C. elegans strains and growth conditions

All C. elegans strains used in this study were maintained at 20ºC on nematode growth medium (NGM) plates seeded with OP50–1 E. coli bacteria as a food source [52]. The names, genotypes, and sources of all strains used in this study are described in S1 Table.

Plasmid construction

The plasmid (pMT686) for whole-body over-expression was constructed by cloning the ubiquitously-expressed eef-1A.1 (formerly eft-3) promotor into the plasmid pPD117.01 (A gift from Andrew Fire, Addgene plasmid #1587) for expression in C. elegans. F44E5.4, C12C8.1, and hsp-1 full-length transcripts were amplified from worm cDNA with primers designed to install an N-terminal HA tag. These inserts were then cloned into pMT686 via Gibson Assembly and the resulting plasmids were screened using Sanger sequencing to validate construction. All primer sequences used in this study are listed in S2 Table.

Generation of transgenic strains

Plasmid constructs for over-expression of F44E5.4, C12C8.1, and hsp-1 were injected into wild-type (N2) hermaphrodite worms alongside the co-injection marker, Pmyo-2::GFP, to generate strains carrying extrachromosomal arrays. All microinjections were performed by SunyBiotech (Fujian, China). Subsequently, extrachromosomal arrays were integrated by UV irradiation. All strains carrying integrated arrays were back-crossed at least 5x to N2s prior to use in experiments.

In vitro AMPylation reactions

Purification of recombinant HSP-3, HSP-4, and FIC-1(E274G)_134–508 was performed according to previously described methods [26,44]. For in vitro AMPylation of HSP-3 and HSP-4, recombinant FIC-1(E274G)_134–508 was first incubated in AMPylation reaction buffer [10 mM HEPES (pH 7.5), 7.5 mM MgCl₂, 150 mM NaCl, 1 mM DTT, 0.5 mM ATP] and incubated at 20ºC for 30 minutes. Then, recombinant HSP-3 or HSP-4 was added to the reaction, and the mixture was incubated at 20ºC for an additional 60 minutes. Fresh samples were immediately submitted to the Proteomics Resource Facility at the University of Michigan for proteomic analysis using their in-solution digestion protocol. Representative figures depicting the identified AMPylated residues were generated in pyMOL using AlphaFold2 [53]-predicted structures obtained from the AlphaFold Protein Structure Database [54,55].

AMPylation mapping by liquid chromatography tandem mass spectrometry (LC-MS-MS)

Briefly, cysteines were reduced with 5 mM DTT for 30 minutes at 45°C. Samples were cooled to room temperature and alkylation of cysteines was achieved by incubating with 65 mM 2-Chloroacetamide, under darkness, for 30 min at room temperature. An overnight digestion with approximately 1 μg sequencing grade modified trypsin was carried out at 37°C with constant shaking in a Thermomixer. Digestion was stopped by acidification and peptides were desalted using SepPak C18 cartridges using manufacturer’s protocol (Waters). Samples were completely dried using a vacufuge. Resulting peptides were dissolved in 9 μl of 0.1% formic acid/2% acetonitrile solution and 2 μls of the peptide solution were resolved on a nano-capillary reverse phase column (Acclaim PepMap C18, 2 micron, 50 cm, ThermoScientific) using a 0.1% formic acid/2% acetonitrile (Buffer A) and 0.1% formic acid/95% acetonitrile (Buffer B) gradient at 300 nl/min over a period of 180 min (2–25% buffer B in 45 min, 25–40% in 5 min, 40–90% in 5 min followed by holding at 90% buffer B for 5 min and requilibration with Buffer A for 30 min). Eluent was directly introduced into Q exactive HF mass spectrometer (Thermo Scientific, San Jose CA) using an EasySpray source. MS1 scans were acquired at 60K resolution (AGC target = 3x10⁶; max IT = 50 ms). Data-dependent collision induced dissociation MS/MS spectra were acquired using Top speed method (3 seconds) following each MS1 scan (NCE ~ 28%; 15K resolution; AGC target 1x10⁵; max IT 45 ms).

Proteins were identified by searching the MS/MS data against E coli BL21 protein database (4156 entries; uniprot-proteome_UP000002032_Ecoli_BL21.fasta, downloaded on 07/17/2019) appended with protein sequences for HSP-3 (WormBase ID: CE08177), HSP-4 (WormBase ID: CE07244), and FIC-1(E274G) using Proteome Discoverer (v2.4, Thermo Scientific). Search parameters included MS1 mass tolerance of 10 ppm and fragment tolerance of 0.2 Da; two missed cleavages were allowed; carbamidimethylation of cysteine was considered fixed modification and oxidation of methionine, phosphoadenosine (329.053 Da) on histidine, lysine, threonine and tyrosine, deamidation of asparagine and glutamine were considered as potential modifications. False discovery rate (FDR) was determined using Percolator and proteins/peptides with a FDR of ≤1% were retained for further analysis.

Worm synchronization

Asynchronous worm populations were washed off of NGM plates with M9 buffer [22.1 mM KH₂PO₄, 42.3 mM Na₂HPO₄, 85.6 mM NaCl, 1 mM MgSO₄] and collected in 15 mL conical tubes. Worms were centrifuged at 215 x g in a Sorvall Legend RT centrifuge for 1 minute, M9 was aspirated, and worms were washed once with 5 mL of M9 buffer prior to bleaching. Worms were dissolved in 5 mL of hypochlorite bleaching buffer at 20ºC while rotating on a nutator (Clay Adams) for exactly 6 minutes and bleaching-resistant embryos were subsequently recovered by centrifugation at 215 x g for 1 minute. After removing the bleaching solution, embryos were washed twice with 5 mL of M9 buffer and centrifuged at 484 x g for 3 minutes before being transferred to fresh NGM plates for downstream experiments. To avoid unwanted selection effects of hypochlorite treatment, synchronized worms were used solely for experimentation and worm stock plates were explicitly spared from bleaching treatments.

RNA interference (RNAi) feeding

RNAi-mediated knock-down by feeding was performed as described previously [56]. Briefly, on the day of the experiment, fresh NGM-RNAi plates supplemented with 1 mM IPTG (Dot Scientific, #DSI5600–25) and 100 μg/mL carbenicillin (GoldBio, #C-103–25) were seeded with HT115 bacteria carrying a pL4440 plasmid expressing double-stranded RNA (dsRNA) against the gene of interest. The pL4440 empty-vector plasmid or pL4440 encoding dsRNA against pos-1 were used as controls. The pos-1 gene encodes a cytoplasmic CCCH zinc-finger protein (POS-1) that is essential for cell fate determination during C. elegans early embryogenesis [57,58], but dispensable in adult animals [59]. Animals fed bacteria expressing dsRNA against pos-1 from hatching are unaffected, but produce non-viable progeny, preventing offspring from interfering with experiments [60]. Embryos or animals were transferred to RNAi plates at indicated time-points and were maintained at 20ºC for all experiments. To achieve combinatorial knock-downs, equal amounts of both HT115-RNAi E. coli strains were seeded onto IPTG plates. RNAi bacterial clones were obtained from either the Vidal [57,58,61] or Ahringer [62] RNAi libraries. Refer to S3 Table for a list of all RNAi clones used in this study and their sources.

Development assays

For each strain of interest, embryos were harvested by hypochlorite treatment of asynchronous worm populations as described above. Embryos were subsequently transferred to small (35mm) NGM-RNAi plates and seeded with HT115-RNAi E. coli bacteria as indicated. The number of embryos on each plate was quantified, and development was scored by visual inspection after 72 hours of incubation at 20ºC. Animals were considered developmentally arrested upon failure to reach the L4 larval stage. Three technical replicates (plates) were scored for each condition and a total of three biological replicates were performed for all experiments.

ER stress assays

For ER stress resistance assays, tunicamycin (Tocris Bioscience, #35-161-0) or thapsigargin (Sigma-Aldrich, #T9033) dissolved in dimethyl sulfoxide (DMSO, Fisher Scientific, #BP231–100) were added ectopically to small (35mm) NGM plates seeded with OP50–1 E. coli 24 hours before use. Final concentrations of 1, 2.5, and 5 μg/mL were utilized for tunicamycin assays, while thapsigargin assay plates contained either 2 or 4.5 μM of the compound. An equivalent amount of DMSO was added to plates and used as a control for each set of experiments. To test for the impact of ER stress, animals were subjected to development assays as described above.

Thrashing assays

Synchronized worm populations were maintained at 20ºC and assessed in thrashing assays at days 3 and 5 of adulthood. Animals were transferred to a 35mm NGM agar plate containing 1 mL of M9 buffer and allowed to acclimatize for 60 seconds. Then, the number of full body bends performed in 30 seconds was scored by visual inspection. Three independent biological replicates were performed with at least 10 animals scored per experiment.

Lifespan assays

Day 1 adult animals were transferred to fresh 60mm NGM-RNAi plates seeded with HT115 E. coli and moved to fresh plates throughout the first 7 days of adulthood. To avoid unwanted effects of FUDR, RNAi against pos-1, a gene whose inactivation prevents egg hatching without affecting adult animals [63], was used at a 50:50 ratio to either empty vector or RNAi against genes of interest. Plates were scored every other day throughout the experiment, and animals were considered dead when repetitive prodding (up to 10x) failed to elicit any detectable body movement. Dead animals were removed immediately from plates upon scoring. All lifespans were conducted at 20ºC.

Imaging of larval-stage worms and polyQ puncta quantification

L2 larvae were washed off of NGM plates with M9 buffer, collected by centrifugation, and washed 3x with M9 supplemented with 0.01% Trition-X-100. Larvae were then fixed in 4% PFA diluted in M9 for 15 minutes, collected, and washed 3x with M9 supplemented with 0.01% Triton-X-100 to remove fixative. Fixed samples were stored in fresh M9 at 4ºC. For imaging, fixed larvae were pipetted onto glass slides and fluorescent images were obtained using a Leica MZ10-F stereomicroscope equipped with a CCD camera (DFC3000 G, Leica) and pE-3000lite LED illuminator (CoolLED). CellProfiler v3.1.9 [64] was used for the semi-automatic quantification of Q40::YFP puncta number and size. Three independent biological replicates were performed and a minimum of 40 animals per genotype and per condition were assessed.

Adult polyQ worm imaging and quantification

Animals were synchronized by bleaching and grown on NGM-RNAi plates from egg at 20ºC until the indicated time-points. For imaging, worms were first transferred to fresh NGM plates with no bacteria under a brightfield dissecting microscope to avoid sampling bias. Animals were then transferred to 2% agarose pads on glass slides, anesthetized in 10 μL of 25 mM tetramisole hydrochloride, and covered with a glass coverslip for imaging. All images were acquired on a Keyence BZ-X700 All-In-One fluorescence microscope. For analysis, Ilastik v1.4.0.post1 [65] was used to train a machine learning-based segmentation algorithm to identify polyQ puncta from a representative set of training images. The resulting segmentation outputs were further processed in Fiji [66] to collate polyQ puncta number per animal as well as puncta size. Three independent biological replicates were performed and a minimum of 50 animals per genotype and per condition were assessed.

Immunoblotting

Animals were washed off of NGM plates with M9 buffer, collected by centrifugation, and resuspended in worm lysis buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 5 mM EDTA, 1% NP-40, 0.1% SDS) supplemented with protease and phosphatase inhibitor cocktail (Thermo Scientific). Worms were transferred to 2 mL reinforced microvials (BioSpec) each containing a 5 mm stainless steel bead (Qiagen) and lysed using a Qiagen TissueLyser II (30 Hz, 5 minutes). Samples were centrifuged at 16,100 x g for 15 minutes at 4ºC to pellet worm debris, and the supernatant was collected. Protein concentrations were normalized using the Pierce bicinchoninic acid (BCA) assay kit (Thermo Scientific, #23225), samples were supplemented with SDS sample buffer, and subjected to SDS-PAGE. Proteins were then transferred to activated 0.2 μm polyvinylidene difluoride (PVDF) membrane using the Trans-Blot Turbo RTA Mini Transfer Kit (Bio-Rad, #1704272) and probed with the indicated antibodies. The following antibodies were used in this study: anti-Thr-AMPylation (Biointron, 17G6, mouse mAb); anti-α-tubulin (DSHB Hybridoma Product 12G10, deposited by Frankel, J./ Nelson, E.M., mouse mAb); anti-GFP (Abcam, ab290, rabbit pAb).

RNA-sequencing and analysis

Animals were synchronized by bleaching and grown at 20ºC on 100mm NGM-RNAi plates seeded with the indicated HT115 E. coli strain until the L4 stage of larval development. Animals were washed off of plates with M9 buffer, collected by gravity sedimentation, and subsequently washed 3 times with M9 to ensure complete removal of bacterial food. Worm pellets were collected by centrifugation and stored at -80ºC. Frozen worm pellets were then thawed on ice, resuspended in 300–600 μL of TRI Reagent (Zymo Research, #R2050-1–50) and lysed using a Qiagen TissueLyser II (30 Hz, 5 minutes). RNA isolation was performed using the Direct-Zol RNA Miniprep Plus Kit (Zymo Research, #R2071) according to manufacturer’s instructions. Library preparation and sequencing were performed by the Advanced Genomics Core at the University of Michigan. Library preparation was performed using the NEBNext Poly(A) mRNA Magnetic Isolation Module (NEB, #E7490L) and NEBNext Ultra II RNA Library Prep Kit (NEB, #E7770L), and samples were sequenced using an Illumina NovaSeq 6000 S4. Three biological replicates were performed. Reads were mapped to the reference genome WBcel235 using STAR v2.7.8a [67] and assigned count estimates using RSEM v1.3.3 [68]. Differential expression analysis to identify significantly up/down-regulated genes was performed with edgeR (v3.19, Bioconductor) [69]. Venn diagrams used to depict the overlap of enriched genes by genotype and/or RNAi condition were generated using InteractiVenn [70]. Gene ontology (GO) and KEGG pathway analyses were performed using ShinyGO (v0.80) [69].

Reverse transcription-quantitative PCR

Sample collection and subsequent RNA isolation was performed as described above (see the RNA-sequencing and analysis section) and total RNA was reverse transcribed using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, #4368814). Reactions were prepared with PowerUp SYBR Green Master Mix (Applied Biosystems, #A25742) and analyzed on a StepOnePlus Real-Time PCR System (Applied Biosystems, #4376600). Changes in gene expression were calculated using the 2-ΔΔCT method from three biological replicates per sample, with the housekeeping genes cdc42 or pmp-3 being used for normalization [71]. All RT-qPCR primers used in this study are listed in S4 Table.

Statistical analysis and graphics

All statistical analyses were performed using GraphPad Prism (v10.2.2) software with the exception of the differential expression analysis of RNAseq data (detailed above). The exact statistical tests used are described in the main text figure panels for each dataset. A p-value of p < 0.05 was used to determine statistical significance. All graphical illustrations were created in Adobe Illustrator (v29.5.1).

The UPR^ER is required to mitigate polyQ protein toxicity in developing C. elegans.

In previous work, we demonstrated that inducible over-expression of the constitutively-active AMPylase, FIC-1(E274G), in C. elegans embryos resulted in complete developmental arrest beyond the L1 larval stage [42]. Changes in endogenous AMPylation levels have also been shown to modulate the aggregation and toxicity of neurodegenerative disease (ND)-associated protein aggregates [41]. We thus wondered if endogenous FIC-1 activity could interfere with proteostasis maintenance in the presence of protein misfolding and aggregation stress during development. To test this, we introduced a fic-1(n5823) null allele into a C. elegans strain expressing a fluorescently-tagged 40-residue polyglutamine (polyQ) tract in body-wall muscle cells (Q40::YFP) [72]. We then used RNAi to knock down major hsp70 chaperone genes in the cytosol (hsp-1), and the ER (hsp-3 or hsp-4), to induce protein unfolding stress in wild-type, fic-1(n5823), Q40::YFP, and Q40::YFP + fic-1(n5823) embryos and assessed their ability to develop into L4 larval animals within 72 hours from hatching (Fig 1A). qPCR confirmed efficient knock-down of the intended target genes by RNAi (S1D - S1F Fig). We found that loss of either hsp-3 or hsp-4 did not interfere with larval development in the absence of polyQ aggregation stress. However, in Q40::YFP animals, hsp-3 or hsp-4 ablation resulted in complete developmental arrest in the L3 larval stage that was rescued in Q40::YFP animals containing the fic-1(n5823) null allele. In contrast, hsp-1 knock-down was embryonic lethal in all tested genetic backgrounds. Combinatorial knock-down of fic-1 + hsp-3 or fic-1 + hsp-4 in wild-type or Q40::YFP animals recapitulated this rescue (S1C Fig). Given the limited effect of RNAi in C. elegans neurons due to a lack of the dsRNA transporter, SID-1 [73], we posit that the rescue observed here is predominantly non-neuronal.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. The UPR^ER is required to mitigate polyQ protein toxicity in developing C. elegans.

(A) Development assay of the indicated strains (legend) depicting the percentage of embryos surviving to at least the L4 stage of larval development at 72 hours when fed RNAi against the indicated genes (X-axis) from hatching. (B) Western blot of lysates from L2 larvae fed the indicated RNAis from hatching and probed for Thr-AMP signal (top) and α-tubulin (bottom) as a loading control. (C) Quantification of Thr-AMP signal intensity normalized to α-tubulin. Each data point represents one lane from one biological replicate. (D) Development assay of the indicated strains depicting the percentage of embryos surviving to at least the L4 stage of larval development at 72 hours when fed RNAi against the indicated genes (X-axis) from hatching. (E) Representative images of L2 larvae following treatment with the indicated RNAis from hatching. Scale bar = 100μm. (F) Quantification of the number of polyQ puncta per animal following RNAi treatment (X-axis). (G) Quantification of L2 larval body size following RNAi treatment. (H-J). Frequency distribution profiles of polyQ puncta sizes in L2 larvae on control (H), hsp-3 (I), or hsp-4 (J) RNAi. Bin size = 20 a.u. In (A, D), translucent data points depict technical replicates, while opaque data points reflect the average for each biological replicate performed (n = 3). In (F-G), each translucent data point represents one worm, and opaque data points reflect the average for each biological replicate (n = 3). Error bars for all plots represent SD. Two-way (A, D) or one-way (F-G) ANOVAs with Tukey’s post-hoc multiple comparisons tests were performed to determine statistical significance. ***p < 0.001; **p < 0.01; *p < 0.05; ns = not significant.

https://doi.org/10.1371/journal.pgen.1011723.g001

Based on these results, we next tested whether endogenous HSP-3 and/or HSP-4 AMPylation occurs and is detectable during early larval development. To this end, we prepared lysates from L2 larvae of wild-type, fic-1(n5823), Q40::YFP, and Q40::YFP + fic-1(n5823) animals and assessed protein AMPylation by immunoblotting using an antibody specific for AMPylated threonine (Thr-AMP) residues [74] (Figs 1B–1C, S1A–S1B). Under control RNAi (pos-1) conditions, we detected robust AMPylation signatures at a molecular weight corresponding to that of HSP-3 and HSP-4. This signal partially decreased upon knock-down of hsp-3 or hsp-4, consistent with AMPylation of both chaperones. To corroborate this approach, we additionally performed in vitro AMPylation reactions with recombinant FIC-1(E274G) and HSP-3 or HSP-4 and submitted these samples for analysis by mass spectrometry. This revealed one AMPylated residue on HSP-3 (T194), and three AMPylated residues on HSP-4 (T368, T455, T576) (S3A–S3B Fig). As expected, fic-1(n5823) null samples were devoid of detectable Thr-AMP signal (Fig 1B–1C). Interestingly, AMPylation levels were unchanged in the presence of Q40::YFP compared to wild-type samples. This suggests that baseline levels of endogenous AMPylation are sufficient to interfere with Q40::YFP larval development when ER homeostasis is perturbed.

Since polyglutamine expansion diseases exhibit threshold-dependent toxicity around 35–40 residues [75,76], we hypothesized that HSP-3/HSP-4 AMPylation may only be detrimental in the context of longer, aggregating polyQ tracts. Using two additional C. elegans strains expressing shorter polyQ tracts, Q24::YFP and Q35::YFP, which do not form puncta during larval development [72], we repeated this assay (Fig 1D) in the presence or absence of fic-1(n5823). In line with our hypothesis, fic-1 deletion had no effect on the development of Q24::YFP or Q35::YFP animals upon hsp-3 or hsp-4 knock-down. Instead, the protective effects of fic-1 loss were specific to Q40::YFP animals, consistent with a polyQ tract length-dependent phenomenon. Expanding upon this, we also tested the impact of two small molecules known to induce ER stress, tunicamycin (Tm), and thapsigargin (Tg) on worm development (S3C–S3D Fig). In the absence of polyQs, fic-1(n5823) null animals were less sensitive to Tm treatment than their wild-type counterparts. In the presence of polyQs, however, fic-1 KO animals exhibited a trend towards increased Tm sensitivity (S3C Fig), suggesting that the benefits of decreased AMPylation decline when multiple proteotoxic insults accumulate.

Having established polyQ toxicity as a driver of embryonic ER stress, we next sought to examine polyQ aggregation dynamics in L2 larvae prior to the onset of developmental arrest. In line with our previous reports in adult worms, fic-1 KO larvae exhibited increased numbers of Q40::YFP puncta (Fig 1E–1F), an observation that occurred independently of hsp-3 or hsp-4 depletion. Interestingly, hsp-3, but not hsp-4 knock-down led to reduced body size in the absence of fic-1 KO (Fig 1G). While there were no significant differences in the distribution of Q40::YFP puncta sizes upon hsp-4 knock-down (Fig 1H and 1J), hsp-3 depletion was associated with a decrease in the relative proportion of smaller puncta in fic-1 KO larvae (Fig 1I). Of note, we did not detect any significant changes in Q40::YFP expression due to RNAi treatment conditions or the presence of the fic-1(n5823) null allele (S2A–S2D Fig). Taken together, these results indicate that the loss of FIC-1-mediated AMPylation impacts polyQ aggregation dynamics and bolsters ER proteostasis during a critical developmental window.

Loss of FIC-1-mediated AMPylation protects against declines in polyQ worm fitness and lifespan in adulthood

With the knowledge that loss of AMPylation confers protection during worm development in the presence of aggregating polyQs, we next assessed if FIC-1 activity is dispensable later in life in adult animals. To avoid lethality in polyQ worms plated on hsp-3/hsp-4 RNAi from hatching, we first synchronized embryos from all strains on control RNAi and transferred animals to experimental RNAi conditions on day 1 of adulthood (Fig 2A). As a general read-out for overall worm fitness and motility [77,78], we performed thrashing assays in day 3 (Fig 2B) and day 5 (Fig 2C) adult animals. Under control conditions, fic-1 KO had no impact on Q40::YFP worm motility at either time-point tested. Interestingly, however, UPR^ER dysregulation through the depletion of either hsp-3 or hsp-4 in adulthood led to a progressive decline in thrashing rates of Q40::YFP animals at day 3 and day 5, while Q40::YFP; fic-1(n5823) worms were protected from this insult. As a secondary measure, we next performed longevity experiments under the same conditions (Figs 2D - 2F, S4A–S4C). Similar to the results observed in the thrashing assay, fic-1 KO did not impact Q40::YFP lifespan under control conditions (Figs 2D, S4A). Upon hsp-3 depletion, however, the lifespan of Q40::YFP animals diminished, while Q40::YFP; fic-1(n5823) worms again remained unaffected (Figs 2E; S4B). Knock-down of hsp-4 decreased lifespan of all strains tested, but here, too, fic-1 KO led to a small but significant increase in lifespan specific to the presence of polyQs (Figs 2F; S4C). Quantifying polyQ puncta at day 3 (Fig 2G–2H) and day 5 (Fig 2I–2J) of adulthood, fic-1 KO increased the number of puncta independent of RNAi condition (Fig 1E–1F). In adult animals, however, Q40::YFP; fic-1(n5823) worms exhibited an increased relative frequency of large polyQ puncta, possibly indicating a shift towards sequestration of toxic polyQ species later in adulthood (S4D -S4F and S4G–S4I Fig). Taken as a whole, our data show that, in the context of protein aggregation stress in the ER, the loss of FIC-1-mediated AMPylation has a beneficial influence on worm fitness beyond larval development, imparting cytoprotective effects well into adulthood.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Loss of FIC-1-mediated AMPylation protects against declines in polyQ worm fitness and lifespan in adulthood.

(A) Schematic outlining RNAi treatment paradigm in adult animals. Embryos were synchronized on control RNAi and transferred at day 1 of adulthood to prevent embryonic lethality in polyQ worms observed upon hsp-3/4 loss. (B-C) Quantification of thrashing rates for day 3 (B) or day 5 (C) adult animals of the indicated genotypes (legend) upon treatment with the indicated RNAis (X-axis) beginning at day 1 of adulthood. (D-F) Lifespan experiments of wild-type and polyQ worms in the presence or absence of the fic-1(n5823) null allele when fed control (D), hsp-3 (E), or hsp-4 (F) RNAi in adulthood. (G-H) Representative images (G) and quantification (H) of polyQ puncta in day 3 adult animals when treated with the indicated RNAis in adulthood. (I-J) Representative images (I) and quantification (H) of polyQ puncta in day 5 adult animals when treated with the indicated RNAis in adulthood. For both (G) and (I), scale bar = 100μm. In (B-C) and (H-J), each translucent data point represents one individual worm, and opaque data points depict the average of each biological replicate (n = 3). Error bars for all plots represent SD. Two-way (B-C) or one-way (H-J) ANOVAs with Tukey’s post-hoc multiple comparisons tests were performed to determine statistical significance. For lifespan studies (D-F), a Mantel-Cox test was used. ***p < 0.001; *p < 0.05; ns = not significant.

https://doi.org/10.1371/journal.pgen.1011723.g002

fic-1 KO initiates a transcriptional response to counteract protein misfolding stress

To gain further insight into the gene(s) and pathway(s) activated in the absence of FIC-1-mediated AMPylation, we performed bulk RNA sequencing analysis of wild-type, fic-1(n5823), and Q40::YFP ± fic-1(n5823) animals under hsp-3 and hsp-4 knock-down conditions. Each strain was grown on control, hsp-3, or hsp-4 RNAi from embryo to the L4 stage of larval development with the exception of Q40::YFP, which was grown solely on control RNAi due to the lethal effects of hsp-3 or hsp-4 knock-down in this strain (Fig 3A). A principal component analysis of the resulting transcriptomes showed separation of the samples based on both the strains’ genetic backgrounds and RNAi treatment conditions (Fig 3B). We first validated the efficacy and specificity of hsp-3 and hsp-4 RNAi, both of which robustly suppressed expression of their intended gene targets (S5C–S5E Fig). While differential expression analysis revealed no significant effect of a given RNAi (e.g., against hsp-3) on the other ortholog (e.g., hsp-4) (S5A–S5B Fig), we did find that knock-down of hsp-3 increased hsp-4 transcripts when assessed by qPCR, as has been shown previously [39] (S5C and S5E Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. fic-1 KO initiates a transcriptional response to counteract protein misfolding stress.

(A) Graphical depiction of experimental design used for RNA sequencing studies. (B) Principle component analysis (PCA) plot showing the separation of samples based on genotype and RNAi treatment conditions. Each data point represents one biological replicate from the indicated group. (C) Venn diagram of genes upregulated upon hsp-3 knock-down across genotypes, with 269 genes specific to Q40::YFP; fic-1(n5823) animals. (D) Over-represented gene ontology (GO) biological process terms under hsp-3 knock-down conditions ordered by fold enrichment. (E) Over-represented GO molecular function terms under hsp-3 (magenta; top) and hsp-4 (teal; bottom) knock-down conditions. (F) Venn diagram of genes upregulated upon hsp-4 knock-down across genotypes, with 371 genes specific to Q40::YFP; fic-1(n5823) animals. (G) Over-represented GO biological process terms under hsp-4 knock-down conditions ordered by fold enrichment. (H) Enriched KEGG pathways upon hsp-3 (magenta; top) or hsp-4 (teal; bottom) knock-down. (I) Curated list of metabolism, lysosomal, protein folding, and stress response-related genes enriched in Q40::YFP; fic-1(n5823) animals upon hsp-3 knock-down. (J) Curated list of ER-associated degradation (ERAD), ER stress and protein processing, glycotransferase, lysosomal, protein folding, stress response, and UPR^ER-related genes enriched in Q40::YFP; fic-1(n5823) animals upon hsp-4 knock-down. For (I-J), gene heat-maps are colored according to log₂FC value (cut-off: log₂FC > 1.5; p < 0.05) with darker colors corresponding to larger FC values.

https://doi.org/10.1371/journal.pgen.1011723.g003

Knock-down of hsp-3 and hsp-4 elicited robust changes in gene expression relative to control conditions, with 415 genes significantly upregulated and 363 genes downregulated [cut-offs: p < 0.05; log₂FC > 1.5 or> -1.5, respectively] upon hsp-3 depletion and 695 genes up- and 784 genes down-regulated under hsp-4 knock-down across all genotypes. Gene ontology (GO) analysis of commonly upregulated genes revealed an activation of protein folding-related genes upon hsp-3 knock-down, while hsp-4 knock-down resulted in the shared upregulation of genes involved in the response to ER stress (S5F–S5G Fig). We then isolated the transcriptomic response specific to Q40::YFP; fic-1 KO animals for further analysis.

Upon hsp-3 knock-down, we identified 269 upregulated and 290 downregulated genes unique to Q40::YFP; fic-1 KO worms (Figs 3C, S6A). Similarly, hsp-4 knock-down elicited the upregulation of 371 genes and the downregulation of 543 genes (Figs 3F, S6D). GO enrichment analysis for over-represented biological process terms revealed a significant enrichment of protein folding and stress response-related genes upon hsp-3 loss in Q40::YFP; fic-1 KO animals, notably including a specific signature related to glutathione metabolism (Fig 3D and 3I). hsp-4 knock-down resulted in a similar enrichment of stress-response and protein folding genes accompanied by genes specific to the endoplasmic reticulum unfolded protein response (UPR^ER) and ER-associated degradation (ERAD) pathways (Fig 3G). Further examination of GO molecular function terms and enriched KEGG pathways in Q40::YFP; fic-1 KO animals on hsp-3 RNAi revealed the upregulation of processes related to unfolded protein binding and ER protein processing, identifying the ER stress response as a shared hallmark of polyQ protein toxicity in the absence of either hsp-3 or hsp-4 (Fig 3E and 3H). Selecting genes for further testing, we identified a suite of significantly upregulated molecular chaperones upon hsp-3 loss, including hsp-90, the HSP-90 ATPase activator ahsa-1, HSP70 family chaperone members F44E5.4 and F44E5.5, the non-canonical small heat-shock protein (sHSP) hsp-17, and the sHSP hsp-16.49 (Figs 3I, S7A). The activation of numerous glutathione S-transferases (gst-1, -5, -8, -14, -24, -31, -32, -44) was unique to hsp-3 loss and suggests a possible role for oxidative stress in polyQ toxicity, as reported previously in C. elegans [79]. Similar to hsp-3 loss, hsp-4 knock-down also resulted in the upregulation of hsp-17 and hsp-16.49, but further involved activation of the sHSP hsp-16.11, DnaJ/HSP40 family proteins dnj-28 and dnj-7, HSP70 chaperone stc-1, and canonical ERAD genes (crt-1, cup-2, der-2, enpl-1) (Figs 3J, S7B). These gene programs activated in response to hsp-3 or hsp-4 knock-down were highly specific to Q40::YFP; fic-1 KO animals, as analysis of upregulated genes unique to wild-type or fic-1 KO animals under either RNAi condition failed to uncover distinct clusters of related genes or further insights via GO analysis (S8A–S8D Fig).

In a limited screen, we next tested whether RNAi against a subset of the genes upregulated in Q40::YFP; fic-1 KO animals, either alone or in combination with hsp-3 or hsp-4 RNAi, could result in developmental arrest. While no single gene’s knock-down fully suppressed larval development, significant impacts were observed for cct-1 and cct-4 in combination with hsp-3 RNAi, and lec-11 and enpl-1 in combination with hsp-4 RNAi (S7C - S7D Fig). Interestingly, genes found to be significantly downregulated upon hsp-3 or hsp-4 knock-down were overwhelmingly associated with eggshell formation, oogenesis, and lysosomal function (S6B–S6C and S6E–S6F Fig). In a second limited screen, we tested RNAi against the top downregulated genes in Q40::YFP; fic-1 KO worms upon hsp-3 or hsp-4 loss to assess if their knock-down in Q40::YFP animals could reverse developmental arrest. Strikingly, RNAi against any of the genes tested (col-135, ilys-5, lys-10, abu-2) in combination with hsp-3 RNAi failed to promote survival (S6G Fig). Results were similar for RNAi against genes tested in combination with hsp-4 RNAi (cpr-8, vit-5, asp-3, clec-53, asah-1), suggesting that downregulation of these genes in Q40::YFP; fic-1 KO animals likely is not a driving survival factor in the face of perturbed ER homeostasis (S6H Fig). Instead, the notable suppression of genes related to reproduction may serve as a complementary protective mechanism or “trade-off” during the final stages of development to cope with protein misfolding stress [80,81].

Taken as a whole, these findings provide a holistic characterization of how the transcriptome of AMPylation-deficient C. elegans expressing aggregation-prone polyQs responds to perturbed ER homeostasis. Our data reveal a broad stress-responsive paradigm, highlighting roles for molecular chaperones, ERAD, UPR^ER, and metabolic genes in mitigating polyQ toxicity in the absence of fic-1.

Loss of fic-1 activates UPR^ER signaling to combat misfolded proteins during development

Following up on the resuts from our RNA sequencing analysis, we first tested for involvement of each UPR^ER branch in polyQ toxicity during worm development. In C. elegans, the UPR^ER consists of three stress sensors, IRE-1, PEK-1, and ATF-6 which are activated in response to protein folding stress and function to initiate downstream upregulation of stress-responsive genes. Given our data implicating UPR^ER signaling in the response to polyQ protein toxicity in the absence of fic-1 (Fig 3E and 3G), we first assessed each branch of the UPR^ER using RNAi-mediated depletion in development assays (S9A - S9E Fig). Upon loss of hsp-3 or hsp-4, we observed the expected developmental arrest phenotype in Q40::YFP animals that was rescued by fic-1 KO. RNAi against ire-1, or its downstream mediator, xbp-1, had no significant effect on development in wild-type or fic-1(n5823) animals in the absence of polyQs (Fig 4A), in line with previous reports [82]. In the presence of Q40::YFP, ire-1 and xbp-1 knock-down significantly impeded worm development, with a modest increase in fic-1 KO animals (Fig 4A), suggesting that the beneficial effects of fic-1 loss are not solely mediated through this arm of the UPR^ER. Next, we tested the two other UPR^ER branches, PEK-1 and ATF-6. Interestingly, RNAi-mediated knock-down of either pek-1 or atf-6 not only impaired development of Q40::YFP-expressing animals, it also abrogated the rescue effects imparted by fic-1 KO (Fig 4B). Combinatorial knock-down of pek-1 and atf-6 did not further reduce worm survival (S9F Fig) suggesting that signaling from either branch is required and sufficient to drive protection from the loss of FIC-1-mediated AMPylation. Given the significant enrichment of UPR^ER-related genes amongst our RNAseq hits for both hsp-3 and hsp-4 knock-down conditions (Fig 3I–3J), we next examined roles for PEK-1’s downstream effector proteins, the transcription factor ATF-4 and the translation initiation factor eIF2. Using a combinatorial RNAi approach, we observed that knock-down of either atf-4 or eif-2A blunted the beneficial role of fic-1 KO during polyQ worm development (Fig 4D), though loss of eif-2A was also observed to increase developmental arrest in a non polyQ-dependent manner. Further, combinatorial knock-down of hsp-3 and either atf-4 or eif-2A was sufficient to restore fic-1(n5823) rescue in Q40::YFP animals, but failed to do so under hsp-4 knock-down conditions. Taken collectively, these data highlight UPR^ER branch-specific roles for coping with protein folding toxicity during worm development and identify contributions from ire-1, pek-1, and atf-6 as key components of this response. Additionally, the hsp-3-specific rescue upon concurrent loss with either atf-4 or eif-2A suggests a chaperone-specific effect of downstream pek-1 signaling, perhaps owed to hsp-4’s reported role as a highly stress-responsive ER chaperone.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Loss of fic-1 activates UPR^ER signaling to combat misfolded proteins during development.

(A) Development assay of the indicated strains assessing for the impact of the IRE-1 branch of the UPR^ER on development. (B) Development assay testing involvement of the PEK-1 and ATF-6 UPR^ER branches. (C) Development assay of downstream pek-1 mediators, ATF-4 and EIF-2A and their role in polyQ worm development. (A-C) All graphs depict the percentage of animals reaching the L4 stage of larval development at 72 hours. Translucent data points reflect technical replicates (plates), while opaque data points represent biological replicates (n = 3). Error bars represent SD. For all graphs, a two-way ANOVA with Tukey’s post-hoc multiple comparisons tests were performed to determine statistical significance. **p < 0.01; *p < 0.05; ns = not significant.

https://doi.org/10.1371/journal.pgen.1011723.g004

UPR^ER signaling upregulates expression of the cytosolic HSP70 family chaperone, F44E5.4, to suppress polyQ toxicity

Having observed a significant upregulation of numerous, predominantly cytosolic, molecular chaperones upon hsp-3 knock-down, we next asked if enhanced cytosolic chaperone activity is sufficient to phenocopy the rescue obtained by the loss of fic-1. To this end, we generated transgenic C. elegans strains expressing three key HSP70 family chaperones – F44E5.4, hsp-1, and C12C8.1 – under the ubiquitous eef1A.1 (formerly eft-3) promoter for whole-body over-expression (Fig 5A). As transcripts of F44E5.4 were significantly upregulated in Q40::YFP; fic-1 KO animals in response to hsp-3 knock-down (Fig 3I), we first asked whether F44E5.4 over-expression could promote the survival of Q40::YFP animals placed on hsp-3 or hsp-4 RNAi from hatching. Indeed, in development assays, F44E5.4 over-expression reversed the developmental arrest of Q40::YFP animals grown on either hsp-3 or hsp-4 RNAi to the same extent as fic-1 deletion (Fig 5B). However, neither the over-expression of hsp-1, nor C12C8.1, imparted the same rescue effect, indicating a unique role for F44E5.4 in suppressing polyQ toxicity (S10A – S10B Fig). To further probe whether F44E5.4 upregulation occurs as a direct consequence of fic-1 loss, we performed development assays knocking-down F44E5.4 alongside hsp-3 or hsp-4 in Q40::YFP; fic-1 KO animals. While F44E5.4 RNAi alone did not impact worm survival, combinatorial knock-down of hsp-3 + F44E5.4 or hsp-4 + F44E5.4 significantly impeded worm development, indicating that F44E5.4 upregulation contributes to, but is not the sole effector mitigating polyQ toxicity in the absence of fic-1 (Fig 5C). Assessment of F44E5.4 knock-down efficacy via qPCR revealed incomplete suppression of F44E5.4 transcripts (S10E Fig), providing a potential explanation for the observed partial rescue. Conversely, we also investigated the impact of fic-1 knock-down in Q40::YFP; F44E5.4 over-expressing (OE) worm development. In this case, fic-1 RNAi alone did not impact worm development, and combinatorial knock-down of hsp-3 + fic-1 or hsp-3 + fic-1 failed to elicit any additive effect on worm development (Fig 5D). Together, these findings suggest that enhanced F44E5.4 expression occurs as a direct consequence of fic-1 loss. Understanding that the protective effects elicited in response to hsp-3 or hsp-4 knock-down are orchestrated through the UPR^ER, we next asked if UPR^ER signaling is required for enhanced F44E5.4 expression. Using quantitative PCR (qPCR), we examined F44E5.4 transcript levels in Q40::YFP; fic-1 KO larvae (L4) grown on hsp-3 RNAi alone or in combination with RNAi against each of the UPR^ER stress sensors. Compared to control RNAi conditions, we again found that F44E5.4 expression is dramatically upregulated in response to hsp-3 knock-down (Fig 5E). In contrast, larvae grown on hsp-3 RNAi combined with RNAi against ire-1 or atf-6 showed significant reductions in F44E5.4 transcript levels (Fig 5E). This effect was most prominent in animals grown on hsp-3 + ire-1 RNAi, which showed a significant reduction (approximately 80%) in F44E5.4 transcripts, while knock-down of atf-6 resulted in an approximate 50% reduction. Interestingly, combinatorial knock-down with pek-1 RNAi had no discernable effects on F44E5.4 expression levels. Correlating this result with our finding that loss of pek-1 prevents the beneficial effects of fic-1 KO in development assays, we speculate that this branch of the UPR^ER may regulate expression of other survival factors independent of F44E5.4. As a whole, these data mechanistically link UPR^ER signaling through all three branches in the absence of fic-1 to increased F44E5.4 expression, blunting misfolded protein toxicity.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. UPR^ER signaling upregulates expression of the cytosolic HSP70 family chaperone, F44E5.4, to suppress polyQ toxicity.

(A) Schematic depicting the design of constructs for whole-body over-expression of select HSP70 chaperones. (B) Development assay of the indicated strains comparing survival rates of Q40::YFP + fic-1 KO and Q40::YFP + F44E5.4 OE animals. (C-D) Development assays of Q40::YFP + fic-1 KO animals (C) or Q40::YFP + F44E5.4 OE animals (D). For all graphs (B-D), X-axes denote the RNAi conditions used. Graphs depict the percentage of animals that have reached the L4 stage of larval development when assessed at 72 hours. (E) Relative mRNA expression levels assessed by qPCR of F44E5.4 in Q40::YFP + fic-1 KO L4 larvae fed the indicated RNAis (X-axis), normalized to control. (F) Representative images of day 1 adult Q40::YFP (left), Q40::YFP + fic-1 KO (middle), and Q40::YFP + F44E5.4 OE (right) animals. (G) Quantification of the number of polyQ puncta in day 1 adult animals visualized in (E). (H) Quantification of thrashing rates of day 3 adult animals of the indicated genotypes (legend) when treated with the indicated RNAi (X-axis) beginning at day 1 of adulthood (see Fig 2A). For (B-D), translucent data points reflect technical replicates, while opaque data points represent the average of each biological replicate (n = 3). For (G-H), each translucent data point represents one individual worm while opaque data points reflect the average of each biological replicate (n = 3). In (G), at least 50 worms were assessed per genotype. For (H), at least 30 worms per genotype, per condition were scored. Error bars for all plots represent SD. For (B and G), a two-way ANOVA with Tukey’s post-hoc multiple comparisons test was performed, and for (C-E, and F) statistical significance was determined using an ordinary one-way ANOVA with Tukey’s post-hoc multiple comparisons tests. ***p < 0.001; **p < 0.01; *p < 0.05; ns = not significant.

https://doi.org/10.1371/journal.pgen.1011723.g005

Expanding our characterization of this pathway beyond the developmental stages, we found that adult Q40::YFP; F44E5.4 OE worms showed an increased number of polyQ puncta on par with that observed in Q40::YFP; fic-1 KO animals, though no differences in puncta size distribution were noted (Figs 5E–5F, S10C). As an additional measure of adult worm fitness, we assessed thrashing rates. F44E5.4 OE worms on a wild-type background showed no difference in thrashing rates compared to wild-type (N2) controls (S10D Fig). In day 3 adults on control RNAi, there was no difference in thrashing between Q40::YFP, those lacking fic-1, or those over-expressing F44E5.4. However, Q40::YFP + fic-1 KO and Q40::YFP + F44E5.4 OE animals were both significantly protected from declines in thrashing rates observed in Q40::YFP animals fed hsp-3 or hsp-4 RNAi in adulthood (Fig 5G). Taken as a whole, these findings indicate that the absence of fic-1 under ER stress conditions facilitates UPR^ER signaling to upregulate cytosolic chaperone levels to mitigate polyQ toxicity. Further, the whole-body over-expression of one of these chaperones, F44E5.4, is sufficient to avert developmental arrest due to compromised ER homeostasis and protect against fitness declines in adulthood.

Limitations of this study

Some C. elegans tissues, such as the nervous system and germline, are insensitive to the effects of RNAi-mediated gene knock-down. As such, we are unable to draw conclusions about dependence on tissue-specificity for the pathway outlined in this study. The model we use for polyglutamine toxicity expresses a Q40 repeat peptide. While pathologically-expanded polyQ tracts are a shared hallmark of polyQ expansion diseases, the protein context surrounding this region is increasingly recognized to modulate cellular responses in a disease-specific manner. As such, we cannot exclude, and further hypothesize that the effects of AMPylation on polyQ toxicity are likely influenced by protein context. To validate the beneficial effects of F44E5.4 upregulation, we generated a whole-body F44E5.4 over-expression strain. While F44E5.4 over-expression indeed rescues polyQ toxicity in our model, this whole-body approach prevents us from gathering further mechanistic insights into the cell types or tissues required for this rescue. Lastly, we have previously identified additional targets of FIC-1-mediated AMPylation in vitro beyond HSP-3 and HSP-4, including translation elongation factors (EEF-1A.2, EEF-1G, EEF-2), histone H3, and the HSP70 family chaperone HSP-1 (ortholog of human HSC70) [26]. While, unlike HSP-3 and HSP-4, these targets have not yet been shown to be AMPylated in vivo, we recognize that fic-1 deletion may potentially impact additional AMPylation targets that could be involved in proteostasis regulation.