• Rds;
  • Peripherin;
  • Glycosylation;
  • Retinal degeneration;
  • Outer segment;
  • Photoreceptor


  1. Top of page
  2. Abstract
  4. Reverse transcriptase PCR analysis of transgenic rds mRNAs
  7. Acknowledgements

Abstract : Rds/peripherin is an integral membrane glycoprotein that is present in the rims of photoreceptor outer segment disks. In mammals, it is thought to stabilize the disk rim through heterophilic interactions with the related nonglycosylated protein rom1. Glycosylation of rds/peripherin at asparagine 229 is widely conserved in vertebrates. In this study, we investigated the role of rds/peripherin N-glycosylation. We generated transgenic mice that expressed only S231A-substituted rds/peripherin in their retinas. This protein was not glycosylated but formed covalent dimers with itself and with glycosylated rds/peripherin. Nonglycosylated rds/peripherin also interacted noncovalently with rom1 homodimers to form a heterooligomeric complex. The glycosylated rds/peripherin ·· rom1 complex bound to concanavalin A-Sepharose, suggesting that the glycan is not directly involved in the interaction between these proteins. In double transgenic mice expressing normal and S231A-substituted rds/peripherin, the mRNA-to-protein ratios were similar for both transgenes, indicating no effect of N-glycosylation on rds/peripherin stability. Finally, expression of nonglycosylated rds/peripherin in transgenic mice rescued the phenotype of outer segment nondevelopment in retinal degeneration slow (rds-/-) null mutants. These observations indicate that N-glycosylation of rds/peripherin is not required for its normal processing, stability, or in vivo function.

Many secreted and transmembrane proteins are modified by N-linked glycosylation. This modification is important for intracellular trafficking, stability, secretion, and plasma membrane expression (Rademacher et al., 1988). N-Glycosylation may also be important for the dimerization of proteins, possibly through interactions with calnexin or other chaperones in the endoplasmic reticulum (Helenius, 1994). Most cell surface adhesion molecules are glycosylated, including the cadherins, integrins, selectins, and immunoglobulin-like cell adhesion molecules (Obrink, 1991 ; Filbin and Tennekoon, 1993 ; Tedder et al., 1995 ; Jones, 1996 ; Walsh and Doherty, 1997). Rds/peripherin is a glycoprotein of 36 kDa apparent molecular mass that contains four transmembrane domains (Travis et al., 1991b). It is located in the rims of photoreceptor outer segment disks (Connell et al., 1991). Outer segments are the functional equivalent of dendrites in other neurons and contain a stack of flattened membranous disks, which are the sites of photon capture and the reactions of visual transduction. Rom1 is a related protein, with an overall identity to rds/peripherin of 37% and a similar predicted membrane topology (Bascom et al., 1992). The distribution of rom1 in outer segment discs is identical to that of rds/peripherin (Bascom et al., 1992). In mammals, both rds/peripherin and rom1 have been shown to exist as covalent homodimers that interact noncovalently with one another to form a heterotetrameric complex (Travis et al., 1991b ; Bascom et al., 1992 ; Goldberg et al., 1995 ; Goldberg and Molday, 1996 ; Kedzierski et al., 1996). The interacting structural elements in rds/peripherin and rom1 are probably located within their large intradiscal (extracellular) D2 loops, between the third and fourth transmembrane domains (Goldberg et al., 1998). D2 loops contain the rds glycan and stretches of very high similarity between rds/peripherin and rom1 (Bascom et al., 1992). Rom1, however, is not glycosylated.

The gene for rds/peripherin was originally defined by the retinal degeneration slow (rds) mouse. This semidominant mutation has a biphasic retinal phenotype. In rds-/- homozygotes, rod and cone photoreceptors completely fail to develop outer segments (van Nie et al., 1978 ; Sanyal and Jansen, 1981 ; Cohen, 1983). Rhodopsin-containing vesicles have been observed in rds-/- retinas scattered throughout the space between photoreceptors and cells of the pigment epithelium (Nir and Papermaster, 1986 ; Jansen et al., 1987 ; Usukura and Bok, 1987), suggesting that outer segment-destined membrane is synthesized but cannot fold into disks without rds/peripherin. Subsequently, photoreceptor cell bodies degenerate in rds mutants (van Nie et al., 1978 ; Sanyal et al., 1980 ; Cohen, 1983). In rds+/- heterozygotes, outer segments are present but shortened and disorganized into whorl structures (Hawkins et al., 1985). Also, the rate of photoreceptor cell death is slower in rds+/- heterozygotes. The rds gene has been identified and shown to be a null allele in mice due to insertion of a 9.2-kb repetitive genomic element into protein-coding exon II (Travis et al., 1989 ; Ma et al., 1995). Mutations in the human RDS gene have been implicated in multiple inherited retinal degenerations, including autosomal dominant retinitis pigmentosa (Keen and Inglehearn, 1996 ; Shastry, 1997). Of the 54 reported disease-causing mutations in RDS, the majority result in single-residue substitutions within the D2 loop (Keen and Inglehearn, 1996 ; Shastry, 1997).

One hypothesis for the function of rds/peripherin and rom 1, consistent with their reported biochemical properties and the ultrastructural phenotype in rds mutants, is that these proteins are members of a new class of adhesion molecules that stabilize the rims of outer segment disks through heterophilic interactions across the intradiscal space (Travis et al., 1991b ; Kedzierski et al., 1996). A prediction of this hypothesis is that the ROM1 gene in humans should also be affected by retinal disease-causing mutations at a frequency similar to that for RDS. Also, loss of the rom1 protein in engineered nullmutant mice should cause a deficit symmetrical to that of rds, resulting in a phenotype of outer segment nondevelopment. It is curious that neither of these predictions has been borne out. No mutation in ROM1 alone has been convincingly associated with any human retinal diseases (Dryja et al., 1997), although a digenic form of retinitis pigmentosa has been reported in patients heterozygous for both a D2 point mutation in RDS (L185P) and a null mutation in ROM1 (Kajiwara et al., 1994). Also, mice have been generated that are homozygous for a null mutation in the rom1 gene (Clarke et al., 1998). Ultrastructural analysis of retinas from these mice showed only mild outer segment dysplasia, in contrast to complete nondevelopment of outer segments in rds-/- mutants. These observations imply that rom1, at least in part, is superfluous to the formation and maintenance of outer segments. Homologues of rds have been cloned from multiple species, including the human (Travis et al., 1991a), cow (Connell et al., 1991), dog (Moghrabi et al., 1995), rat (Begy and Bridges, 1990), cat (Gorin et al., 1993), Xenopus (Kedzierski et al., 1996), skate (Li et al., 1997), and chicken (Weng et al., 1998). A predicted motif conserved in all homologues of rds/peripherin is the site for N-linked glycosylation at asparagine 229. It has no predicted sites for O-glycosylation (Hansen et al., 1998). Rds/peripherin has been shown to be N-glycosylated in several species (Connell and Molday, 1990 ; Travis et al., 1991b ; Kedzierski et al., 1996). In this work, we investigated the possibility that the critical difference between rds/peripherin and rom1 function lies in N-glycosylation of rds/peripherin. According to this model, rom1 cannot function alone to stabilize the disk rim because it is not glycosylated.

When isolated Xenopus retinas in culture were treated with tunicamycin, which blocks N-glycosylation of all proteins, vesicular and tubular structures were observed at the bases of outer segments in the place of flattened disks (Fliesler et al., 1985). This pattern is similar to the vesicular debris observed in rds-/- mutant mice (Sanyal et al., 1980 ; Cohen, 1983) and indicates that the formation of outer segment disks in vertebrates is dependent on N-glycosylation of at least one protein species. Rds/peripherin has been suggested as a candidate protein for this effect (Travis et al., 1991b). To test the possibility that N-glycosylation of rds/peripherin is required for outer segment disk formation, we generated transgenic mice that only express a nonglycosylated form of rds/peripherin. We analyzed the function of this nonglycosylated protein biochemically and by its ability to complement in vivo the rds null phenotype.


  1. Top of page
  2. Abstract
  4. Reverse transcriptase PCR analysis of transgenic rds mRNAs
  7. Acknowledgements

Generation of transgenic mice

The transgenic construct contained 6.5 kb from upstream of the mouse rhodopsin gene as a transcriptional regulator, including 80 bp from the 5′-untranslated region, fused to a mouse rds cDNA containing partial 5′- and 3′-untranslated regions plus the complete protein-coding region, followed by the SV40 t-intron and early polyadenylation signal. This construct was identical to that used for transgenic rescue of rds-/- mutant mice (Travis et al., 1992), except that here, the rds cDNA was first subjected to site-directed mutagenesis using the Sculptor system (Amersham-Pharmacia Biotech, Piscataway, NJ, U.S.A.). The first base of codon 231 was altered from T to G, changing the encoded residue from serine to alanine. The construct was confirmed by DNA sequence analysis before oocyte injection. ICR strain transgenic founders were outcrossed with wild-type C57BL/6 mice or with rds homozygous mutants on a C57BL/6 background. To reduce possible photic damage, only pigmented (nonalbino) animals were studied (Noell et al., 1966). Mice were maintained on a 12-h light—dark cycle and killed between 4 and 6 h after light onset.

Genotype analysis

Mice were analyzed for the presence of the transgene by PCR with the primers 5′-CCTGGAGTTGCGCTGT and 5′-GTCTTTTTCATGAAGCACC, from the mouse rhodopsin promoter (Baehr et al., 1988) and rds coding region (Travis et al., 1989), respectively. Mice were analyzed for the presence of a wild-type rds gene by PCR using the primers 5′-CCTCCATGCCCTGCCTCT, from the first intron of rds (Ma et al., 1995), and 5′-AGCAGAGCGGCCCTGCA, from the rds coding region, and for the presence of the mutant rds gene with 5′- CGCATCAAGAGCAAC, from the rds coding region, and 5′ -CACATTACTCTTAAGGCC, from the rds inserted-element (Ma et al., 1995). PCR was done on DNA from tail cuts under standard conditions.

Nuclease protection analysis

Total RNA was extracted from individual eyecups using TRI Reagent (Molecular Research Center, Cincinnati, OH, U.S.A.), according to the manufacturer's protocol. RNA samples were hybridized to an [α-32P]UMP-labeled RNA probe of 280 nucleotides, which contained 209 nucleotides complementary to the transgenic mRNA and 169 nucleotides complementary to rds mRNA. After hybridization and S1 nuclease digestion, protected fragments were analyzed by electrophoresis on 8% polyacrylamide gels containing 8 M urea, as described (Kedzierski and Porter, 1990). The 209- and 169-nucleotide protected fragments were quantified on a Molecular Dynamics model 425F Phosphorimager and normalized for their content of radioactive nucleotide.

Immunoblot analysis

Polyclonal antisera were prepared by immunizing rabbits with a glutathione S-transferase fusion protein containing the carboxy-terminal domains of mouse rds/peripherin (residues 296-346) or rom1 (residues 296-351) expressed in Escherichia coli and purified by chromatography on glutathioneagarose (Sigma, St. Louis, MO, U.S.A.). For immunoblots, dissected retinas from 3-week-old mice were homogenized on ice in 10 mM Tris-HCl (pH 7.5), 1% sodium dodecyl sulfate (SDS), and protease inhibitors (Boehringer Mannheim, Indianapolis, IN, U.S.A.) and cleared by a 2-min centrifugation in an Eppendorf microcentrifuge at 4°C. SDS extracts were mixed with an equal volume of 2 × sample buffer (2% SDS, 10% glycerol, and 150 mM Tris, pH 6.8), with or without 5% β-mercaptoethanol, for analysis of protein monomers or dimers, respectively, and incubated at 65°C for 3 min. Samples were resolved by SDS-polyacrylamide gel electrophoresis (PAGE) in 10% gels (0.02 retina equivalents per lane) and electrotransferred to Immobilon P membranes (Millipore, Bedford, MA, U.S.A.). For immunolabeling of blots, primary antisera were diluted 1:1,000. Peroxidase-labeled goat anti-rabbit IgG (Kirkegaard and Perry Laboratories, Gaithersburg, MD, U.S.A.) was used at a 1:20,000 dilution as a secondary antibody, and LumiGLO-chemiluminescent substrate (Kirkegaard and Perry Laboratories) was applied for signal detection.

Enzymatic deglycosylation of rds/peripherin

SDS homogenates of retina containing 5% β-mercaptoethanol were heated at 50°C for 3 min and diluted fivefold with 60 mM Tris-HCl (pH 8.0), 1.2% Triton X-100, and 6 mM EDTA. After a 2-min clearing centrifugation, samples were incubated at 37°C for 6 h with cloned (protease-free) peptide N-glycosidase F (Boehringer Mannheim) at 2.5 U per retina equivalent of total protein. Samples were diluted with 5 × Laemmli buffer and subjected to SDS-PAGE for immunoblot analysis, as described above.

Concanavalin A (ConA)-Sepharose binding of rds/peripherin covalent dimers

Urea (final concentration, 8 M) was added to SDS homogenates of retina containing no reducing agent, to dissociate noncovalent rds ·· rom1 complexes without affecting the rds/peripherin and rom1 covalent dimers (Bascom et al., 1992). After a 5-min incubation at room temperature, homogenates were cleared by a 2-min centrifugation, diluted 1:20 with ConA buffer [20 mM Tris-HCl (pH 7.5), 350 mM NaCl, 1 mM CaCl2, 1 mM MnCl2, and 1% Triton X-100], and mixed with preswollen ConA-Sepharose (Amersham-Pharmacia Biotech) for 2 h at 4°C. The beads were washed five times with ConA buffer at 4°C. Bound proteins were eluted with Laemmli buffer and subjected to SDS-PAGE for immunoblot analysis.

ConA-Sepharose binding of rds ·· rom1 complexes

To test binding of rds ·· rom1 complexes to ConA, retinas from C57BL/6 mice were homogenized in 1% Triton X-100 and 0.05% SDS. The homogenate was cleared by a 2-min centrifugation and added to preswollen ConA-Sepharose. The beads were washed five times with the same buffer, and bound proteins were eluted with Laemmli buffer for immunoblot analysis.

Immunoprecipitation of rds/peripherin and rom1

Dissected retinas were homogenized in Triton buffer [50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 5 mM EDTA, 1% Triton X-100, 0.05% SDS, 2.5% glycerol, and protease inhibitors]. Homogenates were cleared by a 2-min centrifugation and incubated with the indicated antiserum at 4°C for 2 h. Samples then were mixed with preswollen protein A-Sepharose (Amersham-Pharmacia Biotech). After washing five times with cold Triton buffer, bound proteins were eluted with Laemmli buffer and subjected to SDS-PAGE for immunoblot analysis.

Reverse transcriptase PCR analysis of transgenic rds mRNAs

  1. Top of page
  2. Abstract
  4. Reverse transcriptase PCR analysis of transgenic rds mRNAs
  7. Acknowledgements

Total RNA was isolated from individual eyecups of 3-week-old male mice that were doubly transgenic with wild-type rds line 113 (Travis et al., 1992) and S231A-rds line 3019 on an rds-/- genetic background, as described for Nuclease protection analysis. First-strand cDNA was synthesized using a cDNA Cycle Kit (Invitrogen, Carlsbad, CA, U.S.A.) and the oligonucleotide primer 5′ -AGCAGAGCGGCCCTGCA complementary to nucleotides 764-748 in the mouse rds coding region (Travis et al., 1989). The cDNA was amplified by PCR in a reaction mixture containing 3 μCi of [α-32P]dCTP for 16 cycles with the sense and antisense rds primers 5′-CCCTGTATCCAGTACCAGCT and 5′-AGCCAGAGGTTGAGCTCC, respectively. In calibration experiments, exponential increases in amounts of reaction products were observed with increasing cycle number over 12-18 cycles of PCR (data not shown). After ethanol precipitation, the PCR products were digested with restriction endonuclease Fnu4HI. The resulting DNA fragments were separated by electrophoresis on a 15% polyacrylamide gel, quantified on a Molecular Dynamics model 425F PhosphorImager, and normalized for content of [32P]dCMP.

Preparation of samples for electron microscopy

Six-week-old mice were anesthetized with sodium pentobarbital (Nembutal ; 75 mg/kg) and were fixed for 5 min by vascular perfusion with 1% formaldehyde and 2% glutaraldehyde in 0.1 M sodium phosphate buffer (pH 7.2). The fixed tissue was then immersed in the same fixative overnight at 4°C and additionally fixed for 1 h in 1% osmium tetroxide. The tissues were dehydrated and embedded in Araldite 502 (CibaGeigy). Ultrathin sections were cut with a diamond knife and stained with uranium and lead salts for electron microscopy.


  1. Top of page
  2. Abstract
  4. Reverse transcriptase PCR analysis of transgenic rds mRNAs
  7. Acknowledgements

Transgene expression

Five transgenic mouse lines were generated with the S231A-rds/peripherin construct. Two lines (3003 and 3019) were selected for further study based on the level of transgene expression. Nuclease protection analysis was performed using a riboprobe of 280 nucleotides. The wild-type endogenous and transgenic rds mRNAs protected 169- and 209-nucleotide fragments of the probe, respectively (Fig. 1). This permitted us to quantify the level of transgenic relative to endogenous rds mRNAs (Kedzierski et al., 1997). The assay was repeated several times on different retinal samples from line 3003 (n = 13) and 3019 (n = 5) mice. After correcting for UMP content in the protected fragments, the mean ± SE expression level for the transgenic compared with the endogenous rds mRNAs was 0.8 ± 0.02 for line 3003 and 3.8 ± 0.2 for 3019.


Figure 1. Nuclease protection analysis of rds mRNAs in S231A-transgenic (TG) and non-TG mouse retinas. TG and endogenous rds mRNAs protected 209- and 169-nucleotide fragments, respectively, of a 280-nucleotide riboprobe. Representative lanes are shown from 3-week-old TG line 3003 and 3019 mice plus non-TG littermates, all on an rds+/+ genetic background. Each lane contains RNA prepared from a single retina.

Download figure to PowerPoint

S231A-substituted rds/peripherin is not N-glycosylated

Immunoblot analysis was done on retinal homogenates from animals of several genotypes using antiserum raised against a carboxy-terminal peptide of mouse rds/peripherin. The carboxy terminus is remote from the glycan-bearing D2 loop of rds/peripherin (Travis et al., 1991b) ; hence, its immunoreactivity should not be affected by its glycosylation state. In homogenates of wild-type retinas, the antibody detected a doublet band with a molecular mass of 36 kDa (Fig. 2A). Doublets on rds immunoblots have been seen before and may represent conformers due to incomplete reduction of intrachain disulfides before SDS-PAGE. Retinas from line 3003 and 3019 transgenic mice on an rds+/+ genetic background contained a second immunoreactive band of 33 kDa, corresponding to S231A-substituted rds/peripherin. Hence, the glycan contributes 3 kDa to the mass of normal rds/peripherin. When the transgenes were placed on an rds-/- background, where no wild-type rds/peripherin is made, only the S231A-substituted protein was detected. The level of S231A-substituted rds/peripherin in retinas from line 3003 on rds-/- was significantly less than in line 3003 retinas on rds+/+. Predigestion of retinal homogenates from wild-type mice with peptide N-glycosidase F, which cleaves asparaginelinked glycans (Chu, 1986), caused a shift in the electrophoretic mobility of rds/peripherin to approximately that of the S231A-substituted protein (Fig. 2B). The small disparity in electrophoretic mobility between enzymatically deglycosylated and S231A-substituted rds/peripherin may be due to slight differences in residual secondary structure. Electrophoretic mobility shifts of similar magnitude between wild-type and mutant forms of rds/peripherin were also seen with the substitutions P216L, L185P, and G167D, none of which affects N-glycosylation (W.K., unpublished data). No shift in electrophoretic mobility after digestion with peptide N-glycosidase F was observed for S231A-substituted rds/peripherin. These data indicate that S231A-substituted rds/peripherin is not glycosylated.


Figure 2. Immunoblot analysis shows nonglycosylation of S231A-substituted rds/peripherin. Retinal homogenates from mice of the indicated genotypes were separated by SDS-PAGE (10% gel) under reducing conditions and immunoblotted with antiserum against rds/peripherin, as described in Materials and Methods. A : Equal amounts of retinal homogenate from transgenic (TG) line 3003 and 3019 mice on rds+/+ (+/+) and rds-/- (-/-) genetic backgrounds. Also included are retinal homogenates from non-TG rds+/+ (+/+), rds+/- (+/-), and rds-/- (-/-) mice. B : Equal amounts of retinal homogenates from TG line 3019 mice on rds+/+ and rds-/- genetic backgrounds, plus non-TG rds+/+ and rds-/- mice. Where indicated, homogenates were digested with peptide N-glycosidase F (N-glyc F) before electrophoresis. The migration positions of protein size standards are indicated to the left in kDa. Note the altered electrophoretic mobility of rds/peripherin from non-TG, rds+/+ mice and the unchanged mobility of rds/peripherin from TG, rds-/- mice after N-glyc F digestion.

Download figure to PowerPoint

Covalent dimerization of glycosylated and nonglycosylated rds/peripherin

To test if the nonglycosylated form of rds/peripherin can form covalent homodimers in vivo, we prepared nonreduced SDS/urea retinal homogenates from mice of different genotypes. These homogenates were separated by SDS-PAGE under nonreducing conditions and subjected to immunoblot analysis. Wild-type rds/peripherin migrated as a dimer with a molecular mass of ~68 kDa, whereas the nonglycosylated protein migrated with a molecular mass of ~60 kDa (Fig. 3). In nonreduced retinal homogenates from S231A (line 3019) transgenics on an rds+/+ genetic background, we observed a broad immunoreactive band, suggesting the presence of heterodimers containing glycosylated and nonglycosylated rds/peripherin, besides both species of homodimer. The poorly resolved, stronger band of intermediate electrophoretic mobility was not present in mixed homogenates from S231A-transgenic on rds-/- and wild-type mice (Fig. 3), indicating that formation of heterodimers only occurs in vivo. No immunoreactive bands were visible on these gels at mobilities corresponding to the molecular weight of rds/peripherin monomers (data not shown).


Figure 3. Immunoblot analysis of rds/peripherin covalent dimers in S231A-transgenic (TG) and non-TG mouse retinas. Retinal homogenates were prepared from mice of the indicated genotypes and separated by SDS-PAGE (10% gel), all under nonreducing conditions. Included are non-TG on rds+/+ (non-TG, +/+), TG line 3019 on rds-/- (TG, -/-), a mixture of homogenates from non-TG on rds+/+ and TG line 3019 on rds-/- mice (non-TG, +/+ & TG, -/-), and TG line 3019 on rds+/+ (TG, +/+). The blot was immunostained with antiserum against rds/peripherin. The migration positions of protein size standards are indicated to the left in kDa. Note the appearance of a poorly resolved band of intermediate size in the homogenate from TG, rds+/+ mice that is not present in the non-TG, rds+/+ & TG, -/- mixed homogenates.

Download figure to PowerPoint

To confirm the existence of glycosylated/nonglycosylated heterodimers, a similar set of retinal homogenates was subjected to chromatography on ConA-Sepharose, which binds only glycosylated proteins. After washing, the bound proteins were eluted with Laemmli buffer and analyzed as monomers by immunoblotting under reducing conditions. A single immunoreactive band was detected with the rds/peripherin antibody in ConA eluates from wild-type homogenates (Fig. 4). In eluates from line 3019 S231A-transgenic, rds+/+ retinas, two immunoreactive bands were detected, corresponding to normal and nonglycosylated rds/peripherin monomers. Only a single band corresponding to the glycosylated monomer was observed when retinal homogenates from wild-type and S231A-transgenic, rds-/- mice were pooled and subjected to ConA-Sepharose chromatography (Fig. 4), confirming that dimerization does not occur in vitro. No rds-immunoreactive bands were detected after ConA chromatography on homogenates of retina that expressed only nonglycosylated rds/peripherin. Also, no rom1 immunoreactivity was detected in any samples after ConA chromatography, because rom1 is not glycosylated and SDS/urea disrupts the rds ·· rom1 complex (Bascom et al., 1992). These observations confirm that rds/peripherin and rom1 do not form covalent heterodimers. Collectively, these data show that nonglycosylated rds/peripherin undergoes covalent dimerization in vivo, both with itself and with glycosylated rds/peripherin monomers.


Figure 4. ConA-Sepharose analysis of rds/peripherin covalent dimers in S231A-transgenic (TG) and non-TG mouse retinas. Retinal homogenates were prepared under nonreducing conditions in the presence of SDS and urea, to eliminate noncovalent interactions but preserve covalent dimerization of rds/peripherin. After dilution in a Triton X-100 buffer, samples were incubated with ConA-Sepharose, washed, and eluted with Laemmli buffer. SDS-PAGE was performed under reducing conditions, followed by immunoblot analysis on equivalent samples of the starting homogenate (homogenate) and ConA-Sepharose eluate (ConA eluate) using antisera against rds/peripherin (rds) or rom1 (rom1), as indicated to the right. The genotypes studied included non-TG on rds+/+ (non-TG, +/+), line 3019 TG on rds-/- (TG, -/-), a mixture of non-TG on rds+/+ and line 3019 transgenic on rds-/- (non-TG, +/+ & TG, -/-), and line 3019 transgenic on rds+/+ (TG, +/+). Note the presence of nonglycosylated rds/peripherin in the ConA eluate from the TG, rds+/+ homogenate but not in the eluate of mixed non-TG, rds+/+ & TG, rds-/- homogenates.

Download figure to PowerPoint

Association of nonglycosylated rds/peripherin with rom1

Immunoprecipitation analysis was performed to determine if nonglycosylated rds/peripherin associates with rom1 homodimers. Triton X-100 homogenates of retinas from mice of several genotypes were reacted with antiserum against the carboxy terminus of rds/peripherin or rom1 and mixed with protein A-Sepharose. After washing, bound proteins were eluted with Laemmli buffer and subjected to immunoblot analysis. With wild-type retinas, the rds and rom1 antibodies both precipitated a complex containing rds/peripherin and rom1 (Fig. 5). A similar pattern was observed with retinas from line 3019 transgenic, rds-/- mice that expressed only S231A-substituted rds/peripherin, suggesting that nonglycosylated rds/peripherin interacts noncovalently with rom1. As expected, no rds/peripherin was detected in nontransgenic rds null mutants after precipitation with either antibody. A small amount of rom1 was detected in extracts of rds-/- mice after immunoprecipitation with the rom1 but not the rds antibody (Fig. 5).


Figure 5. Immunoprecipitation analysis of noncovalent interactions between rds/peripherin and rom1. Triton X-100 retinal homogenates were reacted with rds/peripherin preimmune serum (preimmune) or antiserum against rds/peripherin (rds) or against rom1 (rom1), as indicated below, and then mixed with protein A-Sepharose. After washing, bound proteins were eluted and resolved by SDS-PAGE (10% gels) under reducing conditions. Transfer blots of these gels were immunostained with antiserum against rds/peripherin (rds) or rom1 (rom1), as indicated on the right. Samples were prepared from 3-week-old mice of genotypes : nontransgenic (non-TG) on rds+/+ (non-TG, +/+), TG line 3019 on rds-/- (TG, -/-), and non-TG on rds-/- (non-TG, -/-). Note the reciprocal immunoprecipitation of nonglycosylated rds/peripherin by the rom1 antiserum and of rom1 by the rds antiserum from retinal homogenates of line 3019, rds-/- mice.

Download figure to PowerPoint

Wild-type rds ·· rom1 complex binds ConA-Sepharose

We performed ConA-Sepharose chromatography of a Triton X-100 homogenate obtained from wild-type mouse retina, to determine if the rds glycan is sterically available to bind ConA within the rds ·· rom1 complex. We observed efficient binding of the complex to ConA, evidenced by the similar immunoblot signals for rds/peripherin and rom1 in the homogenate and ConA eluate (Fig. 6). The slight difference in the electrophoretic mobilities of these proteins between the homogenate and eluate is probably due to the higher salt concentration in the eluate.


Figure 6. ConA-Sepharose binds rds ·· rom1 complexes. Triton X-100 homogenates of retinas from nontransgenic, rds+/+ mice were bound to ConA-Sepharose, washed, and eluted with Laemmli buffer. Samples of starting homogenate (homogenate) and ConA-Sepharose eluate (eluate) were analyzed by immunoblotting with antisera against rds/peripherin (rds) or rom1 (rom1), as indicated below. Twice as much ConA eluate was loaded compared with the starting homogenate. Note approximately equivalent binding of rds/peripherin and rom1 to the ConA column.

Download figure to PowerPoint

Stability of nonglycosylated rds/peripherin is comparable to that of the normal protein

To test if the stability of nonglycosylated rds/peripherin is similar to that of the normal protein, we generated mice that were doubly transgenic for wild-type rds line 113 (Travis et al., 1992) and S231A line 3019. First, we directly compared the levels of the 113 and 3019 transgenic mRNAs. Because these mRNA species could not be distinguished by nuclease protection analysis, we used a PCR-based assay. The 113 and 3019 transgenes are identical except for one T to G nucleotide substitution in codon 231. This substitution generated a new Fnu4HI recognition sequence (GCNGC). We performed PCR on first-strand cDNA prepared from retinal RNA samples, followed by Fnu4HI restriction digestion of the amplification product, and PAGE (Fig. 7). By quantitative autoradiography, after correcting for the content of labeled nucleotide in the PCR products, the 3019 and 113 transgenic mRNA levels were nearly identical. Next, we did immunoblot analysis on retinal samples from 113 plus 3019 double-transgenic mice. Immunoblot signals of comparable strength were observed for the glycosylated and nonglycosylated protein products of the two transgenes (Fig. 8). These observations suggest that glycosylated and nonglycosylated rds/peripherins have similar stabilities.


Figure 7. Quantification of wild-type and S231A transgenic (TG) mRNAs. Retinal RNA samples were prepared from 3-week-old mice of the genotypes : non-TG, wild-type TG line 113 (w/t TG), S231A TG line 3019 (S231A TG), and wild-type line 113 plus S231A line 3019 double-TG (w/t + S231A TG), all on an rds-/- background. After reverse transcription of the TG rds mRNAs, exponential-phase PCR was performed with rds primers in the presence of 32P-labeled dCTP. Where indicated below, the amplification products were digested with restriction endonuclease Fnu4HI, which cleaves S213A-mutant but not wild-type rds cDNA. The predicted undigested amplification product is 80 bp, and Fnu4HI restriction products of S231A cDNA are 48 and 32 bp. The migration positions of DNA size standards are shown to the left in bp.

Download figure to PowerPoint


Figure 8. g Immunoblot analysis of retinal homogenates from wildtype and S231A transgenic (TG) mice. Equivalent samples of retinal homogenates from 3-week-old mice were separated by SDS-PAGE (10% gel) under reducing conditions and immunoblotted with antiserum against rds/peripherin. Samples were prepared from non-TG, rds-/- (non- TG, -/-), wild-type TG line 113 on rds-/- (w/t TG, -/-), S231A TG line 3019 on rds-/- (S231A TG, -/-), wild-type line 113 plus S231A line 3019 double-TG on rds-/- (w/t + S231A TG, -/-), non-TG, -/-), non-TG on rds+/+ (non-TG, +/+), and S231A line 3019 TG on rds+/+ (S231A TG, +/+) mice. The migration positions of protein size standards are indicated to the left in kDa. Note approximately equal levels of glycosylated and nonglycosylated rds/peripherin in the double-TG retinal homogenate.

Download figure to PowerPoint

Outer segment morphology in mice expressing only nonglycosylated rds/peripherin

We studied the ultrastructure of photoreceptor outer segments in retinas from S231A-transgenic mice on an rds-/- null background. In line 3019 mice, the outer segment morphology was similar but not identical to that of wild-type outer segments (Fig. 9A). Most of the disks were normally aligned, and outer segment length and disk stacking density were comparable to those of wild-type outer segments. Moreover, the disk perimeters had the hairpin loop profiles that are typically observed in ultrathin sections. The normal morphology in line 3019 was, however, interrupted occasionally with groups of misaligned disks and occasional vesicles. Although these features are sometimes observed in wild-type outer segments, the frequency was higher in transgenic line 3019. In contrast, line 3003 had grossly abnormal outer segments (Fig. 9B). The disks were replaced by whorl structures, and total outer segment membrane area was reduced. A similar phenotypic pattern was observed in nontransgenic rds+/- heterozygotes, but here the outer segment dysplasia was milder (Fig. 9C).


Figure 9. Electron microscope analysis of mouse photoreceptor outer segments (os). All images in this panel are oriented with the retinal pigment epithelium (rpe) at the top. Photoreceptor os are immediately beneath the rpe and immediately above the inner segments (is), the sites of protein synthesis. A: Transgenic line 3019 on rds-/-. os in these mice are nearly normal, with wellordered, stacked membranous disks of equal diameter. B: Transgenic line 3003 on rds-/-. In these mice, the os contain whorls of oversized disks. Extreme examples of outer segment dysmorphogenesis are seen as collections of vesicles (arrows). C: Nontransgenic rds+/- heterozygotes. In these mice, the os consist of layered whorls of oversized disks. This is a milder phenotype than that seen in transgenic line 3003 (B). × 3,000.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  4. Reverse transcriptase PCR analysis of transgenic rds mRNAs
  7. Acknowledgements

Using transgenic mice as a model system, we examined the role of rds/peripherin N-glycosylation on its biochemical properties and its in vivo function in photoreceptor cells. The common posttranslational modification of N-linked glycosylation occurs within the tripeptide motif N-X-S/T (Marshall, 1974 ; Shakin-Eshleman et al., 1996). In the current study, we altered this motif by substituting alanine for serine, eliminating the highly conserved N-glycosylation site at asparagine 229 within the intradiscal D2 loop of rds/peripherin. Besides the site eliminated by the S231A substitution, mouse rds/peripherin contains a second predicted site for N-glycosylation within the D1 loop (Travis et al., 1989). However, this second site is not conserved in many homologues of rds/peripherin, including the closely related rat ortholog (Begy and Bridges, 1990). The data presented in Figs. 2B and 4 suggest that this second potential site for N-glycosylation in D1 is not used. We made three observations concerning the glycosylation state of S231A-substituted rds/peripherin : (a) Its electrophoretic mobility is reduced compared with that of normal rds/peripherin ; (b) unlike normal rds/peripherin, it does not display a shift in mobility after digestion with peptide N-glycosidase F ; and (c) it does not bind to ConA-Sepharose, again unlike normal rds/peripherin. Collectively, these data show that S231A-substituted rds/peripherin is not N-glycosylated in vivo.

All biochemical and functional properties of nonglycosylated rds/peripherin tested in this study were similar to those of the wild-type protein. First, nonglycosylated rds/peripherin promiscuously dimerized with itself and with normal rds/peripherin. Under extraction conditions that preserved noncovalent interactions, nonglycosylated rds/peripherin was equally capable of interacting with rom 1 homodimers as was the glycosylated protein. The observed efficient binding of the rds ·· rom1 complex to ConA-Sepharose suggests that the rds glycan is sterically unhindered and may be physically separate from the regions of interaction between the two proteins. These data indicate that the rds glycan does not participate in covalent homodimerization of rds/peripherin or noncovalent heterooligomerization of rds/peripherin and rom1. Finally, in line 3019 transgenic mice, we observed virtually complete rescue of the outer segment phenotype in rds null mutants.

An ultrastructural phenotype of outer segment dysplasia with formation of membranous whorls was seen in rds+/- heterozygotes (Fig. 9C), as previously reported (Hawkins et al., 1985). Because rds is a null allele in mice, the phenotype in rds+/- heterozygotes is due to haploinsufficienty (Ma et al., 1995). The ultrastructural phenotype in rds+/- heterozygotes may represent a reduced number of greatly enlarged disks arranged in whorls to compensate for the reduction in amount of available rds/peripherin (Kedzierski et al., 1997). As expected, nontransgenic, rds+/- heterozygous mice express the wild-type rds mRNA in retina at 50% the level of rds+/+ mice (W.K., unpublished data). The level of rds/peripherin in rds+/- heterozygotes, however, was <50% (Fig. 2A), reflecting the reduced membrane area and disorganization of disks in these mice. The level of nonglycosylated rds/peripherin in 3003 mice on rds-/- was significantly less than in nontransgenic rds+/- heterozygotes (Fig. 2A), consistent with the more profound reduction and disorganization of disk membranes (Fig. 9B). The genotypes (a) transgene line 3003 on rds-/-, (b) nontransgenic rds+/-, and (c) transgene line 3019 on rds-/- seemingly represent three points in a “titration series” of the rds null phenotype. Based on the biochemical results, we cannot offer an explanation for the mild dysmorphogenesis of ROS in line 3019. It is possible that the level of rds/peripherin protein expression is not commensurate with normal outer segment morphology. Due to semidominance of the gene, protein expression must be precisely controlled. Alternatively, it is possible that glycosylation of rds/peripherin is required for its interaction with some other, unidentified protein. If this is the case, the absence of this interaction produces a very subtle phenotype.

To test if the stability of nonglycosylated rds/peripherin was different from that of the normal protein, we compared the protein levels in a double-transgenic mouse that produced both glycosylated and nonglycosylated rds/peripherin as the products of two nearly identical transgenes (Fig. 8). First, we showed that line 113 and 3019 transgenes express their mRNAs at similar levels (Fig. 7). We observed no discordance between the protein levels in these lines, suggesting that the stability of nonglycosylated rds/peripherin is similar to that of normal rds/peripherin. Significant structural differences exist between the transgenic and endogenous rds mRNAs. The observed disparity between the endogenous and transgenic mRNA-to-protein ratios (Figs. 1 and 2A) may be due to reduced translation efficiency of the transgenic compared with the endogenous rds mRNAs. This interpretation also explains the unexpectedly severe phenotype in line 3003 transgenic mice on rds-/-.

In summary, we have generated transgenic mice that express only nonglycosylated rds/peripherin. It is surprising that despite absolute conservation of this posttranslational modification in a wide range of vertebrates, N-glycosylation of rds/peripherin is not required for its normal function. The critical difference between rds/peripherin and rom1, evidenced by the lack of disease-causing mutations in the ROM1 gene and the mild outer segment phenotype in rom1-/- knockout mice, cannot be attributed to nonglycosylation of rom1. Outside their D2 loops, rds/peripherin and rom1 show significant sequence divergence (Bascom et al., 1992). The structural elements that define the observed functional difference between rds/peripherin and rom1 thus may reside outside the D2 domain. The region containing these elements might be identified by generating transgenic mice that express rds/rom1 chimeric proteins. Another conclusion of the data presented here is that the observed tunicamycin effect on outer segment disc formation (Fliesler et al., 1985) must be due to inhibited N-glycosylation of another protein species besides rds/peripherin, possibly rhodopsin. The total number of known glycoproteins in outer segments that may affect structure is limited (Molday and Molday, 1987). Finally, line 3019 transgenic mice may be useful in further studies to define the biochemical lesion in RDS point mutations associated with human retinal degenerations. In double transgenic experiments with line 3019 plus other transgenes that expresses a mutant form of rds/peripherin, the nonglycosylated state may serve as a biochemical “tag” for the normally functioning protein. This will allow further dissection of rds/peripherin biochemistry in an in vivo setting.


  1. Top of page
  2. Abstract
  4. Reverse transcriptase PCR analysis of transgenic rds mRNAs
  7. Acknowledgements

The authors gratefully acknowledge Marcia Lloyd, Walid Moghrabi, Roxana Radu, and Alice Van Dyke for their outstanding technical assistance, Bobby Korn for his help with site-directed mutagenesis, Jane Johnson for her help with oocyte injection, and Sassan Azarian for his insightful comments on the manuscript. This work was supported by grants EY08043, EY00444, and EY00331 from the National Eye Institute and a grant from the Foundation Fighting Blindness. D.B. is the Dolly Green Professor of Ophthalmology at UCLA and a Research to Prevent Blindness Senior Scientific Investigator.

  • 1
    Baehr W., Falk J.D., Bugra K., Triantafyllos J.T., McGinnis J.F. (1988) Isolation and analysis of the mouse opsin gene.FEBS Lett. 238,253256.
  • 2
    Bascom R.A., Manara S., Collins L., Molday R.S., Kalnins V.I., McInnes R.R. (1992) Cloning of the cDNA for a novel photoreceptor membrane protein (rom-1) identifies a disk rim protein family implicated in human retinopathies.Neuron 8,11711184.
  • 3
    Begy C. & Bridges C.D. (1990) Nucleotide and predicted protein sequence of rat retinal degeneration slow (rds).Nucleic Acids Res. 18,3058.
  • 4
    Chu F.K. (1986) Requirements of cleavage of high mannose oligosaccharides in glycoproteins by peptide N-glycosidase F.J. Biol. Chem. 261,172177.
  • 5
    Clarke G.A., Rossant J.A., McInnes R.R. (1998) Rom1 is required for outer segment morphogenesis and photoreceptor viability. (Abstr.)Invest. Ophthalmol. Vis. Sci. 39 (Suppl.) , S962.
  • 6
    Cohen A.I. (1983) Some cytological and initial biochemical observations on photoreceptors in retinas of rds mice.Invest. Ophthalmol. Vis. Sci. 24,832843.
  • 7
    Connell G.J. & Molday R.S. (1990) Molecular cloning, primary structure, and orientation of the vertebrate photoreceptor cell protein peripherin in the rod outer segment disk membrane.Biochemistry 29,46914698.
  • 8
    Connell G., Bascom R., Molday L., Reid D., McInnes R.R., Molday R.S. (1991) Photoreceptor peripherin is the normal product of the gene responsible for retinal degeneration in the rds mouse.Proc. Natl. Acad. Sci. USA 88,723726.
  • 9
    Dryja T.P., Hahn L.B., Kajiwara K., Berson E.L. (1997) Dominant and digenic mutations in the peripherin/RDS and ROM1 genes in retinitis pigmentosa.Invest. Ophthalmol. Vis. Sci. 38,19721982.
  • 10
    Filbin M.T. & Tennekoon G.I. (1993) Homophilic adhesion of the myelin PO protein requires glycosylation of both molecules in the homophilic pair.J. Cell Biol. 122,451459.
  • 11
    Fliesler S.J., Rayborn M.E., Hollyfield J.G. (1985) Membrane morphogenesis in retinal rod outer segments : inhibition by tunicamycin.J. Cell Biol. 100,574587.
  • 12
    Goldberg A.F. & Molday R.S. (1996) Subunit composition of the peripherin/rds-rom-1 disk rim complex from rod photoreceptors : hydrodynamic evidence for a tetrameric quaternary structure.Biochemistry 35,61446149.
  • 13
    Goldberg A.F., Moritz O.L., Molday R.S. (1995) Heterologous expression of photoreceptor peripherin/rds and Rom-1 in COS-1 cells : assembly, interactions, and localization of multisubunit complexes.Biochemistry 34,1421314219.
  • 14
    Goldberg A.F.X., Loewen J.R., Molday R.S. (1998) Cysteine residues of photoreceptor peripherin/rds—role in subunit assembly and autosomal dominant retinitis pigmentosa. Biochemistry 37,680685.
  • 15
    Gorin M.B., Snyder S., To A., Narfstrom K., Curtis R. (1993) The cat RDS transcript : candidate gene analysis and phylogenetic sequence analysis.Mamm. Genome 4,544548.
  • 16
    Hansen J.E., Lund O., Nilsson J., Rapacki K., Brunak S. (1998) O-GLYCBASE version 3.0 : a revised database of O-glycosylated proteins.Nucleic Acids Res. 26,387389.
  • 17
    Hawkins R.K., Jansen H.G., Sanyal S. (1985) Development and degeneration of retina in rds mutant mice : photoreceptor abnormalities in the heterozygotes.Exp. Eye Res. 41,701720.
  • 18
    Helenius A. (1994) How N-linked oligosaccharides affect glycoprotein folding in the endoplasmic reticulum.Mol. Biol. Cell 5,253265.
  • 19
    Jansen H.G., Sanyal S., De Grip W.J., Schalken J.J. (1987) Development and degeneration of retina in rds mutant mice : ultraimmunohistochemical localization of opsin.Exp. Eye Res. 44,347361.
  • 20
    Jones E.Y. (1996) Three-dimensional structure of cell adhesion molecules.Curr. Opin. Cell Biol. 8,602608.
  • 21
    Kajiwara K., Berson E.L., Dryja T.P. (1994) Digenic retinitis pigmentosa due to mutations at the unlinked peripherin/RDS and ROM1 loci.Science 264,16041608.
  • 22
    Kedzierski W. & Porter J.C. (1990) Quantitative study of tyrosine hydroxylase mRNA in catecholaminergic neurons and adrenals during development and aging.Mol. Brain Res. 7,4551.
  • 23
    Kedzierski W., Moghrabi W.N., Allen A.C., Jablonski-Stiemke M.M., Azarian S.M., Bok D., Travis G.H. (1996) Three homologs of rds/peripherin in Xenopus laevis photoreceptors that exhibit covalent and non-covalent interactions.J. Cell Sci. 109,25512560.
  • 24
    Kedzierski W., Lloyd M., Birch D.G., Bok D., Travis G.H. (1997) Generation and analysis of transgenic mice expressing P216L-substituted rds/peripherin in rod photoreceptors.Invest. Ophthalmol. Vis. Sci. 38,498519.
  • 25
    Keen T.J. & Inglehearn C.F. (1996) Mutations and polymorphisms in the human peripherin-RDS gene and their involvement in inherited retinal degeneration.Hum. Mutat. 8,297303.
  • 26
    Li C., Al-Ubaidi M.R., Naash M.I. (1997) Isolation and characterization of the skate peripherin/rds gene. (Abstr.)Invest. Ophthalmol. Vis. Sci. 38 (Suppl.) , S219.
  • 27
    Ma J., Norton J.C., Allen A.C., Burns J.B., Hasel K.W., Burns J.L., Sutcliffe J.G., Travis G.H. (1995) Retinal degeneration slow (rds) in mouse results from simple insertion of a t haplotype-specific element into protein-coding exon II.Genomics 28,212219.
  • 28
    Marshall R.D. (1974) The nature and metabolism of the carbohydrate-peptide linkages of glycoproteins.Biochem. Soc. Symp. 1726.
  • 29
    Moghrabi W.N., Kedzierski W., Travis G.H. (1995) Canine homolog and exclusion of retinal degeneration slow (rds) as the gene for early retinal degeneration (erd) in the dog.Exp. Eye Res. 61,641643.
  • 30
    Molday R.S. & Molday L.L. (1987) Differences in the protein composition of bovine retinal rod outer segment disk and plasma membranes isolated by a ricin-gold-dextran density perturbation method.J. Cell Biol. 105,25892601.
  • 31
    Nir I. & Papermaster D.S. (1986) Immunocytochemical localization of opsin in the inner segment and ciliary plasma membrane of photoreceptors in retinas of rds mutant mice.Invest. Ophthalmol. Vis. Sci. 27,836840.
  • 32
    Noell W.K., Walker V.S., Kang B.S., Berman S. (1966) Retinal damage by light in rats.Invest. Ophthalmol. Vis. Sci. 5,450473.
  • 33
    Obrink B. (1991) C-CAM (cell-CAM 105)—a member of the growing immunoglobulin superfamily of cell adhesion proteins. Bioessays 13,227234.
  • 34
    Rademacher T.W., Parekh R.B., Dwek R.A. (1988) Glycobiology.Annu. Rev. Biochem. 57,785838.
  • 35
    Sanyal S. & Jansen H.G. (1981) Absence of receptor outer segments in the retina of rds mutant mice.Neurosci. Lett. 21,2326.
  • 36
    Sanyal S., De Ruiter A., Hawkins R.K. (1980) Development and degeneration of retina in rds mutant mice : light microscopy.J. Comp. Neurol. 194,193207.
  • 37
    Shakin-Eshleman S.H., Spitalnik S.L., Kasturi L. (1996) The amino acid at the X position of an Asn-X-Ser sequon is an important determinant of N-linked core-glycosylation efficiency.J. Biol. Chem. 271,63636366.
  • 38
    Shastry B.S. (1997) Signal transduction in the retina and inherited retinopathies.Cell. Mol. Life Sci. 53,419429.
  • 39
    Tedder T.F., Steeber D.A., Chen A., Engel P. (1995) The selectins : vascular adhesion molecules.FASEB J. 9,866873.
  • 40
    Travis G.H., Brennan M.B., Danielson P.E., Kozak C.A., Sutcliffe J.G. (1989) Identification of a photoreceptor-specific mRNA encoded by the gene responsible for retinal degeneration slow (rds).Nature 338,7073.
  • 41
    Travis G.H., Sutcliffe J.G., Bok D. (1991a) The retinal degeneration slow (rds) gene production is a photoreceptor disc membrane-associated glycoprotein.Neuron 6,6170.
  • 42
    Travis G.H., Christerson L., Danielson P.E., Klisak I., Sparkes R.S., Hahn L.B., Dryja T.P., Sutcliffe J.G. (1991b) The human retinal degeneration slow (RDS) gene : chromosome assignment and structure of the mRNA.Genomics 10,733739.
  • 43
    Travis G.H., Groshan K.R., Lloyd M., Bok D. (1992) Complete rescue of photoreceptor dysplasia and degeneration in transgenic retinal degeneration slow (rds) mice.Neuron 9,113119.
  • 44
    Usukura J. & Bok D. (1987) Changes in the localization and content of opsin during retinal development in the rds mutant mouse : immunocytochemistry and immunoassay.Exp. Eye Res. 45,501515.
  • 45
    Van Nie R., Ivanyi D., Demant P. (1978) A new H-2-linked mutation, rds, causing retinal degeneration in the mouse.Tissue Antigens 12,106108.
  • 46
    Walsh F.S. & Doherty P. (1997) Neural cell adhesion molecules of the immunoglobulin superfamily : role in axon growth and guidance.Annu. Rev. Cell Dev. Biol. 13,425456.
  • 47
    Weng J., Beleckyadams T., Adler R., Travis G.H. (1998) Identification of two rds/peripherin homologs in the chick retina.Invest. Ophthalmol. Vis. Sci. 39,440443.