Erik M. Jorgensen is at the Department of Biology and the Brain Institute, University of Utah, Salt Lake City, Utah 84112, USA. jorgensen@biology.utah.edu
In C. elegans, dopamine signaling regulates locomotion behavior. Chase and colleagues report that this signaling occurs through extrasynaptic and antagonistically acting receptors coexpressed in motor neurons. These results provide surprising insights into the G-protein pathways mediating this antagonism, with implications for dopamine signaling across species.
In 1986, John White and colleagues published the complete reconstruction of Caenorhabditis elegans nervous system connectivity1. For years C. elegans biologists have pored over this wiring diagram of the worm, hoping that rabbinical study of its pages would reveal how the nervous system generates behavior. In this issue, Chase et al.2 demonstrate that those late nights with a yad and a candle were doomed to failure and could never reveal the motivations or even explain the movement of a worm.
First, the authors show that the neurotransmitter dopamine can have opposite effects on locomotion, instructing the worm to stay or to go, depending on the type of receptor activated. Second, the authors report that the dopamine inputs are not synaptic but rather humoral. Because control of locomotion is nonsynaptic, a complete map of the synaptic connectivity of the worm nervous system will never fully explain its behavior. Third, they find that antagonistic receptor types are coexpressed on the motor neurons that control locomotion, suggesting that the antagonism is battled out within a single cell, rather than between opposing circuits.
Conflicting actions of dopamine have long been known in vertebrates, and Chase et al. are able to demonstrate the effects of antagonistic signaling pathways on behavior in a physiological context. Intriguingly, the G-proteins mediating this antagonism differ from those previously implicated in dopamine signaling. Dopamine receptors can be divided into D1-like and D2-like subclasses. The antagonistic actions of these two classes have been thought to occur through activation of Gs and Gi signaling pathways. However, Chase et al. show that in the nematode, D1 and D2 antagonism appears to be mediated through Gq and Go. This finding, along with recent results in mice, might force a re-evaluation of the dogma in the dopamine signaling field.
Dopamine modulates a worm's response to food. A worm swims actively without food, but slows when it encounters the edge of a bacterial lawn. Mutants lacking dopamine do not slow when they encounter food3, and worms exposed to exogenous dopamine are paralyzed and stop moving on or off food. However, it was unclear where dopamine acts in the network between the sensory neuron and the motor nervous system. Five dopamine receptors in the worm have been identified by molecular criteria: two in the D1 class and three in the D2 class of receptors. Chase et al. now report that mutations in two of these genes affect locomotion.
Mutants null for dop-3, which encodes a D2-like receptor behaved like mutants lacking dopamine synthesis: that is, they did not slow down when they encountered food, and they were largely resistant to the paralyzing effects of exogenous dopamine exposure (Fig. 1). Knocking out dop-1 caused no obvious defects on its own, but it suppressed the defect of a dop-3 mutant: the dop-3 dop-1 double mutant paused when encountering food and was paralyzed by exogenous dopamine just like the wild type. Analogous effects on slowing in response to food and paralysis by exogenous dopamine suggest that exogenous and endogenous dopamine act on the same signaling pathways. Because the dop-3 dop-1 phenotype does not resemble that of a dopamine synthesis mutant, the other three dopamine receptors must make additional contributions to the locomotory response to dopamine. Because mutations affecting all these receptors are not available, further research will be required to fully characterize the multiple actions of dopamine on locomotion. In the meantime, Chase et al. have determined the cellular focus of DOP-1 and DOP-3 receptors.
Figure 1. Dopamine regulation in a cholinergic motor neuron in C. elegans.
(a) In the wild type, dopamine activation of DOP-3 predominates, and the worm stops swimming. (b) In the dop-3 mutant, activation of DOP-1 predominates and the worm continues to swim. (c) dop-1 dop-3 double mutants respond to dopamine and stop swimming. Thus, there must be another dopamine receptor, which we have called DOP-4. It could act in another cell, or in the cholinergic motor neuron as shown.
The cellular site of action of these receptors is complicated by the known synaptic connections formed by neurons that make and release dopamine. Most dopamine neurons are found in the head ganglia of the worm, where the decisions to stay or go are presumably made. These neurons act redundantly to control pausing on food3, suggesting a target in the head for the effects of dopamine on locomotion. However, dop-3 dopamine receptors are not expressed in the command interneurons in the head, but rather in the cholinergic motor neurons of the ventral nerve cord. Chase et al. expressed the DOP-3 receptor in dop-3 mutants specifically in these motor neurons, so that they were the only cells in the animal expressing DOP-3. The mosaic animals were rescued from the dop-3 defect and were paralyzed by exogenously applied dopamine, indicating that dopamine acts on these distant motor neurons to control locomotion. However, because there are no synaptic inputs on the cholinergic motor neurons from any dopamine-expressing neurons, dopamine must function extrasynaptically as a hormone. Although dopamine is unlikely to function as a hormone in the mammalian brain, it does act in a paracrine fashion4.
Like the D1 and D2 receptors in vertebrates, DOP-1 and DOP-3 have antagonistic effects on behavior. This antagonism could occur in the same cells or in different cells in the behavioral circuitry. For the most part dop-1 and dop-3 are expressed in nonoverlapping cells of the head and tail ganglia. However, DOP-1 and DOP-3 receptors are coexpressed in the cholinergic motor neurons of the ventral nerve cord. To show that DOP-1 was antagonizing the DOP-3 receptor in the same cells, Chase et al. expressed DOP-1 in cholinergic motor neurons of a dop-3 dop-1 double mutant, which caused the animals to be resistant to the paralyzing effects of exogenous dopamine (Fig. 1b). Thus, dopamine acts antagonistically on the actions of a single cell.
To define the downstream components of D2-class DOP-3 receptor signaling, the authors screened for mutants that were resistant to the paralyzing effects of dopamine. These screens identified components of the Go pathway known to act in locomotion: Go itself and G5. Moreover, they identified the RGS protein, which antagonizes Gq. Previous studies have characterized the antagonistic pathways for Go and Gq in the nematode (Fig. 2). Activation of the Gq pathway leads to stimulation of synaptic vesicle exocytosis. Go antagonizes Gq function, leading to reduced synaptic transmission5,
6,
7. Thus, if we were to put in all of the pieces of this puzzle where they seem to fit, dopamine would bind D1-like DOP-1 receptors and activate Gq, and also bind D2-like DOP-3 receptors, which would antagonize D1 function through activation of Go.
Figure 2. Model for Gq and Go pathways in the nematode.
Gq stimulates PLC to convert PIP2 into DAG. DAG recruits UNC-13, which stabilizes syntaxin in the open state. Open syntaxin can prime synaptic vesicles for exocytosis by forming a SNARE complex (not shown). The Gq pathway is antagonized by the Go pathway. The D2 dopamine receptor acts either by activating the RGS protein, which inhibits Gq, or by activating Go, which stimulates DAG kinase. DAG kinase depletes DAG, UNC-13 is inactivated, and synaptic strength is weakened. Dotted lines indicate pathways for which the genetic and physical data are incomplete.
The pathway put forward by Chase et al. conflicts with the canonical idea of dopamine function. A large body of work, starting with the classic work of Greengard and colleagues in the early 1970s, established that D1 receptors activate adenylyl cyclase via Gs, whereas D2 receptors inhibit adenylyl cyclase via Gi8. However, recent work in mice suggests that D1- and D2-class receptors may also activate Gq and Go in the mammalian brain9,
10,
11. If true, these findings suggest that our concept of how dopamine acts in the brain must be revised. But is this model true? First, is Gq the major pathway for D1 signaling in the nematode? A direct link has not yet been established. Moreover, the alternative possibility that D1 receptors act via Gs has not been explored. A worm homolog of Gs activates adenylyl cyclase in the motor neurons and could be responsible for D1 signaling12,
13. Second, does dopamine activate Gq signaling in the mammalian brain? It is necessary to assay dopamine responses in mutants lacking Gq function. Unfortunately, Gq family members are redundant in the mouse brain and double mutants are synthetically lethal14. Tissue-specific knockouts will be required to evaluate the role of Gq proteins in dopamine responses in the mouse. Thus, further work is required in both the worm and mouse to sort out the signaling pathways.
Finally, there is a logical problem with having dopamine both stimulate and inhibit a cell; it seems unfair to jerk the cell around like that. However, antagonistic effects on the same cells can provide the cell with more sensitive regulation. First, altering the ratio of receptor expression levels in these competing pathways could result in opposite behavioral responses. Second, if the receptors have very different affinities for dopamine, simply altering dopamine secretion could reverse the response of the cell. Thus, a fixed circuit can respond in very different ways to a stimulus—and that is why religious study of the wiring diagram of the worm will never lead to enlightenment.
s and G
5. Moreover, they identified the RGS protein, which antagonizes G
q and G
