(Received for publication, March 17, 1997, and in revised form, May 14, 1997)
From the Department of Physiology and Biophysics and the ¶ Department of Neurology, Howard Hughes Medical Institute, University of Iowa College of Medicine, Iowa City, Iowa 52242
The role of the extracellular domain of the
voltage-dependent Ca2+ channel
2
subunit in assembly with the
1C
subunit was investigated. Transiently transfected tsA201 cells
processed the
2
subunit properly as disulfide
linkages and cleavage sites between the
2 and
subunits were shown to be similar to native channel protein. Coimmunoprecipitation experiments demonstrated that in the absence of
subunits,
2 subunits do not assemble with
1 subunits. Furthermore, the transmembrane and
cytoplasmic sequences in
can be exchanged with those of an
unrelated protein without any effect on the association between the
2
and
1 proteins. Extracellular
domains of the
2
subunit are also shown to be
responsible for increasing the binding affinity of
[3H]PN200-110
(isopropyl-4-(2,1,3-benzoxadiazol-4-yl)-1,4-dihydro-2,6-dimethyl-5-([3H]methoxycarbonyl)-pyridine-3-carboxylate)
for the
1C subunit. Investigation of the
corresponding interaction site on the
1 subunit revealed
that although tryptic peptides containing repeat III of native
1S subunit remain in association with the
2
subunit during wheat germ agglutinin
chromatography, repeat III by itself is not sufficient for assembly
with the
2
subunit. Our results suggest that the
2
subunit likely interacts with more than one extracellular loop of the
1 subunit.
The 2
subunit has been identified in every
voltage-dependent Ca2+ channel purified to date
from various mammalian tissues, including skeletal muscle (1, 2), brain
(3, 4), and heart (5, 6). Structurally, the
2
subunit
is a heavily glycosylated 175-kDa protein that is encoded by a single
gene that is post-translationally cleaved to yield the disulfide-linked
2 and
proteins (7, 8). Experimental evidence
supports a single transmembrane topology of the
2
subunit in which all but the transmembrane sequence and 5 carboxyl-terminal amino acids are extracellular (9-11).
Coexpression of mRNA encoding the Ca2+ channel
2
subunit has been shown to modify many properties of
the
1 subunit, including increasing the macroscopic
current amplitude (12, 13), accelerating the activation (14) and
inactivation kinetics, and shifting the voltage dependence of
activation to more hyperpolarizing
potentials.1 However, the physical
structures and molecular interactions that mediate these effects are
entirely unknown.
Interaction sites on the pore-forming subunit have been identified
for several voltage-dependent ion channel auxiliary subunits, including
the Ca2+ and K+ channel
subunits. The
binding site of the Ca2+ channel
subunit has been
localized to a region of approximately 18 amino acids in the
1 subunit I-II cytoplasmic linker (15), and the
corresponding interaction site on the
subunit has also been
described (16). Interaction sites between K+ channel
and
subunits have been mapped to the amino-terminal A and B box
(17, 18) near the cytoplasmic region that is also responsible for the
subfamily-specific assembly of
subunit multimers (19, 20).
Unlike the previously described cytoplasmic interactions, assessment of
interactions between two transmembrane proteins has generally been more
challenging. Transmembrane proteins such as the Ca2+
channel 2
subunit are often extensively glycosylated,
which may preclude the use of bacterial, insect, or in vitro
expression systems because glycosylation is frequently
species-dependent. Likewise, the expression and correct
formation of disulfide linkages is also difficult to reproduce in an
in vitro expression system. Also, although there are reports
of successful uses of the two-hybrid yeast expression system to map
interaction sites of two transmembrane proteins (21), these have often
been performed on a more limited basis after initial investigations
localized interaction domains using mammalian expression systems. Using
transiently transfected human tsA201 cells, we have implicated the
extracellular domain of the
2
subunit in the assembly
with the
1 subunit and have also shown that this region
is responsible for modulation of dihydropyridine binding affinity to
the
1 subunit.
tsA201 cells (SV40 large T antigen transformed HEK 293 cells) (Cell Genesis, Foster City, CA) were maintained at 5% CO2 in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, 2 mM L-glutamine, 50 units/ml penicillin, and 50 µg/ml streptomycin. Transfections were performed using the calcium phosphate method on 50-70% confluent cells. Generally 30 µg of each channel subunit DNA (for 150-mm dish) was added to 1.25 ml of 250 mM sterile filtered CaCl2. An equal volume of 2 × sterile HEBS (274 mM NaCl, 40 mM HEPES, 12 mM dextrose, 10 mM KCl, 1.4 mM Na2HPO4, adjusted to final pH 7.05) was added drop by drop to the Ca2+/DNA mixture with constant agitation. The precipitate was allowed to form for 30 min and added dropwise to the plated cells. The medium was changed the next day.
Construction of Plasmids for Mammalian Cell TransfectionThe cDNA encoding the rat brain
2
subunit and truncated forms were all transferred to
pcDNA3 (Invitrogen) and have been described previously (10). The
1S repeat III was created by polymerase chain reaction
utilizing a forward primer beginning at nucleotide 2544 and a reverse
primer beginning at nucleotide 3489. A Kozak initiation start
consisting of CCACCATGG (where the methionine start site is
underlined) was created in the forward primer along with a
KpnI site for insertion into polylinker of pcDNA3. The
reverse primer contained an in frame termination site and a
XbaI site for ligation.
tsA201 cells were harvested 48 h after transfection by washing two times with 10 ml of phosphate-buffered saline and collected by centrifugation at 3,000 rpm for 5 min. Cell membranes were prepared immediately by resuspending cell pellet from one 150-mm plate in 20 ml of ice-cold hypotonic lysis buffer (10 mM Tris, pH 7.4 with 0.64 mM benzamidine, and 0.23 mM PMSF).2 After a 15-min incubation on ice, swollen cells were disrupted by five strokes with a Dounce homogenizer. Lysed cells were centrifuged at 3,000 rpm for 10 min at 4 °C. The supernatant was then centrifuged at 35,000 rpm for 37 min to collect the membranes. The membrane pellet was resuspended in 1 ml of Buffer I (0.3 M sucrose, 20 mM Tris, pH 7.4, 1.0 mM PMSF, and 0.75 mM benzamidine) and passed through a 28 gauge needle.
Binding AssaysBinding assays were performed in 50 mM Tris, pH 7.4, 0.23 mM PMSF, 0.64 mM benzamidine, and 1.0 mg/ml bovine serum albumin (binding buffer) in a final assay volume of 500 µl. For saturation analysis, 0.05-2 nM (+)-[3H]PN200-110 (Amersham Corp.) were incubated in the dark with 80 µg of membrane protein for 60 min at 37 °C. Nonspecifically bound ligand was determined by the addition of 50 µM nitrendipine. Specific binding sites were determined by subtracting nonspecific binding from total binding. Radiolabeled membranes were washed three times with 5 ml of ice-cold binding buffer on a GF/B glass fiber filter (Whatman) using a Brandel cell harvester (Brandel, Gaithersburg, MD). Data were fitted by a single-site binding model applying nonlinear regression analysis using GraFit software (Trithacus Software, Staines, UK).
ImmunoprecipitationCell membranes (500 µg) were
solubilized in 1 ml of total volume of 1% (w/v) digitonin and 1 M NaCl (final concentrations) for 1 h at 4 °C on a
rolling platform. Protease inhibitors were added at a concentration of
0.23 mM PMSF and 0.64 mM benzamidine. Solubilized protein was isolated by centrifugation at 50,000 rpm for 15 min at 4 °C and subsequently diluted 2-fold with ice-cold double
distilled H20. Solubilized protein was added to 30 µl of protein G-Sepharose that had been preincubated overnight with sheep 41 antiserum (1C II-III loop) (4) or IIF7 ascites (2).
Rabbit skeletal muscle triads (22) were resuspended
to a concentration of 4 mg/ml in Buffer I. Trypsin concentrations were varied from 1:1000 to 1:10 (trypsin:protein), and the incubation time
was variable between 0 and 160 min at 37 °C. The reaction was
terminated by the addition of 1 mM PMSF. Trypsin digested triads were then solubilized in 1% digitonin and 0.5 M
NaCl for 1 h at 4 °C followed by centrifugation at 70,000 rpm
for 30 min. Solubilized protein was added to WGA-Sepharose (Sigma) and
incubated for 3 h at 4 °C. Bound protein was eluted in 300 mM N-acetylglucosamine in Buffer I. For sucrose
gradient fractionation, samples were concentrated to 600 µl in a
YM100 (Amicon) concentration unit. Samples were loaded onto 16-ml
linear gradients of 5-20% (w/w) sucrose in 100 mM NaCl,
50 mM Tris, pH 7.4, and 0.1 mM PMSF. Gradients were centrifuged in a Beckman SW 28 rotor with SW 28.1 buckets for
2 h at 150,000 rpm at 4 °C. Gradients were then fractionated (1.2 ml) from the top with an Isco gradient fractionator, and 100 µl
of each was analyzed on a 5-16% SDS-polyacrylamide gel electrophoresis and stained with Coomassie Blue or immunoblotted with
monoclonal antibody IIF7 or rabbit 136 polyclonal against the
Ca2+ channel 2 subunit (23).
Monoclonal antibodies IIF7 and IIC12 (2) were used to
screen 2 × 104 clones of 1S subunit
epitope library in Y1090 Escherichia coli. Inserts were
amplified from pure phage positives by polymerase chain reaction using
primers directed to
gt11 phage arms. These were directly inserted
into a T-vector (made from Bluescript Sk
plasmid) for
sequencing. All inserts were sequenced using either the dideoxy chain
termination method Sequenase II (U. S. Biochemical Corp.) or automated
sequencer (Applied Biosystems, Inc.).
Because of the extensive post-translational processing events
involved in the formation of the 2
subunit
(N-linked glycosylation, disulfide linkages, and subunit
cleavage), we chose to utilize the mammalian tsA201 cell line for
expression. The endogenous proteolytic cleavage between the
2 and
subunits was investigated by Western blot
analysis of membranes from cells transfected with the full-length
2
subunit. In nonreducing conditions, an antibody directed against the
2 subunit recognized a protein of
175 kDa in both skeletal muscle triads and in membranes prepared from tsA201 cells transfected with the full-length
2
subunit (Fig. 1). One apparent difference between native
and transfected
2
protein is that the transfected
2
protein ran as a broader band. This may reflect
larger amounts of incompletely processed forms including untrimmed
glycosylation and immature noncleaved protein that often result from
transient expression. Cells expressing only the
2
subunit produced a protein migrating at 150 kDa in both the presence
and the absence of reducing agents, which is consistent with the
addition of more than 40 kDa of N-linked oligosaccharide. When identical cell membranes were electrophoresed in reducing conditions, native skeletal muscle
2
subunit shifted
to an apparent molecular mass of 150 kDa. Likewise, there was a
noticeable shift in molecular mass of the transfected
2
protein, suggesting that the cleavage site and
disulfide linkages are similar to native protein.
To test the involvement of the extracellular domain of the
2
subunit in the interaction with the
1 subunit, coimmunoprecipitation experiments were
performed using cells transfected with both
1C subunit
and either the full-length
2
subunit or any one of
the truncated
2
subunit constructs (Fig.
2A). Cell membranes were solubilized in 1%
digitonin and 1 M NaCl prior to immunoprecipitation. The
full-length
2
subunit assembles with the
1C subunit as demonstrated by its coimmunoprecipitation
with an anti-
1C antibody and detection by Western blot
analysis with an anti-
2 antibody (Fig. 2B).
No
2
protein was immunoprecipitated from control untransfected cells. In addition, no
2
protein was
precipitated from cells in the absence of the
1 subunit
(data not shown). Truncation of 450 extracellular amino-terminal amino
acids of the
2
subunit abolished the ability of this
protein to assemble with the
1C subunit, despite its
abundance in the starting material. Likewise, the
2
subunit expressed in the absence of the
subunit was also unable to
coimmunoprecipitate with the
1C subunit.
We also investigated the role of the transmembrane domain of the subunit in assembly with the
1C subunit (Fig.
2B). Substitution of the transmembrane domain from adhalin,
an unrelated type I transmembrane protein (recently renamed
-sarcoglycan), did not appear to alter the ability of the protein to
assemble with the
1C subunit. In this chimera, the 5 cytoplasmic amino acids of the
2
protein were also
substituted with adhalin sequence. Therefore, we conclude that neither
intracellular or transmembrane sequences of the
2
subunit are required for interaction with the
1
subunit.
The region of the 2
subunit responsible for
modulation of dihydropyridine binding to the
1C subunit
was also investigated. Although [3H]PN200-110 binding to
whole cell tsA201 cell membranes was often low and nonsaturable when
1C was transfected in the absence of any auxiliary
subunit, several experiments resulted in significant and saturable
binding that allowed us to determine the binding affinity
(Kd) and binding capacity
(Bmax) using saturation analysis (Table
I). Cells expressing
1C alone had an
average Bmax of 94.6 ± 51.7 fmol/mg
(n = 4).
|
Coexpression of the full-length 2
subunit with the
1C subunit resulted in a significant increase in
binding, most of which could be accounted for by a significant mean
increase in the binding affinity (Table I). Binding was saturable in
all experiments. There appeared to be little effect of the
2
subunit on Bmax (Bmax = 133 ± 73.5 fmol/mg), although
there was significant error between experiments in the
Bmax depending on the transfection efficiency.
Likewise, Western blot analysis on whole cell membranes from
transfected cells showed no effect of coexpression of the
2
subunit on the protein expression of the
1 subunit (data not shown). The binding affinity,
however, was not affected by the differences in transfection
efficiency.
As expected, when the 2 subunit was coexpressed with the
1C subunit, there was no effect on
[3H]PN200-110 binding. This is consistent with
coimmunoprecipitation experiments that demonstrated the inability of
the
2 subunit to associate with
1 in the
absence of the
subunit. However, coexpression of the
2
Ad chimera, in which the transmembrane domain of the
2
subunit was replaced with that of adhalin,
increased [3H]PN200-110 binding affinity to approximately
the same extent as full-length
2
protein.
Because the 1 subunit is very large and difficult to
express, we chose an alternative approach to identify regions
interacting with the
2
subunit. Our approach was to
trypsinize skeletal muscle microsomes containing native dihydropyridine
receptors and follow the
1S subunit fragments remaining
in association with the
2
subunit during WGA affinity
chromatography. By taking advantage of the selective ability of the
glycosylated
2
subunit to bind WGA, any
1S fragment identified is presumed to bind WGA only
through its interaction with the
2
subunit. With
increasing concentrations of trypsin, an
1S
subunit-specific monoclonal antibody that recognizes an epitope within
the first extracellular loop of the IIIS5-IIIS6 linker (amino acid
955-1005) (IIF7) detected 28- and 18-kDa
1S subunit
fragments eluted from a WGA-Sepharose column (Fig. 3).
Sucrose gradient fractionation was subsequently used to demonstrate
cosedimentation of the
1S subunit fragments with the
intact full-length
2
subunit (Fig. 4).
Multiple tryptic fragments of
1 did not bind WGA and
were identified in the starting material, including carboxyl-terminal
fragments (identified by monoclonal antibody IIC12) and fragments of
repeat I and II and the II-III loop (identified by polyclonal antibody
sheep DHPR) (data not shown).
To test the ability of 1S repeat III to associate with
the
2
subunit, we cotransfected tsA201 cells with
constructs containing only repeat III and the full-length
2
subunit. Although the
2
subunit
and repeat III were well expressed, we were unable to detect stable
interactions between these two proteins using coimmunoprecipitation
assays after solubilization in 1% digitionin and 1 M NaCl
(data not shown). This suggests that expression of the
1S subunit repeat III by itself is not sufficient to
form stable interactions with the
2
subunit.
Our data support a model whereby the interaction sites between the
2
and
1 subunits are entirely
extracellular, because transmembrane modifications of the
2
subunit did not appear to alter coassembly with the
1 subunit. Morever, our data suggest a requirement for
nontransmembrane domains of the
subunit in determining a stable
association between the
2
and
1
proteins, because the
2 protein by itself could not
support interaction.
may contain the interaction site, or the
tertiary structure it confers on
2 through its disulfide
linkages may enable
2 to directly interact with the
1 subunit. Low expression of
expressed alone
resulted in our inability to distinguish between these possibilities.
However,
was shown to be able to compete with full-length
2
protein in Xenopus oocytes and inhibit
its stimulatory effects on current amplitude (10), and expression studies in tsA201 cells demonstrated that coexpression of
can significantly modulate the biophysical properties of the
1C subunit.1
Interestingly, our data are consistent with reports regarding a
functionally related pair of proteins, the
Na+,K+-ATPase and auxiliary
subunit, in
which the interaction sites have also been localized to extracellular
domains (24, 25). In this case, the yeast two-hybrid assay was
successful in further localizing the site of interaction on the
subunit to the 61 amino acids most proximal to the membrane (21).
Although the interaction sites between the
voltage-dependent Na+ channel
and
1 or
2 subunits have not been mapped, it
is interesting to note that deletion of the
1
intracellular domain does not alter functional effects of
1 subunit coexpression (26), suggesting that this
interaction may also be in the extracellular domain.
Repeat III of the Ca2+ channel 1S subunit
appears to interact strongly with the
2
subunit after
extensive trypsinization, although we cannot exclude the involvement of
other unidentified fragments (especially in repeat IV) based on our
inability to recognize small tryptic fragments with specific
antibodies. Interestingly, whereas repeat III remains in association
with the
2
subunit after extensive trypsinization, we
were unable to reconstitute the interaction between this small region
and
2
in an expression system. This suggests that
multiple regions of the
1 subunit may be involved in
assembly with the
2
subunit. In analogous studies on
the voltage-dependent Na+ channel, multiple
domains within the carboxyl-terminal half of the skeletal muscle
Na+ channel
subunit were shown to be required for
functional response of the coexpressed Na+ channel
1 subunit on inactivation kinetics (27).
Association of the 2
subunit with the
carboxyl-terminal half of the
1 subunit is consistent
with the significant effects that we and others have measured of the
2
subunit on dihydropyridine binding affinity (28).
The membrane spanning segments IIIS6 and IVS6 of the
1S
subunit have recently been shown to contain amino acids critical for
dihydropyridine binding (29), although the S5-S6 extracellular linkers
of the III and IV repeats also confer dihydropyridine sensitivity (30).
The extracellular domain of the
2
subunit, which is
capable of modulating dihydropyridine binding, may be interacting with
sites at or near these dihydropyridine binding sites within the III and
IV repeats.
Based on several observations regarding the extracellular regions of
the 1 subunit, we can speculate on the exact sites of interaction. Most extracellular loops of the
1 subunit
are small in size, the smallest being only 7 amino acids. The largest
extracellular loops, and thus the regions with the highest probability
of interacting with the
2
subunit, are the S5-S6
linkers, which also contain the pore. Experimental evidence regarding
the folding pattern of the
1 subunit suggests that the
S5-S6 regions of all four repeats closely interact to form the central
pore (31). These amino acids near the pore, while neighboring each
other in tertiary structure, are far apart in primary structure, and
thus it may be difficult to reconstitute the structure of this region
by expression of a single repeat. Based on the substantial effects of
the
2
subunit on dihydropyridine binding affinity, we
predict that the smallest regions of interaction may be within the
S5-S6 extracellular loops, particularly of repeat III, because this
region copurifies with the
2
subunit on WGA
chromatography.
We thank the late Dr. Xiangyang Wei (Medical
College of Georgia) for providing the 1C cDNA clone
and Dr. Terry Snutch (University of British Columbia) for the
2
cDNA clone. We acknowledge the University of
Iowa Diabetes and Endocrinology Research Center, which is funded by
National Institutes of Health Grant DK25295, for cell culture
reagents and the University of Iowa DNA Core Facility for DNA
sequencing.