Verselis VK and Srinivas M (2008), Divalent cations regulate connexin hemichannels...

XB-ART-38884

J Gen Physiol 2008 Sep 01;1323:315-27. doi: 10.1085/jgp.200810029.

Show Gene links Show Anatomy links

Divalent cations regulate connexin hemichannels by modulating intrinsic voltage-dependent gating.

Verselis VK , Srinivas M .

???displayArticle.abstract???
Connexin hemichannels are robustly regulated by voltage and divalent cations. The basis of voltage-dependent gating, however, has been questioned with reports that it is not intrinsic to hemichannels, but rather is derived from divalent cations acting as gating particles that block the pore in a voltage-dependent manner. Previously, we showed that connexin hemichannels possess two types of voltage-dependent gating, termed V(j) and loop gating, that in Cx46 operate at opposite voltage polarities, positive and negative, respectively. Using recordings of single Cx46 hemichannels, we found both forms of gating persist in solutions containing no added Mg(2+) and EGTA to chelate Ca(2+). Although loop gating persists, it is significantly modulated by changing levels of extracellular divalent cations. When extracellular divalent cation concentrations are low, large hyperpolarizing voltages, exceeding -100 mV, could still drive Cx46 hemichannels toward closure. However, gating is characterized by continuous flickering of the unitary current interrupted by occasional, brief sojourns to a quiet closed state. Addition of extracellular divalent cations, in this case Mg(2+), results in long-lived residence in a quiet closed state, suggesting that hyperpolarization drives the hemichannel to close, perhaps by initiating movements in the extracellular loops, and that divalent cations stabilize the fully closed conformation. Using excised patches, we found that divalent cations are only effective from the extracellular side, indicative that the binding site is not cytoplasmic or in the pore, but rather extracellular. V(j) gating remains essentially unaffected by changing levels of extracellular divalent cations. Thus, we demonstrate that both forms of voltage dependence are intrinsic gating mechanisms in Cx46 hemichannels and that the action of external divalent cations is to selectively modulate loop gating.

???displayArticle.pubmedLink??? 18695008
???displayArticle.pmcLink??? PMC2518728
???displayArticle.link??? J Gen Physiol
???displayArticle.grants??? [+]

Species referenced: Xenopus laevis
Genes referenced: dtl gja3 isyna1

???attribute.lit??? ???displayArticles.show???

	Figure 1. Regulation of macroscopic Cx46 currents by external Ca2+ and Mg2+. (A) Representative current–voltage (I-V) relationships (top) at different external Ca2+ concentrations ranging from nominal (0 added) to 1.8 mM were obtained by applying slow (3 min) voltage ramps from +50 to −100 mV to a Cx46-expressing oocyte. The holding potential before and after the ramp was −40 mV. For each curve, external Mg2+ was maintained at 1 mM. The corresponding conductance–voltage (G-V) relationships are plotted below and show that conductance is maximum at or near 0 mV and decreases asymmetrically with voltages of either polarity. Conductance decreases robustly with hyperpolarization, but only modestly with depolarization. Increasing the extracellular Ca2+ concentration caused a significant decrease in overall conductance and led to a substantial rightward shift in voltage sensitivity at negative, but not positive potentials. (B) I-V (top) and corresponding G-V (bottom) relationships obtained at different external Mg2+ concentrations ranging from 1 to 20 mM. External Ca2+ was maintained at nominal levels. Increasing external Mg2+, like Ca2+, decreased overall conductance and caused a significant positive shift in voltage sensitivity at negative potentials. However, these changes required considerably higher concentrations of Mg2+. (C) Shown are representative currents obtained at different concentrations of external Ca2+ and Mg2+. Oocytes were clamped to a holding potential of −40 mV and 5-s voltage steps were applied from +60 to −100 mV followed by a 5-s step to −120. The currents were normalized to a constant prepulse (+10 mV) that preceded each episode. Both Ca2+ and Mg2+ had similar effects, suppressing the magnitude of the current, slowing activation and speeding up deactivation.
	Figure 2. Divalent cations selectively affect gating of Cx46 hemichannels at negative membrane potentials. (A and B) Multichannel cell-attached patch recordings from a Cx46-expressing oocyte recorded with (A) 1 mM Mg2+ present in the patch pipette and (B) no added Mg2+ in the patch pipette. In both cases, Ca2+ was buffered to very low levels with EGTA (Materials and methods). Shown are currents in response to 10-s voltage steps from −70 and +50 mV applied in 20-mV increments. No endogenous channel activity was present these patches. In the presence of 1 mM external Mg2+, Cx46 hemichannels essentially remained open at modest voltages (see traces at −10 and +10 mV) and tended to close for either polarity of membrane voltages. Hyperpolarization led to full and robust closure (see trace at −70 mV), whereas depolarization left a residual current (see current trace at +50 mV). The current records were leak subtracted as described in the Materials and methods. When there was no added external Mg2+, gating at positive voltages did not change, whereas gating at negative voltages changed significantly. In the example shown, gating was essentially abolished at hyperpolarizing voltages up to −70 mV. (C) Cell-attached patch recordings of single Cx46 hemichannels highlighting the two different types of gating at negative (−70 mV) and positive (+40 mV) voltages, termed loop gating and Vj gating, respectively. Recordings were obtained with standard IPS (containing 1 mM Mg2+). Loop gating is characterized by gating transitions between fully closed (C) and open (O) states (left), whereas Vj gating transitions occur between the open state (O) and a long-lasting subconductance state (right). Currents were leak subtracted as described in Materials and methods. All currents were filtered at 1 kHz and data were acquired at 5 kHz.
	Figure 3. Voltage gating is intrinsic to Cx46 hemichannels. Shown are multichannel cell-attached patch recordings from a Cx46-expressing oocyte with (A) 1 mM Mg2+ present in the patch pipette and (B) with no added Mg2+ in the patch pipette. In both cases, Ca2+ was buffered to very low levels with EGTA. When the extracellular solution (patch pipette solution) contained 1 mM Mg2+, membrane hyperpolarization to −70 mV and depolarization to +90 mV produced reductions in current characteristic of loop gating and Vj gating, respectively. Currents were leak subtracted (dotted lines designate the 0 current level). When the extracellular solution was essentially devoid of external divalent cations, Vj gating at +90 mV was unaffected and loop gating at −70 mV was abolished, but application of a very large hyperpolarizing step to −160 mV caused a robust and rapid decrease in current. Although the current clearly declined toward zero, it remained very flickery and noisy. All currents were filtered at 1 kHz and data were acquired at 5 kHz.
	Figure 4. External divalent cations stabilize loop-gating closures induced by hyperpolarization. (A–C) Recordings of a single Cx46 hemichannel in an excised outside-out patch. The patch pipette solution contained standard IPS solution (Materials and methods). (A) Recordings obtained with no added Mg2+ on the extracellular side (bath) of the hemichannel. Shown are 30-s segments of current recorded at membrane potentials of −30, −50, −70, −90, −100, and −130 mV. All points histograms obtained from longer (60–120 s) recordings are shown to the right of each trace. Dashed lines designated C and O in the current traces and the corresponding histograms represent fully closed and fully open states, respectively (currents were not leak subtracted). In the absence of divalent cations, the hemichannel clearly exhibits voltage dependence with hyperpolarization tending toward closure. At large hyperpolarizing voltages, the hemichannel rarely resides in the fully open state and exhibits very noisy flickering (see spread in the histograms away from the fully closed state). (B) Addition of 1 mM Mg2+ to the bath dramatically changed the behavior of the hemichannel. Stable, long-duration closures became evident and considerably more modest voltages produced closure (compare histograms at −90 mV with and without added Mg2+). (C) View of the current trace recorded at −130 mV with no added Mg2+ at an expanded time scale. The solid bar indicates the segment of the trace that was expanded below. The noisy flickering associated with closure can be seen to be occasionally interrupted by full closures that are quiet (indicated by arrows). Asterisk denotes a brief opening event. (D) Plots of Po vs. Vm obtained from the recordings in A and B. Po represents the fraction of time spent in the fully open state (see Materials and methods). These data illustrate that in solutions essentially devoid of divalent cations, voltage dependence remains robust although significantly shifted in the hyperpolarizing direction. Currents were filtered at 2 kHz and data were acquired at 10 kHz.
	Figure 5. The binding site for divalent cations is extracellular. (A) Recording of a single hemichannel in an outside-out patch configuration with 1 mM Mg2+ on the extracellular side and no added Mg2+ on the cytoplasmic side. Openings observed at −70 mV became more sporadic and brief upon hyperpolarization to −90 mV. (B) Recording of a single hemichannel in an outside-out patch configuration with no added Mg2+ on the extracellular side and 1 mM Mg2+ on the cytoplasmic side. In the presence of 1 mM Mg2+ on the cytoplasmic side, the hemichannel largely remained open at −70 mV. Closure increased with hyperpolarization to −90 mV and was robust at −130 mV. These large hyperpolarizing voltages would tend to prevent Mg2+, which is present only on the cytoplasmic side, from entering the pore inconsistent with voltage-dependent entry of Mg2+ leading to block. (C) Patch recording containing three active Cx46 hemichannels in an outside-out patch configuration. Membrane potential was held constant at −50 mV. Starting in symmetric conditions, with 1 mM Mg2+ present on both sides, perfusion of a solution containing 10 mM Mg2+ into the bath (as indicated), which exposed the extracellular side to a high concentration of Mg2+, produced rapid and robust closure. (D) The same experimental paradigm as in C applied to an inside-out patch containing a single hemichannel. Perfusing a solution containing 10 mM Mg2+ into the bath, which now exposed the cytoplasmic side to a high concentration of Mg2+, did not appreciably affect hemichannel behavior except for a small reduction in the magnitude of the unitary current. Shown below the recording are histograms taken from the indicated segments (solid bars) before and after application of 10 mM Mg2+. All currents were leak subtracted. Dashed lines indicate fully open (O) and fully closed (C) states. Currents were filtered at 1 kHz and data were acquired at 5 kHz.
	Figure 6. Examination of the effects of Mg2+ over a voltage range. (A) Single hemichannel currents obtained with 8-s voltage ramps from −70 to +70 mV applied to excised patches. Shown are examples of three current traces in response to voltage ramps applied in succession to a patch in which the Mg2+ concentration was 1 mM on both sides (left), elevated to 10 mM on the cytoplasmic (middle) and elevated to 10 mM on the extracellular (right). The currents were leak subtracted. (B) Current–voltage relationships of Cx46 hemichannels constructed from ensemble averaged currents from multiple patches in response to ±70 mV, 8-s voltage ramps applied to excised patches as shown in A. The I-V relationship with 10 mM Mg2+on the cytoplasmic side was essentially similar to that obtained in symmetric 1 mM Mg2+, except for a small reduction in unitary current and a small negative shift in Erev. In contrast, the same concentration of Mg2+ applied to the extracellular side showed strong voltage dependence with hemichannels closing at inside negative voltages and opening at positive voltages. The strong voltage dependence of closure observed only when Mg2+ is elevated extracellularly but not cytoplasmically argues against a pore-blocking mechanism and is consistent with an extracellular modulatory site, which when bound with divalent cations causes a positive shift in activation to more positive voltages.

References [+] :

Baumgartner, Two-microelectrode voltage clamp of Xenopus oocytes: voltage errors and compensation for local current flow. 1999, Pubmed, Xenbase