| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Biophys J, January 1999, p. 198-206, Vol. 76, No. 1
*Department of Physiology and Biophysics, Finch University of Health Sciences/The Chicago Medical School, North Chicago, Illinois 60064, and #Department of Pediatrics, University of Chicago, Chicago, Illinois 60637 USA
 |
ABSTRACT |
|---|
 Top
 Abstract
 Introduction
Materials and methods
Results
Discussion
References
|
|---|
Lens fiber cells contain two gap junction proteins (Cx56
and Cx45.6 in the chicken). Biochemical studies have suggested that these two proteins can form heteromeric connexons. To investigate the
biophysical properties of heteromeric lens connexons, Cx56 was
co-expressed with Cx45.6 (or its mouse counterpart, Cx50) in
Xenopus oocytes. Whole-cell and single-channel currents
were measured in single oocytes by conventional two-microelectrode voltage-clamp and patch clamp techniques, respectively. Injection of
Cx56 cRNA induced a slowly activating, nonselective cation current that
activated on depolarization to potentials higher than 
10 mV. In
contrast, little or no hemichannel current was induced by injection of
Cx50 or Cx45.6 cRNA. Co-expression of Cx56 with Cx45.6 or Cx50 led to a
shift in the threshold for activation to 40 or 70 mV, respectively.
It also slowed the rate of deactivation of the hemichannel currents.
Moreover, an increase in the unitary conductance, steady state
probability of hemichannel opening and mean open times at negative
potentials, was observed in (Cx56 + Cx45.6) cRNA-injected oocytes
compared with Cx56 cRNA-injected oocytes. These results indicate that
co-expression of lens fiber connexins gives rise to novel channels that
may be explained by the formation of heteromeric hemichannels that
contain both connexins.
| |
INTRODUCTION |
|---|
|
Top
Abstract
Introduction
Materials and methods
Results
Discussion
References
|
|---|
Gap junctional channels allow the intercellular
transfer of ions and metabolites between adjacent cells. They are
oligomeric assemblies of proteins called connexins (Bruzzone et al.,
1996
). The functional properties of gap junctional channels composed of
one connexin (homomeric/homotypic channels) have been investigated extensively (Barrio et al., 1991; Paul et al., 1991; White et al.,
1994; Veenstra et al., 1992). These studies have shown that channels
formed from different connexins have distinct gating and permeability
properties. However, most cells express more than one connexin
(Nicholson et al., 1987; Kanter et al., 1992; Paul et al., 1991)
raising the possibility of mixing. One kind of connexin mixing involves
heterotypic gap junctional channels in which each hemichannel is
composed of a different connexin. Electrophysiological studies have
shown that some heterotypic channels exhibit properties that are novel
compared with those predicted by examination of the homotypic channels
(Barrio et al., 1991; Bruzzone et al., 1994; White et al., 1994). A few
studies have also been published in which the data can be interpreted according to a second kind of connexin mixing through the formation of
gap junctional channels made of heteromeric connexons (more than one
connexin forms the hemichannel) (Barrio et al., 1991; Brink et al.,
1997; Bevans et al., 1998).
The lens is an avascular organ in which gap junctions provide a pathway
for intercellular communication across the organ supporting cell
survival and thus lens transparency. This organ may provide an
excellent system to examine the functional consequences of assembly of
heteromeric connexons. Three different types of connexins have been
identified in the chicken lens: Cx43 (Musil et al., 1990), Cx56 (Rup et
al., 1993; Berthoud et al., 1994), and Cx45.6 (Jiang et al., 1994).
Cx43 is found exclusively in the epithelial cell layer, whereas Cx56
and Cx45.6 colocalize in lens fiber-fiber gap junctions. Biochemical
studies by Konig and Zampighi (1995) and Jiang and Goodenough (1996)
suggest that the mammalian lens fiber connexins, Cx46 and Cx50,
associate to form heteromeric hemichannels.
In the present study, we have examined the physiological consequences of formation of heteromeric connexons by the lens fiber connexins taking advantage of the ability of some of them to form open hemi-gap-junctional channels in single Xenopus oocytes. The data presented in this paper show that co-expression of Cx56 and Cx45.6 or Cx56 and Cx50 induced the formation of functional hemichannels with properties that differ from those expected based on observations of the two homomeric hemichannels.
| |
MATERIALS AND METHODS |
|---|
|
Top
Abstract
Introduction
Materials and methods
Results
Discussion
References
|
|---|
Preparation of cRNA
cDNAs encoding chicken Cx45.6 and mouse Cx50 in the pSP64T
vector (Krieg and Melton, 1984) were provided by Dr. D. A. Goodenough (Harvard University, Boston, MA). We have previously studied
expression of chicken Cx56 (Rup et al., 1993) in Xenopus
oocytes using the same vector (Ebihara et al., 1995). The constructs
were linearized with the restriction enzyme, SalI. cRNAs
were transcribed in vitro with SP6 RNA polymerase using the mMessage
mMachine (Ambion Inc., Austin, TX). The transcripts were purified on a
G-50 Sephadex column (Boehringer Mannheim, Indianapolis, IN) to remove
unincorporated rNTPs and precipitated with isopropanol. The transcripts
were then resuspended in diethyl pyrocarbonate-treated water, and
aliquots were stored at 80°C. The integrity and yield of RNA were
assessed by formaldehyde gel electrophoresis and absorbance at 260 nm, respectively.
Expression of lens connexins in Xenopus laevis oocytes
Stage V and VI oocytes were obtained by partial ovariectomy of
female Xenopus laevis (Nasco, Ft. Atkinson, WI).
The oocytes were defolliculated by incubation with collagenase type IA
(Sigma Chemical, St. Louis, MO) and injected with oligonucleotides
antisense to mRNA for Xenopus Cx38 as previously described
(Ebihara, 1996). They were stored overnight at 17-18°C in modified
Barth's solution (MBS) containing: 88 mM NaCl, 1 mM KCl, 2.4 mM
NaHCO3, 15 mM HEPES, 0.3 mM
Ca(NO3)2, 0.41 mM
CaCl2, and 0.82 mM MgSO4,
pH 7.4. Healthy oocytes were selected and injected with 0.2 to 3 ng RNA
and incubated for 12-24 h. To assess levels of expression of homomeric
and heteromeric channels, oocytes from a single donor frog were
injected with the same amount of each individual connexin cRNA unless
otherwise noted and studied at similar times following injection.
For immunoblot analysis of connexins, plasma membrane-enriched
preparations were obtained as previously described (White et al., 1994;
Gupta et al., 1994; Ebihara et al., 1995). Proteins were resolved by
polyacrylamide gel electrophoresis on sodium dodecyl sulfate-containing
9% acrylamide gels and transferred to Immobilon P membrane (Millipore,
Bedford, MA). Immunoblots were performed as previously described
(Berthoud et al., 1994) using previously characterized rabbit anti-Cx56
antibodies (Berthoud et al., 1994) or anti-Cx45.6 antibodies (kindly
provided by Drs. Jean Jiang and Daniel Goodenough; Jiang and
Goodenough, 1996). Binding of primary antibodies was detected
using alkaline phosphatase-conjugated goat anti-rabbit IgG antibodies
(Boehringer Mannheim, Indianapolis, IN) and the nitro blue
tetrazolium/BCIP substrate kit (Promega, Madison, WI). For
detection of Cx50, a mouse monoclonal antibody kindly provided by Dr.
J. Kistler (Kistler et al., 1985) and alkaline phosphatase-conjugated
goat anti-mouse Ig (Boehringer Mannheim) were used.
Two-electrode voltage-clamp recording of macroscopic hemi-gap-junctional current
Two-microelectrode voltage-clamp recording of
hemi-gap-junctional currents in single oocytes was performed using an
Axoclamp 2A or a Geneclamp 500 amplifier (Axon Instruments, Foster
City, CA). The current and voltage electrodes had resistances of
0.1-0.2 M
, respectively, when filled with 3 M KCl. To prevent KCl
from leaking out of the electrodes, the tips of the electrodes were back-filled with a cushion of 1% agar in 3 M KCl (Schreibmayer et al.,
1994). The membrane potential was usually recorded differentially between the intracellular electrode and a second electrode located in
the bath just outside the oocyte to reduce voltage drops across the
series resistance of the external solution. In the low calcium experiments, a bath clamp was used. The current signal was filtered at
20-50 Hz using a 4-pole Bessel filter. Pulse generation and data
acquisition was performed using an 80386 computer equipped with PCLAMP6
software and a TL-1 acquisition system (Axon Instruments). All
experiments were done at room temperature (22-25°C). Unless otherwise noted, the recordings were performed in MBS containing 0.7 mM
Ca+2. Results are presented as mean ± SD.
Single channel current recording
For patch clamp experiments, the vitelline membrane was removed
using the technique of Methfessel et al. (1986). Single channel currents were measured from cell-attached and inside-out patches with
an Axopatch 200A amplifier (Axon Instruments). Patch pipettes were
pulled from Corning #7052 glass capillaries, 1.5 mm (OD)/1.0 mm (ID)
(W-P Instruments, New Haven, CT) using a Brown-Flaming Micropipette
puller (Sutter Instruments, San Francisco, CA). The patch electrodes
were coated with Sylgard (Dow-Corning, Midland, MI) and fire polished.
The bath solution contained: 100 mM KCl, 5 mM HEPES, 1 mM EGTA, 1 mM
MgCl2, pH 7.4. This solution brought the
oocyte's resting potential close to zero. Therefore, the voltages reported here are approximately equal to the actual transmembrane potential. The pipette solution was identical to the bath solution except that TMACl was usually substituted for KCl to eliminate inward
flow of current through endogenous stretch activated channels. Current
records were filtered at 1 kHz, digitized, and stored on a Tahoe 230 magnetic optical drive (Pinnacle Micro, Irvine, CA) for later analysis.
Single channel analysis
In voltage-step experiments, the ohmic leakage current and
uncompensated capacitive current were digitally subtracted using leak
templates made from sweeps with no openings. In voltage-ramp experiments, the leak current was digitally subtracted using leak templates constructed from segments of records with no openings. For
kinetic analysis of single channels, continuous 2-10 min recordings were idealized using a K-segmental K-means program (written by Dr. Qin,
SUNY at Buffalo, Buffalo, NY). A maximal log-likelihood method was used
for kinetic modeling of idealized data (Qin et al., 1996; Qin et al.,
1997). To verify that the kinetic parameters estimated by this method
were valid, the optimized rate constants were used to construct open
and closed time histograms and compared with experimentally obtained
histograms. Results are presented as mean ± SD.
| |
RESULTS |
|---|
|
Top
Abstract
Introduction
Materials and methods
Results
Discussion
References
|
|---|
Co-expression of Cx56 with Cx45.6 or Cx50 modifies the functional behavior of hemi-gap-junctional currents
To verify that the connexins were being efficiently translated and transported to the plasma membrane, immunoblots were performed using oocyte plasma membrane-enriched preparations (Fig. 1). Oocytes injected with Cx45.6 cRNA contained a major anti-Cx45.6 immunoreactive band of an apparent molecular mass (Mr) of 58 kDa (A, lane 1). Cx56 cRNA-injected oocytes contained the expected Cx56 immunoreactive band of Mr 66 kDa (B, lane 2). No cross-reactivity was observed between antibodies and extracts from oocytes injected with the noncorresponding cRNA (A, lane 2; B, lane 1). Both proteins were identified in membrane-enriched preparations from oocytes co-injected with cRNA for Cx45.6 and Cx56 (lane 3). The presence of Cx50 in oocytes injected with Cx50 cRNA alone or in combination with Cx56 cRNA was similarly confirmed by immunoblot analysis (data not shown).
|
Macroscopic hemi-gap-junctional currents (Ih)
recorded in oocytes injected with cRNA encoding Cx56 or Cx45.6 were
compared with currents recorded from oocytes co-injected with cRNA for Cx56 and Cx45.6 using extracellular solutions containing 0.7 mM calcium. Fig. 2 A shows
families of current traces recorded in response to a series of
depolarizing voltage-clamp steps between 30 and +30 mV from a holding
potential of 40 mV. Oocytes injected with Cx56 cRNA displayed a
large, slowly activating outward current, Ih,
which activated at potentials positive to 10 mV as previously reported (Ebihara et al., 1995). No hemi-gap-junctional current was
observed in oocytes injected with cRNA for Cx45.6 alone. Co-injection of Cx56 with Cx45.6 induced a slowly activating outward current that
had a more negative threshold for activation than Cx56 and a slower
rate of deactivation.
|
The voltage dependence of activation of the hemi-gap-junctional
currents was determined by measuring the initial amplitude of the tail
current at 40 mV following 24-s prepulses to different potentials.
Averaged data showed that the activation curve for Ih in oocytes expressing Cx56 (Fig. 2
B, open squares) was shifted to more negative
potentials by co-injection with Cx45.6 (Fig. 2 B,
solid circles). The threshold for activation (the most
negative potential at which Ih could be observed)
decreased from ~10 mV (n = 3) for oocytes
expressing Cx56 to 40 mV (n = 6) for oocytes co-expressing Cx56 and Cx45.6. In addition, co-injection of Cx56 with
Cx45.6 reduced the slope of the activation curve as might be expected
for a heterogeneous population of channels that activate over a wider
range of voltages. The magnitude of the hemi-gap-junctional currents in
(Cx56 + Cx45.6) cRNA-injected oocytes was significantly smaller than
that recorded in oocytes injected with Cx56 alone (Fig. 2
C). The outward current (measured at the end of a 24-s step
to +30 mV from a holding potential of 40 mV) progressively decreased
when the ratio of Cx56 to Cx45.6 subunits was reduced by injecting
increasing amounts of Cx45.6 cRNA into the oocytes. The reductions in
current suggest that Cx56 associates with Cx45.6 to form heteromers
that are less efficiently transported to the cell surface or have
reduced function.
Similar but more dramatic changes in gating properties were observed
when Cx56 was co-expressed with Cx50. Fig.
3 A compares representative
families of current traces recorded from oocytes expressing Cx56, (Cx56 + Cx50), or Cx50. Little or no detectable current was present in
oocytes injected with Cx50 cRNA alone. Co-expression of Cx50 and Cx56
led to a negative shift in the threshold for activation of
Ih, accelerated the rate of activation, and
slowed the rate of deactivation. Unlike (Cx56 + Cx45.6) oocytes, co-expression of Cx56 and Cx50 increased the amplitude of the resultant
hemi-gap-junctional current (measured at the end of a 24-s step to +30
mV from a holding potential of 40 mV) relative to that of Cx56 alone
(Fig. 3 C). The activation curves for (Cx56 + Cx50)- and
Cx56-induced currents are shown in Fig. 3 B. (Cx56 + Cx50)
channels activated at much more negative voltages and the slope of the
activation curve was decreased. The threshold for activation shifted
from ~10 mV for oocytes expressing Cx56 (n = 3) to
70 mV for oocytes co-expressing Cx56 and Cx50 (n = 4). Furthermore, co-expression of Cx56 with Cx50 reversed the direction
of rectification of the instantaneous I-V relationship from outward to
inward. This can be observed in Fig. 3 A as a marked
reduction in the ratio between the activated outward current and the
tail current. The ratio between the activated outward current at 30 mV
and the initial tail current at 80 mV was 0.19 ± 0.09 (n = 6) for oocytes co-expressing Cx56 and Cx50 and
1.6 ± 0.4 (n = 4) for oocytes expressing Cx56.
|
The effect of co-expression of Cx56 with Cx45.6 (or Cx50) on ion
selectivity was examined by measuring changes in the reversal potential
of Ih when oocytes were exposed to external
solutions containing different monovalent cations or anions. The
reversal potential was measured using a ramp protocol as previously
described (Ebihara and Steiner, 1993). In some experiments, the
reversal potential was also estimated from the reversal of
Ih on step depolarization or from the reversal of
the tail currents. All of these methods gave comparable values for
reversal potential. The results of these experiments are summarized in
Table 1. Assuming
[Na+]i = 20 mM and
[K+]i = 150 mM (Kusano et
al., 1982), in sodium chloride saline (98 mM NaCl), the estimated
reversal potential for Na+ or
K+ in Xenopus oocytes would be +40 mV
and < 100 mV, respectively. The Cx56 hemi-gap-junctional
currents reversed polarity at 12.68 ± 3.6 mV (mean ± SD,
n = 5) in sodium chloride saline. Replacement of all
Na+ by K+ shifted
Erev to 4.8 ± 2.6 mV (mean ± SD,
n = 4). Substitution of
TMA+ for Na+ shifted
Erev to 22.0 ± 7.5 mV (mean ± SD,
n = 4). No significant change in reversal potential was
observed when KGluconate was substituted for KCl. These results
are in agreement with previous results (Ebihara et al., 1995) and
suggest that Cx56 hemi-gap-junctional channels are permeable to
monovalent cations and discriminate poorly between
Na+ and K. Co-expression of Cx56 with Cx50 or
Cx45.6 did not appear to modify the ionic selectivity of the
hemi-gap-junctional channels.
|
Effect of external calcium
In previous work, reduction of external calcium was shown to
increase Cx56 hemi-gap-junctional currents and shift the voltage dependence of activation to more negative potentials (Ebihara et al.,
1995). We confirmed this result using Cx56 cRNA-injected oocytes and
tested for a similar effect of calcium on Cx45.6 and (Cx56 + Cx45.6)
cRNA-injected oocytes. Fig. 4
A-C shows the effect of reducing external calcium on
macroscopic current traces and I-V curves. Independently of the
connexin cRNA injected, reducing the Ca2+
concentration to nominally zero caused a marked increase in
Ih and shifted the threshold of activation to
more negative voltages. However, the Cx45.6- and (Cx56 + Cx45.6)-induced currents activated over a more negative range of
potentials than the Cx56-induced current. Furthermore, the
Cx45.6-induced currents could only be detected in zero calcium MBS and
were considerably smaller in magnitude than the currents induced by
injecting oocytes with similar amounts of cRNA for Cx56 or (Cx56 + Cx45.6). No calcium-sensitive membrane currents were observed in
control, antisense-treated oocytes (data not shown).
|
The activation and deactivation kinetics and voltage sensitivity of the
hemi-gap-junctional currents in zero calcium MBS are shown in greater
detail in Fig. 5. Fig. 5 A
compares families of current traces recorded from oocytes expressing
Cx56, (Cx56 + Cx45.6) or Cx45.6 in response to a series of voltage
clamp steps between 100 and 20 or 30 mV from a holding potential of
10 mV. In the case of Cx56, the hemichannels were not fully activated at 10 mV and application of depolarizing voltage clamp steps elicited
a slowly activating outward current. In the case of Cx45.6 or (Cx45.6 + Cx56), the hemichannels were almost completely activated at 10 mV and
application of large depolarizing voltage clamp steps appeared to cause
the channels to slowly inactivate. For all the connexins, the
hemi-gap-junctional current deactivated with hyperpolarizing steps.
However, the Cx56-induced current deactivated more rapidly and to a
lower steady-state level than the Cx45.6- or (Cx45.6 + Cx56)-induced
currents. To quantitate these differences in steady-state voltage
dependence, the rising phase of the normalized steady-state conductance
(G
) versus voltage curves in zero
calcium MBS (Fig. 5 B) was fit to a Boltzmann equation of
the form:
|
V curve for
Cx56 (open squares, n = 3) had a midpoint
(Vo) of 16.4 mV and a slope factor
(A) of 0.11. We were unable to accurately determine the
G V curve at potentials
>30 mV because Ih was contaminated by a large
endogenous sodium current. The rising phase of the
G V curve for Cx45.6
(solid circles, n = 2) had a
Vo of 65 mV and an A of 0.056. The current was fully activated between 10 and +10 mV and declined at
more positive potentials. The G V curve for the mixed (Cx45.6 + Cx56) channels had a voltage
sensitivity that was intermediate between that of Cx45.6 and Cx56 with
a Vo of 45.2 mV and an A of
0.06 (solid stars, n = 2).
|
Co-expression of Cx45.6 with Cx56 modifies single hemi-gap-junctional channel gating properties
To investigate the effect of heteromeric assembly on the voltage-dependent gating properties of the hemichannels in more detail, single hemichannel currents produced in (Cx56 + Cx45.6)-injected oocytes were recorded from cell-attached and inside-out patches and compared with those of homomeric Cx56 channels.
An example of single channel currents recorded from a cell-attached
patch containing multiple Cx56 channels is shown in Fig. 6 A. In this experiment, the
membrane potential was initially held at 0 mV, and then a 4-s
hyperpolarizing step was applied to the indicated membrane potentials.
Several channels were open at the beginning of the voltage step. When
the membrane potential was stepped to 40 or 60 mV, the channels
reopened multiple times before entering a relatively long-lived closed
state. The time course of deactivation can be more clearly seen in the
ensemble averages obtained from the same patch (Fig. 6 B).
The effect of voltage on the normalized conductance-voltage curve at
4 s in macropatches is shown in Fig. 6 C (solid
circles). The normalized conductance was estimated by dividing the
ensemble averaged current at 4 s by the initial current. The
normalized conductance was maximum at potentials between 10 and +40
mV and declined at more negative potentials. At potentials negative to
40 mV, the channels were mostly closed. Comparison of the normalized
G V curve determined from
single channel data with the normalized G V curve determined from macroscopic currents recorded under
two-microelectrode voltage-clamp conditions in zero calcium MBS (solid
line) demonstrated that the steady-state probability of opening
determined by these two methods were in good agreement.
|
The single channel I-V relationship for Cx56 was determined using a
voltage ramp protocol in cell-attached patches. With KCl in the
pipette, the I-V relationship was nearly linear and had a slope
conductance of ~300 pS (not shown). A similar value for unitary
conductance has been previously reported for Cx46 channels at
potentials negative to Erev (Trexler et
al., 1996). When TMACl was substituted for KCl in the pipette, the I-V
relationship became slightly outwardly rectifying with a slope
conductance of 163 ± 18.3 pS (mean ± SD, n = 4) between 20 and 80 mV and a reversal potential of 19.2 ± 4.8 mV (mean ± SD, n = 4) (Fig. 6
D).
The kinetics of opening and closing of the Cx56 hemichannels were
investigated in more detail at voltages between 30 and 50 mV in
several patches containing 1-4 channels. Fig.
7 A shows a segment from a
typical 10 min, continuous recording at 30 mV. There were at least
three channels in this patch. An all-points amplitude histogram of the
current distribution from this record (Fig. 7 B) shows that
the peaks in the amplitude histogram were spaced uniformly with a
peak-to-peak current of 4.24 ± 0.09 pA. The channels tended to
open as clusters. Within a cluster, there were two types of closures,
one brief and the other long lived. This suggests that there are at
least two closed states with considerably different mean dwell times.
We used a maximum log-likelihood method (Qin et al., 1996) to obtain
estimates for mean channel open and closed times as it contains an
algorithm for dealing with multiple channels. The simplest scheme that
could explain the data contained two closed states and one open state.
Open and closed time duration histograms closely followed curves
generated using the rate constants indicated in Fig. 7 C.
The mean open time was 27.2 ms in the experiment shown in Fig. 7. The
average mean open time obtained from four other experiments at 40 mV
was 23 ± 8.8 ms (mean ± SD).
|
To test if co-expression of Cx45.6 and Cx56 affected the single channel
characteristics, similar experiments were performed in oocytes
co-injected with Cx45.6 and Cx56 cRNA. Representative single channel
current traces and open channel I-V relationship are shown in Fig.
8 A and D. The open
channel I-V relationship for (Cx56 + Cx45.6) channels was determined in
cell-attached patches using a voltage ramp protocol. With TMACl in the
pipette, the slope conductance (determined between 20 and 80 mV)
was 206 ± 31.7 pS (mean ± SD, n = 5) and
the reversal potential was 13.9 ± 2.4 mV (mean ± SD,
n = 5). This value for single channel conductance was
larger than that determined for the Cx56 channels under identical conditions. The variability in the size of the single channels may
reflect the existence of many possible heteromeric forms of the
channel. Another difference between the Cx56 channels and the (Cx56 + Cx45.6) channels was that the (Cx56 + Cx45.6) channels typically did
not deactivate during a 6-s voltage step to 40 or 60 mV; after
hyperpolarization, the channel maintained a high probability of opening
and underwent only rare transitions to the fully closed state. These
single channel properties are reflected in the ensemble averages of
(Cx56 + Cx45.6) channel activity (Fig. 8 B). The ensemble
averaged current had a larger steady-state component. Fig. 8
C shows the normalized conductance-voltage curve determined
at 6 s from ensemble averaged data. The (Cx56 + Cx45.6) channels
activated over a more negative voltage range than the Cx56 channels,
and the slope of the curve was less steep. The voltage dependence of
the (Cx56 + Cx45.6) channels was nearly identical to that predicted
from whole cell recordings (solid line).
|
In addition, we made an attempt to examine the single channel properties of homomeric Cx45.6 channels. Although these experiments were complicated by the fact that the Cx45.6 channels were present in the plasma membrane at a lower density, we were able to successfully record single Cx45.6 channels in two experiments. These channels had a similar conductance, ~220 pS, but were even less sensitive to transmembrane voltage than (Cx56 + Cx45.6) channels (data not shown).
| |
DISCUSSION |
|---|
|
Top
Abstract
Introduction
Materials and methods
Results
Discussion
References
|
|---|
In this study we have presented electrophysiological evidence for
the formation of heteromeric connexons between the lens fiber
connexins, Cx56 and Cx45.6/Cx50. Co-expression of Cx56 with Cx45.6 or
Cx50 led to the formation of hemichannels the characteristics of which
differed from those of the homomeric connexons. The biochemical data
previously published by Konig and Zampighi (1995) and Jiang and
Goodenough (1996) support the presence of these heteromeric channels in
the lens in vivo.
Previous studies in oocyte pairs have shown that Cx46 is able to form
heterotypic gap junctional channels with both Cx43 and Cx50 but that
Cx50 does not form functional heterotypic channels when paired with
Cx43 (White et al., 1994). In addition, an attempt was made to study
heteromeric interactions between Cx46 and Cx50 in the
Xenopus oocyte pair expression system by pairing oocytes co-injected with Cx46 and Cx50 cRNAs with Cx43-expressing oocytes (White et al., 1994). These experiments demonstrated that oocytes co-expressing Cx46 and Cx50 were able to form functional gap junctional channels with Cx43. Unfortunately, because of the lack of resolution of
the two-cell voltage-clamp approach used, the properties of these
currents were indistinguishable from those of Cx46/Cx43 oocyte pairs
and thus, failed to provide evidence for heteromeric association of
Cx46 with Cx50. The approach used in our study was to examine
heteromeric interactions more directly by examining the functional
properties of hemichannels in single oocytes. Our results show that,
whereas Cx56-expressing oocytes can form open hemichannels (Ebihara et
al., 1995; and this study), Cx50 or Cx45.6-expressing oocytes show
little or no detectable connexin-induced currents in solutions
containing 0.7 mM calcium. Thus, if there is no formation of
heteromeric connexons, one would expect to record only Cx56-induced currents in oocytes co-injected with Cx56 and Cx50 or Cx45.6. Instead,
co-expression of the fiber connexins led to a negative shift in the
threshold for activation and slowed the rate of deactivation of the
connexin-induced current. It also led to alterations in the magnitude
of the current and to an increase in single channel conductance and
open times of (Cx45.6 + Cx56) channels compared with Cx56 channels.
Co-expression of Cx45.6 (or Cx50) with Cx56 did not result in any
obvious changes in ionic selectivity. The ionic selectivity of the
hemi-gap-junctional currents was similar to that previously reported
for Cx46 (Trexler et al., 1996; Ebihara and Steiner, 1993).
The primary structure of chick Cx45.6 is most similar (>70% amino
acid sequence identity) to mouse Cx50 (Jiang et al., 1994; White et
al., 1992) and to human Cx50 (Church et al., 1995). Thus, it is not
surprising that when chick Cx45.6 and mouse Cx50 are expressed in
oocyte pairs, they form homotypic gap junctional channels with similar
voltage gating properties (Jiang et al., 1994; White et al., 1992).
Nevertheless, our results show that Cx56 associates with Cx45.6 or Cx50
to form heteromeric connexons with distinct electrophysiological
properties. In particular, (Cx56 + Cx45.6) channels have a more
positive threshold for activation and exhibit a greater degree of
outward rectification than (Cx56 + Cx50) channels.
The most important consequence of formation of heteromeric
hemi-gap-junctional channels may be of physiological relevance. The
mixed channels become less sensitive to block by external calcium or
voltage and their mean open times are longer. Thus, the formation of
heteromeric hemichannels would tend to increase their contribution to
the resting membrane conductance of lens fiber cells. In addition, the
formation of heteromeric gap junctional channels in the lens would be
expected to alter the ability of lens fiber cells to communicate with
each other. The differential properties of heteromeric channels might
be crucial for lens transparency, as cataracts form in mice when either
of the fiber cell connexins is absent (Gong et al., 1997). But, most
importantly, our data suggest a mechanism to explain why autosomal
dominant mutations of only one fiber cell connexin, Cx50, (White et
al., 1997; Shiels et al., 1998; Steele et al., 1998) can lead to
disease. The mutant Cx50 might interfere not only with wild type Cx50
channels, but through heteromeric mixing, it might also interfere with
the formation of Cx46 containing channels.
| |
ACKNOWLEDGMENTS |
|---|
We thank Mary Beth A. Latayan for performing some of the preliminary experiments on (Cx45.6 + Cx56) cRNA-injected oocytes.
This study was supported by the National Institutes of Health Grant EY10589 (to L. Ebihara) and Grant EY08368 (to E. C. Beyer).
| |
FOOTNOTES |
|---|
Received for publication 23 January 1998 and in final form 2 October 1998.
Address reprint requests to Dr. Lisa Ebihara, Department of Physiology and Biophysics, Finch University of Health Sciences/The Chicago Medical School, North Chicago, IL 60064. Tel: 847-578-3424; Fax: 847-578-3265; E-mail: ebiharal{at}mis.finchcms.edu.
| |
REFERENCES |
|---|
|
Top
Abstract
Introduction
Materials and methods
Results
Discussion
References
|
|---|
3 connexin gene leads to proteolysis and cataractogenesis in mice.
Cell.
91:833-843[Medline].
Biophys J, January 1999, p. 198-206, Vol. 76, No. 1
© 1999 by the Biophysical Society 0006-3495/99/01/198/09 $2.00
This article has been cited by other articles:
 |
|
 |
|
  Y. V. Panchin Evolution of gap junction proteins - the pannexin alternative J. Exp. Biol., April 15, 2005; 208(8): 1415 - 1419. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. J. Martinez-Wittinghan, C. Sellitto, L. Li, X. Gong, P. R. Brink, R. T. Mathias, and T. W. White Dominant cataracts result from incongruous mixing of wild-type lens connexins J. Cell Biol., June 9, 2003; 161(5): 969 - 978. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Gu, X. S. Yu, X. Yin, and J. X. Jiang Stimulation of Lens Cell Differentiation by Gap Junction Protein Connexin 45.6 Invest. Ophthalmol. Vis. Sci., May 1, 2003; 44(5): 2103 - 2111. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Ebihara, X. Liu, and J. D. Pal Effect of External Magnesium and Calcium on Human Connexin46 Hemichannels Biophys. J., January 1, 2003; 84(1): 277 - 286. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Rong, X. Wang, I. Niesman, Y. Wu, L. E. Benedetti, I. Dunia, E. Levy, and X. Gong Disruption of Gja8 ({alpha}8 connexin) in mice leads to microphthalmia associated with retardation of lens growth and lens fiber maturation Development, January 1, 2002; 129(1): 167 - 174. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Verheule, M. J. A. van Kempen, S. Postma, M. B. Rook, and H. J. Jongsma Gap junctions in the rabbit sinoatrial node Am J Physiol Heart Circ Physiol, May 1, 2001; 280(5): H2103 - H2115. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Stebbings, M. G. Todman, P. Phelan, J. P. Bacon, and J. A. Davies Two Drosophila Innexins Are Expressed in Overlapping Domains and Cooperate to Form Gap-Junction Channels Mol. Biol. Cell, July 1, 2000; 11(7): 2459 - 2470. [Abstract] [Full Text]
|
|
|
|
 |
|
 D. S. He and J. M. Burt Mechanism and Selectivity of the Effects of Halothane on Gap Junction Channel Function Circ. Res., June 9, 2000; 86 (11): e104 - e109. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. S. He, J. X. Jiang, S. M. Taffet, and J. M. Burt Formation of heteromeric gap junction channels by connexins 40 and 43 in vascular smooth muscle cells PNAS, May 25, 1999; 96(11): 6495 - 6500. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Gu, H. L. Roderick, P. Camacho, and J. X. Jiang Identification and characterization of an amino acid transporter expressed differentially in liver PNAS, March 28, 2000; 97(7): 3230 - 3235. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||
1) used
to fit the distributions is shown on top. The binned data are from the
same patch shown in A idealized with a K-segmental
K-means method and the continuous lines were calculated directly from
the model. There appeared to be four amplitude levels in this
experiment corresponding to 0, 1, 2, and 3 simultaneously open
channels.
