Purpose: Rod-cone gap junctions permit transmittal of rod visual information to the cone pathway. A recent report has shown that this transfer does not occur in mice in which the gap junction protein connexin 36 is knocked out indicating that rod-cone gap junctions are assembled from this protein. It remains unresolved, however, whether rods, cones or both express connexin 36. We have tried to address this question with the use of transgenic rod-less and cone-less mice.
Methods: Deletion of Nrl, a transcription factor, results in a complete loss of rods with a concomitant increase in S-cones. We used this as the rod-less (cone-only) model. Cone-less (rod-only) retinas were from mice expressing an attenuated diphtheria toxin gene under the control of a promoter selective for cones. Nearly all long wavelength cones and 95% of short wavelength cones are missing in this model. Fixed retinal sections from these two models and age matched controls were used to detect connexin 36 gap junctions by immunofluorescence.
Results: Punctate immunofluorescence, indicating the presence of gap junctions, was observed in the inner and outer plexiform layers of both wild type and cone-less and rod-less retinas. Our assumption was that immunofluorescence due to photoreceptor gap junctions would be observed in the outer plexiform layer. In all the animals, most of the immunofluorescence was in the inner plexiform layer, with only a marginal reaction in the outer plexiform layer. In cone-only (rod-less) retina, immunofluorescence in the outer plexiform layer increased by more than 20 fold compared to wild type. In rod-only (cone-less) retina, the outer plexiform layer showed about a 30% decrease in immunofluorescence. In both rod-less and cone-less retinas, immunofluorescence in the inner plexiform layer was higher than in the wild type by 25-50%.
Conclusions: Cones constitute only about 3% of photoreceptors in the wild type retina while they make up 100% of the photoreceptors in cone-only retina. This increase in their numbers coincided with a 20 fold increase in immunofluorescence in the outer plexiform layer, strongly suggesting that cones express connexin 36. Conversely, when the cone numbers went down from 3% to near zero in cone-less retina, immunofluorescence decreased by about 30% in the outer plexiform layer, suggesting again that cones express the connexin and that they contribute to its presence disproportionately more than their numbers indicate. The results from both rod-less and cone-less animals are strongly indicative of cones expressing connexin 36, but are not sufficient to conclude whether rods express the protein. An unexpected observation from our experiments is that immunofluorescence increases slightly in the inner plexiform layer in both rod-less and cone-less retina for reasons that need further investigation.