How misfolded proteins from ER are processed and the impact of ERAD disruption in the context of a whole organism is still elusive. In this study we used genetic analysis to investigate the mechanism that links ERAD defects to cellular responses. We specifically investigated the role of EDEM-1, EDEM-2, and EDEM-3 in C. elegans development and aging. We found that loss of EDEM-2 causes pleiotropic defects that are not further aggravated by EDEM-1 and/or EDEM-3 loss under physiological conditions, indicating a major role for EDEM-2 in targeting misfolded proteins for proteasomal degradation under physiological conditions. Moreover, EDEMs respond differently to ER stress stimuli and both EDEM-1 and EDEM-3 functions become evident under acute ER stress. In addition, our data provide the first evidence that preconditioning to edem loss activates a hormetic XBP-1 independent adaptive program that enhances ER stress proteotoxic responses to promote organism survival under acute ER stress.
The C. elegans genome encodes all three orthologues of the ER α-mannosidase-like subgroup of glycosyl hydrolase family 47, which we hereafter referred to as C. elegans edem-1 (for C47E12.3), edem-2 (for F10C2.5) and edem-3 (for ZC506.1) (Fig 1A). The putative C. elegans EDEM-1, EDEM-2 and EDEM-3 proteins share 46, 39 and 41% amino acid identity and present similar structural organization as their human orthologues, with a notable difference, EDEM-3 does not have the conserved sequence for ER-retention signal (RS) (Fig 1B). Outside of the α-mannosidase domain there is little amino acid identity between the C. elegans EDEM sequences and human counterparts. To elucidate their role in a multicellular organism we first investigated the expression pattern of edem-1, edem-2 and edem-3 in C. elegans wild type (WT) animals, using transcriptional and translational fluorescent reporters. We found that both edem-1 and edem-2 are constitutively expressed in the gut, with stronger expression in anterior and posterior gut cells (Fig 1Ca and 1Da-b). EDEM-1 was also detected in hindgut (Fig 1Cb) and a few neurons in the head (Fig 1Cc), whereas EDEM-2 was detected in hypodermis (Fig 1Da), hindgut (Fig 1Dc), pm6 muscle cells of the pharynx (Fig 1Dd), body wall muscle (Fig 1De), vulval muscle (Fig 1Df), pharyngeal epithelial cells (Fig 1Dg), and in a few neurons in the head (Fig 1Dh) and tail (Fig 1Di). In addition, under ER stress conditions, Pedem-1::GFP was detected in the embryos (Fig 1Cd) and uterus (Fig 1Ce), whereas Pedem-2::GFP was ubiquitously detected (Fig 1Dj). Microinjection of the Pedem-3::EDEM-3::GFP construct in WT animals produced uncoordinated progeny that failed to generate stably transmitting transgenic line, suggesting a possible toxic effect caused by overexpression, misfolding or subcellular mislocalization of the EDEM-3::GFP fusion protein. Therefore edem-3 expression was established using the transcriptional GFP reporter (Fig 1E). Pedem-3::GFP expression was detected in the pharynx, nervous system, body wall muscle (mosaic expression) (Fig 1Ea), coelomocytes (Fig 1Ec), hindgut and tail structures (Fig 1Ee), sensory neurons (Fig 1Ek-n), and CAN neurons (visible in Fig 1Ef and 1Eg). In young adults the expression became apparent in vulva muscle and vulval epithelium (Fig 1Eb), uterus (Fig 1Ed), distal tip cells (Fig 1Ed), and hermaphrodite specific neuron (HSN) (Fig 1Ef). Pedem-3::GFP expression in the gut was transiently activated in L1 and L2 larvae (Fig 1Ej), then faded and reappeared in older animals (mosaic expression) (Fig 1Ea). In addition, Pedem-3::GFP became visible in seam cells (Fig 1Eg), intestine (Fig 1Eh), and excretory canals (Fig 1Ei) upon ER stress. We conclude that, although the expression of each edem (edem-1, edem-2 and edem-3) is constitutively detected in specific cells and tissues under physiological conditions, their expression can be upregulated in additional cells and tissues upon ER stress.
The activation of edem genes by the ER stressors was confirmed by quantifying the endogenous edem mRNA levels through qRT-PCR. In these experiments, the efficacy of the applied stressors to induce ER stress and activate the IRE-1 pathway was evaluated by monitoring xbp-1 mRNA splicing (Fig 2D upper panel). An active form of the transcription factor XBP-1 (XBP-1s), which induces expression of many downstream ERAD genes, is generated by IRE-1-mediated unconventional splicing of xbp-1 mRNA . All three edem genes responded to tunicamycin treatment by increasing the mRNA level, but only the mRNA level of edem-1 and edem-2 significantly increased in response to heat stress (Fig 2D lower panel). No difference in mRNA levels was seen upon osmotic stress treatments suggesting that edem genes are specifically activated by ER stress. Next, to confirm that in C. elegans activation of edem genes by ER stress is IRE-1-dependent, as in mammals, we measured edem mRNA levels in IRE-1-deficient worms exposed to a high dose of tunicamycin. As expected, the upregulation of edem mRNAs by tunicamycin was abolished in ire-1 mutants (Fig 2E). Taken together, sequence analysis, ER accumulation of the misfolded CPL-1* upon EDEM depletion, induction of gene expression by ER stressors and control of expression by the IRE-1/XBP-1 pathway indicate that the C. elegans EDEMs are required for the quality control in the ER and have an evolutionarily conserved role in disposal of misfolded proteins.
Next, we analyzed various phenotypes, including morphological defects, growth rate, fertility and lethality of the following edem deletion alleles: edem-1(tm5068), edem-2(tm5186) and edem-3(ok1790), as well as sel-1(tm3901) (Fig 3A). The edem-1(tm5068) is likely a null allele since the deletion causes a frameshift from nucleotide 91 that removes almost the entire amino acid sequence. The deletion in edem-2(tm5186) removes the TATA box and the nucleotides encoding the first 148 amino acids including the ER targeting signal and therefore is likely a null allele. The deletion in edem-3(ok1748) causes an early frameshift in the reading frame, the gene encoding only for an incomplete ER targeting signal and hence, edem-3(ok1790) is also likely a null allele. The sel-1(tm3901) is a deletion allele that removes the sequence encoding for the first three SEL-1-like repeats.
(A) Genomic organization of edem-1, edem-2 and edem-3 genes. Exons are indicated by blue boxes and introns by lines; the ATG initiation (arrow) and termination codons (arrowheads) are indicated. The lines underneath indicate the position of deletion in the corresponding allele. (B) Growth rate of the indicated strains at 20C. Eggs were laid on the plates and the percentage of eggs that reached L4 stage 52h later was scored. Each strain was scored on three replicate plates in four independent experiments. (C) Brood size of the indicated strains at 20C. Each point represents the total brood of one hermaphrodite. (D) Percentage lethality of the indicated strains grown at 20C. Each point represents % lethality of progenies derived from one hermaphrodite. Lethality encompasses both embryonic and larval arrest. (E) Growth rate of the indicated strains at 25C. Eggs were laid on the plates and the percentage of eggs that reached L4 stage 48h later was scored. Each strain was scored on three replicate plates in five independent experiments. (F) Brood size of the indicated strains at 25C. Each point represents the total brood of one hermaphrodite. (G) Percentage lethality of the indicated strains grown at 25C. Each point represents % lethality of progenies derived from one hermaphrodite. Lethality encompasses both embryonic and larval arrest. (H) Growth rate under mild tunicamycin stress. Eggs were laid on the plates with 2 μg/ml tunicamycin and the percentage of eggs that developed into L4 larvae after 3 days was scored. Each strain was scored on three replicate plates in four independent experiments. The red bars in C-D and F-G indicate the mean SEM. The values inside the columns in B, E and H represent the total number of eggs analyzed. *P
Representative Nomarski images of edem-2 mutant animals showing: (A) phase liquid accumulation into the body cavity, (B) large vesicles in intestine, (C) torn embryos in the uterus, (D) torn oocytes in the proximal gonad, (E) unfertilized oocytes with a large nucleus in the uterus, (F) undeveloped uterus, (G) cytokinesis defects (2 cells embryo with a binucleated cell), and (H) oocytes accumulation in proximal gonad reflecting defective ovulation. (I) Nomarski images of WT embryos and edem-2 herniated embryos. Arrows indicate the position where chitin is broken, and the cytoplasm of the embryo is extruded. Images were acquired with Zeiss stereo microscope Discovery V20. (J) WT and edem-2 embryos stained with FM4-64 dye. While in WT embryo the chitin formed an effective barrier that blocked dye entrance, the dye penetrated the chitin of edem-2 embryos. Arrow points to an embryo in which the dye penetrated and stained the cell membranes. Scale bar: 20 μm.
Next, we examined the contribution of each EDEM to the clearance of misfolded CPL-1* from intestinal cells using the deletion alleles, which by contrast to RNAi silencing allows investigation of epistatic effects on CPL-1* clearance. We first tested the response of WT transgenic worms to increasing concentrations of kifunensine and we found that treatment with 30 μM kifunensine was sufficient to induce a maximal CPL-1* fluorescence in WT transgenic animals (S2A Fig). We used this treatment as a positive control to assess the effect of EDEM loss. Quantification of CPL-1* fluorescence showed increased fluorescence in both edem-1 and edem-2 mutants compared with WT control, and their effects were not synergistic nor additive (Fig 5A). This suggests that EDEM-1 and EDEM-2 rather work in a common degradation pathway to clear misfolded CPL-1* from the ER. sel-1 mutants showed a higher fluorescence compared with WT, probably because SEL-1 works downstream of EDEM proteins in CPL-1* clearance and most likely is important for CPL-1* retrotranslocation. In contrast, edem-3 mutants showed a CPL-1* fluorescence similar to WT control (Fig 5A). Also, edem-3 loss decreased the fluorescence of both edem-1; CPL-1* and edem-2; CPL-1* single mutants, as well as edem-1; edem-2; CPL-1* double mutants (Fig 5A), indicating that edem-3 loss has actually a beneficial effect on CPL-1* clearance. The simplest interpretation is that edem-3 loss triggers activation of a mechanism that clears accumulation of CPL-1* in the intestine and this mechanism is partially dependent on functional EDEM-2 and EDEM-1. Treatment of WT animals with kifunensine showed similar fluorescence as for the edem-TKO triple mutant; CPL-1* mutants, indicating that there was no contribution to fluorescence from other ER mannosidases, i.e., ER manI. 041b061a72