The Zebrafish pob Gene Encodes a Novel Protein ... - Semantic Scholar

2 downloads 0 Views 590KB Size Report
We also show that the degeneration of red cone photoreceptors in the mutants occurs independently of light. .... Localization studies show the protein to be intracellular and markers ... For endoplasmic reticulum (ER) staining, enhanced green.
Copyright © 2005 by the Genetics Society of America DOI: 10.1534/genetics.104.036434

The Zebrafish pob Gene Encodes a Novel Protein Required for Survival of Red Cone Photoreceptor Cells Michael R. Taylor,*,1 Satoshi Kikkawa,*,2 Antonio Diez-Juan,† Visvanathan Ramamurthy,* Koichi Kawakami,‡ Peter Carmeliet† and Susan E. Brockerhoff*,3 *Department of Biochemistry, University of Washington, Seattle, Washington 98195, †Center for Transgene Technology and Gene Therapy, Flanders Interuniversity Institute for Biotechnology (VIB), University of Leuven, B-3000 Leuven, Belgium and ‡ Division of Molecular and Developmental Biology, National Institute of Genetics, Mishima, Shizuoka 411-8540, Japan Manuscript received September 16, 2004 Accepted for publication January 24, 2005 ABSTRACT The zebrafish mutant, partial optokinetic response b (pob), was isolated using an N -ethyl N -nitrosourea (ENU)-based screening strategy designed to identify larvae with defective optokinetic responses in red but not white light. Previous studies showed that red-light blindness in pob is due to the specific loss of long-wavelength photoreceptor cells via an apoptotic mechanism. Here, we used positional cloning to identify the mutated pob gene. We find that pob encodes a highly conserved 30-kDa protein of unknown function. To demonstrate that the correct gene was isolated, we used the Tol2 transposon system to generate transgenic animals and rescue the mutant phenotype. The Pob protein contains putative transmembrane regions and protein-sorting signals. It is localized to the inner segment and synapse in photoreceptor cells, and when expressed in COS-7 cells it localizes to intracellular compartments. We also show that the degeneration of red cone photoreceptors in the mutants occurs independently of light. On the basis of our findings, we propose that Pob is not involved in phototransduction but rather plays an essential role in protein sorting and/or trafficking.

R

OD and cone photoreceptors are highly polarized cells designed to transmit visual information with amazing fidelity. Proper function of these compartmentalized cells requires a large number of molecules that are specific to photoreceptors or that have specialized functions within photoreceptors. The most apical part, the outer segment, is composed of stacks of membranous disks containing the phototransduction machinery, 80% of which is the opsin protein. The photoreceptor synapse is a ribbon type with structural and functional characteristics designed for rapid, continual transmission of visual information over several orders of magnitude of light illumination (Dowling 1987). While much is known about the molecules involved in photoreceptor function, relatively little is known about the mechanisms and regulation of the proper sorting and targeting of these molecules. The nonessential aspect of vision has made it possible to identify mutations in a large number of genes that

Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under accession no. AY745978. 1 Present address: Department of Neurological Surgery, Box 0520, University of California, San Francisco, CA 94143. 2 Present address: Division of Developmental Neurobiology, Department of Neuroscience, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650-0017, Japan. 3 Corresponding author: Department of Biochemistry, Box 357350, University of Washington, Seattle, WA 98195. E-mail: [email protected] Genetics 170: 263–273 (May 2005)

are critical for the proper function and/or viability of photoreceptor cells. Over 80 genes that are essential for normal visual function have been identified in humans (Pacione et al. 2003; see http://www.sph.uth.tmc.edu/ Retnet/). In addition, genetic screens in Drosophila have uncovered many genes necessary for invertebrate visual function and development (Pak 1995). These genetic strategies have identified proteins (1) unique to photoreceptors, such as phototransduction components; (2) unique to photoreceptors, but not easily identified using other experimental strategies; and (3) essential to photoreceptor function with widespread patterns of expression. The latter two categories include premRNA splicing factors, molecules involved in fatty acid biosynthesis, tissue inhibitor metalloproteinases, inosine monophosphate dehydrogenase, and other important cellular components (Pacione et al. 2003). Defects that cause subtle visual phenotypes are not likely to be identified in humans since identification of mutations requires clinical observation by an ophthalmologist and extensive genetic and molecular analysis. Furthermore, recessive mutations that are rare may also be missed because large inbred pedigrees within most human populations are uncommon. A more suitable genetic model would benefit from a relatively short generation time, rapid development of the visual system, large brood size, and ease in identifying visual function defects. While Drosophila sufficiently meets these criteria, the photoreceptors of invertebrates are very differ-

264

M. R. Taylor et al.

ent from vertebrates (Yarfitz and Hurley 1994). In contrast, the zebrafish provides a vertebrate model that has the qualities needed for identifying genes necessary for visual function and has been used in recent years to isolate genes essential for this process (Li and Dowling 1997; Kay et al. 2001; Brockerhoff et al. 2003; Taylor et al. 2004). The zebrafish mutant, partial optokinetic response b (pob), was isolated using an N-ethyl N-nitrosourea (ENU)based screening strategy designed to identify larvae with defective optokinetic responses in red (⬎640 nm) but not white light (Brockerhoff et al. 1997). Previous studies demonstrated that the red-blind defect was caused by the selective degeneration of long-wavelength sensitive cones and was not due to mutations in the most likely candidate, red opsin (Brockerhoff et al. 1997). pob is of particular interest from a vision standpoint because mutations that disrupt a single-cone type are rare (for review see Michaelides et al. 2004). Mutations that cause degeneration of photoreceptors fall into two categories, light dependent and light independent. Lightdependent mutations often affect the phototransduction cascade, whereas light-independent mutations are more likely structural components of photoreceptors or molecules involved in opsin protein transport (Pak 1995). Here we show that pob photoreceptors degenerate in a light-independent manner due to a mutation in a highly conserved protein of unknown cellular function. Localization studies show the protein to be intracellular in COS cells and in the inner segment and synapse of photoreceptors. Furthermore, we find conserved sequence motifs consistent with a role for Pob in protein sorting/trafficking. Our study demonstrates the utility of zebrafish for identifying novel conserved genes with critical roles in vertebrate photoreceptor survival.

MATERIALS AND METHODS Zebrafish strains and maintanence: AB* and WIK strain zebrafish were obtained from the University of Oregon and maintained as an inbred stock in the University of Washington Zebrafish Facility. The ENU-induced mutant pob a1 was previously isolated (Brockerhoff et al. 1997). Heterozygous AB* strain adults carrying the pob mutation were outcrossed with the WIK strain. Carriers of the pob mutation in the F1 generation were identified and inbred to produce a mapping panel of 2000 mutants. Adult fish and larvae were maintained at 28.5⬚ in reverse-osmosis distilled water reconstituted for fish compatibility by addition of salts and vitamins (Westerfield 1995). Transducin in situ binding assay: A transducin in situ binding assay was performed as described (Taylor et al. 2000). Larvae at 6 days postfertilization (dpf) from an adult heterozygous cross were raised in the absence of light, anesthetized on ice for 10 min, and then transferred to optimal cutting temperature (OCT)-filled 15 ⫻ 15 ⫻ 5-mm vinyl molds (Miles, Kankakee, IL). The larvae were submerged and their heads were aligned along one edge of the mold using fine-tip forceps. The OCT blocks were frozen in a dry ice/ethanol bath and stored at

⫺80⬚ in darkness. Transverse cryosections were cut at 10 ␮m through the eyes and applied to nitrocellulose membrane filters (25 mm, 45-␮m pore size). To perform the assay, individual filters were placed on a filter support, preincubated in 50 ␮l PBS buffer containing 100 ␮m GDP and 2 mm NADPH for 1 min at room temperature, and filtered by applying a vacuum. Immediately, reaction buffer [50 ␮l of PBS buffer containing 100 ␮m GDP, 2 mm NADPH, 100 nm GTP␥S, [35S]GTP␥S (ⵑ1 ␮Ci/reaction)] was added, and the sections were exposed to a flash of 640 nm red light (ⵑ300 ␮W/cm2) for 1 sec followed by a 1-min incubation at room temperature. Reactions were stopped by washing the filters under vacuum five times with 5 ml of PBS. After drying for 1 hr at room temperature, the filters were exposed to Kodak XAR film overnight. All steps, when necessary, were performed in complete darkness or under infrared illumination while wearing night vision goggles. Amplified fragment length polymorphism analysis: Amplified fragment length polymorphism (AFLP) markers were generated as described (Vos et al. 1995; Ransom and Zon 1999). Briefly, pooled DNA from 30 wild-type larvae or 30 pob mutants from an AB*/WIK mapping panel was digested with Eco RI or Mse I to generate fragments of varying lengths. Oligonucleotide anchors were ligated to the digested fragments and PCR was performed with DNA primers specific to the anchor but with an additional nucleotide at the 3⬘-end. Another round of PCR was performed with 33P-labeled primers that had two additional nucleotides at the 3⬘-end. These products were denatured in formamide at 100⬚ and run on a 4% denaturing sequencing gel. Following autoradiography, DNA fragments found to cosegregate with the mutation were subcloned and sequenced, and specific DNA primers were designed. Radiation hybrid panel mapping: AFLP markers were mapped using radiation hybrid (RH) panels as described (Geisler et al. 1999; Hukriede et al. 1999). To map genes and markers, DNA primers were first tested on zebrafish and hamster DNA to demonstrate that the primers were specific to the zebrafish sequence. Positive clones were scored and the data were submitted for RH map positions at http://mgchd1. nichd.nih.gov:8000/zfrh/beta.cgi for the Ekker panel or at http://134.174.23.167/zonrhmapper/instantMapping.htm for the Goodfellow panel. Positional cloning and gene identification: Recombination analysis was performed as described (Talbot and Schier 1999). Tightly linked polymorphic markers were generated or identified and were used to identify recombinants by PCR and agarose gel electrophoresis or single-strand conformational polymorphisms (SSCP) (Fornzler et al. 1998). Bacterial artificial chromosome (BAC) or P1 artificial chromosome (PAC) libraries were screened as described (Amemiya et al. 1999). BAC and PAC ends were sequenced from purified clones using SP6 and T7 primers. To identify genes located on PAC 105e21, a gridded zebrafish adult retina cDNA library containing 27,000 clones, constructed by the Resourcen-Zentrum Primary Database (RZPD) in Germany, was screened. For probe generation, PAC 105e21 was partially digested with Eco RI, denatured at 100⬚, and random primers were annealed. DNA polymerase was added in the presence of [␣-32P]dCTP for 10 min at 37⬚ and the unincorporated nucleotides were removed with a G-50 column. The probe was hybridized to the gridded library in triplicate, and the positive clones were identified and ordered from the RZPD. Clones were sequenced on both ends with T3 and T7 primers and BLAST searches were performed to identify the expressed sequence tag (EST). Candidate genes were further analyzed by RH panel mapping, recombination analysis, and DNA sequencing from wild-type and mutant larvae. Tol2 -mediated transgenic rescue of pob : A DNA fragment

Cloning of the Zebrafish pob Gene containing the full-length coding region of the wild-type pob gene was amplified by PCR using the cDNA clone as a template. The primers used were 5⬘-CTACCGGTCGCCACCATG GCTGAGCCCGAGC-3⬘ and 5⬘-ACATCGATTTAAAAGATGC CCGTTGGC-3⬘, which contain restriction sites for Age I and Cla I, respectively. The resulting fragment was digested with the restriction enzymes and cloned into the Age I/Cla I sites of pT2KXG, a derivative of pT2KXIG that lacked the intronic sequence, to replace the eGFP gene (Kawakami and Shima 1999). The resultant construct, pT2KX-pob, has the nonautonomous Tol2 transposable element and drives expression of the wild-type Pob protein under control of the EF1␣ promoter. Tol2 transposase mRNA was transcribed in vitro as described previously (Kawakami and Shima 1999). The pT2KX-pob plasmid and Tol2 mRNA were mixed at final concentrations of 30 ng/␮l in RNase-free water containing 0.05% phenol red as a tracking dye. The mixture was microinjected into one- to two-cell-stage embryos from two adult pob heterozygotes using a PV 820 pneumatic picopump (World Precision Instruments, Sarasota, FL) and a micromanipulator (Narishige International, Greenvale, NY). F0 founder fish were identified by PCR analysis of genomic DNA pools of 2- to 4-day old F1 embryos with primers specific for the Tol2 element sequence (Kawakami and Shima 1999). We identified four founder fish, among which two males and one female were heterozygous for the pob mutation. Then, a cross was conducted between a heterozygous transgenic male and a heterozygous nontransgenic female to determine if the wild-type pob transgene could rescue pob homozygous mutants. Larval fish from the cross were scored for optokinetic responses in red light and subjected to genomic PCR analyses for the transgene and DNA sequencing for the pob mutation. Cellular localization of Pob: For heterologous expression, wild-type and mutant pob constructs were made as follows. The wild-type cDNA was PCR amplified from the start to the amino acid prior to the stop codon with primers containing an Nco I site (5⬘-end) and an XhoI site (3⬘-end) from a zebrafish retinal cDNA library using Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA). After restriction enzyme digestion, this fragment was ligated into the similarly digested pTriEx-4 vector (Novagen). This introduced a His-Tag at the C-terminal end of Pob. The pob mutant cDNA was PCR amplified and cloned into pCR4-TOPO (Invitrogen, San Diego). To introduce the pob mutation into the pTriEx-4 construct, a HindIII/Pst I digest of pob mutant cDNA, containing the mutation, was inserted into the corresponding HindIII/Pst I site in the wild-type pTriEx-4 construct. COS-7 cells were grown in four-well tissue culture slides in Dulbecco’s modified Eagle’s medium/nutrient mixture F-12 (50/50) (Invitrogen), 10% heat-treated fetal calf serum, 10 mm HEPES, pH 7.3. Five hundred nanograms of the wild-type or mutant construct was transiently transfected into 50–60% confluent cells using Fugene-6 reagent (Roche Molecular Biochemicals). After 48 hr, the cells were washed with PBS, fixed with 4% paraformaldehyde/PBS, permeabilized with PBS/ 0.1% Triton-X-100 for 15 min at room temperature, and incubated with blocking buffer (PBS containing 5% goat sera and 0.1% Triton-X-100) for 1 hr. After the removal of blocking buffer, cells were incubated with a 1:1000 dilution of anti-his monoclonal (Novagen) in blocking buffer for 1 hr at room temperature. After repeated washing, cells were incubated with a 1:1000 dilution of goat anti-mouse antibody conjugated to Alexa Fluor 594 (Molecular Probes, Eugene, OR) in blocking buffer for 1 hr at room temperature. For trans -Golgi staining, cells were incubated for 30 min with 7-nitrobenz2-oxa-1,3-diazole (NBD) ceramide (Molecular Probes) conjugated to BSA (diluted 1:100 in PBS), immediately following the removal of the secondary antibody. Cells were then re-

265

blocked for 30 min prior to final washing and the addition of 4⬘,6-diamidino-2-phenylindole (DAPI) (Molecular Probes). After rinsing with PBS, cells were mounted with FluoromountG (Southern Biotechnology Associates) and viewed with either a Nikon Microphot-FX fluorescent microscope or a Zeiss fluorescent microscope. For endoplasmic reticulum (ER) staining, enhanced green fluorescent protein (EGFP)-tagged pob was generated by PCR amplification, using the primers 5⬘-TCTCGAGCTCCCTGG AAACAGTGCG-3⬘ and 5⬘-GGTGGATCCAATATGGACGTCTG TAGTTC-3⬘ from the mouse pob IMAGE:3157147 clone (MRC Geneservice, Cambridge, UK). The 5⬘ primer was flanked with a Sal I restriction site and the 3⬘ primer was designed to remove the stop codon and insert a Bam HI restriction enzyme site in frame with the EGFP sequence. The PCR product was cloned into the EGFP-N1 expression vector (BD Biosciences). COS7 cells were cotransfected using Fugene-6 reagent (Roche Molecular Biochemicals) with 1:1 pob -EGFP-N1 and pDsRed2ER (CLONTECH, Palo Alto, CA; gift of Dirk Snyders, University of Antwerp) to label ER. Images were collected with a Zeiss Axiovert 100M and LSM-510 software by exciting EGFP at 488 nm and DsRed fluorescence by exciting at 568 nm and using tetramethylrhodamine isothiocyanate filters for fluorescence emission. We used an objective of ⫻63 with a pinhole of 0.83 Airy units that gives an optical slice thickness (Z direction) of ⵑ0.9 ␮m. Pixel width was between 0.21 and 0.36 ␮m. Several pictures were taken from different confocal planes to follow the pob EGFP localization. Antibody production and immunoblotting: GST fusion proteins encoding amino acids 38–100 or 181–261 of Pob were constructed in the pGEX-2T vector and expressed and purified in E. coli as described by the manufacturer (Amersham, Buckinghamshire, UK). The purified proteins were used independently to generate polyclonal antibodies in rabbits. In addition, a synthetic 19-mer peptide from Ser 249 to the C-terminal end of Pob was used to generate mouse monoclonal antibodies. Immunoblotting was performed according to standard protocols. Whole eye, body, and brain homogenates from 24 pob or 24 wild-type siblings at 5 days dpf were subjected to 15% PAGE. After transfer to nitrocellulose, the membranes were probed simultaneously with G5H3 (1:1000) or Pob (1:3000) monoclonal antibodies for 3 hr at room temperature. The G5H3 antibody was originally made against a Drosophila G protein ␤-subunit and recognizes several ␤-subunits in Drosophila (Yarfitz et al. 1991). In zebrafish, G5H3 recognizes a single polypeptide at ⵑ34 kDa that was used as a loading control. Secondary goat anti-mouse conjugated to adaptor protein (AP) (Sigma) and nitroblue tetrazolium/5-bromo4-chloro-3-indoyl-phosphate (Roche) were used for detection according to the manufacturers’ protocols. All antibodies produced similar results when used for immunoblot analysis or immunohistochemistry. Immunolocalization of Pob: Zebrafish larvae at 6 dpf were fixed in 4% paraformaldehyde/PBS for 10 min at room temperature (overfixation resulted in a significant reduction of Pob immunoreactivity). Fixed larvae were embedded in OCT and cryosectioned at 10 ␮m thickness. The sections were collected onto Superfrost/Plus slides (Fisher), dried for 30 min at 37⬚, washed briefly in PBS, and blocked in PBS containing 5% goat serum, 1% BSA, and 0.3% Triton X-100 for 1 hr at room temperature. The sections were incubated overnight at 4⬚ with the anti-Pob monoclonal antibody (1:500) and with anticone opsin polyclonal antibodies (1:200) kindly provided by Tom Vihtelic. After three washes in PBS containing 0.3% Triton X-100 for 10 min each at room temperature, sections were incubated with goat anti-mouse Alexa Fluor 594 and goat anti-rabbit Alexa Fluor 488 (Molecular Probes) for 3 hr at

266

M. R. Taylor et al.

room temperature. After three washes in PBS for 10 min each, sections were incubated with DAPI (1 ␮g/ml) for 1 min at room temperature. Finally, sections were briefly rinsed in PBS, coverslipped with Fluoromount-G (EMS), and visualized on a Nikon Microphot-FX fluorescence microscope. Whole-mount in situ hybridization: The pob riboprobe was generated by PCR amplification using total cDNA from 26hr-postfertilization (hpf) embryos as the template; forward primer was 5⬘-GCAGTGGCTCTATGGTGTCGCA-3⬘ and reverse primer was 5⬘-AAAGATGCCCGTTGGCAAGT-3⬘. After cloning into the pGEM-T Easy vector (Promega, Madison, WI), the plasmid was linearized with Not I and transcribed with SP6 polymerase (antisense probe) or linearized with Sal I and transcribed with T7 polymerase (sense probe). Hybridization was performed on 24-hpf embryos under standard conditions (Broadbent and Read 1999).

RESULTS

Red cone photoreceptor degeneration in pob is light independent: We used an in situ binding assay to determine if red cone degeneration depends on light. Light stimulates cone opsins to catalyze the exchange of GTP for bound GDP on transducin ␣-subunits. By using a nonhydrolyzable analog of GTP, GTP␥S, transducin activation can be examined by measuring the light-dependent incorporation of radiolabeled GTP␥S. Larvae at 6 dpf from a cross of pob heterozygous adults were raised in total darkness and subjected to the [35S]GTP␥S in situ binding as described in materials and methods. As shown in Figure 1, one-quarter of the larval retinas (arrowheads) did not incorporate radioactivity in response to red light. The same result had previously been demonstrated on pob larvae raised on a normal light cycle of 14 hr on/10 hr off (Taylor et al. 2000). These data demonstrate that red cones are not biochemically functional in pob mutants regardless of whether they were reared under light or dark conditions. The lightindependent degeneration of red cones was further confirmed by whole-mount in situ hybridization using a DIG-labeled RNA probe generated against zebrafish red opsin (data not shown). pob encodes a novel 30-kDa protein: We performed positional cloning to identify the gene responsible for the selective degeneration of long-wavelength sensitive photoreceptors in the red-blind zebrafish mutant pob. Initially, we produced a mapping panel and generated AFLP markers to map the pob mutation (see materials and methods). Using an RH panel, the AFLP marker, AFLPF2/R2, mapped pob to linkage group (LG) 6 and was found to be 0.06 cM from the pob gene (see Figure 2). We also found that rod transducin-␣, a gene that we had previously mapped (Brockerhoff et al. 2003), resides 0.2 cM from pob. By examining the RH panels, previously mapped EST markers were considered as candidate genes and as tightly linked markers. Using SSCP analysis (Fornzler et al. 1998), we found that EST fb37d04, corresponding to the zebrafish homolog of the sec61 gene, was one recombinant or 0.03 cM from

Figure 1.—Light-independent degeneration in pob. A transducin in situ binding assay was performed to determine whether red cone degeneration in pob is dependent on light. (A) Autoradiography on wild-type and mutant sections from fish grown in darkness and then subjected to the in situ binding assay. One-quarter of the larvae did not incorporate [35S]GTP␥S, suggesting that red cone degeneration in pob is light independent. (B) Adjacent section stained with eosin. Arrowheads indicate pob mutants.

the pob gene. A chromosomal walk was initiated from AFLPF2/R2 on the proximal side and EST fb37d04 on the distal side. Two rounds of “walking” were performed on the proximal side using a PCRable BAC library. On the distal side, EST fb37d04 was used to identify a single BAC, but this BAC was not useful due to chimerism. A gridded PAC library was screened by hybridization with EST fb37d04-specific primers, which isolated three PACs. PAC 105e21 was found to span the critical interval and therefore contained the pob gene (Figure 2A). To identify genes located on PAC 105e21, we used two approaches. First, we produced 32P-labeled DNA probes from regions of the PAC using random primers, screened a gridded cDNA library of adult zebrafish retina, and found 28 positive clones (see materials and methods). Clones were placed into seven categories on the basis of sequence alignments. As a second approach, we sequenced fragments of PAC 105e21 that were generated through subcloning of random fragments. The only additional gene identified was highly homologous to a deubiquitination protease (UnpEL). To determine if any of these clones were potential

Cloning of the Zebrafish pob Gene

267

Figure 2.—Positional cloning of the pob gene. (A) AFLP analysis and RH panel mapping placed the marker, AFLP F2/R2, on LG 6. AFLP F2/R2 and EST fb37d04 were used for recombination analysis and to initiate a chromosomal walk. PAC 105e21 was found to span the critical interval. The number of recombination events between markers is shown with regular type on the proximal side and italic type on the distal side. (B) The translated sequence of the 30-kDa gene and the alignment of the zebrafish, human (AF157321), and mouse (AAH04641) proteins. The putative signal peptide is underlined; the putative transmembrane regions are shown with a bar above the sequences; YXXØ and [DE]XXXL[LI] motifs are shown in boldface italic type; amino acid differences among species are shown in boldface type; and the conserved site of mutation is boxed. (C) The pob mutation was found in a splice acceptor site (g ⬎ t transversion) that resulted in a twoamino-acid deletion (T138_ K139del) and a missense mutation (V140I).

candidates for the pob gene, we performed RH panel and recombination analysis with primers designed against the 5⬘- and 3⬘-UTRs. Inositol hexakisphosphate kinase 2 and calcium-binding protein 5 were eliminated as candidates because they were located outside the critical interval, while other candidates turned out to be false positives because they mapped to different genomic locations. The two candidates that remained were the gene for UnpEL and a novel gene encoding a highly conserved 30-kDa protein of unknown function (Figure 2B). We designed primers against the coding regions of these genes and used these primers to sequence cDNAs from both wild-type larvae and pob mutants. We did not find any mutations within the coding region of UnpEL.

We did, however, find a 6-bp deletion within the cDNA of the 30-kDa gene. Since ENU most commonly introduces single nucleotide changes (Shibuya and Morimoto 1993; Anderson 1995), we sequenced the genomic DNA surrounding the putative deletion. This uncovered a splice acceptor-site mutation that results in a two-amino-acid deletion and a single missense mutation in the pob protein (Figure 2C). The presence of a linked nonconservative sequence change within 30 kDa only in the mutants suggested that we had identified the correct gene. Sequence analysis of the Pob protein revealed a putative signal sequence, potential transmembrane regions, and protein-sorting motifs (see Figure 2B). Using the

268

M. R. Taylor et al.

SignalP algorithm (http://www.cbs.dtu.dk/services/Sig nalP/), a signal peptide was predicted at the N terminus with cleavage between amino acids 28 and 29. Two to three transmembrane regions were detected using TMHMM (http://www.cbs.dtu.dk/services/TMHMM/) and TMpred (http://www.ch.embnet.org/software/tm base/TMPRED_doc.html), respectively. However, if the signal peptide is functional, then the first predicted transmembrane region would not exist, resulting in only one or two transmembrane regions. We also recognized two YXXØ motifs and one [DE]XXXL[LI] motif that are conserved between species (X represents any amino acid; Ø represents a hydrophobic amino acid with a bulky side chain). These motifs are recognized by adaptor protein complexes, which function in protein sorting and transport (Bonifacino and Traub 2003). In addition, YXXØ signals have been found in type I and type II transmembrane proteins and are often situated near the C terminus of transmembrane domains (for review see Bonifacino and Traub 2003). Rescue of the pob phenotype: We rescued the mutant by generating a stable transgenic fish line that expresses the wild-type Pob protein to demonstrate that the pob phenotype is due to the mutation in this gene. To generate a stable transgenic line, we used the Tol2 transposon system that has been recently described (Kawakami and Noda 2004; Kawakami et al. 2004). Embryos from a cross between two pob heterozygotes were co-injected at the one- to two-cell stage with the Tol2 vector containing the wild-type pob cDNA downstream of the EF1␣ promoter and the Tol2 transposase mRNA. Expression of the wild-type Pob protein in the injected fish was confirmed by immunoblot analysis of larval extracts (data not shown). Of 17 injected fish screened, 4 showed germline transmission of the transgene, and 3 of these were heterozygous for the pob mutation. We next mated a founder fish heterozygous for the pob mutation with a nontransgenic fish that was also heterozygous for the pob mutation. We screened fish from this cross at 5 dpf using our behavioral screening apparatus and scored the fish according to whether they were red blind or not. Of 126 fish screened, 23 were red blind and 103 were not. None of the red-blind fish carried the transgene whereas 33 of the 103 red-positive fish were identified as carrying the transgene. We sequenced the genomic region of the pob gene near the mutation in carriers of the transgene; 12 were heterozygous for the pob mutation, 11 were wild type, and 10 were homozygous for the splice acceptor-site mutation. Therefore, the presence of the transgene was able to rescue the mutant phenotype, verifying that the pob phenotype is indeed due to the mutation in the gene encoding the 30-kDa protein. Pob localizes to intracellular compartments: To determine the cellular localization of Pob, we expressed it in a heterologous tissue culture system. A His-tagged version of the zebrafish wild-type and mutant protein

Figure 3.—Intracellular localization of Pob protein in COS-7 cells. (A) Wild-type pob construct with a C-terminal Histag was transiently transfected into COS-7 cells. Wild-type Pob protein (red) showed an intracellular localization and an asymmetric distribution around the nucleus (blue). (B–F) Colocalization of Pob with markers for the trans-Golgi (B–D) and ER (E and F). (B–D) Fluorescence microscopy of COS-7 (B) stained with NBD ceramide (Molecular Probes) to mark the trans-Golgi (green), (C) transiently transfected with wild-type pob construct with a C-terminal His-tag (red), and (D) an overlay of ceramide, His-tag, and DAPI (blue). Note some colocalization with the trans-Golgi, but also an extension of possible emanating vesicles. (E and F) Two confocal sections of COS-7 cells transfected with pDsRed2-ER (red) and pEGFPN1 mouse pob (green). (E) Image on the left is at the level of the nucleus (N); the inset shows a digital zoom from the marked zone. (F) The image on the right is at a plane above the nucleus. Similar to the trans-Golgi staining with the zebrafish clone, we see some colocalization between mouse Pob and the ER.

was expressed in COS-7 cells and localized by immunofluorescence at 48 hr. The wild-type Pob protein is distributed asymmetrically around the nucleus in a pattern

Cloning of the Zebrafish pob Gene

that resembles Golgi or ER (Figure 3A). A similar distribution was found with mutant Pob protein (data not shown). To further examine the cellular distribution, wild-type Pob-transfected COS-7 cells were stained with NBD ceramide to label the trans-Golgi network. These data indicated that Pob, at least partially, colocalizes with trans-Golgi (Figure 3, B–D). In addition, we coexpressed a mouse Pob-EGFP fusion protein with pDsRed2-ER. pDsRed2-ER is a mammalian expression vector that encodes a fusion consisting of Discosoma sp. red fluorescent protein (DsRed2) and the ER-targeting sequence of calreticulin fused to the 5⬘-end. Using confocal imaging on several different optical slices, Pob appears to partially colocalize with pDsRed2-ER (Figure 3, E and F), demonstrating ER localization. Moreover, outside of the reticulum Pob is localized in small vesicles (presumed to be transitional vesicles), adjacent to the ER, and also shows budding of vesicles from pDsRed2 positive ones (Figure 3E, inset), suggesting that positive Pob vesicles are released by or fused to ER. Taken together, these results suggest that Pob is present in multiple intracellular compartments of COS-7 cells and may indicate a dynamic distribution of the protein. Pob expression in wild-type and pob mutant zebrafish: We generated both monoclonal and polyclonal antibodies against Pob and performed immunoblot analysis of wild-type and pob eye homogenates (Figure 4A). Although the Pob protein levels are similar in wild-type and mutant extracts, the mutant protein migrates slightly faster than wild type, likely due to the two-amino-acid deletion. This indicates that the mutant protein is expressed and that the deletion and the missense mutation do not dramatically destabilize it. We also analyzed wild-type homogenates from body, eye, and brain separately to determine if Pob is expressed outside of the retina. As shown in Figure 4B, Pob is found in tissues other than the retina, indicating an extensive distribution of the protein. Next we analyzed the cellular localization of Pob in zebrafish by immunohistochemistry. The localization of Pob within wild-type zebrafish retinas at 6 dpf is shown in Figure 5, A and B. Pob appears to be concentrated in the inner segment and synapse of photoreceptor cells. Consistent with our immunoblot analysis, we detected a similar level and distribution of mutant Pob within the mutant retina (Figure 5, C and D). Note in Figure 5D that the pob photoreceptor layer is thinner due to the absence of red cones (see Brockerhoff et al. 1997). The presence of mutant Pob in mutant photoreceptors indicates that Pob is not specific to red cone photoreceptors. Pob is also abundant in the brain as well as in other tissues (Figure 5 and data not shown). To further examine the expression of the pob, we performed whole-mount in situ hybridization on wildtype embryos. Using an antisense riboprobe, the pob transcript was detected in the eye, brain, and along the body of the embryo (Figure 5F). As a negative control,

269

Figure 4.—Immunoblotting of Pob. (A) Eye homogenates from wild-type and mutant (Mut) larvae at 5 dpf were examined by immunoblotting with a monoclonal antibody against zebrafish Pob and monoclonal antibody G5H3 as a loading control. Pob migrates at ⵑ30 kDa as predicted and appears to be expressed at similar levels between wild-type and mutant homogenates. Note a slightly faster migration in the mutant lane due to the two-amino-acid deletion. (B) Zebrafish larvae at 5 dpf were fractionated into body, eye, and brain samples. Duplicate samples indicate that Pob is widely expressed in larvae.

the sense riboprobe did not show any staining (Figure 5E). These data confirm the widespread distribution of Pob as shown by immunoblot analysis (Figure 4B). DISCUSSION

A number of different screening strategies based on analyzing zebrafish visual behavior have been used to successfully identify mutations that cause subtle and rare defects in the visual system (Li and Dowling 1997; Kay et al. 2001; Brockerhoff et al. 2003; Taylor et al. 2004). The aim of our screen has been to identify mutations that disrupt cone photoreceptors. We have relied on a two-phase screening strategy that uses visual behavior of young larvae for initial identification, followed by electroretinogram and histological analysis for confirmation. The zebrafish pob mutant was first identified using a behavioral screening strategy in which fish were

270

M. R. Taylor et al.

Figure 5.—Immunolocalization and in situ hybridization of Pob. (A–D) The cellular localization of Pob was analyzed by immunohistochemistry using 6-dpf sections from wild-type and mutant (pob) larvae. (A and C) In both wild-type and mutant, Pob appears to be present in the inner segment and synapse of photoreceptors. It is clearly abundant in the brain as well. (B and D) Overlay showing Pob staining (red) in comparison to opsins (green) and nuclei position (blue). Note that the pob photoreceptor layer is thinner due to the specific lack of red cones (D); however, the Pob protein staining remains, suggesting that Pob is expressed in several photoreceptor types. (E–G) In situ hybridization analyzing pob expression in wild-type zebrafish larvae at 24 hpf. (E) The sense probe shows no staining. (F) The antisense probe shows staining in the eye, brain, and faint staining along the body. (G) Close-up view detailing pob staining in the early zebrafish eye.

screened for defects in visual tracking of rotating stripes illuminated with red or white light (Brockerhoff et al. 1997). This strategy had the potential for success because the cones sensitive to the longest wavelength in zebrafish have an absorbance maximum of ⵑ80 nm redshifted from the next cone type (i.e., 564 nm vs. 473 nm) (Robinson et al. 1993). Using this paradigm, we identified pob mutants as larvae that respond to white light but not red light. Initial studies demonstrated that long-wavelength photoreceptors selectively degenerate in pob mutants and that this was not due to a mutation in the most likely candidate, the red opsin gene (Brockerhoff et al. 1997). In humans, cone dysfunction syndromes compose a clinically and genetically diverse group of disorders. The genetic cause has been determined for several color vision defects, achromatopsia (the loss of all color vision), blue monochromatism, protanopia, deuteranopia, and enhanced S-cone syndrome. Defects that af-

fect all photoreceptors or that are positioned on the X chromosome are more common. For example, since many of the phototransduction components are shared between the different cone types, several genes involved in phototransduction have been shown to cause achromatopsia. The majority of blue monochromatism defects are caused either by defects in the locus control region or by rearrangement of the X chromosome L and M pigment gene array. Similarly, protonopia and deuteranopia are due to rearrangements, mutations, or polymorphisms in the red/green opsin gene array (for review see Michaelides et al. 2004). One rare defect causing color blindness, enhanced S-cone syndrome, has been identified as the NR2E3 transcription factor (Haider et al. 2000). Whereas achromatopsia and blue monochromatism typically do not lead to the degeneration of cones, enhanced S-syndrome does eventually lead to generalized cone photoreceptor degeneration in humans (Milam et al. 2002). Human disorders involving

Cloning of the Zebrafish pob Gene

single-cone photoreceptor dysfunction or degeneration have been identified rarely and have not been well characterized. For these very rare disorders, animal models are critical for gene identification and therapeutic exploration. In the current study, we demonstrate that red blindness in pob is due to a two-amino-acid deletion and missense mutation in a highly conserved 30-kDa protein of unknown function. Although the Pob protein appears expressed in many cell types in the retina, brain, and other tissues, the mutant phenotype appears very specific. One possible explanation for this is that Pob interacts with a protein that is specific to long wavelength photoreceptors, such as red opsin (see below). Another explanation could be that the pob mutation generates a hypomorphic allele and is more penetrant in red cone photoreceptors. Future work will examine these issues. This study represents the first presentation of the identification of a gene that can selectively disrupt a single-cone type when mutated in zebrafish. It demonstrates the usefulness of zebrafish as a model organism for identifying novel, highly conserved genes that play specific and critical roles in retinal function. To rescue the red-blind phenotype, we generated transgenic animals with the wild-type pob gene using the Tol2 transposon system. The Tol2 transposable element was identified in the genome of medaka, Oryzias latipes, and it encodes the only known autonomous transposase in vertebrates (Kawakami and Shima 1999). Previous work demonstrated that Tol2 transposition into the zebrafish germline is possible by co-injecting fertilized eggs with the Tol2 transposase transcript along with a plasmid containing the necessary cis sequences (Kawakami et al. 2000). More recently, this system was also shown to be functional in mouse embryonic stem cells and it was used as a gene trap to identify developmentally regulated genes in zebrafish (Kawakami and Noda 2004; Kawakami et al. 2004). The primary advantage of this system is that the insertion efficiency appears significantly higher than with injection of linear DNA, leading to higher germline transmission frequencies. In our study we found that nearly 25% of the fish developed from embryos injected with the wild-type pob construct showed germline transmission. Since the red-blind phenotype of pob does not appear until ⵑ4 dpf, and red photoreceptors are distributed throughout the retina at this stage, we concluded that it would be difficult to achieve a complete rescue with transient expression of wild-type Pob by injection of either a DNA construct or mRNA into fertilized eggs. We used the EF1␣ promoter, a strong ubiquitous promoter in zebrafish, to drive expression of the protein in the transgenic embryos. Since the pob phenotype is recessive and lethal at the larval stage, we generated wild-type pob transgenic fish in a heterozygous background. This allowed us to obtain homozygous mutants carrying the wild-type transgene in the next generation. It was expected that a cross between a

271

transgenic, heterozygous founder and a nontransgenic, heterozygous founder would produce one-quarter homozygous mutant siblings, of which up to a half would contain the transgene, depending upon the mosaicism of the transgenic founder fish. With low mosaicism, as expected for the Tol2 transposon system, it was possible to obtain an adequate number of homozygous mutants carrying the wild-type transgene to rescue the red-blind phenotype. Of 126 fish tested, 10 fish were identified as both homozygous for the pob mutation and carrying the wild-type pob transgene. All of these fish showed normal optokinetic response in red light as well as in white light. Furthermore, none of 23 red-blind siblings carried the wild-type pob transgene. The results clearly indicate that the pob phenotype was rescued by exogenous expression of wild-type Pob and verify that the mutation that we found in the 30-kDa gene is indeed responsible for the red cone degeneration in pob. Genetic defects that result in photoreceptor cell loss fall into two classes. The first class includes mutations that cause light-dependent photoreceptor degeneration. These mutations are often found in genes that encode several key components of the phototransduction cascade, such as arrestin and rhodopsin kinase (C. K. Chen et al. 1999; J. Chen et al. 1999). The second class includes mutations that induce the light-independent degeneration of photoreceptor cells. The pob mutation falls into this second category. These mutations are typically found in genes encoding structural components of photoreceptors or in genes required for the synthesis, folding, or transport of rhodopsin (Pak 1995). Within this class, several Drosophila mutants have been identified including ninaA, ninaC, and ninaE. The ninaA gene encodes a tissue-specific cyclophilin homolog required for the proper transport of opsin from the ER to the rhabdomeres, demonstrating a likely role in protein folding and/or trafficking (Colley et al. 1991; Stamnes et al. 1991). Interestingly, the vertebrate cyclophilin protein, RanBP2, has been shown to function as a chaperone for red and green opsins (Ferreira et al. 1996). Cyclophilins are ubiquitous and abundant proteins, yet only specific isoforms, such as RanBP2, are expressed in cone photoreceptor cells. More recently, intraflagellar transport (IFT) proteins have been shown to be essential for the differentiation and survival of vertebrate sensory neurons. In the zebrafish oval mutant, a mutation that results in the absence of photoreceptor outer segments has been identified in the ift88 gene (Tsujikawa and Malicki 2004). In addition, morpholino knockdown of the IFT complex B polypeptides, IFT52 and IFT57, results in a phenotype similar to oval (Tsujikawa and Malicki 2004). To start to determine the function of Pob, we localized the protein in COS-7 cells and in the retina. In the retina, Pob was clearly localized in the inner segment (IS) and synapse of photoreceptors. The presence of putative transmembrane regions indicates that Pob

272

M. R. Taylor et al.

likely spans a membrane and may localize to vesicles involved in protein or transmitter transport. Although overexpression of Pob in COS-7 cells had a variable distribution, we have found a distinct colocalization with markers of the trans-Golgi and ER. As the pob mutation is located near one of the transmembrane regions and the mutant protein is still localized to the IS and synapse of photoreceptors, it is possible that the deletion of two conserved amino acids may disrupt the proper topology of the protein within the membrane. Sequence analysis revealed the presence of both YXXØ and [DE]XXXL[LI] motifs in Pob, further suggesting that it may play a role in protein sorting/transport. AP complexes AP-1, AP-2, AP-3, and AP-4 recognize these signals and function in the sorting of transmembrane proteins to endosomes and lysosomes (Bonifacino and Traub 2003). AP complexes have also been implicated in targeting proteins to the basolateral membrane in polarized epithelial cells. Studies of viral membrane protein sorting in cultured neurons suggest that the neuronal somatodendritic compartment may be analogous to the epithelial basolateral region (Dotti and Simons 1990). In fact, mutations in Caenorhabditis elegans UNC-101, the ␮1 subunit of AP-1, result in the mislocalization of odorant receptors to olfactory cilia (Dwyer et al. 2001). Odorant receptor localization is also mediated by ODR-4, a novel membrane-associated protein (Dwyer et al. 1998). Interestingly, we also identified YXXØ and [DE]XXXL[LI] motifs in ODR-4, although it has not been determined if these signals are functional in this protein. Abnormal protein transport of rhodopsin leads to photoreceptor degeneration. Mutations within rhodopsin that cause misfolding and also mutations within the C-terminal transport signal are known to lead to accumulation within the IS and to subsequent photoreceptor degeneration (Sung and Tai 2000; Luo et al. 2004). Therefore, one possible explanation for the mutant phenotype is that Pob is involved in protein transport, specifically red opsin transport, and that the mutation leads to an accumulation of opsin within the inner segment and photoreceptor degeneration. Previous work has demonstrated that photoreceptors degenerate in pob, very soon after their formation, at ⵑ3–4 dpf. Future experiments will be directed toward determining the potential role of Pob in protein sorting/transport, identifying molecules that interact with Pob, and determining the direct cause of red cone photoreceptor apoptosis. The authors thank Irina Ankoudinov, Greg Niemi, and George Stearns for technical assistance with pob cloning and immunoblot analysis. We also thank Karen Rutherford for conducting pob transfection experiments and James B. Hurley for a critical reading of this manuscript as well as several helpful discussions. We also thank Camila Esguerra and Alex Crawford for helping A.D.J. with fish methodology. This work was supported by National Institutes of Health grant EY12373 (S.E.B.), a Molecular Cellular Biology training grant (M.R.T.), and a National Eye Institute core grant (EY01730).

LITERATURE CITED Amemiya, C. T., T. P. Zhong, G. A. Silverman, M. C. Fishman and L. I. Zon, 1999 Zebrafish YAC, BAC, and PAC genomic libraries. Methods Cell Biol. 60: 235–258. Anderson, P., 1995 Mutagenesis. Methods Cell Biol. 48: 31–58. Bonifacino, J. S., and L. M. Traub, 2003 Signals for sorting of transmembrane proteins to endosomes and lysosomes. Annu. Rev. Biochem. 72: 395–447. Broadbent, J., and E. M. Read, 1999 Wholemount in situ hybridization of Xenopus and zebrafish embryos. Methods Mol. Biol. 127: 57–67. Brockerhoff, S. E., J. B. Hurley, G. A. Niemi and J. E. Dowling, 1997 A new form of inherited red-blindness identified in zebrafish. J. Neurosci. 17: 4236–4242. Brockerhoff, S. E., F. Rieke, H. R. Matthews, M. R. Taylor, B. Kennedy et al., 2003 Light stimulates a transducin-independent increase of cytoplasmic Ca2⫹ and suppression of current in cones from the zebrafish mutant nof. J. Neurosci. 23: 470–480. Chen, C. K., M. E. Burns, M. Spencer, G. A. Niemi, J. Chen et al., 1999 Abnormal photoresponses and light-induced apoptosis in rods lacking rhodopsin kinase. Proc. Natl. Acad. Sci. USA 96: 3718–3722. Chen, J., M. I. Simon, M. T. Matthes, D. Yasumura and M. M. LaVail, 1999 Increased susceptibility to light damage in an arrestin knockout mouse model of Oguchi disease (stationary night blindness). Invest. Ophthalmol. Vis. Sci. 40: 2978–2982. Colley, N. J., E. K. Baker, M. A. Stamnes and C. S. Zuker, 1991 The cyclophilin homolog ninaA is required in the secretory pathway. Cell 67: 255–263. Dotti, C. G., and K. Simons, 1990 Polarized sorting of viral glycoproteins to the axon and dendrites of hippocampal neurons in culture. Cell 62: 63–72. Dowling, J. E., 1987 The Retina, An Approachable Part of the Brain. Harvard University Press, Cambridge, MA. Dwyer, N. D., E. R. Troemel, P. Sengupta and C. I. Bargmann, 1998 Odorant receptor localization to olfactory cilia is mediated by ODR-4, a novel membrane-associated protein. Cell 93: 455– 466. Dwyer, N. D., C. E. Adler, J. G. Crump, N. D. L’Etoile and C. I. Bargmann, 2001 Polarized dendritic transport and the AP-1 mu1 clathrin adaptor UNC-101 localize odorant receptors to olfactory cilia. Neuron 31: 277–287. Ferreira, P. A., T. A. Nakayama, W. L. Pak and G. H. Travis, 1996 Cyclophilin-related protein RanBP2 acts as chaperone for red/ green opsin. Nature 383: 637–640. Fornzler, D., H. Her, E. W. Knapik, M. Clark, H. Lehrach et al., 1998 Gene mapping in zebrafish using single-strand conformation polymorphism analysis. Genomics 51: 216–222. Geisler, R., G. J. Rauch, H. Baier, F. van Bebber, L. Brobeta et al., 1999 A radiation hybrid map of the zebrafish genome. Nat. Genet. 23: 86–89. Haider, N. B., S. G. Jacobson, A. V. Cideciyan, R. Swiderski, L. M. Streb et al., 2000 Mutation of a nuclear receptor gene, NR2E3, causes enhanced S cone syndrome, a disorder of retinal cell fate. Nat. Genet. 24: 127–131. Hukriede, N. A., L. Joly, M. Tsang, J. Miles, P. Tellis et al., 1999 Radiation hybrid mapping of the zebrafish genome. Proc. Natl. Acad. Sci. USA 96: 9745–9750. Kawakami, K., and T. Noda, 2004 Transposition of the Tol2 element, an Ac-like element from the Japanese medaka fish Oryzias latipes, in mouse embryonic stem cells. Genetics 166: 895–899. Kawakami, K., and A. Shima, 1999 Identification of the Tol2 transposase of the medaka fish Oryzias latipes that catalyzes excision of a nonautonomous Tol2 element in zebrafish Danio rerio. Gene 240: 239–244. Kawakami, K., A. Shima and N. Kawakami, 2000 Identification of a functional transposase of the Tol2 element, an Ac-like element from the Japanese medaka fish, and its transposition in the zebrafish germ lineage. Proc. Natl. Acad. Sci. USA 97: 11403–11408. Kawakami, K., H. Takeda, N. Kawakami, M. Kobayashi, N. Matsuda et al., 2004 A transposon-mediated gene trap approach identifies developmentally regulated genes in zebrafish. Dev. Cell 7: 133–144. Kay, J. N., K. C. Finger-Baier, T. Roeser, W. Staub and H. Baier,

Cloning of the Zebrafish pob Gene 2001 Retinal ganglion cell genesis requires lakritz, a zebrafish atonal homolog. Neuron 30: 725–736. Li, L., and J. E. Dowling, 1997 A dominant form of inherited retinal degeneration caused by a non-photoreceptor cell-specific mutation. Proc. Natl. Acad. Sci. USA 94: 11645–11650. Luo, W., N. Marsh-Armstrong, A. Rattner and J. Nathans, 2004 An outer segment localization signal at the C terminus of the photoreceptor-specific retinol dehydrogenase. J. Neurosci. 24: 2623–2632. Michaelides, M., D. M. Hunt and A. T. Moore, 2004 The cone dysfunction syndromes. Br. J. Ophthalmol. 88: 291–297. Milam, A. H., L. Rose, A. V. Cideciyan, M. R. Barakat, W. X. Tang et al., 2002 The nuclear receptor NR2E3 plays a role in human retinal photoreceptor differentiation and degeneration. Proc. Natl. Acad. Sci. USA 99: 473–478. Pacione, L. R., M. J. Szego, S. Ikeda, P. M. Nishina and R. R. McInnes, 2003 Progress toward understanding the genetic and biochemical mechanisms of inherited photoreceptor degenerations. Annu. Rev. Neurosci. 26: 657–700. Pak, W. L., 1995 Drosophila in vision research. The Friedenwald Lecture. Invest. Ophthalmol. Vis. Sci. 36: 2340–2357. Ransom, D. G., and L. I. Zon, 1999 Mapping zebrafish mutations by AFLP. Methods Cell Biol. 60: 195–211. Robinson, J., E. A. Schmitt, F. I. Harosi, R. J. Reece and J. E. Dowling, 1993 Zebrafish ultraviolet visual pigment: absorption spectrum, sequence, and localization. Proc. Natl. Acad. Sci. USA 90: 6009–6012. Shibuya, T., and K. Morimoto, 1993 A review of the genotoxicity of 1-ethyl-1-nitrosourea. Mutat. Res. 297: 3–38. Stamnes, M. A., B. H. Shieh, L. Chuman, G. L. Harris and C. S.

273

Zuker, 1991 The cyclophilin homolog ninaA is a tissue-specific integral membrane protein required for the proper synthesis of a subset of Drosophila rhodopsins. Cell 65: 219–227. Sung, C. H., and A. W. Tai, 2000 Rhodopsin trafficking and its role in retinal dystrophies. Int. Rev. Cytol. 195: 215–267. Talbot, W. S., and A. F. Schier, 1999 Positional cloning of mutated zebrafish genes. Methods Cell Biol. 60: 259–286. Taylor, M. R., H. A. Van Epps, M. J. Kennedy, J. C. Saari, J. B. Hurley et al., 2000 Biochemical analysis of phototransduction and visual cycle in zebrafish larvae. Methods Enzymol. 316: 536– 557. Taylor, M. R., J. B. Hurley, H. A. Van Epps and S. E. Brockerhoff, 2004 A zebrafish model for pyruvate dehydrogenase deficiency: rescue of neurological dysfunction and embryonic lethality using a ketogenic diet. Proc. Natl. Acad. Sci. USA 101: 4584–4589. Tsujikawa, M., and J. Malicki, 2004 Intraflagellar transport genes are essential for differentiation and survival of vertebrate sensory neurons. Neuron 42: 703–716. Vos, P., R. Hogers, M. Bleeker, M. Reijans, T. van de Lee et al., 1995 AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res. 23: 4407–4414. Westerfield, M., 1995 The Zebrafish Book. University of Oregon Press, Eugene, OR. Yarfitz, S., and J. B. Hurley, 1994 Transduction mechanisms of vertebrate and invertebrate photoreceptors. J. Biol. Chem. 269: 14329–14332. Yarfitz, S., G. A. Niemi, J. L. McConnell, C. L. Fitch and J. B. Hurley, 1991 A G beta protein in the Drosophila compound eye is different from that in the brain. Neuron 7: 429–438. Communicating editor: D. J. Grunwald