Serum Regulates Na'/H' Exchange in Caco-2 Cells by a Mechanism Which Is Dependent on F-actin*

Regulation of Na+/H+ exchange by fetal bovine serum was studied in Caco-2 cells, an established cell line derived from a human colon carcinoma. Cells were grown as polarized monolayers on collagen-coated filters and intracellular pH measured fluorometrically with 2',7'-bis(2-carboxymethyl)-5,6-carboxyfluores-cein. Na+/H+ exchange was reduced 64% when cells were deprived of serum for 4 h. In contrast to other cell types, readdition of serum for 10 min did not activate Na+/H+ exchange; however, readdition of serum for 4 h restored Na+/H+ exchange to control values. This long-term effect of serum on Na+/H+ exchange activity could not be explained by changes in intracellular buffering capacity or intracellular [Na+]. 4-h serum deprivation reduced the K, of the exchanger for external Na+ from 21 to 6 mM, and reduced the V,,, by 67%, but did not alter the ICso for amiloride in the presence of 140 mM Na'. Inhibition of protein synthesis with cycloheximide (5 MM) did not alter the effect of removal or readdition on Na+/H+ exchange. "C) vented exchange was at 37 OC, maintaining cells at 13

K, of the exchanger for external Na+ from 2 1 to 6 mM, and reduced the V,,, by 67%, but did not alter the ICso for amiloride in the presence of 140 mM Na' .
Inhibition of protein synthesis with cycloheximide (5 MM) did not alter the effect of serum removal or readdition on Na+/H+ exchange. Low temperature (13 "C) completely prevented the inhibition of Na+/H+ exchange caused by the removal of serum. In addition, once Na+/H+ exchange was inhibited by serum removal at 37 OC, maintaining cells at 13 "C also blocked the recovery of Na+/H+ exchange caused by serum readdition. Conversely, cytochalasin D (0.1-20 PM) blocked the reduction of Na+/ H+ exchange which occurred due to 4-h serum deprivation, but did not block the restoration of Na+/H+ exchange when the cells were re-exposed to serum for a further 4 h. Colchicine (20 PM) did not alter the effect of serum removal or readdition. These data suggest that serum regulates Na+/H+ exchange activity by a posttranslational mechanism which is dependent on Factin.
The regulation of Na+/H+ exchange has been studied extensively in a wide variety of epithelial and nonepithelial cells. Broadly, two patterns of regulation have emerged (1,2). First, * 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. activation or inactivation of Na+/H+ exchange can be rapid in onset (3)(4)(5)(6)(7)(8). This rapid regulation is believed to be mediated by phosphorylation of the Na+/H+ exchanger or a closely associated molecule (9). Activation of the transporter either by serum, a-thrombin, epidermal growth factor, or osmotic shrinkage (4-7) is thought to shift the intracellular pH sensitivity of the transporter to more alkaline values and decrease the affinity for internal Na+ (without changes in the affinity of the exchanger for external Naf or H+ or change in the V,,,,,) (1,2,10). A second class of regulation requires hours or days to develop. Although less studied, this type of regulation occurs in the kidney proximal tubule in response to thyroxine, glucocorticoids, or chronic acidosis (11)(12)(13). In the cases studied, there is a change in the transport Vmax but no change in the affinity for external Na+, the Hill coefficient, or the intracellular pH sensitivity (2).
In epithelial cells, studies of the regulation of Na+/H+ exchange can be complicated by the presence of multiple Na+/ H+ exchange proteins. In some epithelial cells (notably ileal villus cells and the pig kidney cell line PKE2O) separate apical and basolateral Na+/H+ exchange have been identified (14)(15)(16). Apical and basolateral Na+/H+ exchangers can be different in amiloride inhibition kinetics and regulation by second messengers (15)(16)(17). These differences are consistent with the different physiologic functions which are likely to be mediated by apical exchangers (transepithelial Na+ absorption) uersus basolateral exchangers (cell pH homeostasis).
Previously, we have characterized Na+/H+ exchange in Caco-2 cells as a potential model for regulation of the human intestinal basolateral exchanger (18). These cells have a basolateral membrane Na+/H+ exchanger which is highly sensitive to amiloride (Ki 3.2 PM), and which has a cytoplasmic H+ modifier site. The transporter is not regulated by addition of phorbol dibutyrate or changes in intracellular cyclic AMP, cyclic GMP, Ca2+, or cell volume (18). These properties can be contrasted with those of intestinal apical membrane Na+/ H+ exchange which is less sensitive to amiloride ( K ,  PM) (19,20) and is inhibited by Ca2+, Ca2+/calmodulin kinase 11, and protein kinase C (21-23). In the present study, the long-term regulation of Caco-2 Na+/H+ exchange by fetal bovine serum (FBS)' is described. We observe that FBS deprivation reversibly decreases the V, , , of Na+/H+ exchange by a process which requires hours to occur and is dependent on actin polymerization. This regulation of Na+/H+ exchange in Caco-2 cells by serum is distinct from the rapid effects of serum and growth factors observed in other cell types (1,4,5).
Intracellular pH Measurement and Calibration-Intracellular pH (pH,) was measured with BCECF as described previously (18,24). Briefly, cells on filters were loaded with BCECF by exposure for 60-90 min to 6.25 p~ of the acetoxymethyl ester at room temperature in Na' medium. Cells to be studied in the microscope were mounted and perfused as described previously (18,27). For study in a standard fluorometer, cells were washed three times in TMA medium to remove extracellular dye and then mounted at 45" in a glass cuvette. In experiments where fetal bovine serum (FBS) was removed, the cells were washed in Hanks' medium (GIBCO) three times and incubated at 37 "C in 5% co2/95% air in culture medium from which FBS had been omitted. Control cells were only removed from FBS for 1 h during the dye-loading procedure. In some fluorometer experiments, a perfusion system was used to facilitate changes in extracellular medium. The flow rate was 5 cuvette volumes per min (12.5 ml/min). As described previously (18), BCECF fluorescence was measured by dual excitation (500/440 nm) ratioing in an SLM spectrofluorometer (SPF-500C, SLM, Urbana, IL) equipped with a stirred cuvette, thermostated at 37 "C. The protocol used to calibrate intracellular BCECF is based on the method of Thomas et al. (18,25) employing nigericin. A calibration curve of intracellular pH was constructed by equilibrating cells in pH-clamp medium with 10 p~ nigericin, and titrating pH with additions of HNO,. Individual experiments were calibrated as follows. In the case of perfused experiments, the cells were equilibrated in pH-clamp medium (titrated to pH 7.30) plus 10 p~ nigericin and a single point calibration used to normalize data to the calibration curve. For nonperfused experiments in the fluorometer, the dye was released with 100 p M digitonin and single point calibration used to normalize data to the calibration curve. We have previously shown that these two calibration methods are equivalent in Caco-2 cells (18).
Measurement of Cellular Buffering Capacity for Hydrogen Ions-Buffering capacity (&) was determined by exposure of the cells loaded with BCECF to NH,CI as described previously (18). Briefly, cell pH, was first set by prepulses of 30 mM NH,CI to various pH; values. These acidified cells were exposed to TMA medium, and subsequently 1 mM NH,CI was added to the medium to induce changes in pH, of 0.26 f 0.03 pH units (n = 32). Buffering capacity was calculated using the medium pH and the pH; before and after addition of NH,Cl, according to the formula; b8 = A[H+];/ApH;. Measurement of Na+/H+ Exchange-Na+/H+ exchange activity was measured after the cells had been acidified by transient exposure to 30 mM NH,Cl. Acidified cells in a Na+-free (TMA medium) had a stable pH;. When the TMA medium was replaced with Na+ medium there was a prompt alkalinization (pH, recovery) due to the basolat-era1 Na+/H+ exchanger (26,27). Na+/H+ exchange rates were determined by multiplication of the initial rate of alkalinization over approximately 30-40 s (ApHJmin) by the buffering capacity of the cells at the initial pH, from which the alkalinization started.
Cell Zon Content and Cell Volume Measurements-Total cell Na+ and K+ of cell monolayers were determined by flame photometry as described previously (26,27) and expressed as nanomoles per mg of protein. Cells were suspended by exposure to 0.1% trypsin in Hanks' medium (GIBCO). Cell volume was measured in a Coulter Counter (model ZM) coupled to a computerized pulse-height analyzer (The Nucleus, Oakridge, TN), which was calibrated using latex beads (Coulter) of defined size (28). The median single cell volume of suspended cells was determined automatically by computer.

Measurement of rHlLeucine Uptake and Incorporation into Pro-
tein-Caco-2 cells grown in 24-well plates were incubated with 500 pl of growth media containing 1 pCi/ml [3H]leucine (Du Pont-New England Nuclear Research Products, MA) plus varying concentrations of cycloheximide. After the incubation period, the medium was aspirated and the cells washed three times with Hanks' medium and then incubated with 500 pl of 10% trichloroacetic acid for 1 h at 4 "C. The trichloroacetic acid was then aspirated and cells incubated in 2 M NaOH overnight at 23 "C. The NaOH was collected, neutralized with concentrated HCI, and counted in a scintillation counter.
Statistical Analysis-Where applicable the data are presented as the mean & S.E. of the mean unless indicated. Comparison of means was performed by Student's unpaired t test. A probability of <0.05 was considered significant.

RESULTS
Effect of Serum Removal on pHi Recovery from an Acid Load-Caco-2 cells recover from an acid load by activation of a basolateral Na'/H+ exchanger (18). Previously, we have demonstrated a number of second messenger pathways do not rapidly alter Na'/H' exchange rates in Caco-2 cells despite the ability of these pathways to regulate C1-secretion (18). Since serum addition rapidly stimulates Na'/H' exchange in several other cell types (1,4,6 ) , the effect of serum addition was evaluated in Caco-2 cells. Prior to exposure to 10% FBS, cells were FBS-deprived for 4 h and allowed to reach a steadystate pHi after recovery from an NH4C1-induced acid load. Using fibroblasts under these experimental conditions as positive controls, a Na+-dependent and amiloride-sensitive alkalinization was consistently observed over a 10-min time course, indicative of Na+/H+ exchange activation (Fig. U).' In contrast, addition of serum to Caco-2 cells under these experimental conditions caused a slight acidification [potentially due to an increase in metabolic acid production (29)], but did not demonstrate an alkalinization above resting pHi (Fig. 1B). These results suggest that Caco-2 Na'/H' exchange is not rapidly activated by serum readdition.
Despite the lack of rapid stimulation of Na'/H' exchange by serum readdition, removal of FBS from Caco-2 cells for 4 h reduced the rate of pHi recovery from an acid load (Fig. 2). As this reduction in alkalinization rate could have been due to an increase in cellular buffering capacity without a change in transport by the Na'/H+ exchanger, the buffering capacity of FBS-starved cells was compared with the buffering capacity of normal cells. As shown in Fig. 3 Prior to the start of the truces, cells were acidified by a 20min pulse of 30 mM NH,Cl. Both truces start when cells were exposed to Na+ medium. The rise in pH; is due to Na+/H+ exchange (18). These truces are representative of 42 experiments. be required to explain the observed decrease in the alkalinization rate, the results imply that a decrease in Na+/ H+exchange activity is responsible for the reduced rate of pHi recovery.
The time course and reversibility of the FBS deprivation effects on Na+/H+ exchange were determined (Fig. 4, Table  I). Serum was removed from Caco-2 cells for 2, 4, or 24 h before Na+/H+ exchange rate was determined. Since the rate of Na+/H+ exchange is sensitive to pH;, data are presented as a scatter plot of individual measurements of Na+/H+ exchange rate uersus the starting pH; values (Fig. 4) and compiled data are categorized by the pH; range under study (Table I). 4-h FBS deprivation was sufficient to obtain a maximal inhibition of Na+/H+ exchange, as no further reduction in Na+/H+ exchange activity was obtained after 24-h FBS starvation (Fig. 4). FBS deprivation for 2 h inhibited Na+/H+ exchange, but produced a submaximal reduction in Na+/H+ exchanger rate (Table I).
After 4-h FBS deprivation, readdition of 10% FBS for 4 h was sufficient to restore Na+/H+ activity to control levels ( Fig. 4), but readdition of 10% FBS for 1 h caused only a partial recovery (data not shown). When Caco-2 cells were FBS-deprived for 24 h, full Na+/H+ activity could be restored by readdition of 10% FBS for a further 24 h (Fig. 4).  24 h (A). Buffering capacity was measured by acute addition of 1 mM NH4Cl after pH, had been set to different values with pulses of 30 mM NH4Cl. Data are presented as the mean buffering capacity in 0.10 pH unit ranges (6.61-6.70, 6.71-6.80, 6.81-6.90, 6.91-7.00) 5 S.E. Where there are no error burs, single measurements are presented. The line is a least-squares fit of the data points between pH; 6.51 and 7.23. This line was used to calculate H+ efflux rates. In cells which were FBS-deprived for 4 h, Na+/H+ exchange activity was reduced by an average of 64% over the entire experimental pHi range (6.61-7.10) with a similar percentage decrease at each pH; value (Table I). This suggests that the Na+/H+ exchanger in FBS-deprived cells has a similar pHi sensitivity compared to control cells. Consistent with this hypothesis, FBS-deprived cells attained a steady-state pHi value after recovery from an acid load (7.22 k 0.06; n = 6) which was not significantly different from the steady-state pH, obtained in control cells (7.27 f 0.01; n = 26). Furthermore, in both FBS-deprived and normal cells, removal of external Na' did not cause acidification of resting pH;, indi- Reduction of Na+/H+ exchange after 2-and 4-h FBS deprivation based on the degree of initial acidification For both control and 4-h FBS-deprived cells, the mean H+ efflux rates were calculated from the data in Fig. 4 for the pHi ranges shown. Na+/H+ exchange rates after 2-h FBS-deprivation are also presented. Data are presented as mean f S.E. of (n) experiments. Errors are not presented when n < 4.  However, as shown in Table 11, FBS deprivation for 24 h caused no increase in cellular Na' content. Direct inhibition of the Na+,K+-ATPase with ouabain shows that changes in Na' and K' content are detectable with this technique. FBS deprivation caused no change in total cellular protein on the 35-mm plates (data not shown), but caused a small (4%) reduction in cell volume (Table 11). When combined with the measured ion content of the cells, this change in volume predicts that intracellular ion concentration may have dropped by 10%. Thus the reduction in Na+/H' exchange activity cannot be accounted for by a decrease in the driving force, as this would have required an increase in cellular [Na']. Effect of FBS Deprivation on Na+/H+ Exchange Kinetics and Amiloride Sensitivity-Another potential explanation for the reduction in Na'/H+ exchange is that FBS deprivation might have altered the Na' activation kinetics of Na+/H' exchange. To address this possibility, the initial rate of Na+dependent H' efflux was measured at external Na' concentrations between 2 and 140 mM (actual Na+ concentrations were determined by flame photometry). These measurements were made at similar initial pH, values: 6.80 f 0.09 (n = 11) for control cells, and 6.66 * 0.10 ( n = 13) for FBS-deprived cells. To control for variability between preparations, a protocol was employed in which H+ efflux rates at low [Na+], were normalized to the efflux rate at 138 mM Na' in the same preparation as described previously (18, 27, 30). As shown the H+ efflux rate with 138 mM Na+. qualitatively in Fig. 5, A and B, and quantitatively in Fig. 5C, FBS removal changed the external Na+ activation kinetics of Na'/H+ exchange. The Kt (Na') was lower in FBS-deprived cells than control cells (6 versus 21 mM, respectively), but this decrease in Kt (Na+) cannot explain the observed reduction in Na'/H'exchange after serum deprivation. Note that the V,,, in absolute H' efflux units cannot be determined directly from the plot in Fig. 5C because the data is normalized to the alkalinization rate at 138 mM Na'. However, because the rate of Na+/H+ exchange at 138 mM Na' is known from Fig. 4, the V,,, in FBS-deprived cells can be calculated to be 43% of that in control cells (0.0013 pH units/s and 0.003 pH units/s, respectively). The reduction in V,,, is the kinetic observation which explains the effect of FBS deprivation.
The reduction in K,(Na+) when Caco-2 cells are deprived of FBS could be explained by the presence of multiple iso-forms of Na+/H+ exchanger molecules which alter their relative functional importance when cells are FBS-deprived. Since Na+/H+ exchangers with higher (16) or lower affinity (20) for amiloride have been identified, the amiloride ICso of Caco-2 Na+/H+ exchange was measured after FBS deprivation (18). As shown in Fig. 6, the ICso was 50 ~L M in FBS-deprived cells, similar to the value of 28 p~ previously measured in cells which had not been FBS-deprived (18). Based on these results there is no evidence to suggest the appearance of a new population of Na+/H+ exchangers with different sensitivity to amiloride.
In a further attempt to resolve heterogeneity of Na+/H+ exchange activity in Caco-2 cells, it was hypothesized that there could be different isoforms of Na+/H+ exchange in separate cell subpopulations after FBS deprivation, and/or that only some cells have reduced Na+/H+ exchange in response to FBS deprivation. To evaluate heterogeneity in cellular expression of Na+/H+ exchange after FBS deprivation, pH; recovery of FBS-deprived cells was determined using a microscope-based image analysis system, as described (18). A 34,000-pm2 field of Caco-2 cells (approximately 600 cells) was studied. Following NH&l acid loading, the mean pHi over the field was 6.84 with a coefficient of variation of 53%. After allowing partial pH; recovery (9-min exposure to 138 mM Na+) there was no significant change in the broadness of the population distribution (mean pHi = 7.04 with a coefficient of variation of 50%), implying that all cells recovered at the same rate. If cells had varying numbers (or types) of Na+/ H+ exchangers, it is predicted that the broadness of the pHi distribution should widen in the total population (31). The constancy in the coefficient of variation suggests that the Na+/H+ exchange activity is homogenous in FBS-deprived Caco-2 cells with no evidence of a subpopulation of cells with different transport rates.
Effect of Protein Synthesis Inhibition on the Regulation of Na+/H+ Exchange Activity by FBS-Cycloheximide was used to determine whether protein synthesis was required to observe the changes in Na+/H+ exchange caused by FBS deprivation or replacement. The efficacy of cycloheximide was determined by measuring the inhibition of [3H]leucine incorporation into protein (Fig. 7A). To determine the time course of cycloheximide action, [3H]leucine incorporation was measured after a 1.5-4-h exposure to 5 ~L M cycloheximide and compared with control cells which had not been exposed to cycloheximide. As shown in Fig. 7B, extrapolation of results from cycloheximide-treated cells back to an intersection with  the control data gives an estimate for the time of onset of cycloheximide action of 7 min. These data demonstrate that exposure of Caco-2 cells to 5 p~ cycloheximide rapidly inhibits protein synthesis by greater than 80%. Exposure of Caco-2 cells to 5 ~L M cycloheximide for 4 h in the presence of 10% FBS had no effect on Na+/H+ exchange ( Fig. 7C), nor did cycloheximide alter the buffering capacity of the cells (data not shown). Cycloheximide did not alter the reduction in Na+/H+ exchange activity which followed 4-h FBS deprivation, suggesting that the reduction in Na+/H+ exchange was not due to the synthesis of a protein inhibitor of Na+/H+ exchange. As shown in Fig. 7C, cycloheximide also did not inhibit the restoration of Na+/H+ exchange when Caco-2 cells are re-exposed to 10% FBS for 4 h. Similar results were obtained when 10 pg/ml actinomycin D was used (data not shown). These latter data suggest that the restoration of Na+/H+ exchange activity is not due to the synthesis of new proteins, including Na+/H+ exchanger molecules. The experiments with cycloheximide suggest that the effect of FBS on Na+/H+ exchange is at a posttranslational level.
Effect of Low Temperature on the Regulation of Na+/H+ Exchange by FBS-To test whether the effect of FBS requires intracellular transport of membrane proteins, the effect of 4h FBS deprivation was studied at 13 "C, a temperature known to slow intracellular membrane traffic (32)(33)(34). When cells were FBS-deprived at 13 "C for 4 h and then Na+/H+ exchange measured at 37 "C, the reduction in Na+/H+ exchange activity was prevented (Fig. 8). Furthermore, when the cells were 4-h FBS-deprived at 37 "C and then exposed to 10% FBS at 13 "C, Na+/H+ exchange was not restored but instead remained at levels found after 4-h FBS deprivation (Fig. 8). These data, together with the kinetic data that the Vmax of the exchanger is reduced, suggest that the number of functional exchanger molecules in the plasma membrane may be reduced after FBS deprivation.
Effect of Cytochalasin D and Colchicine o n the Regulation of Na+/H+ Exchange by FBS-To obtain further evidence that FBS deprivation might alter the number of functional Na+/ H+ exchanger molecules in the plasma membrane, the effects of cytochalasin D and colchicine were studied. Cytochalasin D is known to disrupt actin microfilaments (35)(36)(37) and previously has been used to inhibit plasma membrane cycling events (38, 39). As shown in Fig. 9, 20 PM cytochalasin D blocked the reduction in Na+/H+ exchange activity normally caused by 4-h FBS deprivation, suggesting that polymerized F-actin is required for this process. Since nonspecific effects of a high cytochalasin D concentration could also explain this observation, experiments were performed using lower concentrations of drug, and transport rates compared at a similar pHi (pH = 6.70 f .01, n = 16). Cells deprived of serum for 4 h in the presence of either 0.5 or 0.1 PM cytochalasin D had transport rates which were, respectively, 119 f 22% ( n = 5 ) and 88 f 11% ( n = 5) of the transport rates observed when cells were kept in FBS. Since FBS deprivation (without drug) was able to lower transport to 35 f 5% ( n = 3) of the same control in the same experimental series, it is clear that low concentrations of cytochalasin D block the effects of FBS deprivation.
In contrast, incubation of FBS-deprived cells with 10% FBS for 4 h in the presence of 20 PM cytochalasin D did not inhibit restoration of Na+/H+ exchange activity; in fact rates were slightly above control values (Fig. 9). The latter observation suggests a dynamic equilibrium between a cytochalasin D-sensitive process which reduces the number of Na+/H+ exchange proteins in the plasma membrane and a separate cytochalasin D-insensitive process which restores Na+/H+ exchange to the plasma membrane. To test whether such a dynamic equilibrium is present in cells which have not been FBS-deprived, Na+/H+ exchange activity was measured in . % 7 5 -

Intracellular pH
FIG. 8. Effect of 13 "C incubation on Na+/H+ exchange activity in FBS-deprived cells. H+-efflux rates were measured as described in the text and Fig. 4. All H+-efflux rates were measured at 37 'C. H+-efflux rates in control cells (W) which were not FBSdeprived are compared with cells which were FBS-deprived for 4 h at (A) 13 "C or 37 'C (0). H+-efflux rates were also determined in cells which had been FBS-deprived for 4 h at 37 'C then incubated in culture medium containing 10% FBS for a further 4 h at 13 "C (A) or 37 "C (0). Control data collected from cells maintained at 37 ' C are from the same data set as that shown in Fig. 4.

H+-efflux rates in cells which had not been FBS-deprived but had
been incubated with 10 rg/ml cycloheximide are also shown (0). Cytochalasin D was made up as a 10 mg/ml stock in dimethyl sulfoxide within 7 days of experimentation.
normal cells which had been incubated with 20 p~ cytochalasin D for 4 h. Under these conditions there was no change in Na+/H+ exchange activity (Fig. 9). The effect of 20 p~ colchicine (which disrupts microtubules) was also studied (40-43). Colchicine blocked neither the reduction of Na+/H+ exchange activity following FBS deprivation nor the restoration of Na+/H+ activity following 4-h FBS readdition (Fig. 9) suggesting that microtubules do not play a role in this regulation of Caco-2 Na+/H+ exchange.

DISCUSSION
This study demonstrates that FBS deprivation decreases Caco-2 Na+/H+ exchange, causing a 60% reduction in the V,,, of the exchanger and a reduction of Kt (Na+) through a mechanism that takes hours to manifest or reverse. This process is distinct from the commonly observed stimulatory effects of serum on Na+/H+ exchange in both time course and kinetic characteristics (1,(4)(5)(6). The effect of serum on Caco-2 cells most closely follows a pattern of slow regulation previously found in renal proximal tubular cells in response to thyroxine, glucocorticoids, and acidosis (12)(13)(14). The response to these hormones requires hours and has been characterized by changes in the V,,, without changes in the pHi sensitivity or the Kt for external Na+ (12)(13)(14).
The observed decrease in K,(Na+) in FBS-deprived cells could have been due to a change in the relative abundance of multiple subtypes of Na+/H+ exchanger proteins. Our studies provided no corroborating evidence for the presence of kinetically distinct isoforms of Na+/H+ exchangers in Caco-2 cells. Evidence suggests that in epithelial cells significant differences exist between apical and basolateral Na+/H+ exchangers with respect to amiloride sensitivity and pHi sensitivity (15,16,27,30). However, no changes in these parameters were induced by FBS deprivation. Similarly, there was no measurable heterogeneity of transport among individual cells in Caco-2 monolayers as there was no change in the population distribution of Na+/H+ exchange rates. Alternative explanations for the alteration in Kt (Na+) include a change in the function of remaining Na+/H+ transporters after FBS removal, perhaps due to changes in membrane lipid composition or cytoskeletal structure (see below).
The process affected by FBS deprivation is unknown. However, the data are most consistent with a model in which the number of functional exchange proteins in plasma membrane is reduced when the cells are FBS-deprived. This is based on three lines of evidence. First, the changes in activity are due to changes in the V,,, of the exchanger, without requiring protein synthesis of either new Na+/H+ exchange proteins or associated inhibitory molecules. Second, incubating the cells at 13 "C blocks the effects of FBS deprivation and FBS reintroduction (Fig. 8). Incubation of cells at low temperature (<20 "C) has been shown to inhibit a number of membrane fusion events including endocytosis (32), fusion of pinocytotic vesicles with lysosomes, and exocytosis (33,34). Third, the actin-disrupting agent, cytochalasin D (35)(36)(37), inhibits the reduction in Na+/H+ exchange activity after FBS deprivation, but does not affect restoration of transport after readdition of FBS. Since colchicine had no effect during either FBS deprivation or reintroduction, the data implicates a role for actin microfilaments but not microtubules in the FBS effect. Evidence suggests that F-actin also participates in the regulation of other membrane transport processes, including renal water channels, an epithelial Na+ channel, and an intestinal Na:K:2C1 transporter (43)(44)(45)(46)(47). Since cytochalasin D did not affect basal Na+/H+ exchange activity in control cells (Fig.  9), the data suggest that removal of FBS activates a cytochalasin D-sensitive process which reduces Na+/H+ exchange.
The data suggest that F-actin is required either to 1) modulate the activity of Na+/H+ exchange proteins which remain in the plasma membrane or 2) mediate endocytosis of Na+/H+ exchange proteins from the plasma membrane to an intracellular compartment. The observation that cytochalasin D does not alter the restoration process ( Fig. 9) argues against a modulation of function via reversible interaction of F-actin filaments with plasma membrane Na+/H+ exchange proteins or associated molecules. The differential sensitivity of Na+/ H+ exchange reduction and restoration to disruption of actin microfilaments is more consistent with an F-actin-dependent endocytosis and F-actin-independent reinsertion of Na+/H+ exchangers into the plasma membrane.