Maintenance of Oxygen Consumption in Resting Silurus Glanis at Different Levels of Ambient Oxygenation

In most water-breathers, oxygen consumption can be maintained constant in spite of large changes in ambient oxygenation. In the crayfish Astacus leptodactylus we reported the mechanisms which permit this maintenance at basal metabolism (Massabuau & Burtin, 1984). We showed that when this crustacean is exposed to different oxygenation levels, ventilation is adjusted so that the O₂ partial pressure in the arterial blood is maintained in a low and narrow range: it increased from l±0·2 to 3·5 ±0·4 kPa when the inspired increased from 3 to 33 kPa. In absolute terms this change appears rather small compared to the increase. Concurrently, in the expired water remained constant around 1 kPa. Other adaptations are a Bohr effect appearing below 10 kPa (Sakakibara et al. 1987) and an increase of blood flow rate below 5 kPa (Massabuau & Burtin, 1984). The relative constancy of in the fluids leaving the gas exchanger is consistent with the existence of O₂ chemoreception in the branchial cavities (Massabuau & Burtin, 1984).

The aim of the present work was to learn whether this strategy for maintaining resting constant - largely based on ventilatory control of - is restricted to A. leptodactylus or part of a more general pattern in water-breathers. As noted by Shelton et al. (1986), there are data suggesting that in fishes with high O₂-affinity respiratory pigments, ‘may be little affected’ by changes (see for example Eddy, 1974, fig. 2 and table 1; Itazawa & Takeda, 1978, table 1). There has been no comprehensive demonstration of this. We present data showing that this strategy exists in the wels (or sheat-fish), Silurus glanis. It is a nocturnal fish living in lakes and slow-flowing streams (Muus & Dahlstrom, 1978). During the daytime wels lie on the bottom in hollows or under stones and, like crayfish, rarely move, so that measurements can be made on animals that spontaneously remain at basal metabolism.

Experiments were performed on 19 male and female wels, Silurus glanis, reared in captivity and acclimated in our laboratory for at least 2 months. Animals were fed with frozen fish and beef heart. During maintenance and experimental periods, the animals were supplied with water from the Strasbourg water table (see Table 1 for water ionic composition; T = 13 °C; partial pressure of carbon dioxide, ⁠;pH ≈= 8·30™8·40; during the maintenance period, variable during experiments; O₂ capacity coefficient in the water, During experiments, acid-base balance in the water was controlled with a pH-CO₂-stat (Dejours et al. 1978). During experiments fishes were unfed. They were maintained under a natural rhythm of light conditions (dim light during the daytime) and could not see the experimenter. Five types of experiments were performed. All values are presented as mean ±1 standard error (S.E.). P< 0 ·05 was taken as the fiducial limit of significance in paired t-tests.

These determinations were performed in winter on six animals weighing 685 ± 89 g. Blood was sampled by puncturing the caudal aorta or vein of anaesthetized fish (urethane 8 g 1⁻¹). The heparinized blood was stored in a rotating system immersed in melting ice. Gas equilibration was performed at 13°C in a bowl-shaped tonometer (Radiometer type) gently shaken for 30 min. Gas mixtures, N₂/O₂/CO₂, were obtained by using gas-mixing pumps (Wösthoff, Bochum). The O₂ concentration of equilibrated blood was measured with a modified Tucker chamber (Tucker, 1967) on 10 μl samples and pH was determined with a Radiometer 6299A capillary electrode at 13 °C.

These experiments were performed in February and March on five Silurus weighing 127 ± 20 g. was measured in an open-flow respirometer, volume 1300ml, using the technique described by Massabuau et al. (1984). These measurements, together with the defined and the measured (see below), permitted calculation of ventilation, using the Fick principle (Saunders, 1962). Because the existence and importance of possible cutaneous oxygen uptake was not taken into account, the actual value of may have been somewhat overestimated. No allometric correction of was made as in the studied range there was no significant difference with the 1kg standard-mass correction. Each animal was placed in the respirometer at least 24 h before measurements began. It was then exposed for periods of 90 min to 24 h to five levels of The order of presentation was 20, 40, 10, 5 and 2 kPa. Because results were independent of the exposure period, all data at each were computed together.

Seven animals weighing 809 ± 43 g were used for this experiment performed in February and March. To sample the expired water, a catheter was fixed on the upper part of the operculum, above the pectoral fin, where the water flows out after having ventilated the apex of the gill arches (Fig. 1). A hole (≈1·5 mm in diameter) was drilled through the cleithrum, 1–2 mm anterior to the thin sheet of tissue that comes into contact with the body and prevents water reflux. A polyethylene catheter (i.d. 0·38mm, o.d. 1·09mm, length 50–55 cm), with the inner end shaped into a collar of 5 mm diameter, was slipped into the hole from the internal face of the operculum. A second catheter (length ≈ 5 mm), with the outer end shaped into a 5–7 mm diameter collar, was slipped over the first from outside the operculum. They were tied together with a thin stainless-steel wire. The inside of the assembly projected less than 0·5-1 mm into the branchial chamber. The surgery took about 5 min and was performed on anaesthetized animals. Fish were |then acclimated for 2–3 days in the experimental tank (43 cm × 35 cm ×16 cm, renewal rate 10 1h⁻¹) before measurements began. The catheter passed freely through the roof of the experimental box (in a 3 mm hole), and a small piece of Teflon prevented it falling into the tank. was continuously recorded during the daytime using a Radiometer polarographic electrode placed in series between the sampling catheter and a Gilson peristaltic pump (flow rate 0·1 ml min⁻¹). At night, when the animal was active, the catheter was disconnected. The injection port of the electrode was equipped with a T-tube operated by remote control so that it could be calibrated before and after every set of measurements without disturbing the fish. The order of presentation was as above and each plateau lasted about 24h. Consequently an experiment with a single animal took about 1 week.

This was performed in June and July on six Silurus weighing 775 ± 43 g. Animals were kept in the same apparatus as above and exposed to the same protocol of plateaus. Arterial blood was sampled following the technique described for crayfish by Massabuau & Burtin (1984). Its advantage is that it is a push-pull system which requires only a single catheter rather than the complete extractor poreal loop. A catheter was implanted in the caudal aorta 3–5 cm anterior to the caudal fin. It consisted of two parts: a silicone tube (i.d. 0·30mm, o.d. 0·64 mm, length 2cm) which was inserted in the aorta and a polyethylene catheter (i.d 0·38 mm, o. d. 1·09 mm, length 50–55 cm). If used every day it remained patent for 3–4 weeks before spontaneously falling out. After a 7-to 10-day recovery period was measured once a day between 10.00 and 11.00 h. In brief, the system consisted of the arterial catheter, a thermostatted electrode, a 2-m polyethylene tube acting as a blood reservoir and a Gilson peristaltic pump (blood flow rate 0·07 ml min⁻¹) placed in series. was read exactly 6 min after the beginning of the sampling period. Before reinjection into the fish, 100μl of blood was anaerobically sampled in capillary tubes for analysis of acid-base balance. This sample was immediately used to determine pHa (with a Radiometer 6299A capillary electrode thermostatted at 13°C) and the total CO₂ concentration (with a modified Cameron chamber; Cameron, 1971). From these values, arterial blood CO₂ partial pressure, and bicarbonate concentration, [HCO₃⁻]_a, were calculated using a CO₂ solubility of 0·396 mmol 1⁻¹ kPa⁻¹, and (J.-L. Rodeau & B. Burtin, unpublished data; throughout the text we use [HCO₃⁻]_a for [HCO₃⁻]_a + 2[CO₃²⁻]_a.

Five Silurus weighing 742 ± 49 g were examined in February and March. Surgery and experimental procedures were the same as those described above except that (i) mixed venous blood was sampled from the ventral aorta and (ii) was measured instead of because of the expected range and the shape of the O₂-binding curve (see Fig. 2). The animals were exposed to the same plateaus as above, and 100 μl of venous blood was sampled once a day between 10.00 and 12.00 h. The O₂ concentration was immediately measured using a modification of Tucker’s method (Tucker, 1967).

Blood flow rate was estimated by the Fick principle, using these values, values obtained by graphical extrapolation on the O₂-binding curve recalculated for a haematocrit (Hct) of 14% (see Results) and the measurements. was estimated by graphical extrapolation of the same recalculated O₂-binding curve.

Mean O₂-binding curves for whole blood in Silurus are presented in Fig. 2 at two and pH values. The curves are hyperbolic and P₅₀ was 0·64 ±0·03 kPa at pH = 7·96 ± 0·02 and it was 0·85 ± 0·04 kPa at pH = 7·69 ± 0·01 and There was a Bohr effect, ΔlogP₅₀/ΔpH = −0·46 ±0·06, but no visible Root effect at the studied pH. The haematocrit of the blood used for these determinations was 25·0 ±1·6%. In chronically cannulated fishes it was always lower, and decreased to 14·0 ±0·8% within 1 week. To take this into account - with the assumption that haemoglobin characteristics did not change - we recalculated O₂-binding curves for Het = 14% (dotted lines in Fig. 2). following Roughton (1964), they are geometrically similar to the curves at Hct = 25 % (scale 14:25) and have the same P₅₀-Sixteen red cell counts in four animals at different times after manipulation showed that Het was linearly related to cell count: number of red blood cell = 67 500Hct + 38000.

Table 2 shows all the respiratory variables measured and calculated in Silurus exposed to selected and fixed levels. Between 40 and 3 kPa, was maintained constant, whereas was greater the lower the We believe that 2 kPa is about the lower limit of the regulation, as in one animal in which was measured at a lower value it decreased linearly below this value. Fig. 3 is a typical example of the changes recorded during the daytime as fishes were kept at fixed values. was typically low but interspersed with transient peaks. The frequency of peaks - which, based on visual observations, corresponded to periodic ‘sighs’ - was independent of but their amplitude was higher when increased. At was constant and at 38 kPa it could remain steady at about 2 kPa for more than 1 h. Resting values were never as low as zero at any ⁠.Values of were sampled every 6 min in all animals. Depending on technical problems the recording period covered between 6 and 8h, i.e. 60–80 values per animal. The frequency distribution of values is shown in Fig. 4. The modal value of was also determined for each animal at every value. The mean of these is presented in Fig. 5A, together with the results of the measurements and estimates. At values of between about 3 and 40 kPa, the modal value of increased from 0·8 ±0·2 to 3·9±1·4kPa (P<0·05; paired t test) whereas and values did not change. The haematocrit remained constant and was independent of The corresponding changes in acid-base balance in the arterial blood are shown in Fig. 5B. Values of pHa were generally constant at varying levels of [HCO₃⁻]_a and - As a consequence of the maintenance of and pHa (assuming no changes in haemoglobin characteristics), remained constant at 2 mmol 1⁻¹. On the O₂-binding curve this corresponds to 85 % saturation. As was also constant, the arteriovenous O₂ concentration difference and the O₂ capacity coefficient in the blood did not vary.

The present study reports steady-state respiratory adaptations in the teleost S. glanis after 1-day acclimation periods at various levels of inspired ⁠. Although many previous studies of water-breathers have been devoted to this subject, homeostatic mechanisms have received little attention (Dejours, 1988). Our aim was to learn whether the principles of breathing control we found in crayfish (see Introduction) could be extended to teleosts. We did not intend to study the acute phases of adaptation but rather the results of the adaptation. In humans and birds (Bouverot, 1985), as in crayfish (Massabuau & Burtin, 1984), it is generally agreed that the early respiratory changes result from the O₂ stimulation of peripheral chemoreceptors. In teleosts there are strong arguments in favour of the existence of such peripheral O₂ chemoreceptors located in, or close to, the branchial cavity (Eclancher, 1972, 1975; Eclancher & Dejours, 1975; Bamford, 1974; Milsom & Brill, 1986).

Our measurements of blood characteristics are comparable to those of Albers et al. (1981) in Silurus glanis and those of Haws & Goodnight (1962) in the related freshwater species Ictalurus nebulosus and Ictalurus punctatus. We found similar hyperbolic O₂-binding curve and P₅₀ values. The oxygen capacity we report is comparable to those of I. punctatus and I. nebulosus, but ΔlogP₅₀/ΔpH is lower than that given by Albers et al. (1981). Our haematocrit values (25·0 ± 1·6 %) and red blood cell count measured on the sample taken in the anaesthesized animals are comparable to the values reported by Albers et al. (1981) in the same experimental conditions. In the resting state they differ little from the 16 ± 2 % reported in chronically cannulated dogfish by Baumgarten-Schumann & Piiper (1968). It is likely that these differences were related to the stress of surgery and anaesthesia, as we observed negligible blood loss. Following severe exercise Educed by chasing, fish can exhibit a Hct increase of 40%, due mainly to contraction of the spleen and a shift of water out of the plasma (Yamamoto et al. 1980). The low values of we report are in the range for other quiescent water breathers, such as resting eels (Steen & Kruysse, 1964) and crayfish (Massabuau & Burtin, 1984), whereas much higher values have been observed in excited animals (Steen & Kruysse, 1964; Baumgarten-Schumann & Piiper, 1968). Notice that (i) is adjusted to a value (≈=2 kPa) very close to the minimum required to ensure intracellular O₂ supply in single-cell suspensions of rat hepatocytes (Jones & Kennedy, 1982) and (ii) give a mean ‘in vivo’ estimate of the intracellular in S. glanis at basal metabolism. In Silurus, P₅₀ = 0·6kPa, the values of and we observed give the same and values as those in ‘normoxic’ dogfish which have higher and values but a P₅₀ of 2·13 kPa (T = 17°C and Baumgarten-Schumann & Piiper, 1968). The value of reported here is similar to the value we reported in A. leptodactylus (13·8 ±0·8 μmolkg⁻¹min⁻¹; Massabuau & Burtin, 1984) kept in identical water conditions. It is also similar to values obtained in eel (16°80 ± 0·79 μmol kg⁻¹ min⁻¹ at 11·5°C by Kirsch & Nonnotte, 1977), tench (20·2 ± 1·16umol kg⁻¹ min⁻¹ at 13°C by Nonnotte, 1981) and dogfish (20·7 ± l·8μmol kg⁻¹min⁻¹ at 15°C) by Butler & Taylor, 1975; 28·4 ± 7·6μmolkg⁻¹min⁻¹ at 15–17°C by BaumgartenSchumann & Piiper, 1968). All experiments were performed in winter, except for the arterial acid-base balance and P_O2 measurements which were performed in summer. This raised the problem of comparing respiratory parameters measured at different times of the year and at potentially different metabolic levels. In resting carp there is no significant variation in between 24·5°C (3·3 ± 1 kPa in Itazawa & Takeda, 1978) and 10°C (3·8 ± 2T kPa or 1 ± 0·6kPa in Garey, 1967). Consequently it is unlikely that a potential increase of resting metabolism in summer interfered with our measurements.

The key point in the respiratory adaptation of Silurus is that when varies between 40 and 3 kPa, steady-state after 1 day of acclimation appears to be maintained exclusively by ventilatory adjustment with no change of blood flow rate or pH (no Bohr effect). This corresponds to an adaptation based on a principle of economy, because even though it increases 16-fold between and - remains close to its minimum possible value at each value. Indeed, the O₂ extraction coefficient is always around 80–90%. As a result of this adaptation, the value of remains constant at about 2 kPa. It is likely that must be the controlled variable, by analogy with what is known from higher vertebrates (Bouverot, 1985). The capacity to function at such low values must be related to the very high haemoglobin O₂-affinity in Silurus (Fig. 2). The effect of the high O₂-affinity on extraction and ventilation in fishes has recently been discussed by Malte & Weber (1987). During inactive periods, Silurus rests in an environment that can be hypoxic, and the problem of O₂ uptake from the medium is obviously a priority. The details of this mechanism should be different in fishes with lower blood O₂-affinity, which presumably facilitates O₂ release at the cellular level (Krogh & Leitch, 1919), both in more active fishes (like trout) and in nonactive fishes (like dogfish) living in nonhypoxic environments where there is no problem of O₂ uptake. However, the principle of an oxygenation status that is independent of over a wide range must remain valid in steady states, either at rest or at a given level of activity. This latter point is illustrated by data from Garey & Rahn (1970), who measured in gas pockets of Salmo gairdneri swimming freely in a fishery (Fig. 6A). The fishery was supplied by a river with a high photosynthetic rate. In these conditions, although varied between 30 and 6 kPa and temperature between 8 and 17 °C, in the gas pockets (which is a closed estimate of in the surrounding tissues and the venous blood draining them, Rahn, 1957; Piiper, 1965) was independent of Trout can live perfectly well in poorly oxygenated waters. In eastern France we found a population of Salmo trutta fario living in the spring of a river in which the yearround is about 6·7 kPa at 10·0 ± 0·2°C (Massabuau & Fritz, 1984). Ott et al. (1980) reported that Salmo gairdneri can maintain its resting constant down to 2·3 kPa, independently of the temperature between 10 and 20°C. Fig. 6B shows data redrawn from Lomholt & Johansen (1979) which corroborate our results on These authors measured oxygen extraction coefficients in carp exposed to hypoxia. We recalculated the original values from their results. It is clear that they remain in a narrow range, although the mean tends to increase slightly with This is because of the use of an arithmetic mean for a probably non-normal distribution (see Fig. 4 in present paper and fig. 6 in Massabuau & Burtin, 1984). These results are consistent with the constancy of in carp gas pockets (Fig. 6A; Garey & Rahn, 1970) and data from Garey (1967), Eddy (1974) and Itazawa & Takeda (1978), who showed that changes of between 3–3·5 and 20 kPa did not alter the lowest measured values in carp and tench.

In the teleost gill, the countercurrent model is generally accepted to describe the functioning of the gas exchanger (Hughes, 1984). In a system of this type, complete equilibration between inspired water and arterial blood and between expired water and venous blood is theoretically possible. In Silurus at rest, at least at the highest our present results show that equilibration between and is far from complete, whereas is close to (Fig. 5A). Although the latter suggests that diffusion limitation must be very low, the former shows that gas exchange is ventilation-limited (Piiper & Scheid, 1984). The functional basis of this limitation can be attributed to a mismatch between ventilatory and perfusive conductance at all studied values (Table 2). This is in agreement with the general strategy of Silurus in hypoxia, which is based exclusively on the reduction of this ventilatory limitation. Some lamellae are likely not to be ventilated at rest but only perfused. This would lead to the equivalent of a ‘mismatch blood shunt’ (Piiper & Scheid, 1984). True shunt bypassing of the gills has not been described in teleosts (Dunel & Laurent, 1980). The magnitude of the shunt, Sb, can be estimated from the ratio which is the amplitude of the nonequilibration divided by the difference between inspired water and venorus blood. In normoxia Sb was 0·9 (Table 2). The decrease of Sb with hypoxia may correspond to an increase in the number of ventilated lamellae. These changes in ventilation-perfusion inhomogeneities affect gas exchange so that can be either lower or higher than or equal to it (see Piiper & Scheid, 1984, for a theoretical analysis). However, this type of observation, based on small changes, must be considered with caution as there are several uncertainties in our measurements (this is also valid for all the calculated variables in Table 2 where has been used). First, although we can be confident in our blood sampling from carefully chosen vessels, this is not true of our sampling of expired water. A perfectly defined channel exists only in a few species. Some problems of expired water mixing or contamination by backward gas diffusion may exist in Silurus, despite the anatomical arrangement of the branchial cavity (Fig. 1). Second, given the variability of (Figs 3 and 4), we chose the modal value as representing the actual value. Although the modal value is satisfactory in a study of a controlled system, it introduces a bias in the analysis of the overall gas exchanges. Indeed, all the water passing over the gills participates in the gas exchanges. The modal value clearly underestimates (at least at the highest ⁠) the ideal measurement that would be performed on all the collected and mixed expired water.

When varied, changed as a consequence of the ventilatory adaptation. In Silurus this led either to a hypocapnic alkalosis or a hypercapnic acidosis, which were fully compensated within 1 day. This was achieved by metabolic means but also possibly - in the hyperoxic direction - by transient ventilatory adjustments that are likely to occur in dogfish exposed to hyperoxia (Heisler et al. 1988). Burtin et al. (1986) demonstrated that can participate in regulation of acid-base balance in water-breathers.

In the teleost S. glanis, as in the crayfish A. leptodactylus, plays a key role in maintaining resting constant while varies. The main result of the adaptation is that remains constant in Silurus and in a narrow range in Astacus. Consequently, the homeostasis of the milieu intérieur, in terms of O₂, is fulfilled. But our data further show that the countercurrent arrangement of the fish gill is more efficient in achieving this result than is the crosscurrent design (Massabuau, 1983) of the crayfish gill. This has already been proposed on theoretical grounds by Piiper & Scheid (1984). Indeed, when decreased from 40 to decreased slightly in crayfish (see Introduction), as to be expected in a crosscurrent system maintaining constant Also, a Bohr effect appeared atvalues below 10 kPa and was increased at 3·3 kPa (Massabuau & Burtin, 1984; Sakakibara et al. 1987). In S. glanis in the same range, stays constant. There is no Bohr effect and no increase. O₂ supply is maintained simply by ventilatory adjustments.

The authors wish to thank Dr D. C. Jackson for help in preparing the English manuscript and E. Pionnier for the kind supply of Silurus. The experiments were partially financed by funds of the programme PIREN-Eau/Alsace (CNRS) and Dr B. Burtin was supported by the Ministère de la Recherche et de la Technologie.