ADARs employ a neural-specific mechanism to regulate PQM-1 expression and survival from hypoxia

Abstract

The ability to alter gene expression programs in response to changes in environmental conditions is central to the ability of an organism to thrive. For most organisms, the nervous system serves as the master regulator in communicating information about the animal’s surroundings to other tissues. The information relay centers on signaling pathways that cue transcription factors in a given cell type to execute a specific gene expression program, but also provide a means to signal between tissues. The transcription factor PQM-1 is an important mediator of the insulin signaling pathway contributing to longevity and the stress response as well as impacting survival from hypoxia. Herein, we reveal a novel mechanism for regulating PQM-1 expression specifically in neural cells of larval animals. Our studies reveal that the RNA binding protein, ADR-1, binds to pqm-1 mRNA in neural cells. This binding is regulated by the presence of a second RNA binding protein, ADR-2, which when absent leads to reduced expression of both pqm-1 and downstream PQM-1 activated genes. Interestingly, we find that neural pqm-1 expression is sufficient to impact gene expression throughout the animal and affect survival from hypoxia; phenotypes that we also observe in adr mutant animals. Together, these studies reveal an important post-transcriptional gene regulatory mechanism that allows the nervous system to sense and respond to environmental conditions to promote organismal survival from hypoxia.

Introduction

Aerobic heterotrophs need to obtain nutrition and oxygen from the environment, the prolonged absence of which can lead to undesirable consequences including death. However, fluctuations in oxygen and nutrient availability are common in nature and during development; thus, organisms must have a means to both sense the environment and respond. At the most extreme, animals can effectively halt developmental and cellular programs resulting in a transient quiescent state [1]. For example, in the model organism Caenorhabditis elegans (C. elegans), the absence of oxygen can lead to a state of “suspended animation” [2], while first larval stage (L1) animals hatched in the absence of food enter a state of halted development commonly referred to as “L1 arrest” [3, 4].

An animal’s ability to enact these responses relies on the presence of a nervous system that can translate environmental information into physiological responses [5]. Neural gene expression programs are critical for organismal survival to many stresses. However, the nervous system must also communicate information about the environment to other tissues to promote diverse outputs, including behaviors and metabolic changes needed for organismal survival [6–8]. This trans-tissue communication is difficult to study in humans, but, altered brain-gut communication has been implicated in both oncogenesis and neurodegenerative diseases [9, 10]. In contrast, studies in model organisms have been instrumental in demonstrating that the nervous system signals to the peripheral tissues to promote survival and longevity in response to stress [11]. Furthermore, recent data indicates that not only do some peripheral tissues receive the stress signals from the nervous system, but tissues like the intestine can also serve as an important regulatory organ, sending signals back to the nervous system to promote health and longevity [12].

The molecular players underlying the response to environmental conditions are conserved signaling pathways. The major signaling pathways that respond to nutrients include the AMP-activated protein kinase and Target of Rapamycin pathways, while the Hypoxia-inducible factor 1 pathway responds to low oxygen (hypoxia) [13, 14]. In addition, there is cross-talk between the insulin signaling pathway and all of these pathways [14–16], thereby making insulin signaling a key determinant of how animals respond to diet and environmental fluctuations in oxygen levels. In C. elegans, the insulin signaling pathway has only one known receptor, DAF-2, which is homologous to both the mammalian insulin receptor and insulin-like growth factor 1 receptor [17]. Despite having only one receptor, there are over 40 insulin-like peptides (ILPs) encoded in the C. elegans genome [18]. Binding to these agonists and antagonists can influence the ability of DAF-2 to signal to downstream kinases that in turn regulate at least two transcription factors, the well-established FOXO homolog DAF-16 and/or the more recently identified zinc finger protein PQM-1 [19, 20].

Genetic dissection of the insulin signaling pathway, particularly the study of temperature-sensitive loss of function mutations of daf-2, has identified a central role for this pathway in regulating development, longevity, metabolism and reproduction [19, 21, 22]. In addition to genomic mutations, which impact signaling throughout the animal, it is well established that the nervous system is the critical site of action for insulin signaling to regulate diverse aspects of C. elegans physiology [23]. For example, neuronal-specific expression of the insulin receptor rescues both the long-lived phenotype observed in adult animals with altered DAF-2 function as well as the formation of dauer larva, a developmentally arrested life stage that is induced by over-crowding and starvation in wildtype C. elegans but constitutively occurs in daf-2 mutant larval animals [24, 25]. Recently a novel regulatory mechanism for altering insulin signaling via alternative splicing of daf-2 in neurons was identified [26]. The resulting DAF-2B protein retains the extracellular domain but lacks the intracellular domains to mediate downstream signaling, which allows the DAF-2B protein to bind ILPs and influence insulin signaling by competing with full-length DAF-2. Consistent with this, the presence of DAF-2B influences dauer entry and recovery as well as lifespan, further supporting the idea that nervous system-specific regulation of the insulin signaling pathway is important.

Previous studies from our lab identified adenosine (A) to inosine (I) RNA editing sites in daf-2 mRNA isolated from neural cells of L1-arrested animals [27]. Due to differences in base-pairing properties of adenosine and inosine, A-to-I editing events can impact gene expression depending on the region of RNA in which the editing event occurs [28, 29]. For example, A-to-I editing within coding sequences of genes can alter the protein encoded by the gene and editing within 3’ untranslated regions (UTRs) can alter small RNA binding [30]. For the daf-2 transcript, the A-to-I editing sites identified are located within an intronic sequence, which could potentially impact the production of daf-2 splice isoforms. To begin to understand if RNA editing influences DAF-2 function, we examined gene expression changes that occur in neural cells in the absence of the enzyme that is responsible for catalyzing the hydrolytic deamination of adenosine to inosine, ADR-2. ADR-2 is a member of the adenosine deaminase that act on RNA (ADAR) family, and the C. elegans genome encodes two ADAR family members ADR-1 and ADR-2 [31]. ADR-2 is the sole A-to-I editing enzyme in C. elegans [32, 33], as ADR-1 lacks essential amino acids required to perform deamination [32]. However, as ADARs are RNA binding proteins (RBPs) that can also regulate gene expression through binding RNA [34], both ADR-1 and ADR-2 may play roles in editing-independent gene regulation in C. elegans. Interestingly, the long-lived phenotype of animals lacking daf-2 is also observed in adr-2(−) mutants [24, 35]. Together, these data suggested the possibility of C. elegans ADARs impacting insulin signaling. While a role for ADARs in regulating insulin signaling is relatively unexplored, recent work in β cells indicated that the pathophysiological environment of type 1 diabetes patients influences RNA editing [36]. In this work, we sought to determine how ADARs can affect the insulin signaling pathway, particularly in the nervous system of L1-arrested C. elegans.

Results

Decreased expression of genes regulated by insulin signaling upon loss of adr-2

As a first step towards addressing whether ADR-2 regulates insulin signaling, the transcriptomes of wildtype and adr-2 deficient animals were compared. As editing of daf-2 was observed in neural cells isolated from synchronized L1 animals [27], differential gene expression was analyzed in RNA isolated from these same types of biological samples. Using datasets from previously performed RNA-sequencing (RNA-seq) of three biological replicates of wildtype and adr-2(−) neural cells from synchronized L1 animals [37], differential gene expression analysis identified 697 genes significantly altered in neural cells from adr-2(−) animals (p value < 0.05 and log2fold change > |0.5|), with nearly three times as many downregulated genes (501) as upregulated genes (196) (Figure 1A, Supplemental Table S1). These misregulated genes were subjected to gene set enrichment analysis using a C. elegans specific software, WormCat [38]. The analysis for genes upregulated in neural cells from adr-2(−) animals revealed only one significantly enriched gene set, extracellular material (Supplemental Figure S1). The downregulated genes were enriched for four gene sets: stress response, proteolysis, metabolism and lysosome (Supplemental Figure S1). As DAF-2-mediated signaling regulates both stress response and metabolism [19], this suggests that loss of adr-2 might result in altered insulin signaling in neural cells. However, it should be noted that we did not observe significant changes in daf-2 mRNA expression in adr-2(−) neural cells (Supplemental Table S1).

Figure 1-. L1 animals lacking adr-2 have decreased expression of genes regulated by insulin signaling.

(A) Volcano plot depicting gene expression in adr-2(−) neural cells compared to wildtype (WT) neural cells. Dots represent individual genes that are upregulated (red; 186, p < 0.05, log2fold > 0.5), downregulated (blue; 502, p < 0.05, log2fold < −0.5) or not significantly different (grey, p ≥ 0.05) between three biological replicates of WT and adr-2(−) neural cells. (B-D) Expression of the indicated genes was determined relative to expression of the housekeeping gene gpd-3. Values were then normalized to wildtype neural cells (B) or wildtype L1 animals hatched in the absence of food (C, D) and the mean of three biological replicates was plotted. Error bars represent standard error of the mean (SEM). Statistical significance was calculated by two-way ANOVA test. **p < 0.005, ****p < 0.0001. ns indicates not significant (p > 0.05). For D, the indicated genotypes are of strains HAH30, HAH31, HAH32 and HAH33.

To independently examine the expression of genes regulated by DAF-2, qPCR was performed for three known downstream targets in three independent biological replicates of neural cells isolated from wildtype and adr-2(−) animals. Consistent with the RNA-seq dataset, all three genes (dod-17, dod-19, and dod-24) examined were significantly downregulated in adr-2(−) neural cells compared to wildtype neural cells (Figure 1B). These results suggest that, within neural cells, there is decreased expression of genes regulated by DAF-2 upon loss of adr-2.

While this data suggests that loss of adr-2 affects genes regulated by insulin signaling in the nervous system, the nervous system is also the master regulator that coordinates gene regulation between tissues [39]. Studies have demonstrated that DAF-2 function in the nervous system can affect phenotypes such as organismal lifespan by signaling to other tissues [25]. This raised the question of whether the decreased expression of DAF-2-regulated genes upon adr-2 loss could be observed in RNA isolated from whole L1 animals. Furthermore, the neural cells were isolated from L1 animals that were synchronized by hatching in the absence of food, and nutrient levels impact insulin signaling [40]. Hence, gene expression was examined in RNA isolated from three independent biological replicates of synchronized wildtype and adr-2(−) L1 animals as well as a subset of these hatched L1 animals that were exposed to bacterial food for six hours. Similar to neural cells, all three genes (dod-17, dod-19, and dod-24) exhibited significantly decreased expression in hatched L1 animals lacking adr-2 compared to wildtype animals (Figure 1C). This suggests that the impacts of loss of adr-2 on altered neural gene expression (Figure 1A, B) may lead to cell non-autonomous effects and/or that adr-2 regulates genes downstream of insulin signaling in several tissues. In contrast to the hatched L1 animals, there was no significant difference in expression of the three genes (dod-17, dod-19, and dod-24) between wildtype and adr-2(−) animals after feeding for six hours (Figure 1C).

Together these data suggest that upon loss of adr-2, reduced expression of genes regulated by insulin signaling occurs in whole animals and is abrogated by the presence of food. Since the observed decreased gene expression was specific to starved adr-2(−) animals and insulin signaling is known to impact starvation responses in C. elegans, these data suggested that adr-2 regulates dod-17, dod-19, and dod-24 through the insulin signaling pathway. To directly test this possibility, genetic mutants of adr-2 and daf-2 [41] were combined and gene expression was examined. As observed consistently in this study, compared to wildtype animals, dod-17, dod-19, and dod-24 were all significantly reduced in adr-2(−) animals (Figure 1D). Consistent with previous studies [42], dod-17, dod-19, and dod-24 all exhibited significantly decreased expression in daf-2(−) animals (Figure 1D). Compared to wildtype animals, there was a significant reduction in mRNA expression of all three insulin signaling regulated genes in the adr-2(−);daf-2(−) double mutants (Figure 1D). Additionally, the expression profile of dod-17, dod-19, and dod-24 was similar between the adr-2(−) and daf-2(−) single mutants and the adr-2(−);daf-2(−) double mutants (Figure 1D). Together, these data indicate that lack of adr-2 leads to decreased expression of insulin signaling regulated genes and the regulation occurs through the DAF-2 pathway.

Regulation of insulin signaling by ADR-2 is cell non-autonomous and editing-independent

In addition to being expressed in the nervous system, genes regulated by insulin signaling are also highly expressed in the intestine [43, 44]. As the RNA isolated from L1 animals hatched in the absence of food shows decreased gene expression similar to neural cells, this suggests that loss of adr-2 can potentially result in altered insulin-signaling in both neural and intestinal cells. However, there is extensive communication between neural cells and the intestine; thus, it is possible that loss of adr-2 in neural cells is sufficient to result in altered insulin signaling in the intestine. To test this possibility, gene expression was monitored in animals that express ADR-2 only in the nervous system [37]. Briefly, these transgenic animals were generated by injecting adr-2(−) animals with a plasmid construct in which a pan-neural promoter rab-3 drives the expression of adr-2 along with a co-injection marker expressing GFP. A similar strain was previously generated by our lab and shown to result in ADR-2 activity in the nervous system [37]. Since transgenes do not exhibit 100% inheritance in C. elegans, the transgenic animals of interest were sorted for GFP expression using the COPAS Select large particle sorter. Optimized COPAS conditions in terms of extinction and time of flight were used to create a gated window for specifically sorting L1 animals. RNA was isolated from the sorted animals and qPCR was performed to assess gene expression. It was observed that compared to adr-2 lacking animals, animals expressing ADR-2 solely in the nervous system significantly rescued expression of genes regulated by insulin signaling (Figure 2A). This data suggests that the presence of ADR-2 in the nervous system regulates insulin signaling throughout the animals.

Figure 2-. Neural ADR-2 regulates insulin signaling cell non-autonomously in an editing independent manner.

(A, C) Gene expression of L1-arrested animals measured by qPCR. Expression of the indicated genes was determined relative to expression of the housekeeping gene gpd-3. Values were then normalized to WT and the mean of three (C) or four (A) biological replicates was plotted. Error bars represent standard error of the mean (SEM). Statistical significance was calculated by two-way ANOVA test. ****p < 0.0001, **p < 0.005, *p < 0.05 and ns indicates not significant (p > 0.05). For A, the indicated genotypes are of strains HAH23, HAH40 and HAH41. For C, the indicated genotypes are of strains N2, BB20 and HAH22 (B) A dashed line was used to outline the whole worm. For all the strains, the images are representative of 7–10 samples imaged in two biological replicates.

To further validate that neural ADR-2 regulates insulin signaling regulated genes throughout the animal, confocal microscopy was performed to monitor expression of one of these genes, dod-24, upon loss of adr-2 in whole animals. Synchronized L1 animals expressing GFP driven by the dod-24 promoter [43] were analyzed. In wildtype animals, transcription from the dod-24 promoter was observed in neural as well as intestinal cells as expected [43] (Figure 2B). Upon loss of adr-2, decreased GFP expression was observed throughout the animal (Figure 2B), which is consistent with the qPCR analysis of dod-24 expression (Figure 1C). Furthermore, in animals expressing ADR-2 solely in the nervous system, GFP expression was similar to wildtype animals (Figure 2B). These results demonstrate that the presence of ADR-2 in the nervous system cell non-autonomously impacts gene expression.

To begin to dissect the molecular function of ADR-2 in the nervous system that contributed to the altered gene regulation, expression of dod-17, dod-19, and dod-24 was examined in animals expressing an ADR-2 mutant (ADR-2 G184R) that can bind RNA, but lacks the ability to edit [27]. As observed consistently in this study, hatched L1 adr-2(−) animals had significantly decreased expression of insulin signaling regulated genes compared to wildtype animals (Figure 2C). In contrast, gene expression in the ADR-2 G184R animals was similar to wildtype animals (Figure 2C). Together, these results indicate that, while the presence of ADR-2 in the nervous system is critical for proper dod-17, dod-19, and dod-24 expression throughout the animal, the editing function of ADR-2 is not required for this gene regulatory function.

Neural pqm-1 levels impact downstream gene expression throughout the animal

As the ADR-2 editing function was not required for altered dod-17, dod-19, and dod-24 expression, editing of daf-2 is unlikely to be causing the decreased expression of downstream insulin signaling regulated genes in neural cells and animals lacking adr-2. As DAF-2 is at the top of the insulin-signaling regulatory cascade and transcriptional output is mediated by at least two different transcription factors, DAF-16 and PQM-1, we sought to examine whether all DAF-2 regulated genes were equally affected by loss of adr-2. The up- and downregulated genes in the adr-2(−) neural RNA-seq dataset were individually overlapped with either DAF-16 activated or PQM-1 activated genes from a published dataset [20] (Supplemental Figure S2). The number of PQM-1 activated genes that were downregulated in adr-2(−) neural cells (156) was nearly three times the number that would be expected by random chance (53) (Supplemental Figure S2) and dod-17, dod-19 and dod-24 are all genes activated by PQM-1 [20]. The number of overlapping PQM-1 activated genes that were upregulated in adr-2(−) neural cells (11) was lesser than that obtained due to random chance (20) (Supplemental Figure S2). Further, the number of overlapping DAF-16 activated genes either up- (23) or downregulated (61) in adr-2(−) neural cells was very close to what would be expected from random chance (Supplemental Figure S2). These results suggest that loss of adr-2 does not impact all genes downstream of the insulin signaling pathway, but instead, leads to specific downregulation of PQM-1-activated genes.

The above data raised the question of whether loss of adr-2 directly impacts pqm-1 expression. To address this question, pqm-1 expression was monitored in both neural cells and L1 animals using qPCR. In the neural cells from adr-2(−) animals, there was a significant decrease in pqm-1 expression compared to neural cells isolated from wildtype animals (Figure 3A). Consistent with this finding, our neural RNA-seq datasets also revealed that loss of adr-2 indeed resulted in significantly decreased neural expression of pqm-1 (Supplemental Table S1). In contrast to neural cells, pqm-1 expression was not significantly altered in RNA isolated from synchronized L1 animals lacking adr-2 (Figure 3B). Together, this data indicates that the lack of adr-2 impacts pqm-1 expression in a tissue-specific manner. The data also suggests that decreased neural expression of pqm-1 upon loss of adr-2 could impact gene expression throughout the animal.

Figure 3-. Neural PQM-1 activity is sufficient to rescue expression of PQM-1 activated genes in adr-2(−) animals.

(A-D) Gene expression of (A) neural cells and (B-D) L1-arrested animals measured by qPCR. Expression of indicated genes was determined relative to expression of the housekeeping gene gpd-3. Values were then normalized to WT and the mean of three biological replicates was plotted. Error bars represent standard error of the mean (SEM). Statistical significance was calculated by (A) multiple unpaired t tests and (B-D) two-way ANOVA test. ****p < 0.0001 and ns indicates not significant (p > 0.05). For A and B, the indicated genotypes are of strains HAH45 and HAH46. For C, the indicated genotypes are of strains HAH35, HAH37, HAH38 and HAH39. For D, the indicated genotypes are of strains HAH24, HAH41, HAH42, HAH43 and HAH44.

To directly address whether the decreased expression of PQM-1-activated genes in adr-2(−) animals is due to decreased pqm-1 expression, gene expression was assessed in animals lacking pqm-1(−) and adr-2(−);pqm-1(−) animals. Briefly, pqm-1(ok485) animals [45] were obtained and backcrossed to wildtype animals before crossing to adr-2(−) animals. RNA was then isolated from wildtype, adr-2(−), pqm-1(−) and adr-2(−);pqm-1(−) animals obtained from the genetic cross. As expected, compared to wildtype animals, adr-2(−) animals showed significantly decreased expression of dod-17, dod-19 and dod-24 (Figure 3C). Similarly, loss of pqm-1 resulted in significantly decreased expression of dod-17, dod-19 and dod-24 compared to wildtype animals (Figure 3C). These PQM-1 activated genes also showed a significant reduction in expression in adr-2(−);pqm-1(−) animals (Figure 3C), which was similar to the expression levels observed in the individual adr-2(−) and pqm-1(−) mutant animals (Figure 3C). These results suggest that the decreased expression of dod-17, dod-19 and dod-24 upon loss of adr-2 occurs via altered PQM-1 function.

As our results clearly indicate that loss of adr-2 leads to downregulation of pqm-1 specifically in the nervous system, but decreased expression of PQM-1 activated genes throughout L1 animals, we sought to test whether expressing pqm-1 only within the nervous system of adr-2(−) animals could restore dod-17, dod-19 and dod-24 gene expression in L1 animals. Transgenic animals were generated by injecting a plasmid in which pqm-1 expression is driven by the neuronal rab-3 promoter. As a control, this plasmid was first injected into pqm-1(−) animals to observe gene expression changes in pqm-1(−) animals. The resulting transgenic animals were crossed with adr-2(−) animals and genotyped for either pqm-1(−) or adr-2(−) animals that specifically express pqm-1 in the nervous system. Compared to wildtype animals and consistent with other results in this study, there was significantly decreased expression of dod-17, dod-19 and dod-24 in pqm-1(−) animals (Figure 3D). However, pqm-1(−) animals expressing pqm-1 only in the nervous system had significantly increased expression of dod-17, dod-19 and dod-24, which rescued the gene expression to near wildtype levels (Figure 3D). As observed throughout this study, adr-2(−) animals exhibited significantly decreased expression of PQM-1 activated genes compared to wildtype animals (Figure 3D). However, adr-2(−) animals carrying the neural pqm-1 transgene exhibited significantly increased expression of dod-17, dod-19 and dod-24 compared to animals lacking adr-2 (Figure 3D). Together, these data indicate that lack of adr-2 leads to global downregulation of dod-17, dod-19 and dod-24 via decreased expression of the PQM-1 transcription factor in the nervous system.

In the absence of adr-2, ADR-1 binds to pqm-1 mRNA and results in decreased expression of PQM-1 activated genes

The above data suggest that loss of adr-2 results in decreased pqm-1 expression in the nervous system, which leads to decreased expression of PQM-1 activated genes throughout the animal. Since the editing function of ADR-2 is not required for the downregulation of these PQM-1 activated genes (Figure 2C), we sought to test what other function of ADR-2 was critical for regulating pqm-1 expression. ADR-2 directly interacts with ADR-1, a deaminase-deficient member of the ADAR family present in C. elegans [46]. The physical interaction between ADR-1 and ADR-2 can both promote ADR-2 binding to RNA [46], as well as influence the RNAs that ADR-1 binds [35]. However, the biological impacts of this latter function are relatively unknown. To assess if PQM-1 activation of gene expression is altered upon loss of adr-1, RNA was isolated from hatched wildtype, adr-1(−), adr-2(−) and adr-1(−);adr-2(−) L1 animals, and qPCR was performed to measure gene expression of dod-17, dod-19 and dod-24 in these animals. Consistent with the data obtained in this study, adr-2(−) animals had decreased expression of the PQM-1 activated genes compared to wildtype animals (Figure 4A). In contrast, expression levels of dod-17, dod-19 and dod-24 were similar between wildtype and adr-1(−) animals (Figure 4A), suggesting that loss of ADR-1 function did not affect PQM-1-mediated gene regulation in wildtype animals. Interestingly, loss of adr-1 in animals lacking adr-2 significantly increased expression of PQM-1 activated genes compared to animals lacking only adr-2 (Figure 4A). These results suggest that ADR-1 has a unique function in the absence of adr-2, which results in decreased expression of PQM-1 activated genes.

Figure 4-. Neural ADR-1 binding of pqm-1 affects expression of PQM-1 activated genes.

(A,C) Gene expression of L1-arrested animals measured by qPCR. Expression of the indicated genes was determined relative to expression of the housekeeping gene gpd-3. Values were then normalized to WT and the mean of three biological replicates was plotted. Error bars represent standard error of the mean (SEM). Statistical significance was calculated by two-way ANOVA test. ***p < 0.0005, **p < 0.01, *p <0.05. For A, the indicated genotypes are of strains N2, BB19, BB20 and BB21. For C, the indicated genotypes are of strains HAH48, HAH49 and HAH50. (B) Western blot depicting immunoprecipitation of neural ADR-1 from the indicated strains. Bar graph represents the fold enrichment determined by dividing IP/Input value from qPCR for the indicated strains divided by that of negative control. Values were then normalized to negative control and the mean of three biological replicates was plotted. Error bars represent SEM. Statistical significance was calculated by multiple unpaired t tests. **p < 0.005.

So far, the data suggests that upon loss of adr-2, decreased pqm-1 expression in neural cells leads to global downregulation of PQM-1 activated genes and loss of adr-1 can rescue these downstream gene expression changes. As ADR-1 is an RNA binding protein, we sought to determine whether ADR-1 directly binds pqm-1 mRNA specifically in the nervous system, and whether that binding is influenced by the presence or absence of adr-2. To examine binding of ADR-1 to pqm-1 in the L1 nervous system, an RNA immunoprecipitation (RIP) assay was performed with animals that express ADR-1 specifically in the nervous system. To generate these animals adr-1(−) animals were injected with a construct in which the neuronal rab-3 promoter drives expression of an N-terminally 3X FLAG adr-1 genomic sequence. A similar epitope tagged construct under the control of the adr-1 endogenous promoter was previously demonstrated to produce functional ADR-1 protein [33]. Hatched L1 animals were subjected to UV crosslinking to stabilize RNA-protein interactions prior to generation of protein lysates. ADR-1 and associated bound RNAs were immunoprecipitated using magnetic FLAG beads. ADR-1 was efficiently immunoprecipitated from animals expressing ADR-1 in the nervous system both in the presence and absence of adr-2, but not from lysates of the negative control adr-1(−) animals (Figure 4B). On assessing pqm-1 mRNA in the assay, compared to the negative control, there was no enrichment for pqm-1 mRNA in IPs from neural ADR-1-expressing animals that expressed wildtype ADR-2 (Figure 4B). However, a 5-fold enrichment of pqm-1 mRNA was observed in the neural ADR-1 RIP in the absence of adr-2 (Figure 4B). This data suggests that in vivo, ADR-1 binds pqm-1 in the nervous system, but only in the absence of adr-2.

To further examine whether the RNA binding function of ADR-1 is contributing to the decreased expression of PQM-1 activated genes observed in adr-2(−) animals, dod-17, dod-19 and dod-24 expression was monitored in adr-2 lacking animals that also have abolished ADR-1 binding. These mutant animals have three mutations within the conserved KKxxK motif (where K is lysine and x is any amino acid) of the first dsRNA binding domain (dsRBD1) of ADR-1 (K223E, K224A, K227A), which was previously shown to disrupt the ability of ADR-1 to bind RNA in vivo [46]. The mutation was introduced in wildtype animals with an integrated 3X FLAG tag at the adr-1 locus via CRISPR by using a guide RNA targeted to the adr-1 locus and a HDR template containing the desired mutations. These animals were then crossed to adr-2(−) animals to generate adr-2(−) animals that also lack the ADR-1 binding function. RNA was isolated from these animals as well as wildtype and adr-2(−) animals and qPCR was performed. Compared to adr-2(−) animals, adr-2(−); ADR-1 dsRBD1 mutants had significantly increased expression of dod-17, dod-19 and dod-24 (Figure 4C). Together, these results indicate that in the absence of adr-2, ADR-1 binds pqm-1 in the nervous system, which leads to decreased expression of PQM-1 activated genes throughout the animal.

PQM-1 functions in the nervous system to regulate hypoxia survival of L1-arrested animals

Previous studies have indicated that PQM-1 is a negative regulator of hypoxic survival in fourth stage larval (L4) animals [47]. As our data indicates that adr-2(−) animals have decreased pqm-1 expression, we sought to determine if these animals also had altered survival to hypoxic exposure. Additionally, since the data so far indicate that ADARs regulate pqm-1 expression specifically in the nervous system, whether neural PQM-1 specifically plays a role in survival to hypoxia was of interest.

To directly test these questions, hatched L1 animals were exposed to varying concentrations of cobalt chloride (CoCl₂), which serves as a hypoxia mimetic [48]. As we wanted to address the contribution of neural PQM-1 in regulating hypoxia, survival of the neural-specific pqm-1 transgenic animals was examined. However, as the neural pqm-1 strains were transgenic, to avoid any non-specific effects of the coinjected transgenes, survival was compared in animals that all carry the prab-3::GFP transgene. Wildtype, pqm-1(−), pqm-1(−) animals expressing pqm-1 in the nervous system, adr-2(−) and adr-2(−) animals overexpressing pqm-1 in the nervous system were used in the hypoxic survival experiment. After obtaining hatched L1 animals, 5000 animals per strain were washed with NaCl and then exposed to varying concentrations of CoCl₂ (0 – 80 mM) for two hours. After the CoCl₂ exposure, three technical replicates of GFP positive 30–40 L1 animals per strain were plated and incubated for twenty-four hours at 20°C in the presence of food. To measure hypoxic survival, alive and dead animals were counted for all the strains and plotted for all concentrations of CoCl₂. At lower concentrations of CoCl₂ (2.5 mM and 5 mM), the number of alive L1s was similar across all worm strains (Figure 5A). However, at 10 mM and higher concentrations of CoCl₂, consistent with previous studies [49], there was a drastic reduction in the survival of wildtype animals (Figure 5A). With increasing concentrations of CoCl₂, the number of alive pqm-1(−) and adr-2(−) animals was significantly higher than wildtype animals (Figure 5A, Supplemental Figure S3), suggesting an increased hypoxic survival of these animals. Strikingly, there was a sharp decline in the survival of pqm-1(−) animals expressing neural pqm-1 compared to pqm-1(−) animals (Figure 5A, Supplemental Figure S3). This data suggests that PQM-1 function within the nervous system is sufficient to regulate hypoxic survival. Interestingly, a similar result was observed with adr-2(−) animals with transgenic expression of pqm-1 in the nervous system (Figure 5A, Supplemental Figure S3). Together, these data suggest that PQM-1 in the nervous system is a critical regulator of hypoxia survival in hatched L1 animals and that this function of PQM-1 impacts the survival of animals lacking adr-2 in hypoxic environments.

Figure 5-. Survival of hatched L1 animals after CoCl₂ exposure.

(A-B) Survival of transgenic (A) and non-transgenic (B) hatched L1 animals under hypoxic conditions induced by CoCl₂ exposure. Data plotted is average of three biological replicates. Error bars represent standard error of the mean (SEM).

As our molecular data suggest that ADR-1 binding to pqm-1 mRNA in the absence of adr-2 results in altered pqm-1 expression, we sought to determine if loss of ADR-1 RNA binding could influence the survival of adr-2(−) animals to hypoxia. An independent set of hypoxia survival experiments were performed using wildtype, pqm-1(−) and adr-2(−) animals along with the adr-2(−);ADR-1 dsRBD1 mutant animals. Consistent with the transgenic animals assayed in our study, compared to wildtype animals, both pqm-1(−) and adr-2(−) animals showed a significantly increased hypoxic survival (Figure 5B, Supplemental Figure S3). Compellingly, adr-2(−) animals lacking ADR-1 RNA binding function showed hypoxic survival similar to wildtype animals (Figure 5B). Together, these results suggest that neural PQM-1 is a key mediator of hypoxic survival and that binding of ADR-1 to pqm-1 mRNA in the nervous system affects the animal’s ability to survive hypoxic stress.

Discussion

In these studies, we determined the tissue-specific contributions of ADAR proteins in regulating the insulin signaling pathway in C. elegans. Our data revealed unique ADR-1 RNA binding that occurs in the nervous system specifically in the absence of adr-2. As a previous study indicated heterodimer formation between ADR-1 and ADR-2 [46], studies have focused on understanding how ADR-1 facilitates ADR-2 binding to mRNAs to promote editing [27, 33, 46]. However, several other studies have also revealed that ADR-1 competes with ADR-2 for binding certain mRNAs, which can lead to decreased editing levels in transcripts [37, 50]. Together these studies suggest that the relationship between ADR-1 and ADR-2 is complex and may vary based on the tissue and developmental stage of the animals as well as on the individual transcript. In previous transcriptome-wide studies of ADR-1 mRNA binding, it was noted that ADR-1 binds nearly 1200 transcripts in wildtype worms and while ADR-1 is bound to nearly 80% of these in animals lacking adr-2, ADR-1 also uniquely bound nearly 400 mRNAs in the absence of adr-2 [35]. The impact of ADR-1 binding to these unique targets has not been investigated. However, as our study identifies important biological consequences of ADR-1 binding to pqm-1 specifically in the absence of adr-2, the impact of ADR-1 binding should be explored further. Additionally, as we observed that, upon feeding, adr-2(−) animals did not exhibit altered expression of genes downstream of the insulin signaling pathway (Figure 1C), it is possible that the presence of nutrients impacts ADR-1 binding to pqm-1. How environmental factors can influence ADAR target recognition is relatively unexplored and is an exciting future direction.

Our studies also revealed that loss of adr-2 did not globally impact insulin signaling in neural cells, but instead, lead to further repression of genes negatively regulated in response to reduced insulin signaling (Supplemental Figure S2). Previous studies have indicated that the promoters of these genes contain an overrepresented sequence (CTTATCA), referred to as the DAE (DAF-16 associated element) [42] due to changes in expression of these genes upon loss of daf-16 but lack of direct binding by DAF-16 [20]. The downregulation of over 150 DAE-containing genes in neural cells lacking adr-2 suggested a global regulator of these genes was altered. Previous studies have identified the transcription factor PQM-1 as an important factor regulating DAE-containing genes in L4 animals [20]. Herein, we found that loss of adr-2 resulted in a neural-specific decrease in pqm-1 expression, but decreased expression of the DAE-containing genes, dod-17, dod-19 and dod-24 in both neural cells and the intestine. Transcription from the pqm-1 promoter was observed in the nervous system and the intestine in early studies surveying transcription factor expression in C. elegans [51]. More recent studies reported that PQM-1:GFP translational fusions exhibited strong intestinal expression, but neural expression was not observed [20, 52]. With these observations, it is not surprising that several recent studies have reported that PQM-1 has important impacts on intestinal gene expression [47, 53]. However, consequences of loss of pqm-1 specifically in neural cells has also been previously reported for transgenic animals that have altered proteostasis networks, specifically transcellular chaperone signaling [52]. Here, we have added to that body of work by identifying molecular changes that occur in neural cells with reduced pqm-1 expression.

Our studies also indicate that pqm-1 expression solely within the nervous system is sufficient to promote expression of dod-17, dod-19 and dod-24 throughout the L1-arrested animal, providing the first evidence that PQM-1 can regulate gene expression in a cell non-autonomous manner. At present, it is unclear whether the cell non-autonomous regulation is dependent upon insulin signaling or if the downstream genes are affected by loss of daf-2 [42] due to the impacts of DAF-2 on PQM-1 function in the nervous system. Our neural RNA-seq data did identify several insulin-like peptides (ILPs), including ins-4, ins-5, ins-26, ins-35 and daf-28 with significantly reduced expression in adr-2(−) neural cells (Supplemental Figure S4). Previous studies have shown that all of these ILPs are DAF-2 agonists [54–59]. Thus, it is possible that PQM-1 promotes expression of signaling molecules in the nervous system that relay information to the intestine to promote expression of the DAE-containing genes.

Another major unanswered question is what transcription factor could be mediating transcriptional control of the DAE-containing genes in the intestine. Previous studies of C. elegans transcription factors revealed an enrichment in DAE-containing sequences within the bound regions of 13 different proteins [20]. While PQM-1 was at the top of this list, it is possible that one of the other 12 transcription factors could be promoting transcription of DAE-containing genes in the intestine. Focusing on the promoters of dod-17, dod-19 and dod-24 specifically, chromatin immunoprecipitation (ChIP) sequencing studies have identified binding sites for three transcription factors, PQM-1, FOS-1 and NHR-28 [60], but only FOS-1 and PQM-1 are present at the promoters of all three genes. FOS-1 expression and function in somatic gonad cells and anchor cells is well established [61] and to date, there is no evidence that FOS-1 is expressed in intestinal cells. However, it is possible that in L1-arrested animals, FOS-1 expression changes and intestinal transcriptional activity could occur. In this same context, it is possible that other transcription factors occupy the dod-17, dod-19 and dod-24 promoters within intestinal tissue to promote transcription during L1-arrest. Future studies should screen for factors needed specifically in the intestine for proper expression of DAE-containing genes; however, this may prove challenging in L1-arrested animals, where standard RNA interference (RNAi) by feeding cannot be employed.

Our work establishes a novel role for PQM-1 in hypoxic survival of L1-arrested animals. While a majority of L1-arrested wildtype animals die from acute exposure to high (≥10mM) doses of CoCl₂, pqm-1(−) animals exhibited significantly increased survival. It was previously reported that loss of pqm-1 resulted in increased survival of L4 animals placed either in a hypoxia chamber for 16 hours or exposed to low (5 mM) CoCl₂ for 20 hours [47]. Together these data indicate that PQM-1 is a negative regulator of hypoxic survival across developmental timescales. However, it is unclear whether the cellular role of PQM-1 is the same in larval and adult animals. Our phenotypic data revealed that PQM-1 function in neural cells is critical for its function as a negative regulator of hypoxic survival in L1-arrested animals. The previous study reported that PQM-1 promoted intestinal lipid levels and yolk protein transport to developing oocytes under oxygen depletion in adult animals [47]. However, it is important to note that direct binding of PQM-1 to the promoters of genes underlying the metabolic changes during adult hypoxic exposure was not demonstrated, thus it would be interesting to determine whether PQM-1 functions within the intestine or cell non-autonomously regulates hypoxic survival of adult animals.

At present, the physiological role for the negative regulation of survival by PQM-1 is unknown. Clearly on a cellular level, the ability to undergo metabolic changes that allow survival to hypoxia is a major aspect of oncogenesis; thus, players that keep this function in check are important. However, it is also well established in model organisms, including C. elegans, that adult animals reared under hypoxic conditions live longer than animals reared in normoxic conditions [62]. While promoting overall survival, the response to hypoxia is an energy intensive process that disrupts cellular proteostasis [63]; thus, to preserve energy and maintain equilibrium, it is likely equally important to control levels of hypoxic responses. Interestingly, a recent study revealed that fasted animals have an altered response to hypoxia and that the DAF-2 pathway, independent of DAF-16, plays an important role in this response [64]. It would be interesting to see if the role we identified for PQM-1 in L1 animals hatched in the absence of food is also involved in the coordinated response of adult animals to nutritional state and hypoxia. Furthermore, as exposure to limiting oxygen or nutrients has been reported to have transgenerational effects on descendants’ metabolic programming, behavior and fecundity in C. elegans [65–67], exploring the role of PQM-1 in these processes may shed light on the physiological role of negative regulation of hypoxic survival.

Materials and Methods

C. elegans strains and maintenance

All worms were maintained under standard laboratory conditions on nematode growth media seeded with Escherichia coli OP50 [68]. The following previously generated strains were used in this study: Bristol strain N2, BB19 (adr-1(tm668)) [69], BB20 (adr-2(ok735)) [69], BB21 (adr-1(tm668);adr-2(ok735)) [69], HAH22 (adr-2(gk777511) [70] agIs6[dod-24::GFP] [43], daf-2(m596) [41], pqm-1(ok485) [45]. Neural cells were isolated from HAH45 (prab3::rfp::C35E7.6 3’UTR; prab3::gfp::unc-54 3’ UTR; unc-119 genomic rescue) and HAH46 (adr-2(ok735); prab3::rfp::C35E7.6 3’UTR; prab3::gfp::unc-54 3’ UTR; unc-119 genomic rescue).

Strains generated in this study include HAH23 (BB20 + blmEx18(Y75B8A.8 3’ UTR hairpin construct in prab3::GFP::unc-54 3’ UTR (pHH340); prab3::3XFLAG ADR-2 cDNA::unc-54 3’ UTR (pHH438)), HAH24 (BB20 + blmEx19(Y75B8A.8 3’ UTR hairpin construct in prab3::GFP::unc-54 3’ UTR Clone vector (pHH340)), HAH25 (BB19 + blmEx20(prab3::GFP::unc-54 3’ UTR (pHH21); prab3::3XFLAG ADR-1::unc-54 3’ UTR (pHH512))), HAH26 (BB21+ blmEx20(prab3::GFP::unc-54 3’ UTR(pHH21); prab3::3XFLAG ADR-1::unc-54 3’ UTR (pHH512))), HAH27 (adr-2(ok735), agIs6[dod-24::GFP]), HAH28 (BB20 + blmEx18(Y75B8A.8 3’ UTR hairpin construct in prab3::GFP::unc-54 3’ UTR (pHH340); prab3::3XFLAG ADR-2 cDNA::unc-54 3’ UTR (pHH438), agIs6[dod-24::GFP]), HAH29 (adr-2(ok735), daf-2(m596)), HAH30 (wildtype), HAH31 (daf-2(m596)), HAH32 (adr-2(ok735)), HAH33 (adr-2(ok735), daf-2(m596)), HAH42 (HAH38+ blmEx19(Y75B8A.8 3’ UTR hairpin construct in prab3::GFP::unc-54 3’ UTR (pHH340))), HAH43 (HAH38+blmEx21(Y75B8A.8 3’ UTR hairpin construct in prab3::GFP::unc-54 3’ UTR (pHH340); prab3::pqm-1::unc-54 3’ UTR (pHH549))), HAH47 (3X FLAG ADR-1 dsRBD1 (K223E, K224A, K227A) CRISPR).

Animals created by microinjection (HAH23-HAH26, HAH28, HAH40-HAH44) used standard microinjection techniques and were passaged by selecting worms that contained the GFP co-injection marker. The injection mix contained 20 ng/μl of the co-injection marker and 1 ng/μl of the transgene of interest.

Animals created by CRISPR modification (HAH47) used standard microinjection techniques and were identified as rolling F1 progeny and non-rolling F2 progeny. Injection mix for the ADR-1 dsRBD1 mutant strain included 1.5 μM Cas9 (IDT, Alt-R Cas9 nuclease V3), 4 μM tracrRNA (IDT), 4μM of crRNA (IDT)(HH3088) (Supplemental Table S2), 37 ng/μl rol-6 plasmid (HAH293) and 4μM of repair template ssODN (HH3089) (Supplemental Table S2) containing the mutation in ADR-1 dsRBD1(KKxxK-EAxxA). Genomic modifications were verified using Sanger sequencing and ADR-1 expression was verified using western blot.

Crosses were performed by putting 9–10 males and 1 hermaphrodite on mating plates and genotyping was performed for the F1 progeny and F2 progeny using primers mentioned in Supplemental Table S2. The specific crosses performed included: creation of HAH27 by crossing agIs6[dod-24::GFP] hermaphrodites to BB20 males, creation of HAH28 by crossing HAH27 hermaphrodites to HAH23 males, creation of HAH29 by crossing daf-2(m596) hermaphrodites to BB20 males, creation of HAH34 by crossing pqm-1(ok485) hermaphrodites to N2 males, creation of HAH35 (wildtype), HAH37 (adr-2(ok735)), HAH38 (pqm-1(ok485)), HAH39 (adr-2(ok735); pqm-1(ok485)) by crossing HAH34 hermaphrodites to BB20 males, creation of HAH40 (BB20 + blmEx19(Y75B8A.8 3’ UTR hairpin construct in prab3::GFP::unc-54 3’ UTR (pHH340))) and HAH41 (blmEx19(Y75B8A.8 3’ UTR hairpin construct in prab3::GFP::unc-54 3’ UTR (pHH340))) by crossing HAH24 hermaphrodites to N2 males, creation of HAH44 (BB20+ blmEx21(Y75B8A.8 3’ UTR hairpin construct in prab3::GFP::unc-54 3’ UTR (pHH340); prab3::pqm-1::unc-54 3’ UTR (pHH549))) by crossing HAH43 hermaphrodites to BB20 males, creation of HAH48 (wildtype), HAH49 (adr-2(ok735)) and HAH50 (ADR-1 dsRBD1 (KKxxK-EAxxA) CRISPR) by crossing HAH47 hermaphrodites to BB20 males

Cloning

To generate the neural ADR-2 complementary DNA (cDNA) expressing animals, the adr-2 cDNA sequence was amplified from a plasmid with primers HH1962 and HH1963 (Supplemental Table S2). This fragment was cut with restriction enzymes BglII and SalI and then cloned into plasmid pHH326 (prab3::GFP::unc-54 3’ UTR) to generate a plasmid expressing neural adr-2 cDNA. The sequence of the adr-2 region cloned into the plasmid was confirmed using Sanger sequencing.

To generate the neural ADR-1 expressing animals, the rab3 promoter sequence was amplified from plasmid pHH326 (prab3::GFP::unc-54 3’ UTR) with primers HH170 and HH2771 (Supplemental Table S2). This fragment was cloned into plasmid pHH99 (pBluescript 3X FLAG genomic adr-1) previously published [33], with restriction enzymes KpnI and PstI to generate plasmid pHH512 with prab3::3XFLAG ADR-1. The sequence of the rab3 promoter cloned into the plasmid was confirmed using Sanger sequencing.

The vector used to express pqm-1 in the nervous system, pWorm[Exp]-rab-3>cel_pqm-1, was constructed by VectorBuilder. The vector ID is VB221230–1025zug, which can be used to retrieve detailed vector information from vectorbuilder.com.

Gene set enrichment analysis

Gene set enrichment analysis was performed using the online WormCat software [38] for upregulated and downregulated genes identified from comparing the wildtype and adr-2(−) neural RNA-seq datasets.

Bleaching

Synchronized first larval stage (L1) animals were obtained by bleaching with 5M NaOH and Clorox solution. After bleach solution was added, animals were incubated on a shaker at 20°C for 7 minutes and then spun down to collect embryos. Collected embryos were washed with 1X M9 buffer (22.0 mM KH₂PO₄, 42.3 mM Na₂HPO₄, 85.6 mM NaCl, 1 mM MgSO₄) solution thrice. The animals were incubated overnight in 1X M9 solution at 20°C. Next day, hatched L1 worms were spun down and washed again with 1X M9 solution thrice.

Neural cell isolation and COPAS sorting

Neural cells were isolated from synchronized first larval stage worms as previously described [27] and filtered into sterile FACS tubes. Briefly, staining with near IR live/dead fixable dye (Invitrogen) of the isolated neural cells was done before performing FACS sorting. The BD FACSAria II sorter was used to separate the GFP+ neural cells from the non-GFP cells, and FACSDiva 6.1.1 software was used to analyze the sort (IU Flow Cytometry Core Facility). Sorted neural cells were collected into conical tubes with TRIzol (Invitrogen), snap-frozen in liquid nitrogen and stored at −80°C. For sorting transgenic animals, the COPAS BioSelect instrument (IU Flow Cytometry Core Facilty) was used to isolate GFP+ animals based on Time of Flight (TOF) and Extinction (Ext). 250 transgenic GFP+ animals were sorted per strain and collected on unseeded 10 cm NGM plates.

Bioinformatics analysis for differential gene expression

N2 and adr-2(−) L1 neural datasets generated in [37] under accession number GSE1151916 were downloaded and analyzed. 75 bp single-end stranded RNA-sequencing reads were subjected to adaptor trimming and aligned to the C. elegans genome (WS275) using STAR (v2.7.8a) with the parameters: [runThreadN 8, outFilterMultimapNmax 1, outFilterScoreMinOverLread 0.66, outFilterMismatchNmax 10, outFilterMismatchNoverLmax: 0.3]. Indexing of the aligned bam files was performed using samtools (v1.3.1) and featureCounts (v2.0.1) was used to generate the raw read counts file. DESeq2 library (v1.26.0) on R studio [71] was used to process the raw read counts and generate the counts.csv file used for differential gene expression analysis.

RNA isolation and quantitative real-time PCR (qPCR)

RNA extraction was performed using TRIzol (Invitrogen) reagent and DNA contamination was removed by treatment with TURBO DNase (Ambion) followed by the RNeasy Extraction kit (Qiagen) and stored at – 80°C. Concentrations of the RNA samples and presence of any contamination with organic and protein components was determined using a Nanodrop (Fisher Scientific). For qPCR experiments using RNA from L1 animals, 2 ug of DNase-treated RNA was reverse transcribed into cDNA using Superscript III (Invitrogen) with random hexamers (Fisher Scientific) and oligo dT (Fisher Scientific) primers. For qPCR experiments using RNA from neural cells or COPAS sorted transgenic animals, the whole 12 ul was reverse transcribed into cDNA. Following reverse transcription of RNA from L1 animals, 20 μL of water was added to the cDNA. For RNA immunoprecipitation experiments, 200 ng RNA for inputs and the whole 12 ul of IP samples was reverse transcribed into cDNA and no water was added to the cDNA. Gene expression was determined using SybrFast Master Mix water and gene-specific primers (Supplemental Table S2) on a Thermofisher Quantstudio 3 instrument. The primers designed for qPCR (Supplemental Table S2) spanned an exon-exon junction to prevent detection of genomic DNA in the samples. Melting curves were generated for all primer pairs used to ensure high quality of qPCR products. For each gene analyzed, a standard curve of eight to ten samples of ten-fold serial dilutions of the amplified product were used to generate a standard curve of cycle threshold versus the relative concentration of amplified product. Standard curves were plotted on a logarithmic scale in relation to concentration and fit with a linear line. Fit (r²) values were around 0.99 and at least 7 data points fell within the standard curve. Each cDNA measurement was performed in three technical replicates, and each experiment was performed in three biological replicates.

Fluorescence microscopy

Synchronized L1 animals were anesthetized on agarose pads containing sodium azide followed by a coverslip. Images were taken using the Leica SP8 Scanning Confocal Microscope (IU Light Microscopy Imaging Core) and the 10x objective. Each image was taken with the intestinal cells and head ganglia region in focus. For each trial, exposure time was calibrated to minimize the number of saturated pixels for that set of animals. ImageJ bundled with Java 8 software was used to quantify the total intensity of fluorescence per worm as measured by intensity of each pixel in the selected area of a frame (i.e. the worm).

RNA immunoprecipitation

Synchronized L1 worms were washed with IP buffer (50 mM HEPES [pH 7.4], 70 mM K-Acetate, 5 mM Mg-Acetate, 0.05% NP-40 and 10% glycerol) containing a mini EDTA-free cOmplete protease inhibitor tablet (Roche) and UV crosslinked (3 J/cm2) using the Spectrolinker (Spectronics). The worms were then frozen into pellets using liquid nitrogen and stored at −80°C. The frozen worm pellets were ground on dry ice with a cold mortar and pestle, and the cell lysate was centrifuged at maximum speed for 10 minutes to remove cellular debris. Protein concentration was measured using Bradford reagent (Sigma) and the entire L1 lysates were added to 25 ul magnetic Protein-G beads (Invitrogen) for preclearing. After incubation for 1 hour on a rotator at 4°C, 500 ug of the L1 lysate from the supernatant was added to 25 ul anti-FLAG magnetic beads (Sigma). After incubation for 1 hour on a rotator at 4°C, protein-bound beads were washed with wash buffer thrice (0.5 M NaCl, 160 mM Tris-HCl [pH 7.5], 0.1% NP-40, 0.25% Triton X-100) containing a mini EDTA-free cOmplete protease inhibitor tablet (Roche). A portion of the IP (2/5) was stored in 2X SDS loading buffer and used for immunoblotting. The remaining beads were incubated with 1 ul RNasin (Fisher) and 0.5 μl of 20 mg/ml Proteinase K (NEB) at 42°C and 1200 rpm for 15 minutes in a thermomixer. RNA was isolated and qPCR was performed as described above.

Immunoblotting

Protein lysates were prepared as mentioned above. Protein lysates were boiled for five minutes and 100 ug of IP lysates and 10 ug of input lysates were subjected to SDS-PAGE. The same immunoblot was treated with antibodies against FLAG (Sigma, M8823) and β-Actin (Cell Signaling, 8457S) after cutting the blot. Protein bands were visualized using enhanced chemiluminescent detection SuperSignal West Femto Maximum Sensitivity Substrate (Fisher). The immunoblot images without saturation were acquired using Image Lab software (version 6.1.0 build 7) in the BIO-RAD ChemiDoc MP imaging system.

Cobalt chloride exposure assays

These assays were performed as previously described [49] with slight modifications. Briefly, cobalt(II) chloride hexahydrate (CoCl₂) powder (Sigma-Aldrich) was used to make a 0.1 M stock solution with distilled water and then working stocks of varying concentrations were prepared in 85 mM NaCl. Bleaching according to above mentioned conditions was performed to obtain L1-arrested animals. After washing the hatched animals twice with 85 mM NaCl, 5000 L1 animals per strain were exposed to CoCl₂ of varying concentrations in a total volume of 500 ul for two hours. After exposure, the worms were washed twice with 85 mM NaCl to remove any residual CoCl₂ and other debris. After washing, 40 worms per strain were transferred to 35 mm NGM plates seeded with OP50 and incubated at 20°C for twenty-four hours. After the incubation, alive and dead worms were counted in triplicates for each strain and three biological replicates for both the transgenic and non-transgenic sets of worms. Genotypes were blinded before the bleaching to reduce bias.