Hatchery-reared brown trout (Salmo trutta L.) were exercised continuously for periods of several weeks at swimming speeds of 1·5, 3·0 and 4·5 body lengths/s and their rates of growth were determined. Changes in the major muscle constituents were determined by biochemical analysis and changes in muscle cells using histochemistry and electron microscopy.
At the lowest speed the fish grew much more rapidly, and converted food into fish flesh much more efficiently than their controls kept in still water. Large stores of glycogen and lipid were built up. Gross changes were observed mainly in red muscle cells, with enlargement of the mitochondria being very noticeable.
The fish swimming at intermediate speed showed greater growth than the controls, although the energy expended in swimming against the water current led to an inefficient food conversion rate. Large stores of glycogen were built up, but lipid levels fell, suggesting that this was the major fuel for swimming at this speed. Changes in all the muscle fibre types were observed.
The energy required to maintain the fish in the water flow at the highest speed was so great as to have serious detrimental effects on the fish, and many did not survive. Those which did survive showed signs of gross depletion.
Exercise, and especially that regarded as athletic training, has received much attention in recent years from mammalian physiologists. Many studies have shown that prolonged exercise causes hypertrophy of muscle fibres (Goldspink, 1964; Walker, 1966; Kowalski et al. 1969; Goldspink & Howells, 1974), changes in fibre type composition (Barnard et al. 1971 ; Kiessling et al. 1975) and produces greater resistance to fatigue (Holloszy et al. 1971 ; Baldwin et al. 1973). Supporting tissues such as the heart and blood system also change to be better able to supply the muscles with nutrients and oxygen (Richter & Kellner, 1963; Rabinowitz & Zak, 1972).
The study of exercise in fish has received much less attention. Several studies have been carried out on sprint exercise (Bainbridge, 1958; Johnston & Goldspink, 1973a; Wardle, 1975). These authors have shown that fish can rapidly accelerate up to extremely fast speeds, but that they can only maintain such velocities for a few seconds. Other studies have looked at prolonged exercise where animals have been forced to swim for several hours. Black et al. (1959, 1962) observed glycogen break-down with exercise as did several other workers (Beamish, 1968; Pritchard, Hunter & Lasker, 1971; Johnston & Goldspink, 1973b). The other type of exercise, that of training animals for several weeks in order to observe adaptive changes caused by the work load, has been much less studied. Several workers have observed the changes occurring in salmonid fishes during the seasonal migration upstream to the spawning grounds, although during this period the animals did not ingest any food, thus any adaptive changes were masked by the effects of depletion (Brett, 1973). Training of fish has mainly been carried out on members of the Salmonidae. Hammond & Hickman (1966) showed that training trout resulted in an increase in the maximum swimming speed, this possibly being due to an increased tolerance to lactic acid build-up, as shown by Hochachka (1961). Trout have been shown to actively choose running water in preference to still water (Davidson, 1949).
The only work on the training of marine fish has been carried out on the coalfish by Greer Walker (1971) and Greer Walker & Pull (1973). These authors reported that exercise produced an increase in growth rate, provided that the water speed was not too great, and suggested roles for the different muscle types based on the degree of hypertrophy of the fibres at different speeds.
As the restocking of ponds and rivers is being carried out on an increasingly large scale, it is important that information is obtained about the effect of transferring fish from one environment to another. The present work was carried out in order to observe the effects of training trout at known water velocities, especially the effects of training hatchery trout which had previously experienced only very slowly moving water.
MATERIALS AND METHODS
Brown trout (Salmo trutta), 12–15 cm in length, were obtained from the Yorkshire Water Authority Trout Hatchery at Pickering in North Yorkshire. They were allowed to acclimatize in an aquarium for 2 weeks before experimentation at 12 °C and a 10/14 h light/dark photoperiod, with a daily diet of chopped liver.
A flume of similar construction to those described by Greer Walker (1971) and Johnston & Goldspink (1973) was used in this study (Fig. 1). This consisted of a trough of dimensions 250 × 250 × 2500 mm through which a controlled flow of water could be moved. Water travelled through the trough from an upper reservoir to a lower one from which it was pumped back up to the upper one. Water speed was controlled by the rate of pumping, the gradient of the tunnel and the height of the weir. Flow rate was determined using a miniflow probe (George Kent (Stroud) Ltd. Temperature was maintained at 12 °C by means of a refrigeration unit situated in the upper reservoir. Fish which fell back on to the lower restraining grid in the flume were induced to swim again by a mild electric shock (3–5 V a.c.) delivered from the grid.
Three sets of experiments were carried out; 28 days at 1·5 body lengths per second (b.l./s), 28 days at 3·0 b.l./s, and 14 days at 4·5 b.l./s. These were started by placing the fish (in groups of ten) into the flume at a low water speed and allowing them to become acclimated to the new environment. This was taken as the time when the animals began to accept food introduced into the flume and was usually 2 or 3 days. The water speed was then steadily increased until the desired speed was attained. At the lowest speed this was achieved after a few hours, at the intermediate speed after 2 days, while at the highest speed, the final water flow was reached only after 7 days. The trout were fed on chopped liver introduced into the top of the flume, and any excess liver was collected at the bottom with a hand net. All animals were fed once per day to satiation. The experiments were terminated by removing the fish from the water flow and killing them by a blow to the head. Blocks of muscle tissue were then quickly removed from a point on the lateral line immediately below the dorsal fin for histochemical studies. The fish were then plunged into liquid nitrogen in order to prevent any post-mortem changes in the composition of the muscle.
Muscle blocks were mounted on metal chucks and rapidly frozen in liquid freon (dichlorodifluoromethane I.C.I. Arcton 12) at −158 °C. Sections 7–12 μm thick were then cut at −20 °C in a cryostat (Bright Instruments Ltd.). These sections, mounted on coverslips, were stained for myofibrillar adenosine triphosphatase by the method of Padykula & Herman (1955) modified by Guth & Samaha (1970). The following modifications were made to the technique to make it suitable for trout muscle. The sections were not prefixed as this was found to denature the enzyme rapidly. Preincubation was carried out at room temperature at a pH of 10-2 for time periods ranging from 2 to 15 min (Johnston, Ward & Goldspink, 1975). Incubation was for 20 min at room temperature. Using this procedure the pink muscle could be differentiated from the white muscle tissue, although the red muscle was always unstained due to the instability of the red muscle enzyme at high pH. This tissue was stained by using techniques for lipid (Sudan Black B) and succinic dehydrogenase (Nachlas et al.1957).
Using the slides thus produced, it was possible to count the total number of red fibres in the superficial muscle layer. This was achieved by using a projection microscope and a square grid system. The numbers of pink and of white fibres were not counted. The diameters of all three types were measured using a graduated eyepiece. The diameter was found by taking the mean of two measurements on each fibre, the second being at right angles to the first. One hundred fibres were measured for each fibre type for each animal.
Frozen fish were allowed to warm up to − 20 °C. Red and white muscle samples were then dissected out while still frozen, after which they were quickly cooled again in liquid nitrogen. Several workers, e.g. Brandes & Dietrich (1953) and Black et al. (1962) have shown that the chemical composition of teleosts alters with position along the length of the body, so all samples were taken from the same area, at the position of the dorsal fin.
Water content of the tissue was found by drying to constant weight in a vacuum oven at 60 °C. Protein nitrogen levels were estimated by the Biuret method of Gornall, Bardawill & David (1949) after initial digestion in 30% KOH. Following extraction by the method of Seifter al. (1949) the Carroll, Longley & Roe (1956) anthrone method was employed to determine the amount of glycogen stored in the muscle tissue. Lipid concentrations were determined by first extracting from the muscle tissue using the chloroform-methanol-water method of Bligh & Dyer (1959) modified by Hanson & Olley (1965), and then assaying with a Biochemica Test Kit (Boehringer Mannheim) no. 15991.
Small pieces of red and white muscle were fixed in glutaraldehyde and osmium tetroxide and then embedded in Araldite. Ultrathin sections were then cut on a Reichart ultramicrotome and mounted on collodion-coated grids. The grids were stained with uranyl acetate and lead citrate and then viewed using a JOEL JEM 7 A electron microscope.
Those fish swimming at 1·5 and 3·0 b.l./s quickly became adjusted to their new environment, with an average of 20 % of the animals becoming exhausted and falling back on to the lower grid within the first 2 days of each experiment. This was not considered to be abnormal, as all experiments of this nature carried out on trout by the authors have yielded figures similar to this. Three experiments were carried out at 4·5 b.l./s. In the first experiment all of the fish had become exhausted by the seventh day. By the seventh day in runs 2 and 3, the day on which the final speed was reached, 50% had been removed. In the second experiment, at 14 days, this figure had risen to 80 % with the remaining fish appearing very unhealthy and so the run was terminated. At 14 days the third experiment had lost 70%.
Growth of the fish is shown in Table 1. Control fish, kept in still water, did not consume much food and did not change much over the experimental period in either length or weight. Fish swimming at the two lower speeds gave very different results from the controls. Those animals at 1·5 b.l./s appeared to be very healthy, which was shown at the end of the experiment by an 11 % increase in length and a 79 % gain in weight. At 3 b.l./s a much smaller length increase was observed (1·4%) and the weight had risen by only 7%. Animals surviving 14 days at the highest speed were very unhealthy, and although they consumed large quantities of food there was a slight decrease in length and a 21 % loss in weight.
The efficiency of conversion of ingested food into fish flesh was recorded at the two lower speeds (Table 2). At the lowest speed, the conversion of food was much more efficient than for the controls, whereas at the intermediate speed the energy required for swimming greatly reduced the efficiency.
At all speeds involvement of the red muscle fibres was indicated by an increase in both number and size (Table 3). A difference in fish from different years can immediately be seen. The lowest-speed experiments were carried out in the spring of 1974 with a single batch of animals. The red muscle was characterized by a relatively low number of fibres with a large mean diameter. Exercise produced hypertrophy in this muscle by a small increase in fibre diameter (14%) and a large increase in the number of fibres (40%). In order to have the experiments at the same time of year, and with fish of equal age, the two higher-speed experiments were carried out in the spring of 1975, again using a single batch of animals. The red fibres in this group were characterized by a large number of small diameter fibres. Exercise produced a much smaller increase in the number, 18% at the intermediate speed and 20% at the highest. Increase in muscle mass at the intermediate speed was associated with a 31% increase in fibre diameter, while at 4·5 b.l./s the result was complicated, since hypertrophy of the fibres was apparently affected by the poor nutritional state of the animals in the group, resulting in only a 13·5% increase.
Pink muscle fibres did not show any hypertrophy at the lowest speed, but at the intermediate speed there was an 11 % increase in fibre diameter. At the highest speed, this result was again apparently a consequence of the poor state of the fish, and a 12% decrease in pink fibre diameter was observed. White muscle fibres increased in size at all swimming speeds; by 7% at the lowest, by 16·5% at the intermediate, and by 6·5 % at the highest. The low value at the highest speed was believed to be due to the lack of available energy matter rather than a direct effect of swimming at this speed.
At the lowest speed the fish almost doubled their body weight. This was reflected in the lateral muscles by large increases in glycogen and lipid, and by a rise in the protein concentration (Table 4). Glycogen stores were very much increased, rising by 419% in the red muscle and by 395% in the white. Lipid changes were smaller at 57% and 33% respectively. Protein concentrations rose by 12% and 40% in red and white muscle. The large increase in lipid stored in the red muscle was reflected by a drop in the water content of this tissue. There was no change in the water content of white muscle.
At the intermediate swimming speed, the glycogen stores were 251% and 290% greater than the controls for red and white muscle respectively. Although a less marked effect than at the lowest speed, these still represented large increases. The major source of energy at this speed appeared to be lipid; although glycogen was being accumulated, the lipid was being utilized, both that ingested and that which was already present stored in the muscle. The amount of lipid stored in the muscle fell by 14% and 27% in red and white tissue respectively. The water content of both muscles did not change and neither did the protein content.
The poor condition of the animals at the highest speed was indicated by the very low concentrations of stored material in the muscles. Glycogen had dropped by 77 % and 84% respectively in the red and white muscles, the levels being hardly detectable. Much of the stored muscle lipid had been mobilized, as shown by an 85% drop in red muscle and a 60 % drop in the white. The water content of both muscles had increased, while protein levels in the red muscle had dropped only slightly compared with the white tissue which had lost 30% of its initial value.
ELECTRON MICROSCOPE STUDIES
The ultrastructure of the red and white myotomal muscles of teleost fish has been well documented (Nag, 1972; Patterson & Goldspink, 1972). The brown trout examined in the present study had a similar structure. At all speeds, exercise did not appear to affect the structure of the white muscle ; that is, any changes that did occur, such as hypertrophy of the fibres, were not noticeable under the electron microscope. In contrast to this, very noticeable changes occurred in the red muscle fibres at all speeds. At the lower speeds, large changes in the amounts of stored metabolites could often be seen. Fig. 2 (a, b), from a control fish (a) and a fish exercised at the lowest speed (b), shows these changes. In addition to this, exercise affected the oxidative capacity of this muscle which was seen by marked enlargement of the mitochondria (Fig. 2a, b). This happened at all speeds, even the highest.
During the course of this study, in almost every experiment, approximately 20% of the trout died. This always occurred during the early stages when the animals were at a low swimming speed. It appeared that these animals were simply unable to adjust to their new environment of continuously flowing water and fell back exhausted on to the lower grid. It is well known that hatchery-reared salmonids do not have a high survival rate when transferred into running water. Needham & Slater (1945) found that 66% and 44% of two populations of rainbow trout were lost after transplantation, and Miller (1951) noted that 33% of a 3-year-old, and nearly all of a 2-year-old cutthroat trout population died within the first 2 weeks. Many factors are involved in the survival of hatchery trout after relocation, such as availability of food, acceptance of the new food, predation, etc. This study indicates that one factor, the change in environment from the relatively still water of the hatchery pond to the flowing water of the stream, probably accounts for about 20% of the population. It would be interesting to study the effects of exercise on very young trout fry to determine the mortality occurring in these fish and to see whether this inability to adjust to a flowing water environment was a feature possessed by specific members of a population.
The control fish kept in still water did not grow well and some of the fish actually lost weight. This was in marked contrast to the animals kept in running water at 1·5 b.l./s, which ingested much food, grew very rapidly and made a much more Efficient use of the food than the controls (Table 2). After the 28-day period they had almost doubled their weight. This appeared to be achieved by much protein deposition in white muscle and by a build-up of large stores of both lipid and glycogen. In addition, the mitochondria increased in size quite significantly. This is because the red muscle is the active tissue at this speed, although the hypertrophy was also observed at the other two speeds indicating that it was also active there.
At the intermediate speed, 3 b.l./s, much food was ingested but very little growth resulted. At this speed, which was sufficient to involve the white muscle, most of the ingested material was being utilized to provide energy for swimming. The majority of energy was apparently being provided by the oxidation of lipid, as indicated by the hypertrophy of the mitochondria. In addition, whereas glycogen was being built up, the amount of lipid was decreasing despite the amounts taken in each day. Thus it would appear that lipid is the major fuel at this speed in both red and white muscle. The large drop in the lipid content of the white muscle is interesting as this muscle type is very anaerobic. However, it may be that at this speed, which is much lower than the critical speed of 4·5 b.l./s for these fish (Davison, 1976), the white muscle is not being used to any great degree and thus if the very small fibres of the mosaic are the ones in use (Davison & Goldspink, 1977), enough oxygen could be reaching the white tissue to allow oxidation of lipid to occur. In a recent report, Cooper & Hudlicka (1976) showed that long-term stimulation of a mammalian muscle increased the capillary density allowing much more blood to reach the muscle fibres. This may well have occurred in trout white muscle at this speed and is at present being investigated.
At the highest speed many of the animals were unable to survive. Depletion of the tissues was very obvious, and at this speed glycogen appeared to be very important, as indicated by the almost non-existent levels. This is presumably because at such a high speed more energy would be required than could be obtained using the amount of oxygen available in the blood, thus anaerobic respiration would be required. Much protein had been mobilized, but only in the white muscle, leaving the red tissue relatively unaffected. This may have been connected with the production of lactate in the white muscle, as protein breakdown is known to have a role in buffering the acidic effects of lactic acid (Kutty, 1972). This, however, is a wasteful process as valuable energy in the form of amino acids is lost by excretion. This process is also very unhealthy for the fish as at this speed the white muscle would be doing most of the work, so if the muscle-protein, which is mainly contractile, was being mobilized, then the fish would become progressively less able to work against the water current. Thus the large drop in protein content is no doubt a final resort as far as energy production is concerned.
Black et al. (1962) have shown that at high speeds the glycogen content of white muscle in salmonids rapidly goes down. As the animals used for analysis were still alive after 14 days of high-speed swimming, then all stored glycogen in the white tissue would have been used up long before the end of the experiment. So to keep functioning, the white muscle must have been obtaining energy from other sources. It is possible that the protein depletion was due to the animal obtaining energy this way, though as breakdown of protein involves the consumption of much oxygen this seems unlikely. It is probable that fuel was received as glucose via the blood system. If this was so, then the proposal of Wittenberger (1972, 1973) that red muscle (in addition to other stores such as the liver) supplies the white with carbohydrate might in this case be valid, as the red muscle would have little locomotory function at this speed.
A weight drop of 21% was observed in trout at the highest speed. In starving fish this is by no means severe, as several authors have noted much greater depletion; e.g. Creach & Serfaty (1965) noted that starved Cyprinus carpio lost 47% of their initial weight, and Boethius (cited by Love, 1970) reported that an eel had lasted more than 4 years without feeding, losing no less than 76% of its body weight. However, the major difference between these studies and the present one is that they were carried out on quiescent animals. Starved fish become very inactive and thus save energy. The fish in the flume were forced to be active, and it is believed that the observed weight loss was critical, and it is unlikely that they would have survived much longer, as indicated by the protein measurements. Certainly it would seem that an active animal cannot afford to lose as much weight as a quiescent one.
The number of fibres in most mammalian muscles is fixed at birth so that any growth of a muscle is achieved by enlargements of the existing fibres (Goldspink, 1972). An increase in fibre number can occur, but only if the muscle is subjected to an extreme load (Rowe & Goldspink, 1968; Hall-Craggs, 1970). This situation does not occur in fish muscle tissue. Greer Walker (1970) studied changes in the red and white myotomal muscles of the cod Gadus morhua with growth and found that the total number of fibres increased with the size of the fish. He also noted that the fibres tended to reach a maximum size after which increase in muscle mass was achieved by an increase in fibre number. In the present study similar observations were also made in the red muscle of the trout. The two populations of fish had very different growth patterns when kept in the still water of the hatchery ponds, one having fewer larger diameter fibres and the other having many smaller diameter fibres. Exercise produced hypertrophy of the red muscle in both groups, but in different ways. The size of the cells appeared to have almost reached maximum size in the first group; although they were smaller than neighbouring white fibres, this was necessarily so, to allow efficient aerobic metabolism. Hypertrophy was thus attained by a large increase in the number of fibres. In the second group the fibres were much below the maximum size and so in this case increase in red muscle mass was achieved mainly by simple increase in the myofibril content of the existing cells. Exactly how the increase in cell number is achieved is not exactly clear. In grossly overloaded mammalian tissue, increase in cell number is thought to be achieved by splitting of existing ones, producing two smaller daughter cells, although this is an abnormal increase. At the light microscope level in trout, it appeared as though some of the cells were in the process of splitting. Under the electron microscope very small fibres could occasionally be found in both red and white muscles. This tends to indicate that normal fibre number increase may be brought about by the production of new fibres and not by the splitting of existing ones.
It is apparent from the results presented here that exercise does have profound effects on the growth of muscle tissue of fish and the quantification of these changes at the cellular level is obviously important. The interplay between diet and the level of exercise is complex and further work is needed to establish the optimum levels for other species of fish, particularly those which are farmed commercially.
This work was carried out while one of the authors (W. D.) was in receipt of an NERC studentship.