Gap junctions have traditionally been characterized as nonspecific pores between cells passing molecules up to 1 kDa in molecular mass. Nonetheless, it has become increasingly evident that different members of the connexin (Cx) family mediate quite distinct physiological processes and are often not interchangeable. Consistent with this observation, differences in permeability to natural metabolites have been reported for different connexins, although the physical basis for selectivity has not been established. Comparative studies of different members of the connexin family have provided evidence for ionic charge selectivity, but surprisingly little is known about how connexin composition affects the size of the pore. We have employed a series of Alexa dyes, which share similar structural characteristics but range in size from molecular weight 350 to 760, to probe the permeabilities and size limits of different connexin channels expressed in Xenopus oocytes. Correlated dye transfer and electrical measurements on each cell pair, in conjunction with a three-dimensional mathematical model of dye diffusion in the oocyte system, allowed us to obtain single channel permeabilities for all three dyes in six homotypic and four heterotypic channels. Cx43 and Cx32 channels passed all three dyes with similar efficiency, whereas Cx26, Cx40, and Cx45 channels showed a significant drop-off in permeability with the largest dye. Cx37 channels only showed significant permeability for the smaller two dyes, but at two- to sixfold lower levels than other connexins tested. In the heterotypic cases studied (Cx26/Cx32 and Cx43/Cx37), permeability characteristics were found to resemble the more restrictive parental homotypic channel. The most surprising finding of the study was that the absolute permeabilities calculated for all gap junctional channels in this study are, with one exception, at least 2 orders of magnitude greater than predicted purely on the basis of hindered pore diffusion. Consequently, affinity between the probes and the pore creating an energetically favorable in-pore environment, which would elevate permeant concentration within the pore and hence the flux, is strongly implicated.

ap junctions, the only channels that allow direct exchange of small metabolites between cells, are composed of a family of integral membrane proteins, called connexins in vertebrates. Different connexins have been implicated in a diverse array of biological processes and diseases 1 , indicating that gap-junction properties may vary with connexin composition. For example, channels composed of different connexins have different conductances 2 and permeabilities to ions 3 and fluorescent dyes 4,5 . However, although the permeability of gap junctions to some nucleotides 6 , cyclic AMP 7 , calcium 8 and possibly inositol-1,4,5-trisphosphate 8 has been documented, the comparative selectivities of different connexins for biologically significant molecules remain an enigma. Here, using new techniques, we show that the rate of permeation of metabolites through gap junctions differs according to connexin composition.One demonstration of the biological consequences of expression of different connexin isotypes is their variable effectiveness in suppressing growth of tumour cells 9 . Although transfection of C6 glioma cells with connexin-43 (Cx43) suppressed their in vitro growth potential 10 , Cx32 did not 11 . Paradoxically, however, Cx32 transfectants are far better coupled to each other, as assayed by calcein-dye transfer (Table 1; for methods see ref. 12). A diffusion constant for each cell interface was obtained from the time course of calcein transfer in two C6 cell transfectants 12 , using quantitative modelling of dye diffusion through cell monolayers 5 . Independently, we determined the ratio of calcein permeabilities of Cx43 and Cx32 channels to be 1.08 from quantitative measurements of calcein diffusion between Xenopus oocyte pairs expressing defined numbers of Cx43 or Cx32 channels, calculated from the junctional and single-channel conductance of each connexin (P.A. Weber, Y.I. Chen, J. Nitsche and B.J.N., unpublished observations). Applying the relative calcein permeability of the two connexins to the diffusion constants calculated above for the transfected C6 cell lines produced an estimate of the relative number of channels in the Cx32 versus the Cx43 transfectants of 16:1. Hence, the differential effects of these two connexins on cell growth were related not to their expression level but rather to some specific property of the channels, such as selective permeabilities for endogenous signals.To address this possibility, we have developed techniques with which to 'capture', identify and quantify the junctional transfer of endogenous metabolites between cells transfected with Cx43 or Cx32 (ref. 13). 'Donor' cells were metabolically labelled overnight with [ 14 C]glucose, fluorescently stained with DiI, and plated with unlabelled 'receiver' cells at a 1:6.25 ratio to yield confluent monolayers. After allowing 2 h for plating and establishment of cell contact and communication, we separated donors and receivers from each other by fluorescence-activated cell sorting (FACS). Potential transjunctional molecules, whi...

8410The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abstract. While a number of different gap junction proteins have now been identified, hepatic gap junctions are unique in being the first demonstrated case where two homologous, but distinct, proteins (28,000 and 21,000 Mr) are found within a single gap junctional plaque