Dopamine Signaling Is Essential for Precise Rates of Locomotion by C. elegans

Dopamine is an important neuromodulator in both vertebrates and invertebrates. We have found that reduced dopamine signaling can cause a distinct abnormality in the behavior of the nematode C. elegans, which has only eight dopaminergic neurons. Using an automated particle-tracking system for the analysis of C. elegans locomotion, we observed that individual wild-type animals made small adjustments to their speed to maintain constant rates of locomotion. By contrast, individual mutant animals defective in the synthesis of dopamine made larger adjustments to their speeds, resulting in large fluctuations in their rates of locomotion. Mutants defective in dopamine signaling also frequently exhibited both abnormally high and abnormally low average speeds. The ability to make small adjustments to speed was restored to these mutants by treatment with dopamine. These behaviors depended on the D2-like dopamine receptor DOP-3 and the G-protein subunit GOA-1. We suggest that C. elegans and other animals, including humans, might share mechanisms by which dopamine restricts motor activity levels and coordinates movement.


Introduction
Dopamine (DA) signaling plays a key role in many aspects of mammalian nervous system function, including learning, motivation, and motor control [1,2]. Abnormal DA function has been associated with a variety of disorders, including attention-deficit hyperactivity disorder (ADHD), addiction and Parkinson's disease [1,3,4]. Despite advances in the understanding of the genetics and pathophysiology of neurodegeneration in Parkinson's disease and the successful development of several dopamine-based therapies, the molecular mechanisms by which impaired DA signaling leads to movement disorders in Parkinson's disease are poorly understood. The numerous roles of DA and complexity of the human nervous system add to the challenge of identifying a specific role for DA in movement disorders.
The nervous system of the nematode Caenorhabditis elegans contains 302 neurons of known connectivity and many evolutionarily conserved neurotransmitters and neuromodulators, including DA [5,6,7]. These attributes and the existence of dopaminedeficient mutants make C. elegans an attractive organism to use to study the molecular and cellular consequences of impaired DA signaling. C. elegans locomotion consists of coordinated dorsalventral muscle contractions that propel the nematode through its environment in a sinusoidal fashion [8]. C. elegans detects temperature, salts, soluble and volatile chemicals, oxygen, mechanical stimuli and other environmental cues that modulate its motor program [9,10,11,12,13,14,15,16,17]. DA in C. elegans controls habituation to mechanical stimuli [18], foraging [19] and transitions between crawling and swimming behavior [20]. Mutants lacking DA or the D2-like receptor DOP-3 move faster than wild-type animals on a bacterial lawn, indicating that DA plays a role in slowing C. elegans in response to food [16,21].
In studying the response of dopamine-deficient C. elegans mutants to food, we observed that these mutants exhibited highly erratic rates of locomotion. To further characterize this unusual locomotor behavior, we used a video-tracking system that allows the efficient collection of large amounts of high-resolution locomotion data. Our studies revealed a general role for DA in C. elegans in restricting the range of average and instantaneous speeds of the animal. This function of DA is behaviorally and genetically distinct from the role of DA in food-dependent slowing.

An automated tracking system enables the analysis of new aspects of C. elegans locomotor behavior
We used a locomotion tracking system that simultaneously determines the positions of multiple animals on an assay plate at one-second intervals and assembles these position recordings into tracks ( Figure 1A) -see materials and methods. Each track contains movement data for an individual animal. By preserving the integrity of each track, this system allows the determination of single-track parameters, such as instantaneous speed, average speed, variance of speed, instantaneous acceleration, duration of accelerations, and magnitude of accelerations. The instantaneous speed recordings of a population of animals can be represented by histograms that show the fraction of time animals move at each speed. These frequency distributions can be overlaid to compare how populations of animals move when exposed to different experimental conditions ( Figure 1B).
To ensure that data collected by our automated tracking system were consistent with our direct observations of C. elegans locomotion, we examined the average speed of well fed and food-deprived animals on a bacterial food source as described by Sawin et al. [16]. On a bacterial lawn, well-fed animals moved at 76%, and food deprived animals moved at 32%, of their locomotion rate in the absence of bacteria ( Figure 1C). These measurements are comparable to those obtained by manually counting the number of body bends an animal makes during a 20 second interval [16]. Using this manual assay, we observed that on a bacterial lawn well-fed animals moved at 79%, and fooddeprived animals moved at 30%, of their locomotion rate in the absence of bacteria ( Figure 1D).
Fluctuations in the speeds of individual animals correlate with average locomotion rates independently of feeding state We examined the magnitude of fluctuations in the locomotion rates of individual animals by calculating the standard deviations of the intra-track speed measurements. We also calculated the coefficient of variation, which is the standard deviation of intratrack speed measurements divided by the average speed of the animal. The coefficient of variation reflects the amount of intratrack speed variation relative to the average speed of the animal.
Although the standard deviation of intra-track speed measurements changed significantly as the average speed of the animal changed, there was not a statistically significant difference between measurements of well-fed and food-deprived animals. Therefore, fluctuations in speed were not specifically affected by whether the animal had been food-deprived ( Figure 1E). As the average speed of animals increased, the instantaneous speed measurements of the animal fluctuated less relative to average speed ( Figure 1F). Since the variances of speeds of individual animals correlated with average locomotion rates independently of feeding state, we pooled our data for well-fed and food-deprived animals on bacteria to examine locomotion variability over the largest range of average locomotion rates. Observations of wild-type movement collected on multiple days over many months resulted in nearly identical measurements of intra-track speed fluctuations (data not shown). DA signaling is required to restrict the range of speeds at which an individual animal travels The gene cat-2 encodes tyrosine hydroxylase [22], which catalyzes the conversion of tyrosine to L-DOPA, the biosynthetic precursor of DA [23]. cat-2 mutants have greatly reduced levels of DA [7,22]. We isolated a new allele of cat-2(n4547) that contains a large deletion in the cat-2 open reading frame and presumably represents a null allele of this gene ( Figure 2). This allele is easy to identify by PCR and allowed us to backcross and create new strains by directly following the molecular lesion.
Using the automated locomotion tracking system we discovered that the instantaneous speeds of cat-2 mutants were much more broadly distributed than those of wild-type animals in both the well-fed and food-deprived conditions ( Figure 3A). To further define the nature of this abnormality, we examined the average speeds of well-fed and food-deprived individual animals and found that, in addition to cat-2 mutants being faster [16], the average speeds of individual cat-2 mutants were also more broadly distributed compared to the average speeds of individual wildtype animals ( Figure 3B). We then examined the fluctuations in the speeds of individual animals by comparing the intra-track speed measurements of wild-type and cat-2 animals. For individual animals moving at the same average speeds, cat-2 individuals exhibited a greater range of intra-track speeds than did wild-type individuals. Overlaying multiple wild-type and cat-2 speed traces with comparable average speeds illustrates that the continuous track speeds of individual cat-2 animals fluctuate more than those of wild-type animals ( Figure 3C).
To quantify this difference, we determined the standard deviations of the intra-track speeds among cat-2 tracks. Since the magnitude of speed fluctuations of wild-type animals varied by average speed, we first grouped animals by average speed and then compared the speed fluctuations of animals moving at the same average speeds. While comparing the speed fluctuations of an entire group of animals could be biased by the average speed of the group, this method allowed us to compare strains with greater numbers of faster or slower moving animals without such bias. We found that the speed fluctuations of cat-2 tracks were much larger than those of wild-type tracks ( Figure 4A). The average speeds of wild-type and cat-2 animals did not change significantly over the assay period, precluding the possibility that the greater standard  deviation of speeds for cat-2 mutant animals might be caused by a significant increase or decrease in average speed over the period of the assay (data not shown). Differences in speed fluctuations were seen with both our newly isolated allele and the canonical cat-2(e1112) allele ( Figure 5B). Mutations in tph-1 and tdc-1, which perturb the synthesis of serotonin and tyramine/octopamine, respectively, did not cause increased fluctuations in speed (Data not shown). These data demonstrate that DA is required to restrict the range of speeds at which an individual animal travels. In other words, DA is required for the precision with which a normal animal maintains its speed.

Dopamine-deficient mutants make unusually large adjustments to their speed
We next examined wild-type and cat-2 acceleration and found that cat-2 mutants exhibited greater total acceleration than did wild-type animals ( Figure 4B). This greater total acceleration was a consequence of the fact that cat-2 mutants achieve greater peak acceleration and accelerate for slightly longer periods of time ( Figure 4CD). Thus, DA functions to restrict the magnitude of the accelerations and decelerations made by an individual, which explains why dopamine-deficient cat-2 individuals exhibited a greater range of locomotion rates than individual wild-type animals.
To determine whether cat-2 mutants exhibited greater speed fluctuations on bacteria because they exhibit a higher average speed, we suppressed the average speed of cat-2 mutants with a second mutation. Mutation of the serotonin reuptake transporter gene mod-5 greatly reduces the average speed of animals on bacteria [24] and suppressed the abnormally high locomotion rate of cat-2 mutants on bacteria ( Figure 5A). Compared to those of wild-type animals, the speed fluctuations of mod-5(n3314) mutants were not strongly reduced, and mutations in mod-5 did not suppress the large fluctuations in speed associated with reduced DA signaling in cat-2 mutants ( Figure 5B). That mod-5 suppressed the cat-2 increased speed on bacteria without also suppressing the increased speed fluctuations demonstrates that speed on bacteria and fluctuations in speed are aspects of C. elegans behavior that are independently modulated by DA signaling.

Treatment with exogenous dopamine restores normal locomotion to mutants defective in dopamine synthesis
Treatment of cat-2 mutant animals with exogenous DA restored their ability to make small adjustments to their speed and to restrict the range of speeds at which they travel ( Figure 6D-F). Thus, cat-2 mutants retain the ability to receive and transduce DA signals, indicating that reduced levels of DA in the adult worm are responsible for the behavioral defects associated with cat-2 mutation. That exogenous DA rescued the excess speed fluctua-tions of cat-2 mutants suggests that DA acts humorally in maintaining a precise speed. Exogenous DA did not rescue these defects in dop-3 and goa-1 mutants ( Figure 6D-F). This finding is consistent with the proposed roles of these genes in the postsynaptic reception and transduction of DA signals, respectively ( Figure 6D-F) [21], although goa-1 also functions in other signaling pathways [30,31] and does not only act directly downstream of dop-3 in modulating speed.
A wild-type copy of the dop-3 gene can restore normal locomotion to dop-3(vs106) mutants A transgene containing a genomic copy of the dop-3 gene suppressed the large speed fluctuations of dop-3(vs106) mutants ( Figure 7B). A dop-3 cDNA construct with identical upstream and downstream sequences did not have rescuing activity, indicating that a full genomic copy of dop-3 is necessary to rescue dop-3(vs106) defects and that sequences outside of the dop-3 open reading frame do not have rescuing activity ( Figure 7B). Previous studies concluded that dop-3 is expressed in the GABAergic and cholinergic motor neurons of the ventral cord as well as in head and tail neurons [21]. Additional studies concluded that unc-47 is expressed only in GABAergic neurons [32], unc-17 is expressed only in cholinergic neurons [33], tax-4 is expressed in a subset of sensory neurons [34], and gcy-5 is expressed in a single sensory neuron [35]. We replaced the dop-3 promoter sequences in our genomic construct with unc-17, unc-47, tax-4, and gcy-5 promoter sequences in four separate constructs and found that all four suppressed the large speed fluctuations of dop-3(vs106) mutant animals. The unc-17 heterologous promoter construct and dop-3 genomic construct were both able to suppress speed fluctuations to below wild-type levels ( Figure 7B).
These observations indicate that the full genomic open reading frame, including introns and exons, are necessary and sufficient to confer rescuing activity to our DNA constructs with all of the promoter constructs we tested.

Discussion
Using an automated locomotion tracking system we discovered that DA is required to make small adjustments to speed to maintain a constant rate of locomotion. The D2-like DA receptor DOP-3 and the G ao protein GOA-1 were essential for the reception and transduction of a DA signal that acts this processes. Treatment of dopamine-deficient mutants with exogenous DA completely restored the ability to make small adjustments to speed. mod-5 mutations modulated food-dependent slowing but did not modulate fluctuations in speed. While the D1-like DA receptor dop-1 plays a role in food-dependent slowing [21], dop-1 did not appear to function in maintaining precise rates of speed. Similarly, dop-2, dop-4, and a number of other biogenic amine receptors we tested did not function in maintaining precise rates of speed. These findings indicate that DA functions independently in fooddependent slowing and maintaining precise rates of locomotion.

Dopamine might act in monitoring the animal's locomotor activity to precisely maintain a steady speed
The only dopaminergic neurons in C. elegans are the two ADE, two PDE and four CEP sensory neurons [7]. These neurons have been proposed to function in the mechanosensory detection of a bacterial lawn and in the proprioceptive detection of deflections in the body of the animal while it is moving [16,36,37]. The ADE and PDE neurons also appear important for transitions from swimming to crawling behavior, a process that appears dependent on mechanosensation [20]. We suggest that the PDE neurons, the processes of which run the length of the animal, might become stimulated and release DA when the animal is moving.
Since exogeneous DA restored the ability of dop-3 mutants to make normal adjustments to their speed, we conclude that the amount of DA released does not signal the speed of the animal. Rather we propose that DA release increases the efficiency by which animals integrate information about their current speed. Efficient monitoring of current speed would allow animals to maintain a more precise rate of locomotion specifically when the animal is moving. Impaired detection of speed would lead to a reduced ability to detect movement that is too rapid or too slow, resulting in greater variability in both average speeds and instantaneous speeds. Such a deficit would also result in large acceleration and deceleration events, since an acceleration or deceleration event would terminate only when the animal has accelerated or slowed to an acceptable speed, the detection of which would be impaired.

Dopamine is required for normal levels of locomotor activity in both C. elegans and humans
We found that in comparison to wild-type C. elegans, dopaminedeficient mutants more frequently exhibited periods of both low and high rates of locomotion and also had a reduced ability to make small adjustments to speed. Parkinson's disease is similarly characterized by seemingly opposite symptoms that include extended periods of abnormally high (rigidity) and low (bradykinesia) motor activity; in addition, Parkinson's patients have a reduced ability to make fine adjustments to movements (uneven gait, postural instability) [38].
We determined that the metabotropic D2 DA receptor homologue DOP-3 was required for the fine coordination of C. elegans locomotion. In humans, metabotropic DA receptors are the targets of agonists (bromocriptine, apomorphine, pramipexole, ropinirole, rotigotine) that constitute a family of Parkinson's therapies [39,40]. In addition, administration of the DA precursor L-DOPA (levodopa) is the most effective treatment for Parkinson's symptoms [41,42]. We discovered that pre-incubation of animals with exogenous DA restored normal locomotion to C. elegans mutants with reduced DA levels. Based on these data and our belief that pre-incubation with DA results in high levels of the neurotransmitter throughout the animal, we conclude that acute DA release is not required to restore normal locomotion to dopamine-deficient mutants.
We have shown that impaired DA signaling can cause a range of defects in locomotor coordination by C. elegans. We note that like C. elegans cat-2 mutants, human patients with Parkinson's disease display a constellation of movement disorders that include both greatly reduced and abnormally high levels of motor activity as well as impaired motor coordination and that these clinical features can be rescued by treatment with the DA precursor L-DOPA [43,44]. We suggest that perhaps some aspects of Parkinson's disease involve mechanisms similar to those responsible for the behavioral abnormalities of cat-2 mutants. Although the anatomical dopaminergic circuitry in humans and worms is quite different, it is possible that further studies of dopaminedependent movement disorders of C. elegans might identify molecular and cellular mechanisms that are disrupted in humans with Parkinson's disease.

Locomotion Assay
Locomotion assays were conducted according to Sawin et al. [16], except that in most cases, locomotion was recorded by an automated locomotion tracking system. Assay plates containing bacteria were made by spreading a drop of HB101 E. coli into the shape of a ring using the smooth bottom of a small glass test tube. Assay plates were incubated at 37uC overnight; only those containing a textured non-glassy-appearing bacterial lawn were used for locomotion assays. Assays were performed in a temperature-controlled room at 22uC. Assays began 5 min after animals were transferred to assay plates and lasted for 2 min. All experiments were performed on multiple days. Data were pooled for evaluation. Data for a wild-type control were collected on each assay day. A modified assay protocol was used to assay transgenic strains. In this protocol, well-fed animals were transferred to assay plates, and their movements were recorded for several 2 min intervals up to 12 min post-transfer.

Automated Tracking System
We used a Sony XCD-X710 firewire camera at 10246768 resolution to capture high-contrast dark-field images of our assay plates illuminated from the side by a ring of LEDs. Images were acquired at 1 Hz. The centers of mass (centroids) of all worms were found, and individual position recordings were assembled into tracks from centroid positions using the methods outlined by Clark et al. [45]. Animal speed was calculated by dividing the path distance an animal traveled by the time it took to travel that path. Each speed recording was calculated by averaging 5 speed measurements observed during a 5 sec sliding window to minimize the noise in centroid displacements. Such instantaneous speed measurements varied from 0 to roughly 0.3 mm per sec. Instantaneous acceleration rates were determined from these speed measurements.
The automated tracking software we used in our data capture and analysis is very similar to other freely available on-line packages that are more robust, inter-operable, better supported and easier to use than ours (MWT, MAGAT). A copy can be downloaded from GITHUB [46].

Data Analysis
The average speed of a population was defined as the mean of all instantaneous speeds calculated for all animals tested.
To analyze the variation in speed recordings within individual tracks, we used only tracks that contained at least 30 seconds of recorded data. Tracks were first separated into bins by average speed. The standard deviation of the intra-track speed recordings was determined for each track and the average standard deviation for each bin was determined from these calculations.
To analyze average total acceleration, acceleration durations, and peak accelerations within individual tracks, we used only tracks that contained at least 30 seconds of recorded data. Tracks were first separated into bins by average speed. Total acceleration was calculated as the root mean square (RMS) of accelerations. The duration of an acceleration event was calculated as the time elapsed between when an animal's acceleration became greater/ less than zero and then returned to zero. Peak acceleration/ deceleration is the highest/lowest acceleration value of each acceleration/deceleration event. Average properties were calculated for each track. The average value for each bin was then determined from these calculations.
In the above analyses, data for each track (standard deviation of speed, RMS of acceleration, duration of acceleration and peak acceleration) were treated as an independent measurement for computing the means and SEMs shown in the graphs. Significance between genotypes was computed using a 2-way ANOVA. Significance between pairs of points within each bin or between average speeds was computed with a Student's t-test.

Transgenic Strain Construction
The dop-3 cDNA construct contained an RT-PCR-amplified cDNA fragment, 10 kb of upstream regulatory sequences that included the beginning sequence TAAACTTGGTTGACC-CACCCC, and 1.5 kb of downstream sequences that included the ending sequence CCAGACGAGACGCTGAGATTGA. The transcriptional start site of the dop-3 gene was determined by 59 RACE and confirmed by SL1-primed isolation of the dop-3 cDNA.
The sequence of the dop-3 cDNA construct was determined and corresponded to sequences that encode the DOP-3 protein (WP:CE39178) annotated in WormBase. The dop-3-rescuing construct contained the genomic sequences between and including the two sequences annotated above.
Four additional constructs were created by replacing dop-3 promoter sequences in the genomic rescuing construct with heterologous promoter sequences. The tax-4 promoter construct contained 5.3 kb of genomic DNA that was amplified with the primers CACACAACGGACAGTAATGGCG and TCTTGAAACATAATTAAATTTTGAGAATGATAGAAG. The gcy-5 promoter construct contained 1.9 kb of genomic DNA that was amplified with the primers TTTCACAGTTGGCTC-CACCTCTGG and TTTTCATCAGAATAAGTAATTTTTC-GAAAAC. The unc-17 promoter construct contained 4.2 kb of genomic DNA that was amplified with the primers GTTTTTGGGATTTTTGCGGG and CTCTCTCTCTCCCCCTGGAATATTTTAT. The unc-47 promoter construct contained 2.9 kb of genomic DNA that was amplified with the primers AATGGGAGAATAAATGGGACGG and CTGTAATGAAATAAATGTGACGCTGTCG. Constructs were injected at 100 ng/mL into a dop-3(vs106) strain with a myo-2::rfp co-injection marker and into a dop-3(vs106) lin-15(n765) strain with a wild-type copy of lin-15. Both groups of transgenic animals yielded identical results; only data for lin-15 rescued transgenic strains are shown in Figure 7.

Drug Pretreatment
A fresh solution of 50 mM DA hydrochloride (Sigma) in M9 buffer [13] was prepared, and 400 ml of this solution was immediately added to a standard plate containing approximately 10 ml NGM agar seeded with a bacterial lawn. Plates were allowed to dry for 1 hr. Control plates were made by adding 400 ml M9 to standard plates containing approximately 10 ml NGM agar seeded with a bacterial lawn. Animals were incubated on dopamine-containing plates or control plates for 4-6 hr prior to assay. Assay plates and food-deprivation plates did not contain dopamine.