Temperature Effects on Haemolymph Acid-Base Status in vivo and in vitro in the Two-Striped Grasshopper Melanoplus Bivittatus

In the majority of poikilothermic animals that have been studied, increasing temperature is accompanied by decreasing extracellular pH (reviewed by Reeves, 1977; Heisler, 1984; Glass et al. 1985). This effect has been attributed to the maintenance of (1) constant fractional dissociation of imidazole groups or ‘alphastat regulation’ (Reeves, 1977) or (2) constant relative alkalinity (Howell et al. 1970). In many organisms the slope of the line describing the decrease in pH with temperature is significantly below that compatible with either of these two hypotheses (Heisler, 1984; Glass et al. 1985), and the physiological significance of pH changes with temperature remains controversial. Examination of pH regulation with temperature in diverse taxa is an important test of these hypotheses, as pH and temperature effects on enzyme action (Wilson, 1977a,b) should apply to all organisms. In this study I examined the effect of body temperature on haemolymph acid-base status in the two-striped grasshopper Melanoplus bivittatus (Say).

Insects are the largest and most diverse ectothermic taxon, and are important models for testing the significance of pH regulatory patterns with temperature. Only one study has examined the relationship of pH to body temperature in insects. Storey et al. (1984) used P-NMR data to examine temperature effects on intra-and extracellular pH in groups of freeze-tolerant gall fly larvae from –12 to 15°C. In gall fly larvae, haemolymph pH remains constant while intracellular pH varies as predicted by alphastat regulation, maintaining relative alkalinity and calculated protein dissociation fraction at all temperatures.

Two general mechanisms have been proposed for the regulation of extracellular pH with temperature in animals. In most air-breathing animals, the decrease of extracellular pH with temperature is caused by increasing due to reduced ventilation relative to metabolic CO₂ production (reviewed by Jackson, 1982). Gill-breathers such as fish and crustaceans, however, often accomplish extracellular pH regulation by transmembrane ion movements (reviewed by Truchot, 1978; Heisler, 1984). A priori, one might expect air-breathing insects such as grasshoppers to utilize variation of for the control of haemolymph pH with temperature.

Grasshoppers thermoregulate behaviourally when possible but, nonetheless, may experience substantial diurnal body temperature variation (Uvarov, 1977), making them excellent insects for testing the effect of temperature on pH regulation. One objective of this study was to examine diurnal changes of body temperature in the field for M. bivittatus, and to relate these data to the pattern of acid-base regulation. To aid in distinguishing among possible haemolymph pH regulatory mechanisms in this grasshopper, I also examined CO₂ solubility, carbonic acid pK_app, haemolymph buffering, and the effect of temperature on pH and [HCO₃⁻ ] in vitro at constant ⁠.

Melanoplus bivittatus were collected from a disturbed grassland site near Lafayette, CO, and were kept in the laboratory for 1–3 weeks before testing. Only adult females were used in the haemolymph studies because of their larger size. Animals were kept in screen cages with a 14h:10h light:dark cycle, with a photothermal gradient during the light cycle to allow behavioural thermoregulation. Temperatures within cages varied from 30 to 45 °C during the light cycle and averaged 23 °C at night. Relative humidity averaged 30%. Animals were fed on wheat sprouts and wheat bran treated with Tri-Sulfa-G (Norden Laboratories) to control amoebic infections. Under these conditions animals lived for 5–7 weeks as adults, and laid numerous eggs.

Observations of field behaviour and diurnal body temperature variations were conducted at the collection site near Lafayette, CO. Observations were made between 05.00 and 20.00 h over 6 days between 25 August and 25 September 1986, during the peak of local abundance of this species. Thoracic temperature (T_th) was measured with a copper-constantan thermocouple threaded through a 26 gauge needle, connected to a Wescor TH-65 thermocouple thermometer. Ambient temperature (T_a) was measured with a second thermocouple in the shade, 6 cm above ground level. For T_th measurement, grasshoppers were captured by hand and held with a cloth while the thermocouple was inserted into the centre of the thorax. Time between animal capture and T_th measurement averaged 10 s, and ranged from 5 to 15 s. Animals measured in less than 8 s did not differ in temperature from those measured in over 12 s.

The apparent pK (pK_app) of carbonic acid in M. bivittatus haemolymph was measured at 10°C, 25°C and 35°C at 2%, 5%, 10% and 18·7% CO₂ (humidified, balance N₂) provided by Wösthoff gas-mixing pumps. Haemolymph was pooled on ice from 20–30 animals kept at room temperature, and centrifuged for 5 min at 1400g. Total protein loss during the process was less than 5%, as haematocrits were under 3% and clotting was minimal if the haemolymph was kept on ice before centrifugation. Clotting in insect haemolymph requires factors from the haemocytes and the plasma; therefore separation of haemocytes from plasma prevents clotting, without alteration of the remaining plasma proteins (Brehelin, 1979). Haemolymph was frozen, and kept at –70°C for less than 1 week before equilibrations. Preliminary experiments indicated that freezing did not affect either the pH or the C_tot of haemolymph samples at a given and temperature.

Haemolymph melanizes quite rapidly in the presence of oxygen. To eliminate possible effects of this on acid-base status, haemolymph equilibration gases were oxygen-free. The concentration of oxygen in insect haemolymph does not exceed its physical solubility (Bishop, 1923; Babers, 1941), so complications due to a Haldane effect are unlikely. With no oxygen present, haemolymph exhibited a relatively constant decrease in pH with time (roughly –0·06 pH units h^{− 1}) which was arrested by the addition of 20 mmol I^{− 1} NaF. Under these conditions, haemolymph pH was stable for at least 8 h. Droplets (100 μl) of haemolymph were equilibrated with 2%, 5%, 10% or 18·7% CO₂ (balance N₂) for at least 1 h and pH and C_tot were measured.

C_tot was measured as described above, except that samples were collected with polyethylene tubing (PE 10) filled with 0·9% NaCl connected to a Hamilton syringe. Approximately 15 μl of haemolymph was drawn into the tubing, 2 μl was discarded and 10μl was injected into the acid-filled syringe, as described above. Preliminary experiments using saline containing [¹⁴C]lactate indicated that haemolymph sampled in this way was substantially diluted with saline (approximately 20%). However, this dilution factor was relatively constant (S.D. = 3%). This technique minimized the volume of haemolymph required and, after correcting for dilution, calculated pK_app values compared favourably with values in the literature (see Results).

Haemolymph pH was measured on 15 μl samples collected with glass microcapillary tubes. Samples were immediately transferred to the capillary electrode of a Radiometer BMSMK2 blood gas analyser thermostatted to the equilibration temperature. Samples were drawn into the pH electrode and followed by a solution of filtered, saturated KC1 which filled the polyethylene transfer tube and provided an electrical bridge to the reference electrode. This technique yielded slightly different results from those if the polyethylene bridge was completely filled with sample, so the electrode was calibrated with 15 μl samples of pH precision buffers (Radiometer S1500 and S1510) followed with saturated KC1, after repeated rinsing of the electrode with calibration buffer. When the pH of 15 μl haemolymph samples was measured after calibration with 15 μl of calibration buffers, the pH equalled that measured for 40 μl haemolymph samples measured after calibration with excess calibration buffers.

Haemolymph from 30–40 animals kept at room temperature was pooled on ice. Haemolymph was centrifuged as described above, and the supernatant split into 100 μl samples, which were frozen at –70°C. Samples of this haemolymph were tonometered as described above at 2%, 5%, 10% and 18·7% CO₂ at 10°C, 25°C and 35°C, and pH and C_tot were measured. Tonometry was performed within 1 week of freezing.

For measurements of temperature effects on haemolymph acid-base status in vivo, animals were kept individually in cardboard cartons at test temperatures for 12–16 h. Preliminary studies showed that haemolymph pH values of animals kept for 1·5 and 16 h at a temperature were identical, suggesting that this time was sufficient for attainment of steady-state conditions. No food was provided, but each carton contained a moistened paper towel to prevent dehydration of the grasshopper. Haemolymph samples were obtained by grasping the animal and making a quick incision in the ventral neck membrane, and sampling from the haemolymph which appeared on the neck. Haemolymph pH and C_tot were measured as described above for pK_app experiments. Owing to haemolymph volume limitations, pH and C_tot were measured on separate animals. was calculated from the average values for pH and C_tot at each temperature.

Fractional imidazole dissociation was calculated according to Reeves (1976). Relative alkalinity was calculated from the K_w for water (Weast, 1977). Haemolymph protein was measured on 10μl samples, diluted 9:1 (w/v) with water, and kept frozen at –70°C until analysis. Haemolymph protein was measured with a biuret reagent (Layne, 1957), using bovine serum albumin as a standard. Haemolymph total ammonia was measured spectrophotometrically on 50 μl samples, using a specific enzymatic assay (Sigma, kit 170-UV) and a Gilford model 260 spectrophotometer. Background rates of NADH oxidation in haemolymph without glutamate dehydrogenase were subtracted from measured absorbance changes. Samples were kept on ice for less than 1h before measurement, as ammonia content was substantially elevated in frozen samples.

The distribution of grasshoppers during the 30 min before sunrise depended on temperature. Below 5°C, grasshoppers were found on the ground under vegetation, with T_th at or just above T_a. At temperatures above 5°C, grasshoppers occupied upper regions of herbaceous vegetation, generally situated in sites which would become sunny microhabitats when the shrub was directly illuminated by the rising sun. In these animals, T_th was equal to or slightly lower than T_a (Fig. 1). After sunrise, at ambient temperatures between 15 and 35°C, grasshoppers were also generally located in the upper regions of vegetation. Body temperatures increased rapidly in the early morning hours as these animals elevated T_th above T_a behaviourally by choosing sunny microhabitats and basking. Above T_a values of 30°C, the slope of T_th on T_a was not significantly different from zero, indicating that these grasshoppers are excellent behavioural thermoregulators.

CO₂ solubility in M. bivittatus haemolymph and in 0·01moll^{− 1} HC1 decreased with increasing temperature (Table 1). The CO₂ solubility coefficient in haemolymph (α_haem) was fitted to the following formula: α= 0·1079 – 0·020955 InT (°C); r² = 0·994.

Carbonic acid pK_app in M. bivittatus haemolymph decreased with temperature, with a slope of –0·0081 pH units °C^{− 1} between 10 and 35 °C (Table 1). There was no significant effect of haemolymph pH on pK_app.

Plots of pH vs [HCO₃⁻ ] were approximately linear from pH 7·1 to 6·65 (⁠⁠; 12–117mmHg). Protein content averaged 45mgml^{− 1} (S.E. =0·89, N = 31), and did not vary significantly in any of the in vitro tests. Buffer value, defined as Δ [HCO₃⁻ ] × ΔpH^{− 1}, did not vary significantly with temperature, and averaged –34·9mequivl^{− 1}pHunit^{− 1} (S.E. = 1·79, N = 31). CO₂ combining curves were shifted downwards with increased temperature (Fig. 2). This was due to decreases in dissolved CO₂ and [HCO₃⁻ ] at any given ⁠. Average CO₂ capacitance from 12 to 60 mmHg did not vary with temperature, and averaged 0·200 mmol I^{− 1} mmHg^{− 1} (S.E. = 0·0066, N = 9).

Haemolymph pH tended to decrease slightly at all values from 10 to 25 °C, with a slight trend towards a lesser temperature effect at higher values (Table 2). From 25 to 35°C, pH decreased at 2% CO₂ (–0·0039 pH units °C^{− 1}), but it increased slightly at 10 and 18·7%. [HCO₃⁻ ] changes with temperature showed a more complex pattern (Table 2). From 10 to 25°C, [HCO₃⁻ ] decreased at all values, with the decrease being roughly proportional to ⁠. From 25 to 35°C, [HCO₃⁻ ] decreased slightly at 2% CO₂, and increased at 10% and 18·7% CO₂. [HC0₃⁻ (₂₅)]/[HCO₃⁻₍₁₀₎] was approximately 0·7 for all values. [HCO₃⁻₍₃₅₎]/[HCO₃⁻ (25)] was approximately 1·1 at 10 and 18·7% CO₂, and just under 1 at 2% CO₂. Changes in pH or [HCO₃⁻ ] in vitro at constant were not evaluated for statistical significance owing to small sample size.

Haemolymph pH in M. bivittatus acclimated to test temperatures for 12–16 h showed a biphasic pattern (Fig. 3). Haemolymph pH was statistically constant from 10 to 25 °C. Above 25 °C, pH decreased with increasing temperature (ANOVA, P< 0·001). Haemolymph pH decreased significantly from 25 to 35 °C, and from 35 to 40°C (Duncan’s a posteriori test, P<0·05). From 25 to 40°C, the slope of the pH change with temperature was –0·017 pH units °C^{− 1}. Total CO₂ content increased significantly from 10·2 to 13·6mmoll^{− 1} between 10 and 20°C (Duncan’s test P<0·05), and then decreased to 7·5 mmol I^{− 1} as temperature increased to 40°C (Duncan’s test, P < 0·05; Fig. 3). Although true standard errors cannot be calculated for the ⁠, S (dissolved CO₂) and [HCO₃⁻ ] data, as pH and C_tot were not measured in the same animal, estimated standard errors were calculated by using values calculated for randomly paired C_tot and pH values (Fig. 4). This procedure is supported by the lack of a significant correlation between pH and C_tot in individual locusts (Harrison, 1987).

Changes in the derived variables (⁠⁠, S, [HCO₃⁻ ]) were analysed statistically using nonparametric rank tests (Mann-Whitney U). at 20°C was significantly higher than at 10°C (P<0·01), but values at 20, 30 and 40°C did not differ significantly. Dissolved CO₂ declined significantly from 20 to 40°C (P < 0·001), but remained statistically constant from 10 to 20°C. [HCO₃⁻ ] at 40°C was significantly lower than values at either 20 or 30°C (P<0·001, P<0·05, respectively).

Haemolymph protein content was unaffected by 12–16 h of temperature acclimation, and averaged 35·7mgml^{− 1} (S.E. = 1·15; N = 30). Haemolymph total ammonia content increased from 139μmoll^{− 1} at 10°C (S.E. = 40·9; N = 8) to 500μmol I^{− 1} at 35°C (S.E. = 79·0; N = 8; Duncan’s test, P<0·05).

The diurnal pattern of field body temperatures in M. bivittatus was similar to that reported for other grasshoppers (reviewed by Uvarov, 1977). M. bivittatus experienced wide fluctuations in body temperature, indicating that these grasshoppers are particularly good models for testing the effect of temperature variation on acid-base physiology.

The CO₂ solubility coefficients determined in acidified water were bracketed by previous values in the literature (Bartels & Wrbitzky, 1960; Murray & Riley, 1971, values revised by Weiss, 1974), indicating that the techniques utilized here provided satisfactory results despite the use of small samples. CO₂ solubility coefficients in M. bivittatus haemolymph were similar to those reported for vertebrate plasma (Bartels & Wrbitzky, 1960; Austin et al. 1963; Boutilier et al. 1985). Carbonic acid pK_app values in M. bivittatus haemolymph were also similar to values reported in the literature for vertebrates and crustaceans, and the dependence of pK_app on temperature (–0·00808pHunits°C^{− 1}) agrees well with that reported for vertebrate plasmas at constant pH (Severinghaus et al. 1956; Albers & Pleschka, 1967; Boutilier et al. 1985). Apparent pK values similar to those of vertebrates have also been reported for Gastrophilus larvae (Levenbrook, 1950) and honeybee larvae (Bishop, 1923).

The haemolymph nonbicarbonate buffer value of M. bivittatus is among the highest reported for any extracellular fluid (–35 mequivl^{− 1} pH unit^{− 1}). The specific buffer value per gram of haemolymph protein, 0·76mequivg^{− 1} pH unit^{− 1}, is exceedingly high when compared with the value obtained for human haemoglobin (0·13mequivg^{− 1}pHunit^{− 1}; Adair, 1925). However, comparisons with vertebrates are difficult as invertebrate protein contents estimated from bovine serum albumin standards underestimate gravimetric measures by 25–90% (Marsh, 1987).

In other species, for example in a crab and in the tuna, extracellular fluid pH decreases from –0·007 to –0·013pHunits°C^{− 1}in vitro at constant (Wood & Cameron, 1985; Perry et al. 1985). As in vivo in M. bivittatus is statistically constant above 20°C, it is important to know the effect of temperature on pH and [HCO₃⁻ ] at constant to distinguish active pH regulation from such passive, physicochemical effects. The change in pH with temperature in vitro at constant did not exceed –0·004 pH units °C^{− 1}, indicating that the excess changes in pH with temperature in vivo must be due to active processes.

Although the haemolymph sampling procedure was clearly not anaerobic, preliminary studies indicated that, if the haemolymph droplet was sampled quickly, C_tot loss was minimal. This is apparently due to the lack of carbonic anhydrase in insect haemolymph (Levenbrook & Clark, 1950). Haemolymph droplets of smaller size than those sampled from the neck (20 μl) were equilibrated to a near the maximum observed in vivo (47 mmHg), and then exposed to air for varying times before sampling for C_tot. If the droplet was sampled in less than 10s, the loss of C_tot was less than 3% (J. M. Harrison, unpublished data).

A second important consideration is whether this sampling procedure allows collection of ‘resting’ haemolymph. In vertebrates, it has been well documented that catheterization is necessary to obtain blood at resting acid-base status (reviewed by Heisler, 1984). This might be expected to be a particular problem with insects, which are known for their capacity for rapid changes in metabolic rate. Grasshopper activity is accompanied by CO₂ accumulation which accounts quantitatively for haemolymph acidosis during activity (Harrison, 1987). In preliminary experiments, I found that elevations in haemolymph C_tot (and depression of pH) were found only after 20 s of handling. Haemolymph samples obtained from grasshoppers handled for 2–15 s were identical in acid-base status (J. M. Harrison, unpublished data). Therefore, it is reasonable to argue that the haemolymph samples obtained from animals with minimal handling time (<10s) approximated ‘resting’ haemolymph in terms of acid-base status. The absolute values measured in this study for haemolymph pH and [HCO₃⁻ ] at 20 °C (7·121, 11·9 mmol I^{− 1}, respectively) agree well with those reported for the locust, Schistocerca gregaria (7·1, 13mmoll^{− 1}, respectively; Phillips et al. 1986). The low absolute value of haemolymph pH appears to be a general characteristic of insect haemolymph (Buck, 1953).

Between 25 and 40°C, the change in pH with temperature conserves the calculated fractional dissociation state of imidazole quite well, as it varies by only 1·4% around its mean of 0·606 between 25 and 40°C. This calculation must be viewed with some caution, however, given the diversity of values for the temperature-dependence of pK among imidazoles (Edsall & Wyman, 1958). Relative alkalinity is also well conserved, varying approximately 6% around the mean value over this temperature range. Preservation of fractional protein dissociation has been shown to preserve the functional and structural properties of imidazole-containing proteins, and to be theoretically important in the preservation of the capacity for control of metabolic flux (Wilson, 1977a,b; White & Somero, 1982). However, between 10 and 25°C, pH is statistically constant and both relative alkalinity and the calculated fractional dissociation of imidazole groups change significantly, with a decrease in calculated fractional dissociation of imidazole groups of 27% at 10°C. Relative alkalinity at 10°C is reduced to 25% of its value at 40 °C.

In most grasshoppers, feeding and locomotory movements only occur above approximately 20°C (Uvarov, 1977), and this is also true for M. bivittatus. The data suggest that M. bivittatus only regulate haemolymph pH in a manner which preserves relative alkalinity and the fractional dissociation of proteins at temperatures corresponding to daily activity temperatures. This result is similar to that reported for the lizard, Dipsosaurus dorsalis (Bickler, 1981, 1982).

The mechanism by which M. bivittatus changes haemolymph acid-base status with temperature is quite different from the ‘constant CO₂’ system described for most air-breathers (Reeves, 1977). changes with temperature are inconsistent with those required to accomplish the observed pH changes with temperature in vivo. The decrease in pH between 20 and 40°C is accompanied by a statistically constant ⁠, whereas the constant pH between 10 and 20°C is accompanied by increasing ⁠. Thus the changes in haemolymph pH with temperature cannot be explained by variation of ventilation relative to metabolic CO₂ production.

These changes in amount of extracellular bicarbonate could be due to changes in metabolic production and utilization of acid-base equivalents within the haemolymph. However, this seems unlikely given the low cell fraction of grasshopper haemolymph (<3%). A more likely explanation is that the changes in haemolymph bicarbonate are due to transmembrane transfer of acid-base-relevant ions, either between the intra-and extracellular compartments and/or between the haemolymph and the lumen of the digestive tract. Locust recta possess a large capacity for H⁺ excretion, which may be due to an electrogenic H⁺-ATPase and a C1⁻ /HCO₃⁻ antiport (Phillips et al. 1986). The faeces of locusts are generally acidic, suggesting that a constant excretion of protons may be necessary for pH homeostasis in these animals (Phillips et al. 1986). Further studies of haemolymph ion activities, the temperature-dependence of net proton release by the intracellular space and net proton elimination by the excretory system will be necessary to determine the exact mechanism of haemolymph pH regulation with temperature in M. bivittatus.

I would like to thank the members of my committee, M. Breed, C. Carey, B. Mazzeo and G. Snyder for their help throughout this study. I wish particularly to thank my advisor T. Gleeson for his extensive advice and assistance. In addition, two anonymous reviewers made many helpful suggestions on the manuscript. Finally, J. Fewell was an invaluable aid throughout this project. This work was supported by grants from Sigma Xi and the Graduate Foundation Fund of the University of Colorado, and a University Doctoral Fellowship.