Introduction

Cystic fibrosis (CF) is caused by genetic mutations that cause loss of function of the CF transmembrane conductance regulator (CFTR), an epithelial cell Cl− channel (Gadsby et al. 2006). Understanding of normal CFTR function, as well as the molecular mechanisms of dysfunction in CF, is currently hampered by a lack of high-resolution structural information on the full-length protein. The structures of isolated, cytosolic domains of the CFTR protein that are involved in channel regulation have been solved at residue-level resolution (Lewis et al. 2004, 2005; Baker et al. 2007). In contrast, only very low-resolution imaging of the membrane-spanning parts of the protein that form the pathway for Cl− movement has been possible (Mio et al. 2008; Zhang et al. 2009); and as a result, the structure of the transmembrane pore region is currently relatively poorly defined.

An important breakthrough in studies of the structure and function of CFTR has been the development of forms of the protein in which all 18 endogenous cysteine residues have been removed by mutagenesis, resulting in a “cys-less” CFTR (Loo and Clarke 2006). This is important to provide a background into which unique cysteine residues can be introduced and probed individually using irreversible thiol-modifying agents. Cys-less CFTR has been used in this way to investigate the structure and function of the cytoplasmic nucleotide binding domains (NBDs) (Cui et al. 2006; Mense et al. 2006; He et al. 2008), cytoplasmic loops (He et al. 2008) and pore region (Serrano et al. 2006; Beck et al. 2008; Alexander et al. 2009; Bai et al. 2010; El Hiani and Linsdell 2010; Zhou et al. 2010). The utility of this approach is demonstrated by the fact that at least three groups have reported independently generating cys-less forms of CFTR (Cui et al. 2006; Mense et al. 2006; Wang et al. 2007).

Cys-less CFTR is difficult to traffic to the cell membrane (Loo and Clarke 2006; Wang et al. 2007); however, once resident in the membrane, it has been reported to have similar functional properties as wild-type CFTR. For example, protein kinase A (PKA)- and ATP-dependent gating at the single-channel level has been reported to be very similar (Cui et al. 2006; Mense et al. 2006). However, one notable functional difference that has been frequently (Cui et al. 2006; Mense et al. 2006; Li et al. 2009; Bai et al. 2010) but not universally (Serrano et al. 2006) reported is a small (~10–20%) increase in single-channel conductance. This functional difference has particularly strong implications for use of cys-less CFTR for studies of CFTR channel pore structure and function as it raises questions concerning how similar the permeation properties of these two channel constructs are and consequently how well a model of the wild-type channel pore is represented by the cys-less channel pore. Although cys-less has been used by several groups to investigate CFTR pore structure and function (Beck et al. 2008; Alexander et al. 2009; Bai et al. 2010; El Hiani and Linsdell 2010), the origin of this difference in conductance has not previously been investigated. In addition, functional differences between wild type and cys-less might point to previously unrecognized roles for native cysteine residues. With these points in mind, the aims of the present study were to explore both the extent and the molecular basis of functional differences between the wild-type and cys-less CFTR channel pores.

Methods

Experiments were carried out on baby hamster kidney cells transiently transfected with different forms of human CFTR. The CFTR backgrounds used were wild type (prepared as described in Gong et al. 2002) and one in which all cysteines had been removed by mutagenesis (“cys-less” CFTR; prepared as described in Mense et al. 2006; Li et al. 2009). Cys-less CFTR also included a mutation in the first NBD (V510A) to increase protein expression in the cell membrane (Li et al. 2009). The V510A mutation itself is not expected to affect single-channel conductance, which has previously been reported to be similarly increased in cys-less CFTR without (Cui et al. 2006; Mense et al. 2006) and with (Li et al. 2009; Bai et al. 2010) the V510A mutation. Transiently transfected cells were maintained at 37 °C (wild-type background) or 27 °C (cys-less background) for 1–4 days prior to patch-clamp experimentation. Mutations were introduced in these backgrounds using the QuikChange sit-directed mutagenesis system (Stratagene, La Jolla, CA) and verified by DNA sequencing.

Single-channel and macroscopic patch-clamp recordings were made using the excised, inside-out configuration of the patch-clamp technique. CFTR channels were activated after patch excision and recording of background currents by exposure to PKA catalytic subunit (5–30 nM) plus MgATP (1 mM) in the cytoplasmic solution. In some experiments using open channel blockers (see Figs. 5, 6) macroscopic currents were recorded following treatment with sodium pyrophosphate (PPi, 2 mM) to maximize channel activity and minimize changes in channel gating (Linsdell and Gong 2002; Gong and Linsdell 2003a; Fatehi et al. 2007). To ensure maximal stimulation by PPi, this substance was added after the PKA- and ATP-activated current had reached a steady amplitude. During most inside-out patch recordings, the intracellular (bath) solution contained (mM) 150 NaCl, 10 N-tris(hydroxymethyl)methyl-2-aminoethanesulfonate (TES) and 2 MgCl2. In a small number of experiments (see Figs. 5b, 7) this same solution was used in the extracellular (pipette) solution; however, in most cases a low extracellular Cl− concentration (4 mM) was used (NaCl replaced by Na gluconate). To estimate channel permeability to different anions (Fig. 7), NaCl in the intracellular solution was replaced by NaBr, NaI or NaSCN. All experimental solutions were adjusted to pH 7.4. All chemicals were obtained from Sigma-Aldrich (Oakville, Canada), except PKA (Promega, Madison, WI).

Current traces were filtered at 50–100 Hz using an eight-pole Bessel filter, digitized at 250 Hz–1 kHz and analyzed using pCLAMP9 software (Molecular Devices, Sunnyvale, CA). Single-channel current amplitudes were estimated from measurements of the amplitudes of individual open–closed transitions, measured using cursors set to the open and closed current levels over short periods of current record (Tabcharani et al. 1997). At least 20 individual current amplitudes were averaged for each patch at each membrane potential. Macroscopic current–voltage (I–V) relationships were constructed using depolarizing voltage ramp protocols (Linsdell and Hanrahan 1996, 1998). Background (leak) currents recorded before addition of PKA and ATP were subtracted digitally, leaving uncontaminated CFTR currents (Linsdell and Hanrahan 1998; Gong and Linsdell 2003a). Membrane voltages were corrected for liquid junction potentials calculated using pCLAMP9 software. As in a previous study (Zhou et al. 2010), the open channel blocking effects of 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB) on macroscopic current amplitude were analyzed using the simplest version of the Woodhull (1973) model of voltage-dependent block:

$$ I/I_{0} = K_{\text{d}} \left( {\text{V}} \right)/\left\{ {K_{\text{d}} \left( {\text{V}} \right) + \left[ {\text{B}} \right]} \right\} $$
(1)

where I is the current amplitude in the presence of blocker, I 0 the control unblocked current amplitude, [B] the blocker concentration and K d(V) the voltage-dependent apparent dissociation constant, the voltage dependence of which is given by

$$ K_{\text{d}} \left( {\text{V}} \right) = K_{\text{d}} \left( 0 \right){ \exp }\left( { - z\delta {\text{V}}F/RT} \right) $$
(2)

where zδ is the effective valence of the blocking ion (actual valence z multiplied by the fraction of the transmembrane electric field apparently experienced during the blocking reaction) and F, R and T have their usual thermodynamic meanings. For block by suramin, the K d was approximated by

$$ K_{\text{d}} = \left[ {\text{B}} \right]/\left\{ {\left[ { 1/\left( {I/I_{0} } \right)} \right] - 1} \right\} $$
(3)

The macroscopic current reversal potential (V REV) was estimated by fitting a polynomial function to the leak-subtracted I–V relationship and used to calculate the permeability of different anions (X−) relative to that of Cl− (PX/PCl) according to

$$ V_{\text{REV}} = \left( {RT/F} \right){ \ln }\left[ {\left( {{\text{P}}_{\text{X}} \left[ {\text{X}} \right]_{i} + {\text{P}}_{\text{Cl}} \left[ {\text{Cl}} \right]_{i} } \right)/\left( {{\text{P}}_{\text{Cl}} \left[ {\text{Cl}} \right]_{o} } \right)} \right] $$
(4)

where [X] i and [Cl] i (intracellular concentrations) are 150 and 4 mM respectively, and [Cl] o (extracellular) is 154 mM.

Experiments were carried out at room temperature, 21–24 °C. Values are presented as mean ± SEM. Tests of significance were carried out using Student’s two-tailed t-test.

Results

Conductance Differences Between Wild-Type and Cys-Less CFTR

The CFTR protein contains 18 cysteine residues, the locations of which are illustrated in Fig. 1. Of these, four are found in membrane-spanning regions: C128 (TM2), C225 (TM4), C343 (TM6) and C866 (TM7) (Fig. 1). Furthermore, one cysteine is located in the second intracellular loop connecting TMs 4 and 5, and this loop has previously been suggested to contribute to the conductance properties of the channel (Xie et al. 1995). These five cysteine residues were considered of highest likelihood to influence channel conductance and were mutated individually to serines (as in cys-less CFTR).

Fig. 1
figure 1

Location of cysteine residues in the CFTR protein. a Overall topology of the CFTR protein, comprising 12 transmembrane regions (TMs 1–12), two cytoplasmic NBDs (NBD1, NBD2) and the cytoplasmic regulatory domain (RD). b Location of the 18 endogenous cysteines in CFTR. Those mutated in the present study were C128 (TM2), C225 (TM4), C276 (intracellular loop 3, between TM4 and TM5), C343 (TM6) and C866 (TM7). Other cysteines are found in cytoplasmic regions of the protein: N terminus (C76), NBD1 (C491, C524), RD (C590, C592, C647, C832), NBD2 (C1344, C1355, C1395, C1400, C1410) and C terminus (C1458)

Full size image

Cys-less CFTR has previously been shown to have a slightly higher single-channel conductance than wild type (Mense et al. 2006; Cui et al. 2006; Li et al. 2009; see Introduction), an effect that is shown in Fig. 2. Using a low extracellular Cl− concentration (4 mM), the unitary slope conductance was 4.75 ± 0.06 pS (n = 11) for wild type and 5.61 ± 0.17 pS (n = 8) for cys-less (P < 0.0001) (Fig. 1d). This increase in conductance (18.3 ± 3.5%, n = 8) is similar to what has been reported previously under both low (Li et al. 2009) and high (Mense et al. 2006; Cui et al. 2006) extracellular [Cl−] conditions.

Fig. 2
figure 2

Single-channel conductance of wild-type, cys-less and mutant forms of CFTR. a Example single-channel currents carried by the different CFTR variants named with low extracellular Cl− concentration (4 mM), at a membrane potential of −50 mV. Arrows to the left indicate the closed and open channel current levels. b, c Mean single channel I–V relationships recorded under these conditions for wild-type (filled circle, b), cys-less (open circle, b), C343S (filled circle, c) and cys-less S343C (open circle, c). d Mean unitary slope conductances for different channel variants measured under these conditions. *Significant difference from wild-type, †significant difference from cys-less (P < 0.05 in both cases). Mean of data from 4 to 11 patches in (b)–(d)

Full size image

Single-channel conductance was also significantly increased in the point mutant C343S (P < 0.0001) (Fig. 2), albeit to a slightly lesser degree than in cys-less (14.5 ± 2.7%, n = 8). In fact, the conductances of cys-less and C343S were not significantly different (P > 0.4). In contrast, other point mutants studied (C128S, C225S, C276S, C866S) had conductances that were not significantly different from wild type but were significantly different from cys-less (Fig. 2d). This suggests that the conductance differences between wild type and cys-less result predominantly from the mutation of C343 in TM6 to a serine in this construct. Consistent with this hypothesis, the reverse mutation S343C in a cys-less background apparently restored the conductance phenotype of wild-type CFTR, significantly reducing cys-less single-channel conductance to a level not significantly different from that measured in wild type (Fig. 2).

Although TM6 is known to be an important determinant of CFTR pore properties (Linsdell 2006), we are unaware of any previous evidence that C343 influences unitary conductance. To investigate further the role of this amino acid, we therefore made additional substitutions at this position. As shown in Fig. 3, single-channel conductance was significantly reduced in C343L, significantly increased in C343A as well as C343S and not significantly changed in C343T and C343V. This pattern of changes suggests a size-dependent effect, with smaller amino acid side chains at this position being associated with higher single-channel conductance (Fig. 3d).

Fig. 3
figure 3

Effect of mutations at C343 on single-channel conductance. a Example single-channel currents carried by the different CFTR variants named with low (4 mM) extracellular Cl− concentrations, at a membrane potential of −50 mV. Line to the left indicates the closed channel current level. b Mean single channel I–V relationships recorded under these conditions for C343A (filled circle) and C343L (open circle). c Mean unitary slope conductances for different channel variants measured under these conditions. *Significant difference from wild type (P < 0.05). d Dependence of unitary conductance [as in (c)] on the volume of the amino acid side chain at position 343; C represents the wild-type cysteine residue. Side chain volumes estimated according to Richards (1974). Mean of data from 4 to 11 patches in (b)–(d)

Full size image

Other Functional Properties of Cys-Less CFTR

The conductance difference between wild-type and cys-less CFTR raises concerns about whether cys-less represents a good model of the CFTR pore. It is important, therefore, to understand the degree of functional similarity or difference in pore properties of these two channel variants. To this end, we compared a number of different, commonly used measures of pore function: binding of permeant and blocking ions, interactions between Cl− and blocking ions inside the channel pore and anion selectivity.

Permeant anion binding inside the pore is relatively sensitive to the effects of mutagenesis (Mansoura et al. 1998; Ge et al. 2004), and we have previously used block of Cl− permeation by the high-affinity permeant anion Au(CN) −2 as a functional measure of permeant anion binding within the pore (Gong et al. 2002; Ge et al. 2004; Fatehi et al. 2007). Because intracellular Au(CN) −2 affects CFTR channel gating as well as Cl− permeation (Linsdell and Gong 2002), we isolated pore-mediated effects using single-channel recording. As shown in Fig. 4, single-channel currents carried by wild-type and cys-less CFTR were both sensitive to block by intracellular Au(CN) −2 ions. In fact, no quantitative differences in sensitivity to Au(CN) −2 were observed (Fig. 4d).

Fig. 4
figure 4

Block of wild-type and cys-less CFTR by intracellular Au(CN) −2 ions. a Example single-channel currents carried by wild-type (left panels) and cys-less CFTR (right panels) with low extracellular Cl− concentration (4 mM), at a membrane potential of −50 mV. Line to the left indicates the closed channel current level. In both cases currents were recorded in the absence or presence of 50 μM Au(CN) −2 in the intracellular solution (as indicated). b, c Mean single I–V relationships for wild-type (b) and cys-less (c) CFTR under these conditions, both in control (open circle) and in the presence of 50 μM Au(CN) −2 in the intracellular solution (filled circle). d Mean fractional unitary current remaining in the presence of this concentration of Au(CN) −2 for wild type (filled circle) and cys-less (open circle). Mean of data from three to eight patches in (b)–(d)

Full size image

Chloride permeation in CFTR is sensitive to a wide range of impermeant open channel blocking anions acting from the cytoplasmic side of the membrane (Linsdell 2006). These inhibitors are thought to act by two different molecular mechanisms: by interacting with a relatively deep binding site in a voltage- and extracellular Cl− concentration-dependent manner or by interacting with a relatively superficial binding site in a more voltage- and Cl−-insensitive manner (St. Aubin et al. 2007). Examples of commonly used blockers acting by these two mechanisms are NPPB (Linsdell 2005) and suramin (St. Aubin et al. 2007).

Block by NPPB is shown in Fig. 5, under conditions of both low external Cl− concentration (4 mM, Fig. 5a) and high external Cl− concentration (154 mM, Fig. 5b). Under these conditions, NPPB causes a voltage-dependent block of CFTR current (Zhang et al. 2000; Linsdell 2005), and both the apparent affinity (K d[0], Fig. 5e) and voltage dependence (z′, Fig. 5f) are dependent on extracellular Cl− concentration. This is thought to reflect negative interactions between extracellular Cl− ions and intracellular blocking ions inside the channel pore (Gong and Linsdell 2003b). Nevertheless, there were no apparent differences in NPPB affinity or voltage dependence between wild type and cys-less under any conditions studied (Fig. 5). This suggests that not only NPPB block but also Cl−–blocker interactions are unaltered in cys-less CFTR.

Fig. 5
figure 5

Block of wild-type and cys-less CFTR by intracellular NPPB. a, b Example leak-subtracted macroscopic I–V relationships for wild-type and cys-less CFTR (as indicated), recorded before (control) and following addition of 10 μM NPPB to the intracellular (bath) solution. Currents were recorded using low (4 mM, a) or high (154 mM, b) extracellular Cl− concentration. c, d Mean fractional current remaining following addition of NPPB as a function of voltage, for wild type (filled circle) and cys-less (open circle), under low (4 mM, c) or high (154 mM, d) extracellular Cl− concentration conditions. e Mean K d(0) for NPPB, quantified as described in “Methods” section (Eqs. 1, 2) under low and high Cl− concentration conditions as indicated. f Mean apparent valence (−z′) for NPPB, quantified as described in “Methods” section (Eqs. 1, 2) under low and high Cl− concentration conditions as indicated. In both e and f, black bars represent wild-type and white bars cys-less CFTR. Mean of data from five to seven patches. *Significant difference from low [Cl−] conditions for the same channel construct (P < 0.05)

Full size image

Suramin block of wild-type and cys-less CFTR are compared in Fig. 6. As described previously (St. Aubin et al. 2007), addition of 10 μM suramin to the intracellular solution caused a potent, practically voltage-independent inhibition of wild type. The blocking effects of this concentration of suramin on cys-less CFTR appeared indistinguishable from those on wild type (Fig. 6).

Fig. 6
figure 6

Block of wild-type and cys-less CFTR by intracellular suramin. a Example leak-subtracted macroscopic I–V relationships for wild-type (left) and cys-less (right) CFTR with low extracellular Cl− concentration (4 mM). In both cases, currents were recorded before (control) and following addition of 10 μM suramin to the intracellular (bath) solution. b Mean fractional current remaining following addition of suramin as a function of voltage, for wild type (filled circle) and cys-less (open circle). c Mean K d for suramin, quantified at −100 mV as described in “Methods” section (Eq. 3). Mean of data from four to seven patches in (b)–(c)

Full size image

Cystic fibrosis transmembrane conductance regulator shows a lyotropic anion selectivity pattern (Linsdell 2006). Anion selectivity of wild type and cys-less were compared using intracellular (bath) solutions containing different test anions (Br−, I−, SCN−). As shown in Fig. 7, macroscopic current reversal potential measurements with different anions present in the bath solution showed no significant difference in the permeabilities of these anions between wild type and cys-less, with both channels showing the anion selectivity sequence SCN− > Br− > Cl− > I−.

Fig. 7
figure 7

Anion permeability of wild-type and cys-less CFTR. a, b Example leak-subtracted macroscopic I–V relationships for wild-type (a) and cys-less (b) CFTR, recorded with Cl−-containing extracellular solutions and intracellular solutions containing Br−, I− or SCN− as indicated. Note that experiments with different anions were carried out on different patches and, as such, no information is contained in the relative current amplitudes with different anions. Currents reversed close to 0 mV with Cl−-containing intracellular solutions as expected (see Fig. 5b). c Mean PX/PCl values calculated from reversal potential measurements under these conditions as described in “Methods” section (Eq. 4). For each anion tested, black bars represent wild-type and white bars cys-less CFTR. Mean of data from five or six patches

Full size image

Discussion

Cys-less CFTR is known to have a higher single-channel conductance than wild-type CFTR (Cui et al. 2006; Mense et al. 2006; Li et al. 2009; Bai et al. 2010), an effect that is shown in Fig. 2. In contrast, a range of experimental protocols designed to probe different aspects of pore function—binding of permeant (Fig. 4) and blocking (Figs. 5, 6) anions, interactions between a cytoplasmic blocker (NPPB) and extracellular Cl− ions (Fig. 5) and permeability of Br−, I− and SCN− ions (Fig. 7)—did not reveal any further significant differences between these two channel constructs. Previous work has identified specific mutations within the CFTR pore region that significantly alter the strength of block by cytoplasmic Au(CN) −2 (Gong et al. 2002; Ge et al. 2004), NPPB (Zhang et al. 2000; Linsdell 2005) and suramin (St. Aubin et al. 2007), as well as mutations that alter the interaction between extracellular Cl− and cytoplasmic blocking anions (Gupta and Linsdell 2002; Zhou and Linsdell 2009), giving some information on the possible location and distribution of anion binding sites within the pore. Furthermore, mutations in several TMs have been shown to alter the permeability of different anions (Linsdell et al. 1998, 2000; Mansoura et al. 1998; McCarty and Zhang 2001; Ge et al. 2004; Fatehi and Linsdell 2009), again leading to suggestions concerning the possible mechanism of anion selectivity (Linsdell 2006). The isolated increase in unitary conductance seen in cys-less therefore suggests that the functional differences between wild-type and cys-less channel pores are highly limited in scope.

The conductance difference between wild type and cys-less appears to result predominantly from a single difference, the replacement of cysteine 343 in TM6 of wild type by a serine in cys-less. Thus, the point mutation C343S causes an increase in conductance to a near cys-less level, whereas the revertant mutation S343C in a cys-less background reduces conductance to a near wild-type level (Fig. 2). This would appear to reflect a size-dependent effect of amino acid side chain size at this position on unitary conductance, with smaller side chains favoring higher conductance (Fig. 3). Given the overwhelming evidence that the native cysteine side chain at this position in TM6 does not line the aqueous lumen of the pore (Cheung and Akabas 1996; Alexander et al. 2009; Bai et al. 2010; El Hiani and Linsdell 2010), we speculate that mutation of this cysteine leads to a subtle change in packing and/or orientation of the TMs, resulting in a minor change in unitary conductance. To put this effect in context, mutation of many other TM6 residues, including R334, K335, F337, T338 and S341, has previously been shown to have a greater effect on conductance (Sheppard et al. 1993; McDonough et al. 1994; Linsdell et al. 1998; Linsdell 2001; Smith et al. 2001; Ge et al. 2004; Gong and Linsdell 2004; Bai et al. 2010).

Overall the functional differences between wild-type and cys-less CFTR channel pores appear to be of limited scale and scope and to reflect predominantly a minor change in TM arrangement due to a small change in volume of a non-pore-lining side chain. Since C343T (unlike C343S) was not associated with a significant change in single-channel conductance (Fig. 3), a cys-less construct in which C343 is replaced by threonine, rather than serine, might be considered to retain wild-type channel properties more faithfully. However, since most other permeation properties of cys-less were found to be indistinguishable between wild-type and cys-less (containing the C343S mutation), it seems likely that in most respects it is functionally inconsequential if this site is a cysteine, serine or threonine. Overall, cys-less would appear to be a good functional and, likely, structural model of the wild-type CFTR pore.