Unveiling the intercompartmental signaling axis: Mitochondrial to ER Stress Response (MERSR) and its impact on proteostasis

Abstract

Maintaining protein homeostasis is essential for cellular health. Our previous research uncovered a cross-compartmental Mitochondrial to Cytosolic Stress Response, activated by the perturbation of mitochondrial proteostasis, which ultimately results in the improvement of proteostasis in the cytosol. Here, we found that this signaling axis also influences the unfolded protein response of the endoplasmic reticulum (UPRER), suggesting the presence of a Mitochondria to ER Stress Response (MERSR). During MERSR, the IRE1 branch of UPRER is inhibited, introducing a previously unknown regulatory component of MCSR. Moreover, proteostasis is enhanced through the upregulation of the PERK-eIF2α signaling pathway, increasing phosphorylation of eIF2α and improving the ER's ability to handle proteostasis. MERSR activation in both polyglutamine and amyloid-beta peptide-expressing C. elegans disease models also led to improvement in both aggregate burden and overall disease outcome. These findings shed light on the coordination between the mitochondria and the ER in maintaining cellular proteostasis and provide further evidence for the importance of intercompartmental signaling.

Fig 1. Inhibition of hsp-6 results in the reduction of ER stress response.

hsp-4p::GFP, the UPR^ER reporter animals were transferred to the RNAi-containing plates L4 adulthood (55 hours post bleaching). On day 1 adulthood (72 hours post bleaching), animals were treated with a) tunicamycin for 4 hours in order to induce ER stress. Following tunicamycin treatment, the worms were transferred onto RNAi-containing plates and grown at 20°C until day 3 of adulthood and were imaged to assess ER stress levels. Other forms of ER stress, including b) nfyb-1 (cu13) mutant and c) VCP (cdc-48.1 and cdc-48.2) knockdown, were tested in the same manner to further validate that our observation was not tunicamycin specific but rather a conserved physiological mechanism. When multiple RNAi treatments were needed, the same amount of dsRNA-expressing bacteria were mixed. The RNAi target genes are as indicated, EV: empty vector control. The graphs show the mean + /-SD of the animals with representative images of animals, n>=6 (three biological repeats). Each RNAi-treated cohort was compared to the EV control to compare GFP induction.

Fig 2. The reduction of ER stress by hsp-6 RNAi is partially regulated through lipid metabolism.

a) hsp-4p::GFP, the UPR^ER reporter animals were used to determine the effects of hyl-1, hyl-2 and lagr-1 on UPR^ER induction as described in Fig 1. b) Other sphingolipid enzymes along the de novo synthesis pathway and the salvage pathways were also tested in the same manner. c) Knockdown of cardiolipin synthetase crls-1 positively regulated induction of UPR^ER. When multiple RNAi treatments were needed, the same amount of dsRNA-expressing bacteria were mixed. The RNAi target genes are as indicated, EV: empty vector control. The graphs show the mean+/-SD of the animals with representative images of animals, n>=6 (three biological repeats). Each RNAi-treated cohort was compared to the EV control to compare GFP induction.

Fig 3. The reduction of ER stress by hsp-6 RNAi is the result of IRE-1 and PERK modulation.

a) Tunicamycin-treated animals induced UPR^ER in an xbp-1 and ire-1 dependent manner. b) hsp-4p::GFP, the UPR^ER reporter strain was crossed with animals that constitutively activate UPR^ER through the IRE-1-XBP-1 pathway (drawn by hand, Jeson J Li) using Microsoft PowerPoint and its icons). Animals expressing constitutively active IRE-1a (344-967aa) that are lacking part of the luminal domain and animals expressing xbp-1s both activate UPR^ER (c and d). c and d) hsp-6 RNAi reduced UPR^ER activation only in the animals expressing constitutively active IRE-1a (c) but not in the animals expressing the xbp-1 spliced form, xbp-1s. MCSR regulates UPR^ER at the IRE-1 level on the ER membrane. Animals were transferred to hsp-6 RNAi plates or the empty vector control plates on late L4/early adult and were imaged on day 3 (72 hours of RNAi treatment). e) Wild-type N2 animals were treated with EV, gcn-2, or pek-1 RNAi from hatch to the late L4 stage, before transferring to double RNAi treatment or to continue single RNAi treatment. Two individual double RNAi treatment groups were set up utilizing a 1:1 ratio of EV (purple) or hsp-6 (teal) combined with gcn-2 or pek-1 for overnight treatment. Hsp-6 RNAi increased the eIFα phosphorylation in a pek-1-dependent manner. f) DAF-28::GFP transgenic animals were treated with indicated RNAi, or the empty vector control from day 1 of adulthood. Animals were imaged for coelomocyte GFP content on day 4. The graph shows mean+/-SD of GFP intensity normalized to empty vector control, n>=6 with three biological repeats.

Fig 4. The knockdown of hsp-6 reduces protein aggregation in poly-glutamine disease model worms.

a) Paralysis of pan-neuronal Aβ expressing animals crossed with neuronal RNAi-enabled line (GRU102;TU3401) at 25°C. Treatment of RNAi against hsp-6, hyl-1, hyl-2 and EV were used to determine improvements on paralysis during heat shock. Log-rank P value was obtained comparing RNAi to empty vector control: P<0.0001 for all hsp-6, hyl-1, and hyl-2. b) Animals expressing polyglutamine aggregates (Q40::YFP) within the neurons (AM101) were crossed with a neuronal RNAi-sensitive worm line (TU3401). Motility was determined using a body bending assay that measures the number of body bends per 30 seconds. The relative motility was plotted by normalizing with empty vector control, and the rate was compared to that of the empty vector control (Three biological repeats with n>12, mean +/- SD). c) Lifespan of the animals expressing Aβ in neurons with neuronal-sensitive RNAi treatment as shown in a), GRU102; TU3401. P<0.001. * TU3401 strain is known to exhibit RNAi effects in non-neuronal tissues at later ages as it becomes “leaky”; a tighter controlled line has been developed recently [42]. Nonetheless, this strain will have neuronal RNAi knockdown of the target genes, however, it is possible to have RNAi effects in non-neuronal tissues at later ages. d) The accumulation of poly-glutamine aggregates is reduced upon MERSR. Animals expressing polyglutamine aggregates (Q35::YFP) in the body wall muscle (AM140) were treated with indicated RNAi or the empty vector control from day 1 adulthood. Animals were collected on day 5. Protein aggregates were measured by applying protein lysate onto a cellulose acetate membrane through a vacuum slot blotter. The membrane was blotted with GFP antibody to detect the Q35::YFP aggregates. 20ug of total protein lysate from each RNAi treatment was applied to SDS-PAGE followed by Western blot of alpha-tubulin, GFP, VCP, and HSP-6 antibody (lower panels). Total protein and total Q35::YFP protein levels were both at equivalent levels across all specimens. The VCP blot and HSP-6 blot showed that the RNAi knockdown of the targeted protein worked efficiently. e) Graph shows mean+-SD of filter trapped Q35::YFP from three biological repeats. f) Representative confocal image of polyQ::YFP (green) and VCP (CDC48, red) localization in the body wall muscle of AM140. Segmentation of polyQ::YFP by emission intensity (low, yellow; high, cyan). g) Colocalization of VCP with polyQ as measured by the Meanders correlation in high (yellow in the segmentation image) and low-intensity (cyan in the segmentation image) areas of polyglutamine (Poly-Q::YFP) in the empty vector control and hsp-6 knockdown worms. Bars show mean ± 95% CI. Two-tailed Mann–Whitney tests show p=0.05 (*). n =13 (EV); n=14 (hsp-6).