Polarized sorting of rhodopsin on post-Golgi membranes in frog retinal photoreceptor cells.

We have isolated a subcellular fraction of small vesicles (mean diameter, 300 nm) from frog photoreceptors, that accumulate newly synthesized rhodopsin with kinetics paralleling its appearance in post-Golgi membranes in vivo. This fraction is separated from other subcellular organelles including Golgi and plasma membranes and synaptic vesicles that are sorted to the opposite end of the photoreceptor cell. The vesicles have very low buoyant density in sucrose gradients (rho = 1.09 g/ml), a relatively simple protein content and an orientation of rhodopsin expected of transport membranes. Reversible inhibition of transport by brefeldin A provides evidence that these vesicles are exocytic carriers. Specific immunoadsorption bound vesicles whose protein composition was indistinguishable from the membranes sedimented from the subcellular fraction. Some of these proteins may be cotransported with rhodopsin to the rod outer segment; others may be involved in vectorial transport.

Immunocytochemical localization of opsin in the cell membrane of developing rat retinal photoreceptors.

Mature retinal rod photoreceptors sequester opsin in the disk and plasma membranes of the rod outer segment (ROS). Opsin is synthesized in the inner segment and is transferred to the outer segment along the connecting cilium that joins the two compartments. We have investigated early stages of retinal development during which the polarized distribution of opsin is established in the rod photoreceptor cell. Retinas were isolated from newborn rats, 3-21 d old, and incubated with affinity purified biotinyl-sheep anti-bovine opsin followed by avidin-ferritin. At early postnatal ages prior to the development of the ROS, opsin is labeled by antiopsin on the inner segment plasma membrane. At the fifth postnatal day, as ROS formation begins opsin was detected on the connecting cilium plasma membrane. However, the labeling density of the ciliary plasma membrane was not uniform: the proximal cilium was relatively unlabeled in comparison with the distal cilium and the ROS plasma membrane. In nearly mature rat retinas, opsin was no longer detected on the inner segment plasma membrane. A similar polarized distribution of opsin was also observed in adult human rod photoreceptor cells labeled with the same antibodies. These results suggest that some component(s) of the connecting cilium and its plasma membrane may participate in establishing and maintaining the polarized distribution of opsin.

Identification of an outer segment targeting signal in the COOH terminus of rhodopsin using transgenic Xenopus laevis.

Mislocalization of the photopigment rhodopsin may be involved in the pathology of certain inherited retinal degenerative diseases. Here, we have elucidated rhodopsin's targeting signal which is responsible for its polarized distribution to the rod outer segment (ROS). Various green fluorescent protein (GFP)/rhodopsin COOH-terminal fusion proteins were expressed specifically in the major red rod photoreceptors of transgenic Xenopus laevis under the control of the Xenopus opsin promoter. The fusion proteins were targeted to membranes via lipid modifications (palmitoylation and myristoylation) as opposed to membrane spanning domains. Membrane association was found to be necessary but not sufficient for efficient ROS localization. A GFP fusion protein containing only the cytoplasmic COOH-terminal 44 amino acids of Xenopus rhodopsin localized exclusively to ROS membranes. Chimeras between rhodopsin and alpha adrenergic receptor COOH-terminal sequences further refined rhodopsin's ROS localization signal to its distal eight amino acids. Mutations/deletions of this region resulted in partial delocalization of the fusion proteins to rod inner segment (RIS) membranes. The targeting and transport of endogenous wild-type rhodopsin was unaffected by the presence of mislocalized GFP fusion proteins.

Immunocytochemical localization of opsin in outer segments and Golgi zones of frog photoreceptor cells. An electron microscope analysis of cross-linked albumin-embedded retinas.

Adult vertebrate retinal cells (rod and cones) continuously synthesize membrane proteins and transport them to the organelle specialized for photon capture, the outer segment. The cell structures involved in the synthesis of opsin have been identified by means of immunocytochemistry at the electron microscope level. Two indirect detection systems were used: (a) rabbit antibodies to frog opsin were localized with ferritin conjugated F(ab')2 of sheep antibodies to rabbit F(ab')2 and (b) sheep antibodies to cattle opsin were coupled to biotin and visualized by means of avidin-ferritin conjugates (AvF). The reagents were applied directly to the surface of thin sections of frog retinal tissues embedded in glutaraldehyde cross-linked bovine serum albumin (BSA). Specific binding of anti-opsin antibodies indicates that opsin is localized in the disks of rod outer segments (ROS), as expected, and in the Golgi zone of the rod cell inner segments. In addition, we observed quantitatively different labeling patterns of outer segments of rods and cones with each of the sera employed. These reactions may indicate immunological homology of rod and cone photopigments. Because these quantitiative variations of labeling density extend along the entire length of the outer segment, they also serve to identify the cell which has shed its disks into adjacent pigment ipithelial cell phagosomes.

Fine structure of a periciliary ridge complex of frog retinal rod cells revealed by ultrahigh resolution scanning electron microscopy.

We studied the junctional region between rod inner segments (RIS) and outer segments (ROS) in frog retinas by high resolution scanning electron microscopy (SEM). Retinas of dark adapted or light exposed Rana pipiens were critical-point-dried and RIS and ROS were split and coated with ultrathin metal films of niobium and chromium--or decorated with gold--and imaged in a new SE-I imaging mode. The connecting cilium (CC) usually broke at the base of the RIS and remained attached to the ROS. The outer part of the CC plasmalemma expanded to form liplike protrusions beyond which disks evaginated with successively larger diameter until they reached the full width of the ROS. The CC rose out from an invagination of the RIS apical plasma membrane (PM). On the lateral walls of this invagination, a highly ordered complex of nine symmetrically arrayed ridges and grooves rose steeply and extended laterally approximately 0.4-1 micron on the adjacent RIS PM. On the apical plasmalemma, the ridges and grooves formed groups of three to four parallel rows that surrounded the invagination. The grooves were bridged by filaments anchored at the top edges of the ridges. This highly ordered structure we term the periciliary ridge complex (PRC). Its ninefold symmetry apparently reflects the 9 + 0 microtubule organization of the CC axoneme. The three-dimensional structure revealed by SEM was confirmed by transmission electron microscopy (TEM) of sections of Epon-embedded retinas. TEM-immunocytochemistry on thin sections of retinas embedded in glutaraldehyde cross-linked albumin suggested that the PRC and the CC may participate in opsin transport and disk morphogenesis.

Immunocytochemical localization of a large intrinsic membrane protein to the incisures and margins of frog rod outer segment disks.

Immunocytochemical techniques have localized a large protein which is an intrinsic membrane component of isolated frog rod outer segments (ROS). This large protein whose apparent mol wt is 290,000 daltons comprises about 1--3% of the ROS membrane mass. Its molar ratio to opsin is between 1:300 and 1:900. Adequate immune responses were obtained with less than 30 microgram (100 pmol) of antigen per rabbit. Antibodies to the large protein were used for its localization on thin sections of frog retina embedded in glutaraldehyde cross-linked bovine serum albumin (BSA). Specifically bound antibodies were detected by an indirect sequence with ferritin-conjugated antibodies. This technique detected the protein which is represented by 1,000--3,000 molecules per disk. This indicates that the procedure is sufficiently sensitive for analysis of membrane components in low molar proportions. The large protein was specifically localized to the incisures of ROS disks which divide the disks into lobes and to the disk margin. Thus, opsin is mobile within the membrane of the disk while the large protein is apparently constrained to the disk edges. This finding raises the possibility that special functions are also localized ot his unusual region of high curvature, and that collisions of bleached opsin with these edges are physiologically important in couter segment function.

Actin in the photoreceptor connecting cilium: immunocytochemical localization to the site of outer segment disk formation.

Actin has been localized in Rana pipiens retinas that were fixed and embedded in aldehyde cross-linked BSA. Thin sections were reacted sequentially with (a) affinity-purified antiactin antibodies induced in rabbits; (b) biotinyl-sheep anti-rabbit antibodies; and (c) avidin-ferritin conjugates. As expected, antiactin labeling density was high in the apical pigment epithelial cell processes and in the calycal processes of photoreceptors. Actin was also localized in a new site. The connecting cilium that joins the inner and outer segments of both rods and cones was heavily labeled by antiactin at its outer segment (OS), or distal, end. In this region of the cilium, the plasma membrane evaginates to form new OS disks and these basal disks were labeled in some instances. Below the new disks in rods, the cytoplasm of liplike expansions of the distal cilium was also heavily labeled. The plasma membrane and interior of the connecting cilium and the remainder of the OS were unlabeled. These findings suggest that actin may participate in the vectorial transport of opsin and other intrinsic membrane proteins that are incorporated into newly forming OS disks. The results also implicate actin in the membrane expansion involved with OS disk formation.

Membrane protein analysis by two-dimensional immunoelectrophoresis.

Water-soluble membrane proteins may be analyzed by a new, rapid technique that combines electrophoresis on high-resolution sodium dodecyl sulfate (SDS) polyacrylamide gels and immunoelectrophoresis. After separation in the first dimension by electrophoresis in SDS, the proteins are subjected to a second electrophoresis at right angles through a two-layered buffered agarose gel. They first pass through a layer containing Lubrol PX which forms complexes with free SDS and then into an antiserum layer where antigen-antibody precipitates form. Precipitin arcs appear at positions corresponding to the antigens separated in the first dimension. The effectiveness of the technique was demonstrated with frog and cattle opsins, human erythrocyte membrane proteins, and their rabbit antiserums and for several water soluble proteins. By this method two fundamental parameters, molecular weight and antigenicity, may be readily used for analysis of membrane proteins.

Mutant rab8 Impairs docking and fusion of rhodopsin-bearing post-Golgi membranes and causes cell death of transgenic Xenopus rods.

Rab8 is a GTPase involved in membrane trafficking. In photoreceptor cells, rab8 is proposed to participate in the late stages of delivery of rhodopsin-containing post-Golgi membranes to the plasma membrane near the base of the connecting cilium. To test the function of rab8 in vivo, we generated transgenic Xenopus laevis expressing wild-type, constitutively active (Q67L), and dominant negative (T22N) forms of canine rab8 in their rod photoreceptors as green fluorescent protein (GFP) fusion proteins. Wild-type and constitutively active GFP-rab8 proteins were primarily associated with Golgi and post-Golgi membranes, whereas the dominant negative protein was primarily cytoplasmic. Expression of wild-type GFP-rab8 had minimal effects on cell survival and intracellular structures. In contrast, GFP-rab8T22N caused rapid retinal degeneration. In surviving peripheral rods, tubulo-vesicular structures accumulated at the base of the connecting cilium. Expression of GFP-rab8Q67L induced a slower retinal degeneration in some tadpoles. Transgene effects were transmitted to F1 offspring. Expression of the GFP-rab8 fusion proteins appears to decrease the levels of endogenous rab8 protein. Our results demonstrate a role for rab8 in docking of post-Golgi membranes in rods, and constitute the first report of a transgenic X. laevis model of retinal degenerative disease.

The proliferative and apoptotic activities of E2F1 in the mouse retina.

The E2F1 transcription factor controls cell proliferation and apoptosis. E2F1 activity is negatively regulated by the retinoblastoma (RB) protein. To study how inactivation of Rb and dysregulated E2F1 affects the developing retina, we analysed wild-type and Rb(-/-) embryonic retinas and retinal transplants and we established transgenic mice expressing human E2F1 in retinal photoreceptor cells under the regulation of the IRBP promoter (TgIRBPE2F1). A marked increase in cell proliferation and apoptosis was observed in the retinas of Rb(-/-) mice and TgIRBPE2F1 transgenic mice. In the transgenic mice, photoreceptor cells formed rosette-like arrangements at postnatal days 9 through 28. Complete loss of photoreceptors followed in the TgIRBPE2F1 mice but not in the Rb(-/-) retinal transplants. Both RB-deficient and E2F1-overexpressing photoreceptor cells expressed rhodopsin, a marker of terminal differentiation. Loss of p53 partially reduced the apoptosis and resulted in transient hyperplasia of multiple cell types in the TgIRBPE2F1 retinas at postnatal day 6. Our findings support the concept that cross-talk occurs between different retinal cell types and that multiple genetic pathways must become dysregulated for the full oncogenic transformation of neuronal retinal cells.

Apoptosis of the mammalian retina and lens.

Although retinal neurons usually last the entire lifetime of an individual, many innate genetic and developmental errors and external stimuli can reduce their longevity leading to loss of visual acuity or blindness. Similarly, the lens, largely composed of denucleated fiber cells must remain transparent for life if vision is to remain clear. Apoptosis of retinal neurons and newly generated lens fiber cells contributes to retinal degeneration and cataract formation, respectively, in both humans and experimental mammals. The apoptosis is triggered by many stimuli in addition to inherited mutations and may be amenable to pharmacologic amelioration. These studies not only provide new clinical insights but also the opportunity to investigate the molecular pathways leading to apoptosis in an organ that is not required for survival. The eye, becomes, therefore, an important organ for evaluation of theories of apoptosis in vivo.

Immunocytochemical reactivity of Xenopus laevis retinal rods and cones with several monoclonal antibodies to visual pigments.

Immunocytochemical reactions with several antibodies to visual pigments were used to study visual cells of the Xenopus laevis retina. Monoclonal antibodies to bovine opsin "E," 1D4, and 4B4 (reactive with the N- and C-terminus and with the loop connecting transmembrane segments 5-6, respectively) and to chicken visual pigments COS-1 and OS-2 (binding to mammalian red/green and blue cones, respectively), as well as a rabbit antifrog opsin serum 11-7, were applied to semithin and thin sections of the retina. The bound antibodies were detected with the peroxidase technique at the light microscopic level; a three-stage immunogold procedure was used for electron microscopic immunocytochemistry. The overwhelming majority of rods were labeled by monoclonal antibodies "E," 4B4, 1D4, OS-2, and serum 11-7. A small fraction (2-3%) of rods did not bind monoclonal antibodies "E" and 4B4, but this minor population of rods was strongly reactive with 1D4 and to a lesser extent with OS-2, indicating the presence of different visual pigment. These rods differ in shape from the major rod type; they are thinner, shorter, and may be comparable to the blue-sensitive ("green") rods of other amphibia. Cones were morphologically heterogeneous: double cones, large single cones, and small single cones were found, and the large single and the double cones were occasionally duplicated. Double cones and large single cones (as well as their duplicated varieties) strongly bound monoclonal antibodies COS-1 and were unlabeled by all other monoclonal antibodies, except OS-2. The small single cone was remarkably unreactive with COS-1 and "E," weakly labeled by 1D4 and 4B4, and most reactive with OS-2 and 11-7. This unique pattern of immunocytochemical reactions in the small cone type indicates the uniqueness of its visual pigment from other cone types in the Xenopus retina. The present study shows the existence of two different opsins in morphologically distinct (thick and thin) rod types and at least two cone pigments in the heterogeneous cone population.

Preembedding labeling with biotinylated antibodies and subsequent visualization of the biotin groups exposed on thin sections.

The feasibility of labeling cell membranes with biotinylated ligands and detecting the biotin groups on thin sections was investigated. Fixed retinal tissue was incubated with biotinyl- antiopsin . Half of the biotinyl-antibody labeled retinal tissue was incubated with avidin-ferritin (AvF) and embedded in Epon (preembedding reaction). The second half was embedded in glutaraldehyde cross-linked bovine serum albumin (BSA). Thin sections of this preparation were incubated with AvF to detect biotinyl-antibodies exposed by the sectioning (postembedding reaction). Biotin groups on the thin section surface could be readily visualized with AvF. Stereoscopic images demonstrated that the ferritin particles were localized only on the exposed surface of the thin section. The labeling was highly specific, with a very low background. Quantitative analysis was employed in order to determine the optimal reaction conditions for maximizing the labeling density with minimizing nonspecific binding. The possibility of using biotinylated molecules in the study of dynamic cellular events and for the subsequent intracellular localization of biotin on thin sections is suggested.

Immunocytochemical binding of anti-opsin N-terminal-specific antibodies to the extracellular surface of rod outer segment plasma membranes. Fixation induces antibody binding.

We have examined the binding of anti-opsin antibodies to the plasma membrane of frog retinal rod outer segments (ROS) by fluorescence light microscopy and electron microscopy. Polyclonal and monoclonal antibodies specific for the N-terminal domain of opsin were observed to bind to the extracellular surface of ROS plasma membrane of aldehyde-fixed but not of unfixed retinas. This reaction was found regardless of whether purified ROS, rhodopsin, opsin, or an N-terminal peptide of opsin was used as the immunogen. The fixation-induced binding of these antibodies contrasts with the more frequently noted loss of antigenicity upon fixation. Concanavalin A, however, binds to unfixed ROS plasma membranes. Its binding sites in the plasma membrane may be oligosaccharides in the N-terminal region of opsin. These results suggest that the N-terminal domain of opsin is latent in the native membrane and that changes in conformation may account for its detectability in fixed membranes.

Biosynthesis and vectorial transport of opsin on vesicles in retinal rod photoreceptors.

Retinal rod photoreceptor cells absorb light at one end and establish synaptic contacts on the other. Light sensitivity is conferred by a set of membrane and cytosol proteins that are gathered at one end of the cell to form a specialized organelle, the rod outer segment (ROS). The ROS is composed of rhodopsin-laden, flattened disk-shaped membranes enveloped by the cell's plasma membrane. Rhodopsin is synthesized on elements of the rough endoplasmic reticulum and Golgi apparatus near the nucleus in the inner segment. From this synthetic site, the membrane-bound apoprotein, opsin, is released from the Golgi in the membranes of small vesicles. These vesicles are transported through the cytoplasm of the inner segment until they reach its apical plasma membrane. At that site, opsin-laden vesicles appear to fuse near the base of the connecting cilium that joins the inner and outer segments. This fusion inserts opsin into the plasma membrane of the photoreceptor. Opsin becomes incorporated into the disk membrane by a process of membrane expansion and fusion to form the flattened disks of the outer segment. Within the disks, opsin is highly mobile, and rapidly rotates and traverses the disk surface. Despite its mobility in the outer segment, quantitative electron microscopic, immunocytochemical, and autoradiographic studies of opsin distribution demonstrate that little opsin is detectable in the inner segment plasma membrane, although its bilayer is in continuity with the plasma membrane of the outer segment. The photoreceptor successfully establishes the polarized distribution of its membrane proteins by restricting the redistribution of opsin after vectorially transporting it to one end of the cell on post-Golgi vesicles.

Rapid embedding of tissues in Lowicryl K4M for immunoelectron microscopy.

Lowicryl K4M (K4M) was recently introduced as an embedding medium for immunocytochemistry at the electron microscope level (BL Armbruster, E Carlemalm, R Chiovetti, RM Garavito, JA Hobot, E Kellenberger, W Villiger (1982):J Microsc 126:77 and E Carlemalm, M Garavito, W Villiger (1982):J Microsc 126:123). While earlier protocols of fixation and embedding required 4-6 days, the present method has reduced the processing time by accelerating both dehydration of tissues and polymerization of K4M so that tissues can be prepared for sectioning within 4 hr. The immunocytochemical labeling density was quantitated in order to determine relative antigen preservation in tissues embedded by the accelerated protocol as compared to slower K4M embedding techniques and to tissues embedded in glutaraldehyde-cross-linked bovine serum albumin (BSA). Thin sections of Bufo marinus kidney were labeled with rabbit antibody to Na+,K+ATPase alpha chain catalytic subunit isolated from B. marinus kidney microsomes (M Girardet, K Geering, JM Frantes, D Geser, BC Rossier, JP Kraehenbuhl, C Bron (1981):Biochemistry 20:6684). B. marinus retinas were labeled with rabbit anti-opsin. After fixation in paraformaldehyde(3%)-glutaraldehyde(3%), tissues were washed in buffer, dehydrated in 50, 75, and 90% dimethyl-formamide (DMF, 10 min each); K4M:DMF, 1:2 (15 min); K4M:DMF, 1:1, (20 min); K4M (25 min); K4M (30 min) at room temperature and transferred in fresh K4M to BEEM capsules for exposure to ultraviolet light (GE 15 watt, Black-lite, 10 cm, 45 min or less) at 4 degrees C. Thin sections were labeled successively with antibody, biotinylated sheep anti-rabbit F(ab')2 and avidin-ferritin. Ferritin labeling densities were determined by point counting. High labeling densities were observed with both antibodies, equaling or exceeding levels of labeling by slower protocols or embedment in BSA.

Immunocytochemical localization of opsin in the inner segment and ciliary plasma membrane of photoreceptors in retinas of rds mutant mice.

Homozygous 020/A mutant mice bearing the rds gene for slow inherited retinal degeneration have been observed to develop normal photoreceptor inner segments connecting cilia and synaptic contacts but fail to form outer segments. Their retinas are responsive to light, however. In order to assess the sources of these physiological responses we investigated the distribution of opsin in photoreceptors by means of immunoelectron microscopy. Opsin was detected in the inner segment plasma membrane and the distal ciliary plasma membrane. Antibody also bound to lamellar and vesicular membranes in the interphotoreceptor space and, in a small fraction of the photoreceptors, to membranes projecting from the distal cilium. These membranes may represent abortive formation of rod discs in this form of retinal degeneration. Failure to form an organized outer segment may contribute to the persistence of opsin in the inner segment plasma membranes of adult mutant mice.

Differential distribution of opsin in the plasma membrane of frog photoreceptors: an immunocytochemical study.

Opsin molecules on the surface of frog photoreceptors were visualized by immunocytochemistry at the ultrastructural level. Isolated retinas were immersed in biotinyl-antibody to bovine opsin followed by avidin-ferritin conjugates. Anti-opsin bound to the plasma membrane and to the surface of the most basal discs of red rod outer segments. Inner segment plasma membranes of red rod photoreceptors were devoid of anti-opsin label except for the apical plasma membrane in the region of the recently described periciliary ridge complex. The connecting cilium surface from its base at the periciliary region to the site of new disc evagination was almost free of anti-opsin binding, an observation in consonance with prior studies of thin sectioned retinas embedded in glutaraldehyde cross-linked bovine serum albumin. These results indicate that the continuous plasma membrane of photoreceptors is highly polarized. Opsin, which is free to diffuse throughout the outer segment plasma membrane and along the discs, does not back-diffuse onto the inner segment plasma membrane. The periciliary ridge complex and the base of the connecting cilium are possible sites of restriction of opsin mobility. This study also has provided new insight into the molecular structure of frog visual pigments. Frog green rod and cone outer and inner segment plasma membranes were not labeled by this sheep antiserum to bovine opsin. In contrast, discs of green ROS and the lamellae of some cones were labeled when these antibodies were applied to albumin embedded thin sections of frog retinas. Apparently, only internal or intramembraneous domains of green ROS and cone visual pigments were recognized by this antibody while both internal and extracellular domain(s) of red ROS opsin were reactive.

Opsin accumulation in photoreceptor inner segment plasma membranes of dystrophic RCS rats.

Opsin localization in photoreceptor plasma membrane was studied in 10- to 30-day-old dystrophic RCS rats. Preembedding cytochemical procedures with antiopsin antibodies and electron microscopy were employed. In the second postnatal week, opsin was sequestered to the outer segment plasma membrane in affected rat retinas. This distribution resembled that observed in the photoreceptors of normal rats at this age. As unphagocytosed debris accumulated in the subretinal space, the outer segments degenerated, and the distribution of opsin in the plasma membrane changed. Opsin reappeared in the inner segment plasma membrane as the outer segments were lost. Loss of the regionalized distribution of opsin was not associated with visible ultrastructural changes in the inner segments or connecting cilium. Severely damaged cells invariably were labeled on their inner segments at high density. Thus, the presence of an outer segment was correlated with the clearance of opsin from the inner segment, while damage to the outer segment was followed by reappearance of opsin in the inner segment plasma membrane.