ABSTRACT
The drinking rate and the oxygen consumption were measured in Aphanius in sea water (SW) at 17, 20 and 24·5 °C and at 20 °C in 2-fold and -fold SW. Both untrained (shocked) and trained fish were used. In this species shock was observed to reduce the drinking rate.
In trained fish the drinking rate was 11·0±1·0μl/g.h (mean μ S.E.) in SW at 20 °C. The oxygen comsumption was 346± 16 μl O2/g.h. These parameters were not significantly changed in 2 SW and SW.
At 17 °C both drinking rate and oxygen consumption were reduced, and at 24·5 °C were increased.
The results suggest that adaptation to waters of high salinity primarily involves a reduced effective osmotic permeability of the integument (the gills) and an increase in the transport rate of NaCl by the intestine, roughly proportional to the salinity.
INTRODUCTION
When euryhaline teleosts are adapted to waters of salinities higher than the osmolality of plasma an increase in drinking rate has been observed (Lotan, 1969; Maetz, 1970; Maetz & Skadhauge, 1968 ; Potts et al. 1967 ; Shehadeh & Gordon, 1969). This is to be expected since drinking, and absorption of salt (NaCl) and water through the intestinal wall, is the only way for the fish to get water to make up for the osmotic losses. The cyprinodont, Aphanius dispar, can tolerate very high salinities of the surrounding water (Lotan, 1971). It would be of interest to investigate what permits the adaptation to high salinity. It has been suggested that tolerance to high salinity is determined by the capacity of the intestine to absorb the NaCl (Maetz & Skadhauge, 1968; Skadhauge, 1969). It has also been suggested that the overall osmotic permeability coefficient of the integument goes down during adaptation to higher salinity, since the drinking rate is not augmented proportional to the osmolality difference (Shehadeh & Gordon, 1969; Motais et al. 1969).
In the work here reported the drinking rate has been studied in Aphanius adapted to higher salinities, because it is a measure both of the effective osmotic permeability of the integument and of the intestinal salt absorption. The latter has already been studied in vitro (Lotan & Skadhauge, 1972). Since it has been reported recently that drinking rates were low in eels at low temperature (Motais & Isaia, 1972) and during inactivation (Kirsch, 1971), probably caused by a low rate of metabolism, the oxygen consumption was also measured.
MATERIAL AND METHODS
Fish
The fish were collected in March 1972, in a spring in Ein Fashkha at the north-western bank of the Dead Sea. The fish were also observed swimming in the estuary zone in the Dead Sea. Analyses of the waters are reported in Table 1. The fish were transferred the same day by direct flight from Tel Aviv to Copenhagen under an oxygen atmosphere; no mortality followed. Other fish were studied in Jerusalem.
Maintenance
The fish were kept in aquaria with artificial sea water (see Lotan & Skadhauge, 1972) of an osmolality of approximately 1050 mOsm (SW), 2000 mOsm (2 SW) and 3500 mOsm ( SW). (In the previous publication the Mg2+ concentration was reported an order of magnitude too low due to a printing error.) The aquaria were well aerated and the water passed through charcoal filters. Representative values of oxygen content (μl O2/ml, mean ± S.D.) and pH of the waters are: SW, 4·93 ± 0·07, 7·6; 2 SW, 3·91 ± 0·04, 7·8; and 3·5 SW, 2·97 + 0·03, 7·9. They were fed on Tetra-Ovin food. The temperature was maintained at 20 ± °C. In other experiments a temperature of 17 ± °C was maintained using a cooling coil, or a temperature of 24·5 ± °C was maintained by a heating board with automatic control.
Experimental procedures
(1) Stressed fish
These fish were taken abruptly by net and transferred to a small vessel containing the water marker (see below).
(2) Trained fish
The trained fish were never taken out of water. They were adapted for at least one week to a subcompartment in a large aquarium. They were daily allowed to swim in and out of a smaller aquarium placed in the subcompartment, the water was gently stirred and an aerator was introduced. When an experiment was run the level of water was gently lowered in the small aquarium to approximately 2 l, containing 4–6 fish, and the water marker was introduced into a small volume of the medium at the same temperature.
Operation
After 2 h (in Jerusalem 1 h) the fish were removed one by one by net, transferred to another net and flushed with 50 ml 0·9 % saline. They were gently blotted with absorbing paper. They were weighed to the nearest 0·01 g, the tail was cut off and blood was sucked into a capillary for determination of plasma concentrations (Table 2). Some fish were anaesthetized in the small aquarium. This change in procedure did not cause a measurable change in drinking rate.
The fish was placed on its back under a stereo microscope. The abdomen was opened and the walls held aside. The instruments were cleaned and the intestine was carefully dissected out in toto after closing the upper end towards the oesophagus at the operculum. Gall-bladder and liver were meticulously dissected free. The whole intestine was dropped into a small tube, which was placed in the vials for counting the activity of the water marker. Bile was collected in many experiments for analysis of osmolality (Table 2).
In Jerusalem the fish were flushed with saline and the scales were scraped off, the gills were removed, and the tail was cut off as described by Lotan (1969). The radio-activity of the remaining fish was counted.
Measurement of oxygen consumption
Single fish were made to swim into 100 ml glass flasks and left overnight in the main aquaria with individual aerators in the flasks and without food. At the start of the experimental period a sample was taken of the water, the aerator was removed, and the stopper was put in. After 45–60 min the experiment was terminated, the flasks were removed, the fish was taken up and weighed, and samples of the water were taken. The fish weighed 0·3–2·1 g.
Water markers
As unabsorbable water markers [51Cr]EDTA (Hoechst, Germany) was used in Jerusalem and [125I]polyvinylpyrrolidon (Radiochemical Centre, Amersham, England) was used in Copenhagen.
Concentration in intestinal fluids
In these experiments the fish was left in a small aquarium with water marker, as mentioned above, for 24–26 h without food. It was killed and handled as described above and the abdomen was opened. Capillaries were then introduced into the most anterior part of the intestine just below the operculum, and through the anus into the lower end of the posterior intestine. The fluid in some of the capillaries was used for determination of osmolality and sodium ion concentration. The fluid in other capillaries was measured by length and the radioactivity of the capillary was counted. One cm of the capillaries corresponds to 1·049 ± 0·056μl (S.D.), n = 10. The activity of the surrounding medium was measured allowing calculation of the fractional resorption of water (volume). The radioactivity of plasma was measured in the same way.
Analyses
Osmolality was measured on the Ramsay & Brown (1955) micro-osmometer. Na+ was measured on an Eppendorf flame photometer. The water markers were measured on the Packard Auto gamma spectrometer. Oxygen content was measured by the Winkler method.
Control experiments
In the Jerusalem experiments the unavoidable surface contamination was judged from the activity present on fish which had been in the experimental solution for 6 min. It amounted to approximately 5 % of the activity of the 1 h fish. In dead fish a similar activity was found. This activity was subtracted.
In the Copenhagen experiments the following plasma concentrations (as percentage of the surrounding water concentration) were found: SW 17 °C, 2·1 ± 1·5, n = 4; SW 20 °C, 1·3 ±0·9, n = 16; SW 24 °C, 0·9 ± 0·8, n =13; 2 SW 2·4±1·1, n = 3; SW 1·0± 0·4, n = 7 (mean ± S.D.).
In the 2 h experimental periods a substantial fraction of the activity did not reach the posterior part of the intestine. In four experiments the following fractions of the total intestinal activity was present in the posterior intestine: 5·8, 6·0, 3·3, 1·2%.
Attempts were made to include the oesophagus at the operculum in the intestine to be counted. Since it was difficult to obtain a clearly defined boundary the activity of this zone was counted separately in three experiments. It amounted to 1·1, 2·1 and 1·0% of the total intestinal activity.
The validity of the calculation of fractional water resorption on the basis of water marker in the capillaries was checked by filling these with the surrounding fluid and calculating the counting rate, which should be unity. In ten capillaries a recovery of 99·6 ± 2·1 % was found (mean ± S.D.).
The pipetting error showed a coefficiency of variation of 0·41 %.
RESULTS
Drinking rate in stressed fish
In untrained fish abruptly transferred to small experimental vessels the drinking rate was at roughly the same level from fresh water (FW) to 2 SW, around 7 μl/g.h (Table 3). These drinking rates were slightly lower than those observed in undisturbed fish (see below). In the high-salinity groups drinking almost stopped. As a sign of stress in the untrained fish rapid movements of the operculum (ventilation) were observed.
Drinking rate in undisturbed fish
The data are reported in Table 4. In SW-adapted fish at 20 °C an average value of 11·0μl/g.h was observed, and this value was not significantly changed by adaptation to 2 SW or to SW.
This rather unexpected finding warrants two conclusions (see further discussion) : first, that the effective osmotic permeability of skin and gill must be reduced in waters of high salinity; secondly, that the intestinal NaCl absorption rate roughly parallels the salinity.
Reducing the temperature to 17 °C, which is close to the lower limit of survival for Aphanius, approximately halved the drinking rate, whereas augmenting the temperature to 24·5 °C doubled the drinking rate on the average. The observations at high temperature were, however, highly variable. This has been also observed at 25 °C by Motais & Isaia (1972) in the eel. These authors also found a quadrupling of the drinking rate by a temperature increase of 8–10 °C. In the undisturbed fish rapid movements of the operculum were not observed. The drinking rates were lower than observed previously (Lotan, 1969) in a different population of Aphanius.
Oxygen consumption
The oxygen consumption was measured in the conventional way during 24 h sojourn in the experimental vessels without food. The results are reported in Table 5.
Fish in SW at 20 °C consumed 346± 16 μl O2/g.h. At higher salinities a reduction was not observed, an increase of the average value was apparent in 2 SW, no change was seen in SW. At a lower (17 °C) and a higher (24·5 °C) temperature a reduction and augmentation respectively of the O2-consumption were observed. The observed change, roughly a doubling per 10 °C, was identical to that found in Fundulus parvipinnis (Wells, 1935a). The absolute values agree well with those observed in other cyprino- donts. Wells (1935b) observed in Fundulus parvipinnes at 16 °C for 0·5 g fish an oxygen consumption of 235 μl/g.h. Bertalanffy (1951) reported for Lebistes reticulatus of 0·5 g a consumption of 300μl/g.h in females and 400μl/g.h in males (temperature not stated).
This agreement suggests that our fish also were observed in an undisturbed resting state.
Composition of the gut fluid
In some preliminary experiments fluid was taken out of the intestine in capillaries after the fish has been in surrounding water with the unabsorbable water marker for 24 h. The results are reported in Table 6. The available data indicate no major difference between the different salinities. The average concentrations in the anterior intestine were (mean ± S.D.): osmolality, 527±116 mOsm (n = 19); Na+, 125±40 m-equiv/1 (n = 10); 125I ratio, 1·37 ± 0·36 (n = 9). This seems to indicate that a major dilution takes place together with a vigorous sodium absorption with solute-linked water flow. In the posterior intestine the osmolality remained close to plasma osmolality, and the water marker concentration indicated approximately 80 % fractional resorption of water. It was difficult to obtain liquid samples from the lower end of the intestine of fish in SW. By visual inspection it was apparent that less precipitate was present in these fish, probably reflecting the lower CaCO3 concentration in the surrounding water as compared to the fish in 2 SW and SW.
DISCUSSION
In this paper the physiological parameters were measured after full (time-independent) adaptation to a given salinity. Full adaptation is in general found a week after transfer to a new salinity (R. Lotan, unpublished experiments). In waters of high salinity the osmoregulation of Aphanius functions remarkably well. The plasma osmolality is only augmented by 12 % by passage from SW to 2 SW and by a further 11 % on adaptation to SW. Bile osmolality was on the average 2 % lower than plasma osmolality.
The essential findings of the study are summarized in Fig. 1.
Water balance
Two findings in this study were surprising: first, that stress seemed to reduce the drinking rate, and secondly, that high salinity did not result in a higher water intake. Disturbed fish have generally been claimed to drink more than undisturbed fish. A high drinking rate might in some way be connected with the increased ventilation which is seen in the disturbed fish (Davis & Cameron, 1970). This might lead to a larger blood flow. A greater gill perfusion than required for uptake of oxygen might lead to a water loss further augmented if a larger fraction of the blood is directed to the respiratory lamellae (Steen & Kruysse, 1964). A low drinking rate might for the same reason be the result of inactivation. Recently very low drinking rates have been observed in various fish at low temperature (Motais & Isaia, 1972; Isaia, 1972) and indirectly inferred from the branchial Cl- exchange in silver eels in an inactive state (Kirsch, 1971). The ‘undisturbed’ fish in this study during their normal life in the aquarium, maintained for months, were probably no less disturbed than during the measurement of drinking rate. Thus the measurement probably reflects the steady-state conditions for the animal. The situation is probably not different from that during real life in the natural habitat where waves and other animals may also disturb them. It is not known why the stressed fish drank less in this study. Since a quite high mortality was found in these fish, it is possible that the stress put them in such a poor condition that they simply could not drink.
The absolute level of drinking, 1100 μl/100 g.h, is identical to that of another fish of the same weight, Tilapia mossambica (Potts et al. 1967).
The lack of increase in water intake in high-salinity water is probably due to a decreased effective osmotic permeability of the gills. It was not associated with a significantly decreased rate of metabolism. This often occurs in higher vertebrates subjected to dehydration. If the same amount of water is drunk in spite of an increase in the osmotic difference from plasma to medium of 140% from SW to 2 SW and of 340 % from SW to SW, it can be concluded that the integumentary water loss must be roughly the same. This indicates a pronounced reduction in the osmotic permeability coefficient. The reason for a lower effective osmotic permeability could either be haemodynamic changes, which might result in a smaller average osmotic difference, or a real change in the permeability of the epithelium. A lowering of the reflexion coefficient to salt of the epithelium is probably less likely for energetic reasons. Such a high permeability to salt would require a real pump-leak mechanism and thus extrusion of much more NaCl than that coming from the intestinal absorption. Active intestinal absorption of the NaCl drunk in 3·5 SW (11 μl/g.h) implies an absorption of 33 μ-equiv/g.h and corresponds to an oxygen consumption of 1·65 μM/g.h assuming 20 Na+ ions transported per molecule O2 (Ussing, 1960). This is an unusual high fraction (11%) of the total oxygen consumption (15·4μM/g.h) used for intestinal transport. Without a leak mechanism the gill extrusion would require the same energy. Although the gills have been reported capable of extruding more NaCl than the intestine can absorb (Meyer & Nibelle, 1970), a leaky membrane with a reflexion coefficient of 0·25 would probably require more energy than available, due to the extra NaCl inflow. The salt intake, at the same drinking rate, in waters of higher salinity is doubled in 2 SW, and increased -fold in SW. The intestinal salt absorption is presumably augmented to the same degree (Kristensen & Skadhauge, 1974). Shehadeh & Gordon (1969) found at higher salinities starting from plasma osmolality an intestinal sodium absorption proportional to the increase in salinity above that of plasma and a higher fractional salt absorption in the intestine. The same degree of adaptation is seen when the yellow European eel (Skadhauge, 1969, 1974) is exposed to waters of high salinity. If, on the other hand, the osmotic permeability was unchanged, 4·8 times more water would have to be drunk in SW than in SW. This would require a tenfold larger intestinal salt transport (Kristensen & Skadhauge, 1974). This would probably be beyond the limit of adaptation.
Composition of gut fluid
The data are preliminary, but there are sufficient measurements to prove that the fluid in the anterior end of the intestine is diluted to close to plasma osmolality even in SW. This is compatible with the expected osmotic inflow into the intestine, but surprisingly the water marker concentration was not below unity. This finding needs confirmation. No other experiments of this type have been performed, but in those fish in which intestinal ionic concentrations have been measured the fractional water absorption can be deduced from the concentration of Mg2+ and This confirms our observation. In the anterior intestine of the goose fish (Smith, 1930), the rainbow trout (Shedadeh & Gordon, 1969) and the southern flounder (Hickman, 1968) the Mg2+ and concentrations were above that of the SW. The luminal sodium concentration has in general been observed to be lower than that corresponding to the dilution suggested by the osmolality, indicating a vigorous sodium absorption. If this sodium absorption is accompanied by a large amount of solute-linked water the lack of dilution of the water marker is understandable. It is possible that Aphanius is particularly adapted to transport more water per sodium ion than other fish. In the absence of an osmotic difference Aphanius (Lotan & Skadhauge, 1972) showed a larger water absorption per sodium ion than the eel (Skadhauge, 1974).
ACKNOWLEDGEMENTS
We thank Dr J. Böetius for stimulating discussions and assistance with the Winkler analysis. Mrs Bente Hessing Jorgensen provided excellent technical assistance.
The work was supported by The Danish Medical Research Council and ‘NOVO’s Fond’. E.S. received a travel grant from ‘Tribute to the Danes’.