Membrane assembly in retinal photoreceptors. II. Immunocytochemical analysis of freeze-fractured rod photoreceptor membranes using anti-opsin antibodies.

We have used a cytochemical technique for labeling freeze-fractured tissues (Pinto da Silva, P., C. Parkison, and N. Dwyer (1981) Proc. Natl. Acad. Sci. U.S.A. 78: 343-347) to examine the distribution of immunoreactive opsin in rod photoreceptor membranes. Aldehyde-fixed retinas of African clawed frogs (Xenopus laevis) embedded in a cross-linked protein matrix were frozen and fractured at -196 degrees C, then thawed and labeled with biotinylated sheep anti-cow opsin IgG followed by avidin-ferritin. In thin sections of plastic-embedded retinas, rod outer segment (ROS) disc membranes exposed by fracturing bound specific antibody intensely and relatively uniformly. However, they differed from membranes of the inner segment as well as those of erythrocytes in that protoplasmic face leaflets did not assume an interrupted bilayer appearance and disc exoplasmic face leaflets were apparently lost during thawing. The disposition of opsin immunoreactivity in the cell membrane was highly asymmetric. Although ROS plasma membranes from which discs are elaborated labeled heavily with anti-opsin after cleavage, fractures passing along inner segment plasma membranes bound very little antibody. In cross-fractures exposing inner segment cytoplasm, we found specific labeling of Golgi complex elements, as well as both perimitochondrial and periciliary vesicles. The latter are presumed to be the vehicle shuttling newly synthesized membrane to the ROS for disc assembly. These results suggest that opsin-containing membrane is sorted out within the cell, being transported from synthetic sites to the immediate periciliary zone where localized insertion into the cell membrane takes place. Furthermore, the close correspondence of the present immunocytochemical analysis with the distribution of opsin deduced from prior quantitative freeze-fracture analysis (Besharse, J. C., and K. H. Pfenninger (1980) J. Cell Biol. 87: 451-463) offers the possibility that fracture-label may be generally useful for study of patterned membrane topography in neuronal cells.