| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
(Received for publication, August 11,
1994; and in revised form, November 2, 1994) From the
The cystic fibrosis transmembrane conductance regulator (CFTR)
Cl The cystic fibrosis transmembrane conductance regulator (CFTR) ( A notable
feature of CFTR and many members of the ABC transporter family is the
presence of two NBDs. In CFTR, the two NBDs share sequence similarity
in certain conserved regions such as the Walker A
(GXXGXGKT/S) and Walker B
(R/KX To better understand the function
of each of the NBDs, we studied CFTR Cl These results
suggested that mutation of the Walker lysines in CFTR may reduce the
rate of ATP hydrolysis (if it occurs) without altering the ability of
the protein to assume a normal tertiary structure, bind ATP, and
undergo ATP binding-induced conformational changes. To investigate the
role of each NBD in CFTR channel regulation, we studied CFTR variants
in which the conserved Walker lysines in NBD1 (Lys
For whole cell
and excised macropatch data, replayed records were filtered at 1 kHz
using a variable 8-pole Bessel filter (Frequency Devices Inc.,
Haverhill, MA) and digitized at 2 kHz. For single channel analysis,
replayed data were filtered at 1 kHz using a variable 8-pole Bessel
filter, digitized at 5 kHz, and digitally filtered at 500 Hz. Idealized
records were created using a half-height transition protocol;
transitions less than 1 ms in duration were not included in the
analysis. Single channel open and closed time histograms were
plotted with a logarithmic x axis with 10 bins/decade, and the
maximum likelihood method was used to fit a one or two component
exponential function, respectively. Burst duration data were plotted as
histograms with a logarithmic x axis with 10 bins/decade and
were fit with both one and two component exponential functions using
the maximum likelihood method, with a lower fitting limit of 2.5 ms. To
determine if the two component function fit statistically better than a
one component fit, the log likelihood ratio test was used and
considered significant at a value of 2.0 or greater(26) . Burst analysis was performed with a t
Figure 5:
Burst
analysis of wild-type CFTR and each Walker lysine mutant. A,
closed time histogram of wild-type CFTR. Data are from an experiment in
which the membrane patch had only one active channel, studied in the
presence of 1 mM ATP and 75 nM catalytic subunit of
PKA. Data are plotted with a logarithmic x axis and linear y axis. Burst duration delimiter, t
Results are means ±
S.E. of n observations. Statistical significance was assessed
using the log likelihood ratio test or a paired or unpaired
Student's t test as appropriate.
Photolabeling was performed by preincubating membranes (50
µg of membrane protein/sample) on ice with the indicated amount of
[
For experiments with excised
inside-out membrane patches, the pipette (extracellular) solution
contained (in mM): 140 N-methyl-D-glucamine,
100 aspartic acid, 35.5 HCl, 5 CaCl
Figure 1:
Properties of K464A/K1250A in
transiently transfected HeLa cells. A, whole cell currents
recorded under basal conditions, 2 min after addition of cAMP agonists
(10 µM forskolin, 100 µM isobutylmethylxanthine, and 250 µM 8-(4-chlorophenylthio)adenosine-3`,5`cyclic monophosphate) to the
bath solution, and 5 min after removing cAMP agonists. Voltage pulse
protocol is shown. Each trace is the average of three voltage steps. B, time course of current from an excised membrane patch. Each
point is average current during 1 s of data collection, with one point
collected every 5 s. Bars indicate presence of ATP (1
mM), PKA (75 nM), and absence of Mg
However, mutation of the Walker A lysines
did alter channel gating. Fig. 2shows examples of single
channel tracings. The top tracing shows the characteristic pattern of
gating in wild-type CFTR, with short bursts of activity (in which the
channel flickers open and closed) separated by longer closings. In
comparison, the duration of the closed intervals between bursts of
activity was considerably increased in variants containing mutations of
the Walker lysines. More strikingly, the duration ofbursts was much
prolonged in variants containing mutation of Lys
Figure 2:
Single-channel traces of wild-type CFTR,
and each Walker lysine mutant studied in excised, inside-out membrane
patches from transiently transfected HeLa cells. All measurements were
made in the presence of 1 mM ATP and 75 nM PKA with
membrane potential clamped at -80 mV. Each tracing is 20 s long. Dashed lines represent channel closed state, and downward
deflections correspond to channel openings. For purpose of
illustration, traces were digitized at 2 kHz and filtered at 200 Hz and
chosen to illustrate changes in closed times between bursts and burst
duration and not changes in single channel open state probability (P
To
quantitate these changes and to determine how these mutations altered
CFTR Cl
Figure 3:
P
To
determine how these mutations decreased P
Figure 4:
Open- and closed-time constants from
patches of membrane containing one channel. Histograms were plotted and
fit as described in ``Experimental Procedures''.
Although alteration of these fast events
contributed to the decrease in P To quantitate how the Walker
lysine mutations were altering the duration of channel opening bursts,
burst analysis was performed using a t Fig. 5B shows the effect of Walker lysine mutations
on channel burst duration. K464A had a small but significant decrease
in burst duration relative to that of wild-type channels (136 ±
7 ms for K464A versus 210 ± 22 ms for wild-type, p = 0.007). In contrast, both variants with mutation of the
Walker lysine in NBD2 had a 4-5-fold increase in the burst
duration relative to that of wild-type (881 ± 96 and 1118
± 217 ms for K1250A and K1250M, respectively, p <
0.001 for wild-type versus either NBD2 mutant). Interestingly,
when the Walker lysines in both NBDs were mutated (K464A/K1250A), the
longer burst duration conferred by mutation of K1250 was dominant over
the shorter burst conferred by K464A (1101 ± 310 ms versus 136 ± 7 ms for K464A, p < 0.001). These
results suggest that the NBDs are important not only in opening the
channel, but also in closing the channel, specifically in ending bursts
of activity. Because mutations of the Walker lysines are expected to
slow the rate of ATP hydrolysis, and because such mutations in NBD2
markedly slowed the rate at which bursts of activity were terminated,
we interpret these data to suggest that termination of a burst is an
active event, mediated by NBD2, which involves ATP hydrolysis.
Mutation of both Walker lysines to alanine, or of Lys
Figure 6:
Binding of 8-N
When we added 1 mM AMP-PNP to excised membrane
patches, most openings appeared to be of normal duration (Fig. 7A). However, there were occasional openings that
had a substantially prolonged burst duration, similar to those observed
in K1250A, K1250M, and K464A/K1250A. This observation was confirmed by
an analysis of burst duration histograms. Prior to addition of AMP-PNP,
there was only one population of bursts (Fig. 7B),
since in five of six experiments a two component exponential function
did not fit the data statistically better than a one component fit.
After addition of AMP-PNP, a second longer, and statistically distinct
population of bursts was evident in six of six experiments (Fig. 7C). Although this second population represented
a small fraction of the total number of bursts, each individual burst
within this population was substantially prolonged (note the log scale
in Fig. 7, B and C). As a result, the mean
burst duration (Fig. 7D) approximately doubled (p = 0.033) and P
Figure 7:
Effect of AMP-PNP on wild-type CFTR
channel activity. A, current traces from a stably transfected
C127 membrane patch with only a single active channel; data are plotted
as in Fig. 3. The presence of both ATP (0.3 mM) and PKA
were necessary to see the effect of AMP-PNP (1 mM). B and C, burst duration histograms of wild-type CFTR
activity in the absence (B) and presence (C) of 1
mM AMP-PNP. Data for each histogram are derived from
successive interventions in one experiment and plotted and fit as
described under ``Experimental Procedures.'' Separate
populations of bursts appear as different peaks. The solid line is the superimposed maximum likelihood fit. D and E, effect of 1 mM AMP-PNP on burst duration (D) and P
Figure 8:
Model of the interaction of the NBDs of
CFTR with ATP and the effect on channel gating. Vertical columns from left to right represent events at NBD1, at
NBD2, and the effect on channel opening and closing.
``C'' and ``O'' represent channel
open and closed states, respectively. See text for a detailed
description.
In state
1, no ATP is bound to the NBDs and the channel is closed. The
requirement of ATP for the channel to open supports this conclusion. In state 2, ATP binds to both NBDs, but the channel remains closed.
We conclude that ATP binding alone is not sufficient to open the
channel based on two types of observation. First, biochemical (9) and functional studies (31, 33, and our present data) have
shown that nonhydrolyzable analogs can bind to and interact with the
channel, but in the absence of ATP, they are not capable of opening
it(11, 12, 24) . Second, biochemical studies
have shown that nucleoside triphosphates can bind to the channel in the
absence of Mg In state 3, hydrolysis has occurred at NBD1,
driving the channel into a bursting state. Within a burst, the channel
flickers open and closed. We suggest that hydrolysis at NBD1 opens the
channel, because mutation of K464 slows the rate of opening; in K464A
the duration of the closed state between the bursts of activity was
increased 5-fold. Our finding that mutation of Walker lysines did not
alter binding of 8-N In state 4,
hydrolysis has occurred at NBD2, actively terminating the burst and
closing the channel. Mutation of the Walker lysine in NBD2 will likely
slow hydrolysis, delay the exit from state 3, and thereby increase the
burst duration. Assigning hydrolysis at NBD2 to occur after that at
NBD1 also explains the observation that mutations of Lys Our observation that AMP-PNP prolongs
the burst duration in a manner similar to mutations in NBD2 suggests
that AMP-PNP is binding to NBD2 and supports the conclusion that
hydrolysis is required to terminate a burst and close the channel.
AMP-PNP appeared to be much less potent than ATP at interacting with
CFTR as suggested by binding studies (AMP-PNP had approximately 5% the
binding potency of ATP) (9) and by the data in Fig. 8C (which shows that the fraction of bursts with a
prolonged duration was small). Nevertheless, once bound, the effect on
burst duration was substantial. In addition to the direct role of
each NBD in opening and closing the channel, we speculate that there
must also be some form of cross-talk between the NBDs to account for
the finding that NBD2 mutations decrease the rate of channel opening.
We indicate this in Fig. 8by the arrow from NBD2 to
NBD1. Thus, it appears that normal function at NBD2 is necessary for
normal function at NBD1, and it appears that events at NBD2 may
regulate NBD1 function. In a similar way, NBD1 may also influence NBD2
function. Although we have not indicated it in Fig. 8because
the effect is small, this is suggested by the observation that mutation
of Lys In addition to the steps shown in the model,
the channel might have additional states where there is either a
phosphorylated intermediate, an ADP-bound state, or other
conformational states which we cannot yet resolve and have therefore
not included. Other steps involving release or removal of bound
nucleotide may also occur in recycling of the channel from state 4 to
state 1, but our studies do not address these steps. In addition, this
model does not take into account the effect of different
phosphorylation states on channel activity. We previously suggested
that hydrolysis is required for channel opening (11) . Our
current findings support that conclusion. Gadsby and colleagues (33, 34) recently proposed that hydrolysis is not only
involved in channel opening but that hydrolysis at one of the NBDs may
be involved in closing. Our current findings support their conclusion
to the extent that the data suggest that hydrolysis at one of the NBDs
(NBD2) closes the channel by terminating a burst of activity. However,
Baukrowitz et al.(34) concluded that hydrolysis at
the NBD involved in closure only occurs at high phosphorylation states (i.e. in the presence of PKA). We found that the NBD2 mutants
always prolonged CFTR bursts, even in the absence of PKA (n = 13, not shown), suggesting that NBD2 always participates
in closing the channel. We previously reported that the N-terminal half
of CFTR (which lacks NBD2) could produce functional
channels(35) . Unfortunately that result does not address
whether or not a single NBD (NBD1) is sufficient for normal gating,
because the active channel was probably a dimer in which a second NBD1
could substitute for NBD2. The requirement for NBD2 function has also
been suggested by the finding of Rich et al.(36) who
found that a CFTR construct in which NBD2 was deleted was properly
localized to the plasma membrane, but was not functional. Clearly, more
studies are required to resolve this issue. The model shown in Fig. 8does not explain some aspects of CFTR gating. One such
aspect is the requirement for the presence of PKA to see a stimulatory
effect of AMP-PNP. We and others (11, 12, 24) previously reported that when
channels had been phosphorylated, but were studied in the absence of
PKA, AMP-PNP had no effect. Using different preparations, Quinton (31, 32) and Hwang (33) found that the
presence of either cAMP agonists or PKA was required to see an effect
of AMP-PNP. These results suggest that PKA phosphorylates a labile site
which in some way alters the ability of AMP-PNP to interact
functionally with the NBDs. If AMP-PNP bound to, and prevented
hydrolysis at NBD1, then we might have expected to observe an increase
in the frequency of long closures between bursts. However that was not
observed. Perhaps the affinity of AMP-PNP for NBD1 is less than that
for NBD2; this difference might be assessed by biochemical binding
experiments. Alternatively, the affinity of AMP-PNP for both NBDs may
be equal, but an infrequent prolongation of the interval between bursts
may well be within the range observed in the absence of AMP-PNP. Another uncertainty of the model is that it assumes channel opening
and closing require hydrolysis of ATP. Our current findings and
previous data are consistent with the notion that these domains
hydrolyze ATP. In fact, it would be difficult to explain the data if
hydrolysis did not occur. Moreover, several other members of the ABC
transporter family have been shown to hydrolyze ATP. However, at
present there have been no biochemical studies that have directly
demonstrated hydrolysis by CFTR, and therefore we cannot completely
rule out mechanisms which do not involve hydrolysis. The model shown
in Fig. 8predicts that NBD2 bound to nucleoside triphosphate
(state 3) is the active state and that hydrolysis to nucleotide
diphosphate causes the transition to an inactive channel (state 4).
This scheme is reminiscent of the way that G proteins function. In G
proteins, the nucleoside triphosphate (GTP) bound state is the active
state, and the protein remains active as long as bound nucleoside
triphosphate remains unhydrolyzed (reviewed in (37) ). Once
hydrolysis occurs, the G protein with bound nucleoside diphosphate
(GDP) becomes inactive and remains inactive until GDP is replaced with
GTP. In this way, the hydrolysis of GTP acts as a timing mechanism.
When the rate of hydrolysis of GTP is reduced, as occurs in oncogenic ras mutations, the protein remains in the active state for a
prolonged duration. Likewise, binding of nonhydrolyzable GTP analogs,
such as GMP-PNP, produce a prolonged activation. These findings appear
to parallel what we observed with Walker lysine mutations in NBD2
(functionally analogous to ras mutations) and with AMP-PNP
(analogous to GMP-PNP). With both interventions, exit from the active
state was delayed and the duration of bursts of activity was prolonged.
This analogy and speculation suggests the intriguing possibility that
other domains of CFTR, or other proteins, might act in a manner similar
to the guanine nucleotide exchange factors or GTPase-activating factors
(GAPs) to modulate CFTR Cl The
suggestion that NBD2 might have a function similar to that of a G
protein might also explain the effect of ADP. Our previous data
indicated that ADP was a relatively potent inhibitor, prolonging the
duration of the long closed state between bursts(29) . Mutation
of Lys In summary, the data
indicate that the two NBDs in CFTR have different functions in
controlling channel opening and channel closing and suggest that
hydrolysis of ATP may be necessary for both processes. The model we
have proposed explains many aspects of the data, but more importantly,
it provides a framework from which future experiments can be designed
to further elucidate the novel regulation of CFTR.
Volume 270,
Number 4,
Issue of January 27, 1995 pp. 1711-1717
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS AND DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
channel contains two cytoplasmic
nucleotide-binding domains (NBDs). After phosphorylation of the R
domain, ATP interacts with the NBDs to regulate channel activity. To
learn how the NBDs regulate channel function, we used the patch-clamp
technique to study CFTR and variants which contained site-directed
mutations in the conserved Walker A motif lysine residues in either
NBD1 (K464A), NBD2 (K1250A and K1250M), or both NBDs simultaneously
(K464A/K1250A). Studies in related proteins suggest that such mutations
slow the rate of ATP hydrolysis. These mutations did not alter the
conductive properties of the channel or the requirement for
phosphorylation and ATP to open the channel. However, all mutations
decreased open state probability. Mutations in NBD1 decreased the
frequency of bursts of activity, whereas mutations in NBD2 and
mutations in both NBDs simultaneously prolonged bursts of activity, as
well as decreased the frequency of bursts. These results could not be
attributed to altered binding of nucleotide because none of the mutants
studied had reduced 8-N
ATP binding. These data suggest that
the two NBDs have distinct functions in channel gating; ATP hydrolysis
at NBD1 initiates a burst of activity, and hydrolysis at NBD2
terminates a burst.
)is a Cl channel with novel regulatory
mechanisms that are attributed to the cytoplasmic R domain and two
cytoplasmic nucleotide-binding domains (NBDs) (for reviews, see (1) and (2) ). Phosphorylation of the R domain by
cAMP-dependent protein kinase (PKA) or by protein kinase C is required
for the channel to open(3) . Once the R domain is
phosphorylated, intracellular ATP interacts with the NBDs to regulate
channel activity. The NBDs of CFTR share sequence similarity with the
NBDs in a family of proteins called either the traffic ATPases or ATP
Binding Cassette (ABC) transporters(4, 5) . Data from
many members of the ABC transporter family have shown that the NBDs
hydrolyze ATP to drive active transport of substrate across cell
membranes. In CFTR, biochemical (6, 7, 8, 9) and functional (10) studies have shown that the NBDs interact directly with
ATP and its analogs. Although CFTR forms a Cl
channel
in which substrate (ion) flow is passive rather than active, functional
studies indicate that hydrolyzable nucleoside triphosphates and
divalent cations are required for the channel to
open(11, 12) . These considerations suggest that the
NBDs of CFTR control channel gating by hydrolyzing ATP.
h
D) motifs (where X refers to any amino acid and h refers to a hydrophobic residue)
and in a short sequence (LSGGQ) that has been called a linker region by
Ames (13) and which shares some similarity to sequences in G
proteins(4) . However, the overall amino acid homology between
the two NBDs of CFTR is only 29%. The difference in primary structure,
as well as functional studies(10) , suggest that the two NBDs
may have different functions. But how ATP controls channel opening and
closing through the NBDs and what roles each of the two NBDs play in
channel regulation is not known. channels
containing mutations of the lysine in the Walker A motif. The
importance of the Walker lysine is reflected by the fact that it is
absolutely conserved in members of the ABC transporter
family(4, 5) . Crystallographic and NMR studies of
adenylate kinase have suggested that the Walker lysine comes into
contact with either the
- or -phosphate of the bound ATP (14, 15, 16) and is important for hydrolytic
function(16) . Functional studies of many ATP binding and
hydrolyzing enzymes demonstrate that mutation of the invariant Walker
lysine decreases the rate of ATP hydrolysis, while often maintaining
ATP binding and the ability to undergo ATP binding-induced
conformational changes. For example, mutation of the Walker lysine in
adenylate kinase to methionine decreased k
over
a thousandfold with only a slight change in K
, and a very small change in
G, suggesting that these changes were not due
to destabilization of overall structure (17) . Experiments with
a member of the ABC transporter family, the mdr1 gene product,
demonstrated that mutation of the Walker lysine in either of the two
NBDs individually or simultaneously abolished drug resistance without
altering binding to 8-N
ATP(18) .) and
NBD2 (Lys
) were mutated either individually or
simultaneously to either alanine or methionine. We studied these
variants using the whole cell and excised inside-out configurations of
the patch-clamp technique.
Site-directed Mutagenesis
CFTR mutants were
constructed in the vaccinia virus expression plasmid pTM-CFTR4 (19) by the method of Kunkel(20) . A mutant containing
two alterations was generated by simultaneously, including two
mutagenic oligonucleotides in the reaction. Single letter amino acid
abbreviations are used to refer to the residues mutated. For example,
in the mutant K464A, the lysine residue at position 464 in wild-type
CFTR is replaced with alanine. Mutants were verified by restriction
enzyme analysis, DNA sequencing around the site of mutation, and in
vivo expression. Some functional aspects of K464A, K1250M, K1250A,
and R-S660A function have been described
previously(10, 21, 22) .Cells and Transfection Procedure
Wild-type CFTR
was studied in either stably transfected NIH 3T3 fibroblasts, C127
mouse mammary epithelia cells, or transiently transfected HeLa cells as
described previously(10) . Similar results were obtained with
all three cell types, and the data are combined. Mutant CFTR was
studied in transiently transfected HeLa cells with the vaccinia
virus/bacteriophage T7 hybrid expression system as described previously (23, 24) . Briefly, cells were infected with a
recombinant viruses (10-20 multiplicity of infection each),
containing the bacteriophage T7 RNA polymerase driven by a viral
promoter. A second recombinant virus containing CFTR cDNA driven by a
T7 promoter was added simultaneously, or a plasmid containing the T7
promoter was transfected using Lipofectin (Life Technologies, Inc.) 1 h
after infection. HeLa cells were plated at approximately 5 
10 cells/cm
on collagen-coated glass coverslips
24 h before infection and were studied 18 h after infection. All
mutants produced fully glycosylated mature protein (band C) as assessed
by immunoprecipitation and phosphorylation.Patch-clamp Technique
Methods for excised,
inside-out, and whole cell patch-clamp recording are similar to those
described previously(11, 24, 25) . An
Axopatch 200 amplifier (Axon Instruments, Inc.) was used for voltage
clamping and current amplification. A microcomputer (IBM AT compatible)
and the pClamp software package (Axon Instruments, Inc.) were used for
data acquisition and analysis. Data were recorded on videotape
following pulse code modulation using a PCM-2 A/D VCR adapter (Medical
Systems Corp., Greenvale, NY). Patch pipettes were fabricated as
described(24) , with seal resistance routinely greater than 5
gigaohms. Voltages are referenced to the extracellular side of the
membrane. Whole cell and excised macropatch experiments were performed
at a holding potential of -40 mV; single channel data were
recorded at a holding potential of -80 mV. Experiments were
conducted at 34-36 °C using a temperature-controlled
microscope stage (Brook Industries, Lake Villa, IL).
(the time
which delineates intraburst from interburst closures) of 20 ms using
pClamp 6.0 software. This value was derived from analysis of single
channel closed time histograms (see ``Results,'' Fig. 5A) and by the method of Sigurdson et
al.(27) . Closures longer than 20 ms were considered to
define gaps between bursts, whereas closures shorter than this time
were considered gaps within bursts. In experiments where the membrane
patch contained five or fewer active channels, bursts in which only one
channel was open, which were separated from other bursts by greater
than 20 ms, and which had no superimposed openings were included in the
analysis. There was no statistical difference between burst durations
derived from patches with only one active channel compared with patches
with more than one active channel, nor was there a consistent trend
toward an increase or decrease of burst values by inclusion of data
from patches with more than one channel.
, was
determined as described under ``Results.'' B, mean
channel burst duration for wild-type CFTR and each Walker lysine
mutant. All measurements were made in the presence of 1 mM ATP
and 75 nM PKA with the membrane voltage clamped at -80
mV. Burst durations were determined with a t of 20
ms, as described under ``Experimental Procedures.'' For
wild-type, K464A, K1250M, K1250A, and K464A/K1250A, n =
15, 12, 5, 16, and 6, respectively; p < 0.01 relative to
wild type for all.
[
Photolabeling of membrane-associated CFTR was
performed as described previously(9) , except CFTR was
expressed in HeLa cells grown on 150-mm dishes. Membranes were prepared
by differential centrifugation and resuspended in 20 mM HEPES,
pH 7.5, 50 mM NaCl, 3 mM MgSO-P]8-N
ATP
Photolabeling, with 2
µg/ml each of leupeptin, aprotinin, and pepstatin.
Membrane-associated CFTR was phosphorylated prior to photolabeling by
incubation with 100 nM PKA and 0.1 mM ATP in buffer
containing 10 mM MgCl and 50 mM PIPES, pH
6.8, for 20 min at 30 °C, then diluted in 12 volumes of buffer
containing 20 mM Tris, pH 7.5, 0.2 mM EGTA, and 1
µM calyculin A. Membranes were collected by centrifugation
and resuspended at a concentration of 2.5 mg of protein/ml in the same
buffer.-P]8-N
ATP (6-12 Ci/mmol)
diluted in 20 mM HEPES, pH 7.5, 50 mM NaCl, 3 mM MgSO, with 2 µg/ml each of leupeptin, aprotinin,
and pepstatin in a total volume of 30 µl. After UV irradiation for
60 s, CFTR was solubilized and immunoprecipitated as described (9, 28) using antibodies raised against the R domain
(M13-1, 0.5 µg/sample) and against the C terminus (M1-4, 5
µg/sample). Immunocomplexes were analyzed by SDS-polyacrylamide gel
electrophoresis and incorporation of
[-P]8-N
ATP was quantitated with
an AMBIS radioanalytic imaging system (AMBIS Systems, Inc., San Diego,
CA).Chemicals and Solutions
Catalytic subunit of PKA
was obtained from Promega Corp. (Madison WI).
[-P]8-N
ATP was from ICN
Radiochemicals (Irvine, CA). Adenosine 5`-triphosphate (ATP; disodium
salt), 5`-adenylylimidodiphosphate (AMP-PNP, lithium salt), and all
other reagents were from Sigma., 2 MgCl, 10
HEPES, pH 7.3, with 1 N NaOH. The bath (intracellular)
solution contained (in mM): 140 N-methyl-D-glucamine, 135.5 HCl, 3 MgCl,
10 HEPES, 4 cesium, and 1 EGTA, pH 7.3, with 1 N HCl
([Ca]
< 10
M). For whole cell experiments, pipette (intracellular)
solution contained (in mM): 120 N-methyl-D-glucamine, 115 aspartic acid, 3
MgCl
, 4 cesium, 1 EGTA, and 1 mM NaATP, pH 7.3, with 1 N HCl
([Ca]
< 10
M). Bath (extracellular) solution contained (in
mM): 140 NaCl, 10 HEPES, 1.2 MgSO
, 1.2
CaCl, and 30 mM sucrose, pH 7.3, with 1 N NaOH.
Effect of Walker A Motif Lysine Mutations on Channel
Gating
Previous reports have shown that the CFTR variants K464A,
K1250M, and K1250A produced functional
channels(10, 22) . To learn whether a variant
containing mutations in both Walker lysines could function, we studied
K464A/K1250A. Cells expressing K464A/K1250A produced cAMP-activated
Cl currents that retained many hallmarks of CFTR
channel activity. Whole cell currents were time- and
voltage-independent (Fig. 1A). Excised membrane patch
experiments showed that K464A/K1250A required PKA phosphorylation and
intracellular Mg
and ATP for activity (Fig. 1B) and had ion conduction properties identical
to that of wild-type CFTR (Fig. 1C). These findings
suggested that mutation of both Walker lysines did not cause global
alterations in CFTR channel structure or a loss of the normal
requirements for function.
in bath (cytosolic) solution. The Mg
-free
solution contained 1 mM EDTA. C, single channel
current-voltage relationship. Single channel conductance was unchanged
relative to wild-type (9.7 ± 0.7 picosiemens, n = 4 versus 10.1 ± 0.4 picosiemens, n = 5, respectively). Data are mean ± S.E., n = 4, except +100 mV, where n = 2. Some
error bars are hidden by data symbols.
.
).
channel gating, we analyzed single channel
open state probability (P
), single channel open
and closed times, and the duration of bursts. Fig. 3shows that
in the presence of 1 mM ATP and 75 nM PKA, all of the
variants had a reduced P compared with that of
wild-type CFTR (0.44 ± 0.02, n = 15). Mutation
of NBD1 decreased P more than mutations in NBD2,
although the difference did not achieve statistical significance
between K464A and K1250A. Interestingly, when the K464A and K1250A
mutations were combined in the same molecule (K464A/K1250A), P (0.27 ± 0.05, n = 7) was
greater than that observed with the K464A mutation alone (0.13 ±
0.02, p = 0.002, n = 12).
for
wild-type CFTR and each Walker lysine mutant. All measurements were
made in the presence of 1 mM ATP and 75 nM PKA with
membrane potential clamped at -80 mV. For wild-type CFTR, K464A,
K1250M, K1250A, and K464A/K1250A, n = 15, 12, 6, 15,
and 9. Asterisks indicate statistical significance (p < 0.001) relative to wild type.
, we
analyzed open and closed time histograms from experiments in which the
membrane patch contained only one channel. As we have described
previously, open and closed time histograms were best fit with one and
two component fits, respectively(29) . The open time constant
(), which describes the openings within a burst of
activity, decreased to a roughly similar extent for all the mutants (Fig. 4A). In addition, the fast closed time constant
(), which represents the brief closures within a
burst of activity, approximately doubled in all the mutants (Fig. 4B).
is the open time constant; and
are the fast and slow time constants from the closed time
histogram, respectively. Data for mutants K1250A and K1250M were
combined (K1250A/M; n = 1 and 2,
respectively). For wild-type, K464A, and K464A/K1250A, n = 6, 6, and 4, respectively. Asterisks indicate
statistical significance (p < 0.05) relative to wild
type.
, a much more
pronounced effect resulted from an increase in the slow closed time
constant (), which describes the long closures
between bursts of activity (Fig. 4C). The NBD1 mutation
increased
5-fold, whereas similar mutations in NBD2
or both NBDs together produced a 30-50-fold increase. The finding
that channels with a mutation in NBD2 have a similar or higher P
than K464A despite the fact that they remained
closed longer is explained by the observation that the bursts of
activity in channels with a mutated NBD2 are much longer than those of
wild-type or K464A (see Fig. 2). (the time
which separates interburst closures from intraburst closures) of 20 ms.
This value was derived from analysis of wild-type CFTR closed time
histograms derived from excised inside-out membrane patches containing
a single channel studied in the presence of 1 mM ATP plus PKA.
When plotted with a logarithmic x axis (Fig. 5A), the two closed states appear as distinct
populations and the time constants differ by greater than 2 orders of
magnitude (average fast closed time constant, ,
= 1.79 ± 0.22 ms, average slow closed time constant,
, = 187 ± 13 ms, n =
6). The burst delimiter, t
, was chosen as the
nadir between the two populations of closures. The large difference
between time constants suggests that misclassification errors made when
defining bursts should be small(30) . A value of 20 ms was also
derived using the method of Sigurdson et al.(27) . Effect of Mutating the Walker A Motif Lysines on
8-N
Mutation of the Walker lysine slows
the rate of hydrolysis in proteins that hydrolyze ATP. Although such
mutations do not alter nucleotide affinity is some systems, in others
they have produced either a decreased or increased affinity. We
considered the possibility that mutation of the Walker lysines in CFTR
might alter ATP binding affinity and thereby alter gating. Therefore,
we measured incorporation of the photoreactive ATP analog
8-NATP BindingATP into membrane-associated wild-type and mutant CFTRs.
Our previous studies showed that membrane-associated wild-type CFTR was
labeled in a specific and saturable way by 8-NATP, that
8-NATP supported channel function, and that ATP competed
with 8-NATP(9) , suggesting that 8-NATP
binding is an appropriate assay of the functional ATP binding sites of
CFTR. to methionine, did not significantly alter photolabeling by
8-N
ATP relative to that of wild-type (Fig. 6).
Because we were not able to quantitate the amount of CFTR present, we
could not determine the stoichiometry of binding. However, the absolute
amounts of radiolabel incorporated were similar, suggesting that
mutation of the Walker lysines did not change the number of ATP binding
sites. These data suggest that the change in gating observed in the
Walker lysine mutants was not due to an altered affinity of CFTR for
ATP, but instead due to a change in some other intrinsic activity of
the NBDs, perhaps ATP hydrolysis.
ATP to wild-type
and variant CFTR. Binding of
[-P]8-N
ATP to
membrane-associated wild-type CFTR, K464A/K1250A, and K1250M, n = 9, 11, and 3, respectively. Data are expressed as amount
of radiolabel incorporated relative to the amount bound at 20
µM to correct for variations in
[-P]8-N
ATP specific activity
among experiments. A similar amount of photolabel incorporation was
observed for all groups. Immunoprecipitation and phosphorylation
suggest that similar amounts of wild-type or mutant protein were
present in each reaction.
Effect of AMP-PNP on Channel Gating
If the
speculation that termination of a burst of activity requires ATP
hydrolysis at NBD2 is correct, then we would predict that a
nonhydrolyzable analog of ATP, such as AMP-PNP, might also interact
with NBD2, inhibit the hydrolysis step, and prolong the duration of
bursts. Previous studies have shown that AMP-PNP cannot support channel
activity on its own(12, 24) . However, Quinton and
co-workers (31, 32) have shown that AMP-PNP in the
presence of ATP and cAMP agonists can increase Cl current through the apical membrane in sweat gland duct and T84
intestinal epithelia. In addition, Hwang et al.(33) have reported that AMP-PNP could increase the
activity of CFTR Cl
channels in guinea pig cardiac
myocytes when the channels were studied in the presence of ATP plus PKA
at 25 °C. Therefore, we tested the hypothesis that AMP-PNP would
prolong the burst duration when it was added in the presence of ATP and
PKA.
increased 35% (p = 0.006) (Fig. 7E).
(E). Asterisks indicate statistical significance, p = 0.033 and p = 0.006, for burst and P,
respectively; n = 6.
A Model of ATP Regulation of CFTR by the NBDs
Our
data allow us to assign very different functions to the two NBDs, and
we conclude that hydrolysis is involved in both channel opening as well
as closing. To interpret our present data, as well as previous
observations, we propose the model shown in Fig. 8, in which
each of the NBDs have distinct functions in channel gating. Note that
this model refers to the gating of CFTR by the NBDs after the R domain
has been phosphorylated. The model shows the interaction of ATP with
NBD1, with NBD2, and the resulting effect on channel gating (on the right). Four states in the gating cycle are shown.
(6, 8, 9) ,
yet Mg
is required for channel opening(11) .
These results suggest that a step subsequent to ATP binding is required
to open the channel.
ATP, and the fact that mutation of
Walker lysines in the NBDs of many other ATPases slows the rate of
hydrolysis, support the speculation that the transition from a closed
to bursting state requires hydrolysis at NBD1. are dominant in determining burst duration. When both
Lys
and Lys
are mutated in the same
protein, the burst duration is the same as that observed in K1250A or
K1250M. That is, once the channel is open in a bursting mode, an active
step is required to close it.
slightly decreased burst duration. Overall, these
results suggest that although each NBD may have a discrete function in
channel gating, the function of one NBD may influence or modulate the
function of the other.
channel activity.
to methionine abolished inhibition by ADP,
suggesting that ADP-dependent inhibition was mediated through NBD2. By
analogy, in G proteins nucleoside diphosphate (GDP) potently inhibits
activity by binding to the protein and thereby limiting the rate of
nucleotide exchange(38, 39) . Thus, we speculate that
in CFTR, ADP bound to NBD2 in state 4 might prevent the transition back
to state 1 and then on to states 2 and 3.
)
We thank Dr. David Gadsby for providing preprints
prior to publication, Dr. John Marshall and Dr. Seng Cheng for the gift
of M13-1 antibody, the University of Iowa DNA Core facility for
oligonucleotide synthesis and DNA sequencing, L. G. DeBerg for the
construction of mutants, S. P. Weber, P. H. Karp, S. R. Struble, J. A.
Cieslak and E. C. Kasik for excellent technical assistance, T. A.
Mayhew and D. R. Vavroch for secretarial assistance, and our laboratory
colleagues for their advice and critical comments.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  H. F. Bao, L. Liu, J. Self, B. J. Duke, R. Ueno, and D. C. Eaton A synthetic prostone activates apical chloride channels in A6 epithelial cells Am J Physiol Gastrointest Liver Physiol, August 1, 2008; 295(2): G234 - G251. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T.-Y. Chen and T.-C. Hwang CLC-0 and CFTR: Chloride Channels Evolved From Transporters Physiol Rev, April 1, 2008; 88(2): 351 - 387. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Csanady, A. C. Nairn, and D. C. Gadsby Thermodynamics of CFTR Channel Gating: A Spreading Conformational Change Initiates an Irreversible Gating Cycle J. Gen. Physiol., November 1, 2006; 128(5): 523 - 533. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Zhou, X. Wang, H.-Y. Liu, X. Zou, M. Li, and T.-C. Hwang The Two ATP Binding Sites of Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) Play Distinct Roles in Gating Kinetics and Energetics J. Gen. Physiol., October 1, 2006; 128(4): 413 - 422. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Csanady Statistical Evaluation of Ion-Channel Gating Models Based on Distributions of Log-Likelihood Ratios Biophys. J., May 15, 2006; 90(10): 3523 - 3545. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Cai, A. Taddei, and D. N. Sheppard Differential Sensitivity of the Cystic Fibrosis (CF)-associated Mutants G551D and G1349D to Potentiators of the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) Cl- Channel J. Biol. Chem., January 27, 2006; 281(4): 1970 - 1977. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z.-R. Zhang, B. Song, and N. A. McCarty State-dependent Chemical Reactivity of an Engineered Cysteine Reveals Conformational Changes in the Outer Vestibule of the Cystic Fibrosis Transmembrane Conductance Regulator J. Biol. Chem., December 23, 2005; 280(51): 41997 - 42003. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Fuller, Z.-R. Zhang, G. Cui, and N. A. McCarty The Block of CFTR by Scorpion Venom is State-Dependent Biophys. J., December 1, 2005; 89(6): 3960 - 3975. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Zhou, X. Wang, M. Li, Y. Sohma, X. Zou, and T.-C. Hwang High affinity ATP/ADP analogues as new tools for studying CFTR gating J. Physiol., December 1, 2005; 569(2): 447 - 457. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G. Bompadre, T. Ai, J. H. Cho, X. Wang, Y. Sohma, M. Li, and T.-C. Hwang CFTR Gating I: Characterization of the ATP-dependent Gating of a Phosphorylation-independent CFTR Channel ({Delta}R-CFTR) J. Gen. Physiol., March 28, 2005; 125(4): 361 - 375. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G. Bompadre, J. H. Cho, X. Wang, X. Zou, Y. Sohma, M. Li, and T.-C. Hwang CFTR Gating II: Effects of Nucleotide Binding on the Stability of Open States J. Gen. Physiol., March 28, 2005; 125(4): 377 - 394. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. L. Berger, C. O. Randak, L. S. Ostedgaard, P. H. Karp, D. W. Vermeer, and M. J. Welsh Curcumin Stimulates Cystic Fibrosis Transmembrane Conductance Regulator Cl- Channel Activity J. Biol. Chem., February 18, 2005; 280(7): 5221 - 5226. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
