A single neuron in C. elegans orchestrates multiple motor outputs through parallel modes of transmission

Abstract

Animals generate a wide range of highly coordinated motor outputs, which allows them to execute purposeful behaviors. Individual neuron classes in the circuits that generate behavior have a remarkable capacity for flexibility, as they exhibit multiple axonal projections, transmitter systems, and modes of neural activity. How these multi-functional properties of neurons enable the generation of highly coordinated behaviors remains unknown. Here we show that the HSN neuron in C. elegans evokes multiple motor programs over different timescales to enable a suite of behavioral changes during egg-laying. Using HSN activity perturbations and in vivo calcium imaging, we show that HSN acutely increases egg-laying and locomotion while also biasing the animals towards low-speed dwelling behavior over longer timescales. The acute effects of HSN on egg-laying and high-speed locomotion are mediated by separate sets of HSN transmitters and different HSN axonal projections. The long-lasting effects on dwelling are mediated by HSN release of serotonin that is taken up and re-released by NSM, another serotonergic neuron class that directly evokes dwelling. Our results show how the multi-functional properties of a single neuron allow it to induce a coordinated suite of behaviors and also reveal for the first time that neurons can borrow serotonin from one another to control behavior.

Figure 1.. The HSN neuron evokes egg-laying, acute speeding, and long-term slowing

(A) Cartoon illustrating the anatomy of the HSN neuron (red), relative to the overall body organization of C. elegans. Pharyngeal and vulval muscles are shown in light gray, and eggs are shown in dark gray. (B) Cartoon depicting the intersectional promoter strategy used to obtain HSN-specific expression of CoChR. The cat-4 promoter drives expression of an inverted CoChR-sl2-GFP expression cassette; due to its inverted orientation, it is not expressed. The egl-6 promoter drives expression of Cre recombinase, which acts on the lox sites (black and white triangles) to invert the CoChR-sl2-GFP, permitting expression in cells that express both cat-4 and egl-6 (only HSN). See Fig. S1A for a fluorescent image of the resulting strain. (C) Behavioral responses to HSN::CoChR activation. Data are shown as an event-triggered average depicting the mean change in velocity (top) and egg-laying (bottom) during light illumination of HSN::CoChR animals. Blue boxes indicate the light illumination period in all HSN::CoChR experiments. There is an increase in reversals at light offset that is apparent in these data as well. We chose not to focus on this, because it was unclear if this reflected a rebound effect not directly under HSN control. Data are shown as means ± standard error of the mean (SEM). Statistics for velocity was performed on the change in velocity: difference of mean velocity during-stimulation period and velocity pre-stimulation (2-minute average baseline). Statistics for egg-laying rate was performed with the mean egg-laying rate during the stimulation period. Due to the extremely transient nature of the egg-laying events, the mean for the egg-laying rate over time is a jagged line. N = 270 stimulation events across 19 animals for the ATR group; N = 42 stimulation events across 6 animals for the no ATR control. ****p<0.0001, unpaired t-test. (D) Behavioral responses to HSN::CoChR activation in egl-1(n487gf) mutants, displayed as in (C). Gray shows WT behavior (same data as in C) for reference. N = 144 stimulation events across 10 animals. ****p<0.0001, unpaired t-test. (E) Behavioral responses to HSN::CoChR activation in animals treated with FUDR, which abolishes egg production. Data are displayed as in (C). Gray shows WT behavior (same data as in C) for reference. N = 140 stimulation events across 10 animals. *p<0.05, ****p<0.0001, unpaired t-test. (F) Effect of HSN:: CoChR activation on speed in animals travelling at high baseline speeds, due to recent transfer to food. The mechanical agitation associated with transfer is sufficient to transiently arouse animals. The pair of two consecutive blue light stimuli were intended to mimic a burst of two HSN activity peaks (see Fig. 2), though the effects on locomotion are already visible after the first stimulus. Data are shown as speed surrounding the stimulation events. Lines depict mean speed and error shading is SEM. N= 195 animals for the ATR group, and 159 animals for the no-ATR control. Statistics was performed on the change of speed: the difference between post-stimulation (one-minute average) and pre-stimulation (one-minute average baseline). ****p<0.0001, unpaired t-test. (G) Event-triggered average showing animal speed surrounding native, spontaneous egg-laying events in wild-type (black) and egl-1(n487gf) (red) animals. Lines depict mean velocity and error shading is SEM. N= 389 egg-laying events across 19 animals for wild-type and 169 egg-laying events across 20 animals for egl-1(n487gf) animals. ****p<0.0001, unpaired t-test. (H) Top: Cartoon depicting the behavioral assay used to quantify exploratory behavior on food This assay is commonly used as a metric of dwelling versus roaming behavior (Flavell et al., 2013). Animals are placed on NGM agar plates seeded with E. coli OP50 bacterial food as L4 animals. The next day, the tracks of the animal’s movement, which are visible in the bacterial lawn, are scored based on how many squares of a superimposed grid they traversed. Bottom: Exploratory behavior of animals of the indicated genotypes, using the assay depicted. Dots are individual animals, bars show means, and error bars indicate SEM. N= 20 animals for each genotype **p<0.01, unpaired t-test.

Figure 2.. HSN activity is acutely correlated with egg-laying and high-speed locomotion

(A) Example dataset from one animal showing HSN GCaMP signal (top) and animal speed (bottom) over a 20-minute recording of HSN calcium in freely-moving animals. Red arrows indicate egg-laying events. Note that HSN activity occurs in discrete peaks and that egg-laying coincides with some of these peaks. (B) Event-triggered average showing average HSN GCaMP signal surrounding egg-laying events. Lines indicate mean and error shading is SEM. N= 16 egg-laying events. **p<0.01, empirical p-value, comparing to random distribution in Fig. S2A. (C) Left: Event-triggered average showing average speed (black) during HSN calcium peaks. HSN calcium is also shown (green). Right: the same datasets, but increased zoom along the x-axis and green here depicts the derivative of HSN GCaMP. Note the precise time alignment of animal speed and derivative of HSN calcium. Lines and error shading are means and SEM, respectively. n= 104 peaks across 15 animals. **p<0.01, increase in speed during HSN peaks, empirical p-value, comparing to random distribution in Fig. S2B. (D) Number of HSN peaks in the minute preceding HSN peaks that either result in egg-laying or not. Dots are individual HSN peaks. *p<0.05, unpaired t-test. (E) Event-triggered averages displayed as in (C), except splitting data based on how many HSN calcium peaks occurred in the minute preceding the HSN calcium peak being examined. N=22-47 calcium peaks per plot. **p<0.01, increase in speed during HSN peaks, empirical p-value, comparing to appropriate random distributions (as in (C)). (F) Example calcium traces of RIB and HSN from a joint calcium recording of both neurons in immobilized animals. (G) Fraction of the total HSN calcium peaks that occurred while RIB was high (i.e. network was in a ‘forward’ state) or low (i.e. network was in a ‘reverse’ state). Dots are individual animals. Bars show means and error bars are SEM. N= 8 animals. *p<0.05, paired t-test.

Figure 3.. HSN evokes acute speeding through its neuropeptidergic outputs

(A-B) Event-triggered averages showing average changes in velocity (top) and egg-laying (bottom) in response to HSN::CoChR activation via light illumination. Gray provides data from WT animals as a control and reference. Mutant alleles were tph-1(mg280) and nlp-3(n4897). Lines show means and error shading shows SEM. N= 140 stimulation events across 7 animals for tph-1 and 180 stimulation events across 11 animals for tph-1;nlp-3 animals. **p<0.01, ****p<0.0001, unpaired t-test. (C) Cartoon illustrating the CRISPR/Cas9-generated conditional knockout allele of unc-17, which encodes the vesicular acetylcholine transporter (VAChT) and is required for cholinergic transmission. (D) Event-triggered averages showing average changes in velocity and egg-laying in response to HSN::CoChR activation via light illumination. HSN-specific unc-17 knockout refers to loxP flanked unc-17(syb5779 syb5987) animals expressing pegl-6::Cre. Gray provides data from WT animals as a control and reference. Lines show means and error shading shows SEM. N= 120 stimulation events across 12 animals. **p<0.01, unpaired t-test. (E) Event-triggered averages showing average changes in velocity and egg-laying in response to HSN::CoChR activation via light illumination in egl-21 (n476) animals. . Gray provides data from WT animals as a control and reference. Lines show means and error shading shows SEM. N=205 stimulation events across 14 animals. ****p<0.0001, unpaired t-test. (F) Cartoon illustrating the strain used for cell-specific disruption of neuropeptide production. A single-copy, floxed egl-21 rescue was introduced into egl-21(n476) null animals. Red line in the top (Wild type) panel indicates the genomic location of the n476 deletion in the genome. Red star in the lower panel indicates the n476 mutation. (G) Event-triggered averages showing average changes in velocity and egg-laying in response to HSN::CoChR activation via light illumination. HSN-specific egl-21 knockout refers to egl-21(n476);kySi61[loxP-egl-21genomic-loxP] animals expressing pegl-6::Cre. Gray provides data from WT animals as a control and reference. Lines show means and error shading shows SEM. N=150 stimulation events across 15 animals. ****p<0.0001, unpaired t-test. (H-K) Event-triggered averages showing average changes in velocity and egg-laying in response to HSN::CoChR activation via light illumination. Gray provides data from WT animals as a control and reference. Mutant alleles were flp-2(gk1039), flp-28(flv11), and flp-26(gk3015), flp-2(flv15);flp-28(flv11) for double mutants. Because the genetic loci of flp-2 and flp-28 are very close, we used CRISPR to generate a 17bp insertion in flp-2 gene in the flp-28(flv11) background. Lines show means and error shading shows SEM. For each genotype, flp-2: N= 180 stimulation events across 18 animals,; flp-26: 160 stimulation events across 16 animals; flp-28: 255 stimulation events across 26 animals; flp-2;flp-28: 175 stimulation events across 19 animals, **p<0.01, ****p<0.0001, unpaired t-test. (L) Event-triggered averages showing average changes in velocity surrounding native egg-laying events. Mutant alleles were flp-2(flv15);flp-28(flv11),. Lines show means and error shading shows SEM. N= 167 egg laying events across 6 animals. ****p<0.0001, unpaired t-test.

Figure 4.. HSN-released serotonin is taken up by the NSM neuron, which re-releases it to evoke long-term slowing

(A) Exploratory behavior in animals of the indicated genotypes. Alleles used are tph-1(mg280) and nlp-3(n4897). Dots are individual animals; bars show mean; and error bars show SEM. N= 19–23 animals for each genotype. ****p<0.0001, one-way ANOVA with Dunnett's multiple comparison test. (B-D) Effect of HSN::CoChR activation on speed in animals travelling at high baseline speeds, due to recent transfer to food. Data are shown as mean speed surrounding light illumination for tph-1(mg280), nlp-3(n4897), and tph-1(mg280);nlp-3(n4897) animals. Lines depict mean speed and error shading is SEM. Gray provides data from WT animals as a control and reference.. N= 84 – 215 animals for each genotype. *p<0.05, ****p<0.0001, unpaired t-test. (E) Exploratory behavior in animals of the indicated genotypes. Promoters used for tph-1 rescue are tph-1 (3kb) promoter for native tph-1 expressing neurons and egl-6 promoter for HSN rescue. Dots are individual animals; bars show mean; and error bars show SEM. N= 11 – 30 animals for each genotype. **p<0.01, ****p<0.0001, one-way ANOVA with dunnett's multiple comparison test. (F) Average speed over time of animals with HSN chemogenetically silenced (red) or not (black). HSN::HisCl refers to pegl-6::HisCl. Note that animals were first exposed to histamine in this experiment at t=0 min of this video. Left: instantaneous speed of histamine-treat and no histamine control over 90-minute recordings. Right: mean speed over every 10 min interval for each condition. N= 202 and 208 animals for regular NGM and histamine NGM plates respectively. *p<0.05, ****p<0.0001. Bonferroni-corrected t-test. (G) Cartoon illustrating serotonin release and re-uptake by NSM and HSN neurons. Note that HSN does not appear to express mod-5, based on single-cell sequence data and GFP reporter expression (Taylor et al., 2021; Duerr et al., 1999). (H) Event-triggered averages depicting the average change in animal speed upon NSM::Chrimson stimulation with red light illumination. This plot compares the effect of NSM stimulation in serotonin-null tph-1 mutants to animals where serotonin production was restored specifically in HSN (tph-1; egl-6::tph-1 cDNA). Animals were starved for 3 hours before the assays, which makes the effects of serotonin on locomotion more pronounced. Lines show means; and error shading shows SEM. Statistics for velocity was performed on the average speed during light illumination period of indicated genotypes. N= 253 for tph-1 animals and 351 for tph-1;HSN::tph-1 cDNA animals. ****p<0.0001, unpaired t-test. (I) Event-triggered averages depicting the average change in animal speed upon NSM::Chrimson stimulation with red light illumination in wild-type and egl-1(n487gf) animals. Lines show means; and error shading shows SEM. N= 105 for wild-type animals and 252 for egl-1(n487gf) animals. ***p<0.0001, unpaired t-test. (J) Event-triggered averages depicting the average change in animal speed upon NSM::Chrimson stimulation with red light illumination for indicated genotypes. Alleles used are egl-1(n487gf) and mod-5(n822). Lines show means; and error shading shows SEM. N= 63-237 animals.

Figure 5.. The distinct outputs of HSN map onto different sub-cellular compartments

(A) Cartoon of HSN neuron: soma, vulval presynaptic region and distal axon are labeled with different shades of green that are consistent with the colors presented in (B) and (C). (B) Example traces showing simultaneous calcium imaging of three subcellular compartments of HSN in an immobilized animal. (C) Event-triggered averages showing average HSN GCaMP signals in each sub-cellular compartment, centered on the time points of HSN calcium peaks in the vulval presynaptic region. (D) Cartoon depicting split GFP strategy to label CAT-1/VMAT in a cell-specific fashion and CRISPR/Cas9-engineered cat-1 allele. (E) Representative images of the head (left) and mid-body (right) region of an animal with CAT-1 labeled via split GFP in HSN. Note that images were collected under identical imaging conditions and are displayed on same color scale. Soma and presynaptic sites are indicated with arrows. Areas with red asterisks are gut autofluorescence. (F) Left: Cartoon of HSN neuron (red) in the context of the animal’s body organization. Right: Illustration of the site of the HSN axotomy. Note that the HSN soma is still connected to the HSN vulval presynaptic region where HSN forms synapses onto VC neurons and vulva muscles, but not to the distal axon, as a consequence of the axotomy. (G) Egg-laying behavior of Mock and HSN-axotomized animals. Number of eggs laid over a 3hr recording is shown. ‘Mock’ animals were mounted for axotomy, but not cut with the laser, and otherwise handled identically to HSN-axotomized animals. Dots show individual animals; bars are means; and error bars are SEM. N= 15 animals for Mock and 14 animals for HSN axotomized group. (H) Event-triggered average showing average animal speed surrounding native, spontaneous egg-laying events. Data are shown for Mock and HSN-axotomized animals. Lines show means and error shading shows SEM. N= 161 egg laying events across 15 mock animals; N= 126 egg laying events across 14 HSN-axotomized animals. **p<0.01, unpaired t-test. (I) Baseline mean velocity of Mock, HSN-axotomized, wild-type and egl-1(n487gf) animals. Dots show individual animals; bars are means and error bars are SEM. N= 15 animals (Mock), 14 animals (HSN-axotomized), 18 animals (wild-type) and 14 animals (egl-1(n487g))). **p<0.01, One-way ANOVA with Sidak's multiple comparison test.

Figure 6.. Aversive sensory input to HSN is transmitted through humoral neuropeptide release, which inactivates HSN

(A) Cartoon depicting the behavioral assay used to measure the effect of high osmolarity on egg-laying behavior of C. elegans. The metric at the bottom is the y-axis in subsequent plots. (B) Percent eggs laid on high osmolarity, compared to control (150 mOsm) osmolarity, shown for the indicated genotypes. Alleles used are tax-2(p691) and ocr-2 (ak47). N= 3 – 4 plates with 10 animals on each plate. Bars are means and error bars are SEM. **p<0.01, ****p<0.0001, two-way ANOVA with dunnett's multiple comparison test. (C) Percent eggs laid on high osmolarity (300 mOsm), compared to control (150 mOsm) osmolarity, shown for the indicated genotypes. Alleles used are tax-2(p691)and tax-4 (p678). Dots are individual plates with 10 animals each; bars are means; and error bars are SEM. N= 4 – 9 plates for each genotype.. ****p<0.0001, one-way ANOVA with dunnett's multiple comparison test. (D) Percent eggs laid on high osmolarity (300 mOsm), compared to control (150 mOsm) osmolarity, shown for the indicated genotypes. Dots are individual plates with 10 animals each; bars are means; and error bars are SEM. N= 2 – 26 plates for each genotype. Exact ablation lines used are: AWB: peIs1715 [str-1p::mCasp-1 + unc-122p::GFP]; AQR/PQR/URX: qals2241[gcy-36::egl-1; gcy-35::GFP; lin-15(+)]; BAG: kyIs536 (flp-17p::p17 domain of human Caspase3::sl2::gfp; elt-2::GFP); kyIs538 (glb-5p::p12 domain of human Caspase3::sl2::gfp; elt-2::mcherry); ASI: oyIs84(gpa-4p::TU#813+gcy-27p::TU#814+gcy-27p::GFP+unc-122p::DsRed) TU#813 and TU#814 are split caspase vector; ASJ: mgIs40([daf-28p::nls-GFP]; jxEx100[trx-1::ICE + ofm-1::gfp]); ASK: qrIs2[sra-9::mCasp1]; AFD: ttx-1(p767); AIA: kyEX4745[gcy-28dp::unc-103(gf)::sl2::mCherry, elt-2::mCherry]; AWC, ASE: ceh-36(ky640); AIB: flvEx356[inx-1::unc-103::sl2GFP; myo-2::mCherry]: ASK, ASI: oyIs84(gpa-4p::TU#813+gcy-27p::TU#814+gcy-27p::GFP+unc-122p::DsRed) ;qrIs2[sra-9::mCasp1]. **p<0.01, one-way ANOVA with dunnett's multiple comparison test. (E) Percent eggs laid on high osmolarity (300 mOsm), compared to control (150 mOsm) osmolarity, shown for the indicated genotypes. Dots are individual plates with 10 animals each; bars are means; and error bars are SEM. N= 6 – 28 plates for each genotype. . Promoters used for tax-4 rescue are: tax-4 (all tax-4 expressing neurons), gcy-33 (BAG), sra-9 (ASK), srg-47 (ASI), srh-11 (ASJ). *p<0.05, **p<0.01, ****p<0.0001, one-way ANOVA with dunnett's multiple comparison test. (F) Percent eggs laid on high osmolarity (300 mOsm), compared to control (150 mOsm) osmolarity, shown for the indicated genotypes. Dots are individual plates with 10 animals each; bars are means; and error bars are SEM. N= 4 – 20 plates for each genotype. Alleles used are tax-4(p678), BAG ablation (kyIs536 (flp-17p::p17 domain of human Caspase3::sl2::gfp; elt-2::GFP); kyIs538 (glb-5p::p12 domain of human Caspase3::sl2::gfp; elt-2::mcherry)), flp-17 (n4894), flp-10 (ok2624). *p<0.05, **p<0.01, ****p<0.0001, one-way ANOVA with dunnett's multiple comparison test. (G) Number of eggs laid on control (150 mOsm) and high (300 mOsm) osmolarity, shown for Mock and HSN-axotomized animals (same site of HSN axotomy as shown in Fig. 5F). Dots show individual animals; bars are means; and error bars are SEM. N= 21 animals for Mock and 16 animals for HSN axotomized group. *p<0.05, two-way ANOVA followed by Sidak's multiple comparison test.