Role of active potassium transport by integumentary epithelium in secretion of larval-pupal moulting fluid during silkmoth development

Curiously little is known about the moulting process in insects, although the nucleic acid and protein metabolism of moulting and its hormonal control have been studied exhaustively (see reviews by Wigglesworth, 1970; Chen, 1971; Wyatt, 1972, and Doane, 1973). After mitosis of the integumentary epithelial cells and prior to formation of a new epicuticle, the moulting gel is formed from zymogen granules produced by these cells and is secreted into the ecdysial space which will persist between the old and new cuticles. The ultrastructure of gel formation has been studied in a Lepidopterous insect (Locke, 1969; Locke & Krishnan, 1973; Hikida & Jungreis, 1975) and the gel has been shown to contain inactive forms of hydrolytic enzymes (Lensky, Cohen & Schneiderman, 1970; Katzenellenbogen & Kafatos, 1971a, b, c; Bade & Shoukimas, 1974)-These latter workers adhered to the classical view of Passonneau & Williams (1953) that moulting fluid is simply a sol form of the moulting gel (Katzenellenbogen & Kafatos, 1970; Lensky et al. 1970). However, Jungreis (1973, 1974) has shown that the fluid present in the exuvial space when the old cuticle is being actively hydrolysed and reabsorbed is a potassium salt solution in which the enzymes of the gel are dissolved and presumably activated (see Lensky et al. 1970). An apt analogy is the formation of cocoonase solution by adult silkmoths in which a crystalline pro-enzyme is secreted by the galea and subsequently is activated by a potassium bicarbonate buffer solution which is secreted by the labial glands (Kafatos, 1968). In the case of the moulting fluid, however, the enzyme-containing gel and the potassium salt solution are both secreted by the same tissue, the integumentary epithelium, but at different times. This paper will deal with the mechanism of formation of the potassium salt solution by the pharate pupal integument (PPI) of Hyalophora cecropia (L): this solution constitutes the bulk of the fluid volume (moulting fluid) in the exuvial space prior to ecdysis.

We will show that the potassium salt solution is formed by the active transport of potassium across the integument followed by a presumed osmotic flow of water. The potassium pump in the integument resembles that in the labial gland (Kafatos, 1968) and indeed it has many properties in common with that in other insect tissues which secrete or reabsorb fluids in conjunction with active potassium transport (reviews by Maddrell, 1971; Berridge & Oschman, 1972). However, the biophysics of the potassium pump has been difficult to study in these small, fluid-secreting tissues and many of its properties have been analysed in the large, isolated, silkworm midgut (Harvey & Wood, 1972, 1973; review by Harvey & Zerahn, 1972). Keynes (1969) has classified the K-pump in Malpighian tubules, labial gland, and midgut as a Type V pump. Its principal properties are that (1) potassium is secreted from blood-side to apical side of an epithelium, rendering the apical side electrically positive to the blood-side, (2) no counter ion is required, (3) the pump is electrogenic, and (4) it is not inhibited by ouabain. However, unlike the other K-transporting insect tissues, the midgut does not secrete a fluid. The pharate pupal integument has great potential as a model for analyzing the role of the K-pump in fluid secretion by insects.

The first clue that active potassium transport plays a role in the formation of the moulting fluid came from the demonstration that potassium is present at virtually the same high concentration in both moulting fluid (132 m-equiv./l) and integument (135 m-equiv./kg), but at a much lower concentration (31 m-equiv./l) (P < 0·001) in the haemolymph of pharate pupae of H. cecropia (Jungreis & Tojo, 1973; Jungreis, 1974). In addition, the osmotic pressure of moulting fluid was 38 % greater than that of blood collected from the same insects. These observations suggested that potassium is actively transported from blood-side (haemolymph) to exuvial side (moulting fluid) of the developing pupal integument, and that fluid movement between the integumentary epithelial cells is restricted (see Hikida & Jungreis, 1975). Water would follow potassium passively, but not to equilibrium, resulting in the formation of a moulting fluid hyper-osmotic to haemolymph. In the present study this model has been tested both in vivo and in vitro and has been confirmed. Preliminary evidence has also been obtained which suggests that resorption of the moulting fluid is not driven by potassium transport, but that it is driven by the active transport of some other ionic species, possibly bicarbonate.

Larvae of the giant American silkmoth, Hyalophora cecropia (L.) were reared either in the laboratory on a synthetic diet (Riddiford, 1968) at 25 °C and 65 % R.H. or out of doors on wild cherry foliage. Pharate pupae were selected for analysis at stages described elsewhere (Jungreis, 1974; Hikida & Jungreis, 1975). The pharate pupal integument was dissected from the larval cuticle at 24 to 1 h before ecdysis, when it consists of a delicate cuticle developing over a one-cell-thick layer of pupal integumentary epithelial cells. Midgut, fat body, and other adhering tissues were scraped gently from the haemolymph-side of the integument and the larval cuticle was then separated from it by inserting blunt-tipped forceps between the two layers and separating them carefully at the major points of attachment, the spiracular junctions. The isolated integument was transferred as a flat sheet to a glass or Plexiglass chamber, was tied in place with fine cotton twine, and an area of 0·75 cm² was exposed to separate, identical bathing solutions on both sides. The standard solution, designated as 32-K-S-Tris, consisted of 32 mm-KCl, 5 mm-MgCl₂, 5 mm-CaCl₂, 5 mm Tris-HCl and 160 mm sucrose, had a pH of 8·60, and had a total osmotic equivalent of 260 milliosmolal (Harvey & Nedergaard, 1964; Wood, 1972). Other solutions are described in the appropriate table or figure.

The potential difference (PD) across the PPI was measured by calomel electrodes, was read on a Keithley Electrometer (Model 602 A), and was recorded on a Servoscribe i S Recorder. The PPI was short-circuited as a flat sheet by Wood’s modification (1972, 1975) of the classical technique of Ussing & Zerahn (1951). Wood’s principal modification was to introduce a third calomel electrode to record the potential gradient in the solution when current was passed through it and to use this value to correct automatically for the resistance of the bathing solutions. The short-circuiting was accomplished manually using a current supply, designed by Mr Dennis Unwin of Cambridge University, which passed current through the solutions via Ag-AgCl electrodes. The short-circuit current, I_sc, was read on a Keithley Picoammeter (Model 410A). Fluxes of potassium, J_K, were measured with either ⁴²K or ⁸⁶Rb by comparing the radioactivity in samples on the cold side to the specific activity of the isotope on the side to which it was added. To confirm that ⁴²K and ⁸⁶Rb could not be distinguished by the. epithelial transport system, both were placed sirriultaneously on the exuvial side of the integument, which was bathed in 32-K-S-Tris, and the unidirectional fluxes were measured over a 2 h period. ⁸⁶Rb was transported at 94–110 % (mean = 105 %) of the rate of ⁴²K, indicating that ⁸⁶Rb is virtually indistinguishable from ⁴²K and can be used as a tracer for potassium fluxes in this solution.

The PD across the integument was measured in vivo in selected body segments at 60-12 h before the larval-pupal ecdysis. For this purpose two tuberculin syringes were fitted with 0-5 in, 27-gauge hypodermic needles and were attached to polyethylene tubing. Each of the two bridges formed in this way was filled with a 3 % agar, 32-K-S-Tris solution and was connected via a calomel electrode to the Keithley electrometer. One needle was pushed through the integument and into the haemolymph, where it served as a reference electrode, whereas the other one was used to probe the exuvial space.

The moulting fluid was, in all regions, invariably positive with respect to the haemolymph (mean = 8·8 mV, n = 20) at 40–24 h before ecdysis, when moulting fluid is being secreted from haemolymph to exuvial space (Table 1). The PD was still positive but smaller in magnitude (mean = +1·2, n = 8) at 18–12 h before ecdysis, when moulting fluid is being resorbed. Significantly, the material in the exuvial space was negative with respect to the haemolymph (approximately − 20 mV) at 60 h before ecdysis, when moulting fluid is not yet abundant (Table 1). When the integument was isolated and bathed with the same physiological saline (32-K-S-Tris) on both sides, the exuvial side was positive to the haemolymph-side in all cases which extended from 30 to i h before ecdysis (Table 2).

The isolated integument was short-circuited and a typical time-course of the I_sc and K-fluxes is shown as Fig. 1. This preparation was bathed in a 32 mm-KCl solution but showed a similar time-course of I_sc and K-fluxes as those preparations bathed in 32-K-S-Tris. The average steady-state value of the I_sc was 0·182 μ-equiv. cm⁻² h⁻¹(n = 5) during the period from 30 to 3 hours before ecdysis (Table 2). The ratio of the potassium influx (J_ln) toward the exuvial side, to efflux (J_in) toward the haemolymph-side, ranged from 5·4 to 64 under open-circuit conditions and from 2·4 to 7·5 under short-circuit conditions at 30 to 3 h before ecdysis (Table 2). Despite values of the PD and I_sc at 2 to 1 h before ecdysis being similar to those at earlier times, the potassium flux ratios at that time were close to unity. The net flux of potassium toward the exuvial side of the integument had an average value of 0·126 μ-equiv. cm^{− 2} h^{− 1}(n = 5) and accounted for 70 % (n = 5) of the I_sc at 30 to 3 h before ecdysis. However, at 2 to 1 h before ecdysis the net potassium flux accounted for only 9 % of the I_sc (Table 2).

The I_B0 of the isolated integument was maintained equally well in chloride or bicarbonate solutions, and persisted even in simple solutions consisting of 32 mm potassium chloride or 32 mm potassium bicarbonate and sucrose, glucose, and Tris (Fig. 2).

When the integument was isolated during the period from 30 to 1 h before ecdysis and bathed in 32 mm potassium on both sides the exuvial side was always positive with respect to the haemolymph-aide, after an occasional temporary negative phase (Figs. 1, 2; Table 2), suggesting that the active potassium transport postulated above to take place in the living insect still continues in the isolated preparation. (The rapid initial decrease in I_sc occurred in all preparations although the I_sc did not always become negative. The phase usually lasted no more than h in solutions free of calcium and magnesium. It might be caused by a reversal of the K-pump, by an activation of an anion pump oriented toward the exuvial side of the PPI, or by other mechanisms such as a loss of tissue K towards the haemolymph-side. It may be induced by the removal of the tissue from a hormonal influence in the haemolymph as discussed on page 364 below. However, we are concerned here only with the steady state I_sc and with the fluxes responsible for it.)

Consistent with our demonstration of active potassium transport is the requirement for oxygen in maintaining the PD across the isolated PPI (Jungreis, 1973). The potential difference cannot be maintained when potassium is replaced either by sodium or by choline (Jungreis, 1973), suggesting that the transport mechanism has a strong preference for potassium, at least in solutions containing calcium. Sodium is not present in the bathing solution and, indeed, the potential and short-circuit current are readily demonstrated in a simple solution consisting of 32 mm-KCl or 32 mm-KHCO₃, sucrose, glucose and Tris buffer (Fig. 2). This lack of counter-ion involvement suggests that the K-transport is electrogenic. Finally, both the PD across the PPI, and the I_sc, are unaffected by ouabain (Jungreis, 1973). In these respects the K-pump in the isolated pharate pupal integument resembles that of the larval midgut (Harvey & Nedergaard, 1964) and can be classified as a Type V Pump (Keynes, 1969).

An apparent discrepancy arises regarding the time before ecdysis at which K transport was demonstrated in vitro compared to the time of secretion and reabsorption of the moulting fluid in vivo. Moulting fluid has been fully resorbed in vivo by 16-12 h before ecdysis (Jungreis, 1974), yet isolated integuments actively transported potassium from haemolymph to exuvial sides until 6 h before ecdysis. Furthermore, it was the time at which the insects were dissected, rather than the time when transport was studied (normally 8 to 6 h after dissection), which determined the transport properties of the preparation. These observations suggest that some factor (perhaps hormonal), present in vivo, turns off the pump approximately 18 h before ecdysis. In the absence of this factor in vitro, the K-pump is not turned off until 2 to 1 h before ecdysis. However, within 15 min after ecdysis has occurred in vivo, the isolated integument has lost its ability to generate a PD in vitro.

We propose that in response to a favourable osmotic gradient, maintained dynamically by the active transport of K, larval-pupal moulting fluid is formed as a bulk flow of water from haemolymph to exuvial sides of the pharate pupal integument. Furthermore, we propose that this potassium salt solution, which comprises the bulk of the moulting fluid, acts as a buffer at an alkaline pH. Moulting fluid is reported to have an alkaline pH (7·8) similar to the alkaline pH optima of the hydrolytic enzymes present in moulting fluid (Lensky et al. 1970; Katzenellenbogen & Kafatos, 1971a, b, c; Bade & Shoukimas, 1974), unlike haemolymph, which has an acidic pH (Passonneau & Williams, 1953; Katzenellenbogen & Kafatos, 1970). Transport of K and water is accompanied by morphological changes in the integumentary epithelium which occur at approximately 48 h before ecdysis, when the secretion of moulting fluid is at its maximum (Hikida & Jungreis, 1975)

Electroneutrality is maintained in vivo by an as yet unidentified anion: chloride present either in blood or moulting fluid can neutralize only a small fraction of the potassium present (Jungreis, Jatlow & Wyatt, 1973; Jungreis, 1974). A possible anion is bicarbonate, which has been qualitatively measured in silkmoth moulting fluid (Jungreis, unpublished data), and which would account for the alkalinity of moulting fluid. Recall the analogy with the labial gland of adult silkmoths which secretes potassium bicarbonate (Kafatos, 1968). However, bicarbonate would have to originate intracellularly from carbonic acid produced by the integumentary cells because it is not added to the bathing solutions. In vitro the short-circuit current is maintained equally well in bicarbonate or chloride solution (Fig. 2). This hypothesis assumes a transport route through the cells and predicts (1) that when K is actively transported from the cells of the integument to its exuvial side, bicarbonate neutralizes it and (2) that there is a passive exchange of K⁺ in the haemolymph for H⁺ in the transporting cells. Although no studies of bicarbonate transport are reported here, the atttractive hypothesis that resorption of moulting fluid could result from bicarbonate transport, from exuvial side to haemolymph-side of the PPI, is supported by the large I_sc in the absence of net K transport 2 to 1 h before ecdysis (Table 2).

The clear result of the present study is that potassium is actively transported from haemolymph-side to exuvial side of the integument when it is isolated from the pharate pupa, rendering the exuvial side electrically positive with respect to the haemolymph-side. An active potassium transport from haemolymph to exuvial space also takes place in the living pupa, because the space is invariably positive to the haemolymph during the period when moulting fluid is being secreted, and because the potassium concentration is elevated in the space which is hyperosmotic to the haemolymph during secretion (Jungreis & Tojo, 1973; Jungreis, 1974). Presumably water flows osmotically from haemolymph to exuvial space, accounting for the increase in volume of the multing fluid during secretion. These observations, together with the demonstration by Jungreis (1973) that the bulk of the moulting fluid is not only formed much later than the time of secretion of the moulting gel but also overlaps the time of gel-sol transformation, forces us to revise current hypotheses regarding a simple gel → sol transformation of moulting gel to moulting fluid. We propose that the formation of the moulting fluid has three steps instead of two: (1) the classical secretion of proteinacious moulting gel, (2) its transformation to a moulting sol in the presence of a potassium salt solution, and (3) the continued addition to the sol of a potassium salt solution which accounts for the bulk of the moulting fluid volume. We speculate that the potassium salt solution has three functions: (1) to accomplish the gel-sol transformation, (2) to activate the gel enzymes, and (3) to buffer the enzyme solution at a pH favourable to the activity of gel enzymes.

This research was supported in part by a research grant (AI-05903) from the National Institute of Allergy and Infectious Diseases, U.S. Public Health Service, and in part by a research grant (GB 43457) from the National Science Foundation. We thank Dr John L. Wood for criticising the manuscript.