Deletion of phenylalanine 508 causes attenuated phosphorylation-dependent activation of CFTR chloride channels

Cell culture and electrophysiology

NIH3T3 mouse fibroblasts stably expressing WT- or ΔF508-CFTR channels (NIH3T3-CFTR and NIH3T3-ΔF508, respectively) were described previously (Berger et al. 1991). Cells were grown at 37°C and in 5 % CO₂ in Dulbecco's modified Eagle's medium-H21 supplemented with 2 mM glutamine and 10 % fetal bovine serum. For some whole-cell experiments, NIH3T3-ΔF508 cells were incubated at 27°C for 2–3 days before use.

Single-channel patch-clamp recording

Cells were passaged and grown on small, sterile glass chips in 35 mm tissue culture dishes. Prior to recording, glass chips were transferred to a continuously perfused chamber located on the stage of an inverted microscope (Olympus Corp., Tokyo, Japan). CFTR channel currents were recorded at room temperature (∼22°C) with a patch-clamp amplifier (EPC9, Heka Electronic, Lambrecht, Germany), filtered at 100 Hz with a built-in three-pole Bessel filter, and stored on videotapes. Data were subsequently refiltered at 25 Hz with an eight-pole Bessel filter (Frequency Devices Inc., Havervill, MA, USA) and captured onto a hard disk at a sampling rate of 50 Hz. Patch-clamp electrodes were made from Corning 7052 glass capillaries (Warner Instruments, Hammed, CT, USA). The pipette resistance was usually 3–5 MΩ, and the seal resistance was >20 GΩ. The pipette solution contained (mM): 140 N-methyl-D-glucamine chloride (NMDG-Cl), 2 MgCl₂, 5 CaCl₂ and 10 Hepes (pH 7.4 with NMDG). The superfusion solution contained (mM): 150 NaCl, 2 MgCl₂, 1 EGTA, 5 glucose and 5 Hepes (pH 7.4 with NaOH). For experiments using excised inside-out patches, the bath solution contained (mM): 150 NMDG-Cl, 10 EGTA, 10 Hepes, 8 Tris and 2 MgCl₂ (pH 7.4 with NMDG). Pipette potential was held at 50 mV relative to the bath. Downward deflections represent channel opening.

Whole-cell patch-clamp recording

Whole-cell currents were recorded with an Axopatch-1D patch-clamp amplifier (Axon Instruments, Foster City, CA, USA). Voltage pulses were generated using Igor software (Wavemetrics, Lake Oswego, OR, USA) and XOP (developed by Dr Richard Bookman at the University of Miami, Miami, FL, USA). Current traces, filtered at 1 kHz with a built-in four-pole Bessel filter, in response to voltage pulses (-100 to +100 mV at 20 mV increments) were digitized (at 2 kHz) directly into the computer hard disk through an ITC-16 interface (Instrutech Corp., Greatneck, NY, USA). The pipette solution contained (mM): 85 aspartic acid, 5 pyruvic acid, 10 EGTA, 10 Hepes, 20 tetraethylammonium-Cl, 5 Tris, 10 MgATP, 2 MgCl₂ and 5.5 glucose (pH 7.4 with CsOH). The bath solution contained (mM): 145 NaCl, 5 glucose, 5 Hepes, 2 MgCl₂, 1 CaCl₂, 5 KCl and 20 sucrose (pH 7.4 with NaOH). A chloride gradient of external concentration 156 mM and internal concentration 24 mM was used to generate an outward CFTR current at a holding potential of 0 mV. Sucrose (20 mM) was added to the bath to avoid activation of swelling-induced chloride conductance. This holding current at 0 mV was normalized with the membrane capacitance to yield a current density. Since in native NIH3T3 cells that are not transfected with CFTR neither forskolin nor forskolin plus genistein induced any current (n= 14), it is concluded that the currents activated in NIH3T3-CFTR or NIH3T3-ΔF508 cells were carried by CFTR.

Reagents

Forskolin, purchased from Calbiochem, was stored as a 20 mM stock solution in dimethyl sulfoxide (DMSO) at 4°C. Genistein, purchased from LC laboratories (Woburn, MA, USA), was stored as a 100 mM stock solution in DMSO at -20°C. MgATP and all salts and buffers were purchased from Sigma, and MgATP was stored as a 250 mM stock solution. The catalytic subunit of protein kinase A (PKA) was purchased from Promega. All [PKA] are reported as assayed by Promega (Flockhart & Corbin, 1984). The same lot of PKA was used to avoid variations in specific activity.

Kinetic analysis

PKA activation time course analysis

After excision of the membrane patch into an inside-out mode, the pipette was moved to a position a few micrometres away from the solution outlet to ensure that the pipette tip was completely immersed in the stream of the perfusate. ATP (2 mM) was perfused to the patch to ensure an absence of channel activity due to basal phosphorylation. Freshly made PKA solution was then continuously perfused together with ATP until a steady-state macroscopic current was reached. The fast solution change perfusion system (< 10 μl dead-space volume and complete solution change within 1 s in the chamber) ensures that the diffusive process of these compounds is not the rate-limiting step of our experiments. The mean current amplitude was measured, using Igor software as described previously (Wang et al. 1998), over every 5 s for the first 2 min of PKA addition and subsequently over every 30 s until a steady state was obtained. These mean current values were then normalized to the maximal steady-state mean current amplitude to yield percentage activation. The PKA activation time course was fitted with a sigmoidal function or an exponential function using Igor software. This sigmoidal function is a product of multiple single exponential functions with the same time constant (see Fig. 2 legend).

Single-channel kinetic analysis

We calculated single-channel P_o from recordings that showed fewer than three simultaneous levels of open channel current over >3 min of continuous traces. All-point amplitude histograms were first generated with the Igor software. The area under individual peaks of the all-point amplitude histogram was then used to calculate P_o using the following formula:

where A₀, A₁, …, A_N are: the area under the closing peak, the first channel opening peak, …, the Nth channel opening peak, respectively. Dwell time analysis was performed for recordings containing a single channel using Igor software as described previously (Wang et al. 1998).

Data are expressed as arithmetic means ±s.e.m. Statistical comparison was made using Student's t test. A value of P < 0.05 was considered to be statistically significant.

PKA-dependent phosphorylation activation of WT- and ΔF508-CFTR

It has been shown previously that ΔF508-CFTR channels respond poorly to cAMP stimulation with characteristically prolonged closed times in cell-attached patches (Dalemans et al. 1991; Haws et al. 1996; Hwang et al. 1997). Figure 1 shows representative current traces in cell-attached patches from NIH3T3 cells stably expressing ΔF508-CFTR. In the presence of 10 μM forskolin only a few opening events were observed over a 3 min time span. Further addition of a maximally effective concentration of CPT-cAMP (200 μM, see Al-Nakkash & Hwang, 1999), a membrane-permeant cAMP analogue, failed to increase the channel activity. In four experiments the mean macroscopic current amplitude in the presence of forskolin and CPT-cAMP was 1.2 ± 0.3 times that in the presence of forskolin alone, suggesting that 10 μM forskolin already maximally activates the cAMP-dependent CFTR activity (He et al. 1998; Al-Nakkash & Hwang, 1999). Contrary to previous reports (Drumm et al. 1991; Kelley et al. 1996), these data suggest that supersaturating concentrations of cAMP fail to rectify the functional defect of the plasma membrane ΔF508-CFTR channels. Thus, it is likely that the functional defect of ΔF508-CFTR resides in the step(s) beyond the production of cAMP and the activation of PKA. Although ΔF508-CFTR channels do not respond to increasing [cAMP], addition of 20 μM genistein in the continued presence of forskolin and CPT-cAMP caused a 28.5 (± 5.6)-fold (n= 12) increase in the mean macroscopic current amplitude. In patches containing one single ΔF508-CFTR channel (Fig. 1, inset), it can be seen that genistein drastically increases the P_o of ΔF508-CFTR. By visual inspection, it appears that genistein prolongs opening bursts and shortens closings as reported previously (Hwang et al. 1997; Wang et al. 1998).

Since CFTR channel activity in an intact cell is determined by both PKA-dependent phosphorylation and ATP-dependent gating, defects in either process could account for the low P_o of ΔF508-CFTR observed in cell-attached patches. We first examined the PKA-dependent activation in inside-out membrane patches excised from NIH3T3 cells expressing either WT-CFTR or ΔF508-CFTR by monitoring the time course for macroscopic current activation upon addition of purified PKA and ATP. After excision of the membrane patch into an inside-out configuration, 2 mM ATP was applied for 3–5 min to ensure that there was no channel activity before addition of PKA. Absence of channel opening prior to application of PKA in most of our patches suggests a negligible basal phosphorylation of the CFTR protein when expressed in NIH3T3 cells.

Figure 2A shows representative traces of the PKA-dependent activation of macroscopic WT-CFTR channel current in excised inside-out patches. As the [PKA] was increased from 6.25 to 62.5 U ml⁻¹, the time taken to reach a steady state was shortened. Activation time courses (Fig. 2B) were quantified by plotting the normalized activation vs. time (see Methods). At lower [PKA] (6.25 or 12.5 U ml⁻¹), the activation time courses show an initial delay. Fitting the data (smooth curves in Fig. 2B) with a sigmoidal function yielded time constants of 310.3 ± 23.4 s (n= 6) and 163.9 ± 10.4 s (n= 7), respectively. At higher [PKA] the activation time courses approximate a single exponential function with smaller time constants of 70.4 ± 6.5 s (n= 6) for 50 U ml⁻¹ PKA and 35.9 ± 4.4 s (n= 6) for 62.5 U ml⁻¹ PKA. Note that even at the highest [PKA] tested, the time constant is two orders of magnitude slower than the time constants for ATP-dependent gating of strongly phosphorylated CFTR (see below, and also Zeltwanger et al. 1999). Figure 3 shows the relationship between PKA concentration and the apparent activation rate (1/time constant). A curvilinear relationship between [PKA] and the apparent activation rate suggests that phosphorylation at multiple sites is required for maximal activation of CFTR channels.

To compare the activation time course of ΔF508-CFTR with WT-CFTR, 62.5 U ml⁻¹ PKA was used to activate ΔF508-CFTR (Fig. 4A). Figure 4B plots the percentage activation of ΔF508-CFTR current vs. time. Compared to WT-CFTR in the presence of the same concentration of PKA, it took a much longer time to reach a steady-state activation for the mutant channel. The smooth curve in Fig. 4B represents the fit of the data with a sigmoidal function (time constant 252.0 ± 17.2 s, n= 7). Corresponding results for WT-CFTR are presented for comparison. Note that the shape and the time constant of the activation time course for ΔF508-CFTR in the presence of 62.5 U ml⁻¹ PKA are similar to those for WT-CFTR in the presence of a 10-fold lower concentration of PKA (Fig. 3). The mean current amplitude of ΔF508-CFTR upon maximal activation was ∼50 % of that for WT-CFTR (Fig. 4C) presumably due to a lower channel density in NIH3T3-ΔF508 cells (see below). These results suggest a 7-fold difference in the PKA-dependent activation rate between WT- and ΔF508-CFTR.

ATP-dependent gating of fully phosphorylated WT-CFTR and ΔF508-CFTR

We next compared the ATP-dependent gating of ΔF508-CFTR and WT-CFTR. ΔF508-CFTR channels were first activated with 62.5 U ml⁻¹ PKA and 2 mM ATP until a steady state was achieved. Different concentrations of ATP were subsequently applied. Figure 5A demonstrates the relationship between [ATP] and macroscopic ΔF508-CFTR channel current. The macroscopic current amplitudes at different concentrations of ATP were normalized to the current level at 2.75 mM ATP. The dose-response relationship for WT-CFTR, previously published (Zeltwanger et al. 1999), was plotted here for comparison. There are no statistically significant differences between the ATP dose-response for WT-CFTR and that for ΔF508-CFTR (Student's t test, P= 0.45, 0.20 and 0.19 for 0.1, 0.5 and 1.0 mM ATP, respectively). Curve fitting of the data with Hill's equation yields K_½ of 128 ± 24 μM and 136 ± 68 μM for WT- and ΔF508-CFTR, respectively. Figure 5B shows the relationship between single-channel P_o and [ATP] for WT- and ΔF508-CFTR. Again, no statistically significant differences in P_o were observed at any [ATP] tested (Student's t test, P= 0.05, 0.31 and 0.22 for 0.1, 1.0 and 2.75 mM ATP, respectively). Curve fitting of the P_o values for ΔF508-CFTR with Hill's equation yields a K_½ of 213 ± 62 μM and a maximal P_o of 0.50 ± 0.07 (186 ± 38 μM and 0.43 ± 0.02 for WT-CFTR). To further examine the ATP-dependent gating of ΔF508-CFTR, dwell time analysis was performed. Figure 6A shows representative single-channel traces of PKA-phosphorylated WT- or ΔF508-CFTR in the presence of 2.75 mM ATP. Events from three patches containing one single ΔF508-CFTR channel were pooled to generate cumulative open and closed time histograms (Fig. 6B). Comparing the open time and closed time constants between WT- and ΔF508-CFTR (τ_o= 0.45 ± 0.03 s and τ_c= 0.52 ± 0.01 s for WT-CFTR, from Zeltwanger et al. 1999) in the presence of 2.75 mM ATP, we found no evidence for different single-channel kinetics between WT and the mutant channels. These single-channel kinetic data and the dose-response relationships shown above strongly suggest that, once fully phosphorylated, the interaction between ATP and the NBDs in ΔF508-CFTR is similar to that of wild-type channels.

Comparison of whole-cell current density between WT- and ΔF508-CFTR

Our previous studies (Hwang et al. 1997) have shown that in cell-attached patches and in the presence of a saturating concentration of forskolin (10 μM), the steady-state single-channel P_o for WT-CFTR is >10 times that for ΔF508-CFTR. However, genistein, in the presence of forskolin, could stimulate both WT-CFTR and ΔF508-CFTR activity to a similar P_o value of ∼0.7. Similar magnitudes of potentiation of cAMP-dependent CFTR channel current by genistein can be observed in whole-cell patch-clamp experiments. Figure 7 shows representative whole-cell recordings from NIH3T3-CFTR and NIH3T3-ΔF508 cells. With 10 mM ATP in the pipette, forskolin (10 μM) induced a significant outward current at a 0 mV holding potential ([Cl]_i= 24 mM, [Cl]_o= 156 mM) from an NIH3T3-CFTR cell, and genistein (20 μM) further increased the current by about 2.3-fold (Fig. 7A, left panel). Figure 7A (centre panel) shows samples of resulting current traces over the entire voltage range of -100 to 100 mV at a, b and c, marked in the left panel of Fig. 7A. In the absence of forskolin (Fig. 7A, centre panel, a), the current amplitudes over the ±100 mV range were very small indicating minimal basal channel activity and a high whole-cell resistance. In the presence of forskolin, as well as in the presence of forskolin and genistein (Fig. 7A centre panel, b and c), currents were dramatically increased and nearly time independent at positive membrane potentials. An outward rectification of the whole-cell CFTR chloride conductance, which is also shown in the current-voltage (I–V) relationship (Fig. 7A, right panel), is seen mostly because of the imposed Cl⁻ gradient across the membrane (see Methods).

In NIH3T3-ΔF508 cells grown at 37°C, forskolin (10 μM) elicited a very small current while genistein potentiated the current dramatically and the whole-cell ΔF508-CFTR currents showed similar properties of outward rectification and time independence (Fig. 7B) to WT-CFTR (Fig. 7A). Compared to WT-CFTR, however, the absolute current amplitude from ΔF508-CFTR cells is much smaller (notice the different current scales in Fig. 7A and B). Figure 8 summarizes the results from these whole-cell experiments. Current density was used to normalize differences in the cell size. When cells were incubated at 37°C, genistein increased forskolin-induced WT-CFTR current density by ∼2-fold (21.3 to 46.9 pA pF⁻¹). For ΔF508-CFTR, genistein gave a 40-fold increase in forskolin-induced current density (0.1 to 4.3 pA pF⁻¹). In the presence of forskolin and genistein, the whole-cell current density of NIH3T3-CFTR cells is ∼10-fold higher than that of NIH3T3-ΔF508 cells. However, when NIH3T3-ΔF508 cells were incubated at 27°C for 2–3 days before the experiment, the whole-cell current density was greatly increased. In the presence of forskolin plus genistein, the current density of NIH3T3-ΔF508 cells is ∼70 % of that from NIH3T3-CFTR cells (although they are not different significantly, P= 0.22, Student's t test). Thus, as reported previously (Hwang et al. 1997), lowering the culture temperature, although increasing the channel density in the cell membrane, does not rectify the defective response of ΔF508-CFTR channels to the cAMP stimulation. Since the whole-cell current density reflects the number of functional channels in the plasma membrane, these results allow us to compare quantitatively channel density between WT and ΔF508 channels (see Discussion).

Incremental activation of the CFTR channel by PKA-dependent phosphorylation

Applying different concentrations of PKA to excised inside-out membrane patches and monitoring CFTR channel current before reaching the steady state allowed us to quantify the PKA phosphorylation-dependent activation of CFTR. Just like using ATP-dependent gating parameters to extract biochemical information about the ATP-hydrolysis reaction (e.g. Zeltwanger et al. 1999; Weinreich et al. 1999), this real-time recording of the PKA-dependent CFTR activation process is another example of using electrophysiological methods to tackle (albeit indirectly) biochemical processes. The time course of PKA-dependent CFTR activation could be best fitted with a sigmoidal function, suggesting that multiple kinetic steps are involved in reaching the maximal function of the CFTR channel. Our previous studies (Wang et al. 1998) have demonstrated that CFTR can assume partially phosphorylated states upon pulse applications of PKA. The partially phosphorylated CFTR channel has a lower P_o with characteristic prolonged closed times even at millimolar [ATP] (cf. Hwang et al. 1993, 1994). Mathews et al. (1998) also reported a prolonged closed time associated with CFTR mutants where some of the PKA consensus serine residues are converted to alanine. Conversely, when the CFTR channel is strongly phosphorylated, the P_o is increased with a concomitant shortening of the closed time (Wang et al. 1998). Since multiple serine residues in the R domain seem to be involved in PKA-dependent phosphorylation activation of the CFTR channel (Rich et al. 1993; Chang et al. 1994; Seibert et al. 1997; Winter & Welsh, 1997; Mathews et al. 1998), it is expected that, in excised inside-out patches, the channel should sojourn through many intermediate states before it reaches a maximal level of phosphorylation. Thus, the time course can be sigmoidal provided that the P_o increases with an increased level of phosphorylation (cf. Wilkinson et al. 1997). The curvilinear relationship between [PKA] and the apparent activation rate (Fig. 3) further supports a co-operative relationship between multiple phosphorylation sites and the CFTR channel activity.

A slower PKA-dependent phosphorylation activation with the ΔF508 mutation

Our data demonstrated that ΔF508-CFTR has a slower apparent PKA-dependent activation rate than WT-CFTR in excised inside-out membrane patches. The simplest interpretation of this observation is that PKA phosphorylates ΔF508-CFTR channels at a slower rate, although we cannot rule out the possibility of a slower conformational change that occurs after phosphorylation. In other words, compared to WT-CFTR, ΔF508-CFTR proteins are not such good substrates for PKA. (Note that the rate of an enzymatic reaction is determined by the concentrations of the enzyme and the substrate as well as the intrinsic properties of the substrate.) This conclusion then predicts that, in an intact cell where kinases are constantly counteracted by cellular phosphatases, the critical phosphorylation site(s) for CFTR function will be phosphorylated to a lesser extent in ΔF508-CFTR than in WT-CFTR. Thus, the steady-state P_o of ΔF508-CFTR in cell-attached patches is much lower than that of WT-CFTR despite a normal ATP-dependent gating with the ΔF508 mutation (see below). This conclusion also predicts that when cellular PKA activity is lowered experimentally (e.g. using lower [forskolin]), WT-CFTR should have a lower P_o with similar kinetic behaviour to ΔF508-CFTR channels in cell-attached patches. Indeed, we have shown previously that in cell-attached patches, the behaviour of ΔF508-CFTR in the presence of a saturating concentration of forskolin mimics that of WT-CFTR in the presence of a submaximal concentration of forskolin (Hwang et al. 1997; Al-Nakkash & Hwang, 1999). In the current report, we show that, in excised inside-out patches, the PKA-dependent activation rate of ΔF508-CFTR is similar to that of WT-CFTR at a much lower [PKA] (Fig. 3). Our conclusion that ΔF508-CFTR is defective in PKA-dependent phosphorylation is interesting since F508 is located in NBD1 while PKA-dependent phosphorylation is believed to happen mostly in the R domain (Cheng et al. 1991; Chang et al. 1994; Seibert et al. 1997). Thus, this conclusion implies a functional perturbation in the R domain secondary to a structural change in NBD1 (i.e. an interdomain interaction). Our results also appear to resolve a long-standing puzzle in which the P_o of purified ΔF508-CFTR is not significantly different from that of WT-CFTR in lipid bilayers (Li et al. 1993) while the mutant channel behaves abnormally in cell-attached patches (Dalemans et al. 1991; Haws et al. 1996).

ATP-dependent gating behaviour of ΔF508-CFTR

In addition to a defective PKA-dependent phosphorylation, is the ATP-dependent gating also defective with the ΔF508 mutation? ATP-dependent gating for phosphorylated ΔF508-CFTR has been studied (Li et al. 1993; Schultz et al. 1999), but results from these two reports are quite different. While Li et al. (1993) showed no difference in channel kinetics at 1 mM ATP between WT- and ΔF508-CFTR, Schultz et al. (1999) demonstrated a ∼2-fold difference in the maximal P_o and a 9-fold difference in K_½ of the ATP dose-response relationship. Our results are more in line with those of Li et al. (1993). We found no differences in P_o, single-channel kinetics or the ATP dose-response relationship between WT- and ΔF508-CFTR channels. Although it is unclear what accounts for these discrepancies, different levels of membrane-associated protein phosphatases in systems used by different groups may be partly responsible.

Li et al. (1993) characterized purified CFTR incorporated in lipid bilayers, where there is likely to be little, if any, phosphatase contamination. The current studies were performed in excised membrane patches from NIH3T3 cells where minimal membrane-associated phosphatase activity was noted (Zeltwanger et al. 1999). On the other hand, Schultz et al. (1999) studied ΔF508-CFTR in excised inside-out membrane patches from L-cells or HEK 293 cells. Schultz et al. (1999) clearly demonstrated a fast dephosphorylation of ΔF508-CFTR upon removal of PKA in excised patches, suggesting a strong phosphatase activity associated with ΔF508-CFTR channels in the membrane patch. Because of this robust dephosphorylation of CFTR, Schultz et al. (1999) could not consistently obtain truly stationary conditions for their kinetic analysis of ΔF508-CFTR. Thus, they considered their estimates of kinetic parameters as ‘first approximations’ (Schultz et al. 1999). Indeed, if the membrane patches contain significant phosphatase activity, the P_o as well as the ATP dose-response relationships will be affected in such a way that the kinetic parameters should reflect an average behaviour of different phospho-isoforms of CFTR (a situation mimicking CFTR in intact cells). This might explain their reported long closed time associated with ΔF508-CFTR.

Although we cannot rule out the possibility that ΔF508-CFTR may have an abnormally higher dephosphorylation rate than the WT channel as suggested by Schultz et al. (1999), our data indicate that ΔF508-CFTR channels still could be strongly phosphorylated by PKA in excised patches due to the fact that the mutant channel activity reached a steady state with a similar P_o to that of WT-CFTR. Furthermore, the conclusion that the ATP-dependent gating of ΔF508-CFTR is not affected by the mutation is consistent with a recent homologous model of NBD1 of CFTR, in which the F508 residue was placed in an α helix that was >10 Å away from the ATP-binding pocket. This model also suggests that deletion of F508 causes little change in the ATP-binding pocket and speculates that this region where F508 is located may be involved in inter-domain interactions (Armstrong et al. 1998; also see Hung et al. 1998).

Relative roles of the trafficking defect and the functional defect with the ΔF508 mutation

It has been well established that deletion of F508 causes a trafficking defect and a functional defect, but the relative roles of these two abnormalities in cystic fibrosis pathogenesis is unknown. In theory, the defective trafficking will result in a reduced number of functional ΔF508-CFTR channels in the plasma membrane, whereas the abnormal response to the cAMP stimulation causes a lower P_o of the existing channels. Measurements of whole-cell CFTR current density could yield quantitative information about the number of functional channels in the plasma membrane since the macroscopic whole-cell current amplitude is determined by the number of channels, the single-channel P_o and the single-channel amplitude. It has been shown that WT- and ΔF508-CFTR chloride channels have identical single-channel conductances (Dalemans et al. 1991) and very similar P_o values in the presence of forskolin and genistein in intact cells (Hwang et al. 1997). Therefore, the whole-cell current density in the presence of forskolin and genistein provides a parameter that can be used to compare plasma membrane channel densities between cells expressing WT- or ΔF508-CFTR channels. With known values for P_o and single-channel current amplitude, one can even estimate precisely the channel density in the cell membrane.

From the data presented in Fig. 8, several conclusions can be made. (1) In the presence of cAMP stimulation alone (i.e. 10 μM forskolin), the current density of NIH3T3-ΔF508 cells is almost negligible (0.1 ± 0.1 pA pF⁻¹ at 37°C and 0.6 ± 0.2 pA pF⁻¹ at 27°C). However, with forskolin and genistein together, significant currents can be elicited (4.3 ± 2.1 pA pF⁻¹ at 37°C and 33.6 ± 14.2 pA pF⁻¹ at 27°C). Thus, using cAMP-dependent chloride current to quantify ΔF508-CFTR in the cell membrane is likely to produce an underestimation of the channel density (e.g. Rich et al. 1993). (2) In the presence of forskolin and genistein, when there is little difference in the P_o between WT- and ΔF508-CFTR, the current density of NIH3T3-ΔF508 cells increases from 9 % (4.3/46.9) to 72 % (33.6/46.9) of that from NIH3T3-CFTR cells as the culture temperature is lowered. This result provides a quantitative estimation of the role of the trafficking defect and the degree of its correction by the low culture temperature. (3) Assuming a single-channel amplitude of 0.2 pA and a P_o of 0.7 in the presence of forskolin and genistein (Hwang et al. 1997), the current density values can be converted to channel density given a specific membrane capacitance of 1 μF cm⁻² (3.35 μm⁻² for WT-CFTR, 0.31 μm⁻² for ΔF508-CFTR at 37°C and 2.40 μm⁻² for ΔF508-CFTR at 27°C). Thus, the trafficking defect caused by the ΔF508 mutation decreases the number of functional ΔF508-CFTR channels to ∼10 % of that of WT-CFTR. However, in the presence of cAMP, the physiological stimulus, alone, the current density from NIH3T3-ΔF508 cells is only 0.4 % that of NIH3T3-CFTR cells because of a combination of the trafficking defect and the functional defect. Therefore, at least in these two NIH3T3 cell lines where similar amounts of total CFTR proteins (mature plus immature) have been demonstrated (Denning et al. 1992), the functional defect of ΔF508-CFTR plays a role of equal importance in the pathophysiology of cystic fibrosis to that of the trafficking defect.

One should note, however, that the relative role of the trafficking defect may vary in different expression systems (c.f. Rich et al. 1993; Lukacs et al. 1994; Haws et al. 1996) and in different tissues in patients with cystic fibrosis (Kälin et al. 1999). Using immunocytochemical methods, Kälin et al. (1999) demonstrated that the amount of CFTR proteins in the cell membrane is drastically reduced in sweat ducts isolated from ΔF508 homozygotes, but the expression of the mutant protein is indistinguishable from the WT-CFTR in airway and intestinal epithelial cells. This result strongly suggests that the abnormal trafficking may not be the major determinant for ΔF508 cystic fibrosis disease. Thus, a better understanding of the molecular mechanism for the functional defect of ΔF508-CFTR will aid in the design of pharmacological agents for therapeutical intervention in cystic fibrosis.