Oxygen and Carbon Dioxide Transport Characteristics of the Blood of the Nile Monitor Lizard (Varanus Niloticus)

Although various studies have been made upon the aerobic performance of varanid lizards (Mitchell, Gleeson & Bennett, 1981; Gleeson, 1981; Gleeson, Mitchell & Bennett, 1980; Bennett, 1972) and upon their cardiopulmonary function (Millard & Johansen, 1974; Burggren & Johansen, 1982; Heisler, Neumann & Maloiy, 1983; Johansen & Burggren, 1984), the information available on the gas transport properties of varanid blood is limited to a few studies closely focused on certain aspects. Data have been provided on: the blood oxygen affinity at the preferred body temperature of the savannah monitor lizard (Varanus exanthematicus) (Wood, Johansen & Gatz, 1977) ; the in vitro oxygen dissociation curve and the Bohr effect at 25°C; and the in vivo O₂ dissociation curve at 30°C during voluntary diving of Varanus niloticus (Wood & Johansen, 1974). The present study was intended to characterize the in vitro O₂ dissociation curves and carbon dioxide transport properties of Varanus blood in more detail and at two different temperatures, providing the basic information required for an extensive quantitative analysis of cardiopulmonary gas transport in this species (J. W. Hicks, A. Ishimatsu & N. Heisler, in preparation).

Nile monitor lizards, Varanus niloticus, were purchased from an animal dealer in the United States and airfreighted to West Germany. They were kept in a large (5×7 m) room with access to a small diving tank. Room temperature was maintained at 30°C. Infrared heat radiators were provided throughout the room to allow behavioural body temperature regulation by the animals towards their preferred body temperature of 35 °C. They were regularly fed on chopped beef liver and kidney, and chicken meat, and supplied with live laboratory mice at intervals of about 1 week. Food was withheld for 3 days prior to surgery.

Nine specimens of Varanus niloticus (mass 2–6 kg) were anaesthetized with halothane and nitrous oxide. Anaesthesia and oxygen supply were maintained during surgery by artificial ventilation (rate: 12–20 min⁻¹) with a humidified gas mixture of 67% N₂O, 30% O₂ and 3% CO₂via a soft rubber tube which was sized to the trachea and then introduced into it. The gas mixture was passed through a halothane vaporizer (Drager, Lübeck, FRG) set to l·5vol%. After an initial period of about 10min the inspired halothane level was reduced to 0·5–1·0%. Five non-occlusive cannulae (PE 50) were implanted: into the right and left aortic arches, right and left atria and pulmonary artery. These sites were required for subsequent studies of pulmonary and cardiovascular function (J. W. Hicks, A. Ishimatsu & N. Heisler, in preparation; A. Ishimatsu, J. W. Hicks & N. Heisler, in preparation). Following surgery each cannula was filled to its final volume with a heparinized (200 i.u. ml⁻¹ sodium heparinate) solution of PVP (polyvinylpyrrolidone; 0·5 g ml⁻¹ in distilled water) (Brown & Breckenridge, 1975). The PVP gel prevents blood from entering the cannula and clotting in it, thus obviating the need for daily flushes with heparinized saline. After surgery and before the experiments the animals were allowed to recover for 72h in a thermostatted (±1°C) chamber with access to drinking water. Each animal was kept at the experimental temperature (25 or 35 °C) for 12 h before samples were taken for determination of O₂ and CO₂ dissociation curves.

Prior to each experiment an arterial blood sample from the left aortic arch was analysed for blood gases, pH, haematocrit (Hct), haemoglobin concentration ([Hb]) and O₂ capacity. The blood required for curve analyses was withdrawn from the same site, and the blood loss during the experiments was kept as low as possible by reinfusion of unused blood. Whole blood O₂ dissociation curves were determined in vitro by one of two methods: either by the mixing technique described by Haab, Piiper & Rahn (1960) and detailed by Scheid & Meyer (1978), or by determination of the oxygen content of whole blood equilibrated at a certain with the Lex-O₂-Con apparatus (Lexington Instruments, Waltham, MA, USA).

Two 1-ml samples of blood were equilibrated for 10 min with humidified gases at or at the same in intermittently rotating cuvettes. The gas mixtures were provided by precision gas mixing pumps (Wösthoff, Bochum, FRG), and was adjusted to values of 21 or 35 mmHg at 25 °C and 35 or 49 mmHg at 35 °C. The lower at each temperature approximates normal in vivo values (Table 1). Predetermined fractions of blood equilibrated with or 220 mmHg were mixed in gas-tight tuberculin syringes and the of the mixture (P_mix) was measured in duplicate in a thermostatted electrode system (±0·05 °C) as described by Bridges (1983). The pH of each mixed blood sample was determined using a Radiometer BMS3 and PHM 64 pH meter. Haematocrit was measured by microcentrifugation, and [Hb] determined by the cyanmethaemoglobin method. The blood oxygen saturation was calculated from the equation of Scheid & Meyer (1978):

where V₁ is the fractional volume of oxygenated blood, P_eq is the equilibrating (in this series 220mmHg), is the oxygen capacity of blood, is the solubility of O₂ in plasma and P_mix is the of the final mixture.

The solubility of O₂ in plasma was determined at both temperatures by equilibration with 50 % O₂ for 10 min and duplicate determination of the O₂ content. As a validity check of our mixing technique, the saturation of each mixture was compared to values determined directly from O₂ content measurements. Saturation values determined using equation 1 were not significantly different from those determined directly.

Determination of P_mix may potentially be biased by two factors. First, the metabolic rate of reptilian nucleated red blood cells is several times higher than that of mammalian cells (Bridges, 1983). Second, nucleated red blood cells may exhibit much higher gas/blood factors than reported for human blood (Bridges, 1983; N. Heisler, personal observation). Both of these factors would result in P_mix values lower than the actual at the time of mixing, and in a comparatively left-shifted oxygen dissociation curve.

To account for these sources of error, additional oxygen dissociation curves were generated at 25 and 35 °C by equilibrating separate 1 -ml heparinized blood samples of a blood pool each for 10 min at a number of different values (29 –101 mmHg) at the same as described above. Following tonometry, the blood was analysed for ⁠, ⁠, pH, [Hb], Het and O₂ content. The saturation of individual samples was determined from O₂ content measurements, taking differences in [Hb] between samples into account. The gas/blood factor determined for each sample was used for correction of raw blood data.

Oxygen dissociation curves were transformed according to the Hill equation (between 10 and 90 % O₂ saturation), and the cooperativity constant (n) and P₅₀ were determined by least-suares linear regression analysis (r>0·95). Significant effects of temperature and on the cooperativity constant n were determined using paired t-tests. The CO₂— Bohr coefficients were determined from Hill plots for each animal from 10 to 90% saturation. Significance between variables at both temperatures was determined using paired or unpaired t-tests where applicable.

In vivo resting acid-base and haematological parameters of varanid blood at 25 and 35 °C are summarized in Table 1. Both Het and [Hb] increased by 30% as temperature increased. The temperature-induced changes in Hct and [Hb] were fully reversible. The was lower at 25°C and there was no change in plasma [HCO₃⁻]. Arterial plasma pH fell with rising temperature by −0·007 units °C⁻¹, which is much less than expected on the basis of the alphastat hypothesis (Reeves,1972), but close to values generally found in reptiles, in particular varanids (reviewed by Heisler, 1986b).

The solubility of oxygen in Varanus plasma was found to be 0·00276 ± 0·0001 ml O₂ 100 ml⁻¹ at 25°C (x̄±S.D., N = 4) and ·00245 ± 0·00006ml O₂ 100ml⁻¹ at 35°C (N = 5). The gas/blood factor of the analytical set-up used in combination with varanid blood of the stated haematological parameters was found to be 1·29 ± 0·02 (x̄ ± S.D., N = 129). Temperature and level of did not affect this value. All raw data were corrected accordingly.

The combined effects of temperature and on the oxygen dissociation curve of Varanus blood are shown in Fig. 1. P₅₀ was 34·3 ± 4·5 mmHg (x̄ ± S.D., N = 4) at 25°C and and rose to 46·2 ±2-4 mmHg (N = 5) at 35°C and ⁠. The cooperativity constant (n) was 2·97 ± 0·40 at 25°C, and increased significantly to 3·40±0·26 at 35 °C (P < 0·001).

The CO₂-Bohr coefficient was independent of the level of oxygenation at 25°C (Fig. 2). At 35°C the CO₂-Bohr coefficient was dependent on saturation, with the Bohr slope doubling between 10 and 90 % saturation (Fig. 2). The ΔlogP₅₀/ΔpH values were −0·300±0·190 at 25°C and −0·662 ± 0·200 at 35 °C.

The carbon dioxide dissociation curve of varanid blood was shifted to the right and downwards because of a fall in blood total CO₂ content with rising temperature (Fig. 3). The relationship between and can be characterized by the general equation (Fig. 3). This type of logarithmic equation provides the best fit of the data, but does not reflect any theoretical background of the CO₂ transport properties. The relationships between and were not significantly different for oxygenated and deoxygenated blood (see Fig. 3). At individual values a marginal significance could only be detected for 7 % CO₂ at 25°C (P< 0·05). Apparently, the blood of V. niloticus has no significant Haldane effect, at least not in the physiological range.

There were no significant differences between buffer values (β) for oxygenated and deoxygenated blood at either temperature. The overall β values were 12·0 ±1·9 at 25°C and increased to 17·5 ± 3·2 mequiv pH⁻¹1⁻¹ at 35°C (x̄±S.D., N = 8) (Fig. 4). The lower buffer value at 25 °C reflects the influence of temperature on Hct and [Hb]. This is evident when the buffer value is expressed as the specific buffer value (β/[Hb]), which is independent of temperature: 2·4 ± 0·8 at 25°C and 2·7 ± 0·4 mequivpH⁻¹ l⁻¹ (g%Hb)⁻¹ at 35°C.

The total CO₂ determined in whole blood originated mainly from the plasma fraction. Only 7–8 % was contained within red blood cells (Fig. 5). These values are close to those expected on the basis of mammalian data, taking the lower Hct into account. There was a tendency for an increase in the erythrocyte fraction with rising ⁠, as well as for higher values in deoxygenated blood. However, statistically significant differences were not found. Also, temperature had no effect on this parameter.

The blood oxygen dissociation curve of the Nile monitor lizard, Varanus niloticus, is highly sigmoidal, has a low affinity for oxygen, and is relatively insensitive to temperature. P₅₀ values of 35·3 (25°C) and 42-4mmHg (35°C) found at in vivo pH values of 7·59 and 7·52 agree with earlier investigations. Correcting the P₅o at 25°C (present study) to a pH of 7·45 results in a value of 40·0 mmHg, which is close to the P₅₀ of 42·4 mmHg reported by Wood & Johansen (1974) for this temperature and pH. Also the Hill coefficient of 31 reported by Wood & Johansen (1974) is compatible with our value of 2·97 at 25 °C.

The cooperativity constant of Varanus is modified by both temperature and CO₂. The increase in n as temperature increases produces a saturation-dependent temperature coefficient ⁠. The increase in with falling saturation levels is similar to the saturation dependence of the temperature coefficient in pig (Willford & Hill, 1986). The CO₂-Bohr factor, at least at 35°C, increases with saturation. This results from the decrease of n with rising (3·4 at to 34 at ⁠). Similar CO₂-induced pH effects on n and resulting saturation-dependence of the CO₂-Bohr coefficients have been reported in blood of sea turtles, Chelonia mydas and Caretta caretta (Lapennas & Lutz, 1982) and freshwater turtles, Pseudemys scripta and Chrysemys picta (Maginniss, Song & Reeves, 1980; Maginniss, Tapper & Miller, 1983). The effects of changing temperature and pH on the shape of reptilian oxygen dissociation curves are in contrast to the relationships in man where CO₂-Bohr slopes and temperature coefficients are independent of saturation (Hlastala, Woodson & Wranne, 1977; Reeves, 1980). The effects of temperature and CO₂ on haemoglobin cooperativity may reflect the multiple haemoglobin systems found in reptiles (Maginniss et al. 1980).

The plasma [HCO₃⁻] at in vino values (35 and 38 mmol l⁻¹ at 25 and 35 °C, respectively) are higher than those reported for most lizards, snakes and crocodilians (10– 25 mmol l⁻¹, e.g. Howell & Rahn, 1976; Weber & White, 1986; Grigg & Cairncross, 1980; Glass & Heisler, 1986; for references see Heisler, 1986b) and more comparable to those for diving turtles and other diving lizards (e.g. Wood & Moberly, 1970; Howell & Rahn, 1976; Jackson & Heisler, 1982, 1983). The in vitro non-bicarbonate buffer value of Varanus niloticus blood is similar to that in other diving reptiles (see Wood & Johansen, 1974), and is little influenced by the blood oxygenation state. Because of the relatively high bicarbonate concentration, the total blood buffer value is relatively high and probably provides an additional reserve for periods of anaerobic diving.

The apparent lack of a significant Haldane effect cannot immediately be understood. Since Bohr and Haldane effects are both considered to be due to negative heterotropic allosteric ligand interaction, any haemoglobin exhibiting a Bohr effect would also be expected to show a Haldane effect. As pointed out by Dill, Edwards & Florkin (1932), however, the differences between of oxygenated and deoxygenated blood would be minimized at low [Hb] values. The lower [Hb] and Hct levels may, therefore, be responsible for the failure to detect a small Haldane effect (see Fig. 5) during the present experiments. Wood & Johansen (1974) have reported a Haldane effect for V. niloticus blood. The average [Hb], however, was significantly lower in our animals (3·6–4·6g 100 ml⁻¹) than in theirs (74 g 100ml⁻¹). The reason for a lower [Hb] and Hct in our study is unclear: the animals were apparently healthy, free of parasite infestation, and surgery resulted in little if any blood loss. Also, [Hb] and Hct remained stable throughout the experiments. Haldane factors in reptiles are reported as being quite variable. As recently reported (Weinstein, Ackerman & White, 1986), oxygenation of Pseudemys blood decreases by 1 mmol l⁻¹ at in vivo levels at both 25 and 35°C ([Hb] = 5·8 g 100 ml⁻¹). This is similar to the mean level of change calculated for varanid blood, and Pseudemys blood also exhibited a temperature-dependent change in at any given ⁠, similar to that for Varanus. In sharp contrast, Grigg & Cairncross (1980) found a fairly large Haldane effect in the Australian crocodile, Crocodylus porosus, with a of 5 mmol l⁻¹ at in vivo of 40mmHg ([Hb] = 8·7g 100ml⁻¹), but no temperature-related shift of the CO₂ dissociation curve, which, as stated, may have been due to their small number of measurements. At present, a final decision as to whether the lack of a significant Haldane effect reflects a unique physicochemical property of varanid haemoglobin, or whether the low [Hb] makes it technically difficult to measure a significant effect, requires further investigation.

Body temperature has a pronounced effect on Hct and [Hb] in Varanus niloticus. In fact, the lower buffer value at 25°C reflects the decrease in [Hb], as shown by the constancy of the specific buffer value (β/[Hb]) at both temperatures. The precise mechanism for the temperature dependence of Hct and [Hb] cannot be deduced from the present study. It is possible that, at lower temperatures with associated lower cardiac output, more red blood cells physically settle within the vascular system. As temperature and cardiac output increase more red cells are physically mixed with the plasma. Kooyman et al. (1980) have observed large fluctuations in the arterial [Hb] in Weddell seals during periods of rest and immediately following a dive, with [Hb] being 30% higher following a dive. They suggest that this increase results from the physical effects of increased cardiac output on erythrocytes which have settled in venous sinuses. Periods of activity may also increase Hct and [Hb] in V. niloticus. An increase in Hct of between 20 and 40 % has been observed at both 25 and 35°C during periods of submaximal activity (J. W. Hicks, A. Ishimatsu & N. Heisler, personal observation). An alternative mechanism would be active mobilization from the reticuloendothelial system during periods of exercise, as occurs in higher vertebrates. Due to the significant effect of [Hb] on both the O₂ and CO₂ transport characteristics of blood, this phenomenon requires further investigation.

Supported by Deutsche Forschungsgemeinschaft.