Na-K-2CI Cotransport in Intestinal Epithelial Cells INFLUENCE OF CHLORIDE EFFLUX AND F-ACTIN ON REGULATION OF COTRANSPORTER ACTMTY AND BUMETANIDE BINDING*

Although CAMP-dependent epithelial chloride secre- tion is largely regulated via apical membrane chloride channels, CAMP also remodels basolateral F-actin and activates basolateral Na-K-2Cl cotransport. Whether activation of cotransport is a primary event or secondary to activation of chloride efflux is not established, and the basis for the cytoskeletal dependence is unknown. We studied cotransport in the intestinal line (which lacks cAMP-regulated chloride efflux) and in its subclone C1.19A (in which this pathway is present). Cotransporter activity was enhanced by forskolin in both lines but to a considerably greater extent in subclone C1.194 in which the number of bumetanide binding sites was also observed to increase. The F-actin stabilizer phalloidin markedly attenuated CAMP-stimulated co- transport in C1.19A monolayers, but the increase in bumetanide binding was preserved. These studies identify two mechanisms for activation of Na-K-2Cl cotransport by CAMP: components independent and dependent of CAMP-elicited chloride efflux. Additional Na-K-2Cl cotransporters become accessible to the cell surface coin-

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. j: Supported by a pilot feasibility study from the Harvard Digestive Diseases Center, a basic research award from the Glaxo Institute for Digestive Health, and National Institutes of Health Grant R29 DK48010-01. To whom correspondence should be addressed:Dept. of Surgery, Beth Israel Hospital, 330 Brookline Ave., Boston, MA 02215. chloride entry must increase to balance the rate of apical chloride exit; the mechanism by which such coordination of apical and basolateral transport events is accomplished is incompletely understood (3,4).
Internalization of chloride across the basolateral membrane of chloride-secreting epithelia occurs primarily via bumetanide-inhibitable Na-K-2Cl cotransport (5). Although in order to meet the demands of sustained secretion, Na-K-2C1 cotransport could increase simply on the basis of altered chemical gradients, evidence from a number of epithelial and non-epi-thelia1 cell types argues for complex and possibly multiple regulatory mechanisms. In shark rectal gland and avian erythrocytes, cotransporter activity appears to correlate with the phosphorylation state of the cotransporter protein (3, 41, suggesting that CAMP may directly activate the cotransporter via a CAMP-dependent protein kinase cascade. Furthermore, in several cell types, cotransport can be activated by hypertonic cell shrinkage (6), implying that, in the case of a secretory epithelium, Na-K-2C1 cotransport could increase as a secondary response t o CAMP-elicited salt and water loss engendered by the activation of apical membrane chloride channels (3,4,7,8). Finally, an additional level of regulation has been identified in several cell types whereby CAMP elicits an increase in the number of specific bumetanide binding sites, an event that may reflect a n increase in the number of functional cotransporters in the plasma membrane (3,4,(7)(8)(9)(10)(11)(12)(13).
We have shown that, in the chloride-secreting human intestinal epithelial cell line T84, CAMP not only activates apical membrane chloride efflux, but also concurrently elicits basolat-era1 microfilament remodeling and enhances basolateral salt uptake via Na-K-2Cl cotransport. Furthermore, these additional basolateral events (but not CAMP-stimulated apical chloride exit) could be largely attenuated by the F-actin stabilizing agent phalloidin, with resultant inhibition of net chloride secretion (14,15). Whether the observed CAMP-elicited activation of cotransport is a primary event or a secondary response to the activation of CAMP-regulated chloride efflux has not been elucidated, and the basis for the microfilament dependence of regulation is unknown.
One possible way to address the importance of CAMP-elicited chloride efflux to the regulation of cotransport by CAMP would be to eliminate the CAMP-regulated chloride efflux pathway. Unfortunately, of the available pharmacologic blockers of chloride channels, none is specific in its action or completely inhibitory (16). Alternatively, one could compare cotransporter regulation in cell lines with variable expression of the presumed CAMP-regulated chloride efflux pathway (CFTR) (17). In this regard, the pluripotent human intestinal epithelial cell line HT29 presents a unique model system in which to study the regulation of cotransport. Parent HT29 cells are unpolarized and undifferentiated and thus resemble fetal colonocytes (18); although parent HT29 cells contain mRNA for CFTR, CFTR protein is not processed into the plasma membrane, and the CAMP-regulated chloride efflux pathway is absent (17). Ion transport pathways known to be present in parent HT29 cells reflect those typically found in the basolateral membranes of chloride secretory epithelial cells and include Na-K ATPase, Na-K-2C1 cotransporter, and several types of potassium channels (19,20). By manipulation of culture conditions (e.g. treatment with sodium butyrate), parent HT29 cells have been induced to express a n apical membrane domain inclusive of chloride efflux pathways (18). One such subclone, a permanently differentiated line designated C1.19A (21), expresses amounts of CFTR protein similar to those expressed by the parental HT29 line, but CFTR protein reaches the plasma membrane and the regulated chloride efflux pathway (and the capacity for vectorial chloride secretion) is present (17).
In this study, we compare parent HT29 cells with C1.19Acells to assess the requirement for chloride emux in the regulation of Na-K-2C1 cotransport by CAMP and to further define the role of F-actin. We find that, in parent HT29 cells, CAMP rapidly increases cotransport despite the absence of a regulated chloride efflux pathway but that activation of cotransport is not accompanied by an increase in bumetanide binding sites. In contrast, in subclone C1.19A, substantially greater activation of cotransport by CAMP occurs and is accompanied by a 2-fold increase in bumetanide binding. Phalloidin inhibited CAMP-stimulated cotransporter activity in C1.19A cells without affecting the number of bumetanide binding sites. The results indicate that CAMP may activate Na-K-2Cl cotransport both by pathways dependent on and independent of regulated chloride efflux. Furthermore, the CAMP-elicited increment in bumetanide binding occurs independently of F-actin rearrangement, whereas activation of ion translocation by the cotransporter appears to require this cytoskeletal remodeling.

MATERIALS AND METHODS
Cell Culture-Human colon adenocarcinoma HT29 cells obtained from American Type Culture Collection and the C1.19A subclone, generously provided by Dr. C. L. Laboisse (211, were maintained in McCoy's 5A medium and Dulbecco's modified Eagle's medium, respectively, with 10% heat-inactivated fetal bovine serum and glucose (4.5 @iter). Cells were grown in a humidified 95% 0,-5% CO, incubator at 37 "C, fed every other day, and split 1:5 every 5 days. Experiments were carried out using cells which were grown on 35-mm Petri dishes or 24-well plates and which were fed every 2-3 days and on the day prior to experiments. Unless specified, experiments were carried out in a HEPES-phosphate-buffered Ringer's solution (HPBR), pH 7.5, which contains 135 rn NaC1,5 m~ KC1,3.33 m M NaH,PO,, 0.83 m M Na,HPO,, 1 m~ CaCI,, 1 m M MgCl,, 10 mM glucose, and 5 m M HEPES. Solutions were prewarmed to 37 "C.
lZ5I Efflux-The presence of a cAMP-regulated chloride efflux pathway was determined by ' ' ' 1 efflux, as previously described (14,22). Briefly, monolayers were loaded in HPBR containing 2.5 pCi x ml-I for 30-90 min. Monolayers were then rapidly washed with unlabeled HPBR to eliminate extracellular tracer. One-minute efflux measurements were obtained by a sample-and-replace technique. Residual radioactivity was extracted with 0.1 N NaOH. The rate constant of efflux, expressed in min", was calculated as before (14,22). @Rb Uptake-86Rb uptake was measured as previously reported (14). Monolayers were incubated in HPBR with or without 10 p~ bumetanide, 0.5 m M ouabain, and 0.033-10 p~ forskolin. This solution was then aspirated, and replaced with 1 ml of HPBR containing 1.5 pCi x ml" 86Rb. Uptake was terminated by rapidly washing the dishes four times with an ice-cold solution containing 100 rn MgCl, and 10 m M Tris, pH 7.4. Radioactivity was extracted as above. Total protein was determined on extracts of representative monolayers utilizing a spectrophotometric assay (Pierce Chemical Co.). Uptakes were expressed as nmol of K+ x mg of protein" x min" (14,23). The activity of the Na-K- Preliminary experiments in both cell lines indicated that specific binding of c3H1bumetanide was consistently less than approximately 40% of total binding if the binding buffer consisted of standard HPBR. As reported by others (24), specific binding was found to be enhanced by using a modified HPBR (120 m~ sodium gluconate, 25 m~ potassium gluconate, 5 m M KCl, and 10 m~ "is-HC1, pH 7.4); binding experiments in the present study utilized this modified HPBR. Binding was terminated by four rapid washes with ice-cold 150 m~ NaCl and 10 m M Tris-HC1, pH 7.4. Radioactivity was extracted and protein determined as above. Specific (saturable) binding represented the difference between total binding and binding measured in the presence of excess unlabeled bumetanide. Kd and B,, were determined by Scatchard analysis using least-squares linear curve-fitting to the transformed data, with Kd representing the negative reciprocal of slope of the line thus obtained and I ? , , representing the x intercept.
Phalloidin Loading-Subsets of monolayers were loaded with the F-actin stabilizing agent phalloidin by overnight incubation in media containing 33-100 p~ phalloidin as previously described (14,15,25).
Control monolayers were incubated in fresh media for an equal length of time. In subsets of experiments, absence of monolayer toxicity was confirmed by lactate dehydrogenase release (26).
Materials and Analysis-Radioisotopes were obtained from DuPont NEN. A i I other chemicaIs were from Sigma. All data are expressed as mean -c S.E. Statistical analysis was performed by paired Student's t test or by two-way analysis of variance (ANOVA). Fig. 1, forskolin (10 J.N, a CAMP agonist) evoked a rapid increase in the rate constant of lz5I e m u in 4-5-day-old monolayers of C1.19A cells but not in parent HT29 cells, thus confirming the presence of a CAMP-regulated chloride emux pathway in the subclone expressing apical membrane components (C1.19A) and its absence in the parent HT29 line, consistent with the findings of

Na-K-2C1 Cotransporter Regulation and F-actin
others (17). The CAMP-regulated chloride efflux pathway remained present in post-confluent C1.19A monolayers; in some HT29 monolayers older than 4-5 days, a small cAMP-elicited increase in ' " I efflux that was delayed in onset could occasionally be detected (data not shown), as also noted by others (17).
Nu-K-2Cl Cotransport: Basal and CAMP-regulated-Under unstimulated conditions, bumetanide-sensitive *'jRb uptake was similar in 4-5-day-old monolayers of subclone C1.19A and parent HT29 cells (13.1 2 1.6 versus 10.6 2 1.5 nmol of K' x mg of protein" x min" for C1.19A versus parent HT29, respectively, each n = 32, n.s.). Uptake was linear over the first 10 min of exposure to tracer in both cell lines (data not shown). In response to 10 PM forskolin, bumetanide-sensitive "jRb uptake (both in the presence or absence of ouabain) increased in timedependent fashion in both cell lines (Fig. 2). In the absence of ouabain, peak stimulation was seen between 5 and 15 min, with decreased but still considerable stimulation evident after a 30-min exposure to forskolin. Dose-response curves in both cell lines measured after 10 min of forskolin treatment indicated maximal stimulation of bumetanide-sensitive 86Rb uptake in both lines at 10 p~ forskolin (Fig. 3). Other investigators have demonstrated that 10 PM forskolin increases intracellular CAMP to equivalent levels in both parental and C1.19A cells (17). For all subsequent experiments discussed below, 10 p~ forskolin was used.
In parallel experiments performed on the same day with monolayers of equivalent age (4-5 days), bumetanide-sensitive "Rb uptake was stimulated approximately 2-fold by 10 PM forskolin in parent HT29 cells but increased more than %fold in subclone C1.19A (Fig. 4). Thus, despite the absence of a CAMPactivated chloride efflux pathway in parent HT29 cells, Na-K-2C1 cotransport was stimulated by forskolin in this cell line. Interestingly, the degree of stimulation appeared to be considerably greater in the polarized subclone C1.19A (p < 0.001 compared to parent HT29, n = 32 for each group), which possesses the CAMP-regulated chloride efflux pathway. This difference between parent HT29 cells and the clonal line was also clearly apparent in the time-course experiments depicted in Fig. 2. PHlBumetanide Binding and Response to a P -S p e c i f i c and nonspecific bumetanide binding characteristics for the two cell lines were similar; a representative experiment for C1.19A cells is shown in Fig. 5. Binding of L3HIbumetanide reached equilibrium within approximately 30-40 min (data not shown), consistent with the experience of others (4,24). Under unstimulated conditions, the number of specific [3Hlbumetanide binding sites and the affinity for bumetanide were similar in parent HT29 and subclone C1.19A cells (Fig. 6). By Scatchard analysis, B,, was 1.28 2 0.17 versus 1.26 0.17 pmol x mg of protein" and Kd = 163 2 21 versus 182 2 44 nM for HT29 versus C1.19A, respectively, for four separate experiments. Thus, consistent with the functional data presented above, the number of functional membrane cotransporters as indicated by the number of available binding sites for bumetanide is similar in unstimulated parent HT29 cells and the C1.19A subclone.
We next sought to determine whether the number of specific bumetanide binding sites increases in response to a CAMP stimulus in these cell lines (Fig. 7). In response to 10 J~M forskolin, specific r3H]bumetanide binding was observed to increase markedly (approximately 2-fold) in subclone C1.19A. Interestingly, under the same binding conditions, and in parallel with experiments demonstrating functional activation of the cotransporter (by bumetanide-sensitive s6Rb uptake), no such increase in specific binding of [3H]bumetanide was observed in HT29 cells. By Scatchard analysis, the increase in bumetanide binding in C1.19A cells largely represented a n increase in B,, (   Parent HT29 cells and subclone C1.19A cells were grown in 24-well dishes, and *'jRb uptake studies were performed as described under "Materials and Methods." Monolayers were bathed in HPBR with or without 10 PM bumetanide for 15 min and then exposed to 0.033-10 p~ forskolin for 10 min. =Rb uptake was initiated by aspirating this solution and replacing it with an identical buffer containing 1.5 pCi x ml" a6Rb. After 3 min, uptakes were terminated and radioactivity extracted as described in Fig. 2  forskolin in the presence or absence of 10 p~ bumetanide for 10 min. 86Rb uptake was initiated by aspirating this solution and replacing it with an identical buffer containing 1.5 pCi x ml" 86Rb. After 3 min, uptakes were terminated and radioactivity extracted. Protein content of representative monolayers treated identically were determined, and data were then converted to nmol of K+ x mg of protein" x min" as described under "Materials and Methods." Bars represent mean -c S.E. of n = 32 monolayers in each group. Does Not Affect PHIBumetanide Binding-We reported previously that in the intestinal epithelial cell line T84, CAMP-mediated increases in Na-K-2C1 cotransport (assessed by bumetanide-sensitive 86Rb uptake) were largely attenuated in monolayers loaded with the actin stabilizer phalloidin (14). We utilized similar conditions to examine cotransporter function and l3H1bumetanide binding in the HT29 and C1.19Acell lines.
HT29 and C1.19Amonolayers were loaded with phalloidin by overnight co-incubation in fresh media containing 33-100 PM phalloidin. 86Rb uptake was then measured under unstimulated and forskolin-stimulated conditions as above. The bumetanide-insensitive component of 86Rb uptake was examined as an estimate of Na+/K+-ATPase activity and an indicator of cell viability. Under both unstimulated and forskolin-stimulated conditions, and in both cell lines, the burnetanide-insensitive component of uptake (the ouabain-sensitive fraction) was not affected by phalloidin loading (data not shown), thus demonstrating the absence of nonspecific toxic effects of phalloidin on membrane transport. Further evidence of the absence of toxicity was obtained by measuring lactate dehydrogenase release (7.0 2 2.5% release uersus 6.2 2 2.4% release for 100 PM phal- loidin-loaded uersus control monolayers, each n = 4). Phalloidin loading did not affect bumetanide-sensitive 86Rb uptake in unstimulated monolayers in either cell line. However, the CAMP-elicited increase in cotransporter activity was inhibited by approximately 70% in C1.19Acells (Fig. 8). Despite the dramatic inhibition of CAMP-stimulated cotransporter activity observed with phalloidin-loaded C1.19A cells, [3H]bumetanide binding in C1.19A cells indicated no effect on either basal or CAMP-stimulated increases in B,,, nor any effect of phalloidin on the Kd for bumetanide (Fig. 9). For example, for a representative experiment performed in triplicate, forskolinstimulated B,, was 2.86 versus 3.12 pmol x mg of protein" and Kd was 124 uersus 122 nM for control versus phalloidin-loaded monolayers, respectively; this experiment was repeated on three more occasions, with similar results. Phalloidin loading did not significantly inhibit CAMP-mediated increases in bumetanide-sensitive "Rb uptake in HT29 cells (0.10 < p < 0.15, Fig. 8). Similarly, in a limited series of experiments, phalloidin did not affect L3H]burnetanide binding in parent HT29 cells (data not shown).

DISCUSSION
The Na-K-2Cl cotransporter is an integral component of the transport apparatus of numerous secretory and absorptive epithelia (5,271. In addition, activation of Na-K-2Cl cotransport is an important means by which many epithelial and non-epithelial cells restore normal cell volume after cell shrinkage (6). The   '"-,+*.

6.
'. interrelationship between the roles of the cotransporter in epithelial chloride secretion and cell volume regulation is not clearly understood. It has been speculated that vectorial transport by polarized epithelial cells could reflect a specialized adaptation of volume-regulatory transport processes present in non-polarized cells (1). Simplistically viewed, a chloride-transporting epithelium accomplishes net secretion by restricting a salt-dumping pathway (chloride channels) to the apical membrane and a salt-loading pathway (the cotransporter) t o the basolateral membrane. However, it remains to be established whether regulation of basolateral chloride uptake through the Na-K-2Cl cotransporter and chloride exit through apical chlo-ride channels are linked via changes in cell volume or changes in intracellular ionic activities or, indeed, whether these transport events are independently regulated.
We have shown that CAMP activates Na-K-2C1 cotransport in the chloride-secreting T84 cell line and that both cAMP-mediated activation of Na-K-2C1 cotransport and net chloride secretion could be largely attenuated by phalloidin (14). This suggested that the basolateral Na-K-2C1 cotransporter could be an important regulatory site for net chloride secretion and that cotransporter activity was functionally modified by dynamic changes in F-actin. Because cell volume perturbations may be associated with actin remodeling (281, one possible explanation for these findings is that the cytoskeleton provides a means for transducing volume-regulatory signals to membrane-bound transport proteins such as the cotransporter. We attempted to further elucidate the role of the apical chloride efflux pathway and the influence of F-actin in the CAMPmediated regulation of Na-K-2C1 cotransport by examining the parent and a clonal derivative of a human intestinal epithelial cell line, in which the clone possesses a regulated chloride efflux pathway while the parent does not. We observed that, under unstimulated conditions, parent HT29 cells and subclone C1.19A display similar functionally defined Na-K-2C1 cotransporter activity and similar bumetanide binding characteristics, and others have shown the parent and subclone generate comparable CAMP signals (17). In response to maximal agonistelicited activation of adenylate cyclase, rapid stimulation of ion translocation by the Na-K-2C1 cotransporter (bumetanide-sensitive "Rb uptake) was evident in both parent HT29 cells and subclone C1.19A. Thus, despite the absence of a CAMP-replated chloride efflux pathway in parent HT29 cells, Na-K-2C1 cotransport doubled, implying that the presence of a regulated anion efflux pathway is not required for cotransporter activation. This finding is consistent with earlier work by Turner and co-workers (19,20), who observed that 30-90-min exposure of HT29 cells to forskolin activated bumetanide-sensitive potassium influx in HT29 cells. We also establish that activation of cotransport by forskolin is rapid, apparent within 2 min, evidence consistent with a direct role of CAMP in cotransporter regulation. However, activation of cotransport by CAMP in subclone C1.19Awas considerably greater than in the parent HT29 cell line, suggesting that the presence of the regulated chloride efflux pathway may augment the response of the cotransporter to a CAMP stimulus.
This difference between parent HT29 cells and subclone C1.19A is consistent with the two general mechanisms for cotransporter regulation by CAMP proposed by Haas et al. (3): a direct mechanism in which cotransport is activated by CAMPdependent phosphorylation and an indirect mechanism in which cotransport is activated as a secondary response to salt loss through CAMP-regulated chloride channels. The exact mechanism by which CAMP is capable of directly activating cotransport is unknown, and it is interesting in this regard to recognize that in some systems, CAMP has been found, in fact, to inhibit cotransport ( 5 ) . Recently reported biochemical characterization of the cotransporter protein of shark rectal gland and avian erythrocytes indicate that activation of cotransport indeed correlates with the state of phosphorylation of the putative cotransporter protein, although whether the cotransporter itself is the substrate for the action of CAMP-dependent protein kinase remains t o be established (10,13).
Whether the apparent augmentation of CAMP-stimulated Na-K-2C1 cotransport in cells possessing a regulated chloride efflux pathway is due to cell shrinkage or to changes in intracellular ion activities is not clear. Little is known about the cell volume regulatory mechanisms of either HT29 or C1.19A cells. However, it is interesting to note that bumetanide-sensitive potassium uptake in HT29 cells does not increase in response to hypertonicity (ZO), suggesting that the smaller CAMP-elicited increase in cotransporter activity observed in parent HT29 cells cannot be attributed solely to the absence of cell shrinkage that would result from CAMP-dependent salt efflux. There is some evidence that cotransporter activity may be modulated by cytoplasmic chloride activity (81, possibly through allosteric inhibition at an internal modifier site; it therefore seems possible that some of the apparent volume-regulatory responsiveness of the cotransporter may, in fact, be due to changes in intracellular chloride activity rather than cell shrinkage per se (29).
One possible explanation for the greater CAMP-elicited increase of cotransport in subclone C1.19A is suggested by the t3H1bumetanide-binding data. In response to CAMP, the numbers of specific binding sites increased markedly in the differentiated subclone, whereas parent HT29 cells showed only a minimal change in bumetanide binding sites. Thus, one could speculate that the number of functional cotransporters in the plasma membrane of C1.19A cells is up-regulated in response to CAMP, with the increase in numbers of cotransporters being associated with the chloride efflux-dependent (indirect) component of CAMP-mediated regulation. The lesser degree of activation of cotransport in parent HT29 cells may thus be due to the absence of this recruitment phenomenon. There may, however, be other explanations. For example, although it is likely that bumetanide binds to the cotransporter itself (probably at a low affinity chloride site; Refs. 5 and 27), it does not necessarily follow that specific bumetanide binding is, in fact, a reflection of the total number of membrane cotransporters; one could speculate that the appearance of new bumetanide binding sites reflects not the recruitment of new cotransporters into the membrane but instead an activation of dormant cotransporters already present, the "active" conformation of the cotransporter being capable of binding bumetanide. If true, however, it does not explain why phalloidin loading would affect cotransporter activity but not the increase in bumetanide binding. One difficulty inherent in bumetanide binding studies is that equilibrium binding requires a 30-40-min incubation with this ligand, making it difficult to correlate precisely the rapid increases in cotransporter activity evident by ion tracer methodology with enhanced specific bumetanide binding. However, the absence of CAMP-induced increases in bumetanide binding in parent HT29 cells despite a 2-fold increase in bumetanide-sensitive Rb uptake suggests that cotransporter activity and number are not necessarily tightly correlated. Slotki (30) recently reported that long term stimulation of HT29 cells by forskolin did not result in an increase in specific bumetanide binding, although cotransporter activity remained increased; the short term effects of forskolin on binding were not addressed in that study.
It has been speculated that the actin-based cytoskeleton may serve to transduce regulatory signals for the Na-K-2C1 cotransporter (5); however, until recently, direct evidence for this proposal has been lacking. In addition to our recent demonstration of the microfilament dependence of Na-K-2C1 cotransporter activity in the T84 and now in the C1.19A cell lines, Jessen and Hoffman (31) reported that disruption of cellular actin with cytochalasin B activates the cotransporter in Ehrlich ascites cells. More recently, activation of Na-K-2C1 cotransport in endothelial cells has been shown t o correlate with phosphorylation of myosin light chain (32). Because of the paradigm that correlates activation of cotransport with numbers of bumetanide binding sites, initially we favored a mechanism in which additional cotransporter units were recruited into the plasma membrane by a microfilament-dependent event, analogous to the insertion of the vasopressin-regulated water channel in toad bladder and mammalian collecting duct (a process that has been shown to be augmented by cytochalasins and inhib-ited by phallotoxins) (33)(34)(35)(36). However, the present study indicates that phalloidin does not affect the CAMP-mediated increase in bumetanide binding in C1.19A cells despite inhibiting bumetanide-sensitive Rb uptake. Therefore, phalloidin appears to inhibit ion translocation by the cotransporter but not the number of active membrane cotransporters. F-actin thus appears to influence the transport characteristics of the cotransporter itself, and this effect is particularly apparent in the C1.19A cell line, which displays the chloride efflux-dependent component of CAMP-mediated regulation.
Whether actin filaments directly associate with the cotransporter and influence its function cannot be determined from the present study. The cytoplasmic aspect of the plasma membrane of all eukaryotic cells is lined by a cortical meshwork consisting not only of actin filaments but also a number of actin-associated proteins. Dynamic changes in submembranous actin polymerization may thus alter the kinetics and equilibria of a number of biochemical reactions involving actin-associated proteins and various membrane proteins linked to the cytoskeleton. Several examples of ion transport proteins that are linked to the actin-based cytoskeleton have now been identified. For example, the Na-K ATPase of renal epithelia has been shown to bind ankyrin and fodrin, thereby forming a metabolically stable complex with actin (37). This association appears to be of importance not only in the maintenance of the polarized distribution of the Na-K ATPase but also in modulation of pump function (38,39). A direct regulatory effect of short actin filaments on the open probability of a sodium channel in the A6 renal epithelial cell line has recently been demonstrated (40,411, and melanoma cells devoid of actin-binding protein have been observed to be unable to volume-regulate or activate potassium channels in response to a hypoosmotic stimulus (42). While the present study provides additional evidence that the cortical cytoskeleton can exert a regulatory influence over the function of the Na-K-2C1 cotransporter, further investigation is needed to define the structural basis for this interaction.