Cellular Expression and Functional Roles of All 26 Neurotransmitter GPCRs in the C. elegans Egg-Laying Circuit

GFP reporter strains to identify cells expressing each neurotransmitter GPCR

Our analysis includes 21 neurotransmitter GPCRs for monoamines and classical neurotransmitters, plus five “orphan” GPCRs (Fig. 1A). Expression patterns for 13 of these have been previously analyzed (Cho et al., 2000; Lee et al., 2000; Tsalik et al., 2003; Chase et al., 2004; Rex et al., 2004; Carnell et al., 2005; Dempsey et al., 2005; Carre-Pierrat et al., 2006; Suo et al., 2006; Wragg et al., 2007; Hapiak et al., 2009; Plummer, 2011; Gürel et al., 2012; Yemini et al., 2020) using GPCR gene promoter fragments of a few kilobases to drive GFP expression. However, recent work shows that obtaining accurate expression patterns often requires larger GFP reporter transgenes (Serrano-Saiz et al., 2013; Pereira et al., 2015; Gendrel et al., 2016) that include the full gene of interest and up to ∼20 kb of flanking genomic DNA on either side (Tursun et al., 2009).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

GFP reporter transgenes for 26 neurotransmitter GPCRs. A, Small-molecule neurotransmitter GPCRs in C. elegans. B, Schematics of GFP reporters used in this study. C-terminal::GFP constructs are illustrated by the reporter for the orphan receptor ADOR-1. The construct provides 21 kb of promoter region DNA 5′ to the coding exons and 9 kb of 3′ DNA. GFP coding sequences were inserted into the last ADOR-1 coding exon so that the transgene expresses the receptor with GFP fused to its C-terminus. SL2::GFP constructs are illustrated by the dopamine receptor DOP-1 reporter. This reporter is similar to that for ADOR-1, except that an SL2 trans-splicing signal followed by GFP coding sequences were inserted immediately downstream of the stop codon so that this reporter coexpresses the receptor and GFP as separate proteins. Analogous schematics of all 26 receptor::GFP reporter transgenes are found in Extended Data Figure 1-1 and described in Extended Data Figure 1-2.

Therefore, we sought to generate large GFP reporter transgenes for the 26 neurotransmitter GPCR genes (Fig. 1; see Extended Data Figs. 1-1, 1-2). Thirteen were already constructed (Sarov et al., 2012), and we generated seven more in a similar fashion. These “C-terminal::GFP” reporters express a GPCR with GFP fused at or very close to the C-terminus of the receptor (Fig. 1B). For a few receptors, these C-terminal::GFP fusion proteins were strongly localized to neural processes and excluded from the cell bodies, making it difficult to identify the GFP-expressing neurons. So, for four such receptors (DOP-1, DOP-4, GAR-2, and F35H10.10), we generated “SL2::GFP” reporters (Fig. 1B), which produce a single primary RNA transcript that is trans-spliced to produce separate GPCR and GFP mRNAs, so that the GFP produced fills out the cells that express the GPCR gene (Tursun et al., 2009). For the remaining two receptors (MGL-2 and SER-4), technical issues precluded use of large fosmid-based reporters, so we analyzed expression of smaller promoter::GFP reporters that had been previously generated (Tsalik et al., 2003; Gürel et al., 2012; Yemini et al., 2020).

The 26 GPCR::GFP reporter transgenes were each transformed into C. elegans to produce high-copy, chromosomally integrated transgenes. Such transgenes are expressed in the same cells as the corresponding endogenous genes (Serrano-Saiz et al., 2013; Pereira et al., 2015; Gendrel et al., 2016), but at high enough levels that the GFP fluorescence produced is sufficient to identify even cells that normally express a GPCR gene at low levels.

Cells of the egg-laying system

We first examined animals carrying transgenes that express mCherry in specific cells of the egg-laying system to generate accurate diagrams of its anatomy (Fig. 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Anatomy of the egg-laying system visualized with mCherry markers. A, Schematic of the C. elegans egg-laying system. Neuron and muscle cells (colored and labeled with cell names) are overlaid on the uterus (gray) containing eggs (white). VC4 and VC5 neurons are at the ventral midline. All other cells depicted are on the left side of the animal, and their equivalents on the right side of the bilaterally symmetric anatomy are not shown. B, Schematic of the egg-laying neurons. C, Confocal image of the egg-laying neurons labeled by the ida-1::mCherry marker. Dashed lines outline the cell bodies. D, Schematic of the egg-laying muscles. E, Confocal image of the egg-laying muscles labeled by the unc-103e::mCherry marker. um1 (outlined) and um2 (not indicated) are rarely and faintly labeled by unc-103e::mCherry; but even so, labeling of vm1/vm2 cells provides landmarks to help identify um1 and um2. F, Schematic of cells other than neurons and muscles of the egg-laying system. Red dashed or solid lines indicate cell junctions labeled by the ajm-1::mCherry marker. In addition to cells of the egg-laying system, the ajm-1::mCherry marker also labels junctions of seam and intestinal cells. G, Confocal image of the egg-laying system labeled by the ajm-1::mCherry marker. The following conventions are used in this and other figures in this work. Scale bars, 10 µm. Anterior is left and ventral down. du, Dorsal uterine cell; sp, spermatheca; um, uterine muscle cell; ut, uterine toroid cell; utse, uterine seam cell; uv, uterine ventral cell; vm, vulval muscle cell; vt, vulval toroid cells. All C. elegans strains used in this work are described in Extended Data Figure 2-1.

Neurons of the egg-laying system were labeled by ida-1::mCherry (Fig. 2B,C). These neurons are as follows: (1) the HSNs, which promote egg laying by releasing serotonin and neuropeptides (Waggoner et al., 1998; Shyn et al., 2003; Hapiak et al., 2009; Emtage et al., 2012; Brewer et al., 2019); (2) the uv1s, which inhibit egg laying by releasing tyramine and neuropeptides (Alkema et al., 2005; Collins et al., 2016; Banerjee et al., 2017); and (3) VC4/VC5, which release acetylcholine onto the vm2 egg-laying muscles (Duerr et al., 2001; Collins et al., 2016).

We refer to the four uv1 cells in this work as “neurons” despite the fact that they are not on the canonical list of 302 neurons established by White et al. (1986). The uv1 (uterine-vulval type 1) cells were originally assumed to be epithelial cells that simply help form the structure connecting the uterus to the vulva through which eggs are laid (Newman et al., 1996). However, a body of evidence now shows that the uv1s regulate egg-laying behavior by acting as neurosecretory cells: (1) the uv1 cells express neuronal markers, including synaptotagmin, syntaxin, and UNC-13 (Jose et al., 2007) as well as rab-3 (our unpublished observations); (2) the uv1 cells release the neurotransmitter tyramine to inhibit egg laying (Alkema et al., 2005; Collins et al., 2016); (3) the uv1 cells also release the neuropeptides NLP-7 and FLP-11 to inhibit egg laying (Banerjee et al., 2017); and (4) the uv1 cells show calcium transients at a specific time after each egg is laid (Collins et al., 2016). Thus, the uv1 cells have properties analogous to those of the more widely studied NSM cells, which are also neurosecretory cells that make no classical synapses (Nelson and Colón-Ramos, 2013), and which are included in the list of 302 canonical neurons (White et al., 1986).

Muscle cells of the egg-laying system were labeled by unc-103e::mCherry, which is expressed in the vm1 and vm2 vulval muscle cells and was used as a landmark to identify the um1 and um2 uterine muscle cells (Fig. 2D,E). Eggs are laid when vm1 and vm2 open the vulva and um1 and um2 muscle cells squeeze eggs out of the uterus through the vulva (Waggoner et al., 2000; Kim et al., 2001; Bany et al., 2003; Ringstad and Horvitz, 2008; Collins et al., 2016). There are also body wall muscle cells surrounding the egg-laying system (not shown in Fig. 2 but referred to as bwm in later figures).

Other cells of the egg-laying system were identified using ajm-1::mCherry, which labels the junctions between cells of the uterus and vulva (Fig. 2F,G). Eggs are fertilized in the spermatheca (sp) and enter the uterus via the spermatheca-uterine (sp-ut) valve. The uterus itself comprises the ut1-ut4 uterine toroid cells and du cells that together form a tube to hold eggs. Eggs pass from the uterus into the vulva through a junction comprising the uterine seam (utse) and uterine-ventral (uv1 and uv3) cells, and the vulva itself comprises several types of vulval toroid (vt) cells (Schindler and Sherwood, 2013; Ghosh and Sternberg, 2014; Ecsedi et al., 2015).

Analyzing GFP expression by neurotransmitter GPCR::GFP transgenes

We sought to (1) accurately identify every cell of the egg-laying system that expresses any of the 26 neurotransmitter GPCR::GFP transgenes; (2) qualitatively record the GFP expression levels for each reporter in each cell, which may reflect expression levels of the endogenous receptors; and (3) record animal-to-animal variability in GFP expression levels from these reporters, which may reflect variability in expression of the endogenous receptors. Thus, we crossed each of the 26 neurotransmitter GPCR::GFP transgenes into the mCherry marker strains and collected 3D confocal images for at least 10 young adult animals per double-labeled strain. In total, images of 809 animals were analyzed.

The process for identifying GFP-expressing cells and scoring GFP levels is illustrated for the octopamine receptor OCTR-1 (Fig. 3; Movie 1). Each image was analyzed using 3D rotations, cuts through the image volumes, and mCherry markers to identify internal features. The strongest GFP-expressing cells in a strain were scored as +++, those with weaker but easily detectable GFP as ++, and those with GFP just above background as +.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Scoring expression of a GFP reporter for the octopamine receptor OCTR-1 in cells of the egg-laying system. A, B, Two-dimensional renderings of a confocal image of the egg-laying system in an octr-1::gfp; ajm-1::mCherry young adult showing GFP (A) or GFP and mCherry fluorescence (B). The ut2 uterine toroid cells are outlined by the ajm-1::mCherry marker, allowing them to be identified as GFP-positive despite not being uniformly filled out by OCTR-1::GFP. Solid line with arrowheads labels the ventral nerve cord (vnc), a bundle of axons that run past the vulva but that are not known to affect egg laying. In these and other figure panels, white asterisks indicate position of the vulval opening, cells labeled with both GFP and mCherry are named with yellow text, cells labeled with GFP only are named with green text, the number of + symbols indicates the relative intensity of GFP labeling, and identified cells are outlined by faint dotted lines. B, White box represents the region shown at higher magnification in C-F. C, D, An octr-1::gfp; ida-1::mCherry animal with GFP fluorescence displayed at high gain to make ++ GFP labeling visible in the HSN and VC4/VC5 neurons. These neurons are double-labeled by ida-1::mCherry in D. E, F, An octr-1::gfp; unc-103e::mCherry animal with GFP fluorescence displayed at high gain and at a different focal plane than in C, D to visualize + GFP labeling in vm1 and vm2 egg-laying muscles. Movie 1 is a video that illustrates in more detail how octr-1::gfp strains were analyzed. Extended Data Figure 3-1 illustrates how animal-to-animal variations in GFP expression were analyzed.

For reporters that expressed a receptor with GFP fused to its C-terminus, the fusion proteins were sometimes subcellularly localized. An example is seen in Figure 3B, where OCTR-1::GFP clusters within the ut2 uterine toroid cells. It is unclear whether such subcellular localization might reflect a similar localization of the corresponding endogenous receptors; thus, we chose not to interpret such subcellular localization in this work.

Neurotransmitter GPCR expression in the C. elegans egg-laying system

Figure 4 shows 2D image renderings illustrating a subset of findings, whereas Extended Data Figure 4-1 shows a larger set of images for each of the 26 neurotransmitter receptors. Figure 5 summarizes all the findings from all 809 images analyzed. In total, our results show 139 cases in which a neurotransmitter GPCR is expressed in a specific cell type of the egg-laying system.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Examples of neurotransmitter GPCR::GFP expression in neuronal, muscle, and other cells of the egg-laying system. A-H, Examples of GPCR::GFP expression in neurons of the egg-laying system. Top panel of each image pair shows reporter GFP reporter fluorescence for the receptor named in green, with all GFP-labeled cells of the egg-laying system indicated by white cell names and arrowheads. Bottom, GFP plus the ida-1::mCherry fluorescence used to confirm identity of egg-laying neurons, with only the double-labeled neurons indicated by yellow names and white arrowheads. I-P, Examples of GPCR::GFP expression in muscle cells of the egg-laying system. Images are labeled as in A-H, but bottom panels in each pair show the unc-103e::mCherry fluorescence used to confirm identity of muscle cells. mCherry labeling of um1 is not visible in J. Q-V, Examples of GPCR::GFP expression in cells of the egg-laying system other than neurons and muscles. Images are labeled as in A-H, but bottom panels in each pair show ajm-1::mCherry fluorescence. sp, ut, utse, uv3, and vt cells are identified by their outlines visualized with ajm-1::mCherry. Extended Data Figure 4-1 presents examples of GPCR::GFP expression for all 26 neurotransmitter GPCRs.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Summary of neurotransmitter GPCR::GFP expression in the C. elegans egg-laying system. A, Summary of neurotransmitter GPCR::GFP expression by cells in and near the egg-laying system. Dark green represents GFP expression that was on average strong (scored as +++, among brightest of any cell in the animal) or easily detectable (++). Light green represents weak GFP expression (+, just above background in at least 2 different animals). White represents cells in which GFP was not reproducibly observed. v, Variable (denotes cells in which GFP detected in more than two but less than half of animals observed). bwm, Body wall muscles; sp-ut, spermathecal-uterine valve. B, Number of strongly and weakly expressed receptors in each cell type of the egg-laying system. Receptors with variable expression are included. Totals are shown broken down by the Gα protein thought to be activated by the receptors in each cell type when that information is known. C, Neurons and muscles of the egg-laying system that express neurotransmitter GPCRs, indicating which cells have the corresponding neurotransmitter synaptically released onto them (synaptic target cells) or not (extrasynaptic target cells), based on the connectome (Albertson and Thomson, 1976; White et al., 1986; Xu et al., 2013) and neurotransmitter maps (Serrano-Saiz et al., 2013; Pereira et al., 2015; Gendrel et al., 2016). Cell names in dark/light green represent strong/weak receptor expression. Italics indicate that receptor expression is variable. Extended Data Figure 5-1 presents a more detailed summary of our GPCR::GFP expression results, including comparison with previously published work and details of variability in GFP expression.

Although 13 of the 26 neurotransmitter GPCRs were previously reported to be expressed in the egg-laying system, the vast majority of the expression we found was not previously reported (Extended Data Fig. 5-1A). In 20 instances, we reproduced a published report of a receptor being expressed in a particular cell type; but in 119 additional instances, we made a novel observation of a receptor being expressed in a cell type. In eight instances, a published case of a receptor being expressed in a cell type was not reproduced in our data (Extended Data Fig. 5-1A).

While GFP reporters were often expressed strongly and reproducibly in a given cell type, we also observed cases of weak and/or variable GFP expression (Fig. 5; Extended Data Fig. 5-1B). This phenomenon is illustrated for octr-1::gfp expression in the HSN, VC4, and VC5 neurons in Extended Data Figure 3-1. We collected data for all detectable expression so that future studies can investigate what types of expression of a GPCR::GFP reporter might reflect functional presence of the corresponding endogenous receptor.

We note that single-cell RNA sequencing (RNAseq) has the potential to map which cells express any gene, including the neurotransmitter GPCR genes we characterize here. Our data are far richer than the currently available RNAseq data for the egg-laying system. Our work scores expression in 19 cell types in the egg-laying system (Fig. 5) versus only four in the current RNAseq data (Taylor et al., 2020). As examples, we scored expression in vm1 and vm2 separately rather than aggregating all vulval muscle cell expression, and we included in our analysis cells other than neurons and muscles, which turn out to express many neurotransmitter GPCRs. Further, we measured variability in expression from animal to animal, and discovered weak expression below the detection threshold of current RNAseq data.

Individual cells express multiple neurotransmitter GPCRs

Individual cells expressed up to 20 different neurotransmitter GPCRs (Fig. 5). Neurons (examples in Fig. 4A-H) each expressed 2-7 receptors at strong, reproducible levels and 4-10 additional receptors at weak and/or variable levels (Fig. 5). Non-neuronal cells also expressed multiple neurotransmitter GPCRs strongly and reproducibly; but unlike neurons, these cells expressed only 0-3 additional receptors at weak and/or variable levels. Figure 4I-P shows examples of GPCR::GFP expression in muscles cells, whereas Figure 4Q-V shows examples of GPCR::GFP expression in non-neuronal, nonmuscle cells of the egg-laying system.

The multiple receptors expressed on a single cell often couple to different G proteins (Fig. 5B). The C. elegans G proteins Gα_o, Gα_q, and Gα_s are each expressed in most or all neurons plus some muscle and other cells (Koelle, 2018). Receptors that couple to Gα_q and Gα_s cooperate to activate neurons and muscles, whereas receptors that couple to Gα_o oppose such activation (for review, see Koelle, 2018). Therefore, our results suggest that individual cells of the egg-laying system have the ability to synthesize information received from multiple neurotransmitters and to compute an appropriate resulting level of activity by integrating signaling from multiple receptors and G proteins.

Neurotransmitter GPCRs are often positioned to receive extrasynaptic signals

As expected, neurotransmitter receptors were often expressed on cells that receive synapses from neurons releasing their activating neurotransmitter. More interestingly, receptors were also expressed on cells that do not receive any synapses at which the corresponding neurotransmitter is released (Fig. 5C).

Serotonin receptors illustrate expression in cells positioned to receive synaptic versus extrasynaptic signals. In the egg-laying circuit, HSN neurons release serotonin to stimulate egg laying at synapses onto both the vm2 muscle cells and VC neurons (White et al., 1986; Collins and Koelle, 2013; Brewer et al., 2019). We observed that serotonin receptors are expressed, as expected, on these vm2 and VC postsynaptic cells (Fig. 5). If we restrict our analysis only to strongly expressed receptors, vm2 cells express three serotonin receptors: the SER-1 and SER-7 receptors that stimulate egg laying by coupling to Gα_q and Gα_s, respectively, and the SER-4 receptor that inhibits egg laying by coupling to Gα_o (Cho et al., 2000; Carnell et al., 2005; Dempsey et al., 2005; Carre-Pierrat et al., 2006; Hapiak et al., 2009; Gürel et al., 2012). The VC neurons express SER-7, which could mediate the ability of the HSN to excite the VCs (Collins et al., 2016). Finally, the HSN variably expresses SER-7, which might serve as an autoreceptor to allow the HSN to excite itself. The above instances of serotonin receptor expression all could mediate synaptic serotonin signaling.

Remarkably, our data show seven additional instances in which serotonin receptors are strongly and reproducibly expressed by cells of the egg-laying system that do not receive synapses from the HSN or any other serotonergic neurons (Fig. 5). For example, the vm1 and um1/2 egg-laying muscle cells, which contract with the vm2 muscle cells to cause egg laying, all express serotonin receptors. Therefore, serotonin released by HSN may travel extrasynaptically to the vm1 and um1/2 cells to signal through serotonin receptors on these cells to help induce their activity and egg-laying events.

Acetylcholine, GABA, and glutamate, like serotonin, each have GPCRs expressed on cells onto which these neurotransmitters are synaptically released. In addition, acetylcholine, GABA, and glutamate each have GPCRs expressed on neurons, muscles, and/or epithelial cells of the uterus that could receive these signals extrasynaptically (Fig. 5A,C).

No cells of the egg-laying system receive synapses from neurons releasing dopamine, octopamine, or tyramine, yet receptors for each are expressed in the egg-laying system (Fig. 5). For example, each of the four dopamine receptors DOP-1 through DOP-4 is strongly and reproducibly expressed in at least one cell type of the egg-laying circuit (Fig. 5A). Indeed, endogenous dopamine has been shown to promote egg laying (Nagashima et al., 2016), and exogenous dopamine has been reported to both promote and inhibit egg-laying (W. R. Schafer and Kenyon, 1995; Vidal-Gadea et al., 2012). Similarly, application of octopamine and tyramine has been shown to inhibit egg laying (Horvitz et al., 1982; Alkema et al., 2005; Collins et al., 2016); and as in the case of dopamine, these effects may occur via extrasynaptic signaling through the receptors for these neurotransmitters we found expressed in the egg-laying system.

Neurotransmitter GPCRs are expressed by cells other than neurons and muscles

In addition to being expressed in neurons and muscles (Fig. 4A-P), GPCR::GFP reporters were also frequently expressed on other cell types not previously known to respond to neurotransmitters (Figs. 4Q-V, 5A,B). Such cells include the uterine toroid epithelial cells ut1-ut4 that form a tube that holds unlaid eggs (Fig. 2F,G), the vulval toroid (vt) epithelial cells that form the opening through which eggs are laid (Fig. 2F,G), and the spermatheca that holds sperm and through which eggs pass to be fertilized (Fig. 2F,G).

The presence of neurotransmitter receptors on epithelial cells has precedent in mammals, as serotonin receptors are found on epithelial/endothelial cells of the gut (Berger et al., 2009; Hoffman et al., 2012) and blood vessels (Ullmer et al., 1995; Berger et al., 2009). The presence of receptors on the uterine toroid epithelial cells may indicate that these ut cells receive neurotransmitter signals to aid in pushing eggs out of the uterus (Newman et al., 1996). Neurotransmitter receptors on the spermatheca may regulate its stretch and contraction before ovulation (Wirshing and Cram, 2017), particularly since downstream effectors of the G protein Gα_q have been shown to be involved in ovulation (Clandinin et al., 1998).

Knockout mutants suggest neurotransmitter GPCRs function redundantly and/or under specific conditions

We analyzed knockout mutants for each of the receptors for changes in the frequency of egg-laying behavior. The mutations used are all deletions that remove large portions of the receptor coding regions and that are predicted to result in a lack of any functional receptor expression. One assay of egg-laying behavior counts unlaid eggs in a worm (Fig. 6A). A WT worm both produces and lays eggs, accumulating ∼18 unlaid eggs in its uterus at steady state. Mutants with a hyperactive or inactive egg-laying circuit continue to produce eggs but accumulate fewer or more unlaid eggs, respectively (Chase et al., 2004). A second assay of egg-laying behavior scores the developmental stage of just-laid eggs. Eggs in WT animals undergo multiple rounds of cell divisions while they wait in the uterus to be laid, but eggs of hyperactive egg-layers are laid shortly after being fertilized and often at earlier stages of development, with fewer than eight cells (Fig. 6A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Neurotransmitter GPCRs knockouts rarely reveal egg-laying defects. A, Micrographs of animals with egg-laying phenotypes. WT adults (top) typically accumulate 15-20 unlaid eggs, whereas hyperactive egg-laying mutants, such as like goa-1 (second from top), accumulate fewer unlaid eggs and egg-laying defective mutants, such as egl-10 (third from top), accumulate more unlaid eggs. Eggs from WT animals have developed to the ∼100 cell stage by the time they are laid (top right), whereas eggs from hyperactive egg-laying mutants are laid earlier, often with fewer than eight cells (bottom right). B, Number of unlaid eggs for neurotransmitter GPCR single knockout strains. egl-10(md176) and goa-1(n1134) served as controls (these control data are replotted in Fig. 7C and Extended Data Fig. 6-1B). n ≥ 30 for each strain. *Statistical significance (p < 0.05) compared with the WT using one-way ANOVA with Bonferroni correction for multiple comparisons. Significant p values were as follows: goa-1 (p < 0.0001); egl-10 (p < 0.0001); ser-5 (p = 0.0002); dop-4 (p = 0.0004); tyra-2 (p = 0.001); gar-3 (p = 0.0021); gbb-1 (p = 0.0253); gbb-2 (p < 0.0001); mgl-1 (p < 0.0001); ador-1 (p < 0.0001); F35H10.10 (p < 0.0001); dop-5 (p = 0.0001); and dop-6 (p < 0.0001). Error bars indicate 95% CIs. Here and in other graphs in this work, different colors are used to plot data for receptors for different neurotransmitters, or for the orphan GPCR class. C, Percent of early-stage eggs laid by neurotransmitter GPCR gene single knockout strains. C, D, goa-1(n1134) is used as a control hyperactive egg-laying mutant (these control data are replotted in Fig. 6D and Extended Data Fig. 7-1A). n = 100 eggs per strain. *Statistical significance (p < 0.05) compared with the WT using the Fisher's exact test. Significant p values were as follows: goa-1 (p < 0.00001); ser-5 (p = 0.0349); and F35H10.10 (p = 0.0349). The Wilson-Brown method was used to determine the 95% CI for binomial data. D, Percent of early-stage eggs laid by strains carrying combinations of KOs in genes encoding Gα_o-coupled neurotransmitter GPCRs. goa-1(n1134) is used as a control hyperactive egg-laying mutant. Red “X” identifies the receptors knocked out in a given strain. Statistical analysis was as in C. The five significant p values (for measurements from left to right that are denoted with asterisks) were p < 0.00001, p = 0.0101, p = 0.0055, p < 0.00001, and p = 0.0184. Extended Data Figure 6-1 presents further analysis of neurotransmitter GPCR knockout strains.

Single KOs of neurotransmitter GPCRs showed at most modest changes in the number of unlaid eggs accumulated (Fig. 6B). While the dop-4(ok1321) null mutant used in Figure 6B appeared WT for egg laying, a second independent allele of dop-4 showed a possible defect (Extended Data Fig. 6-1A). However, we were unable to rescue the egg-laying defect of this second mutant with a WT dop-4 transgene (data not shown), suggesting that its egg-laying defect was an artifact of the genetic background and not because of loss of dop-4. Twelve GPCR KOs accumulated significantly fewer unlaid eggs than did the WT (Fig. 6B); however, only two of these laid significantly more early-stage eggs than did the WT control, suggesting that the other 10 may have accumulated fewer unlaid eggs simply because of decreased egg production, not a change in egg laying. Even the two that laid significant early-stage eggs did so at levels far below the 50% early-stage egg threshold considered biologically significant (Bany et al., 2003). In summary, loss of no single neurotransmitter causes strong defects in the standard egg-laying assays used in Figure 6.

Because the goa-1 mutant for the G protein Gα_o has a strong hyperactive egg-laying phenotype (Fig. 6), we hypothesized that the effects of knockouts for receptors that signal through this G protein were masked by functional redundancy. Thus, we crossed together Knockout mutations for Gα_o-coupled neurotransmitter GPCRs to see whether this would reveal a strong hyperactive egg-laying phenotype. We generated a series of combination knockouts, from double knockouts to a quintuple knockout, but none of these revealed a hyperactive egg-laying phenotype approaching the strength of the goa-1 mutant (Fig. 6D), nor did they show an abnormal accumulation of unlaid eggs (Extended Data Fig. 6-1B). One possibility is that Gα_o signals through even more Gα_o-coupled receptors, such as neuropeptide receptors (Ringstad and Horvitz, 2008), which would need to be knocked out to phenocopy the Gα_o mutant.

Another interpretation of the lack of strong egg-laying defects in neurotransmitter GPCR knockouts is that we only assayed egg-laying behavior under standard laboratory growth conditions. Egg-laying behavior is known to be regulated by many different conditions, such as environmental CO₂ (Fenk and de Bono, 2015), the presence or absence of food (Dong et al., 2000), and whether the animal is in the roaming versus dwelling state of locomotion (Cermak et al., 2020). Specific neurotransmitter GPCR knockouts might be required to regulate egg laying in these or other specific circumstances.

Overexpression of neurotransmitter GPCRs reveals egg-laying defects

We next analyzed strains that overexpress individual neurotransmitter receptors, reasoning that this could result in gain-of-function phenotypes that might reveal receptor functions, even if individual receptors function redundantly with each other. Our high-copy GPCR::GFP transgenes should overexpress neurotransmitter receptors. The 20 C-terminal::GFP reporters overexpress a receptor with GFP fused to its C-terminus, while the four SL2::GFP reporters overexpress a receptor without any such modification (Extended Data Fig. 1-1). Two of our GPCR::GFP transgenes (for SER-4 and MGL-2) were constructed for technical reasons in a manner that would not cause receptor overexpression, so we did not analyze overexpression of these two receptors.

We piloted the neurotransmitter receptor overexpression strategy with the serotonin receptor SER-1, which had previously been reported to stimulate egg laying by acting partially redundantly with the serotonin receptor SER-7 (Hapiak et al., 2009). In the context of this experiment, we refer to the ser-1::gfp transgene as “ser-1(oe)” to indicate that the ser-1 gene is overexpressed. The experimental strategy is schematized in Figure 7A; we expected SER-1 overexpression to increase serotonin stimulation of egg laying and result in hyperactive egg laying, and that knocking out serotonin biosynthesis would suppress this effect since the overexpressed SER-1 would no longer be activated by serotonin. As expected, the ser-1(oe) strain displayed a strong hyperactive egg-laying phenotype (Fig. 7B). The same phenotype was also observed in animals carrying an extrachromosomal ser-1(oe) transgene and in two independent chromosomal integrants for the ser-1(oe) transgene (Extended Data Fig. 7-1B), suggesting that the phenotype was not an artifact of chromosomal integration. The hyperactive egg-laying phenotype of ser-1(oe) was completely suppressed in animals lacking tryptophan hydroxylase enzyme (Fig. 7B), an enzyme required for serotonin biosynthesis (Sze et al., 2000). This control suggests that the effect of the high-copy ser-1 transgene is because of the overexpressed SER-1 serotonin receptor mediating increased serotonin signaling. This control is crucial because the high copy transgenes we are using to overexpress neurotransmitter GPCRs each contain a ∼40 kb stretch of C. elegans genomic DNA that often carries one or more flanking genes in addition to the GPCR gene, and these flanking genes are also overexpressed. Our results overexpressing ser-1 were consistent with the previous data suggesting that SER-1 is one of two partially redundant receptors through which serotonin signals to stimulate egg laying (Hapiak et al., 2009).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Neurotransmitter GPCR overexpressors reveal strong egg-laying defects. A, Strategy to reveal and validate egg-laying defects caused by increased signaling through overexpressed neurotransmitter GPCRs. B, Overexpression of the serotonin GPCR SER-1 (ser-1(oe)) causes hyperactive egg-laying that is suppressed by knocking out serotonin biosynthesis with a null mutant in the tryptophan hydroxylase enzyme (TPH-1). n = 100 eggs per strain. Error bars indicate 95% CIs determined by the Wilson-Brown method. *Statistical significance (p < 0.05) compared with the WT (Fisher's exact test). p values for the two asterisked comparisons were as follows: p < 0.00001 (left comparison) and p < 0.00001 (right comparison). C, Number of unlaid eggs for neurotransmitter GPCR gene overexpressor strains. egl-10(md176) and goa-1(n1134) served as controls. n ≥ 30 for each strain. Error bars indicate 95% CIs. *p < 0.05, compared with the WT (one-way ANOVA with Bonferroni correction for multiple comparisons). p values for the comparisons to the WT were as follows: goa-1 (p < 0.0001); egl-10 (p < 0.0001); ser-1(oe) (p = 0.0011); ser-5(oe) (p = 0.0025); ser-7(oe) (p = 0.0189); dop-2(oe) (p < 0.0001); dop-3(oe) (p < 0.0001); dop-4(oe) (p = 0.0019); octr-1(oe) (p = 0.0223); ser-6(oe) (p < 0.0001); ser-2(oe) (p = 0.009); tyra-2(oe) (p < 0.0001); gar-1(oe) (p < 0.0001); gar-2(oe) (p < 0.0001); gbb-1(oe) (p < 0.0001); gbb-2(oe) (p < 0.0001); mgl-1(oe) (p < 0.0001); mgl-3(oe) (p < 0.0001); ador-1(oe) (p = 0.0006); dop-5(oe) (p < 0.0001); and dop-6(oe) (p = 0.0389). D, Overexpression of the glutamate GPCR MGL-1 causes defective egg laying that is suppressed by knocking out glutamate signaling with a mutation in the vesicular glutamate transporter EAT-4. D-G, n ≥ 30 for each strain. Error bars indicate 95% CIs. *p < 0.05 (one-way ANOVA with a Tukey's test to determine statistical significance for multiple comparisons). p values for the two asterisked comparisons in D were as follows: p < 0.0001 (left comparison) and p < 0.0001 (right comparison). E, Overexpression of the glutamate GPCR MGL-3 causes defective egg laying that is suppressed by knocking out glutamate signaling with a mutation in the vesicular glutamate transporter EAT-4. p values for the two asterisked comparisons were as follows: p < 0.0001 (left comparison) and p < 0.0001 (right comparison). F, Overexpression of the tyramine GPCR SER-2 causes defective egg-laying that is suppressed by knocking out the tyramine biosynthesis enzyme tyrosine decarboxylase (TDC-1). p values for the two asterisked comparisons were as follows: p < 0.0001 (left comparison) and p < 0.0001 (right comparison). G, Overexpression of the octopamine GPCR SER-6 causes defective egg laying that is suppressed by knocking out the octopamine biosynthesis enzyme tyramine β-hydroxylase (TBH-1). p values for the two asterisked comparisons were as follows: p < 0.0001 (left comparison) and p = 0.0257 (right comparison). Extended Data Figure 7-1 presents additional validation tests of neurotransmitter GPCR overexpression phenotypes.

We examined all 24 neurotransmitter GPCR overexpressor transgenes and found six more instances of strong effects on egg laying. We applied the same validation tests we used for ser-1 overexpression to each of these 6 additional cases: we required that multiple independent overexpressor transgenes for a GPCR produce similar effects on egg laying, and that these effects are suppressed by a mutation that blocks synthesis or transport of the corresponding neurotransmitter. We found we could not validate two of these 6 cases: the GABA receptor GBB-1 and the octopamine receptor SER-3 each had one overexpressor transgene that resulted in changes to egg laying, but these phenotypes were not reproduced using other independent overexpressor transgenes (Extended Data Fig. 7-1A,C,H). However, the remaining 4 cases were validated, and each suggests novel functional effects of neurotransmitter signaling in the egg-laying system, as described below.

Overexpression of the glutamate GPCRs mgl-1(oe) and mgl-3(oe) caused strong accumulation of unlaid eggs (Fig. 7C). Extrachromosomal mgl-1(oe) and mgl-3(oe) transgenes also caused similar egg-laying defects (Extended Data Fig. 7-1E,F), and abolishing glutamate signaling by knocking out the vesicular glutamate transporter EAT-4 (Lee et al., 1999) completely suppressed the egg-laying defects of glutamate receptor overexpression (Fig. 7D,E). In the egg-laying system, mgl-1 is weakly and variably expressed in uv1 and HSN neurons but shows strong and consistent expression in specific epithelial cells of the uterus, whereas mgl-3 shows its strongest (albeit variable) expression in the VC4 neuron (Fig. 5). These glutamate receptors are also expressed outside the cells of the egg-laying system (Yemini et al., 2020), and it remains to be determined where they act to affect egg laying. Our results represent the first time glutamate signaling has been shown to affect C. elegans egg laying, and provide a starting point for future analysis of this issue.

Overexpression of the octopamine receptor SER-6 caused accumulation of unlaid eggs (Fig. 7C). We saw similar phenotypes using other extrachromosomal and chromosomally integrated ser-6(oe) transgenes (Extended Data Fig. 7-1D) and could suppress this egg-laying defect by knocking out the octopamine biosynthetic enzyme TBH-1 (Alkema et al., 2005) (Fig. 7G). Previous studies had shown that exogenous octopamine inhibits egg-laying behavior (Horvitz et al., 1982), but the mechanism and receptor(s) mediating this effect had remained unknown. SER-6, along with other octopamine receptors, is strongly expressed in the uv1 neuroendocrine cells and in the epithelial cells of the uterus (Fig. 5). To date, the uterine epithelial cells have not been implicated in regulation of egg laying. However, given that SER-6 is coupled to the excitatory G protein Gα_s, we hypothesize that SER-6 could activate the uv1s, which in turn inhibit egg laying (Jose et al., 2007; Banerjee et al., 2017).

Overexpression of the acetylcholine receptor GAR-1 similarly caused accumulation of unlaid eggs (Fig. 7C). We reproduced this defect with an extrachromosomal gar-1(oe) transgene (Extended Data Fig. 7-1G) but could not attempt to further validate this effect since a null allele for the acetylcholine biosynthetic enzyme is lethal (Rand, 1989). Inhibition of egg laying via GAR-1 signaling would be interesting as acetylcholine has been shown to both activate and inhibit egg laying (Bany et al., 2003; W. F. Schafer, 2006). GAR-1 is expressed principally in the uv1 neuroendocrine cells and the vm1/2 vulval muscles (Fig. 5). As a Gα_o-coupled receptor, GAR-1 is expected to reduce neural/muscle activity. Thus GAR-1 could activate egg laying by inhibiting the uv1 cells, which in turn inhibit egg laying, and GAR-1 could inhibit egg laying by inhibiting the vm1/2 muscles, which execute egg laying.

The final receptor that displayed a validated egg-laying defect when overexpressed was the tyramine receptor SER-2. The integrated ser-2(oe) transgene caused a mild retention of unlaid eggs (Fig. 7C) that was reproduced with an extrachromosomal ser-2(oe) transgene (Extended Data Fig. 7-1I) and suppressed when we knocked out the tyramine biosynthetic enzyme tyrosine decarboxylase (Alkema et al., 2005) (Fig. 7F). Tyramine was previously known to inhibit egg laying (Alkema et al., 2005; Collins et al., 2016). SER-2 is weakly and/or variably expressed in neurons (HSN, uv1) and muscle cells (um2) of the egg-laying system, but strongly expressed in certain non-neuronal/nonmuscle cells, such as epithelial cells of the vulva (Fig. 5). SER-2 is coupled to the inhibitory G protein Gα_o, and we hypothesize that it could inhibit egg laying by acting in the HSN and um2 cells that act to stimulate egg laying.

Rich and functionally robust neurotransmitter GPCR signaling controls neural circuit activity

This study provides, for the first-time, analysis of the expression and functions of every neurotransmitter GPCR in a model neural circuit. The C. elegans egg-laying circuit was already one of the most intensively studied small neural circuits, yet this study provides new insights into this circuit that may apply to neural circuits in general.

The first insight is that there may be far more neurotransmitter signaling within neural circuits than had been suggested by previous work. All cells in the egg-laying system express multiple neurotransmitter receptors, and each cell type expresses its own unique combination of up to 20 such receptors. Cells of the egg-laying circuit express receptors for neurotransmitters not previously implicated in control of egg laying, such as glutamate and GABA. Cells previously thought to have merely structural roles, such as the epithelial cells that make up the uterus and vulva, also express neurotransmitter receptors. Expression and function of neurotransmitter receptors in cells other than neurons and muscles have not been widely studied, yet our results suggest that this phenomenon may be widespread.

A second insight is that circuits may be remarkably robust systems because of an apparent high level of redundancy and compensation among neurotransmitter GPCRs. While mutations that affect downstream signaling by the neural G proteins Gα_o, Gα_q, and Gα_s each show dramatic effects on egg laying (W. F. Schafer, 2006; Koelle, 2018), no single neurotransmitter GPCR receptor knockout causes significant defects. The receptor knockout strains we analyzed in our study are grossly healthy and move relatively normally (data not shown), suggesting that functional redundancy of neurotransmitter GPCRs extends beyond the egg-laying system. Another interpretation of the lack of obvious defects in GPCR knockouts is that GPCRs may function to regulate egg-laying only under specific circumstances that we did not test. An example of neurotransmitter GPCRs functioning to control egg-laying both redundantly and only in specific circumstances came to light very recently. Cermak et al. (2020) found that the DOP-2 and DOP-3 dopamine receptors function redundantly to promote egg laying, but only during the ∼20% of the time worms in standard laboratory conditions spend in the “roaming” state of locomotion. We found both dop-2 and dop-3 were expressed in the egg-laying system, but we did not assay the dop-2; dop-3 double mutant or assay animals specifically in the roaming state, so we could not detect the function of dopamine receptors in egg laying discovered by Cermak et al. (2020).

We found that overexpressing neurotransmitter GPCRs generate gain-of-function phenotypes that can reveal receptor functions despite the issue of redundancy. This strategy reveals novel potential functions of dopamine, octopamine, acetylcholine, GABA, and glutamate, and orphan receptors in controlling the egg-laying circuit. Receptor overexpression may similarly be useful in revealing redundant neurotransmitter signaling functions in other circuits.

Our results generate a rich set of hypotheses to direct future studies of the egg-laying circuit. For example, prior work has shown that egg laying is activated when the HSN releases a combination of serotonin and the neuropeptide NLP-3 (Brewer et al., 2019). These two signals act partially redundantly to turn on the egg-laying circuit, with serotonin itself acting through partially redundant receptors (Hobson et al., 2006; Xiao et al., 2006; Hapiak et al., 2009). The multiple layers of redundancy in this circuit activation step lead to circuit activity being highly robust, but they also complicate efforts to understand the mechanism of circuit activation. Our results show that the serotonin receptors that activate the circuit, principally SER-1 and SER-7, with SER-5 playing a lesser role (Hapiak et al., 2009), are each expressed in different overlapping sets of cells in the egg-laying system. SER-1 and SER-7 are both expressed in the vm2 muscle cells, whose serotonin activation is crucial for executing egg laying (Brewer et al., 2019). SER-7 is also expressed in the VC neurons, which synapse onto vm2. The HSNs activate the VCs, which fire every time an egg is laid (Collins et al., 2016), and our findings suggest the hypothesis that HSNs activate the VC neurons via the SER-7 serotonin receptor. We also found that the SER-1, SER-5, and SER-7 receptors are expressed on the HSNs themselves, suggesting the hypothesis that serotonin from the HSN could act in an autocrine feedback loop to maintain HSN activity. Prior work has shown that the uv1 cells release tyramine and a combination of neuropeptides (Collins et al., 2016; Banerjee et al., 2017), with these signals acting redundantly to inhibit the egg-laying circuit, and that tyramine itself acts via multiple redundant receptors. Our results show which cells of the circuit express each of the three tyramine receptors, setting the stage for studies to tease out the redundant mechanisms by which the circuit is inactivated.

The strategies and tools applied in this work for analyzing neurotransmitter signaling in the egg-laying circuit can be applied to any neural circuit in C. elegans or other organisms to generate sets of testable hypotheses for how signaling among the cells of a neural circuit control its activities and function.