Water Transport Across the Short-Circuited Midgut of the American Silkworm

The midgut of the American silkworm transports potassium actively from the blood-side to the lumen (Harvey & Nedergaard, 1964) and when the midgut is short-circuited, the short-circuit current (I_SC) is approximately equivalent to the active K transport. The rate of movement of water has been considered to be low. It has been measured as 9 ±70μl h⁻¹ (N=3) from blood-side to lumen in an open system (Nedergaard, 1972). In the closed system described by Harvey, Haskell & Zerahn (1967) the water flow could be monitored from the shape and size of the midgut and it was obvious that the movement of water was small.

In this paper the movement of water across the short-circuited gut is compared with the active K flux; the water flux follows the K flux from the blood-side to the lumen and depends on the K and sucrose concentrations in the bathing solution. When the K flux is stopped by lack of oxygen in the solution the water flux also stops.

Fifth instar larvae (weight about 10 g) of Hyalophora cecropia (L) grown on artificial diet (Riddiford, 1968) or willow were used. The larvae were chilled on ice for at least 1 h before the gut was removed and placed in the apparatus described by Harvey & Zerahn (1972). The lumen side was open, so it was easy to take samples and get circulation with a slow flow of oxygen saturated with water vapour. Control experiments in which the gut was substituted with a piece of plastic tubing showed that no change of ²²Na concentration in the lumen took place during 1 h, provided that there was no evaporation or addition of water during the experimental period. The blood-side was closed and stirred with a magnetic flea. All bathing solutions contained (in mmol 1⁻¹) NaCl, 30; CaCl, 1; MgCl₂, 1; KHCO₃, 2; plus different concentrations of KC1 and sucrose. The concentrations of K were 16, 32, 64 and 128 mmol 1⁻¹. The concentrations of sucrose were 100, 166 and 332 mmol 1⁻¹. All experiments were performed on short-circuited guts, and the active K flux was assumed to be equal to the short-circuit current (I_sc)

The water flux was measured by monitoring the decrease in concentration of ²²Na in the lumen solution. This radioisotope (half-life 2·6 years) was chosen because it emits gamma rays as well as hard beta particles (β⁺). Both kinds of radiation can be easily measured. The penetration of ²²Na from the lumen side to the blood-side was small and was corrected for. Possible leaks in the preparation were checked by adding the very powerful dye Amaranth (Merck) to the lumen solution so that a leak would quickly be revealed by a pink-coloured blood-side solution. Leaky preparations were discarded because the leak would make the flux from lumen to blood-side too high.

The activity of ²²Na was measured with a statistical deviation of 0·1%, so the amount of radioisotope had to be 0 ·5–1 μCi in the 10 ml lumen solution to perform a single measurement in about 10 min. Samples were taken in duplicate from lumen and blood-side after a 1-h experimental period. All samples were 1 ml, and were taken with a Carlsberg pipette. From the weights of test samples, a standard deviation of ± 0·1% was found (N = 10). The same tested pipette was used for initial and final samples. The radioactivity measurements were also tested by taking 10 measurements of 1-ml samples for both gamma and beta counting, and the accuracy was found to be close to 0·1%. Radioactivity was measured at least twice for every sample: the early measurements were made with gamma counting and the later with beta counting. Gamma ray emission from ²²Na was measured with a Nal crystal, ‘Selectronic’ amplifiers and scalers and a pulse height analyser. The beta particles were measured with a Packard Tricarb liquid ion spectrometer.

The gut (about 100 mg) was placed in the apparatus with 30 mmol 1⁻¹ Na in the solution and then short-circuited. After a few min, ²²Na was added and after 5 min the two starting samples were taken allowing equilibration of the gut Na with the luminal Na.

Na is not transported actively to any significant degree when the solutions contain Ca and Mg with a K concentration as high as used here (Harvey & Zerahn, 1972). The concentrations of Ca and Mg were not made higher than 1 mmol 1⁻¹ because these ions can be actively transported to a certain extent by the midgut (Wood, Jungreis & Harvey, 1975; Wood & Harvey, 1976).

The results are expressed as μmol of water transferred in relation to μmol of K⁺ transported, evaluated from the graphical picture given by the recorder during the experimental period of 1 h. From Table 1 it can be seen that the ratio between water transport and potassium transport was dependent on the concentration of K in the bathing solution and had a minimum around 32mmoll⁻¹ and 166mmoll⁻¹ sucrose. When the concentration of sucrose was doubled, from 166 to 332 mmol 1⁻¹, the transport of water per transported K was decreased from 75 ±3 to 14 ± 7.

The active K transport (I_sc) is dependent on the K concentration (| K |): in a Lineweaver-Burk plot, l/I_scvs 1/| K | gives a straight line (K_m= 7·5 ±1·7; Zerahn, 1982). As can be seen from Table 1, I_SC was not much affected by increasing the sucrose concentration to 332 mmol 1⁻¹, even if the water flux showed a pronounced decrease.

The I_sc decreased drastically when metabolism was reduced by replacing oxygen in the bathing solution with nitrogen, and water movements decreased (Table 2). The small I_sc left during anoxia could result from the transport of some K followed by some water. The rate of water movement was calculated from the ratio μl H₂Oμequiv K to be 16·3 μl Using this correction on the observed value of 10·6 gave a total value of − 5·7 μl, which is too small to be determined with any reasonable accuracy.

Experiments to determine the osmotic movement of water across the midgut were performed on larvae grown on willow (Salix babylonica). The gut was bathed with 200 mmol 1⁻¹ sucrose saline on the lumen-side and 100 mmol 1⁻¹ sucrose saline on the blood-side. The K concentration was 64 mmol 1⁻¹. In five experiments, the mean I_K was 2800μA, and mean water transfer was 281/dh⁻¹. Control experiments with identical solutions on both sides with 100 mmol 1⁻¹ sucrose gave a ratio for H₂O/K of 31 ± 16 (N = 3). With 200 mmol 1⁻¹ sucrose the ratio H₂O/K was 37 ± 14 (N = 4), mean value 34. From Table 1 a value of about 80 could be expected, so it seems that diet may affect the ratios.

As 2800 μA will be equal to 104 μequiv h⁻¹ of K transported, 34 water molecules per K corresponds to 3552μequiv of water, equal to 18 × 3552μg, or 64μlof water. The rest of the water, 281 – 64 = 217 μl, is osmotically transferred.

104μequiv of K⁺ were actively transported across the gut, and would have been neutralized by chloride from the AgCl electrode. In 8 ml solution (10 – 2 ml, for the first samples) 104μequiv of K⁺ would give rise to a concentration of 104/8 = 13 mmol 1⁻¹ or 26 mosmol KC1. Starting with no osmotic difference and building up to 26 mosmol at the end, the mean value during 1 h will be 13 mosmol.

A difference of 100 mosmol moves 217μl of water, so 13 mosmol will move 217 × 13/100 = 28μl. A maximum of 28 μl of the transported water can be explained as moving as a result of increased osmotic pressure in the lumen, and this will be less if K penetrates the gut passively. The rest, at least 64 –28μl = 36μ1, must come from another source, viz. the water following the active K transport.

In all experiments where the midgut is bathed in solutions with the same composition on both sides, no net flow of water due to osmotic differences across the gut will take place; if there is a net flow of water it must be derived from the processes in the gut, active K transport being the most likely. This is in agreement with the observation that if the K transport is reduced by lack of oxygen the water flux is also stopped (Table 2).

The low value for the rate of water movement observed by Nedergaard (1972) (9 ± 70 μh⁻¹) is probably because the the guts were not short-circuited, but had their natural potential difference, which will give a low rate of K transport and a large back-flux of K with unknown influence on water flux.

Table 1 shows that the concentration of K and sucrose may both be important for the water flux. It may be that the water is following the K transport, and the water molecules and K⁺ ions are both passing a transport route or channel, the size of which should be large enough to contain the amount of water co-transported with K⁺, and small enough to let the K⁺ ion have its influence: a magnitude of some 100 water molecules seems likely. With high sucrose concentration the channel is narrowed and the amount of water is small, so only little can follow the active K transport. No explanation for the minimum value of the H₂O/K ratio with approx. 32 mmol 1⁻¹ K and 166mmoll⁻¹ sucrose has been found (Table 1). With 100mmoll⁻¹ sucrose a minimum for the ratio was not found but the ratio was higher at all K concentrations. This may agree with the concept that the transport channel is wider with the lower sucrose concentration.

The number of water molecules following one atom of K is about 100 in 64 mmol 1⁻¹ K solution. There is 56 mol 1⁻¹ water in this solution, giving a ratio between water molecules and K atoms of 875, so the transport of K is very far from being isosmotic with the solution.

In Malpighian tubules or gall bladder the transported solution is approximately isosmotic. In the isolated and short-circuited frog skin, however, there is a similar low rate of transport of water with the active Na transport (Capraro & Marro, 1963), compared to isosmotic transport.

The water transport is probably only possible in connection with the active transport, but the K transport will continue even though the water transport may vary greatly. A model for a K pump should include these processes.

I thank Dr V. Koefoed-Johnsen and Professor S. O. Andersen for reading and discussing the paper, and Miss Lene Dam and Mrs Susanne Munk Jensen for excellent technical assistance.