Isolation of the Mouse Nyctalopin Gene Nyx and Expression Studies in Mouse and Rat Retina

Katrin Pesch,1 Christina Zeitz,2,3,4 Julia E. Fries,5 Stefanie Mu¨nscher,2,3 Carsten M. Pusch,1 Konrad Kohler,5 Wolfgang Berger,2,3 and Bernd Wissinger1

PURPOSE. It has been shown recently that mutations in NYX and impaired cone function and is due to mutations in the (nyctalopin on chromosome X), encoding a novel protein CACNA1F gene encoding the ␣-1-subunit of an L-type calcium associated with the leucine-rich repeat (LRR) protein superfam- channel.3,4 In the complete form (CSNB1), function of the rod ily, are responsible for the complete form of X-linked congen- system is completely absent, and cone-evoked signals are ital stationary night blindness (CSNB1). This study describes mostly normal in amplitude on standard ERG recordings. We the isolation and molecular characterization of the mouse or- and others have recently shown that CSNB1 is caused by thologue Nyx and its expression pattern in the retina. mutations in a novel gene of unknown function, designated 5,6 METHODS. Nyx was isolated by conventional DNA library NYX. screening and polymerase chain reaction (PCR)–based ap- The human NYX gene is composed of three exons and proaches. Gene expression in different mouse tissues was spans approximately 28 kb of genomic DNA in Xp11.4. Its studied by RT-PCR. Subsequently, the expression pattern of full-length transcript of 2713 bp encodes an open reading Nyx and its gene product in mouse and rat retinas was inves- frame (ORF) for a deduced protein of 481 amino acids (desig- tigated by RNA in situ hybridization and immunohistochemis- nated nyctalopin) that shares sequence similarities with mem- try with Nyx-specific antibodies. bers of a protein superfamily characterized by tandem arrays of leucine-rich repeat (LRR) motifs.7,8 Additional sequence fea- RESULTS. The Nyx gene encodes a protein of 476 amino acids tures suggest that nyctalopin is a glycosylphosphatidylinositol that contain 11 consecutive LRR motifs flanked by amino- and (GPI)-anchored extracellular protein with 11 typical and 2 carboxyl-terminal cysteine-rich LRRs. At the amino acid level, cysteine-rich LRRs. The LRR motif plays a role in protein– Nyx is highly homologous to its human orthologue (86% iden- protein interactions, and members of the LRR superfamily are tity). The gene is expressed in the eye but also, at lower levels, involved in various processes in development, signal transduc- in brain, lung, spleen, and testis. Nyx expression was found tion, DNA repair and recombination, and RNA processing, but during all stages of postnatal retinal development and was also cell adhesion and axon guidance.9,10 However, until now, confined to cells of the inner nuclear layer and the ganglion the precise function of NYX and particularly its pathophysio- cell layer in adult mouse and rat retinas. logical relationship with CSNB1 was unknown. CONCLUSIONS. These data suggest an important function of the Herein, we report on the identification and characterization Nyx protein in the inner retina and provide evidence that of the mouse orthologous gene Nyx and its expression pattern CSNB1 is based on a defect in the inner retinal circuitry. (Invest at the mRNA and protein levels during postnatal retinal devel- Ophthalmol Vis Sci. 2003;44:2260–2266) DOI:10.1167/ opment as well as in the adult mouse and rat tissue. iovs.02-0115

he Schubert-Bornschein type of congenital stationary night MATERIAL AND METHODS Tblindness (CSNB) is characterized by a so-called “negative ERG,” in which the amplitude of the a-wave is larger or equal Isolation of Genomic Nyx Sequences to that of the b-wave.1 Two X-chromosomal forms of CSNB can High-density filters of a genomic P1-derived artificial chromosome be distinguished, both clinically and genetically.1,2 The incom- (PAC) library from mouse strain 129 (Library 711; Resource Centre/ plete form (CSNB2) is characterized by residual rod function Primary Database [RZPD], Berlin, Germany) were hybridized with a 32P-dCTP–labeled human NYX probe (800 bp of exon 3). DNA from positive clones was isolated with a kit (Plasmid Midi Kit; Qiagen, Hilden, Germany); digested with EcoRI, HindIII, HincII, and PstI; From the 1Molekulargenetisches Labor, Universita¨ts-Augenklinik Tu¨bingen, Germany; 2Max-Planck-Institut fu¨r Molekulare Genetik, Ber- blotted on a nylon membrane; and rehybridized with the human NYX lin, Germany; 4Freie Universita¨t, Berlin, Germany; and 5Labor fu¨r Neu- probe. A single positive 7-kb EcoRI fragment was preparatively isolated rohistologie und Zellbiologie, Universita¨ts-Augenklinik Tu¨bingen, Ger- on an agarose gel, purified with a gel extraction kit (Quiaquick; Qia- many. gen), and cloned into a vector (pBluescript II SK; Stratagene, La Jolla, 3Present affiliation: Abteilung Medizinische Molekulargenetik und CA). The 7-kb EcoRI insert was then digested with PstI, and the Gendiagnostik, Institut fu¨r Medizinische Genetik, Universita¨t Zu¨rich, fragments were further subcloned in the same vector. Sequence anal- Switzerland. ysis of subclones was performed using standard M13 forward (5Ј- Supported by Grant Fo¨. 01KS9602 from the Federal Ministry of GTTTTCCCAGTCACGACG-3Ј) and reverse primers (5Ј-CAGGAAA- Education, Science, Research and Technology; the Interdisciplinary CAGCTATGACC-3Ј). Analysis of sequence data revealed that one of the Center of Clinical Research, Tu¨bingen; and Grant T-GE-0900-0030 from The Foundation Fighting Blindness. subclones contained part of exon 3 of the orthologous mouse gene Submitted for publication February 4, 2002; revised August 6 and (Nyx). This nucleotide sequence information was used to design December 30, 2002; accepted January 30, 2003. mouse-specific primers for RT-PCR and rapid amplification of cDNA Commercial relationships policy: N. ends (RACE)-PCR. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertise- Synthesis of cDNA and RACE-PCR ment” in accordance with 18 U.S.C. §1734 solely to indicate this fact. Corresponding author: Bernd Wissinger, Molekulargenetisches La- cDNA was synthesized from total mouse eye RNA by using a commer- bor, Universita¨ts-Augenklinik, Auf der Morgenstelle 15, D-72076 Tu¨- cial technology (SMART; Clontech, Palo Alto, CA). 3Ј-RACE experi- bingen, Germany; [email protected]. ments were performed with a forward primer derived from the mouse

Downloaded from jov.arvojournals.org on 09/28/2021 IOVS, May 2003, Vol. 44, No. 5 The Murine Orthologue of NYX 2261

genomic sequence (5Ј-GCTACAAGGCCACGTTTCTCTTC-3Ј). The 2-kb (pBluescript II SKϩ; Stratagene). The plasmid pRO4 containing the rat RACE product was cloned into a vector (pCR2.1-TOPO; Invitrogen, rhodopsin cDNA was kindly provided by Armin Huber, Institut fu¨r Groningen, Germany) and sequenced with M13 and a gene-specific Zoologie, Universita¨t Karlsruhe, Germany. primer (5Ј-GCACTCTGCGTCTTAACTCTG-3Ј). 5Ј-RACE experiments One microgram of plasmid DNA was linearized by restriction en- were only successful with the implementation of the polymerase zyme digestion and purified by phenol-chloroform extraction. Digoxi- provided by the manufacturer (Advantage-GC 2 PCR Polymerase; Clon- genin-UTP–labeled sense and antisense riboprobes were generated by tech) for PCR amplification and the application of a nested PCR in vitro transcription of linearized plasmids with a kit (DIG RNA strategy with reverse primers in the 3Ј-untranslated region (UTR) Labeling Kit; Roche Molecular Biochemicals, Mannheim, Germany). (5Ј-CATGAGTTACGTGCTGAGCCCGCC-3Ј) and the coding sequence Before the in situ hybridization, sections were treated with protein- of exon 3 (5Ј-GGCCATTCCGGTCCAGATCGATG-3Ј). ase K buffer (0.1 M Tris-HCl [pH 8] and 0.05 M EDTA) for 5 minutes at 37°C and digested with 0.3 ␮g/mL proteinase K (Sigma, Diesenhofen, RNA Isolation and RT-PCR Germany) for 8 minutes at 37°C. Slides were then washed two times Total RNA was isolated from various mouse tissues (RNeasy Midi and for 3 minutes each in diethyl pyrocarbonate (DEPC)-treated water, Mini Kit; Qiagen). Tissues were homogenized (PT3100 Polytron; Brink- postfixed for 15 minutes in paraformaldehyde (PFA; 4% PFA in 0.2 M man Instruments, Westbury, NY) in lysis buffer. The RNA preparation PB), washed again three times with DEPC-treated water, and air dried. was treated with DNase I and tested for the absence of DNA contam- Sixty microliters of hybridization solution (50% deionized form- ϫ ϫ ination by control PCR amplification with two mND gene intron amide [Sigma], 5 SSC, 5 Denhardt’s solution, 0.5 mg/mL tRNA ␮ primers in combination with two neo primers.11 Reverse transcription [Fluka, Buchs, Switzerland]) and 0.2 ng/ L of the digoxigenin-labeled of total RNA was performed by random hexanucleotide priming with riboprobe were denatured for 5 minutes at 80°C and applied to the reverse transcriptase (Omniscript; Qiagen). cDNA amplifications were sections. The slides were then incubated in a humidified chamber for performed with a forward primer in exon 2 (5Ј- GCCAAGGGATGCT- 16 hours at 64°C. Posthybridization washing steps were performed ϫ GARCCTG-3Ј) and a reverse primer in exon 3 (5Ј-CATTCCGGTCCA- two times for 30 minutes each in 0.1 SSC at 64°C. After a 10-minute GATCGATG-3Ј), using Taq DNA polymerase in a kit (HotStar with wash in Tris-buffered saline (TBS; 0.15 M NaCl and 0.1 M Tris-HCl [pH Q-solution; Qiagen), applying 35 cycles of 45 seconds at 94°C, 45 7.5]) at room temperature (RT), the slides were incubated for 30 seconds at 60°C, and 1 minute at 72°C. Primers for the amplification of minutes with blocking solution (10% blocking reagent in 0.1 M maleic Gapdh cDNA were 5Ј-GACCACAGTCCATGCCATCACT-3Ј (forward) acid, 0.15 M NaCl [pH 7.5]; Roche Molecular Biochemicals). Drained and 5Ј-TCCACCACCCTGTTGCTGTAG-3Ј (reverse). slides were incubated with alkaline-phosphatase–conjugated anti- digoxigenin antibody (1:500, in 10% blocking solution, 0.15% Triton Animals and Tissue Preparation X-100 in TBS; Roche Molecular Biochemicals) for 45 minutes at 37°C. for Histological Studies Sections were then briefly rinsed two times for 15 minutes each in TBS and preincubated for 10 minutes in substrate buffer (0.1 M Tris-HCl All experiments performed in this study were in accordance with the [pH 9.5], 1 mM MgCl2, 10% tetramisole-hydrochloride [Fluka]). Four ARVO Statement for the Use of Animals in Ophthalmic and Vision microliters nitroblue tetrazolium salt (NBT; 30 mg/mL; Bio-Rad Labo- Research. To determine cell specificity of Nyx mRNA expression in the ratories, Munich, Germany) and 4 ␮L 5-bromo-4-chloro-3-indolyl phos- rodent retina, in situ hybridization was performed on mouse and rat phate (BCIP; 15 mg/mL; Bio-Rad Laboratories) were mixed with 1 mL retina tissue. For the mouse experiments retinal tissue from adult of substrate buffer, and each section was incubated in 200 ␮L of this animals at postnatal day (P)76 and from the developmental stages P3, solution, in a humidified chamber in the dark for 24 to 72 hours at RT. P5, P10, P15, P20, and P30 were used. The rat retina was mature at Color reaction was stopped with stop buffer (0.1 M Tris-HCl [pH 7.5] P44, and the developmental stages were equal to the mouse stages. and 0.01 M EDTA), and covered with sorbitol (Merck, Darmstadt, The day of birth was designated as P0. Nyx protein was localized Germany). immunohistochemically in adult rat retina.

C57BL/6 mice and Brown Norway rats were killed by a short CO2 incubation and decapitation. The eyes were removed and dissected Immunohistochemistry along the ora serrata, and the posterior eyecups were fixed in 2% (in To raise antibodies against NYX, two peptides comprising the amino situ hybridization) or 4% (immunohistochemistry) paraformaldehyde acid sequences LTTSSPGPSPEPAATTV and ASLSDSLSSRGVG were pre- in phosphate buffer (PB; 0.1 M, pH 7.4) for 13 to 30 minutes at 4°C. pared by solid-phase peptide synthesis, using the Fmoc/But-strategy.12 After washing in PB, tissues were cryoprotected by immersion in 30% The peptides were purified to a homogeneity of more than 95% by (wt/vol) sucrose in PB overnight at 4°C. Eyecups were then embedded HPLC, and their identity was confirmed by electrospray mass chroma- in cryomatrix (TissueTek; Leica, Nussloch, Germany) and frozen, and tography. The peptides were coupled to keyhole limpet hemocyanin ␮ radial sections (10–12 m) were cut on a cryostat. For in situ hybrid- (KLH) by the glutaraldehyde method13 and used as antigens to raise ization, sections were mounted in tissue adhesive (Vectabond; Vector polyclonal antibodies in New Zealand White rabbits, according to Laboratories, Burlingame, CA) for immunohistochemistry on silane- standard immunization protocols (Charles River Service Laboratory, coated glass slides and dried at 60°C for 2 to 3 hours. Slides were stored Sulzfeld, Germany). The resultant antiserum was purified by affinity Ϫ at 80°C until further use. chromatography (Protein A Sepharose; Amersham Biosciences, Samples processed for either in situ hybridization or immunohisto- Freiburg, Germany) and subsequent by immunoaffinity chromatogra- chemistry were examined by microscope (AX70; Olympus, Tokyo, phy (applying the peptides used for immunization) and finally concen- Japan) with Normarski optics. Images were adjusted for brightness and trated by ultrafiltration on a 20-kDa cut-off membrane. contrast on computer (Photoshop; Adobe Systems, San Jose, CA). In sections used for immunohistochemistry endogenous peroxi-

dase was blocked with 3% H2O2 in 40% methanol. To reduce back- In Situ Hybridization ground staining, slides were preincubated for 1 hour in 10% normal As the probe template, a 520-bp PCR fragment of the human NYX goat serum (NGS, Sigma) and 0.3% Triton X-100-PBS (PBST). The (nucleotide positions 1887-2406, GenBank accession number primary antibody was diluted 1:1000 in PBST containing 10% NGS and AJ278865; http://www.ncbi.nlm.nih.gov/Genbank; provided in the incubated for 3 hours at RT or overnight at 4°C. After a wash with public domain by the National Center for Biotechnology Information, phosphate-buffered saline (PBS), the samples were incubated for 1 Bethesda, MD) was subcloned into an SmaI-linearized vector (pBlue- hour with a biotin-conjugated secondary antibody (dilution 1:200, script II SKϩ; Stratagene) by blunt-end ligation, and a 551-bp fragment Vectastain Elite Kit; Vector Laboratories) in PBST with 5% NGS. After a of the murine Nyx (corresponding to nucleotide positions 557-1107 in rinse in PB, retinal sections were processed with an avidin-biotin the human sequence) was subcloned into a PstI-linearized vector peroxidase complex, and immunoreaction was visualized with a dia-

Downloaded from jov.arvojournals.org on 09/28/2021 2262 Pesch et al. IOVS, May 2003, Vol. 44, No. 5

minobenzidine-nickel solution as the chromogen (1 mg/mL diamino- as for the NYX probe in Figures 3 and 4. Rhodopsin signals

benzidine, 0.2% glucose 0.004% NH4Cl, 0.09% (NH4)2Ni(SO4)2, and 1 were observed solely in the photoreceptor layer (Fig. 3B)— ␮L/mL glucose oxidase in PB). After terminating the reaction in PB the that is, in the cell bodies and the myoid regions of the photo- slides were coverslipped with glycerol/PBS (9:1). To determine the receptor cells. specificity of the antigen-antibody reaction, negative control experi- A clear and pronounced expression of Nyx was found in the ments were performed, either by omitting the primary antibody or inner nuclear layer (INL) and the ganglion cell layer (GCL) in preabsorbing it with the appropriate peptide. fully differentiated retinas (rat: P44, mouse: P76; Figs. 3A, 3C). Staining of the INL was mostly confined to the inner row of cells, which largely corresponds to the localization of amacrine RESULTS cells. In addition, some stained cells were occasionally local- Cloning of the Mouse Orthologue of NYX ized superior to this innermost row. However, the most intense Nyx staining was observed in the GCL (Figs. 3A, 3C). In As a prerequisite for gene expression studies, we isolated the rodents, the GCL consists of only slightly more than 50% of mouse orthologue of NYX. A genomic PAC library from mouse ganglion cells, whereas the remaining represent displaced am- strain 129 was screened with a human probe corresponding to acrine cells. Using a size criterion (soma diameter larger than Ј part of exon 3 (codons 242-418 and 3 -UTR). In this way, three 15 ␮m), at least a subset of the stained cells can be considered PAC clones were identified and characterized in more detail. ganglion cells.14 Those cells showed the most intense staining, Analysis of subclone sequences revealed a 1410-bp segment suggesting a particularly high level of Nyx expression. Our with 85% identity to exon 3 of the human NYX gene. Based on hybridization protocol was adjusted in such a way that a back- this partial genomic mouse sequence, primers were designed ground-free and unquestionable signal could be obtained with and used for the isolation of the full-length Nyx cDNA from the Nyx probe that can be clearly allocated to single cells. This reverse-transcribed total mouse eye RNA by RACE-PCR. Se- was nicely achieved in the inner retina where the signal filled quence analysis of amplification products and their compari- the entire cell body (Figs. 3A, 3B). Compared with this staining, son with the human nucleotide sequence revealed 85% identity the signal in the outer retina was extremely faint and not in the ORF. In addition, the splice site within the ORF (be- assigned to cell bodies in the same way as in the inner retina. tween amino acid residues 12 and 13 in the human sequence) A weakly stained band was found at approximately the level of is conserved in the mouse, as shown by the alignment of the outer limiting membrane, along the border of the outer genomic and cDNA sequence data. The cDNA sequence of Nyx nuclear layer (ONL) with the myoid region of the photorecep- has been deposited in the GenBank database (Accession-No. tors (Fig. 3, PhR), and where the photoreceptors terminate, at AY114303). The mouse gene encode 476 amino acid residues, the border to the outer plexiform layer (OPL; Figs. 3A, 3C, 4). whereas a 481-amino-acid protein is predicted from the human Ͻ 5,6 At early postnatal stages ( P5), the retina is not fully devel- sequence (Fig. 1). Computational protein sequence analysis oped, and only the GCL and the adjacent inner plexiform layer and motif predictions of NYX identified a characteristic domain (IPL) have already separated from the neuroblast layer (NBL). structure: The core sequence consists of 11 LRRs that are 6 Strong Nyx expression was observed in the GCL early, at P3. At flanked by two cysteine-rich LRRs. This core segment is pre- that stage, the GCL is still multilayered. Nyx-positive cell bodies ceded by a putative signal sequence and followed by a GPI in the GCL, at that time, were found primarily in the innermost anchor at the very C terminus. Amino acid sequence identity row of the GCL, adjacent to the axon fiber layer. In the between human and mouse is much higher in the LRR core undifferentiated NBL of P3 animals, distinctly labeled cells Ͼ ( 90%) than in the signal sequence and GPI anchor (62% and were also present near the NBL–IPL border. In addition, more 52%, respectively). However, virtually all previously identified diffuse staining was scattered over those parts of the NBL that mutations affect conserved amino acid residues (Fig. 1). later form the INL (Fig. 4). The same pattern was present at P5. Aligments of the Nyx cDNA with our PAC sequences and At P10, when the INL had separated from the photoreceptors, the draft sequence of the mouse genome indicate that the staining was clearly confined to the GCL and INL. With eye murine gene is split into 4 exons. The ORF is confined to the opening, which occurred around P15 in both species, the final two exons, analogous to the human gene. In silico analysis expression pattern of Nyx was essentially the same as in the with the draft mouse genome assembly placed Nyx close to the retina of adult animals, with the exception of a few labeled centromere of the murine X chromosome in the vicinity of cells in the more proximal part of the INL. Later developmental cask. The proximity of these two genes was also demonstrated stages reflected the murine Nyx expression pattern of the by positive hybridization of human CASK sequences to the PAC mature rat retina. clones harboring Nyx. Thus, Nyx localizes to a region on the To examine cell-specific protein localization we performed murine X chromosome that is syntenic to the p11.4 segment of immunohistochemistry on retinal sections of adult rats, with the human X chromosome. polyclonal antibodies against the NYX carboxyl terminus. Im- Gene expression in different mouse tissues was explored by munoreaction to the antibodies was found in the GCL, IPL, INL, RT-PCR with forward and reverse primers in exons 2 and 3, and OPL. The photoreceptor region—the ONL and the inner respectively (Fig. 2). Transcripts were detected in the eye, and outer segments—were devoid of immunoreactivity, ex- brain (cerebrum and cerebellum), lung, spleen, and testis, but cept for weak staining at the level of the ONL–myoid border, not in the kidney, heart, and liver. similar to the situation found with the mRNA probe (Fig. 3E). The immunoreaction in the GCL and in the inner row of the Expression of Nyx in Mouse and Rat Retinas INL resembled the pattern of mRNA expression. Cells of dif- To examine cell-specific mRNA expression in retinal neurons, ferent size and more than 50% of the somata were labeled in in situ hybridization was performed on retinal sections from rat the GCL (Fig. 3E, large arrows), and individual cells were and mouse, using human and mouse NYX antisense ribo- distinguished along the INL-IPL border (Fig. 3E, arrowheads). probes. Besides an overall punctuate labeling of the entire IPL two Hybridization with a rat rhodopsin cDNA antisense ribo- bands of enhanced immunoreaction were present in sublayers probe was performed to verify the quality and reliability of our 1 and 3, within which horizontally running processes were in situ hybridization protocol and the specificity of the staining observable (Fig. 3E, asterisks). Neurons along the outer margin in the different retinal layers (compare also Bech-Hansen et of the INL were also labeled (Fig. 3E, small double arrows) and, al.5). The reaction time of the rhodopsin probe was the same in addition, the OPL showed immunoreactivity with more

Downloaded from jov.arvojournals.org on 09/28/2021 IOVS, May 2003, Vol. 44, No. 5 The Murine Orthologue of NYX 2263

FIGURE 1. Amino acid sequence comparison of NYX and its mouse orthologue. The putative signal sequence at the N terminus and the C-terminal GPI anchor are shaded. Eleven typical LRRs (alternately de- picted by shaded boxes) are ﬂanked by N- and C-terminal cysteine-rich LRRs (LRRNT and LRRCT, under- lined). Vertical bars: amino acid identities; dots: similar residues. The overall sequence similarity is 86%. The position of intron 2 within the coding region is identical in both human and mouse, and indicated by an arrow above the murine amino acid sequence. Bold letters: previously identiﬁed mutations in the human sequence.5,6 Amino acid substitu- tions as well as deletions (triangles) and insertions (arrows) are listed below the human amino acid sequence.

pronounced staining of fibers running through the proximal conservation at the sequence level and a similar structural part of the OPL above the INL somata (Fig. 3E, small arrow). composition of the protein in both species. All hitherto iden- No labeling was observed when the primary antibody was tified mutations in patients with CSNB1 affect conserved amino either omitted (Fig. 3F) or preabsorbed with the peptide used acids. The deduced amino acid sequence of the mouse gene is to raise the antibody. five residues shorter at the N terminus. The putative start codon in the mouse gene coincides with a second in-frame DISCUSSION ATG codon in the human sequence. Thus, this second ATG Cloning of the mouse orthologue of NYX and sequence com- codon may represent the functional translation initiation in the parison to its human counterpart revealed a high degree of human gene, too.

Downloaded from jov.arvojournals.org on 09/28/2021 2264 Pesch et al. IOVS, May 2003, Vol. 44, No. 5

various developmental stages showed early expression of Nyx (at least as early as P3) and no gross differences in the temporal and spatial expression pattern during postnatal retinal development. The early expression in the developing retina may indicate that NYX at this stage plays a role in synaptogenesis and neuronal circuit formation. Continuous expression in the

FIGURE 2. Analysis of Nyx expression by RT-PCR in different mouse tissues. The highest level of expression was observed in the eye. RT-PCR of Gapdh was used to show that equal amounts of cDNA were used as template. Control lane: result of RT-PCR without the template.

In the adult mouse, Nyx is expressed in several tissues (brain, lung, spleen, and testis) with highest expression levels observed in the eye (Fig. 2). Similarly, expression of NYX in neural and several non-neural tissues was found in humans.5,6 Expression in kidney, consistently shown in humans, is absent or largely reduced in the mouse (Fig. 2). On RNA in situ hybridization, we observed Nyx expression in the cells of the GCL and the inner part of the INL in the mouse and rat retinas. This pattern is consistent with expression of Nyx in amacrine and ganglion cells in the rodent retina. Localization of the Nyx protein by immunohistochemical analysis showed predominant staining in the inner retina, from the GCL up to the terminals of the photoreceptors. However, the photoreceptor layer itself was free of anti-Nyx immunoreactivity, except for a very faintly stained band along the border of the ONL toward the inner segments. In addition to the localization of Nyx mRNA in the cells of the GCL and the inner INL, numerous cells in the outer INL along the INL–OPL border were immunoreactive. These Nyx protein-positive cells out- numbered the sporadically detectable mRNA-expressing cells in the outer region of the INL considerably, thus supporting the idea that Nyx may also be vertically transported to cells that are not able to produce it themselves. Immunostaining in the outer part of the INL raises the question of the identity of these cells. There is evidence from ERG recordings that the function of depolarizing bipolar cells is impaired in CSNB1.15–18 However, location, size, and shape of these immunoreactive neurons make it likely that they are horizontal rather than bipolar cells, an assumption further supported by single horizontally oriented fibers running in the OPL close to these cells. Even though the identity of these cells still has to be determined, it is obvious from the in situ hybridization data that most of them do not express Nyx and therefore depend on Nyx produced in the proximal retina. Our results clearly exclude a pronounced expression of Nyx in photoreceptors in the rodent retina. Even though there was a FIGURE 3. Nyx localization in the adult rodent retina. (A) Antisense very faint signal in both in situ hybridization and immunohis- riboprobe of murine Nyx labeled cells in the GCL (large arrows) and tochemistry approximately at the level of the external limiting the inner half of the INL (arrowheads) in the rat retina. A few stained cells were visible in the more distal INL (small arrow); (B) rhodopsin membrane, we never observed (in any developmental stage antisense probe labeling the photoreceptor ONL; (C) Nyx antisense examined) staining in the cell bodies of the ONL or the myoid probe labeling in the mouse retina (GCL: large arrows, inner row of regions of the photoreceptors comparable to the lucid NYX the INL: arrowheads); (D) Nyx sense control in the mouse retina; (E) expression in the INL and GCL. In the human retina, NYX immunohistochemical localization of NYX protein in the rat retina. expression of similar intensity has been reported in all nuclear Small double arrows: reactive cells in the distal INL close to the layers, including the ONL and the inner segments.5 This dis- INL–OPL border; small arrow: a stained fiber in the OPL. (✽) and (✽✽) crepancy may reflect species differences in NYX expression. A Enhanced staining in sublayers 1 and 3, respectively, in the IPL. Large similar difference was found in the kidney, where NYX expres- arrows and arrowhead correspond to descriptions in (A) and (C). sion was consistently shown in humans,5,6 but was absent, or Note the massive overall staining in the inner retina. In the photoreceptor layer (ONL and PhR) only a small band along the ONL–PhR at least largely reduced, in the mouse (Fig. 2). transition, approximately at the level of the outer limiting membrane, It has been argued that, analogous to the Drosophila LRR was faintly marked. A similar staining was found with the riboprobe in proteins chaoptin and capricious, NYX may be involved in the rat and mouse (A, C). (F) Negative control, with the NYX antibody formation of synaptic connections between neurons during preabsorbed. The section is free of any staining. (A, C, E) ✖ marks the development and maturation of the retina. Our in situ results at ONL-PhR transition.

Downloaded from jov.arvojournals.org on 09/28/2021 IOVS, May 2003, Vol. 44, No. 5 The Murine Orthologue of NYX 2265

FIGURE 4. In situ hybridization with digoxigenin-labeled Nyx riboprobe. Radial sections of mouse retinas at various developmental stages (P3– P30). Strong Nyx expression was present at P3 in the innermost row of the multilayered GCL adjacent to the axon fiber layer. Labeled cells were present in the undifferentiated NBL near the NBL–IPL in P3 retinas. The same pattern was observed at P5. At P10, when the INL was completely separated from the photoreceptors, staining was confined to the GCL and INL. In P15 the expression pattern of Nyx was essentially the same as in the retina of adult mouse retinas, with the exception of a few labeled cells in the more proximal part of the INL. Later developmental stages (P20, P30) reflect the Nyx expression pattern of the mature rat retina. Scale bar, 50 ␮m.

adult retina suggests a functional relevance throughout life— Because there is no evidence that amacrine or ganglion for example, in the maintenance of the extracellular matrix or cells contribute to the rod b-wave or the cone ON response, cell interaction processes. its loss in patients with CSNB1 cannot be readily explained CSNB1 was initially thought to be caused by a defect in if the function of NYX is restricted to the inner retina. signal transmission from rods to rod bipolar cells,1 but more However, it might be assumed that the absence of a func- recent studies have shown that there is a general impairment of tional defect of nyctalopin in amacrine and ganglion cells the retinal ON-pathway that involves both rod and cone sig- will impair the formation of regular synaptic contacts with naling15–19 and is apparently due to a functional defect 18 their input bipolar cells and thus indirectly have also an postsynaptic to the photoreceptors. adverse effect on the functional differentiation of the bipolar In the mammalian retinal circuitry, the main (sensitive) cells themselves. Further studies on the neuroretinal cir- signaling pathway of rods involves a single type of depolarizing cuitry in patients with CSNB1 are necessary to solve this bipolar cell (rod ON-bipolar), which in turn contact AII ama- question. crine cells through a sign-preserving glutamate synapse. Signals Of note, a naturally occurring mouse model nob (no b- from the AII amacrine cells then infiltrate the cone signaling wave) has been described that resembles CSNB1 in humans: pathways by exciting the cone ON bipolar cells through gap stationary course, preserved a-wave, and absent b-wave and junction electrical contacts and inhibiting OFF cone bipolar 30 cells through glycinergic synapses.20 The available electrophys- oscillatory potentials in ERG recordings. Moreover the nob iological data suggest that function of rod bipolar cells and also trait displays X-linked recessive inheritance, and recent linkage the cone ON pathway through the depolarizing bipolar cells analysis excludes the gene involved in the incomplete form of are compromised in CSNB1. Our histologic analyses in the CSNB (CSNB2) but maps the nob locus to a region syntenic to 31 rodent retina localize Nyx expression to cells of the inner half the human CSNB1 locus. of the INL and the GCL, which probably represents amacrine Our in silico mapping results localize Nyx to the centro- cells and ganglion cells, respectively. This expression pattern meric region of the murine X chromosome within the nob- contrasts with the principal electrophysiological findings in critical interval. This localization further supports the idea that CSNB1, which propose a main defect in the depolarizing bipo- a mutation in Nyx gives rise to the nob phenotype. Thus, lar cells.15–18 further studies of the nob mouse model may provide more However, there is also evidence for impairment of more insight into the complex pathophysiology of CSNB1 in humans distal neuroretinal function in patients with CSNB1. Miyake et and may help to elucidate the relationship between the prin- al.21 showed that the scotopic threshold response (STR) is not cipal electrophysiological findings and the restricted expres- recordable in these patients. Intravitreous aspartate injections sion pattern of NYX in the retina. indicate that the STR origin is located postsynaptically to the photoreceptors,22 and microelectrode recordings in the cat retina show that the STR is maximal around the IPL and the Note Added in Proof GCL.23,24 Furthermore, the oscillatory potentials are extremely small or absent in patients with CSNB1.25 Although the exact After final submission of this paper, Gregg and coworkers published site of their origin is still unknown, they are probably gener- an article in which they showed that nob mice do indeed have a ated in or near the IPL,26–28 possibly by depolarizing amacrine mutation in the Nyx gene (Gregg RG, Mukhopadhyay S, Candille SI, cells.29 In light of our expression data, we think that the et al. Identification of the gene and the mutation responsible for the analysis of such distal retinal function in CSNB1 patients de- mouse nob phenotype. Invest Ophthalmol Vis Sci. 2003;44:378– serves further attention. 384).

Downloaded from jov.arvojournals.org on 09/28/2021 2266 Pesch et al. IOVS, May 2003, Vol. 44, No. 5

Acknowledgments 15. Miyake Y, Yagasaki K, Horiguchi M, Kawase Y. On- and off-responses in photopic electroretinogram in complete and incom- The authors thank Eckart Apfelstedt-Sylla for providing expertise in plete types of congenital stationary night blindness. Jpn J Ophthal- clinical electrophysiology; Hubert Kalbacher for peptide synthesis and mol. 1987;31:81–87. antibody purification, Thomas Wheeler-Schilling for thoughtful advice 16. Sieving PA. Photopic ON-and OFF-pathway abnormalities in retinal and consultation on in situ hybridization; Gudrun Ha¨rer for valuable dystrophies. Trans Am Ophthalmol Soc. 1993;91:701–773. technical assistance in immunohistochemistry; and Ulrike Pesch for 17. Quigley M, Roy MS, Barsoum-Homsy M, Cheverette L, Jacob JL, helpful comments and critical reading of the manuscript. Milot J. On- and off-responses in the photopic electroretinogram in complete-type congenital stationary night blindness. Doc Ophthal- References mol. 1996;92:159–165. 18. Kim SH, Bush RA, Sieving PA. Increased phase lag of the funda- 1. Miyake Y, Yagasaki K, Horiguchi M, Kawase Y, Kanda T. Congen- mental harmonic component of the 30 Hz flicker ERG in Schubert- ital stationary night blindness with negative electroretinogram: a Bornschein complete type CSNB. Vision Res. 1997;37:2471–2475. new classification. Arch Ophthalmol. 1986;104:1013–1020. 19. Young RSL. Low-frequency component of the photopic ERG in 2. Boycott KM, Pearce WG, Musarella MA, et al. Evidence for genetic patients with the X-linked congenital stationary night blindness. heterogeneity in X-linked congenital stationary night blindness. Clin Vision Sci. 1991;6:309–315. Am J Hum Genet. 1998;62:865–875. 20. Sharpe LT, Stockman A. Rod pathways: the importance of seeing 3. Bech-Hansen NT, Naylor MJ, Maybaum TA, et al. Loss-of-function nothing. Trends Neurosci. 1999;22:497–504. mutations in a calcium-channel 1-subunit gene in Xp11.23 cause 21. Miyake Y, Horiguchi M, Terasaki H, Kondo M. Scotopic threshold incomplete X-linked congenital stationary night blindness. Nat response in complete and incomplete types of congenital station- Genet. 1998;9:264–267. ary night blindness. Invest Ophthalmol Vis Sci. 1994;35:3770– 4. Strom T, Nyakatura G, Apfelstedt-Sylla E, et al. An L-type calcium 3775. channel gene is mutated in incomplete X-linked congenital station- 22. Wakabayashi K, Gieser J, Sieving PA. Aspartate separation of the ary night blindness. Nat Genet. 1998;19:260–263. scotopic threshold response (STR) from the photoreceptor a-wave 5. Bech-Hansen NT, Naylor MJ, Maybaum TA, et al. Mutations in NYX, of the cat and monkey ERG. Invest Ophthalmol Vis Sci. 1988;29: encoding the leucine-rich proteoglycan nyctalopin, cause X-linked 1615–1622. complete congenital stationary night blindness. Nat Genet. 2000; 23. Sieving PA, Frishman LJ, Steinberg RH. Scotopic threshold re- 26:319–323. sponse of proximal retina in cat. J Neurophysiol. 1986;56:1049– 6. Pusch CM, Zeitz C, Brandau O, et al. The complete form of x-linked 1061. congenital stationary night blindness is caused by mutations in a 24. Frishman L, Sieving PA, Steinberg RH. Contributions to the elec- gene encoding a leucine-rich repeat protein. Nat Genet. 2000;26: troretinogram of currents originating in proximal retina. Vis Neu- 324–327. rosci. 1988;1:307–315. 7. Hockin AM, Shinomura T, McQuillan DJ. Leucin-rich repeat glyco- 25. Heckenlively JR, Deidre A, Martin BS, Arthur L, Rosenbaum MD. proteins of the extracellular matrix. Matrix Biol. 1998;17:1–19. Loss of electroretinographic oscillatory potentials, optic atrophy, 8. Kobe B, Deisenhofer J. The leucine-rich repeat: a versatile binding and dysplasia in congenital stationary night blindness. Am J Oph- motif. Trends Biol Sci. 1994;19:415–421. thalmol. 1983;96:526–534. 9. Krantz DE, Zipursky SL. Drosophila chaoptin, a member of the 26. Odgen TE. The oscillatory waves of the primate electroretinogram. leucin-rich repeat family, is a photoreceptor cell-specific adhesions Vision Res. 1973;13:1059–1074. molecule. EMBO J. 1990;6:1969–1977. 27. Yanagida T, Koshimizu M, Kawasaki M, Yonemura D. Microelec- 10. Nose A, Takeichi M, Goodman CS. Ectopic expression of connec- trode depth study of the electroretinographic oscillatory potentials tin reveals a repulsive function during growth cone guidance and in the frog retina. Doc Ophthalmol. 1987;67:355–361. synapse formation. Neuron 1994;13:525–539. 28. Wachtmeister L. Oscillatory potentials in the retina: what do they 11. Berger W, van de Pol D, Ba¨chner D, et al. An animal model for reveal. Prog Retin Eye Res. 1998;17:485–521. Norrie disease (ND) gene targeting of the mouse ND gene. Hum 29. Djammgoz MBA. Common features of light-evoked amacrine cell Mol Genet. 1996;5:51–59. responses in vertebrate retina. Neurosci Lett. 1986;71:187–191. 12. Chan WC, White PD. Fmoc Solid Phase Synthesis. Oxford, UK: 30. Pardue MT, McCall MA, LaVail MW, Gregg RG, Peachey NS. A Oxford University Press; 2000. naturally occurring mouse model of X-linked congenital stationary 13. Chana VB. Current Protocols in Protein Science. Vol. 2. New York: night blindness. Invest Ophthalmol Vis Sci. 1998;39:2443–2449. John Wiley & Sons; 1995. 31. Candille SI, Pardue MT, McCall MA, Peachey NS, Gregg RG. Local- 14. Guenther E, Schmid S, Grantyn R, Zrenner E. In vitro identification ization of the mouse nob (no b-wave) gene to the centromeric of retinal ganglion cells in culture without the need of dye label- region of the X chromosome. Invest Ophthalmol Vis Sci. 1999;40: ing. J Neurosci Methods. 1994;51:177–181. 2748–2751.

Downloaded from jov.arvojournals.org on 09/28/2021