The Electrical Properties of the Anterior Stomach of the Larval Mosquito (Aedes Aegypti)

The ion-transport properties of the insect midgut are critically important for maintaining the appropriate ionic concentrations and pH for digestion and absorption. The pH of the midgut also plays an important role in determining susceptibility to biological control agents such as Bacillus thuringienis (Knowles, 1994). Despite the importance of this system, most of our information about insect midgut ion transport is from studies on the apparently unique midgut of lepidopteran larvae (for reviews, see Dow, 1986; Clements, 1992).

The anterior stomach region of the larval mosquito is known to be an alkalizing organ and apparently regulates the alkalizing process via a neural or hormonal mechanism (Dadd, 1975, 1976). Recently, cDNA clones of two V-ATPase subunits have been isolated from the midgut of Aedes aegypti larvae (Gill et al., 1998), and a high level of expression of one of these subunits was observed in anterior midgut (Filippova et al., 1998). However, the relationship between the expression of V-ATPase and transepithelial ion transport is unclear.

In an initial study of the ion-transport properties of the larval mosquito midgut, Clark et al. (1999) established that the anterior stomach of Aedes aegypti generates a lumen-negative transepithelial potential that is relatively large in magnitude (approximately 60 mV) immediately after dissection, but which rapidly decays to −10 mV or less. At this time, the biogenic amine 5-hydroxytryptamine hyperpolarized the tissue to a value approximately 50 % of its initial value when applied at concentrations previously reported to be present in hemolymph (Clark and Bradley, 1997; Clark et al., 1999). This lumen-negative potential stands in marked contrast to the lumen-positive potential across the larval lepidopteran midgut (Klein et al., 1996) and appears to be driven by different primary motive forces. Other mechanisms, such as those responsible for the apparently homologous ability to generate extremely alkaline conditions, seem to be shared between the two groups (Clark, 1999).

Aedes aegypti (Vero Beach strain) eggs were provided by Dr Marc Klowden, University of Idaho, USA, from a continuously maintained colony. Eggs were hatched in deionized water, and larvae were maintained in deionized water at 26 °C on a 16 h:8 h L:D photoperiod. The water was replaced each morning, and the larvae were fed at that time with ground Tetramin flakes (Tetrawerke, Melle, Germany).

The saline used in all experiments was based on the description by Edwards (1982a,b) of larval Aedes hemolymph composition and consisted of the following (in mmol l⁻¹): NaCl, 42.5; KCl, 3.0; MgSO₄, 0.6; CaCl₂, 5.0; NaHCO₃, 5.0; succinic acid, 5.0, malic acid, 5.0; L-proline, 5.0; L-glutamine, 9.1; L-histidine, 8.7; L-arginine, 3.3; dextrose, 10.0; Hepes, 25. pH was adjusted to 7.0 with NaOH.

Perfusion pipettes were made from glass capillary pipettes (100 μl, VWR). An initial pull on a David Kopf Instruments vertical pipette puller (model 700B) was followed by hand-forging to form a pipette with a bulbous tip approximately 200 μm in diameter and 400 μm in length that could be inserted into the gut lumen and tied in place using a fine human hair.

A Harvard Apparatus multi-speed transmission infusion/ withdrawal perfusion pump (model 902) was modified for simultaneous infusion and withdrawal, allowing perfusion of the gut section with a negligible distending pressure gradient across the epithelium. Micromanipulators (Brinkmann) were used to hold and manipulate the infusion and withdrawal pipettes. Sections of the anterior stomach were isolated in a 1 ml bath. The withdrawal-side pipette was then inserted into the lumen and tied in place. The infusion-side pipette was inserted into the other end of the gut section and also tied. Gut sections that showed a sudden drop in transepithelial potential (TEP) as the perfusion/voltage-sensing pipette was tied, and those with obvious leaks, were rejected. The perfusion pump and bath flow were then started. The bath was perfused by gravity at a rate of approximately 15 ml h⁻¹ with oxygen-saturated saline (100 % O₂), while the lumen of the isolated gut section was perfused from anterior to posterior with identical but non-oxygenated saline at a rate of 62.3 μl h⁻¹.

A Ag/AgCl electrode was inserted into the perfusion pipette to within 5 mm of the tip. In later experiments, a second Ag/AgCl electrode was inserted into the outflow pipette. Conventional microelectrodes were pulled from Omega dot filament glass (1.5 mm o.d., 0.75 mm i.d.; Frederick Haer), filled with 500 mmol l⁻¹ KCl and connected to Ag/AgCl electrodes. Tip resistances were not determined. The reference potential for these measurements was the bathing solution, which was connected via a KCl/agar bridge to a Ag/AgCl electrode.

In early experiments, voltage deflection after current injection at the input pipette was compared with that at the microelectrode. In later experiments, the voltage deflection at the outflow pipette was also measured.

Current was injected briefly through the perfusion-side electrode, and voltage deflections were recorded at the injection pipette and in an intracellular or intraluminal microelectrode. In later experiments, the voltage deflection at the outflow pipette was also recorded. The direction of current flow was then reversed and the injection repeated.

In some experiments, after successful luminal or intracellular positioning of the microelectrode and initial measurements, bath flow was stopped, and 5-hydroxytryptamine (5-HT) added to the bath and mixed to a final concentration of 10⁻⁴ mmol l⁻¹.

Data were acquired through high-impedance amplifiers using data acquisition software (Sable Systems, Las Vegas, NV, USA). Data were sampled at 0.11 s intervals, with each sample an average of 112 readings. Results are reported as means ± S.E.M. followed by the number of experiments in parentheses. Mean values were compared using a two-sided Student’s t-test. All times are measured from the time the gut was cannulated.

Theory

Measurement of the electrical parameters of a tubular organ using cable properties has been discussed previously including in the work of Boulpaep and Sakin (1979), Leyssens et al. (1992) and, of course, Hodgkin and Rushton (1946). We shall go over it briefly here.

The possible pathways of current flow in the gut in these experiments as well as the symbols used are shown in Fig. 1A. Longitudinal current enters the lumen and either proceeds along the lumen or crosses the tubular wall to the outside solution. By the downstream end, all the current must have crossed the walls. Since the tubular walls are made up of cells with tight junctions, some of the radial current may conceivably travel longitudinally from cell to cell. Most of the radial current, however, may be expected to pass through the basal cell membrane to the outside bathing solution. The equations for this system are:

Preliminary computations indicated that, even if the entire lateral cell surface allowed free conduction, the results would be very little different from the simplified case. The simplified case were therefore used for the treatment of all data. The data provide a check on whether the simplified case can be used. If the value of the space constant λ obtained from luminal punctures is the same as that derived from cellular punctures, then there is no measurable longitudinal current flow through the cells.

Preliminary experiments showed that repeated punctures led to progressively lower resistance values, suggesting that the gut did not seal well after removal of the microelectrode. Hence, each tissue was impaled only once. It was impaled soon after cannulation, and the electrode was left then in place so that the voltage and voltage responses could be followed over time.

In later experiments, the voltage deflection at the outflow pipette was also measured. There were two criteria to be met for an acceptable reading of luminal voltage. The first was that all three readings, at the input perfusion pipette, at the outflow perfusion pipette and at the microelectrode, were all within 2 mV of one another whenever current was not passed. The second criterion was that the voltage deflection from the microelectrode should be smaller than that from the input pipette and greater than that from the outflow pipette. It became clear during these experiments that the voltage deflection from a cellular penetration was much lower than that from a luminal penetration, normally less than 20 % of it. Hence, the criteria used for the early experiments were (i) identity of the perfusion pipette and microelectrode voltages at rest and (ii) that the microelectrode voltage deflection be at least 20 % of that at the perfusion pipette. Twenty-one experiments fulfilled these criteria and were used for the determination of the space constant, λ.

The resistivity of the solution used both for perfusing and bathing the tissue was 43.1 Ω cm. The mean distance between the inflow and outflow pipettes (W in Fig. 1) was 733±41.9 μm (N=21). The distance beyond the outflow pipette and the tie (Y in Fig. 1) was estimated to average 367 μm. The external diameter, as determined by optical measurements, was 287±5.1 μm (N=22). The tissues are usually translucent upon mounting, and it is possible obtain an estimate of cell height from the differences in light transmission. Height, measured this way, was approximately 45 μm. Hence, luminal diameter was estimated to be 197 μm. Microscopic analysis of fixed slides cut perpendicular to the long axis of the gut gave a value of cell height of 63±0.68 μm (N=166). If this value is used, luminal diameter (a) is computed to be 162 μm.

As reported previously, the transepithelial potential (TEP) decays with time (Clark et al., 1999). There is an early phase during which TEP falls rapidly (becomes less negative). This ends after approximately 5 min and is followed by a second phase during which depolarization is much slower (see Fig. 2). When tissues were cannulated and perfused rapidly, the input impedance (as measured from the input perfusion pipette) of the gut followed a biphasic response. It was initially high and fell during the early period when the voltage was falling rapidly. Thereafter, it rose. Of the 21 experiments reported here, 14 provided readings within the first 5 min after cannulation. We detected the initial fall in nine of these. All tissues showed a subsequent rise in input impedance during the time that TEP was either falling slowly or constant. Since we do not know the total current in the region between the pipette tips (see above), we cannot convert the voltage deflection to a resistance. However, in any preparation, the fraction of the current that flowed out beyond the pipette tips, in the regions labeled Z and Y in Fig. 1A, would be expected to remain constant. Hence, changes in the voltage deflection seen at the input pipette reflect changes in the input impedance. We shall use the term ‘input pipette impedance measure’ for this voltage deflection. Fig. 2 shows the changes in potential and impedance in a typical experiment.

All measurements were taken between 23 and 28 min. The space constant, λ, was estimated in each experiment from the ratio of the voltage deflection of the microelectrode to the voltage deflection at the input perfusion pipette. Each experimental ratio gave a value of λ from equation 6. These values were then averaged to give a mean value of λ of 936±84.5 μm (N=21). Fig. 3 depicts the values of the voltage deflection ratio as a function of distance from the input perfusion pipette. The curve is the voltage decrement computed from the mean value of λ. It should be noted that this is the mean of the space constants, not the mean of the individual points.

Two populations of cells were detected. One group had stable basal potentials, whereas in the second, the basal potential decayed along a time course similar to that of the TEP.

The usual criteria for an acceptable cell puncture were used; entry had to be sharp, the voltage had to remain stable for at least 1 min and had to return to within 2 mV of the baseline value upon removal. In addition, the voltage had to be different from the luminal voltage. One hundred successful impalements were recorded. All showed a basal resistance that was much lower than observed in the luminal puncture experiments. The mean value of the basal voltage (V_b) was 66.1±1.76 mV (N=100). With the exception of a single cell that had a stable potential of −139 mV, all appeared to belong to a single population. Fig. 4 shows the distribution of basal voltages in this group of cells. Fig. 5 shows that these cells were stable not only for an individual puncture but that V_b stayed constant as the TEP decayed from its initial value of approximately −60 mV to near zero. Fig. 6 shows that these cells were distributed throughout the gut with no apparent localization.

Basal voltage deflection, in response to an injected current (basal voltage divider), although uniformly low, varied slightly as a function of distance along the gut. Fig. 7 shows that this measure of basal resistance can be fitted adequately with a curve that is 7 % of the luminal voltage decrement.

Criteria for an acceptable puncture of this population of cells included sharp entry and return of the voltage to within 2 mV of the value measured in the bath upon removal. The voltage at early times was required to be more negative than the simultaneously measured luminal voltage. Finally, the voltage deflection upon current injection was required to be less than 20 % of the deflection at the input pipette. Fig. 8 illustrates the puncture of a decaying cell and compares the time course of this decay of V_b with that of the TEP. In Fig. 8, the voltage deflections from current injection in the input pipette recording have been removed to allow a better comparison of the two voltages. Fig. 8 also illustrates a frequent characteristic of the decaying cells; late in time, V_b is often less negative than is TEP.

Fig. 9 depicts the frequency distribution of the earliest values of V_b obtained in these decaying cells. All values were collected within 4 min of cannulation. The mean time after cannulation at which these voltages were collected was 1.75±0.10 min (N=45). At that time, the mean basal membrane voltage was −100±3.2 mV (N=45) and the mean TEP was −38.8±3.3 mV (N=45). The high voltage values of this group are clearly characteristic of these cells. Because of the rapid changes in TEP and V_b, it is not clear whether the cluster exhibiting lower voltages (between −87 and −55 mV) are accurate measurements within a single voltage distribution, whether they represent a second population of decaying cells or whether they simply come from cells already showing significant decay. Other than their initial voltage, this group of lower-voltage decaying cells did not differ in any way from the group exhibiting higher initial voltages. The value of V_b given, of −100 mV, is the mean of all the decaying cells. The decaying cells were distributed uniformly throughout the anterior stomach (Fig. 10).

At late times after cannulation, V_b of the decaying cells was characteristically slightly less negative to the hemolymph than was the lumen. Of the 19 punctures for which data at late times were obtained, 15 showed this phenomenon. The time of the late collections averaged 15.35±1.68 min, at which time TEP averaged −10.23±2.26 mV and V_b averaged −9.28±2.03 mV.

The basal voltage divider ratio was indistinguishable from that of the stable cells. It was uniformly low and decreased slightly with increasing distance along the gut. The same curve that described basal voltage deflection in the stable cells, 7 % of λ, also described basal voltage deflection of this population (Fig. 11).

5-HT (10⁻⁴ mmol l⁻¹) caused an immediate hyperpolarization in stable cells (N=13). Prior to administration, V_b averaged −83.0±6.8 mV. Within 3 min of administration, V_b averaged −104±9.63 mV, an average percentage change of 24.4±2.9 %. These values are statistically different (P<0.001). At that time, there was no significant effect either on input pipette impedance measure (mean input impedance measure 114±16.3, N=11, before 5-HT and 113±18.0, N=11, after 5-HT) or on the voltage divider ratio (mean voltage divider ratio 0.076±0.017, N=10, before 5-HT and 0.0809±0.024, N=9, after 5-HT). There was a small increase in TEP; it increased from −4.95±1.06 mV (N=12) before to −8.04±2.07 mV (N=12) after 5-HT treatment. These values differ statistically (P<0.001). The full response of the cells to 5-HT was evident by 3 min; the mediator was washed off after 5 min. Interestingly, Clark et al. (1999) noted that the transepithelial hyperpolarization reached a maximum 15 min after application of the agent. Fig. 12 illustrates the response of a stable cell to 5-HT. There was no change in input impedance or voltage divider ratio after 5-HT administration. 5-HT had no effect on V_b in decaying cells (N=9), regardless of the time of application, although TEP increased in these cells as well.

The present work corroborates the previous work from this laboratory showing that the anterior stomach of the mosquito exhibits an inside-negative voltage that rapidly decays upon isolation. The previous publication also showed that administration of 5-HT produced a transepithelial hyperpolarization (Clark et al., 1999). The present work extends our understanding of the electrical properties of this epithelium and its cells.

Theoretical calculations of the complete model shown in Fig. 1A suggested that there was no reasonable possibility of appreciable longitudinal current between cells. The data collected here corroborate that statement. The finding that the voltage divider ratio was proportional to the space constant in all cells indicates that, in each cell, the same current density passed through the apical and basal membranes. This leads to the simplified model of Fig. 1B.

The value of Y, the distance beyond the outflow pipette and the tie, was crudely estimated to be approximately 370 μm. No systematic study of this was made, but the value of Y was within 20 % of that estimate in all cases. The value of Y does have an effect on that calculated space constant, but this level of uncertainty is trivial. A 20 % error in the value of Y would have led to an error of only 3 % in the computed value of λ.

Cell impalements that decay are obviously suspect. We have two reasons to believe that these are accurate measurements of undamaged cells. First, they are seen only in the first few minutes after a tissue has been isolated. If the decay of V_b was because of damage, we would expect to damage cells later on in the experiment too. Second, the time course of the decay of V_b in these cells closely follows the time course of both the initial fall in input pipette impedance measure and the initial rapid fall in TEP. Since all the decaying cells had similar basal voltage divider ratios and a similar time course of decay, we believe that they were all samples from the same population and that the results were collected at slightly different times after isolation. Finally, the decaying cells did not respond to 5-HT, regardless of their initial V_b.

The space constant is 9.36×10⁻² cm and ρ is 43.1 Ω cm. If we use the visual estimates of internal radius (a) of 98 μm and of cell height (b) of 45 μm, the estimate of the combined wall resistance r_m, as computed from λ₂=ar_m/2ρ, is 73.8 Ω cm². If we use the cell height from microscopic analysis of 63 μm, then we compute an internal radius of 81 μm and the estimate of r_m is 82.2 Ω cm². We regard these two estimates as in reasonably good agreement and will use the average values of a=93 μm, b=53 μm and r_m=78 Ω cm² in subsequent calculations. We need to know the paracellular resistance r_p to calculate the apical and basal membrane resistances. We have no information on its value.

We shall make the calculation on the assumption that r_p=∞. The error made this way is that the values that we calculate for r_a and r_b, the resistance of the tubule lumen and the cells, respectively, will be too low by an amount depending on the true value of r_p. The ratio of these values will, however, be accurate. Two equations describe the two unknowns, r_a and r_b. From equation 5, we have: 78=r_a+(93r_b/146). From the voltage divider data, we obtain the relationship: (r_b/146)/(r_a/93)=0.07. Simultaneous solution of these equations gives r_a=72.9 Ω cm² and r_b=8.0 Ω cm².

These are the values at long times after isolation. Earlier in time, we believe that r_a is somewhat lower and that r_b is appreciably higher. It was not possible to obtain a satisfactory estimate of λ in the first minutes. This is probably because resistances were changing quite rapidly and at somewhat different rates in different preparations. However, since the input pipette impedance measure is nearly the same at very late times as it is upon cannulation, the value of λ would be similar at the two times.

Further studies of the transport physiology of the stomach would be greatly facilitated by an ability to voltage-clamp the tissue. The present study shows that acceptably uniform clamping could not be accomplished by injection of current through the perfusion pipette. Instead, an axial internal electrode will be required to achieve uniform current density throughout the length of the perfused gut.

A significant question for epithelial transport is whether this is a tight or a leaky tissue. There are at least three different ways of defining the degree of tightness. The first and simplest would be simply to compare r_m with that of other tissues. A second way, suggested by Guggino et al. (1982), would be to determine whether the resistance of the paracellular pathway is lower than that of the transcellular pathway. A low-resistance paracellular pathway is the defining characteristic of a leaky epithelium in this view. A third way would be to examine the ratio of λ to the length of the tubule. This measure defines whether a voltage change in one region of the tubule can have an effect in other regions. Again, a low ratio would define a leaky epithelium.

We shall use mammalian renal proximal tubule as an archetype for a leaky epithelium and mammalian renal distal tubule as an archetype for a tight epithelium. Typical values of r_m in renal proximal tubule are approximately 5 Ω cm² and for renal distal tubule are 600 Ω cm² (Boulpaep and Seely, 1971). In these terms, the mosquito anterior stomach appears to be intermediate.

We cannot directly apply the criterion suggested by Guggino et al. (1992) since we did not measure the paracellular resistance. However, the data suggest that a significant fraction of the injected current did pass through the cells. In a similar study of ant Malpighian tubules, Leyssens et al. (1992) did not detect any deflection in V_b upon current injection. They concluded that that tissue was a leaky epithelium. We did observe a deflection in V_b and, hence, current did pass through the cells. According to this criterion, we would decide that mosquito anterior stomach is probably a tight epithelium.

To use the third discriminator, we must know the length of the tubule in vivo. We shall estimate the lengths of the reference tubules from figures given for humans (1.4 cm for proximal convoluted tubule and 0.5 cm for distal convoluted tubule; Bloom and Fawcett, 1968) and divide by 2.6, a scaling factor derived from the relative glomerular filtration rates of human and dog kidney. The estimates of length so derived are 0.54 cm and 0.19 cm for canine proximal and distal convoluted tubules respectively. The space constant of canine proximal tubule is 58 μm, and the ratio of space constant to length is 58/5400=0.0108. The space constant of distal convoluted tubule is 426 μm, and the ratio of space constant to length is 426/1900=0.224. Both space constants were calculated from the data of Boulpaep and Seely (1971). These estimates are all somewhat imprecise but they might be expected to be approximately correct. The estimates of the ratio of λ to length differ by a factor of more than 20 between the two archetypal epithelial types. For the mosquito anterior stomach, the space constant is 936 μm and the total length is approximately 1600 μm. The ratio of λ to length is therefore 936/1600=0.59. According to this criterion, the mosquito anterior stomach is clearly a tight epithelium. The finding that there was detectable current passing through the cells and the finding of a large ratio of space constant to organ length led us to conclude that this organ is a tight epithelium.

This conclusion seems consistent with what we know of the function of the anterior stomach. Although fat storage, presumably leading to fat absorption, has been observed (Wigglesworth, 1942), the only documented transepithelial movement is alkalization of the gut contents (Clements, 1992). A tight epithelium would seem necessary for the establishment and maintenance of the high pH gradient across the anterior stomach.

The data allow us to construct voltage profiles of both types of cell and to look at how they change both over time and as a result of 5-HT application. Fig. 13 shows these profiles. Immediately after isolation, TEP is approximately −60 mV. The decaying cells have a basal transmembrane potential V_b of −100 mV and an apical transmembrane potential V_a of −40 mV.

Within 5 min, V_b collapses to a few millivolts and the potentials of the lumen, cell fluid and bathing solution are all within a few millivolts of each other. Later, TEP may be more negative than V_b. The stable cells start with a V_b of −65 mV and a V_a of only −5 mV. Soon (5 min) after isolation, V_a starts to rise so that, 25 min after isolation, V_b is still −65 mV, but V_a is now −60 mV. At that time, TEP is −5 mV. The effect of 5-HT is to cause an immediate hyperpolarization of the basal membrane of the stable cells.

We do not have sufficient information to describe the processes that lead to these complex changes. We can, however, construct a scenario that is consistent with these changes. Upon isolation, the basal membranes of the decaying cell lose a trophic hormone. They immediately become very leaky. This accounts for the collapse of V_b in these cells, the collapse of TEP and the initial reduction of input impedance measure. Somewhat later, starting 4 or 5 min after isolation, the apical membranes of the stable cells become less conductive. Since there is no change in basal pump activity, V_b of these cells stays constant, but V_a rises. Since TEP is some sort of weighted average of the transepithelial voltage from both cell types, there is a further slow reduction in TEP, this time accompanied by a parallel increase in input impedance measure. This accounts for the second, slower phase of depolarization of TEP. The reversal of polarity of V_a in the decaying cells late in time is surely a peculiar finding. It could be a reflection of the reversal of net flux in an electrogenic antiport situated in the apical membrane. Finally, 5-HT enhances the activity of an electrogenic pump located in the basal membrane of the stable cells. Again, since TEP is an average of the activity of these two cell types, the response of V_b in the stable cells is more pronounced than is the response of TEP.

In summary, the present study shows that the anterior stomach of the mosquito midgut epithelium is a tight epithelial tissue. It contains two electrically distinct cell types. Changes in the properties of one cell type are associated with the decay of the transepithelial potential after isolation. Although, from a macroscopic viewpoint, 5-HT opposes the decline in transepithelial potential, its action is manifest in the stable cells rather than the decaying cells. This suggests that the 5-HT present in the hemolymph of the insect is not the sole factor that sustains the transport activities of the gut in vivo.

This research was supported by grants NRICGP 97-35302-4919 and NRICGP 99-35302-8371 to T.M.C. The authors would like to thank Mark Klowden and Gail DeSantis for eggs and Joan Folwell for preparation of the histological slides.