The Regulation of Haemocyanin Oxygen Affinity During Emersion of the Crayfish Austropotamobius Pallipes : III. The Dependence of Ca²⁺-Haemocyanin Binding on the Concentration of L-Lactate

During emersion the concentrations of calcium and L-lactate ions in the haemolymph of the crayfish Austropotamobius pallipes increase significantly (Taylor & Wheatly, 1981; Morris, Tyler-Jones, Bridges & Taylor, 1986a). It has, in addition, been shown that these changes in Ca²⁺ and L-lactate affect the oxygen affinity of the crayfish haemocyanin (He) in an agonistic manner (Morris, Tyler-Jones & Taylor, 1986a). Implicit in these reports was the suggestion that the binding, to the haemocyanin, of one of these ion species affected the binding of the other.

The increase in haemocyanin oxygen affinity caused by elevated [Ca²⁺] has been documented by a number of workers (Larimer & Riggs, 1964; Truchot, 1975; Weiland & Mangum, 1975; Arisaka & van Holde, 1979; Graham, Mangum, Terwilliger & Terwilliger, 1983), and the effect has been shown to be large in A. pallipes (Morris et al. 1986a,b). The potentiating effect of L-lactate on haemocyanin oxygen affinity (Truchot, 1980) has been found throughout the Crustacea although not in all species (Booth, McMahon & Pinder, 1982; Graham et al. 1983; Mangum, 1983a; Bridges, Morris & Grieshaber, 1984; Bridges & Morris, 1986; Morris & Bridges, 1986a). The direct binding of Ca²⁺ to haemocyanin has been demonstrated in several studies (e.g. Morimoto & Kegeles, 1971; Brouwer, Bonaventura & Bonaventura, 1978; Kuiper et al. 1979). There are, however, few investigations of the binding of L-lactate to haemocyanin (Graham et al. 1983; Johnson, Bonaventura & Bonaventura, 1984; Graham, 1985).

There is no evidence that Ca²⁺ and L-lactate in crustaceans interact directly or bind interactively to the haemocyanin (Graham et al. 1983). A study of turtle blood (Jackson & Heisler, 1982) indicates, however, that some calcium lactate should be formed in physiological solutions. In the present study, we determine to what extent the binding of Ca²⁺ to A. pallipes haemocyanin is dependent on the concentration of L-lactate in a species with a large calcium effect. It is suggested that the existence of such interdependent binding of these ions may result in a reduction of the observed potentiation of O₂ affinity by any one of these ions.

The haemolymph used in these investigations was obtained by sampling (approx. 500μd) from the pericardial space of several (N>10) Austropotamobius pallipes (Lereboullet). The animals had been collected and maintained, and the blood handled as described previously (Morris et al. 1986a,b). As all the experimental material was to be dialysed, the ion concentrations in the pooled sample were not determined; instead, the dialysis solution was prepared on the basis of previous measurements made in this species (Morris et al. 1986b) and, except when explicitly mentioned, had the following composition (in mmol I⁻¹): NaCl, 181; KC1, 4·7; CaCl₂, 17; MgCl₂, 1·0; NaHCO₃, 3·0 (pH 7·9). In those cases where CaCl₂ was either 9 or 45 mmol 1⁻¹, the concentration of NaCl was adjusted so that [Cl⁻] remained constant. These two values are representative of maximum and minimum [Ca²⁺] measured during a 24-h emersion period (Morris et al. 1986a). The use of ‘hot’ or ‘cold’ refers to the inclusion or absence of ⁴⁵Ca²⁺ in the experiment.

Samples (2·5 ml) of pooled A, pallipes haemolymph were dialysed at 4°C for 24h against a Ringer’s solution (5 1, pH 7·9) containing 9, 17 or 45 mmol 1⁻¹ CaCl₂. The concentration of haemocyanin in the resulting solutions was determined spectro-photometrically by measuring the absorbance peak near 335 nm and calculating the concentration using an extinction coefficient of In subsequent experimental series protein concentration was also measured according to the method of Bradford (1976) (BioRad, protein test kit). This consistently produced values between 4 and 7 % greater than the calculated values for haemocyanin, indicating the presence of a non-haemocyanin protein in the haemolymph. The total concentrations of Ca²⁺ in the pooled sample and in all subsequent sub-samples in this series were determined in a u.v.-visible spectrophotometer (Kontron, Uvikon 710) using a colourimetric test (test kit 1028, Roche, Basel, Switzerland). After these determinations had been made, each dialysed haemocyanin solution was divided into two and centrifuged at 136000 g for 30 min in a precooled rotor on an Air-Fuge (Beckman, California, USA) to pellet the haemocyanin. From each preparation (1 ml) 100 μl of supernatant Ringer’s solution was removed and replaced by 100 μl of the same Ringer (pH 7·9) containing either 10 or 105 mmol 1⁻¹ L-lactate to give final concentrations in the remixed haemocyanin solutions of 1 or 10·5 mmol 1⁻¹, respectively.

Each of the preparations was then introduced into the upper chamber of a Centricon 30 microconcentrator (Amicon GmbH, Witten, FRG). The microconcentrator is a styrene-acrylonitrile tube holding a polycarbonate membrane filter (M_r cut-off 30000; haemocyanin M_r > 68 000) at the mid-point. The tube can be capped at both ends and placed in a centrifuge. Centrifugation at 5000g for 20 min forced approximately 25 % of the haemocyanin-free Ringer through the membrane into the lower chamber (filtrate). The amount was determined by weighing the collecting vessel before and after collection. A sub-sample was then taken from the filtrate to determine the total Ca²⁺ concentration ([Ca²⁺]_f) for each preparation. The filtrate and the retained solution containing the haemocyanin were then proportionally recombined to produce a reconstituted sample and the [Ca²⁺] was measured. This determined how much, if any, Ca²⁺ had bound irreversibly to the microconcentrator parts. Taking these controls into account, it was then possible to compare the [Ca²⁺]_f of the haemocyanin solutions containing 10 and 1 mmol 1⁻¹ L-lactate. In addition, by subtracting [Ca²⁺]_f from the Ca²⁺ concentration measured in the original dialysed samples ([Ca²⁺]_d), it was possible to estimate the bound Ca²⁺ in relation to [haemocyanin]. The procedure was repeated in triplicate.

In the second series of determinations, four haemolymph samples were pre dialysed (24 h) against one of four Ringer ‘s solutions (pH 7·9) containing approximately 0, 1, 5 and 10 mmol 1⁻¹ L-lactate and 17 mmol 1⁻¹ Ca²⁺. A sample was removed from each dialysis for the determination of [haemocyanin] and [lactate⁻], and for the construction of oxygen equilibrium curves (see below). L-lactate was determined by the method of Gutmann & Wahlefeld (1974) modified according to Engel & Jones (1978).

⁴⁵CaCl₂ (Amersham Buehler GmbH & Co KG) was then added to the dialysis Ringer ‘s solution so that 100μl of Ringer contained 114000 d.p.m. (Ca²⁺= 17 mmol l⁻¹). This involved the addition of 20 μl of ⁴³Ca²⁺ solution to each dialysis solution, which produced an insignificant change in [Ca²⁺], and dialysis was then continued for a further 12h.

The resulting haemocyanin-⁴³Ca²⁺ solutions were introduced into microconcentrators and treated as in the method described above. The ^4;,Ca²⁺ activities of the dialysis Ringer, dialysed blood, filtrate, retentate and reconstituted mixture were determined in a programmable scintillation counter (Beckman, Model LS1801, Irvine, CA, USA) which corrected for quenching and counting efficiency. The quenching correction factor was installed by counting known ⁴⁵Ca²⁺ activities in the absence and presence of various amounts of A. pallipes haemocyanin. The difference between [Ca²⁺]_f in the filtrate and [Ca²⁺]_d in the dialysed haemolymph represented the calcium bound to haemocyanin.

The introduction of EDTA (250mmol l⁻¹) into the dialysed haemolymph prior to ultrafiltration resulted in the chelation of practically all Ca²⁺ (17 mmol l⁻¹) and thus negated any effect of the haemocyanin on the distribution of Ca²⁺. All samples were generated and measured as described above.

The oxygen equilibrium curves constructed for the haemocyanin dialysed against four different L-lactate concentrations in the second experimental series were prepared, using a spectrophotometric method, on samples in a diffusion chamber (Sick & Gersonde, 1969). This method, modified according to Bridges, Bicudo & Lykkeboe (1979) was that used previously in the investigation of Ca²⁺ and lactate effects on A. pallipes haemocyanin (Morris et al. 1986a,b). The half-saturation tension (P₅₀) and the cooperativity of O₂ binding at this tension (n₅₀) were calculated from regression equations for values between 25 and 75 % saturation (see Bridges et al. 1984).

Unless otherwise stated, all means are expressed ±S.D. The means of the sample groups were compared using Student’s t-test, and significance levels are stated at appropriate points in the text and figure legends.

In these experiments conducted without the use of ⁴⁵Ca²⁺, the concentration of haemocyanin was determined to be 50±6 mg ml⁻¹ For the purposes of this experiment this was assumed to represent constant [haemocyanin]. The concentrations of Ca²⁺ measured in the various fractions from the microconcentrators are given in Fig. 1. Dialysis against the two extreme Ca²⁺ concentrations employed (9 and 45 mmol l^{− 1}) resulted in haemocyanin solutions (A in Fig. 1) containing apparently greater concentrations of Ca²⁺ than had been present in the dialysis Ringer. A similar trend is observed at all three Ca²⁺ levels; the [Ca²⁺] of the filtered solution containing 1 mmol 1^{− 1} L-lactate is significantly lower than that of the dialysed haemocyanin solution. The magnitude of this difference is related to the absolute concentration of Ca²⁺. It was not possible to demonstrate that the filtrate (C in Fig. 1) from preparations containing 10 · 5 mmol 1^{− 1} L-lactate contained concentrations of Ca²⁺ that differed significantly from those of the corresponding dialysed haemolymph (A in Fig. 1), but in all cases the [Ca²⁺] of the high-lactate solutions were higher than those in the low-lactate solutions (Fig. 1).

The proportional recombination of filtrate and retained material produced three solutions (D in Fig. 1), each of which contained Ca²⁺ at a similar concentration to that of the respective dialysed haemolymph (A in Fig. 1), thus indicating that there was no significant binding of Ca²⁺ to the microconcentrator. The amounts of Ca²⁺ calculated (see Materials and Methods) to be bound to the haemocyanin in the presence of 1 mmol 1^{− 1} L-lactate were 94, 8 · 70 and 4 · 8 mmol Ca²⁺ to one molecule of haemocyanin at [Ca²⁺] of 45, 17 and 9 mmol 1^{− 1}, respectively. Increasing the concentration of L-lactate to 10 · 5 mmol 1^{− 1} reduced the amount of bound Ca²⁺ by 63 · 5, 53 · 5 and 25% for the same three Ca²⁺ concentrations, respectively.

The haemocyanin concentrations of the four preparations had a mean value of 53 ± 5 mg ml⁻¹. The different L-lactate concentrations of the four solutions are given in Figs 2 and 3. It is immediately apparent from Fig. 2 that the concentration of Ca²⁺ in the dialysed haemolymph preparations is significantly greater than that in the dialysis Ringer’s solution (17 mmol 1⁻¹). The filtrate, as hypothesized, contained Ca²⁺ at concentrations similar to that in the Ringer. Statistical analysis determined, however, that three of the four filtrates contained Ca²⁺ in amounts significantly greater (P<0·05) that that of the Ringer ‘s solution. The retained haemocyanin solution held, in all four cases, Ca²⁺ at a higher concentration than the original dialysed haemolymph, which is consistent with the removal of 25% of the Ringer (filtrate) with a lower [Ca²⁺].

To clarify the role of L-lactate in influencing the binding of Ca²⁺ to haemocyanin, the contribution of bound Ca²⁺ to the [Ca²⁺]a measured in dialysed haemolymph was corrected to a constant haemocyanin concentration of 50 mg ml⁻¹ (Fig. 2). This was achieved by determining what the total calcium content of the solution was and adjusting this for a slightly higher or lower protein content. The amount of Ca²⁺ bound was also calculated as the mole ratio (Table 1; Fig. 2), which at near 0 m moll⁻¹ lactate approached 9mmolCa²⁺mmol⁻¹ haemocyanin. It can be seen from Table 1 that increasing L-lactate from 0·08 to 8·6 mmol 1⁻¹ brought about a reduction in bound Ca²⁺ of nearly 30%. A regression line was calculated for these data and can be described by the following relationship:

The individual oxygen equilibrium curves are not presented, but the P₅₀ and n₅₀ values for each of the curves made using the solutions with four different [lactate^-] are shown (Fig. 3). The slope of the plots of log P₅₀ pH and hence the Bohr effect (ϕ) did not differ significantly (P< 0·05 ; analysis of covariance). The mean value for all the data was calculated as ϕ = 0 ·46 ± 0 ·02. The relationship between oxygen infinity and [lactate⁻] could be expressed by the regression equation at pH 7·8: log P₅₀⁼ 0· 589– 0·041 log[lactate⁻]. This describes a very slight effect of [lactate⁻] on oxygen affinity which was, however, still significant over the range [lactate⁻] = 0·08– 8·6mmol 1⁻¹ (analysis of covariance, P<0·05). It was not possible to demonstrate a dependence of n₅₀ on either the concentration of L-lactate or the pH within the physiological range.

The objective of this experiment was to determine whether calcium ions would distribute non-preferentially between the various microconcentrator fractions when haemocyanin-bound calcium was released. Although there was some variation in the amount of ⁴⁵Ca²⁺ in each of the dialysed solutions containing the four different concentrations of L-lactate, the same trend was observed throughout (Fig. 4). The differences in [Ca²⁺] previously seen in the comparison of dialysed haemolymph with the corresponding haemocyanin-free filtrate, and also with the retained material, were not observed in the presence of EDTA. The Ca²⁺ was indeed distributed evenly between all fractions, as would be expected for a small ion with no tendency to bind to the macromolecular components of the solution.

The physiological consequences of variation in the haemolymph concentration of Ca²⁺ and L-lactate for the oxygen affinity of A. pallipes haemocyanin have been discussed previously (Morris et al. 1986a,b). These investigations, although showing the interactive effect between calcium and lactate ions in potentiating haemocyanin oxygen affinity, demonstrated neither the actual binding of Ca²⁺ to assembled haemocyanin polymers nor that the extent of this binding was dependent on the concentration of L-lactate.

Previous studies on the role of Ca²⁺ in crustacean haemocyanin function have considered primarily the assembly requirements for this ion (e.g. Morimoto & Kegeles, 1971; Kuiper et al. 1979) or have concentrated on the role of Ca²⁺ as an allosteric effector of the binding of O2 to haemocyanin (e.g. Truchot, 1975 ; Brouwer et al. 1978; Miller & van Holde, 1981 ; Wheatly & McMahon, 1982; Mason, Mangum & Godette, 1983; Taylor, Morris & Bridges, 1985).

Kuiper et al. (1979) concluded that the number of calcium ions binding to Panulirus interruptus haemocyanin was pH-dependent and also that two types of oxygen-linked calcium-binding sites existed. These sites effected preferentially the T or the R state. Klarman & Daniel (1980) concluded that a major role of both Mg²⁺ and Ca²⁺ was to increase cooperative binding of O₂ to arthropod haemocyanin. Differences in the binding of Mg²⁺ and Ca²⁺ were, however, demonstrated for the haemocyanin of Callianassa califomiensis by Arisaka & van Holde (1979), who identified approximately eight binding sites for Ca²⁺, some of which had a very high affinity. The number of binding sites identified by Kuiper el al. (1979) in Panulirus interruptus was at least 10 at physiological pH.

Previous measurements of calcium binding by haemocyanins have usually been performed with isolated, purified haemocyanins using Tris buffers and EDTA treatment initially to remove all Ca²⁺. The present study used only native haemocyanin, with the only manipulation being a dialysis against a Ringer ‘s solution of a similar composition to that of the plasma.

The estimated amount of Ca²⁺ bound by A. pallipes haemocyanin in the present study was dependent on the free Ca²⁺ concentration, although apparently less so at higher concentrations, as indicated by the cold experiments. This suggests that the haemocyanin of A. pallipes may, much like that of C. californiensis (Arisaka & van Holde, 1979), have some high-affinity and a number of lower-affinity Ca²⁺ binding sites. The Ca²⁺ sites that are lactate-labile may be regarded as low-affinity sites, whereas those still occupied by Ca²⁺ at high lactate concentrations may prove to be of higher affinity. Taking account of the results from both cold and hot experiments, a value of nine Ca²⁺-binding sites per molecule of A. pallipes haemocyanin can be calculated, which is within the same order of magnitude as reported previously. In addition, using the same methods as in the present study, we showed that the haemocyanin of Callinectes sapidus had 4– 5 Ca²⁺-binding sites per molecule of haemocyanin (C. R. Bridges, in preparation); suggesting that nine sites for A. pallipes haemocyanin is not an overestimate.

The direct binding of L-lactate to crustacean haemocyanin has been more difficult to demonstrate. The steric nature of this molecule and the postulated binding site have been discussed (Graham et al. 1983; Mangum, 1983b; Graham, 1985; Bridges & Morris, 1986), but only Johnson et al. (1984) have determined the number of binding sites. In their study of C. sapidus haemocyanin, only 2·8 lactate-binding sites could be assigned per six oxygen-binding sites; this indicates a ratio of <1·0 between L-lactate and the haemocyanin oxygen-binding site.

The present study determined that log P₅₀/log [lactate⁻] (the effect of L-lactate on the oxygen affinity of the A. pallipes haemocyanin) had a low value (−0·04) in the presence of 17 mmol 1⁻¹ Ca²⁺, confirming that the haemocyanin was functionally similar to the haemocyanin used in previous investigations. The value was, however, lower than that previously reported (Morris et al. 1986b), but it still represents a significant effect of L-lactate. Evidence has been presented that the magnitude of the lactate effect is dependent on [Ca²⁺] as is the allosteric effect of Ca²⁺ on [lactate⁻] (Morris et al. 1986b). It is now possible to conclude that this interdependence is due to interaction, not necessarily direct competition, in the binding of the two ions to the haemocyanin of A. pallipes. Graham et al. (1983) also attempted to demonstrate a similar phenomenon in solutions of Cancer magister haemocyanin, by measuring the change in the ‘free’ calcium concentration with a selective electrode in diluted samples after the addition of L-lactate. No change was detected in these experiments, which is interesting in view of the reduced activity of calcium at high lactate concentrations. A. pallipes haemocyanin also exhibits a high calcium sensitivity of CÉ binding (Morris et al. 1986b). Results for Callinectes sapidus haemocyanin (C. R. Bridges, in preparation) suggest that lactate-calcium interactions may be reconciled with haemocyanins with a large calcium sensitivity of O₂ binding (Mason et al. 1983).

The formation of calcium lactate has been shown to occur in aqueous solutions to an extent defined by the association constant (Gosh & Nair, 1970; Jackson & Heisler, 1982). The value of [Ca²⁺/Hc]/[lactate⁻] in respect of possible calcium lactate formation was estimated to be approximately −0·35. This is slightly different from the value of −0·32, calculated on the assumption of no formation of calcium lactate. The net result of this difference would be a reduction (at [lactate^-] = 8·6 mmol I⁻¹) in Ca²⁺/Hc from 6·2 to 5·9. This refinement of the data represents a ‘worst case’ estimate, as it was not possible to determine the role of other ions such as Mg²⁺ and haemocyanin acting as counterions for L-lactate. Additionally, any formation of calcium lactate would not influence the distribution of the calcium in such a way as to simulate haemocyanin binding, since the Ca²⁺-lactate⁻ complex could easily pass the filter membrane and distribute non-preferentially, as did the EDTA-Ca²⁺ chelate.

This report demonstrates that increasing the concentration of L-lactate reduces Ca²⁺ binding. In one experiment, an increase in [lactate⁻] from 1 to 10·5 mmol 1⁻¹ reduced Ca²⁺ binding by approximately 50%. The second series, using ⁴⁵Ca²⁺ and increasing [lactate⁻] from 0·08 to 8·60 mmol 1⁻¹, resulted in a reduction of nearly 30%. The dependence of Ca²⁺ binding on the concentration of L-lactate was determined to have a coefficient of −0·35. This would suggest that an increase in [lactate⁻] of approximately 3 mmol I⁻¹ would be required to reduce by one the number of calcium ions bound to a haemocyanin molecule, when [Ca²⁺] = 17 mmol I⁻¹.

We cannot at present make firm conclusions about the mechanism of this interaction. Guesnon, Poyart, Bursaux & Bohn (1979) have discussed the binding of lactate⁻ and Cl⁻ to human haemoglobin and whether this occurs at a common site. In this case the ions have at least the same charge sign, whereas it is more difficult to envisage competition at the same site between oppositely charged lactate^- and Ca²⁺. These authors (Guesnon et al. 1978) also demonstrated interaction between lactate^-, Cl⁻ and 2,3 diphosphoglycerate, which all have the same charge sign. A similar interaction, but with both positively and negatively charged ions, is observed in solutions of A. pallipes haemocyanin between lactate, calcium and urate ions (Morris, Bridges & Grieshaber, 1986; Morris & Bridges, 1986b). Manwell (1961) has suggested that the intramolecular interactions within the haemocyanin could be due to configurational changes rather than to the presence of any specific group within the protein. This argument may be applied to the interaction of Ca²⁺ and L-lactate. L-lactate induces a configurational change that decreases the affinity of the haemocyanin for Ca²⁺. Kuiper, Zoila, Finazzi-Agro & Brunori (1981) have speculated that Ca²⁺-binding sites on P. intemiptus haemocyanin may be clustered, but there is no absolute requirement for lactate-and Ca²⁺-binding sites to be in close proximity if the competitive effects are mediated by steric changes in protein configuration. Such a conclusion would not be counter to the hypothesis of Johnson et al. (1984) that L-lactate binding sites may be between haemocyanin subunits.

The interdependence of the potentiating effects of lactate and Ca²⁺ binding to A pallipes haemocyanin has been shown to have particular physiological consequences (Morris et al. 1986a). The binding of these ions potentiates O₂ affinity to varying extents, the effect increasing in a non-linear manner when the effector concentration is increased (see Tables in Bridges et al. 1984; Bridges & Morris, 1986). The reduction in the binding of one effector ion species brought about by the increased binding of the second effector species results in an apparent reduction in the potentiation due to the second effector. This creates the situation of apparent modulation of modulator function. This feature may be adaptive, however, as pointed out by Mangum (1983b) for C. sapidus. The simple addition of the separate effects of lactate and Ca²⁺ would increase haemocyanin oxygen affinity to an extent that O₂ extraction by the tissues would become impaired. In A pallipes, preventing haemocyanin oxygen affinity from increasing to too great an extent, whilst maintaining the ability to modulate O2 binding, may be a key adaptive feature.

We should like to thank Drs E. W. Taylor and R. Tyler-Jones for providing samples of haemolymph from /I. pallipes and Ms A. Lundkowski for technical assistance. Financial support was provided by the Royal Society, London (SM) and by the Deutsche Forschungsgemeinschaft (CRB and MKG, Gr 456/10-1).