ON THE INTERACTION OF NH 4 + AND Na + FLUXES IN THE ISOLATED TROUT GILL

Transcription

1 J. exp. Bio/. (1976), 64, With a figures fruited in Great Britain ON THE INTERACTION OF NH 4 + AND Na + FLUXES IN THE ISOLATED TROUT GILL BY THEODORE H. KERSTETTER AND MICHAEL KEELER Marine Laboratory, Humboldt State University, Arcata, California 95521, U.S.A. {Received 13 August 1975) SUMMARY 1. Sodium influx was measured in isolated, previously perfused gill arches of rainbow trout, Salmo gairdneri, by measuring incorporation of M Na into gill tissue following timed exposure to a 1 mm ^NaCl medium. Transport rates approximated those estimated for intact fish and were linear for at least one min. 2. NH^Cl-containing perfusates at ph 7 and 8 stimulated Na+ influx equally, indicating that only ionized ammonia is important in the transport process. A Na + /NH 4 + exchange at basal and/or lateral membranes of the transporting cells is suggested. 3. Low-sodium Ringer perfusate augmented Na + influx; in one group of gills the transport rate was more than double that of controls. The increase in transport induced by internal NH 4 + was not additive with the low sodium augmentation. A reduction in intracellular [Na + ] is postulated as the mechanism operating in both cases. 4. Ouabain had no appreciable effect on Na + influx, either with or without NH 4 + in the perfusate. Diamox partially blocked the augmented Na + influx induced by NH 4 +. Amiloride completely inhibited Na+ influx, both with and without NH 4 + in the perfusate. INTRODUCTION Recent studies on ion uptake in the freshwater teleost gill have demonstrated a number of interesting characteristics of the transport systems (reviewed by Maetz, 1971). A model developed by Maetz and Garcia-Romeu (1964) proposed an exchange of endogenous NH + 4 and HCO 8 ~ for Na+ and Cl~ in the medium; intracellular hydration of CO 2 and subsequent dissociation of the carbonic acid supplies unlimited HCO 3 ~ for exchange and H+ for reaction with NH 3. Subsequently, work with rainbow trout and goldfish has indicated that H+ as well as NH^ can serve as an exchanging cation (Kerstetter, Kirschner & Rafuse, 1970; Maetz, 1973). But many details of the transport systems, particularly the nature of the transfer processes across both mucosal and serosal borders, remain unclear. One question of particular interest concerns the movement of ions across the serosal cell membranes. The above model incorporates the well-known Na +,

2 518 T. H. KERSTETTER AND M. KEELER K+-ATPase to complete sodium movement into the blood. But there is little evidence for this in the trout. In fact, the gills of freshwater teleosts have much lower levels of this enzyme than do their salt-water-dwelling counterparts (Kamiya & Utida, 1969; Zaugg & McLain, 1970). A second question relates to the effect of ammonia on sodium movement. Is the Na^Nr!^ exchange located on the apical border, the basal border (in an obvious analogy to the Na+-K+ pump), or is it a feature of both membranes? Ideally, investigations of transport mechanisms should incorporate experimental manipulation of bathing media on both sides of the epithelium. This is done routinely in investigations of isolated frog skins and toad bladders, but it is more difficult with the teleost gill for two reasons. First, the gill is not a flat sheet which can be used to separate internal and external solutions, and, secondly, the perfused gill has not been an unqualified success. Although Shuttleworth (1972, 1974) has reported good results with both freshwater- and sea-water-adapted eels {Angxdlla dieffenbachu), both Richards & Fromm (1970) using trout, and Kirschner (1969) using eels, reported negative net sodium fluxes in isolated perfused gills. Kirschner also described a constantly increasing pressure in his constant-flow system. Bellamy (1961) studied isolated gills incubated in tap and sea water, but he did not attempt to perfuse the arches. Most investigators have consequently chosen to use intact fish for gill ion transport studies. This choice puts severe restrictions on ways the internal compartment can be manipulated and precludes the use of precisely known concentrations of inhibitors, hormones and other introduced materials. The above difficulties together with the desirability of permuting the internal bathing solution of the gill made a new approach seem attractive and potentially worth while. Briefly, it is based on the incorporation of labelled sodium into gill arches which had been perfused in situ, removed from the fish, and individually exposed to an NaCl medium. The procedure is based on methods devised by Biber & Curran (1970) in their studies of ion transfer across the mucosal border of frog skins. It has two advantages over the perfused gill: first, it is far less time-consuming, and second, it enables one to use fish much smaller than required for a successful gill perfusion. METHODS Rainbow trout (Salmo gairdneri), g body weight, were supplied by the Humboldt State University experimental fish hatchery. They were kept unfed at 11 C in 0-2 mm-nacl tap water for at least 6 days prior to use. Preparation of the animals and experimental procedures were done in a constant temperature room at 9 C. Perfusion of the gills was done before individual arches were removed. Fish were first anaesthetized in o-i % tricaine methane sulphonate buffered with KHC0 3. The ventral aorta was exposed by a mid-ventral incision the length of the body cavity, and the dorsal aorta was then cut. The perfusion solution was injected into the ventral aorta with moderate pressure by means of a 20 ml syringe and a 23 gauge needle. The effectiveness of the perfusion was easily checked by watching the colour change of the gills. Occasional incomplete perfusions were discarded. Individual gill arches were removed by cutting through the arch at the dorsal and ventral roots. They were then rinsed twice in distilled water and placed on

3 Ammonium and sodium fluxes in isolated trout gill Time of exposure to 1 mm-naci (s) Fig. 1 Sodium influx vs. time in isolated gills. Perfusates are as noted, and N is given for each point. Vertical bars are S.E.M. damp filter paper. Generally, the three anterior arches on each side were used. Following removal of all the arches, each one in turn was grasped with a haemostat, rinsed rapidly in distilled water, and exposed to a 1 -o mm- w NaCl solution with gentle agitation for a timed interval. After this exposure and with a delay of less than 1 s, the arch was rinsed in a sequence of three distilled water vessels; each rinse lasted about 0-5 s. The arch was then gently blotted between two layers of absorbent paper. The filament portion of each arch was removed, weighed to the nearest o-ooi g ( g, range of 10 weighings), and placed in a tube for gamma counting. Previous tests indicated that no appreciable increase in accuracy was gained from solubilizing the tissue as long as care was taken to ensure the filaments were all at the bottom of the tube. The influx of sodium was calculated from counts per min in the filaments of each arch, and the specific activity of the exposure medium. Statistical significance was tested by Student's t test. The basic perfusion medium was Cortland's Ringer solution (Wolf, 1963). Sucrose was substituted for NaCl to give a low sodium solution for some experiments, and ph changes for other experiments were achieved by modifying phosphate and bicarbonate concentrations of the basic Ringer solution. Ouabain, Diamox and NH 4 C1 were added for appropriate experiments. All solutions were filtered before use.

4 520 T. H. KERSTETTER AND M. KEELER Internal solution io-* M-NH«C1 io" 1 M-NI^Cl io" 1 M-NH,C1 IO" 1 M-NH^CI Table i. Sodium influx in isolated gills: control values, stimulation by NH t +, and inhibition by amuoride External solution i mm-nacl i mm-nacl i mm-nacl 15» 5-3 ±o-3 (12) 72 ±o-6 (2) 13-4! 17 (6) Exposure time 30 a 81 ±03(8) 130±3-6 (3) 237 ±30 (6) i mm-nacl, amiloride i mm-nacl, amiloride 2-6±0-3 (3) (3) 3-2±o-6( 3 ) 4-o±o-9(3) i mm-nacl, 28 ±02(3) amiloride 4-3 ±o-s (3) Sodium influx in n-equiv. x o-i g~ 1 ±s.e.m. 60 s 14-1 ±o-8 (10) 2i-2±3-3(3) 36-6±46 (6) 5-3±03 (3) 6-7±i-9(3) 46±03 (3) RESULTS The initial control measurements, using gill arches perfused with unmodified Cortland's Ringer solution, were made for 15, 30 and 60 sec periods of exposure to the M NaCl external solution. Sodium uptake was linear with time (Fig. 1), and is expressed as n-equiv. Na+xo-i g" 1 of gill tissue. Table 1 summarizes control data. It is worth noting at this point that a 100 g rainbow trout has about 1 g wet weight of gill filaments, and that the sodium uptake (Table 1) for 1 min translates to an influx of about 8-5 /i-equiv. x tr 1 x 100 g" 1 body weight. Kerstetter et al. (1970) gave a Na + influx value of about 20 /t-equiv. x h" 1 x 100 g" 1 for intact rainbow trout, a figure not too different from the one given above. Stimulation and inhibition of the transport system Ammonium chloride, added to the perfusate in concentrations of i-o and io-o mm, markedly stimulated the Na + uptake (Fig. 1). At io" 3 M-NH^Cl, uptake was still linear with time at 1 min, but with io~ 2 M-NH 4 C1 there was an apparent loss of linearity. Note also the obvious concentration dependence. Amiloride, known to inhibit Na + influx in intact trout (Kirschner, Greenwald & Kerstetter, 1973), was added to the external medium at a concentration of 6x io^gxml" 1. Gill arches were first exposed for 5 s to the inhibitor in a non-labelled NaCl solution, then to the ^NaCl medium also containing amiloride. Sodium uptake, as expected, was markedly depressed (Table 1). The unequivocal inhibition shown in these experiments, both with and without added NH 4 +, indicated that the Na + uptake measured here is the result of a transport system, and not a simple diffusion process. Internal ph effects In the free base form, ammonia penetrates cell membranes rapidly, and in fact the model of Maetz (1971) has NH 3 diffusing across serosal membranes, accepting a proton in the cell, and exchanging for Na + at the mucosal border as NH 4 +. The

5 Ammonium and sodium fluxes in isolated trout gill 521 Table 2. The stimulation of Na+ influx by internal NH + t in relation to ph Internal solution 10-' M-NH.C1 lo" 1 M-NH.C1 PH ±o-i (4) 4'3 ± '4 (4) 7'S±I-I (6) 6i±o- S (6) Exposure time Sodium influx in n-«quiv, x o-i g -1 ±s.b.m. 30 s 8-3 ±07 (4) 7-4 ±0-3 (4) i2-o±i-6(6) (6) Table 3. Stimulation of Na+ influx by low-na+ perfusate, and the effect of adding NH t + Exposure time Internal solution Month of experiment 15 s 30 s Sucrose Ringer February (6) (6) Sucrose Ringer, February (8) (8) io-«m-nh.c1 Sucrose Ringer April '8io-4(9) (8) Sucrose Ringer, April (6) (8) io-» M-NH.C1 Sucrose Ringer August 1074 II-61I-I(6) (6) Sucrose Ringer, August (6) (6) io"» M-NH4CI Sodium influx in n-equiv, xo-i g ph dependence of the NH 3 ^ NH 4 + equilibrium (pk ~ 9-3) oflered a ready means of testing that hypothesis, since at ph 8 ten times as much NH 3 should be available as at ph 7 for any given concentration of NH 4 C1. Accordingly, Na+ uptake at ph 7 and ph 8 was tested with and without IO~ 3 M-NH 4 C1 in the perfusate. Table 2 summarizes the results. Clearly there was no stimulation of the transport system at the higher ph. In fact the results, although statistically not significant, indicate a depression of Na + uptake at the higher ph and would tend to support a conclusion that in all stages of the transport process, ammonia participates only in the protonated form. Effects of low-sodium perfusate A stimulation of Na + influx was also evident in gills with the low-sodium perfusate (Table 3). One puzzling result of these experiments was the difference in the degree of transport augmentation between the 2-75 and 4-75 groups. Although the reason is unknown, it should be emphasized that such seasonal differences only appeared in the sucrose-ringer perfused groups. Results with and with all NH 4 +-perfusates were remarkably constant. The differences between the 2 sucrose- Ringer perfused groups helped reveal another interesting result, that ammonia stimulation and the low-sodium effect are not additive. With io~ 2 M-NI^Cl in the perfusate, Na+ influx reached the same level, not only in the 2 sucrose-ringer groups

6 522 T. H. KERSTETTER AND M. KEELER Table 4. Sodium influx in isolated gills following a 15 or 20 min delay before exposure to the ^NaCl medium Delay before Exposure time Internal exposure solution (min) 15 s ±0-2 (a) 8-4 ±03 (2) ±0-4 (2) (2) (6) (6) io" 1 M-NH 4 C '(3) 127! 1-9' (3) io~«m-nr^cl Sucrose Ringer I3 - O± 1-2(6) (6) Sucrose Ringer li-6*(3) '(3) Sucrose Ringer, (8) (8) Sucrose Ringer, 20 ii'2lo-s 1 (4) 20-4! 1-5' (4) Sodium influx in n-equiv. xo-i g * * ' Paired differences from controls significant at o-oi, 0-05, o-io levels respectively. (Table 3), but also in the NaCl-Ringer groups (Tables 1, 5). At a lower internal ammonia concentration, io" 3 M, there was no apparent NH^4" effect with sucrose Ringer (Table 3), even though that level of NH 4 C1 stimulated transport when added to an. These results indicate that the stimulation of Na + uptake by ammonium ions is not superimposed upon the low-sodium stimulation. The two treatments thus may share, in part, a common mechanism. Effects of delayed exposure to the external medium The short-term viability of isolated gill arches was tested by delaying, for min, the exposure of some arches to the ^NaCl medium. Gill arches from the same fish, exposed to the external medium immediately after removal, served as controls. Table 4 summarizes the results. The controls showed no loss of activity following the 15 min delay. But, when IO~ 2 M-NH 4 + was added to the perfusate, some of the NH^-induced increment was lost in the delayed arches. Note that this group still retained greater transport activity than the undelayed NaCl controls had. The low-sodium, NH 4 +-free gills lost activity after a 20 min delay, as did their io~ 2 M-NH 4 + counterparts, but the decrement was about the same in both cases. Effects of ouabain and acetazolamide Ouabain (widely recognized as an inhibitor of the Na + -, K+-stimulated ATPase in a variety of tissues) was tested at 5 x io~* M with and without NH 4 + in the perfusate. With one possible exception, it was without effect (Table 5). The exception was in a group of NH^-free gill arches delayed 15 min before exposure to the ^NaCl. A paired t test, comparing immediately exposed gills to delayed gills from the same fish, showed that the slight decline in activity was significant (P < o*oi). But this was true only in the 15 sec exposure group, hence interpretation of that result is difficult. Diamox (acetazolamide), also known to inhibit Na + influx in the trout (Kerstetter et al. 1970), was tested at concentrations up to io" 3 M in NaCl

7 Ammonium and sodium fluxes in isolated trout gill 523 Table 5. Effects of ouabain anddiamox on sodium influx in isolated gill arches Tntprnfll ( XlllCillul \ ixoosurc solution (min) * ouabain * IS uuuijain * io-» M-NH.+ ouabain * io" 1 M-NH,+ ouabain io~«m-nh 4 C1 io"«m-nh 4 C1 diamox Sucrose Ringer io- 1 M-NH 4 C1 diamox diamox Delay before IS IS ±0-8(3) 5- ±o-7t(3) 11-3 ±1-8(4) 70 ±0-8(4) ia-3±i-a(6) 7-4±o- 4 t(6) i3'3±o-7 (5) 7-4 ±0-6(4) 6-6±o-s( 5 ) Exposure time x ±06 (3) 94 ±2-0(3) 21-2 ±2-7 (4) 14-2 ± 1-5(4) i9-8±i- 7 (6) n-3±o-8t(6) 23'4±i-8(s) ia-4±o-5 (4) io-i±o-3t (5) See Tables I and < I for ouabain-free controls. t Paired differences, o-8 ±0-5, P < X Significantly different from diamox-free controls, P < o-oi. See Table: 3 for control values. Fish used for this : single experiment were from a different genetic stocl Sodium influx in n-equiv. xo-i g~ 1 ±s.e.m IS 12 6os o±i-4(3) 6 ±0-9 (3) 36 8 ±3-6 (4) 25' 9 ±4-5 (4) perfusate with io~ 2 M-NH 4 +. An incomplete but very apparent inhibition of Na+ uptake resulted (Table 5). Diamox was tested with NH 4 +-free perfusate in another experiment. Inhibition was detectable only in the 30 sec exposure group, and was slight. (Control values in that experiment are different; the fish were of a different genetic stock than those used in all other experiments, and this probably accounts for the difference.) The third type of Diamox test was with sucrose Ringer, io~ a M-NH 4 C1. There was no effect, even at io" 3 M-Diamox (Table 5). DISCUSSION This study represents a new approach to investigations of the freshwater teleost gill, so it is necessary first to establish the validity of the preparation. The most convincing point in this regard is Na + influx in isolated gills in relation to the influx in intact fish, assuming 1 g gill tissue (wet weight) per 100 g body weight. With ammonia-free perfusate, the estimated value is 8-5/i-equiv. x 100 g^xh" 1 ; with io" 3 M-NH 4 C1 in the perfusate, the influx is about 12. Sodium influx in anaesthetized, intact rainbow trout in a 1 mm-nacl medium varied between 18 and 23 /t-equiv. x 100 g -1 x h -1 in a previous study (Kerstetter et al. 1970). Amiloride reduced Na+ uptake in isolated gill arches to 4nanequivalentsxmin~ 1 xo > i g" 1, which translates to 2'4/i-equiv. x h -1 x 100 g" 1 in intact fish. Kirschner et al. (1973) gave a value of

8 524 T. H. KERSTETTER AND M. KEELER about 6 # 5/<-equiv. xh -1 x 100 g" 1 for intact trout exposed to that inhibitor, they used a much lower concentration. Finally, the lack of deterioration in control gill arches delayed for 15 min before exposure to the ^NaCl indicates that we are not dealing with a dying preparation, at least in the time of these experiments. It appears that the isolated trout gill, used as described in this report, can validly be used to answer questions relating to ion transport mechanisms in teleost gills. The first of these questions is whether or not a Na+/K + exchange at the basal and lateral membranes is an integral link in the transepithelial transport process. Our data indicate that if such is the case, it is not ouabain-sensitive (Table 5). Since other workers have shown that the Na +, K + ATPase of salmonid gills is inhibited by ouabain (Pfeiler & Kirschner, 1972), it seems safe to conclude that Na+/K + exchange is not involved. It can be argued that the short incubation times of the gill arches in the ^NaCl medium did not allow for significant throughput of sodium, and that we really measured only uptake across the mucosal border. But the high rates of sodium uptake, especially with NH 4 + present internally, and the linearity shown in Fig. 1 both indicate near normal transport behaviour and thus, by implication, a transepithelial movement. The delivery of ouabain via the vascular system should rule out doubts about accessibility of transport sites to the inhibitor, and the 15 min delay incorporated into some of the ouabain work ensured adequate time for an effect to be manifested. Shuttleworth, Potts & Harris (1974) achieved immediate results with ouabain in the perfused gills of a saltwater fish, Platichthys flesus. Our results disagree with two recent reports, both of which give evidence for a Na + /K + exchange component in the transepithelial sodium movement of freshwater teleost gills. Richards & Fromm (1970) reported ouabain inhibition in continuously perfubed trout gills, and Shuttleworth & Freeman (1974) reported an apparent inhibition of sodium transport in perfused freshwater eel gills when K + -free perfusate was used. But neither of those investigations used radioactive tracer, so they measured only net fluxes. Teleost gills are complex structures with several different cell types and a large area of respiratory epithelium. Blocking Na + /K + exchange throughout the gill would almost certainly lead to sodium loading of all cells and consequently to an increased sodium leak through their apical borders. The result would be a decrease in net flux even if the influx via the specialized transport cells were unaffected. The second question concerns the form in which ammonia crosses the inner epithelial border. The presently accepted model of the freshwater teleost gill sodium transport system is based on diffusion of NH 3 across the basal membranes (Maetz & Garcia-Romeu, 1964; Maetz, 1971). Thus with a given concentration of NH4CI in the extracellular compartment one would expect a higher transmembrane movement at a higher ph, since more of the total ammonia would be in the free base form. The extension of this argument is that if more NH 3 enters the transporting cells, more NH^ will be available for Na+/NH 4 + exchange. The data in Table 2 clearly indicate that free base concentration is not a consideration, and this is especially evident from the wide ph range (one unit) employed. It should be emphasized here that the data in Table 2 were from experiments using io" 3 M-NH 4 C1, not a saturating concentration. Thus any increases in available ammonia should have been reflected by an increase in transport rate. We must conclude that it is NH 4 + which is entering the cells. But the ionic form cannot do so in measurable amounts unless accompanied by an

9 Ammonium and sodium fluxes in isolated trout gill 525 Interstitial and vascular compartment /" Transport cell \ External medium 3 2 /l-s Fig. 2. Idealized diagram of gill transport cell. See text for discussion. anion or exchanged for a cation. It is unlikely that Cl~ is accompanying NH 4 +, since the gill transports Cl~ in the opposite direction. The other possible candidate, bicarbonate, is also tentatively ruled out by the experiments employing perfusates of different ph. The ph 7 Ringer solution was low in bicarbonate (less than half the normal concentration) yet no difference in Na + influx resulted. We are left with the tentative conclusion that NH + 4 is exchanging for Na+ at the serosal membrane. Although the comparative rates of NH 4 +-stimulated transport at ph 7 and ph 8 are the most compelling evidence for a serosal membrane exchange, data from an additional experiment, the treatment with low-na + (sucrose) Ringer solution with and without NH + 4, indirectly support that conclusion. As noted above, ammonia-free sucrose Ringer stimulated Na + uptake (Table 3). The most likely explanation for this is lowered cell sodium resulting from reduction of the internal leak pathway, faz (Fig- 2). This could conceivably affect the throughput of tracer sodium in two ways. The first is by increasing influx, f x -*%, at the outer membrane, and the second by an apparent increase in the basal membrane step, /a-^, because of reduced competition for transport sites by non-labelled sodium. The second possibility implies that f^z is the limiting step in the transport system, and if this is the case then ammonia must be acting on the system at the basal membrane. If the first choice is true, then / 1 _ >2 is the limiting step and would seem to be sensitive to intracellular [Na + ] in order to account for the observed stimulation of Na+ influx under low internal sodium conditions. But the stimulation by NH + 4 is not additive with that of low-sodium perfusate. When io~ 2 M-NH 4 C1 was added, Na+ uptake reached the same level as in the Cortland's Ringer, io~ 2 M-NH 4 Cl-perfused, gills. This was true of both groups of the sucrose Ringer experiments, high and low stimulation (Table 3). One explanation is that the maximum possible transport rate was reached for gills in a 1 mm-nacl external medium. An alternative is that both internal NH + 4 and a low-sodium perfusate stimulate transport by the same mechanism, reducing intracellular [Na+]. In further support of this argument, internal NH 4 C1 at io~ 8 M had no stimulatory effect when added to sucrose Ringer (Table 3), even though that concentration of ammonia in stimulates transport. This implies that the ammonia effect was completely masked by the low-sodium effect, giving additional evidence that the two forms of transport stimulation are not additive and hence may have similar operational modes. The most probable means for ammonia to reduce cell sodium would seem to be a Na+/NH 4 + exchange at the basal membrane.

10 526 T. H. KERSTETTER AND M. KEELER In two recent reports, Maetz (1972, 1973) showed that ammonia fluxes through goldfish gills could not be adequately explained by diffusion gradients, except when sodium was unavailable for transport, and that when Na+ was available the gills were permeable to NH 4 +. He also confirmed once more the stimulation of Na+ influx which results when the animal is given an ammonium salt load. But he was not able to demonstrate a stoichiometric relationship between sodium influx and ammonia efflux. Specifically, neither he nor Kerstetter et al. (1970) could show variations in ammonia output when sodium influx was changed by varying external [Na + ]. And this was true for both increases and decreases in Na+ influx. If our hypothesis of a Na+/NH d + exchange at the basal membrane is correct, the absence of stoichiometry between the two ions might be easier to explain. Ammonium ions, entering the transport cells, need not a priori be passed directly outward to the medium. One could speculate that at least some of the NH 4 + is incorporated into the amino acid pool via the amination of a keto glutarate. The high levels of glutamic dehydrogenase reported for teleost gills (Goldstein & Forster, 1961) provide a basis for such speculation, and indeed Schoffeniels (1968) has stated that in crustacean gills the equilibrium is in the direction of glutamic acid formation. Glutamate synthesis need not account for all incoming NH 4 +, but even if only a fraction were so used, the Na + -NH 4 + stoichiometry would be lost. Finally, the effects of Diamox should be considered. The greatest inhibition was with gills perfused by with NH 4 + added. Ammonia-free with Diamox showed a much smaller depression of transport, and the gills perfused by sucrose Ringer with NH 4 C1 were unaffected by the inhibitor. The results are puzzling in view of the unequivocal inhibition of Na + influx in intact teleosts (Maetz & Garcia-Romeu, 1964; Kerstetter et al. 1970). The explanation may be related to the duration of these experiments. A delay period, unfortunately not employed in these treatments, might be necessary for the inhibitor to achieve sufficient intracellular levels to fully manifest its effects. The total absence of an effect in the sucrose Ringer-NH 4 C1~ perfused gills is even more puzzling. Although one might speculate that a more negative intracellular potential in these gills effectively blocked entry of the weakly acidic acetazolamide molecule, the high pk (7-4) makes this seem unlikely. The paradox thus remains unexplained. Supported by NSF grant GB to T.H.K. REFERENCES BELLAMY, D. (1961). Movements of potassium, sodium, and chloride in incubated gills from the silver eel. Comp. Biochem. Pkyiiol. 3, BIBER, T. U. L. & CURRAN, P. F. (1970). Direct measurement of uptake of sodium at the outer surface of the frog skin. J. gen. Pkytiol. 56, GOLDSTEIN, L. & FORSTER, R. P. (1961). Source of ammonia excreted by the gills of the marine teleost Myoxocepkalut icorpius. Am. J. Phytiol. 200, KAMIYA, M. & UTIDA, S. (1969). Sodium-potassium-activated adenosinetriphosphatase activity in gills of fresh-water, marine and euryhaline teleosts. Comp. Biochem. Phytiol. 31, KERSTETTER, T. H. ( KIRSCHNER, L. B. & RAFUSE, D. (1970). On the mechanisms of sodium ion transport by the irrigated gills of rainbow trout (Salmo gavrdneri). J. gen. Phytiol. 56, KIRSCHNER, L. B. (1969). Ventral aortic pressure and sodium fluxes in perfused eel gills. Am. J. Pkytiol. 317,

11 Ammonium and sodium fluxes in isolated trout gill 527 KIRSCHNBR, L. B., GRBBNWALD, L. & KBRSTBTTER, T. H. (1973). Effect of amiloride on sodium transport across body surfaces of freshwater animals. Am. J. Pkysiol. 324, MAETZ, J. (1971). Fish gills: mechanisms of salt transfer in fresh water and sea water. Phil. Tram. R. Sot. B 36a, MAETZ, J. (1972). Branchial sodium exchange and ammonia excretion in the goldfish Carasrius auratus. Effects of ammonia-loading and temperature changes. J. exp. Bio!. 56, MAETZ, J. (1973). Na + /NH + 4, Na + /H + exchanges and NH, movement across the gill of Carattius auratus. J. exp. Biol. 58, MAETZ, J. & GARCIA-ROMEU, F. (1964). The mechanism of sodium and chloride uptake by the gills of a freshwater fish, Carastius auratus. II. Evidence for NH + 1 /Na + and HCO,~/C1~ exchanges. J. gen. Pkysiol. 47, PFEILEH, E. & KIRSCHNEB, L. B. (1972). Studies on gill ATPase of rainbow trout (Salmo gairdneri). Biochim. biopkys. Acta 383, RICHARDS, B. D. & FROMM, P. O. (1970). Sodium uptake by isolated-perfused gills of rainbow trout (Salmo gairdneri). Comp. Biochem. Pkysiol. 33, SHUTTLEWORTH, T. J. (197a). A new isolated perfused gill preparation for the study of the mechanisms of ionic regulation in teleosts. Comp. Biochem. Pkysiol. 43A, SHUTTLEWORTH, T. J. & FREEMAN, R. F. H. (1974). Factors affecting the net fluxes of ions in the isolated perfused gills of freshwater Angidlla dieffenbackii. J. comp. Pkysiol. 94, SHUTTLEWORTH, T. J., POTTS, W. T. W. & HARRIS, J. N. (1974). Bioelectric potentials in the gills of the flounder Platichtkys flesus. J. comp. Pkysiol. 94, SCHOFFENIALS, E. (1968). The control of intracellular hydrogen transport by inorganic ions. Archs int. Pkysiol. 76, WOLF, K. (1963). Physiological salines for freshwater teleosts. Progve Fish Cult. 35, ZAUGC, W. S. & MCLAIN, L. R. (1970). Adenosinetriphosphatase activity in gills of salmonids: seasonal variations and salt water influences in coho salmon, Onchorhynchus kisutch. Comp. Biochem. Pkysiol. 35,