Oxygen Consumption and Swimming Performance in Hypoxia-Acclimated Rainbow Trout Salmo Gairdneri

Fish swimming at high sustained speeds may experience an oxygen limiting situation similar to that incurred during ambient hypoxia. At swimming velocities approaching the critical swimming speed, the metabolic demand for oxygen may be greater than can be provided by ventilatory and circulatory systems (Jones & Randall, 1978). When hypoxia and exercise are combined, this oxygen limitation must be enhanced. Jones (1971) demonstrated that either moderate hypoxia or anaemia resulted in a reduction of critical swimming speed in trout. Presumably, hypoxia and exercise together act to reduce the ability of the gills to take up oxygen, while exercise and anaemia will affect the oxygen exchange as well as the oxygen transport system. Hypoxia-acclimated fish have been demonstrated to have blood respiratory properties resulting in improved oxygen loading capabilities (Wood & Johansen, 1972). It is not known whether this property is advantageous during exercise in normoxic or hypoxic water. The aim of this study was to determine whether hypoxia acclimation of trout results in an improved ability to maintain oxygen consumption and swimming performance, compared to normoxic fish, when both groups are tested at variable ambient and swimming speed in a specially designed swimming respirometer. A similar study on goldfish (Kutty, 1968) demonstrated that hypoxia acclimation had no significant influence on swimming performance.

Rainbow trout (Salmo gairdneri Richardson) of 250−350 g weight and 29−30 cm total length, were used. Fish were held in several well-aerated tanks at 15 °C and fed trout pellets regularly. Three days before the first swimming trial, 10 fish were selected and moved to a smaller aquarium under similar conditions, but not fed. A separate group of 10 fish was acclimated to hypoxia in an aquarium in which of the water was controlled (±3 mmHg) by an oxygen regulating system (Wood, Johansen & Weber, 1975). Oxygen tension was lowered over a 4-day period to the final oxygen tension of 40 mmHg. Fish were acclimated to stable conditions for at least 2 weeks before the first experimental run was commenced. Feeding was also ended 3 days prior to the first run for the hypoxia-acclimated fish.

Oxygen consumption was measured in a Brett-type tunnel respirometer, previously described by Christiansen, Lomholt & Johansen (1982). Oxygen tension, temperature and water velocity in the respirometer were monitored and recorded on magnetic tape by a Hewlett-Packard 9825 A calculator. Temperature and were also recorded on a Hewlett-Packard 7132 A two-channel chart recorder. The HP-computer was programmed automatically to sample -data, and control and shift between preset water values.

Oxygen tension of the water was measured continuously by a Radiometer oxygen electrode (E-5046) mounted in a cuvette (D616) and thermostatted at the experimental temperature of 15 °C. A constant flow past the electrode was maintained by an Ismatec Mini-s roller pump mounted downstream of the cuvette. The oxygen electrode was repeatedly calibrated with zero solution and air-saturated water.

Oxygen tension in the water feeding the respirometer section of the swimtunnel was maintained and controlled by shunting a fraction of the water leaving this section through a Sci-Med Kolobow membrane gas exchanger (lung) with a 3·5 m² surface area.

of control or normoxia-adapted fish was determined at four water speeds (5, 20, 35 and 50 cms⁻¹) and three oxygen tensions, normoxia (145 mmHg) and hypoxia (40 and 60mmHg).

Swimming velocities of fish in the respirometer were corrected for solid blocking effects of fish body cross-sectional area (Webb, 1975). Actual swimming velocities at the four pre-selected water speeds were therefore 5·5, 21·9, 38·3 and 54·8 cms⁻¹. These corrected swimming velocities have been used in all further discussions.

The fish to be tested was acclimated overnight to the swimming chamber at the first trial swimming speed of 5·5 cm s⁻¹. Oxygen tension was kept higher than 145 mmHg throughout this period. was measured by starting at the lowest swimming speed (5-5 cms⁻¹) and increasing the speed in a stepwise manner to 54·8 cms⁻¹. In order to ensure that oxygen consumption was at a steady state, the fish were made to swim for 30 min at the chosen swimming speed before measurements began. At this time, air and water flow through the gas exchanger was stopped and the fish allowed to swim for an additional 40 min to measure ⁠. Water speed was then increased, and was brought back to the level of 145 mmHg. At the conclusion of measurements at 54·8cms⁻¹ in normoxia, swimming speed was reduced to 5·5 cms⁻¹ whereafter nor-moxic conditions were maintained for 30 min. was then lowered over the next 30 min by flushing the membrane gas exchanger with nitrogen. When had stabilized at the pre-determined oxygen tension (40 or 60 mmHg) for 5 min, measurements were again started. Oxygen tension during the 30 min steady state swimming period was kept at the hypoxic test condition. In the case of animals swimming at 40 mmHg, was raised to 45 mmHg 3−4 min before measurements began, and was never allowed to fall below 30 mmHg during this phase. An animal was judged to have collapsed when it could no longer keep itself off the electrified grid at the rear of the swimming chamber.

In a separate experiment to measure possible oxygen debt, fish from the control group were put through the same experimental protocol used to measure at the four swimming speeds in normoxia. At the conclusion of the swimming trial at 54·8cms⁻¹, water velocity was reduced to 5·5 cms⁻¹ and monitored for the next hour.

Hypoxia-acclimated fish were tested in a manner similar to control animals (normoxia-acclimated). During overnight acclimation to the chamber, was kept at 40 mmHg by the oxygen regulating system that controlled water and air flow through the membrane gas exchanger. At the end of this period, oxygen consumption kvas measured at the acclimation (40mmHg) and swimming speed (5·5cms⁻¹). Oxygen tension was then raised to air saturation and measurements in normoxia were begun at 5·5 cm s⁻¹ and proceeded through the four speeds in normoxia and hypoxia. Therefore, unlike the control group, two measurements of at 5·5 cm s⁻¹ in hypoxia were made for the hypoxia-acclimated group. The first, taken at the conclusion of the overnight chamber acclimation and before was raised for normoxia swimming trials, will be referred to in the text and table as ‘prenormoxia’ ⁠. The second was made when was reduced to hypoxic levels at the beginning of hypoxia trials and is similar in its temporal relationship to the measurement made for control animals.

Maximum swimming velocity (V_max) was calculated for all fish failing to swim at the highest velocity. An interpolation was performed to determine swimming velocity if the fish did not collapse exactly at the beginning or end of the swimming period by adjusting the last velocity increment in proportion to the amount of time spent at that velocity (Brett, 1964; Webb, 1971).

V_max for fish completing trials at 54·8 cm s⁻¹ could not be calculated since maximum swim-velocity was never attained due to water speed limitations of the respirometer. For the purpose of calculation and discussion, however, V_max for these animals was taken to be 54·8 cm s⁻¹. Vm., reported for conditions at which some fish completed all trials (normoxia and 60 mmHg) will therefore probably be underestimates of the true maximum swimming velocity (critical swimming speed).

For blood sampling, fish were netted and held between two wet towels to facilitate rapid (30 s) sampling. A heparinized plastic syringe was used to draw approximately 2ml blood, percutaneously from a caudal vessel. Blood Hct, pH, oxygen affinity (P₅₀), ATP and GTP concentrations were determined as described by Tetens & Lykkeboe (1981).

Statistical analysis of results was based on the Student’s t-test. All values in the text, table, and figures are given as mean ± standard deviation.

Results of oxygen consumption determinations in normoxia (control) and hypoxia-acclimated fish in normoxia and hypoxia are summarized in Table 1 and Figs 1 and 2. Neither group of fish had any apparent difficulty completing swimming trials through all four velocities in normoxia. Differences in swimming performance began to emerge in hypoxia trials. At water O₂ tensions of 60 mmHg, three out of eight control fish could not complete the test at 54·8 cm s⁻¹. Swimming trials at a water of 40 mmHg resulted in two out of eight fish failing at 38·3 cm s⁻¹, while all failed at 54·8cms⁻¹. All eight hypoxia-acclimated fish successfully completed the 38·3 ems-¹ trial in hypoxia (40mmHg), but none were successful at 54·8cms⁻¹.

Maximum swimming velocity, shown in Table 1, in control fish decreased from a value of ⩾54·8 cm s⁻¹ in normoxia to 51·5 ± 6·0 cm s⁻¹ at 60 mmHg. A further reduction of to 40 mmHg resulted in a decrease of critical swimming speed to 41·4 ± 5·7 cm s⁻¹, which was significantly lower (P< 0·001) than Vmu recorded at 60 mmHg. Results were similar for hypoxia-acclimated animals. Vmo fell from ⩽54·8 cm s⁻¹ in normoxia to 40·6 ± 2·2 cm s⁻¹ at 40 mmHg.

Standard metabolic rates (SMR) listed in Table 1 were calculated from linear regressions of data from individual fish. SMR of control fish was 82·5 ± 27·4mg O₂ kg⁻¹h⁻¹ and did not change significantly when was lowered to 40 mmHg. Hypoxia-acclimated animals in normoxia had lower, but insignificantly different SMRs. Reduction of during swimming trials in this group increased SMR by 87% (P<0·001) to 124·87 ± 30·5mg O₂ kg⁻¹ h⁻¹, which is significantly higher (P < 0·05) than control fish at 40 mmHg. Recalculation of SMR in hypoxia-acclimated fish in hypoxia, utilizing the ‘prenormoxia’ values measured at 5·5 cm3⁻¹, considerably alters the slope of the regressions and lowers the SMR to 105·04 ± 24·0 mg O₂ kg⁻¹ h⁻¹. This is not significantly different from controls in hypoxia, but is still significantly higher than SMR of hypoxia-acclimated fish in normoxia (P< 0·001).

Oxygen consumption of control fish in normoxia averaged 97·5 mg O₂ kg⁻¹ h⁻¹ at 5·5cms⁻¹ and increased to 318·5mg O₂ kg⁻¹ h⁻¹ at the highest velocity (Fig. 1).

Swimming trials at 60 and 40 mmHg resulted in values that were not significantly different from those recorded in normoxia (Fig. 2).

The ‘prenormoxia’ oxygen consumption of hypoxia-acclimated fish at 40 mmHg (5·5 cms⁻¹) was 110·1 mg O₂ kg⁻¹h⁻¹ and fell significantly (P< 0·005) to 75·9mg O₂ kg⁻¹h^−l when was raised to normoxia. Increasing the swimming velocity in normoxia to 54·8 cm s⁻¹ increased to an average of 345·7 mg O₂ kg⁻¹ h⁻¹ (Fig. 2). There were no significant differences in normoxic values between hypoxia-acclimated fish and controls swimming at any swimming speed.

Reduction of from normoxia to 40 mmHg following normoxic swimming trials, resulted in an 88 % increase in in the hypoxia-acclimated animals at 5·5 cm s⁻¹. This differed significantly (P < 0·001) from measured in normoxia, as well as measured under similar hypoxic conditions before normoxia swimming trials during the ‘prenormoxia’ phase (P<0·025). Oxygen consumption increased through the remaining two swimming speeds, but did not differ significantly from in normoxia or of control animals in normoxia or hypoxia (Fig. 2).

No oxygen debt resulting from the swimming protocol in normoxia was noted in the values. Oxygen consumption prior to swimming trials was 99·2 ± 24·4mg O₂ kg⁻¹ h⁻¹ (5·5 cms⁻¹). Following the conclusion of swimming at the highest water velocity, at 5·5 cms⁻¹ had returned to 99·3 ± 23·6mg O₂ kg⁻¹ h⁻¹ within 30 min and was 94·7 mg O₂ kg⁻¹ h⁻¹ after 60 min.

Haematocrit and blood O₂ capacity in normoxia and hypoxia-acclimated trout, were not significantly different: average heamatocrits were 26·3 ±6·5% and 25·2 ± 6·0 % and O₂ capacities 9·25 ± 2·28 vol % and 8·69 ± 3·37 vol %, respectively, for normoxia- and hypoxia-acclimated fish. Red cell ATP concentration was, however, markedly higher in the normoxic group, averaging 4·86 ± 0·54 mmol 1⁻¹ packed red blood cells against 2·51 ± 0·56 mmol 1⁻¹ packed red blood cells (P < 0·001). Blood O₂ affinity was elevated in hypoxic fish expressed by a P₅₀ value of 19·5 mmHg compared to 22·9 mmHg in normoxia-acclimated fish. The Bohr factor and n-value did not vary between the groups: ∅= −0·49 and n = 1·92 ±0·25 in normoxic fish, and ∅ = −0·45, n = 2·34 ± 0·45 in hypoxia-acclimated fish.

The maximum velocity a fish can maintain for a precise period of time is termed the critical swimming speed (Brett, 1964), and will vary with the magnitude of the speed increment and the amount of time spent at each speed. Usually, a decrease in test period duration will result in an increase in critical swimming speed (Webb, 1971).

These factors, in combination with differing acclimation temperatures and size of fish will result iri a wide range of values. Critical swimming speeds are, however, a useful way of quantifying swimming performance in the same species of fish. It should be pointed out that since all of the fish in this study completed trials at the highest swimming speeds in normoxia, V_max for these fish will not be as high as their critical swimming speeds. However, in cases where fish were unable to complete the swimming trials (40 mmHg), V_max, by definition, will be the same as the critical swimming

Critical swimming speeds reported for rainbow trout at 15 °C (1-h swimming periods) range between 39·2cms⁻¹ (4·3 Ls⁻¹) (Beamish, 1978) and 58·1 cms⁻¹ (2·0Ls⁻¹) (Webb, 1971). Our study showed that V_max for rainbow trout in normoxia is at least 54·8cms⁻¹ (1·8 Ls⁻¹), since no fish failed to complete the swimming trial at this speed. This is in agreement with values of 58·1 cms⁻¹ (2Ls⁻¹) published by Webb (1971) using fish of similar size.

Reduction of ambient decreased V_max to 51·5cms⁻¹ (1·7Ls⁻¹) at 60mmHg and 41·4cms⁻¹ (l·4Ls⁻¹) at 40mmHg in control fish. A decrease in critical swimming speeds with severity of hypoxia has earlier been noted in bass (Micropterus salmoides), coho salmon (Oncorhyncus nerka), goldfish (Carassius auratus) and rainbow trout (Salmo gairdneri) (Kutty, 1968; Dahlberg, Shumway & Doudoroff, 1968; Jones, 1971). Hypoxia acclimation in the present study had no effect on V_max. Kutty (1968) found the same to be true of hypoxia-acclimated goldfish.

The failure of fish to swim at reduced was attributed by Kutty (1968) to be due to a peripheral or central oxygen sensing system, not exhaustion, since fish resumed steady state swimming when was raised. This contention is not supported by our experiments on rainbow trout. Following collapse in hypoxia, several fish had difficulty orientating themselves in a normal swimming posture even though water velocity had been reduced and ambient increased above 100 mmHg. On two occasions, was measured continuously following collapse in hypoxia. Oxygen uptake of these fish in normoxia was elevated above pre-exercise values, in one case by 85 % and by 200 % in another. These observations suggest that fish were making maximal effort under the experimental conditions and ceased swimming due to exhaustion.

Standard metabolic rate (SMR) extrapolated for control fish in normoxia (82·5mg O₂ kg⁻¹ h⁻¹) falls within the range of 72·5mg O₂ kg⁻¹ h⁻¹ (Webb, 1971) to 112mg O₂ kg⁻¹ h⁻¹ (Rao, 1968) reported for rainbow trout at the same temperature. SMR of hypoxia-acclimated fish was somewhat lower. As severity of hypoxia increased, SMR of control animals remained stable. This was not the case in hypoxia-acclimated animals, since a significant increase in SMR was noted at water of 40 mmHg. Ott, Heisler & Ultsch (1980) also noted a similar increase in nonacclimated rainbow trout during hypoxic exposure. In our experiments, SMR in control fish remained stable at reduced water of 40 mmHg. According to Ott et al. (1980) SMR in rainbow trout increased between 20 and 100 % at this before falling when the ambient was lowered below the critical point (22 mmHg at 15 °C) and oxygen uptake became dependent on ⁠. It is difficult to understand, however, why in our study the increase in SMR was apparent only in the hypoxia-acclimated fish and not in the control group.

Oxygen consumption of normoxia-acclimated fish at each swimming speed remained remarkably independent of environmental ⁠. This is, however, in agreement with previous work on less or minimally active rainbow trout exposed to hypoxic water levels of 90 mmHg (Smith & Jones, 1982) and 30 mmHg (Holeton & Randall, 1967). It contrasts, however, with the large increase in resulting from a decreased ambient (90 mmHg) which was reported by Hughes & Saunders (1970).

Oxygen uptake of hypoxia-acclimated trout in normoxia was similar to that of control fish at all swimming speeds in normoxia. With the onset of hypoxia, a significant increase (P < 0·001) in oxygen uptake at 5·5 cm s⁻¹ was recorded but did not persist at the remaining swimming speeds since at 21·8 cm s⁻¹ and 38·3 cm s⁻¹ were the same as in normoxia. measured in hypoxia at the lowest swimming speed could increase because of an oxygen debt incurred during the preceding swimming trials in normoxia. We believe, however, this to be unlikely since there was no significant elevation of in control fish when swimming in 60 or 40 mmHg at 5·5 cm s⁻¹. This was further substantiated by the results of tests conducted to determine the possibility of an oxygen debt resulting from the experimental protocol followed in the normoxia experiments. These tests showed that oxygen consumption in normoxia, measured at 5·5 cm s⁻¹ before and after the swimming protocol used in the experiment, were the same. Finally, a similar change in was noted when ‘prenormoxia’ was compared to normoxic In this case, when fish were taken from an ambient of 40 mmHg to normoxic levels, rather than vice versa, decreased 31 %. Despite the opposite direction in the change of ambient ⁠, in both cases was higher in hypoxia than in normoxia. An increased of hypoxia-acclimated fish in hypoxia at low levels of activity, relative to non-acclimated fish, was reported in carp (Lomholt & Johansen, 1979) and flounder (Kerstens, Lomholt & Johansen, 1979). Oxygen consumption of acclimated carp was 40% higher while acclimated flounders maintained at 100 % above controls in hypoxia. It should be noted, however, that the increased values were significantly lower than recorded in normoxia, even though ambient was above the critical of 21 mmHg earlier reported for carp (Ott et al. 1980).

The acclimation history of the fish in our study may have influenced the results by giving an excitement or stress factor not present in control animals. Hypoxia-acclimated animals experienced two dramatic changes in during the course of the experiment. Following 3 weeks of acclimation to of 40 mmHg, water tension was raised to air saturation and returned to 40 mmHg 4 h later. Such changes in of water may affect the physiological responses (blood O₂ transport properties) of hypoxia-acclimated fish differently from fish facing a single change in ambient water conditions (Soivio & Nikinmaa, 1981).

An increased oxygen uptake in response to decreasing recorded in lamprey, carp, goldfish, brook trout and rainbow trout has been attributed by many authors to increased cost of ventilation during certain phases of exposure to hypoxia (Beamish, 1964; Marvin & Heath, 1968; Hughes & Saunders, 1970; Nikinmaa & Weber, 1984). Gill water flow measured directly in trout (Randall & Jones, 1973; Smith & Jones, 1982) was shown to double when environmental was reduced from normoxia to 90 mmHg, while a reduction to 40 mmHg doubled VG in plaice and flounder (Steffensen, Lomholt & Johansen, 1982) and increased it six times in carp (Lomholt & Johansen, 1979). It is tempting to apply the same argument to explain the decreased in hypoxia-acclimated trout when was raised to normoxic levels. Perhaps, in response to suddenly increased ⁠, hypoxia-acclimated fish respond with a reduction in ventilation volume similar to the reduction in gill water flow recorded in carp, goldfish, and trout, when exposed to hyperoxia (Dejours, 1973). A potential lowering in metabolism from a lowered ⁠, however, is not supported by the values measured in control fish. Ventilation volume of control fish in hypoxia (60 or 40 mmHg) should be much higher than in normoxia, yet at the lowest swimming hpeed (5·5 cm s⁻¹) was unchanged.

The differences in blood respiratory properties between the normoxic and hypoxic groups obtained from in vitro analysis in our study confirm those from earlier studies (Wood & Johansen, 1972; Tetens & Lykkeboe, 1981).

The overall consequence of the blood changes for blood O₂ transport must be that gill blood O₂ loading is favoured in the hypoxia-acclimated fish as a result of the O₂ affinity increase caused by the reduced red cell organic phosphate concentration. Why the O₂ -Hb affinity increase measured in this study is less than predictable from the reduction in red cell ATP concentration is not clear. In this respect, it may be that sampling via caudal vessels may affect O₂ binding properties of the red blood cells as a result of swelling (Nikinmaa, 1981).

It is possible that exercise as such causes rapid alterations in the blood respiratory function (increase in blood oxygen affinity etc.) (Nikinmaa, 1983; Nikinmaa, Cech & McEnroe, 1984) which result in maximal and similar O₂ loading in both normoxic and hypoxia-acclimated fish. Overall, however, it seems clear that the swimming performance gauged by the experimental conditions does not in any crucial way depend on altered blood respiratory properties brought about by hypoxia acclimation. While a slight shift in the oxygen dissociation curve will increase maximal O₂ delivery to the tissues, it may be that the back-up internal O₂ transport capacity depends more heavily on an increase in cardiac output and tissue perfusion.

This study was conducted while one of the authors (PGB) was supported by a Marshall Fund Fellowship from the Denmark-America Foundation. The authors would also like to thank Ms S. Kornerup and Mr E. Larsen for their technical assistance.