The effects of human cytomegalovirus (HCMV) infection on Cl−/ exchanger activity in human lung fibroblasts (MRC-5 cells) were studied using fluorescent, ion-sensitive dyes. The intracellular pH (pHi) of mock- and HCMV-infected cells bathed in a solution containing 5% CO2-25 mM were nearly the same. However, replacement of external Cl−with gluconate caused an H2DIDS-inhibitable (100 μM) increase in the pHi of HCMV-infected cells but not in mock-infected cells. Continuous exposure to hyperosmotic external media containing CO2/ caused the pHi of both cell types to increase. The pHi remained elevated in mock-infected cells. However, in HCMV-infected cells, the pHi peaked and then recovered toward control values. This pHirecovery phase was completely blocked by 100 μM H2DIDS. In the presence of CO2/ , there was an H2DIDS-sensitive component of net Cl− efflux (external Cl− was substituted with gluconate) that was less in mock- than in HCMV-infected cells. When nitrate was substituted for external Cl− (in the nominal absence of CO2/ ), the H2DIDS-sensitive net Cl− efflux was much greater from HCMV- than from mock-infected cells. In mock-infected cells, H2DIDS-sensitive, net Cl− efflux decreased as pHi increased, whereas for HCMV-infected cells, efflux increased as pHi increased. All these results are consistent with an HCMV-induced enhancement of Cl−/ exchanger activity.
- cell volume
- hydrogen 4,4′-diisothiocyanostilbene-2,2′-disulfonic acid
- sodium/hydrogen exchanger
cytomegaly, the enlargement of the host cell, is a major hallmark of human cytomegalovirus (HCMV) infection (see Fig. 1 in Ref. 1). Evidence suggests that cytomegaly and viral replication are closely linked (25), and both may be linked to enhanced Na+ entry into the host cell (13). We previously showed that 72 h after HCMV infection the activity of the Na+/H+exchanger (NHE) is enhanced in two ways (10):1) the intracellular pH (pHi) operating range for the NHE is shifted toward higher, or more alkaline, pHi values, and2) in the absence of CO2/ , HCMV infection makes the NHE much more responsive to a challenge with hyperosmotic solutions.
Many cells are able to homeostatically regulate their cell volume in response to shrinkage by use of a combination of increased NHE and Cl−/ exchanger activities. The increase of NHE activity due to cell shrinkage tends to increase pHi. In turn, the increase of pHistimulates Cl−/ exchanger activity. The enhanced Na+ uptake (via the NHE) and Cl− uptake (via the Cl−/ exchanger) result in an increased intracellular osmolyte content, which causes net water entry. We previously suggested that the pathological increase of host cell size (volume) caused by HCMV infection may be due, at least in part, to the physiologically inappropriate activation of such a pHi-linked mechanism (10). Our recent finding of enhanced NHE activity after HCMV infection is consistent with this hypothesis. The next obvious question is whether HCMV infection also enhances Cl−/ exchanger activity of the host cell.
The anion exchanger or Na+-independent Cl−/ exchanger (hereafter referred to as the Cl−/ exchanger) has several distinctive properties permitting its functional characterization. For instance, because it exchanges Cl− for , removal of extracellular Cl− will cause the transport mechanism to mediate a net exchange of intracellular Cl− for extracellular . This exchange will result in a net loss of intracellular Cl− as well as intracellular alkalinization. Another characteristic of the Cl−/ exchanger is that it is inhibited by disulfonic acid stilbene derivatives such as H2DIDS when they are presented to the extracellular face of the cell membrane (6). These agents also inhibit the external Na+-dependent Cl−/ exchanger (28). However, this latter exchanger is inhibited as pHi increases, whereas the Cl−/ exchanger is stimulated by alkaline pHi (4, 24, 26, 32). Finally, this exchanger can exchange Cl−for (3, 14, 22), and this exchange, unlike that of Cl−for , will have no pHi consequences, inasmuch as is the salt of a strong acid and, hence, is a very weak base. We exploited all these properties to investigate whether HCMV infection increased Cl−/ exchanger activity over the same time period it enhances NHE activity.
Cell Cultures and HCMV Infection
Details of cell culture and HCMV infection protocols have been presented previously (10). Briefly, a cell line (MRC-5) derived from human embryo lung fibroblasts, passages 18–30, was cultured in MEM with Earle’s salts, supplemented with 2 mM glutamine and 10% heat-inactivated FCS. The cells were grown in an incubator with a humidified atmosphere of 5% CO2-95% air at 37°C. A stock of HCMV (strain AD169) was generated in confluent MRC-5 cells (see Ref.2 for more details).
Three days after cells were seeded on 6 × 24-mm glass coverslips, confluent MRC-5 cells were exposed for 1 h to a suspension containing HCMV at a multiplicity of infection of approximately three plaque-forming units per cell or a mock-infecting, virus-free suspension (see Ref. 2 for details of mock infection). Two days after exposure to HCMV, the FCS was reduced to 1% (10).
Standard Solutions and Reagents
Standard HEPES-buffered solution contained (in mM) 128 NaCl, 5 KCl, 1 MgCl2, 1 CaCl2, 10 glucose, and 20 HEPES. The pH was adjusted to 7.4 (at 37°C) withN-methyl-d-glucamine, and the osmolality was 285 ± 5 mosmol/kgH2O. The standard CO2/ solution had the following composition (in mM): 123 NaCl, 25 NaHCO3, 5 KCl, 1 MgCl2, 1 CaCl2, and 10 glucose. The pH of the standard CO2/ solution was 7.4 when bubbled with 5% CO2-95% air. Cl−-free solutions were the same as the standard solutions, except NaCl was replaced with sodium gluconate or NaNO3. Hyperosmotic solutions were made by addition of 78 mM NaCl to the standard CO2/ solution; this increased the osmolality to ∼415 mosmol/kgH2O (146% of normal osmolality).
The high-K+ HEPES solution used in the studies of pH dependence of Cl− efflux contained (in mM) 20 sodium gluconate, 20 potassium gluconate, 100 KNO3, 1 magnesium gluconate, 3 calcium gluconate, 10 glucose, and 20 HEPES, pH 7.4. Diethyl amiloride (DEA; Molecular Probes, Corvallis, OR), a 5-amino-substituted derivative of amiloride (19), was prepared as a 10 mM stock solution in distilled water and used at a final concentration of 5 μM. This concentration was sufficient to block virtually all NHE activity (10). H2DIDS (Molecular Probes) was added directly to the saline solution at a final concentration of 100 μM. Tributyltin (Fluka, Milwaukee, WI) and nigericin (Sigma Chemical, St. Louis, MO) were made as 50 mM stock solutions in ethanol and added directly to the saline solutions just before use to obtain a final concentration of 10 μM. Valinomycin (Sigma Chemical) was also prepared as an ethanol stock solution (9 mM) and added to the saline solution to obtain a final concentration of 5 μM.
Our methods for measuring pHi with the pH-sensitive fluorescent probe 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF; Molecular Probes) are described in detail elsewhere (10). Briefly, cell populations were loaded with BCECF by exposure to 5 μM BCECF-AM, the permeant ester, for 10–30 min. An SLM-Aminco spectrofluorometer (model DMX-1000) was used to measure the dye fluorescence. The pH is proportional to the ratio of light emitted at 535 nm when BCECF is excited at two wavelengths (450 and 495 nm). The dye was calibrated intracellularly by use of the high-K+-nigericin technique (29). The pH calibration solution contained 130 mM potassium gluconate, 20 mMN-methyl-d-glucamine chloride, 2 mM MgCl2, 20 mM HEPES, and 10 μM nigericin. All pHiexperiments were conducted at 37°C.
Intracellular Cl− Concentration Measurements With Use of N-(6-Methoxyquinolyl)acetoethyl Ester
The fluorescent dyeN-(6-methoxyquinolyl)acetoethyl ester (MQAE; Molecular Probes) was used according to a method developed by Verkman et al. (30) and refined by Koncz and Daugirdus (20). When excited at 365 nm, MQAE emits light at 450 nm. The emitted light intensity is inversely related to the Cl− concentration ([Cl−]) of the MQAE-containing solution, i.e., its output is quenched by Cl− but unaffected by or (30). Cells were loaded with MQAE by bathing in a 10 mM solution of MQAE (dissolved in MEM with 0% FCS) in the incubator (37°C) for 2–3 h.
All MQAE experiments were performed in a spectrofluorometer at room temperature to minimize the rate of dye loss from the cells. In addition, the fluorescence at an intracellular [Cl−] ([Cl−]i) of 0 mM (F0) was determined in every experiment by use of the double-ionophore technique. This involves bathing the cells with a Cl−-free solution ( replaced Cl−) that contained 10 μM tributyltin (a Cl−/OH−exchanger) and 10 μM nigericin (7, 20). At the end of every experiment, fluorescence readings were corrected for background by use of an SCN−-containing solution to maximally quench the MQAE ion-sensitive signal (Fig.1 A). This solution contained 150 mM K+, 150 mM SCN−, 20 mM HEPES, pH 7.4, and 5 μM valinomycin (20). [Cl−]iis directly proportional to the ratio F0/F (where F is the fluorescence reading corrected for background; see above) according to the following equation whereK SV is the Stern-Volmer constant.
K SV was experimentally determined for mock- and HCMV-infected cells (Fig.1 B). The double-ionophore technique (see above) was used to completely deplete the cells of Cl−, and the fluorescence intensity was noted as F0. Then, in the continuous presence of the ionophores, the cells were exposed to an external solution with a known [Cl−], and the assumption was made that, in the presence of the ionophores, extracellular [Cl−] ([Cl−]o) was equal to [Cl−]i. Finally, the cells were exposed to the KSCN solution to obtain the maximally quenched fluorescent signal (Fig.1 A). This procedure was repeated over a range of [Cl−]ovalues, with a new coverslip of cells used for each [Cl−]o. The [Cl−]icalibration solutions were made by mixing high-K+ HEPES solution with Cl− or as the anion (seeStandard Solutions and Reagents). After correction for dye leakage (Fig. 1 A, bottom trace), F0/F values at each [Cl−]iwere determined. The resultant F0/F values are plotted in Fig.1 B, andK SV was 25.7 and 19.7 M−1 for the mock- and HCMV-infected cells, respectively. As noted in the legend of Fig.1 B, the fits of the slopes of the lines used to determine theK SV indicate that, with 95% confidence, the two cell treatments indeed have different values. The basis for the difference is unknown, but given the profound effects of the virus on the host cell, it is not surprising that such a difference might exist. If, however, there is no difference between theK SV values for mock- and HCMV-infected cells, the apparent increase of [Cl−]icaused by HCMV infection would be 30% less than we report here. Both values of K SV are in reasonable agreement with those reported by other workers for other cell types (20). The efflux rate constants were calculated from the experimental data with use of a monoexponential function and with the assumption that [Cl−]iasymtotically approached 0 mM.
To vary pHi over the range 6.4–7.8, we used the high-K+-nigericin technique (29). Briefly, we exposed cells to a nigericin solution similar to that described above for calibration of the pHi indicator BCECF, except all the Cl− was replaced by . This solution ensures that pHi is equal to extracellular pH (pHo) and effectively “clamps” the pHi while permitting the intracellular Cl− to exchange with extracellular . Using the MQAE technique described above, we measured the rate constant for the decrease of [Cl−]i. In a separate series of studies (not shown), we found that pHi ≅ pHo within 2–3 min of exposure to the nigericin solution (cf. Fig. 4 in Ref. 2). Therefore, using data taken between 5 and 25 min after exposure to the nigericin “clamping” solution, we calculated the rate constants for Cl− loss.
All pooled data are means ± SE. Paired and unpairedt-tests were used, as appropriate, to test for statistical significance.
The present investigation studied two facets of Cl−/ exchange: its role in the determination of pHi and its role in the determination of [Cl−]i.
Effect of CO2/ on Resting pHi in Mockand HCMV-Infected Cells
We previously measured the resting pHi of mock- and HCMV-infected cells (72 h after exposure to HCMV) when they had been equilibrated in standard HEPES solution (10). Under the conditions of CO2/ -free external fluid, we observed a significantly more alkaline resting pHi in the HCMV- than in the mock-infected cells (7.49 vs. 7.40) (10). Figure2 shows what happens to pHi for mock- and HCMV-infected cells when the external fluid is changed from standard HEPES-buffered solution to standard CO2/ buffered solution. The pHi of mock- and HCMV-infected cells initially acidified as the CO2 entered and was hydrated. After the initial acidification, the mock-infected cells increased their pHi significantly above that measured when they were bathed in the HEPES-buffered solution (Fig. 2). The initial acidification was more pronounced for the HCMV- than for the mock-infected cells. However, after this initial acidification, the pHi of the HCMV-infected cells also increased, to a value very close to that attained by the mock-infected cells (Fig. 2). Thus CO2/ exposure results in net acid loading in the HCMV-infected cells and acid extrusion in mock-infected cells. As the results from the following experiments will show, it is highly likely that at least one of the CO2/ -dependent acid loaders in HCMV-infected cells is the base-extruding Cl−/ exchanger.
Effect of External Cl− Removal on pHi of Cells Exposed to CO2/
Removal of extracellular Cl−in the presence of CO2/ would be expected to result in a net cellular uptake of (in exchange for the net cellular loss of Cl−) in cells containing functional Cl−/ exchangers (i.e., “reverse” Cl−/ exchange). The increase of intracellular concentration ([ ]i) would manifest itself by an increase of pHi. Figure3 shows the pooled results obtained from a number of experiments on mock- and HCMV-infected cells. When the composition of the external solution was changed from standard (Cl−-containing) CO2/ solution to a Cl−-free (gluconate) CO2/ solution, the mock-infected cells responded with an acidification of ∼0.10 pH unit over 5 min. We observed similar acidification when external Cl− was replaced in the absence of CO2/ (unpublished observations). The pHi effect of external Cl− removal was fully reversible on return of the Cl− (not shown). The basis of the acidification in the mock-infected cells is unknown but has been noted by other workers (5, 15, 31). Possible causes that have been suggested include anion effects on the activity of a vacuolar H+-ATPase (31) and novel properties of the multidrug resistance protein (15).
In contrast, the HCMV-infected cells responded by alkalinizing ∼0.3 pH unit within 5 min of external Cl− removal. This alkalinization could be prevented by pretreating the cells with 100 μM H2DIDS (data not shown), which suggests that this pHichange might be the result of net uptake in exchange for intracellular Cl− mediated by the Cl−/ exchanger. We noted that after reaching a maximum the pHi of HCMV-infected cells decreased. Whether this secondary acidification is related to the acidification observed in mock-infected cells is unknown. On return of Cl− to the external solution, pHi returned to pretreatment values (not shown). Characterization of the Cl−-dependent acidification processes in mock- and HCMV-infected cells awaits further investigation.
Effect of CO2/ on the pHi Response to a Hyperosmotic Challenge
We previously reported that, in the absence of CO2/ , treatment with hyperosmotic external fluids causes HCMV-infected cells (and, to a much lesser degree, mock-infected cells) to exhibit a significant DEA-sensitive alkalinization of pHi. We interpreted this observation to mean that the volume sensitivity of the NHE of the HCMV-infected cells was more enhanced than that of the mock-infected cells under this condition. In the present study we investigated the response to hyperosmotic treatment under the more physiological situation when CO2 and are also present.
Figure 4 shows the results of increasing the osmolality of the standard CO2/ bathing solution to 146% of the normal, control osmolality. For the mock- and HCMV-infected cells, hyperosmotic challenge resulted in an alkalinization of pHi. This alkalinization appeared to have two components: an initial, small, but relatively rapid, DEA- and H2DIDS-insensitive alkalinization and a somewhat slower, but much larger, DEA-sensitive alkalinization. The DEA-insensitive alkalinization was not observed in the absence of CO2/ ; in fact, a slight acidification was observed (10). The basis of this rapid, CO2/ -dependent, DEA-, H2DIDS-insensitive alkalinization is uncertain. However, consider that a reduction of cell volume would increase [ ]ias well as intracellular . Because CO2 will rapidly reequilibrate across the cell membrane, the net effect of the cell volume reduction would be an increase of [ ]irelative to intracellular , which would translate into a rise of pHi.
Although mock- and HCMV-infected cells exhibited a secondary, DEA-sensitive alkalinization, the peak pHi reached by the mock-infected cells was somewhat greater than that reached by the HCMV-infected cells (∼0.2 vs. 0.14 pH unit; not significant). In addition, Fig. 4 shows that mock- and HCMV-infected cells reached this peak pHi ∼6 min after the onset of the hyperosmotic challenge. After that peak, however, there were significant differences between the pHi behaviors of the two cell treatments in the continued presence of the hyperosmotic challenge. The pHi of mock-infected cells changed very little over the next 8 min (Fig. 4). However, the pHi of the HCMV-infected cells decreased ∼0.1 pH unit (Fig. 4) over this same period. This “recovery” of pHi in the HCMV-infected cells was statistically significant (P < 0.0001, pairedt-test) and was abolished by treatment with 100 μM H2DIDS (Fig. 4, 6 min). In addition, treatment with H2DIDS caused the peak pHi to be significantly more alkaline for HCMV-infected cells (P< 0.05, unpaired t-test). As noted above, DEA treatment prevented the slow, secondary phase of alkalinization for both mock- and HCMV-infected cells.
Effect of CO2/ on [Cl−]i in Mockand HCMV-Infected Cells
We used MQAE to measure the steady-state [Cl−]iin the presence and absence of CO2/ . Table 1 summarizes these results for mock- and HCMV-infected cells. Two important points emerge from these data. First, we found a significantly higher [Cl−]iin HCMV- than in mock-infected cells. This was true without regard to whether the fluid bathing the cells contained CO2/ . The [Cl−]iof the HCMV-infected cells bathed with HEPES standard solution (nominally 0 mM ; calculated [ ] with the assumption that solution is in equilibrium with room air = 0.13 mM) was 48% higher than that of the mock-infected cells identically treated and was 63% higher when the cells were bathed with the CO2/ standard solution (25 mM ). Thus HCMV infection acted to raise [Cl−]iby -independent as well as dependent mechanisms. Second, bathing the cells in the standard CO2/ solution increased the [Cl−]iof HCMV-infected cells more than mock-infected cells. The [Cl−]iincrease for HCMV-infected cells caused by bathing them in CO2/ -containing fluids was ∼17 mM, whereas the same treatment increased the [Cl−]iof mock-infected cells ∼5 mM (Table 1).
Effect of HCMV Infection on Net Cl−Efflux in CO2/ -Free Media
In this series of studies we measured net Cl− efflux caused by replacement of extracellular Cl− with gluconate. Gluconate, being a large organic anion, has a limited ability to mediate the anion exchange (17). Therefore, in the absence of , the net Cl− loss observed on the replacement of external Cl− with this anion is assumed to be mediated via mechanisms other than Cl−/ exchange (e.g., Cl−channels, Na+-K+-Cl−cotransport, Na+-Cl−cotransport, K+-Cl−cotransport).
Cells were bathed with a nominally CO2/ -free medium for at least 15 min before removal of extracellular Cl−. This ensures that pHi is at its new steady state and that [ ]i≅ 0 mM. Figure 5 shows that, under these conditions, when extracellular Cl− is replaced (with gluconate), the [Cl−]iof HCMV-infected cells decreased much more slowly than that of mock-infected cells. The data points between 2 and 20 min after the Cl− replacement were fitted with a single-exponential function with the assumption that the [Cl−]iasymptotically approaches 0 mM. The resultant calculated rate constants are summarized in Table 2. We previously showed a 1.44 times greater cell volume-to-plasma membrane surface area ratio of HCMV- than mock-infected cells (10). To meaningfully compare the relative rates of Cl−movement of the mock- and HCMV-infected cells, this morphological effect needs to be taken into consideration. To facilitate this comparison of the relative rates of Cl− movement, we have multiplied the experimentally determined rate constants in Table 2 by a normalized volume-to-surface area ratio. This ratio is defined as 1 for the mock-infected cells and is therefore equal to 1.44 for HCMV-infected cells 72 h after exposure (10). After applying this correction (Table 2), we see that 1) in the absence of CO2/ , Cl− is lost about twice as fast from mock-infected cells as from HCMV-infected cells and2) in the absence of CO2/ , treatment with 100 μM H2DIDS does not reduce the rate of Cl− loss by either cell treatment (in fact, some stimulation was noted).
Effect of HCMV Infection on Net Cl−Efflux in CO2/ -Containing Media
We have already shown that when extracellular Cl− was replaced by gluconate in the presence of , there is a significant increase in pHi of HCMV-infected cells, whereas the pHi of mock-infected cells was acidified (Fig. 3). This pHi alkalinization was blocked by H2DIDS, suggesting that it was mediated by Cl−/ exchange, e.g., a net uptake of extracellular in exchange for intracellular Cl−. Thus the operation of such a mechanism would be expected to result in an increased rate of intracellular Cl− loss (relative to that observed in the absence of CO2/ ), and the increased rate of Cl− loss should be H2DIDS sensitive. We therefore measured net Cl− losses from mock- and HCMV-infected cells (72 h after exposure) in the presence of CO2/ . Figure 6 shows examples of the effect of replacing external Cl− by gluconate on the rate of change of [Cl−]iin mock- and HCMV-infected cells. On average, treatment with 100 μM H2DIDS reduced the rate constant of net Cl− efflux for mock- and HCMV-infected cells, but the effect was much more pronounced in the HCMV-infected cells (cf. Table 2). The collated data in Table 2 show an indexed rate of Cl− loss in mock-infected cells bathed in standard HEPES solution that is only 70% of that in cells bathed in CO2/ solution. On the other hand, the same solution change decreased the indexed rate for Cl− loss to 25% in HCMV-infected cells. Thus the -stimulated increase in the net rate of Cl− efflux (corrected for the change in surface-to-volume ratio) was ∼3.3-fold greater in the HCMV- than in the mock-infected cells treated identically (0.142 vs. 0.043). These data clearly show that, in the presence of CO2 and , there is an H2DIDS-sensitive Cl− efflux pathway for mock- and HCMV-infected cells. However, this pathway is minimal in mock-infected cells but is greatly increased after exposure to the virus. These results are in good qualitative agreement with the results in Fig. 3, which shows that in HCMV- but not mock-infected cells the pHi alkalinizes when external Cl− is removed.
H2DIDS-Sensitive Net Cl− Efflux Stimulated by External
Further evidence was sought that the H2DIDS-sensitive Cl− efflux pathway represents Cl−/ exchange by examining the effects of replacing extracellular Cl− with . has been reported to serve as an excellent exchange partner for the Cl−/ exchanger (3, 14, 17, 22). Moreover, because it is a very weak base, its entry is expected to cause little change of pHi. We therefore tested the effect of extracellular on net Cl− efflux. Figure7 A shows the effect on the [Cl−]iof mock-infected cells caused by replacing extracellular Cl− with . In eight such experiments the rate constant for Cl− loss from the mock-infected cells was 0.161 ± 0.011 min−1. Treatment with 100 μM H2DIDS reduced the rate constant for Cl− loss by 17%, to 0.134 ± 0.021 min−1.
Figure 7 B, in contrast, shows the effect of the same protocols on HCMV-infected cells 72 h after infection. In six such experiments, replacement resulted in a loss of intracellular Cl− with a rate constant of 0.182 ± 0.013 min−1. Treatment with 100 μM H2DIDS reduced the rate constant by 54%, to 0.083 ± 0.004 min−1. In three experiments we examined the effects of replacement on cells infected 24 h before their assay. In this case the control rate constant was 0.152 ± 0.005 min−1, and after H2DIDS treatment it decreased 35%, to 0.099 ± 0.017 min−1. The fact that the net Cl− efflux caused by substitution was much greater in the HCMV- than in the mock-infected cells and that Cl− loss in the HCMV-infected cells was largely prevented by H2DIDS treatment suggests that exchanges with Cl− via the Cl−/ exchange mechanism and that this mechanism is much more active in the HCMV- than in the mock-infected cells.
Figure 8 compares the H2DIDS-sensitive rate constants for Cl− loss in exchange for in mock- and HCMV-infected (24 and 72 h after exposure) cells. It indicates that HCMV induces an infection time-dependent increase in the rate constants for H2DIDS-sensitive Cl− efflux. Thus, within 24 h of exposure to the virus, the activity of the H2DIDS-sensitive pathway, presumably the Cl−/ exchanger, is almost doubled, and it increases further between 24 and 72 h after exposure. Application of the normalized volume-to-surface area ratio correction for the 72-h-postexposure cells shows that HCMV infection increased the index for the rate of Cl− loss for H2DIDS-sensitive Cl− by ∼5.3 fold, from ∼0.027 to 0.143.
pHi Sensitivity of H2DIDS Net Cl− Efflux
Cl−/ exchange in a number of preparations has a characteristic pHi sensitivity, being progressively stimulated by increases in pHi between ∼6.0 and 7.8 (24,26, 32). To test whether the H2DIDS-sensitive Cl− efflux pathways identified above are stimulated over the same pHi range, we used the high-K+-nigericin technique (29), as described in methods, to establish steady-state pHi values between 6.4 and 7.8. Because exposure of such “pHi-clamped” cells to CO2/ containing media would cause rapid changes in pHi, we had to find another means to activate net Cl− efflux via the Cl−/ exchanger to examine its pHisensitivity. , being the conjugate base of a strong acid, does not cause pHi changes (see above), and it appears to be transported by the Cl−/ exchanger of MRC-5 cells (Fig. 7).
Figure 9 shows the results of varying the pHi (accomplished by varying pHo; seemethods) on the rate constant of net Cl− efflux caused by replacing external Cl− with . In the case of the mock-infected cells, the H2DIDS-sensitive component of the net Cl− efflux decreased as pHi was increased. However, in HCMV-infected cells the H2DIDS-sensitive component was more than doubled as pHi was increased from 6.4 to 7.8. Thus Cl−/ exchange in HCMV-infected cells displays a pHi sensitivity suggestive of its being mediated by a Cl−/ exchanger.
HCMV Infection Causes Increased Cl−/ Exchanger Activity
We have obtained several results that lead us to the conclusion that HCMV infection results in a significant enhancement of Cl−/ exchanger activity. Removal of external Cl− creates a favorable thermodynamic gradient for the outward net movement of cellular Cl−. If this outward net movement is obligatorily linked to an inward net movement of via the Cl−/ exchanger, an H2DIDS-sensitive intracellular alkalinization will result. When we removed external Cl− in the presence of CO2/ , the pHi of HCMV-infected MRC-5 cells (72 h after exposure) became ∼0.3–0.5 pH unit more alkaline (Fig. 3). In contrast, the pHi of mock-infected cells is made acidic by the same treatment (Fig. 3). The alkalinization induced in the HCMV-infected cells by external Cl− removal was prevented by treatment with 100 μM H2DIDS, a well-known inhibitor of the Cl−/ exchanger (6). Thus these observations regarding the effects of removing external Cl− on pHi are consistent with the view that HCMV infection strongly enhances Cl−/ exchanger activity in MRC-5 cells.
A second important line of pHievidence that HCMV infection enhanced Cl−/ exchanger activity comes from the results of our studies on the pHi response of mock- and HCMV-infected cells to challenge with a hyperosmotic external fluid. We challenged them with a moderately hyperosmotic solution (146% of normal osmolality) buffered with CO2/ (Fig. 4). We observed that HCMV- and mock-infected cells became substantially more alkaline within 6 min of the initiation of the challenge. This pHi increase was due to the osmosensitivity of the NHE (21), since it could be prevented in mock- and HCMV-infected cells by treatment with DEA, an inhibitor of NHE. However, in the continued presence of the hyperosmotic fluid, the pHi of HCMV-infected cells (but not that of mock-infected cells) substantially recovered from this osmotically induced alkalinization. This recovery was most likely due to activation of the Cl−/ exchanger activity, since 1) it was not observed in the absence of CO2/ (10) and 2) it was prevented by treatment with 100 μM H2DIDS (Fig. 4). The fact that the maximal alkalinization on cell shrinkage for HCMV-infected cells was significantly greater (P < 0.05, unpairedt-test) in the presence of H2DIDS than in its absence, whereas for mock-infected cells there was no significant difference, is also consistent with much more Cl−/ exchanger activity after HCMV infection.
Finally, there was the effect on the resting pHi of bathing the cells in CO2/ solution. We earlier demonstrated that HCMV-infected cells bathed with HEPES-buffered solutions had a resting pHi ∼0.1 pH unit more alkaline than the identically treated mock-infected cells (10). This difference was due, in part, to the HCMV-induced stimulation of the NHE. In contrast, we report here that, when bathed in CO2/ buffered solutions, HCMV- and mock-infected cells had a similar pHi (Fig. 2). Thus, when CO2/ solution substituted for HEPES, the enhanced Cl−/ exchanger activity in the HCMV-infected cells “shunted” some of the alkaline pHi caused by HCMV infection.
Further evidence that HCMV enhances the Cl−/ exchanger mechanism comes from studies on [Cl−]i. Using MQAE, we found that the steady-state [Cl−]iof HCMV-infected cells bathed in CO2/ -containing solution was 63% higher than that of mock-infected cells (Table 1). Part, but not all, of this difference can be attributed to -dependent mechanisms. Thus treatment with H2DIDS (Figs.6 B and7 B) reduced the resting [Cl−]ifor HCMV-infected cells. In addition, bathing mock- and HCMV-infected cells in HEPES-buffered solutions resulted in a reduction of [Cl−]i(Table 1). For mock-infected cells the decrease was relatively small, declining from 58 to 53 mM; for HCMV-infected cells, [Cl−]ifell from 95 to 78 mM. In keeping with our pHi results, the apparent contribution of a CO2/ -dependent process is much greater for the HCMV- than for the mock-infected cells. However, even in the nominal absence of CO2/ , the two cell treatments have significantly different [Cl−]i.
Another approach used to characterize the effects of HCMV infection on Cl− transport mechanisms was measurement of the net rate of intracellular Cl− loss caused by the replacement of extracellular Cl−. In two series of studies we used gluconate, an anion that has a very limited ability to serve as an exchange partner for the Cl−/ exchanger (17). When external Cl− was replaced by gluconate in the absence of CO2/ , intracellular Cl− decreased in mock- and HCMV-infected cells. The rate of intracellular Cl− loss (corrected for changes in cell surface area and volume) was about two times faster from mock- than from HCMV-infected cells (Table 2). However, for neither cell was the rate substantially affected by treatment with H2DIDS.
The presence of CO2/ completely changed this pattern of Cl− movement. Addition of CO2/ increased the rate constants for loss of Cl− by both cell treatments but that by HCMV-infected cells increased the most (1.4- vs. 4-fold; Table 2). When changes in cell surface area and volume are taken into account, the relative rate of Cl− efflux is 32% greater from HCMV-infected cells in the presence of CO2/ than from mock-infected cells treated identically. However, treatment with H2DIDS reduced the rate of loss for HCMV-infected cells by >50% while slowing the rate from mock-infected cells by only ∼20% (Table 2). Thus this is another example of the HCMV-infected cells being much more responsive to the presence of CO2/ and being more sensitive to H2DIDS treatment.
is an anion known to exchange with Cl− via the Cl−/ exchanger (3, 14, 17, 22). When we replaced all the external Cl− with , we found that the H2DIDS-sensitive Cl− efflux was 5.3-fold greater in HCMV- than in mock-infected cells (Fig. 7). This provides further evidence of a much-enhanced Cl−/ exchanger activity in the HCMV-infected cells.
The Cl−/ exchanger involved with regulatory volume increase is thought to be the AE2 isoform (18). This isoform has a characteristic pHi sensitivity, being progressively stimulated as pHi is increased from ∼7.0 to 7.5 (32). At such alkaline pHi values, it functions as a base extruder. In contrast, AE1 has a relatively flat pHi dependence and does not respond to cell shrinkage when expressed inXenopus oocytes (18, 32). Our previous report shows that exposure to a moderate hyperosmotic challenge in the absence of CO2/ will increase the pHi of HCMV-infected cells (10). Our present results show that, in the presence of CO2/ , mock- and HCMV-infected cells initially alkalinize when challenged with a moderate hyperosmotic solution (Fig 3). However, the HCMV-infected cells recover by extruding base equivalents (Fig. 4). We have suggested that this recovery of pHi by the HCMV-infected cells in the continuous presence of the hyperosmotic challenge could be the result of the enhanced activity of the Cl−/ exchanger resulting from the osmotically induced pHi alkalinization.
If enhanced Cl−/ exchanger activity is responsible for the base extrusion, it would be expected that its activity would be increased by increases of pHi. We tested this hypothesis by measuring net Cl− efflux (using MQAE) from cells in which pHi was set to 6.4–7.8 with use of an ionophore to clamp pHito pHo. Because pHi was the independent variable for this study, we could not use as an exchange partner for Cl−. Instead, we took advantage of the fact that , an extremely weak base, can readily exchange for Cl− via the Cl−/ exchanger (3, 14, 17, 22). Our studies demonstrated a striking difference between the pHisensitivities of Cl− efflux between mock- and HCMV-infected cells. The Cl− efflux from HCMV-infected cells was nearly abolished at pHi 6.4 but reached a broad maximum as pHi was increased to ∼7.5. In contrast, Cl−efflux from mock-infected cells was maximal at pHi 6.4 and fell linearly as pHi was increased. These contrasting patterns of pHisensitivity are consistent with the conclusion that the HCMV-infected cells have much more Cl−/ ( ) exchanger activity (probably via an AE2 isoform) than the mock-infected cells.
Effects of HCMV Infection of Other Cl− Transport Pathways
The preceding discussion shows that HCMV infection caused a significant enhancement (3- to 5-fold) of anion exchange (Cl−/ exchanger) activity in MRC-5 cells. The mock-infected cells display only very slight evidence of such exchanger activity but clearly possess other Cl− transport mechanisms. Potential candidates include Na+-K+-Cl−cotransport, K+-Cl−cotransport, Na+-Cl−cotransport, Na+-dependent Cl−/ exchange, and Cl− channels. We have preliminary evidence that Na+-K+-Cl−and Na+-Cl−cotransport are functionally present in mock-infected cells but are absent or nearly absent from HCMV-infected cells (23). The Na+-dependent Cl−/ exchanger is an acid extruder. It is inhibited by disulfonic stilbene derivatives such as H2DIDS. However, unlike the activity we measured, it is stimulated by decreases in pHi (4). Thus, although this mechanism may exist in mock- and HCMV-infected cells, its operation cannot explain the -dependent acidification of pHi that we observed in HCMV-infected cells under several conditions (Figs.2-4).
For the HCMV-infected cells, Cl−/ exchange is an important means of Cl− transport into and out of the cell. In the absence of CO2/ , Cl− efflux from HCMV-infected cells is extremely slow, despite the fact that [Cl−]iis much higher than in the mock-infected cells. This may suggest that HCMV infection has decreased the electrochemical driving force (by causing membrane potential depolarization; see below). That would cause decreased net Cl− movement through channels. Alternatively, HCMV infection might result in inhibition of the Cl−channels.
Why Is [Cl−]iSo High in HCMV-Infected Cells?
Table 1 shows that HCMV infection increased the [Cl−]iin cells bathed in CO2/ -containing solution even more than in cells bathed in HEPES-buffered solution. Our data indicate that 31–46% of the overall increase can be ascribed to the addition of CO2/ to the external solution. Such an effect is possible, since the Cl−/ exchanger is thermodynamically poised to increase [Cl−]i([Cl−]o= 132 mM and the intra- and extracellular concentrations of are nearly equal). Furthermore, H2DIDS treatment reduced the [Cl−]iof HCMV-infected cells (Figs. 6 and 7). The above estimate of a 31–46% contribution of the Cl−/ exchanger to the [Cl−]imay be an underestimate, because the nominally “ -free,” HEPES-buffered solution is likely to contain as much as 0.13 mM (with the assumption that the solution was in equilibrium with room air). Thus, in addition to enhancing the activity of the Cl−/ exchanger, HCMV infection might also act to increase the activity of other “intracellular Cl−-loading” processes.
In addition to effects on possible intracellular Cl−-loading processes discussed above, HCMV infection might reduce the membrane resting potential so that the higher [Cl−]iwould result from some voltage-sensitive pathway for Cl− transmembrane movement. Finally, HCMV infection might reduce the activity of Cl− transport processes, which would lower [Cl−]i, such as K+-Cl−cotransport and/or Cl− channels. Final resolution of the basis of the high [Cl−]iin HCMV-infected cells awaits further studies.
CO2/ Treatment Induces NHE Osmosensitivity in Mock-Infected Cells
We previously reported that exposure to moderately hyperosmotic, HEPES-buffered solutions did not stimulate NHE activity in mock-infected cells, whereas the present report clearly shows a large DEA-sensitive alkalinization in mock-infected cells bathed in CO2/ -containing solution (Fig. 3).
In addition to the presence of CO2/ , there is another difference in the experimental conditions of the two sets of studies. The earlier study used sucrose to increase the osmolality, whereas we used NaCl in the present study. However, the difference in the osmolyte used to increase the external fluid osmolality is unlikely to explain the difference. Control experiments performed in the presence of CO2/ showed that increasing the osmolality of the solution bathing mock-infected cells with sucrose also stimulated a DEA-sensitive intracellular alkalinization (not shown). Thus we must consider that it is the presence of CO2/ that has, in some way, changed or greatly enhanced the response of the NHE to the hyperosmotic challenge.
Changes of [Cl−]ihave been shown to play a modulatory role in the activity of the NHE (16, 27). However, our results (Table 1) show only a modest increase of [Cl−]iin the mock-infected cells caused by bathing in CO2/ solution. Chen and Boron (8) suggested that there could be an extracellular “receptor” that activates the NHE in a way to make it much more sensitive to cell volume changes. A clear resolution of this intriguing effect on the behavior of the NHE of changing from a CO2/ -free to a CO2/ containing fluid awaits further study.
HCMV Infection Causes Concurrent Increases in Activities of NHE and Cl−/ Exchange
We previously reported that HCMV infection causes an increased activity of the NHE characterized by an alkaline shift of the pH optima and a much increased sensitivity to being stimulated by the cells’ exposure to hyperosmotic fluids (10). Thus, at pHi 7.45, we showed that the flux via the NHE was doubled 72 h after the cells had been exposed to the virus (10).
As discussed above, our present results provide strong evidence that HCMV infection, in addition to its effects on NHE activity, also substantially increases the activity of the Cl−/ exchanger. The combination of increased activity of these two ion exchangers, coupled by the NHE-induced pHi alkalinization, has often been demonstrated to effect an increase in cell volume after cell shrinkage (18, 24). It is possible that their combined increased activity may be responsible, at least in part, for the cell swelling so characteristic of HCMV infection. In support of that view is the fact that the increase in Cl−/ exchanger activity is already apparent 24 h after exposure to the virus (Fig. 8), well in advance of the increase of cell size that begins at ∼36–48 h after exposure (1).
We gratefully acknowledge the excellent technical assistance of Joshua C. Russell, Charles Rassier, and Kenneth Wilson.
Address for reprint requests: J. M. Russell, Dept. of Physiology, Allegheny University of the Health Sciences, 2900 Queen Ln., Philadelphia, PA 19129.
This work was supported by National Institute of Neurological Disorders and Stroke Grant NS-11946 (J. M. Russell).
Present address of A. A. Altamirano: Dept. de Microbiologı́a, Facultad de Medicina, Universidad de Buenos Aires, Paraguay 2155, 1121 Buenos Aires, Argentina.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 1998 the American Physiological Society