pH-Dependent Dimerization of the Carboxyl Terminal Domain of Cx43

Abstract

Previous studies have demonstrated that the carboxyl terminus of the gap junction protein Cx43 (Cx43CT) can act as an independent, regulatory domain that modulates intercellular communication in response to appropriate chemical stimuli. Here, we have used NMR, chemical cross-linking, and analytical ultracentrifugation to further characterize the biochemical and biophysical properties of the Connexin43 carboxyl terminal domain (S255-I382). NMR-diffusion experiments at pH 5.8 suggested that the Connexin43 carboxyl terminus (CX43CT) may have a molecular weight greater than that of a monomer. Sedimentation equilibrium and cross-linking data demonstrated a predominantly dimeric state for the Cx43CT at pH 5.8 and 6.5, with limited dimer formation at a more neutral pH. NMR-filtered nuclear Overhauser effect studies confirmed these observations and identified specific areas of parallel orientation within Cx43CT, likely corresponding to dimerization domains. These regions included a portion of the SH3 binding domain, as well as two fragments previously found to organize in alpha-helical structures. Together, these data show that acidification causes Cx43CT dimer formation in vitro. Whether dimer formation is an important structural component of the regulation of Connexin43 channels remains to be determined. Dimerization may alter the affinity of Cx43CT regions for specific molecular partners, thus modifying the regulation of gap junction channels.

Full text

pH-Dependent Dimerization of the Carboxyl Terminal Domain of Cx43
Paul L. Sorgen,* Heather S. Duffy,
y
David C. Spray,
y
and Mario Delmar
z
*Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha, Nebraska;
y
Department of
Neuroscience, Albert Einstein College of Medicine, Bronx, New York; and
z
Department of Pharmacology,
Upstate Medical University, Syracuse, New York
ABSTRACT Previous studies have demonstrated that the carboxyl terminus of the gap junction protein Cx43 (Cx43CT) can
act as an independent, regulatory domain that modulates intercellular communication in response to appropriate chemical
stimuli. Here, we have used NMR, chemical cross-linking, and analytical ultracentrifugation to further characterize the
biochemical and biophysical properties of the Connexin43 carboxyl terminal domain (S255-I382). NMR-diffusion experiments at
pH 5.8 suggested that the Connexin43 carboxyl terminus (CX43CT) may have a molecular weight greater than that of
a monomer. Sedimentation equilibrium and cross-linking data demonstrated a predominantly dimeric state for the Cx43CT at
pH 5.8 and 6.5, with limited dimer formation at a more neutral pH. NMR-filtered nuclear Overhauser effect studies confirmed
these observations and identified specific areas of parallel orientation within Cx43CT, likely corresponding to dimerization
domains. These regions included a portion of the SH3 binding domain, as well as two fragments previously found to organize in
a-helical structures. Together, these data show that acidification causes Cx43CT dimer formation in vitro. Whether dimer
formation is an important structural component of the regulation of Connexin43 channels remains to be determined.
Dimerization may alter the affinity of Cx43CT regions for specific molecular partners, thus modifying the regulation of gap
junction channels.
INTRODUCTION
Gap junctions allow for the direct exchange of ions and other
small molecules (,1kDa) between neighboring cells. These
intercellular channels are formed by oligomerization of con-
nexin proteins. Six connexins form a hemichannel, or
connexon. Two connexons, one provided by each apposing
cell, bind across the extracellular space to form a gap junction
channel. Recent studies show that not only the presence, but
also the regulation of gap junctions is essential for normal
tissue function. Our laboratories have been interested in
understanding the changes in high-order structure that
associate with the regulation of the major cardiac gap junction
protein, Connexin43 (Cx43)
1
(Duffy et al., 2002,2004;
Sorgen et al., 2002).
One intracellular stimulus that has long been known to
regulate gap junctions is pH. Acidification of the intracellular
space leads to a loss of intercellular communication (Spray
and Bennett, 1985; Francis et al., 1999; Stergiopoulos et al.,
1999). This phenomenon, commonly known as ‘‘pH
gating’’, is proposed to act as a substrate for the development
of malignant ventricular arrhythmias consequent to myocar-
dial ischemia (Kleber et al., 1987; Janse and Wit, 1989;
Cascio et al., 1995; Peters et al., 1997). Work preformed in
Xenopus oocytes showed that truncation of the carboxyl
terminal region of Cx43 (Cx43CT) prevented acidification-
induced channel closure (Morley et al., 1996). Coexpression
with the carboxyl terminal (CT) partially recovered pH
gating. This result led to the hypothesis that pH gating results
from a ‘‘ball-and-chain’’ type mechanism in which a gating
particle (the CT) binds to a separate region of the protein (a
‘‘receptor’’) to close the channel. This model also applies to
the regulation of Cx43 channels by other factors (Homma
et al., 1998; Zhou et al., 1999; Moreno et al., 2002). In
general, the CT region of Cx43 is seen as a regulatory
domain that can interact with the cytoplasmic microen-
vironment and modify intercellular communication ac-
cordingly.
A number of studies show that the CT domain remains
functional even if it is not fused with the pore-forming region
of the connexin protein. Indeed, the isolated CT domain
retains its ability to regulate Cx43 channels in response to pH
(Morley et al., 1996), insulin and insulin-like growth fac-
tor (Homma et al., 1998), c-Src (Zhou et al., 1999), and
transjunctional voltage (Moreno et al., 2002). Moreover,
several studies show that the Cx43CT fragment acts in vitro
as a substrate for a number of kinases thought to
phosphorylate Cx43 channels (Swenson et al., 1990; Gold-
berg and Lau, 1993; Saez et al., 1993; Moreno et al., 1994;
Loo et al., 1995; Cooper et al., 2000; Lampe and Lau, 2000),
and it can bind to known molecular partners of Cx43 such
as zonula ocludens-1 (Giepmans and Moolenaar, 1998;
Toyofuku et al., 2001), tubulin (Giepmans et al., 2001), and
the SH3 domain of Src (Kanemitsu et al., 1997; Duffy et al.,
2004). These results suggest that the CT domain retains at
least some of its functional properties when in isolation. In
vitro studies on the biophysical and biochemical properties
of the CT domain are therefore expected to shed light on the
molecular mechanisms regulating Cx43 gap junctions.
Submitted December 29, 2003, and accepted for publication March 23, 2004.
Address reprint requests to Paul L. Sorgen, Dept. of Biochemistry and Mo-
lecular Biology, University of Nebraska Medical Center, Omaha, NE 68198.
Tel.: 402-559-7557; Fax: 402-559-6650; E-mail: psorgen@unmc.edu.
Ó2004 by the Biophysical Society
0006-3495/04/07/574/08 $2.00 doi: 10.1529/biophysj.103.039230
574 Biophysical Journal Volume 87 July 2004 574–581

We have initiated a characterization of the high-order
structure of the Cx43CT domain, and backbone resonances
have been recently assigned (Sorgen et al., 2002). Moreover,
we have used NMR translational diffusion analysis to study
the mobility of Cx43CT using magnetic gradients (Duffy
et al., 2002). These data showed reduced diffusion velocity
of the Cx43CT fragment at low pH, which could be
explained by the formation of a higher molecular weight
species, perhaps consequent to oligomerization of the
fragment. In this study, we used cross-linking, analytical
centrifugation, and NMR to show that the Cx43CT domain
exists in a predominantly monomeric state at pH 7.5 and
a predominantly dimeric state at pH 6.5 and 5.8. We have
further identified the specific residues involved in dimeriza-
tion. These include a fragment of the SH3 binding domain,
as well as two regions thought to organize as a-helical
structures (Sorgen et al., 2002). We hypothesize that
dimerization of the CT domain may occur in functional
Cx43 channels and play a role in the regulation of gap
junctions in response to low pH.
MATERIALS AND METHODS
Expression and purification of recombinant
GST-Cx43CT
The Cx43CT,
15
N-Cx43CT, and
15
N
13
C-Cx43CT polypeptides were
expressed and purified as described previously (Duffy et al., 2002). All
polypeptides were confirmed for purity and analyzed for degradation by
SDS-PAGE, NMR, and mass spectroscopy. The analysis showed that
Cx43CT kept at 25°C and at pH 7.2 and 7.5 was stable for up to ;30 h; after
this time, protein cleavage was detected near residue G321, possibly due to
autoproteolyisis. All experiments were performed within a 24-h window.
Cross-linking
The cross-linking of Cx43CT was carried out for 1 h at room temperature
using 10 mM of 1-Ethyl-3-(3-Dimethylaminopropyl) carbodiimide Hydro-
chloride (EDC). The reactions occurred in phosphate-buffered saline (PBS)
at pH 5.8, 6.5, and 7.5 and were quenched by the addition of ethanolamine-
HCL to a final concentration of 100 mM. Complete quenching was achieved
by leaving the reactions standing for 10 min at room temperature followed
by heating in SDS sample buffer. The products of the reaction were then run
on 15% SDS-PAGE gels.
Immunoblot analysis
Complexes within the gel were electrophoretically transferred to nitrocel-
lulose membrane (Schleicher and Schuell, Keene, NH) and probed for Cx43
as previously described (Thi et al., 2003). Briefly, membranes were blocked
in 5% skim milk in phosphate-buffered saline (PBS) and probed for 1 h at
room temperature using polyclonal antibodies directed against the carboxyl
terminal domain of Cx43 (Zymed, South San Francisco, CA, and 181A, gift
from Elliot Hertzberg, AECOM, Bronx, NY) diluted in 5% skim milk in
PBS. After rinses in PBS with 0.05% Tween20 (PBST) membranes were
incubated with horseradish peroxidase-conjugated rabbit secondary IgGs
(Santa Cruz Biotech, Santa Cruz, CA). Protein bands were detected using
Amersham ECL detection kit (Amersham Biosciences, Piscataway, NJ) and
exposed on Fuji X-Ray film.
Analytical centrifugation
Sedimentation equilibrium experiments were performed using a Beckman
Optima XL-I analytical ultracentrifuge and an AN-60Ti rotor. The Cx43CT
was analyzed at 25°C in PBS buffer (pH 5.8, 6.5, 7.2, and 7.5). For each
condition, data were collected at three concentrations (A
280
¼0.3, 0.5, and
0.9) and two rotor speeds (18,000 rpm and 26,000 rpm). Absorbance scans at
280 nm were taken after 22 h and 24 h at each speed; it was assumed that
equilibrium was reached if the scans were unchanged. Data analysis was
performed using the Beckman XL-A/XL-I software package within
Microcal, ORIGIN v4 using values of the buffer density and protein partial
specific volume determined as described below. Each analysis consisted of
the six-absorbance scans taken at three different nominal concentrations and
at each of the two rotor speeds.
Buffer densities were determined using a Mettler DE40 density meter
operated at the experimental temperature and data were analyzed with the
program Sednterp v1.03. Partial specific volume was determined from
amino acid residue composition and calculated in Sednterp.
Nuclear magnetic resonance
NMR data were acquired at 7°C using a Bruker DRX-600 spectrometer
fitted with a triple resonance probe and triple axis gradients. All
experimental data to determine backbone sequential assignments have been
described (Sorgen et al., 2002). Intermolecular nuclear Overhauser effects
(NOEs) were observed in a 3D
13
C F1-edited, F3-
13
C/
15
N-filtered NOE
spectrum (Lee et al., 1994), with a mixing time of 125 ms.
13
C
15
N-Cx43CT
alone was used as the control for leakage through the filter.
13
C
15
N-Cx43CT
was titrated with unlabeled Cx43CT to a 1:1 molar ratio at pH 5.8. NMR
spectra were processed using NMRPipe (Delaglio et al., 1995) and analyzed
using NMRView (Johnson and Blevins, 1994).
RESULTS
Cx43CT chemical cross-linking
To assess the possibility of pH-dependent dimerization of
Cx43CT, solutions of purified, recombinant Cx43CT at
differing pH values were combined with the irreversible
cross-linking agent EDC. EDC, a zero-length cross-linker,
reacts with carboxylic acid and primary amine-containing
molecules. Protein species were separated by SDS-PAGE
(10 mg per lane) and stained with Coomassie blue (Fig. 1 A).
Lanes 1 and 11 correspond to the molecular weight
standards. The 14 kDa bands seen across the gel correspond
to the monomeric form of Cx43CT. Lanes 2, 5, and 8
correspond to samples that were not exposed to the cross-
linking agent. The pH of the solvent varied from 5.8 to 7.5.
An ;29 kDa band was apparent in lanes 3 and 4, 6 and 7,
and—very faintly—lane 9. The cross-linked bands had the
expected mobility for dimers of Cx43CT, and an immuno-
blot experiment confirmed that the product was indeed
Cx43CT (Fig. 1, Band C).
In addition to the dimer at pH 5.8 and 6.5, immunoblot
analysis revealed higher molecular weight aggregates. The
molecular weights for these bands correlated to tri-, tetra-,
penta-, and hexamers of Cx43CT polypeptides (numbers 3,
4, 5, and 6in Fig. 1 B) cross-linked. Although dimeric and
some trimeric species were seen for the Cx43CT polypeptide
at pH 7.5, oligomerization was much more prevalent at the
Cx43CT Dimer Formation 575
Biophysical Journal 87(1) 574–581

low pH, suggesting a pH dependence to the dimerization
process. Interestingly, it can be seen in Fig. 1 Athat equal
amounts of protein were loaded for all lanes; yet, the two
antibodies were not as reactive toward the monomer as they
were toward the higher molecular weight aggregates.
Though the conditions are far from those present when
a connexin molecule oligomerizes into a functional gap
junction in a mammalian cell, Cx43CT oligomerization did
not exceed that which occurs in a connexon.
Molecular mass determination
The observation of pH-dependent oligomerization of
Cx43CT was confirmed by sedimentation equilibrium
analysis. This method also allowed us to more accurately
quantify the stoichiometry of oligomerization at different pH
values. Plots of optical density at equilibrium as a function of
radius for different pH values at two rotor speeds are shown
in Fig. 2. Each plot was best fit by a function derived from
a self-association model to determine the fraction of protein
in specific oligomeric states (Beckman XL-A/XL-I software
package). A monomer-dimer-trimer model showed excellent
convergence, as demonstrated by the minimum deviation
seen in the residuals (top plot on each panel) and a weighted
variance approaching unity (Fig. 3). In contrast, fits of data to
single component (i.e., nonoligomeric), monomer-dimer,
and monomer-dimer-tetramer models had weighted variance
values significantly greater than one. The results obtained
from the monomer-dimer-trimer model were used to
calculate the fraction of total protein that existed in the
dimer form at each pH value. At a pH of 7.5, only 12% of the
total protein content was in a dimer conformation. A slight
reduction in the pH of the solvent to 7.2 increased the
fraction of dimers to 40%. When the pH of the solvent was
reduced to 6.5 and 5.8, more than 84% and 86%,
respectively, of the Cx43CT molecules were dimerized
(percent based on subunit concentration; extinction co-
efficient ;12,450 cm
1
M
1
). Other than a small amount of
trimer (,1%), the nondimerized form of Cx43CT was found
as a monomer at all pH values tested. These results strongly
support the notion of a pH-dependent dimerization of
Cx43CT. Additional experiments were conducted to iden-
tify the specific residues involved in this process of self-
association.
Filtered NOE
We have recently used NMR to determine the resonance
assignments for Cx43CT (Sorgen et al., 2002). Here, we
conducted a 3D
13
C F1-edited, F3-
13
C/
15
N-filtered NOE
experiment (Lee et al., 1994) to confirm intermolecular self-
association and identify the resonance peaks that were
affected by dimerization (Fig. 4). This approach further
allowed us to determine whether the interaction between
individual molecules was parallel; that is, whether energy
was transferred from the amino acid in one polypeptide to
the amino acid in the equivalent position in the homolo-
gous molecule. The two thick vertical lines across the map
correspond to resonances originating from the water
molecules, as well as a small trace of glycerol in the solvent.
A control map, shown by the red contours, was first acquired
using only a double-labeled species of the protein (
13
C
15
N-
labeled Cx43CT). Since resonance peaks originating from
the labeled protein were filtered, red contours represent
‘‘background signals’’ originating from the small fraction of
the protein that did not incorporate the labeled amino acids
and/or from nonoptimized pulse-field gradients (Lee et al.,
1994). A second map, shown by the black contours, was
obtained when the unlabeled species of Cx43CT was mixed
with the
13
C
15
N-labeled species at a 1:1 molar ratio. Signals
that overlapped (black overlapping on red) were identified as
background signals and not actual energy transfers between
the species. However, the black contours that did not overlap
FIGURE 1 Cross-linking recombinant Cx43CT with EDC. (A) Cx43CT,
at concentrations of 7.0 (lanes 2, 3, 5, 6, 8, and 9) and 1.0 (lanes 4, 7, and 10)
mg/mL in PBS buffer (pH indicated above each panel). Equal amounts of
protein (10 mg) were run on a 15% SDS-PAGE. The molecular mass
standards for the 10, 15, 20, 25, and 37 kDa bands have been labeled.
Proteins in the gel presented in panel Awere transferred to nitrocellulose
membrane and analyzed by immunoblot analysis using Zymed polyclonal
anti-Cx43 antibody (B) or 181A polyclonal anti-Cx43 antibody (C).
576 Sorgen et al.
Biophysical Journal 87(1) 574–581

FIGURE 2 Distribution of recombi-
nant Cx43CT at sedimentation equilib-
rium. The concentrations of Cx43CT in
PBS buffer (pH indicated in each box)
at equilibrium (A, 18,000 rpm; B,
26,000 rpm) are shown as a function
of radius. The solid lines are the
theoretical curves. The calculated sub-
unit molecular mass for Cx43CT at all
pH values (5.8, 6.5, 7.2, 7.5) was 14.4
kDa, which is in excellent agreement
with that deduced from the sequence of
Cx43CT (14.2 kDa).
Cx43CT Dimer Formation 577
Biophysical Journal 87(1) 574–581

corresponded to resonance peaks resulting from the transfer
of energy between a labeled molecule and one that was
unlabeled. Given that the resonance peaks had been already
assigned for each amino acid in the Cx43CT sequence
(Sorgen et al., 2002), this experiment allowed us to identify
the specific amino acids involved in the intermolecular
interaction. These results confirmed the presence of self-
association. Moreover, it permitted the assignment of the
specific areas of dimerization. These were: M281–N295,
R299–Q304, S314–I327, and Q342–A348. A sampling of
the resonance peaks (14 from a total of 30) from each region
are identified and labeled in Fig. 4 (see arrows). In-
terestingly, region M281–N295 includes amino acids
thought to be part of the SH3 binding domain of Cx43
(Kanemitsu et al., 1997; Zhou et al., 1999), whereas regions
S314–I327 and Q342–A348 have been identified as
containing a secondary, a-helical structure (Sorgen et al.,
2002).
DISCUSSION
We have used NMR, cross-linking, and sedimentation
equilibrium to characterize the process of pH-dependent
self-association of the carboxyl terminal domain of Con-
nexin43. Our data strongly suggest that acidification of the
solvent leads to dimerization of the protein. We further show
that the areas of dimerization include a putative SH3 binding
domain, as well as two regions found to contain secondary
structure (Sorgen et al., 2002). These results lead to the
hypothesis that dimerization of the CT domain may be one of
the structural changes involved in the chemical regulation
of Cx43. However, some technical aspects need to be
considered.
Technical considerations
All experiments presented in this article were obtained in
vitro from isolated protein fragments in solution. The
conditions are therefore substantially different from those
likely to be present in the microenvironment of the Cx43CT
domain when integrated in a gap junction plaque. Therefore,
the results presented in this article need to be interpreted with
caution. However, a number of previous studies indicate that
the isolated Cx43CT domain retains at least some of its
functional properties. For example, the Cx43CT fragment
can be phosphorylated in vitro by kinases known to modify
the behavior of Cx43 gap junctions (Cooper et al., 2000;
Lampe and Lau, 2000). This same isolated domain can bind
FIGURE 3 Validation of the mono-
mer-dimer-trimer model to describe the
self-association of Cx43CT. Cx43CT in
PBS buffer (pH 5.8) at equilibrium was
used to identify the correct self-associ-
ation model to fit the sedimentation
equilibrium data. The same data were
fit to (A) monomer-dimer-trimer, (B)
monomer-dimer-tetramer, (C) mono-
mer-dimer, and (D) monomer self-
association models. The best fit, as
indicated by the lowest variance value,
is the monomer-dimer-trimer model.
578 Sorgen et al.
Biophysical Journal 87(1) 574–581

in vitro to known Cx43 molecular partners such as ZO-1
(Toyofuku et al., 2001), tubulin (Giepmans et al., 2001), and
the SH3 domain of Src (Kanemitsu et al., 1997; Duffy et al.,
2004). Finally, coexpression experiments in cells show that
the CT fragment can act as an independent domain to rescue
the pH sensitivity of truncated Cx43 channels (Morley et al.,
1996). The same applies to the ability of Cx43 to be
regulated by insulin and insulin-like growth factor (Homma
et al., 1998), c-Src (Zhou et al., 1999) and transjunctional
voltage (Moreno et al., 2002). These studies show that both
in vitro and in vivo the isolated Cx43CT domain retains
biochemical and functional properties consistent with those
found in the full-length channels. Furthermore, our results
show that dimerization occurs in vitro even at neutral pH
(7.0). It is therefore likely that in those studies where the CT
fragment has been used, at least a fraction of the protein
species was present in a dimerized state. Whether the ability
of Cx43CT to act as a kinase substrate, a molecular partner
and/or a functional ‘‘gating particle’’, is affected by the
dimerization state remains to be determined.
Dimerization of other channel proteins
Our data show that the CT domain of Cx43 tends to dimerize
in response to low pH. This is not the first study showing that
intracellular domains of channel proteins undergo dimeriza-
tion. Tetrameric channel proteins such as the HCN channels,
the IP3 receptor, and the NMDA receptor have been dubbed
‘‘dimers of dimers’’ because of the self-association of specific
subunit domains (Galvan and Mignery, 2002; Sun et al., 2002;
Schorge and Colquhoun, 2003). In some cases, dimerization
has been shown in vitro using isolated protein fragments
and then corroborated in functional channels (Galvan and
Mignery, 2002; Leach et al., 2003). Dimerization seems to be
more than a simple biochemical phenomenon. Indeed, in the
case of cyclic nucleotide gated channels, dimerization of
a regulatory domain of the protein substantially modifies
channel function (Matulef and Zagotta, 2002; Ulens and
Siegelbaum, 2003). Analogous to the structure of tetrameric
channels as ‘‘dimers of dimers’’, we speculate that Cx43
hemichannels may exist as ‘‘trimers of dimers’’ and their
oligomeric state may have a regulatory role on the function of
Cx43. It is worth noting that connexin dimerization on
Western blots has been observed ever since the initial gap
junction isolation studies on liver (see Traub et al., 1982;
Green et al., 1988). Although most apparent for Cx32
(VanSlyke and Musil, 2000), antibody-recognized bands of
the approximate dimeric mobility have also been seen for
Cx40 (Matesic et al., 2003) and for Cx43 (Hossain et al.,
1994; VanSlyke and Musil, 2000; Roger et al., 2004), thus
supporting the possibility that Cx43 dimerization may occur
in the setting of an assembled channel.
Dimerization of Cx43CT and regulation of
Cx43 gap junctions
Analysis of the specific residues involved in dimerization
identified four specific areas: M281–N295, R299–Q304,
S314–I327, and Q342–A348. Region M281–N295 is in-
teresting from the point of view of Cx43 regulation. Serine
282 is a substrate for MAPK phosphorylation (Warn-Cramer
et al., 1996,1998). An additional MAPK phosphorylation site,
Serine 279 (Warn-Cramer et al., 1996,1998), is very close to
the dimerization site and its structure may be affected by the
oligomeric state of the protein. Moreover, region 271–287 is
thought to act as an SH3 binding domain, critical for the
interaction of Cx43 with v-Src (Kanemitsu et al., 1997). It is
therefore possible that regulation by these two kinases
(MAPK and v-Src) may either change the dimerization state
of the protein or involve access to a binding site modified by
dimerization. It is also worth noting that deletion of amino
acids 281–300 renders Cx43 channels less sensitive to
acidification-induced uncoupling (Ek-Vitorin et al., 1996).
A similar result is obtained when a peptide corresponding to
amino acids 271–287 of Cx43 is injected in the intracellular
space of Cx43-expressing cells (Calero et al., 1998). We
speculate that dimerization may be a part of the pH gating
process and the 271–287 peptide competitively inhibits the
pH-dependent dimerization of the native CT domains.
Finally, regions 314–327 and 342–348 correspond to areas
of the primary sequence where high-order structure has been
found (Sorgen et al., 2002). We speculate that dimerization
may involve formation of a coiled-coil structure between the
two subunits. Further structural studies will aim at solving the
structure of the Cx43CT dimer and characterizing the effects
of dimerization on other intramolecular interactions, such as
FIGURE 4 Mapping inter-Cx43 carboxyl terminal interactions from a 3D
13
C F1-edited, F3-
13
C/
15
N-filtered NOE experiment.
13
C
15
N-Cx43CT was
titrated with unlabeled Cx43CT to a 1:1 molar ratio at pH 5.8 (black).
13
C
15
N-Cx43CT alone was used as the control for leakage through the filter
(red). Labeled is a sample of the intermolecular interactions. a,b,d,g,e,
and Hsymbolize the alpha, beta, gamma, delta, and epsilon protons.
Cx43CT Dimer Formation 579
Biophysical Journal 87(1) 574–581

the pH-dependent binding of Cx43CT to the cytoplasmic loop
domain (Duffy et al., 2002).
In summary, the work described here has enabled us to
further understand pH-dependent changes in the structure of
Cx43. We show in vitro dimerization of the Cx43CT domain,
a region of the Cx43 protein often used for in vitro studies of
biochemical modifications of the Cx43 gap junctions. We
further show that the fraction of the total protein present in
dimer form is a function of the pH of the solvent. Finally, we
show that some of the regions of dimerization are also
involved in regulation of Cx43 channels, thus opening the
possibility that dimerization may be a structural component of
the regulation of gap junctions.
We thank Mark Girvin and Sean Cahill for their teaching, insight, and
helpful suggestions about the NMR experiments performed for this project.
We would also like to thank Wanda Coombs for purification of the Cx43CT
used in this study and Michael Brenowitz for assistance with the
sedimentation equilibrium experiments.
This work was supported by United States Public Health Service grants F32
GM20504, HL39707, NS41282, MH65495, and GM5769.
REFERENCES
Calero, G., M. Kanemitsu, S. M. Taffet, A. F. Lau, and M. Delmar. 1998. A
17mer peptide interferes with acidification-induced uncoupling of
connexin43. Circ. Res. 82:929–935.
Cascio, W. E., T. A. Johnson, and L. S. Gettes. 1995. Electrophysiologic
changes in ischemic ventricular myocardium: I. Influence of ionic, meta-
bolic, and energetic changes. J. Cardiovasc. Electrophysiol. 6:1039–1062.
Cooper, C. D., J. L. Solan, M. K. Dolejsi, and P. D. Lampe. 2000. Analysis
of connexin phosphorylation sites. Methods. 20:196–204.
Delaglio, F., S. Grzesiek, G. W. Vuister, G. Zhu, J. Pfeifer, and A. Bax.
1995. NMRPipe: a multidimensional spectral processing system based
on UNIX pipes. J. Biomol. NMR. 6:277–293.
Duffy, H. S., A. W. Ashton, P. O’Donnell, W. Coombs, S. M. Taffet, M.
Delmar, and D. C. Spray. 2004. Regulation of connexin43 protein
complexes by intracellular acidification. Circ. Res. 94:215–222.
Duffy, H. S., P. L. Sorgen, M. E. Girvin, P. O’Donnell, W. Coombs, S. M.
Taffet, M. Delmar, and D. C. Spray. 2002. pH-dependent intramolecular
binding and structure involving Cx43 cytoplasmic domains. J. Biol.
Chem. 277:36706–36714.
Ek-Vitorin, J. F., G. Calero, G. E. Morley, W. Coombs, S. M. Taffet, and
M. Delmar. 1996. PH regulation of connexin43: molecular analysis of
the gating particle. Biophys. J. 71:1273–1284.
Francis, D., K. Stergiopoulos, J. F. Ek-Vitorin, F. L. Cao, S. M. Taffet, and
M. Delmar. 1999. Connexin diversity and gap junction regulation by
pHi. Dev. Genet. 24:123–136.
Galvan, D. L., and G. A. Mignery. 2002. Carboxyl-terminal sequences
critical for inositol 1,4,5-trisphosphate receptor subunit assembly. J. Biol.
Chem. 277:48248–48260.
Giepmans, B. N., and W. H. Moolenaar. 1998. The gap junction protein
connexin43 interacts with the second PDZ domain of the zona
occludens-1 protein. Curr. Biol. 8:931–934.
Giepmans, B. N., I. Verlaan, T. Hengeveld, H. Janssen, J. Calafat, M. M.
Falk, and W. H. Moolenaar. 2001. Gap junction protein connexin-43
interacts directly with microtubules. Curr. Biol. 11:1364–1368.
Goldberg, G. S., and A. F. Lau. 1993. Dynamics of connexin43 phos-
phorylation in pp60v-src-transformed cells. Biochem. J. 295:735–742.
Green, C. R., E. Harfst, R. G. Gourdie, and N. J. Severs. 1988. Analysis
of the rat liver gap junction protein: clarification of anomalies in its
molecular size. Proc. R. Soc. Lond. B Biol. Sci. 233:165–174.
Homma, N., J. L. Alvarado, W. Coombs, K. Stergiopoulos, S. M. Taffet,
A. F. Lau, and M. Delmar. 1998. A particle-receptor model for the
insulin-induced closure of connexin43 channels. Circ. Res. 83:27–32.
Hossain, M. Z., L. J. Murphy, E. L. Hertzberg, and J. I. Nagy. 1994.
Phosphorylated forms of connexin43 predominate in rat brain: dem-
onstration by rapid inactivation of brain metabolism. J. Neurochem. 62:
2394–2403.
Janse, M. J., and A. L. Wit. 1989. Electrophysiological mechanisms of
ventricular arrhythmias resulting from myocardial ischemia and in-
farction. Physiol. Rev. 69:1049–1169.
Johnson, B. A., and R. A. Blevins. 1994. NMRView: a computer program
for the visualization and analysis of NMR data. J. Biomol. NMR. 4:
603–614.
Kanemitsu, M. Y., L. W. Loo, S. Simon, A. F. Lau, and W. Eckhart. 1997.
Tyrosine phosphorylation of connexin 43 by v-Src is mediated by SH2
and SH3 domain interactions. J. Biol. Chem. 272:22824–22831.
Kleber, A. G., C. B. Riegger, and M. J. Janse. 1987. Electrical uncoupling
and increase of extracellular resistance after induction of ischemia in
isolated, arterially perfused rabbit papillary muscle. Circ. Res. 61:
271–279.
Lampe, P. D., and A. F. Lau. 2000. Regulation of gap junctions by
phosphorylation of connexins. Arch. Biochem. Biophys. 384:205–215.
Leach, R. N., M. R. Boyett, and J. B. Findlay. 2003. Expression,
purification and spectroscopic studies of full-length Kir3.1 channel
C-terminus. Biochim. Biophys. Acta. 1652:83–90.
Lee, W., M. J. Revington, C. Arrowsmith, and L. E. Kay. 1994. A pulsed
field gradient isotope-filtered 3D 13C HMQC-NOESY experiment for
extracting intermolecular NOE contacts in molecular complexes. FEBS
Lett. 350:87–90.
Loo, L. W., J. M. Berestecky, M. Y. Kanemitsu, and A. F. Lau. 1995.
pp60src-mediated phosphorylation of connexin 43, a gap junction
protein. J. Biol. Chem. 270:12751–12761.
Matesic, D., T. Tillen, and A. Sitaramayya. 2003. Connexin 40 expression
in bovine and rat retinas. Cell Biol. Int. 27:89–99.
Matulef, K., and W. Zagotta. 2002. Multimerization of the ligand binding
domains of cyclic nucleotide-gated channels. Neuron. 36:93–103.
Moreno, A. P., M. Chanson, S. Elenes, J. Anumonwo, I. Scerri, H. Gu,
S. M. Taffet, and M. Delmar. 2002. Role of the carboxyl terminal
of connexin43 in transjunctional fast voltage gating. Circ. Res. 90:
450–457.
Moreno, A. P., J. C. Saez, G. I. Fishman, and D. C. Spray. 1994. Human
connexin43 gap junction channels. Regulation of unitary conductances
by phosphorylation. Circ. Res. 74:1050–1057.
Morley, G. E., S. M. Taffet, and M. Delmar. 1996. Intramolecular
interactions mediate pH regulation of connexin43 channels. Biophys. J.
70:1294–1302.
Peters, N. S., J. Coromilas, N. J. Severs, and A. L. Wit. 1997. Disturbed
connexin43 gap junction distribution correlates with the location of
reentrant circuits in the epicardial border zone of healing canine infarcts
that cause ventricular tachycardia. Circulation. 95:988–996.
Roger, C., B. Mograbi, D. Chevallier, J. F. Michiels, H. Tanaka, D.
Segretain, G. Pointis, and P. Fenichel. 2004. Disrupted traffic of
connexin 43 in human testicular seminoma cells: overexpression of Cx43
induces membrane location and cell proliferation decrease. J. Pathol.
202:241–246.
Saez, J. C., V. M. Berthoud, A. P. Moreno, and D. C. Spray. 1993. Gap
junctions. Multiplicity of controls in differentiated and undifferentiated
cells and possible functional implications. Adv. Second Messenger
Phosphoprotein Res. 27:163–198.
Schorge, S., and D. Colquhoun. 2003. Studies of NMDA receptor function
and stoichiometry with truncated and tandem subunits. J. Neurosci.
23:1151–1158.
Sorgen, P. L., H. S. Duffy, S. M. Cahill, W. Coombs, D. C. Spray, M.
Delmar, and M. E. Girvin. 2002. Sequence-specific resonance assign-
ment of the carboxyl terminal domain of Connexin43. J. Biomol. NMR.
23:245–246.
580 Sorgen et al.
Biophysical Journal 87(1) 574–581

Spray, D. C., and M. V. Bennett. 1985. Physiology and pharmacology of
gap junctions. Annu. Rev. Physiol. 47:281–303.
Stergiopoulos, K., J. L. Alvarado, M. Mastroianni, J. F. Ek-Vitorin, S. M.
Taffet, and M. Delmar. 1999. Hetero-domain interactions as a mechanism
for the regulation of connexin channels. Circ. Res. 84:1144–1155.
Sun, Y., R. Olson, M. Horning, N. Armstrong, M. Mayer, and E. Gouaux.
2002. Mechanism of glutamate receptor desensitization. Nature.
417:245–253.
Swenson, K. I., H. Piwnica-Worms, H. McNamee, and D. L. Paul. 1990.
Tyrosine phosphorylation of the gap junction protein connexin43 is
required for the pp60v-src-induced inhibition of communication. Cell
Regul. 1:989–1002.
Thi, M. M., T. Kojima, S. C. Cowin, S. Weinbaum, and D. C. Spray. 2003.
Fluid shear stress remodels expression and function of junctional proteins
in cultured bone cells. Am. J. Physiol. Cell Physiol. 284:C389–C403.
Toyofuku, T., Y. Akamatsu, H. Zhang, T. Kuzuya, M. Tada, and M. Hori.
2001. c-Src regulates the interaction between connexin-43 and ZO-1 in
cardiac myocytes. J. Biol. Chem. 276:1780–1788.
Traub, O., U. Janssen-Timmen, P. M. Druge, R. Dermietzel, and K.
Willecke. 1982. Immunological properties of gap junction protein from
mouse liver. J. Cell. Biochem. 19:27–44.
Ulens, C., and S. A. Siegelbaum. 2003. Regulation of hyperpolarization-
activated HCN channels by cAMP through a gating switch in binding
domain symmetry. Neuron. 40:959–970.
VanSlyke, J. K., and L. S. Musil. 2000. Analysis of connexin intracellular
transport and assembly. Methods. 20:156–164.
Warn-Cramer, B. J., G. T. Cottrell, J. M. Burt, and A. F. Lau. 1998.
Regulation of connexin-43 gap junctional intercellular communication
by mitogen-activated protein kinase. J. Biol. Chem. 273:9188–9196.
Warn-Cramer, B. J., P. D. Lampe, W. E. Kurata, M. Y. Kanemitsu, L. W.
Loo, W. Eckhart, and A. F. Lau. 1996. Characterization of the mitogen-
activated protein kinase phosphorylation sites on the connexin-43 gap
junction protein. J. Biol. Chem. 271:3779–3786.
Zhou, L., E. M. Kasperek, and B. J. Nicholson. 1999. Dissection of the
molecular basis of pp60(v-src) induced gating of connexin 43 gap
junction channels. J. Cell Biol. 144:1033–1045.
Cx43CT Dimer Formation 581
Biophysical Journal 87(1) 574–581