Innexins Form Two Types of Channels

Oocytes

Oocytes were prepared as described previously [15]. In brief, Xenopus laevis oocytes were isolated by incubating small pieces of ovary in 2 mg ml⁻¹ collagenase in calcium free oocyte Ringer’s solution (OR2) and stirring at 1 turn/second for 3 hours at room temperature. After being thoroughly washed with regular OR2, oocytes devoid of follicle cells and having uniform pigmentation were selected and stored in OR2 at 18 °C.

In vitro transcription of mRNAs

Hirudo innexins 2 and 3 (Hminx2 and Hminx3) had been cloned into the expression vector pCR-BluntII-TOPO. mRNA was transcribed by SP6 (Hminx2) or T7 (Hminx3) RNA polymerase from 10 μg of XbaI- (Hminx2) or SpeI- (Hminx3) linearized plasmid using the mMessage mMachine kit (Ambion). mRNA was quantified by absorbance (260nm), and the proportion of full-length transcripts was checked by agarose gel electrophoresis. 20 nl of mRNA (50 ng/μl) was injected into oocytes.

Solutions

OR2 solution in mM: 82.5 NaCl, 2.5 KCl, 1.0 MgCl₂, 1.0 CaCl₂, 1.0 Na₂HPO₄, 5.0 HEPES, antibiotics (Penicillin, 10,000 units/ml; Streptomycin, 10 mg/ml), pH7.5). Patch pipette KGlu solution: 140 mM potassium gluconate, 10 mM KCl, 5.0 mM TES, pH 7.5)

Voltage clamp

Whole cell voltage clamp recording was performed with two intracellular microelectrodes as described [15].

Patch clamp technique

Single innexon channels were studied by the patch-clamp technique [16] using a WPC-100 amplifier (E.S.F. Electronics, Goettingen, Germany). Currents were filtered at 5 kHz, digitized using a VR-10B digital data recorder, and stored on video tape. The recordings were transferred to a Power Macintosh (Apple) computer using an ITC-18 Computer Interface (Instrutech Corporation) and analyzed. Acquisition and analysis were done with the Acquire and TAC programs (both from Bruxton Corporation, Seattle, WA).

Details of patch clamp techniques used, including application of negative pressure, can be found elsewhere [17,18].

Extracellular ATP measurements from oocytes

ATP assay solutions (Luciferin/Luciferase, Sigma, St. Louis) were mixed with supernatants collected from hm-inx2 injected and uninjected cells treated with oocyte Ringer’s solution (OR2) or potassium gluconate (KGlu) in the presence of 0–100 μM carbenoxolone (CBX). Oocytes were 4 days post-injection. Innexin expression and cell viability were confirmed electrophysiologically. Cells were pretreated for 10 minutes in experimental solutions and then isolated for 10 minutes in 150 μL of the same experimental solutions. 100 μL supernatant was obtained for each condition. Each condition was done in quintuplicate. Luminescence readings were obtained with a Victor 1420 multilabel counter (PerkinElmer, Waltham, MA) on a 96 well culture plate. ANOVA with Post-Hoc Fisher test (Statistica, StatSoft Inc., Tulsa, OK) was used for comparison and statistical analysis of luminescence.

In situ dye release

6-Carboxyfluorescein was injected into a single glia cell of the leech connective and dye release was followed under a fluorescence microscope. Details of the technique are given in Supplement 1 and Supplemental Figure 1.

Innexins form innexons

Both Hminx2 and Hminx3 formed patent “hemichannels” (innexons) in the non-junctional plasma membrane (Fig. 1a, b). Like pannexin 1, but in contrast to “hemichannels” formed by some connexins, no change of extracellular calcium was required to observe innexon channel activity. The channels were closed at the resting membrane potential (−40 mV) and opened when the membrane was depolarized to –20 mV or higher (Fig. 1a, b). Cytoplasmic acidification, induced by perfusing the oocytes with CO₂-saturated Ringer’s solution, closed the channels (Fig. 1c, d). This effect slowly reversed upon washout (Supplemental Figure 2). Moreover, carbenoxolone closed the channels in a dose dependent way (Fig. 1e, Supplemental Figure 3). In Figure 1b, as shown previously [19], sham injected oocytes did not exhibit a current that was activated in this voltage range and that was sensitive to cytoplasmic acidification.

To test whether the formation of innexons might be a general property of innexins, two other innexins, Hminx1 and Hminx6, which are found in leech neurons, were also expressed in oocytes. Both formed non-junctional membrane channels with properties similar to those of the channels formed by Hminx2 and Hminx3 (Supplemental Figure 4).

To test for ATP permeability of innexons, the efflux of ATP from oocytes expressing Hminx2 was measured by a luminometric luciferase assay. As reported previously, uninjected control oocytes released some ATP when depolarized by a high potassium solution (KGlu) (Figure 1f), [18,20] which probably represents the Brefeldin sensitive vesicular release [20]. Hminx2 expressing oocytes had the same basal release as control oocytes. However, when depolarized with high potassium solution a significant increase in ATP release was observed (Figure 1f). Carbenoxolone inhibited this ATP release in a dose dependent way (Figure 1f).

Consistent with the macroscopic membrane currents, expression of Hminxs1,2,3 and 6 in single Xenopus oocytes resulted in the appearance of a novel type of membrane channel, with a large unitary conductance of about 500 pS (Hminxs2,3 and 6) and about 250 pS (Hminx1) in symmetrical 150 mM KCl (Fig. 2a, Supplemental Figure 4). The channels were closed at negative potentials and open at positive potentials. The voltage gate was slow, and consequently the channel remained open for tens of seconds after switching from positive to negative potential. The channel activity was attributable to innexins because such large conductance channels were not observed in control oocytes and because of the distinct features of the channel, which were quite similar to pannexin 1 channels expressed in oocytes [23]. Like pannexin 1 channels, all leech innexin channels tested here, when active, exhibited multiple subconductance states at which they mainly dwelled, while sojourns to the full open and closed states were infrequent (Fig. 2a, Supplemental Figure 4). For a more detailed characterization of the channels we concentrate hereafter on those formed by Hminx2.

Innexon channels are mechanosensitive

It has been reported that pannexon channels are sensitive to mechanical stress [18]. To test whether innexons are also mechanosensitive, we used cell-attached and excised membrane patches from oocytes expressing Hminx2. After identification of the innexons, based on single channel conductance, voltage dependence and gating properties (archetypal subconductance levels), the channels were subjected to mechanical stress by suction applied to the patch pipette [21]. Fig. 2b shows a typical response of Hminx2 channels to stretch. When the patch, containing at least three channels, was held at −50 mV, mechanical stress caused activity to increase. The channel activity initially was in the form of subconductance states before full openings were discerned. For quantitative analysis, open probability was determined before and during stretch (Fig. 2b). During prolonged stretch, channel activity often subsided spontaneously (not shown).

Oocytes are known to express an endogenous stretch activated channel [22–24]. However, the channel properties including unit conductance (~50pS), kinetics, and calcium block of this endogenous channel [17,24] are distinct from those of the large conductance channel seen only in innexin mRNA injected oocytes (Figure 2).

Innexon channels are activated by cytoplasmic calcium

Pannexin1 channels can be activated by cytoplasmic calcium. We tested whether the functional similarity of pannexin and innexin channels also extend to this property. Oocytes endogenously express a calcium activated chloride channel that could potentially interfere with the analysis. However, the innexon channels can be discriminated easily from the endogenous calcium activated chloride channel based on their unitary conductances of ~500 pS and 3 pS [25], respectively.

After identification of innexon channels at potentials permissive for channel opening, the channels were held at −50 mV. At this potential, channels were closed or stayed in low subconductance states. Application of calcium to the intracellular side of the channels contained in excised inside-out membrane patches resulted in channel activity that ceased upon washout of the calcium containing solution (Supplemental Fig. 5). Free calcium concentrations in the micromolar range were sufficient to elicit this effect. Current recordings from inside-out patches containing multiple Hminx2 channels were used to determine the Ca²⁺ dose-response relationship (Supplemental Fig. 5b). When the free [Ca²⁺] was 10⁻⁶ M or less, the channels opened briefly and resided in low subconductance states. When the free [Ca²⁺] was 10⁻⁵ M or higher, multiple channels opened, resulting in rapid increases in patch currents.

Repeated applications of 10⁻⁵ M [Ca²⁺] to the intracellular side resulted in decreased current amplitudes and slower activation kinetics (Supplemental. Fig. 6). It is unclear whether this adaptation phenomenon was due to complete inactivation of individual channels or whether the majority of channels were driven into lower subconductance states.

Innexons function in vivo

Although measurements using frog oocytes show that Hminx1, 2, 3 and 6 form innexons, direct study of the leech nervous system is required to determine whether innexins normally form innexons in vivo. One measure of the presence of pannexons and other “hemichannels” has been the uptake or loss across the plasmalemma of small molecules of dye that can be specifically blocked by drugs such as carbenoxolone [26], which we showed above blocks Hminx2 innexon activity in frog oocytes (Fig. 1e,f).

A single ~5 mm-long, glial cell expressing both Hminx2 and Hminx3 ensheathes all the axons of each connective between ganglia along the nerve cord of the leech (Supplementary Figure 1a) [27]. To determine whether Hminx2 and Hminx3 form non-junctional innexon channels in the living nervous system as they do in oocytes, dye release studies were done on live connective glial cells. The dye 6-carboxyfluorescein (6-CF), which is gap junction permeable for leech glia, was directly injected into the connective glial cell using small negative currents (−2 nA). Injection was restricted to one of the two glial cells of the connective; the second served as control. Both cells were imaged at 5 minute intervals to determine the rate of dye loss (see Methods). In leech saline solution, approximately 15% of the dye was lost from the connective every 5 minutes. This dye loss was almost completely blocked in saline with 10 μM carbenoxolone. This effect was reversed by rinsing the carbenoxolone from the tissue, as shown for a representative example in Figure 3a. Figure 3b shows the average of 6 experiments. Carbenoxolone’s attenuation of dye release demonstrated that non-junctional channels were involved in the passage of dye across the membrane.