GPA-14, a Gαi subunit mediates dopaminergic behavioral plasticity in C. elegans

Background Precise levels of specific neurotransmitters are required for appropriate neuronal functioning. The neurotransmitter dopamine is implicated in modulating behaviors, such as cognition, reward and memory. In the nematode Caenorhabditis elegans, the release of dopamine during behavioral plasticity is in part modulated through an acid-sensing ion channel expressed in its eight dopaminergic neurons. A D2-like C. elegans dopamine receptor DOP-2 co-expresses along with a Gαi subunit (GPA-14) in the anterior deirid (ADE) pair of dopaminergic neurons. Findings In follow-up experiments to our recently reported in vitro physical interaction between DOP-2 and GPA-14, we have behaviorally characterized worms carrying deletion mutations in gpa-14 and/or dop-2. We found both mutants to display behavioral abnormalities in habituation as well as associative learning, and exogenous supply of dopamine was able to revert the observed behavioral deficits. The behavioral phenotypes of dop-2 and gpa-14 loss-of-function mutants were found to be remarkably similar, and we did not observe any cumulative defects in their double mutants. Conclusion Our results provide genetic and phenotypic support to our earlier in vitro results where we had shown that the DOP-2 dopamine receptor and the GPA-14 Gαi subunit physically interact with each other. Results from behavioral experiments presented here together with our previous in-vitro work suggests that the DOP-2 functions as a dopamine auto-receptor to modulate two types of learning, anterior touch habituation and chemosensory associative conditioning, through a G-protein complex that comprises GPA-14 as its Gα subunit.


Findings
Neural plasticity is dependent upon various neurotransmitters, including the catecholamine dopamine [1,2] and abnormal dopaminergic transmission is associated with memory disorders [3]. Dopamine receptive neurons have dopamine receptors mainly localized to the morphologically plastic dendritic spine regions [4]. In humans, these seven-transmembrane G-protein coupled dopamine receptors are classified into D1-and D2-types. Activated D1-type receptors couple to Gα s and activate adenylyl cyclase, and D2-receptors tend to act antagonistically to D1receptors, mediating the signal transduction through Gα i [5]. Dopamine released into the synaptic cleft by presynaptic neurons interacts with its receptors after which it is either degraded by monoamine oxidase or taken up through a dopamine transporter [6]. The release of dopamine from acidic vesicles is accompanied by increase in H + ion concentrations stimulating presynaptic acidsensing ion channels (ASICs) that are proposed to modulate levels of dopamine in the synaptic cleft [7,8]. Some mammalian D2-receptors are localized to dopamine releasing neurons in a pre-synaptic configuration and thereby act as auto-receptors [9].
Caenorhabditis elegans is an ideal invertebrate model to study genes involved in behavioral plasticity [10][11][12]. In the adult hermaphrodite C. elegans, dopamine is synthesized in eight neurons: two anterior deirid neurons (ADEs), two posterior deirid neurons (PDEs) and four cephalic neurons (CEPs) [13]. Four dopamine receptor genes have been identified in the C. elegans genome: dop-1, dop-2, dop-3 and dop-4. Based on pharmacological properties of their protein products and their sequence profiles, DOP-1 is classified as a D1-type receptor, while DOP-2 and DOP-3 are classified as D2-type receptors and DOP-4 is invertebrate specific [14]. Loss-of-function mutants for dop-1 tend to habituate faster [15,16] and loss-of-function dop-2 mutants display associative learning deficits [8]. DOP-3 and DOP-4 have both been implicated in response to aversive soluble repellents [17,18]. Seventransmembrane receptors such as the dopamine receptors typically transduce their signal through G-proteins. The C. elegans genome encodes for 21 Gα, 2 Gβ and 2 Gγ genes; one particular gene, gpa-14, codes for a Gα i subunit and shows expression overlap with DOP-2 as well as with ASIC-1 in the ADE dopaminergic neurons [8,19,20]. We have recently reported physical interaction between DOP-2 and GPA-14 and that the third intracellular loop of DOP-2 is essential for binding to GPA-14 in vitro [21]. To study the functional significance of the above molecular interaction in the intact organism, results from follow-up behavioral and genetic experiments are presented here.

Hypothesis
Considering the role of dopamine in plasticity and the interaction of DOP-2 with GPA-14, we hypothesized that deletion of gpa-14 will cause behavioral abnormalities similar to dop-2 mutants [8,21]. Towards this end, we report that both dop-2(vs105) and gpa-14(pk347) lossof-function mutants display associative learning deficits as well as faster habituation at remarkably similar rates. Additionally, the phenotype of the gpa-14(pk347);dop-2 (vs105) double mutant is virtually identical to either single-mutant, and exogenous dopamine tends to revert the mutant behavior.

gpa-14 and dop-2 mutants display similar behavioral deficits
We carried out a behavioral profiles for gpa-14(pk347) and dop-2(vs105) mutants obtained through the Caenorhabditis Genetic Center, and cultured on standard nematode growth media with E. coli OP50 at 20°C. Compared to wild type N2 animals, individuals of both strains are phenotypically normal in terms of body size, shape, growth, movement and locomotion and they respond normally to gentle touch, and display normal chemotaxis to both soluble and volatile chemicals. Dopamine has been reported to mediate movement and food sensing behaviors in that when C. elegans encounter food, they move more slowly [2]. Our basal slowing assays did not reveal any significant difference between wild type animals and the deletion mutants (Additional file 1). A modified habituation assay was used in which a forward moving worm was gently touched on the anterior with an eyelash hair causing the worm to move backwards [22]. The mechanical stimulus was repeated with 10 sec intervals until the animal no longer responded to the stimulus. In order to confirm that the novel anterior touch based assay used here was in fact measuring habituation, we initially provided 20 gentle anterior touches (with 10 sec inter-stimulus intervals) to each worm and their response was scored after each stimulus as 1 or 0 (1 = worm moves away, and 0 = no response or worm continues moving in same direction). We observed that increasing number of touches decreased the probability of the worms' response and we did not observe abrupt disruption in their ability to respond ( Figure 1A). The latter would have indicated sensory fatigue. We also plotted the same data in terms of the average point at which worms stopped responding to the anterior touch stimulus ( Figure 1B). Based on assay similarities and substantial neural circuitry overlap of the observed behavior with the extensively studied mechanical tap habituation, we consider that a decrease in response to anterior touch is a form of habituation, although additional tests will allow necessary verification [22,23]. Our subsequent assays recorded the number of mechanical stimuli repeated with 10 second intervals until the response failed as a measure of habituation of individual worms. The results are noteable in that both dop-2(vs105) and gpa-14(pk247) mutants exhibited significantly faster habituation rates compared to wild-type ( Figures 1A and B, 2A).
It has been previously reported that dop-2(vs105) mutants are deficient in chemotaxis based associative learning paradigms [9]. We tested the performance of gpa-14(pk347) in a learning assay in which the chemo-attractant isoamyl alcohol (IAA) was paired with starvation [9]. For conditioning, the animals were exposed to 3 μl of IAA for 90 minutes. IAA (2μl) diluted to 1/100 in ethanol was applied to the gradient spot and ethanol (2μl) was applied to the diluent point, and worms were placed equidistant to the two points. Plates were left undisturbed for one hour, after which the animals were counted to calculate a chemotaxis index for each plate [9,24]. Naive worms showed strong attraction towards to 1:100 dilution of IAA. After conditioning, N2 worms displayed significantly reduced attraction to isoamyl alcohol compared to both gpa-14(pk347) and dop-2(vs105) ( Figure 3A). Additionally, there was no significant difference between the learning capacity of the gpa-14(pk347) or the dop-2(vs105) strains. Similar associative learning results were obtained using paradigms that used non-volatile sodium chloride as unconditioned stimulus (Additional file 2). In control experiments we did not observe any decrease in attraction to isoamyl alcohol when the animals had been previously exposed to isoamyl alcohol in the presence of food/E. coli, confirming that the decrease in response observed upon conditioning in the absence of E. coli is a learned response and not due to adaptation [ Figure 3C].
In order to test whether DOP-2 and GPA-14 exert their influence on the observed adaptive behavior through the same pathway, a gpa-14(pk347);dop-2(vs105) double mutant was generated. This double mutant adapted significantly faster than N2 at rates similar to those observed for gpa-14(pk347) or dop-2(vs105) single mutants (Figure 2A). No cumulative abnormalities were observed in the double mutant indicating a genetic interaction between dop-2 and gpa-14.

Dopamine supplements revert behavioral deficits in mutants
It has been shown that dopamine deficiency results in faster habituation rates while abundance in dopamine has been correlated with slower habituation rates [16,17]. Behavioral deficits attributed to low levels of dopamine have been ameliorated by providing exogenous dopamine [8,15,16]. Upon assaying in the presence of 5mM exogenous dopamine, we were able to bring habituation rates of both the single-mutants [gpa-14(pk347) and dop-2 (vs105)] as well as the double mutant [gpa-14(pk347); dop-2(vs105)] to levels similar to the habituation rate of wild type N2 animals ( Figure 2B). Addition of exogenous dopamine was also able to rescue the conditioned associative learning deficits for the gpa-14 (pk347) and dop-2(vs105) deletions as well as the gpa-14(pk347);dop-2(vs105) double deletion mutants ( Figure 3B).  4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Response Probablity  Figure-1A was plotted as a bar graph to illustrate that the average habituation point of dop-2(vs105) and gpa-14(pk347) deletion mutants was significantly lower than WT animals (p < 0.05, two tailed student's t-test, n = 30 for each strain). Error bars represent SEM values.

Discussion
We provide genetic and phenotypic support to our previously reported in vitro interaction between the DOP-2 dopamine receptor and the GPA-14 Gα i subunit [21]. The gpa-14(pk347) and dop-2(vs105) mutants display faster rates of habituation and diminished associative learning capacities. Faster mechanosensitive habituation has been reported for mutants in dop-1, which codes for a D1-like receptor [15], and crosstalk between AVM touch neurons (which express a D1-like receptor, DOP-1) and ADE dopaminergic neurons (which express the D2-like DOP-2 receptor) has been proposed previously [16]. It will be interesting to explore whether DOP-1 and DOP-2 work as an antagonistically coordinating pair as indicated by the pharmacological properties of D1-and D2-receptor types [5]. Supplementary exogenous dopamine in either of the single mutant or the double mutant used in this study reverted the observed phenotypic abnormalities, indicating that DOP-2 and GPA-14 play an upstream role in regulating the dopaminergic pathway of behavioral plasticity in worms. In addition to its role in learning and memory, dopamine has been reported to regulate  Figure 3 Deletion mutants in gpa-14 and dop-2 display limited associative learning that is rescued by exogenous dopamine. Associative learning chemotaxis assay. Three-day-old synchronized worms were conditioned with 3 μl of isoamyl alcohol on freshly prepared chemotaxis plates and tested with 1:100 isoamyl alcohol in the absence of food, as described previously [8,24]. Plates were left undisturbed for one hour and then kept at −10°C for 3 minutes. Chemotaxis index was calculated by subtracting the number of worms at the diluent sector from the number of worms at the chemical gradient sector and dividing by the total number of worms on the plate. Tests were then repeated in the presence of plates supplemented with 60 μl of 5 mM exogenous dopamine. (A) While naïve strains [wild type, dop-2(vs105) and gpa-14(pk347)] are equally attracted to isoamyl alcohol but after conditioning, wild type worms display significantly greater decrease in attraction towards isoamyl alcohol than dop-2(vs105) and gpa-14(pk347) (p = 0.35 for naive and p < 0.05 for conditioned animals, ANOVA). Additionally, the conditioned response of the gpa-14;dop-2 double mutant is the same as either single mutant (p = 0.46, One-Way ANOVA). (B) When conditioned in the presence of exogenous dopamine, no statistical significance was noted between the wild types and the mutants (p = 0.65 for naives and p = 0.88 for conditioned, One Way ANOVA). (C) As a control, 3-day-old synchronized worms were exposed to 3 μl of isoamyl alcohol in the presence of E. coli for 90 minutes and tested for their attraction to 1:100 isoamyl alcohol. No significant difference in attraction was observed for wild type or either of the mutants [dop-2(vs105) and gpa-14(pk347)] before or after treatment (p = 0.85 for naives and p = 0.89, One Way ANOVA). Bars represent SEM values n ≥ 3 assays for each strain.
food encounter response of C. elegans [2]. According to our results, the deletion of dop-2 or gpa-14 has no effect on their food encounter response. This could imply that, although dopamine activity is reduced in these mutants, its reduction may not be significant enough to affect their food encounter response. Another possibility is that basal slowing may not be as sensitive to dopamine levels as mechanosensation and chemosensensation. Absence of any additional or cumulative abnormalities in the gpa-14(pk347);dop-2(vs105) double mutant provides genetic support towards an interaction between dop-2 and gpa-14 and that their protein products participate in the same molecular pathway during behavior plasticity. Apparently, the observed habituation and associative learning behaviors in both mutants [gpa-14(pk347) and dop-2(vs105)] may give an impression of contradictory phenotypes. While diminished associative learning clearly correlates with poorer neuronal summation, faster habituation does not necessarily draw a parallel with an antagonistic explanation. On the other hand faster habituation may correlate with ignoring a cue earlier than optimum assuming that WT habituation represents an optimal rate. In summary, examining the behavioral profile of dop-2(vs105) and gpa-14(pk347) deletion mutants has revealed that both show similar aberrations in plasticity, and our results with the dop-2;gpa-14 double mutant indicate that GPA-14 and DOP-2 work together in the same pathway. In order to rule out the contribution of background mutations, follow-up experiments using RNAi knockdowns or transgenic rescue of the mutants can provide required validation of the results presented here.
These results are significant in that neuroimaging studies in humans indicate that deficient feedback monitoring in the cortex is associated with learning deficits due to D2 receptor polymorphisms [25,26]. The ASIC and DAT-1 pathways are also known to modulate dopamine release in the dopaminergic neurons including ADEs and have been proposed to work in conjunction with DOP-2 ([8] and Additional file 3). However, the nature of the upstream or parallel role of DOP-2 in the plasticity pathway is not understood at the molecular level. It is conceivable that DOP-2 may act as an auto-receptor in the ADE neurons and upon stimulation it transduces the extracellular dopamine signal through the GPA-14 Gα i subunit [8,21]. Alternately, DOP-2 may exert its influence at the somatodendritic membrane, given the dual localization of mammalian D2 receptors at both cellular compartments of dopaminergic neurons [27]. Further investigations are needed and it will be interesting to identify downstream components of the ASIC, DAT-1 and DOP-2 pathways and their crosstalk in modulating precise levels of neurotransmitter in the synaptic cleft.