ABSTRACT
Measurements of electrical potential difference (PD), short-circuit current (SCC) and unidirectional fluxes of sodium and chloride were made across portions of the intestine. Based on the results, the intestine can be divided into at least four physiologically distinct regions.
These four physiological regions, designated from anterior to posterior as R I–fl, R III A, R III B and R IV, do not completely correspond to the four anatomically distinct regions of the intestine.
The PD (serosal side positive) in R I–II, R III A, R III B and R IV is 1·08, 12·4, 5·61 and 31·7 mV, respectively.
The SCC in these same regions is 9·9, 50·4, 49·7, and 16·4μA cm2, respectively.
When short-circuited, net sodium and net chloride fluxes in the above regions are –0·36 and –0·27, 1·46*** and –0·92*, 1·74*** and –0·06 and 0·01*** and 0·07 μmol cm-2 h-1, respectively. Positive fluxes indicate net mucosal to serosal movements and asterisks indicate significant net fluxes (* P < 0·05, *** P < 0·001).
There is good agreement between the SCC and net sodium transport in R III B. In the other regions of the intestine the ionic basis of the SCC has not been completely explained.
The properties of the intestine in vitro appear to make the intestine well suited for the task of conserving sodium, a function which the intestine performs in vivo.
INTRODUCTION
In Lumbricus terrestris there are potentially three major sites of salt and water exchange: the integument, the nephridia and the gut. Some of the properties of the integument have been studied in the intact animal (Kamemoto, 1964; Dietz, 1974; Kirschner, Greenwald & Kerstetter, 1973; Carley, 1975) and in isolated preparations (Tercafs, 1965; Prusch & Otter, 1977). The nephridia have been studied with micropuncture techniques (see review by Zerbst-Boroffka & Haupt, 1975). The properties of the gut, however, remain poorly understood in Lumbricus and other annelids. Some evidence for the involvement of the gut in salt and water balance in annelids has been summarized by Oglesby (1978). There have been few prior attempts to study the gut in vitro and most of our knowledge of it comes from whole animal studies. In Lumbricus, the most compelling evidence of the importance of the-gut in salt and water balance is the difference in the sodium concentration in the anterior and posterior parts of the gut. In pond-water-adapted Lumbricus, Dietz & Alvarado (1970) found that the sodium concentration in the crop and rectal fluids were 60 and 6 mM, respectively. Thus, it would appear that the crop produces an isosmotic sodium solution and, as this solution passes through the gut, sodium is reabsorbed so that a dilute sodium solution is eliminated. Some similarities between the roles of the gut and the nephridia for animals in pond-water have been suggested by Dietz & Alvarado.
In the present study, sodium transport and chloride transport have been examined in isolated portions of the intestine of animals adapted to pond-water. The results indicate that the intestine is differentiated into at least four physiologically distinct regions which are organized in a manner similar to that found in many vertebrates. The results support the idea that the intestine plays an important role in sodium conservation and they are consistent with the results found in the whole animal. A preliminary report of this research has been presented (Cornell, 1979).
MATERIALS AND METHODS
Specimens of Lumbricus terrestris were obtained from The College Biological Supply Co., Bothell, Washington, U.S.A., or collected in Pullman, Washington, U.S.A. Animals were maintained at room temperature (18-25 °C) in the laboratory in artificial pond-water (see Dietz & Alvarado, 1970) for at least a week prior to their use. The results reported here were obtained at 23–25 °C during the months of July to January, inclusive.
Portions of the intestine were dissected from animals submerged in ice-chilled saline. In the anterior regions of the intestine, where the typhlosole is present, a shallow incision was made with fine scissors in the sagittal plane along the dorsal surface of the intestine. This operation eliminated the internal ridge formed by the typhlosole. A second incision along the midline of the newly created dorsal surface allowed the intestine to be opened and laid back as a flat sheet (Fig. 1A). In the posterior region of the intestine, which lacks the typhlosole, a single incision along the dorsal surface allowed the intestine to be opened.
The tissue was mounted between small, two-part lucite holders with rounded, elongated apertures of from 0·144 to 0·196 cm2 (Fig. 2). The technique for mounting the tissue was designed to minimize edge-damage. First, a uniform layer of silicone grease was applied to the inner surfaces of the two parts of the holder. The tissue was then mounted on the six locating pins of the first part of the holder and the second part of the holder was positioned and held to the first with two finely threaded screws (56 threads/in.). The screws were adjusted so that the tissue, which was visible through the lucite holder, was just barely compressed by the silicone grease. The screws allowed adjustments in the distance between the two parts of the holder of less than 50 μm, a distance which was about a quarter of the combined thickness of the silicone grease. Thus, the mechanical strains on the tissue were taken by the locating pins and the silicone grease, rather than by the edge of the aperture. The tissue was kept under physiological saline during mounting and all of the above procedures were performed with the aid of a dissecting microscope (10× ). Once the tissue was mounted in the holder it was unaffected by the assembly of the chamber. The first part of the holder, being of greater diameter than the chamber openings, was firmly clamped between the half-chambers. The second part of the holder, being of smaller diameter than the chamber openings, remained free inside one of the halfchambers. Thus, the mounting of the tissue was independent of the assembly of the chambers and the initial adjustment of the holder was undisturbed during the remaining part of the experiment. When tissue was removed from a holder after 4 or 5 h, no markings caused by the aperture were visible. In all experiments involving the anterior regions of the intestine, the mucosa covering the typhlosole was not included within the aperture.
Each chamber half contained 1·9 ml of physiological saline, which was oxygenated and circulated by a ‘bubble lift’ (Fig. 2). Periodically, 0·5 ml samples were taken from the ‘cold’ side to determine the amount of isotope present. Prior to sampling, the fluid levels in both chamber halves were lowered simultaneously by withdrawing fluid into two separate reservoirs (not shown in Fig. 2). The actual sample was taken from one of the reservoirs, the fluid replaced, and then the contents of the reservoirs were emptied back into the half-chambers. This procedure allowed relatively large fluid samples to be taken without subjecting the tissue to a hydrostatic pressure difference.
Physiological saline, partly based on the ionic composition of Lumbricus blood (Boroffka, 1965), contained the following: Na+, 71 HIM; Cl-, 74 HIM; K+, 4·0 mM ; Ca2+, 3·0 mM; Mg2+, 0·35 mM; , 2·0 mM; , 0·43 mM; dextrose, 10 mM. This solution was buffered to pH 7 ·3 with 5 mM MOPS (morpholinopropanesulphonic acid; Cal. Biochem), and gassed with moist air. In some experiments a bicarbonate-based solution was used. This contained 25 mM bicarbonate, which partly replaced the chloride in the above solution. The bicarbonate saline was gassed with a moist mixture of 95 % O2 and 5 % CO2. In all experiments, similar solutions were used on both sides of the chamber.
Unidirectional fluxes of Na+ and Cl- were measured with 22Na and 38C1 (New England Nuclear) at concentrations of 0·5–5·0 μCi/ml of physiological saline. In experiments involving only 22Na, count rates were determined with either a liquid scintillation counter (Nuclear Chicago, system 722) or a gas flow, beta counter (Nuclear Chicago, model D-47). In experiments involving both of these isotopes, 22Na was counted on a gamma counter (Nuclear Chicago, model 1810) and 22Na plus 36C1 were counted together on the above liquid scintillation counter, 36C1 being determined by difference. Flux rates of Na+ and Cl- were calculated after accounting for the amounts of these isotopes lost from the chamber due to sampling.
Voltage measurements were made with a pair of calomel electrodes connected as follows: calomel electrode M-KCI: 1·0 M-NaCl-agar bridge : physiological saline, the last connexion being made at a distance of 2-3 mm from the tissue. Electrode asymmetries were balanced to within 01 mV after the bridges were allowed to equilibrate for about 30 min in a common solution of physiological saline. Currentpassing electrodes consisted of a pair of chlorided silver wires connected as follows: chlorided silver: 10 M-NaCl: 1·0 M-NaCl-agar bridge : physiological saline, the last connexion being made at the apex of the conical interior of the half-chamber, which was about 15 mm from the tissue. The current pathways were not obstructed by the voltage electrodes. Amplifier-voltage-clamps, built by the author, allowed the voltage and short-circuit current from two preparations to be multiplexed on to the single channel of a Varian model G-10 chart recorder. These signals were also monitored with a Tektronix model 5110 oscilloscope. The voltage-clamp allowed for automatic correction of the clamping error caused by the resistance of the saline. The technique by which this was accomplished is similar in principle to that described by Wood & Moreton (1978); however, the ‘third electrode’ is inside the electrical box, rather than in the chamber. To correct for solution resistance the IR drop across a calibrated tenturn resistor, which is set equal to the solution resistance, is determined by the automatic voltage-clamp. Since all of the current passing through the voltage-clamp must pass through the ten-turn resistor, the voltage across this resistor is equal to the voltage error caused by the solution resistance. The error is corrected automatically at the error amplifier, an operational amplifier configured as a summing amplifier, which is balanced when electrode voltage (Transepithelial Potential Difference plus the IR drop across the solution) – correction voltage (IR drop across the ten-turn resistor) = clamp command voltage, which is zero when short-circuiting.
The sodium concentrations in the coelomic fluid and the luminal fluids in various regions of the gut were measured. Coelomic fluid was obtained by puncturing the body wall with a drawn-out pipette. Luminal fluids were obtained after the body wall was carefully opened. The exposed gut was covered with mineral oil and the fluid was removed by puncturing the gut with a drawn-out pipette. Sodium concentration was determined with a Perkin-Elmer atomic absorption spectrophotometer using standard procedures which have been previously described (Cornell, 1980 a).
RESULTS
Regions of the intestine
The intestine of Lumbricus can be divided into four anatomically distinct regions which are readily distinguishable with the aid of a dissecting microscope (see Arthur, 1963). For present purposes these regions will be designated, from anterior to posterior, as R I to R IV. The intestine can also be divided into four regions on the basis of the physiological results in this paper, but these regions differ somewhat from the anatomical regions. The physiological regions have been expressed in terms of the anatomical regions and are here designated, from anterior to posterior, as R I–II, RIIIA, RIIIB, and R IV (Fig. 1B). R I–II (the anterior intestine) extends back about 30 segments from the posterior margin of the crop. The surface of the typhlosole in this region is marked by extensive transverse folding, which diminishes as it extends posteriorly. Together, RIIIA and RIIIB (the mid-intestine) extend back about 31 segments from the posterior boundary of R I–II, a boundary which is not particularly distinct. Unfortunately, the boundary between R III A and R III B is not distinguishable by visual means. As an approximation, R III A may be considered to represent the anterior two-thirds of the mid-intestine. R IV (the posterior intestine) is easily recognized by the thin walls and the lack of a typhlosole. R IV extends back about 22 segments from the distinct posterior boundary of RIII B and it terminates at the anus. The number of segments varies so that the above data on segment numbers should be regarded only as a rough guide.
Electrical measurements
The average electrical potential difference (PD) across all regions of the intestine was serosal-side positive. In most preparations a stable PD was reached during the first hour after the tissue was mounted in the chamber. In a number of cases, however, the PD and the short-circuit-current (SCC) oscillated, the cause for this being unknown. Apart from these preparations, the PD and the SCC were reasonably stable during the usual 2-3 h period of an experiment and the useful lifetime of some preparations exceeded 24 h. In all cases, PD and SCC were averaged over periods of at least 45 min. The DC resistance (R) was determined independently, from currentvoltage plots, current being measured after the tissue was clamped for 25 s. These plots were usually linear in the range between ± 50 mV, but were often curvilinear when carried out to +100 mV.
The data for PD, SCC, and R were skewed in a positive direction, as indicated by the coefficient for skewness (g1), and were heteroscedastic, as indicated by Bartlett’s test. Consequently, they were coded for negative values and logarithmically transformed, which eliminated these problems. The back-transformed and decoded data for the four physiological regions appear in Table 1. Single classification analysis of variance indicates that there are significant differences (P < 0·001) among the means for all three variables. In addition, the Student-Newman-Keuls test (see Table 1) indicates that the means for PD in all four regions are significantly different from one another (P < 0·05). The results are a little more complex in the case for SCC and R; however, taken together they also indicate that there are at least four regions, especially when the linear arrangement of the intestine is taken into account when interpreting the results.
Tracer fluxes
Unidirectional fluxes of Na+ and Cl- in the mucosal to serosal direction ( and ) and in the serosal to mucosal direction ( and ), net fluxes, and observed and calculated flux ratios for the four regions of the intestine are presented in Tables 2–5 for both open- and short-circuited preparations. Each flux determination is based on a minimum of three 15 min sample periods. For reasons previously stated, the flux data were logarithmically transformed before statistical tests were made. The t test was used to test for significant net fluxes and to test for differences between the observed and expected flux ratios in short-circuited preparations, these two tests being equivalent. For fluxes measured under open-circuit conditions, equality of observed and expected flux ratios were tested by Scheffé’s multiple contrast test (Zar, 1974, p. 159). An approximate version of this test was used since the variances for the flux and electrical terms were not equal. In this version the three sample group variances and F ratios were weighted by sample size. Expected fluxes were calculated from the Ussing criterion (see Ussing, 1949; Koch, 1970) which predicts the flux ratio expected from a passive system, taking into account the electrochemical potential difference.
In R I-II, sodium and chloride both move passively (Table 2). Unidirectional sodium fluxes, about 4–5 μmol cm-2 h-1, are greater in this region than in other regions of the intestine while unidirectional chloride fluxes are not relatively large, 0·6–0·9 μmol cm-2 h-1. The tracer and the electrical data suggest that this region is ‘leaky’, particularly to sodium.
In R IIIA (Table 3), sodium and chloride are both actively transported. Under short-circuit conditions the net fluxes (fms–fms) are 1·46 and –0·92 μmol cm-2 h-1 for sodium and chloride, respectively (note that chloride is secreted). Similar results have been found in the rat jejunum (Munck, 1970). In the mammalian intestine, chloride and bicarbonate transport are often reciprocally related (Binder, 1975), and this may also be the case in R IIIA. In two sets of experiments a change from low (2 mM) to high (25 mM) bicarbonate saline (on both sides of the tissue) resulted in a threefold increase in (from –0·43 to - 1·28μmol cm-2 h-1), and a small increase in (from 0·88 to 0·98 μmol cm-2 h-1), but a decrease in SCC (from 68 to 47 μA cm-2). These data yield no direct information about bicarbonate absorption, but they are compatible with such absorption and clearly show that chloride secretion is affected by bicarbonate.
In R IIIB (Table 4), sodium is actively transported, fms–fsm = 1 · 74 μ mol cm-2 h-1 under short-circuit conditions. Chloride, however, appears to move passively in this region. There is considerable variation in the unidirectional chloride fluxes and since these are large, about 8 μ mol cm-2 h-1, a small net transport of chloride might not be readily apparent. R IIIA and R IIIB are similar when anatomical criteria are used; however, both the chloride fluxes and the electrical data indicate that they are physiologically distinct regions.
In R IV (Table 5), sodium is also actively transported, fms–fsm = 1 · 01 μ mol cm-2 h-1 under short-circuit conditions. When open-circuited, there is a significant net transport of chloride in this region. Chloride transport is passive, however, indicated both by the calculated flux ratio for the open-circuited preparation, and by the fact that the net transport approaches zero in the short-circuited preparation. The PD in R TV is substantial (about 30 mV) and its effects on the fluxes of sodium and chloride are clearly seen in the differences between these fluxes in open- and short-circuited preparations. Both the tracer and the electrical data suggest that R IV is a relatively ‘tight’ epithelium.
Short-circuit currents and sodium and chloride fluxes
For each region, a correlation coefficient (r) was determined for each set of unidirectional fluxes and their associated short-circuit currents. This was of some use in identifying significant relationships between net fluxes and short-circuit currents since there should be a positive correlation between short-circuit current and unidirectional fluxes in the active direction for rheogenic transport processes. In addition, the unexplained, or residual SCC (SCCR) was calculated as follows: , but when a net flux did not differ from zero, the appropriate unidirectional fluxes were removed from the equation. Values of SCCR were tested with Scheffé’s multiple contrast test to determine if they differed from zero (Table 6). A significant negative SCCR indicates an unexplained net transport of either anions in the mucosal to serosal direction, or cations in the serosal to mucosal direction.
In R I-II, where no significant net fluxes were found, there were no significant correlations between the unidirectional fluxes and short-circuit currents. Thus, the origin of the SCC, about 0 · 4 μ equiv cm-2 h-1, is not explained. In R III A, where sodium and chloride are both actively transported, but in opposite directions, a significant correlation between a sodium flux and SCC was not found. There is in this region, however, a significant correlation for versus SCC (r = 0 · 93, P < 0 · 01). Despite this fact, it is clear that SCC is not a reliable measure of in this region. For example, the transfer from low to high bicarbonate saline resulted in an increase in but a decrease in SCC. Furthermore, SCCR in this region is –0 · 58 μ cm-2 h-1, significantly different from zero (P < 0 · 05).
In both R IIIB and R IV there are significant correlations between and SCG (r = 0 · 96, P < 0 · 01, and r = 0 · 95, P < 0 · 01, respectively). In R IIIB SCCR = 0· 14 < mol cm-2 h-1, which is not significantly different from zero, and thus SCC is a good measure of in this region, at least under the present conditions. In R IV, however, SCCR = – 0 · 46 μ mol cm-2 h-1, a value significantly different from zero (P < 0 · 01 ). Thus, there are strong relationships between SCC and either or in three of the four regions of the intestine, but only in R IIIB is SCC a good estimate of a net flux.
The effects of some inhibitors on sodium transport
In all regions of the intestine either KCN (10−3M)or ouabain (10−8 to 10−5 M) added to the serosal side inhibited SCC and PD. Maximum inhibition, i.e. an 80–100% reduction in SCC, usually took place within 30–60 min for either inhibitor. The effects of KCN, but not ouabain, were reversible. Inhibition by ouabain was sometimes preceded by an initial stimulation of SCC.
The effects of amiloride, a ‘sodium channel’ blocker, were dependent on the region of the intestine. In general, amiloride was added to the mucosal side to give a final concentration of 10−4 M. In all cases where io-4 M was ineffective, a greater concentration, up to 10−3 M, also failed to produce a response. When effective, amiloride rapidly (within 45 s) and reversibly reduced PD and SCC, inhibited net sodium transport, and caused an increase in resistance. The effects of amiloride were dose-dependent and at a concentration of 10−4 M the SCC was usually inhibited by 60–90%.
In R I–II, amiloride inhibited SCC and PD in two out of four preparations. Flux measurements were not made on amiloride-treated preparations in this region. In R IIIA, amiloride did not affect PD or SCC in nine experiments; furthermore, in three pairs of experiments it had little or no effect on sodium fluxes. In R III B, however, amiloride inhibited PD and SCC in five preparations and it reduced . but not in two pairs of experiments. Similarly, in R IV, amiloride inhibited PD and SCC in 16 preparations and it reduced but not in three pairs of experiments, e.g. see Fig. 3. In general, amiloride was ineffective when added to the serosal side at 10−4 M, but in a few cases it reduced SCC by 10–15 % 1n R IHB and R IV. Amiloride has similar effects on the serosal side of toad bladder (Sharp, 1979). These results may be related to the fact that amiloride reportedly enters cells from the serosal side, but not from the mucosal side (Robbie, 1971, cited by Sharp, 1979).
Some potential methodological problems
Some potential problems associated with the use of small non-circular apertures are worth considering. Perhaps the most obvious one concerns the question of edge-damage, a problem which is aggravated by high-resistance epithelia, such as R IV. Although considerable attention was given to the mounting procedure, edge-damage may have occurred. We can, however, eliminate the possibility of severe edge-damage by the following considerations. The addition of 10−4 M amiloride to the mucosal side of five preparations of R IV caused the resistance to increase from 2430 to 13 680 Ω cm2. Resistance measurements were generally not made at higher concentrations of amliloride ; however, increasing the concentration to 10−3 M caused a further reduction in PD in several preparations, and thus probably resistance as well. Nevertheless, if we assume that the resistance in the presence of io-4 M amiloride is due to edgedamage, i.e. the tissue resistance is infinite, we can calculate a maximum value for the resistance of normal tissue, which is (13680 × 2430)7(13 680 –2430) = 2955 Ω cm2, or 22 % greater than the apparent value for normal tissue. It is clear that edge-damage in carefully mounted preparations of R IV is substantially less than in known cases of edge-damaged frog skin, for example, where the true resistance may be 200 % greater than the apparent value (Helman & Miller, 1973). Furthermore, considering the assumption of infinite tissue resistance in the above calculation, which ignores the resistance of paracellular shunts and the increased effect of 10−3 M amiloride, it seems probable that the true resistance is closer to the measured value than the maximum calculated value. The effects of edge-damage on resistance, PD, SCC, and sodium fluxes have been examined in other ‘tight’ epithelia, e.g. frog skin (Helman & Miller, 1973, 1974). In general, edge-damage causes resistance, PD, and the flux ratio to be underestimated, but it has little effect on the net flux or the associated SCC. Edgedamage would not cause passive transport to be misidentified as active transport, either through the use of the short-circuiting technique, or by the comparison of observed and calculated flux ratios under open-circuit conditions. Thus, even if we were to accept the maximum calculated resistance as the true resistance, and considered the effects that edge-damage would have on the other data, no changes in the conclusions drawn from the data would be required. Naturally, in the low-resistance regions of the intestine (R I-II, R IIIA, and R IIIB) edge-damage does not pond a significant problem.
Another concern is the possibility of non-uniform current distribution in the elongated apertures. This possibility was tested by passing current (300 μ A cm-2) across a chamber fitted with an empty tissue holder while moving a pair of voltage electrodes simultaneously along the longitudinal axis of the aperture, parallel to the aperture. No variation in current was found although a 1 % variation in current would have been easily detected. In addition, no differences in electrical measurements were found between preparations of either R III B or R IV mounted in either circular or elongated apertures (N was equal to or greater than ten in all four categories).
A question of more general concern, namely whether or not the preparation was effectively short-circuited, also warrants some discussion. The aim in short-circuiting a tissue preparation, like the one currently used, is to provide an electrical potential difference of zero across the entire tissue, as opposed to just the active membrane (see Rehm, 1975). One difficulty in achieving this is the error introduced by the voltage drop across the solutions between the voltage electrodes when current is passed through the chamber. The error can become appreciable in low-resistance epithelia, such as R III B, for example, where the true short-circuit current may be greater than 1 · 5 times the uncorrected value. In the present study the error caused by solution resistance was automatically corrected by the voltage-clamp, once the solution resistance was determined. In test runs with a model low resistance ‘tissue’, composed of resistors and a battery, the voltage clamp maintained the voltage across the model tissue at 0 ± 0 · 1 mV for extended periods of time and through realistic ranges of SCC, when the ‘tissue’/solution resistance ratio was 1 · 0, a ratio which was more demanding than the usual ratio of 2 · 0 or more, encountered with actual preparations. One internal test of the short-circuiting technique comes from the comparison of the observed and expected flux ratios in open- and short-circuited preparations. Thus, transport which is identified as active transport in the open-circuited condition should also be identified as active transport in the short-circuited condition, and similarly with passive transport. In every case there is agreement between these two criteria for distinguishing between active and passive transport (see Tables 2 – 5). Thus, there is some reassurance that short-circuit conditions were achieved.
DISCUSSION
The intestine of Lumbricus terrestris is a complex organ, composed of at least four physiologically distinct regions, designated as R I-II, R IIIA, RIIIB, and R IV (Fig. 4). The classical criteria for active transport (i.e. an observed flux ratio which differs significantly from either unity, in short-circuited preparations, or from the flux ratio calculated from Ussing’s equation (Ussing, 1949), in open-circuited preparations) indicate that sodium is actively transported in three regions (R IIIA, R IIIB, and R IV), and chloride, in one region (R IIIA). A summary of the net ion fluxes in short-circuited preparations and some electrical measurements for the four regions of the intestine are shown in Fig. 4.
In R I-II, there are no significant net fluxes of sodium or chloride. Some additional data, which will be discussed below, suggest that sodium may be transported in this region; however, if this is the case, then the rate of transport must be low. In contrast to R I-II, there is clear evidence that sodium is actively transported in the mucosal to serosal direction (i.e. it is absorbed) in R IIIA, R IIIB and R IV, the respective rates being 146, 1 · 74 and 1 · 01 μ mol cm-2 h-1 under short-circuit conditions. In no region of the intestine is there evidence of active chloride absorption; however, chloride is actively secreted in R IIIA at 0 · 92 μ mol cm-2 h-1 under short-circuit conditions. Chloride secretion is stimulated by an increase in bicarbonate concentration (from 2 to 25 mM). It is tempting to speculate that R III A mayabsorb bicarbonate (or secrete hydrogen ions) and thus play a role in acid-base regulation. Bicarbonate absorption in R IIIA would help explain the discrepancy between the SCC and the sum of sodium and chloride transport (Table 6). In R IIIB and R IV an increase in bicarbonate had little or no effect on sodium or chloride transport. R I–II was not examined in high bicarbonate saline. In R I–II, R IIIA, and R IV the origin of the SCC has not been completely explained; although, in RIIIB the SCC can be explained, within experimental error, by sodium absorption ( equals 93 % of the SCC, Table 6).
Amiloride, a ‘sodium channel’ blocker which inhibits some mechanisms of sodium entry at the apical (mucosal) border, has differing effects on the four regions. In half of the experiments with R I-II, amiloride (mucosal side) inhibited SCC, but in the remaining experiments it had no effect. The reason for these mixed results is not known. However, these data, along with the fact that ouabain (serosal side) uniformly inhibited SCC, raise the possibility of active sodium absorption. The SCC in R I-II is small (about 0 · 4 μ equiv cm-2 h-1 of charge transfer) and a small net sodium flux may have been overlooked because of the relatively large unidrectional fluxes of sodium (4 – 5μ mol cm-2 h-1). In RIIIA, a low-resistance region which absorbs sodium and secretes chloride, amiloride had no effect. This is a typical result with ‘leaky’ intestinal preparations (Bentley, 1979). The effects of amiloride on RIIIB represent an unusual case and will be dealt with below. In R IV, a high-resistance region where sodium is absorbed, amiloride inhibits sodium transport and SCC (Fig. 3).
The resistance in R IV in the presence of amiloride (13 680 Ω cm2, which represents a five to sixfold increase) has a bearing both on the extent of edge-damage and on the paracellular shunt resistance. If the resistive pathways in R IV may be described by a typical model for tight epithelia and the apical membrane resistance approaches infinity in the presence of amiloride (see review by MacKnight, DiBona & Leaf, 1980), then the value of 13 680 Ω cm2 must represent the summed parallel resistance of both edge-damage and shunt resistance. Thus, the above value provides a maximum estimate of edge-damage, but equally important, it provides a minimum estimate of the shunt resistance. It is clear that edge-damage, if present, is not extensive and that R IV is truly a ‘tight’ preparation. It is worth pointing out that the shunt resistance in toad bladder (Bufo marinus, 11890 – 12170 Ω cm2, Reuss, Gatzy & Finn, 1978) is comparable to the minimum estimate in R IV.
The present data were obtained during the months of July-January, a period in which the properties of the intestine are fairly constant. However, during January-April there are some significant changes in the properties of at least two of the regions, RIIIA and RIIIB. These changes have been described in a preliminary report (Cornell, 1980b) and will not be dealt with here in detail. However, it should be clear that the values reported here are not yearly averages. For the most part, the changes are of a quantitative nature and while the changes are significant, the overall view of the intestine is similar when the seasonal changes are taken into account. Relevant to the present discussion is the change in amiloride sensitivity which occurs in R III B, perhaps the most striking of the seasonal changes. Amiloride inhibits sodium transport during July-January, but it has no effect in April. This is the first known case of a seasonal change in amiloride sensitivity and the results suggest the existence of at least two mechanisms of sodium entry at the mucosal border, i.e. one which is sensitive to amiloride and one which is not. Whether both mechanisms are present simultaneously is not known. It may be recalled that R I-II also showed some variation in amiloride sensitivity; however, there are insufficient data to establish a correlation between amiloride sensitivity and season in this region. R IIIB and R I-II are the only ‘leaky’ preparations of which I am aware which are sensitive to amiloride. In mammals, the small intestine is not sensitive to amiloride, but the colon, which has Miigher resistance, is sensitive to amiloride (Bentley, 1979).
For purposes of comparison, there are few data of a similar nature available for annelids. In perhaps the only previous study of the isolated intestine, Sylvia & Boettiger (1967) deduced from electrical measurements that sodium was absorbed from the intestine of Lumbricus. The low PD (1-9 mV) and SCC (0 · 35 μ A for the perfused preparation consisting of R I-II through R IIIB) was most likely caused by the shunting of current through R I-II, a region of low resistance and PD.
In Table 7 the rates of sodium and chloride transport and some electrical poperties of a number of intestinal preparations are compared. The rates of sodium absorption in the intestines (especially the mid-intestines) of Lumbricus, the bullfrog (Rana catesbeiana), and the freshwater prawn (Macrobrachium rosenbergii) are similar. By contrast, the rate of sodium absorption in the sculpin (Cottus scorpius), a marine teleost, is considerably higher. The high rate in the sculpin (a hypotonic regulator which drinks sea water and subsequently absorbs salts and water from the intestine) reflects the important role of its intestine in overall salt and water balance (see review by Kirschner, 1979). In mammalian preparations the rates of sodium transport vary considerably, depending on the region and the experimental conditions. The jejunum usually secretes sodium while the ileum and colon usually absorb sodium. The rates of sodium absorption in the ileum and colon are usually greater than those found in Lumbricus, but less than in marine fishes, such as the sculpin.
Active chloride absorption does not seem to occur in the isolated intestine of Lumbricus, although it does occur in other preparations, e.g. bullfrog small intestine. In mammalian preparations, active chloride transport has often been difficult to demonstrate. Some possible reasons for this have been discussed by Frizzell et aL (1973). In the ileum, for example, which may spontaneously exhibit either active chloride secretion or active chloride absorption, chloride transport is thought to be regulated at the brush border by intracellular factors which are as yet poorly defined (Frizzell et al. 1973). Until the intestine of Lumbricus has been examined under a broader range of conditions the possibility of active chloride absorption should not be excluded. Active chloride secretion, however, occurs in Lumbricus in RIIIA, a region which actively absorbs sodium. Under certain conditions, for example in the presence of choleragen, the rabbit ileum secretes chloride, but in this case sodium is also secreted (Powell, Binder & Curran, 1973). The rat jejunum secretes chloride and absorbs sodium when bathed in saline containing 15 mM lysine (Munck, 1970). These results are similar to those found in R IIIA when 10 mM glucose is present in the bathing media; however, in the jejunum the removal of lysine results in sodium secretion, but the removal of glucose in R IIIA does not affect sodium absorption. In both cases chloride continues to be secreted.
There appears to be a common pattern to the organization of the intestine in Lumbricus and in many vertebrates. This pattern, which consists of low resistance regions which generate a small PD, followed by a terminal region of higher resistance which generates a larger PD, may be seen in Lumbricus, the bullfrog, and the rabbit (Table 7). The higher resistance of the terminal region probably reflects an adaptation for salt conservation.
The available data from isolated preparations and from whole animals support the view that the intestine of Lumbricus plays an important role in salt conservation. Dietz & Alvarado (1970) found that in pond-water-adapted Lumbricus the concentration of sodium in the crop was 62 mM compared with 6 mM in the rectum ( = R IV). My own measurements in five animals indicate that sodium is 60 – 80 mM in the crop, gizzard, and the three anterior regions of the intestine (roughly the same concentration as in the coelomic fluid or blood), but it is only 5 – 10 mM in R IV. The rate of crop fluid formation is not known ; however, it would appear that significant quantities of sodium could be lost if this fluid were not modified before it left the animal. The data from the isolated preparations suggest that little fluid modification would occur in R I–II. In both R IIIA and R IIIB, sodium would be expected to be absorbed. The low resistance of these two regions is consistent with the possibility of isosmotic fluid transport, a postulate which would explain the lack of change in sodium concentration in these regions in the whole animal. In R IV, where sodium absorption would also occur, the high resistance of this region suggests that it is capable of maintaining the large sodium concentration gradient which is found in the whole animal. R IV appears to be a very specialized region. The high resistance (2095 Ω cm2) and the ability to absorb sodium against a io-to-i concentration gradient are characteristics which are more generally associated with tight epithelia, such as frog skin or toad bladder, than with intestinal preparations. The specializations found in R IV would be of benefit in environments with low sodium availability, such as pond-water or moist soil.
Based on the results from the isolated preparations, RIIIA, RIIIB, and R IV together would be expected to absorb sodium at a total rate of about 1 · 0 μ mol g-1 h-1 in the whole animal, this rate being derived from the open-circuit net fluxes and the surface area in a typical 3 g animal. This rate is considerably higher than the rate of sodium influx in whole animals in pond-water, which occurs across the integument about 0 · 03 μ mol g-1 h-1 for animals in a steady state, Dietz & Alvarado, 1970). If the calculated rate of intestinal absorption is correct, then a small decrease in this in the whole animal would require, on a percentage basis, a much larger compensatory increase in sodium uptake in order to maintain the steady state. This suggests that overall sodium balance may be greatly influenced by the intestine. The calculated rate of intestinal sodium absorption also suggests that most of the sodium which is absorbed from the intestinal lumen is secreted into the gut of animals in pond-water, since the rate of absorption (1 · 0 μ mol g-1 h-1) is much greater than the rate which can be accounted for by drinking pond-water (0 · 0003 μ mol g-1 h-1, Dietz & Alvarado, 1970). The crop might be the site for this sodium secretion.
It appears that the intestine does not conserve chloride by active means. Dietz & Alvarado (1970) report that the chloride concentrations in the crop and rectal fluids are 2 and 13 mM, respectively, the blood concentration of chloride being about 39 mM (Boroffka, 1965). It should be noted that an unknown anion, or anions, is present in the blood and crop fluid, the sodium concentration in each being about 60 mM. Unfortunately the chloride concentrations in the luminal fluids between the crop and the rectum are unknown. However, based on the in vitro results, it would seem probable that the 20-to-i chloride concentration gradient found across the crop would not be maintained in R I-II. In R III A, which secretes chloride in vitro, the chloride concentration might equal or exceed the blood concentration. If this were the case, some chloride might be passively absorbed in R III B (a region which is relatively permeable to chloride) if a favourable electrical potential existed, such as exists in vitro. In R IV, chloride would also be expected to be passively absorbed. It is interesting to note that the in vitro PD (32 mV) is more than sufficient to account for by passive means the 3-to-1 concentration gradient (blood to intestinal fluid) which is found across R IV in vivo.
The observations made on the whole animal may be incorporated with the in vitro results to form a consistent view of the physiological role of the intestine in salt and water balance. Nevertheless, extrapolating from in vitro to in vivo conditions can clearly be misleading. For example, in frogs the values of fmax and Km for sodium are ten times greater in the isolated skin than for the skin on the intact animal (Kirschner, 1978). Consequently, the ideas expressed here regarding the function of the intestine in vivo must be regarded as very speculative.
The present study is based on animals adapted to pond-water. A number of justifications for this are discussed by Dietz & Alvarado (1970). However, Carley (1978) points out that although Lumbricus displays many characteristics of freshwater animals, in the field it is a terrestrial animal, more often subject to desiccation than to hydration. Indeed, there are physiological differences between animals adapted to pond-water and soil, for example, a change in integumental water permeability (Kamemoto, 1964; Carley, 1975). The properties of the intestine may also differ. However, it would seem that the basic organization of the intestine in Lumbricus could accommodate both salt and water absorption and thus fulfil some of the needs of both aquatic and terrestrial existences. The present data establish the ability of the intestine to absorb sodium. The ability to absorb water remains to be demonstrated. That this function may be performed is suggested by the fact that the posterior nephridia of some of the most xeric earthworms (e.g. Pheretima posthuma) open into the intestine rather than to the exterior (Bahl, 1945).
ACKNOWLEDGEMENTS
I thank Dr Leonard B. Kirschner for many helpful discussions during the course of this work and during the preparation of the manuscript. I would also like to thank Dr Ralph I. Smith for his critical evaluation of the manuscript. This work was supported by a National Institutes of Health postdoctoral training grant, GM 01276.