Serotonin promotes exploitation in complex environments by accelerating decision-making
- Shachar Iwanir†1,
- Adam S. Brown†2,
- Stanislav Nagy1,
- Dana Najjar3,
- Alexander Kazakov4,
- Kyung Suk Lee5,
- Alon Zaslaver4,
- Erel Levine5 and
- David Biron1, 3Email author
© Iwanir et al. 2016
Received: 6 August 2015
Accepted: 21 January 2016
Published: 4 February 2016
Fast responses can provide a competitive advantage when resources are inhomogeneously distributed. The nematode Caenorhabditis elegans was shown to modulate locomotion on a lawn of bacterial food in serotonin (5-HT)-dependent manners. However, potential roles for serotonergic signaling in responding to food discovery are poorly understood.
We found that 5-HT signaling in C. elegans facilitates efficient exploitation in complex environments by mediating a rapid response upon encountering food. Genetic or cellular manipulations leading to deficient serotonergic signaling resulted in gradual responses and defective exploitation of a patchy foraging landscape. Physiological imaging revealed that the NSM serotonergic neurons responded acutely upon encounter with newly discovered food and were key to rapid responses. In contrast, the onset of responses of ADF serotonergic neurons preceded the physical encounter with the food. The serotonin-gated chloride channel MOD-1 and the ortholog of mammalian 5-HT1 metabotropic serotonin receptors SER-4 acted in synergy to accelerate decision-making. The relevance of responding rapidly was demonstrated in patchy environments, where the absence of 5-HT signaling was detrimental to exploitation.
Our results implicate 5-HT in a novel form of decision-making, demonstrate its fitness consequences, suggest that NSM and ADF act in concert to modulate locomotion in complex environments, and identify the synergistic action of a channel and a metabotropic receptor in accelerating C. elegans decision-making.
KeywordsPatchy environments Resource distribution Foraging systems Decision-making Serotonin (5-HT) C. elegans Serotonin 5-HT Foraging Slowdown Re-feeding
The spatial distribution of resources in natural environments can be non-uniform. Patches of food sources generically result from non-linear interactions between multiple species and inhomogeneous environmental conditions [1–3]. As a result, foraging behavior in patchy environments is extensively studied both theoretically and experimentally in diverse species and habitats [4–10]. Yet, key questions such as the fitness consequences of efficient foraging and the neural basis of these behaviors are not sufficiently understood [11–15].
Food availability can affect locomotion patterns [16–19]. In the absence of food, Caenorhabditis elegans predominantly roam, a behavioral state characterized by fast, directional locomotion. In contrast, on a standard bacterial lawn C. elegans exhibit mostly non-directional dwelling and mean velocities are an order of magnitude lower than off food [18–21]. In addition, several minutes after being placed on food, starved C. elegans exhibit slower motion than well-fed animals. This was termed “enhanced slowdown”. The biogenic amine serotonin (5-hydroxytryptamine; 5-HT) modulates behaviors of C. elegans including locomotion, feeding, and egg laying [22–27] as well as the enhanced slowdown [16, 19, 28].
The dynamics of C. elegans locomotion during an encounter with newly discovered food have not been previously characterized. Under standard laboratory conditions, C. elegans typically forages on a large and homogeneously food-covered landscape. In such an environment responses to newly discovered food are not easily assayed and potential deficiencies may not incur a significant impact. In contrast, on a patchy foraging landscape responding quickly to a newly discovered patch of food may be crucial to efficient exploitation. Delayed reactions could potentially prove as detrimental as a deficiency in navigating to a patch in the first place.
Three neuronal cell types display robust serotonin biosynthesis in the hermaphrodite: the amphid sensory neuron ADF; the pharyngeal neurosecretory-motor neuron NSM; and the hermaphrodite-specific neuron HSN . NSM was implicated in mediating an enhanced slowdown of locomotion on food after a period of starvation [16, 28, 30] and in decision-making during steady-state transitions between roaming and dwelling on food [20, 31]. The sensory neuron ADF has been primarily associated with navigation [32–36] and pharyngeal pumping [33, 37]. However, the specific roles of serotonergic neurons in mediating responses to newly encountered food are not well understood.
Here we show that serotonergic signaling accelerated the slowdown of animals upon encountering food, such that they could abruptly pause at the edge of a bacterial lawn. To address the biological relevance of an abrupt slowdown, we assayed exploration and resource exploitation of animals in patchy environments. Under these conditions, serotonergic signaling afforded a substantial advantage in exploitation. The pharyngeal neurosecretory-motor neuron NSM responded physiologically to the actual encounter and was the primary driver of the abrupt slowdown. In contrast, the onset of activity in the chemosensory serotonergic neuron ADF occurred prior to the encounter with food. Correspondingly, ADF affected locomotion during this time. Finally, we found that a 5-HT-gated chloride channel (MOD-1) and a 5-HT metabotropic receptor (SER-4) act together to accelerate C. elegans decision-making.
The slowdown of C. elegans upon encountering food is abrupt
The most striking feature of the observed dynamics was an abrupt slowdown upon encountering the edge of the bacterial lawn: a nearly complete halt reached within 3 sec (τ = 1.73 ± 0.12 sec, Fig. 1b), where the timescale τ was obtained from a fit to a single exponential decay. After a brief pause, wild-type animals typically advanced about one body length into the large bacterial lawn during a 1-min period. During the following 10–15 min, animals predominantly dwelled before reaching steady-state transition rates between roaming (fast, mostly forward locomotion) and dwelling [20, 38]. The transiently augmented dwelling behavior was consistent with previous reports of enhanced slowing [16, 18].
Immediately preceding the encounter with the edge of the bacterial lawn, the roaming animals exhibited a mild but significant deceleration. During the 30-sec period preceding the encounter, the mean deceleration was 0.66 ± 0.04 μm/sec2, P <0.01 (Fig. 1b and Additional file 1: Movie S1). Despite being small in magnitude, this mild preemptive slowdown was observed in >70 % of the individual wild-type encounters (Additional file 2: Figure S1D). The overall structure of the response to encountering food was conserved across a range of experimental conditions. In particular, characteristics of motion during the encounter and the following 15 min were only weakly affected by or independent of the time the animals spent off food prior to the assay (Additional file 2: Figure S1). Thus, a robust and eminent feature of the response to newly discovered food was a rapid slowdown upon encounter with the edge of the lawn.
Serotonergic signaling accelerates slowdown upon re-feeding
Animals in which vesicle exocytosis was blocked by the tetanus toxin light chain gene [40, 41] in all of the serotonergic neurons (Ptph-1::TeTx) were phenotypically similar to tph-1 mutants (τ = 24.90 ± 0.56 sec, Fig. 2 and Additional file 3: Figure S2). Mutants in the mod-5 gene, encoding a transporter required for 5-HT re-uptake, exhibited a lower mean speed both off and on food (see also [16, 42]) and an abrupt slowdown upon encounter (τ = 3.24 ± 0.14 sec, Fig. 2 and Additional file 3: Figure S2). To test the impact of differences in baseline locomotion, we confirmed that the changes in relative velocity maintain the phenotypic trends observed in the wild-type, mutant, functional ablation, and rescue strains assayed (Additional file 4: Figure S3). Taken together, these results indicate that 5-HT signaling accelerates slowdown upon encountering food and acts on a timescale of 2–3 sec.
An abrupt slowdown supports exploitation of a small patch of food
Under what circumstances could an abrupt slowdown be advantageous? A gradual modulation of locomotion can suffice when the spatial scale of the feeding area is much greater than the length scale associated with the slowdown. In contrast, in environments where food is spatially confined to small patches, an abrupt slowdown may increase exploitation [3, 14, 43]. The ratio between the typical speed of a roaming animal and the timescale of slowdown determined the relevant length scale, which was comparable to one body length.
To test the utility of abrupt slowdown, we placed animals on plates with patches of bacterial food that were 0.5–1 mm in diameter, termed “micro-patches”. This assay was repeated using different concentrations of food and various patch arrangements (see Methods). We measured the fraction of time spent on a patch of food during the 5 min following the first encounter with this patch. Responses were scored visually and divided into four categories, where the first two resulted in favorable exploitation of the food patch (see Methods).
We assayed tph-1 mutants and Ptph-1::TeTx transgenics in the presence of micro-patches. In both cases the animals spent significantly less time on a patch of food during the 5 min following the first encounter with this patch, as compared to wild-type (Additional file 5: Figure S4A and Additional file 6: Movie S2). The chances of exploitation increased with the concentration of food on the patch in a dose-dependent manner in wild-type animals, tph-1 mutants, and Ptph-1::TeTx transgenics. Serotonin-deficient animals were completely unable to stop at patches of low food concentrations and performed poorly as compared to wild-type under all of the conditions tested. Thus, a deficiency in either the synthesis or the release of 5-HT reduced the efficiency of translating an encounter with a small patch of food to successful exploitation in a concentration-dependent manner.
Serotonergic signaling promotes exploitation in a complex environment
To examine the role of 5-HT signaling in exploiting a complex environment, we assayed wild-type animals, tph-1 mutants, and quintuple mutants lacking the function of the five serotonin-binding receptors that have been identified in C. elegans [25, 26]. Two simple measures of the efficiency of food exploitation are: 1) comparing the numbers of patches that were exploited with those merely encountered; and 2) comparing the overall fraction of time spent at patch locations to the time spent exploiting. Both measures were quantified during the first 2 hours of the experiment, when the environment was mostly intact, and throughout the experiment, as patches were progressively consumed (Fig. 3b, c and Additional file 5: Figure S4B–D).
While wild-type animals exploited as many patches as they encountered, tph-1 mutants were unable to exploit the vast majority of encountered patches. Moreover, these mutants spent significantly less time than wild-type at patch locations, whether they were regarded as exploiting or not. Thus, even if animals fed efficiently without slowing, for example by frequent turns resulting in multiple passes through a patch, serotonergic signaling remained advantageous.
Next, we measured the fractions of the area of a food patch that was encountered or exploited (Fig. 3d and Additional file 5: Figure S4C). When all encountered patches (throughout the experiment) were considered, wild-type animals covered a significantly larger fraction per patch than tph-1 mutants (Fig. 3d, i). When we limited the analysis to exploited patches, the difference between wild-type and tph-1 mutants was smaller but remained significant (Fig. 3d, ii), indicating that tph-1 mutants were capable of exploiting patches once exploitation was initiated.
Possible contributors to deficient exploitation could have been a smaller coverage of the patch per event due to enhanced patch leaving , an exploration defect resulting in fewer encounters, or the defect in initiating exploitation upon encounter demonstrated above. As depicted in Fig. 3d, iii, tph-1 mutants exploited a similar area per event, suggesting that elevated patch leaving did not appreciably contribute to reduced exploitation. The time spent between encountering patches was also similar between tph-1 and wild-type animals (Fig. 3d, iv), indicating that the mutants did not exhibit exploration/navigation defects. Combined, our various assays suggested that a major contributor to the tph-1 exploitation defect was their inability to slow down abruptly, and that this defect was not sensitive to experimental details. Moreover, mutants lacking the function of all five C. elegans serotonin receptors were significantly defective as compared to wild-type and similar to tph-1 mutants (Fig. 3b–d). Thus, serotonergic signaling conferred an exploitation advantage in a complex environment by accelerating decision-making.
The serotonergic neurons ADF and NSM respond prior to and upon encountering food, respectively
Spontaneous physiological activity of NSM on food is sporadic  and chemical cues can evoke slow changes in NSM cytosolic calcium . In ADF, familiar food and salt can evoke calcium transients [33, 46]. However, the activity of serotonergic neurons was not examined during encounters with newly discovered food. To characterize the physiological activity that promotes the rapid responses we observed, we monitored calcium levels in freely behaving animals using the genetically encoded indicator GCaMP3.0 . Our behavioral assays indicated that egl-1 mutants, lacking the HSN serotonergic neuron type [48, 49], were indistinguishable from wild-type (data not shown). Therefore, to study cellular mechanisms, we focused on the physiological activity in NSM and ADF prior to, during, and after re-feeding encounters.
To further address the activation of ADF, we assayed its physiological activity in osm-6 mutants. The osm-6 gene encodes an intraflagellar transport component. Its absence results in defects to the ultrastructure of sensory cilia and, as a result, in chemosensory and mechanosensory defects. Expression of osm-6 was reported in many ciliated neurons, including ADF, but not in NSM [50, 51]. In these assays, we observed a prominent activation of ADF prior to the encounter in control animals and no activation of ADF at any time in osm-6 mutants. Correspondingly, the preemptive slowdown was significant in the control group and absent in the mutants (Fig. 4b). In contrast, NSM responses were not prominently affected by the loss of OSM-6 function (Fig 4c). Thus, chemosensation by ADF contributes to the preemptive slowdown.
Physiological responses of ADF to bacteria supernatant were detected in a microfluidic device . To demonstrate ADF responses to diffusible chemicals from food in freely behaving animals on a standard assay plate, we recorded from animals placed on the face of an agar pad opposing the side supporting the bacterial lawn (see Methods). The minimal possible distance between animals thus placed and the bacterial lawn was 2–3 mm, as determined by the thickness of the pad. Although this assay does not reproduce the geometry of a gradient near the edge of a lawn, it effectively serves to isolate the diffusible chemicals from the bacteria that produce them. A mock encounter was defined as the time when the position of the animal was directly above the edge of the lawn, and activity during the 20 sec surrounding the mock encounter was compared to baseline (Additional file 11: Figure S5A). Consistent with chemosensation, we observed activation of ADF, but not of NSM, in the vicinity of the mock encounter in the reverse patch assay.
The abrupt slowdown was completed in 2–3 sec, while the calcium transients in NSM and ADF lasted for tens of seconds. However, tph-1 mutants and Ptph-1::TeTx transgenics slowed down over >30 sec (Additional file 11: Figure S5B). Moreover, NSM calcium transients predicted dwelling states for several minutes on bacterial lawns  and the average time spent by a wild-type animal on a small patch was similarly long (see Fig. 3b, c). Taken together, our measurements demonstrated that ADF and NSM responded physiologically to the proximity and encounter with bacterial food, respectively, and continued to affect locomotion for tens of seconds after the encounter.
Is NSM required for exploiting food resources or for generic motor control? To address this we monitored the activity of NSM when the animal paused for a reason other than food encounters, i.e. in order to avoid an obstacle. We positioned a standard platinum pick near the anterior of a forward moving animal, such that the animal gently touched it due to its own motion. These collisions typically evoked abrupt pauses followed by brief reversals. In this case, NSM neurons were not detectably activated (Fig. 4a, left inset). Thus, NSM is not part of a general motor circuit that controls slowing down. Rather, it underlies a specific subset of behaviors including responses that facilitate exploitation of localized food resources.
ADF and NSM play complementary roles in mediating the dynamics of slowdown
We observed that ADF, but not NSM, was activated before the animal reached the edge of the bacterial lawn. The timing of ADF activation resembled the timing of the mild preemptive slowdown prior to the encounter (Fig. 1b and 4). Both the abrupt slowdown upon encounter and the mild preemptive slowdown were affected by changes in 5-HT signaling (Fig. 2 and Additional file 12: Figure S6), suggesting that 5-HT may regulate distinct aspects of the overall response. To explore the individual contributions of different serotonergic neurons, we assayed transgenics expressing TeTx in either NSM or ADF.
To assay the sufficiency of ADF and NSM, we expressed tph-1 specifically in NSM or ADF on a tph-1(mg280) mutant background. Rescuing the function of TPH-1 in NSM mostly restored the abrupt slowdown upon encounter (τ = 2.49 ± 0.16 sec) but not the preemptive slowdown. ADF-specific rescues resulted in an opposite phenotype (Fig. 5e, f and Additional file 12: Figure S6). High expression levels typical of such experiments may have contributed to the observed phenotypes by allowing one neuronal type to partially compensate for the absence of the other. Cell-specific tph-1 deletion transgenics  exhibited similar, albeit weaker, phenotypes (Fig. 5g, h, Additional file 4: Figure S3, Additional file 5: Figure S4D, and Additional file 12: Figure S6). Incomplete penetrance of Cre expression or inactivation of the floxed tph-1 gene could not be ruled out as a possible contributor to these weaker phenotypes. Taken together, these results indicate that NSM and ADF act through 5-HT signaling to mediate slowdown prior to and during encounters.
Optogenetic activation of serotonergic neurons induces a rapid slowdown off food
Optogenetic activation of serotonergic neurons was previously shown to affect locomotion on food  and off food . To assess the sufficiency of 5-HT signaling in mediating abrupt slowdown, we optogenetically activated the serotonergic neurons of freely moving starved animals using the light-activated cation channel channelrhodopsin (ChR2). Innate responses of C. elegans to blue light were avoided by performing the assays on a light-insensitive (lite-1) mutant background [53, 54].
To further investigate the role of 5-HT, we assayed mod-5 mutants. Exposure to blue light reduced the locomotor activity of mod-5; lite-1; Ex[Ptph-1::ChR2] animals by 70 %. Locomotion of mod-5 mutants declined throughout the 30-sec light stimulus (τ = 5.1 ± 1.1 sec) and returned to baseline post-stimulation slower than wild-type (τ = 37.3 ± 2.1 sec, P <0.01, see Fig. 6). This suggested that the uptake of endogenous 5-HT also contributed to shaping the modulation of locomotion on short timescales. Combined with our earlier results, these observations associate the dynamics of 5-HT signaling and of slowdown.
Efficient exploitation on a patchy landscape is mediated through the synergistic action of the 5-HT-gated channel MOD-1 and the 5-HT metabotropic receptor SER-4
Five serotonin-binding receptors have been identified in C. elegans [25, 26]. Of these, the Go-protein coupled receptor SER-4 (an ortholog of mammalian 5-HT1 receptors) and the serotonin-gated chloride channel MOD-1 have been implicated in slowing locomotion in response to exogenous 5-HT . However, endogenous 5-HT acted appreciably only through MOD-1 to increase steady-state dwelling at the expense of exploration on bacterial food .
To test the efficiency of foraging of these mutants in a complex environment we assayed them for 1.5–2.5 hours on a 5 × 5 patch arena. In this assay, the quintuple receptor mutants and ser-4; mod-1 double mutants exhibited severe defects in the number of exploited patches (as compared to the number of patches encountered) and in the total time spent on exploitation during the assay. The phenotypes of these two strains were identical and resembled that of tph-1 mutants (Fig. 7c, d). In contrast, mod-1 and ser-4 single mutants exhibited a milder defect and no significant defect, respectively. Combined, these results implicate MOD-1 and SER-4 in acting synergistically to mediate rapid decision-making.
Foraging and exploitation in patchy environments is a well-studied neuroethological paradigm for the enhancement of fitness afforded by decision-making circuits. Similarities in the challenges facing different species yield repeated algorithmic aspects of their solutions and conserved underlying mechanisms. As a result, decision-making is associated with fundamental functions of nervous systems [31, 43, 56]. Recently, the activity of dorsal raphe neurons was recorded in behaving mice that predicted and responded to rewards and punishment. The results indicated that serotonergic neurons signal about reward and punishment on short (hundreds of milliseconds) and long (minutes) timescales .
We have shown that in C. elegans 5-HT signaling accelerates the attenuation of locomotion upon encountering newly discovered food. We address the importance of this acceleration by showing that deficiencies in serotonergic signaling reduced the efficiency of foraging in a patchy environment. Previously, NSM was the major neuron implicated in affecting locomotion on food. It was shown to release serotonin extrasynaptically to diffuse in the vicinity of the nerve ring and activate a distributed circuit of target cells (as opposed to post-synaptic partners) . Here, the serotonergic neuron types NSM and ADF were found to play complementary roles, where one neuron could partially compensate for a deficiency in another. The combined action of the 5-HT-activated chloride channel MOD-1 and the 5-HT receptor SER-4 was required for the accelerated response to discovering food.
Serotonergic modulation of locomotion has been examined under various sets of conditions [16, 18, 20, 22, 23, 28–30, 42, 58–61]. In contrast, fitness consequences of these effects on locomotion have not been studied in detail [31, 38]. Exogenous 5-HT robustly attenuates locomotion [22, 23] and stimulation of endogenous 5-HT release can cause slowdown [20, 28]. In addition, serotonergic signaling accelerates food-dependent aversive responses [30, 60, 61].
Prior to the isolation of tph-1 mutants, Sawin et al. assayed locomotion by manually scoring body bends 5 min after placing animals on a bacterial lawn. Under these conditions, they showed that the slowdown of locomotion on food is enhanced in starved nematodes as compared to well-fed controls . Enhanced slowdown was more severe on a mod-5 mutant background and required the action of the MOD-1 receptor [28, 42, 58, 59]. The same assay was used to show that SER-4 contributed to post-starvation slowing on food  and to assay tph-1 mutants [28, 62]. Both studies found that the slowdown of tph-1 mutants was only mildly reduced as compared to wild-type and Gürel et al. did not report on the basal slowing of well-fed animals in their hands. These assays resulted in highly active, well-fed animals on food (compare basal slowing in [16, 19, 62] to steady-state locomotion on food in [20, 21]) suggesting that the transfer to the assay lawn may have been a factor in determining behavioral outcomes (as noted in ).
Unless specified otherwise, animals in the current study were deprived of food prior to the assay. However, in our hands starvation did not appreciably alter the preemptive slowdown, the abrupt slowdown upon encounter, nor the velocity 15 min post-encounter (Additional file 2: Figure S1). Moreover, 5-HT signaling deficiencies affected the approach to the edge of a lawn and the abruptness of responding to the encounter, but not center of mass motion 10–20 min thereafter. Additional file 2: Figure S1 and Additional file 3: Figure S2 demonstrate that the 5-min interval used in [28, 42, 58, 59, 62] is intermediate between the abrupt response to the encounter and the relaxation time to steady-state locomotion on food. Since the transient adaptation of locomotion upon encounter is not completed within 5 min, we suggest a different interpretation of the existing collective body of data: upon encountering food, rather than determining the target locomotion activity on food, serotonergic signaling accelerates the response to the sudden change in the external environment. The resulting abrupt response contributes to the efficiency of foraging in complex landscapes.
Flavell et al. assayed well-fed animals for 90 min on food, where wild-type animals predominantly dwell (as opposed to predominantly roaming off food). They found that the serotonergic NSM neurons are active during dwelling periods and inhibit roaming through 5-HT secretion. Moreover, they found that tph-1 mutants roamed 3–5 fold more than wild-type on a bacterial lawn. Thus, serotonergic signaling can affect steady-state transitions between dwelling and roaming . We did not observe large differences between the center of mass velocities of wild-type animals and tph-1 mutants 5–20 min post-encounter (Additional file 3: Figure S2). However, since the focus of this work was on the encounter with food, the duration of our assays was limited as compared to the 90-min recordings with no prior starvation described in .
The NSM neuron is physically separated from the locomotor circuit by the basal membrane . Based on its location and structure, NSM was hypothesized to play a role in reporting the availability of resources . Nevertheless, it was implicated in modulating locomotion in [20, 28] and in the current study. Moreover, our data reveals prominent activation of NSM upon encountering bacteria. This raises the question of whether food signals onto NSM to drive locomotion through sensory neurons or whether NSM may directly monitor the state of the pharynx. Indirect activation of NSM was demonstrated in . However, a detailed anatomical dissection of NSM identified a putative sensory process that may monitor pharyngeal activity . We observed NSM activation in osm-6 mutants, where cilium structure (and hence chemosensory function) is severely disrupted (Fig. 4c). Although our data cannot support a definitive answer, it is consistent with the notion that NSM monitors the state of the pharynx.
Significant physiological activity in ADF was detectable prior to the encounter with the food, while NSM responded detectibly upon encounter. Thus, ADF may act to refine the dynamics of the slowdown response. This can conceivably occur in several, not mutually exclusive, manners: 1) early release of 5-HT may prime the downstream neural circuit such that an abrupt slowdown can take place before NSM is fully activated; 2) the mild ADF-dependent early slowdown may positively affect detection (e.g. by affecting the angle of attack or providing more time for sampling the environment in the vicinity of food); and 3) responses to potentially competing stimuli may be suppressed in the vicinity of a patch of food. In addition, 5-HT released by ADF may contribute to the robustness of the response if it can be scavenged and used by mod-5-expressing neurons (NSM and/or others ).
Rather than identifying distinctive roles of individual serotonergic neuronal types, our combined physiological and behavioral data suggest partial functional redundancy. The phenotypes of wild-type animals, tph-1 mutants, and Ptph-1::TeTx transgenics were consistent across conditions. Disrupting the function of ADF or NSM resulted in partial defects as compared to tph-1 mutants. These findings suggest that different neurons capable of synthesis and/or uptake of 5-HT may act redundantly to modulate locomotion upon encountering food. Jafari et al., using the assays described in , reported that synaptic release of 5-HT from uptake neurons (that do not synthesize 5-HT themselves) was not required for locomotor responses to food deprivation. Rather, these neurons prevent exaggerated responses by scavenging extrasynaptic 5-HT . A detailed analysis of neurons that promote exploitation in complex environments would require the dissection of 5-HT uptake, storage, and synaptic release in subsets of serotonergic and 5-HT uptake cells.
The SER-4 receptor and the MOD-1 chloride channel were shown to affect locomotion through GOA-1 activation in target cells or hyperpolarizing them, respectively [20, 28, 65, 66]. Head sensory- and inter-neurons were implicated as sites of action of GOA-1 and MOD-1 for inhibiting locomotion [20, 28]. It was further proposed that additional identified effectors act to facilitate the MOD-1 or SER-4 signaling pathways  or to oppose them . While our data implicate SER-4 and MOD-1 in the abrupt response, the phenotypes of the mod-1; ser-4 double and the quintuple receptor mutants fall short of recapitulating in full the deficits of tph-1 mutants and Ptph-1::TeTx transgenics. Thus, our data suggests that an effect of a serotonin noncanonical receptor, metabolite, or precursor is yet to be accounted for.
When C. elegans navigate on shallow chemical or thermal gradients, typical responses to external cues include gradual changes in the curvature of their trajectories [67, 68] or modulation of the rate of infrequent sharp turns [17, 69, 70]. As a result, sensory stimuli can be integrated over timescales of tens of seconds, allowing the animal to filter out perceived abrupt (noisy) stimuli. In contrast, in a patchy environment an abrupt change in the availability of food is information of the utmost importance . Correspondingly, behavioral responses must be rapid—analogous to a filter with a high cut-off frequency. The typical scales of the problem are expected to determine the required abruptness.
Typical roaming velocities off food are in the 100–200 μm/sec range. In our hands, patches as small as 400–500 μm in diameter merited exploitation. An order of magnitude estimate yields a timescale of slowdown of 2–3 sec, in agreement with our experimental observations and with the 2 Hz frequency of exploratory head motion . Additional factors, such as concentration and quality, can affect the detailed responses to discovering food. However, abruptness remains advantageous under a range of conditions. Taken together, the robust activation of serotonergic neurons, the contribution of 5-HT re-uptake to proper locomotion dynamics, and the importance of this pathway to efficient foraging suggest that responding to newly discovered food is a key role of serotonergic signaling in C. elegans.
This study suggests that a key role of serotonergic signaling in C. elegans is to accelerate decision-making and promote efficient exploitation of resources in complex environments. It primarily implicates the serotonergic neuron type NSM, the serotonin-gated chloride channel MOD-1, and the ortholog of mammalian 5-HT1 metabotropic serotonin receptors SER-4 in mediating this process. In addition, it demonstrates how different cells can use a common modulator to affect locomotion in complementary manners.
Behavioral assays: large bacterial lawn
To assay re-feeding on a large bacterial lawn, 72–96-hour-old adults were washed twice in a drop of M9 buffer and transferred to unseeded NGM plates for 1–2 hours. Animals were assayed on 35 mm NGM plates seeded with a 25 μl drop of 5x concentrated overnight culture of OP50 bacteria. The bacteria were spread in an oval shape of an area 100–150 mm2 on one side of the plate and assay plates were incubated overnight at 37 °C and kept at 4 °C prior to the assay. In each assay, 8–10 animals were transferred to M9 droplets on the empty side of the assay plate and started to crawl once the droplets were absorbed. Unless stated otherwise, animals that reached the bacterial lawn in less than 150 sec (<20 % of the total number of animals) were not scored in order to minimize the effect of the transfer on the data. Images were captured using a cooled CCD camera (CoolSNAP HQ2, Photometrics, Tucson, AZ, USA), an Olympus SZX16 stereomicroscope equipped with an SDF PLAPO 1XPF Objective (Olympus America Inc., Center Valley, PA, USA), and Micro-Manager Open Source Microscopy Software. Imaging was performed at 5 frames per sec, using a magnification of 0.7x and 4 × 4 binning. Under these conditions, the area of a single animal was 40–45 pixels. The centers of mass of the animal bodies were tracked using custom MATLAB scripts (MathWorks Inc., Natick, MA, USA). Center of mass velocities were computed using a temporal resolution of 1 sec. Each genotype was assayed at least on three different days and control animals were recorded each day. In our hands, day-to-day variability was typically small. To assess whether the preemptive slowdown was statistically significant, a linear function was fitted to the speeds of individual animals during the 50 sec prior to encountering food using the MATLAB Curve Fitting Toolbox. Post-encounter decay constants (τ) were based on a single exponential fit of the 0–150-sec data. The goodness of fit (R2) was higher than 0.9 in all cases. Experiments were performed on at least three independent days and day-to-day inconsistencies were not identified.
Behavioral assays: micro-patches
For Additional file 5: Figure S4A, the 1x concentration was defined as OD600 = 1.5. Patches were arranged in concentric rings, 20 animals were placed on the assay plate, and only first encounters with a single patch were scored. Four ordinal categories were scored by visual inspection: 1) a full stop at the edge of the patch; 2) a noticeable slowdown of locomotion at the edge without a full stop; 3) a delayed slowdown of locomotion; and 4) the absence of a noticeable response. The former two response types resulted in feeding on the patch of food and the latter two did not.
For long-term assays in a complex environment, patches contained OP50 bacteria at a concentration of OD600 = 2.5. Assay plates were freshly seeded prior to the experiment. Individual food patches were seeded by gently touching preloaded thin gel loading tips to a standard agar plate, typically resulting in a patch diameter of 0.8–1 mm. Unless stated otherwise, the bacteria mix was supplemented with 10x kanamycin to prevent growth. Prolonged experiments (5–10 hours, Fig. 3) were performed on a 7 × 7 square lattice. Shorter experiments (2.5 hours, Fig. 7) were performed on a 5 × 5 hexagonal lattice. In both cases the distance between nearest neighbors was 2.5 mm. Food deprivation was performed as described above.
A single animal per plate was imaged at 5 frames per sec using a 5MP scientific CCD camera (Prosilica GC2450, Allied Vision Technologies, Stadtroda, Germany) at magnifications of 0.33–0.65x. A 3D printed square frame (18 mm inner and 24 mm outer diameter) lined with copper tape was used to contain the animal in the field of view. Custom LabVIEW (National Instruments Inc., Austin, TX, USA) and MATLAB scripts were used for image acquisition and data analysis.
Images were analyzed using the posture-based method as described previously . Patch outlines were traced manually for each experiment, and frames where the nose-tip of the worm was within 5 % of the body length from a patch were labeled “encounter”. If, during an encounter, the mean velocity of the animal was at the lowest 15th percentile of measured roaming velocities then the frames of the encounter were labeled “exploitation”. Brief (exploratory or accidental) excursions out of the patch, in which the distance between the edge and the nose-tip was no greater than 20 % of the body length, were consolidated. Thus, motion on the order of 100 μm beyond the patch edge did not fractionate a continuous event. Patch coverage was quantified by determining the area occupied during each exploitation event. Experiments were performed on at least three independent days and day-to-day inconsistencies were not identified.
Calcium imaging was performed on freely behaving animals on large bacterial lawn, 60 mm diameter, assay plates. Here, 50 μl overnight bacterial culture was seeded in a ring and incubated overnight at 37 °C. The transgenic strains assayed expressed the calcium indicator GCaMP3.0 in the target neuron on a lite-1 mutant background [53, 54] (see Additional file 14: Table S1). Food deprivation was performed as described above, several animals were initially placed at the center of the ring, and a single animal approaching the inner diameter of the ring was continuously imaged. To image NSM during evoked pauses (Fig. 4, left inset), a standard platinum pick was manually placed near the anterior of a forward moving animal such that it gently touched it due to its own motion.
The analysis of neuronal calcium transients was performed similarly to the previously described procedure in . Data shown in Fig. 4a and Additional file 11: Figure S5 was acquired using the setup described above for large bacterial lawn assays. Imaging was performed at 5 frames per sec at a magnification of 10x for up to 25 min. In our hands, bacterial lawns had clearly visible sharp edges. First-day adults were transferred to assay plate and allowed several minutes to equilibrate. Animals were tracked manually until they were approximately 15 body lengths from the edge of the food patch, at which point the assay plate remained in place. For reverse patch imaging (Additional file 11: Figure S5A), assay plates were prepared with 2 ml agar; the resulting thickness was 2–3 mm. Prior to the assay, the agar was carefully flipped such that the animals crawled on the agar surface opposite to the lawn. The minimal possible distance of the animal from the bacteria was thus 2–3 mm.
For data shown in Fig. 4b, c, imaging was acquired at a frame rate of 11 frames per sec. Images were captured using a cooled EMCCD camera (Evolve 512, Photometrics), an Olympus 83x inverted microscope equipped with a UPLSAPO 10x objective (Olympus America Inc.), and Micro-Manager Open Source Microscopy Software. Animals were tracked automatically using a custom worm tracker (coded in MATLAB).
Image analysis was performed using custom MATLAB scripts, which identified the fluorescent neuron (the only bright particle in the head), measured the background fluorescence in its vicinity, and calculated the background subtracted mean fluorescence intensity of the identified particle. Velocities in Fig. 4b were calculated based on the displacement of the imaged cell. Experiments were performed on three independent days and day-to-day inconsistencies were not identified.
Optogenetic assays were performed in 1.1 × 1.1 mm microfluidic chambers in which the animals could move freely. Blue light illumination (λ = 475 ± 15 nm) was supplied by a Luxeon Star 7-LED assembly (Luxeon Star LEDs, Brantford, ON, Canada) with a diffused optic array driven by a 700 mA FlexBlock driver (LEDdynamics, Inc., VT, USA). The LED assembly was mounted to the scopes approximately 7 cm from the sample location. Light intensity at the location of the animals was set to 0.8 mW/mm2. Transgenic animals expressing the light-gated ion channel channelrhodopsin (ChR2) in all serotonergic neurons on a lite-1 mutant background (see Additional file 14: Table S1) were illuminated for 30 sec in the presence of 1 mM ATR. Control animals were assayed in the absence of ChR2 and presence of ATR (Fig. 6). The level of activity of the animals was assayed using the frame subtraction method as previously described . In order to obtain the timescales of slowdown and recovery, as described in the text, a single exponential function was fitted to each dataset using the MATLAB curve fitting toolbox. The goodness of fit (R2) was higher than 0.9 in all cases except for the recovery of the faster recovering lite-1; Ex[Ptph-1::ChR2] line (shown in orange), where R2 = 0.73. Experiments were performed on three independent days and day-to-day inconsistencies were not identified.
Data analysis was performed using custom MATLAB scripts. Individual statistical tests are detailed in the corresponding figure legends.
Some strains were provided by the Caenorhabditis Genetics Center (CGC), which is funded by National Institutes of Health (NIH) Office of Research Infrastructure Programs (P40 OD010440). Additional strains were kindly provided by the Horvitz Laboratory at the Massachusetts Institute of Technology, the Komuniecki laboratory at the University of Toledo, the Ringstad laboratory at the New York University School of Medicine, and the Bargmann laboratory at Rockefeller University. We thank Vera Hapiak and Richard Komuniecki for useful discussions. This work was supported by the Burroughs Wellcome Fund Career Awards at the Scientific Interface (DB), the Searle Scholars Program (DB), the National Science Foundation (IOS 1256989, DB), I-core (AZ), and the European Research Council (ERC Starting Grant 336803, AZ).
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- McCann KS, Rasmussen JB, Umbanhowar J. The dynamics of spatially coupled food webs. Ecol Lett. 2005;8:513–23.View ArticlePubMedGoogle Scholar
- Durham WM, Climent E, Barry M, De Lillo F, Boffetta G, Cencini M, et al. Turbulence drives microscale patches of motile phytoplankton. Nat Commun. 2013;4:2148.View ArticlePubMedGoogle Scholar
- Abu Baker MA, Brown JS. Foraging in space and time structure an African small mammal community. Oecologia. 2014;175:521–35.View ArticlePubMedGoogle Scholar
- Driessen G, Bernstein C. Patch departure mechanisms and optimal host exploitation in an insect parasitoid. J Anim Ecol. 1999;68:445–59.View ArticleGoogle Scholar
- Davidson DL, Morris DW. Density-dependent foraging effort of Deer Mice (Peromyscus maniculatus). Funct Ecol. 2001;15:575–83.View ArticleGoogle Scholar
- Cresswell JE, Osborne JL. The effect of patch size and separation on bumblebee foraging in oilseed rape: implications for gene flow. J Appl Ecol. 2004;41:539–46.View ArticleGoogle Scholar
- Sims DW, Southall EJ, Humphries NE, Hays GC. Scaling laws of marine predator search behaviour. Nature. 2008;451:1098–102.View ArticlePubMedGoogle Scholar
- Zhang Z, Zhan X, Yan L, Li M, Hu J, Wei F. What determines selection and abandonment of a foraging patch by wild giant pandas (Ailuropoda melanoleuca) in winter? Environ Sci. 2009;16:79–84.Google Scholar
- Birk MA, White JW. Experimental determination of the spatial scale of a prey patch from the predator’s perspective. Oecologia. 2014;174:723–9.View ArticlePubMedGoogle Scholar
- Shaffer CA. Spatial foraging in free ranging bearded sakis: traveling salesmen or Lévy Walkers? Am J Primatol. 2014;76:472–84.View ArticlePubMedGoogle Scholar
- Lemon WC. Fitness consequences of foraging behaviour in the zebra finch. Nature. 1991;352:153–5.View ArticleGoogle Scholar
- Vergassola M, Villermaux E, Shraiman BI. “Infotaxis” as a strategy for searching without gradients. Nature. 2007;445:406–9.View ArticlePubMedGoogle Scholar
- Stephens GJ, Johnson-Kerner B, Bialek W, Ryu WS. Dimensionality and dynamics in the behavior of C. elegans. PLoS Comput Biol. 2008;4:e1000028.PubMed CentralView ArticlePubMedGoogle Scholar
- Hayden BY, Pearson JM, Platt ML. Neuronal basis of sequential foraging decisions in a patchy environment. Nat Neurosci. 2011;14:933–9.PubMed CentralView ArticlePubMedGoogle Scholar
- Kvitsiani D, Ranade S, Hangya B, Taniguchi H, Huang JZ, Kepecs A. Distinct behavioural and network correlates of two interneuron types in prefrontal cortex. Nature. 2013;498:363–6.PubMed CentralView ArticlePubMedGoogle Scholar
- Sawin ER, Ranganathan R, Horvitz HR. C. elegans locomotory rate is modulated by the environment through a dopaminergic pathway and by experience through a serotonergic pathway. Neuron. 2000;26:619–31.View ArticlePubMedGoogle Scholar
- Gray JM, Hill JJ, Bargmann CI. A circuit for navigation in Caenorhabditis elegans. Proc Natl Acad Sci U S A. 2005;102:3184–91.PubMed CentralView ArticlePubMedGoogle Scholar
- Ben Arous J, Laffont S, Chatenay D. Molecular and sensory basis of a food related two-state behavior in C. elegans. PLoS One. 2009;4:e7584.PubMed CentralView ArticlePubMedGoogle Scholar
- Omura DT, Clark DA, Samuel ADT, Horvitz HR. Dopamine signaling is essential for precise rates of locomotion by C. elegans. PLoS One. 2012;7:e38649.PubMed CentralView ArticlePubMedGoogle Scholar
- Flavell SW, Pokala N, Macosko EZ, Albrecht DR, Larsch J, Bargmann CI. Serotonin and the neuropeptide PDF initiate and extend opposing behavioral states in C. elegans. Cell. 2013;154:1023–35.PubMed CentralView ArticlePubMedGoogle Scholar
- Gallagher T, Bjorness T, Greene R, You Y-J, Avery L. The geometry of locomotive behavioral states in C. elegans. PLoS One. 2013;8:e59865.PubMed CentralView ArticlePubMedGoogle Scholar
- Horvitz HR, Chalfie M, Trent C, Sulston JE, Evans PD. Serotonin and octopamine in the nematode Caenorhabditis elegans. Science. 1982;216:1012–4.View ArticlePubMedGoogle Scholar
- Ségalat L, Elkes DA, Kaplan JM. Modulation of serotonin-controlled behaviors by Go in Caenorhabditis elegans. Science. 1995;267:1648–51.View ArticlePubMedGoogle Scholar
- Srinivasan S, Sadegh L, Elle IC, Christensen AGL, Faergeman NJ, Ashrafi K. Serotonin regulates C. elegans fat and feeding through independent molecular mechanisms. Cell Metab. 2008;7:533–44.PubMed CentralView ArticlePubMedGoogle Scholar
- Chase DL, Koelle MR. Biogenic amine neurotransmitters in C. elegans. WormBook. 2007; Feb 20:1–15.
- Hapiak VM, Hobson RJ, Hughes L, Smith K, Harris G, Condon C, et al. Dual excitatory and inhibitory serotonergic inputs modulate egg laying in Caenorhabditis elegans. Genetics. 2009;181:153–63.PubMed CentralView ArticlePubMedGoogle Scholar
- Song B, Avery L. Serotonin activates overall feeding by activating two separate neural pathways in Caenorhabditis elegans. J Neurosci. 2012;32:1920–31.PubMed CentralView ArticlePubMedGoogle Scholar
- Gürel G, Gustafson MA, Pepper JS, Horvitz HR, Koelle MR. Receptors and other signaling proteins required for serotonin control of locomotion in Caenorhabditis elegans. Genetics. 2012;192:1359–71.PubMed CentralView ArticlePubMedGoogle Scholar
- Jafari G, Xie Y, Kullyev A, Liang B, Sze JY. Regulation of extrasynaptic 5-HT by serotonin reuptake transporter function in 5-HT-absorbing neurons underscores adaptation behavior in Caenorhabditis elegans. J Neurosci. 2011;31:8948–57.View ArticlePubMedGoogle Scholar
- Harris G, Korchnak A, Summers P, Hapiak V, Law WJ, Stein AM, et al. Dissecting the serotonergic food signal stimulating sensory-mediated aversive behavior in C. elegans. PLoS One. 2011;6:e21897.PubMed CentralView ArticlePubMedGoogle Scholar
- Yapici N, Zimmer M, Domingos AI. Cellular and molecular basis of decision-making. EMBO Rep. 2014;15:1023–35.PubMed CentralView ArticlePubMedGoogle Scholar
- Pocock R, Hobert O. Hypoxia activates a latent circuit for processing gustatory information in C. elegans. Nat Neurosci. 2010;13:610–4.PubMed CentralView ArticlePubMedGoogle Scholar
- Song B, Faumont S, Lockery S, Avery L. Recognition of familiar food activates feeding via an endocrine serotonin signal. Elife. 2013;2:e00329.PubMed CentralPubMedGoogle Scholar
- Bargmann CI, Horvitz HR. Chemosensory neurons with overlapping functions direct chemotaxis to multiple chemicals in C. elegans. Neuron. 1991;7:729–42.View ArticlePubMedGoogle Scholar
- Chang AJ, Chronis N, Karow DS, Marletta MA, Bargmann CI. A distributed chemosensory circuit for oxygen preference in C. elegans. PLoS Biol. 2006;4:e274.PubMed CentralView ArticlePubMedGoogle Scholar
- Hukema RK, Rademakers S, Dekkers MPJ, Burghoorn J, Jansen G. Antagonistic sensory cues generate gustatory plasticity in Caenorhabditis elegans. EMBO J. 2006;25:312–22.PubMed CentralView ArticlePubMedGoogle Scholar
- Cunningham KA, Hua Z, Srinivasan S, Liu J, Lee BH, Edwards RH, et al. AMP-activated kinase links serotonergic signaling to glutamate release for regulation of feeding behavior in C. elegans. Cell Metab. 2012;16:113–21.PubMed CentralView ArticlePubMedGoogle Scholar
- Shtonda BB, Avery L. Dietary choice behavior in Caenorhabditis elegans. J Exp Biol. 2006;209:89–102.PubMed CentralView ArticlePubMedGoogle Scholar
- Sze JY, Victor M, Loer C, Shi Y, Ruvkun G. Food and metabolic signalling defects in a Caenorhabditis elegans serotonin-synthesis mutant. Nature. 2000;403:560–4.View ArticlePubMedGoogle Scholar
- Schiavo G, Benfenati F, Poulain B, Rossetto O, Polverino de Laureto P, DasGupta BR, et al. Tetanus and botulinum-B neurotoxins block neurotransmitter release by proteolytic cleavage of synaptobrevin. Nature. 1992;359:832–5.View ArticlePubMedGoogle Scholar
- Macosko EZ, Pokala N, Feinberg EH, Chalasani SH, Butcher RA, Clardy J, et al. A hub-and-spoke circuit drives pheromone attraction and social behaviour in C. elegans. Nature. 2009;458:1171–5.PubMed CentralView ArticlePubMedGoogle Scholar
- Ranganathan R, Sawin ER, Trent C, Horvitz HR. Mutations in the Caenorhabditis elegans serotonin reuptake transporter MOD-5 reveal serotonin-dependent and -independent activities of fluoxetine. J Neurosci. 2000;21:5871–84.Google Scholar
- Adams GK, Watson KK, Pearson J, Platt ML. Neuroethology of decision-making. Curr Opin Neurobiol. 2012;22:982–9.PubMed CentralView ArticlePubMedGoogle Scholar
- Bendesky A, Tsunozaki M, Rockman MV, Kruglyak L, Bargmann CI. Catecholamine receptor polymorphisms affect decision-making in C. elegans. Nature. 2011;472:313–8.PubMed CentralView ArticlePubMedGoogle Scholar
- Li Z, Li Y, Yi Y, Huang W, Yang S, Niu W, et al. Dissecting a central flip-flop circuit that integrates contradictory sensory cues in C. elegans feeding regulation. Nat Commun. 2012;3:776.View ArticlePubMedGoogle Scholar
- Thiele TR, Faumont S, Lockery SR. The neural network for chemotaxis to tastants in Caenorhabditis elegans is specialized for temporal differentiation. J Neurosci. 2009;29:11904–11.PubMed CentralView ArticlePubMedGoogle Scholar
- Tian L, Hires SA, Mao T, Huber D, Chiappe ME, Chalasani SH, et al. Imaging neural activity in worms, flies and mice with improved GCaMP calcium indicators. Nat Methods. 2009;6:875–81.PubMed CentralView ArticlePubMedGoogle Scholar
- Trent C, Tsuing N, Horvitz HR. Egg-laying defective mutants of the nematode Caenorhabditis elegans. Genetics. 1983;104:619–47.PubMed CentralPubMedGoogle Scholar
- Roy SH, Tobin DV, Memar N, Beltz E, Holmen J, Clayton JE, et al. A complex regulatory network coordinating cell cycles during C. elegans development is revealed by a genome-wide RNAi screen. G3 (Bethesda). 2014;4:795–804.PubMed CentralView ArticlePubMedGoogle Scholar
- Perkins LA, Hedgecock EM, Thomson JN, Culotti JG. Mutant sensory cilia in the nematode Caenorhabditis elegans. Dev Biol. 1986;117:456–87.View ArticlePubMedGoogle Scholar
- Collet J, Spike CA, Lundquist EA, Shaw JE, Herman RK. Analysis of osm-6, a gene that affects sensory cilium structure and sensory neuron function in Caenorhabditis elegans. Genetics. 1998;148:187–200.PubMed CentralPubMedGoogle Scholar
- Zaslaver A, Liani I, Shtangel O, Ginzburg S, Yee L, Sternberg PW. Hierarchical sparse coding in the sensory system of Caenorhabditis elegans. Proc Natl Acad Sci. 2015;112:201423656.View ArticleGoogle Scholar
- Edwards SL, Charlie NK, Milfort MC, Brown BS, Gravlin CN, Knecht JE, et al. A novel molecular solution for ultraviolet light detection in Caenorhabditis elegans. PLoS Biol. 2008;6:e198.PubMed CentralView ArticlePubMedGoogle Scholar
- Ward A, Liu J, Feng Z, Xu XZS. Light-sensitive neurons and channels mediate phototaxis in C. elegans. Nat Neurosci. 2008;11:916–22.PubMed CentralView ArticlePubMedGoogle Scholar
- Nagy S, Raizen DM, Biron D. Measurements of behavioral quiescence in Caenorhabditis elegans. Methods. 2014;68:500–7.PubMed CentralView ArticlePubMedGoogle Scholar
- Pearson JM, Watson KK, Platt ML. Decision making: the neuroethological turn. Neuron. 2014;82:950–65.PubMed CentralView ArticlePubMedGoogle Scholar
- Cohen JY, Amoroso MW, Uchida N. Serotonergic neurons signal reward and punishment on multiple timescales. Elife. 2015;4:e06346.Google Scholar
- Ranganathan R, Cannon SC, Horvitz HR. MOD-1 is a serotonin-gated chloride channel that modulates locomotory behaviour in C. elegans. Nature. 2000;408:470–5.View ArticlePubMedGoogle Scholar
- Dernovici S, Starc T, Dent JA, Ribeiro P. The serotonin receptor SER-1 (5HT2ce) contributes to the regulation of locomotion in Caenorhabditis elegans. Dev Neurobiol. 2007;67:189–204.View ArticlePubMedGoogle Scholar
- Harris GP, Hapiak VM, Wragg RT, Miller SB, Hughes LJ, Hobson RJ, et al. Three distinct amine receptors operating at different levels within the locomotory circuit are each essential for the serotonergic modulation of chemosensation in Caenorhabditis elegans. J Neurosci. 2009;29:1446–56.PubMed CentralView ArticlePubMedGoogle Scholar
- Hapiak V, Summers P, Ortega A, Law WJ, Stein A, Komuniecki R. Neuropeptides amplify and focus the monoaminergic inhibition of nociception in Caenorhabditis elegans. J Neurosci. 2013;33:14107–16.PubMed CentralView ArticlePubMedGoogle Scholar
- Omura TD, Horvitz HR. C. elegans integrates food, stress, and hunger signals to coordinate motor activity. In: PhD Thesis. Cambridge, MA: Massachusetts Institute of Technology; 2008.Google Scholar
- White JG, Southgate E, Thomson JN, Brenner S. The structure of the nervous system of the nematode Caenorhabditis elegans. Philos Trans R Soc Lond B Biol Sci. 1986;314:1–340.View ArticlePubMedGoogle Scholar
- Axäng C, Rauthan M, Hall DH, Pilon M. Developmental genetics of the C. elegans pharyngeal neurons NSML and NSMR. BMC Dev Biol. 2008;8:38.PubMed CentralView ArticlePubMedGoogle Scholar
- Lochrie MA, Mendel JE, Sternberg PW, Simon MI. Homologous and unique G protein alpha subunits in the nematode Caenorhabditis elegans. Cell Regul. 1991;2:135–54.PubMed CentralPubMedGoogle Scholar
- Olde B, McCombie WR. Molecular cloning and functional expression of a serotonin receptor from Caenorhabditis elegans. J Mol Neurosci. 1997;8:53–62.View ArticlePubMedGoogle Scholar
- Luo L, Clark DA, Biron D, Mahadevan L, Samuel ADT. Sensorimotor control during isothermal tracking in Caenorhabditis elegans. J Exp Biol. 2006;209(Pt 23):4652–62.View ArticlePubMedGoogle Scholar
- Iino Y, Yoshida K. Parallel use of two behavioral mechanisms for chemotaxis in Caenorhabditis elegans. J Neurosci. 2009;29:5370–80.View ArticlePubMedGoogle Scholar
- Pierce-Shimomura JT, Morse TM, Lockery SR. The fundamental role of pirouettes in Caenorhabditis elegans chemotaxis. J Neurosci. 1999;19:9557–69.PubMedGoogle Scholar
- Ryu W, Samuel A. Thermotaxis in Caenorhabditis elegans analyzed by measuring responses to defined thermal stimuli. J Neurosci. 2002;22:5727–33.PubMedGoogle Scholar
- Chittka L, Skorupski P, Raine NE. Speed-accuracy tradeoffs in animal decision making. Trends Ecol Evol. 2009;24:400–7.View ArticlePubMedGoogle Scholar
- Alkema MJ, Hunter-Ensor M, Ringstad N, Horvitz HR. Tyramine Functions independently of octopamine in the Caenorhabditis elegans nervous system. Neuron. 2005;46:247–60.View ArticlePubMedGoogle Scholar
- Brenner S. The genetics of Caenorhabditis elegans. Genetics. 1974;77:71–94.PubMed CentralPubMedGoogle Scholar
- Nagy S, Wright C, Tramm N, Labello N, Burov S, Biron D. A longitudinal study of Caenorhabditis elegans larvae reveals a novel locomotion switch, regulated by Gαs signaling. Elife. 2013;2:e00782.PubMed CentralView ArticlePubMedGoogle Scholar
- Sanders J, Nagy S, Fetterman G, Wright C, Treinin M, Biron D. The Caenorhabditis elegans interneuron ALA is (also) a high-threshold mechanosensor. BMC Neurosci. 2013;14:156.PubMed CentralView ArticlePubMedGoogle Scholar