ABSTRACT
A method is described for investigating the rate of loss of water and carbon dioxide from terrestrial insects by absorbing tritiated water and 14CO2 from a gas stream flowing past the insect. The loss of water and carbon dioxide can be studied simultaneously with a time resolution (nominal) down to 2 min.
The theoretical and experimental bases of analysing the data are considered in detail. The determination of the efflux rate constant for water is straightforward and, if an estimate of surface area is available, the efflux rate constant can be converted to a permeability coefficient. In the case of 14CO2 loss, the interpretation is complicated by the presence of other compartments within the body that can be labelled with 14C. A multicompartment model of 14C exchange is developed and a method of obtaining the efflux rate constant of 14CO2 is described. The efflux rate constant for 14CO2 can be used to estimate CO2 output.
Introduction
Several methods have been used to study water loss of insects in air. Most studies have been made by gravimetric methods, in which loss of mass in dry air has been followed. An early example is the classic study of Ramsay (1935a,b). Other studies have been made in air of different humidities (e.g. Loveridge, 1968). Gravimetric methods can only be applied under non-steady-state conditions, where there is a mass loss to measure. Movements of living insects result in balance instability (e.g. Noble-Nesbitt, 1969; Machin et al. 1991; Lighton, 1992), limiting the accuracy of measurement. Continuous mass recording improves the accuracy which can be attained with the gravimetric method to some extent (Noble-Nesbitt, 1969, 1991; Kestler, 1985; Machin et al. 1991).
The early use of isotopic water (deuterated water, DHO) to investigate water permeability in animals is outlined by Krogh (1939) and includes an aquatic insect example (Libellula). Isotopic water exchange in terrestrial arthropods was first studied by Govaerts and Leclercq (1946), also using DHO. Subsequently, the availability of radioactive tritiated water (THO) has led to its wide use in investigating permeability in aquatic arthropods (including Smith, 1969; Lockwood et al. 1973; Nicolson and Leader, 1974) and also to a number of investigations in terrestrial arthropods (including Wharton and Devine, 1968; Coenen-Staß and Kloft, 1976; Nicolson et al. 1984; Appel et al. 1986; Machin et al. 1992). Another elegant technique for measurement of water loss involves infrared absorbance studies of air flowing past animals. This technique has the advantage that CO2 can also be measured (e.g. Hadley and Quinlan, 1993; Lighton et al. 1993). These methods have a high sensitivity and fast time resolution but also some experimental limitations.
Isotopic methods have a number of advantages for investigating water and CO2 exchange in small terrestrial animals, including insects. High specific activities can be used (particularly for THO) to increase sensitivity. The methods can be used in the presence of high ambient and , including steady-state conditions. They allow simultaneous measurement of influx and efflux using two isotopes or one isotope combined with weighing (Appel et al. 1986). Moreover, isotopic methods enable a direct comparison to be made with measurements on aquatic animals. These features provide advantages over both the gravimetric and infrared absorbance methods. Although there is only one radioactive form of water, in principle H2O, DHO, THO and H218O could be resolved using mass-spectrometry techniques (Nagy, 1983) and H218O could also be investigated following conversion of 18O to radioactive 18F using the 18O(p, n)18F reaction (e.g. Wood et al. 1975; Cooper, 1982; Nagy, 1983).
Although there are differences between the diffusion coefficients of the various isotopic forms of water (Kohn, 1965), House (1974) concludes that it seems unlikely that this is a serious source of error in determinations of water permeability for biological membranes. However, House (1974) does point out that the assumption that both the inner and outer media are well mixed can be a source of error.
Loss of THO from Periplaneta americana and Blaberus trapezoideus into a flowing airstream was investigated by Coenen-Staß and Kloft (1976), who monitored the THO loss continuously by passing the gas through a Geiger-type detector, but they were unable to calibrate water loss in any absolute unit. Nicolson et al. (1984) absorbed THO from an airstream directly into a scintillation cocktail in a series of vials in an ice bath. This method has also been used by Machin et al. (1992). The method of Nicolson et al. (1984) forms the basis of the technique described in the present paper.
The loss of CO2 can be studied using similar methods. If the insect is loaded with 14C-labelled substrates, 14CO2 in the air phase can be measured and the CO2 output estimated. Rilling and Steffan (1972) have used this method to study the CO2 output from a phylloxerid, continuously feeding on 14C-labelled sucrose. This approach has been continued by Rhodes (1992), who investigated the contributions of various substrates to CO2 production in aphids. Interesting information might be obtained if a mixture of T- and 14C-labelled substances were used.
Materials and methods
Flow-through apparatus
Isotope fluxes were determined using a flow-through system (Fig. 1A). The experimental chamber consisted of a modified 60 ml plastic syringe with gas-tight plunger. A small (2 mm diameter) hole was drilled through the plunger and polythene (0.7 mm i.d.) tubing was threaded through the hole and secured with Araldite to ensure a gas-tight seal. Air was drawn through a drying column of 6-16 mesh self-indicating silica gel (Fisons) with a peristaltic pump (Minipuls 2, Gilson Medical Electronics) or passed through two gas bottles in series, each containing appropriate saturated solution (O’Brien, 1948; Winston and Bates, 1960) to humidify it. A large gas volume above the solution in the first bottle in series increases the time constant and facilitates equilibrium. A four-way valve allowed switching of the airstream. The airstream then passed through the chamber. A Vaisala temperature–humidity probe was used to check chamber humidity. Flow rates were checked with a bubble flowmeter. Air flowing through the chamber and over the unrestrained insect then passed through polythene tubing and an attached 0.6 mm diameter (23 gauge) hypodermic needle cemented with Araldite in the side arm of a Pettenkofer tube. The length of tubing between the syringe and Pettenkofer tube was kept as short as possible. The Pettenkofer tube consisted of an approximately 520 mm long by 8 mm i.d. glass burette equipped with a PTFE stopcock (Interflon) and a 1 mm i.d. glass side-arm attached about 20 mm above the stopcock. The Pettenkofer tube was held at an angle of approximately 45 ° from vertical in an ice-filled polystyrene container. The side-arm, stopcock and tip extended beyond the ice container, where they remained accessible for ease of manipulation.
Depending upon the isotopes studied, either 10 ml of scintillant (Pico-Fluor 15, Packard) (to absorb tritiated water vapour) or 10 ml of scintillant plus 1 ml of tissue solubilizer (NCS-II, Amersham International) (to absorb 14CO2 also, as tissue solubilizer is a highly alkaline solution) was poured into the Pettenkofer tube. The height of the liquid in the tube could be adjusted by variable insertion of a glass rod. Bubbles emerging from the needle in the side-arm travelled up through the column of liquid. In trials using tubes in series, it was found that more than 99 % of the tritiated water vapour was absorbed in the first tube and a similar high absorbancy was found for 14CO2 with this methodology (Rhodes, 1992). Nicolson et al. (1984) and Machin et al. (1992) found a similar high efficiency of THO absorbance with a much higher flow rate. Sampling periods of 15, 10 or 2 min were used, followed by the rapid draining of the liquid into a scintillation vial and refilling the tube. In the case of the longer collection times, there was a very short period of a few seconds when the Pettenkofer tube was empty. However, the tube was still wet with absorbent and this should minimise losses. In the case of 2 min changes, another four-way valve was used to switch the air flow immediately to another previously filled Pettenkofer tube, avoiding any loss of isotope (Fig. 1B). An air flow was always maintained into each Pettenkofer tube to avoid any backflow of scintillant into the polythene tubing.
An air flow rate of 10 ml min-1 was used in all experiments. A higher flow rate is unnecessary since, at this flow rate, the loss of H2O and CO2 will be rate-limited by the animal. This flow rate should not be greatly different from air movement in the natural environment of the insect. The volume of the chamber was adjusted to either 30 or 20 ml. The theoretical time constant of the chamber (τ) is given by volume/flow rate. Thus τ=3 or 2 min, respectively. In reality, the air in the chamber is not completely mixed and the real time constant must exceed these figures. The time constant, of course, limits the time resolution of this method. Although the time constant could be reduced by higher flow rates, a low rate should minimise Bernoulli effects that might result in forced ventilation of the tracheal system. This might have been an effect with the high flow rates used in previous studies (e.g. Ramsey, 1935a,b; Hadley and Quinlan, 1993).
The apparatus has been used to investigate the rate of H2O and CO2 loss from unrestrained Periplaneta americana (Noble-Nesbitt et al. 1995). In principle, the apparatus could be used to change the nature of the gas mixture (pCO2, pH2O, etc.) flowing over the insect during the experiment. It should also be possible to modify the apparatus so that there is a continuous countercurrent of absorbent flowing down the Pettenkofer tube into a fraction collector or into a continuous-flow radioactivity detector.
At the end of an experiment, the insect was anaesthetised with CO2 or ether (in the case of experiments involving 14CO2), removed from the chamber and weighed. The majority of the abdomen was removed, weighed and ground up in 5 ml of tissue solubilizer. After extraction overnight and centrifugation, 40 μl samples of the supernatant were added to 10 ml of scintillant and counted to obtain the residual activity. It was assumed that the cockroach contained 72 % (w/w) water (Noble-Nesbitt and Al-Shukur, 1987) and that the abdominal sample was representative of the animal as a whole. The residual activity (R) is given by:
The THO and 14CO2 activities were determined by β-scintillation counting (Packard Tricarb 300), using appropriate gate settings and quench corrections to obtain separate activities. In most cases, the samples were counted to a standard error of ±1 %.
In the experiments described here and in Noble-Nesbitt et al. (1995), Periplaneta was labelled with THO by allowing it to feed on a small piece of food to which 5 μl of THO in water (0.37 MBq μl-1; 1.85 MBq) was added using a Hamilton syringe. This non-invasive method of loading was used in preference to injection. However, preliminary experiments using loading by injection showed that broadly consistent results are obtained with both methods of loading, although injection has the attendant risk of damage and trauma. The loaded insect was left overnight for the THO to equilibrate with the body fluids. In some experiments, the loss of 14CO2 was studied simultaneously with THO loss. The H14CO3− pool was labelled by injecting 10 μl of 14CO32- solution (3.7 kBq μl-1; 37 kBq) between two abdominal sternites of a CO2- or cold-anaesthetised Periplaneta, using a Hamilton syringe. The site of injection was sealed with a drop of Newskin (Beecham) as the needle was removed. As the loss of 14CO2 was so rapid, the insect had to be injected shortly before the experiment commenced following overnight THO equilibration.
Data processing and discussion
An insect can be labelled with THO by injection or feeding or by exposure to THO vapour, as was done with DHO by Govaerts and Leclercq (1946). Feeding is the simplest method. In Fig. 2A, the values of the rate constant are given for an experiment starting soon after the animal was fed. The apparent rate constant rises with time. This suggests that the THO has not had time to distribute uniformly throughout the various water compartments in the system. There is other evidence (Pichon, 1970) that mixing of haemolymph in the cockroach is a slow process. When, after feeding, the insect was left overnight to equilibrate, the rate constant was broadly stable (Fig. 2B). Similar results were obtained with loading by injection. Any fluctuations then represent real variations of k0 with time. Incidentally, equation 4 could also be used when rates of loss are studied by a gravimetric method. The calculation of absolute rate of water loss is easy:
The CO2 output of an insect can be studied by similar procedures but in this case the efflux rate constant will presumably be determined by diffusion from, or ventilation of, the tracheal system rather than by cuticular permeability. One method would be to load with 14C-labelled substrate and absorb 14CO2 from the air. In the present study, a more direct method was used, loading with labelled 14CO32-. This would convert to H14CO - at physiological pH and label the HCO - pool in the body fluids. As metabolically produced CO2 will pass through this pool (Fig. 3A), the efflux rate constant, calculated in the same way as the THO efflux rate constant, would give information on the rate of CO2 loss. This can be compared with the H218O method for estimating CO2 loss (Lifson and McClintock, 1966). As 14CO32- does not label the water pool, it has advantages over H218O. However, following injection of 14CO32-, the apparent efflux rate constant falls rapidly with time (Fig. 4A). A similar rapid fall in the efflux rate constant of 45Ca2+ from cells was interpreted in terms of Ca2+ uptake into a slowly exchanging compartment (Tomlinson et al. 1991). A similar model can be applied to the interpretation of 14CO2 loss. Some of the injected 14C will move back up the metabolic pathways and thus be present in a non-H14CO3- form. Consider the model given in Fig. 3B. The fluxes between the compartments can be described by the equations:
The rate of excretion of CO2 (JCO2) is given by:
where B is the quantity of HCO3- in the body fluids. This can be estimated from the volume of water in the animal and the concentration of HCO3- estimated using the Henderson–Hasselbalch equation. For the insect illustrated in Fig. 4, taking pH=7.1 and pCO2 = 40 mmHg (5.3 kPa), then HCO3-=20 mmol l-1 and the CO2 output was estimated as 396 μl g-1 insect h-1. This estimated value for CO2 output is within the range of values for Periplaneta determined by Kestler (1971). Using the values above, the mean ventilation rate can also be estimated as 125 μl g-1 insect min-1. Of course, the actual ventilation rate will fluctuate with time about this mean value (see Noble-Nesbitt et al. 1995).
The technique outlined enables the efflux of water and CO2 to be investigated in small animals with a time resolution down to a few minutes. If required, the loss of THO and 14CO2 can be studied simultaneously. In the accompanying paper (Noble-Nesbitt et al. 1995), use of the techniques to investigate the loss of water and CO2 from Periplaneta americana under various conditions is described.
ACKNOWLEDGEMENTS
We are grateful for the assistance of Dr M. Williams and Mr. G. Thomas in preparing computer-generated figures and equations.