The functioning of the haemocyanin of the terrestrial christmas island red crab gecarcoidea natalis and roles for organic modulators

The major respiratory pigment in the haemolymph of decapod crustaceans is the copper-based protein haemocyanin. Haemocyanin markedly increases the O₂-carrying capacity of the haemolymph and is now appreciated to be under the control of a complex suite of effectors and organic modulators that regulate O₂-binding by haemocyanin (for reviews, see Mangum, 1983; Morris, 1990, 1991; Burnett, 1992; Truchot, 1992; Morris and Bridges, 1994). Thus, haemocyanin functioning is apparently controlled in a manner analogous to vertebrate haemoglobin O₂-binding. A significant role for haemocyanin in O₂ transport in quiescent crabs has been questioned, and the involvement of haemocyanin in O₂ transport varies among species. The importance of O₂ delivery by haemocyanin increases in response to increased O2 demand and during environmental hypoxia (e.g. McMahon et al. 1979; Greenaway et al. 1988; Lallier and Truchot, 1989; Morris et al. 1996b).

Recently, a dichotomy in the functioning of haemocyanin in terrestrial and aquatic decapod crustaceans has become apparent: terrestrial decapods appear to have haemocyanin with little or no sensitivity to most modulators. In some cases, this appears to be due to specific blocking of binding sites by dialysable compounds (e.g. Morris et al. 1988) and in others to an absence of binding (Morris and Bridges, 1994). The haemocyanins of terrestrial crustaceans tend to become readily saturated with O₂ at the gas-exchange surfaces, despite a haemolymph acidosis, partly as a result of the efficiency of the gas-exchange organs (e.g. Wood and Randall, 1981; Greenaway et al. 1988; Adamczewska and Morris, 1994a, 1998). In addition, the lower viscosity of air, compared with water, coupled to the relatively greater abundance of O₂ over CO₂ in air compared with water allows, a relative hypoventilation in air-breathers and reduces the costs of O₂ extraction from air. Consequently, the need for modulators to increase haemocyanin O₂-affinity in terrestrial crabs appears to be largely redundant (Morris, 1990). Instead, terrestrial crustaceans appear to rely on altering ventilation and circulation rates to maintain O₂ delivery by haemocyanin (Morris, 1991; Morris and Bridges, 1994). Haemocyanin is also the main protein in the haemolymph of crustaceans (Chen and Cheng, 1993), and higher concentrations of haemocyanin in terrestrial compared with aquatic species are thought to provide increased buffering for the higher and consequent CO₂ content in air-breathers (for reviews, see Mangum, 1983; Morris and Bridges, 1994).

The ability to increase the rate and capacity of O₂ delivery is essential if the increased metabolic rate is to be supported aerobically and, thereby, to increase the scope for activity. While increased activity can also be supported by anaerobic metabolism, anaerobiosis is very inefficient compared with aerobiosis and results in acidosis and an O₂ debt that must be repaid later (Herreid and Full, 1988).

The affinity of haemocyanin for O₂ can be affected by temperature, pH and salinity (for a review, see Truchot, 1992) as well as by neurohormones (Morris and McMahon, 1989; McMahon and Morris, 1990), L-lactate, urate (Truchot, 1980; Morris et al. 1985a) and various inorganic molecules (Mangum, 1983; McMahon, 1986; Morris, 1990, 1991). Changes in the O₂-affinity of respiratory pigments in response to temperature and salinity represent perturbing effects rather than regulatory mechanisms (Jokumsen and Weber, 1982; Morris, 1991; Lallier and Truchot, 1997). In contrast, L-lactate, urate and Ca²⁺ are thought to modulate haemocyanin O₂- affinity to optimise O₂ delivery to the tissues during exercise or environmental hypoxia (for a review, see McMahon, 1985; Morris, 1990; Burnett, 1992; Truchot, 1992). For example, anaerobiosis in crustaceans generates L-lactate, and the concomitant acidosis results in decreased haemocyanin O₂- affinity. In many aquatic crustaceans, L-lactate can partially offset the Bohr shift, thereby assisting O₂ uptake at the gas- exchange surfaces by increasing the O₂-affinity of haemocyanin (Booth et al. 1982; Morris et al. 1986a,b; McMahon, 1985; Morris, 1990). It is these organic modulator effects that seem to be specifically absent in land crabs.

The condition of greatest potential demand on haemocyanin functioning is likely to be either during moderate environmental hypoxia or during exercise at close to the maximum aerobic speed or during repeated sprinting. Thus, functionally important modulations of O₂-binding by haemocyanin in terrestrial crabs might be most important during exercise since they rarely encounter environmental hypoxia.

The Christmas Island red crab Gecarcoidea natalis provides a model to assess haemocyanin function in land crabs since, with the exception of a 3 week larval phase, it is completely terrestrial. In addition, these red crabs undertake an annual breeding migration during which they may walk distances of 1 km per day for a number of consecutive days. Importantly, this species does not exhibit exceptional aerobic capacity (ability to increase exercise intensity) compared with that of other crustaceans, and O₂ shortage at the tissue level during exercise results in anaerobiosis (Adamczewska and Morris, 1998). A characterisation of O₂ binding by the haemocyanin of G. natalis was carried out to determine the response to any effectors of haemocyanin O₂-affinity. This characterisation of the haemocyanin O₂-affinity and the sensitivity of O₂-binding to haemocyanin to modulators in vitro was used, together with in vivo data (Adamczewska and Morris, 1998), to model the functioning of the pigment in vivo in exercising G. natalis. The models of haemocyanin O₂-function could then be used to assess whether anaerobiosis during exercise occurs as a result of insufficient O₂ delivery to the tissues or whether it is due to other limitations, such as the aerobic capacity of the tissues.

Red crabs Gecarcoidea natalis (Pocock) with a body mass ranging from 114 to 232 g (190±6 g; mean ± S.E.M., N=50) were collected from Christmas Island and maintained in the laboratory as described previously (Adamczewska and Morris, 1994a). To determine the functioning of the haemocyanin of G. natalis in oxygen transport, in vitro oxygen equilibrium curves were constructed using a spectrophotometric diffusion chamber method (Sick and Gersonde, 1969; Morris et al. 1985b). Samples of venous haemolymph were obtained by sampling via the arthrodial membrane at the base of the last walking leg. In the diffusion chamber, a subsample of haemolymph (15 μl) was deoxygenated with CO₂/N₂ mixtures (Wösthoff gas-mixing pumps), and the change in absorbance was determined after stepwise addition of O₂. Changes in absorption at 335 nm were proportional to the O₂ saturation of the haemocyanin. The pH of the haemolymph was controlled by changing the proportion of CO₂ (0.1–3.6 %) to span the pH range measured in vivo. The pH of the haemolymph was measured near the P₅₀ (the partial pressure of O₂ at which the haemocyanin is 50 % saturated) with a capillary electrode (G299a, Radiometer) housed in a BMS2 (Blood Micro System) and thermostatted at the experimental temperature of 25±0.1 °C, unless otherwise specified.

Haemolymph from at least eight red crabs was pooled to produce each ‘batch’ of haemolymph used for the construction of oxygen equilibrium curves (OECs). Because of the number and duration of treatments, it was necessary to use several different batches of haemolymph. The effects of all treatments on the O₂-binding properties of a batch of haemolymph were compared with the O₂-binding properties of that batch of whole haemolymph. All OECs were constructed using fresh haemolymph refrigerated at 4 °C or held on ice. The individual crabs used for haemolymph sampling were different for each batch of haemolymph. At the end of a series of determinations, and for each batch of haemolymph, the haemocyanin O₂- affinity was checked to ensure that the haemolymph had not deteriorated.

where R is the gas constant (in kJ) and T is temperature (in K).

The concentrations of L-lactate and urate in the haemolymph were manipulated to determine their effect on the haemocyanin O₂-affinity and binding cooperativity at 25±0.1 °C. A 500 μl sample of haemolymph was enriched with uric acid by equilibration with 4 mg of uric acid (Sigma, no. U-2625) overnight. The haemolymph was then centrifuged for 10 min at 10 000 g to precipitate any excess solid urate. Haemolymph enriched with urate was then diluted with the original whole haemolymph to obtain different concentrations of uric acid. The use of uric acid resulted in slight acidification of the haemolymph samples (see Results), but sodium urate did not appear to bind to the haemocyanin since it was not possible to increase the urate concentration in the haemolymph using the sodium salt. Urate concentrations in the haemolymph were determined using a Sigma Diagnostics test kit (no. 685).

The effect of L-lactate was investigated using four separate batches of haemolymph, each taken from a different group of eight red crabs, which had been collected from Christmas Island on one of two different occasions. The L-lactate concentration in whole haemolymph withdrawn from the crabs was always less than 1 mmol l⁻¹ ; in haemolymph enriched with L-lactate, the concentration of L-lactate ranged from 12.3 to 26.5 mmol l⁻¹ (see Table 1). To determine the effect of L- lactate, haemolymph was enriched with L-lactate by the ultracentrifugation and replacement method (Bridges and Morris, 1986). Briefly, in the morning of the day of construction of the OEC, a sample of whole haemolymph (300 μl) was centrifuged (Air-Fuge, Beckman, USA) for 20 min at 160 000 g to pellet the haemocyanin. A proportion (10 %) of the plasma solution was removed and replaced with an equal volume of concentrated L-lactate (Sigma, no. L-2250 lithium lactate) solution in G. natalis Ringer, consisting of (in mmol l⁻¹ ): NaCl, 335; KCl, 8.5; MgSO₄, 1; MgCl₂, 10.4; CaCl₂, 17.5; and NaHCO₃, 1. The salt concentrations of the Ringer’s solution were based on the measured salt concentration in the pooled whole haemolymph sample (see below). The sample was remixed and used for OEC determination. The concentration of L-lactate in each sample was determined using a Boehringer Mannheim test kit (no. 138 084). Control OECs were constructed on a different pooled haemolymph sample containing 20 mmol l⁻¹ LiCl.

The combined effect of L-lactate and urate on the haemocyanin O₂-affinity was also examined. A series of OECs was constructed using haemolymph enriched with three different concentrations of uric acid and containing 21 mmol l⁻¹ L-lactate.

Haemolymph samples (400 μl) were dialysed for 20 h at 4 °C in two washings, each consisting of 2 l of G. natalis Ringer’s solution at pH 8.1. The effects of Mg²⁺ and Ca²⁺ on the haemocyanin O₂-binding properties were investigated by dialysing haemolymph in Ringer’s solutions containing three different concentrations of Mg (5.2, 10.4 and 16 mmol l⁻¹ ) and Ca (10, 17.5 and 24 mmol l⁻¹ ). The concentrations in the dialysed haemolymph were determined as total elemental concentration by the Atomic Adsorption Spectrophotometer (AAS) (GBC 906). To suppress interference, determinations of Na and K were made in the presence of 5.9 mmol l⁻¹ CsCl₂ and of Mg and Ca in the presence of 7.2 mmol l⁻¹ LaCl₃.

Separate haemolymph samples were similarly dialysed but in G. natalis Ringer’s solution containing sufficient L-lactate to raise the concentration in the sample to 21 mmol l⁻¹. Some of these samples were further manipulated to alter the urate concentration (see above) so as to produce dialysed haemolymph containing 21 mmol l⁻¹ and a range of urate concentrations. The maximum concentrations of L-lactate used were well within the physiological range for G. natalis which, during intensive exercise, may have haemolymph levels in excess of 30 mmol l⁻¹ (Adamczewska and Morris, 1994b). The upper limit for the urate concentration used was determined by the maximum carrying capacity of the haemolymph and thus the maximum effect achievable.

The specific effect of CO₂ on haemocyanin O₂-binding was determined by manipulating haemolymph pH by the addition of either 0.01 mol l⁻¹ of HCl-enriched or 0.01 mol l⁻¹ NaOH- enriched Ringer’s solution to induce a ‘fixed acid’ Bohr shift rather than a CO₂ Bohr shift. A subsample of haemolymph (150 μl) was ultracentrifuged for 20 min at 160 000 g to sediment the haemocyanin (Beckman, Air-Fuge) and 40 μl of the plasma was removed. The plasma was then mixed with different volumes (range 10–40 μl) of the acidic or basic red crab Ringer’s solution, and 40 μl of this mixture was returned to the haemolymph, and the haemocyanin was resuspended (Bridges and Morris, 1986). The OECs were constructed at constant CO₂ levels of either 0.4 % or 4 % CO₂. Analysis of covariance (ANCOVA) was used to test for heterogeneity in the variances (slopes) between the data sets to determine the dependence of logP₅₀ on pH. For data sets with homogeneous variances, the elevations of the regression lines were compared to determine whether changes in haemocyanin O₂-affinity had occurred. The significance level of P=0.05 was applied to all the resulting F values. Values are presented as means ± S.E.M.

For each OEC, the P₅₀ at the physiological pH of 7.6 was calculated from the regression equations describing the dependence of logP₅₀ on pH and used for comparisons among treatments. The P₅₀ and cooperativity of O₂-binding (n₅₀) were calculated using saturation values between 25 % and 75 %.

The affinity of red crab haemocyanin for oxygen was clearly dependent on temperature (Fig. 1A). The ANCOVA showed that the dependence of logP₅₀ on pH (Bohr factor) was the same for haemocyanin at 20, 25 and 35 °C (i.e. homogeneous variances) and that the reduction in haemocyanin O₂-affinity induced by increased temperature was statistically significant (significant change in elevation of slopes). The affinity of haemocyanin for O₂ across the range 15–35 °C decreased from a P₅₀ of 0.80 kPa to 3.31 kPa at pH 7.6 (Fig. 1A). The haemocyanin sensitivity to temperature was lowest in the range 30–35 °C (ΔH=−31.9 kJ mol⁻¹ ) and highest between 15 and 20 °C (ΔH=−62.2 kJ mol⁻¹ ), with an overall ΔH of −59.0 kJ mol⁻¹ over the temperature range examined (15–35 °C). The magnitude of the Bohr shift (ϕ=ΔlogP₅₀/ΔpH) was lowest at the lowest temperature of 15 °C (ϕ=−0.13) but increased to a maximum at 30 °C (ϕ=−0.49) (Fig. 1A). The cooperativity of haemocyanin O₂-binding (n₅₀=3.16±0.07, N=27) of the red crab haemocyanin was not dependent on either pH or temperature (Fig. 1B).

There was clearly no difference between the ‘fixed acid’ and CO₂ Bohr shifts, and ANCOVA did not reveal any specific effect of CO₂ on haemocyanin O₂-affinity. The relationship between logP₅₀ (measured in kPa) and pH could be described by a common equation for both 0.4 and 4.0 % CO₂: logP₅₀=2.03−0.235pH (r² =0.943). Similarly, there was no obvious difference between haemocyanin O₂-binding cooperativity at 0.4 % or 4 % CO₂, with a mean n₅₀ of 3.13±0.06 (N=19).

Haemolymph dialysed against three concentrations of Mg showed no change in either affinity or cooperativity of haemocyanin O₂-binding. The relationship between logP₅₀ (measured in kPa) and pH could be described by a common equation for Mg concentrations between 6.1 and 19.7 mmol l⁻¹ : logP₅₀=1.892−0.220pH (r² =0.946) with a mean n₅₀ of 3.32±0.04 (N=15). In contrast, dialysing haemolymph against increasing concentrations of Ca resulted in a progressive increase in the haemocyanin O₂-affinity (Fig. 2A). An increase in Ca concentration from 8.7 to 23 mmol l⁻¹ resulted in a decrease in P₅₀ at pH 7.6 from 2.29 to 1.26 kPa. The change in the affinity of haemocyanin for O₂ at different concentrations of Ca could be described by the following equation: logP₅₀=−0.61log[Ca]+0.91 (r² =0.958) (Fig. 2A). The Bohr factor remained similar at all concentrations of Ca used (mean ϕ=−0.24±0.02, N=3). Similarly, the n₅₀ value of 3.30±0.04 (N=16) was not dependent on either Ca concentration or pH (Fig. 2B).

Increased concentrations of urate did not alter the mean Bohr shift of whole haemolymph (ϕ=−0.26±0.03, N=4 at 25 °C) but did produce a progressive increase in haemocyanin O₂-affinity (Fig. 3A). At pH 7.6, P₅₀ decreased from 1.82 kPa at 0.08 mmol l⁻¹ uric acid to 1.00 kPa in the presence of 4.32 mmol l⁻¹ uric acid (Fig. 3A). Interestingly, the haemocyanin O₂-affinity changed linearly with increasing [urate] rather than as log[urate] as described for other haemocyanins (Morris, 1991). Therefore, the effect of urate on the affinity of haemocyanin for O₂ at pH 7.6 could be quantified by the following relationship: logP₅₀=0.267−0.060[urate] (r² =0.997). The cooperativity of haemocyanin O₂-binding, mean n₅₀=3.08±0.05 (N=23), was not affected by different urate concentrations (Fig. 3B).

In the presence of 21 mmol l⁻¹ L-lactate, increasing the amount of urate in the haemolymph from 0.08 to 3.8 mmol l⁻¹ influenced both the affinity of haemocyanin for O₂ and the Bohr effect (Fig. 4A). However, there was no clear relationship between the Bohr effect and the concentration of urate: ϕ varied between −0.47 and −0.27. In the presence of 21 mmol l⁻¹ L-lactate, increasing the urate concentration increased the affinity of haemocyanin for O₂ (Fig. 4A) from a P₅₀ (at pH 7.6) of 1.82 kPa at 0.08 mmol l⁻¹ uric acid to a P₅₀ of 1.23 kPa (at pH 7.6) at 3.8 mmol l⁻¹ uric acid. The dependence of haemocyanin O₂-affinity on urate concentration in the presence of 21 mmol l⁻¹ L-lactate could be described by the following equation: logP₅₀=0.248−0.045[urate] (r² =0.898).

While high concentrations of either L-lactate (see below) or urate had no effect on the cooperativity of O₂-binding by haemocyanin, the combined effect of 21 mmol l⁻¹ L-lactate and urate in whole haemolymph resulted in a significant decrease in the mean O₂-binding cooperativity of the haemocyanin from 3.18±0.09 (N=7) to 2.60±0.05 (N=5) (Fig. 4B).

The effect of L-lactate on the haemocyanin O₂-affinity of G. natalis was very unusual: increasing the concentration of L- lactate in the haemolymph resulted in a decrease in the haemocyanin O₂-affinity. Furthermore, the magnitude of the effect did not appear to be simply dependent on L-lactate concentration, but rather on the ageing of the haemolymph sample (Fig. 5; Table 1). The decrease in haemocyanin O₂- affinity in haemolymph enriched with L-lactate was greater in ‘fresh haemolymph’ (1 day after sampling) than in aged haemolymph (4–21 days old; Table 1). The O₂-affinity of haemocyanin in the presence of 20 mmol l⁻¹ Li⁺ (P₅₀=1.92 kPa at pH 7.6) was indistinguishable from that of the original whole haemolymph (P₅₀=1.93 kPa at pH 7.6).

Haemolymph enriched with L-lactate had a significantly lower haemocyanin O₂-affinity than the corresponding whole haemolymph in all cases examined (Table 1; Fig. 5). Haemocyanin O₂-affinity in the haemolymph enriched with 20 mmol l⁻¹ l-lactate determined 1 day after collection was less than that of whole haemolymph (Fig. 5B–D; Table 1): at pH 7.6, the P₅₀ increased by 0.31±0.14 kPa (N=3). In contrast, in haemolymph enriched with L-lactate several days after collection, the increase in P₅₀ was only 0.14±0.02 kPa (N=4). Thus, the effect of L-lactate on haemocyanin O₂- affinity (ΔlogP₅₀/Δlog[lactate]) was time-dependent and decreased from a maximum value of 0.044 on day 1 to 0.001 after 4 days of storage at 4 °C. The range of concentrations of L-lactate in the haemolymph (0.1–26.5 mmol l⁻¹ ) did not affect the binding cooperativity of the haemocyanin (mean n₅₀=3.16±0.02, N=67).

The Bohr shift of dialysed haemolymph (ϕ=−0.25) was greater than that of the corresponding whole haemolymph (ϕ=−0.18; Fig. 6A) but did not differ from the overall mean value (ϕ=−0.26). However, increasing the L-lactate concentration in the dialysed haemolymph (to 21 mmol l⁻¹ ) partially reversed this effect (ϕ=−0.22). Dialysed haemolymph enriched with L-lactate had a lower haemocyanin O₂-affinity (P₅₀=1.82 kPa) than either dialysed haemolymph (P₅₀=1.26 kPa) (ANCOVA, elevation: P<0.001) or whole haemolymph (P₅₀=1.74 kPa; ANCOVA, elevation: P<0.001; Fig. 6A).

Increasing the concentration of urate in dialysed haemolymph reduced the Bohr shift (ϕ=−0.12) in comparison with that of both dialysed and whole haemolymph (Fig. 6A). The haemocyanin O₂-affinity in dialysed haemolymph enriched with 4.32 mmol l⁻¹ urate (P₅₀=0.87 kPa) was significantly higher than that in dialysed haemolymph (P₅₀=1.26 kPa) at the physiological pH 7.6 (Fig. 6A). The mean n₅₀ of whole haemolymph (n₅₀=3.23±0.07, N=5) was not significantly different from that of dialysed haemolymph (Fig. 6B). However, the O₂-binding cooperativity of dialysed haemolymph enriched with urate (n₅₀=2.8±0.19, N=8) was significantly lower than that of dialysed haemolymph (n₅₀=3.31±0.06, N=5).

In G. natalis under resting conditions (pH 7.6 at 25 °C), the affinity of haemocyanin for O₂ (P₅₀=1.77 kPa) and the Bohr shift (ϕ=−0.26) were comparable with those of other terrestrial crustaceans (Morris, 1991). Attempts to correlate haemocyanin O₂-affinity with terrestriality have not been very successful. The mean haemocyanin affinities for O₂ at pH 7.6 are not substantially different when aquatic (P₅₀=2.21 kPa), amphibious (P₅₀=2.03 kPa) and terrestrial (P₅₀=2.10 kPa) species (data from Mangum, 1983; Morris, 1991; Truchot, 1992; Morris and Bridges, 1994) are compared, but such comparisons are phylogenetically confounded. Comparison among closely related species, for example within the Grapsidae, do show a progressive increase in haemocyanin O₂- affinity with increased reliance upon air breathing (Morris and Bridges, 1994). Within the Gecarcinidae, however, the haemocyanin O₂-affinity of red crabs is similar to that of the closely related G. lallandi and Gecarcinus lateralis (Morris and Bridges, 1994; Redmond, 1968) but lower than that of Cardisoma spp. (P₅₀<1 kPa; Redmond, 1962; Dela-Cruz and Morris, 1997) which, while obligate air-breathers, are still restricted in their distribution to areas with access to open water (e.g. Morris and Adamczewska, 1996).

The affinity of G. natalis haemocyanin for O₂ was quite sensitive to temperature changes (ΔH=−59 kJ mol⁻¹ ), but the tropical habitat of red crabs presents an environment of constant temperature with little selection pressure for the evolution of a haemocyanin of reduced temperature-sensitivity (Burnett et al. 1988; Morris, 1991; Eshky et al. 1996). The cooperativity of haemocyanin O₂-binding (n₅₀) of 3.2 in G. natalis was well within the general range of 2–4 reported for other decapods (Mangum, 1983; Truchot, 1992; Morris et al. 1996a) and was not affected by manipulating the concentrations of inorganic ions or by temperature changes. However, high concentrations of urate added to either whole haemolymph in the presence of 21 mmol l⁻¹ L-lactate or to dialysed haemolymph reduced n₅₀ to 2.6 and 2.8, respectively. Allosteric binding to haemocyanin of modulators of O₂- affinity can change the binding cooperativity of the respiratory pigment by preferentially affecting either the oxy- or the deoxy-conformation state of the haemocyanin (Zeis et al. 1992) or by disrupting the interactions between subunits (Decker et al. 1989). Since both urate and L-lactate alter the conformational structure of haemocyanin by binding to the subunits (Johnson et al. 1984; Nies et al. 1992), simultaneously high concentrations of both these modulating substances may change the cooperativity of O₂-binding.

A reduction in the O₂-binding cooperativity in dialysed haemolymph enriched with urate, but not in whole haemolymph enriched with urate, provides evidence for the presence of some dialysable factor in whole haemolymph which stabilised the quaternary structure of the haemocyanin. The presence of dialysable factors in the haemolymph which affect the response of haemocyanin to modulators or to O₂-binding have been postulated previously in the terrestrial anomuran Birgus latro (Morris et al. 1988) and more recently in other species of decapod (Bridges et al. 1997; Lallier and Truchot, 1997).

The concentrations of urate measured in the haemolymph of crustaceans range from 0.01 to 0.7 mmol l⁻¹ (Henry and Cameron, 1981; Lallier et al. 1987; Lallier and Truchot, 1989; Dela-Cruz and Morris, 1997; Morris et al. 1996b; Morris and Callaghan, 1998). Since the effect of urate on haemocyanin O₂-affinity in red crabs was relatively small (ΔlogP₅₀/Δ[urate]=−0.06) and linearly dependent on the urate concentration, the increase in haemolymph [urate] in G. natalis after 20 min of intermittent exercise (0.04 mmol l⁻¹ ; Adamczewska and Morris, 1998) would result in an increase in haemocyanin O₂-affinity of only 0.05 kPa and have minimal significance in optimising the function of the haemocyanin in O₂ transport in vivo.

The sensitivity of decapod crustacean haemocyanins to Ca²⁺ (ΔlogP₅₀/Δlog[Ca]) ranges from 0 to −0.82, with terrestrial crustaceans showing a marked reduction in sensitivity (for reviews, see Morris, 1990; Morris and Bridges, 1994). In comparison with other terrestrial crustaceans, the sensitivity of G. natalis haemocyanin to Ca²⁺ (ΔlogP₅₀/Δlog[Ca]=−0.61) was quite large. During O₂ shortage, Ca²⁺ concentrations increase relatively slowly in the haemolymph over a period of hours, but the concentration of L-lactate may increase within minutes (Wood and Randall, 1981; Morris et al. 1986b, 1996b; Lallier et al. 1987; Lallier and Walsh, 1990). While L-lactate and urate may generally be of reduced utility in land crabs, retaining Ca²⁺ as a modulator of haemocyanin O₂-affinity may be advantageous for air-breathing crabs experiencing chronic acidosis (for a review, see Morris and Bridges, 1994).

The reverse effect of L-lactate, decreasing haemocyanin O₂- affinity in G. natalis, is unique among the Crustacea (Truchot, 1980; Morris et al. 1985b; McMahon, 1985; Morris, 1990, 1991; Morris and Bridges, 1994; Truchot, 1992). Additionally, the effect of L-lactate on the haemocyanin O₂-affinity was not related to the L-lactate concentration but instead to the time that the haemolymph had been stored in vitro. As a result of this time-dependent effect, the magnitude of the effect of L-lactate on haemocyanin O₂-affinity was difficult to determine precisely. However, on the day after the haemolymph had been withdrawn from the animals, ΔlogP₅₀/Δlog[lactate] ranged from 0.013 to 0.044, with a mean value of 0.029±0.009 (N=3) (Fig. 5A–D; Table 1) but declined to 0.001 thereafter.

The fact that storage of haemolymph in vitro resulted in a change in the sensitivity of haemocyanin to L-lactate modulation within 4 days while the haemocyanin O₂-binding properties of whole haemolymph remained unchanged for up to 11 days (Table 1) suggests that the effect of L-lactate is independent of the fundamental haemocyanin O₂-binding properties of the haemocyanin in red crabs. Allosteric binding of L-lactate to haemocyanin is quite specific (Graham, 1985) and results in conformational changes in the haemocyanin that affect the affinity of O₂-binding by this protein (Johnson et al. 1984; Nies et al. 1992). The presence of an unstable and dialysable unidentified cofactor in the haemolymph could account for the diminished effect of L-lactate on haemocyanin O₂-affinity with time. The current data provide further evidence supporting the presence of unknown factors in the haemolymph of crustaceans which affect the O₂-binding properties of haemocyanin (Morris et al. 1985a, 1988; Bridges et al. 1997; Lallier and Truchot, 1997).

The ‘reverse’ lactate effect represents a unique and novel potential adaptation to life on land among the Crustacea. Given that the availability of O₂ in air is more than sufficient to maintain haemocyanin oxygenation at the gas-exchange surfaces in red crabs (Farrelly and Greenaway, 1994; Adamczewska and Morris, 1994a, 1998), the decrease in haemocyanin O₂-affinity will assist in O₂ unloading at the tissues without compromising loading (for reviews, see McMahon, 1986; Morris, 1990). Fish and aquatic crustaceans possess feedback mechanisms to increase haemoglobin or haemocyanin O₂-affinity and thereby to maximise O₂ uptake from a relatively O₂-poor environment (Weber, 1980; Booth et al. 1982; Lallier and Truchot, 1989; Morris, 1990; Burnett, 1992). In contrast, the ease of O₂ extraction from air allows mammals to decrease the haemoglobin O₂-affinity during increased O₂ demand to maximise O₂ delivery to the tissues but still maintain O₂ uptake from air. Like mammals, red crabs appear to possess mechanisms to reduce the affinity of haemocyanin for O₂ during the initial stages of O₂ shortage during exercise to optimise O₂ unloading at the tissues at an unchanged ⁠.

The binding sites for L-lactate in crustacean haemocyanin have been estimated at 0.2–0.5 per O₂-binding site, with a high affinity for L-lactate (Johnson et al. 1984; Nies et al. 1992). The rate of L-lactate production in red crabs can reach 1.8 mmol l⁻¹ min⁻¹ (Adamczewska and Morris, 1994b) and, assuming that the number and affinity of L-lactate binding sites on the haemocyanin of G. natalis are similar to those of other species, haemocyanin would become saturated with L-lactate within the first few minutes of exercise. Thus, the increase in L-lactate concentration during an initial period of exercise would assist O₂ off-loading at the tissues by decreasing the O₂- affinity of haemocyanin; subsequent increases in L-lactate concentration would not reduce haemocyanin O₂-affinity further and thus would not compromise O₂ loading.

To describe the functional role of haemocyanin in G. natalis, model OECs were constructed, at the haemolymph pH measured in vivo and using the haemocyanin O₂-binding properties determined in vitro, for crabs at rest and for crabs after 20 min of intermittent exercise (Adamczewska and Morris, 1998). Superimposed on the model OECs were the and the concentration of haemocyanin-bound O₂ measured in vivo for individual crabs (Adamczewska and Morris, 1998). After 20 min of intermittent exercise, the Bohr shift arising from the haemolymph acidosis predicts a decrease in the O₂- affinity of haemocyanin (0.43 kPa increase in P₅₀) compared with that of the crabs at rest (Fig. 7). This predicted Bohr shift would increase O₂ unloading at the tissues by 70 % (Fig. 7, filled bars); however, the in vivo data show that the haemocyanin O₂ delivery to the tissues increased by more than 100 % (Fig. 7, hatched bars).

In an attempt to describe the haemocyanin O₂ functioning in vivo more precisely, the modulating effects of Ca²⁺, urate and L-lactate were incorporated into the modelled OEC (Fig. 8). An increase in haemolymph Ca (of 2 mmol l⁻¹ ) and urate (of 0.04 mmol l⁻¹ ) concentration after 20 min of intermittent exercise (Adamczewska and Morris, 1998) could theoretically increase the haemocyanin O₂-affinity at the P₅₀ by up to 0.14 kPa. However, the increase in L-lactate concentration after exercise would predict a decrease in affinity of approximately 0.3 kPa, thus partially opposing the effect of Ca²⁺ and urate and decreasing the haemocyanin O₂-affinity at the P₅₀ by 0.17 kPa more than predicted by the Bohr effect alone (Fig. 8).

It is important to note that the effects of modulators on haemocyanin O₂-affinity are not necessarily additive (Morris et al. 1986a,b, 1987; Zeis et al. 1992), and model curves for in vivo functioning of haemocyanin in O₂ transport must be used with caution. However, the in vitro modelling of O₂ transport by haemocyanin in G. natalis appeared to approximate closely the data collected in vivo regarding O₂ transport (Adamczewska and Morris, 1998). The unique effect of L-lactate, to decrease the binding of O₂ by haemocyanin beyond that predicted by the Bohr shift, appeared to be adaptive in red crabs by maximising the delivery of O₂ to the tissues during exercise. This role for L-lactate is dependent on maintained O₂ loading and thus the exchange efficiency at the lungs, and these two features must have evolved in concert.

In red crabs, up to 80 % of the haemocyanin-bound O₂ was released to the tissues during exercise; thus, the O₂ partial pressure and O₂ reserves in the venous haemolymph were relatively low and did not provide a large O₂ diffusion gradient into the tissues (Adamczewska and Morris, 1998). During intermittent exercise, red crabs were operating at the limit of the O₂-transport system and any additional work would have to be supported by anaerobic metabolism. Whether the O₂- binding properties of haemocyanin limit exercise during the migration of red crabs awaits field evaluations of the in vitro models generated in the present study.

This work was supported by grants of the Australian Research Council to S.M. and was carried out while A.M.A. was in receipt of an Australian postgraduate award.