- Email: [email protected]
[16] Cation-anion cotransport
280
RED BLOOD CELLS
[16]
tion to net flux rates, anion-exchange rates also can be measured by applying the Jacobs-Stewart cycle in ammonium solutions and the osmotitration method at constant cell volume. Acknowledgment I would like to thank Drs. K. J. Netter, H. Passow, W. Schwarz, and E. S. Vesell for their valuable comments on the manuscript and Miss D. Wfrsd6rfer for her excellent secretarial assistance.
[ 16] C a t i o n - A n i o n C o t r a n s p o r t B y M A R K HAAS a n d THOMAS J. M C M A N U S
Introduction Specific pathways for the coupled transport of cations and anions in the same direction across cell membranes have been described in a wide variety of cells, including reabsorptive and secretory epithelia, 1,z excitable c e l l s , TM cultured cell lines, 4-7 Ehrlich ascites tumor c e l l s , 8'9 and red blood cells from mammalian and avian species. J0-13The majority of these pathways involve the cotransport of the alkali cations sodium and/or potassium with chloride, although cotransport systems involving other ions (e.g., sodium plus phosphate j4) have also been described. In general, J R. Greger, E. Schlatter, and F. Lang, Pfluegers Arch. 396, 308 (1983). 2 H. C. Palfrey, P. Silva, and F. H. Epstein, Am. J. Physiol. 246, C242 (1984). 3 j. M. Russell, J. Gen. Physiol. 81, 909 (1983). 3a S. Liu, R. Jacob, D. Piwnica-Worms, and M. Lieberman, Am. J. Physiol. 253, C721 (1987) 4 j. A. McRoberts, S. Erlinger, M. J. Rindler, and M. H. Saier, J. Biol. Chem. 257, 2260 (1982). 5 N. E. Owen and M. L. Prastein, J. Biol. Chem. 260, 1445 (1985). 6 C. D. A. Brown and H. Murer, J. Membr. Biol. 87, 131 (1985). 7 K. Amsler, J. J. Donahue, C. W. Slayman, and E. A. Adelberg, J. Cell. Physiol. 123, 257 (1985). 8 p. Geck, C. Pietrzyk, B. C. Burckhardt, B. Pfeiffer, and E. Heinz, Biochim. Biophys. Acta 600, 432 (1980). 9 E. K. Hoffmann, C. Sjoholm, and L. O. Simonsen, J. Memhr. Biol. 76, 269 (1983). ~0M. Haas, W. F. Schmidt III, and T. J. McManus, J. Gen. Physiol. 80, 125 (1982). " P. W. Flatman, J. Physiol. (London) 341, 545 (1983). t2 p. K. Lauf, J. Membr. Biol. 77, 57 (1983). ~3M. Canessa, C. Brugnara, D. Cusi, and D. C. Tosteson, J. Gen. Physiol. 87, 113 (1986). 14 p. Gmaj and H. Murer, Physiol. Rev. 66, 36 (1986).
METHODS IN ENZYMOLOGY.VOL. 173
Copyright© 1989by AcademicPress, Inc. All rightsof reproductionin any formreserved.
[16]
CATION-ANION COTRANSPORT
281
the ion specificities of (Na + + CI-), (K + + CI-), and (Na + + K + + CI-) cotransport pathways are fairly stringent; rubidium can substitute for potassium, lithium for sodium (though in most cases with lower affinity), and bromide for chloride. ~°,15,~6Other cations and anions are usually not transported by these pathways, though some may serve as inhibitors as discussed below. In this chapter, we describe in some detail two methods for the study of cation-chloride cotransport in intact red blood cells which are directed at establishing the stoichiometric relationship between cotransported ions, the effect of cell volume on this stoichiometry and on the magnitude of cotransport, and the electrogenicity or electroneutrality of the cotransport process. It should be noted that these methods, as described, were developed specifically for use in red blood cells in order to take advantage of (or circumvent) certain properties of red cell membranes including their low conductance to sodium and potassium, their relatively high chloride conductance, and the rapid anion-exchange system which also appears to mediate a substantial fraction of the chloride conductance. The experiments presented to illustrate these methods were performed with duck red cells, which exhibit an (Na ÷ + K + + CI-) cotransport that is stimulated either by cell shrinkage ~7,~8 or/3-adrenergic catecholamines, ~°,~5 and a (K + + C1-) cotransport system that is stimulated by cell swelling. 19-21 The methods described below have also been successfully employed in studies of red cells of other species that have lower cotransport rates than those of duck red cells. In addition, these methods may be adapted for use in the study of cation-chloride cotransport in other types of cells in suspension, such as cultured cells or Ehrlich ascites cells. Before presenting the specific methods, it is important to include a brief discussion of how cation fluxes occurring via cation-chloride cotransport are defined. It is common practice to define cotransport as the ouabain-insensitive influx or efflux of sodium and/or potassium that is inhibited by "loop" diuretics such as furosemide or bumetanide. Although in most cases this is quite acceptable, there are several potential problems associated with using diuretic-sensitive fluxes as the sole criterion for cotransport. First, furosemide, at a concentration (1 mM) needed i5 W. F. Schmidt III and T. J. McManus, J. Gen. Physiol. 70, 81 (1977). 16 M. Canessa, I. Bize, N. Adragna, and D. Tosteson, J. Gen. Physiol. 80, 149 (1982). 17 W. F. Schmidt III and T. J. McManus, J. Gen. Physiol. 70, 59 (1977). 18 F. M. Kregenow, J. Gen. Physiol. 58, 396 (1971). 19 F. M. Kregenow, J. Gen. Physiol. 58, 372 (1971). 20 M. Haas and T. J. McManus, J. Gen. Physiol. 85, 649 (1985). 21 T. J. McManus, M. Haas, L. C. Starke, and C. Y. Lytle, Ann. N.Y. Acad. Sci. 456, 183 (1985).
282
RED BLOODCELLS
[16]
to fully inhibit most (Na + + K ÷ + CI-), (Na ÷ + C1-), and (K + + CI-) cotransport systems also partially inhibits other transport pathways in red ceils 22 including sodium-dependent glycine transport and anion exchange (which in the presence of bicarbonate can mediate monovalent cation transport via cation-CO3 ion pairs23). Second, while bumetanide inhibits (Na ÷ + K ÷ + CI-) cotransport systems with high affinity, this is generally not true with (K ÷ + C1-) systems. Thus, (K + + CI-) cotransport systems have been occasionally termed "bumetanide insensitive" (e.g., Ref. 24) on the basis of the finding that they are not significantly inhibited by 10-5 M bumetanide (a concentration which fully inhibits (Na + + K + + C1-) cotransport under physiological ionic conditions22), when in fact higher concentrations of bumetanide (1 mM) do inhibit (K + + CI-) cotransport (e.g., Refs. 25 and 25a). Third, the inhibitory potency of bumetanide and furosemide toward (Na ÷ + K + + CI-) and (Na + + C1-) cotransport, and in at least one case (K ÷ + C1-) cotransport as well, is dependent on the ionic composition of the extracellular medium. Bumetanide inhibits (Na ÷ + K ÷ + 2C1-) cotransport in duck red cells by apparent competition for a chloride sitefl6 and furosemide likewise competes with chloride in inhibiting (Na ÷ + CI-) cotransport in toad cornea. 27 In contrast, increasing extracellular sodium, [Na]o, and potassium, [K]o, enhances inhibition of (Na + + K + + 2C1-) cotransport by " l o o p " diuretics, 22 and binding of radiolabeled bumetanide to duck red cells 28 and membranes from dog kidney outer medulla29is likewise enhanced by increasing [Na]o and [K]o. In the complete absence of [Na]o and [K]o, 2 × l0 -4 M bumetanide (or 1 mM furosemide) is required for complete inhibition of(Na + + K + + 2C1-) cotransport in duck red cells. In the case of (K + + CI-) cotransport in sheep red cells treated with the sulfhydryl reagent N-ethylamaleimide, furosemide inhibition is also enhanced by increasing [K]o. t2 An alternative to using diuretic-sensitive fluxes as a measure of cation-chloride cotransport is to measure chloride-dependent cation fluxes, i.e., the difference in ouabain-insensitive sodium or potassium flux in the presence of chloride and in the complete absence of chloride. Most, if not 22 H. C. Palfrey, P. W. Feit, and P. Greengard, Am. J. Physiol. 238, C139 (1980). 23 j. Funder and J. O. Wieth, Acta Physiol. Scand. 71, 168 (1967). 24 C. Brugnara, A. S. Kopin, H. F. Bunn, and D. C. Tosteson, J. Clin. Invest. 75, 1608 (1985). 25 L. R. Berkowitz and E. P. Orringer, Am. J. Physiol. 252, C300 (1987). zsa D. Kaji, J. Gen. Physiol. 88, 719 (1986). 26 M. Haas and T. J. McManus, Am. J. Physiol. 245, C235 (1983). 27 j. H. Ludens, J. Pharmacol. Exp. Ther. 22,3, 25 (1982). M. Haas and B. Forbush III, J. Biol. Chem. 261, 8434 (1986). 29 B. Forbush III and H. C. Palfrey, J. Biol. Chem. 258, 11787 (1983).
[16]
CATION--ANION COTRANSPORT
283
all cation-chloride cotransport systems are completely inhibited when chloride is totally replaced by a permeant anion other than bromide. Ideally, chloride-dependent fluxes should be equal to furosemide- or bumetanide-sensitive fluxes when a fully inhibitory concentration of diuretic is used, and in avian red cells we find this to be the case when the anions methyl sulfate or methane sulfonate are used to replace chloride. However, relatively lipophilic anions such as nitrate and thiocyanate increase passive, nonspecific fluxes of sodium and potassium across red cell membranes, 23 and thus ouabain-insensitive cation fluxes in the presence of these anions will be higher than in a chloride medium containing a dose of furosemide or bumetanide that fully inhibits cation-chloride cotransport. It is also advisable to avoid the use of nitrate in studies of cation-chloride cotransport for another reason: in several different cell types it appears that nitrate may compete with chloride for sites on (Na + + K + + C1-) cotransport 5,6,28 and (K + + C1-) 3° cotransport pathways. Stoichiometry of Cotransport: The Zero-Trans Efflux Method Measurement of net sodium and potassium effluxes into media free of these cations has proved to be a useful way of determining stoichiometric relationships between cotransported cations in red blood cells. This method is particularly applicable to cells such as red cells in which the baseline permeability of the cell membrane to monovalent cations is low, and most of the transport of sodium and potassium occurs via specific mediated pathways. In duck red cells in particular, >50% of sodium and potassium effluxes into isotonic choline or tetramethylammonium (TMA) media are inhibited either by 1 mM furosemide or by replacement of intraand extracellular chloride with methyl sulfate; this percentage increases when (Na t + K ÷ + 2C1-) cotransport or (K ÷ + CI-) cotransport is stimulated by shrinking or swelling the cells, respectively (see below). Choline and TMA do not appear to interact with either (Na t + K ÷ + 2CI ) or (K ÷ + CI-) cotransport systems in a variety of cell types, and are thus the cations of choice for use in the incubation medium in zero-trans efflux experiments. We have used TMA in most of our experiments simply because TMA salts of several monovalent anions other than chloride are available commercially or can be easily prepared from TMA hydroxide and the appropriate acid. Efflux determinations can be made by measuring either cell or medium sodium and potassium contents before and after given time periods of incubation in the zero-trans medium. When the effluxes of sodium and potassium are large, as in duck red cells, measure3op. K. Lauf, J. Membr. Biol. 82, 167 (1984).
284
RED BLOOD CELLS
[16]
ment of cell cation contents gives good results. However, it is important to note that when substantial net salt efflux occurs, the cells will also lose water and shrink during the incubation period. Therefore, cell sodium and potassium levels (and values for net effluxes) should be determined and expressed as total contents (millimoles/kilogram cell solid or millimoles/ liter original cells) rather than as concentrations (millimoles/liter cell water). When the effluxes of sodium and potassium are small, as in human red cells, it is preferable to measure the accumulation of these cations in the extracellular medium, either directly or by isotopic tracer methods using cells preloaded with Z2Na, 24Na, 42K, or 86Rb (the latter being a potassium congener). If the hematocrit and cell contents of sodium and potassium (and specific activities of their tracers, if applicable) are known, measurements of sodium and potassium accumulation in the medium will yield values for their effluxes. An example of a zero-trans efflux experiment using duck red cells is shown in Fig. 1. In this experiment, the cells were loaded using the nystatin method 3~ (see Fig. I legend) to contain a constant level of potassium and three different sodium contents. They were then incubated 10 min in sodium- and potassium-free TMA chloride media of different osmolalities, with or without 1 mM furosemide (the range of osmolalities employed was different for the three groups of cells in order to bring cells of three different sodium and total salt contents into the same range of cell volumes). In Fig. 1, furosemide-sensitive net sodium efflux (-ANal) and potassium efflux (-AKc) are plotted as a function of the average cell water content over the 10-min incubation period; this average was used because the cells were progressively shrinking as they lost sodium, potassium, and chloride into the TMA chloride media. The isotonic water content of duck red cells is 1.50 kg H20/kg cell solid, or 60% (w/w). In swollen cells with Wc > 1.5, the cells lose potassium without a concomitant sodium loss, indicative of the (K + + C1-) cotransport seen in swollen duck red cells. 2°,2l This swelling-induced (K + + CI-) cotransport is independent of [Na]c, at least over the range of values tested. In contrast, cell shrinkage (We < 1.5) stimulates net sodium and potassium efflux with an N a + : K ÷ stoichiometry of 1:1, indicative of the (Na + + K + + 2C1-) cotransport activated by shrinkage of duck red cells. 1°:7 Increasing cell sodium from 41.1 to 73.6 mmol/kg cell solid increases the level of outward (Na + + K ÷ + 2C1-) cotransport; the effect of increasing cell sodium on this process appears to saturate above the latter level. The zero-trans experiment in Fig. 1 nicely displays the two volumesensitive cation-chloride cotransport systems in duck red cells, and dem31 A. Cass and M. Dalmark, Nature (London), New Biol. 244, 47 (1973).
[16]
CATION--ANION COTRANSPORT
~O
oA
~ ~
285
4o Z
A
w> o30
•
•
o
o
[]
J
-J zw L,I '0 O3 ~
20
0
D
\\\
I W
0
~ \\L.ff
~
• •
o o
~,.,._o.~M~ 73.6+0.,
,
U_
, -//I.25
-
1.50
1.75 Wc
(Kg H 2 0 / K g
2.00
2.25
2.50
CELL SOLID)
FIG. 1. Effects of cell water and sodium content on furosemide-sensitive net sodium and potassium effluxes from duck red cells into sodium- and potassium-free media. Cells were prepared to contain 202 mmol potassium/kg cell solid and the three listed sodium contents by the nystatin method as follows: Cells were incubated for 1 hr at 0°, 2% hematocrit in media containing 20/~g/ml nystatin, 50 mM sucrose, 100 mM KC1, and either 20, 35, or 50 mM NaCI. This procedure was then repeated by centrifuging and resuspending the cells in the same media without nystatin. Nystatin was eluted from the cells, restoring the membrane to its normal state of low cation permeability, by washing the cells six times at 25°, 2% hematocrit in the same solution as above (without nystatin), but containing 0.1 mM ouabain and 0.25% bovine serum albumin (fraction V). The albumin-containing solutions were titrated to pH 7.4 at 25° with KOH. Test incubations were performed for 10 min at 3% hematocrit, 41 ° (body temperature for ducks) in media containing varying concentrations of TMA chloride, with and without 1 mM furosemide. The ranges of osmolalities over which the cells were incubated were as follows: 140-350 mOsm/kg (O), 170-380 (11), and 200-410 (A), all in increments of 30 mOsm/kg. Open symbols represent furosemide-sensitive net potassium efflux (-AKc), closed symbols represent furosemide-sensitive net sodium efflux
(-ANal).
onstrates the l : l N a + : K ÷ stoichiometry of the (Na + + K + + 2 0 - ) cotransport pathway. Whether experiments such as that in Fig. 1 can be routinely used to s h o w c a t i o n : c h l o r i d e stoichiometries of cotransport systems in red cells remains an open question; this is because a small but significant fraction o f anion exchange remains even in cells treated with a maximal inhibitory concentration o f D I D S , a stilbene derivative that is a potent inhibitor o f the red cell anion exchanger. B e c a u s e chloride fluxes
286
RED BLOOD CELLS
[16]
via the anion exchanger are so rapid, even when this exchanger is 99.99% inhibited, the residual chloride fluxes (at 37°) would be expected to be of the same order of magnitude as those occurring via (Na ÷ + K ÷ + 2C1-) cotransport in maximally shrunken duck red cells. Lytle32 has circumvented the anion-exchange problem by performing measurements of net Na ÷, K ÷, and CI- effluxes from duck red blood cells incubated in TMAgluconate medium. The cells were first treated with DIDS. Under these conditions, bumetanide (1 mM)-sensitive efflux stoichiometries of approximately 1 N a ÷ : l K + : 2 C1- and 1 K ÷ : I CI- were obtained in shrunken and swollen cells, respectively. Substitution of impermeant, nonexchangeable anions such as gluconate for extracellular chloride does markedly alter the membrane potential of red blood cells, and is thus best applied in studies of transport processes known to be electroneutral (see below). Treatment of the cells with DIDS is also warranted to minimize changes in intracellular pH resulting from CI-/OH- exchange or (H ÷ + CI-) cotransport via the anion-exchange pathway. 3aa Hall and Ellory33 have reported experiments in which the stoichiometry of bumetanide-sensitive tracer sodium, potassium, and chloride influxes in ferret red cells (which exhibit bumetanide-sensitive fluxes similar in magnitude to those in duck red cells H) was found to be 2 Na ÷ : 1 K ÷ : 3 CI-, similar to the cotransport stoichiometry determined in squid axon. 3 In these experiments of Hall and Ellory33 anion exchange was minimized not only by treating the cells with two inhibitors of this exchange (DIDS plus SITS), but also by lowering the temperature to 21 ° . The rationale for lowering the temperature in these experiments is that the red cell anionexchange pathway has a greater sensitivity to temperature (higher Arrhenius activation energy) than does (Na ÷ + K ÷ + C1-) cotransport. 33 A major advantage of the zero-trans efflux method with regard to determination of stoichiometries of cotransported cations is that potential errors due to K÷/K + and Na÷/Na ÷ exchanges are avoided. Both d u c k 10'20'32'34 and h u m a n j3,34a red cells exhibit bumetanide-sensitive, chloride-dependent, 1 : 1 K+/K ÷ and Na÷/Na + exchanges, and kinetic analysis of these exchanges has provided strong evidence for their being partial reactions of (Na + + K + + 2C1-) cotransport. 3~'34,34a When bumetanidesensitive tracer influxes of sodium and potassium are measured in either 32 C. Lytle, Ph.D. thesis, Duke University (1988). 32a M. L. Jennings, J. Membr. Biol. 40, 365 (1978). 33 A. C. Hall and J. C. Ellory, J. Membr. Biol. 85, 205 (1985). 34 C. Lytle, M. Haas, and T. J. McManus, Fed. Proc., Fed. Am. Soc. Exp. Biol. 45, 548 (1986). 34a j. Duhm, J. Membr. Biol. 98, 15 (1987).
[16]
CATION-ANION COTRANSPORT
287
duck or human red cells, the potassium influx exceeds that of sodium because the high intracellular potassium concentration of these cells favors K+/K ÷ exchange relative to Na÷/Na ÷ exchange. In duck red cells prepared by the nystatin method to contain 150 mM sodium and 10 mM potassium, bumetanide-sensitive sodium influx now exceeds potassium i n f l u x . 32,34 Indeed, it is quite possible that the stoichiometry of 2 Na ÷ : 1 K ÷ : 3 CI- reported for bumetanide-sensitive influxes in ferret red cells 33 reflects a component of Na÷/Na + exchange, since ferret red cells have very low ouabain-sensitive cation fluxes and thus contain sodium as their major intracellular cation. ~ In zero-trans efflux experiments, K+/K + and Na÷/Na + exchanges cannot occur, and the true stoichiometry of (Na + + K ÷ + C1-) cotransport processes is determined. Electrogenicity vs Electroneutrality of Transport: The Valinomycin Method A decade ago it was shown that partial replacement of extracellular chloride, [Cl]o, by an impermeant anion such as gluconate markedly inhibited furosemide-sensitive net sodium and potassium influx into duck red cells which were shrunken or treated with catecholamines. 35 Because in red cells the membrane potential (Era) is equal to the Nernst potential for chloride (Ecl), replacement of [Cl]o by gluconate at constant intracellular chloride, [C1]c, causes Em to become more positive. It was thus initially thought that the decrease in (Na ÷ + K ÷) uptake when [Cl]o was partially replaced by gluconate was the result of this depolarization. 35 In this model, (Na ÷ + K ÷) was thought to cross the membrane in an electrogenic manner, with chloride (or presumably, any other permeant anion) following passively to maintain electroneutrality. However, the finding that the coupled, furosemide-sensitive movements of sodium and potassium in duck red cells have a specific anion requirement for chloride (or bromide) and are completely inhibited when chloride is fully replaced by other permeant anions such as nitrate 1°,36 led to the consideration that chloride might in fact move across the membrane in an electrically neutral complex with the cations, i.e., (Na ÷ + K ÷ + 2C1-). In this latter case, the inhibitory effect of partial replacement of [C1]o at constant [Cl]c can be attributed to a change in the direction of the chloride chemical potential gradient, independent of Em. However, as long as the ratio of [Cl]c/[Cl]o affects both Em and the chloride chemical gradient, as is the case in red 3s W. F. Schmidt III and T. J. McManus, J. Gen. Physiol. 70, 99 (1977). 36 F. M. Kregenow and T. Caryk, Physiologist 22, 73 (1979).
288
RED BLOOD CELLS
[16]
cells, it is not possible to distinguish between these two possible mechanisms of transport [electrogenic (Na + + K +) cotransport versus electroneutral (Na + + K + + 2C1-) cotransport]. This problem was resolved by employing the potassium ionophore valinomycin to increase the potassium conductance (PrO of the membrane sufficiently so that PK exceeded Pcl. Upon further addition of DIDS, which reduces Pc~ in red cells by -50% (in addition to inhibiting electroneutral anion exchange37), PK sufficiently exceeds P a so that Em can be held constant near the potassium equilibrium potential (EK) while [Clio is varied at constant [Cl]c.~° It is actually possible to estimate PK/Pcl in valinomycin-treated cells by measuring 1- or 2-min potassium effluxes from these cells into NaCI and RbCI media. If sufficient valinomycin is present (10 -7 M or more in human red cells38), the potassium efflux into NaC1 gives an estimate of Pcl, since the rate of this efflux is limited only by the conductance of the major permeable anion available to follow potassium out of the cell to maintain electroneutrality. 38 If, however, the external medium contains only RbCI, valinomycin promotes a K+/Rb + exchange which is limited by PK, since valinomycin preferentially selects rubidium over potassium. 39 Thus, potassium efflux into RbC1 gives an estimate of PK, and PK/Pc~in valinomycin-treated cells can be estimated by the ratio of K ÷ efflux into RbCI divided by K + efflux into NaCI (see Table I of Ref. 10 for further details). Knowing PK/Pa and assuming Pya is small relative to PK and Pcl (a safe assumption in valinomycin-treated red cells), Em can be estimated from the constant-field (GoldmanHodgkin-Katz) equation. With Em held constant near EK by the use of valinomycin and DIDS, it is possible to vary [Cl]o (at constant [Cl]c) independently of Era. Figure 2 shows an experiment in which net changes in cell sodium were measured in DIDS-treated duck red cells as a function of [C1]o, at constant [Cl]c. Methyl sulfate, which has a conductance about one-half that of chloride in DIDS-treated cells, J0 was used to replace chloride; with DIDS present the potential exchange of internal chloride for external methyl sulfate was inhibited and [Cl]c remained essentially constant over the 5-min incubation period. The cells were incubated with norepinephrine (1 /zM) to stimulate (Na ÷ + K ÷ + 2C1-) cotransport, in the presence and absence of 2/zM valinomycin and 1 mM furosemide. In the presence of valinomycin, Em was held constant near EK (about - 8 mV in this experiment), whereas Ern became progressively more negative in the absence of valinomycin as 37 p. A. Knauf, G. F. Fuhrmann, S. Rothstein, and A. Rothstein, J. Gen. Physiol. 69, 363 (1977). 38 M. J. Hunter, J. Physiol. (London) 268, 35 (1977). 39 T. E. Andreoli, M. Tieffenberg, and D. C. Tosteson, J. Gen. Physiol. 50, 2527 (1967).
[16]
CATION--ANION COTRANSPORT SYMBOL
VALINOMYCIN ( 2 / . t M )
on
--
o
~
+
• •
289
0
tf)
-~ _J
o
<3 ~
-4
o Z t,O
- /
"~---+NO~PINEPHRINE (Na)0 = 5 0 m M (K)0 • l O O m M
-6
r-(Na)c = 4 8 m M initial~ ( K ) c 137 rnM
E
L_(CI)c
-12 I 25
I 50
I 75 (CI)o mM
I I00
50 m M
I 125
I 150
FJc. 2. Effect of the chloride gradient on norepinephrine-stimulated, furosemide-sensirive net sodium m o v e m e n t s in duck red cells at constant and variable Era. Cells were preincubated 8 hr at 4 ° in a medium containing 80 m M sodium methyl sulfate (MeSO4), 60 m M NaCI, and 20 m M KCI to increase cell sodium and also replace - 5 0 % of cell chloride with MeSOn. Next, an additional preincubation was done tbr 90 rain at 41 ° (body temperature for ducks) in media of the same ionic composition but also containing 0. I m M ouabain (to further increase [Na],) and 10 5 M DIDS. Test incubations were done at 41 °, 3% hematocrit in media containing 50 m M sodium, 100 m M potassium, 10 ~ M DIDS, 10 " M norepinephrine, and 0.1 m M ouabain, with and without 2 × 10 6 M valinomycin and 1 m M furosemide. Methyl sulfate substituted for external chloride, maintaining isotonicity, in the presence of valinomycin, PK/Pct was estimated to be 7.5 (see text), and Em (as estimated from the constant-field equation) was constant at - 7 . 9 -+ 0.5 mV (SEM, n = 8) over the entire range of [CI],, from 2.5 to 150 m M . Em in the absence of valinomycin varied from - 0 . 4 mV at 2.5 m M [CI],, to - 1 8 . 3 mV at 150 m M [CI],, as estimated from the constant-field equation with PM~so4= 0.5 P o (as determined in DIDS-treated duck red cells ~") and a s s u m i n g Pcl >>PK ,PN,. (Reproduced from The Journal of General Physiology, 1982, volume 80, pp. 125-147, by copyright permission of the Rockefeller University Press.)
[Cl]o was increased (see legend for Fig. 2). Figure 2 shows that at low [Cl]o there was a furosemide-sensitive net loss of sodium, whereas at high [Cl]o, when the chloride gradient was inwardly directed, the cells gained sodium. This effect of [C1]o on furosemide-sensitive net sodium movements was identical in the presence and absence of valinomycin, showing that the parameter affecting this sodium transport is the chemical potential gradient of chloride and not Era. The experiment shown in Fig. 2 strongly supports the idea that catecholamine-stimulated, furosemide-sensitive sodium transport in duck red
290
RED BLOOD CELLS
[16]
cells occurs via an electrically neutral cotransport of (Na + + K ÷ + 2C1-) rather than an electrogenic (Na + + K ÷) cotransport. Further evidence that this is the case has come from experiments similar to that in Fig. 2, only in which a significant sodium gradient was present and furosemidesensitive net transport of sodium against this gradient was driven by a chloride gradient at constant Em (see Fig. 7 of Ref. 10). Finally, two methods can and have been used to confirm that the cotransport pathway is indeed electrically neutral. First, in cells treated with DIDS, the effect of Em on furosemide-sensitive net sodium transport can be studied by varying [K]o at constant [K]c in the presence of valinomycin (varying Em ; Em = EK) and in its absence (constant Em ; Em =Ecl). When this experiment was done in duck red cells, the same effect of varying [K]o was seen whether or not valinomycin was present, confirming a lack of effect of Em on the cotransport pathway and strongly suggesting its electroneutrality. 10 In addition, changes in Em upon activation or inhibition of transport pathways can be detected using potential-sensitive, fluorescent carbocyanine dyes such as diS-C3[5] (for a detailed study of the use of such dyes in red cells, see Ref. 40). Briefly, these dyes are taken up by red cells upon hyperpolarization of the membrane, and released upon depolarization. Because the fluorescence of the dye is quenched when it is within (or associated with) the cells, changes in dye fluorescence can be used to monitor changes in Em .40 This dye method was used to demonstrate the electrogenicity of the Na÷/K ÷ pump in red cells, as a depolarization of the membrane (increase in diS-C3[5] fluorescence) was seen upon addition of ouabain. 41 This result is not surprising, considering that the Na÷/K ÷ pump transports three sodium ions out of the cell in exchange for two potassium ions. 42 However, in duck red cells incubated in a TMA chloride medium containing I mM KC1 and 1 mM NaCI [conditions designed to maximize outward net (Na + + K ÷ + 2C1-) cotransport], we found no effect of either adding norepinephrine (to stimulate cotransport) or subsequently increasing [K]o on diS-C3[5] fluorescence, confirming the electroneutrality of (Na + + K + + 2C1-) cotransport in these cells 2j (also see Ref. 43). Likewise, the (K ÷ + CI-) cotransport stimulated by cell swelling was found to be electrically neutral using the dye method in duck red cells, 2j as well as 4o p. j. Sims, A. S. Waggoner, C. H. Wang, and J. F. Hoffman, Biochemistry 13, 3315 (1974). 4~ j. F. Hoffman, J. H. Kaplan, and T. J. Callahan, Fed. Proc., Fed. Am. Soc. Exp. Biol. 38, 2440 (1979). 42 R. L. Post and P. C. Jolly, Biochim. Biophys. Acta 25, 108 (1957). 43 F. M. Kregenow, in "Membrane Transport in Red Cells" (J. C. Ellory and V. L. Lew, eds.), p. 383. Academic Press, New York, 1977.
[16]
CATION--ANION COTRANSPORT
291
in human red cells, 44 where Em was determined from the ratio of [H+]c/ [H+]o in cells treated with the proton-specific ionophore CCCP and DIDS. 45 Summary Two methods have been described for the study of cation-chloride cotransport systems. The zero-trans efflux method is designed to determine stoichiometric relationships between cotransported ions under conditions where ion exchanges cannot occur. These exchanges (e.g., Na+/ Na +, K+/K ÷) may occur as partial or incomplete reactions of a cotransport process and can lead to erroneous determinations of the stoichiometry of the cotransport process. The zero-trans efflux method can also be used to study the effects of cell volume, pH, and intracellular ion concentrations on cotransport processes. The valinomycin method is used to determine the electrogenicity or electroneutrality of transport, and in this regard can be used in conjunction with other methods such as those employing potential-sensitive dyes or microelectrodes. Other, more recently developed ionophores with specificity for lithium rather than potassium 46 have now been used to study the effect of Em on the ATP-dependent Na+/K + p u m p . 47 It may be possible to use such ionophores to confirm the suspected electroneutrality of (K ÷ + CI-) cotransport systems, as well as for other studies of specific potassium transport processes in which valinomycin obviously cannot be used. Both methods discussed in detail in this chapter, and particularly the valinomycin method, were originally devised for use in red blood cells in order to take advantage of (or circumvent) properties of the red cell membrane, such as its low permeability to sodium and potassium and relatively high permeability to chloride. However, valinomycin has been used successfully to demonstrate the electroneutrality of (Na + + K ÷ + 2C1-) cotransport in MDCK cells, 4 and the zero-trans efflux method should be applicable to the study of transport processes in other types of cells in suspension, so long as the transport system being studied can be accurately defined (e.g., as an inhibitor-sensitive or chloride-dependent cation flux) and comprises a significant fraction of the total salt efflux. 44 C. Brugnara, T. Van Ha, and D. C. Tosteson, J. Gen. Physiol. 92, 42a (1988). 45 R. I. Macey, J. S. Adorante, and F. W. Orme, Biochim. Biophys. Acta 512, 284 (1978). 46 R. Margalit and A. Shanzer, Biochim. Biophys. Acta 649, 441 (1981). 47 R. Goldshlegger, S. J. D. Karlish, A. Rephaeli and W. D. Stein, J. Physiol. (London) 387, 331 (1987).
RED BLOOD CELLS
[16]
tion to net flux rates, anion-exchange rates also can be measured by applying the Jacobs-Stewart cycle in ammonium solutions and the osmotitration method at constant cell volume. Acknowledgment I would like to thank Drs. K. J. Netter, H. Passow, W. Schwarz, and E. S. Vesell for their valuable comments on the manuscript and Miss D. Wfrsd6rfer for her excellent secretarial assistance.
[ 16] C a t i o n - A n i o n C o t r a n s p o r t B y M A R K HAAS a n d THOMAS J. M C M A N U S
Introduction Specific pathways for the coupled transport of cations and anions in the same direction across cell membranes have been described in a wide variety of cells, including reabsorptive and secretory epithelia, 1,z excitable c e l l s , TM cultured cell lines, 4-7 Ehrlich ascites tumor c e l l s , 8'9 and red blood cells from mammalian and avian species. J0-13The majority of these pathways involve the cotransport of the alkali cations sodium and/or potassium with chloride, although cotransport systems involving other ions (e.g., sodium plus phosphate j4) have also been described. In general, J R. Greger, E. Schlatter, and F. Lang, Pfluegers Arch. 396, 308 (1983). 2 H. C. Palfrey, P. Silva, and F. H. Epstein, Am. J. Physiol. 246, C242 (1984). 3 j. M. Russell, J. Gen. Physiol. 81, 909 (1983). 3a S. Liu, R. Jacob, D. Piwnica-Worms, and M. Lieberman, Am. J. Physiol. 253, C721 (1987) 4 j. A. McRoberts, S. Erlinger, M. J. Rindler, and M. H. Saier, J. Biol. Chem. 257, 2260 (1982). 5 N. E. Owen and M. L. Prastein, J. Biol. Chem. 260, 1445 (1985). 6 C. D. A. Brown and H. Murer, J. Membr. Biol. 87, 131 (1985). 7 K. Amsler, J. J. Donahue, C. W. Slayman, and E. A. Adelberg, J. Cell. Physiol. 123, 257 (1985). 8 p. Geck, C. Pietrzyk, B. C. Burckhardt, B. Pfeiffer, and E. Heinz, Biochim. Biophys. Acta 600, 432 (1980). 9 E. K. Hoffmann, C. Sjoholm, and L. O. Simonsen, J. Memhr. Biol. 76, 269 (1983). ~0M. Haas, W. F. Schmidt III, and T. J. McManus, J. Gen. Physiol. 80, 125 (1982). " P. W. Flatman, J. Physiol. (London) 341, 545 (1983). t2 p. K. Lauf, J. Membr. Biol. 77, 57 (1983). ~3M. Canessa, C. Brugnara, D. Cusi, and D. C. Tosteson, J. Gen. Physiol. 87, 113 (1986). 14 p. Gmaj and H. Murer, Physiol. Rev. 66, 36 (1986).
METHODS IN ENZYMOLOGY.VOL. 173
Copyright© 1989by AcademicPress, Inc. All rightsof reproductionin any formreserved.
[16]
CATION-ANION COTRANSPORT
281
the ion specificities of (Na + + CI-), (K + + CI-), and (Na + + K + + CI-) cotransport pathways are fairly stringent; rubidium can substitute for potassium, lithium for sodium (though in most cases with lower affinity), and bromide for chloride. ~°,15,~6Other cations and anions are usually not transported by these pathways, though some may serve as inhibitors as discussed below. In this chapter, we describe in some detail two methods for the study of cation-chloride cotransport in intact red blood cells which are directed at establishing the stoichiometric relationship between cotransported ions, the effect of cell volume on this stoichiometry and on the magnitude of cotransport, and the electrogenicity or electroneutrality of the cotransport process. It should be noted that these methods, as described, were developed specifically for use in red blood cells in order to take advantage of (or circumvent) certain properties of red cell membranes including their low conductance to sodium and potassium, their relatively high chloride conductance, and the rapid anion-exchange system which also appears to mediate a substantial fraction of the chloride conductance. The experiments presented to illustrate these methods were performed with duck red cells, which exhibit an (Na ÷ + K + + CI-) cotransport that is stimulated either by cell shrinkage ~7,~8 or/3-adrenergic catecholamines, ~°,~5 and a (K + + C1-) cotransport system that is stimulated by cell swelling. 19-21 The methods described below have also been successfully employed in studies of red cells of other species that have lower cotransport rates than those of duck red cells. In addition, these methods may be adapted for use in the study of cation-chloride cotransport in other types of cells in suspension, such as cultured cells or Ehrlich ascites cells. Before presenting the specific methods, it is important to include a brief discussion of how cation fluxes occurring via cation-chloride cotransport are defined. It is common practice to define cotransport as the ouabain-insensitive influx or efflux of sodium and/or potassium that is inhibited by "loop" diuretics such as furosemide or bumetanide. Although in most cases this is quite acceptable, there are several potential problems associated with using diuretic-sensitive fluxes as the sole criterion for cotransport. First, furosemide, at a concentration (1 mM) needed i5 W. F. Schmidt III and T. J. McManus, J. Gen. Physiol. 70, 81 (1977). 16 M. Canessa, I. Bize, N. Adragna, and D. Tosteson, J. Gen. Physiol. 80, 149 (1982). 17 W. F. Schmidt III and T. J. McManus, J. Gen. Physiol. 70, 59 (1977). 18 F. M. Kregenow, J. Gen. Physiol. 58, 396 (1971). 19 F. M. Kregenow, J. Gen. Physiol. 58, 372 (1971). 20 M. Haas and T. J. McManus, J. Gen. Physiol. 85, 649 (1985). 21 T. J. McManus, M. Haas, L. C. Starke, and C. Y. Lytle, Ann. N.Y. Acad. Sci. 456, 183 (1985).
282
RED BLOODCELLS
[16]
to fully inhibit most (Na + + K ÷ + CI-), (Na ÷ + C1-), and (K + + CI-) cotransport systems also partially inhibits other transport pathways in red ceils 22 including sodium-dependent glycine transport and anion exchange (which in the presence of bicarbonate can mediate monovalent cation transport via cation-CO3 ion pairs23). Second, while bumetanide inhibits (Na ÷ + K ÷ + CI-) cotransport systems with high affinity, this is generally not true with (K ÷ + C1-) systems. Thus, (K + + CI-) cotransport systems have been occasionally termed "bumetanide insensitive" (e.g., Ref. 24) on the basis of the finding that they are not significantly inhibited by 10-5 M bumetanide (a concentration which fully inhibits (Na + + K + + C1-) cotransport under physiological ionic conditions22), when in fact higher concentrations of bumetanide (1 mM) do inhibit (K + + CI-) cotransport (e.g., Refs. 25 and 25a). Third, the inhibitory potency of bumetanide and furosemide toward (Na ÷ + K + + CI-) and (Na + + C1-) cotransport, and in at least one case (K ÷ + C1-) cotransport as well, is dependent on the ionic composition of the extracellular medium. Bumetanide inhibits (Na ÷ + K ÷ + 2C1-) cotransport in duck red cells by apparent competition for a chloride sitefl6 and furosemide likewise competes with chloride in inhibiting (Na ÷ + CI-) cotransport in toad cornea. 27 In contrast, increasing extracellular sodium, [Na]o, and potassium, [K]o, enhances inhibition of (Na + + K + + 2C1-) cotransport by " l o o p " diuretics, 22 and binding of radiolabeled bumetanide to duck red cells 28 and membranes from dog kidney outer medulla29is likewise enhanced by increasing [Na]o and [K]o. In the complete absence of [Na]o and [K]o, 2 × l0 -4 M bumetanide (or 1 mM furosemide) is required for complete inhibition of(Na + + K + + 2C1-) cotransport in duck red cells. In the case of (K + + CI-) cotransport in sheep red cells treated with the sulfhydryl reagent N-ethylamaleimide, furosemide inhibition is also enhanced by increasing [K]o. t2 An alternative to using diuretic-sensitive fluxes as a measure of cation-chloride cotransport is to measure chloride-dependent cation fluxes, i.e., the difference in ouabain-insensitive sodium or potassium flux in the presence of chloride and in the complete absence of chloride. Most, if not 22 H. C. Palfrey, P. W. Feit, and P. Greengard, Am. J. Physiol. 238, C139 (1980). 23 j. Funder and J. O. Wieth, Acta Physiol. Scand. 71, 168 (1967). 24 C. Brugnara, A. S. Kopin, H. F. Bunn, and D. C. Tosteson, J. Clin. Invest. 75, 1608 (1985). 25 L. R. Berkowitz and E. P. Orringer, Am. J. Physiol. 252, C300 (1987). zsa D. Kaji, J. Gen. Physiol. 88, 719 (1986). 26 M. Haas and T. J. McManus, Am. J. Physiol. 245, C235 (1983). 27 j. H. Ludens, J. Pharmacol. Exp. Ther. 22,3, 25 (1982). M. Haas and B. Forbush III, J. Biol. Chem. 261, 8434 (1986). 29 B. Forbush III and H. C. Palfrey, J. Biol. Chem. 258, 11787 (1983).
[16]
CATION--ANION COTRANSPORT
283
all cation-chloride cotransport systems are completely inhibited when chloride is totally replaced by a permeant anion other than bromide. Ideally, chloride-dependent fluxes should be equal to furosemide- or bumetanide-sensitive fluxes when a fully inhibitory concentration of diuretic is used, and in avian red cells we find this to be the case when the anions methyl sulfate or methane sulfonate are used to replace chloride. However, relatively lipophilic anions such as nitrate and thiocyanate increase passive, nonspecific fluxes of sodium and potassium across red cell membranes, 23 and thus ouabain-insensitive cation fluxes in the presence of these anions will be higher than in a chloride medium containing a dose of furosemide or bumetanide that fully inhibits cation-chloride cotransport. It is also advisable to avoid the use of nitrate in studies of cation-chloride cotransport for another reason: in several different cell types it appears that nitrate may compete with chloride for sites on (Na + + K + + C1-) cotransport 5,6,28 and (K + + C1-) 3° cotransport pathways. Stoichiometry of Cotransport: The Zero-Trans Efflux Method Measurement of net sodium and potassium effluxes into media free of these cations has proved to be a useful way of determining stoichiometric relationships between cotransported cations in red blood cells. This method is particularly applicable to cells such as red cells in which the baseline permeability of the cell membrane to monovalent cations is low, and most of the transport of sodium and potassium occurs via specific mediated pathways. In duck red cells in particular, >50% of sodium and potassium effluxes into isotonic choline or tetramethylammonium (TMA) media are inhibited either by 1 mM furosemide or by replacement of intraand extracellular chloride with methyl sulfate; this percentage increases when (Na t + K ÷ + 2C1-) cotransport or (K ÷ + CI-) cotransport is stimulated by shrinking or swelling the cells, respectively (see below). Choline and TMA do not appear to interact with either (Na t + K ÷ + 2CI ) or (K ÷ + CI-) cotransport systems in a variety of cell types, and are thus the cations of choice for use in the incubation medium in zero-trans efflux experiments. We have used TMA in most of our experiments simply because TMA salts of several monovalent anions other than chloride are available commercially or can be easily prepared from TMA hydroxide and the appropriate acid. Efflux determinations can be made by measuring either cell or medium sodium and potassium contents before and after given time periods of incubation in the zero-trans medium. When the effluxes of sodium and potassium are large, as in duck red cells, measure3op. K. Lauf, J. Membr. Biol. 82, 167 (1984).
284
RED BLOOD CELLS
[16]
ment of cell cation contents gives good results. However, it is important to note that when substantial net salt efflux occurs, the cells will also lose water and shrink during the incubation period. Therefore, cell sodium and potassium levels (and values for net effluxes) should be determined and expressed as total contents (millimoles/kilogram cell solid or millimoles/ liter original cells) rather than as concentrations (millimoles/liter cell water). When the effluxes of sodium and potassium are small, as in human red cells, it is preferable to measure the accumulation of these cations in the extracellular medium, either directly or by isotopic tracer methods using cells preloaded with Z2Na, 24Na, 42K, or 86Rb (the latter being a potassium congener). If the hematocrit and cell contents of sodium and potassium (and specific activities of their tracers, if applicable) are known, measurements of sodium and potassium accumulation in the medium will yield values for their effluxes. An example of a zero-trans efflux experiment using duck red cells is shown in Fig. 1. In this experiment, the cells were loaded using the nystatin method 3~ (see Fig. I legend) to contain a constant level of potassium and three different sodium contents. They were then incubated 10 min in sodium- and potassium-free TMA chloride media of different osmolalities, with or without 1 mM furosemide (the range of osmolalities employed was different for the three groups of cells in order to bring cells of three different sodium and total salt contents into the same range of cell volumes). In Fig. 1, furosemide-sensitive net sodium efflux (-ANal) and potassium efflux (-AKc) are plotted as a function of the average cell water content over the 10-min incubation period; this average was used because the cells were progressively shrinking as they lost sodium, potassium, and chloride into the TMA chloride media. The isotonic water content of duck red cells is 1.50 kg H20/kg cell solid, or 60% (w/w). In swollen cells with Wc > 1.5, the cells lose potassium without a concomitant sodium loss, indicative of the (K + + C1-) cotransport seen in swollen duck red cells. 2°,2l This swelling-induced (K + + CI-) cotransport is independent of [Na]c, at least over the range of values tested. In contrast, cell shrinkage (We < 1.5) stimulates net sodium and potassium efflux with an N a + : K ÷ stoichiometry of 1:1, indicative of the (Na + + K + + 2C1-) cotransport activated by shrinkage of duck red cells. 1°:7 Increasing cell sodium from 41.1 to 73.6 mmol/kg cell solid increases the level of outward (Na + + K ÷ + 2C1-) cotransport; the effect of increasing cell sodium on this process appears to saturate above the latter level. The zero-trans experiment in Fig. 1 nicely displays the two volumesensitive cation-chloride cotransport systems in duck red cells, and dem31 A. Cass and M. Dalmark, Nature (London), New Biol. 244, 47 (1973).
[16]
CATION--ANION COTRANSPORT
~O
oA
~ ~
285
4o Z
A
w> o30
•
•
o
o
[]
J
-J zw L,I '0 O3 ~
20
0
D
\\\
I W
0
~ \\L.ff
~
• •
o o
~,.,._o.~M~ 73.6+0.,
,
U_
, -//I.25
-
1.50
1.75 Wc
(Kg H 2 0 / K g
2.00
2.25
2.50
CELL SOLID)
FIG. 1. Effects of cell water and sodium content on furosemide-sensitive net sodium and potassium effluxes from duck red cells into sodium- and potassium-free media. Cells were prepared to contain 202 mmol potassium/kg cell solid and the three listed sodium contents by the nystatin method as follows: Cells were incubated for 1 hr at 0°, 2% hematocrit in media containing 20/~g/ml nystatin, 50 mM sucrose, 100 mM KC1, and either 20, 35, or 50 mM NaCI. This procedure was then repeated by centrifuging and resuspending the cells in the same media without nystatin. Nystatin was eluted from the cells, restoring the membrane to its normal state of low cation permeability, by washing the cells six times at 25°, 2% hematocrit in the same solution as above (without nystatin), but containing 0.1 mM ouabain and 0.25% bovine serum albumin (fraction V). The albumin-containing solutions were titrated to pH 7.4 at 25° with KOH. Test incubations were performed for 10 min at 3% hematocrit, 41 ° (body temperature for ducks) in media containing varying concentrations of TMA chloride, with and without 1 mM furosemide. The ranges of osmolalities over which the cells were incubated were as follows: 140-350 mOsm/kg (O), 170-380 (11), and 200-410 (A), all in increments of 30 mOsm/kg. Open symbols represent furosemide-sensitive net potassium efflux (-AKc), closed symbols represent furosemide-sensitive net sodium efflux
(-ANal).
onstrates the l : l N a + : K ÷ stoichiometry of the (Na + + K + + 2 0 - ) cotransport pathway. Whether experiments such as that in Fig. 1 can be routinely used to s h o w c a t i o n : c h l o r i d e stoichiometries of cotransport systems in red cells remains an open question; this is because a small but significant fraction o f anion exchange remains even in cells treated with a maximal inhibitory concentration o f D I D S , a stilbene derivative that is a potent inhibitor o f the red cell anion exchanger. B e c a u s e chloride fluxes
286
RED BLOOD CELLS
[16]
via the anion exchanger are so rapid, even when this exchanger is 99.99% inhibited, the residual chloride fluxes (at 37°) would be expected to be of the same order of magnitude as those occurring via (Na ÷ + K ÷ + 2C1-) cotransport in maximally shrunken duck red cells. Lytle32 has circumvented the anion-exchange problem by performing measurements of net Na ÷, K ÷, and CI- effluxes from duck red blood cells incubated in TMAgluconate medium. The cells were first treated with DIDS. Under these conditions, bumetanide (1 mM)-sensitive efflux stoichiometries of approximately 1 N a ÷ : l K + : 2 C1- and 1 K ÷ : I CI- were obtained in shrunken and swollen cells, respectively. Substitution of impermeant, nonexchangeable anions such as gluconate for extracellular chloride does markedly alter the membrane potential of red blood cells, and is thus best applied in studies of transport processes known to be electroneutral (see below). Treatment of the cells with DIDS is also warranted to minimize changes in intracellular pH resulting from CI-/OH- exchange or (H ÷ + CI-) cotransport via the anion-exchange pathway. 3aa Hall and Ellory33 have reported experiments in which the stoichiometry of bumetanide-sensitive tracer sodium, potassium, and chloride influxes in ferret red cells (which exhibit bumetanide-sensitive fluxes similar in magnitude to those in duck red cells H) was found to be 2 Na ÷ : 1 K ÷ : 3 CI-, similar to the cotransport stoichiometry determined in squid axon. 3 In these experiments of Hall and Ellory33 anion exchange was minimized not only by treating the cells with two inhibitors of this exchange (DIDS plus SITS), but also by lowering the temperature to 21 ° . The rationale for lowering the temperature in these experiments is that the red cell anionexchange pathway has a greater sensitivity to temperature (higher Arrhenius activation energy) than does (Na ÷ + K ÷ + C1-) cotransport. 33 A major advantage of the zero-trans efflux method with regard to determination of stoichiometries of cotransported cations is that potential errors due to K÷/K + and Na÷/Na ÷ exchanges are avoided. Both d u c k 10'20'32'34 and h u m a n j3,34a red cells exhibit bumetanide-sensitive, chloride-dependent, 1 : 1 K+/K ÷ and Na÷/Na + exchanges, and kinetic analysis of these exchanges has provided strong evidence for their being partial reactions of (Na + + K + + 2C1-) cotransport. 3~'34,34a When bumetanidesensitive tracer influxes of sodium and potassium are measured in either 32 C. Lytle, Ph.D. thesis, Duke University (1988). 32a M. L. Jennings, J. Membr. Biol. 40, 365 (1978). 33 A. C. Hall and J. C. Ellory, J. Membr. Biol. 85, 205 (1985). 34 C. Lytle, M. Haas, and T. J. McManus, Fed. Proc., Fed. Am. Soc. Exp. Biol. 45, 548 (1986). 34a j. Duhm, J. Membr. Biol. 98, 15 (1987).
[16]
CATION-ANION COTRANSPORT
287
duck or human red cells, the potassium influx exceeds that of sodium because the high intracellular potassium concentration of these cells favors K+/K ÷ exchange relative to Na÷/Na ÷ exchange. In duck red cells prepared by the nystatin method to contain 150 mM sodium and 10 mM potassium, bumetanide-sensitive sodium influx now exceeds potassium i n f l u x . 32,34 Indeed, it is quite possible that the stoichiometry of 2 Na ÷ : 1 K ÷ : 3 CI- reported for bumetanide-sensitive influxes in ferret red cells 33 reflects a component of Na÷/Na + exchange, since ferret red cells have very low ouabain-sensitive cation fluxes and thus contain sodium as their major intracellular cation. ~ In zero-trans efflux experiments, K+/K + and Na÷/Na + exchanges cannot occur, and the true stoichiometry of (Na + + K ÷ + C1-) cotransport processes is determined. Electrogenicity vs Electroneutrality of Transport: The Valinomycin Method A decade ago it was shown that partial replacement of extracellular chloride, [Cl]o, by an impermeant anion such as gluconate markedly inhibited furosemide-sensitive net sodium and potassium influx into duck red cells which were shrunken or treated with catecholamines. 35 Because in red cells the membrane potential (Era) is equal to the Nernst potential for chloride (Ecl), replacement of [Cl]o by gluconate at constant intracellular chloride, [C1]c, causes Em to become more positive. It was thus initially thought that the decrease in (Na ÷ + K ÷) uptake when [Cl]o was partially replaced by gluconate was the result of this depolarization. 35 In this model, (Na ÷ + K ÷) was thought to cross the membrane in an electrogenic manner, with chloride (or presumably, any other permeant anion) following passively to maintain electroneutrality. However, the finding that the coupled, furosemide-sensitive movements of sodium and potassium in duck red cells have a specific anion requirement for chloride (or bromide) and are completely inhibited when chloride is fully replaced by other permeant anions such as nitrate 1°,36 led to the consideration that chloride might in fact move across the membrane in an electrically neutral complex with the cations, i.e., (Na ÷ + K ÷ + 2C1-). In this latter case, the inhibitory effect of partial replacement of [C1]o at constant [Cl]c can be attributed to a change in the direction of the chloride chemical potential gradient, independent of Em. However, as long as the ratio of [Cl]c/[Cl]o affects both Em and the chloride chemical gradient, as is the case in red 3s W. F. Schmidt III and T. J. McManus, J. Gen. Physiol. 70, 99 (1977). 36 F. M. Kregenow and T. Caryk, Physiologist 22, 73 (1979).
288
RED BLOOD CELLS
[16]
cells, it is not possible to distinguish between these two possible mechanisms of transport [electrogenic (Na + + K +) cotransport versus electroneutral (Na + + K + + 2C1-) cotransport]. This problem was resolved by employing the potassium ionophore valinomycin to increase the potassium conductance (PrO of the membrane sufficiently so that PK exceeded Pcl. Upon further addition of DIDS, which reduces Pc~ in red cells by -50% (in addition to inhibiting electroneutral anion exchange37), PK sufficiently exceeds P a so that Em can be held constant near the potassium equilibrium potential (EK) while [Clio is varied at constant [Cl]c.~° It is actually possible to estimate PK/Pcl in valinomycin-treated cells by measuring 1- or 2-min potassium effluxes from these cells into NaCI and RbCI media. If sufficient valinomycin is present (10 -7 M or more in human red cells38), the potassium efflux into NaC1 gives an estimate of Pcl, since the rate of this efflux is limited only by the conductance of the major permeable anion available to follow potassium out of the cell to maintain electroneutrality. 38 If, however, the external medium contains only RbCI, valinomycin promotes a K+/Rb + exchange which is limited by PK, since valinomycin preferentially selects rubidium over potassium. 39 Thus, potassium efflux into RbC1 gives an estimate of PK, and PK/Pc~in valinomycin-treated cells can be estimated by the ratio of K ÷ efflux into RbCI divided by K + efflux into NaCI (see Table I of Ref. 10 for further details). Knowing PK/Pa and assuming Pya is small relative to PK and Pcl (a safe assumption in valinomycin-treated red cells), Em can be estimated from the constant-field (GoldmanHodgkin-Katz) equation. With Em held constant near EK by the use of valinomycin and DIDS, it is possible to vary [Cl]o (at constant [Cl]c) independently of Era. Figure 2 shows an experiment in which net changes in cell sodium were measured in DIDS-treated duck red cells as a function of [C1]o, at constant [Cl]c. Methyl sulfate, which has a conductance about one-half that of chloride in DIDS-treated cells, J0 was used to replace chloride; with DIDS present the potential exchange of internal chloride for external methyl sulfate was inhibited and [Cl]c remained essentially constant over the 5-min incubation period. The cells were incubated with norepinephrine (1 /zM) to stimulate (Na ÷ + K ÷ + 2C1-) cotransport, in the presence and absence of 2/zM valinomycin and 1 mM furosemide. In the presence of valinomycin, Em was held constant near EK (about - 8 mV in this experiment), whereas Ern became progressively more negative in the absence of valinomycin as 37 p. A. Knauf, G. F. Fuhrmann, S. Rothstein, and A. Rothstein, J. Gen. Physiol. 69, 363 (1977). 38 M. J. Hunter, J. Physiol. (London) 268, 35 (1977). 39 T. E. Andreoli, M. Tieffenberg, and D. C. Tosteson, J. Gen. Physiol. 50, 2527 (1967).
[16]
CATION--ANION COTRANSPORT SYMBOL
VALINOMYCIN ( 2 / . t M )
on
--
o
~
+
• •
289
0
tf)
-~ _J
o
<3 ~
-4
o Z t,O
- /
"~---+NO~PINEPHRINE (Na)0 = 5 0 m M (K)0 • l O O m M
-6
r-(Na)c = 4 8 m M initial~ ( K ) c 137 rnM
E
L_(CI)c
-12 I 25
I 50
I 75 (CI)o mM
I I00
50 m M
I 125
I 150
FJc. 2. Effect of the chloride gradient on norepinephrine-stimulated, furosemide-sensirive net sodium m o v e m e n t s in duck red cells at constant and variable Era. Cells were preincubated 8 hr at 4 ° in a medium containing 80 m M sodium methyl sulfate (MeSO4), 60 m M NaCI, and 20 m M KCI to increase cell sodium and also replace - 5 0 % of cell chloride with MeSOn. Next, an additional preincubation was done tbr 90 rain at 41 ° (body temperature for ducks) in media of the same ionic composition but also containing 0. I m M ouabain (to further increase [Na],) and 10 5 M DIDS. Test incubations were done at 41 °, 3% hematocrit in media containing 50 m M sodium, 100 m M potassium, 10 ~ M DIDS, 10 " M norepinephrine, and 0.1 m M ouabain, with and without 2 × 10 6 M valinomycin and 1 m M furosemide. Methyl sulfate substituted for external chloride, maintaining isotonicity, in the presence of valinomycin, PK/Pct was estimated to be 7.5 (see text), and Em (as estimated from the constant-field equation) was constant at - 7 . 9 -+ 0.5 mV (SEM, n = 8) over the entire range of [CI],, from 2.5 to 150 m M . Em in the absence of valinomycin varied from - 0 . 4 mV at 2.5 m M [CI],, to - 1 8 . 3 mV at 150 m M [CI],, as estimated from the constant-field equation with PM~so4= 0.5 P o (as determined in DIDS-treated duck red cells ~") and a s s u m i n g Pcl >>PK ,PN,. (Reproduced from The Journal of General Physiology, 1982, volume 80, pp. 125-147, by copyright permission of the Rockefeller University Press.)
[Cl]o was increased (see legend for Fig. 2). Figure 2 shows that at low [Cl]o there was a furosemide-sensitive net loss of sodium, whereas at high [Cl]o, when the chloride gradient was inwardly directed, the cells gained sodium. This effect of [C1]o on furosemide-sensitive net sodium movements was identical in the presence and absence of valinomycin, showing that the parameter affecting this sodium transport is the chemical potential gradient of chloride and not Era. The experiment shown in Fig. 2 strongly supports the idea that catecholamine-stimulated, furosemide-sensitive sodium transport in duck red
290
RED BLOOD CELLS
[16]
cells occurs via an electrically neutral cotransport of (Na + + K ÷ + 2C1-) rather than an electrogenic (Na + + K ÷) cotransport. Further evidence that this is the case has come from experiments similar to that in Fig. 2, only in which a significant sodium gradient was present and furosemidesensitive net transport of sodium against this gradient was driven by a chloride gradient at constant Em (see Fig. 7 of Ref. 10). Finally, two methods can and have been used to confirm that the cotransport pathway is indeed electrically neutral. First, in cells treated with DIDS, the effect of Em on furosemide-sensitive net sodium transport can be studied by varying [K]o at constant [K]c in the presence of valinomycin (varying Em ; Em = EK) and in its absence (constant Em ; Em =Ecl). When this experiment was done in duck red cells, the same effect of varying [K]o was seen whether or not valinomycin was present, confirming a lack of effect of Em on the cotransport pathway and strongly suggesting its electroneutrality. 10 In addition, changes in Em upon activation or inhibition of transport pathways can be detected using potential-sensitive, fluorescent carbocyanine dyes such as diS-C3[5] (for a detailed study of the use of such dyes in red cells, see Ref. 40). Briefly, these dyes are taken up by red cells upon hyperpolarization of the membrane, and released upon depolarization. Because the fluorescence of the dye is quenched when it is within (or associated with) the cells, changes in dye fluorescence can be used to monitor changes in Em .40 This dye method was used to demonstrate the electrogenicity of the Na÷/K ÷ pump in red cells, as a depolarization of the membrane (increase in diS-C3[5] fluorescence) was seen upon addition of ouabain. 41 This result is not surprising, considering that the Na÷/K ÷ pump transports three sodium ions out of the cell in exchange for two potassium ions. 42 However, in duck red cells incubated in a TMA chloride medium containing I mM KC1 and 1 mM NaCI [conditions designed to maximize outward net (Na + + K ÷ + 2C1-) cotransport], we found no effect of either adding norepinephrine (to stimulate cotransport) or subsequently increasing [K]o on diS-C3[5] fluorescence, confirming the electroneutrality of (Na + + K + + 2C1-) cotransport in these cells 2j (also see Ref. 43). Likewise, the (K ÷ + CI-) cotransport stimulated by cell swelling was found to be electrically neutral using the dye method in duck red cells, 2j as well as 4o p. j. Sims, A. S. Waggoner, C. H. Wang, and J. F. Hoffman, Biochemistry 13, 3315 (1974). 4~ j. F. Hoffman, J. H. Kaplan, and T. J. Callahan, Fed. Proc., Fed. Am. Soc. Exp. Biol. 38, 2440 (1979). 42 R. L. Post and P. C. Jolly, Biochim. Biophys. Acta 25, 108 (1957). 43 F. M. Kregenow, in "Membrane Transport in Red Cells" (J. C. Ellory and V. L. Lew, eds.), p. 383. Academic Press, New York, 1977.
[16]
CATION--ANION COTRANSPORT
291
in human red cells, 44 where Em was determined from the ratio of [H+]c/ [H+]o in cells treated with the proton-specific ionophore CCCP and DIDS. 45 Summary Two methods have been described for the study of cation-chloride cotransport systems. The zero-trans efflux method is designed to determine stoichiometric relationships between cotransported ions under conditions where ion exchanges cannot occur. These exchanges (e.g., Na+/ Na +, K+/K ÷) may occur as partial or incomplete reactions of a cotransport process and can lead to erroneous determinations of the stoichiometry of the cotransport process. The zero-trans efflux method can also be used to study the effects of cell volume, pH, and intracellular ion concentrations on cotransport processes. The valinomycin method is used to determine the electrogenicity or electroneutrality of transport, and in this regard can be used in conjunction with other methods such as those employing potential-sensitive dyes or microelectrodes. Other, more recently developed ionophores with specificity for lithium rather than potassium 46 have now been used to study the effect of Em on the ATP-dependent Na+/K + p u m p . 47 It may be possible to use such ionophores to confirm the suspected electroneutrality of (K ÷ + CI-) cotransport systems, as well as for other studies of specific potassium transport processes in which valinomycin obviously cannot be used. Both methods discussed in detail in this chapter, and particularly the valinomycin method, were originally devised for use in red blood cells in order to take advantage of (or circumvent) properties of the red cell membrane, such as its low permeability to sodium and potassium and relatively high permeability to chloride. However, valinomycin has been used successfully to demonstrate the electroneutrality of (Na + + K ÷ + 2C1-) cotransport in MDCK cells, 4 and the zero-trans efflux method should be applicable to the study of transport processes in other types of cells in suspension, so long as the transport system being studied can be accurately defined (e.g., as an inhibitor-sensitive or chloride-dependent cation flux) and comprises a significant fraction of the total salt efflux. 44 C. Brugnara, T. Van Ha, and D. C. Tosteson, J. Gen. Physiol. 92, 42a (1988). 45 R. I. Macey, J. S. Adorante, and F. W. Orme, Biochim. Biophys. Acta 512, 284 (1978). 46 R. Margalit and A. Shanzer, Biochim. Biophys. Acta 649, 441 (1981). 47 R. Goldshlegger, S. J. D. Karlish, A. Rephaeli and W. D. Stein, J. Physiol. (London) 387, 331 (1987).