Identified neurons in C. elegans coexpress vesicular transporters for acetylcholine and monoamines

Janet S. Duerr, Jennifer Gaskin, James B. Rand


We have identified four neurons (VC4, VC5, HSNL, HSNR) inCaenorhabditis elegans adult hermaphrodites that express both the vesicular acetylcholine transporter and the vesicular monoamine transporter. All four of these cells are motor neurons that innervate the egg-laying muscles of the vulva. In addition, they all express choline acetyltransferase, the synthetic enzyme for acetylcholine. The distributions of the vesicular acetylcholine transporter and the vesicular monoamine transporter are not identical within the individual cells. In mutants deficient for either of these transporters, there is no apparent compensatory change in the expression of the remaining transporter. This is the first report of neurons that express two different vesicular neurotransmitter transporters in vivo.

  • vesicular acetylcholine transporter
  • vesicular monoamine transporter
  • Caenorhabditis elegans
  • cotransmission
  • nematode

biochemical and pharmacological studies of vertebrate neuronal and neuroendocrine tissues have shown four distinct vesicular transmitter transport activities: a monoamine transporter (biogenic amines), an acetylcholine transporter, a γ-aminobutyric acid (GABA)/glycine transporter, and a glutamate transporter (reviewed in Ref. 9). The vesicular acetylcholine transporters (VAChTs) and the vesicular monoamine transporters (VMATs) have been shown to be members of the same subfamily of 12-transmembrane domain transport proteins (2, 10,13, 24, 42). VAChT is expressed in the same cells as choline acetyltransferase (ChAT, the enzyme for acetylcholine synthesis) and has been shown to transport acetylcholine in vitro (13). Although Caenorhabditis elegans appears to have a single VMAT gene (7), two related VMAT proteins (and genes) have been identified in mammals: VMAT1 is often found in neuroendocrine cells, and VMAT2 is primarily neuronal (10, 12, 24, 44). Recombinant VMATs have been shown to mediate the transport in vitro of dopamine, norepinephrine, epinephrine, serotonin, and histamine (VMAT2 only) (7, 11, 12). Antibodies to the vesicular transporters have been used to study vesicular localization and maturation (25, 30, 44, 45) as well as cell identity (28, 36, 37). The particular neurotransmitter used by VMAT-positive cells must be determined using additional specific markers.

We now report that several identified neurons of the nematode C. elegans contain both VAChT and VMAT; some of these cells contain serotonin, suggesting that they may be both cholinergic and serotonergic. This represents the first reported example of two different vesicular transporters being expressed in the same neurons in vivo.


Nematode strains.

Wild-type C. elegans (N2 Bristol) was grown on solid medium (3, 19). The bas-1(ad446),cat-1(e1111), cha-1(p1152),egl-1(n487), lin-39(n709), tph-1(mg280), andunc-17(e245) mutations have been described (3, 4, 17,27, 33, 35, 39).


Goat antiserum to C. elegans VMAT has been described (7). To stain the VAChT/UNC-17 protein, we used a mouse monoclonal antibody (1403), which stained similarly to the previously described rabbit serum (R383) (2). Purchased antibodies included mouse anti-dopamine (Biodesign International, Kennebunk, ME), rabbit anti-histamine, rabbit anti-tyramine (both from Sigma, St. Louis), rabbit anti-octopamine (Biogenesis, Sandown, NH), and rabbit anti-serotonin (H. Steinbusch, Maastricht University, The Netherlands).

For staining with antibodies to the vesicular transporters or choline acetyltransferase, mixed populations of nematodes were prepared by freeze cracking (1, 7). Fixation consisted of 2 min in methanol followed by 4 min in acetone. For staining with anti-neurotransmitter antibodies, nematodes were prepared with two different methods. In one method, the freeze-crack protocol was followed except that fixation consisted of 12–24 h in 4% formaldehyde in 0.1 M phosphate buffer. A second method was based on the protocol of Garriga et al. (15). Mixed populations of nematodes were collected and placed in tubes containing 4% formaldehyde in 0.1 M phosphate buffer, pH 7.2. Tubes were placed in a dry-ice ethanol bath for 1 min, then thawed, refrozen, thawed, and placed on a shaker at 4°C for 24 h. After rinsing in phosphate-buffered saline (PBS), fixed nematodes were incubated overnight with shaking at 37°C in a solution containing 2-mercaptoethanol (5% 2-mercaptoethanol, 1% Triton X-100, and 120 mM Tris, pH 7.0). Nematodes were rinsed in PBST (PBS with 0.5% Triton X-100), and then 10–50 μl of nematodes were incubated in 100 μl of collagenase solution (100 units of Sigma type VII collagenase, 1 mM CaCl2, 0.1% Triton X-100, and 100 mM Tris, pH 7.4) on a shaker at 37°C for 2–24 h. Finally, fixed nematodes were rinsed with PBST.

After fixation, nematodes were rinsed in PBS and then blocked with 10% donkey serum in antibody buffer (0.5% Triton X-100, 1 mM EDTA, and 0.1% BSA in PBS with 0.05% sodium azide) for 1 h. Primary antibody incubations (1:50–1:200) were done overnight. After thorough rinsing with antibody buffer, nematodes were incubated in secondary antibodies for 4 h. Unlabeled and Cy3-labeled secondary antibodies were obtained from Jackson ImmunoResearch (West Grove, PA). Oregon green 488 was coupled to secondary antibodies using a kit from Molecular Probes (Eugene, OR). After rinsing, nematodes were mounted in antibleaching medium (14).

Stained nematodes were visualized with a four-laser Leica TCS NT confocal microscope.


Colocalization of VAChT and VMAT in specific neurons.

We cloned the VAChT gene (unc-17) and the VMAT gene (cat-1) in C. elegans and prepared specific antibodies (2, 7). VAChT and VMAT are associated with synaptic vesicles; generally, there is little immunoreactivity in cell somas. As expected, the cellular expression patterns of the two transporters are quite different. However, we have found six neurons that are immunopositive for both proteins: a bilaterally symmetrical pair of cells in the head and four cells in the body (Fig.1). We have identified the four neurons in the body as VC4, VC5, HSNL, and HSNR.

Fig. 1.

Distribution of immunoreactivity of vesicular monoamine transporter (VMAT; red), vesicular acetylcholine transporter (VAChT; green), or both in wild-type and mutantCaenorhabditis elegans. Top row: VMAT and VAChT immunoreactivities imaged (left) or diagramed (right) in the midbody region of a wild-type adult. Immunoreactivity is seen in the ventral nerve cord (VNC), dorsal nerve cord (DNC), and the dorsal and ventral sublateral nerve cords (sub). VMAT and VAChT immunoreactivities are present around the vulva (indicated on the diagram by a blue line and seen in the fluorescent image as a thin line of autofluorescent cuticle). Although most neuronal cell bodies (gray ovals in diagram) are not immunoreactive, the cell bodies of the VC4 and VC5 are visible. Yellow regions in the ventral nerve cord are generally due to red (VMAT) and green (VAChT) positive regions that are very close, but are not overlapping, whereas the yellow region around the vulva is due to colocalization (see higher magnification views below). Second row: close-up of the region around the vulva of a wild-type C. elegans. The cell bodies of VC4 and VC5 are indicated; these cells contain both VAChT and VMAT immunoreactivity. The VNC outside of the region of the vulva contains generally nonoverlapping spots of VAChT or VMAT immunoreactivity. A ventral sublateral nerve cord (sub) that contains VAChT immunoreactivity is indicated. Third row: higher magnification of a portion of the vulva innervated by the VC4/5 and HSNL. Bottom row: the region around the vulva in aunc-104 mutant, in which synaptic vesicles and VAChT and VMAT immunoreactivity are abnormally concentrated in cell bodies. The cell bodies of VC4/5 and HSNL/R are indicated; all 4 neurons contain both VAChT and VMAT immunoreactivity. Several motor neurons that contain only VAChT immunoreactivity are seen in the VNC. An embryo (emb) inside the adult hermaphrodite shows VMAT immunoreactivity in a few developing neurons. All images are maximum projections of confocal series with anterior to the left and dorsal on top. Scale bars are 5 μm.

The class of VC neurons consists of six motor neurons with cell bodies in the ventral nerve cord (47). Although the six VC cells are derived from parallel postembryonic lineages (40), the mature cells show considerable morphological diversity (47,48). The VC4 and VC5 neurons flank the developing vulva and send highly arborized processes to innervate the vulval muscles; the other VCs have less extensive output to these muscles. In addition, all of the VC cells have sparse output to the ventral body muscles and to other motor neurons (47) and contain Phe-Met-Arg-Phe-NH2 (FMRF-amide) immunoreactivity (38). VC4/5 contain weak and variable serotonin immunoreactivity and induced fluorescence (7), but do not contain tryptophan hydroxylase, the synthetic enzyme for serotonin (41). We found that VC4/5 contain significant punctate VAChT immunoreactivity and VMAT immunoreactivity (Fig. 1).

The HSN neurons are bilaterally symmetrical with posterior lateral cell bodies. They receive input from several interneurons and have extensive output to the vulval muscles. In addition, the HSNs have an anterior process that travels up the ventral nerve cord and into the nerve ring, with minor synapses onto motor neurons and interneurons (47,48). The HSNs are necessary for normal egg laying; ablation leads to significant defects in egg laying (6, 17). The HSNs contain strong serotonin immunoreactivity (6) and FMRF-amide immunoreactivity (38).

The HSN somas occasionally stained faintly for both of the vesicular transporters, whereas the major region of innervation by the HSNs, the vulva, stained brightly. However, this output region is shared with the output regions of VC4/5. Therefore, we examined staining inunc-104 mutants, which are deficient for a kinesin-related protein required for the transport of synaptic vesicles from neuronal cell bodies to synapses (16, 32). In unc-104mutants, synaptic vesicles are found in large clusters in cell bodies (16), and a number of synaptic vesicle-associated antigens, such as synaptotagmin, VAChT, and VMAT, are also mislocalized to neuronal cell bodies (2, 7, 31). In unc-104mutants, somas of the HSNs (as well as VC4/5) stained for both VMAT and VAChT (Fig. 1).

We used lineage mutants to eliminate specific cells. Inegl-1 mutants, the HSNs undergo programmed cell death (17). In lin-39 mutants, there are variable defects in the lineages that give rise to the VC cells so that some individuals are missing VC4 and VC5 (4). In both mutants, the output fields of the remaining cells are positive for both VMAT and VAChT (n > 20, Fig. 2). Thus in all four of these neurons, both types of vesicular transporters are expressed and are present in the synaptic output zones.

Fig. 2.

Distribution of VMAT (red) and VAChT (green) or both in lin-39 and egl-1mutants. Top row: the region around the vulva in alin-39 mutant. Due to the (variable) lineage defects oflin-39, VC4 and VC5 are absent in this individual; the remaining immunoreactivity at the vulva is due to HSNR and HSNL (with, perhaps, minor contributions from VC1, VC2, VC3, and VC6). The VNC and a ventral sublateral (sub) are indicated. Middle row: the region around the vulva in an egl-1 mutant. HSNL/R are absent in egl-1 mutants; the remaining immunoreactivity at the vulva is due to innervation by VC4 and VC5, with minor contributions from VC1, VC2, VC3, and VC6. Bottom row: higher power view of the same egl-1 mutant shows immunoreactivity at the vulva arising from the VCs. Top andmiddle rows are maximum projections of confocal series; thebottom row is a single confocal section. Anterior is to the left and dorsal is on top. Scale bars are 5 μm.

Differences between VC4/5 and HSNs.

VMAT immunoreactivity is more intense in the HSNs than in VC4 and VC5, whereas VAChT (and ChAT) immunoreactivity is more intense in VC4 and VC5 than in the HSNs. Overall, the HSNs appear to be more aminergic and less cholinergic than VC4 and VC5.

The HSNs contain the synthetic enzyme for serotonin, tryptophan hydroxylase (41), as well as high levels of serotonin immunoreactivity and serotonin-like induced fluorescence. Therefore, they are likely to use serotonin as their aminergic neurotransmitter. In contrast, VC4/5 do not express detectable levels of tryptophan hydroxylase (41), and they contain very low and variable levels of serotonin immunoreactivity and serotonin-like induced fluorescence (Fig. 3). Antibodies to dopamine, octopamine, tyramine, and histamine (all of which work only poorly in our preparation) did not give a detectable signal in VC4/5. VC4/5 may contain a low level of one of these amines or may use some other aminergic neurotransmitter.

Fig. 3.

Serotonin immunoreactivity in mutant C. elegans. Serotonin immunoreactivity (green) and nuclei (blue) near the vulva in an adult unc-104 mutant. A maximal projection of a confocal series through the right half of an adult shows strong serotonin immunoreactivity in the cell body and process of HSNR. (The out-of-focus HSNL soma is not seen in this partial projection.) Serotonin immunoreactivity is only weakly present in the cell bodies of VC4 and VC5 (arrows) and is undetectable in the processes of these two neurons. The opening of the vulva can be seen because of the autofluorescence of the cuticle in that region. Scale bar is 10 μm.

The serotonin detected in VC4/5 could arise from uptake of serotonin synthesized by other cells. For example, the HSNs should release serotonin near the VC4/5 terminals and cell bodies; this serotonin might be taken up by VC4/5. To test this hypothesis, we examined serotonin immunoreactivity in VC4/5 in mutants that lack HSNs. We found that the faint serotonin immunoreactivity present in VC4/5 in some unc-104 animals (16/46 or 35% cells positive) was not detected in unc-104;egl-1 mutants (1/62 or 2% cells positive). Therefore, the low level of serotonin sometimes present in VC4/5 may normally originate in the HSN cells.

The different cell types show slight differences in the development of expression of these different transporters. The six VCs are born at the end of the first larval stages (40). VAChT (and ChAT) immunoreactivity is generally undetectable until the fourth larval stage. At this time, the vulva develops and the VC4/5s begin to arborize and innervate the vulval muscle (23). VAChT immunoreactivity increases slightly before VMAT immunoreactivity increases. The highest levels of both VAChT and VMAT immunoreactivity are reached as the vulva becomes completely differentiated in the early adult and are maintained throughout adulthood.

The HSNs are born and migrate into the middle of the body during embryogenesis. They begin to extend axons during the second or third larval stage (15). The axons continue to grow and differentiate through the late fourth larval stage; final differentiation is not complete until early adulthood. The HSNs do not show detectable VMAT or VAChT immunoreactivity (nor serotonin immunoreactivity) (15) before the late fourth larval stage. Both vesicular neurotransmitter transporters are detectable in the early adult at the same time and increase in intensity in concert. This coordinate expression contrasts with the more complex timing of expression seen in VC4/5.

Subcellular localization of VAChT and VMAT.

The subcellular patterns of immunoreactivity for VAChT and VMAT are not identical. The slightly different distributions within somas suggest different patterns of synthesis or trafficking. In addition, different regions of the terminal fields sometimes have different relative amounts of the two transporters (Fig. 2). This is true not only in wild-type animals (in which the terminal fields of the VC4/5 and the HSNs overlap) but also in egl-1 mutants in which all of the innervation is derived from the VCs. This suggests that VAChT and VMAT are present in different vesicles and/or different vesicle types, although there may also be some membrane compartments containing both transporters.

Interactions between neurotransmitter systems.

We have been unable to demonstrate any compensatory interactions between the two transmitter systems. Thus the expression of ChAT and VAChT in the HSNs and in VC4/5 (measured by immunostaining) is unchanged in cat-1 null mutants, which lack VMAT. Also, the expression of VMAT in these four cells is unchanged in cha-1or unc-17 mutants, which have decreased cholinergic neurotransmission. The expression of both VAChT and VMAT was unchanged in tph-1, which lacks serotonin (41), andbas-1 mutants, which have reduced levels of serotonin and dopamine (27, 35). Furthermore, neither the HSNs nor the VCs are required to induce the expression of the multitransmitter phenotype in the other cells.


Evidence for multiple transmitter function.

Previous functional and pharmacological studies have suggested that the HSNs use both acetylcholine and serotonin (46). In addition, serotonin and acetylcholine affect egg-laying behavior in different ways (43). Our immunocytochemical results support and extend these conclusions. In addition, although direct physiological analysis of C. elegans vulval muscles is not yet reported, in the nematode Ascaris suum, the vulval muscles respond directly to both acetylcholine and serotonin (34).

If both cell types (HSN and VC) release acetylcholine and possibly serotonin on the same set of muscles, why do decreases in either of the transmitters have only a small effect on egg laying, while loss of the HSNs (but not the VCs) has a very strong effect on egg laying (17, 43)? We hypothesize that there is some element of HSN structure or function not present in VC4/5. This may involve the synaptic output from the HSNs, such as yet another neurotransmitter specifically released by the HSNs. For example, the HSNs and VC4/5 all contain FMRF-amide immunoreactivity, but they may release different neuroactive peptides (29).

Subcellular localization of vesicular transporters.

In mammals, it has been shown that in cell somas and axons, VAChT is localized to the precursors of small synaptic vesicles (45), whereas VMAT2 is preferentially associated with the precursors of large dense core vesicles (25). At synapses, rat VMAT2 can be found in both small clear vesicles and large dense core vesicles (30). In culture, the subcellular localization of these two transporters can be similar or different depending on cell type (26). It thus appears that VAChT and VMAT2 can be present in different vesicles and/or vesicle types, although there may also be vesicles or pools of vesicles containing both transporters. Both large and small vesicles are present inC. elegans neurons (48); it is quite possible that C. elegans VAChT and VMAT are also in different vesicle types.

Dale's principle and multiple neurotransmitters.

As originally formulated, Dale's principle stated that a mature neuron makes use of the same transmitter substance at all of its synapses (5, 8). Although Dale did not specifically address the question of multiple transmitters within a single neuron, over the years Dale's principle acquired a widely accepted corollary that an individual neuron secretes a single neurotransmitter. After the discovery and characterization of neuropeptides, this principle was further limited to apply only to classical neurotransmitters.

In the last fifteen years, there have been numerous reports of more than one classical neurotransmitter localized to the same mature neuron by immunocytochemical means (reviewed in Ref. 22). Reports include colocalization of serotonin and ChAT immunoreactivities (21). Robust physiological and pharmacological evidence for the release of multiple transmitters (GABA and glycine or GABA and ATP) from single vertebrate neurons has recently been reported (18, 20).

We have now presented evidence for the simultaneous presence of two vesicular transporters in identified neurons. Although this result was not unexpected for the HSNs (based on previous physiological and immunocytochemical studies), it provides further evidence for synaptic vesicular release of different classes of classical neurotransmitters from single identified neurons. Because the ratio of the VAChT and VMAT proteins varies in different parts of the terminal fields of each of these cells, it is likely that VAChT and VMAT are present in different vesicles and/or different vesicle types, and perhaps even at different terminal types. We are now examining the ultrastructure of the VC and HSN terminals by electron microscopy to further study this question.


We thank Lee Eiden, Ken Miller, Dennis Frisby, and Becky Eustance Kohn for advice and suggestions. Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center, which is funded by the National Institutes of Health National Center for Research Resources. tph-1(mg280) was a gift of Ji Ying Sze.


  • This research was supported by National Institute of General Medical Sciences Grant GM-38679 and by Oklahoma Center for the Advancement of Science and Technology Grant HN3-023.

  • Address for reprint requests and other correspondence: J. S. Duerr, Program in Molecular and Cell Biology, Oklahoma Medical Research Foundation, 825 NE 13th St., Oklahoma City, OK 73104 (E-mail:Janet-Duerr{at}

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract