UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Proton-atpase in fish gills Lin, Hong 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1993_spring_phd_lin_hong.pdf [ 4.22MB ]
JSON: 831-1.0086373.json
JSON-LD: 831-1.0086373-ld.json
RDF/XML (Pretty): 831-1.0086373-rdf.xml
RDF/JSON: 831-1.0086373-rdf.json
Turtle: 831-1.0086373-turtle.txt
N-Triples: 831-1.0086373-rdf-ntriples.txt
Original Record: 831-1.0086373-source.json
Full Text

Full Text

PROTON-ATPASE IN FISH GILLSByHONG LINB.A.Sc., Zhejiang University, 1986M.Sc., The University of British Columbia, 1989A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDEPARTMENT OF ZOOLOGYWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAMARCH 1993© Hong Lin, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of ^tro o(3yThe University of British ColumbiaVancouver, CanadaDate (t) ^ /3DE-6 (2/88)ABSTRACTThe cellular mechanisms responsible for branchial protonexcretion and sodium absorption in freshwater rainbow trout,Oncorhynchus mykiss (Walbaum), were investigated by monitoringthe proton excretion rate and determining the proton ATPaseactivity in gill tissue under different inhibitor treatments andenvironmental conditions. Evidence for the existence of anelectrogenic proton pump in fish gills was accumulated.Branchial proton excretion rate was estimated by measuringthe total CO 2 , total ammonia, pH and buffer capacity of theinspired and expired water of the fish. The net proton excretionacross fish gills was resistant to 0.1 mM amiloride which couldcompletely abolish the branchial sodium uptake, indicating that aNa +/H+ exchanger was not responsible for this proton transport.Branchial proton excretion, however, was sensitive to vanadateand acetazolamide, suppressed by low external water pH and sodiumlevels, and stimulated by elevated ambient Pco 2 . All thesecharacteristics are typical for proton transport mediated by anelectrogenic proton pump, as demonstrated in frog skin, turtlebladder and mammalian kidney.N-ethymaleimide-sensitive ATPase activity was measured incrude homogenates of gill tissue from rainbow trout using acoupled-enzyme ATPase assay in the presence of EGTA, ouabain andazide. This NEM-sensitive ATPase activity, determined to be about1.5 umol/mg.pr./h. at 15 °C for freshwater trout, is alsoiiinhibited by other proton-ATPase blockers such as DCCD, DES,PCMBS and Bafilomycins. It was concluded, therefore, that theNEM-sensitive ATPase activity was generated by a proton-translocating ATPase. Since this NEM-sensitive ATPase was alsosensitive to the plasma membrane ATPase inhibitor vanadate, the11+-ATPase in fish gill was speculated to be a plasma membranetype. Sodium concentration in the external media was the primaryregulator of the H+-ATPase in fish gills, with low water sodiumlevels associated with high H +-ATPase activity. High externalcalcium concentration and plasma cortisol levels had a markedstimulating effect on H +-ATPase activity in fish gills only whenthe water sodium level was low. Thus the major role of the H +-ATPase in the gill epithelium is to facilitate Na+ uptake fromfresh water. The H +-ATPase in the gills also plays a role inacid-base regulation.It was concluded that an electrogenic proton pump (H +-ATPase) indirectly coupled to a sodium conductive channel in thegill epithelium is the ion transport pathway which mediatesproton excretion and energies sodium absorption in freshwaterfish.iiiTABLE OF CONTENTSPageAbstract^ iiTable of Contents^ ivList of Tables vList of Figures viList of Abbreviations viiiAcknowledgements^ ixGeneral Introduction 1Section I: Branchial proton excretion in rainbow trout^26mediated by an electrogenic proton pump.Chapter 1. Inhibitor sensitivity of branchial proton^26excretion and in vivo  evidence for theexistence of a proton pump.Summary^ 27Introduction 28Materials and Methods^ 30Results 38Discussion^ 47Chapter 2. Effects of external water pH, Pco 2 and^57ion levels on branchial proton excretionmediated by proton pump.Summary^ 58Introduction 59Materials and Methods^ 61Results 64Discussion^ 76Section II: Proton-ATPase activity in gill tissue^80of rainbow trout.Chapter 3. Inhibitor sensitivity and classification^80of the proton-ATPase in gill tissue.Summary^ 81Introduction 82Materials and Methods^ 84Results 90Discussion^ 97Chapter 4. Environmental and hormonal regulation of^101the proton-ATPase in fish gills.Summary^ 102Introduction 103Materials and Methods^ 105Results 108Discussion^ 125General Discussion 130References^ 145ivLIST OF TABLESTableTable1.2.Total CO2 excretion by fish.Comparison between the reduction in the rateof ammonia excretion and the reduction in therate of net proton excretion under amiloridetreatment.4453Table 3. Test of inhibitor interference with PK or LDHin the NADH oxidation reaction.88Table 4. Effects of inhibitors on ATPase activity incrude homogenate of trout gill tissue.95Table 5. Plasma pH and NEM-sensitive ATPase activity ingill tissue of NH4C1-injected rainbow trout.124vLIST OF FIGURESFigure 1. Simplified cross section of gill epithelium of^2freshwater fish with the gas exchange and iontransport pathways.Figure 2. A schematic view of the outermost living cell^23layer of frog skin, with two distinct cellularpathways for Na + absorption.Figure 3. The recirculating system with a black chamber.^32Figure 4. Bicarbonate concentration differences between^39inspired and expired water of rainbow trout undercontrol, amiloride, vanadate and acetazolamidetreatments.Figure 5. Net proton excretion across the gill epithelium 41of rainbow trout under control, amiloride, vanadateand acetazolamide treatments.Figure 6. Branchial ammonia excretion rates of rainbow^45trout under control, amiloride, vanadate andacetazolamide treatments.Figure 7. Schematic representation of gas and ion^50transport across the gill epithelium of rainbowtrout.Figure 8. The relationship between the branchial net^65proton excretion and the inspired water pH.Figure 9. The relationship between branchial net proton^67excretion and expired water carbon dioxide levels.Figure 10. Branchial proton excretion rate of rainbow^69trout in control and sodium-free water.Figure 11. Branchial proton excretion rate of rainbow^72trout in control, sodium-free water andsodium-free water plus 0.1 mM 9-anthroic acid.Figure 12. Branchial proton excretion rate of rainbow^74trout in control and high calcium (10 mM) water.Figure 13. NEM-sensitive ATPase activity in the crude^91homogenates of trout gill tissue in response tovarious concentrations of NEM.viFigure 14. DCCD-sensitive ATPase activity in the crude^93homogenates of trout gill tissue in response tovarious concentrations of DCCD.Figure 15. NEM-sensitive ATPase activity in the gill^109tissue of rainbow trout acclimated to variousNa+ and Ca ++ levels in the external media for10-14 days.Figure 16. Ouabain-sensitive ATPase activity in the^111gill tissue of rainbow trout acclimated tovarious Na + and Ca ++ levels in the externalmedia for 10-14 days.Figure 17. Plasma cortisol concentration in freshwater^113and seawater rainbow trout after 7 days ofchronic cortisol treatment.Figure 18. NEM-sensitive ATPase activity in the gill^115tissue of freshwater and seawater rainbow troutafter 7 days of chronic cortisol treatment.Figure 19. Ouabain-sensitive ATPase activity in the^118gill tissue of freshwater and seawater rainbowtrout after 7 days of chronic cortisol treatment.Figure 20. NEM-sensitive ATPase activity in the gill^120tissue of freshwater rainbow trout during 48hours of hypercapnia treatment and 24 hoursrecovery.Figure 21. NEM-sensitive ATPase activity in the gill^122tissue of freshwater rainbow trout duringcontrol and 16 days exposure to soft andhard alkaline water.Figure 22. Hypothetical model of the gas and ion^131transport pathways in gill epithelium offreshwater rainbow trout.Figure 23. The relationship between the NEM-sensitive^140ATPase activity and the ouabain-sensitive ATPaseactivity in freshwater rainbow trout underdifferent treatments.Figure 24. The relationship between the NEM-sensitive^142ATPase activity and the ouabain-sensitive ATPaseactivity in seawater-adapted rainbow trout undercontrol and cortisol treatments.viiLIST OF ABBREVIATIONS9-AA^9-anthroic acidCA^carbonic anhydraseCHAPS^3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonateDCCD^N,N'-dicyclohexylcarbodiimideDES^diethylstilbestrolEGTA^ethyleneglcol-bis-(B-aminoethylether)N,N,N',N'-tetraacetic acidLDH^lactic dehydrogenaseNBD-Cl^7-chloro-4-nitrobenz-2-oxa-1,3-diazolePCMBS^p-chloromercuri-benzenesulfonatePK^pyruvate kinaseSITS^4-acetamido-41-isothiocyanatostilbene-2-21-disulfonicacidviiiACKNOWLEDGEMENTSI would like to thank Dr. David Randall, my supervisor, forhis inspiration and support during these studies. It has trulybeen an enjoyable experience. I am grateful to my supervisorycommittees Dr. J.E. Phillips, Dr. G.K. Iwama, Dr. A.M. Perks andDr. J.D. McPhail for valuable comments and discussion on themanuscript. My appreciation extends to all the members in Dr.Randall's lab, specially Mark Shrimpton, Colin Brauner, Sumi Aotaand Nick Bernier, for their pleasant company and help. I alsolike to thank Dr. K. Iwata and Mr. J. Wilson who collaborated onthe alkaline study in Chapter 4. Appreciation is also expressedto Dr. Hochachka's Lab for generous use of theirspectrophotometer.I thank my husband Lee, and my parents for always beingthere for me.ixGENERAL INTRODUCTIONGas exchange and ion transport in fish gill epitheliaThe gills of the fish are the primary site of gas exchange,acid-base regulation and osmoregulation (Maetz and Garcia-Romeu,1964; McDonald et al, 1989; Randall, 1990). The gill lamellarepithelium is permeable to 02 , CO2 and NH3 (Figure 1), and 100% ofoxygen uptake, 90% of carbon dioxide and at least 60% of ammoniaexcretion relies on passive diffusion of the gases (Randall,1990). Ion transfer across the gill epithelium, on the otherhand, is usually mediated by active or passive ion transportprocesses (Figure 1). In freshwater teleosts, branchialabsorption of Na + and Cl - from the hypo-osmotic environmentcompensates for the constant loss of ions from the body bydiffusion. The mechanisms for Na + and Cl - uptake are independent(Maetz and Garcia-Romeu, 1964) and the counterions for Na + andCl - , presumably fr(N11 4+) and HCO3 - respectively, are extruded tothe water simultaneously (Perry and Randall, 1981; McDonald etal, 1989). These electroneutral ion exchange pathways have beensuggested to be involved in acid-base regulation and a portion ofcarbon dioxide and ammonia excretions.Carbon dioxide (CO 2) excretion acidifies water as it passesover the gills, due to the hydration of CO 2 forming HCO3 - and H.(Wright et al, 1986; Lin & Randall, 1990). Ammonia (NH3 )excretion on the other hand, can alkalize the expired water offish due to the conversion of NH3 to NH4+ , consuming protons. The1Figure 1. Simplified cross section of gill epithelium offreshwater fish with the gas exchange and ion transportpathways. ATP-driven pumps are denoted by half-filledcircles, ion exchanger by opened circles and passivediffusion by arrows. (Adapted from McDonald et al, 1989).2WATER PLASMA,E3RANCHIAL EPITHELIUMNa+ ^XH4 4-, H +CI -MCCD,z°,^©0°degree of acidification or alkalinization in expired waterdepends on the pH and buffer capacity of the external water, theCO2 and NH3 excretion rate and the amount of acid that isexcreted into the water by the fish.The gill lamellar epithelium separating the blood from theexternal water consists mainly of three cell types - mucous (alsocalled goblet), epithelial (also called pavement) and chloridecells (Laurent and Dunel, 1980). Mucous cells secrete mucous thatforms a thin layer on the gill surface. Epithelial cells arepermeable to respiratory gases such as 0 2 , CO 2 and NH 3 and play aprominent role in gas exchange. Chloride cells, also known asmitochondria-rich cells, house most of the energy consuming iontransport pathways such as Na +-K+-ATPase and are involved inosmoregulation by euryhaline teleost (Pisam, et al, 1987). Infreshwater fish the epithelium and chloride cells are joined bytight junctions, which act as a minimal barrier for diffusion ofgases, but has a high resistance to the transfer of ions andwater. Thus gill epithelia in freshwater fish are considered tobe "tight" epithelia (Sardet, 1980).This tight epithelium consists of two cell membranes.Studies of turtle bladder epithelia showed that the apicalmembrane has a very low permeability to H + , OH - and is capable ofgenerating an electrochemical potential gradient for protonseither by a proton pump or by an exchange that is driven by aconcentration gradient for another ion. The second cell membraneis the basolateral cell membrane which is permeable to HCO 3 - , OH -4and H+ and permits the passive movement of these ions due to theexistence of transport proteins specialized in the transfer ofanions (Steinmetz, 1985).Chloride cells in the gill epithelia manage the role ofacid-base regulation and osmoregulation because they have highmitochondrial activity and contain primary and secondary activeion transport pathways. These pathways (see Figure 1) have beenpostulated to include:(1) Na +-K+-ATPase in the basolateral membrane. This primaryactive transport pathway maintains a low intracellular Na +concentration by continuously pumping Na + out to serosal side inexchange for K. For each ATP consumed, 3 Na + is pumped out and 2K+ is pumped in. K + can be replaced by H + or NH4+ (Evans et al,1989) and enter the cell through this pathway. Ouabain is thespecific inhibitor for Na +-K+-ATPase (Pedersen and Carafoli,1987). Ouabain applied to the extracellular surface of themembrane blocks potassium dependent Na +-K+-ATPase by competingfor the K + binding site. Ammonia excretion in Opsanus beta perfused head was reduced by 22% when 0.1 mM ouabain was added tothe perfusate, possibly due to the blockage of Na +-K+ (NH4+ )-ATPaseby ouabain (Evans et al, 1989).(2) C1 -/HCO3" exchange is located in the apical membrane,through which chloride is absorbed and bicarbonate excreted.Since the gill epithelium is impermeable to HCO 3 - (Perry et al.1982) and Cl - , it has been suggested that a portion of the carbondioxide entering the gill epithelium is hydrated, forming5bicarbonate, which is then exchanged for chloride across theapical membrane. The addition of 4-acetamido-4'-isothiocyanatostilbene-2-2'-disulfonic acid (SITS), (known to block C1 -/HCO3 -exchange in red blood cells), to the water results in a rise inblood pH in trout (Perry et al. 1981) and a 71% reduction in Cl -uptake (Perry & Randall, 1981). Although no more than 10% of thetotal carbon dioxide is transferred to water by this means, it isa primary pathway for chloride uptake in freshwater fish. SITSinhibits this anion transport by binding to a specific membraneprotein.Considerable evidence has been gathered to support theexistence of C1 -/HCO3 - exchangers in the basolateral cellmembranes of tight epithelia such as in turtle bladder and froggastric mucosa (Steinmetz, 1985). In turtle bladder, this Cl -/HCO3 - exchanger is paralleled by a Cl - conductive channel whichpermits the Cl - to recycle across the basolateral membrane sothat the exit of HCO3 - is associated with the transfer of anelectron.(3) Na +/H+ (NH4 + ) exchange in the apical membrane wasconcluded to be the major pathway for sodium uptake and protonexcretion and an optional excretory pathway for ammonia (Wrightand Wood, 1985). Amiloride is a very potent and specificinhibitor of sodium transport, including both the sodiumconductive channels and Na -E/H+ (NH 4 + ) exchange, in a wide varietyof cellular and epithelial transport systems, by competing forthe Na + transport site (Benos, 1982). Addition of amiloride to6water reduced sodium uptake in the gill epithelium of trout by84% (Perry and Randall, 1981) and 94% (Wright and Wood, 1985).The trout gill NH3 permeability coefficient of 6X10 -3 cms -1(Avella & Bornancin, 1989) is intermediate between valuesreported for the toad bladder and mammalian kidney tubule. It hasbeen suggested that ammonia excretion, although dominated by NH 3diffusion (Hillaby & Randall, 1979; Cameron & Heisler, 1983), isalso mediated by Na +/NH4+ exchange on the apical surface (Payan,1978; Wright & Wood, 1985). The stimulating effect of NH 4+ onsodium flux, however, could be explained in terms of a pH effectof the ammonia addition (Cameron & Kormanik, 1982), and Avellaand Bornancin (1989) and Heisler (1990) concluded that thebalance of evidence was against the presence of Na +/NH4+ exchangeacross the apical surface of trout gills. Avella and Bornancin(1989) considered the trout gill to be similar to other tightepithelia, such as frog skin and toad bladder, in that passivesodium uptake from water is indirectly coupled to an activeelectrogenic proton transport system.Possibility of the existence of le-ATPase in fish gillsThe most recent evidence for the existence of Na +/H+ (NH4+ )exchange in the fish gills comes from the in vivo studies ofWright and Wood (1985), who demonstrated that sodium influx andammonia efflux is coupled with approximately 1:1 stoichiometry inthe external water pH range of 4-9 and concluded that a flexiblecombination of NH 3 diffusion and Na +/NH4 + exchange was the major7mechanism of ammonia excretion. 94% of the Na + uptake by thegills of freshwater trout was inhibited by 0.1 mM amiloride inthe external media. However, without altering APNH 3 or A[NH 4+ ]across the epithelium, which means no ammonia transfer shouldhave been shifted from Na +/NH4+ exchange to NH 3 diffusion,amiloride exposure caused only 23% reduction in ammoniaexcretion. These authors provided no clear explanation for thecontrary result. The degree of coupling between sodium absorptionand ammonia excretion was also found to be rather loose ingoldfish, carp and trout in vivo (Payan, 1978). Sodium uptakefrom a diluted medium such as freshwater, will usually require anactive transport process. Although intracellular sodiumconcentration in the gill epithelium cell may be lowered by theNa+-K+-ATPase in the basolateral membrane to about 10 mM(intracellular sodium ion activity in frog skin epithelium wasmeasured to be 6.2 mmo1/1 using a double-barrelled ion sensitivemicroelectrodes, by Harvey and Ehrenfeld, 1986), it is stillhigher than that in fresh water (usually < 1 mM) and the sodiumelectrochemical gradient across the apical membrane could notpossibly drive the Na +/H+ exchange (Avella and Bornancin, 1989).Therefore, the argument for a Na +/H+ (NH 4+) exchange is not strong.The NH4+ concentration gradient might provide the driving forNa +/H+ (NH 4+) exchange. However, Avella and Bornancin (1989) re-examined the mechanism of ammonia excretion and sodium absorptionusing an isolated-perfused head preparation. They found thatammonia excretion was basically dependent on passive NH 38diffusion, sodium absorption was indirectly modified byintracellular pH and sodium uptake and ammonia excretion wereuncoupled. An indirect coupling of an electrogenic proton pumpand a passive sodium entry, instead of a NaVir(NH4+) exchange,was proposed by these authors as the mechanism of sodium andproton transport across the fish gill epithelium.NaVH+ exchange is only one of the two fundamentallydifferent mechanisms that has been proposed to account for Naand H' transport in opposite directions in ion-transportingepithelia. Na +/H+ countertransport, a passive and electroneutralprocess, exists in certain isolated cells such as sea urchin eggsand red blood cells, and in "leaky" epithelia such as rabbitgall-bladder, small intestine and renal proximal tubule. Theother mechanism is that of an active proton transport, which hasbeen described in "tight" epithelia such as the turtle and toadurinary bladder (Steinmetz, 1986; Al-Awqati, 1978), frog skin(Ehrenfeld et al, 1985) and mammalian renal collecting tubule(Ait-Mohamed et al, 1986). This proton transport is initiated bya electrogenic proton-translocating-ATPase which pumps hydrogenion to one side of the membrane and generates a negativepotential in the other side of the membrane that, in some cases,drives sodium flux via a sodium conductive channel (Ehrenfeld etal, 1985).The gill epithelium of freshwater fish is considered to be atight epithelium (Sardet, 1980; Avella and Bornancin, 1989). Itresembles freshwater frog skin and turtle urinary bladder9epithelia in many features. Functionally they are all capable ofacid-base regulation and electrolyte transport. Both freshwaterfrogs and freshwater fish are hyperosmotic to their aqueoussurroundings and face the problem of continual loss of body saltto the environment. The lost salts have to be replaced through anactive transport system across the epithelia in the skin of frogsor the gills of fish. Morphologically they contain analogous celltypes. The epithelium of turtle urinary bladder consists of basalcells, granular cells and carbonic anhydrase (CA)-rich cells. CA-rich cells, which contain numerous mitochondria andtubulovesicular membrane structures, are responsible for frsecretion in turtle bladder (Madsen and Tisher, 1985). It couldbe equivalent to the chloride cell in the gill epithelium.The apical membrane of frog skin comprises stratumgranulosum firmly interconnected by tight junctions that form abarrier separating the apical bathing solution from thebasolateral solution (Nielsen, 1982), similar to those in gilllamellae. The outermost living cell layer of the frog skinepithelium is composed of cuboidal granular (GR) cells andmitochondria-rich (MR) cells (Ehrenfeld et al, 1989). The lattercells can be identified by their long flask-like shape and narrowapical pole beneath the stratum corneum, the exclusivelocalization of carbonic anhydrase and the rod-shaped intra-plasma membrane particles. The intercalated cells in mammalianrenal collecting tubule also share similar characteristicsdescribed above for the MR cell of frog epithelium and10responsible for proton transport across kidney tubular epithelium(Brown et al, 1988). Bartels (1989) suggested that, in lamprey,the morphological characteristics of the epithelial cell wereidentical to those of the frog skin GR cell. Thus, it is logicalto suggest that the same ion transport mechanism, namely H +-ATPase indirectly coupled with a sodium channel, will exist ingill epithelium to account for the proton and sodium transport.Differences do exist between the gill epithelium and othertight epithelia. Carbonic anhydrase for example, is generallydistributed in the chloride cells, epithelial cells and mucouscells in fish gills (Rahim et al, 1988), but restricted to MRcells of frog skin (Rosen and Friedley, 1973) and CA-rich cellsof turtle bladder (Madsen and Tisher, 1985). The percentage ofchloride cells in gill epithelium of freshwater fish is only 1%(Perry and Walsh, 1989), whereas the ratio of MR to GR cells infrog skin is much higher and MR cells can represent between 13 to60% of the exposed apical area depending on salt adaptation(Ehrenfeld et al, 1989), since frog skin plays a minor role ingas transfer. The location and distribution of ion transportpathways in different cell types in fish gill epithelium,therefore, might not be exactly the same as that in frog skinepithelium.Background information about H +-ATPaseProton-Translocating-ATPases are integral membrane proteinsthat vectorially translocate W. from one surface to the other1 1(Pedersen and Carafoli, 1987). They can be classified into threecategories: Mitochondria H +-ATPase (F-type), vacuolar H +-ATPase(V-type) and plasma membrane H +-ATPase (P-type). Mitochondria H+-ATPase utilizes the proton gradient generated by the cytochromechain in the inner mitochondrial membrane for ATP synthesis andprovides the energy source for other ATPases in the cell.Vacuolar H+-ATPase and mitochondria H +-ATPase share a number ofimportant structural properties, including complexity of subunitcomposition and probably are derived from a common evolutionaryancestor (Forgac, 1989). Vacuolar H +-ATPase consumes ATP andactively pumps protons against an electrochemical gradient intothe vacuoles. The function of plasma membrane H +-ATPase issimilar to that of vacuolar H+-ATPase but it has a lowermolecular weight and a simpler subunit structure, with only 2subunits instead of 18 as in mitochondria H +-ATPase and 16 as invacuolar H+-ATPase (Forgac, 1989).The H +-ATPase in the plasma membrane of eukaryotic cells areclassified as phosphorylated ion motive enzymes because they forma covalent phosphorylated intermediate as part of the reactioncycle (Pedersen and Carafoli, 1987). Na +-K+-ATPase and Ca ++-ATPaseare also phosphorylated ion motive ATPases. Vanadate, atransition state analog of phosphate, inhibits P-type ATPase byblocking the formation of phosphorylated intermediates in theATPase.Plasma membrane H +-ATPase and vacuolar H +-ATPase couple tomitochondrial H +-ATPase in a master-slave relationship.12Mitochondria H +-ATPase functions obligatorily in the direction ofATP synthesis and supplies ATP to the plasma membrane H +-ATPaseand vacuolar H +-ATPase which function obligatorily in thedirection of ATP hydrolysis to transport protons actively(Pedersen and Carafoli, 1987).Many epithelial membranes have the capacity to transporthydrogen ions. In the "tight" epithelial membranes, such as inmammalian renal collecting tubules (Gluck and Al-Awqati, 1984;Ait-Mohamed et al, 1986), turtle urinary bladder (Steinmetz andAndersen, 1982; Steinmetz, 1985) and frog skin (Ehrenfeld et al,1990), the P-type or V-type proton-translocating-ATPase, alsotermed electrogenic proton pump, is responsible for the transportof protons. The proton pump is characterized by tight coupling tothe energy of ATP hydrolysis but not directly dependent onmovement of other ions. It plays a vital role in acid-base andosmotic homeostasis in animals.Proton pumps located in the luminal membrane of freshwaterturtle urinary bladder and apical membrane of freshwater frogskin have been extensively studied because the turtle bladder andfrog skin permit some major simplifications, eg. they can bestretched as a flat epithelium sheet between two bulk solutions,and the passive 1-1' permeability in these epithelia is so low thatnet 1-1 + transport rates can be used directly to study thebehaviour of the proton pump (Steinmetz and Andersen, 1982).The proton transport mediated by this active pump hasseveral important properties:13(1) The proton pump is electrogenic. In the short-circuitedturtle bladder, proton secretion takes place against a steep pHgradient and generates 40 mV negative intracellular potential(Steinmetz and Andersen, 1982). Proton excretion in frog skingenerates a negative potential (about -50 mV) which is a drivingforce for passive sodium uptake via a conductive Na + entrychannel from dilute mucosal solutions. 0.05 mM amiloride added tothe mucosal solution inhibited sodium influx in frog skin undervoltage clamp conditions but had no effect on proton excretion(Ehrenfeld et al, 1985). However, addition of 0.01 mM amilorideor substitution of external Na + by Mg++ or K+ caused a hyper-polarization of apical membrane potential and inhibited protonexcretion under open circuit conditions in frog skin (Harvey andEhrenfeld, 1986). Proton excretion in frog skin does not requiresodium transport but they can be electrically coupled with a 1:1stoichiometry.(2) Proton pump operation is regulated primarily by theproton electrochemical gradient across the membrane. Protonsecretion in turtle bladder as well as the proton excretion infrog skin increases with increasing mucosa-negative voltage andincreasing lumina' pH (Steinmetz and Andersen, 1982; Ehrenfeld etal, 1985). ApH (serosal pH - mucosal pH) and Alli (the electricpotential difference between the serosal side and the mucosalside of the epithelium) have equivalent inhibitory or stimulatingeffects on proton flux. Frog skin proton transport is annulled atApH of 1.59 unit or A* of -80 mV, but proton secretion in turtle14bladder continues until a ApH of up to 2.7 pH units or Alp down to-180 mV. Studies of turtle urinary bladder showed that theinstantaneous changes in the proton transport rate induced byvarying pH involved changes of intrinsic properties of the protonpump but not the number of pumps (Steinmetz, 1986).(3) The cellular acid-base conditions are the second majordeterminant of the rate of proton transport. Cell [Fr] is thefinal common pathway by which changes in serosal or mucosal Pco2and HCO3 - can exert their effects on the pump (Cannon et al,1985). Proton transport rate could be related to cell [Fr] as aMichaelis-Menten function in which hydrogen ions serve as thesubstrate for the proton pump. The proton pump in turtle bladderis far more sensitive to pH changes on its cytoplasmic side thanon the luminal side. The regulatory pH range on the cell side isabout 1 pH unit, whereas it is at least 3 pH units on the luminalside (Steinmetz, 1985).(4) Protons transported by the proton pump have anintracellular origin. Protons are generated in the cellularcompartment from CO 2 hydration under the catalysis of carbonicanhydrase (CA). CA-rich cells, which constitute 10-15 percent ofluminal surface area, are responsible for proton secretion inturtle bladder (Steinmetz and Andersen, 1982). Proton transportvia the proton pump is stimulated by an increase in the ambientCO 2 in both turtle bladder and frog skin (Cannon et al, 1985;Ehrenfeld and Garcia-Romeu, 1977). Addition of 1 mM acetazolamide(CA inhibitor) to the serosal side of frog skin caused15considerable inhibition of the Na + and fr fluxes without changingthe relationship between the two (Ehrenfeld and Garcia-Romeu,1977). Administration of 0.1 mM ethoxyolamide, a lipophilic CAinhibitor, to either the serosal or the mucosal side of frogskin, induced a 70% reduction in proton excretion which in turnreduce sodium influx (Ehrenfeld et al, 1985). It is suggestedthat CA inhibition affects proton transport indirectly byinhibiting the intracellular CO 2 hydration reaction whichsupplies hydrogen ions to the proton pump. The by-product,bicarbonate ion, passes into the serosal medium by C1 -/HCO3 "exchange in the basolateral membrane. SITS applied to the serosalsite of turtle bladder caused secondary inhibition on Fr -transport which was associated with increased alkalinity of thecell. Cl - was required in the serosal compartment for II +transport in the urinary bladder epithelium (Steinmetz andAndersen, 1982).(5) H +-ATPase has been noted to be either bound to membranesor packaged in cytoplasmic vesicles of CA-rich cells in turtlebladder (Arruda et al, 1990) and intercalated cells in rat kidney(Brown et al, 1988). Proton pumps can be inserted in the membraneby exocytosis of vesicles containing the ion motive enzyme orremoved from the membrane by endocytosis of segments of themembrane in which H +-ATPases are concentrated, and the processcan be induced by environmental stimuli (Brown, 1989; Schwartzand A1-Awqati, 1985; Stetson, 1989). Increased ambient CO 2 forexample, causes cytoplasmic acidification in both turtle bladder16and mammalian renal tubules which in turn raises cell calcium.Calcium causes rapid fusion of the vesicles to the luminalmembrane and the proton pump is inserted exocyticly. The pumpsthen turnover in the luminal membrane and pump protons out of thecell, recovering the cell pH towards its original level (Cannonet al, 1985; Arruda et al, 1990; Schwartz and Al-Awqati, 1985).(6) Proton pump activity relies on the energy of ATPhydrolysis. The stoichiometry of the pump is estimated to be 3 1-1 +/ ATP (Al-Awqati and Dixon, 1982). The CA-rich cells, MR cellsand intercalated cells in which H+-ATPase is located, haveabundant mitochondria (Madsen and Tisher, 1985). Turtle urinarybladder is capable of using both aerobic and anaerobic metabolismas energy sources to drive active proton transport. Oligomycin, aclassical inhibitor of mitochondrial ATPase, has no effect onturtle bladder H+-ATPase (Steinmetz and Andersen, 1982). However,the proton pump in frog skin depends completely on oxidativemetabolism, since anaerobic conditions totally block protonexcretion, and 83% of the proton pump activity in frog skin wasinhibited by 14 ug/ml oligomycin in either mucosal or serosalsolutions, presumably due to an indirect effect on mitochondrialrespiration (Ehrenfeld et al, 1985).(7) Inhibitor sensitivity:Orthovanadate is a specific inhibitor of plasma membraneATPase. Micromolar concentrations of vanadate can completelyabolish the plasma membrane ATPase of Neurospora, but evenmillimolar concentrations had no effect on mitochondrial ATPase17from the same organism (Goffeau and Slayman, 1981). Vanadate,however, inhibits not only 13 +-ATPase in plasma membrane, but manyother transport ATPases, including Na +-W-ATPase and Ca'-ATPase.Orthovanadate ion, VO 43- , acting as a phosphate transitionanalogue, blocks the formation of phosphorylated intermediates inthese ATPase (Pedersen and Carafoli, 1987). Vanadate in theserosal solution of toad bladder (Beauwens et al, 1981) andturtle bladder decreases proton secretion markedly underanaerobic and aerobic conditions (Arruda et al, 1981). Protonexcretion through frog skin is completely abolished when 1 mMvanadate is applied to the serosal site (Ehrenfeld et al, 1985).However, vanadate exerts no inhibitory effect on either theproton-ATPase activity or the proton transport in mammalian renaltubules (Gluck and Caldwell, 1987; Turrini et al, 1989) becausethe H+-ATPase in mammalian kidney belongs to the vacuolar type.N-ethylmaleimide (NEM) is another metabolic inhibitor thateffects V-type and P-type H +-ATPase, with much more potentinhibition on vacuolar H+-ATPase (1-2 uM) than plasma membraneH+-ATPase (0.1-1 mM). Mitochondrial ATPase is virtually resistantto NEM (Forgac, 1989). NEM is an alkylating agent that isrelatively selective for sulfhydryl groups (SH-) and inhibits H +-ATPase in an ATP-protetable manner. So-called NEM-sensitiveATPase is found in all segments of mammalian kidney (Ait-Mohamedet al, 1986; Gluck and Al-Awqati, 1984; Gluck and Caldwell,1987). Proton excretion across frog skin is also inhibited by 1mM NEM (Ehrenfeld et al, 1990). PCMBS is also a SH-group reagent18and affects proton-ATPase in rat and bovine kidney (Turrini etal, 1989; Gluck and Al-Awqati, 1984).Dicyclohexylcarbodiimide (DCCD) can bind to a subunit (theDCCD binding protein) of the hydrophobic channel portion andinhibit H +-ATPase in mitochondria, vacuolar and plasma membrane(Pedersen and Carafoli, 1987). The sensitivity was highest in F-type, followed by V-type and then P-type (Forgac, 1989). H +-ATPase in mammalian kidney (Ait-Mohamed et al, 1886), turtlebladder (Steinmetz and Andersen, 1982) and frog skin (Ehrenfeldet al, 1985) is sensitive to DCCD. Diethylstibestrol (DES) has avery similar effect on H +-ATPase as DCCD (Pedersen and Carafoli,1987) and 0.1 mM DES was found to inhibit proton excretion infrog skin (Ehrenfeld et al, 1990).7-chloro-4-nitrobenz-2-oxa-1,3-diazole (NBD-Cl) is also analkylating agent that potently inhibits H +-ATPases of all types(Forgac, 1989), by affecting the NH 2-group in an ATP-protetablemanner. Proton-ATPase in mammalian kidney was reported to besensitive to 10-20 uM NBD-C1 (Gluck and Caldwell, 1987; Turriniet al, 1989).Bafilomycins is a macrolide antibiotic that appears to be avery specific and potent inhibitor to vacuolar H +-ATPase (Bowmanet al, 1988). Mitochondrial H +-ATPase is resistant to it andplasma membrane H +-ATPase is moderately sensitive to it. It canbe used as a valuable tool for distinguishing among the threedifferent types of H +-ATPases. Another reagent demonstratingunique specificity to V-type H +-ATPase is potassium nitrate19(KNO3) (Bowman, 1983). Vacuolar H +-ATPase from Neurospora wasinhibited by KNO 3 with a half-maximal inhibition seen at 50 mM,whereas plasma membrane and mitochondria H +-ATPase from the sameorganism was completely resistant to KNO 3 up to 100 mM (Bowman,1983).(8) Hormonal regulation:Among the hormonal factors regulating proton secretion, themineralocorticoid hormones play a central role. The rate ofacidification by the urinary bladder is stimulated by aldosteronein toad (Ludens et al, 1974) and turtle (Al-Awqati et al, 1976)and the effects were independent of the stimulation of sodiumtransport. The stimulation of proton secretion is not associatedwith a change in the apparent proton-motive force (the intrinsicproperty of the pump; Al-Awqati et al, 1976). Instead,aldosterone has been shown to be preferentially bound to MR cellsof urinary bladder (Sapirstein and Scott, 1975) and frog skin(Harvey, 1992) and produce rapid exocytotic insertion of theproton pump (Harvey, 1992). Aldosterone was also reported to haveboth a long-term and short-term stimulating effect on proton-ATPase in renal collecting tubules of mammals (Garg and Narang,1988; Khadouri et al, 1989; Mujais, 1987).Deoxycorticosterone acetate (DOCA), a potentmineralocorticoid, elicits a stimulation of proton transport inamphibian skin (Ehrenfeld et al, 1989), which also inducesmorphological changes in the MR cells of frog skin epithelia(Voute et al, 1972; Voute et al, 1975). Serosal 10 -6 M DOCA20addition induced a 31% proton secretion and a 28% sodiumabsorption. The action of DOCA on H +-ATPase appears to beindependent of the sodium effect since stimulation of protonefflux was also observed in the absence of sodium transport(Ehrenfeld et al, 1989).Neurohypophyseal hormones, including arginine vasotocin(AVT) and oxytocin, enhance transepithelial sodium transport,osmotic response and permeability to a variety of solutes in frogor toad urinary bladder and frog skin (De Sousa and Grosso,1981). Sodium transport in frog skin was only enhanced by AVT(10 -6 M) or oxytocin (100 mU/ml) when the skin was bathed on itsapical side with high sodium containing solution (115 mmo1/1),while sodium influx is independent of proton secretion sincechloride ions can provide the permeant counterion for sodiumabsorption (Ehrenfeld et al, 1989). Conversely, in dilute NaClsolutions, where sodium absorption is tightly coupled to protonsecretion, neither oxytocin nor AVT stimulated sodium transportor proton secretion (Ehrenfeld et al, 1989). This indicates thatneurohypophyseal hormones increase sodium permeability in frogskin, but exert no effect on the H +-ATPase in MR cells. Studieson three amphibian species by Brown et al (1981) showed thatthere was no relationship between MR cell morphology and oxytocintreatments, and suggested that there was a lack ofneurohypophyseal receptors on the MR cell membranes.(9) Cellular model for Na + absorption and 1-1 + excretion infrog skin:21Two types of sodium absorption kinetics have been describedin frog skin under open-circuit conditions (Ehrenfeld et al,1989; Ehrenfeld et al, 1990). The first one of high capacity andlow affinity, is predominant in high NaCl-containing water, andis mainly mediated by the granular (GR) cells (Figure 2). Cl - isthe accompanying anion and proceeds through voltage-gated Cl -channels and Cl ./HCO3 - exchangers (not represented in the figure)located on the apical membranes of the mitochondria-rich (MR)cells and/or through a paracellular pathway. An acid load blocksthe sodium transport by the gating effects of protons on the Na'and K+ conductance of the GR cells. Basolateral exchangers (Na +/H+and C1 -/HCO3 - ) regulate cell pH and thereby participate in thecontrol of Na + transport. The second sodium absorption mechanism,of high affinity and low capacity, is predominant in diluteapical NaC1 solutions, which is the physiological condition forfreshwater frogs. Sodium absorption here is energized by afavourable apical electrical potential generated by theelectrogenic proton pump located in the MR cells, and thiselectrical effect should occur in both cell types. After an acidload, Na + transport and H + secretion are increased and Na +transport via MR cells may be favoured, since the Na + and IC'conductances of the GR cells are inhibited by intracellularprotons. The two cell types are targets for aldosterone whereasthe neurohypophyseal hormone, AVT and oxytocin affect only the GRcells.22Figure 2. A schematic view of the outermost living cell layer offrog skin, with two distinct cellular pathways for Na +absorption. (Adapted from Ehrenfeld et al, 1990).23Na + H+GR. cellOxytocineAVTExchangersNear andCI1HCCipMR.cellkw:bangers (Na'/H' sod a 7HCO 3)A24The key question I tried to answer in this study was: DoesH+-ATPase exist in fish gills as suggested by Avella andBornancin (1989)? Most investigators have supported Maetz's model(1964) of Na +/H+ and C1 -/HCO3 - , but the H+-ATPase model has notbeen investigated. Gills of freshwater fish behave like frog skinin many ways: branchial sodium influx is sensitive to amilorideand low pH (Wright and Wood, 1985); proton efflux is inhibited bylow pH (McDonald and Wood, 1981); transepithelium potentialchanges with external water pH (McWilliams and Potts, 1978; Ye etal, 1991). These are all consistent with the presence of a protonpump in fish gills. I have attempted to gather evidence for theexistence of a proton pump in fish gills, and to study itsinhibitor sensitivities, function, regulation and relationshipwith other ion transport pathways in gill epithelium.Due to the structural complexity of fish gills, gillepithelium can not be stretched out as a flat sheet; an isolatedgill preparation does not behave in the same way as in vivo (Perry et al, 1984); many conventional methods for studying iontransport pathways such as the Ussing chamber, cannot be appliedto gills. My approach to the problem therefore, was limited toexamining the branchial proton excretion of intact fish, andassaying H +-ATPase activity in homogenates of gill tissue.25SECTION I: BRANCHIAL PROTON EXCRETION IN RAINBOW TROUT MEDIATEDBY AN ELECTROGENIC PROTON PUMPChapter 1: Inhibitor sensitivity of branchial proton excretionand in vivo evidence for the existence of a proton pump26SUMMARYTotal CO2 , total ammonia, pH and buffering capacity of theinspired and expired water in rainbow trout were measured andbranchial proton excretion rate was calculated from these data.Ion transport inhibitors such as amiloride, vanadate andacetazolamide were added to the water to determine their effectson branchial proton excretion. There was a substantial protonexcretion across the gill epithelium which caused dehydration ofbicarbonate in expired water and was responsible for the expiredwater acidification. Proton excretion across the gills isinsensitive to 0.1 mM amiloride but sensitive to vanadate,acetazolamide and higher concentration of amiloride; thus, weconclude that proton excretion is probably mediated by an activeproton pump, instead of a Na ±/H+ exchanger, on the apicalmembrane of gill epithelium similar to that reported for the frogskin and turtle bladder.27INTRODUCTIONWater is acidified as it passes over the gills of fish(Wright et al, 1986; Lin and Randall, 1990). The acidification ofexpired water is inhibited by acetazolamide, a carbonic anhydraseinhibitor in external water. It was concluded therefore, that theacidification of water was caused by the hydration of excretedCO 2 , resulting in increased concentration of bicarbonate andprotons (Wright et al, 1986). Carbonic anhydrase in the gillmucus (Wright et al, 1986) and on the apical surface of the gillepithelium (Rahim et al, 1988) will catalyse the CO 2 hydrationreaction.Acidification of expired water could also be caused byexcretion of acid equivalents across the gills. Excretion ofprotons, NH4+ or absorbtion of HCO 3 - or any combination of thethree would also result in a decrease of expired water pH. HCO 3 "absorbtion is unlikely to be the cause of expired wateracidification because this acidification was not affected bySITS, a C1 -/HCO3 - exchange inhibitor (Lin and Randall, 1991).Proton transport in some epithelia is inhibited by acetazolamide(Steinmetz and Anderson, 1982; Ehrenfeld and Garcia-Romeu, 1977),and this could be an alternate interpretation of theacidification of expired water.Na +/H+ (NH4+) exchange in the apical membrane of gillepithelia has been concluded to be the major pathway for sodiumuptake and proton excretion and an optional excretory pathway for28ammonia (Maetz, 1964; Wright and Wood, 1985). This conclusion waschallenged by Avella and Bornancin (1989) who demonstrated thatsodium uptake and ammonia excretion were uncoupled. An indirectcoupling of an electrogenic proton pump and a passive sodiumentry was proposed by these authors as the mechanism of sodiumand proton transport across the fish gill epithelium. In fact,proton transport in frog skin and turtle bladder is mediated by aproton pump and it is sensitive to acetazolamide and vanadate.The objective of this study is first, to clarify whether theacidification of expired water in fish can be attributed toexcreted CO 2 hydration, by analyzing the bicarbonate level inboth inspired and expired water; and secondly, to identify thepathway through which protons are excreted by applying iontransport inhibitors such as amiloride, vanadate andacetazolamide to external water.29MATERIALS AND METHODSAnimals and PreparationRainbow trout Oncorhvnchus mykiss  (Walbaum), weighing202-592g, were maintained in outdoor fibreglass tanks suppliedwith flowing dechlorinated Vancouver tap water (pH 6-7; [CaTh0.03 mM; [Na + ], 0.89 mM; [C1 - ], 0.92 mM; [W], 0.03 mM, [Mg'],0.01 mM). Fish were fed daily with commercial trout pellets andfeeding was suspended for at least 48 h prior to experimentation.Surgery was performed on each fish under general anaes-thesia (1:10,000 MS222 solution, pH adjusted to 7.5 with NaHCO 3 )to fix an opercular cannula for sampling expired water. Fish werethen confined, but not physically restrained, in a black chamberto recover for at least 24 h. This black chamber was suppliedwith aerated dechlorinated tap water during the recovery period.Three hours prior to the experiment, the water supply wasswitched to the aerated test solution (40 mM NaC1 and 0.5 mMCaC1 2 in dechlorinated tap water, from Wright et al. 1986) with abuffering capacity (B) of 81 uequiv.L - '.pH unit -1 . The testsolution had the same ionic strength as the buffer solution usedto calibrate the pH electrodes. By using this test solution, wereduced the response time of the pH electrode, increased itsstability and thus obtained more precise water pH measurements.Temperature was regulated to that of tap water with a coolingcoil.Experimental Protocols30Experiments were carried out using a recirculating systemconnected to a black chamber (Figure 2). The volume of therecirculating system was 6 litres and it was aerated andcontrolled to ambient temperature. A magnetic stirring bar wasused in the reservoir to ensure complete mixing. Ion transportinhibitors such as amiloride, vanadate and acetazolamide wereadded to the recirculating system and their effects on CO 2 ,ammonia and net proton excretion determined. Inspired and expiredwater samples (approximately 5 ml each) were withdrawn from theoutlets of the glass electrode chambers for total CO 2 , ammoniaand buffer capacity analysis.(1) Amiloride treatmentThree concentrations of amiloride were utilized. Eachexperiment started with a one hour control period, with the fishrested in the recirculating system containing the test solutionalone and inspired and expired water were sampled at 30 minutesand 60 minutes. The system was flushed with fresh test solutionand amiloride was added to the system to give a finalconcentration of 0.1, 0.5 or 1 mM. Recirculation was restored andthe amiloride treatment lasted for another hour. Water sampleswere taken at 30 minutes and 60 minutes. The results of the 30minutes and 60 minutes sampling were pooled for data analysissince there was no significant difference between them.(2) Vanadate treatmentExperiment started with a one hour control period followedby a one hour treatment period. Freshly made 3 mM sodium31Figure 3. The recirculating system with a black chamber. Fishwere prepared with an opercular cannula. Inspired andexpired water samples were withdrawn from the outlets of theglass electrode chambers.32orthovanadate (Na3VO4) solution was boiled and the cooledsolution was neutralized with 0.1 moll -1 HC1. After the controlperiod, 200 ml of the 3 mM vanadate solution was added to the 6litre recirculating system to obtain a final concentration of 0.1mM. Since the ammonia accumulation in the system was very low(less than 100 uM after 2 hours), flushing the system with freshtest solution at the beginning of the treatment period wasconsidered unnecessary. External water pH was adjusted to thecontrol value during the two 30 minutes treatment periods.Inspired and expired water samples were taken at 30 minutesintervals and the 30-min and 60-min samples were pooled for dataanalysis since no significant differences were observed.(3) Acetazolamide treatmentFor the acetazolamide treatment, acetazolamide was added tothe system to give a final concentration of 0.1 mM after a 30-mincontrol period. Inspired and expired water were sampled at theend of control period and 30 minutes and 60 minutes afteracetazolamide treatment. Again the 30-min and 60-min treatmentsamples were pooled for analysis.Analytical techniques and calculationsInspired and expired water pH were monitored during thewhole experimental period with combination glass pH electrodeshoused in two water-jacketed glass chambers (Wright et al. 1986).Inspired pH (pH in ) and expired pH (pH ex ) values were recorded ateach sampling.Total carbon dioxide contents of inspired water [CO 2 ] in and34expired water [CO2]ex were measured immediately with a Carle gaschromatograph (model III) containing a CO 2 discriminating column(porapak Q) (Boutilier et al. 1985; Lenfant & Aucutt, 1966).Total ammonia contents of inspired water [Amm] in and expiredwater [Amm] ex were measured by a micro-modification of thesalicylate-hypochlorite assay with frozen water samples (Verdouwet al. 1978). To ensure that there was no ammonia loss from waterin the recirculating system, two experiments were carried out inwhich known amounts of NH 4C1 were added to the system withoutfish. No loss from the system occurred.The ammonia excretion rate of the fish was calculated as:Ammonia excretion rate = ([Amm] i ,. f - [Amm],„ . ")*y/t*W,where i and f refer to the initial and final ammoniaconcentration in inspired water in umoli 1 , V is the volume ofthe system (6 1 in this study), t is the elapsed time in hoursand W is the mass of the fish in kilograms.Bicarbonate concentrations in inspired water [HCO 3 ] 1, andexpired water [HCO 3 - ] ex were calculated from [CO 2 ],„, pH i„ and[CO2 ] ex , pHex , respectively, by the Henderson-Hasselbalchequation, using the pK, o2 and a, 02 values from Boutilier et al.(1985). Carbonate formation is negligible over this pH range.Ammonium ion concentrations in inspired water [NH 4+ ] 1, and expiredwater [N114+ ] ex were calculated from [Amm] i„, pH, and [Amm] ex ,respectively, by the Henderson-Hasselbalch equation, using thepKm, value from Cameron & Heisler (1983).If we assume all the carbon dioxide is excreted as CO 2 and35all the ammonia is excreted as NH 3 , the pH change occurring inthe water passing over the gills of fish could be due to CO 2hydration, ammonium ion formation or acid excretion.CO2 + H2O ^ HCO3- + H+NH3 + 1-1+ ^ NH4The proton added by CO 2 hydration can be estimated from thedifference of bicarbonate levels in inspired and expired water:[HCO3 ] e.-[HCO3l in . The proton consumed by ammonium ion formationcan be estimated from the difference of ammonium ion levels ininspired and expired water: [NH:] e.-[NH:] In .The total acid equivalents added to the water, [11 + ] total / canbe calculated from the appropriate buffer curve and pH in and pHex •The buffer capacities of inspired and expired water were measuredby titrating the stored water samples with 0.1 moll -1 HC1. 20 mlof water sample was held in a glass beaker with a water jacketfor temperature control and its pH was increased to approximately8.0 by addition of 0.1 moll -1 NaOH as the sample was aerated. HC1then was manually added to the aerated sample with a syringe andpH was recorded 2 -3 minutes after each addition to ensurecomplete equilibrium. Buffer curves were constructed from the pHand the amount of acid added for each individual water sample.The buffer curves of inspired and expired water were very similarand only inspired water buffer curves were used for subsequentcalculations.Since [H+ ] total = ([HCO31 e,-[HCO3 - ] ) - UNH:L x-[NH:] ) + netproton excretion, the net branchial proton excretion rate,36therefore, is equal to: [Fr] total - ([HCO31 e,-[HCO31 10 +([NH4+ ],,,-[NH4+ ] in ). All the measurements and calculations wereconducted with each individual inspired and expired watersamples.Data are presented as means ± standard error. Student'stwo-tailed t-test and analysis of variance (ANOVA) was used totest for significant differences between means. Tests ofsignificance were conducted at the 5% level of rejection.37RESULTSWhen water pH was approximately neutral, [HCO3 ],„ was alwaysgreater than [HCO31 e, ( Figure 4, [HCO 3-] - [HCO31,,, givespositive values), indicating that there was no CO 2 converted toHCO3 - as water passed over the gills. Instead, bicarbonate levelwas reduced due to HCO3 - dehydration induced by proton excretionor bicarbonate absorption. When proton excretion was inhibited by0.1 mM vanadate, 0.1 mM acetazolamide or 0.5 or 1 mM amiloride(see Figure 5), the extent of bicarbonate dehydration decreasedand [HCO3 - ] ex increased. Therefore, [HCO3 ]in - [HCO3- ]ex wassignificantly lower than the control value. This suggests thatbranchial proton excretion not only caused expired wateracidification but also dehydrates HCO3 - as water passed over thegills. The control [HCO3 ]in - [HCO3 )ex value at pH 7.6 was greaterthan that at pH 6.6 because the excretion of acid equivalents washigher at pH 7.6 (Figure 5).There is a substantial net proton excretion across the gillof freshwater rainbow trout when external water pH was aboutneutral. This proton excretion was not affected by 0.1 mmo11 -1amiloride (Figure 5). Increasing amiloride concentration in theexternal medium induced a reduction in proton excretion, but morethan 50% of the net proton excretion was still sustained even in1 mM amiloride (Figure 5).0.1 mM vanadate treatment resulted in reductions of netproton excretion by 58% (Figure 5). Acetazolamide treatment also38Figure 4. Bicarbonate concentration differences between inspiredand expired water of rainbow trout under control, amiloride,vanadate and acetazolamide treatments. * indicates asignificant difference between the control and treatmentvalues (P<0.05); Bars show standard errors; N=6.391501200I - 90-JO 5600I3000.1mM 0.5mM 1mM 0.1mM 0.1mMAmiloride Amiloride Amiloride Vonodote AcetazolomidepH6.6 pH6.6 pH6.6 pH7.6 pH7.0ControlTreatmentTFigure 5. Net proton excretion across the gill epithelium ofrainbow trout under control, amiloride, vanadate andacetazolamide treatments. * indicates a significantdifference between the control and treatment values(P<0.05); Bars show standard errors; N=6.41-0.1mM 0.5mM 1mM 0.1mM 0.1mMAmiloride Amiloride Amiloride Vonodote AcetazolomidepH6.6 pH6.6 pH6.6 pH7.6 pH7.0caused a 48% reduction in net proton excretion (Figure 5). Thedifferences between control proton excretion values were due todifferent pH in external water.The differences in total CO 2 content between inspired waterand expired water with different treatments are presented inTable 1. [CO2 ] ex-[CO2 ],„ represents the CO2 excretion rate if weassume that ventilation rate is constant. There was nosignificant difference in [CO2 ],,-[CO2 ],„ between control and drug-treated animals except in the case of acetazolamide. CO 2excretion increased when fish were exposed to 0.1 mMacetazolamide in water.Ammonia excretion of fish was not significantly inhibited by0.1 mM amiloride (Figure 6). However, higher concentrations ofamiloride induced a reduction in ammonia excretion, by 58% with0.5 mM amiloride and by 87% with 1 mM amiloride (Figure 6).Vanadate and acetazolamide had no significant effect on ammoniaexcretion.43Table 1. Total CO2 Excretion By Fish (umol/l.kg)0.1 mM Amiloriden=60.5 mM Amiloriden=61 mM Amiloriden=60.1 mM Vanadaten=100.1 mM Acetazolamiden=4[CO2]ex -Control[CO2]inTreatment39.73 ± 6.57 32.75 ±^5.3071.00 ±^8.04 71.48 ±^10.4362.46 ±^8.11 57.77 ±^7.4976.94 ± 4.41 63.68 ±^5.6055.11 ± 9.91 82.90 ±^10.32** indicates a significant difference from the control value.44Figure 6. Branchial ammonia excretion rates of rainbow troutunder control, amiloride, vanadate and acetazolamidetreatments. * indicates a significant difference betweencontrol and treatment values (P<0.05); Bars show standarderrors; N=6.45400350-(7)^300ccO _c• ' 2501.)(;),^200o•F^1500E• 1 0050_0 1mM 0.5mM 1mM 0.1mM 0.1mMAmiloride Amiloride Amiloride Vonodate AcetazolomidepH6.6 pH6.6 pH6.6 pH7.6 pH7.0DISCUSSIONThe fact that the bicarbonate level in inspired water offish was always higher than that in expired water indicateseither bicarbonate was converted to CO 2 as water passed over thegills, or bicarbonate was absorbed from the water, probably viaC1 -/HCO3 - exchange. The later possibility was eliminated by theobservation that SITS, an inhibitor of C1 -/HCO3 - exchange, had noeffect on acid excretion across fish gills (Lin and Randall,1991). Bicarbonate could be converted to CO 2 when CO 2 wasexcreted to the water by the fish if a substantial amount of acidwas also excreted simultaneously into the water. Therefore, theexpired water acidification reported by Wright et al (1986) andLin and Randall (1990) was not due to the hydration of excretedCO2 but due to acid excretion across the gills. The excreted acidwas partially consumed by HCO 3 - dehydration and ammoniaprotonation, and the rest contributed to expired wateracidification.It has long been hypothesized that Na +/H+ (NH 4+ )electroneutral exchange is the principal mechanism of sodiumuptake and proton excretion in the gill epithelium of fish(Wright & Wood, 1985). This antiport exchange process is blockedby 0.1 mM amiloride, a very potent and relatively specificinhibitor of sodium transport in a wide variety of cellular andepithelial transport systems (Benos, 1982). 84% and 94% reductionof the Na + uptake by the gills of intact freshwater rainbow trout47exposed to 0.1 mM amiloride in the external media were reportedby Perry & Randall (1981) and Wright & Wood (1985), respectively.In our studies, this concentration of amiloride had no effect oneither net proton excretion or ammonia excretion when comparedwith control values from the same animals (Figure 5 and 6). Ourexperimental conditions are similar to those of Perry & Randalland Wright & Wood. This indicates that sodium influx and protonion efflux are not directly coupled (see also Avella andBornancin, 1989).It is well documented that proton transport in mammaliankidney (Steinmetz, 1985), amphibian urinary bladder (Al-Awqati,1978; Steinmetz, 1986) and frog skin (Ehrenfeld et al. 1985) ismediated by an electrogenic proton pump. The gill epithelium infreshwater fish is considered to be "tight" (Sardet, 1980) andresembles frog skin and turtle bladder epithelia functionally andmorphologically. My studies demonstrated that proton excretion introut was unaffected by low concentrations of amiloride but wasinhibited by vanadate and acetazolamide. The same have beenreported for the proton transport in frog skin (Ehrenfeld et al.1985; Ehrenfeld & Garcia-Romeu, 1977) and turtle bladder (Al-Awqati, 1978; Steinmetz, 1986). I therefore agree with Avella andBornancin (1989) and similarly postulate that the fish gill hasan electrogenic proton pump in the mucosal membrane, similar tothat reported for frog skin and toad bladder, rather than aNa +/H+ exchange mechanism.The electrogenic proton pump, or H +-translocating ATPase,48on the apical membrane removes protons from the cell andgenerates a negative potential in the inner side of the apicalmembrane (Figure 7). Sodium influx, driven by the negativepotential, occurs via a sodium channel that is highly sensitiveto amiloride. Na +/K+ (NH4+)-ATPase in the basolateral membranepumps sodium out of the cell into the blood. Thus, protonexcretion and sodium uptake are intimately, but indirectly,linked. Since Avella et al. (1987) showed that branchial sodiumuptake was proportional to the number of chloride cells in thegills and, since proton pumps consume energy and chloride cellsare rich in mitochondrial and can supply the energy demand, Ispeculate that the electrogenic proton pump is located in thechloride cell.Ammonium ions can replace potassium on the Na +/K+-ATPaseand thus enter the cell and form NH3 and protons (Evans et al.1989). The deprotonation of NH 4 + could supply the proton pump andNH3 could diffuse passively across the apical membrane into thewater. Although much less sensitive than the Na + channel,Na +/K+-ATPase in the basolateral membrane can be inhibited byamiloride that has entered the cell when applied in highconcentrations to the mucosal side (Knauf et al. 1976; Kleyman &Cragoe, 1988). Thus, the reduced proton and ammonia excretion in0.5 and 1 mM amiloride treatments could be accounted for by theinhibitory effect of amiloride on Na +/K+ (NH4 + )-ATPase in thebasolateral membrane. In support of this contention, Evans et al (1989) showed that amiloride did not affect ammonia excretion if49Figure 7. Schematic representation of gas and ion transportacross the gill epithelium of rainbow trout. ATP-drivenpumps are denoted by filled circles, ion exchanger by openedcircles and passive diffusion by arrows. See text fordetails.5051the perfused head of the toadfish was pretreated with ouabain,which blocks Na +/K+ (NH4+)-ATPase. Proton excretion in frog skinwas inhibited by 0.5 mM amiloride by 35% but was not affected by0.05 mM amiloride, whereas sodium uptake was completelyabolished. If we assume that the ventilation rate of the fish was100 ml/min (Lin and Randall, 1990), we can compare the ammoniaexcretion rate with the net proton excretion rate under amiloridetreatments (Table 2). The reduction in ammonia excretion wasequivalent to that in proton excretion, indicating thepossibility that NH3 and protons were both originating from NH 4 +transported into the epithelium via Na +/K+ (NH 4+ )-ATPase in thebasolateral membrane (Figure 7).Ammonia elimination was not affected by vanadate oracetazolamide, indicating that proton and ammonia efflux from thegill epithelium are through different pathways. Thus, ammoniumentry into the gill epithelium may affect proton excretion (Table2), but variations in proton excretion do not appear to affectammonia excretion. Ammonium cannot be the sole source of protonshowever, because proton excretion can be more than twice ammoniaexcretion in some instances. Acetazolamide, a traditionalcarbonic anhydrase inhibitor, inhibits proton excretion in fishgills (Figure 5), as demonstrated in frog skin and turtle bladder(Ehrenfeld & Garcia-Romeu, 1977; Steinmetz, 1986). This suggeststhe possible contribution of intracellular CO 2 hydration to theproton supply. The apical addition of 0.01 mM ethoxzolamide, alipid soluble CA inhibitor, blocked net H + excretion as well as52Table 2. Comparison between the reduction in the rate of ammoniaexcretion and the reduction in the rate of net proton excretionunder amiloride treatment.Amiloride^Reduction in the rate of Reduction in the rate ofconcentration^ammonia excretion^net proton excretion(mM) (umol/kg.h) (umol/kg.h)0.1 47.968 ± 20.568* 60.806 ± 36.178*0.5 130.542 ±^9.346 143.169 ± 61.0591 280.560 ±^35.561 298.105 ± 86.911Net proton excretion was converted to umol/kg.h by assumingventilation rate = 100 ml/min (Lin & Randall, 1990). * indicatesa value not significantly different from zero.53Na+ absorption in frog skin (Harvey and Ehrenfeld, 1986). Whileinhibiting proton excretion, acetazolamide also elevates carbondioxide excretion across fish gills (Table 1). Because CO 2excretion is not diffusion limited (Perry et al,  1982), elevatedCO2 excretion usually involves an increase in CO 2 production or adecrease in CO 2 storage in the fish. There is not enoughinformation from this study to explain it.The amount of proton excretion in this study is of the samemagnitude as that reported by Avella and Bornancin (1989), but inexcess of that measured by McDonald and Wood (1981). Comparisonwas made assuming the ventilation rate of the fish is 100 ml/min(see Lin and Randall, 1990). The method I used in this study toestimate branchial proton excretion rate is different from thatused by the above mentioned workers, which deals with wholeanimals in a 30-60 min time period. In my method I focus on theinstantaneous proton excretion across fish gill epithelium. Thedetailed analysis of inspired and expired water chemistry wasdesigned to reveal changes occurring in the gill surface thatmight not be picked up in whole animals studies. I accounted forthe proton addition / consumption from the CO 2 hydration /dehydration by measuring the bicarbonate levels in inspired andexpired water. However, during the titration process a portion ofthe CO 2 component was also involved, and the proton excretionrate might have been over-estimated. In the method used by theother authors, the hydrogen ions consumed by HCO 3 - dehydrationwas not accounted for and proton excretion rate might had been54under-estimated. The other reason that the proton excretionreported here is high is that, proton excretion was probablystimulated under the high salt experimental condition. 40 mM NaC1was added to the external water in order to obtain precise pHmeasurements. The operation of the putative proton pump wasprobably stimulated by the sudden increase of Na + levels in thewater and 3 hours was not long enough for endocytotic retreat ofproton pump. When blood pH was measured under similar condition(Lin and Randall, 1990), fish were suffering from a slightalkalosis (blood pH was 7.95-8.00 compared to the normal value of7.80).Vanadate has a nonspecific inhibitory effect on ATPases andcould be acting on Na +-K+-ATPase on the basolateral border of thefish gill. In my studies, more than 50% of the net protonexcretion across the gill epithelium was inhibited by 0.1 mMvanadate applied to the mucosal membrane. De Sousa & Grosso(1979) showed that applying 1 mM vanadate to the outer surfacedid not affect the Na +-K+-ATPase in the basolateral membrane offrog skin. Arruda et al. (1981) showed that vanadate had noeffect on the backleak of proton or bicarbonate secretion but hada direct effect on H +-translocating ATPase in turtle bladder.Thus, I conclude that the reduction in proton excretion observedin my studies was induced by the inhibitory effect of vanadate onH+-translocating-ATPase in the apical membrane. The reason thatthe proton excretion was not completely abolished was,presumably, because of the difficulty of vanadate reaching the55action site from the mucosal side (Arruda et al. 1981).0.1 mM amiloride had no effect on the putative fish gillproton pump in open-circuit conditions. In frog skin epitheliumunder open-circuit conditions however, application of 0.01 mMamiloride caused a hyper-polarization of membrane potential, afall in intracellular sodium ion activity, an inhibition of 1-1 +excretion and a decrease in pHi (Harvey and Ehrenfeld, 1986).Inhibition of sodium influx should have increased membranepotential and reduced proton excretion in fish gill epithelium.This did not happen; therefore, if the proton pump does exist,there must be some other counter-ion that can replace sodium.Perry and Randall (1981) found that 0.1 mM amiloride inhibitedchloride influx in the fish gill. Inhibition of both chloride andsodium influx, when fish are exposed to 0.1 mM amiloride, wouldtend to ameliorate any rise in potential across the apicalmembrane and, therefore, permit continued functioning of theproton pump.In conclusion, my results provide preliminary evidence thatthe acidification of expired water in rainbow trout in neutralwater is mainly caused by a net proton excretion, probablymediated by an active proton pump on the apical membrane of gilllamellae. This proton pump is sensitive to vanadate andacetazolamide, and resembles the electrogenic proton pump in frogskin and turtle bladder.56Chapter 2: Effects of external water pH, Pco 2 and ion levels onbranchial proton excretion mediated by proton pump57SUMMARYBranchial proton excretion was sensitive to external waterpH. When water pH is below 5.5 (about 2.5 units lower than bloodpH), proton excretion was inhibited. In water pH range of 5.5-8.0, proton excretion increased linearly with water pH. In pHabove 8.0, proton excretion appeared saturated. Elevated waterPco2 levels stimulated branchial proton excretion, indicatingintracellular CO 2 hydration provides hydrogen ion to the putativeproton pump. Sodium-free water suppressed proton excretion acrossthe gills but addition of the chloride channel inhibitor 9-anthroic acid transiently lessened the suppression, whichindicates that both sodium and chloride influx across the gillepithelium probably play a role in justifying the electricalpotential generated by active proton excretion.58INTRODUCTIONIt was concluded from the studies in Chapter 1 that theacidification of expired water across fish gills was induced by anet branchial proton excretion. Inhibitor sensitivity studiesruled out the possibility that a Na -E/H+ exchanger was responsiblefor the proton flux, and indicated the existent of a primaryproton pump in the gill epithelium that mediated the measuredproton excretion, but many questions remain unanswered. Forexample: Proton transport mediated by a proton pump in turtlebladder (Steinmetz, 1986) and frog skin (Ehrenfeld et al, 1985)was regulated by the transepithelial pH gradient. The fact thatexpired water acidification was sensitive to external water pH(Lin and Randall, 1990) suggests that branchial proton excretionmay also be pH-sensitive, but it remains to be tested. Ammoniumion deprotonation in the intracellular compartment was consideredto be the supply of hydrogen ions for the proton pump but couldonly account for 50% or less of the proton excretion. What is theother source of protons? Is it from carbon dioxide hydration asobserved in frog skin (Ehrenfeld and Garcia-Romeu, 1977) and toadbladder (A1-Awqati, 1978)? What is the counterion for theelectrogenic proton pump? The sodium channel inhibitor,amiloride, appeared to have no effect on branchial protonexcretion but the application of this drug (and also SITS, a C1 -/HCO3 - exchanger inhibitor) was complicated by its double effecton both Na* and Cl - flux across fish gills (Perry and Randall,591981). Other approaches, such as removing sodium from externalwater and / or employing a chloride channel inhibitor, have to beused to clarify these points.Experiments reported in this chapter were designed toexamine these problems. Attempts were also made to investigatethe effect of high external calcium levels on the protonexcretion across fish gills. Fish appear to be able to maintainacid-base and osmotic homeostasis better in hard water (McDonaldet al, 1980; McDonald 1983; McDonald et al, 1983), it isinteresting to know whether the proton pump was playing a role inthis process.60MATERIALS AND METHODSExperiments were carried out on rainbow trout of both sexes,weighing between 200 and 500 g. Fish were maintained in largeoutdoor tanks and regularly fed. Feeding was suspended two daysbefore surgery. To sample expired water from fish gills, anopercular cannula was fixed right next to the opercular openingin the fish using the method described in chapter 1. The fish wasallowed to recover from anaesthesia for 24 hours in a blackchamber supplied with aerated dechlorinated tap water. Two hoursprior to the experiment, this black chamber was hooked up to arecirculating system and the water supply was switched to theaerated test solution (40 mmoll -I NaC1 and 0.5 mmo11 -1 CaC1 2 , pH 6-7, see Chapter 1). Inspired and expired water samples were takenfor pH, total CO 2 , total ammonia and buffer capacity analysis andnet branchial proton excretion rate can be calculated with thesedata. Detailed analytical techniques and calculations aredescribed in Chapter 1. Control water samples were obtained twohours after the fish was acclimated to the test solution andexperimental water samples were obtained after one of thefollowing treatments.(1) Varying external water pHThe pH of the test solution was adjusted to the 4, 5, 8 or9 by adding 0.1 mmo11 -1 NaOH or 0.1 mmo11 -1 HC1 at the beginningof each different environmental pH exposure. A magnetic stirringbar was used in the reservoir to ensure quick and complete61mixing. The water pH of the recirculating system changed slightly(<0.3 unit) over the 30 minutes experimental period, and noattempt was made to stabilize pH during this period. Inspired andexpired water samples were taken 30 minutes after the fish wereexposed to the desired external pH.(2) Sodium-free external waterSodium-free test solution was prepared by replacing 40mmo11 -1 NaC1 with 40 mmoli 1 choline chloride. The recirculatingsystem was flushed with sodium-free test solution after controlsampling and fish were maintained in the recirculating sodium-free test solution for one hour. Water samples were taken at 30minutes and 60 minutes. The recirculating system was then flushedwith regular test solution and inspired and expired water weresampled 30 minutes later.(3) 9-anthroic acid treatmentFish were maintained in sodium-free test solution for 30minutes after the control period. 9-anthroic acid (9-AA) was thenadded to the recirculating system to a final concentration of 0.1mM. Water samples were taken 30 minutes and 60 minutes after thefish were exposed to 9-AA in sodium-free external water.(4) High external calcium levelsCalcium chloride was added to the recirculation system to afinal concentration of 10 mM after control sampling. Inspired andexpired water samples were obtained 30 minutes and 60 minutesduring high calcium treatment.All branchial net proton excretion rate are expressed as62umol/L.kg and are reported as mean values ± standard error.Statistical comparisons were made with Student's t-test forpaired samples. Regression analyses were used to describerelationships between variables.63RESULTSThe relationship between net proton excretion across fishgills and the inspired water pH is illustrated in Figure 8. Whenwater pH is below 5.5, there is no net proton excretion; that is,any change in water pH can be accounted for by CO 2 hydration /HCO3 - dehydration and / or NH 3 protonation. Net proton excretionwas completely inhibited at low external water pH. When water pHis between 5.5 and 8.0, net proton excretion increased linearlywith inspired water pH. Branchial net proton excretion wasmaximal when inspired water pH was above 8.0, which is higherthan blood pH in the fish.Figure 9 shows the relationship between net proton excretionand expired water carbon dioxide levels in a neutral environment.The figure contains all the control net proton excretion ratesand the associated Pco 2 values in expired water. There is ageneral tendency for increased Pco 2 in expired water to beassociated with an increase in net proton excretion.The removal of sodium from the external water caused amarked decrease in branchial proton excretion (Figure 10). Thisreduction persisted during the total 60 minutes of the cholinechloride treatment period and proton excretion was completelyrecovered 30 minutes after sodium was returned to the externalwater. Addition of 9-AA to sodium-free external water caused atransient recovery in net proton excretion across fish gills, but60 minutes after the onset of the 9-AA treatment64Figure 8. The relationship between the branchial net protonexcretion and the inspired water pH, expressed by aregression curve (R2 = 0.853).656 6Figure 9. The relationship between branchial net proton excretionand expired water carbon dioxide levels. The linearregression line has a R 2 value of 0.77. The slope of theregression line is significantly different from zero(P<0.05).67,,68Figure 10. Branchial proton excretion rate of rainbow trout incontrol and sodium-free water. * indicates a significantdifference between control and treatment values (P<0.05);Bars show standard errors; N=6.69300250200XW JC^150O oo- E0--1-, 100500^Control^Choline^Choline Recovery^Omin^30min^60min 90minproton excretion was again significantly lower than controllevels (Figure 11).Branchial proton excretion rate was slightly elevated duringthe one hour exposure of fish to high calcium water (Figure 12).The differences, however, are not statistically significant.71Figure 11. Branchial proton excretion rate of rainbow trout incontrol, sodium-free water and sodium-free water plus 0.1 mM9-anthroic acid. * indicates a significant differencebetween control and treatment values (P<0.05); Bars showstandard errors; N=6.72C0. --6-,X• -11-1O 0O EQ.)z4003503002502001501 00500^Control^Choline^9 —AA^9—AAOmin^30min^60min 90minFigure 12. Branchial proton excretion rate of rainbow trout incontrol and high calcium (10 mM) water. Bars show standarderrors; N=6.74C. _0a)o- CT,X -YL.L1CO 0O EQ_200150100500Control^Calcium^CalciumOmin 30min 60minDISCUSSIONThe relationship between net proton excretion across fishgills and inspired water pH, shown in Figure 8, is very typicalof proton transport mediated by an electrogenic proton pump. Atconstant serosal pH, net proton secretion in turtle bladderincreased linearly with luminal pH over the physiological rangeof urine pH (4.4 - 7.4). Proton secretion was maximal at higherpH (Steinmetz, 1986). A linear relationship between protonexcretion and mucosal pH over a limited range was also reportedin frog skin by Ehrenfeld et al (1985). This indicates that theelectrochemical gradient for protons across the membrane is afundamental regulator of proton transport.When external water pH was lower than 5.5, the ApH acrossthe epithelial membrane is > 2.4 pH unit (assuming blood pH offish was 7.9, see Lin and Randall, 1990). The electrochemicalgradient for protons was too great for the proton pump to workagainst, therefore no hydrogen ions excluded across the gillepithelium. The reversal of gill transepithelial potentialassociated with low water pH (McWilliams and Potts, 1978; Ye etal, 1991) probably resulted from the shut-down of the protonpump. Within the pH range of 5.5 to 8.0, more protons weretransported by the active pump as the proton electrochemicalgradient apposing on the pump decreased. Above inspired water pHof 8.0, the proton electrochemical gradient was favourable toproton excretion, however, branchial proton excretion plateaued,76probably because the electrogenic proton pump was now operatingat its maximal capacity.Elevated carbon dioxide levels in expired water, which is anindicator of the CO 2 level in venous blood, appear to enhanceproton excretion (Figure 9). This was also observed in toadbladder (Al-awqati, 1978), turtle bladder (Arruda et al, 1990)and frog skin (Ehrenfeld and Garcia-Romeu, 1977). In all theseepithelia carbon dioxide provides a source of hydrogen ions forthe proton pump when it is hydrated, the reaction catalysed bycarbonic anhydrase in the intercellular compartment. The sametheory probably applies to the fish gill epithelium becauseacetazolamide, a carbonic anhydrase inhibitor, also inhibitsproton excretion in fish gills (Chapter 1, Figure 5).Removal of sodium from the external water resulted in areduction of branchial proton excretion (Figure 10), probably dueto the lack of a counterion to diminish the negative potentialgenerated by the proton pump on the inner side of the apicalmembrane. This then increases the electrochemical gradientagainst which the proton pump must operate and accordingly protonexcretion is reduced. Substitution of external Na + by Mg' or K*was observed to cause a hyper-polarization of the apical membraneand inhibited proton excretion in frog skin under open circuitconditions (Harvey and Ehrenfeld, 1986).Accumulation of negative potential in the apical membrane offish gills was also expected when 0.1 mM amiloride was added tothe external water and blocked the sodium conductive channels.77However 0.1 mM amiloride also inhibited chloride influx acrossfish gills (Perry and Randall, 1981), which might alleviate thebuild-up of the negative potential against the proton pump. Thisexplanation was supported by the fact that addition of 9-anthroicacid (9-AA) caused a transient recovery of branchial protonexcretion under sodium-free conditions (Figure 11). 9-AA wasshown to inhibit a Cl - conductance in the apical membrane ofcanine tracheal epithelium (Welsh, 1984) and frog kidneyperitubular cell (Oberleithner et al, 1983). Luminal addition of0.1 mM 9-AA inhibited the electrogenic HCO 3 - secretion initiatedby the proton pump in turtle bladder by blocking the recycling ofCl - via a chloride channel in the luminal membrane (Stetson etal, 1985). Application of 9-AA to the mucosal side of fish gillsmight inhibit chloride influx and therefore temporary reduce theelectrical gradient against which the proton pump must operate.Elevated calcium levels in external water had no significantstimulating effect on proton excretion across fish gills duringthe one hour experimental period (Figure 12). It was demonstratedby Avella et al (1987) that high calcium levels in freshwater (10mM) stimulated the proliferation of chloride cells and the sodiuminflux by 2.5 times. But this stimulating effect only appearedafter a long-term acclimation (7-15 days). Insertion of a protonpump into the apical membrane of turtle bladder epithelium wasmediated by cell calcium (Adelsberg and Al-Awqati, 1986). Thetime course of the study reported here was probably too short tocause any change in chloride cell numbers or proton pump activity78in fish gills.79SECTION II: PROTON-ATPASE ACTIVITY IN GILL TISSUE OF RAINBOWTROUTChapter 3: Inhibitor sensitivity and classification of theproton-ATPase in gill tissue80SUMMARYN-ethymaleimide-sensitive ATPase activity was measured incrude homogenates of gill tissue from rainbow trout using acoupled-enzyme ATPase assay in the presence of EGTA, ouabain andazide. This NEM-sensitive ATPase activity, determined to be about1.5 umol/mg.pr ./h. at 15 °C for freshwater trout, was alsoinhibited by other proton-ATPase blockers such as DCCD, DES,PCMBS and Bafilomycins. It is concluded, therefore, that the NEM-sensitive ATPase activity was generated by a proton-translocatingATPase. Since this NEM-sensitive ATPase was also sensitive to theplasma membrane ATPase inhibitor vanadate, I speculate the H +-ATPase in fish gill is a plasma membrane type.81INTRODUCTIONThe Na +/H+ exchanger in fish gill epithelium was postulatedto be the major pathway for Na + uptake and acid excretion (Wrightand Wood, 1985). The sodium concentration in fresh water,however, is usually lower than 1 mM, and the intracellular sodiumconcentration in the gill epithelial cell, although lowered bythe Na +-K+-ATPase in the basolateral membrane, is much higherthan 1 mM. Intracellular sodium ion activity in frog skinepithelium was reported to be 6.2 mmo1/1 (Harvey and Ehrenfeld,1986). The sodium electrochemical gradient across the apicalmembrane could not drive Na +/H+ exchange. An alternativemechanism which will account for Na + and 1-1 + transport in oppositedirections is an electrogenic H +-translocating-ATPase coupledwith a sodium conductive channel (Chapter 1), as demonstrated infreshwater frog skin (Ehrenfeld et al. 1985). This so calledproton pump will consume ATP, actively exclude hydrogen ionsacross the membrane and generate a negative potential inside theapical membrane, which will then drive sodium influx via thesodium channel.The existence of an H +-ATPase is well documented not only infreshwater frog skin, which has the same Na + uptake function asFW fish gills, but also in other tight epithelia such as turtleurinary bladder (Steinmetz and Andersen, 1982) and mammalianrenal collecting tubule (Gluck and A1-Awqati, 1984; Ait-Mohamedet al. 1986). N-ethymaleimide, a covalent SH-reactive reagent, is82a H+-ATPase inhibitor commonly used to identify H +-ATPase indifferent organisms (Pedersen and Carafoli, 1987).My previous in vivo studies (Chapter 1 and 2) showed thatproton excretion across the gill epithelium of freshwater troutwas sensitive to external pH, Pco2, vanadate (a plasma membraneATPase inhibitor) and acetazolamide (a carbonic anhydraseinhibitor), but was not sensitive to 0.1 mM of amiloride whichwill block the Na' influx across the gills of rainbow troutcompletely (Wright and Wood, 1985). All these characteristics aretypical for II + transport mediated by H +-ATPase in other tightepithelia and thus indicated the presence of a proton pump in thegill epithelium. The objective of these studies was to measureH+-ATPase activity directly in crude homogenates of gill tissueand examine its sensitivity to proton-ATPase inhibitors includingNEM, DCCD, PBMBS and bafilomycins. These drugs are usually tootoxic for in vivo application.83MATERIALS AND METHODSExperimental animalsRainbow trout, Oncorhynchus mykiss (Walbaum), weighing 300-500g were kept in aerated, dechlorinated Vancouver tap water([Na +], 0.89 mM; [CaTh 0.03 mM; [C1A, 0.92 mM) at 10-15 °Cambient temperature. Animals were fed commercial trout pelletstwice a week and terminated at least one day before usage.Preparation of gill tissue homogenatesA crude homogenate of rainbow trout gill tissue was preparedusing a method modified from Zaugg (1981). Fish were killed witha blow to the head. The gills were perfused through the heartwith heparinized saline in order to clear red blood cells. Gillfilaments (approximately 1 g wet weight) were trimmed fromsupporting arches and immersed in 2 ml of a cool homogenatemedium I containing 300 mM sucrose, 2 mM EGTA, 1 mMdithiothreitol and 100 mM tris-HC1 at pH 7.3. Tissue was thenhomogenized with a Kontes micro ultrasonic cell disrupter for 20strokes, 2m1 distilled water was added to the homogenates andanother 20 strokes were employed to ensure all filaments weredisintegrated. The diluted homogenates were centrifuged for 7 minin a Janetzli laboratory table centrifuge (model T32c) at about4000 rpm (2000 RCF). Supernatant solutions were discarded, andpellets suspended in 1 ml of homogenate medium II (homogenatemedium I containing 6% CHAPS, a zwitterionic detergent) werehomogenized twice for 20 strokes each. The resulting homogenates84were centrifuged as before and supernatant solutions were removedfor ATPase assay usage. We found the ATPase activity in thesupernatant was stable for at least one month when stored in a -80 °C freezer. Protein concentration of the supernatant was around6-8 mg/ml.Determination of NEM -sensitive ATPase activityNEM-sensitive ATPase activity was measured by a modifiedcoupled-enzyme ATPase assay used for determination of Na+-K+-ATPase activity (Scharschmidt et al. 1979). The formation of ADPcatalyzed by ATPase is coupled with NADH oxidation in thepresence of excess pyruvate kinase (PK), lactate dehydrogenase(LDH) and phosphoenolpyruvate.ATPaseATP^ ADP + Pi^(A)PKADP + phosphoenolpyruvate ^ pyruvate + ATP (B)LDHPyruvate + NADH + H + lactate + NAW^(C)Stock reaction buffer containing 130.9 mM Tris (pH 7.4 at15 °C), 1.05 mM EGTA and 13.09 mM KC1 was prepared in advance. Onthe day of assay NADH, phosphoenolpyruvate, ouabain and sodiumazide were added to the reaction buffer in an amount necessary tobring their concentrations to 0.52 mM, 2.62 mM, 2.12 mM and 5.24mM, respectively. NEM was added to half of the reaction buffer toa concentration of 1.06 mM. Tris ATP was dissolved in 200 mMMgCl 2 solution to yield a 200 mM concentration. LDH-PK enzymemixture (1000 units each per ml) was purchased from Sigma. All85reagents were kept on ice. To perform the assay, 0.945 mlreaction buffer containing NADH and phosphoenolpyruvate, 0.025 mlATP-MgC1 solution and 0.01 ml LDH-PK mixture were added to a 1.5ml cuvette. The reaction was begun by adding 0.02 ml crudehomogenate and mixing the contents of the cuvette by inversion.The final 1 ml reaction mixture thus contained 125 mM Trisbuffer, 1 mM EGTA, 12.5 mM KC1, 5 mM NaN 3 , 2 mM ouabain, 5 mMMgC1 2 , 5 mM ATP, 2.5 mM phosphoenolpyruvate, 0.5 mM NADH and 10units each of LDH and PK (with or without 1 mM NEM). Theoxidation of NADH was continuously monitored at 340 nm at 15 °C inthe temperature-controlled cuvette compartment of a continuouslyrecording spectrophotometer (Perlin-Elmer Lambda 2). ATPaseactivity was calculated from the slope of the linear portion ofthe tracing, the NADH mM extinction coefficient, the volume ofthe reaction mixture, and the milligrams of crude homogenateprotein added:ATPase activity (umol pi/mg.pr.h.)slope (OD units/h.)^1 mlx(OD units.m1) protein (mg)6.22 (^umolProtein concentrations in the crude homogenates were determinedby the method of Bradford (1976). The differences between theATPase activity with and without NEM represent the NEM-sensitiveATPase activity.Preliminary tests were conducted to determine theappropriated amount of enzyme or homogenate for the assay and the86relationship between activity and temperature in the range of 10-15 °C was tested to be linear. pH optimum was not tested butaccording to the literatures, H +-ATPase has a broad pH optimalrange 7.0-7.5. My reaction mixture was well buffered to 7.3.Application of other inhibitorsIn experiments with the inhibitors DCCD, DES and PCMBS,stock solutions were made by dissolving the drugs in 100% ethanoland then adding them to the reaction mixture to the requiredconcentration. Stock solution of bafilomycins (purchased from Dr.Altendorf, Fachbereich Biologie/Chemie, U. Osnabruck, FRG) wasprepared in dimethylsulfoxide (Bowman et al. 1988). Controlsamples containing the proper amount of solvent were assayedsimultaneously. KNO 3 , acetazolamide and sodium vanadate are watersoluble and assays were performed in the same way as for NEM.Since the effect of all these inhibitors on PK and LDH wasunknown, an experiment was designed to test their interferencewith the NADH oxidation reaction. To omit the reaction (A), Iused ADP to replace ATP in the original assay. The initial O.D.of the 1 ml reaction mixture containing 125 mM Tris buffer, 1 mMEGTA, 12.5 mM KC1, 5 mM NaN 3 , 2 mM ouabain, 5 mM MgCl 2 , 2.5 mMphosphoenolpyruvate, 0.5 mM NADH and 10 units each of LDH and PKand appropriate amount of inhibitor was recorded. Then 25 ul 200mM ADP was added to the reaction mixture and final O.D. wasrecorded 1 minute later. A significant decrease of O.D. indicatesa complete reaction without interference. If O.D appeared to beunchanged, the drug being tested was effecting the NADH oxidation87Table 3. Test of inhibitor interference with PK or LDH in theNADH oxidation reaction. - indicates no interference; + indicatesminus interference; ++ indicates serious interference.1 mM NEM^ ---+5 mM Vanadate^++1 mM DCCD0.025 mM DES0.01 mM PBMBS0.025 Bafilomycins100 mM KNO30.1 mM Acetazolamide0.25 mM NDB-C1^+0.5 mM NDB-Cl ++0.1 mM Amiloride^+0.5 mM Amiloride ++1 mM Amiloride ++1.2 mM NEM0.1 mM Vanadate0.5 mM Vanadate88reaction itself and can not be used in the ATPase assay. Table 3summarized the result of this drug test. The effect of NDB-Cl, aspecific inhibitor for H+-ATPase, and amiloride, which appearedto inhibit branchial proton excretion (Chapter 1), on gill tissue1-1+-ATPase activity cannot be examined. But all other inhibitorscan be trusted to appose their effect on ATPase only.Statistical analysisAll ATPase assays were performed in triplicate. In Figuresand Tables, data are presented as mean ± standard error. Resultshave been statistically analyzed using unpaired Student's t-testbetween appropriate sample means. 5% was taken as the fiduciallimit of confidence.89RESULTSA substantial amount of the ATPase activity in the crudehomogenates of gill tissue, in the presence of azide, ouabain andEGTA, was sensitive to NEM. Figure 13 shows the ATPase activityof gill tissue in response to various concentration of NEM. NEMcauses a dose-dependent inhibition of ATPase activity. Maximalinhibition is observed in 1 mM NEM, which accounted for more than70% of the total ATPase activity (Table 4). The difference ofATPase activity with and without 1 mM NEM is referred to as theNEM-sensitive ATPase activity and, at 15 °C, it was determined tobe about 1.5 ± 0.09 umol/mg.pr.h. for freshwater adapted trout.Vanadate (0.1 mM) suppressed 60% of the ATPase activity ingill tissue (Table 4). When 1 mM NEM and 0.1 mM vanadate wereapplied together, the percentage of ATPase affected increasedslightly only. The combination of 1 mM of NEM and 0.1 mM vanadatesuppressed 80% of the ATPase in crude homogenates, whichindicates that of the 60% ATPase that was sensitive to vanadate,50% is from the NEM-sensitive ATPase.The effect of DCCD was also examined, a maximum of 52% ofthe total ATPase activity was suppressed by 1 mM of DCCD (Table4). A similar dose response curve to that produced by NEM wasobserved with DCCD (Figure 14). The sensitivity profiles of fishgill ATPase towards both NEM and DCCD are similar to those foundfor the inhibition of proton-ATPase in rat kidney (Ait-Mohamed etal, 1986).90Figure 13. NEM-sensitive ATPase activity in the crude homogenatesof trout gill tissue in response to various concentrationsof NEM. Activity is expressed in % assuming the ATPaseactivity is 0 with 1 mM of NEM and 100% without NEM. Eachpoint is the mean ± S.E. of four fish.919 2Figure 14. DCCD-sensitive ATPase activity in the crudehomogenates of trout gill tissue in response to variousconcentrations of DCCD. Activity is expressed in % assumingthe ATPase activity is 0 with 1 mM of DCCD and 100% withoutDCCD. Each point is the mean ± S.E. of four fish.939 4Table 4. Effects of inhibitors on ATPase activity in crudehomogenate of trout gill tissue.Inhibitor^Concentration^Relative activitymM^ %None^ 100NNEM 1 27.08^±^2.36 6Vanadate 0.1 38.18^±^3.01 6NEM + Vanadate 1 + 0.1 20.87^±^2.37 6DCCD 1 47.43^±^0.65 4DES 0.025 36.88 ±^0.82 3PCMBS 0.010 55.35 ±^4.36 3Bafilomycins 0.025 58.02^±^2.25 3KNO3 100 72.25 ±^2.61 3Acetazolamide 0.1 97.19^±^5.66 395DES and PCMBS had maximal inhibitory effects at much lowerconcentrations and also accounted for inhibition of 63% and 45%of the total ATPase activity, respectively (Table 4).Bafilomycins, a very specific and potent inhibitor of vacuolarH+-ATPase (Bowman et al, 1988), significantly inhibited ATPaseactivity of fish gills only at concentrations above 25 uM (Table4), whereas vacuolar H +-ATPase was completely blocked atconcentrations as low as 0.1 uM (Bowman, et al., 1988). Potassiumnitrate, another inhibitor used to distinguish vacuolar H +-ATPasefrom plasma membrane H +-ATPase (Bowman, 1983), caused less than a30% reduction in fish gill ATPase activity at a concentration of100 mM (Table 4), a dosage that is sufficient to inhibit 80% ofthe vacuolar H +-ATPase. 0.1 mM of acetazolamide however, had noeffect on the ATPase activity of fish gills.96DISCUSSIONThese studies demonstrate for the first time the existenceof a NEM-sensitive ATPase in crude homogenates of fish gilltissue. H +-ATPase has been noted to be either bound to membranesor packaged in cytoplasmic vesicles (Arruda et al. 1990). Thecrude homogenates prepared with the current method containedmainly the membrane faction of gill cells, and the NEM-sensitiveATPase that I detected in this study was released by a proteinsolubilizer from the membrane faction. Soluble cell material,mitochondria, cytoplasmic vesicles and other organelles wouldhave been discarded in the supernatant of the firstcentrifugation (2000g) because much higher relativecentrifugation forces are required to spin down this materialLittle ATPase activity was found in the discarded supernatant butprotein solubilizer was never applied. If there was H +-ATPasepackaged in cytoplasmic vesicles, perhaps protein solubilizer wasrequired to release them for subsequent detection.NEM is an alkylating agent that is relatively selective forsulfhydryl groups (SH-). It inhibits vacuolar H +-ATPase in anATP-protetable manner in concentrations under 10 uM (Forgac,1989; Pedersen and Carafoli, 1987). Phosphorylated ATPases(including Na +-K+-ATPase, Ca'-ATPase and plasma membrane H +-ATPase) are sensitive to higher concentrations (100 uM - 1 mM) ofNEM (Forgac, 1989). Since the assay was carried on in thepresence of EGTA, a Ca ++ chelator which should abolish Ca ++-ATPase97activity; azide, a mitochondrial H +-ATPase inhibitor; andouabain, a Na +-K+-ATPase inhibitor; the contribution of unrelatedATPase activity was minimal. Thus, the ATPase activity in thecrude homogenate of gill tissue that was sensitive to 1 mM of NEMprobably originated from plasma membrane H +-ATPase.30% of the ATPase activity was NEM-insensitive and is ofunknown origin. Unidentified NEM-insensitive ATPase was alsodetected in mammalian kidney (Ait-Mohamed et al. 1986; Garg andNarang, 1988). Bornancin et al. (1980) presented evidence of C1 - -HCO3 - -ATPase in microsomes from the gill plasma membrane ofrainbow trout. This might account for the NEM-insensitive ATPaseactivity in the gill tissue crude homogenates.The gill ATPase sensitive to NEM was also sensitive tovanadate. Orthovanadate ion, VO 43 , acting as a phosphatetransition analogue, blocks the formation of phosphorylatedintermediates in all P-type ATPases. 0.1 mM vanadate was reportedto inhibit branchial proton excretion in freshwater trout(Chapter 1). Urinary acidification by turtle bladder (Arruda, etal. 1981) and proton transport across freshwater frog skin(Ehrenfeld, et al. 1985), both mediated by H +-ATPase, are alsovanadate-sensitive. However, vanadate fails to inhibit NEM-sensitive ATPase activity and proton transport in mammaliankidney (Gluck and Al-Awqati, 1984; Ait-Mohamed et al, 1986) whichwas believed to be a vacuolar H +-ATPase (Forgac, 1989).DCCD inhibits mitochondrial, vacuolar and plasma membraneH+-ATPase by binding to the c-subunit of the hydrophobic channel98portion (Pedersen and Carafoli, 1987). Mitochondrial H +-ATPasehas the highest sensitivity to DCCD (0.1-0.5 uM), followed byvacuolar H +-ATPase (1-10 uM) and then plasma membrane H +-ATPase(10-100 uM). The dose response curves of the ATPase activity ingill tissue towards NEM and DCCD are very similar to thosereported for rat kidney H +-ATPase (Ait-Mohamed, et al., 1986).0.05 mM of DCCD inhibited proton excretion and decreased pHi infrog skin (Harvey and Ehrenfeld, 1986). DES was also reported toinactivate H +-ATPase in the F o moiety level (Pedersen andCarafoli, 1987). 0.1 mM DES markedly reduced the proton excretionacross frog skin. Proton transport mediated by H +-ATPase inbovine kidney medulla was completely blocked by 10 uM PCMBS, aSH-group reagent, (Gluck and Al-Awqati, 1984), which partiallyinhibited H+-ATPase activity in fish gills.Bafilomycins is a macrolide antibiotic that has a specificand potent inhibitory effect on vacuolar H +-ATPase (Bowman et al.1988). Mitochondrial H+-ATPase is resistant to up to 1 mM ofbafilomycins whereas vacuolar H +-ATPase is completely inhibitedby <0.1 uM of the antibiotic. Phosphorylated ATPase exhibitsintermediate sensitivities with 4 0 values between 10-100 uM. Thesensitivity level of the ATPase in gill tissue to bafilomycins iswithin this intermediate range.Vacuolar H+-ATPase also demonstrates a unique sensitivity toKNO3 with a 4 0 value of about 50 mM (Bowman, 1983). Theresistance of fish gill ATPase to nitrate indicates that the H +-ATPase we measured is not a vacuolar type. The balance of all99these pharmacological properties indicates that H +-ATPase in fishgills is a plasma membrane type and not a vacuolar type.Acetazolamide was demonstrated to inhibit luminalacidification in turtle bladder by stimulating the endocytosis ofapical membrane (Dixon et al. 1988; Graber et al. 1989). Theinhibition appeared to be independent of cell pH, which ruled outthe possibility of a secondary affect due to carbonic anhydraseinhibition. The inhibitory effect of acetazolamide on in vivo proton excretion (Lin and Randall, 1991) can not be reproduced inthe in vitro ATPase assay, indicating that acetazolamide has nodirect effect on H +-ATPase itself.100Chapter 4: Environmental and hormonal regulation of the proton-ATPase in fish gills101SUMMARYThe effects of sodium, calcium and CO 2 levels inenvironmental water on the proton-ATPase activity in fish gillswere investigated. Sodium concentration in the external media wasthe primary regulator of the H +-ATPase in fish gills, with lowsodium levels associated with high H +-ATPase activity. It wasconcluded therefore, that the major role of the H +-ATPase in thegill epithelium was to facilitate Na + uptake from fresh water.High external calcium concentration had a marked stimulatingeffect on H +-ATPase activity in fish gills when the sodium levelwas low. Environmental hypercapnia induced a 70% increase in theH+-ATPase activity in fish gills. The effect of the steroidhormone cortisol was also investigated. H +-ATPase activity waselevated in freshwater fish after chronic cortisol infusion, butnot in seawater fish. The H +-ATPase in fish gills also plays arole in acid-base regulation.102INTRODUCTIONThe results from Chapter 3 confirm the existence of a protonpump, or proton-ATPase in the gill epithelium of freshwaterrainbow trout. The first apparent function of the proton-ATPasein fish gill is to generate an electrical gradient favourable forsodium uptake from freshwater, and the effective operation ofthis active pump will be critical for the maintenance of osmotichomeostasis in fish living in a very dilute media. For seawaterfish, however, the negative potential generated by H +-ATPase willretard the animal's ability to excrete sodium in seawater.Euryhaline fish such as rainbow trout, therefore, should be ableto regulate the proton-ATPase activity in the gill epitheliumaccording to environmental salinity.Influx of sodium in freshwater fish is effected by differentenvironmental factors such as hypercapnia (Goss et al. 1992) andwater Ca" levels (Avella et al. 1987). If the proton pump doesprovide the driving force for the sodium uptake in freshwaterfish, these factors could act on 1-1 +-ATPase and variation of Na +influx would be the secondary outcome. The objective of thesestudies, therefore, was to examine the effect of environmentalvariations on H +-ATPase activity in fish gills.It has also been suggested that the molecular target of themineralocorticoid hormone aldosterone on urine acidification isH+-ATPase in the kidney collecting tubule of mammals (Mujais,1987; Garg and Narang, 1988; Khadouri et al. 1989). Cortisol, the103equivalent steroid hormone in fish, increased in rainbow troutwhen they were transferred from freshwater to seawater (Foskettet al, 1983) and when they were acclimated to deionized water(Perry and Laurent, 1989). The effect of cortisol on gill H +-ATPase activity has yet to be investigated.H+-ATPase in fish gills might also play a role in the acid-base balance of the whole animal and in the regulation ofintracellular pH, as reported in frog skin (Ehrenfeld et al,1990), turtle bladder (Cohen and Steinmetz, 1980) and mammaliankidney (Sabatini et al, 1990). An acidosis, caused byenvironmental hypercapnia or NH 4C1 injection and an alkalosis,caused by alkaline water exposure, was induced in trout andeffects on H+-ATPase activity was determined to illuminate thesignificance of gill proton-ATPase in acid-base regulation.104MATERIALS AND METHODSFreshwater rainbow trout weighing between 150-500 g weremaintained in large fibreglass tanks supplied with flowingdechlorinated Vancouver tap water. Seawater adapted rainbow trout(400-600 g) were obtained by acclimating freshwater rainbow troutto seawater (34-38% 0) for 8 to 10 weeks in the large fibreglasstanks at DFO West Vancouver laboratory. After the fish weresubjected to one of the treatments described below, they weresacrificed and crude homogenates of gill tissue containing mainlythe membrane faction, were obtained by the method described inChapter 3.The H+-ATPase activity in gill tissue of fish underdifferent treatments was determined by the modified couple-enzymeassay in a temperature-controlled spectrophotometer using N-ethymaleimide as a specific inhibitor (see Chapter 3 fordetails). NEM-sensitive ATPase is equivalent to H +-ATPaseactivity (Chapter 3). Ouabain-sensitive ATPase, which isequivalent to Na +-K+-ATPase, was measured in gill tissue usingthe method from Scharschmidt et al (1979). All ATPase assays wereperformed in triplicate and data are presented as mean ± standarderror. Student's two-tailed t-test and one-way ANOVAs (analysisof variance) were used to test for significant differences at the5% level rejection between means.Acclimation to various external sodium and calcium concentrationsFour kinds of external media were prepared by dissolving105NaC1 and/or CaC1 2 into dechlorinated Vancouver tap water: 100 mMNaCl; 100 mM NaCl plus 1 mM CaC1 2 ; 1 mM CaC1 2 and 10 mM CaC1 2 .Fish were placed into a 100 litre opaque fibreglass tanks(density < 25g/L) filled with different external media foracclimation period of 10-14 days. The tanks were well-aerated andtemperature was maintained at ambient levels with cooling coils.The external media were changed daily to prevent ammoniaaccumulation.Chronic cortisol treatmentPlasma cortisol levels were elevated in both freshwater andseawater adapted trout by implantation of Alzet mini osmoticpumps containing cortisol (Reid and Perry, 1991). Mini osmoticpumps were loaded with cortisol (hydrocortisone 21-hemisuccinate,Sigma) or saline (in sham treatments) and surgically implantedinto the peritoneal cavity of anaesthetized fish. The nominalcalculated plasma cortisol concentration was 200 ngm1 -1 and fishwere sampled for gill tissue and blood (by caudal puncture) 7days after the implantation. Plasma cortisol level was measuredusing a Gammacoat 1251 cortisol radioimmunoassay kit (IncstarCorp.).Hypercapnia treatmentRainbow trout were placed into a 100 litre opaque fibreglasstank supplied with flowing aerated dechlorinated Vancouver tapwater (water pH 5.8-6.2). Fish were allowed to acclimate for 24h.Control fish were sampled right before the 48h hypercapniatreatment (2% CO 2 in air, mixed using a Wosthoff gas mixing pump,106water pH 5.0-5.5) and fish were sampled at 6, 24 and 48hfollowing hypercapnia. Fish were then allowed to recover for 24hand samples were taken at 6 and 24h of recovery.NH4C1 injectionRainbow trout were fitted with dorsal aortic catheters underMS-222 anaesthesia (1:10000 in NaHCO 3-buffered freshwater). Fishwere then allowed to recover for 24h in a sectioned plexiglasbox. 2 mlkg -1 body mass of saline (control group) or 1 moll -1NH4C1 in saline was injected daily for 2 days into the dorsalaorta of the fish over a period of approximately 5 min. 48h afterthe first injection, blood samples were taken for pH measurementusing a microcapillary pH electrode (Radiometer G279/G2) coupledto a PHM84 pH meter, and fish were sacrificed for gill tissuesampling.Exposure to alkaline waterFreshwater trout were held in a 200 litre fibreglass tanksupplied with flowing aerated dechlorinated tap water (pH 6-7).After 24 hours of acclimation, control fish were sampled. Thenwater pH was increased to 10 by metering a concentrated NaOHsolution into the tank using a peristaltic pump. In another setof experiments, water hardness was also increased to 50 ppm([Ca ++ ] = 1.25 mM) by metering a concentrated CaC1 2 solution intothe tank with a separate peristaltic pump. Fish were sampled onthe first, second, third, fourth and sixteenth day of thealkaline soft water or alkaline hard water exposure.107RESULTSIn rainbow trout acclimated to different external levels ofNa+ and / or Ca", gill NEM-sensitive ATPase activity decreased asNa+ acclimation level increased (Figure 15). NEM-sensitive ATPaseactivity was significantly lower in fish acclimated to 100 mM ofNa+ (with or without Ca"). Seawater adapted rainbow trout hadonly 1/3 of the NEM-sensitive ATPase activity of their freshwatercounterparts. The addition of Ca" to high Na + water made nodifference to NEM-sensitive ATPase activity. Increased Ca' levelin low Na + media, however, had a marked stimulating effect andresulted in a two fold increase on NEM-sensitive ATPase activityin fish gills (Figure 15).Ouabain-sensitive ATPase activity, which is equivalent toNa +-K+-ATPase activity, decreased when trout were acclimated tofreshwater containing high sodium and high calcium levels, butincreased when they were acclimated to low sodium but highcalcium water (Figure 16).Chronic cortisol infusion in freshwater rainbow trout causeda 170% increase in plasma cortisol level (Figure 17) and a 30%increase in NEM-sensitive ATPase activity in gill tissue (Figure18). Seawater adapted animals on the other hand, showed noresponse of NEM-sensitive ATPase activity (Figure 18) to asimilar cortisol treatment although their plasma cortisol levelsincreased four-fold (Figure 17). There was no significantdifference in gill tissue Na +-K+-ATPase activity between sham and108Figure 15. NEM-sensitive ATPase activity in the gill tissue ofrainbow trout acclimated to various Na + and Ca ++ levels inthe external media for 10-14 days. Mean ± S.E. * indicates asignificant difference from the control value (P < 0.05).Number in brackets indicates the sample size.109No1mMCa100mM 100mMNa4. 2.00E^1.5C1.00.50.0ControlFW1mM 1 OmMCo^CoSWCa.)(r)0CLH-Q. _cr)GUU)L.1zFigure 16. Ouabain-sensitive ATPase activity in the gill tissueof rainbow trout acclimated to various Na' and Ca ++ levels inthe external media for 10-14 days. Mean ± S.E. * indicates asignificant difference from the control value (P < 0.05).Number in bracket indicates the sample size.111ControlFWa)—-C0_2.5(7).E 2.0-6E 100mMNa^Na1mMCaSW^1mM 1 O m MCa^Co112Figure 17. Plasma cortisol concentration in freshwater andseawater rainbow trout after 7 days of chronic cortisoltreatment. Mean ± S.E. * indicates a significant differencefrom the sham treatment value (P < 0.05). Number in bracketindicates the sample size.113600500Q)400OE›.„3000 200(r)0_1000Sham Cortisol^Sham CortisolFreshwater SeawaterFigure 18. NEM-sensitive ATPase activity in the gill tissue offreshwater and seawater rainbow trout after 7 days ofchronic cortisol treatment. Mean ± S.E. * indicates asignificant difference from the sham treatment value (P <0.05). Number in bracket indicates the sample size.115E 1.503 02 52 0E1.00.50.0Sham Cortisol^Sham Cort'solFreshwater Seawatercortisol treatments in either freshwater or seawater animals(Figure 19).Hypercapnia treatment in freshwater fish induced a rapidincrease in the NEM-sensitive ATPase activity which stabilized ata level twice that of normocapnia (Figure 20). The elevated NEM-sensitive ATPase activity returned to control levels after 24hrecovery.A blood acidosis was associated with NH 4C1 injection (Table5). NEM-sensitive ATPase activity, however, was not altered byNH4C1 injection. Daily injection of NH 4C1 was performed on twofish for five days and no change of NEM-sensitive ATPase activitywas observed (data not shown).Short term (4 days) exposure to alkaline soft water inducedno change in NEM-sensitive ATPase activity in fish gill tissue(Figure 21). But long term (16 days) exposure resulted in asignificant decrease of H+-ATPase activity. Alkaline hard water,on the other hand, stabilized the H +-ATPase activity in thecontrol level for the whole period except for a transientdecrease in activity at the beginning of the exposure (Figure21). The fish subjected to soft water and hard water treatmentswere from different fish stocks, which may account for thedifferent control levels of H +-ATPase activity observed.117Figure 19. Ouabain-sensitive ATPase activity in the gill tissueof freshwater and seawater rainbow trout after 7 days ofchronic cortisol treatment. Mean ± S.E. Number in bracketindicates the sample size.1183 02 50.50.0Shorn Cortisol^Shorn Cort . solFreshwater Seawater119Figure 20. NEM-sensitive ATPase activity in the gill tissue offreshwater rainbow trout during 48 hours of hypercapniatreatment and 24 hours recovery. Mean ± S.E. (N=6) *indicates a significant difference from the control value (P< 0.05).1200 6^24^48 6^24HourControl2.0 —0.0RecoveryHypercapniaFigure 21. NEM-sensitive ATPase activity in the gill tissue offreshwater rainbow trout during control and 16 days exposureto soft (N=6) and hard (N=4) alkaline water (pH=10). Mean ±S.E. * indicates a significant difference from the controlvalue (P < 0.05).122. _1 . 401,73ci) 2^1 .2cnQ_• C12_ 1.0_ 0 0.8• E_C00 . 60. 4• Softwater0 Hardwater0.20.0^"^I I CON 1 2 3 4 168DayTable 5. Plasma pH and NEM-sensitive ATPase activity in gilltissue of NH 4C1-injected rainbow trout (n=6).Plasma pH^NEM-sensitive ATPase Activityumol/mg.pr.hSaline-Injected 7.70 ±^0.12 0.985 ±^0.123NH4C1-Injected 7.37 ±^0.22* 1.020 ±^0.224*Indicates a significant difference between saline and NH 4C1injection values.124DISCUSSIONEhrenfeld et al. (1985) demonstrated that the sodiumabsorption across freshwater frog skin was mediated by an activeproton pump indirectly coupled with a sodium channel, instead ofa Na +/1-1+ exchange. The proton pump in frog skin was inhibited byDCCD and vanadate. Since the gill epithelium in freshwater fishis a tight epithelium similar to that of freshwater frog skin, itis reasonable that they have the same mechanism to solve the sameosmoregulatory problem. The fact that fish acclimated to waterwith low Na + level have higher W-ATPase activity suggests thatthe functional significance of the H +-ATPase in fish gills is togenerate a electrochemical gradient for Na + uptake from a dilutemedia. When external Na + level is high, H +-ATPase is down-regulated, possibly by endocytosis of membrane protein intointracellular vesicles (Schwartz and Al-Awqati, 1986). Theresidual H +-ATPase in fish gills may play a role in acid-baseregulation, althought it is not required for sodium absorption orsodium excretion in fish living in a high sodium environment.The stimulating effect of Ca" on H +-ATPase could beexplained using the cellular model of Wendelaar Bonga et al (1992), who proposed the existence of an apical Ca' channel.When the external medium has a high Ca' but low Na + level, Ca'could enter the cell via the Ca' channel and reduce thepotential gradient generated by the H +-ATPase and therefore Na +influx. The resulting high Ca' concentration in the cell might125stimulate H +-pump insertion from intracellular vesicles into theapical membrane in order to maintain Na + influx. Exocytosis ofproton pump containing vesicles into the cell membrane is Ca'dependent in turtle bladder epithelium (Adelsberg and Al-awqati,1986).Variations in Ca' level in freshwater environments werereported to affect the gill morphology and sodium influx inrainbow trout (Avella et al. 1987). Na + influx increased 2.5times in fish acclimated to FW + 10 mM CaC1 2 for 15 days and newglobular chloride cells appeared and proliferated in thesecondary lamellae. Fish acclimated to FW + 5 mM CaC1 2 for 5 daysshowed no change in gill morphology or sodium flux, perhapsbecause the duration was too short for morphologicalmodification. In earlier studies (Chapter 2) I found no effect ofcalcium exposure on branchial proton excretion after 2 hours.I have also investigated the effects of chronic infusion ofcortisol on H+-ATPase activity in gill tissue in FW and SWrainbow trout. Aldosterone treatments, either long term (7 days)or short term, were reported to stimulate 1-1 + secretion mediatedby H+-ATPase in the collecting duct of the mammalian kidney (Gargand Narang, 1988; Mujais, 1987). The functionally parallelsteroid hormone in fish is cortisol. Whether aldosterone ispresent in teleosts is uncertain, and little information isavailable about its function in fish (Butler, 1973). Perry andLaurent (1989) have shown that plasma cortisol levels rosetransiently in fish exposed to deionized water. Daily126intramuscular injections of cortisol for 10 days caused anincrease in Na + uptake (Perry and Laurent, 1989). The 30%increase of H+-ATPase activity observed in freshwater troutfollowing chronic cortisol treatment is probably responsible forthis increased Na + uptake. Similar cortisol treatment had noeffect on SW acclimated rainbow trout, indicating that Na +concentration is the predominant regulator of the H +-ATPase infish gills.Another possible function of the H +-ATPase in the fish gillepithelium is acid-base regulation. I induced respiratoryacidosis through hypercapnia treatment and observed an increasedH+-ATPase activity in fish gills under this treatment. Similarhypercapnia treatment in FW catfish was reported to cause amarked increase in Na + influx, which might be correlated with theincreased H+-ATPase activity. The elevated H +-ATPase activitycould be induced by CO 2 mediated proton pump insertion viaexocytosis, as demonstrated in turtle bladder epithelium (Cannonet al. 1985; Arruda et al. 1990). High CO 2 levels reduced theintracellular pH of the proton secreting cell, which increasedthe intracellular Ca ." concentration and in turn stimulated thefusion of cytoplasmic vesicles containing proton pump into theapical membrane. This would then correct the intracellularacidosis.Chronic metabolic acidosis was induced in the fish by NH 4C1injection and resulted in no significant change in H+ -ATPaseactivity. This indicates that high plasma hydrogen ion levels127alone can not stimulate H +-ATPase activity in gill tissue orproton pump insertion to the epithelium. I concluded thatelevated CO 2 levels increased H +-ATPase activity via a depressionof epithelial pH. A metabolic acidosis will only activate H + -ATPase activity if the acidosis is transferred into the gillintracellular compartment. If I follow this argument, thenpresumably NH 4C1 infusion has little or no effect on epithelialpH because H+-ATPase activity was unchanged. Unfortunately, I wasnot able to measure the effects of NH 4C1 injection on gillepithelial pH. The operation of the existing proton pumps wasprobably sufficient to correct the acidosis.The alkalosis, reported to occur immediately in rainbowtrout exposed to pH 10 alkaline water (Yesaki and Iwama, 1992),is more severe in soft water than in hard water. Fish in hardwater were able to maintain acid-base balance because externalcalcium helps to stabilize biological membranes, maintain theintegrity of cell to cell junctions and control ion and waterpermeability across epithelial tissue (McDonald et al, 1983). Itis possible that high gill intracellular calcium levels (due tothe high calcium level in the external water) prevented thewithdrawal of proton pumps from the apical membrane in the faceof a mild gill epithelial alkalosis. Proton-ATPase activitydecreased in fish maintained in soft alkaline water, probably dueto the intracellular alkalosis, but the process occurred veryslowly. This adjustment of W-ATPase activity in the gillepithelium might lessen the acid-base disturbance in the fish by128reducing proton excretion and contribute to the their long termsurvival. Apparently proton pump withdrawal was unnecessary forfish in alkaline hard water when Ca +1- reduced the permeability ofgill epithelium to base equivalents. But reduction of H +-ATPaseactivity in gill apical membrane might be a critical adaptationalresponse to fish in alkaline soft water in order to retainhydrogen ions. Yesaki and Iwama (1992) reported a 100% mortality4 days after fish were exposed to soft alkaline water. No loss offish occurred in my studies, possibly because no surgery waspreformed on these animals.129GENERAL DISCUSSIONThe studies reported in this thesis provide us, for thefirst time, with direct in vivo and in vitro evidence of theexistence of a active electrogenic proton pump, or H + -translocating-ATPase, in fish gills. This proton pump, indirectlycoupled with a sodium conductive channel, instead of aNa +/H+ (NH4+) exchanger, is responsible for the sodium uptake andproton excretion in freshwater fish. As discussed in the GeneralIntroduction, the operation of a Na +/H + (NH 4+) exchanger in theapical membrane of the gill epithelia of freshwater fish isenergetically improbable, because the Na + electrochemicalgradient is unfavourable for the exchanger. Studies on ammoniaexcretion in fish found ammonia excretion was not always coupledwith sodium absorption (Payan, 1978; Avella and Bornancin, 1989)and when it was, the coupling could be explained in terms of anintracellular pH effect (Cameron and Kormanik, 1982; Avella andBornancin, 1989). My studies confirm the absence of theNa +/H+ (NH4+) exchange in the gills of freshwater fish andestablish a new model for sodium and proton transport across fishgill epithelia (Figure 22).The relationship of the proton pump and the other iontransport pathways in the gill epithelium of freshwater fish issummarized in Figure 22. The primary electrogenic proton pump inthe apical membrane consumes ATP and excludes hydrogen ion fromthe intracellular compartment to the external water, and130Figure 22. Hypothetical model of the gas and ion transportpathways in gill epithelium of freshwater rainbow trout.ATP-driven pumps are denoted by filled circles, ionexchanger by opened circles and passive diffusion by arrows.See text for details.131generates a negative potential in the inner side of the apicalmembrane. This negative potential will serve as a driving forcefor sodium uptake from freshwater via a sodium conductivechannel. CO 2 hydration in the intracellular compartment providesone of the sources of protons to the proton pump. HCO 3 - generatedby CO2 hydration can either be excreted to the external water bya C1 -/HCO3 - exchanger or recycled to the serosal side via an anionchannel. The deprotonation of NH 4+ which entered the cell fromthe serosal side through the Na +-K+ (NH4 +)-ATPase pathway in thebasolateral membrane provides another source of protons to thepump. The resulting NH 3 produced will diffuse across the apicalmembrane and be excreted to the water. Na + and sometimes Ca ++enter the cell from the mucosal side via the conductive channelsand act as the counterion for the proton pump. Cl - transferredacross the apical membrane into the cell can also affect theelectric potential generated by the pump and therefore affect theproton pump's operation.The proton pump model solves the energetic problem forsodium uptake from a very dilute medium in fish, but stillexplains the 1:1 stoichiometry for Na + uptake and proton/ammoniaexcretion as demonstrated in freshwater fish by Wright and Wood(1985). Proton excretion via the proton pump can facilitateammonia excretion either by converting NH 3 to NH4+ in theunstirred water layer right next to the gill epithelium or bylowering the intracellular pH and enhancing NH 4 + deprotonation inthe epithelial compartment. The operation of the proton pump133however, is unfavourable to CO 2 excretion because it facilitatesHCO3 - dehydration in the external water and CO 2 hydration in theintracellular compartment. Since the gill epithelial membrane isvery permeable to CO 2 and CO 2 is very soluble in water, fish canmaintain a CO 2 excretion despite this obstruction.The studies of the inhibitor sensitivity, environmental andhormonal regulation of the H +-ATPase in fish gills show that theH+-ATPase in gill tissue is very similar to the H +-ATPase inturtle bladder and mammalian kidney and that branchial protonexcretion is very similar to the proton transport mediated by theproton pump in frog skin and turtle bladder. H +-ATPase in fishgills is sensitive to NEM, DCCD, DES, PCMBS and bafilomycins, toan extent that is characteristic for plasma membrane H +-ATPase(Chapter 3). Both the H+-ATPase in gill tissue and the branchialproton excretion are inhibited by vanadate (Chapter 1 and 3),indicating that the H+-ATPase in fish gills is probably a P-typeas in frog skin and not a V-type as in mammalian kidney.Branchial proton excretion is sensitive to external water pH(Chapter 2), or the electrochemical gradient of hydrogen ion (thepH change in fish blood is negligible), which is a standardfeature of proton transport via proton pumps (Steinmetz, 1985).The effect of ambient CO 2 levels and acetazolamide (Chapter 1 and3) on the branchial proton excretion is also typical for protontransport mediated by proton pumps. The effect of acetazolamidecan be explained either by its direct inhibition on carbonicanhydrase in the gill epithelium or by stimulating the134endocytosis of apical membrane (Dixon et al, 1988; Graber et al,1989), but not by its direct effect on H +-ATPase (Chapter 3). Theeffects of amiloride, 9-anthroic acid and sodium-free water onbranchial proton excretion reveal the electrical linkage betweenNa + , Cl - and H .' in the gill epithelial compartment (Chapter 1 and2). Under open circuit conditions the negative potentialgenerated by the proton pump will slow down its own operation,unless counterions are available to balance the potential. Na + isusually playing this role whereas Cl - rewards the problem, andthe balance of the two effects determines the electric potentialimposed on the proton pump.The major role of the H +-ATPase in the gill epithelium is tofacilitate Na uptake from freshwater. Thus the external Na +level is the primary regulator of H +-ATPase activity in gilltissue (Chapter 4). The higher the external Na + level, the lowerthe H +-ATPase activity in gill tissue. The proton pump appearedto be removed from the membrane as external sodium levelsincreased, through an unknown mechanism. This phenomenon has notbeen reported in other H +-ATPase transport systems, not even infreshwater frog skin which shares the same sodium uptake functionas freshwater fish gills. Ca ++ and cortisol stimulate therecruitment of H +-ATPase to the apical membrane when external Na +levels are low (Chapter 4), and this is probably associated withthe gill morphological changes induced by Ca ++ (Avella et al,1987) and cortisol (Perry and Laurent, 1989) under similarexperimental conditions. Avella et al (1987) discovered that135after 5 days of acclimation in distilled water or 5 to 15 days ofacclimation in freshwater plus 10 mM of CaC1 2 , there was asignificant proliferation of new protruding chloride cells in thesecondary lamellae of rainbow trout. The maximal rate ofbranchial sodium uptake (V..) was also increased in both cases,suggesting an increase of the number of carriers that mightlocate in chloride cells. Perry and Laurent (1989) demonstratedthat the increased sodium and chloride absorption capacity acrosstrout gills after long term cortisol treatment and deionizedwater exposure was also correlated to the proliferation ofchloride cells. It is possible that the proton pump is located inthe branchial chloride cell, and the recruitment of H +-ATPaseinduced by Ca ++ or cortisol stimulates the branchial sodiumabsorption.The H+-ATPase in fish gills is also involved in acid-baseregulation. It appears that the acid-base disturbance of theanimal has to be transferred to the intracellular compartment inthe gill epithelium to cause any adjustment of H +-ATPase activityin gill tissue (Chapter 4). The acidosis induced by NH 4C1injection for example, did not result in an increase of H +-ATPaseactivity in gill tissue because gill epithelium pH was probablyunchanged (Chapter 4). Distinction must be made between thebranchial proton excretion rate, and the H +-ATPase activity ingill tissue, which refers to the amount of enzyme incorporated inthe apical membrane of gill epithelium. When fish were exposed tohigh pH water, an alkalosis was induced in the epithelial cells.136H+-ATPase activity in gill tissue was lower due to the removal ofproton pump from the apical membrane in order to retain hydrogenions (Chapter 4). The branchial proton excretion under thiscondition, on the other hand, was maximal because the protonelectrochemical gradient imposed on the pump favoured theoperation of the remaining proton pumps (Chapter 1). I was notable to examine branchial proton excretion and H +-ATPase activityat the same time, since fish subjected to surgery can not survivepH 10 exposure over 4 days (Yesaki and Iwama, 1992), but protonpump withdrawal by membrane endocytosis did not occur in thefirst 4 days of exposure (Chapter 4). Nevertheless the problemcan be illuminated by the current approaches. The reverse isexpected in fish exposed to acid water. Branchial protonexcretion was completely suppressed by low environmental pH(Chapter 2), but even intact fish had difficulty surviving achronic exposure of pH 4 water and the H +-ATPase activity in gilltissue under this condition can not be examined.The localization of the proton pump in fish gills is aninteresting but unanswered question. It is not difficult tospeculate that the pump is in the apical membrane, because thatis where the H +-ATPase will function as in frog skin, turtlebladder and mammalian kidney. Experimentally drugs and treatmentsapplied to the apical side of fish gills affect the operation ofthe pump (Chapter 1 and 2). External water pH, for instant,exerts an immediate effect on the operation of the proton pumpand the apical membrane is known to be impermeable to hydrogen137ion. Whether the proton pump is housed in the chloride orepithelium cell is difficult to predict. Attempts have been madeto separate chloride cells from epithelial cells using a FACStarcell sorter, but the size and density differences of these twocell lines in freshwater fish gills are too small to permitseparation. Results of morphological studies by other workers arecontroversial. Freshwater fish exposed to deionized water, highcalcium water or treated cortisol shown proliferation andenlargement of branchial chloride cells and enhanced NaC1transporting capacity (Avella et al, 1987; Laurent and Perry,1990; Perry and Laurent, 1989). Environmental hypercapnia,however, caused density increases of epithelium cells infreshwater fish gills, which was apparently associated withelevated Na + uptake rate (Goss et al, 1992). All the abovementioned treatments were tested for H +-ATPase activity infreshwater trout in my studies and all resulted in an increase ofH+-ATPase activity (Chapter 4). Membrane protein with a surfacestructure similar to the proton pump in turtle bladder andmammalian kidney as described by Brown et al (1987) was observedin epithelium cells but not in chloride cells of freshwater fishgills under electron microscope (Perry, personal communications).It is possible that H+-ATPase and sodium channels are housed indifferent cell types, as long as gap junctions exist laterally soall the cells are electrically coupled, as in frog skin. It isalso possible that sodium channels are present in both chloridecells and epithelium cells, and under certain conditions (eg.138respiratory acidosis) those in one particular cell type (eg.epithelium cell) become predominant, as demonstrated in frog skin(Ehrenfeld et al, 1990). More powerful tools such asimmunocytochemistry have to be employed before any conclusion canbe drawn in this matter.Na+-K+-ATPase plays an important role in osmoregulation ofseawater fish and the cellular localization of this primary iontransport pathway was concluded to be in the basolateral membraneof chloride cells in fish gills (Karnaky et al, 1976). There is aweak correlation between the H +-ATPase activity and the Na +-K+-ATPase activity in freshwater rainbow trout (Figure 23) but thiscorrelation disappeared in seawater animals (Figure 24). Themajor morphological change in gill epithelium when fish aretransferred from freshwater to seawater is an increase in thesize and number of chloride cell (from 1% to 13%, Perry andWalsh, 1989). The lack of correlation between H +-ATPase and Na +-K+-ATPase in seawater fish argues against the notation that H+-ATPase is housed in the chloride cell, as suggested by the weakcorrelation between the two in freshwater fish. Ca ++-ATPase infish gills was also located in the basolateral membrane ofchloride cells and provides the major driving force for calciumuptake (Perry and Wood, 1985; Wendellaar Bonga et al, 1992). Aslong as gap junctions exist between chloride cells and epitheliumcells, all the primary and secondary ion transport pathways canbe coupled and the gill epithelium can be viewed as a generalcompartment (Figure 22). When external sodium and calcium levels139Figure 23. The relationship between the NEM-sensitive ATPaseactivity and the ouabain-sensitive ATPase activity infreshwater rainbow trout under different treatments. Thelinear regression line has a R2 value of 0.626. The slope ofthe regression line is significantly different from zero(P<0.05).140Figure 24. The relationship between the NEM-sensitive ATPaseactivity and the ouabain-sensitive ATPase activity inseawater-adapted rainbow trout under control and cortisoltreatments.142are not too low, Na +-K+-ATPase and Ca'-ATPase in the basolateralmembrane may be sufficient to energize the uptake of sodium andcalcium from external water. However, when external sodium andcalcium levels are very low, Fr-ATPase is needed in addition togenerate an immediate driving force from apical membrane and thedeterioration of the intracellular ion concentration on thedriving capacity can be avoid.In summary, substantial evidence was accumulated in thesestudies for the existence of an electrogenic proton pump in fishgill epithelium. Branchial proton excretion exhibitscharacteristics typical of proton transport mediated by protonpump in frog skin, turtle bladder and mammalian kidney. Thebalance of pharmacological studies indicates that the H +-ATPasein fish gills is probably a plasma membrane type. The major roleof the proton pump in fish gills is to facilitate sodium uptakefrom fresh water. Thus the external sodium level is the primaryregulator of its activity and it could also be affected byexternal calcium or cortisol level of the fish. The proton pumpin fish gills is also involved in acid-base regulation. The oldmodel of Na +/H+ (NH 4+) exchanger is rejected because it isenergetically improbable and branchial proton excretion isresistant to 0.1 mM amiloride.144REFERENCESAdelsberg, J.V. and Al-Awqati, Q. (1986). Regulation of cell pHby Ce +-mediated exocytotic insertion of H +-ATPases. J. Cell Biol. 102:1638-1645.Ait-Mohamed, A.K., Marsy, S., Barlet, C., Khadouri, C. andDoucet, A. (1986). Characterization of N-ethylmaleimide-sensitive proton pump in the rat kidney. J. Biol. Chem. 261(27):12526-12533.Al-Awqati, Q. (1978). W . transport in urinary epithelia. Am. J. Physiol. 235: F77-F88.A1-Awqati, Q. and Dixon, T. E. (1982). Effect of electrochemicalgradients on active H transport in an epithelium. Current topics in membrane and transport. 16:163-174.Al-Awqati, Q., Norby, L. H., Mueller, A. and Steinmetz, P.R.(1976). Characteristics of stimulation of H + transport byaldosterone in turtle urinary bladder. J. Clin. Invest. 58:351-358.Arruda J.A.L., Dytko, G. and Talor, Z. (1990). Stimulation of Hsecretion by CO 2 in turtle bladder: role of intracellularpH, exocytosis, and calcium. Am. J. Physiol. 258:R222-R231.Arruda J.A.L., Sabatini S. and Westenfelder C. (1981) Vanadateinhibits urinary acidification by the turtle bladder. KidneyInt.  20:772-779.Avella,M. and Bornancin,M. (1989). A new analysis of ammonia andsodium transport through the gills of the freshwater rainbowtrout (Salmo gairdneri). J. exp. Biol. 142:155-175.Avella,M., Masoni,A., Bornancin,M. and Mayer-Gostan,N. (1987).Gill morphology and sodium influx in the rainbow trout(Salmo gairdneri) acclimated to artificial freshwaterenvironments. J. exp. Zool. 241: 159-169.Bartels, H. (1989). Freeze-fracture study of the pavement cell inthe lamprey gill epithelium. Analogy of membrane structurewith the granular cell in the amphibian urinary bladder.Biol. Cell. 66:165-171.Beauwens, B., Crabbe, J. and Rentmeesters, M. (1981). Effects ofvanadate on the functional properties of the isolated toadbladder. J. Physiol. 310:293-305.Benos, D.J. (1982). Amiloride: a molecular probe of sodiumtransport in tissues and cells. Am. J. Physiol. 242: C131-145C145.Bornancin, M., De Renzis, G. and Naon, R. (1980) C1 - -HCO3-ATPasein gills of the rainbow trout: evidence for its microsomallocalization. Am. J. Physiol. 238:R251-R259.Boutilier, R.G., Iwama, G.K., Heming, T.A. and Randall, D.J.(1985). The apparent pK of carbonic acid in rainbow troutblood plasma between 5 and 15 °C. Respir. Physiol. 61:237-254.Bowman, E.J. (1983). Comparison of the vacuolar membrane ATPaseof Neurospora crassa with the mitochondrial and plasmamembrane ATPases. J. Biol. Chem. 258(24):15238-15244.Bowman, E.J., Siebers, A. and Altendorf, K. (1988). Bafilomycins:A class of inhibitors of membrane ATPases frommicroorganisms, animal cells, and plant cells. Proc. Nalt. Acad. Sci. USA 85:7972-7976.Bradford, M.M. (1976). A rapid and sensitive method for thequantization of microgram quantities of protein utilizingthe principle of protein-dye binding. Analytical Biochem. 72:248-254.Brown, D. (1989). Membrane recycling and epithelial cellfunction. Am. J. Physiol. 256:Fl-F12.Brown, D., Gluck, S. and Hartwig, J. (1987). Structure of thenovel membrane-coating material in proton-secretingepithelial cells and identification as an H +-ATPase. J. Cell Biol. 105:1637-1648.Brown, D., Grosso, A. and De Sousa, R.C. (1981). The amphibianepidermis: distribution of mitochondria-rich cells and theeffect of oxytocin. J. Cell Sci. 52:197-213.Brown, D., Hirsch, S. and Gluck, S. (1988). Localization of aproton-pumping ATPase in rat kidney. J. Clin. Invest. 82:2114-2126.Butler, D.C. (1973). Structure and function of the adrenal glandof fishes. Am. Zool.  13:839-880.Cameron, J.N. and Heisler, N. (1983). Studies of ammonia in therainbow trout: physicochemical parameters, acid-basebehaviour, and respiratory clearance. J. exp. Biol. 105:107-125.Cameron, J.N. and Kormanik, G.A. (1982). The acid-base responsesof gills and kidneys to infused acid and base loads in thechannel catfish, Ictalurus punctatus. J. exp. Biol. 99:143-146160.Cannon C., Adelsberg J.V., Kelly S. and A1-Awqati Q. (1985)Carbon-dioxide-induced exocytotic insertion of H + pumps inturtle-bladder luminal membrane: role of cell pH andcalcium. Nature 314(4):443-446.Cohen, L.H. and Steinmetz, P.R. (1980). Control of active protontransport in turtle urinary bladder by cell pH. J. Gen. Physiol. 76:381-393.De Sousa, R.C. and Grosso, A. (1979). Vanadate blocks cyclic AMP-induced stimulation of sodium and water transport inamphibian epithelia. Nature. 279:803-804.De Sousa, R.C. and Grosso, A. (1981). The mode of action ofvasopressin: membrane microstructure and biologicaltransport. J. Physiol. (Paris). 77:643-669.Dixon, T.E., Clausen, C. and Coachman, D. (1988). Constitutiveand transport-related endocytotic pathways in turtle bladderepithelium. J. Membr. Biol. 102, 49-58.Ehrenfeld, J. and Garcia-Romeu, F. (1977). Active hydrogenexcretion and sodium absorption through isolated frog skin.Am. J. Physiol. 233: F46-F54.Ehrenfeld,J., Garcia-Romeu,F. and Harvey,B.J. (1985).Electrogenic active proton pump in Rana esculenta skin andits role in sodium ion transport. J. Physiol., Lond. 359:331-355.Ehrenfeld, J., Lacoste, I. and Harvey, B.J. (1989). The key roleof the mitochondria-rich cell in Na and H + transport acrossthe frog skin epithelium. Pflugers Arch. 414:59-67.Ehrenfeld, J., Lacoste, I. Garcia-Romeu, F. and Harvey, B.(1990). Interdependence of Na + and H + transport in frogskin. In: Animal nutrition and transport processes. 2. Transport, respiration and excretion: comparative andenvironmental aspects.  ed. Truchot, J.P. and Lahlou, B.Comp. Physiol. 6:152-170.Evans, D.H., More, K.J. and Robbins, S.L. (1989). Modes ofammonia transport across the gill epithelium of the marineteleost fish Opsanus beta. J. exp. Biol. 144:339-356.Forgac, M. (1989). Structure and function of vacuolar class ofATP-driven proton pumps. Physiol. Reviews. 69(3):765-796.Foskett, J.K., Bern, H.A., Machen, T.E. and Conner, M. (1983).Chloride cells and the hormonal control of teleost fish147osmoregulation. J. exp. Biol.  106:255-281.Garg, L.C. and Narang, N. (1988). Effects of aldosterone on NEM-sensitive ATPase in rabbit nephron segments. Kidney Int. 34:13-17.Gluck, S. and A1-Awqati, Q. (1984). An electrogenic proton-translocating adenosine triphosphatase from bovine kidneymedulla. J. Clin. Invest. 73:1704-1710.Gluck, S. and Caldwell, J. (1987). Immunoaffinity purificationand characterization of vacuolar H +-ATPase from bovinekidney. J. Biol. Chem. 262(32):15780-15789.Goffeau, A. and Slayman, C.W. (1981). The proton-translocationATPase of the fungal plasma membrane. Biochininica et Biophysica Acta. 639:197-223.Goss, G.G., Laurent, P. and Perry, S.F. (1992). Evidence for amorphological component in acid-base regulation duringenvironmental hypercapnia in the brown bullhead (Ictalurusnebulosus). Cell Tissue Res. 268:539-552.Graber, M., Dixon, T., Coachman, D. and Devine, P. (1989).Acetazolamide inhibits acidification by the turtle bladderindependent of cell pH. Am. J. Physiol. 256:F923-F931.Harvey, B.J. (1992). Energization of sodium absorption by the H +-ATPase pump in mitochondria-rich cells of frog skin. J. exp. Biol. 172:289-309.Harvey, B.J. and Ehrenfeld, J. (1986). Regulation ofintracellular sodium and pH by the electrogenic H + pump infrog skin. Pflugers Arch. 406:362-366.Heisler, N. (1990). Mechanisms of ammonia elimination in fishes.In: Animal Nutrition and transport processes. 2. Transport, respiration and excretion: comparative and environmental aspects. ed. Truhot, J.P. and Lahlou, B. Basel, Karger.Comp. Physiol. 6:137-151.Hillaby, B.A. and Randall, D.J. (1979). Acute ammonia toxicityand ammonia excretion in rainbow trout (Salmo gairdneri). J.Fish. Res. Bd. Can. 36:621-629.Karnaky, K.J.J R ., Kinter, L.B., Kinter, W.B. and Stirling, C.E.(1976). Teleost chloride cell. II. Autoradiographiclocalization of gill Na,K-ATPase in killifish Fundulus heteroclitus  adapted to low and high salinity environments.J. Cell Biol. 70:157-177.Khadouri, C., Marsy, S., Barlet-Bas, C. and Doucet, A. (1989).148Short-term effect of aldosterone on NEM-sensitive ATPase inrat collecting tubule. Am. J. Physiol. 256:F177-F181.Kleyman, T.R. and Cragoe, E.J.Jr. (1988). Amiloride and itsanalogs as tools in the study of ion transport. J. Membr. Biol. 105:1-21.Knauf, H., Simon, B. and Wais, U. (1976). Non-specific inhibitionof membrane-ATPase by amiloride. Naunyn-Schmiedeberg's Arch. Pharmac. 292:189-192.Laurent, P. and Dunel, S. (1980). Morphology of gill epithelia infish. Am. J. Physiol. 238: R147-159.Laurent, P. and Perry, S.F. (1990). Effects of cortisol on gillchloride cell morphology and ionic uptake in the freshwatertrout, Salmo gairdneri. Cell Tissue Res.  259:429-442.Lenfant, C. and Aucutt, C. (1966). Measurement of blood gases bygas chromatography. Respir. Physiol. 1:398-407.Lin, H. and Randall, D.J. (1990). The effect of varying water pHon the acidification of expired water in rainbow trout. J.exp. Biol. 149:149-160.Lin, H. and Randall, D.J. (1991). Evidence for the presence of anelectrogenic proton pump on the trout gill epithelium. J.exp. Biol. 161:119-134.Ludens, J.H. and Fanestil, D.D. (1974). Aldosterone stimulationof acidification of urine by isolated urinary bladder of thecolombian toad. Am. J. Physiol. 226:1321-1326.Madsen, K.M. and Tisher, C.C. (1985). Structure-functionrelationships in H+-secreting epithelia. Federation Proc. 44:2704-2709.Maetz, J. and Garcia-Romeu, F. (1964). The mechanism of sodiumand chloride uptake by the gills of a freshwater fish,Carassius auratus. II. Evidence for NH:/Na + and HCO3 -/C1 -exchanges. J. Gen. Physiol.  47:1209-1227.McDonald, D.G. (1983). The interaction of environmental calciumand low pH on the physiology of the rainbow trout, Salmogairdneri. I. Branchial and renal net ion and Wfluxes. J.exp. Biol. 102:123-140.McDonald, D.G., Hobe, H. and Wood, C.M. (1980). The influence ofcalcium on the physiological responses of the rainbow trout,Salmo gairdneri, to low environmental pH. J. exp. Biol. 88:109-131.149McDonald, D.G., Tang, Y. and Boutilier, R.G. (1989). Acid and iontransfer across the gills of fish: mechanisms andregulation. Can. J. Zool. 67:3046-3054.McDonald, D.G., Walker, R.L. and Wilkes, P.R.H. (1983). Theinteraction of environmental calcium and low pH on thephysiology of the rainbow trout, Salmo gairdneri. II.Branchial ionoregulatory mechanisms. J. exp. Biol. 102:141-155.McDonald, D.G. and Wood, C.M. (1981). Branchial and renal acidand ion fluxes in the rainbow trout, Salmo gairdneri, at lowenvironmental pH. J. exp. Biol.  93:101-118.McWilliams, P.G. and Potts, W.T.W. (1978). The effects of pH andcalcium concentrations on gill potentials in the browntrout, Salmo trutta. J. comp. Physiol. 126:277-286.Mujais, S.K. (1987). Effects of aldosterone on rat collectingtubule N-ethylmaleimide-sensitive adenosine triphosphatase.J. Lab. Clin. Med. 109:34-39.Nielsen, R. (1982). A coupled electrogenic Na +-K+ pump formediating transepithelial sodium transport in frog skin.Current topics in membrane and transport. 16:87-108.Oberleithner, H., Ritter, M., Lang, F. and Guggino, W. (1983).Anthracene-9-carboxylic acid inhibits renal chloridereabsorption. Pfluegers Arch. 398:172-174.Payan, P. (1978). A studies of the Na +/NH 4 + exchange across thegill of the perfused head of the trout (Salmo gairdneri). J.comp. Physiol. 124:181-188.Pedersen, P.L. and Carafoli, E. (1987). Ion motive ATPases. I.Ubiquity, properties, and significance to cell function.Trends in Biochemistry 12:146-150.Perry, S.F., Davie,P.S., Daxboeck, C.; Ellis, A.G. and Smith,D.G. (1984). Perfusion methods for the study of gillphysiology. In: Fish Physiology. Vol. 10B. ed: Hoar, W.S.and Randall, D.J. Academic press, New York. pp.325-389.Perry, S.F., Davie,P.S., Daxboeck, C. and Randall, D.J. (1982). Acomparison of CO 2 excretion in a spontaneously ventilatingblood-perfused trout preparation and saline-perfused gillpreparation: contribution of the branchial epithelium andred blood cell. J. exp. Biol. 101:47-60.Perry, S.F., Haswell, M.S., Randall, D.J. and Farrell, A.P.(1981). Branchial ionic uptake and acid-base regulation inthe rainbow trout, Salmo gairdneri. J. exp. Biol. 92:289-150303.Perry, S.F. and Laurent, P. (1989). Adaptational response ofrainbow trout to lowered external NaC1 concentration:contribution of the branchial chloride cell. J. exp. Biol. 147:147-168.Perry, S.F. and Randall, D.J. (1981). Effects of amiloride andSITS on brachial ion fluxes in rainbow trout, (Salmogairdneri). J. exp. Zool. 215:225-228.Perry, S.F. and Walsh, P. (1989). Metabolism of isolated fishgill cells: contribution of epithelial chloride cells. J.exp. Biol. 144:507-520.Perry, S.F. and Wood, C.M. (1985). Kinetics of branchial calciumuptake in the rainbow trout: effects of acclimation tovarious external calcium levels. J. exp. Biol.  116:411-433.Pisani, M., Caroff, A. and Rambourg, A. (1987). Two types ofchloride cells in the gill epithelium of a freshwater-adapted euryhaline fish: Lebistes reticulatus; theirmodifications during adaptation to saltwater. Am. J. Anat. 179:40-50.Rahim, S., Delaunoy, J.P. and Laurent, P. (1988). Identificationand immunocytochemical localization of two differentcarbonic anhydrase isoenzymes in teleostean fish erythrocyteand gill epithelia. Histochemistry. 89: 451-459.Randall, D.J. (1990). Control and co-ordination of gas exchangein water breathers. In: Advances in comparative and environmental physiology, Vol.6, edited by R.G. Boutilier.Springer-Verlag Berlin Heidelberg, p.253-278.Reid, S.D. and Perry, S.F. (1991). The effects and physiologicalconsequences of raised levels of cortisol on rainbow trout(Oncorhychus mykiss) erythrocyte B-adrenoreceptors. J. exp. Biol. 158:217-240.Rosen, S. and Friedley, N.J. (1973). Carbonic anhydrase activityin Rana pipiens  skins: Biochemical and histochemicalanalysis. Histochemie. 36:1-4.Sabatini, S., Laski, M.E. and Kurtzman, N.A. (1990). NEM-sensitive ATPase activity in rat nephron: effect ofmetabolic acidosis and alkalosis. Am. J. Physiol. 258:F297-F304.Sapirstein, V.S. and Scott, W.N. (1975). Bonding of aldosteroneby mitochondria-rich cells of the toad urinary bladder.Nature 257:241-243.151Sardet, C. (1980). Freeze fracture of the gill epithelium ofeuryhaline teleost fish. Am. J. Physiol. 238: R207-212.Scharschmidt, B.F., Keeffe, E.B., Blankenship, N.M. and Ockner,R.K. (1979). Validation of a recording spectrophotometricmethod for measurement of membrane-associated Mg ++ and Na + -K+-ATPase activity. J. Lab. Clin. Med. 93(5):790-799.Schwartz. G.J. and A1-Awqati. Q. (1985). Carbon dioxide causesexocytosis of vesicles containing H + pumps in isolatedperfused proximal and collecting tubules. J. Clin. Invest. 75:1638-1644.Schwartz, G.J. and Al-Awqati, Q. (1986). Regulation oftransepithelial H + transport by exocytosis and endocytosis.Ann. Rev. Physiol. 48:153-161.Schwartz, G.J., Barasch, J. and Al-Awqati, Q. (1985). Plasticityof functional epithelial polarity. Nature. 318(28):386-317.Steinmetz, P.R. (1985). Epithelial hydrogen ion transport. In TheKidney: Physiology and Pathophysiology (ed. D.W.Seldin andG.Giebiach). pp.1441-1458. New York: Raven.Steinmetz, P.R. (1986). Cellular organization of urinaryacidification. Am. J. Physiol. 251: F173-F187.Steinmetz, P.R. and Andersen, O.S. (1982). Electrogenic protontransport in epithelial membranes. J. Membr. Biol. 65, 155-174.Stetson, D.L. (1989). Turtle urinary bladder: regulation of iontransport by dynamic changes in plasma membrane area. Am. J. Physiol. 257:R973-R981.Stetson, D.L., Beauwens, R., Palmisano, J., Mitchell, P.P. andSteinmetz, P.R. (1985). A double-membrane model for urinarybicarbonate secretion. Am. J. Physiol. 249:F546-F552.Turrini, F., Sabolic, I., Zimolo, Z., Moewes, B. and Burckhardt,G. (1989). Relation of ATPases in rat renal brush-bordermembranes to ATP-Driven H + secretion. J. Membr. Biol. 107:1-12.Verdouw,H., van Echted, C.J.A. and Dekkers, E.M.J. (1978).Ammonia determination based on indophenol formation withsodium salicylate. Water Res. 12:399-402.Voute, C.L., Hanni, S. and Amman, E. (1972). Aldosterone inducedmorphological changes in amphibian epithelia in vivo. J.Steroid Biochem. 3:161-165.152Voute, C.L., Thummel, J. and Brenner, M. (1975). Aldosteroneeffect in the epithelium of the frog skin. A new story aboutan old enzyme. J. Steroid Biochem. 6:1175-1179.Welsh, M.J. (1984). Anthracene-9-carboxylic acid inhibits apicalmembrane chloride conductance in canine tracheal epithelium.J. Membr. Biol. 78:67-71.Wendelaar Bonga, S.E., Flik, G. and Verbost, P.M. (1992). Calciumregulatory processes in fish. In: New trends in basic andapplied research in aquaculture. Ed. Lahlou, B. andVitiello, P. Springer-Verlag, Berlin Herdelberg. (In press).Wright, P.A., Heming, T.A. and Randall, D.J. (1986). DownstreampH changes in water flowing over the gills of rainbow trout.J. exp. Biol. 126:499-512.Wright, P.A. and Wood, C.W. (1985). An analysis of branchialammonia excretion in the freshwater rainbow trout: effectsof environmental pH change and sodium uptake blockage. J.exp. Biol.  114:329-353.Ye, X., Randall, D.J. and He, X. (1991). The effect of acid wateron oxygen consumption, circulating catecholamines and bloodionic and acid-base status in rainbow trout (Salmo qairdneri, Richardson). Fish Physiol. Biochem. 9(1):23-30.Yesaki, T.Y. and Iwama, G.K. (1992). Survival, acid-baseregulation, ion regulation, and ammonia excretion in rainbowtrout in highly alkaline hard water. Physiol. Zool. 65(4):763-787.Zaugg, W.S. (1982). A simplified preparation for adenosinetriphosphatase determination in gill tissue. Can. J. FishAquat. Sci. 39:215-217.153


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items