Severed molecules functionally define the boundaries of the cystic fibrosis transmembrane conductance regulator's NH(2)-terminal nucleotide binding domain

Abstract

The cystic fibrosis transmembrane conductance regulator is a Cl(-) channel that belongs to the family of ATP-binding cassette proteins. The CFTR polypeptide comprises two transmembrane domains, two nucleotide binding domains (NBD1 and NBD2), and a regulatory (R) domain. Gating of the channel is controlled by kinase-mediated phosphorylation of the R domain and by ATP binding, and, likely, hydrolysis at the NBDs. Exon 13 of the CFTR gene encodes amino acids (aa's) 590-830, which were originally ascribed to the R domain. In this study, CFTR channels were severed near likely NH(2)- or COOH-terminal boundaries of NBD1. CFTR channel activity, assayed using two-microelectrode voltage clamp and excised patch recordings, provided a sensitive measure of successful assembly of each pair of channel segments as the sever point was systematically shifted along the primary sequence. Substantial channel activity was taken as an indication that NBD1 was functionally intact. This approach revealed that the COOH terminus of NBD1 extends beyond aa 590 and lies between aa's 622 and 634, while the NH(2) terminus of NBD1 lies between aa's 432 and 449. To facilitate biochemical studies of the expressed proteins, a Flag epitope was added to the NH(2) termini of full length CFTR, and of CFTR segments truncated before the normal COOH terminus (aa 1480). The functionally identified NBD1 boundaries are supported by Western blotting, coimmunoprecipitation, and deglycosylation studies, which showed that an NH(2)-terminal segment representing aa's 3-622 (Flag3-622) or 3-633 (Flag3-633) could physically associate with a COOH-terminal fragment representing aa's 634-1480 (634-1480); however, the latter fragment was glycosylated to the mature form only in the presence of Flag3-633. Similarly, 433-1480 could physically associate with Flag3-432 and was glycosylated to the mature form; however, 449-1480 protein seemed unstable and could hardly be detected even when expressed with Flag3-432. In excised-patch recordings, all functional severed CFTR channels displayed the hallmark characteristics of CFTR, including the requirement of phosphorylation and exposure to MgATP for gating, ability to be locked open by pyrophosphate or AMP-PNP, small single channel conductances, and high apparent affinity of channel opening by MgATP. Our definitions of the boundaries of the NBD1 domain in CFTR are supported by comparison with the solved NBD structures of HisP and RbsA.

Scheme S1

Figure 1

(A–C) Representative current recordings from oocytes held at their initial resting membrane potential in Ca²⁺-free oocyte Ringer's solution. Oocytes were injected with (A) water, (B) 6 ng of WT CFTR RNA, or (C) 2.5 ng of Flag3-633 RNA plus 2.5 ng 634-1480 RNA. Horizontal bars mark periods of exposure to 50 μM forskolin plus 1 mM IBMX. Basal and activated conductances were monitored by applying ±60-mV voltage steps ≤1-s long. Magnitudes of basal and activated conductances and rates of activation were similar in oocytes injected with 2.5 or 6 ng of WT CFTR RNA. (D) I-V relationships obtained at times a-f in A–C; empty symbols, resting (basal); filled symbols, activated. Steady current levels averaged over ∼200 ms near the end of 1-s voltage steps from −100 to +80 mV, in 20-mV increments, are plotted against voltage.

Figure 2

Defining the COOH-terminal boundary of NBD1 by coexpressing severed CFTR segments. Bars show mean (±SEM) basal and activated conductances (by 50 μM forskolin plus 1 mM IBMX) of oocytes injected with cRNAs (2.5 ng per construct) of various CFTR segments cartooned at left: NH₂-terminal Flag epitope (black zigzag); transmembrane domains (blue, cyan); NBD1 (red); R domain (green); NBD2 (yellow). Conductances were calculated from linear fits to steady state I-V data between −60 and −20 mV; average values are from five or more oocytes. The COOH terminus lies between residues 622 and 634.

Figure 3

Defining the NH₂-terminal boundary of NBD1 by coexpressing severed CFTR channel segments. Methods and symbols are as in Fig. 2. The NH₂ terminus was found to lie between residues 432 and 449.

Figure 8

Representative excised-patch current recordings showing WT, 1-432+433-1480, 1-633+634-1480, Flag-WT, Flag3-432+433-1480, and Flag3-633+634-1480 CFTR channels. All constructs displayed hallmark characteristics of WT CFTR including requirement for phosphorylation by PKA (here 300 nM) and exposure to MgATP (here 1 mM for WT, 2 mM for others) for channel activity, slow burst kinetics, and locking in the open state by a mixture of ATP (0.1 mM) and PP_i (2 mM).

Figure 4

Negligible basal and activated conductances of oocytes injected with 2.5 ng cRNA encoding the indicated single truncated CFTR channel segments. Methods and symbols are as in Fig. 2.

Figure 5

Protein expression of CFTR constructs severed near the COOH terminus of NBD1. (A and B) Immunoblots of membrane proteins from oocytes expressing individual CFTR segments or pairs of segments, as indicated, resolved by SDS-PAGE, transferred, and blotted with anti–R-domain antibody (A) or with anti–Flag M2 antibody (B). (C) Coexpressed constructs Flag3-633 plus 634-1480 were digested with N-glycosidase-F (left) or endoglycosidase-H (right), and the products were identified with anti–R-domain antibody. Arrows in A and C mark fully glycosylated (∼150 kD, fat arrow), core glycosylated (∼95 kD, thin arrow), and deglycosylated (∼90 kD, thin arrow) forms of the 634-1480 protein.

Figure 6

Protein expression of CFTR constructs severed near the NH₂ terminus of NBD1. (A) Immunoblot of membrane proteins from oocytes expressing individual CFTR segments or pairs of segments, as indicated, blotted with anti–R-domain antibody. (B) Coexpressed constructs Flag3-432+433-1480 were digested with N-glycosidase-F (left) and endoglycosidase-H (right), and the products were identified with anti–R-domain antibody. Arrows in A and B mark fully glycosylated (∼160 kD, fat arrow), core glycosylated (∼125 kD, thin arrow), and deglycosylated (∼120 kD, thin arrow) forms of 433-1480 protein. *Partially degraded form (∼70 kD).

Figure 7

Coimmunoprecipitation of severed CFTR constructs. Digitonin-solubilized membrane proteins were immunoprecipitated with anti–Flag M2 affinity beads, eluted, resolved with SDS-PAGE on 7.5–10% gradient gels, and blotted with anti–R-domain antibody (top) or anti–NH₂-terminus antibody (bottom). (A) Constructs severed near the COOH terminus of NBD1: uninjected (lane 1), Flag3-622+634-1480 (lane 2), Flag3-633+634-1480 (lane 3); arrows are as in Fig. 5 A. (B) Constructs severed near the NH₂-terminus of NBD1: Flag3-432+449-1480 (lane 1), Flag3-432+433-1480 (lane 2); arrows are as in Fig. 6 A.

Figure 9

Kinetic analysis of WT, 1-432+433-1480, 1-633+634-1480, Flag-WT, Flag3-432+433-1480, and Flag3-633+634-1480 channels. Kinetic parameters were extracted from records (filtered at 100 Hz, sampled at 1 kHz) of patches containing one to seven channels, exposed to 2 mM MgATP and 300 nM PKA. Mean burst durations of constructs severed between NBD1 and the R domain (1-633+634-1480 and Flag3-633+634-1480) were significantly shorter than those of WT or Flag-WT. All constructs containing the Flag epitope (Flag-WT, Flag3-432+ 433-1480, and Flag3-633+634-1480) showed significantly prolonged interburst durations and, hence, significantly reduced P _o compared with WT. Significance levels: *P < 0.1, **P < 0.05, and ***P < 0.01.

Figure 11

Apparent affinities for activation of channel P _o by MgATP. Representative macropatch currents are shown for WT and Flag-WT (A), 1-633+634-1480 and Flag3-633+634-1480 (B), 1-432+433-1480 and Flag3-432+433-1480 (C), and Flag3-414+433-1480 (D) channels. After prephosphorylation by PKA, and removal of both PKA and ATP, once all channels had closed, 50 μM ATP was applied for 10–30 s, bracketed between exposures to 2 mM ATP. (E) The ratios of the mean steady currents, I_50μM/I_2mM, were 0.50 ± 0.02 (WT), 0.56 ± 0.03 (Flag-WT), 0.51 ± 0.02 (1-633+634-1480), 0.42 ± 0.01 (Flag3-633+634-1480), 0.49 ± 0.02 (1-432+433-1480), 0.44 ± 0.01 (Flag3-432+433-1480), and 0.45 ± 0.01 (Flag3-414+433-1480). None of the constructs differed significantly from WT at P < 0.05.

Figure 12

Alignment of CFTR's NBD1 with HisP. PSI-Blast aligned CFTR residues 422–637 with corresponding residues in HisP. Secondary structure elements shown below (β-strands, magenta arrows; α-helices, orange boxes) correspond to those of the solved crystal structure of HisP (Hung et al. 1998). CFTR residue numbers are centered above the corresponding letters. Teal boxes identify the three consensus sequences conserved in most ABC proteins. Red boxes in the CFTR sequence identify nontolerated, green boxes tolerated, cut sites. (Cuts occurred between the two highlighted residues in each case.) Gray boxes identify sections flanking the NBD1 sequence that could be omitted without destroying channel function. *Residue F508.

Figure 10

Single-channel conductances of WT, 1-432+433-1480, 1-633+634-1480, FlagWT, Flag3-432+433-1480, and Flag3-633+634-1480 channels in symmetrical 140 mM Cl⁻. (A and B) Segments of representative CFTR channel current records, illustrated for WT and 1-432+433-1480, at holding potentials of −80, −40, 0, +40, and +80 mV, were used to create amplitude histograms, fitted by sums of Gaussians (right). The distances between adjacent peaks, plotted against potential gave linear I-V plots. (C) Fitted slopes gave conductances of 6.8 ± 0.3 (WT), 7.1 ± 0.2 (1-432+433-1480), 6.3 ± 0.2 (1-633+634-1480), 6.6 ± 0.3 (Flag-WT), 7.0 ± 0.2 (Flag3-432+433-1480), and 7.0 ± 0.1 (Flag3-633+634-1480), none of which differed significantly from WT at P < 0.1.