|
Figure 1. Schematic representation of the dual oocyte perfusion apparatus that allows both cytoplasmic access to gap junction channels and monitoring of intercellular conductance. The chambers, designed by F. Cao, B.J. Nicholson, and G. Nottingham (Cao, 1997), are assembled from two halves to form compartments that remain separated by a thin coverslip with a small (0.7-mm diam) hole drilled centrally (A). Oocytes, in 600 µl of L15 media per compartment, establish coupling through the hole in the coverslip, with which the oocyte membrane forms a low resistance seal. During perfusion experiments, one oocyte is cut open with fine dissecting scissors, and at this time, the seal resistance replaces the membrane resistance of the open cell as shown in the equivalent circuit (B and C).
|
|
Figure 2. Membrane topology of Cx32 showing all sites mutated to cysteine and analyzed in the current study. Cysteine substitutions were made at the indicated sites using overlapping PCR. Nonfunctional mutants are indicated with an “X.” At seven sites, cysteine substitution resulted in functional channels with a reverse-gating response to voltage (bold circles). Other sites where cysteine substitution altered channel properties (shaded circles) include A88 in M2, where cysteine substitution appears to induce hemichannel formation, and E146 in M3, where cysteine substitution results in a disulfide bond with C201 in M4 (Fig. 3, E and F). The topological plot is modified from Milks et al. (1988) to accommodate the empirical data demonstrating this proximity of E146 and C201. Extension of transmembrane spans beyond the margins shown here would be required for helices that are significantly tilted from the normal to the bilayer (helices C and D in Unger et al. [1999], corresponding to M3 and M4; Fig. 9).
|
|
Figure 3. Transjunctional currents recorded from paired Xenopus oocytes expressing wtCx32 and several representative cysteine substitution mutants. Oocytes expressing mutants V136C (B), A196C (C), and E146CC201S (E) were paired homotypically (mutant/mutant), and showed gating responses minimally modified from wild-type Cx32 (A). Note that the E146C mutant alone was nonfunctional, and currents could only be recorded for the double mutant or after DTT perfusion of the single mutant (F). (D) M34C was paired heterotypically with wtCx32, with positive polarity defined relative to the mutant-expressing oocyte. These channels, which are closed at rest and open in response to positive transjunctional voltages, are an example of the reverse-gating phenotype discussed in Fig. 2 and the text. In all cases, transjunctional currents were recorded from oocytes clamped at −40 mV, whereas the partner was pulsed in 20-mV increments from the same holding potential.
|
|
Figure 4. Transjunctional currents recorded from paired Xenopus oocytes expressing Cx32V37C show minimal changes on perfusion. Oocytes were paired homotypically (mutant/ mutant) in the perfusion chamber, and voltage sensitivity of the junctions was assessed in the intact (A) and perfused (B) configuration. Junctional currents were altered only slightly after perfusion. Transjunctional currents were recorded from the intact oocyte clamped at −20 mV, whereas the partner was pulsed in 20-mV increments from the same holding potential.
|
|
Figure 5. Structure of the thiol reagent maleimidobutyryl biocytin (MBB), and detection of its reactivity with Cx32 in the perfusion chamber. (A) Irreversible reaction occurs between the SH group of a cysteine and the maleimide ring, with the biotin moiety providing easy detection. (B) Biochemical determination that MBB reacts with Cx32 in the dual-oocyte perfusion system. Radiolabeled Cx32, immunoprecipitated from intact oocytes, shows a full-length and partially degraded form just below the 30-kD marker, and an aggregated dimer at ∼50 kD (left lane). Cx32 detected by avidin blot after intracellular perfusion of a paired oocyte with 1 mM MBB and immunoprecipitation of Cx32 (second lane from left) reveals the same bands, although the increased sensitivity of avidin detection revealed additional aggregated forms, both dimeric (50–60 kD) and trimeric (∼90 kD). No protein is detected for noninjected oocytes exposed to MBB or for perfused Cx32-injected oocytes not exposed to MBB (left two lanes). (C and D) Changes in junctional conductance (Gj) during typical SCAM experiments. In most cases, conductance increased steadily before and after addition of the reagent as shown for wtCx32 (C). In cases where a substituted cysteine was accessible, a decrease in conductance occurred after the addition of MBB, as shown for Cx32Y135C (D). The effect of MBB on each oocyte pair was determined by comparison of the conductance immediately before addition of the reagent, to that 20 min after its application (ΔGj).
|
|
Figure 6. Mean percent changes in conductance after MBB perfusion. For Cx32 and the majority of cysteine substitution mutants, a small increase in conductance occurs after addition of the reagent, which represents a continuation of the trend in conductance before the application of MBB (open bars). A significant decrease in conductance compared to wt was observed after application of MBB to a number of cysteine substitution mutants (black bars, P < 0.01; shaded bars, 0.01 < P < 0.05). In all of these cases, block was also observed in ≥50% of perfusions (>60% for black bars). Bolded open bars indicate sites where cysteine substitution altered the gating properties of the channel. Hence, reactivity does not necessarily infer that the native amino acid lies in an accessible position. Error bars represent SEM. Numbers adjacent to the bars indicate the number of experiments where conductance was reduced versus the number of experiments conducted.
|
|
Figure 7. Changes in conductance after extracellular MBB application to intact oocyte pairs where access to the pore was not possible. As in the perfusion experiments in Fig. 6, wtCx32 and cysteine substitution mutants in M2 and M3 continued to show small increases in conductance after addition of the reagent. In contrast, significant decreases in conductance compared to wt (P ≤ 0.05; t test) were observed after application of MBB to a number of cysteine substitution mutants in M1 and M4 (filled bars). In all of these cases, block was also observed in >50% of perfusions. Reactivity suggests that the designated cysteines lie within an MBB-accessible environment, continuous with the extracellular space but not in the pore. Data are presented in the same format as Fig. 6.
|
|
Figure 8. Position of all significantly reactive cysteines in Cx32. Solid filled circles in M2 and M3 indicate sites where channel block by MBB was highly significant (P < 0.01), whereas boldly circled residues with lighter shading (V84 and L89 in M2; F141 and Y151 in M3) showed conductance changes that were slightly less significant (0.01 < P < 0.05). Lightly shaded residues in M1 (I33) and M4 (A196 and S198) showed block in response to extracellular MBB application, and are not thought to line the pore. Boldly circled open sites indicate reactive positions where cysteine substitution induced reversed gating. Reactivity of these residues suggests that they are accessible in the closed state, but not necessarily in the native, open conformation.
|
|
Figure 9. Models of the pore-lining helices of connexins showing reactive sites and the proposed assignment of helices in the gap junction structure. (A) Reactive sites span the length of the M3 helix, which appears to lose its helical periodicity at its extracellular end (displayed as a “break” in the helix). The positioning of M4 is based on the observation that a disulfide bond is capable of forming between C201 in M4, and a substituted cysteine at position 146 in M3. (B) Helical assignments to the gap junction structure of Unger et al. (1999). The main pore-lining helix (M3) is assigned with certainty based on the accessibility of M3 residues in both open and closed channels. The M4 assignment is based on evidence for disulfide formation. There is less certainty regarding the assignments of M1 and M2. It has been proposed (see Discussion; Helical assignments) that the structure shown in B is in the closed state, where M1 would contribute to the pore at the widened cytoplasmic mouth. However, our data indicate that M2 would contribute to the pore mouth in the open state.
|
|
Figure 10. Aligned amino acid sequences for the third transmembrane domain of Cx32 and several other connexins. Highlighted residues indicate pore-lining residues in Cx32 and their corresponding sites in other characterized connexins (open-boxed site only tested in the closed state). The periodicity in the NH2-terminal two-thirds of M3 suggests α-helix, but this breaks down at the COOH-terminal end. Many sites are strictly conserved, whereas others show relatively minor changes despite reports of significant differences in permeability between connexins. This suggests that subtle differences in residue length or branching may define the affinity of binding sites within the pore.
|
