser2481 site. (c) Western blot experiments on hippocampal slices from PERK KO or GCN2 KO mice showed no change in p70S6K phosphorylation at Thr389. (d) 4E-BP1 phosphorylation was not altered in hippocampal slices of either GCN2 KO or PERK KO mice. (e) Hippocampal levels of eEF1A were not affected in either GCN2 KO or PERK KO mice. (f) Treatment with the PERK inhibitor GSK2606414 (1 µM) did not affect phosphorylation of p70S6K (Thr389) in hippocampal slices. (g) Treatment with the PERK inhibitor GSK2606414 (1 µM) did not affect phosphorylation of 4E-BP1 in hippocampal slices. (h) CREB phosphorylation at Ser129 site was unaffected in hippocampal slices from either GCN2 KO or PERK KO mice. (i) Phosphorylation of PSD95 at Thr19 site was unaltered in hippocampal slices from either GCN2 KO of PERK KO mice. n=4–6 mice for each group with up to 5 technical replicates. Cumulative data are shown in bar graphs.

Fig. 5

Genetic removal of PERK decreases eEF2 phosphorylation. (a–b) Western blot experiments on hippocampal slices from GCN2 KO mice or hippocampal slices treated with PERK inhibitor GSK2606414 (1 µM) showed no change in eEF2 phosphorylation. (c) Genetic removal of PERK resulted in reduced levels of eEF2 phosphorylation in hippocampal slices. Unpaired student t-test, *p=0.0195. (d) Immunofluorescence/confocal microscopy showed that neuronal phospho-eEF2 staining was decreased in area CA1 of hippocampal slices from PERK KO mice. Scale bar, 50 µm. Representative images of three independent experiments. (e–g) Western blot experiments showed that AMPK phosphorylation was increased in hippocampal slices of GCN2 KO mice but unaltered in slices treated with GSK2606414 (1 µM) or hippocampal slices from PERK KO mice. Unpaired student t-test, *p=0.0318. n=4–6 mice for each group with up to 5 technical replicates. Cumulative data were shown in bar graphs.

Fig. 6

Rapamycin treatment induced repression of the m-TORC1-AMPK-eEF2K signaling pathway in GCN2 KO and PERK KO mice. (a) Hippocampal slices treated with rapamycin (1µM) from GCN2 KO and PERK KO mice showed significant reduction of p70S6K phosphorylation at Thr389 site. Unpaired student’s t-test, **p=0.0018 and 0.0026. (b) Rapamycin (1µM) reduced hippocampal AMPK phosphorylation in GCN2 KO and PERK KO mice. Unpaired student’s t-test, *p=0.0168 and 0.0339. (c) Rapamycin (1µM) treatment reduced hippocampal phosphorylation of eEF2 in GCN2 KO and PERK KO mice. Unpaired student’s t-test, *p=0.0111, and **p=0.0027. n=3 mice for each group with up to 6 technical replicates. Cumulative data shown in bar graphs.

Fig. 7

Rapamycin-induced L-LTP failure is prevented by genetic removal of eEF2 kinase. (a) HFS-induced L-LTP was blocked in hippocampal slices from WT mice treated with rapamycin (1 µM, closed squares, n=5) compared with WT slices treated with vehicle (open squares, n=10). In contrast, in slices from eEF2K KO mice, L-LTP sustained in the presence of rapamycin (filled circles, n=12). Hippocampal L-LTP in eEF2K KO mice was not altered (open circles, n=10). Arrow denotes HFS. (b) Representative traces of fEPSPs before and after HFS for LTP experiments shown in a. (c) Cumulative data showing mean fEPSP slopes 90 minutes after delivery of HFS based on LTP experiments in a. Unpaired student t-test, *p=0.0429. (d) Western blot experiments demonstrating reduction of eEF2 phosphorylation in hippocampi of eEF2K KO mice. Representative gels of three independent experiments. (e) De novo protein synthesis assessed by SUnSET was significantly increased in hippocampal slices of eEF2K KO mice, compared to WT control mice. Unpaired student’s t-test, *p=0.0464, n=5 mice for each group with 2 technical replicates.