Ion Transport Across the Midgut of the Tobacco Hornworm (Manduca Sexta)

The physiology of the larval lepidopteran midgut has been studied in a variety of species such as Manduca sexta, Bombyx mori, Hyalophora cecropia, Philosamia cynthia and Spodoptera littoralis. Of these, Manduca sexta has received extensive attention as a model for the study of epithelial ion transport because it is easily raised in the laboratory on an artificial diet and the midgut of this large insect engages in vigorous rates of K⁺ secretion (Dow, 1986). In addition, this species is widely used to study the hormonal control of development (Bollenbacher, 1988), feeding behaviour (Timmins et al. 1988) and energy metabolism (Gies et al. 1988), and thus there is a great deal of information on the biology of this organism. Despite the interest in ion transport across the tobacco hornworm midgut (see reviews by Harvey et al. 1981, 1983; Harvey, 1980; Blankemeyer and Harvey, 1977; Wolfersberger et al. 1982; Dow, 1986), there have been surprisingly few studies of solute fluxes across this epithelium. Except for a recent report of Cl⁻ transport by Chao et al. (1989), only the transepithelial transport of K⁺, under short-circuit conditions, has been measured (Cioffi and Harvey, 1981; Wolfersberger and Giangiacomo, 1983; Moffett et al. 1983; Moffett and Koch, 1985). The transepithelial transport of other solutes has not been measured in this species.

The tobacco hornworm midgut is composed of three morphologically distinct regions (Cioffi, 1979) which transport K⁺ at different rates (Cioffi and Harvey, 1981). Recent studies (Chamberlin, 1990) indicate that the rate of luminal alkalinization also differs among the midgut regions. These studies clearly indicate that the midgut is divided into functional as well as morphological regions, yet there have been no other studies examining regional ion transport. In this communication the transepithelial transport of K⁺, Na⁺ and Cl⁻ was measured in the three midgut sections using a saline which contains physiological levels of haemolymph amino acids, sugars and ions. This saline maintains midgut function for several hours (Chamberlin, 1989) and more closely reflects haemolymph composition than the saline used in previous transport studies (see Cioffi and Harvey, 1981). Ion fluxes were monitored under short-circuit conditions to determine if the ions are actively transported. These fluxes were compared to those obtained under open-circuit conditions to determine if the large transepithelial potential, which would be present in vivo, affects the rate of ion transport.

Tobacco hornworms were grown from eggs or larvae purchased from Carolina Biological Supply Company (Burlington, North Carolina). The larvae were maintained at 28°C on a 16 h: 8 h light:dark cycle and fed an artificial diet (Carolina Biological Supply Co.). Fifth-stage larvae weighing 3–5.3g were used in all experiments.

Larvae were decapitated and a dorsal incision made to expose the midgut. The anterior, middle or posterior midgut was then cleaned of adhering tracheae and Malpighian tubules, opened, removed from the animal, and stretched over the collar of a modified Ussing chamber (Hanrahan et al. 1984). The opening (0.159 cm²) of the collar was backed with fine nylon screen to support the tissue and the midgut was secured to the collar with a cotton thread. Each half-chamber was filled with 5 ml of control saline (Table 1) and vigorously bubbled with 100 % oxygen.

The PD was measured with calomel electrodes connected to KCl-agar bridges which were inserted into the saline on either side of the midgut. PD (lumen positive) and I_sc, with compensation for saline resistance, was measured with a custom-made voltage clamp (Duke University Physiology Department Technical Shop) and monitored on Soltec recorders. When measuring ion fluxes under shortcircuit conditions, the current clamp was turned off briefly during sampling and the PD noted. The specific epithelial resistance was calculated from the I_sc and this potential.

Unidirectional K⁺, Na⁺ and Cl⁻ fluxes were measured under open-circuit and short-circuit conditions. The tissues were allowed to equilibrate in the chambers for approximately 90min prior to the measurement of isotopic fluxes. ⁴²K (119–244 mmol l⁻¹ KC1, 1.53–3.43 mCi ml⁻¹), ²²Na (0.015–0.018 mmol l⁻¹, carrier-free, 1000 μCi ml⁻¹) and ³⁶Cl (1.2 mol l⁻¹ HC1, 650 μCiml⁻¹) were purchased from Dupont/New England Nuclear (Wilmington, Delaware). Unidirectional isotopic fluxes were measured by adding stock ⁴²K, ²²Na or ³⁶Cl to one halfchamber (referred to as side I). After 15 min, a 1ml sample was taken from the other half-chamber (side II) and the volume replaced with 1 ml of isotope-free saline. This sampling procedure was repeated every 15 min for a total of 90 min. A 50 μl sample from side I was diluted with 950 μl of saline and counted to determine the specific activity of side I. ²²Na and ⁴²K were counted on a Beckman 5500 gamma counter and samples containing ³⁶Cl were added to 10 ml of scintillation fluid (scintiverse Bio-HP) and counted on a Beckman LS8000 scintillation counter.

The results of the flux studies indicated that all three measured ions were actively transported by the midgut tissue. To determine if deletion of these ions would affect the Isc, another set of experiments was performed in which the tissues were mounted in control, K⁺-free, Na⁺-free or Cl⁻-free saline (compositions in Table 1) and the monitored after 77 –103 min. The I_sc of ion-depleted midguts was compared to that of tissues that had been bathed continuously in control saline for 90 min (control in Table 3). Ions were reintroduced bilaterally to ion-depleted midguts in the form of a methylsulphate or N-methyl-D-glucamine salt to the final concentration found in control saline. After 15 min the I_sc was again noted.

The unidirectional flux for each 15-min period was calculated as described by Hanrahan et al. (1984). The radioactivity on side II never exceeded 0.5 % of that of side I and therefore no correction for backflux was necessary. The declined slightly with time and therefore a comparison between tracer fluxes with the I_sc measured at the time of sampling would underestimate the I_sc over the entire flux period. A comparison of I_sc with tracer fluxes was made by integrating the I_sc chart recordings with the use of a Hewlett Packard digitizing pad and calculating the area using a computer program. All values are presented as means±standard errors with N indicating the number of midguts. Statistical analyses were conducted using a Student’s t-test. P ⩽ 0.05 was considered to represent a significant difference.

Throughout this paper I_sc and ion fluxes are reported in units of μequivcm⁻²h⁻¹. The midgut epithelium, however, varies in thickness along its length and the amount of cellular material per unit area will be different for the three midgut regions. Surface area can be converted to mass by using the factors reported by Cioffi and Harvey (1981; anterior: 68 mg cm⁻²; middle: 52 mg cm⁻²; posterior: 112 mg cm⁻²).

Fig. 1 shows the time course of I_sc and PD for the three midgut sections. Note that there is an initial rise then a fall in these measurements. During the 90-min experimental period, however, the I_sc and PD declined only slightly (see Figs 2 –7). The resistances were 200.2±7.7, 226.5±15.6 and 216.6±9.2Ωcm² (N=43) for the anterior, middle and posterior sections, respectively. Steady-state fluxes were achieved 30 min after addition of isotope, indicating that, at most, 30 min was necessary for isotopic mixing in the extracellular and intracellular ion pool. Net K⁺ and Cl⁻ secretion and net Na⁺ absorption occurred in all three midgut sections under open- and short-circuit conditions (Figs 2–7). These net rates of ion transport remained constant throughout the entire 90 min experimental period.

Unidirectional fluxes measured between 30 and 45 min after isotope addition are summarized in Table 2. The flux ratios of K⁺, Na⁺ and Cl⁻ were all greater than 1 under short-circuit conditions, indicating active transport. Under open-circuit conditions there were no regional differences in the K⁺ flux from haemolymph to lumen or lumen to haemolymph ⁠.In contrast, under short-circuit conditions, the anterior section had a significantly greater and significantly smaller than the other sections. The middle section had the greatest under open-circuit conditions, but there were no regional differences in when the tissue was short-circuited. There was a significant increase in of the anterior section when this flux was measured under short-circuit conditions and this flux was significantly greater than that measured in the middle and posterior sections. The of the posterior midgut was significantly higher than that of the other sections under short-circuit conditions, but only greater than that of the anterior section under open-circuit conditions.

Omission of K⁺ caused a marked drop in I_sc in all midgut sections (Table 3). Reintroduction of K⁺ stimulated the I_sc in all regions, but control levels of I_sc were regained only in the middle and posterior sections. The omission of Na⁺ and Cl⁻ also caused a depression of the I_sc. Reintroduction of Na⁺ significantly stimulated the I_sc,but control values were not reached. Reintroduction of Cl⁻ had no effect.

A comparison of integrated with isotope fluxes is shown in Table 4. Note that the I_sc is not equivalent to the sum of net K⁺, Na⁺ and Cl⁻ fluxes. The disparities between net ion transport and the I_sc for the anterior, middle and posterior sections were 14.6, 4.7 and 3.4 μequivcm⁻²h⁻¹, respectively.

Previous measurements of transepithelial ion transport across tobacco hornworm midgut involved bathing the tissue with a simple saline which contained K⁺ as the only monovalent cation and sucrose as the only exogenous metabolic substrate (see Cioffi and Harvey, 1981; Crawford and Harvey, 1988). In this saline, the I_sc would rise and then fall in a biphasic fashion. The results presented in this study indicate that midguts bathed in a more complex saline also display an initial rise and subsequent fall in PD and I_sc. After approximately 90 min, however, the rates of PD and decline were less than 2%h⁻¹ (calculated from values in Figs 2 – 4) and 0.5 % h⁻¹ (calculated from values in Figs 5 – 7), respectively. These rates are far slower than the ‘pseudo steady-state’ decline of 20 % h⁻¹ reported by Crawford and Harvey (1988). These results, along with those of Chamberlin (1989), indicate that the midgut requires the presence of metabolic substrate and/or other ions to sustain in vitro performance.

Researchers, using the simple saline described above, have reported the I_sc to be as low as 243 μ Acm⁻² (9.1 μequivcm⁻²h⁻¹; Mandel et al. 1980) and as high as 1159μAcm⁻² (43.2μequivcm⁻²h⁻¹, Cioffi and Harvey, 1981). Similarly, resistances have been reported from 78.3 Ωcm² (Moffett, 1980) to 230 Ωcm² (Harvey and Wolfersberger, 1979). These disparities may arise from the use of differentsized larvae and reflect the changing properties of the midgut epithelium as larval growth proceeds. In this study the larvae were generally smaller and the saline more complex than those used in previous studies and therefore quantitative comparisons between this and previous studies should be made judiciously.

The midgut of the tobacco hornworm is composed of three distinct regions which differ in the gross epithelial folding pattern as well as in the structure of the columnar and goblet cells (Cioffi, 1979). This structural heterogeneity is reflected in regional differences in I_sc, PD and rates of ion transport. The anterior section consistently had the highest I_sc and PD, while the middle section had the lowest. Whereas there were no regional differences in K⁺ transport under open-circuit conditions, under short-circuit conditions the rate of K⁺ secretion was greatest in the anterior section. This contrasts with the findings of Cioffi and Harvey (1981), in which the anterior section was shown to have the lowest rate of K⁺ secretion and I_sc. The in the posterior region increased significantly under open-circuit conditions, indicating that this process is energized, to some extent, by the PD.

In previous studies the net K⁺ flux was virtually equivalent to the I_sc. In this study the K⁺ flux exceeded the I_sc in all three midgut sections. Omission of K⁺ produced a large fall in I_sc, indicating that K⁺ transport is electrogenic, although allosteric effects of K⁺ on other transporters cannot be discounted. Reintroduction of K⁺ stimulated the I_sc, although, in the anterior section, the control values of Isc were not achieved by this manoeuvre. This may indicate that ion substitution can cause non-specific damage to the epithelium, such as perturbation in cell volume regulation or metabolic impairment.

The mechanism of electrogenic K⁺ secretion across lepidopteran midgut is not well understood. Several workers have proposed that electrogenic K⁺ secretion occurs via an apically located K⁺-ATPase (reviewed by Dow, 1986). The evidence for this model is based upon measurements of the K⁺ electrochemical gradient across the apical membranes and the isolation of the K⁺-stimulated ATPase from goblet cell membranes (Wieczorek et al. 1986). Recently, however, evidence has been put forward to support a different model of K⁺ secretion. Schweikl et al. (1989) and Wieczorek et al. (1989) have shown that the midgut cells have a H⁺-ATPase which secretes hydrogen into the goblet cell lumen. The H⁺ is then exchanged for K⁺. The function of the H⁺-ATPase is not to acidify the lumen, but to generate a large electrical gradient across the apical membrane which will drive H⁺ absorption and K⁺ secretion via the H⁺/K⁺ exchanger. Assuming an intracellular pH of 7.2, under the conditions of previous studies (saline pH=8.3; see Cioffi and Harvey, 1981) only the electrical component of the electrochemical gradient would drive the H⁺/K⁺ exchanger. Therefore, the current generator (the H⁺-ATPase) would determine the rate of K⁺ secretion. In this study, the saline was adjusted to haemolymph pH of 6.7. This pH is lower than the assumed intracellular pH and thus there would be a chemical as well as an electrical driving force for the H⁺/K⁺ exchanger. Under these conditions, the rate of K⁺ secretion via the H⁺/K⁺ exchanger may exceed that of the electrogenic H⁺-ATPase.

In the study by Wieczorek et al. (1989) the presence of anions stimulates apical membrane H⁺ secretion, but dissipates the membrane potential. Thus, an anion conductance must exist on the apical membrane. In support of this view, Cl⁻ is secreted by all regions of the midgut. Omission of Cl⁻ causes a significant drop in Zsc. This may be due to inhibition of apical anion conductance and therefore limit the secretion of H⁺ by the apical H⁺ pump. Alternatively, the removal of Cl⁻ may inhibit entry of K⁺via basal KC1 cotransporters (Zeiske and Schroder, 1988) and thus suppress the I_sc. It should be noted, however, that introduction of Cl⁻ to tissue bathed in Cl⁻-free saline did not stimulate I_sc. This may indicate nonspecific damage caused by long-term Cl⁻ deletion, as has been described in renal epithelia (Chamberlin et al. 1984).

Recently Chao et al. (1989) have demonstrated Cl⁻ absorption across the posterior midgut of M. sexta. This observation contrasts with the observations made in the present study. The conditions of the two studies differ in the composition of saline and size of larvae used and these differences may account for the conflicting observations. Interestingly, Chao et al. (1989) postulate that there may be more than one transepithelial Cl⁻ transport system and speculate that Cl⁻ secretion may also occur. The conditions of the present study appear to have been appropriate for revealing this secretory process. This secretion of Cl⁻ may account for the higher luminal Cl⁻ concentration in the posterior versus anterior sections of the midgut observed in vivo (Dow et al. 1984).

Dow et al. (1984) failed to detect intracellular levels of Na⁺ in the midgut of M. sexta, yet there is Na⁺ in the midgut contents and haemolymph. Since there is a favourable electrochemical gradient for Na⁺ entry from either the haemolymph or lumen (see Moffett and Koch, 1988a,b;Dow et al. 1984) and since it is unlikely that the membranes are absolutely impermeable to Na⁺, it would seem plausible that Na⁺ is transported by the midgut. This hypothesis was confirmed in this study in which all three midgut sections actively transported Na⁺. The was negligible, indicating that Na⁺ permeability was low and edge damage was minimal. Sodium may be cotransported across the apical membranes with amino acids since it has been shown that Na⁺ gradients can energize amino acid transport in membrane vesicles isolated from the midgut of M. sexta (Hennigan and Wolfersberger, 1989) as well as that of another caterpillar, Philosamia cynthia (Hanozet et al. 1984). It is not clear, however, what mediates Na⁺ efflux, although a Na⁺ pump seems unlikely since efforts to isolate a Na⁺/K⁺-ATPase from this tissue have been unsuccessful (Jungreis and Vaughan, 1977).

It is clear that Na⁺ transport by M. sexta midgut differs from that of Hyalophora cecropia midgut. In H. cecropia net Na⁺ secretion has been demonstrated and complete replacement of K⁺ with Na⁺ can maintain a substantial (Zerahn, 1971). This is clearly not the case for M. sexta. Moffett (1980) showed that complete substitution of K⁺ with Na⁺ results in a negligible I_sc. In this study, the removal of K⁺ led to a precipitous fall in I_sc, although 17.6 mmol l⁻¹ Na⁺ was present. Although there is net absorption of Na⁺ by M. sexta midgut, omission of Na⁺ from the saline inhibited the I_sc. This may reflect metabolic inhibition due to impaired substrate uptake. Alternatively, the presence of Na⁺ may stimulate K⁺ entry and I_sc, as has been suggested by Thomas and May (1984).

Under open-circuit conditions, net K⁺ secretion is not balanced by Na⁺ absorption and Cl⁻ secretion (see Fig. 8). Charge neutrality may be kept by secretion of anions such as HC0₃⁻, phosphate, citrate or succinate. In addition, absorption of cations such as basic amino acids, H⁺, Mg²⁺ or Ca²⁺ could maintain charge balance. The presence of carbonic anhydrase in midgut tissue (Ridgway and Moffett, 1986), high luminal pH in vivo (Dow, 1984) and luminal alkalinization in vitro (Chamberlin, 1990) indicate that HCO₃⁻ secretion or H⁺ absorption is plausible. The possibility of polyvalent or organic cation absorption is strengthened by the observations that these processes occur in other lepidopteran larvae (Hanozet et al. 1989; Wood and Harvey, 1976; Wood et al. 1975). The midgut of M. sexta actively secretes anionic dyes (Nijhout, 1975) and therefore secretion of polyvalent organic anions present in the haemolymph (e.g. succinate, citrate) is conceivable. Elucidation of these transport processes in this species awaits further experimentation.

I wish to thank Dr J. Thomason for writing the computer program used for integration in this study. I am grateful to J. Martin and R. B. Thomson for their critical reading of the manuscript. This work was supported by an Ohio University Research Challenge Grant and funds from the Ohio University College of Arts and Sciences and College of Osteopathic Medicine.