CDK5 and MEKK1 mediate pro-apoptotic signalling following endoplasmic reticulum stress in an autosomal dominant retinitis pigmentosa model

Abstract

Chronic stress in the endoplasmic reticulum (ER) underlies many degenerative and metabolic diseases involving apoptosis of vital cells. A well-established example is autosomal dominant retinitis pigmentosa (ADRP), an age-related retinal degenerative disease caused by mutant rhodopsins. Similar mutant alleles of Drosophila Rhodopsin-1 also impose stress on the ER and cause age-related retinal degeneration in that organism. Well-characterized signalling responses to ER stress, referred to as the unfolded protein response (UPR), induce various ER quality control genes that can suppress such retinal degeneration. However, how cells activate cell death programs after chronic ER stress remains poorly understood. Here, we report the identification of a signalling pathway mediated by cdk5 and mekk1 required for ER-stress-induced apoptosis. Inactivation of these genes specifically suppressed apoptosis, without affecting other protective branches of the UPR. CDK5 phosphorylates MEKK1, and together, they activate the JNK pathway for apoptosis. Moreover, disruption of this pathway can delay the course of age-related retinal degeneration in a Drosophila model of ADRP. These findings establish a previously unrecognized branch of ER-stress response signalling involved in degenerative diseases.

Figure 1. Cdk5 and its regulatory subunit p35 (Cdk5alpha) are required for Rh-1^G69D-induced apoptosis

(A–C) External adult eye phenotypes caused by overexpressing Rh-1^G69, together with lacZ (A), or with inverted repeat (IR) transgenes to knockdown lacZ (B) and cdk5 (C). Note a partial recovery of eye size upon cdk5 knockdown. (D–F) cdk5 knockdown does not affect the rough eye phenotype caused by p53-overexpression (E, F). (D) is a control fly eye with normal morphology. (G–J) Apoptosis in larval eye discs assessed through TUNEL (magenta). Misexpression of Rh-1^G69D led to massive apoptosis (G), which was suppressed by knocking down cdk5 (H). Rh-1^G69D-triggered apoptosis (I), was also suppressed in a p35 (cdk5alpha) −/−background (J). (K–M) cdk5 knockdown does not affect the degree of ATF4 protein induction (red) in response to Rh-1^G69D misexpression. Shown are; a control eye disc (genotype, y, w) (K), a disc misexpressing Rh-1^G69D together with a control lacZ transgene (L), or with a cdk5-IR transgene (M). (N–P) cdk5 knockdown does not affect the degree of XBP1 pathway activation, as assessed through the XBP1-EGFP reporter (green). Shown are; a control eye disc expressing XBP1-EGFP alone (N), or together with Rh-1^G69D and lacZ (O), or with Rh-1^G69D and cdk5-IR (P). (Q, R) Rh-1^G69D-induced apoptosis in ATF4 (crc) −/− discs. TUNEL (magenta) shows Rh-1^G69D-triggered apoptosis in crc+ (Q) and crc¹ −/− discs (R). (S, T) Rh-1^G69D-induced apoptosis (magenta) in ire1 −/− clones (marked by the absence of green). The image is a magnified view of the region overexpressing Rh-1^G69D. TUNEL positive cells are found within the ire1 −/− clones. The scale bar in (G) represents 100 μm for panels (G–R). (S, T) 20 μm. Error bars show ±SEM. Genotypes: gmr-Gal4, UAS-Rh-1^G69D/UAS-lacZ;UAS-dicer2/+ (A, G, L), gmr-Gal4, UAS-Rh-1^G69D/UAS-lacZ-IR;UAS-dicer2/+ (B), gmr-Gal4, UAS-Rh-1^G69D/UAS-cdk5-IR;UAS-dicer2/+ (C, H, M), gmr-Gal4/+ (D), gmr-Gal4,UAS-p53/+;UAS-dicer-2/+ (E), gmr-Gal4, UAS-p53/UAS-cdk5-IR;UAS-dicer2/+ (F), gmr-Gal4/UAS-Rh-1^G69D;+/+ (I), gmr-Gal4, Df(p35)^C2;UAS-Rh-1^G69D, Df(p35)^20C (J), y,w (K), gmr-Gal4/+;UAS-xbp1-EGFP/+ (N), gmr-Gal4, UAS-Rh-1^G69D/UAS-lacZ;UAS-dicer2/UAS-xbp1-EGFP (O), gmr-Gal4, UAS-Rh-1^G69D/UAS-cdk5-IR;UAS-dicer2/UAS-xbp1-EGFP (P), gmr-Gal4/UAS-Rh-1^G69D (Q), gmr-Gal4, crc¹/crc¹, UAS-Rh-1^G69D (R), gmr-Gal4, ey-flp/+; UAS-Rh-1^G69D/+; FRT82, ire1^f02170/FRT82, ubi-GFP (S, T).

Figure 2. Drosophila Mekk1 is required for Rh-1^G69D to trigger apoptosis

(A–C) External adult eyes. A control adult eye with wild type morphology is shown in (A). The degree of eye ablation as a result of Rh-1^G69D misexpression (B), was suppressed in a mekk1^ur-36 −/− background (C). (D–F) Apoptosis in eye discs as assessed through TUNEL labeling (magenta). A control eye disc shows little apoptosis (D). Massive apoptosis caused by Rh-1^G69D misexpression (E), is strongly suppressed in a mekk1^ur-36 −/− background (F). (G–I) The degree of ER-stress as estimated through the xbp1-EGFP reporter (green). Control eye discs show little signs of ER stress (G). The degree of xbp1-EGFP reporter activation by Rh-1^G69D misexpression is similar between mekk1+ discs (H) and mekk1^ur-36 −/− discs (I). (J–L) anti-ATF4 antibody labeling (red) is not affected by mekk1. A control disc (J). ATF4 is induced in Rh-1^G69D expressing discs (K), and is not affected in a mekk1 −/− background (L). (M) mekk1 mutants are more resistant to Tunicamycin (Tm) feeding. 4–5 days old male flies (20 –25 flies in each vial) were allowed to feed for 7 days with standard cornmeal medium supplemented with 5 ug/ml Tm. The percentage indicates the number of flies survived from feeding with Tm (n = 3, p = 0.0062). The scale bar in (D) represents 100 μm for all panels. Error bars show ± SEM. Genotypes: gmr-Gal4/UAS-Rh-1^G69D;+/+ (B, E, K), gmr-Gal4/UAS-Rh-1^G69D;mekk1^ur-36/ mekk1^ur-36 (C, F, L), y, w (D), gmr-Gal4 /UAS-xbp1-EGFP;+/+ (G), gmr-Gal4, UAS-Rh-1^G69D/UAS-xbp1-EGFP;+/+ (H), gmr-Gal4, UAS-Rh-1^G69D/UAS-xbp1-EGFP;mekk1^ur-36/mekk1^ur-36 (I), gmr-Gal4/+ (J).

Figure 3. Mekk1 and Cdk5 mediate JNK signaling activation in response to stress

(A) Thapsigargin (Tg) treatment induces JNK phosphorylation dependent on cdk5 and mekk1. Cells pre-treated with dsRNA against EGFP (negative control) show anti-phospho JNK blots after 2 hours of Tg treatment. Pre-treatment of dsRNAs against cdk5 or mekk1 reduces JNK phosphorylation, while dsRNAs against ire1, traf4, atf6 and perk do not have obvious effects. (B) The role of cdk5 was further assessed by pre-treating cells with the Cdk5 inhibitor, Roscovitine (Lanes 4–9). Control cells were pre-treated with DMSO (Lanes 1–3). (C–F) cdk5 and mekk1 mediates JNK phosphorylation in response to H₂O₂ treatment. (C) H₂O₂ induced JNK phosphorylation analyzed in cells pretreated with dsRNAs either against EGFP (Lanes 1–3) or cdk5 dsRNA (Lanes 4–6). Anti-Cdk5 blot (middle panel) shows the degree of Cdk5 knockdown by RNAi. (D) Quantification of anti-phospho JNK bands after 1 hour of H₂O₂ treatment, with either a control dsRNA against EGFP or against cdk5, shows a statistically significant change (n=3, p=0.0018). (E) H₂O₂ induced JNK phosphorylation in cells pre-treated with dsRNAs against either EGFP (lanes 1–3) or mekk1 (4–6). Anti-Mekk1 blot (middle panel) shows the degree of Mekk1 knockdown by RNAi. (F) Quantification of phospho JNK bands after 1 hour of H₂O₂, from cells pretreated with dsRNA against either EGFP or mekk1, shows a statistically significant difference (n=3, p=0.0023). (G–I) Rh-1^G69D expression (green) activates JNK signaling in eye imaginal discs, as evidenced by the JNK reporter puc-lacZ (magenta) (H, I). (H) shows the anti-betaGal single channel of (I). (G) is a negative control without Rh-1^G69D expression. (J–L) The requirement of bsk (Drosophila JNK) in Rh-1^G69D-induced apoptosis. A control bsk+ disc expressing Rh-1^G69D shows many TUNEL positive cells (magenta) (J), which is suppressed in discs with bsk −/− mosaic clones (K, L). bsk −/− clones are marked by the absence of GFP (green). The scale bar in (G) represents 100 μm. Genotypes: gmr-Gal4/+; puc^E69/+ (G), gmr-Gal4,UAS-Rh-1^G69D/+; puc^E69/+ (H, I), gmr-Gal4, ey-flp/+;UAS-Rh-1^G69D/ ubi-GFP, FRT40;+/+ (J), gmr-Gal4, ey-flp/+;bsk^170B, FRT40, UAS-Rh-1^G69D/ ubi-GFP, FRT40;+/+ (K, L).

Figure 4. Cdk5 phosphorylates Mekk1

(A) Conserved Cdk5 consensus phosphorylation sites within Mekk1 of various Drosophila species. (B) Cdk5 phosphorylates Mekk1^WT on the S1127 residue in vitro. Immunopurified Mekk1 was incubated with recombinant Cdk5 and p35, and subsequently probed with an antibody against phospho S1127 residue of Mekk1 (anti-p-Mekk1^S1127). Anti-Flag blots show total Flag-tagged Mekk1 levels (lower panel). (C) Phosphorylation at Mekk1 in transfected cells. The Flag-tagged Mekk1 was immunoprecipitated and probed with the anti-p-Mekk1^S1127 antibody. The average intensities of phospho Mekk1 bands are shown in a graph underneath the blot. Only Cdk5 tranfected cells show a statistically significant increase in Mekk1 phosphorylation after H₂O₂ treatment (n=3, p=0.0003). (D) Validation of the Mekk1 phosphorylation sites. HEK 293T cells were transfected with Cdk5, together with the indicated expression plasmids marked above the blot. Phospho-Mekk1 bands do not appear when a mutant Mekk1 plasmid lacking the putative phosphorylation sites are transfected (lanes 5, 6). Anti-Flag blots show Flag-tagged Mekk1 levels (middle blot), while anti-HA bands show transfected Cdk5 levels (lower blot). (E) Mekk1 phosphorylation bands disappear after phosphatase treatment (lane 4). Immunoprecipitated complex were either untreated or treated with λ-phosphatase (PP) prior to western blot analysis. (F) Coimmunoprecipitation of Cdk5 and Mekk1. 293T cells were transfected with Flag-tagged Mekk1^WT, together HA-tagged Cdk5^WT. The protein complexes were immunoprecipitated using anti-Flag antibody and analyzed by western blotting using anti-HA antibody. The interaction between Cdk5^WT and Mekk1^WT was enhanced when cells were pretreated with H₂O₂ (lanes 2).

Figure 5. The course of late onset retinal degeneration of ninaE^G69D/+ flies is delayed upon knockdown of Cdk5, or in the mekk1^ur-36−/− background

(A) Quantification of the degeneration process using the Rh1>GFP fluorescence. For each genotype, the percentage indicates the number of flies with intact ommatidial arrays as evidenced by Rh>GFP pattern, from an average of eight independent crosses. Loss of mekk1 function delays the course of retinal degeneration of ninaE^G69D/+ flies (n = 8, p = 0.0062). (B–E) Representative images of 20 day old adult eye tangential sections. Genotypes are as indicated in the panels. Wild type flies show clusters of seven rhabdomeres (stained as black circles) in a trapezoidal pattern within each ommatidia. While this pattern is disrupted in the ninaE^G69D/+ retina, this degenerative phenotype is suppressed in ninaE^G69D/+ with a mekk1^ur-36 −/− background. (F) The knockdown of Cdk5 suppresses late onset retinal degeneration of ninaE^G69D/+ flies (n = 5, p = 0.0004). (G, H) Representative images of 20 day old adult retina downregulating Cdk5 in the ninaE^G69D/+ background.