Behavior^*

C. elegans senses a variety of mechanical stimuli. These stimuli include gentle touch stimulus delivered to the body (Chalfie and Sulston, 1981; Sulston et al., 1975), harsh touch to the midbody (Way and Chalfie, 1989), harsh touch to the head or tail (Chalfie and Wolinsky, 1990), nose touch (Kaplan and Horvitz, 1993) and texture (Sawin et al., 2000). Gentle touch to the body is sensed by six sensory neurons (ALML/R, PLML/R, AVM, and PVM). An important feature of any assay for gentle touch mediated by these cells is that it not be confused with the response to other mechanical signals. Thus, care should be taken to touch the animals in a way that will not stimulate other mechanosensory neurons. In practice one should avoid too harsh a stimulus and not touch the animals too near the tip of the head or tail.

The above assays treat touch sensitivity as an all-or-none phenomenon. Often, however, animals give a partial response to the touch stimulus. In addition, even touch-insensitive animals sometimes respond to the first touch. Two general methods provide a more quantitative measure of the touch response: counting the responses to multiple touches and touching with defined stimuli.

The tap-withdrawal protocol measures the responses of a worm to a single tap or a series of successive taps (trains of taps), given to the side of a 4 cm Petri plate filled with 10 ml of NGM agar. The taps/trains cause the worm's environment to vibrate briefly, and this causes the adult worm to swim backwards for some distance (termed a reversal). This technique produces a quantifiable measure of the magnitude of this reversal response. For this assay worms are placed on the center of a Petri plate and are videotaped through the lens of a dissecting microscope. The mechanical tapper is arranged to hit the center of one side of the plate. The tapper is composed of an electromagnetic relay run by a stimulator. (Rankin et al., 1990) The tapper delvers a force of approximately 1 to 2 Newtons per tap which are transmitted through the dish and the agar to stimulate the worm. These experiments are usually run with the lids off of the plates while taps are being delivered. The effects of drugs on the tap response can be assessed by adding the drug plus vehicle to the agar and comparing the results to a group that was tested with just the vehicle for the drug. The length of the reversal to tap can be affected by several variables: 1) The intensity of the tap; Chiba and Rankin (Chiba and Rankin, 1990) showed that the tap response is graded and that stronger stimuli produce larger reversals. 2) The age of the worm; Chiba and Rankin (Chiba and Rankin, 1990), showed that as worms got larger so too did reversals. 3) The agar in the plates; as agar plates age they dry out and lead to differences in the ease with which worms can move. This can be controlled for by only using plates within a certain age range, or by weighing plates prior to use. 4) The reversal response can also be altered by the ambient temperature and humidity in the room where the testing is done. This is probably the result of changes in the agar and ease of movement; it is an important variable to control. At the very least, experimental and control animals must always be run together, on the same day, at about the same time under all of the same conditions. By taking note of temperature and humidity conditions for each animal it will be possible to determine whether variations in either variable affected the outcome of the experiment.

Prior to undertaking tap reflex experiments, it would be advisable to also read the protocols in the Learning, Adaptation and Habituation section on the tap reflex.

Once the experiment has been run and videotaped, then the behavior must be scored. Using stop-frame video analysis, each worm's behavior is scored by tracing its body from the videosimage onto a transparent acetate sheet on the video monitor, going frame by frame through the tape. The adult worm will usually (90% to 95% of the time) respond to the tap with a reversal response, with the size and number of the responses decreasing as stimuli are repeated through the habituation run. In a typical reversal response, a stopped or forward-moving worm moves backwards for a distance (usually less than 1 or 2 worm lengths) and then either remains still or re-initiates forward movement in a new direction. To score the response, a person notes the pre-tap position of the worm, and then traces the total distance the worm reversed (i.e. track length) onto the acetate sheet. Because the head of the worm often makes small side-to12 side foraging movements, reversals are most accurately scored by tracing the path of the tail. In order to have consensus a scoring manual should be developed with rules for scoring variations from this behavior (i.e. the reversal must occur within 1 second of the tap, a pause is scored as 0 backward movement, if the worm is already swimming backwards when the tap is delivered this is a “missing data point” etc). In some cases, a worm will respond with acceleration forward rather than a reversal. In a typical acceleration response, a stopped or forward-moving worm either initiates forward movement or increases its speed, respectively. Generally, in experiments with adult wild-type worms, these acceleration responses are simply scored as missing data points (an adult worm will accelerate to tap only about 5% of the time). Under other circumstances or protocols, such as with larval, mutant or laser-ablated animals, the acceleration response occurs more often and is therefore scored. If a worm accelerates after the tap, the acceleration is scored by measuring the distance the worm moves (i.e. track length) during the 1-sec interval before the tap and subtracting it from the distance the worm moves during the 1-sec interval after the tap. If the measurement for the second after the tap is ≥1.75 X the measurement for the second before the tap then the response is considered an acceleration. It is important to note that the reversal response and the acceleration response are two qualitatively different outcomes, which cannot easily be compared. These “scored” tracings are then scanned into a computer and the length of each is measured using NIH image. The values from this program, representing the reversal/acceleration magnitude, can then be directly transported into a statistical package for data analysis. The frequency of reversals and accelerations is also an important measure of the behavior of the worm, and are scored simply by counting the number of instances of either response.

Avoidance of high osmolarity (concentrated solutions) by C. elegans was first described by Culotti and Russell (1978). and revisited by various researchers over subsequent years. The original assay, described in an updated format below, tests the rate at which C. elegans will cross an osmotic barrier on the surface of the agar. The barrier is usually circular in shape-referred to as a ring. Animals are placed in the center of the ring (an area of low osmolarity) so they are surrounded by the high osmolarity ring which will they avoid as a noxious stimulus. Animals escape the ring by crossing the high osmolarity barrier. Outside the ring is normal NGM agar, a region of low osmolarity. In practice, the rate at which animals escape is dependent on several parameters. Particularly critical are plate dryness, the number of animals in the ring versus the size of the ring, and the rate of diffusion of the osmotic barrier. The assay is dependent on the number of times that animals encounter the osmotic barrier; special consideration must be taken when assessing osmotic avoidance for uncoordinated or lethargic animals. Hermaphrodites are assayed.

Control strains must be used each day. N2 animals are generally used as positive controls as they robustly avoid the ring barrier. Generally, che-3 or osm-10 animals are used for negative controls. Uncontaminated, uncrowded cultures must be used. Animals that have been starved during their life, those close to running out of food, animals that have passed through dauer or animals from culture plates that are very dry should not be assayed.

The number of animals to be tested determines the size of the ring. A 5/8^th inch diameter ring holds 30 to 100 animals for population assays; a 3/8^thinch diameter ring holds 1 to 5 animals. Too many animals for a specific ring size leads to aberrantly high escape rates; too large a ring for the number of animals leads to aberrantly low escape rates.

The dryness of the assay plates is a critical consideration and can be the critical variable to control. The objective is to obtain a large number of plates of uniform dryness. Plate dryness is assessed by weighing the plates. Once you establish the correct plate weight in your lab, then (if the plate volume poured remains unchanged) finding good plates for the assay is relatively straight-forward.

NGM agar plates are allowed to sit out on the bench overnight after a large batch is poured. It is important that an equal volume of media is poured into each plate; an automated media dispenser is ideal. The next day the plates are inverted and then placed in a plastic box. The length of time required for plates to dry depends on atmospheric humidity in the laboratory and how the box is covered. In winter in Boston it can take only a few days for plates to dry in a box with a closed (but not tightly sealed) lid. In summer, it can take weeks. To identify plates of uniform dryness, plates are removed from the box and individually weighed. If a constant amount of plate media is dispensed during plate pouring, then the weight of each plate is directly related to how dry the plate is (how much evaporation has taken place). If the amount of media is a constant over weeks or months, then plates can be often selected based on weight alone-once the correct weight is determined in behavioral assays. Be sure to remove any condensation from the plate lid before weighing.

The only definitive test to determine the correct weight (and dryness) is to run the osmotic avoidance assay with the control strains. A good rule of thumb to start with here is that animals crawling on the agar surface of an appropriately dried plate should have a visible, but small meniscus.

Alternative approaches: The drying process can be hastened by covering the box with a large Kimwipe (as opposed to a lid) although uniformity of plate dryness may decrease. Or, plates can be dried in a sterile hood for a few hours to hasten the drying process-but the plates must be allowed to equilibrate with the lids closed for several hours afterward.

Prepare 8M glycerol in water. This corresponds to a 60% solution of pure glycerol in water; “pure glycerol” from Sigma is a 13M solution. Add a tiny bit of bromophenol blue to color the 8M glycerol and the resulting rings blue. 8M glycerol aliquots can be stored in the −20°C freezer for years; aliquots are stable at room temperature for days.

Depending on the number of animals to be assayed, either ultra fine point Sharpie lids (∼3/8^th inch) or metal test tube caps (∼5/8^th inch) are obtained. C. elegans S Basal is required for population assays.

Osmotic barrier rings must be printed onto the agar of the assay plate so they will be ready as soon as the animals finish rinsing in the S Basal (see next paragraph). Usually four assay plates can be run in one batch. Obtain 4 additional 6cm Petri dish lids and four metal test tube caps. Pipette about 25 to 30μl of 8M glycerol into the surface of each inverted lid. Place the open end of the cap down onto the inverted lid bisecting the drop of glycerol. Spin the cap against the surface of the lid to spread the glycerol around the circumference of the cap. Lift the cap. If there is a bubble of glycerol closing the cap, then pop it. Place the lid with glycerol very gently onto the agar of an assay plate. Try to avoid marring the surface of the agar. Let the glycerol soak into the agar; this usually takes between 2 and 5 minutes. Lift the cap off. Check to make sure the annular osmotic barrier is uniform. Always run the control animals in duplicate (e.g., N2 and osm-10) first to confirm that the assay plates are working. It is a good idea to run control interspersed with experimental animals during the assays.

Rinse animals off culture plates and into a 1.5ml microfuge tube with 1ml of S Basal using a Pasteur pipette. No need to close the tube lids. Let animals settle by gravity for 5 minutes, remove excess S Basal and add another 1ml. After 5 to 10 minutes, move animals in a minimal amount of liquid to the center of the osmotic ring. Remove excess liquid by very gently blotting with a rolled up Kimwipe. Once the excess liquid is removed and animals start crawling, the timing begins for that assay plate. Record the time the assay starts. Replace plate lid and avoid disturbing animals by dropping or tapping the plate during the assay. Repeat till all four assays are underway. When the assays are done, wipe off the caps and lids so they can be reused.

Animals should move actively within the ring during the assay period. N2 animals should avoid entering the osmotic barrier for at least 10 minutes; 50 to 80% of osm-10 or other control animals should either die in the osmotic barrier or cross it and escape the ring. Those who die in the ring are counted as escapers. If too many N2 animals escape, then the plates are likely too wet. If too few osm-10 animals escape, then the plates are likely too dry.

The number of animals that have escaped at 10 minutes is recorded along with the number of animals that remain in the ring. The % escaping animals is calculated. At least 5 assays per strain or genotype are suggested; these should be determined on at least two different days.

The rate at which animals encounter the osmotic barrier is important in determining their response. If limited animals are available-or individual animals need to be tested, then smaller rings should be used.

Rings can be generated using ultra fine point Sharpie lids and 20 μl of 8M glycerol as described above. Use the open end of the Sharpie lid. Animals to be assayed are gently (and rapidly) picked from the bacterial lawn to an empty plate (to eliminate food) to the assay plate. Up to 5 animals can be used in a ring of this size. At least 7 rings with at least 4 animals for each genotype is a good staring point. As described above, control strains should be run before experiments begin.

If the osmotic response of individual animals must be assessed (e.g., laser ablation experiments), then animals should be assayed at least 4 times. Once an animal crosses the osmotic barrier, then the animal is immediately returned to the bacterial lawn for at least 15 minutes.

If small numbers of animals are used, then it is convenient, and perhaps more accurate, to record the time the animals cross the barrier, not the % escaping. While the assay is running, plates are checked every minute and the number of animals who have escaped is recorded. The average time to escape is then calculated.

Genetic screens for animals defective for avoidance of aversive compounds were carried out on assay plates with 10³-10⁴ worms per plate. The assay plates were prepared by pouring 5 ml of chemotaxis agar (1.5% agar, 75 mM NH₄Cl, 10 mM MOPS, pH 7.2 with NH₄OH, and 0.25% Tween 20) into a 9 cm Petri dish. A thin line (25 μl) of an aversive compound (e.g., copper as 150 mM CuSO₄, 2 % SDS, or high osmotic pressure as 4 M fructose; Ward, 1973) was poured across the midline of the Petri dish and another 25 μl traced in a semicircle around one half of the dish (Figure 2). Young adult worms were washed twice with S-basal buffer to remove bacteria. Immediately after the compound was absorbed into the chemotaxis agar, worms were placed on one side of the assay plate in a small volume of buffer. Excess buffer was carefully removed with tissue paper. A volatile attractant (2 μl of isoamyl alcohol 1:10 dilution in ethanol) was placed across the assay plate 1 cm from the opposite edge with the aversive semicircle. Normal, healthy worms approached the attractant, but stopped abruptly at the boundary formed by the repulsive compound. Individuals that passed through the aversive barrier were isolated and allowed to produce self progeny. To quantify the behavior of a given strain, an aversion index (A.I.) was calculated by dividing the number of worms that, one hour after placing the worms on the plate, crawled through the aversive compound by the total number of worms on the plate (approximately 250–500). This index includes all animals that attempt to crawl up the side of the plate.

Nematodes are grown on E. coli strain HB101 at 20°C (Brenner, 1974). Raise all animals with plentiful food in uncrowded conditions (ideal-place about 3 adults on one 6 cm plate with food, or 10 adults on one 10 cm plate with food, wait 4 days at 20 and test their progeny; for mutants you will need to place down more adults at the beginning-about twice as many for osm-6 as for N2). Wild-type animals are C. elegans variety Bristol, strain N2. Test animals in their first or second days as adults, larvae do not perform as well.

Chemotaxis assays are based on the assay developed by Ward (1973).

Unless indicated otherwise, all males come from the him-5 (e1490) mutant, which segregates XO male progeny by X chromosome nondisjunction during meiosis (Hodgkin et al., 1079). To construct cue source regions, the muscle mutant unc-52 (e444), individuals of which become relatively motionless by the time they are young adults, was used. For the vulvaless experiment we constructed a let-23 (lsy1) unc-52 (e444); dpy-20 (e1282)lin-3 (n378); him-5 (e1490) strain (PS3980; Aroian and Sternberg, 1991; Ferguson and Horvitz, 1985). lov-1 (sy582δ) and pkd-2 (sy606δ) have been described (Barr et al., 2001). The sensory mutants osm-5 (p813) and osm-6 (p811), as well as the diverse isolates CB4932 (Taunton, England), CB4555 (Pasadena, CA), and CB4856 (Hawaii) were obtained from the Caenorhabditis Genetics Center (St. Paul); isolates are described by Hodgkin and Doniach (Hodgkin and Doniach, 1997) and further characterized by de Bono and Bargmann (de Bono and Bargmann, 1998). Similar to wild-type [Bristol N2; him-5 (+)] strains, males from isolates occur spontaneously at an inconveniently low frequency in the self-progeny of hermaphrodites. To obtain a constant supply of males, isolates were heat shocked and maintained by backcrossing. To test males from osm-6, him-8 (e1489); osm-6 (p811) strains were made, and the presence of osm-6 was verified by use of fructose avoidance and dye-fill assays (http://cobweb.dartmouth.edu/_ambros/worms/16.html; Culotti and Russell, 1978). All animal stocks were stored at 18°C. All stocks and animals harvested for upcoming trials were grown on standard 5-cm diameter NG agar plates inoculated with the Escherichia coli strain OP50, grown in Luria-Bertani media (LB), as a food source.

Bacterial lawns were grown on standard 5-cm diameter agar plates for all trials: three drops of OP50-inoculated LB, spread thinly, with a ∼0.25-centimeter gap between the edge of the lawns and the walls of the plates, to dissuade test animals from leaving trials. Plates were then stored at ∼22°C for 2 days until used for trials. Source and test animals were harvested daily at the fourth larval stage (L4), and stored at ∼18°C overnight with 10-20 animals per same-sex plate to be used the following day in trials as young adults. The muscle mutant unc-52 was used to condition agar plates for at least 3 and up to 8 h (see Figure 1a in Simon and Sternberg, 2002). Five to 10 min before the onset of a particular trial, source animals were removed. Trials consisted of the introduction of a single test animal on to a single trial plate (see Figure 1a in Simon and Sternberg, 2002), and movement of the animal was then documented for 5 min. Effort was made to orient animals in the direction of the 3 × 7-mm scoring regions. A motor-driven stage attached to a computer joystick was used to help videotape trials through a dissecting microscope. Trials were videotaped, and the number of reversals per crossing and the time animals spent in conditioned region were scored and are on record. Backwards movement equal to or greater than a body length was counted as “1 reversal”; backward movement less than a body length was counted as “0.5 reversal.” We used the frequency of reversals as indication that males detect a cue; it seems likely that when males pass over a mate-finding cue that they will move backwards, and by repeated reversals, home in on a cue's source. Equal numbers of unconditioned and conditioned agar trials were run in parallel on ∼5 plates each per day; moreover, trials were run blindly and interspersed at random. Isolate trials were run in parallel alongside trials of the laboratory strain (Bristol N2). Unconditioned agar trials were meant to parallel conditioned agar trials as closely as possible: (i) made from same day bacterial lawns, (ii) kept at the same temperature, and (iii) had bacteria picked on and off to mimic introduction of source animals as on conditioned agar trials. Average temperature during trials was ∼22°C.

The holding assay was identical to the response assay with the following modifications: (i) source animals were introduced in a pile in area A, (ii) test animals were introduced directly on this conditioned spot (see Figure 1b in Simon and Sternberg, 2002); (iii) in addition to area A, number of reversals was scored in outwardly stacked, concentric ring-areas (with radii of 1 mm, 5 mm, 16.5 mm, and ∼25 mm; unequal sized areas were normalized by dividing outer ring-areas by 24, 247, and 330, respectively); and (iv) movements of test animals were documented for 10 min.

Design of the attraction assay was similar to the holding assay with the following modifications: (i) source animals were left to condition a point source for 24 h; (ii) at the onset of each trial, individual hermaphrodites were placed atop conditioned regions; and (iii) test males were introduced ∼1.5 cm from the conditioned point source/unconditioned scoring region (see Figure 1c in Simon and Sternberg, 2002).

The working bench should be horizontal. Avoid sources of wind or vibration-such as a centrifuge machine or a draft chamber nearby. Do not use an incubator for the assay because it is strong source of wind of fan and vibration from the compressor.

Room temperature should be kept at 25 ± 1 °C.

6-cm seeded NGM plates. Use a 6-cm plate containing 14 ml of NGM, which is minimal volume required for getting sufficient bacterial lawn. The NGM plates were spread with overnight-cultured E. coli, incubated for a day at room temperature (or for overnight at 37°C), and stored at room temperature. Do not use old seeded NGM plates (over 2 weeks).

C. elegans: Use the healthy adult animals from uncrowded plates. Animals should be maintained by transferring 2 to 3 animals to a new NGM plate before starving. Do not use the starved animals or those from a starved agar chunk.

Assay plate: The medium consists of 2% agar, 50mM NaCl, 1mM CaCl₂, 1mM MgSO₄, and 10mM pH 7.0 HEPES. After autoclaving, pour 10 ml of the medium exactly into a 9-cm plate. Store at 4°C and use within 2 weeks. Spread 25 ml of 100 mM (CH₃COO)₂Cu solution on the midline of an assay plates, and place at room temperature for 18-22 hr to allow diffusion before the assay.

Wash buffer: 0.02% gelatin, 50mM NaCl, and 25mM pH 6.0 potassium phosphate.

A number of decisions need to be made prior to running any experiment on habituation to tap: 1) The age of the worms to be used; Chiba and Rankin (1990) showed that it is only after the end of the larval stages that worms consistently reverse to tap. It is best to use worms of the same age for the experiments because for wild-type worms the length of the reversal response is directly related to the length of the worm. 2) How many stimuli will be delivered; Rankin and Broster (1992) showed that the higher the number of stimuli the longer it took for spontaneous recovery to begin. 3) The interstimulus interval (ISI or frequency) of the taps; over a series of experiments we have shown that habituation at short ISIs has different properties from habituation at long ISIs and is probably mediated by some different mechanisms (reviewed in Rose and Rankin (2001)). 4) Whether there will be food (E. coli) on the plate or not; for short-term habituation experiments there is not usually food on the plate; for long-term habituation experiments that last 24 hours there usually is a small amount of food. And 5) at the end of the experiment will there be tests for either spontaneous recovery or for dishabituation? Whatever decisions you make, the most important thing is to be consistent. In addition it is important to always run control animals at the same time and under exactly the same conditions as the experimental animals. That means run one experimental and then one control throughout the experiment. If you are testing mutant strains the best control is the N2 strain from the lab where the mutant was made or into which it was backcrossed. Alternatively a rescue strain might be a good control-however the high level of variability in some rescue strains makes this sometimes problematic.

When worms that have been habituated to tap are left unstimulated for a period of time they show spontaneous recovery of the tap response. To look at spontaneous recovery worms are left on the microscope stage and are given test stimuli at various intervals after habituation training to determine the extent to which the response has recovered. Spontaneous recovery can be used to distinguish habituation from sensory adaptation or motor fatigue. Worms habituated at short ISIs show great habituation, but recover more rapidly (within ∼10 min) than worms habituated at long ISIs (at least 30 min), which show less overall habituation. This rapid recovery following habituation at short ISIs is the opposite of what would be predicted for adaptation or fatigue (where the most decremented group should recover the most slowly).

Dishabituation can also be used to distinguish habituation from sensory adaptation or motor fatigue. In dishabituation the habituated response is rapidly recovered by the delivery of a dishabituating stimulus that is usually a strong novel or noxious stimulus. In this protocol, the dishabituating stimulus is a brief train of electric shocks, generated by the Grass S88 stimulator. The shocks are delivered using a hand-held spanning electrode (a bipolar electrode used for stimulating rat brains) with wires placed into the agar on either side of the worm, about 3 to 4 worm widths apart. The 2 forks of the electrode thus span the animal, piercing the agar about 1 worm width away on each side of the cuticle, at about the middle of the worm's length. Neither fork of the electrode actually contacts the cuticle--the shock is transmitted solely through the agar. The train of shocks usually consists of 10 ms shocks of 60 V at 10 PPS for 600 ms. It is important to decide on the appropriated delivery time for the shock-the worms require about 30 seconds to recover from the shock, and so it is important to give them time to recover before the first tap. However, if there is a long interval between the last tap of habituation and the first tap following shock it is important to control for any spontaneous recovery that might have occurred in the interval by having a control group in which the same interval is given, but no shock delivered. Following this dishabituating stimulus an additional set of 10 taps at the same ISI as training, is given to assess the effect of the shock on the response that was habituated. Dishabituation is shown by the increase in response amplitude above habituated levels following the shock.

To assess long-term memory for habituation tap stimuli are delivered in a distributed or spaced training procedure. For training approximately 20 young adult worms are pre-plated on an agar plate containing a thin lawn of E. coli 24 hours prior to training. A second plate for control worms is pre-plated at the same time. Both plates are left on a vibration-free shelf in the lab over night. Twenty-four hours after pre-plating the experimental plate is placed on the stage of the dissecting scope and taps are delivered. The worms receive four blocks of 20 taps each. With in a block taps are delivered at a 60 s ISI, and between blocks there is a 1-hour rest period (Rose et al., 2002). The control plate of worms receives a single tap at the end of the training period. One hour after the end of training worms from the experimental and control plates are transferred to individual Petri plates (again with a thin lawn of E. coli) and left on the vibration free shelf for approximately 22 hours. To test for long-term memory for habituation each worm (both experimental and control) is given a set of 10 taps at a 60 s ISI. Testing can begin about 22 hours after training and continue until 26 hours after training. Responses to the test taps are scored and long-term memory is considered to be present when the mean response magnitude of the trained or experimental worms is significantly smaller than the mean response magnitude of the control worms. In these experiments it is usual to have at least 20 worms per group. It is possible to test drug effects on memory by delivering the drug through the agar that the worm is on during training (for drug experiments worms can be pre-plated on drug containing plates 1 hour prior to training). For drug experiments there are four groups of worms tested: drug trained and control groups, and vehicle trained and control groups. Worms are removed from the drug plates 1 hour after training is completed and tested approximately 24 hours later on plates without drug. The drug is considered to have blocked memory if the vehicle group shows normal long-term memory with trained worms having smaller responses to the 10 test taps than control worms, and if the drug group shows no difference between trained and control worms. In addition, to show that the drug did not damage the worms in some way it is important that the control worms for the drug and vehicle groups do not differ from one another.

Adaptation plates are prepared as follows:

10 mLs of adaptation agar (3% agar, 5mM KPO₄ [pH6], 1 mM CaCl₂, 1 mM MgSO₄) or chemotaxis assay agar (2% agar, 5mM KPO₄ [pH6], 1 mM CaCl₂, 1 mM MgSO₄) is aliquoted into 10 cm Petri plates, and allowed to dry overnight. The day of the experiment, all plates are dried at 37°C without lids for 1 hour, and a representative plate of each type is melted to determine volume. Ethanol is added to appropriate plates to a concentration of 200 mM, the plates are Parafilmed and allowed to soak for 1.5–2 hours. Odorant is aliquoted onto 5 agar plugs on the lid of adaptation plates (for benzaldehyde adaptation, 2 μL of 100% benzaldehyde is used; for butanone, 5 μL of 100% butanone is used). Between 100 and 300 animals are washed twice in S Basal and once in assay buffer (5mM KPO₄ [pH6], 1 mM CaCl₂, 1 mM MgSO₄), then placed on an adaptation plate, and the plate is sealed with Parafilm. Animals are incubated in adaptation conditions for 90 minutes, then washed twice with S Basal and once with assay buffer and placed on chemotaxis plates. Chemotaxis is allowed to proceed for 1 hour, worms are counted, and a chemotaxis index is calculated. Adaptation is observed as a decrease in chemotaxis to an odorant after pretreatment with the odorant.

Chemotaxis assay plates are prepared as follows: 10 mLs of assay agar (2% agar, 5mM KPO₄ [pH6], 1 mM CaCl₂, 1 mM MgSO₄ is aliquoted into 10 cm Petri plates, and allowed to dry overnight. The day of the experiment, all plates are dried at 37°C without lids for 1 hour. If ethanol is to be used in the experiment, one representative plate is melted to determine final volume, and ethanol is added to the desired final concentration (generally 200 mM). The ethanol plates are sealed with Parafilm, and ethanol is allowed to soak into the agar for 1.5 -2 hours. Between 100 and 300 animals are washed twice in S Basal and once in assay buffer (5 mM KPO₄ [pH6], 1 mM CaCl₂, 1 mM MgSO₄). A spot of diluted odorant (for benzaldehyde, 1 μL of 1:200 benzaldehyde: ethanol) is placed on one side of the plate. Exactly opposite is placed a spot of diluent (1 μL ethanol). To each spot is added the anesthetic sodium azide (1 μL of 1 M sodium azide) to immobilize the worms when they reach a spot. Washed worms are then placed onto a spot on the plate that is exactly between the odorant and diluent spots and slightly off center. Excess liquid is wicked up with the corner of a Kimwipe. After 1 hour, worms are counted, and a chemotaxis index is calculated as follows: CI= (# of worms at the odorant spot - # of worms at the diluent spot)/ # of worms in the assay. A high chemotaxis index (close to 1) indicates that the odorant acts as a strong attractant, a lower chemotaxis index indicates that the odorant is a less effective attractant or does not act as an attractant.

We demonstrated that C. elegans is able to form an association between the presence of the odorant benzaldehyde and the food content of its environment. When exposed to 100% benzaldehyde for 1 h in the absence of food the naive attractive response is reduced. In contrast, the benzaldehyde attractive response is strengthened when animals are exposed to 100% benzaldehyde for 1 h in the presence of food. These results suggest that animals can form an association between the odorant benzaldehyde and food status (Nuttley et al., 2002).

Nematode strains were obtained from the Caenorhabditis Genetics Center at the University of Minnesota (St. Paul). All experiments used well fed adults cultivated at 20°C on standard nematode growth medium (NGM plates) seeded with E. coli strain OP50.

Assays were performed with standard 100-mm Petri dishes containing 6 m20/15g/liter agar). Where necessary, odorants were diluted in ethanol and l of CTX medium (10 mM Mops, pH 7.2/0.25% (vol/vol) Tween reported as percentages by volume. Plates were sealed with a strip of Parafilm around the edge during all odorant exposures. Briefly, 15 min before the assay 1 ml of 1 M NaN₃ was applied to the centers of two test spots that were 6 cm apart. This acts as an anesthetic to immobilize any animals that reach the spot during the assay. Individuals (100-300) then were placed at the center of the plate between the two spots, 1 ml of the test odorant was placed at one spot, and 1 ml of ethanol was applied to the control spot. After 1 h of chemotaxis, animals within a 2-cm radius of either spot were counted, and a chemotactic index (CI) was calculated as the number of animals at the test spot minus the number of animals at the control spot and divided by the total number of animals on the plate. A positive CI indicates an attraction to the odor, and a negative CI indicates an aversion.

To measure the effects of pre-exposing animals to starvation-benzaldehyde pairing, 500-1,000 animals were placed on a CTX plate, and then 2 ml of the odorant were placed on a piece of Parafilm on the lid of the plate. The plate was sealed with Parafilm and then left inverted for 1 h. The animals then were rinsed from the test plate in approximately 2 ml of water and transferred to a conical centrifuge tube, where they were allowed to settle to the bottom of the tube. Within 5 min, the concentrated animals were collected and transferred to a fresh CTX plate for testing.

For assays involving food-benzaldehyde association on the CTX plates, cells from a 6X concentration of logarithmically growing E. coli culture were spread onto the plate and allowed to dry immediately before pre-exposure. A 1X concentration of E. coli was defined as an amount equivalent to 100 ml of culture diluted to an OD600 of 0.5. As with many food-related behavior, exogenous serotonin (5-HT) can substitute for food in this assay. For preparation, 5-HT salt (Sigma) was diluted in M9 buffer for a final concentration of 100mM. 100μL of this solution was spread onto the plate and allowed to dry immediately before benzaldehyde pre-exposure.

Assays are undertaken with well-fed, 1 day old adult animals raised at 20 degrees Celsius. Plates are closed during the assay che-1 and N2 animals are used as controls. The chemotaxis assay were performed essentially as described by Dusenbery et al. (1975) with some modifications (Uchida et al., 2003). A radial concentration gradient of NaCl is established by spotting 2μl of 5M NaCl to the center of 9cm plate containing 8 ml of agar medium [2% agar, 0.25% Tween20, 10mM HEPES(pH7.2)], and leaving the plate at room temperature for 12-16 hr. For the assay, the animals are placed on the surface of agar, 1 cm distant from the periphery and allowed to move freely for 1hr.

Put the agar assay plate 1cm away from frosted glass and light it from agar plate side using a slide projector set 2-3m distant. Worm tracks appear as darkened tracks on the frosted glass. These are photographed using a digital camera.

The working bench should be horizontal. Avoid sources of wind and vibration such as a centrifuge machine or a draft chamber nearby. Do not use an incubator for the assay because it is strong source of wind of fan and vibration from compressor.

Room temperature should be kept at 25±1°C.

6-cm seeded NGM plates. Use a 6-cm plate containing 14 ml of NGM, which is minimal volume required for getting sufficient bacterial lawn. The NGM plates were spread with overnight-cultured E. coli, incubated for a day at room temperature (or for overnight at 37°C), and stored at room temperature. Do not use old seeded NGM plates (over 2 weeks).

Use the healthy adult animals from uncrowded plates. Animals should be maintained by transferring 2 to 3 animals to a new NGM plate before starving. Do not use the starved animals or those from a starved agar chunk.

Conditioning plate: The medium consists of 3% agar, 50mM NaCl, and 10mM pH 7.2 MOPSxNH₄. After autoclaving, pour 8 ml of the medium exactly into a 9-cm plate. Store at 4°C and use in 3 weeks. (See Figure 3.)

Mock conditioning plate: The medium consists of 3% agar and 10mM pH 7.2 MOPSxNH₄. After autoclaving, pour 8 ml of the medium exactly into a 9-cm plate. Store at 4°C and use in 3 weeks.

Assay plate: The medium consists of 2% agar and 10mM pH 7.2 MOPSxNH₄. After autoclaving, pour 3∼3.5 ml of the medium exactly into a 6-cm plate. Store at 4°C and use in 3 weeks.

NaCl plug: The medium consists of 2% agar, 50mM NaCl, and 10mM pH 7.2 MOPSxNH₄. After autoclaving, pour 6 ml of the medium exactly into a 6-cm plate. Store at 4°C and use in 3 weeks. To make NaCl-gradient on an assay plate, cut a plug (5 mm in diameter) from the 50mM NaCl plate and put a plug on one side of an assay plate for 19∼23 hr (Figure 4). After making NaCl gradient, discard the plug from assay plate before chemotaxis assay. You may use the Pasteur pipette (5 mm in diameter) to cut and put a NaCl-plug. Wash buffer: 0.5g per liter gelatin, 1mM CaCl₂, 1mM MgSO₄, and 25mM pH 6.0 potassium phosphate. Store at room temperature.

We demonstrated that olfactory adaptation (benzaldehyde-starvation learning) is taste context dependent. (Law et al., 2004) Animals adapted in a sodium acetate context require the same sodium acetate taste context for adaptation response. The same context effects can be observed if animals are trained in an ammonium chloride context or the standard NGM context.

All experiments used well-fed adults cultivated at 20°C on normal growth medium (NGM; 50 mM NaCl, 15 g/l agar, 20 g/l peptone, 1 mM cholesterol, 1 mM CaCl₂, 1 mM MgSO4, 1 mM KPi [pH = 7.0]) seeded with the OP50 E. coli strain. Minimal medium was 15 g/l agar, 20 g/l peptone, 1mM KPi (pH = 7.0). CTX medium was 10 mM Mops (pH 7.2), 0.25% v/v Tween 20, 15 g/l agar with NaAc or NH₄Cl added to a final concentration of 25 mM where indicated.

N2 worms received benzaldehyde adaptation training in either the NaAc or NH₄Cl added to a CTX medium for a final concentration of 25 mM. Plates were poured one day before experiments and allowed to dry.

In a second set of experiments, we defined the contexts using the standard NGM context and a minimal context (with all the cues in the NGM context removed except for 15 g/l agar and 20 g/l peptone). The olfactory adaptation learning assay was carried out as described by Nuttlet et al. (2002).

Thermotaxis is one of the interesting behaviors in C. elegans. This behavior, reported first by Hedgecock and Russell (1975), should require some mechanisms for sensing and memorizing temperature, which are still largely unknown. Thermotaxis is also a behavior difficult to assay, because this behavior is influenced by the environmental conditions including temperature, population density, and feeding status. Animals avoid temperatures at which they have been starved and accumulate at temperatures associated with food. However, the critical control of the environmental conditions enable us analyze this challenging behavior. This protocol describes a series of tips for thermotaxis (Ttx) assays (Mori and Ohshima, 1995).

Equipment and reagents: The working bench should be horizontal. There should be no source of wind and vibration nearby-such as a centrifuge machine or a draft chamber. Do not use an incubator for Ttx assays because it is strong source of wind of fan and vibration from the compressor.

Room temperature should be kept at 25 ± 1 °C. Avoid local temperature changes caused by sunlight. Low humidity (20 ∼ 30%) is preferred for assay.

Incubator to grow animals: A small volume (100 ∼ 150 L) incubator is preferred to get stable temperature.

6-cm seeded NGM plates. Use a 6-cm plate containing 14 ml of NGM, which is minimal volume required for getting sufficient bacterial lawn. The NGM plates were spread with overnight-cultured E. coli, incubated for a day at room temperature (or for overnight at 37°C), and stored at room temperature. Do not use old seeded NGM plates (over 2 weeks).

C. elegans: Use the healthy adult animals from uncrowded plates. Animals should be maintained by transferring 2 to 3 animals to a new NGM plate before starving. Do not use the starved animals or those from a starved agar chunk. To avoid crowding, you may transfer ∼ 15 of L4 larvae to a fresh and pre-incubated NGM plate. Grow those for 8 ∼ 24 hr (depend on cultivation temperature) and use for assays.

9-cm Ttx plates. Ttx medium consists of 2% agar, 0.3% NaCl and 25 mM potassium phosphate buffer (pH 6.0). After autoclaving, pour 8 ml of the medium exactly into a 9-cm plate. Store at 4°C and use within 2 weeks.

Vials filled with frozen glacial acetic acid (stored at 4°C). Glass vials are 10-cm long and 2.7-cm diameter. Freeze glacial acetic acid at -20°C and store at 4°C.

C. elegans crawling is characterized by a series of forward movement punctuated by occasional backward movements and abrupt turns. A swimming worm suspended in a microdroplet will also execute these movements spontaneously. There are several advantages to studying worm behavior by examining swimming patterns in microdroplets: 1) The movements of a single worm can be monitored for an extended period of time using a stationary video camera. 2) Thermal stimuli are easily delivered to the microdroplet. 3) The movements of the swimming worm can be quantified objectively using machine vision software.

We have developed a microdroplet-based thermotactic assay that effectively quantifies the behavioral output of individual swimming worms responding to defined thermal stimuli. (Ryu and Samuel, 2002; Satterlee et al., 2004; Figure 5).

The thermal stage is designed for use on an inverted microscope. The worm swims freely in a microdroplet sandwiched between a cover slip from above and a round sapphire window from below. The typical microdroplet contains ∼0.5 μL of NGM buffer (same salt concentrations as NGM media) that flattens to a circular droplet ∼2 mm in diameter and ∼300 micrometers in thickness within which we place a single young adult worm. (For monitoring the swimming behavior of a single worm in the microdroplet, a low magnification objective (1-4x) is useful. But this thermal stage can also be used for physiological measurements of neuronal activity using high-numerical aperture objectives.)

The single-crystal sapphire window (15.9 mm diameter, 1 mm thick; Rolyn Optics, Inc) is embedded in a slab of copper such that the upper surfaces of the sapphire and copper are flush. We use copper and single-crystal sapphire because of their excellent thermal conductivities. A microthermocouple - either embedded in the copper near the window or in contact with either the sapphire or the droplet itself - is used to continuously monitor the actual temperature and to provide feedback to the control electronics. A suitable thermocouple is the IT-23 Teflon-coated Type T thermocouple made by Physitemp Instruments Inc.

A Peltier-effect 85 Watt thermoelectric element (5 cm x 5 cm) sandwiched between the copper stage and a brass, water-cooled heat sink provides feedback-controlled heating/cooling. The thermoelectric controller we use is the Thorlabs TEC2000 (Thorlabs, Newton, NJ), which provides adequate power and straightforward computer interfacing, but many other options exist. To minimize the temperature gradient across the Peltier element (and thus the power the thermoelectric controller has to deliver), we also use a temperature-controlled water bath (Model 1160S Heated/Refrigerated Circulator, VWR) to fix the temperature of the heat sink to an intermediate value of the actual experiments.

An alternate microdroplet configuration consists of suspending 50 μL of NGM buffer from a 4 mm diameter hole in an temperature controlled anodized aluminum plate. The top of the hanging droplet is covered with a glass coverslip and the bottom of the droplet is left exposed. The microdroplet can be imaged from above using a video camera with a zoom lens or a stereomicroscope. Worms swim naturally but remain in the center of the droplet due to the droplet's curvature. There is some evaporative loss, but if the droplet volume is maintained, a single worm can be studied for many hours.

We have written software in the LabVIEW programming environment (National Instruments, Austin, TX) to simultaneously control and monitor the temperature of the droplet and also capture images of the worm. Thermal stimulation of the swimming worm is achieved via defined voltage waveforms sent to the thermoelectric controller, using a multifunction data acquisition card [PCI-6024E]. Movements of the worm swimming in the microdroplet are imaged with a CCD camera and captured to the computer with a PCI-1407 Image Acquisition card. We find that a simple and effective metric of behavioral performance is the nose-to-tail distance of the swimming worm. During forward swimming, bending waves propagate from nose-to-tail and the nose-to-tail distance oscillates just below the contour length of the worm. During reorientations (either sharp turns or reversals), the nose-to-tail distance decreases significantly. (See Figure 6 and Figure 7).

We wrote image-processing software to calculate the nose-to-tail distance of a swimming worm using the LabVIEW Vision Development Module. Since the software works best with a high-contrast image, we typically used dark-field illumination for the microdroplet supplied by an LED ring or a fiber-optic gooseneck lamp set at an oblique angle. For each video frame, a background subtraction and threshold was used to eliminate debris or uneven illumination from each image. The centerline of the worm was then calculated using a skeletonize algorithm. The absolute distance between the end-to-end of the centerline is then calculated for each video frame. Typically, we use a frame rate of 1-5 Hz. High frame rates are suitable if one wants to calculate the rate of bending waves, which will emerge as the high frequency component in the oscillation of end-to-end distance. Reorientation events are detected in each experiment by flagging significant dips in the nose-to-tail distance over the time course of each experiment.

The cryophilic response of the worm thermotaxis is straightforward to measure using the microdroplet assay. In earlier work, we found that when worms navigate at temperatures warmer than their cultivation temperature (T_cult), increasing temperature stimulates reorientations and decreasing temperature suppresses reorientation (1). The cryophilic response dominates worm movements when worms swimming in a microdroplet are subjected to thermal oscillations.

In order to set the memory of T_cult, we grow worms overnight on NGM plates containing an ample lawn of OP50. A single worm is picked from the plate, rinsed in an NGM buffer (containing the same salt concentrations as the plate), and then picked to the microdroplet. Thermal cycles around ambient temperatures warmer than T_cult (e.g., cycling temperatures between 23-24°C for T-_cult-=20°C) will stimulate reorientations during the warming phase and suppress reorientations during the cooling phase. By cycling the temperature several times for each worm, it is possible to perform a robust measurement of the stimulus-response relationship for an individual worm. By measuring the ambient temperatures at which the cryophilic drive is active, it is also possible to use the microdroplet in a 'memory assay' to quantify the set-point of T-_cult in a single animal.

The main limitation of the microdroplet assay for thermotaxis is that a correlate of isothermal tracking has not yet been detected. During isothermal tracking, an individual worm persistently zigzags along each isotherm in a spatial thermal gradient. One possibility is that the uniform temperature of the microdroplet could not support a correlate of isothermal tracking that might require a spatial thermal gradient. Another possibility is that the behavioral response is subtle and might also require a stimulus pattern that more closely matches the thermal experience of a worm tracking an isotherm (e.g., alternating upward and downward steps of ∼0.1°C every 10-20 seconds).

The automated analysis of behavioral responses of single swimming worms has accelerated studies of thermotaxis (2). The machine-vision analysis of movements of single worms in microdroplets is a general technique, which can easily be implemented in other studies, e.g., measurements of spontaneous reorientation, bending wave frequency.

Conventional methods for studying worm thermotactic migration utilize frozen glacial acetic acid as a cold source and an incubator as a hot source to create spatial thermal gradients on agar plates. (See preceding protocol, Mori and Ohshima, 1995). We have developed an alternative approach utilizing a simple linear thermal gradient apparatus that utilizes Peltier-effect thermoelectric elements to produce highly reproducible and stable gradients on agar plates with simultaneous monitoring of worm movements using video microscopy.

We establish a thermal gradient across an anodized aluminum slab (4"x9"x1/4") by affixing 160W thermoelectric elements on each end. Each thermoelectric element is sandwiched between the aluminum slab and its own brass heat sink. Water is continuously circulated through both heat sinks by a temperature-controlled water bath (Model 1160S Heated/Refrigerated Circulator, VWR). The thermoelectric elements create a thermal gradient across the middle of the slab, atop which can be placed a 9 cm agar plate. A few drops of glycerol between the plate and the top surface of the slab ensure thermal contact. The agar plate is covered to isolate the surface from air currents. To prevent condensation build-up, the cover of the agar plate is coated with a thin layer of liquid detergent (Figure 8).

Separate feedback loops are needed for both thermoelectric elements. We use separate calibrated thermistors embedded in the agar slab at the “hot” and “cold” edges of the gradient. Each thermistor and Peltier-effect thermoelectric element is connected to its own thermoelectric controller. The amount of power needed for steep gradients (1°C/cm) exceeds the operating range of two Thorlabs TEC2000s (Thorlabs, Newton, NJ). One option is to use two Oven Industries Model 5C7-195 thermoelectric controllers (Oven Industries, Mechanicsburg, PA) with external power supplies with the proper power rating (>160W). Another option is to use Wavelength electronics MPT10000 Chassis Mounted Thermoelectric modules (Wavelength electronics, Bozeman, MT), although these modules are not 'stand-alone' devices and require assembly into an electronic instrument (e.g., RIS-549 made by the Electronic Engineering Laboratory at the Rowland Institute at Harvard).

The temperature gradient at the agar surface will be less steep than the gradient across the aluminum slab, and needs to be verified with each setting of the thermoelectric elements. This can easily be done using a hand-held digital thermometer (e.g., Fluke 50 Series) with a T-type thermocouple.

Uniform illumination and high-contrast images are necessary to visualize worms navigating the linear thermal gradients. We have engineered a simple dark-field illuminator using a ring of 48 superbright light-emitting diodes in the same plane as the agar plate. Using fresh, clean plates without surface defects, the only light-scattering objects are the worms themselves. These high-contrast images of the worms can be recorded using a CCD camera (e.g., Panasonic WV-BP554) equipped with a zoom lens (25 mm) that captures the image of the whole plate.

We use a PCI-1407 Image Acquisition card to grab frames from the CCD camera of worms executing thermotactic migrations along the surface of the agar plate. In this way, we can measure overall thermotactic drift upwards or downwards on the thermal gradient as well as score individual movements like forward crawling, backward crawling, and turns. We have written simple particle-tracking algorithms using LabVIEW Vision Assistant to measure individual crawling trajectories.

Measurement of thermotactic migration requires setting the worms to a particular cultivation temperature (T_cult). When placed on a spatial thermal gradient, worms will track isotherms in a band of temperatures near T_cult. Worms can be set to a specific T_cult by growing them overnight at T_cult on an NGM plate with OP50. The day of the experiment, several young adult worms should be rinsed in NGM buffer and picked to a fresh NGM plate, which should then be transferred immediately to the linear thermal gradient. It is necessary to cover the plate. Usually, the thermal gradient will equilibrate after 3-4 minutes, and isothermal tracks should emerge among the worms navigating at temperatures near their T_cult (Figure 9).

Reversal behavior refers to the tendency of the worms to change their mode of locomotion from forward to backward by changing the direction of propagation of the sinusoidal body waves. Reversals can be induced, as when an acute stimulus such as touch produces an immediate reversal, or spontaneous as when the worm reverses in the absence of an acute stimulus. Spontaneous reversals reflect the integration of many aspects of both the environment and the internal state of the worm and therefore can be a sensitive measure of behavioral state.

There are two methods for scoring reversals: human observation and computer vision. Human observation involves watching a worm, typically under a dissecting microscope and counting the number of times they reverse. (Chiba and Rankin, 1990; Croll, 1975) The scoring can be aided by making a video record of the assay. Reversals run the gamut from a stutter in the forward locomotion that lasts less than a second to several seconds of backward movement, so one must operationally define what constitutes a reversal. The specific definition of a reversal is less important then its consistent application. One convenient definition for human observation is any retrograde movement that is greater than the length of the pharynx. This has the virtue of being easy to gauge and, because the pharynx is proportional to the size of the worm, is less dependent on the size of the worm.

Computer vision applications typically involve a camera mounted on a microscope, connected to a frame-grabber and image analysis software, which are in turn connected to a motorized stage. (Pierce-Shimomura et al., 1999) The image analysis software identifies the silhouette of the worm, determines how far the worm has moved since the last image was taken and commands the stage to re-center the worm in the field of view. The record of the movements of the stage is also a record of the movements of the worm. Updating the stage every second is generally a high enough sample rate to catch most reversals although some very brief reversals may be missed. (Pierce-Shimomura et al., 1999; Zhao et al., 2003) Alternatively, the worms can be recorded under low magnification with no movement of the stage and their position in the field of view recorded at each time point. To reduce computational demands, the position of the worm is usually abstracted to its centroid. Reversals are defined as changes in the bearing of the centroid (determined by the angle between the line segments defined by successive pairs of data points) greater than a certain angle, e.g., >50° (Pierce-Shimomura et al., 1999). However, the centroid can be misleading, especially when the worm is significantly bent as in the case of an omega turn, resulting in false positives and negatives. It is important to visually confirm the result of at least a sample of the reversals scored by the computer. More recently, software has been developed that abstracts the worm as a series of line segments along the midline (Baek et al., 2002; Feng et al., 2004). In principle, this should result in more accurate scoring of reversals. Various combinations of hardware have been employed for worm tracking (Hardaker et al., 2001; Pierce-Shimomura et al., 1999; Zhao et al., 2003). Software is often written by the authors and available on request (Feng et al., 2004).

Although a researcher can score one worm out a population on a plate, it is often advantageous to place worms individually on an assay plate and score them one at a time. This is because, whether scoring by eye or by computer, it is essential to control the state of the worm and its environment as much as possible. Among the variables that can affect reversal frequency are: temperature, humidity, wetness of the agar, presence of food, age of the worm, sex of the worm and gravidity of the worm (Zhao et al., 2003). A worm will also reverse if it bumps into another animal. Moreover, computer vision applications used so far only track one worm at a time. By scoring L4 larvae one can minimize the effects of age and gravidity. To control for other effects, it is useful to assay control and experimental populations on the same day using the same agar plate, preferably by alternating experimental/mutant and control worms (Zhao et al., 2003; Zheng et al., 1999).

Although one can measure reversal frequency simply as the number of reversals in a given period of time, it can be more informative to score the reversal interval, i.e. the time between reversals. By plotting a histogram of reversal intervals, one finds that the resulting distribution can best be fit by a lognormal function (Hardaker et al., 2001; Zhao et al., 2003), or the sum of two exponentials (Brockie et al., 2001; Pierce-Shimomura et al., 1999). The parameters of these distributions can vary significantly with the experimental conditions, providing information about how the probability of reversal varies over time or among populations in different behavioral states.

Reversal frequency can be measured in response to various experimental manipulations. These include: the presence of a source of chemo-attractive compounds (Pierce-Shimomura et al., 1999), the presence of drugs (in the agar) and mechanical stimulation (Zhao et al., 2003). Mechanical stimulation with an eyelash or pick requires that the lid be left off the plate; in this case care must be taken to avoid air currents that can stimulate the worms to reverse.

Reversal duration/distance, how long the worm spends in retrograde locomotion for a given reversal and how far it moves backward, are also sensitive indicators of the state of the worm. A video record of the animal's behavior traced onto acetate or a computer vision application can be used to measure the duration of a reversal and its distance (Chiba and Rankin, 1990). Reversal duration can also be scored visually with a timer and the reversal distance estimated by counting the number of body bends generated while the worm is reversing. Body bends are generally an accurate measure of distance for worms of approximately equal size. As above, control of the environment is essential to get consistent results.

Ivermectin is an antiparasitic drug that affects nematodes including C. elegans. It activates glutamate-gated chloride channels which in turn inhibit a variety of excitable cells including the pharyngeal muscle and neurons involved in locomotion (Cully et al., 1994; Dent et al., 2000). Because ivermectin prevents pharyngeal pumping, one effect of ivermectin is to cause the animals to starve. Ivermectin-induced paralysis of locomotion appears to be an effect of ivermectin on the neuronal circuit controlling locomotion rather than on the muscles themselves. Both the locomotion and eating phenotypes have been used to measure sensitivity to ivermectin.

Prepare stock a solution of the ivermectin (e.g., 10 mg/ml, Sigma-Aldrich, St. Louis, MO.) in 100% DMSO. Prepare drug plates at desired concentrations (EC₅₀ for wild type worms is ∼1 ng/ml) by adding 1% DMSO and then ivermectin from a stock solution to melted NGM agar. For dose response curves, which require preparing plates at many drug concentrations, we prefer to pour 35 mm plastic Petri dishes. Seed plates with bacteria a day or two in advance. If only testing a few drug concentrations, pick eggs from the strain of interest directly onto the drug plate. If many drug concentrations are to be tested, egg prep the strain of interest using alkaline bleach and resuspend eggs from one 60 mm plate in ∼1 ml sterile M9 buffer. Estimate the concentration of eggs by placing ∼30 ml of suspension from a well shaken tube (always re-shake the suspension because eggs will settle), put on an agar plate and count the number of eggs. Try to have between 50 and 100 eggs per 30 ml; dilute suspension if necessary. Pipette 30 ml of egg suspension per drug plate. Be sure to include control plates lacking drugs (ideally with 1 % DMSO although in practice the DMSO seems not to have much effect). Allow worms to grow for 3-4 days and score growth. A simple method of scoring growth is to count the number of hermaphrodites that have become gravid adults, although there is any number of criteria one could use to measure growth. It is essential however to include negative controls (no drug) against which to normalize the data. We typically score 6 plates per ivermectin concentration using at least two different egg preps. The growth assay is most sensitive to pharyngeal function and primarily reflects the effect of ivermectin on feeding (Dent et al., 2000).

To measure effects of ivermectin on locomotion, a motility assay can be used. (Arena et al., 1995) Wash worms off of plates with 25 mM HEPES buffer and incubate in liquid with drug up to 16 hours. Score the percentage of worms that are still thrashing in liquid.

C. elegans locomotion involves bouts of forward locomotion punctuated about twice a minute by episodes of turning that have been termed “pirouettes” (Pierce-Shimomura et al., 1999). Pirouettes appear to be the final common path for approach behaviors such as chemotaxis, and avoidance behaviors such as the response to high osmolarity. Because the efficiency of either type of behavior depends on the change in orientation produced by individual pirouettes, it is often of considerable importance to quantify the effects of pirouettes on the direction of locomotion.

Pirouettes are defined in terms of omega turns and reversal turns, here collectively referred to as “sharp turns”, and the intervals between them called “runs”. The study that originally proposed the concept of a pirouette (Pierce-Shimomura et al., 1999), noted that the distribution of run durations was well fit by the sum of two exponentials, suggesting that there are two distinct kinetic states, one that emits mainly short runs and another that emits mainly long runs. Pirouettes were defined as either a single sharp turn or a series of sharp turns each separated from its neighbors by short runs. Because either type of run state is theoretically capable of generating a short or long run, it is not possible to definitively assign a run of a particular duration to one state or the other. However, one can minimize the number of classification errors by computing a critical run duration, called t_crit, below which a run is assumed to be generated by the short run state. The equation for t_crit is shown in Figure 10 where A_l and A _s are the respective amplitudes of the long-and short-run exponential components and _l and _s are their respective time constants. In the original study, t_crit was found to be approximately 6 sec, but t_crit is likely to be different in different types of experiments. A first step therefore in any study of turning behavior will be to obtain the distribution of run durations under appropriate experimental conditions, and to compute the value to t_crit to be used in the analysis of data generated under these conditions.

The following summarizes the steps required to identify pirouettes and to measure their effects on orientation.

The number of eggs that accumulate inside an adult is a function both of the rate of egg production and the rate of egg-laying. If you want to use this assay to quantify differences in the rates of egg-laying between various strains, you have to first be certain that the strains you're assaying have a similar rates of egg production. In practice, I do this by measuring brood sizes on all the strains being analyzed to make sure they are similar.

This assay can be used to determine if various compounds can stimulate egg-laying in liquid (Weinshenker et al., 1995). M9 buffer inhibits the egg-laying behavior of wild-type animals. Serotonin and other drugs can overcome this inhibition to stimulate egg-laying in M9 buffer. This assay was originally used to determine if egg-laying defective (Egl) animals had neuronal, vulval or muscle defects. (Trent et al., 1983) Egl animals with vulval muscle defects will not lay eggs when exposed to serotonin. If the muscles of an Egl animal could be induced to contract to release eggs when exposed to exogenous serotonin then the Egl animal was thought to have functioning muscles but defective HSN neurons that did not release serotonin.

Treating animals with drugs to block serotonin removal from the synapse, such as fluoxetine, can potentiate serotonin signaling to stimulate egg-laying. Mutant animals that do not lay eggs when exposed to fluoxetine are thought to have HSN neurons defective in serotonin release.

Imipramine was originally used potentiate serotonin, signaling but fluoxetine has since been shown to be a more selective serotonin reuptake inhibitor (SSRI) and should be used instead of imipramine (Weinshenker et al., 1995).

Data can be presented as the mean eggs laid per animal.

N2 animals laid and average of 13 eggs in serotonin compared to very few in M9 (Desai and Horvitz, 1989).

N2 animals in 1 mg/ml of fluoxetine laid 7.94±1.16 eggs (Weinshenker et al., 1995).

Alternatively data can be presented by binning animals into four categories based on the number of eggs laid, >7 eggs, 4-7 eggs, 1-3 eggs or zero eggs (Trent et al., 1983).

Wells of a flat-bottomed 96-well plate received a total of 200 ul medium, including 167 ul culture medium (M9 buffer + E. coli broth + 6 ul DMSO +/− drug as appropriate) and 33 ul worm suspension (50–90 worms in M9 buffer). Chitinase activity was measured after a significant proportion of the worms were gravid and L1 larvae were apparent (this is not time-dependent to any degree; the enzyme activity is very stable in the culture supernatants).

To measure enzyme activity, 10 ul of a 0.8 mM solution of chitinase substrate (4-methylumbelliferyl-beta-D-N,N′N″-triacetylchitotrioside, Sigma Chemical Co.) is added to each well. Plates are incubated at 37 C for 1 hr. The reaction is terminated by addition of 100 ul alkaline buffer (1 M glycine/1 N NaOH, pH 10.6). Fluorescence is determined on a 96-well fluorimeter at excitation 360/40, emission 460/40.

Egg-laying behavior provides an important phenotypic assay for genetic analysis of nervous system function. In particular, egg-laying is a useful behavioral readout for the activity of the Go/Gq signaling network in neurons and muscle cells, and for neuromodulation by serotonin, acetylcholine, and neuropeptides. The most common assays for egg-laying phenotypes count the number of eggs retained in the uterus, an indirect measure of overall egg-laying rate,. More detailed assays of the temporal pattern of egg-laying can in principle make it possible to distinguish effects on different egg-laying signal transduction pathways. In the presence of abundant food, wild-type animals lay eggs in a specific temporal pattern: egg-laying events tend to be clustered in short bursts, or active phases, which are separated by longer inactive phases during which eggs are retained. This egg-laying pattern can be accurately modeled as a three-parameter probabilistic process, in which animals fluctuate between discrete inactive, active, and egg-laying states. Mutations affecting different cellular or molecular components of the egg-laying circuitry have different effects on this pattern. However, because egg-laying events are infrequent, direct monitoring of the temporal dynamics of egg-laying requires the use of long-term video recording and analysis.

The essential piece of equipment is a microscope equipped for video recording of worms at approximately 50 X magnification (i.e. high enough to easily see eggs on a video image). Some mechanism for recording video (e.g., a VCR) is also required. Most applications of this technique have also utilized a motorized stage controlled by an automated tracking system that allows the worm to be continuously maintained in the field of view during a long recording. Several systems appropriate to this purpose have been described. In the absence of such a tracking system, it is in principle to manually maintain the animals within the field of view throughout the recording, though the required durations of these recordings would make this a tedious task.

Assay conditions and growth media: Unless otherwise noted, nematodes are grown and assayed at room temperature on standard nematode growth medium (NGM) seeded with E. coli strain OP50 as a food source. Typically first-day adult hermaphrodites (obtained by picking L4 hermaphrodites the previous day) are used. Drugs (e.g., serotonin or levamisole) can be added directly to the agar before pouring.

Recording and analysis of C. elegans egg-laying behavior: Single animals are observed as they crawled across a standard 8 cm agar plate using a stereoscopic dissecting microscope equipped with a computer-controlled motorized stage. A video camera records the animal's behavior at standard video frame rate, and a time stamp is used to indicate the time of egg-laying events on the resulting videotape. For the tracking systems commonly used for these analyses, egg-laying times are determined by manually replaying the videotapes. An egg-laying event is defined as a muscle contraction that leads to the expulsion of one or more eggs. An automated algorithm for egg-detection has been described (Geng et al., 2005), though it has not been extensively used for actual behavioral analysis.

ML estimates of model parameters from real data. For wild-type N2 and most N2-derived mutant strains, egg-laying occurs in short bursts, or active phases. The onset of the active phase can be modeled as an exponential random variable with a time constant on the order of 20-30 minutes, while egg-laying events within the active phase model as an exponential random variable with a time constant of around 20 seconds.

For egg-laying events that conform to this pattern, we can prove that the random variable governing the egg-laying interval times has the following probability density function (p. d. f.) (Zhou et al., 1998):

One way these parameters can be estimated for a given set of data is using the maximum likelihood method. This involves maximizing the likelihood function for the above probability density function for the observed interval data. Given N observations of the intervals x = [x₁,x₂, …, x_N], the likelihood function is given by:

which is a function of the parameter vector θ = [p, λ₁, λ₂]. The maximum likelihood 'estimate' is defined as the θ that maximizes f(x|θ) over all possible θ's. The bi-modal peak information provided by the histogram of the log intervals was used to obtain a rough estimate of θ and initialize the non-linear ML algorithm, which adjusted the parameters to maximize the likelihood function.

For some cases (e.g., N2 animals on serotonin, some Egl-c mutants, some non-N2 nematode strains) a pattern resembling a homogeneous Poisson process is observed. The time constant for egg-laying in these cases can be estimated by performing a weighted least-squares linear regression to the log tail distribution (see below).

Log tail distribution: One useful method for assessing the temporal pattern of egg-laying is the log tail distribution. It can be shown that the log tail probability (i.e., the log of the probability that a given interval will be longer than time x) for intervals generated by animals fluctuating between active and inactive egg-laying states is given by:

When the interval duration t is plotted against the log of the frequency of intervals greater than t, a biphasic plot will be obtained. For the long intervals (x >5/λ₁), ln Pr(X≥x) becomes approximately linear with respect to x, with slope -pλ₂. Thus, the slope of this plot is inversely proportional to the average duration of the inactive phase. If the pattern follows a single exponential distribution, the log tail distribution will be linear with a negative slope equal to the time constant for egg-laying.

Histogram of log intervals: Another way to visualize the overall distribution is with a histogram of interval times. Since the intervals between eggs laid are clustered at short intervals and sparse at long intervals, it is more meaningful for reasons of better dynamic range to study the distribution of the log intervals (denoted by Y = ln X). Animals showing an N2-like clustered pattern exhibit a bimodal distribution, which can be shown to be represented by the following p. d. f. of Y:

We can show that when λ1 and p λ2 are sufficiently different, fY(y) peaks at y = ln(1/λ1) and ln(1/pλ2) with corresponding peak heights k1/e and k2/e. When the values of λ1 and pλ2 are close (or when egg-laying events are unclustered), only a single peak is observed.

Advantages: One significant advantage of this type of analysis is the ability to more precisely classify the phenotypes of egg-laying defective mutants. Some mutants with slow egg-laying rates (e.g., mutants defective in HSN function) show a lengthened time constant for the long interburst intervals but normal or even a shorter time constant for the short, intraburst intervals. In contrast, other mutants (e.g., mutants affecting muscle calcium channels) exhibit an unclustered, or less clustered, egg-laying pattern. A second advantage is a greater sensitivity in detecting subtle egg-laying phenotypes. For example, mutants defective in vulval muscle nicotinic acetylcholine receptors show increased time constants for the short intraburst intervals, even though they are not visibly Egl. Finally, this method has the advantage of being a direct measure of egg-laying behavior. Unless the uterus is depleted of eggs, it does not appear that the number of eggs in the uterus significantly alters the probability of an egg-laying event; thus, as long as a mutant retains nearly normal numbers of eggs there is less concern that effects on the rate of egg-production will confound the scoring of an egg-laying phenotype.

Disadvantages: As noted above, mutants that do not retain many eggs are difficult to analyze with this method. Practically, the analysis of video recordings is relatively time- and labor-intensive. To obtain reliable timing parameters, it is usually necessary analyze at least 100 intervals between egg-laying events, a number that generally requires analysis of at least 30 hours of recording. Thus, phenotyping methods based on evaluating the number or developmental stage of eggs retained in the uterus are far more effective assays for identifying egg-laying defective mutants in screens.

See also “Leaving food by males, not hermaphrodites” in Chemosensation section.

A “good” turn occurs when the male tail remains in contact with the hermaphrodite throughout the turn and continues backing on her opposite side after the turn is completed. A “sloppy” turn occurs when the male's tail temporarily loses contact, but regains contact on the opposite side quickly because of a favorable trajectory (weak or late curling occurred, but was sufficient). A “missed” turn occurs when the male's tail sails off the end of the hermaphrodite, completely losing contact. This can occur either because the tail curl is not as tight or because the tail curl is late. With missed turns, the male often moves off away from the hermaphrodite and only resumes “turning” upon contact with this or another hermaphrodite in the mating spot (See Figure 4 of Loer and Kenyon, 1993).

Plastic Petri plates (6-cm-diameter), NGM agar medium, Escherichia coli strain OP50 and compound microscope with moveable stage, video camera, and monitor.

The day before, put 5 μl of bacteria from an overnight culture of E. coli strain OP50 in the center of each plate, creating a 5 mm circle of bacteria. Leave the plates on the bench overnight. Move individual larval (L4 stage) worms to separate plates about 12 hours prior to the assay. At the beginning of the experiment, place a single hermaphrodite and a single male on opposite sides of a bacterial patch. Put the whole Petri dish on the microscope stage. Trials can be videotaped through the 5X objective of a compound microscope. The male is kept in the field of view using the moveable stage, and the behaviors of the worms are videotaped for later analysis.

Hermaphrodite sprinting is defined as a visually distinct increase in the velocity of hermaphrodite swimming (at least 2 seconds in duration) while the male is in contact but prior to sperm transfer.

Same as for the Sprinting assay (above).

The setup is the same as for the Sprinting assay above. Video tape the activities of the worms for one hour or 15 minutes after successful sperm transfer. Fifteen minutes is sufficient to allow sperm to migrate beyond the opening of the vulva making sperm ejection unlikely. After sperm transfer is observed, keep the hermaphrodite vulva in the field of view for 15 minutes. The male ejaculate will displace eggs in the hermaphrodite uterus and can be seen through the body wall as an area filled with clear fluid. Individual sperm cells may be visible as small round cells. Score sperm ejection when fluid is expelled from the vulva. Often a singly mated hermaphrodite ejects sperm multiple times, therefore the number of sperm ejection bouts can also be recorded.

An isolated male on a restricted food source on an agar plate eventually leaves the food and wanders about the agar surface. This is a sex-specific and age-specific behavior not expressed by adult hermaphrodites or juveniles of either sex. Moreover, males do not leave a food source if hermaphrodites are present (providing a means for assaying male detection and response to hermaphrodites, see below). Since wandering behavior is not expressed in the presence of hermaphrodites, it appears to be a mate-searching behavior. Mate-searching is assayed by measuring the rate at which isolated males leave a food spot in the Leaving Assay (Adapted with permission from Lipton et al., 2004, Copyright 2004 by the Society for Neuroscience). Mutant males that do not leave food are said to have a Las phenotype (leaving assay defective).

Adapted with permission from Lipton et al. (2004). Contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

Contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

Males can detect hermaphrodites and chemical substances (pheromones) produced by hermaphrodites. This ability can be assayed using a variation of the Leaving assay described above (Lipton et al., 2004) or by observing male behavior on agar conditioned by hermaphrodites (Simon and Sternberg, 2002). These assays may reveal responses to multiple signals, including non-sexual chemical cues, for example those involved in social interactions (de Bono and Bargmann, 1998), as well as sex pheromones.

Adapted with permission from Lipton et al. (2004), Copyright 2004 by the Society for Neuroscience. Contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

From Simon and Sternberg (2002), contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

Contributed by A. Barrios and Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

Compiled and contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

The Holding Assay (Simon and Sternberg, 2002) is identical to the Response to Pheromone Assay, Variation 1, with the following modifications:

Compiled and contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

(Similar to the Tendency to Mate Assay below.)

The Attraction Assay (Simon and Sternberg, 2002) is similar to the Holding Assay with the following modifications:

Compiled and contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

Male copulatory behavior can be broken down into a series of sub-behaviors (Barr and Sternberg, 1999; Emmons and Sternberg, 1997; Liu and Sternberg, 1995; Loer and Kenyon, 1993). Assays for each of the main subdivisions of the behavioral series are given below. These subdivisions may themselves be composed of further definable sub-steps. In addition to copulation itself, the time it takes for a male to begin copulation when in the presence of hermaphrodites (Tendency to Mate) can be determined.

Compiled and contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

This assay really measures male fertility rather than behavior. It gives an overall indication of mating ability if the gonad and morphology of the male are normal. It has been traditionally used to score the mating ability of mutant males (Hodgkin, 1983) and is useful for determining whether a particular genotype may be used as the male in a genetic cross (Hodgkin et al., 1988).

(Similar to the Attraction assay above and the Response assay below.)

Contributed by L. Jia and Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

By K. Knobel and M. Barr, a variation of Liu and Sternberg (1995) and Barr and Sternberg (1999) (Lov phenotype), contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

Response is a complex behavior. It begins when a male touches a hermaphrodite with his tail and stops his forward locomotion. He next assumes a specific posture in which the posterior body curves slightly ventrally and the fan is pressed against the side of the hermaphrodite. He then begins to back up by propagating anteriorly-moving waves through the anterior part of the body, while maintaining the posterior body pressed against the hermaphrodite. All of these steps must be achieved to score a positive response.

By E. Peden and M. Barr (Peden and Barr, 2005), contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

By E. Peden and M. Barr, contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

By T. Liu and M. Barr, contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

Turning is a complex behavior expressed during the response phase of copulation when the male encounters either end of the hermaphrodite. Males first must detect the end (often they detect its approach before they actually reach it, at about 10% or 90% of body length). Detecting the end stops backing and triggers a deep ventral bend or arch of the posterior body region so that the fan slides around to the other side of the hermaphrodite. Next, the deep ventral bend or arch is propagated along the body to the head, while the posterior body region is held in the posture typical of backing and the fan slides along the body. When the entire body is around the hermaphrodite, normal backing is resumed.

Contributed by R. Garcia and Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

Spicule insertion is probably the most difficult step of mating behavior for the male to accomplish. This behavioral step consists of two motor behaviors: the spicules prodding the vulval slit rhythmically, and the spicules penetrating the vulva. Both of these behaviors require high magnification and cooperative hermaphrodites for consistent measurement.

Contributed by R. Garcia and Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

Normally, a male keeps his copulatory spicules within his tail prior to and between mating events. Certain mutations may disrupt this regulation and cause males to protract their spicules spontaneously in the absence of hermaphrodites. The spicule muscles and neurons are quite sensitive cells and precocious spicule protraction can provide a readout to determine if a mutation can enhance cell excitability.

Contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

Acetylcholine agonists such as levamisole, arecoline, nicotine and oxotremorine M can cause adult males to protract their spicules. These drugs act on the spicule muscles and some drugs such as oxotremorine M also act on the neurons that innervate them. Each of these drugs stimulate spicule protraction via a different mechanism; thus drug-induced spicule protraction can be used as a readout to determine if a certain mutation may interrupt different types of pre-and post synaptic cholinergic signaling.

By G. Schindelman and P. Sternberg.

Compiled and contributed by Scott Emmons, Albert Einstein College of Medicine, Bronx, NY, USA, May, 2005.

I thank the contributors listed above for their help in assembling this list of behavioral assays (Scott Emmons).