|Home | About | Journals | Submit | Contact Us | Français|
Stress, or threats to homeostasis, is a universal part of life. Organisms face changing and challenging situations everyday, and the ability to respond to such stress is essential for survival. When subjected to acute stress, the body responds molecularly and behaviorally in order to recover a steady state. We developed a simple and robust assay of behavioral plasticity for Drosophila larvae in which well-defined behavioral responses and recovery can be observed and quantified. After experiencing different control and bright light treatments, populations of photophobic fly larvae were placed a defined distance from a food source to which they crawled. Half-times (t½), or times at which half the total number of larvae reached the food, were used to compare different treatments and larval populations. Repeated control treatments with a main experimental strain gave tight, reproducible t½ ranges. Control treatments with the wild type strains Oregon R and Canton S, the “rover” and “sitter” alleles of the forager locus, and eyeless mutants gave comparable results to those of the experimental strain. Exposure to bright light for a defined time period resulted in a reproducible slowing of locomotion. However, given a defined recovery period, the larvae recover full, normal locomotion. In addition, bright light treatments with Canton S gave comparable results to those of the experimental strain. Eyeless mutants, which are partially blind, do not show a response to bright light treatment. Thus, our assay measures the behavioral responses to bright light in Drosophila larvae and therefore might be useful as a general assay for studying behavioral plasticity and, potentially, adaptation to a stressful stimulus.
Hans Selye (1956) first coined the term “stress” in his book The Stress of Life and described it as the “non-specific response of the body to any demand.” By definition, stress is a state in which homeostasis is threatened in either a perceived or physical manner (Pacak and Palkovits, 2001). When the cause of stress is uncontrollable, unpredictable, and of short duration (acute stress), the body reacts in an adaptive, compensatory manner in order to regain or maintain its homeostatic state. This natural stress response can be molecular and/or behavioral, and the organism may recover from the acute event without any lasting effects. This process involving the adaptive physiological response to acute stress is referred to as allostasis (Sterling and Eyer, 1988).
The ability to adapt to changing, challenging situations and environments is integral to an organism's survival (McEwen, 1999). The body's molecular and behavioral responses to stressful circumstances are advantageous because they allow for brisk CNS changes followed by rapid restoration of homeostasis. However, these responses are a “double-edged sword” (McEwen, 1998) – while they promote survival, they can also have long-term, detrimental effects on neuronal function. When the uncontrollable stressor is repeated or of longer duration, the stress becomes chronic and can lead to allostatic load and molecular changes in the brain. The stress becomes remembered and learned, and the stress response can be provoked by non-threatening events, such as in post-traumatic stress disorder. In Aplysia, a transient shock is translated to a chronic ‘anxiety’ state, both behaviorally and molecularly. In the three forms of learning examined in the Aplysia—habituation, sensitization, and classical conditioning—two stages of memory storage were observed: a transient memory that lasts minutes and an enduring memory that could last days or even weeks (Pinsker et al., 1970, 1973; Carew et al., 1972; Frost et al., 1985). Short-term memory stems from changes in synaptic strength between interconnected neurons (Castellucci et al., 1970; Kupfermann et al., 1970), while the conversion of a short-term memory to a long-term one requires protein synthesis and the formation of new neural connections (Castellucci et al., 1989). In studying the molecular biology behind this phenomenon, it was found that the neurotransmitter serotonin plays a key role in learning and the formation of both short and long term memories. Serotonin increases presynaptic cAMP, which activates PKA and leads to synaptic strengthening (Byrne and Kandel, 1996). Repeated puffs of serotonin activate PKA and lead to a tightly controlled cascade of gene activation that gives rise to the growth of new synaptic connections (Schacher et al., 1988; Dash et al., 1990; Glanzman et al., 1990; Bailey et al., 1992; Bacskai et al., 1993; Bailey and Kandel, 1993; Kaang et al., 1993; Martin et al., 1997a), and long-term changes in synaptic function and structure are confined to synapses stimulated by serotonin (Martin et al., 1997b; Casadio et al., 1999).
The Drosophila melanogaster larva undergoes two stages before pupation and metamorphosis: foraging and wandering (Sokolowski et al., 1984). The foraging stage spans most of the larva's life, from the beginning of first instar to late third instar, in which it is feeding and burrowed deep into the food substrate. During this time, Drosophila larvae are photophobic and will actively move away from bright light (Lilly and Carlson, 1990; Gordesky-Gold et al., 1995; Sawin-McCormack et al., 1995). Approaching late third instar, larvae enter the wandering stage where they leave the food to find an appropriate pupation site. At the onset of wandering, their repulsion to light decreases until the larvae behave indifferently towards bright light stimuli (Sawin-McCormack et al., 1995).
The Drosophila larva possesses a simple olfactory system (Python and Stocker, 2002). The major components of the larval chemosensory system consist of the dorsal organ, the terminal organ, the ventral organ, and a series of pharyngeal sensilla (Stocker, 1994; Cobb, 1999). The dorsal and terminal organs together form the antennomaxillary complex and are involved in olfaction and taste, respectively (Singh and Singh, 1984; Heimbeck et al., 1999; Oppliger et al., 2000). The dorsal organ, which consists of the larval antenna and main olfactory organ, contains 21 odorant receptor neurons, while the terminal organ contains roughly 80 gustatory neurons (Tissot et al., 1997; Heimbeck et al., 1999; Python and Stocker, 2002).
Bolwig's organ, the larval eye, is the light-sensing organ of the Drosophila larva and comprises the larval visual system (Bolwig, 1946). It is composed of two bilateral clusters of 12 photoreceptor cells in the larval mouth hooks (Steller et al., 1987; Hofbauer and Campos-Ortega, 1990). The larval optic nerve is formed by the photoreceptors' axons and innervates the optic lobe primordium portion of the brain lobes (Green et al., 1993; Campos et al., 1995).
Many different assays have been developed and used to study behavior in Drosophila larvae. For example, choice assays have been used to study photobehavior (Lilly and Carlson, 1990), olfactory response (Shaver et al., 1998), gustatory response (Heimbeck et al., 1999), visual learning (Gerber et al., 2004), olfactory learning (Scherer et al., 2003), and thermobehavior (Liu et al., 2003). Path length as-says have been used to examine foraging behavior (Pereira and Sokolowski, 1993; Pereira et al., 1995; Sokolowski et al., 1997) and photobehavior (Busto et al., 1999). Locomotion, crawling, and turning behavior have been studied using touch-sensitive assays (Caldwell et al., 2003; Tracey et al., 2003) and plate assays (Heiman et al., 1996; Yang et al., 2000; Suster et al., 2003).
Hypergravity exposure (Le Bourg and Minois, 1999) and starvation/desiccation (Hoffmann and Harshman, 1999)have been used to examine stress responses in adult flies. However, a method has not been developed to examine stress responses in Drosophila larvae. Here we have developed a locomotion assay and scoring method that is not only useful in studying behavior, but can also be used in conjunction with bright light to examine behavioral responses in Drosophila larvae.
Fly strains were maintained at room temperature (25±2 °C) in plastic vials or glass bottles containing a standard cornmeal/molasses Drosophila medium. Eggs from adult flies 1–10 days old were collected on fresh egg plates (molasses-agar media in 35 mm×10 mm dishes) with a small amount of yeast paste in the center. The plates were replaced after 24-hour incubation periods and kept at room temperature, while the hatched larvae were allowed to grow. Early third instar larvae (72–78 h) from these plates were tested in the experiments.
Homozygous strains of UAS-mCD8-GFP; ddc-GAL4 flies were used for the treatments and assays. In addition, the following strains were also examined as controls: wild type strains Canton S and Oregon R, the “rover” (forR/forR) and “sitter” (forS/forS) alleles of the forager locus, eyeless (Drosophila pax-6 homolog, ey/ey).
Because larvae spend most of their lives burrowed in food, they are covered in food substrate when immediately removed from the medium. This poses a problem in crawling assays, as larvae covered in yeast will leave yeast trails as they crawl, causing other larvae to follow their paths or be attracted to them. To avoid this problem, the larvae were washed in distilled water after collection.
Using a small moistened paintbrush, approximately 200–400 larvae were collected from the molasses agar plates and placed in a small amount of distilled water. After gently stirring the water with the brush to aid in washing the larvae, the water was removed and drained using a 1000 μL Pipetman. Clean distilled water was added again, and the washing procedure was repeated two to three times until the larvae were clean of yeast.
The effects of the bright light treatments on larvae crawling were assessed and quantified using a locomotion assay. The apparatus was a 100 mm 15 mm dish composed of 2.3% agar with a circular hole (25×mm diameter) dug out in the center. A small amount of cold yeast-water paste (50:50, yeast from Lesaffre Yeast Corporation) was spread along the edges of the hole prior to running the assay. In addition, the larvae were gathered and put onto a spatula for transfer onto the plate with a brush. At the start of the assay, the larvae were placed and spread out 5 mm from the edge of the plate. The assay was run for 60 min. To allow multiple simultaneous runs of the assay and faster counting of the larvae, the assays were recorded in Quicktime movie format (.mov) using Apple iSight webcams and SecuritySpy software on a Macintosh computer.
The larvae were scored by counting the number to reach the edge of the yeast within each minute of the assay. Larvae that crawled out of the yeast were scored only once. Those larvae that did not make it to the yeast but were still mobile within the 60 min were marked with an infinity time. Those that were not crawling (from possible injury during collection/washing or treatment) were disregarded from the assay.
No density-dependent effects were observed in any of the assays. Increasing or decreasing the number of larvae tested in the assays did not affect crawling speed or arrival times to the yeast.
For each assay, an arrival-time or distribution plot (number of larvae scored over time) was drawn. In addition, a “half-time”, or t½, was manually determined by interpolation from the raw data and used as a comparison tool. The half-time corresponds to the time at which half the total number of larvae in the assay reached the yeast.
Average arrival times and logarithmic slopes were also determined in these experiments as potential comparison measures. Although all gave similar statistical results, the t½ was used as the main comparison measure.
All data were normally distributed and were analyzed using one-way Analyses of Variance (ANOVA). To determine which data sets had significantly different means, the Tukey-Kramer Multiple Comparisons Test was performed as a post-test.
Larvae from homozygous strains of UAS-mCD8-GFP; ddc-GAL4 flies were used as the main experimental strain in the treatments and assays. This particular strain was chosen because it is isogenic and well-characterized. As serotonergic and dopaminergic neurons are labeled with green fluorescent protein in this strain, it will be used for future anatomical investigations.
Because larvae are repulsed by light, we hypothesized that bright light from a Fostec high intensity light source applied directly onto the larvae would be an effective cause of stress for the animals. The larvae were kept in approximately 500 μL of distilled water during the light exposure, both to buffer environmental temperatures changes and to prevent migration away from the light source.
In a set of pilot experiments, the duration of light to use was determined. Periods of 0 min (no light, wild type), 10, 20, and 30 min of light were applied onto the larvae immediately following washing, after which the larvae were observed in the locomotion assay and compared. Ten minutes of light gave the maximum behavioral response. However, increasing the duration of light gave a response that reverted back to a wild type, no-light response. The larvae's loss of response to increased amounts of light is probably due to desensitization to the light after 10 min of exposure. After the initial 10 min, the larvae may start recovering from the light and thus begin to show more wild-type responses.
Because we were looking to develop an assay that characterizes both stress and behavioral plasticity in fruit fly larvae, we hypothesized that the larvae would be able to recover after light exposure. In another set of pilot experiments, we tested the larvae's recovery from the bright light and determined an amount of delay time that resulted in a fully recovered response. After exposure to 10 min of light, the larvae were given rest periods in which they were allowed to roam freely in a covered and empty 35 mm × 10 mm Petri dish. A 40-min delay or rest period after light was sufficient to give responses that were consistent with wild type responses.
Initially, an assay time of 30 min was sufficient to account for almost all of the larvae and minimize the number marked with an infinity time. However, after determining bright light and recovery times and incorporating them into the methods, a much longer assay time was needed, especially for the bright light assays. An assay time of 60 min was found to be sufficient.
An overview of the complete experimental protocol is shown in Fig. 1.
In order to verify that the washing and periods in distilled water had no significant effects on larval behavior in our assay, we conducted several sets of control experiments without light. Before being tested in the behavioral assay, the larvae underwent a 50-min treatment period that included 10 min in distilled water (approximately 500 μL) and a total of 40 min of rest, all of which were conducted in partial dark. These intervals were determined based on the observations and results from the pilot experiments described above. The 10 min in distilled water were administered at three different time points: 40 min (rest before water), 20 min (intermediate water), and 0 min (water before rest). During the rest periods, the larvae were allowed to roam freely in a covered and empty 35 mm × 10 mm dish. Small amounts of distilled water were used to help collect the larvae after rest periods. A diagram of the control treatments can be seen in Fig. 2A.
Using larvae from the homozygous UAS-mCD8-GFP; ddc-GAL4 lines, each control treatment was tested in the behavioral assay multiple times. A raw data plot from a typical control assay is shown in Fig. 3A. The half-times, or t1/2 values, of all the control assays are listed in Table 1. The t1/2 values of all sets of experiments gave a tight range and reproducible data (Fig. 4). The data yielded comparable results and were not significantly different from one another (one-way ANOVA, P > 0.05). Therefore, the washing and treatments do not significantly affect larval behavior in this assay.
Bright light treatment experiments followed the same procedures as the control treatments outlined in Section 3.2.1, except larvae were exposed to bright light instead of partial dark while kept in distilled water. The larvae were still kept in partial dark during rest periods. Three modes of bright light and delay were examined. To examine these modes, bright light periods for the duration of 10 min were administered at three different time points within a 50-min window: at 40 min (no delay), 20 min (intermediate delay), and 0 min (long delay). Therefore, all larvae experience matching amounts of light and rest but at different times and orders. A diagram of the bright light treatments can be seen in Fig. 2B.
Populations that endured bright light exposure with no delay yielded half-times that were significantly different from those of the control populations. A raw data plot from a typical bright light, no delay assay is shown in Fig. 3B. These differences subsided with a long delay period from the light. Populations with intermediate delay gave results that were midway between no and long delay. Therefore, the intermediate and long delay periods administered after bright light exposure resulted in intermediate and full recovery from the light, respectively. Populations with no delay experienced no recovery. These results are listed in Table 2 and are shown in Fig. 4.
All controls yielded comparable results and were not significantly different from one another (one-way ANOVA, P > 0.05). The differences among the bright light, no recovery populations and the controls were significant (one-way ANOVA, Tukey comparison, P < 0.001). Bright light populations that experienced full recovery were not significantly different from the controls (P > 0.05). The bright light, intermediate recovery populations were significantly different from the controls (P < 0.001) as well as from the bright light, no recovery populations (P < 0.001).
The control treatments “rest before water” and “water before rest” were roughly tested with the wild type strains Canton S and Oregon R and the forager alleles forR and forS. The results from all the strains, listed in Table 3, yielded a narrow data range (values between 4.043 and 4.573) which fit nicely into the UAS-mCD8-GFP; ddc-GAL4 controls range (Fig. 4).
The bright light treatments with no recovery and full recovery were each tested five times with the wild type strain Canton S. These results are listed in Table 4. The half-times from the bright light with no recovery Canton S populations were significantly different from those of the UAS-mCD8-GFP; ddc-GAL4 control populations (one-way ANOVA, Tukey comparison, P < 0.001) but not significantly different from those of the UAS-mCD8-GFP; ddc-GAL4 bright light, no recovery populations (P > 0.05). This indicates that the no recovery bright light populations from both Canton S and UAS-mCD8-GFP; ddc-GAL4 yield comparable results.
In addition, the half-times from the full recovery light populations of Canton S were not significantly different from those from the controls and full recovery light populations of UAS-mCD8-GFP; ddc-GAL4 (P > 0.05 for both). This indicates that the full recovery bright light populations from both Canton S and UAS-mCD8-GFP; ddc-GAL4 yield comparable results.
Drosophila larval populations show reduced migration toward food after bright light exposure. To confirm the bright light as the source of the behavioral responses seen in the UAS-mCD8-GFP; ddc-GAL4 line, eyeless mutant strains were also tested. Due to an impaired visual system, eyeless mutants are partially blind. If the reduced migration resulted from bright light exposure, eyeless mutants should not exhibit as large of an adaptive response to the bright light.
In a rough preliminary phenotype test to demonstrate their partial blindness, approximately 100–150 eyeless larvae were placed (after washing) on a 100 mm × 15 mm agar plate on which a beam of light 15 mm in diameter was shone. This was also done for UAS-mCD8-GFP; ddc-GAL4 larvae, which were used as the control. The number that crossed the beam of light, as measured by the number of trails left in the agar, was much lower for the control than for the eyeless mutants. This confirms a reduced visual input in the eyeless larvae. The phenotype tests were conducted in partial dark.
In the light-treated groups, the eyeless mutants showed greatly reduced behavioral responses compared to those of the UAS-mCD8-GFP; ddc-GAL4 line. The results are listed in Table 5 and shown in the last two columns of Fig. 4. The “rest before water” control and the “bright light with no recovery” treatments were each repeated five times, while the other four treatments were tested once. The control treatments gave half-times (4.382 ± 0.0991) that were comparable to those from the control treatments of the wild type strains, the “rover” and “sitter” alleles of the forager locus, and the UAS-mCD8-GFP; ddc-GAL4 line. The bright light, no recovery eyeless populations produced half-times (4.733 ± 0.1259) that were somewhat increased, but not significantly different, from the control treatments and significantly less than the bright light UAS-mCD8-GFP; ddc-GAL4 populations. The bright light eyeless population with intermediate recovery gave a t1/2 value (4.023) comparable to that of the eyeless intermediate control (3.898). The bright light eyeless population with full recovery gave a t1/2 value (4.500) comparable to that of the “water before rest” eyeless control population (4.085).
The “rest before water” control and the “bright light with no recovery” populations with the eyeless mutants were not significantly different from one another (one-way ANOVA, P > 0.05) or from the control and full recovery populations of the UAS-mCD8-GFP; ddc-GAL4 line (P > 0.05) but were significantly different from the bright light, no recovery and bright light, intermediate recovery populations of the UAS-mCD8-GFP; ddc-GAL4 line (P < 0.001).
All treatments of the eyeless mutant produced results that emulated those of the control treatments for the UAS-mCD8-GFP; ddc-GAL4 line, the wild type strains, and the “rover” and “sitter” alleles of the forager locus. Exposure to light did not give a significant behavioral response. Therefore, the behavioral response of non-visually impaired larvae is due to a light responsivity and not due to high temperature and other potential effects of the treatment.
In our experiments, we developed a locomotion assay for Drosophila larvae that can be used to assess behavioral effects and various alterations in the nervous and sensory systems during Drosophila development. Using an agar plate with a yeast paste hole dug out in the center, large quantities of larvae were tested in a behavioral assay after undergoing different control and bright light treatments. The treatments were compared using the half-times, or times at which half the total number of larvae reached the yeast, of the populations.
Our treatments and locomotion assay test the behavioral response and recovery from bright light in Drosophila larvae. The assay is simple and proved to be robust. In our experiments, a well-defined behavioral response was observed – a significant increase in the t1/2 value of the bright light populations from the non-light populations. Full recovery from the light was also observed. Bright light as the source of the behavioral response was also confirmed and demonstrated using eyeless mutants. One caveat with the eyeless experiments is that the eyeless gene may have other minor functions in addition to larval vision, although they are not revealed in our experiments with the eyeless mutant larvae.
Control treatments with the experimental strains, the wild type strains Oregon R and Canton S, the “rover” and “sitter” alleles of the forager locus, and the eyeless mutants all gave comparable results. The data from the strains collectively fit into a tight range and were very reproducible, demonstrating the assay's steadiness and giving a stable baseline without alterations in behavior. In Canton S, bright light treatments with no recovery gave comparable behavioral responses to those observed with the experimental strain UAS-mCD8-GFP; ddc-GAL4. Bright light with full recovery in Canton S also gave fully recovered responses that were comparable to those observed with the experimental strain. Thus, the assay was not background specific. There was no genetic contribution to the results, which did not differ among different strains and genotypes undergoing the same assay treatments.
Drosophila larvae in the foraging stage prefer dark areas and are repulsed by light. When a control population of larvae is left to crawl towards yeast, the majority of the larvae (approximately 80%) reach the yeast within the first 10 min, while the rest trickle in during the remaining time. This leaves a severely left-skewed distribution in a plot of larvae scored versus time (Fig. 3A). When a population of larvae is left to crawl towards yeast after exposure to bright light, there is an increase in the number of larvae that reach the yeast later in the assay time period; thus, there is more noise in the tail of the distribution (Fig. 3B). This leads to the significant increase in the t1/2 value, which corresponds to the behavioral response exhibited in our experiments. It is not at all clear what causes the behavioral change in response to light. The fact that larvae rapidly adapt to this noxious condition may indicate that the response is a form of allostasis, or adaptation to stress.
Our assay may be a behavioral model for stress. It quantifies the adaptive behavioral responses to bright light in Drosophila larvae and is a measure of behavioral plasticity. Given the robustness of our assay and its simplicity, it could be used in a genetic screen for mutants in behavioral plasticity with respect to their abilities to adapt to bright light. While a well-defined behavioral response is observed in the assay, we do not know what it is or what changes are taking place in the larvae to cause the response. Is it a loss of appetite or motor function due to stress? In addition, because our experiments involve populations of larvae, it is not clear whether all larvae respond in the same manner. Are the larvae stressed at the molecular level despite not showing a behavioral response to the light? Conversely, are the few larvae from control populations that reach the yeast later in the assay undergoing stress, or are there always a select few that will always be slow? Further investigations using this assay may answer these questions and give additional insights into behavioral plasticity and adaptation to stress.
We would like to thank Claire Cronmiller, Cedric Williams, Jay Hirsh, Custis Hawkwood, and members of the lab for help with the manuscript and for many helpful discussions. We would also like to thank the Bloomington Drosophila Stock Center and Jay Hirsch for fly strains. This work was supported by a grant from NINDS R29 37322, The Jeffress Foundation, and The Keck Foundation to BGC.