ABSTRACT
Secretion of chloride from blood to lumen is accomplished in the rectal gland of elasmobranchs by a process of secondary active transport involving the co-transport of Cl− with Na+ across the basolateral membranes of rectal gland cells. Energy is provided by ATP via membrane Na-K-ATPase, which establishes an electrochemical gradient favouring Na+ influx into the cell. The involvement of K+ in the co-transport mechanism, so as to provide a ratio of 1 Na+: 1 K+: 2 Cl− entering the cell, would increase the energetic efficiency of the process, and is consistent with the CI/O2 ratio of 27–30 observed in secreting rectal glands. Secretion is stimulated by cyclic AMP (cAMP) and by vasoactive intestinal peptide (VIP) and adenosine, which activate adenylate cyclase. Activation of the gland in vivo probably occurs via VIP-secreting nerves as well as circulating agents; it is inhibited by somatostatin. Cyclic AMP probably stimulates chloride secretion by at least three mechanisms: (1) increasing chloride conductance across the luminal cell membrane, (2) enhancing the co-transport pathway for transmembrane movements of Na+, K+ and Cl− and (3) activating Na-K-ATPase.
INTRODUCTION
Although elasmobranch fishes, including sharks and rays, are said originally to have developed in fresh water, they have lived for millions of years in the salty ocean where they are subjected to osmotic stresses that menace the composition of their body fluids. The shark swims in a sea containing about 500 mequiv 1−1 of sodium chloride, or 1000 mosmoll−1. The serum sodium of the shark is approximately 260–290 mequiv 1−1, but sodium and its associated anions in the extracellular fluid of the shark amount to only half the external osmotic pressure of the ocean. In order to counter balance the hypertonicity of the sea, which if unopposed would rapidly dehydrate the animal, elasmobranchs synthesize urea, which circulates in high concentration in the blood and also permeates all cells. The concentration of mineral salts plus urea precisely balances the osmotic concentration of sea water, preventing the depletione body water. Nevertheless, sodium chloride tends to diffuse inward through the gills and is also absorbed when food is swallowed. Thus, the shark must combat a tendency to become hypernatremic. The mechanism that elasmobranchs have devised for getting rid of unwanted increments of salt is the rectal gland (Burger & Hess, 1960; Burger, 1962).
The rectal gland of the spiny dogfish shark, Squalus acanthias, is a relatively simple structure that looks like a human appendix. It is composed of tubules packed tightly together looking superficially in histologic section like mammalian kidney cortex, and well supplied with capillaries. Electron microscopy reveals a uniform population of cells with extensive basolateral infoldings (Fig.1). These are rich in Na-K-ATPase, as demonstrated autoradiographically by the binding of radioactive ouabain in histological sections (Eveloff et al. 1979). The rectal gland has a single artery, vein and central duct. It is therefore a simple matter to isolate the gland, remove it from the shark and perfuse it in the laboratory at the temperature of sea water with a shark Ringer ‘s solution resembling an ultrafiltrate of shark plasma.
RECTAL GLAND SECRETION IN VIVO
The properties of fluid secreted by the rectal gland of Squalus acanthias, are shown in Table 1. In comparison with shark plasma, the fluid is concentrated with respect to sodium chloride, but isosmotic by virtue of the fact that it contains almost no urea. The concentration of sodium and chloride therefore approximates that of sea water, or about 500 mequiv 1−1. The isolated perfused gland secretes a fluid that is entirely similar in composition to that produced in vivo. When perfused in the laboratory at a pressure comparable to the blood pressure of the shark (20–25 mmHg), the potential difference across the gland, serosal (blood) side to apical (duct lumen), is orientated with the duct negative. Since the concentration of chloride in the duct is almost twice that in the perfusate, chloride is transported against both an electrical and a chemical gradient, while sodium moves down an electrical gradient that roughly balances the opposing chemical gradient (Silva et al. 1977).
The intact, free-swimming shark secretes rectal gland fluid at a highly variable rate. Small intravenous infusions of hypertonic sodium chloride stimulate profuse section. An equally effective stimulus to rectal gland secretion is the isotonic expan sion of body fluids with shark Ringer ‘s solution (Solomon et al. 1980). It seems likely that in live fish, swimming in the ocean in osmotic equilibrium with their surround ings, changes in the osmotic activity of body fluids are immediately translated into changes in body volume, so it is not surprising that isotonic expansion should be a potent stimulus to the rectal gland. Intravenous infusions stimulate rectal gland secretion promptly, within the first half-hour of collection. The volumes attained [2– 4 ml(g gland wet weight)−1 h−1] are comparable to the rates of secretion obtained when the isolated gland is maximally stimulated in vitro. Since the mean arterial blood pressure is not altered by these intravenous infusions, it seems likely that the afferent stimulus for rectal gland secretion lies on the venous side of the circulation.
CHARACTERISTICS OF RECTAL GLAND SECRETION
Some key characteristics of rectal gland secretion are listed below, as elucidated by studies of isolated glands perfused with shark Ringer ‘s solution containing 5 mm-glucose and aerated with 99% O2, 1% CO2 (pH7-5-7-6) (Silva et al. 1977).
Chloride is actively secreted against an electrical and chemical gradient, from blood to duct.
Secretion is inhibited by ouabain and by omitting potassium from the perfus ing solutions. It is therefore likely to depend on the activity of Na-K-ATPase, which lines the basolateral border of rectal gland cells.
Secretion of chloride depends on the presence of sodium in the perfusate, and secretion of sodium depends on the presence of chloride in the perfusate.
Secretion is inhibited by furosemide and furosemide analogues, agents which block coupled sodium-chloride transport in other tissues. Furthermore, the potency of ‘loop diuretics’ in inhibiting rectal gland secretion follows the same rank order as their potency in blocking sodium-chloride co-transport in avian erythrodays or in the thick ascending limb of canine kidneys (Palfrey et cd. 1979).
The intracellular concentration of chloride as deduced from chemical measurements, as well as the activity of chloride in cellular cytoplasm as determined directly by micropuncture with chloride electrodes, is 60–80 mequiv 1−1 (Duffey et al. 1978; Welsh, Smith & Frizzell, 1980). This is two to four times higher than that predicted by the Nemst equation from the simultaneously measured electrical poten tial inside cells, which has been variously measured at –60 to –100mV (Table 2). Chloride must therefore be transported uphill into the cell across the basolateral border.
These features of rectal gland secretion strongly suggest the hypothesis of a ‘secon dary active’ co-transport of NaCl across basolateral membranes, in which chloride is transported into the cell against an electrochemical gradient, coupled to the downhill movement of Na+. A downhill electrochemical gradient for Na+ is maintained by activity of Na-K-ATPase (Fig. 2). More direct evidence for this hypothesis has been obtained from experiments on membrane vesicles.
DIRECT EVIDENCE FOR SODIUM CHLORIDE CO-TRANSPORT IN MEM-BRANE VESICLES FROM RECTAL GLAND
A membrane fraction rich in basolateral plasma membranes (as evidenced by the specific activity of Na-K-ATPase) was prepared from the rectal gland of the spiny dogfish and the uptake of “Na into the plasma membrane vesicles investigated by a rapid filtration technique (Eveloff et al. 1978). Sodium uptake was greatest in the presence of a potassium chloride gradient directed into the vesicles. It was strikingly reduced when chloride was replaced with nitrate and was even slower with gluconate (Fig. 3). If the membrane vesicles were pre-equilibrated with potassium chloride or potassium nitrate plus valinomycin, to minimize the influence of any electrical driving forces on sodium movement, the uptake of sodium was still greatest in the presence of chloride and remarkably decreased in the presence of nitrate. Furthermore, satura tion of sodium uptake by increasing sodium chloride concentrations was observed, suggesting a saturable carrier. Furosemide at concentrations of 10−3M and 10−4M decreased sodium uptake into the vesicles in a dose-dependent manner only in the presence of chloride. Analogues of furosemide were effective in proportion to their comparative diuretic potency (Kinne et al. 1979). On the other hand, sodium uptake was not inhibited by amiloride, an inhibitor of sodium-proton exchange, or by 4-amino-4 ‘-isothiocyanostilbene-2,2 ‘-disulphonate (SITS), which blocks anion-exchange mechanisms in other systems (Kinne, Friedman & Kinne-Saffran, 1982), nor do these agents inhibit secretion by perfused rectal glands (unpublished data). Interestingly, the transport properties of membranes isolated from glands stimulated by cAMP and theophylline do not differ significantly from those isolated from glands in the basal state (Hannafin et al. 1981).
These results provide direct evidence for a coupling of sodium and chloride fluix across the plasma membrane of the rectal gland via a co-transport system sensitive to furosemide. They are strong support for the co-transport hypothesis of chloride secretion illustrated in Fig. 2.
HORMONAL CONTROL OF RECTAL GLAND SECRETION
The addition of theophylline or of dibutyryl cAMP to solutions perfusing the isolated rectal gland results in a burst of secretory activity (Stoff et al. 1977). When secretion is stimulated in this way the electrical potential difference between duct and perfusate becomes more negative (–5 to –15 mV), the flow of duct fluid is increased (from about 0·3 to 2–4mlh−1g−1), and the concentration of sodium chloride in the secretion frequently increases slightly (from 450 to 500mequivI−1). Thus, cAMP stimulates the movement of chloride against an even higher electrochemical gradient than existed previously, underlining the active nature of chloride transport in rectal gland secretion. A relatively constant rate of secretion in vivo (about 1200/tequiv h−1g−1) can be obtained by perfusing with 0·25 mm-theophylline and 0·05 mm of dibutyryl cAMP. The addition of cyclic GMP to the perfusate has no effect, either to stimulate secretion, or to block secretory stimuli.
An extensive search for the ‘first messenger’ of rectal gland secretion was carried out in perfused glands (Stoff et al. 1979). A number of hormones have been studied at pharmacological concentrations and found to have no significant effect on chloride secretory rate. Vasoactive intestinal peptide was the only hormone studied that stimulated chloride secretory rate.
The effect of VIP to increase chloride secretory rate appears to be mediated by cAMP. The hormone induces a significant rise in intracellular cAMP over the entire concentration range studied (10−6–10−8M). Theophylline, an inhibitor of phosphodiesterase, enhances the cAMP response to VIP at each concentration of the hormone tested. These findings indicate that the effect of VIP in this tissue occurs via activation of adenylate cyclase. The effect of theophylline on cAMP metabolism parallels its augmentation of VIP-stimulated volume and chloride secretory rate and supports the view that intracellular cAMP mediates active chloride transport in the rectal gland. Although secretin and glucagon share marked structural homologies with VIP, neither of these hormones produces a significant increase in secretion or in intracellular cAMP. These observations suggest the presence of a highly specific cell receptor for VIP at the contraluminal surface of rectal gland epithelial cells.
It is likely that since VIP functions as a neurotransmitter in other animals, at least one mode of stimulation of the rectal gland in vivo will be found to involve neurogenic stimulation via nerve terminals containing VIP. VIP has, in fact, recently been identified by immunohistochemical means in nerve fibres of isolated perfused glands (R. Lechan&J. Stoff, unpublished observations). Veratridine (or the cruder prepara tion, veratrine), which depolarizes nerves, stimulates secretion when added to the medium perfusing isolated glands. The stimulation by veratridine is prevented by tetrodotoxin, which blocks neural impulses (Erlij, Lodenquai & Rubio, 1981). Furthermore, veratridine stimulation requires the presence of nerves, since prepara tions of isolated rectal gland cells devoid of neural elements cannot be stimulated desrease their oxygen consumption by veratridine, whereas their respiration is in creased by VIP or cAMP (Silva et al. 1982).
Adenosine, known to activate adenylate cyclase in other tissues, also stimulates gland secretion at a concentration of 10−5 M, presumably via activation of A2 (excita tory) membrane receptors (Erlij, Silva & Reinoch, 1978; Forrest, Rieck & Murdaugh, 1980). At lower concentrations, adenosine might also activate inhibitory (A1) recep tors, thereby retarding secretion. Both the inhibitory and excitatory actions of adenosine are inhibited by theophylline (Daly, 1982).
Studies of electrolyte secretion by mammalian gastrointestinal tract and pancreas indicate that somatostatin may inhibit secretory transport (Carter, Bitar, Zfass & Makhlouf, 1978). Both somatostatin and VIP have been identified by immunohisto chemical methods at similar sites in several species, including Squalus acanthias. The effect of somatostatin on VIP-induced secretion was therefore studied in the isolated perfused rectal gland. Somatostatin ( 1 · 4 × 10−7 M) had no effect alone on the secretory rate for chloride, but completely inhibited chloride secretion induced by VIP. The effect of somatostatin was reversible; i.e., chloride secretion induced by VIP could be restored by removal of somatostatin from the perfusion medium (Stoff et al. 1979).
Somatostatin blocks the transient stimulation of isolated rectal glands perfused with veratrine (Fig. 4), consistent with the surmise that veratrine depolarizes VIP-releasing nerves in this preparation. Interestingly, somatostatin does not prevent stimulation of the gland by adenosine or by forskolin, a compound that directly activates adenylate cyclase. The fact that somatostatin markedly inhibits VIP but exerts little or no inhibition on adenosine and forskolin suggests a specific point of action on the VIP-receptor-adenylate cyclase complex (Epstein et al. 1982).
The primacy of neural as opposed to hormonal regulation of rectal gland secretre in live sharks has been called into question by the experiments of Solomon et al. (1981) on explanted glands. In these experiments the rectal gland of a spiny dogfish was excised completely and attached via its artery to the circulation of another fish which was maintained alive at a normal arterial pressure and in running sea water. Following infusion of 150 ml of isotonic shark Ringer ‘s solution, secretion by the explanted gland increased dramatically, accompanied by an increase in blood flow to the gland, although there was no change in the pressure of perfusion. The increases in secretion and blood flow and the reduction in vascular resistance equalled those produced by similar infusions in rectal glands in situ with intact neural connections. The effect of volume expansion on secretion could be inhibited by infusing somato statin into the circulation of the explanted gland, but not by theophylline, implying that VIP rather than adenosine is likely to have been the excitatory agent (Solomon et al. 1981). Despite these suggestive findings, we have thus far been unable to demonstrate a convincing and consistent correlation between arterial blood levels of VIP (as determined by a rather difficult radioimmunoassay) and the rate of rectal gland secretion in intact sharks.
MECHANISMS BY WHICH CYCLIC AMP ACTIVATES SECRETION BY THERECTAL GLAND
Increase in apical chloride conductance
The activation of secretion by rectal gland cells involves changes in the permeability and enzymatic characteristics of both luminal and basolateral cell membranes. The chloride content of stimulated cells is consistently slightly lower than in the basal or resting state. Furthermore, measurements of intracellular potential indicate that after cAMP stimulation it becomes less negative (i.e., cAMP depolarizes the membrane) (Table 2). At the same time, duct potential becomes more negative, so that the voltage difference across the apical membrane falls. Thus, the electrochemical force driving chloride across the apical membrane into the duct lumen is diminished during excita tion at the same time that the secretory flux of chloride is greatly increased. Chloride conductance across the luminal membrane of the cell must therefore be increased by stimulation. This appears to be analogous to the augmented chloride conductance produced by epinephrine across the apical membrane of corneal epithelium (Klyce & Wong, 1977; Nagel & Reinach, 1980) and induced by secretagogues in the luminal (apical) membrane of canine tracheal epithelium (Welsh, Smith & Frizzell, 1983). It should be noted that a change in luminal permeability as the result of activation of adenylate cyclase located in basolateral membranes is characteristic of many other hormone-induced alterations in epithelial transport, such as those of vasopressin and parathyroid hormone.
Increase in conductance of basolateral membranes
In addition to facilitating the exit of Cl− from the cell across its luminal border, it seems likely that the permeability of basolateral membranes to the movements of other ions is also increased when rectal gland secretion is stimulated. It is not yet cleaether this is a primary or a secondary effect. When the active sodium-potassium pump (Na-K-ATPase) of perfused rectal glands is blocked by ouabain, the concentra tion of intracellular ions changes in the direction of equilibrium with extracellular fluid. These changes were examined when isolated perfused glands were in the basal state and also when they were stimulated to secrete with cAMP and theophylline, to see whether stimulation affected the passive movement of sodium, potassium and chloride across cell membranes (Silva, Spokes & Epstein, 1979). In basal glands ouabain induced an increase of 30 mequiv I “1 in intracellular Na+ concentration and a decrease of 50 mequiv 1−1 in intracellular K+ after 30 min, while the intracellular concentration of Cl− was unchanged. In stimulated glands, these movements were exaggerated. The increase in intracellular Na+ averaged 112 mequiv l−1 and the decrease in intracellular K+ averaged 96 mequiv 1−1 (P<0·01), while mean intracellular Cl−concentration rose by 80mequivI−1. Furosemide, 10−4M, partially reversed the accelerated changes in intracellular electrolytes that are seen after ouabain is added to stimulated glands (Fig. 5). Similar results have been obtained with slices of the rectal gland of the European dogfish, Scyliorhinus canicola (Shuttleworth & Thompson, 1980). These data are consistent with an action of Camp upon a ouabain-insensitive transport of sodium, potassium and chloride in the rectal gland analogous to that described in avian erythrocytes (see Palfrey & Greengard, 1982). An increase in coupled NaCl entry across the basolateral cell border is intuit ively necessary to permit the large transcellular secretory flux of chloride that is stimulated by cAMP, and an increase in basolateral potassium permeability has been postulated to accompany increased Na/K pump activity in Na-absorbing epithelia (Schultz, 1981). In a system analogous to the rectal gland, the tracheal epithelium of the dog, an increase in basolateral conductance with the onset of active Cl− secretion has now been demonstrated by more rigorous means (Welsh et ai. 1983).
Direct activation of Na-K-ATPase
Finally, it is probable that hormonal stimulation of the rectal gland activates Na-K-ATPase directly. The oxygen consumption of stimulated glands is more than four-fold higher than in the basal state. Ouabain-inhibitable oxygen consumption, presumably a reflection of Na-K-ATPase activity, is more than six-fold greater after stimulation (Silva, Stoff & Epstein, 1979). Fig. 7 shows the estimated intracellular concentrations of Na, K and Cl in unstimulated and stimulated perfused rectal glands. With stimulation, intracellular concentrations of sodium and potassium change reciprocally, resulting in a fall in intracellular sodium and in many experi ments but not all, a rise in intracellular potassium. Thus, it is difficult to ascribe the activation of Na-K-ATPase to an increase in intracellular sodium content.
Ouabain binding to rectal gland slices is increased after cAMP stimulation (Shuttleworth & Thompson, 1978, 1980) but the Na-K-ATPase activity of cell homogenates is unchanged (Silva, Stoff & Epstein, 1979). In an attempt to examine the mechanism of enzyme activation, the binding of ouabain to rectal gland cells of Squalus acanthias has been measured before and after stimulation with cAMP and theophylline (J. Epstein et al. 1981). Stimulation significantly accelerates the rate of binding and increases the amount of ouabain bound at equilibrium in the presence of low concentrations of ouabain (10−9M to 10−7M) (Fig. 7). Scatchard plots suggest at least two classes of binding sites, one of high and one of low affinity. Stimulation with cAMP and theophylline appears to increase the ouabain affinity of the high-affinity site. Ouabain binding at this high-affinity site is increased by cAMP and theophylline even when rectal gland secretion is inhibited by furosemide or bumetanide, and when Li+ is substituted for Na+, or NO3− for Cl−. The changes in ouabain binding induced by cAMP and theophylline in this species do not appear, therefore, to be secondary to secretory activity. In this respect they differ from the findings in Scyliorhinus canicola, where ouabain binding was thought to vary with Na-K-ATPase turnover (Shuttleworth, 1980). Furthermore (again unlike Scyliorhinus, Shuttleworth et al. 1982), ouabain binding is not increased in Squalus rectal gland cells when cell ATP hydrolysis is stimulated by the addition of a sodium ionophore that increases sodium entry into the cell (Silva et al. 1982). Thus, the increase in ouabain affinity produced by cAMP in this species seems not to be a simple consequence of increased Na-K-Pase turnover nor of an increase in internal cell sodium concentration, but presumably reflects a change produced by cAMP in the configuration, environme or location of existing enzyme so as to enhance its activity.
HIGH-ENERGY PHOSPHATE AND TRANSPORT IN RECTAL GLAND
The shark rectal gland contains substantial amounts of phosphocreatine, creatine phosphokinase, and adenine phosphonucleotides, and therefore provides a useful model in which to study the sequence of events involving high-energy phosphate turnover when secretory work is abruptly stimulated. Kinetic sequential studies of this sort can be done in the isolated perfused gland by nuclear magnetic resonance (n.m.r.), as described for other organs like heart, liver and kidney (Gadian & Radda, 1981; Ross, 1981). Fig. 8 illustrates the effect of stimulation with cAMP and theophylline on the 31P n.m.r. spectrum of an isolated rectal gland perfused at 16 ° C within the bore of a superconducting magnet (Oxford Instruments) in the laboratory of Dr George Radda. Stimulation produced little or no change in ATP, but a distinct fall in phosphocreatine and a reciprocal rise in the inorganic phosphate peak (F. H. Epstein, K. Thulbom & R. S Balaban, unpublished data). Thus, in the shark rectal gland as in skeletal muscle or mammalian heart, sudden substantial energy drains are met by changes in cellular levels of creatine phosphate in such a way as to tend to conserve the level of ATP within the cell.
STOICHIOMETRY OF CHLORIDE TRANSPORT IN RECTAL GLAND
The relationship of oxygen uptake to chloride transport in isolated perfused rectal glands has stimulated speculation that more than one Cl− is coupled with the move ment of one Na+ into the cell across the basolateral membrane (Silva et al. 1980). The molar ratio of net Cl− secretion to oxygen consumption in stimulated glands (Cl/O2 ratio) consistently averages 27 to 30 (a ratio, it may be remarked, that is also charac teristic of the mammalian kidney). If 3 mol of Na+ are transported per mol of ATP hydrolysed by rectal gland membrane Na-K-ATPase, and if the P/O ratio of rectal gland mitochondria is 3:1, then the ratio of Na+ transported to O2 consumed should be 18: 1, and if 1 mol of Cl− were transferred for every mol of Na+ transported, the ratio of Cl− to O2 would also be 18:1. The ratio of 30:1 observed in the experiments of Silva et al. (1980) clearly suggests the possibility that two or more Cl− may be linked to the movement of one Na+. Fig. 2 illustrates that from a thermodynamic standpoint, the electrochemical potential driving Na+ into the cell is more than enough to account for the simultaneous entry of more than one Cl−.
One mechanism by which more than one Cl− could be transported into the cell with one Na+ is provided in the recent demonstration (Hannafin, Kinne-Saffran, Fried man & Kinne, 1983) that K+ is required for the co-transport of Na+ and Cl− by rectal gland membrane vesicles. Moreover, radioactive rubidium (86Rb+), a marker for potassium, is accumulated by such vesicles by a process that depends on the presence of Cl− and is blocked by loop diuretics. This suggests that, as in avian erythrocytes (Palfrey & Greengard, 1982), Ehrlich ascites cells (Geek et al. 1980) and mammalian thick ascending limb (Greger & Schlatter, 1981), Cl− co-transport in the rectal gland might involve the linkage of one Na+, one K+ and two Cl−. Note that little additional thermodynamic work is required for the inward transport of K+; since K+ is in approximate electrochemical equilibrium across the basolateral membrane, the chemical gradient opposing influx is roughly balanced by the electrical gradient favouring entry. Potassium entering the cell in this fashion would diffuse back out across the basolateral membrane. Sodium would be pumped out by Na-K-ATPase into the intercellular space, and diffuse passively into the duct lumen across inter cellular junctions. Cl− would leave the cell chiefly across the luminal membrane, to be joined in the lumen by Na+. The net result would be an amplification of the efficiency of the cell, permitting two NaCl to be secreted into the duct lumen at the cost of transporting only one Na+via Na-K-ATPase (Fig. 9).
The importance to this system of K+ movement across the basolateral membrane is supported by the effects of Ba2+ on rectal gland secretion. A competitive inhibitor of K+ conductive pathways, Ba2+ (2–5 HIM) reversibly depressed rectal gland secretion while reducing transglandular electrical potential (Silva et al. 1981). Potassium recycling from cell to interstitial space was presumably inhibited by Ba2+. The model depicted in Fig. 9 predicts that such a change would lower the potential difference across the luminal membrane and thereby reduce the driving force for chloride secretion.
The general features of this model for active chloride transport linked to Na+ and K+ appear to be applicable to a variety of transporting epithelia in addition to the rectal gland. These include canine tracheal epithelium (Welsh et al. 1983), teleost intestine (Field, Kimberg, Orellana & Frizzell, 1981) and the thick ascending limb of Henle ‘s loop (Greger & Schlatter, 1981). A modification of the model is required for epithelia in which Cl− is actively reabsorbed (as in intestinal mucosa and the thick ascending limb of Henle) rather than secreted (as in canine tracheal epithelium and rectal gland). In reabsorbing epithelia, the Na-K-Cl transporter must be on the apical surface of the cell (as evidenced by the site at which furosemide inhibits), as is the chief Ba2+-sensitive channel for K+ exit. These relationships are illustrated in Fig. 10. Further delineation of the precise molecular steps by which ion movements are activated in the rectal gland of the shark should provide important clues to the way in which transport is controlled by hormones in other tissues.
ACKNOWLEDGEMENTS
Supported by grants from the U.S. Public Health Service (AM 18078 and SO7-RR-05764) and the National Science Foundation (DEB 7826821 and PCM-77-01146).