Localization of HRG4, a Photoreceptor Protein Homologous to Unc-119, in Ribbon Synapse

Tomomi Higashide,1 Margaret J. McLaren,2 and George Inana1

PURPOSE. TO characterize further HRG4, a novel photoreceptor protein recently identified by subtractive cDNA cloning, by sequence analysis and immunolocalization. METHODS. The rat homolog of HRG4, RRG4 was expressed and used to prepare an antibody. The antibody was used in Western blot analysis, and immunofluorescent localization at the light and electron microscopic levels of HRG4-RRG4 protein. The HRG4-RRG4 sequence was also analyzed for homologies.

RESULTS. HRG4-RRG4 showed 57% homology with unc-119, a Caenorhabditis elegans neuropro- tein causing defects in locomotion, feeding, and chemosensation when mutated. By Western blot analysis, the HRG4-RRG4 protein was demonstrable only in retina and was soluble in nature. Immunofluorescence microscopic study of human and rat retinas, using the HRG4-RRG4 antibody, and other rod and cone photoreceptor-specific antibodies showed that the HRG4-RRG4 protein is localized in the outer plexiform layer of the retina in the synaptic termini of rod and cone photoreceptors. Electron microscopic immunolocalization showed the protein in the cytoplasm and on the presynaptic membranes of the photoreceptor synapses. CONCLUSIONS. The homology to unc-119 and localization to the photoreceptor synapse are suggestive of a function for HRG4-RRG4 in photoreceptor neurotransmission. HRG4 is the first photoreceptor-enriched synaptic protein to be reported, suggesting that its function may be unique to the specialized ribbon synapses formed between photoreceptors and the horizontal and bipolar cells of the retina. (Invest Ophthalmol Vis Set. 1998;39:690-698)

he recently identified HRG4 is a novel human photorecep- Recently, it has been determined that HRG4 is related to a tor cDNA encoding a 240 amino acid protein.1 The HRG4 gene discovered in Caenorhabditis elegans. The C. elegans cDNA was one of the clones isolated by a subtractive gene, unc-119 was isolated in a mutant nematode with defects T 3 cDNA cloning strategy to obtain novel human retina-specific cD- in locomotion, feeding behavior, and chemosensation. The NAs.2 The expression of HRG4, determined by Northern blot defect in locomotion was shown to be based in the nervous analysis, appears only in the retina among the 14 different human system, most likely involving sensory and motor neurons. Ex- and rat tissues tested and is correlated temporally with the devel- pression of unc-119 was seen in many neurons, consistent opment of photoreceptors in the rat. When studied by in situ with the abnormal phenotype of the worm. Because unc-119 hybridization, HRG4 mRNA is seen to be localized to the photo- appears to be involved in nematode neuronal function, includ- receptor inner segments of rods and cones. The HRG4 sequence, ing chemosensation, the homology of HRG4 with unc-119 is which is 92% conserved between human (HRG4) and rat (RRG4), suggestive that the novel photoreceptor protein might also be shows a two-domain structure: The first is an amino-terminal involved in neuronal signal induction or transmission in the region of approximately 60 amino acids, rich in proline and photoreceptors. In this report, the production of an antibody glycine, and moderately conserved (67%) between human and to HRG4-RRG4 and its localization to the synaptic termini of rat. The second is the latter three fourths of the sequence, which rod and cone photoreceptors are described. Of the many known synaptic proteins expressed in photoreceptors, so far is 100% conserved between the two species. When firstreported , only HRG4 has been found to be mainly in photoreceptors, HRG4 did not show significant homology with any of the known suggestive of a specialized function in visual neurotransmis- protein or gene sequences.' sion.

From the 'Bascom Palmer Eye Institute and the department of Physiology and Biophysics, University of Miami School of Medicine, MATERIALS AND METHODS Florida. Supported by the National Institutes of Health, Bethesda, Mary- Production of a Recombinant RRG4 Protein land, research grant EY10848 (GI); and by Research to Prevent Blind- 1 ness, New York, New York. The open-reading frame of the RRG4 cDNA was amplified by Submitted for publication July 1, 1997; revised October 7, 1997; polymerase chain reaction with a 5' primer, including an extra accepted December 18, 1997. ZtaraHI-restriction site immediately upstream of the translation Proprietary interest category: N. initiation site and a 3' primer in the 3' noncoding region down- Reprint requests: George Inana, Laboratory of Molecular Genetics, Bascom Palmer Eye Institute, Department of Ophthalmology, Univer- stream of the BamHl site on the antisense strand (5' primer, sity of Miami School of Medicine, 1638 NW 10th Avenue, Miami, FL 5 '-CGTAGGATCCATGAAGGTGAAGAAGG-3', the translation start 33136. site is underlined; 3' primer, 5'-TATACAGTTTGTCCAGGTC-3',

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, April 1998, Vol. 39, No. 5 Photoreceptor HRG4 Localizes to Ribbon Synapse 691

nucleotide positions 1210 to 1228 in the RRG4 cDNA1). The phenylmethylsulfonyl fluoride, and 1 /xg/ml aprotinin and was polymerase chain reaction product was digested with BamHl centrifuged at l,000g for 10 minutes at 4°C to remove cellular and ligated into a bacterial expression vector, pGEX4T-2 (Phar- debris. The supernatant was centrifuged at 105,00Qg for 1 hour at macia Biotech; Uppsala, Sweden). Transformants with the cor- 4°C to separate the cytosolic fraction from the membrane pellet. rectly oriented insert were selected according to the pattern of Ten micrograms of protein from the supernatant fraction was restriction fragments. Possible point mutations introduced in used for Western blot analysis. The membrane fraction was sus- the coding region by Taq polymerase were excluded by DNA pended in the same voliune of sample buffer (50 mM Tris-HCl sequencing. Six hundred milliliters of bacterial culture with an (pH 6.8), 100 mM dithiothreitol, 2% SDS, 0.1% bromphenol blue, 4 OD6oo of 1 was incubated at 30°C for 5 hours with 0.17 raM and 10% glycerol) for SDS-PAGE as the supernatant fraction, and isopropyl j3-i>thiogalactopyranoside (IPTG) to induce the ex- an amount equivalent to that used for the supernatant fraction pression of a glutathione-S-transferase-RRG4 fusion protein. was subjected to Western blot analysis. The bacterial pellet was suspended in 20 ml phosphate-buff- Triton X-114 phase separation was also performed to sep- ered saline (PBS; 150 mM NaCl, 16 mM Na2HPO4, and 4 raM arate the hydrophilic proteins from the amphiphilic membrane 6 NaH2PO4 [pH 7.3]) containing 2 mM ethylenediaminetetraace- proteins, as described by Bordier. Briefly, adult rat retina was tic acid, 0.1% /3-mercaptoethanol and 0.2 mM phenylmethyl- homogenized in a buffer containing 10 mM Tris-HCl (pH 7.4), sulfonyl fluoride and lysed on ice by mild sonication. After 150 mM NaCl, and 100 /xg/ml phenylmethylsulfonyl fluoride solubilization with 1% Triton X-100, the fusion protein was and centrifuged at l,000g for 10 minutes at 4°C to remove adsorbed to 2 ml glutathione-Sepharose beads (Pharmacia Bio- tissue debris. Proteins in the supernatant were solubilized with tech). The beads were washed with PBS and were incubated 1% Triton X-l 14 (Sigma) on ice and overlaid on a cushion of 6% with 8 NIH units of thrombin (Sigma, St. Louis, MO) at room (wt/vol) sucrose, 10 mM Tris-HCl (pH 7.4), 150 mM NaCl, and temperature for 30 minutes, to release the recombinant RRG4 0.06% Triton X-114, and incubated at 30°C for 5 minutes to protein from the fusion partner binding to the glutathione- induce phase separation. After centrifugation at 300g for 3 Sepharose (Pharmacia Biotech). The resultant protein contains minutes, the upper aqueous phase and the detergent phase two extra amino acids (Gly and Ser) at the aminoterminus of were supplemented with Triton X-114 and buffer, respectively, the RRG4 protein. to obtain equal concentrations of salt and detergent. Ten micrograms of protein from the aqueous phase and an equivalent Production and Affinity Purification of the amount of the detergent phase were used for Western blot Polyclonal Anti-RRG4 Antibody analysis. All procedures using animals were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic Western Blot Analysis and Vision Research. A New Zealand White rabbit was anes- Protein samples were suspended in the sample buffer and thetized and injected intradermally at multiple sites with 150 fractionated by SDS-PAGE (12.5% acrylamide).4 Equal amounts /xg of the recombinant RRG4 protein, emulsified in Freund's of each sample were loaded in two gels, except for the recom- complete adjuvant (Sigma). One hundred micrograms of the binant RRG4 protein (approximately 200 ng for Coomassie antigen, emulsified in Freund's incomplete adjuvant (Sigma), blue staining, 10 ng for Western blot analysis). One gel was was given as a booster injection every 2 weeks thereafter. The stained with Coomassie blue (Bio-Rad; Hercules, CA) to check activity of the antisera was tested on Western blots. Antisera the quantity of the protein in each sample, and the duplicate with high activity were purified using an affinity column with was electroblotted onto a membrane (Immobilon-P; Millipore, the recombinant RRG4 protein attached to CNBr-activated Bedford, MA). The blot was incubated at 4°C overnight in a Sepharose 4B (Pharmacia Biotech), according to the manufac- blocking buffer containing 5% (wt/vol) nonfat dried milk in 50 turer's protocol. The antibodies were adsorbed to the antigen mM Tris-HCl (pH 7.5), 200 mM NaCl, and 0.05% Tween 20. Sepharose gel and were eluted with 100 mM glycine (pH 3). After the blocking of nonspecific binding sites, the blot was The concentration of the antibody was estimated from the incubated, first with 1 /xg/ml affinity-purified anti-RRG4 anti- 4 OD280 (1 OD28O = 0.75 mg/ml) with a spectrophotometer body in the blocking buffer for 2 hours at room temperature, (DU Series 60; Beckman Instruments; Fullerton, CA). then with the complex of biotinylated goat antirabbit IgG (immunoglobulin G; Life Technologies, Gaithersburg, MD) and Protein Extracts from Tissues streptavidin-alkaline phosphatase conjugate (Life Technolo- Adult rat (Long Evans) tissues (retina, brain, lung, heart, mus- gies) in 50 mM Tris-HCl (pH 7.5), 200 mM NaCl, and 0.05% cle, liver, and kidney) were homogenized in a buffer contain- Tween 20 for 1 hour at room temperature. The antibody- ing 50 mM Tris-HCl (pH 8), 150 mM NaCl, 0.02% sodium azide, binding sites on the blot were visualized by incubation with 100 /xg/ml phenylmethylsulfonyl fluoride, 1 /xg/ml aprotinin, nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-in- and 1% Triton X-1004 and were centrifuged at 10,000g for 10 dolylphosphate /Holuidine salt. For the control experiment, minutes at 4°C. Protein concentration of the supernatant was preimmune serum (1:1000 dilution or 2 ju-g/ml immunoglobu- measured by the modified Lowry method5 using a kit (Micro lin, affinity-purified with a protein A-Sepharose CL-4B column Protein Determination; Sigma). Ten micrograms of protein per [Pharmacia Biotech]), and the flowthrough fraction from the lane was used for sodium dodecyl sulfate-polyacrylamide gel affinity purification of anti-RRG4 antibody (1:5000 dilution) electrophoresis (SDS-PAGE) and Western blot analysis. were used for the substitution for the primary antibody.

Protein Fractionation of the Rat Retina Immunofiuorescent Staining Adult rat retina was homogenized in a buffer containing 50 mM Adult rat retinas and human donor retinas were fixed in 4% Tris-HCl (pH 8), 150 mM NaCl, 0.02% sodium azide, 100 /xg/ml paraformaldehyde in PBS (10 mM phosphate, 130 mM NaCl [pH

Downloaded from iovs.arvojournals.org on 09/26/2021 692 Higashide et al. IOVS, April 1998, Vol. 39, No. 5

7.2]) at 4°C for 2 hours and were immersed in 20% sucrose in PBS epoxy resin (EMBED 812; Electron Microscopic Sciences, Ft. overnight. The tissue was frozen in O.C.T. compound (Miles Washington, PA) on the slide by placing an inverted embed- Diagnostics, Elkhart, IN) and sectioned at 6-jum thickness. ding capsule filled with resin over the section. After curing at For immunostaining, the sections were postfixed in 4% 60°C for 24 hours, the sections in the capsules were detached paraformaldehyde, washed in PBS, and blocked with 10% nor- from the slides with heat.'' Ultrathin sections were cut and mal goat serum in PBS and 0.1% Triton X-100 at room temper- stained with uranyl acetate and lead citrate and were examined ature for 30 minutes, followed by incubation with the affinity- with an electron microscope (JEM-100 CXII; JEOL, Japan). purified rabbit anti-RRG4 antibody (2 jug/ml) in PBS with 1% bovine serum albumin (BSA) at room temperature for 1.5 hours. After washing in PBS, die sections were treated with RESULTS fluorescein isothiocyanate-conjugated (FITC) goat antirabbit Figure 1 shows the alignment of unc-119 widi the sequences IgG polyclonal antibody (1:50 dilution; Life Technologies) in of the human (HRG4) and rat (RRG4) photoreceptor proteins. PBS with 1% BSA at room temperature for 30 minutes. Slides The overall homology of unc-119 with the photoreceptor were then washed in PBS and mounted in glycerol-PBS. Pre- proteins was 57%. As previously demonstrated, the homology immune serum (1:100 dilution) and the flowthrough fraction between HRG4 and RRG4 reflects the presence of a two- from the affinity purification of anti-RRG4 antibody (1:100 domain structure in this protein, in that the region of the first dilution) were also used to check the background staining and domain (proximal one fourth) is moderately homologous, the specificity of the anti-RRG4 antibody. whereas the region of the second domain (distal three fourths) For the double immunofluorescent staining of HRG4 and is 100% conserved between the two sequences. It is interesting S-antigen in the human retina, sections were first stained for to note that unc-119 showed the same pattern of homology HRG4 as described earlier, except rhodamine-conjugated goat with HRG4-RRG4: The homology was at most 29%, with huge antirabbit IgG polyclonal antibody (1:50 dilution; Boehringer gaps in the proximal one fourth, whereas it was 59% in the Mannheim, Indianapolis, IN) was used as a secondary antibody. distal three fourths. After washing, sections were again blocked with goat serum A prokaryotic expression clone of RRG4 was constructed and reacted with 1:100 antibovine S-antigen monoclonal anti- using the glutathione-5-transferase gene fusion system, and body, A9-C6 (a generous gift from L Donoso7) followed by recombinant RRG4 was produced and purified by glutathione FITC-conjugated antimouse IgG antibody (1:50 dilution; Life affinity chromatography and thrombin cleavage. The recombi- Technologies). nant RRG4 was used to produce a polyclonal rabbit antiserum, For the double immunofluorescent staining of HRG4 and which was subsequently affinity-purified. In Western blot anal- red-green cone opsin in the human retina, sections were first ysis of human and rat retinal proteins, the affinity-purified stained for HRG4, using rhodamine-conjugated goat antirabbit antibody demonstrated a specific reaction with a 35- and 33- IgG polyclonal antibody, followed by postnxation in cold meth- kDa species for HRG4 and RRG4, respectively, along with a anol to mask the epitopes of the anti-RRG4 antibody that are strong reaction with the 33-kDa recombinant RRG4 (Fig. 2). recognizable by the second polyclonal antirabbit IgG anti- The antibody to the rat protein showed a strong cross-reactiv- body.8 The sections were washed, blocked with goat serum, ity with the human antigen. The preimmune serum gave no and incubated with the polyclonal antihuman red-green cone reaction with the retinal proteins or the recombinant RRG4. opsin antibody (generous gift from J. Saari)910 and FITC-con- Significantly, the sizes of the immunoreactive native HRG4 and RRG4 were identical with those of the in vitro-expressed jugated antirabbit IgG antibody. All slides were examined using 1 a Zeiss photomicroscope III (Carl Zeiss, Oberkochen, Ger- products, confirming that the larger-than-calculated sizes are many) by phase contrast and fluorescence microscopy, using real characteristics of these proteins. The size of the recombi- filter sets selective for fluorescein and rhodamine. nant RRG4 expressed through the glutatiiione-5-transferase gene fusion system also matched that expressed by in vitro Electron Microscopic Immunocytochemistry transcription and translation. Adult rat retina was fixed as described earlier and frozen- To determine the subcellular distribution of RRG4 protein, sectioned at 16 /am thickness. The sections were postfixed in rat retinal extract was fractionated by high-speed centrifuga- 4% paraformaldehyde, washed in PBS, immersed in 0.3% hy- tion, and the membrane fraction and supernatant were exam- drogen peroxide for 10 minutes to inactivate the endogenous ined by Western blot analysis. The result demonstrated the peroxidase, and blocked with 10% normal goat serum in PBS immunoreactivity to be in the supernatant (Fig. 2, lanes 4, 5). and 0.1% Triton X-100 for 30 minutes at room temperature. Fractionation of the retinal proteins into membrane and non- The affinity-purified anti-RRG4 antibody (2 /xg/ml) or the pre- membrane portions by phase separation, using Triton X-114, immune serum (1:100 dilution) in PBS, containing 1% BSA and was also performed.6 Western blot analysis of the resultant 0.1% Triton X-100, were applied to the sections for 1.5 hours aqueous and detergent phases also indicated RRG4 to be in the at room temperature. After washing in PBS, sections were soluble, nonmembranous fraction (Fig. 2, lanes 6, 7). incubated with 0.5 ju-g/ml of biotinylated antirabbit IgG sec- The expression of HRG4-RRG4 was previously shown to ondary antibody (Life Technologies) in PBS with 1% BSA, fol- be only in the retina by Northern analysis of RNAs from 14 lowed by incubation with streptavidin- horseradish peroxidase different tissues, including the retina, iris, cornea, and retinal conjugate (Life Technologies) in PBS with 1% BSA. Color de- pigment epithelium from the eye.1 Tissue expression at the velopment of the antibody-binding sites was performed with protein level was examined by Western blot analysis, using the 0.5 mg/ml 3,3'-diaminobenzidine tetrahydrochloride in 50 mM antibody. Analysis of proteins from seven different rat tissues Tris-HCl (pH 7.6) with 0.03% hydrogen peroxide. The sections demonstrated the presence of 33-kDa RRG4 protein only in the were postfixed with 2% glutaraldehyde in PBS, osmificated, retina (Fig. 3), consistent with the results of the mRNA expres- and dehydrated with ethanol. Each section was embedded in sion studies.

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, April 1998, Vol. 39, No. 5 Photoreceptor HRG4 Localizes to Ribbon Synapse 693

unc-119 1 MK aeqqqqsiApgSAtfPS qmPrPP pvteqaitteaeLlaKnqlt II E II II III III HRG4 1 MKVKKGGGGaGtAtESAPGPSgqSVaPiPqPPAEsESGSESEPdaGPGPRpGPLQrKQPIG I I I ! I I I I I I I III III I I I I II I I I I I I I II I I I I I II I I I RRG4 1 MKVKKGGGGtGpgaEpvPGaSnrSVePtrePgAEaESGSESEPepGPGPRlGPLQgKQPIG consensus MKvkkgggg-g-a-esapgpS—sv-p-p-Ppae-esgsesep—gpgpr-gpLq-KqpIg

unc-119 46 PnDVLaLpglTqgfLCSPsaNvYnleFtkFqIRDlDtehVLFEIaKP Eneteenlqaq I I I I I II I I I I I I I I I I I I I I I I I I I I I HRG4 62 PEDVLGLQRITGDYLCSPEENIYKIDFVRFKIRDMDSGTVLFEIKKPPVSERLPINRRDLD I I I I || I I I I I I I I I I I II II I I I I I I I I I I I I I I I I II I I I I I I I I I I I I I I I I I I I I I I RRG4 62 PEDVLGLQRITGDYLCSPEENIYKIDFVRFKIRDMDSGTVLFEIKKPPVSERLPINRRDLD

consensus PeDVLgLqrlTgdyLCSPeeNiYkldFvrFklRDmDsgtVLFEIkKPpvsErlpinrrdld

unc-119 104 aesaRyVRYrFaPnFLkLktVGATVEFkVGDvPithFRMIERHfFkdrLLKcFDFeFGFCm I I I I I I II I I M I I I I I I I I I I I II I I I I I I I I I I I II HRG4 123 PNAGRFVRYQFTPAFLRLRQVGATVEFTVGDKPVNNFRMIERHYFRNQLLKSFDFHFGFCI I II I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I RRG4 123 PNAGRFVRYQFTPAFLRLRQVGATVEFTVGDKPVNNFRMIERHYFRNQLLKSFDFHFGFCI

consensus pnagRfVRYqFtPaFLrLrqVGATVEFtVGDkPvnnFRMIERHyFrnqLLKsFDFhFGFCi

unc-119 165 PnSrNnCEHIYeFPqLSqqLmddMInnPnETrSDSFYFVenkLVMHNKADYSYda I I I I I I I I I I I I I II I I I I I I I I I I I I I I I I I I I I I HRG4 184 PSSKNTCEHIYDFPPLSEELISEMIRHPYETQSDSFYFVDDRLVMHNKADYSYSGTP I I I I I I I I I I II I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I RRG4 184 PSSKNTCEHIYDFPPLSEELISEMIRHPYETQSDSFYFVDDRLVMHNKADYSYSGTP FIGURE 1. Homology between unc-119 and HRG4-RRG4. Genalign program of the IntelliGe- netics software package (Mountain View, CA) was used to align the protein sequences. The overall homology of unc-119 to the photoreceptor proteins is 57%.

Indirect immunofluorescence microscopic analysis with rod outer segments and the remainder of the rod photorecep- the anti-RRG4 antibody was used to determine the localization tor cells were reactive to a lesser extent, as reported previous- of HRG4-RRG4 protein in sections of rat and human retina. ly7 (Fig. 5B). When the fluorescent signals from the two probes The results in rat demonstrated a thin band of prominent were superimposed (by photographic double exposure), the immunoreactivity at the level of the outer plexiform layer image revealed dramatic yellow double staining of the HL-OPL, (OPL), with a fainter reaction also seen in the inner segments indicative of colocalization of HRG4 and S-antigen in the rod of the photoreceptors (Fig. 4A). In the para- and perifoveal axons, in the termini, or in both, with some pockets of red human retina, the OPL is substantially thickened by the pres- staining, representing sites of localization of HRG4 alone in the ence of bundles of horizontally oriented photoreceptor axons cone terminals (Fig. 5C). The same pattern was also seen in the termed Henle's fiber layer (HL). In this region, HRG4 immuno- outer portion of the photoreceptors, with yellow immunoflu- reactivity in the OPL was observed in a wide band extending orescence of the rod inner segments indicating the presence of directly from the inner boundary of the outer nuclear layer, HRG4 and S-antigen, and red immunofluourescence of the which is suggestive that this protein was localized in the cone inner and outer segments indicating the presence of photoreceptor axons as well as in their synaptic termini (Fig. HRG4 alone. The positive staining of cone outer segments for 4D). No immunostaining was observed in control sections HRG4 may be because of the shortness and the more open reacted with the preimmune serum (Figs. 4C, 4F). The OPL is access of the cone outer segments for diffusible HRG4 com- composed of synapses between the photoreceptors and the pared with that of rod outer segments. The presence of HRG4 horizontal and bipolar cells. HRG4-RRG4 has been shown by in in the rod and cone inner/outer segments could also be appre- situ hybridization to be expressed in rod and cone photore- ciated in sections stained with the HRG4 antibody alone (Fig. ceptors.1 To confirm that the observed HRG4-RRG4 immuno- 5A). This result thus indicates that HRG4 might be present in reactivity within the OPL was contributed by both types of termini immunoreactive for S-antigen (rod termini) and in photoreceptors, double immunofluorescent studies were per- those nonimmunoreactive for S-antigen (cone termini). formed, using this antibody in combination with antibodies As previously shown,8 human retinal sections immuno- recognizing proteins specific to rod and cone photoreceptors. stained with an antibody against red-green cone opsin10 dis- Double staining of human retina was first performed with the play intense staining of the cone outer segments, as well as HRG4-RRG4 antibody in combination with A9C6 antibody faint staining of the cone axons and termini (Fig. 5F); thus, the against the rod-specific protein S-antigen.7 With the RRG4 distribution of cone termini within the OPL can be visualized antibody and visualization with rhodamine, the HL-OPL was with this antibody. To take advantage of this, a comparison was labeled as before, in addition to the outer portions of photo- made of the immunoreactivity observed in the HL-OPL of receptors, probably because of postmortem diffusion of HRG4 human retinas double stained with antibodies to HRG4-RRG4 (Fig. 5A); with the S-antigen antibody visualized with FITC, the (Fig. 5E) and cone opsin (Fig. 5F). The result of this combina-

Downloaded from iovs.arvojournals.org on 09/26/2021 694 Higashide et al. IOVS, April 1998, Vol. 39, No. 5

12 3 4 5 6 7 12 The precise localization of HRG4 within the photoreceptor synaptic termini was investigated by electron microscopic A C immunocytochemistry. A preembedding method of electron microscopic immunocytochemistry1' was used, because RRG4 kDa antigenicity was lost after fixation with glutaraldehyde. Analy- 45- sis of ultrathin sections from immunoreacted rat retinal samples showed electron-dense particles in the presynaptic space of rod photoreceptor spherules in the outer plexiform layer, 31- whereas the horizontal and bipolar cell processes were devoid of reaction (Fig. 6A). Within the spherules, the particles were homogeneously distributed in the presynaptic space and were localized to the presynaptic membranes of the horizontal and bipolar cell processes. In the inner plexiform layer, the termini B kDa D were devoid of reaction (data not shown). Control sections 97- ^S reacted with preimmune serum were devoid of immunoreac- 66 - a— S3 tivity (Fig. 6B). Because of the paucity of cones in rat retina,12 m w m m these tissues were not suitable for electron microscopic immu- 45 rT SS — nocytochemical study of this protein in cone termini.

22- 1 2 3 4 5 fi 7 14- FIGURE 2. Western blot analysis of HRG4 protein. (A) Protein samples were fractionated in a 12.5% sodium dodecyl sulfate- kDa polyacrylamide gel and electroblotted onto a polyvinylidenedi- fluoride membrane. The blot was reacted with the affinity- 43- purifiecl anti-RRG4 antibody (1 /zg/ml), and color development was performed with nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolyIphosphate /j-toluidine salt. Lanes: 1, pro- 30- tein extract of adult human retina (10 /xg); 2, recombinant RRG4 protein (10 ng); 3, protein extract of adult rat retina (10 /Ltg); 4, supernatant of the homogenate of adult rat retina centrifuged at 105,000g for 1 hour at 4°C (10 /Ltg); 5, pellet of the homogenate of adult rat retina (membrane fraction) centrifuged at 105,000g for 1 hour at 4°C and resuspended in the same volume of the sample buffer as the supernatant; 6, aqueous phase of Triton X-l 14 phase separation of the homogenate kDa of adult rat retina (10 /xg); 7, detergent phase of Triton X-l 14 B phase separation of the homogenate of adult rat retina diluted back to the same volume as the aqueous phase. (B) Coomassie blue staining of the duplicate gel of (A). A larger amount of recombinant RRG4 (200 ng) was loaded in lane 2 for the 43- visualization of the protein. For other lanes, the same amount of protein was loaded as in (A). (C) Protein samples were electrophoresed and blotted as described. The blot was re- 30- acted with the preimmune serum (2 /xg/ml immunoglobulin affinity-purified with a protein A-Sepharose [Pharmacia Bio- tech] CL-4B column). Lanes: 1, protein extract of adult rat 20- retina (10 jug); and 2, recombinant RRG4 (10 ng). (D) Coomas- sie blue staining of the duplicate gel of (C). A larger amount of recombinant RRG4 (200 ng) was loaded in lane 2 for the visualization of the protein. FIGURE 3. RRG4 protein expression in various tissues of the adult rat. (A) Protein extracts (10 /xg) from adult rat tissues (lanes: 1, retina; 2, brain; 3, lung; 4, heart; 5, muscle; 6, liver; tion of antibodies, when viewed in a double exposure, dem- and 7, kidney) were separated in a 12.5% sodium dodecyl onstrated that termini positive for cone opsin were also immu- sulfate-polyacrylamide gel and electroblotted. The blot was noreactive for HRG4, evidenced by the yellow double reacted with the affinity-purified anti-RRG4 antibody (1 /xg/ml), immunofluorescence in the OPLs closely matching that seen and color development was performed witli nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolylphosphate p-tolu- with cone opsin (Figs. 5F, 5G). Thus, HRG4 was localized to idine salt. Arrowhead indicates the position of the band (ap- the rod and the cone photoreceptor synaptic termini. This proximately 33 kDa) seen only in the retina lane. (B) localization did not change during the 12 light: 12 dark diurnal Coomassie blue staining of the duplicate gel of (A) was per- cycle (data not shown). formed to check the quantity of the protein in each sample.

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, April 1998, Vol. 39, No. 5 Photoreceptor HRG4 Localizes to Ribbon Synapse 695

A C

FIGURE 4. Immunofluorescent staining of HRG4-RRG4 protein in rat and human retina. Adult rat (A, B, C) and human (D, E, F) retina was fixed in 4% paraformaldehyde and frozen-sectioned. The sections were incubated with 2 /xg/ml affinity-purified anti-RRG4 antibody (A, D) or the preimmune serum diluted at 1:100 (C, F) followed by reaction with FITC-conjugated secondary antibody. Phase contrast images of rat (B) and human (E) retina are shown for the identification of retinal layers. The paracentral region of the fundus is shown. The immunoreactivity is mainly localized to the OPL in both species, whereas inner segments of rods and cones (human) show weak staining (A, D), not seen in control specimens (C, F). In the paracentral human retina, the outer plexiform layer is tliickened by the presence of elongated photoreceptor axons forniing Henle's fiber layer. The human retina shows autofluorescence in the retinal pigment epithelium (D, F). GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; HL-OPL, Henle's fiber layer-outer plexiform layer; ONL, outer nuclear layer; IS, inner segment; OS, outer segment; RPE, retinal pigment epithelium. Scale, 50 /xm.

DISCUSSION calization in the photoreceptor synapse, however, has not been demonstrated.16 The results indicate that HRG4-RRG4 is a very abundant synaptic protein in photoreceptors. As such, it would be the first What might be the function of HRG4? Results of the of its kind because, although a number of proteins have been electron microscopic immunocytochemical study showed identified in the OPL or in the photoreceptor ribbon synapse, that HRG4 is localized in the cytoplasm and on the presyn- none qualifies as a protein that is predominantly, if not exclu- aptic membrane of the horizontal and bipolar processes sively, in the photoreceptor synapse. Bl6 is an 88-kDa antigen within the photoreceptor synapse (rod spherule). The im- present in the synaptic ribbon of photoreceptor terminals, but ages produced by this technique are very similar to those it is also present in the brain.1314 Dystrophin has also been observed in retinal synapses immunostained with antibodies localized to the OPL presynaptic region in which it is thought against the synaptic vesicle proteins synaptotagmin, synap- to control the calcium channel and neurotransmission, but it is sin, and SV2.17'18 Thus, HRG4 certainly could be associated also present in other tissues, including brain.15 RET52 is a with synaptic vesicles, which are present in the cytoplasm 52-kDa protein, detected in the inner segments of photorecep- and on the presynaptic membrane during docking and exo- tors and the OPL and is known to be related to the actin- cytosis. HRG4 may play a role in the mechanism of photo- bundling erythrocyte membrane component, dematin; its lo- receptor neurotransmitter release through the synaptic ves-

Downloaded from iovs.arvojournals.org on 09/26/2021 696 Higashide et al. IOVS, April 1998, Vol. 39, No. 5

(0 o HPE H

FIGURE 5- Double immunofluorescent staining of HRG4 with S-antigen and red-green cone opsin. (A, B) Frozen sections of adult human retina (paracentral region) were stained with anti-RRG4 antibody and rhodamine-conjugated secondary antibody, then with anti-S-antigen antibody and FITC-conjugated secondary antibody. Color photographs were taken with appropriate sets of filters for HRG4 (A) and S-antigen (B). (C) Double exposure showing superimposition of images seen in (A, B). (D) Phase contrast image of (A, B, C). (A) HRG4 immunoreactivity (red) is prominent in Henle's fiber layer-outer plexiform layer and is discernible in the inner segments of rods and cones, and the outer segments of cones, but not of rods. Fluorescence in the retinal pigment epithelial layer (A, B, C) is unrelated to the immunostaining and is caused by autofluorescence of lipofuschin in these cells. (B) Prominent S-antigen immunoreactivity is seen in the rod outer segments and in the Henle's fiber layer-outer plexiform layer. (C) In the double exposure, areas of overlap of HRG4 and S-antigen fluorescent signals in rod photoreceptors are indicated by yellow immunofluorescence, seen in a punctate pattern in the Henle's fiber layer-outer plexiform layer and in the rod inner segments. A double-labeled (yellow) rod inner segment is indicated (R). The inner segments of most cones (C), by contrast, label only with HRG4 antibody, and appear red. (E, F) Double immunofluorescent staining of HRG4 and red-green cone opsin. Adjacent sections from the human retina in (A, B, C, D) were stained with anti-RRG4 antibody and rhodamine-conjugated secondary antibody, followed by anti-red-green cone opsin antibody and FITC-conjugated secondary antibody. Photographs were taken with appropriate filtersfo r HRG4 (E), and red-green cone opsin (F), and with both filter sets (G). (H) Phase contrast of (E, F, G). (F) Red and green sensitive cones show immunoreactivity for red-green cone opsin predominantly in their outer segments. (B) indicates an unlabeled blue cone also marked in (E, G). Red-green cones also exhibit weak opsin staining throughout the cell, enabling visualization of their synaptic terminals in the Henle's fiber layer-outer plexiform layer (F). In the double exposure (G), areas of colocalization of the signal for HRG4 and red-green cone opsin appear as yellow-orange immunofluorescence, brightest in the Henle's fiber layer-outer plexiform layer, and weaker in the inner segments of red and green cones. Inner segments of blue cones display red fluorescence due to labeling of HRG4 alone. In the double exposures, superimposition of the two images is not perfect, because of slight misalignment of the two filter sets. INL, inner nuclear layer; HL-OPL, Henle's fiber layer-outer plexiform layer; ONL, outer nuclear layer, IS, rod and cone inner segments; OS, rod and cone outer segments; RPE, retinal pigment epithelium; CIS, cone inner segments; COS, cone outer segments; ROS, rod outer segments. Scale, 24 ^tm.

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, April 1998, Vol. 39, No. 5 Photoreceptor HRG4 Localizes to Ribbon Synapse 697 A

FIGURE 6. Electron microscopic immunocytochemical study of RRG4. Frozen sections of adult rat retina were reacted with 2 /xg/ml affinity-purified anti-RRG4 antibody (A) or with preimmune serum diluted at 1:100 (B), followed by sequential incubation with biotinylated secondary antibody, streptavidin-horseradish peroxidase conjugate and 3,3'-diaminobenzidine tetrahydrochloride. Ultrathin sections of the immunoreacted tissues were prepared from the paracentral region of the fundus and were counterstained with uranyl acetate and lead citrate. (A) The rod presynaptic terminal is homogeneously filled with electron-dense reaction product, some of which is present on the presynaptic membrane of the bipolar and horizontal cell processes, but the processes themselves are devoid of staining. (B) Control section devoid of immunoreactivity. B, bipolar cell process; H, horizontal cell process; R, synaptic ribbon; RS, rod spherule. Scale, 0.5 /nm.

icle cycle by interaction with other synaptic proteins in the brane phosphoprotein that binds to synaptic vesicles and actin, rod and cone synapse. The homology of HRG4 with unc-119 and is thereby postulated to control their interaction and re- is consistent with the possibility that HRG4 may play an lease, leading to exocytosis.22 This control is thought to be important role in the synaptic vesicle cycle. Although local- mediated by phosphorylation: A carboxyl-terminal, proline-rich ization within the synapse or the function of unc-119 in C. region with phosphorylation sites is important in binding to, elegans has not yet been demonstrated, the existence of a and possibly in controlling the exocytosis of, synaptic vesi- defect in this gene leading to significant problems in move- cles.23'24 Synapsin is notable in that it is absent in photorecep- ment, feeding, and chemosensation is consistent with a tor ribbon synapses,25'26 whereas HRG4 appears to be present defect in neurotransmission.3 abundantly in this location. Considering the importance of the What type of synaptic protein could HRG4 be? HRG4 is a postulated functions of synapsin, however, it seems likely that 240-amino acid, 35-kDa, acidic (pi 5.96), hydrophilic protein a synapsin homolog would exist in ribbon synapses. It is with three different types of putative phosphorylation sites possible that HRG4 could be such an entity. Similarly, an (five casein kinase II, two protein kinase C, and two tyrosine essential synaptic protein, syntaxin 1, is absent in the retinal kinase) and a proline- and glycine-rich amino-terminal region of ribbon synapse; however, a homolog, syntaxin 3, is present approximately 60 amino acids.L Proline-rich regions in proteins instead.27 are important for protein-protein interactions, many of which The phenotype of unc-119 C. elegans indicates that the 19 also involve phosphorylation. The proline-rich amino termi- effects of this gene are multisystem, involving neurons in nal region of HRG4, which has several casein kinase II phos- various locations in the worm.3 On the basis of levels of HRG4 phorylation sites, may also be important in such interactions. mRNA and protein expression detectable by Northern and The absence of a hydrophobic membrane-insertion sequence Western blot analyses of multiple tissues, it is interesting that tends to argue against HRG4 as an integral synaptic membrane the current data indicated that HRG4, the putative mammalian protein, such as synaptotagmin, synaptobrevin, synaptophysin, homolog of unc-119 was expressed predominantly in the pho- 20 or SNAP-25 It could, however, be a peripheral membrane- toreceptors. There are at least three possible explanations for associated protein, and the association with membrane could this: After divergence, the gene may have evolved to have a be through an A'-myristoyl group that may be attached to HRG4 more limited expression in mammals; HRG4 may be a distant at a consensus site, if the amino-terminal 5 to 8 residues, which homolog of unc-119, and there may be a closer homolog of are conserved between human and rat, are cleaved off to unc-119 with a wider range of neuronal expression; and HRG4 21 expose the glycine residue. The electron microscopic pic- may be expressed at very low levels in other neurons, unde- ture is consistent with association with synaptic vesicles and tectable by the methods used in this study. the presynaptic plasma membrane, as mentioned earlier. In summary, HRG4 appears to be a novel synaptic protein HRG4 shares several molecular features in common with a present in ribbon synapses of rod and cone photoreceptors. It number of known synaptic proteins involved in exocytosis and is likely to be involved in some aspect of the synaptic vesicle endocytosis.20 Among these, synapsin I is a peripheral mem- cycle. It will be of interest to elucidate its precise function in

Downloaded from iovs.arvojournals.org on 09/26/2021 698 Higashide et al. IOVS, April 1998, Vol. 39, No. 5

the mammalian visual system, through biochemical, physiolog- 13. Balkema GW. A synaptic antigen (Bl6) is localized in retinal syn- ical, and molecular genetic analyses, including the production aptic ribbons. / Comp Neurol. 1991;312:563-583. of transgenic and knockout mice. 14. Bachman KM, Balkema GW. Developmental expression of a synaptic ribbon antigen (Bl6) in mouse retina./ Comp Neurol. 1993; Acknowledgments 333:109-117. 15. Ueda H, Kobayashi T, Mitsui K, Tsurugi K, Tsukahara S, Ohno S. The authors thank Hirohiko Kato and Jose Perez for their technical Dystrophin localization at presynapse in rat retina revealed by advice and assistance. immunoelectron microscopy. Invest Ophthalmol Vis Sci. 1995;36: 2318-2322. References 16. Roof D, Hayes A, Hardenbergh G, Adamian M. A 52 kD cytoskeletal protein from retinal rod photoreceptors is related to erythrocyte 1. Higashide T, Murakami A, McLaren MJ, Inana G. Cloning of the dematin. Invest Ophtbalmol Vis Sci. 1991;32:582-593. cDNA for a novel photoreceptor protein. J Biol Cbem. 1996;271: 17. Koontz MA, Hendrickson AE. Comparison of immunolocalization 1797-1804. patterns for the synaptic vesicle proteins p65 and synapsin I in 2. Murakami A, Yajima T, Inana G. Isolation of human retinal genes: macaque monkey retina. Synapse. 1993; 14:268-282. recoverin cDNA and gene. Biochem Biophys Res Comm. 1992; 187:234-244. 18. Okada M, Erickson A, Hendrickson A. Light and electron microscopic analysis of synaptic development in macaca monkey retina 3. Maduro M, Pilgrim D. Identification and cloning of unc-119, a gene expressed in the Caenorhabditis elegans nervous system. Genet- as detected by immunocytochemical labeling for the synaptic ics. 1995; 141:977-988. vesicle protein, SV2.J Comp Neurol. 1994;339:535-558. 4. Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Labora- 19- Williamson MP. The structure and function of proline-rich regions tory Manual. Cold Spring Harbor, New York: Cold Spring Harbor in proteins. Biochem J. 1994;297:249-260. Laboratory; 1989. 20. Siidhof TC. The synaptic vesicle cycle: A cascade of protein- 5. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement protein interactions. Nature. 1995:375:645-653. with the Folin phenol reagent. / Biol Cbem. 1951;193:265-275. 21. Grand RJA. Acylation of viral and eukaryotic proteins. Biochem J. 6. Bordier C. Phase separation of integral membrane proteins in 1989;258:625-238. Triton X-114 solution. /Biol Chem. 1981;256:l604-l607. 22. Siidhof TC, Jahn R. Proteins of synaptic vesicles involved in exo- 7. Donoso LA, Merryman CF, Edelberg KE, Naids R, Kalsow C. S- cytosis and membrane recycling. Neuron. 1991;6:665-677. antigen in the developing retina and pineal gland: a monoclonal 23. Siidhof TC, Czernik AJ, Kao H-T, Takei K, Johnston PA, Horiuchi antibody study. Invest Ophtbalmol Vis Sci. 1985;26:56l-567. A, Kanazir SD, Wagner MA, Perin MS, De Camilli P, Greengard P. 8. Sakuma H, Inana G, Murakami A, Higashide T, McLaren MJ. Immu- Synapsins. Mosaics of shared and individual domains in a family of nolocalization of X-arrestin in human cone photoreceptors. FEBS synaptic vesicle phosphoproteins. Science. 1989; 24 5:1474 -1480. Lett. 1996;382:105-110. 24. Greengard P, Valtorta F, Czernik AJ, Benfenati F. Synaptic vesicle 9. Lerea CL, Bunt-Milam AH, Hurley JB. Alpha transducin is present in phosphoproteins and regulation of synaptic function. Science. blue-, green-, and red-sensitive cone photoreceptors in the human 1993;259:780-785. retina. Neuron. 1989;3:367-376. 25. Mandell JW, Townes-Anderson E, Czernik AJ, Cameron R, Green- 10. Kelley MW, Turner JK, Reh TA. Ugands of steroid/thyroid receptors induce cone photoreceptors in vertebrate retina. Develop- gard P, De Camilli P. Synapsins in the vertebrate retina: absence ment. 1995;121:3777-3785. from ribbon synapses and heterogeneous distribution among con- 11. Kitazawa R, Kitazawa S, Fukase M, et al. The expression of para- ventional synapses. Neuron. 1990;5:19-33. thyroid hormone-related protein (PTHrP) in normal parathyroid: 26. Mandell JW, Czernik AJ, De Camilli P, Greengard P, Townes- histochemistry and in situ hybridization. Histochemistry. 1992;98: Anderson E. Differential expression of synapsins I and II among rat 211-215. retinal synapses. / Neurosci. 1992; 12:1736-1749. 12. Carter-Dawson LD, LaVail MM. Rods and cones in the mouse 27. Morgans CW, Brandstatter JH, Kellerman J, Betz H, Wiissle H. A retina: structural analysis using light and electron microscopy. SNARE complex containing syntaxin 3 is present in ribbon syn- J Comp Neurol. 1979;188:245-262. apses of the retina./Neurosci. 1996;l6:6713-6721.

Downloaded from iovs.arvojournals.org on 09/26/2021