| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The Leslie and Susan Gonda (Goldschmied) Multidisciplinary Brain Research Center and Faculty of Life Sciences, Bar Ilan University, Ramat Gan, Israel
Submitted 5 July 2005; accepted in final form 31 August 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
The investigation focused on Aplysia feeding, a model system for examining how a relatively simple neural circuit controls complex behavior. Of particular interest is modulation of consummatory feeding responses (biting, swallowing, rejection) because the neural circuitry controlling these responses has been explored (for review, see Elliott and Susswein 2002
). Previous studies examining feeding patterns in controlled laboratory conditions identified stimuli that affect consummatory responses and thereby would be thought to control ad libitum patterns of feeding. The effects of many of these stimuli on the nervous system have been characterized (Elliott and Susswein 2002). This study examined feeding in freely behaving A. californica to determine whether feeding patterns in this condition are explained by data from more controlled conditions and, if not, to identify additional variables affecting feeding and the neural circuitry underlying feeding.
Four major variables regulating feeding have been identified. 1) Food arousal is initiated by touching food to the lips and decays when food is removed. It causes animals to bite more quickly (Kupfermann 1974; Susswein et al. 1978) and with greater intensity (A. J. Susswein, K. R. Weiss, and V. I. Kupfermann, unpublished data) in response to food. Food arousal is partially mediated by identified neurons, particularly neurons C-PR, C2, and MCC, which respond to touch of food to the lips as well as by release of peptide modulators from motor neurons innervating the musculature producing consummatory responses (for review, see Kupfermann et al. 1991). 2) Satiation causing inhibition of consummatory responses is signaled by activating mechanoreceptors in the anterior gut (Susswein and Kupfermann 1975a,b; Susswein et al. 1976). Satiation interacts with food arousal: it becomes progressively more difficult to arouse animals to eat, and arousal decay more rapidly, as the gut is filled (Susswein et al. 1978). Thus a partially satiated animal that briefly loses contact with food may not begin eating again when food is re-encountered because arousal will decay more quickly, and will be initiated more slowly, as a result of the effects of the partial satiation on arousal (Susswein et al. 1978). 3) Sensory adaptation or habituation of chemoreceptors (Schwarz et al. 1988) can terminate a meal before the anterior gut is filled (Horn et al. 2001). 4) In A. fasciata, pheromones regulating sexual behavior also modulate feeding, in part via modulating the amplitude of swallowing (Blumberg and Susswein 1998; Blumberg et al. 1998; Botzer et al. 1991; Teyke and Susswein 1998; Ziv et al. 1991a).
Previous data provided contradictory expectations for the current experiments. A hungry A. californica fed a single meal eats 
15% of its body weight in a 2- to 3-h meal. The quantity eaten, and the patterning of feeding, are explained by the volume required to fill the anterior gut (Susswein and Kupfermann 1975b; Susswein et al. 1976). Twenty-four hours later animals are unresponsive when the lips are stimulated with food for 2 min. Responses to food are restored over the next few days (Kupfermann 1974). However, if the lips are stimulated for 15 min 24 h after a satiating meal, the animals respond (Susswein et al. 1978), eating a smaller meal than that eaten the previous day. The amount eaten, and the patterning of feeding, are explained quantitatively by the volume of food present in the anterior gut from the previous meal (Susswein and Kupfermann 1975b; Susswein et al. 1976). These data suggest that in ad libitum conditions, Aplysia may eat every few days, when the gut becomes empty, perhaps appended by brief meals initiated after long contact with food. By contrast, experiments on another Aplysia species, A. fasciata, showed that animals ate a number of small meals daily (Ziv et al. 1994). Field studies on both A. californica and on A. fasciata found that some animals eat large meals in natural conditions (Kupfermann and Carew 1974; Susswein et al. 1984).
Our data indicated that patterns of feeding in A. californica with constant access to food were different from those expected and could not be explained on the basis of the effects of the previously identified variables on consummatory movements. First, feeding was regulated by controlling the initiation of a feeding bout rather than via regulating the consummatory behaviors within a bout. Thus the portions of the neural circuit governing feeding that have been described in greatest detail are unlikely to have a major role in the modulation of ad libitum feeding. Second, the constant presence of food was found to initiate a previously undescribed state in which Aplysia eat very little. This newly defined state operates at least in part via circulating factors that inhibit the ability of an otherwise adequate stimulus to initiate feeding. The major stimulus condition causing A. californica to eat was found to be the introduction of food after a period of time without food.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Experiments were performed on A. californica weighing 50–120 g that were purchased from Marinus (Long Branch, CA) and from Marinus Scientific (Garden Grove, CA). The animals were stored in 600-l tanks of aerated, filtered Mediterranean seawater maintained at 17°C. Lighting was light:dark 12:12. Animals were fed two to three times weekly with Ulva lactuca, which was collected at various sites along the Mediterranean coast of Israel and then stored frozen. Experiments were generally performed during the winter months (from October through May). Animals were not observed mating or laying eggs, indicating that they were probably sexually immature.
Food
The food used in all experiments was the seaweed U. lactuca, which was gathered at various sites along the Mediterranean coast of Israel and then stored frozen. Food was thawed before use in an experiment.
Experimental conditions
Animals were transferred from the storage tanks to 5- or 10-l aerated experimental aquaria 24 h prior to an experiment. The aquaria were kept at 23°C. Animals were generally kept one to a 5-l aquarium. In one experiment, animals were kept two to a 10-l aquarium with the two animals separated by a partition that prevented contact between the animals but that allowed free water flow. In some experiments, a single animal was kept in a 10-l aquarium with a partition with food behind the partition that animals were able to smell but could not contact.
STEADY-STATE FOOD ACCESS.
In animals examined in this condition, food (U. lactuca) was available ad libitum for 
1 wk prior to the experiment both in the storage tanks and in the experimental aquaria.
FOOD DEPRIVED. Animals examined in this condition were food-deprived for 5 days prior to the experiment.
PARTIALLY DEPRIVED.
These experiments were designed to test the effects of specific periods of food deprivation on feeding. Prior to the experiment animals were kept with steady-state access to food. The food was then removed for periods of a few hours to a few days, and the food was then restored to the animals, and the quantity of food eaten was examined. When the deprivation period was 
12 h, to prevent a large temperature change just prior to the experiment, animals were transferred 24 h before the experiment to experimental aquaria at 23°C with food and were then transferred to aquaria without food for the period appropriate to the experiment.
FOOD IN THE ENVIRONMENT. After a period with steady-state access to food in the storage tanks, animals were transferred to a 10-l aquarium in which a partition separated the animal from food on the other side of the partition. To test feeding, food was then added to both sides of the partition.
In a second experiment, after 24 h in a 10-l aquarium with food behind the partition, all of the water and food were removed, and fresh water without food placed in the aquarium. Food was added to the aquarium 3 h later.
Experimental measures
PERCENT TIME SPENT FEEDING.
Two procedures were used to measure the percentage of time spent feeding. Both procedures have been used in previous experiments describing time budgeting and bout patterning of A. fasciata behavior (Ziv et al. 1991b, 1994) and have been described in detail previously. In one, animals were continuously observed and the time of onset and offset of all feeding bouts was noted. In the other, animals were sampled every 5 min, and feeding was noted. The number of times that feeding was observed per unit time was used to estimate the percent time spent feeding during that time period (Susswein et al. 1983; Ziv et al. 1991b). Use of each procedure is noted in the text. In some experiments, feeding was sampled during the dark phase of the day or spanned a number of hours that included the transitions from light to dark or the reverse. For these experiments, the timers controlling the laboratory lights were adjusted a number of days previously, so that the day and night were reversed or were offset by 6 h. As in previous experiments (Ziv et al. 1991b), animals were observed during the dark phase by using dim incandescent lighting that pointed down and away from the animals. Such lighting provided illumination of <0.5 lux to the animals but was nonetheless sufficient to observe feeding.
WEIGHT OF FOOD EATEN.
In some experiments, a preweighed quantity of food was placed in the experimental aquarium at the start of the observation period, and the food remaining was weighed again at the end of the experiment. The difference in weight was attributed to consumption during the observation period. For animals in steady-state conditions, the food already in the aquarium was removed before adding the preweighed food. Previous experiments (Botzer et al. 1991) have shown that the food used in these experiments retains its integrity over the experimental period and the weight change is a reliable estimate of the quantity of food eaten.
Surgical procedures
Animals were placed in a 5-l chamber partially filled with seawater and with ice made from seawater. When animals stopped moving and lost contact with the substrate because of lack of muscle tone, the tips of the chemosensory rhinophores were cut off. Animals were then restored to the holding tanks until used in an experiment. Control animals were treated in the same way as were the animals without rhinophores, except that the tip of the rhinophores was not cut.
Extracellular recording
The cerebral and buccal ganglia connected via the cerebrobuccal connectives were dissected from animals that previously had steady-state access to food. Before dissection, animals were injected with 50% of their body weight of isotonic MgCl2. The cerebral and buccal ganglia were removed and placed in a solution of 1:1 ASW, isotonic MgCl2. To limit possible damage, the ganglia were not desheathed. The two ganglia were pinned and then separated from one another by building a wall around the cerebral ganglion with petroleum jelly (Vaseline). The cerebral-buccal connectives crossed the wall isolating the cerebral and buccal ganglia. Possible leak of the Vaseline seals was determined by filling the cerebral ganglion chamber almost to overflow and observing if the level changed over 2–5 min. In addition, the fluid in the cerebral ganglion chamber was stained with Fast Green to detect possible leakage of the dye to the other chamber. Suction electrodes were used to record extracellular nerve activity and were placed on the radula nerve and buccal nerve 2 (Gardner 1971) [these are nerves 1 and 5 in the terminology of Scott et al. (1991)]. The volume of the cerebral ganglion chamber was 1 ml, whereas the buccal ganglia chamber contained 10 ml. After completing the dissection and pinning of the ganglia, the bathing solution was initially changed to artificial seawater (ASW) with the following composition (in mM): 460 NaCl, 10 KCl, 11 CaCl2, 55 MgCl2, and 5 NaHCO3 at pH = 7.64.
To induce activity related to feeding the nonhydrolyzable cholinergic agonist carbachol (CCh, 2.5 x 10–3 M) was applied to the cerebral ganglion for 10 min. Previous data (Susswein et al. 1996) has already shown that CCh applied to the cerebral ganglion induces organized buccal motor patterns. These were recorded via the suction electrodes. A total of seven trials of CCh application were run. These were separated from one another by 10-min intervals in which the cerebral ganglion was bathed in ASW. To determine how hemolymph from either hungry animals or from animals in the steady state affects the activity induced by CCh, hemolymph was applied along with CCh during the fourth and the fifth runs. After the fourth run, the CCh was washed and was replaced with hemolymph alone. Thus the ganglion was exposed to hemolymph from the start of the fourth run in CCh, throughout the 10-min period between the fourth and the fifth runs, and was removed only after the fifth run. Data were tabulated and are shown from the exposure to carbachol immediately preceding the exposure to hemolymph and from the second run in carbachol in the hemolymph. The effects of hemolymph from hungry versus steady-state animals on buccal motor program were tested using a blind procedure: the experimenter was unaware of whether the hemolymph applied had been extracted from previously hungry animals or from animals with steady-state access to food.
Hemolymph was extracted by pricking animals with an empty syringe that penetrated through the body wall into the hemocoele. Animals were 100 ml in volume, and 5–7 ml of hemolymph were extracted. Hemolymph from two to four animals was extracted just prior to use in an experiment. The hemolymph from the different animals was combined. Both the buccal and cerebral ganglia were bathed with the hemolymph.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Regulation of consummatory behaviors does not contribute to regulation of food intake
An initial experiment was designed to determine how animals regulate the total quantity of food eaten, and the patterning of feeding, when they have ad libitum access to food. Feeding was examined in animals that had been food deprived for 1 wk as well as in animals that had had steady-state access to food for a number of days prior to the experiment. Feeding behavior was observed for 6 h from 3–9 h after the onset of light. Animals were observed continuously, thereby allowing us to determine the total time devoted to feeding, the length of all feeding bouts, and the length of inter-bout intervals. The weight of the food consumed during the 6-h observation was also measured. The 6-h observation was divided into 12 half-hour intervals, and the percent time spent feeding was calculated for each interval.
REGULATION OF INTER-BOUT INTERVALS CAUSES A DECREASE IN FEEDING DURING MEALS. Previously food-deprived animals ate well when allowed steady-state access to food. There was a significant decrease in the percent time spent feeding over the 6 h of observation (Fig. 1A). The decrease presumably reflects satiation as the hungry animals consumed progressively larger quantities of food. It is important to note that the percent feeding decreased gradually over the 6-h observation.
|
|
The possible effect of feeding on interbout intervals was directly assessed (Fig. 2C). For this analysis, the 6-h observation was divided into 1-h bins because some interbout intervals were >30 min. There was a significant increase in interbout intervals over the 6-h observation, particularly in the latter part of the observation.
REGULATION OF INTER-BOUT INTERVALS CAUSES A DECREASE IN FEEDING IN STEADY-STATE CONDITIONS. Aplysia that had been maintained with steady-state access to food prior to the observation ate remarkably little. Only four of six animals ate at all during the 6-h observation. The six animals devoted to feeding a mean of 9.87 ± 4.15 min (or 2.75 ± 1.15% of the 6-h time period; Fig. 1B). As would be expected for animals in steady-state conditions, in the four animals that fed, there was no significant change in the time spent feeding during the 6-h observation period.
The difference in time spent feeding in steady-state conditions and in previously hungry animals could arise from shorter bout lengths, from longer inter-bout intervals, or from both. Feeding bouts were significantly longer in steady-state conditions than in previously hungry animals (Fig. 2B). These data indicate that the time spent feeding in steady-state conditions is decreased because of longer intervals between feeding bouts, in spite of an increase in mean bout length. To estimate the mean interbout interval in steady-state conditions, the total time not spent feeding in the six animals was divided by the total number of feeding bouts observed. This gave a value of 67.7 min. A survey of the interbout intervals showed that shortest interval was 14 s and the longest was longer than the 6-h observation period. Thus changes in either food-finding behavior, or in the likelihood of initiating feeding after food is encountered, account for the difference in time spent feeding between previously deprived animals and animal with steady-state access to food.
REGULATION OF THE QUANTITY OF FOOD EATEN.
Animals that were in steady-state conditions consumed significantly less food than did previously hungry animals (Fig. 3A). This is not surprising because previously hungry animals spent significantly more time spent feeding than did steady-state animals (Fig. 3B). However, the difference in the quantity eaten could also arise in part via a difference in the efficacy of feeding behavior. When animals are hand-fed individual meals, the amplitude of biting responses decreases (Susswein et al. 1976). To determine whether the efficacy of consummatory responses is regulated, we divided the time spent feeding by the quantity eaten to obtain a measure of the active feeding time needed to consume 1 g of food. Calculating the time needed to consume 1 g of food in previously hungry and in steady-state animals showed that there was no significant difference in efficacy of feeding (Fig. 3C). Both previously hungry and steady-state animals required a mean of 20 min of active feeding to consume a gram of food. These data indicate that changes in the amplitude of biting and swallowing movements produced by changes in the feeding state are unlikely to contribute greatly to changes in the quantity of the food eaten because such changes would have altered the efficacy of feeding bouts.
|
The preceding data suggest that regulating the amplitude or efficacy of feeding movements contributes relatively little to the regulation of how much is eaten. However, it is possible that the regulation of amplitude occurs only at the start of a meal in previously hungry animals. When animals begin to eat, feeding movements may be vigorous, but the amplitude of feeding movements may then decrease rapidly and reach the steady-state values very quickly. To test this possibility feeding was observed continuously for 1 h in six previously food-deprived animals. These animals required a mean of 28.04 ± 2.41 (SE) min to eat a gram of food. Thus feeding was not more efficient during the first hour of feeding than during the steady state in spite of the greater time spent feeding.
FEEDING IS INHIBITED IN ANIMALS WITH STEADY-STATE ACCESS TO FOOD.
Our data indicated that animals ate relatively little in steady-state conditions. If the measured rate of feeding in the six animals that were observed continuously was maintained over the entire day, animals would spend a mean of 39.48 min/day feeding. However, a number of previous studies demonstrated that in field conditions Aplysia may spend many hours a day feeding (Kupfermann and Carew 1974; Susswein et al. 1983, 1984). We examined a number of factors that could have caused the observation of a low rate of feeding even if more time were actually devoted to feeding. We found that none of these factors operated and that the observation that animals with steady-state access to food eat little was confirmed.
TIME SPENT FEEDING AT OTHER TIMES OF DAY.
Because A. californica feeding is primarily a diurnal activity (Kupfermann 1974; Lyons et al. 2005), and the observations reported in the preceding text were during the daylight hours, feeding should have been observed. Nonetheless, it is possible that relatively little feeding was seen because of significant differences in diurnal versus nocturnal feeding between animals with steady-state access to food (as in the present experiments) and animals that are intermittently exposed to food (as in previous experiments showing diurnal feeding). In steady-state conditions, feeding may be more prominent at other times of day. To examine this possibility, feeding was sampled every 5 min in conditions of steady-state access to food during 6 h of night (from 3 to 9 h after the lights went out), during the 6 h of transition from day to night (3 h before and 3 h after the lights extinguished), and during the 6 h of transition from night to day (3 h before and 3 h after the lights turned on). Together with the observations in the preceding text during the daylight hours, these observations would reveal feeding at all hours of the day, if it was present.
The rates of feeding were extremely low during all hours of the day (Fig. 4). There was no significant difference between the rates of feeding seen during the day or night or during the transition periods. Thus animals with steady-state access to food eat very little throughout the day-night cycle.
|
To test this possibility, feeding was sampled every 5 min in seven Aplysia with steady-state access to food that were observed for 6 h during the daytime, for 3 days in succession, providing a larger data set to catch an animal eating a large meal. No large meals were observed. Animals ate a mean of 0.50 ± 1.1% of the time. The longest meal seen was 3.3% of the total observation period.
IS THE ANTERIOR GUT FULL?
Previous data on animals fed once daily showed a strong correlation between the volume that fills the gut and the satiation state of an animal (Susswein and Kupfermann 1975b). At satiation, the anterior gut is close to full. The best predictor of satiation is the ratio of the weight of the gut contents divided by the weight of the empty anterior gut (Susswein and Kupfermann 1975b). In satiated animals, this ratio is 8.8 ± 0.9 (SE). To determine whether animals with steady-state access to food acted as though satiated or close to satiated because the gut was full, or close to full, animals (n = 7) were dissected, and the gut contents and the empty gut were weighed. The mean ratio of gut contents to the empty gut was found to be 3.47 ± 2.66 (SD), with a range of 7.55 (close to full) to 0.52 (empty). Thus the animals were very heterogeneous, with some animals having gut contents indicating that they were close to satiated and others with a virtually empty gut. Nonetheless, all of the animals were not eating before being dissected and would probably have eaten very little over the subsequent few hours.
DOES THE STEADY STATE DEPEND ON TEMPERATURE?
Previous behavioral experiments on A. californica feeding were performed on animals that were cooled. In addition, A. californica are found in relatively cool waters. Previous studies have shown that temperature can have major effects on the modulation of the feeding musculature (Vilim et al. 1996; Zhurov and Brezina 2005). In our experiments, animals were stored at 17°, but experiments were performed at room temperature. To test whether the steady-state conditions arose as a result of the relatively warm temperature at which experiments were performed, ad libitum feeding was observed in 16 animals (8 hungry, 8 steady-state) at 17° for 4.5 h. Hungry animals ate a mean of 2.55 ± 0.56 g, whereas animals in the steady-state ate only 0.3 ± 0.44 g. Hungry animals spent a mean of 22.05 ± 5.18% of the time eating, whereas the steady-state animals spent a mean of 1.14 ± 1.35% time eating. These data indicate that the steady-state inhibition of feeding is intact at cooler temperatures. Nonetheless, some modulation by the temperature was seen in hungry animals. Patterning in the hungry animals was different from that seen in the previous experiments at room temperature. Animals ate more vigorously at the start of the meal and became satiated more rapidly. Thus during the first half hour, animals ate a mean of 77.08 ± 18.52% of the time, but by the start of the second hour, animals were eating only 4.17 ± 7.22% of the time.
Is a period without food a signal to eat?
Our data have shown that animals having steady-state access to food eat little, but animals that have been food-deprived eat a large meal. This suggests that the signal to begin eating may be a period of food-deprivation. How long a deprivation is needed to induce an increase in feeding similar to that seen in hungry animals? To examine this question, animals were allowed steady-state access to food for a number of days. The animals were then food-deprived for 1, 2, 3, 4, 5, 6, or 8 days, and feeding was then sampled every 5 min for 4.5 h (Fig. 5), enough time to sample a large meal after the deprivation. In all of the groups, the rate of feeding gradually declined during the first 2 h. For all periods of food deprivation, there were significant differences between the percent time spent feeding during the first half hour of exposure to food and the last half hour of the test (P < 0.05, paired t-test for all 7 groups). These data confirm that a period of food deprivation is a signal to begin feeding when food is restored. The data also indicate that a 24-h period of food deprivation is sufficient to induce feeding.
|
Previous data had shown that in steady-state conditions in different experiments animals ate for a mean of 0.5–4% of the time (see preceding text). In the present experiment, the mean percent time spent feeding in the last hour of the observation in all groups was 3.37 ± 0.98%, suggesting that animals had reached the steady state by the end of the observation. How long does it take the animals to reach the steady state? For successive hourly intervals, we tested whether the 95% confidence interval of the percent time spent feeding over the hour was significantly >3.37%. Animals were assumed to reach the steady state when the lower limit of the 95% confidence interval of the percent feeding over an hour reached 3.37%. To increase the resolution of the measurements, hourly intervals were calculated at 15-min offsets (Fig. 6). After 1, 3, 5, and 6 days of food deprivation, animals reached a subsequent hourly feeding level not significantly different from that in the steady state 2 h after the introduction of food (Fig. 6). After 2 days without food, animals reached the steady state in 1.75 h, and after 4 days without food, animals reached the steady state after 1.5 h (Fig. 6). Thus there was no trend for a gradual increase in time to reach steady state from day to day. However, after 8 days of food deprivation, animals reached the steady state after 2.75 h, suggesting that more than a week of food deprivation might be needed to allow animals to become fully hungry.
|
|
|
in the figure that marks when the 95% confidence interval overlaps the steady state). With only 12 h of deprivation, there was also a decrease in the percent time spent feeding between the first and last 30-min observation period, and the percent time spent feeding was significantly higher than in steady state conditions for the first 1.75 h after the introduction of food. By contrast, with 6 h of deprivation, there was no significant difference between the first and last 30 min of the observation, and the percentage of time spent feeding was not significantly higher than that in steady-state conditions at any time after the food was introduced (note the lack of arrow), although there was still a tendency for more time to be spent feeding just after the food was introduced). For 3 h of deprivation, there was no significant difference between the first and last 30 min of the observation. However, the percent time devoted to feeding was increased over the steady-state value during the first hour after the introduction of food (note ). The percentage of time devoted to feeding during the first 2 h after food was introduced was compared in the five groups of animals (Fig. 7). A one-way ANOVA showed that the difference in time spent eating between the groups approached significance. A post hoc test showed that there was a significant difference between feeding after a 24-h food deprivation and an immediate test with no deprivation, but feeding after 3, 6, or 12 h of deprivation was not significantly different from feeding after no deprivation or after 24 h of deprivation.
These data indicate that over a period of 3–24 h without food, the effect of food deprivation increases and reaches a maximal effect by 24 h. Periods of deprivation of <24 h produce increases in feeding using some measures but not others.
FOOD IN THE ENVIRONMENT SIGNALS THE STEADY-STATE. After animals are in a condition of steady-state access to food, a period of 24 h without food will initiate a large meal. The 24-h period without food removes a number of separable stimuli that are provided to the animals by the steady-state access to food. In the presence of steady-state food, animals sense the food in the environment, which stimulates chemoreceptors. In addition, animals contact the food, and touch of food will stimulate mechanoreceptors as well as contact chemoreceptors. Finally, the occasional consumption of food will activate a variety of postingestion stimuli. The continued or intermittent presence of any or all of these stimuli could contribute to the animal's maintained steady-state behavior, and the removal of one or all of these stimuli could contribute to the initiation of feeding after the 24 h without food. To test the possible contribution of food in the environment that is not contacted to the maintenance of the steady-state, animals were first given access to steady-state food for 3 days. They were then transferred for 24 h to an environment in which food was present behind a partition, allowing the animals to smell the food, but the partition prevented them from touching or eating the food. At the end of the 24 h in the presence of inaccessible food, animals were given 4.5 h of access to food, and feeding was examined. During the test of feeding after 24 h in the presence of inaccessible food, animals ate a mean of 1.6 ± 4.2% of the time (mean ± SD; n = 15), similar to that seen in animals with steady-state access to food (Fig. 9A). These data indicate that food in the environment is sufficient to maintain the steady-state condition, even if the animals never contact the food or consume it.
|
40 to 2% of the time spent feeding. These data indicate that the rhinophores sense food in the water that is a sufficient stimulus for maintaining the steady state. The data also show that stimuli not sensed by the rhinophores are sufficient to initiate the steady state when previously hungry animals encounter food. ARE THE RHINOPHORES NECESSARY? The preceding experiment showed that chemostimuli sensed by the rhinophores are sufficient to maintain the steady state. Are they also necessary or can other receptors responding to food substitute for the rhinophores in maintaining the steady state? To test whether other receptors may also contribute to maintaining the steady state, animals without rhinophores and sham-operated controls were allowed steady-state access to food for 3 days, and feeding was then measured for 4.5 h. There was no significant difference between the two groups. Both groups showed little feeding (Fig. 9C), similar to that seen previously in animals in steady-state conditions. These data, coupled with the previous experiment, indicate that the rhinophores are not necessary for either initiating or maintaining the steady state, even though they are a sufficient input pathway for maintaining the steady state. Other input pathways are also able to signal the presence of food and thereby initiate and maintain the steady state.
POSTINGESTION CUES ARE USED TO DETERMINE MEAL SIZE. The preceding data indicate that the maintenance of the steady-state condition of food access can be explained entirely by the presence of food stimuli in the water, which are sensed by the rhinophores. When the food stimuli are removed, and food is then restored, animals will eat a meal. What determines the meal size when food is restored? To examine this question, animals were maintained for 24 h with food in the environment but inaccessible as in the previous experiment. However, the food was then removed for 3 h, before testing feeding by restoring accessible food to the animals for 4.5 h. Previous studies (see Figs. 7 and 8) had already shown that 3 h without food causes Aplysia to eat minimally when food is restored. However, after 24 h in which food was present but inaccessible, 3 h without food initiated feeding that was similar to that observed after a 24-h period of food deprivation (Fig. 9D). Thus the absence of postingestive stimuli over the previous 24 h led to a large increase in feeding when food was restored. These data indicate that postingestion stimuli contribute to the meal size after feeding is restored.
Neural correlates of the steady state
To determine whether correlates of the steady state can be measured by recording from the Aplysia nervous system, we took advantage of a finding that buccal motor programs that are correlates of ingestive movements can be elicited by bathing the cerebral ganglion in the cholinomimetic carbachol (Susswein et al. 1996). Carbachol is thought to be effective in inducing buccal motor programs because sensory neurons sensing food on the lips are thought to be cholinergic (Susswein et al. 1996). The buccal motor programs are measured by extracellular recordings from buccal nerves, which innervate the buccal muscles producing consummatory feeding movements. Preliminary studies (A. J. Susswein and I. Kupfermann, unpublished results) had suggested that hemolymph from animals in the steady state has a role in modulating the ability of carbachol to induce buccal motor programs. We examined the buccal motor programs induced by a 10-min exposure to carbachol when the ganglia were bathed in hemolymph taken from animals that had steady-state access to food, as well as from animals that had been food deprived (Fig. 10). The presence of hemolymph reduced the ability of carbachol to induce buccal motor programs with respect to programs induced in the presence of artificial seawater. In the absence of hemolymph, carbachol induced a mean of 9.50 ± 11.18 (SD) buccal motor programs with a mean latency of 50.334 ± 81.76 s (compare this value to those in Fig. 10). However, hemolymph taken from animals that had been in the steady state produced a much larger inhibition of feeding than did hemolymph from previously hungry animals. Indeed, in six of seven preparations in hemolymph from animals that had been in the steady state, the carbachol was unable to induce any buccal motor programs. These data indicate that inhibition of feeding in steady-state conditions occurs in part via release of chemical factors into the hemolymph. These factors in turn reduce the ability of stimuli that normally initiate consummatory feeding patterns to do so.
|
The preceding data indicate that A. californica with steady-state access to food eat between 2 and 4% of the time. However, previous data on another Aplysia species, A. fasciata, showed that animals with steady-state access to food eat close to 20% of the time (Susswein et al. 1983). These animals also had access to conspecifics, and to egg cordons, which release pheromones into the water that strongly modulate A. fasciata feeding (Blumberg and Susswein 1998; Blumberg et al. 1998; Botzer et al. 1991; Teyke and Susswein 1998; Ziv et al. 1991a). It is possible that the difference between the large time investment in feeding observed in A. fasciata and the small investment observed in the current experiments stems from the fact that the experiments in A. californica were on isolated animals. Had other animals been present, the time devoted to feeding might be increased to values observed previously in A. fasciata. The possible modulation of feeding by pheromones has not been previously examined in A. californica.
To examine the possibility that pheromones secreted by conspecifics might increase feeding in A. californica, animals were maintained either one to an experimental aquarium or two to an aquarium with a partition separating the two animals from one another. Feeding was observed for 6 h, both in steady-state conditions of access to food, as well as in previously hungry animals (Fig. 11). There was no significant difference in the percent time spent feeding between isolated animals and those maintained together in steady-state conditions. However, in previously hungry animals the presence of a conspecific significantly increased the percent time spent feeding over that in an isolated animal. These data indicate that pheromones secreted by conspecifics affect feeding in A. californica but cannot explain the difference in percent feeding in steady-state conditions observed in these experiments and in those on A. fasciata.
|
40% of the time. The time devoted to feeding declined rapidly over the first 2–3 h of access to food. By contrast, in animals maintained with conspecifics the decline in the percent time spent feeding was much slower, and animals had not reached steady-state values by the end of the 6 h observation (Fig. 12). Even during the last 2 h of the observation, animals that were in the presence of conspecifics spent significantly more time feeding than did the isolated animals [P = 0.04, t(19) = 2.11, 2-tailed t-test].
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
15% of their body weight. Feeding causes decreases in consummatory response amplitude and increases in latency (Susswein et al. 1976). The size and dynamics of a meal are set by passive fill of the anterior gut (Susswein and Kupfermann 1975a,b). Sensory adaptation also contributes to the initiation and termination of consummatory responses (Horn et al. 2001). In addition, food arousal is initiated by touching food to the lips; arousal decays when animals lose contact with food. Food arousal interacts with satiation: as animals satiate, it becomes more difficult to arouse them, and arousal declines more rapidly (Susswein et al. 1978). It was suggested (Susswein et al. 1978) that the interaction between satiation and arousal could contribute to spacing of feeding bouts. When animals are hand-fed a single daily meal they consume 60% less than when they are hungry, and the anterior gut contents are 60% full at the start of the meal, from food remaining in the gut from previous meals. Patterning of feeding is similar to that in the last 40% of the large meal eaten by hungry animals (Susswein et al. 1976). These experiments suggested that in ad libitum conditions, animals would eat large meals that are terminated when the anterior gut is close to full or as a result of sensory adaptation. During the meals, consummatory feeding responses would become less effective, thereby causing a progressive decrease in food consumption per unit time spent feeding. Meals would be initiated as the gut becomes empty or as sensory adaptation decays. We found that in ad libitum conditions, feeding is not easily explained by the previously identified factors regulating consummatory responses. First, consummatory responses were minimally regulated. Second, in steady-state conditions, the quantity of food eaten is unrelated to the quantity of material in the anterior gut. Third, a newly identified factor, the presence of food in the environment, is a major variable inhibiting feeding.
Lack of regulation of consummatory responses
Feeding is regulated primarily by regulating interbout intervals rather than by regulating feeding bout lengths. Thus in steady-state conditions, animals eat less than after food deprivation, but bouts are longer. In addition, in previously deprived animals, the time spent feeding decreases over a number of hours, but bout length decreases only during the first half hour. These findings contrast with data showing that regulation of bout length contributes to learning affecting feeding behavior (Chiel and Susswein 1993; Susswein et al. 1986). Thus Aplysia can regulate bout lengths in other behavioral contexts. Regulation of consummatory behavior amplitude also did not contribute to regulation of the quantity eaten. Consummatory response amplitude is modulated by learning (Chiel and Susswein 1993) and by pheromones (Blumberg and Susswein 1998). In addition, a previous study showed that the amplitude of consummatory movements is decreased as Aplysia satiate (Susswein et al. 1976). It is likely that in our experiments consummatory response amplitude was regulated, but this regulation was difficult to pick up in experiments not explicitly designed to find it.
Regulation of interbout intervals indicates that regulating bout initiation is more important than is the regulating consummatory responses within a bout. However, much more is known about the neural circuitry giving rise to consummatory behaviors, i.e., repetitive biting, swallowing, and rejection (for review, see Elliott and Susswein 2002) than about the neural circuitry influencing the decision to initiate consummatory behaviors. It will be important to gather more information on the neural mechanisms underlying the decision to initiate a bout.
Steady-state conditions
Aplysia with constant access to food are in a novel state in which they eat little. The little eaten is distributed into bouts 150 s with 70 min separating bouts.
ENTRY, MAINTENANCE, AND EXIT FROM THE STEADY STATE. In previously hungry animals a 2- to 3-h meal causes animals to enter the steady state. We did not explore which aspects of the meal initiate the steady state. Possible contributors include the presence of food, touch of food to the animals, the performance of feeding behaviors, or postingestion stimuli. However, initiating the steady state is not dependent on the rhinophores because animals without rhinophores enter the steady state as well as do intact animals.
The steady state is maintained by the presence of food, which is sensed by the rhinophores, even if the animals do not contact the food. However, the effect of only 24 h of food in the water was examined. Animals might eventually leave the steady state, even if food is present, if the period without food contact is longer. In addition, the steady state can be maintained in the absence of the rhinophores, indicating that other stimuli can also maintain the steady state.
Animals leave the steady state after a period without food. Three hours without food was sufficient to terminate the steady state when animals had not contacted food for 24 h. However, termination of the steady state is influenced by postingestion stimuli because 3 h without food was not effective in terminating the steady state in animals that had been in contact with food over the previous 24 h. The nature of the postingestion stimuli affecting steady-state termination remains to be determined.
HEMOLYMPH AND NEURAL CORRELATES OF THE STEADY STATE.
The steady state is maintained in part via humoral factors. Thus application of the cholinomimetic carbachol onto the cerebral ganglion either failed to elicit buccal motor program, or elicited fewer programs, with a longer delay, in preparations bathed in hemolymph from animals that had had steady-state access to food with respect to buccal motor program elicited in hemolymph from hungry animals. The identity of the hemolymph factors inhibiting feeding has not been explored. In addition, the mechanism by which stimuli signaling the steady state cause an eventual change in the hemolymph has not been explored. The finding that a humoral factor following a meal affects Aplysia feeding behavior is consistent with a previous study suggesting that humoral factors cause an increase in heart rate that is correlated with decreased feeding when Aplysia are fed once every 3 h over 3 days (Dieringer et al. 1978).
OTHER BEHAVIORS AFFECTED BY FOOD.
The finding that food stimuli maintain the steady state is reminiscent of a previous finding that steady-state food access inhibits sexual behavior in A. fasciata (Nedvetski et al. 1998; Susswein 1984). However, inhibition of sexual behavior required animals to touch the food occasionally (Nedvetski et al. 1998), presumably because without touch, the animals become adapted to the food. Food may cause continued inhibition of feeding, but not sexual behavior, by activating two sensory pathways with only one displaying adaptation without food contact. Food stimuli could also habituate in motor systems controlling feeding but not in those controlling sexual behaviors. Food could also differentially affect feeding and mating because a hemolymph factor is an intermediary in the inhibition of feeding but not sexual behavior. The hemolymph factor may maintain the block of feeding after adaptation of the receptors responding to food.
NUTRIONAL FACTORS AND CONTROL OF FEEDING.
In the steady state, feeding is not controlled by variables that monitor nutritional state directly or indirectly. Previous studies have shown that Aplysia regulate nutritional variables such as hemolymph glucose concentrations, but these do not affect feeding behavior (Horn et al. 1998). Our data now show that gut fill also does not affect feeding behavior in the steady-state. The dissociation of Aplysia feeding from monitors of nutritional state is superficially surprising. However, this dissociation is seen when Aplysia are in an environment of abundance with food constantly available. In this condition, the small amount of food eaten is presumably adequate to support the metabolic needs. It is unlikely that a 24-h deprivation, which induces a large meal, increases the metabolic costs of the animal so greatly that it must now compensate for the deprivation. The large meal probably reflects a change in strategy in response to a patchy food supply. When food is only sometimes present, a useful strategy may be to feed maximally because the source and time of the next meal is not predictable. Previous studies have shown that in nature, A. californica eat large meals (Kupfermann and Carew 1974). If our interpretation is correct, these were seen in an environment with a patchy food distribution. Natural environments harboring Aplysia are variable; sometimes food is constantly available, at other times it is variably present (Susswein et al. 1984). Previous studies (Susswein et al. 1976) showed that animals hand-fed daily also eat sizeable meals. Our data suggest that these meals provided energy well beyond that needed to support the metabolism. The large daily meals may have arisen from an experimental protocol that maximized the total quantity of food eaten.
It is important to note that the small time investment in feeding was observed in sexually immature animals. Previous studies on ad libitum feeding in sexually mature A. fasciata showed that they budgeted 15–20% of their day to feeding (Susswein et al. 1983). However, sexually mature isolated A. fasciata devote no more than 4% to feeding, similar to our present findings. In A. fasciata, pheromones released by conspecifics and by egg cordons significantly increase feeding both in hungry animals and in an
= 0.5, there was a significant difference between feedings after a 24-h food deprivation, and an immediate test with no deprivation but feeding after 3, 6, or 12 h of deprivation was not significantly different from feeding after no deprivation or after 24 h of deprivation.
