Neuronal transcriptome analyses reveal novel neuropeptide modulators of excitation and inhibition imbalance in C. elegans

Neuropeptides are secreted molecules that have conserved roles modulating many processes, including mood, reproduction, and feeding. Dysregulation of neuropeptide signaling is also implicated in neurological disorders such as epilepsy. However, much is unknown about the mechanisms regulating specific neuropeptides to mediate behavior. Here, we report that the expression levels of dozens of neuropeptides are up-regulated in response to circuit activity imbalance in C. elegans. acr-2 encodes a homolog of human nicotinic receptors, and functions in the cholinergic motoneurons. A hyperactive mutation, acr-2(gf), causes an activity imbalance in the motor circuit. We performed cell-type specific transcriptomic analysis and identified genes differentially expressed in acr-2(gf), compared to wild type. The most over-represented class of genes are neuropeptides, with insulin-like-peptides (ILPs) the most affected. Moreover, up-regulation of neuropeptides occurs in motoneurons, as well as sensory neurons. In particular, the induced expression of the ILP ins-29 occurs in the BAG neurons, which were previously shown to function in gas-sensing. We also show that this up-regulation of ins-29 in acr-2(gf) animals is activity-dependent. Our genetic and molecular analyses support cooperative effects for ILPs and other neuropeptides in promoting motor circuit activity in the acr-2(gf) background. Together, this data reveals that a major transcriptional response to motor circuit dysregulation is in up-regulation of multiple neuropeptides, and suggests that BAG sensory neurons can respond to intrinsic activity states to feedback on the motor circuit.


Introduction
Neural circuits are dynamic, changing their properties in response to experience. These changes are critical for maintaining circuit homeostasis and in processes like memory. Many factors such as c-Fos and BDNF, are activated early in response to increased neural activity and further regulate the expression of downstream genes [1]. These early-acting genes are also involved in many neurological diseases. For example, mutations in the activity-dependent transcriptional repressor gene Mecp2 are associated with Rett's syndrome [2]. Mecp2 is competent [18]. Although BAG neurons interact with the motor circuit to respond to environmental cues, our data indicates that intrinsic activity states modulate expression of genes in the sensory BAG neuron, which then feed-back on the motor circuit. Finally, we present functional evidence supporting that the concerted action of several neuropeptides underlies motor circuit hyperactivity.

Expression profiling of adult cholinergic neurons
We performed cell-type specific RNA-seq analyses, using the Pacr-2::gfp reporter juIs14 to isolate GFP expressing neuronal cells by FACS followed by RNA-seq (Materials and methods, S1 Table) [10,19,20,21]. Besides expression in the ventral cord cholinergic motoneurons (VA, VB, DA, and DB), GFP expressed from juIs14 is also present in several unidentified neurons in the head and tail [10]. RNA-seq data from isolated wild-type neurons was analyzed with the Cufflinks program to identify expressed genes (Materials and methods, S2 Table). The gene list for the neurons labeled by Pacr-2::gfp from wild type animals was compared with a recent study that profiled major tissue-types in adult C. elegans [22]. We found that almost all of the genes identified in our dataset (~95%) were also detected in a pan-neuronal analyses of expressed genes ( Fig 1A), but shared very little overlap with tissue-specific expression profiles identifying enriched transcripts in hypodermis and muscle ( Fig 1A). This comparison suggests our samples were relatively free of contamination from surrounding non-neuronal tissue. We also compared our gene list to previous analysis of a subset of cholinergic motoneurons (VA and DA) labeled with Punc-4::gfp [23]. Although the cell-types analyzed in these studies are not identical (A-type only vs. neurons expressing Pacr-2::gfp), and they were performed at different life stages (larval vs. young adult in this study) using different techniques (microarray vs. RNA-seq in this study), over half of the genes (~75%) from our dataset were also identified by microarray in larval type A motoneurons ( Fig 1B). These include core cholinergic genes such as unc-17/VAChT (Avg. FPKM = 484.1) and pan-neuronal genes such as unc-13 (Avg. FPKM = 90.9), required for synaptic vesicle priming [24,25]. Additionally, both of these datasets were relatively free of transcripts enriched in GABAergic motoneurons. Thus, this comparison shows the genes identified in our sample are highly enriched for those that are validated to be expressed and function in cholinergic neurons.

Differential expression analyses between wild-type and acr-2(gf) neurons
We are interested in the genes differentially expressed in response to altered neuronal activity. The transcriptome of neurons labeled by Pacr-2::gfp in adult acr-2(gf) mutants was compared to those from wild type animals for changes in gene expression using DESeq2. This analysis identified 234 genes as significantly mis-expressed in acr-2(gf) mutants (S2 Table) [26,27]. We found flp-18 to be significantly up-regulated in acr-2(gf), as predicated from our previous studies [16]. Analysis of the distribution of the fold change data using a histogram showed that a majority of the changes were up-regulation, at around 2-fold ( Fig 1C). GO-term analyses [28] indicated that genes involved in neuropeptide signaling were highly enriched in our gene list ( Fig 1D). In total, the expression of 21 neuropeptide genes were identified as significantly upregulated in acr-2(gf) neurons, representing approximately 18% of the estimated 113 neuropeptide genes identified in C. elegans (Fig 2A) [12]. Of these, several insulin-like peptides (ILP) were the most up-regulated (Fig 2A, S2 Table). In addition, none of these ILP genes were detected as expressed in wild type samples (S2 Table). Together, these data indicate that the major response to altered motor circuit activity at the transcriptional level is to increase neuropeptide gene expression.
To validate the RNA-seq analyses, we analyzed transcriptional reporters for the most upregulated neuropeptide gene of each class. flp-12 and nlp-1 were the most up-regulated neuropeptides of their respective classes. These genes have been implicated in context-dependent behaviors. flp-12 is involved in male locomotion, and nlp-1 functions in food-evoked turning behaviors [29,30]. 2kb upstream sequences of flp-12 and nlp-1 were used to generate GFP reporters, and both reporters showed expression in neurons in the head, similar to published studies of these neuropeptides (Fig 2B-2E) [31,32]. In acr-2(gf), the flp-12 reporter showed a different expression pattern than wild type, with ectopic expression in a pair of neurons close to the ganglion (Fig 2B and 2D). Similarly, the nlp-1 reporter was up-regulated and expressed in additional cells in acr-2(gf) animals compared to wild type (Fig 2C and 2E). Therefore, this analysis provides an explanation for the increased RNA levels detected in RNA-seq of acr-2 (gf), and also shows that increased locomotor excitation may cause mis-expression of these neuropeptide genes.  [22]. In contrast, almost all of the genes identified as expressed in cholinergic neurons labeled by Pacr-2::gfp are identified in pan-neuronal data-sets. Shown are Venn Diagrams with overlaps between the indicated data sets. The number in each region indicates the number of genes in that category. Hypodermis, muscle, and neuronal datasets are from Kaletsky et al. (2018). Statistics for significance of overlap were performed using a hypergeometric distribution with the phyper function in R. B. Genes identified in previous expression profiling of larval cholinergic motoneurons, including core neuronal and cholinergic markers, are identified by RNA-seq of Pacr-2::gfp expressing neurons [23]. The Venn Diagram displays that over half of genes identified by microarray as expressed in larval A-type motoneurons are also identified by RNA-seq in adult cholinergic neurons that express Pacr-2::gfp. The number in each region indicates the number of genes in that category. Larval A-type motor neuron expression data is from Von Stetina et al. (2007). Statistics for significance of overlap were performed using a hypergeometric distribution with the phyper function in R. C. Histogram of log 2 (Fold Change) for significantly different genes. Most genes that were different in acr-2(gf) compared to wild type were upregulated at around 2-fold. D. GO-term analyses of genes significantly different in acr-2(gf) neurons compared to wild type (see Materials and methods). Genes involved in neuropeptide and G-protein signaling (which included neuropeptide genes) were the most affected. https://doi.org/10.1371/journal.pone.0233991.g001

acr-2(gf) increases expression of ins-29 and acr-2 in BAG sensory neurons
Among all up-regulated neuropeptides, ins-25 and ins-29 showed the most dramatic increase (Fig 2A). As the expression pattern and function of these ILPs is unknown, we investigated their expression patterns in further detail. ins-25 and ins-29 are in an operon on Chromosome I, with ins-29 being upstream of ins-25, separated by 697bp intergenic sequence (Materials and methods, Fig 3A). We used 2kb upstream sequences of ins-29 to drive GFP expression. In wild type animals, the Pins-29 transcription reporter was sometimes weakly expressed in two neurons in the head, but often expressed in just one neuron or no GFP expression was detected ( Fig 3B). However, in acr-2(gf) animals, consistent strong expression of GFP was observed in the same two head neurons (Fig 3C). These neurons are likely sensory, as they extend dendrites out to the nose of the animal.
The dendrite morphology and cell position of cells expressing ins-29 suggested that they were likely to be the BAG neurons, which are known for their role in gas-sensing, particularly CO 2 avoidance ( Fig 3C) [33,34,35]. To confirm that the cells expressing ins-29 were indeed BAG neurons, animals were generated co-expressing the Pins-29::ins-29::SL2::mKate reporter and an integrated GFP marker for BAG [Pgcy-33::gfp] (Fig 4A-4F). We observed that in acr-2 (gf) animals, where the ins-29 reporter expression was consistently detected, the expression of mKate2 completely overlapped with GFP ( Fig 4D-4F). For those wild type animals that expressed the ins-29 reporter, mKate2 also overlapped with the BAG GFP marker (Fig 4A-4C).
The transcription factor ets-5 is required for BAG development and function [18]. To determine if ets-5 is involved in the expression of ins-29 in acr-2(gf), an ets-5(0) acr-2(gf) double mutant strain with Pins-29::gfp was generated. Penetrance of expression was quantified by scoring for presence of GFP expression in zero, one, or both BAG neurons ( Fig 4G). Very few wild-type animals expressed the transgene in both BAG neurons (31%), while most acr-2(gf) animals (99%) did. We found a marked decrease in the expression of the Pins-29::gfp reporter in ets-5(0) acr-2(gf) mutants, similar to wild type. Altogether, these data indicate that in acr-2 (gf) animals, ins-29 is up-regulated in sensory BAG neurons in an ets-5-dependent manner.

ins-29 expression in BAG neurons is regulated by systemic acr-2(gf) activity
To address whether the up-regulation of Pins-29::gfp in response to acr-2(gf) is dependent on neuronal activity, we next tested if reduction in motor circuit hyperactivity could restore expression of ins-29 to wild type. The TRPM channel gtl-2 is expressed in non-neuronal tissues, such as the hypodermis, and regulates motor circuit hyperexcitation via systemic ion homeostasis [36]. Null mutation of gtl-2 almost completely suppresses acr-2(gf) convulsion and locomotion phenotypes. The expression of Pins-29::gfp was analyzed in both gtl-2(0) and

Fig 3. Expression of ILP genes and acr-2 is up-regulated in two head neurons in acr-2(gf) animals.
A. Shown is the gene structure of ins-29 and ins-25 loci. Also shown are diagrams of the ins-29 transcription reporters and expression constructs used in this study (Listed in S1 Table). Two ins-29 expression constructs were made, one that contained SL2:: mKate2 sequence to monitor transgene expression, the other has mKate2 in-frame fused to INS-29, without the SL2 trans-splice sequence. B-C. Pins-29::gfp is expressed in two head neurons in acr-2(gf).  is similar to wild type in acr-2(gf). (F) Complete overlap is observed between the two reporters in acr-2(gf). GFP-only signal is from the Pmec-4::gfp co-injection marker. Scale bar:10μm. In all images, animals are rolled in 10% agarose to facilitate visualization of BAG neuron pairs. Beading observed in some images (i.e. the Pgcy-33::gfp and Pmec-4::gfp in B) is due to the mounting procedure. G. Almost all acr-2(gf) animals express Pins-29::gfp(juEx7742) in both BAG neurons, however mutation of ets-5 causes animals to exhibit more similar Pins-29::gfp expression patterns as wild type. To assay Pins-29::gfp expression, animals were scored for detectable GFP in 0, 1, or 2 BAG neurons. N for each genotype is labeled in the bar. H. Loss-of-function of gtl-2(0); acr-2(gf) backgrounds. Animals were scored for GFP expression in zero, one, or both neurons (Fig 4H). gtl-2(0) animals alone resembled wild type. The expression pattern of Pins-29::gfp in gtl-2(0); acr-2(gf) animals also resembled wild type. This result supports the idea that the changes in neuropeptide expression are likely due to systemic motor activity.
In light of the finding that acr-2 expression was up-regulated in BAG neurons (Fig 3D-3I), we wished to discriminate whether ins-29 expression was regulated by acr-2(gf) activity in BAG neurons or the motor circuit. We expressed acr-2(gf) cDNA in a subset of motor neurons or in BAG neurons and assessed Pins-29::gfp expression. As previously reported, expression of acr-2(gf) using the unc-129 promoter (which drives expression in the ventral cord DA and DB motoneurons) caused animals to display an Unc phenotype as well as convulsions, albeit at a lower frequency than acr-2(gf) mutants (Fig 4I) [37]. In contrast, BAG neuron-specific expression of acr-2(gf) driven by the gcy-33 promoter did not cause convulsion. We quantified penetrance of Pins-29::gfp reporter expression in animals expressing acr-2(gf) in cholinergic motoneurons or BAG neurons and found that almost all animals expressing acr-2(gf) in this small set of cholinergic motoneurons showed expression of the Pins-29::gfp reporter in both BAG neurons, similar to acr-2(gf) mutant animals ( Fig 4J). However, animals with the Pgcy-33::acr-2(gf) transgenes had much lower penetrance of Pins-29 reporter expression. Taken together, this data support that increasing motor circuit activity results in a cell non-autonomous effect on ins-29 expression in BAG neurons.
Neuropeptides are typically secreted from the cell in which they are produced and can act at a distance (endocrine signaling) as well as on post-synaptic or nearby cells (paracrine signaling). We next sought to address whether INS-29 can be secreted, by generating a transgene expressing INS-29::mKate2 fusion protein. Secreted molecules are present in the coelom (or interstitial fluid), which are generally taken up by the coelomocytes. We found mKate2 in the vacuoles of posterior coelomocytes in both wild type and acr-2(gf) animals (S1 Fig). This observation shows that INS-29 is secreted and may therefore act on multiple cells and tissues within the animal.
We further asked whether these neuropeptide genes affected synaptic transmission in the locomotor circuit using pharmacological assays. Aldicarb is an acetylcholinesterase inhibitor which leads to buildup of acetylcholine at the synaptic cleft and eventual paralysis in wild type animals [38]. Mutants that increase or decrease cholinergic activity display increased or decreased sensitivity to aldicarb. acr-2(gf) animals are more sensitive to aldicarb, consistent with their hyperactivity as described by previous electrophysiology and pharmacology analysis [10]. ins-29(0) ins-25(0); flp-12(0) acr-2(gf) animals were similar to acr-2(gf) alone on aldicarb, indicating that overall neurotransmission was not strongly affected by these mutations (Fig 6A, S3 Table). ins-29(0) ins-25(0); flp-12(0) mutants were also not significantly different from wild type animals on aldicarb (S4 Table). Levamisole is an agonist for the post-synaptic cholinergic receptors on muscle [39,40]. Resistance or sensitivity to levamisole can represent changes in muscle responsiveness. In these experiments, we found that ins-29(0) ins-25(0); flp-12(0) acr-2(gf) animals were less sensitive to levamisole than acr-2(gf) alone (Fig 6B, S5 Table). However, ins-29(0) ins-25(0); flp-12(0) triple mutants showed similar levamisole sensitivity to wild type (S6 Table). This result suggests that these neuropeptides may function post-synaptically to affect motor circuit function, and their function is only observed in the acr-2(gf) background, consistent with these peptides being differentially expressed in response to acr-2(gf) hyperactivity.
We also asked if transgenic over-expression of ins-29 could rescue the levamisole phenotype of ins-29(0) ins-25(0); flp-12(0) acr-2(gf) animals. Animals expressing the Pins-29::ins-29::SL2:: mKate2 transgene in wild-type or mutant backgrounds were assessed for levamisole sensitivity. We observed that over-expression of ins-29 in wild type exhibited a levamisole resistance phenotype, indicating that ins-29 can inhibit post-synaptic activity in the absence of acr-2(gf) when over-expressed (Fig 6C, S7 Table). On the other hand, and in contrast to the convulsion analyses, ins-29 over-expression did not affect the levamisole hypersensitivity of acr-2(gf) animals, and acr-2(gf) animals with or without the transgene had similar levamisole sensitivity profiles. ins-29 over-expression, however, significantly rescued the suppression of acr-2(gf) by ins-29(0) ins-25(0); flp-12(0) mutations. These results of ins-29 over-expression indicate that the function of these neuropeptides in motor circuit regulation may be context-dependent, with different activities depending on the genetic background and state of the circuit.

Discussion
Using neuronal-type specific RNA-seq, we identified over 200 genes whose expression levels are altered in response to hyperactivity in the motor circuit of C. elegans. Genes encoding neuropeptides were over-represented in this list, and we validated that fluorescent reporters for ins-29, flp-12, and nlp-1 were over-and/or ectopically-expressed in acr-2(gf) animals compared to wild type. These data support the conclusion that the major transcriptional response to A. Aldicarb sensitivity shown as percentage of animals not paralyzed after 1hour on the drug. No significant difference was observed (Two-way ANOVA followed by Dunnett's test, compared to acr-2(gf) alone.). Each data point represents one trial and 10 animals per genotype were scored per trial. Data is also shown in S3 Table. B. Levamisole sensitivity of neuropeptide mutants in the acr-2(gf) background at 15min. Deletion of insulin-like peptides with flp-12(0) significantly reduces levamisole sensitivity compared to acr-2 (gf) at this timepoint. Each data point represents one trial and 10 animals per genotype were scored per trial ( � P<0.05, Two-way ANOVA followed by Dunnett's test compared to acr-2(gf) single mutant.). Data is also shown in S5 Table. C. Effect of ins-29 over-expression on levamisole sensitivity. Shown are a subset of mutant strains analyzed in Fig 6B also carrying the ins-29 over-expression transgene (juEx7966). Over-expression of ins-29 can restore acr-2(gf) levels of levamisole sensitivity to that of ins-29(0) ins-25(0); acr-2(gf) strains, but also inhibits levamisole sensitivity in wild-type circuits. Each data point represents one trial and 10 animals per genotype were scored per trial ( ��� P<0.001, � P<0.05. Two-way ANOVA followed by Bonferroni test.). Data is also shown in S7 Table. https://doi.org/10.1371/journal.pone.0233991.g006

PLOS ONE
motor circuit hyperactivity in C. elegans is by altering neuropeptide gene expression. Bioinformatic analyses of upstream sequences did not identify common motifs in the promoter regions of theses neuropeptides, suggesting that multiple transcriptional pathways may be involved in these changes. Indeed, the fluorescent reporters analyzed here showed diverse expression patterns as well as varied changes in expression, from over-expression in the same cells as wild type, to ectopic expression patterns.
ILPs were the most up-regulated genes identified by RNA-seq in acr-2(gf). We have previously shown that loss of function mutations in unc-31/CAPS, which is required for the release of the majority of neuropeptides, reduce convulsion rate of acr-2(gf) mutants, indicating that the release of unidentified neuropeptides modulates motor circuit hyperactivity [16,17]. Here, we show that loss of the ILP genes ins-29 and ins-25, along with flp-12, can reduce convulsion of acr-2(gf) animals. This effect is modulatory, as mutation of these neuropeptides partially suppresses, but does not block, convulsion. The effects of these peptides are in contrast to our previous studies on a different neuropeptide flp-18, which is also up-regulated in cholinergic motoneurons of acr-2(gf) (this study), yet loss of flp-18 function, together with loss of flp-1, increased convulsion of acr-2(gf) animals [16]. These data are consistent with many studies showing that neuropeptide modulation involves combinatorial and redundant action of different peptides that can have varied effects on epileptic states [5]. Paradoxically, we further find that overexpression of wild type ins-29 or flp-12 in acr-2(gf) reduced convulsion. We speculate that the observed effects from overexpressing each peptide reflect promiscuity of receptors for the neuropeptides. Indeed, the variety of effects observed by ins-29 over-expression would be consistent with it affecting multiple pathways. Future studies will involve defining specific receptors for each peptide to clarify their function.
The expression and function of ins-29 has not been previously characterized. Examination of ins-29 transcriptional reporters showed expression in the gas-sensing BAG neurons, with increased intensity and more penetrant expression in both BAG neurons detected in acr-2(gf) compared to wild type. Furthermore, co-labeling also showed up-regulation of acr-2 itself in these neurons. We show that the regulation of ins-29 expression in BAG neurons is activitydependent. For example, genetic suppression of acr-2(gf) by null mutation of gtl-2 restored Pins-29::gfp expression to wild type patterns. BAG neurons are necessary for C. elegans to respond to changes in environmental CO 2 [33]. The observation that ins-29 is expressed in the BAG sensory neurons, rather than motoneurons or pre-motor interneurons, was surprising. For example, whereas flp-18 is expressed from the cholinergic motoneurons to affect acr-2(gf) activity and motor function, ins-29 expression is increased in sensory neurons in the head and INS-29 is secreted. Together, these results indicate that genes involved in cholinergic neurotransmission including the acr-2 cholinergic receptor subunit gene, as well as the insulin-like peptide gene ins-29 are up-regulated in BAG neurons due to motor circuit hyperactivity in the acr-2(gf) background. Neuropeptide signaling from BAG has been shown previously to interact with another cholinergic circuit, the egg-laying circuit [41]. flp-17 and flp-10 secreted from BAG act in parallel with cholinergic signaling to inhibit egg-laying. It is proposed that signaling from BAG integrates favorable environmental signals with the egg-laying circuit.
The BAG-specific transcription factor ets-5 also regulates Pins-29::gfp expression. Analysis of the ins-29 promoter did not reveal a clear ets-5 biding site, and it is possible this effect in indirect. ets-5-dependent pathways, therefore, are necessary for the increased expression of Pins-29::gfp in response to acr-2(gf). This result also indicates a function for ets-5 beyond development and maintenance of BAG neuron identity, but in modifying transcription in the BAG neurons under different physiological conditions in mature animals.
In humans, seizure is a common co-morbidity of diabetes. In vitro analysis has also found, for example, that IGF signaling can be neuroprotective in response to injury, but can also promote epileptogenesis [42]. Insulin peptide signaling may also play a role in Alzheimer's Disease (AD) [6]. Although the role for insulin signaling in the brain is still unclear, levels of insulin and insulin receptors are markedly decreased in the brains of AD patients. Our data show that in C. elegans, ILP signaling is altered by aberrant cholinergic activity to modulate circuit function.

Sample preparation for RNA-seq
Samples were prepared for dissociation and FACS essentially as described [19,44]. Animals were synchronized by hypochlorite treatment. Eggs were isolated from wild type N2, CZ631, and CZ5808 gravid adults and allowed to hatch overnight and arrest at the L1-stage. Synchronized L1s were plated onto 15cm NGM plates seeded with OP50 bacteria. Approximately 8--10,000 L1s were plated to 15-20 plates for each strain. As N2 cells are only needed to control for the FACS sort, only 5 plates were prepared. These plates were incubated at 20˚C for 72hr to reach adulthood prior to collection.
The entire process of cell preparation and sort was completed in a single day. Animals were washed off plates and spun down and then washed in M9 media to remove bacteria. Typically, pellets for each strain would be~500μl in volume, and these would be split into two tubes. 750μl of lysis buffer (200μM DTT, 0.25% SDS, 20mM HEPES pH8.0, 3% sucrose) was added to each tube and samples were incubated for approximately 6-7 minutes. Worms were then rapidly washed in M9 five times. Next, 500μl of 20mg/ml freshly made pronase solution was added. Samples were incubated in pronase approximately 20 minutes. Every 2-3 minutes, each sample was disrupted by pipetting with a P200, and samples were monitored for dissociation with a dissection microscope. When large worm chunks were no longer visible under the dissection microscope, the reaction was stopped by adding 250μl ice cold PBS-FBS solution (1X PBS solution with 2% Fetal Bovine Serum). Samples were then centrifuged in the cold at top speed in a microcentrifuge for 10 minutes, and pelleted cells were resuspended in 500μl FBS. Cells were then syringe-filtered (5μm pore). Samples were spun again in the cold and resuspended in ½ the starting volume of FBS. 80,000-100,000 GFP+ cells were collected for each sample at the UCSD Flow Cytometric Core in Moore's Cancer Center. Cells were sorted directly into Trizol LS and stored at -80˚C until preparation. RNA was isolated using Qiagen RNAeasy kit.

RNA-seq analyses
RNA library preparation and sequencing were performed at the Institute for Genomic Medicine at UCSD. RNA libraries were prepared with Illumina TruSeq kit. Libraries were sequenced on an Illumina HiSeq4000. Data analyses were performed using the Galaxy platform [27]. We obtained 50-70 million single-end reads/sample. Reads for each sample were mapped and aligned using TopHat [45]. Identification of expressed genes in wild type was determined using Cufflinks. "Expressed" genes were selected by filtering for genes with an FPKM >10 in both replicates. This analysis produced 1,812 transcripts (S2 Table). Using the BioVENN site, our list of "expressed genes" from wild type neurons was compared with those identified as enriched in adult epidermis and muscle and expressed in neurons by RNA-seq, as well as larval A-type motoneurons by microarray [22,23,46]. For differential expression analyses between wild type and acr-2(gf) neurons, BAM files were loaded into the HTSEQ program and then analyzed by DESeq2 for differential expression analyses (S2 Table) [26,47]. GO-term enrichment analysis was performed using the GOrilla algorithm for enrichment of Biological Process terms [28]. Data analyses were performed in Microsoft Excel, R, and Graphpad Prism. Sequencing datasets have been deposited in the Gene Expression Omnibus (Accession GSE139212).

Convulsion and pharmacological analyses
All behavior observations reported here were made on mutations that were outcrossed with N2 for at least 4 times. Convulsions were defined as simultaneous contraction of the body wall muscles producing a rapid, concerted shortening in body length. The convulsion frequency for day-1 adult animals was calculated during a 90-second period of observation. For levamisole sensitivity, ten day-1 adult animals were transferred to fresh plates containing 1 mM levamisole. Animals were scored for paralysis ever 15 minutes for 1 hour. Sensitivity to 500 μM or 1mM aldicarb was assessed by transferring ten day-1 adults to fresh aldicarb plates and by monitoring worms for paralysis every 30 minutes by gently touching the animal with a platinum wire. Drug sensitivity was quantified from at least three independent experiments.

Imaging and microscopy
Images of fluorescent reporter lines were taken on a Zeiss LSM 700 or 800 confocal microscope. Animals were mounted in thick agarose (10%) and rolled with the ventral side up for consistent imaging. Hyperstacks were processed in ImageJ. (For images in Fig 4, Pgcy-33::gfp strains, gains were set at 500V for mKate2 and 500V for GFP. For images in Fig 3, of juIs14 strains, gains were set at 550V for mKate2 and 600V for GFP.). All images are taken using the 63X objective. Scoring of Pins-29::gfp expression in Fig 4 was performed using a Zeiss Axioplan 2 compound microscope at the 63X objective unless otherwise indicated. A neuron was scored as "expressed" if GFP signal was clearly visible in the cell body and dendrites through the eyepiece (with expression in acr-2(gf) as a baseline comparison). Images of INS-29::mKate2 in coelomocytes was were captured using a Zeiss AxioImager compound microscope with the 100X objective.
C. elegans transgenic multi-copy arrays were generated using standard protocols [See S1 Table for list of transgenic strains made in this study] [49]. Reporter DNA was typically injected at 25ng/μl concentration. The flp-12 over-expression plasmid was injected at 5ng/μl, as injection at 25ng/μl failed to yield transgenics.