N-Ethylmaleimide–Sensitive Factor b (nsfb) Is Required for Normal Pigmentation of the Zebrafish Retinal Pigment Epithelium

Abstract

Purpose: Despite the number of albinism-causing mutations identified in human patients and animal models, there remain a significant number of cases for which no mutation has been identified, suggesting that our understanding of melanogenesis is incomplete. Previously, we identified two oculocutaneous albinism mutations in zebrafish, au13 and au18. Here, we sought to identify the mutated loci and determine how the affected proteins contribute to normal pigmentation of the retinal pigment epithelium (RPE).

Methods: Complementation analyses revealed that au13 and au18 belonged to a single complementation group, suggesting that they affected the same locus. Whole-genome sequencing and single nucleotide polymorphism (SNP) analysis was performed to identify putative mutations, which were confirmed by cDNA sequencing and mRNA rescue. Transmission electron microscopy (TEM) and image quantification were used to identify the cellular basis of hypopigmentation.

Results: Whole-genome sequencing and SNP mapping identified a nonsense mutation in the N-ethylmaleimide–sensitive factor b (nsfb) gene in au18 mutants. Complementary DNA sequencing confirmed the presence of the mutation (C893T), which truncates the nsfb protein by roughly two-thirds (Y297X). No coding sequence mutations were identified in au13, but quantitative PCR revealed a significant decrease in nsfb expression, and nsfb mRNA injection rescued the hypopigmentation phenotype, suggesting a regulatory mutation. In situ hybridization revealed that nsfb is broadly expressed during embryonic development, including in the RPE. Transmission electron microscopy analyses indicated that average melanosome density and maturity were significantly decreased in nsfb mutants.

Conclusions: au18 and au13 contain mutations in nsfb, which encodes a protein that is required for the maturation of melanosomes in zebrafish RPE.

Albinism is a congenital disorder that affects roughly 1 in 17,000 people worldwide^1–3 and is characterized by a complete or partial deficiency of melanin. Ocular defects associated with albinism include nystagmus, foveal hypoplasia, low acuity, and photophobia.^4,5 Hypopigmentation can occur in the eyes and skin (oculocutaneous albinism [OCA]), eyes only (ocular albinism [OA]), or can be accompanied by more severe phenotypes in the relatively rare syndromic forms of albinism, such as Hermansky-Pudlak and Chediak-Higashi syndromes. In all known forms of albinism, hypopigmentation results from mutations in genes encoding proteins involved in melanin synthesis (melanogenesis)^4,6 or in melanosome transport (e.g., Refs. 7, 8). 

In the eye, albinism affects the retinal pigment epithelium (RPE), a layer of pigmented cells closely apposed to the photoreceptors at the back of the retina. Proper pigmentation of the RPE is critical for normal retinal development.⁹ Melanosomes are specialized lysosome-related organelles that contain pheomelanin and eumelanin pigments. Melanosomes in the RPE aid the visual system by absorbing stray photons to prevent light scatter and focus light onto the photoreceptors.¹⁰ Melanogenesis involves the trafficking of melanogenic enzymes to forming melanosomes, facilitating pigment synthesis and deposition, a process that occurs through several well-defined stages.^11–14 In mammals, stage I premelanosomes, still a part of the endocytic pathway, undergo membrane invaginations to form intralumenal vesicles and begin to develop a proteinaceous fibril meshwork. This meshwork becomes fully formed in stage II melanosomes. Melanin precursors and enzymes are trafficked to stage II melanosomes, and melanin aggregates on the surface of the meshwork, generating a stage III melanosome with characteristic pigmented striations. Melanin aggregation continues until the organelle is completely filled in and becomes a mature stage IV melanosome. This process is well conserved amongst vertebrates, the only notable difference being that, in fish, melanin aggregates on the surfaces of intralumenal vesicles instead of on the striated meshwork.¹⁵ 

Melanogenesis depends on proper vesicular targeting and membrane fusion. Indeed, mutations in the genes encoding a number of components of the endocytic pathway result in albinism.^6,16 Despite the number of albinism causing mutations identified in human patients and animal models, there remain a significant number of cases for which no mutation has yet been identified, suggesting that our understanding of the proteins involved in melanogenesis may be incomplete. Forward genetic screens in animal model systems can be very useful in this regard by utilizing a high-throughout and unbiased screening approach to identify mutations in a process of interest. Highlighting this utility, our laboratory previously reported a forward genetic screen that identified ∼20 mutations with defects in eye development.¹⁷ Of this collection, two mutants presented with oculocutaneous albinism, and in this study, we report the cloning and characterization of these mutations and identify them as mutant alleles affecting the N-ethylmaleimide–sensitive factor (nsfb) gene. 

N-ethylmaleimide–sensitive factor (Nsf) is a critical regulator of membrane fusion; NSF was initially identified in yeast as an enzyme required for protein transport between Golgi stacks^18,19 and has since been identified as a AAA superfamily (ATPase associated with various cellular activities) protein that consists of three domains: two ATPase domains (D1 and D2) and an N-terminal domain (NSF-N).^20–22 The NSF protomers form a hexamer, and this constitutes part of the soluble NSF attachment protein (SNAP) complex, which mediates vesicular fusion by disassembling SNARE (SNAP receptor) complexes. SNARE complexes form when SNARE proteins on different membranes interact and provide the requisite energy for membrane fusion by forming four-helix bundles and bringing the membranes into close apposition.²³ As a part of the SNAP complex, NSF couples hydrolysis of ATP with the disassembly of SNARE complexes following membrane fusion, freeing individual SNARE proteins to participate in future vesicular fusion events.^22,24,25 Zebrafish possess two nsf genes: nsfa and nsfb.²⁶ nsfa is required for myelination of axons,²⁷ the secretion of trophic factors during the maturation of hair cell synapses,²⁸ and neuropeptide-dependent visually mediated background adaptation.²⁶ In contrast, very little is known about nsfb. Here we demonstrate that nsfb mutants display hypopigmentation of the RPE, resulting from a defect in melanogenesis. 

Methods

Zebrafish Husbandry

Zebrafish were maintained at 28.5°C on a 14-hour light/10-hour dark cycle. Embryos were obtained from the natural spawning of mutant or wild-type parents in pairwise crosses. According to established protocols,²⁹ embryos were collected and raised at 28.5°C in the dark until they reached appropriate age for experimentation. All animals were treated in accordance with provisions established by the University of Texas at Austin Institutional Animal Care and Use Committee, and experimentation conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Bright field images of live embryos were captured with a Leica MZT 1600F stereomicroscope (Leica, Wetzlar, Germany). 

Single Nucleotide Polymorphism Mapping and Genotyping

Heterozygous au18 carriers in an AB background were outcrossed with wild-type Tübingen (TU) fish to generate a polymorphic mapping line. Two hundred homozygous mutant embryos from four parental mapping pairs were collected, genomic DNA was isolated (DNeasy Blood & Tissue Kit; Qiagen, Valencia, CA, USA), and 1 μg was utilized in whole-genome sequencing at the University of Texas Genomic Sequencing and Analysis Facility. On an Illumina HiSeq 2000 machine (Illumina, San Diego, CA, USA), 82 million paired-end 100-bp sequences were generated for an average genome coverage of 10.9×. To identify putative mutations, sequencing reads were analyzed via the MegaMapper single nucleotide polymorphism (SNP)-mapping pipeline.³⁰ Complementary DNA sequencing was performed to confirm the putative nsfb mutation. To genotype nsfb^c893t, we utilized dCAPS Finder 2.0³¹ to analyze the genomic sequence of nsfb proximal to the SNP and generate the following primers: forward, 5′-agattttgaacaaatacgtgggaggat-3′ and reverse, 5′-tgactgttgacagaggaacc-3′. These primers amplify a 260-bp product with a MboI restriction site present in wild-type genomic DNA (gDNA), but not in nsfb^c893t. Amplicons from gDNA generated from individual embryos were digested with MboI and run on a 3% agarose gel to distinguish wild-type (260 bp) from nsfb^c893t (234 bp) bands. 

In Situ Hybridization

In situ hybridizations were performed as described.^32,33 For nsfb probe synthesis, a partial cDNA fragment of the coding sequence of nsfb was cloned by RT-PCR using the followng primers: 5′-agagaccgccaaatgtcagg-3′ and 5-ctaaaaagccacatgtttatcagggc-3′. 

Phylogenetic Analysis

Protein sequences of nsf were obtained from Ensembl release 80 (http://www.ensembl.org, in the public domain),³⁴ the Saccharomyces genome database (SGD project, http://www.yeastgenome.org, in the public domain),³⁵ WormBase (http://www.wormbase.org, release WS248, in the public domain),³⁶ and FlyBase (http://flybase.org, in the public domain).³⁷ These sequences are yeast nsf (SEC18 and YBR080C), Drosophila Nsf1 (FBgn0000346) and Nsf2 (FBgn0266464), C. elegans NSF (WBGene00003818), Xenopus NSF (ENSXETG00000031364), Zebrafish nsfa (ENSDARG00000007654) and nsfb (ENSDARG00000038991), mouse nsf (ENSMUSG00000034187), chimpanzee nsf (ENSPTRG00000009320), and human NSF (NSG00000073969), with yeast AAA protein ORC1 as an outgroup (YML065W). A phylogenetic tree was generated via Geneious alignment (Geneious version 7.1.5, http://www.geneious.com, in the public domain)³⁸ and bootstrapped 1000 times for statistical analysis. 

Quantitative PCR

Quantitative PCR was performed on wild-type, au18, and au13 embryos at 48 hours post fertilization. Ten embryos were collected per each biological replicate (n = 3), mRNA was extracted with Trizol Reagent (Life Technologies, Carlsbad, CA, USA) according to the manufacturer's protocol, and potential genomic DNA contaminants were digested with RNAse-free DNAse I (Roche Diagnostics, Mannheim, Germany) before being reverse transcribed into cDNA with the iScript Reverse Transcriptase Kit (Bio-Rad, Hercules, CA, USA). Real-time quantitative PCR was conducted on five technical replicates per biological replicate using SYBR Green and a Via 7 Real-Time PCR System (Life Technologies). 

act2b was used as the reference gene, using primers provided as a generous gift by Johann Eberhart (University of Texas at Austin, Austin, TX, USA): forward, 5′-cgagcaggagatgggaacc-3′ and reverse, 5′-caacggaaacgctcattgc-3′, which produce an amplicon of 128 bp. Primers for nsfb were as follows: forward, 5′-ggagctttcttggaggctttca-3′ and reverse, 5′-cgctctgtcctttcacctgtt-3′, which amplify 68 bp.²⁶ Fold change in nsfb transcript expression among au18, au13, and WT was calculated using the relative gene quantification method.³⁹ 

Histology

Embryos were fixed in fresh 4% glutaraldehyde and 2% paraformaldehyde overnight at room temperature, stained in 2% OsO₄, 2% potassium ferrocyanide, and 2% uranyl acetate, and microwave embedded with a modified reduced viscosity Spurr-Quetol-651 resin using a BDMA accelerator (Electron Microscopy Services, Hatfield, PA, USA) via a 30%, 50%, 75%, 100%, and 100% resin/acetone infiltration series. Samples were sectioned using a Leica Ultracut UC7 ultramicrotome at a thickness of 1 μm, stained with 1% toluidine blue and 1% borax, and imaged on a Leica DM2500 at 400× and 1000× magnification. 

Transmission Electron Microscopy

Samples were embedded as above, and cut on a Leica Ultracut UC7 ultramicrotome at a thickness of 70 nm. These sections were imaged on a FEI Tecnai transmission electron microscope (FEI, Hillsboro, OR, USA). Reference images were acquired at 390× magnification to locate the optic nerve and ultrastructure of the RPE. Images (n = 3) of the RPE were collected from three regions of the RPE, dorsal, central, and ventral, from n = 3 fish at 11,500× magnification. Central RPE was defined as the RPE contained within the region 10° to 30° clockwise from a line connecting the middle of the optic nerve head and the lens core. Dorsal RPE was defined as the region included between 80° and 100° clockwise and ventral RPE between 30° and 40° counterclockwise. Images were rotated to orient the RPE on a horizontal line, and the number, size, and maturity of melanosomes were analyzed using ImageJ software (http://imagej.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA). 

Statistical Analysis

Levene's test found that the homoscedasticity assumption is not satisfied in any dataset. Therefore, to determine whether there was a significant difference between sibling and mutant data during each time point, as well as the change over time of each variable within an individual group, the nonparametric Kruskal-Wallis test was applied, and the Mann-Whitney U test was performed for pairwise comparisons. Analyses were completed using GraphPad Prism version 6.0f for Mac (GraphPad Software, San Diego, CA, USA; www.graphpad.com, in the public domain). 

Results

au13 and au18 Mutants Display Oculocutaneous Albinism

au13 and au18 mutants were isolated from an N-nitroso-N-ethylurea mutagenesis screen for recessive mutations affecting eye development in zebrafish.¹⁷ Both au13 and au18 mutations are fully penetrant and strikingly similar in phenotype (Fig. 1); au13 fails to complement the au18 mutation (data not shown), suggesting that the mutations affect the same gene. Phenotypically, hypopigmentation is noticeable in mutant embryos by 48 hpf (Figs. 1A, 1C), and the degree of hypopigmentation varies in severity, with mild and severe mutants present in the same clutch of embryos. Both severity classes are globally hypopigmented; in mild mutants, small areas of the RPE are completely devoid of pigment, whereas others appear largely normal. Meanwhile, RPE in severely affected mutants contains numerous large regions lacking pigmentation. By 72 hpf, severely affected mutants begin to display early signs of degeneration, and this includes brain necrosis and a slightly curved body axis. 

Figure 1

 

au18 and au13 mutants are oculocutaneous albinos. (A, B) Lateral and dorsal whole embryo images of phenotypically wild-type sibling and mild and severe au18 mutants at 48hpf and 72hpf. (C, D) Images of phenotypically wild-type sibling and mild and severe au13 mutants at 48hpf and 72hpf. In both lines, severe mutants display large regions devoid of pigment in the RPE.

Given the likely allelic relationship between au13 and au18 mutants, and their phenotypic similarity, we focused our efforts on further characterizing the eye phenotype of au18 mutants at 24, 48, 58, and 72 hpf. au18 mutant embryos are histologically indistinguishable from wild-type siblings at 24 hpf (Supplementary Fig. S1). At 48 hpf, au18 retinas are slightly microphthalmic and appear to have fewer melanosomes in their RPE compared with wild-type siblings (Figs. 2A–F). This difference in pigmentation becomes more pronounced over time; by 72 hpf, the mutant RPE contains markedly fewer melanosomes than that of wild-type siblings (Figs. 2M–R). Beyond RPE hypopigmentation and mild microphthalmia, some of the most severely affected mutants possessed lens defects where the posterior of the lens was epithelialized (Figs. 2L, 2R). All other aspects of eye development appeared to be relatively normal in both mutants. 

Figure 2

 

The RPE is disrupted in au18 mutants. Images of sibling and mutant embryos show that the RPE of au18 embryos is hypopigmented at 48 (A, C, E), 58 (G, I, K), and 72 hpf (M, O, Q). Transverse histologic sections of sibling and mutant embryos at 48 (B, D, F) 58 (H, J, L), and 78 hpf (N, P, R) reveal that au18 eye is microphthalmic, but overall retina structure appears to be normal. Areas devoid of pigment are marked by white arrowheads. Scale bar: 20 μm. (B′, D′, F′, H′, J′, L′, N′, P′, R′) Magnified images of RPE show that RPE proximal to the optic nerve was consistently depigmented in au18 and confirm that au18 RPE appears to have fewer melanosomes than wild-type siblings. The posterior lens also appears to be cataractous in severe au18 mutants (F, L, R). Scale bar: 10 μm.

au13 and au18 Phenotypes Are Caused by Mutations in nsfb

To identify the causative mutation in au18, we performed next-generation sequencing followed by SNP mapping using the MegaMapper analysis pipeline.³⁰ Whole-genome sequencing of au18 mutants generated ∼11× average coverage of the zebrafish genome. MegaMapper analysis identified a 6-Mb window on chromosome 12 as the likely location of the causative mutation. Within this window, a C-to-A transversion mutation at nucleotide 893 (C893T) of the nsfb coding sequence was predicted to be the mutation underlying the au18 phenotype. This mutation is predicted to cause a premature stop codon at amino acid 297, truncating nsfb by 60% (S297X). Complementary DNA sequencing confirmed the presence of this mutation in homozygous au18 mutant embryos and its absence in wild-type embryos (Fig. 3A). Nsfb contains three major domains: an N-terminal binding domain required for its association with SNAP-complex proteins and two ATPase domains, D1 and D2 (Fig. 3A). The first ATPase domain is required for overall ATPase activity of the protein, whereas the second ATPase mediates hexamerization during complex formation.^22,40,41 The early stop codon in au18 is predicted to truncate the protein within the first ATPase domain (red dotted line, Fig. 3A), likely resulting in nonfunctional nsfb protein. To confirm that this mutation is causative of the au18 phenotype, we performed an mRNA rescue experiment by injecting full-length nsfb^WT and nsfb^au18 mRNA into homozygous mutant au18 embryos. Injection of nsfb^WT rescued pigmentation in embryos derived from heterozygous au18 crosses. (131 of 132), whereas nsfb^au18 failed to rescue the mutant phenotype (74 of 94; Figs. 3B, 3C). Phenotypic rescue of homozygous nsfb^au18 mutants was confirmed by genotyping in a subset of injected embryos (data not shown).⁴² Based on these data, we conclude that au18 possesses a nonsense mutation in nsfb and hereafter refer to this allele as nsfb^au18 in accordance with zebrafish nomenclature guidelines. 

Figure 3

 

au13 and au18 phenotypes are likely caused by mutations in nsfb. (A) Schematic of the nsfb protein. There are three major domains: the Nsf-N, ATPase domain 1 (D1), and ATPase domain 2 (D2). The mutation occurs at nucleotide 893 (C893T), and the dotted red line indicates the location of the premature stop codon at amino acid 297 (S297X) in au18 mutants. This mutation is predicted to truncate the protein by 60%. (B) Graph illustrating the percentage of phenotypically mutant embryos in each clutch following injection of vehicle, nsfb^WT, and nsfb^au18 mRNA. Both au13 and au18 mutant phenotypes are rescued by injection of nsfb^WT, whereas injection of nsfb^au18 mRNA failed to rescue either the au13 or au18 mutant phenotype. (C) Representative images of WT, au18^+/−, and au18^−/− embryos following nsfb^WT injection. (D) Quantitative PCR quantification of the relative fold change of nsfb in nsfb^au13 and nsfb^au18 embryos at 48 hpf. Transcript levels were normalized to act2b. Error bars indicate SEM of five technical replicates for each experiment. n = 3 biological replicates.

As mentioned above, au13 fails to complement the nsfb^au18 mutation, suggesting that they possess mutations in the same gene. However, we were unable to identify a causative mutation in the coding sequence of nsfb in au13 mutants (data not shown). Injection of nsfb^WT mRNA rescued the mutant phenotype in embryos from an au13 heterozygous incross (145 of 148), whereas injection of nsfb^au18 failed to rescue the pigmentation phenotype in these clutches (46 of 60). Given the noncomplementation between nsfb^au18 mutants and au13 mutants, and these rescue data, the most parsimonious explanation is that au13 possesses a mutation in a regulatory region of nsfb that alters transcript expression. To test this possibility, we performed quantitative PCR analysis of nsfb in au13 and nsfb^au18 mutants, which revealed that nsfb expression was reduced ∼2.4-fold in both nsfb^au18 and au13 relative to wild-type embryos (Fig. 3D). Thus, we conclude that the au13 phenotype also results from a mutation in nsfb, and hereafter refer to this line as nsfb^au13. 

nsfb Is a Ubiquitously Expressed Paralog of Human NSF

The zebrafish genome contains two paralogs of human NSF, nsfa and nsfb, which likely arose from a genome duplication event in teleosts.⁴³ Phylogenetic analysis of nsf genes suggests that nsfb is more distantly related to human NSF than nsfa (Fig. 4A). The nsfa gene is primarily expressed in the nervous system,²⁷ but the expression pattern of nsfb is unknown. In situ hybridization revealed widespread expression, with apparently higher levels of expression in the somites at 24 hpf. At 48 hpf, nsfb mRNA expression appears to be more restricted to the anterior half of the embryo, with some enrichment in the otic vesicle and pectoral fins (Fig. 4B). 

Figure 4

 

nsfb is a paralog of human NSF and is ubiquitously expressed during early embryo development. (A) Phylogenetic analysis of the relationship between nsf in vertebrates. This analysis was constructed using Geneious Tree Builder based on an alignment of nsf amino acid sequences from human, chimpanzee, mouse, Xenopus, zebrafish, Drosophila, C. elegans, and yeast, using yeast Origin recognition complex subunit 1 (ORC1) as an outgroup. Numbers at nodes represent bootstrap values after 1000 replicates. (B) In situ hybridization reveals that nsfb expression is widespread at 24 hours after fertilization, with higher levels of staining in the somites and nervous system. At 48 hpf, nsfb remains widely expressed in the head, with high levels of expression in the otic vesicle and pectoral fins, but is less apparent in the somites and posterior trunk. Lateral and dorsal views of embryos stained for antisense probe are on top, and sense controls on bottom. Scale bar: 200 μm.

Melanosome Formation and Maturation Are Disrupted in nsfb^au18 Mutants

Due to the importance of membrane trafficking and vesicular fusion during melanogenesis, we hypothesized that loss of nsfb function results in an arrest in melanosome maturation in nsfb^au18 RPE. To test this hypothesis and assess melanogenesis in the nsfb^au18 RPE, we performed transmission electron microscopy (TEM) of nsfb^au18 mutant and sibling RPE at 48, 58, and 72 hpf, quantifying several aspects of melanosome formation from TEM images taken from the dorsal, central, and ventral regions of the RPE (Figs. 5A, 5B). Because hypopigmentation varies in nsfb^au18 mutants, we continued to separate nsfb^au18 embryos into mildly and severely hypopigmented cohorts to determine whether there were quantitative differences in phenotypes between these groups. Quantification of melanosome density revealed that density increased between 58 and 72 hpf in phenotypically wild-type sibling embryos (P = 0.001). The average density of nsfb^au18 mutants was significantly lower than sibling at all time points and did not display a significant change over time (Fig. 5C). Moreover, severely hypopigmented nsfb^au18 mutants contained significantly fewer melanosomes than mild nsfb^au18 mutants at 48 and 72 hpf (P < 0.0001). 

Figure 5

 

Melanosome formation and maturation is disrupted in nsfb^au18 mutants. (A) Cartoon schematic showing the dorsal, central, and ventral regions of the RPE that were analyzed. RPE between 20° and 40° clockwise from an imaginary line (dotted line) drawn between the center of the lens and the center of the optic nerve was defined as central, whereas RPE 80° to 100° was defined as dorsal. Finally, RPE contained within a box 30° to 50° counterclockwise was defined as ventral. (B) Representative TEM images of central RPE for sibling, mild, and severe nsfb^au18 RPE at 48, 58, and 72 hours after fertilization. Images oriented distal (choroid) up, proximal (retina) down. Scale bar: 1 μm. (C) Quantification of the average number of melanosomes per square micrometer of RPE indicates that nsfb^au18 mutants contained significantly fewer melanosomes than sibling at all time points, and severe nsfb^au18 mutants contained significantly fewer melanosomes than mild nsfb^au18 mutants at 48 and 72 hpf. (D) The graph depicting the average percentage of melanosomes containing pigment reveals that nsfb^au18 melanosomes were significantly less mature than sibling at all time points and that severe mutants were less mature than mild. (E) Graph depicting the average perimeter of melanosomes. Although the average perimeter of melanosomes in sibling and mild nsfb^au18 mutants was not significantly different, severe nsfb^au18 melanosomes were significantly smaller than sibling. Furthermore, although melanosomal perimeter increased in sibling and mild RPE, severely affected mutant melanosomes decreased over time. (Mann-Whitney U test, *P < 0.01, **P < 0.001, ***P < 0.0001).

Because melanosomes accumulate more pigment as they mature,¹¹ the percentage of each melanosome containing pigment was quantified, and this measure was used as a proxy for melanosome maturity. These analyses revealed that nsfb^au18 melanosomes were significantly less mature than phenotypically wild-type siblings at all time points examined (Fig. 5D), and melanosomes in severe nsfb^au18 mutants were less mature than those in mild nsfb^au18 mutants at 58 and 72 hpf. While phenotypically wild-type melanosomes grew more mature over all time points (P < 0.0001), overall melanosomal maturity decreased between 48 and 58 hpf in mildly affected (P = 0.0004) and severely affected embryos (P < 0.0001). 

Finally, quantification of the melanosomal perimeter (Fig. 5E) revealed that the average perimeter of melanosomes in phenotypically wild-type siblings did not significantly differ from mild nsfb^au18 mutants. Melanosomal perimeter also increased significantly in sibling and mildly affected mutants between 48 and 58 hpf (P < 0.0001). However, the perimeter of melanosomes in severe nsfb^au18 mutants was significantly reduced compared with sibling at 58 and 72 hpf (P < 0.0001). Also, unlike in phenotypically wild-type and mildly affected embryos, the melanosomal perimeter decreased significantly between 48 and 58 hpf in severely affected mutants (P < 0.0001; Fig. 5E). 

Discussion

Albinism results from mutations in a diverse collection of genes that regulate melanogenesis or melanosome trafficking⁴⁴; however, data from human and animal studies suggest that we do not yet know the full complement of genes required for normal pigmentation. For example, although seven genetic subtypes of OCA have thus far been identified, OCA1A and OCA1B (Mendelian Inheritance in Man [MIM] 203100, 606952), OCA2 (MIM 203200), OCA3 (MIM 20390), OCA4 (MIM 606574), OCA5 (615312), OCA6 (MIM 113750), and OCA7 (MIM 615179), the genes underlying OCA5-7 were discovered only very recently,^45,46 highlighting the probability that there are other undiscovered loci responsible for albinism. Here we report two novel zebrafish oculocutaneous albino mutants and establish a role for nsfb in melanogenesis. 

Patients with albinism disorders often present with varying degrees of ocular defects, and these include foveal hypoplasia, nystagmus, and improper retinal axon decussation.^4,5 Despite our growing understanding of the genetic etiology of OA, OCA, and the syndromic types of albinism, we know comparatively less about the molecular and cellular underpinnings of these diseases and how pigmentation affects the visual system. Zebrafish are a powerful model organism for gene discovery and mechanistic analysis of protein function during ocular development and disease due to tractable genetics and the ability to perform a diverse array of molecular, biochemical, anatomic, and behavioral studies.^47,48 Many pigmentation mutants have been identified and characterized in zebrafish,^49–58 some of which directly model existing human pathologies,^59,60 while others presage the identification of human mutations. For example, the causative mutation of OCA6, SLC24A5, was identified in a zebrafish pigmentation mutant⁶¹ prior to in the mouse⁶² or in human patients.⁶³ Thus, the zebrafish system represents a potent tool for studying the molecular and cellular regulation of melanogenesis, as well as identifying novel albinism genes in vertebrates. 

Here, we report the cloning and characterization of two novel oculocutaneous albino mutants that result from mutations in nsfb: nsfb^au18, which results from a nonsense mutation, and nsfb^au13, whose causative mutation is unknown but is likely located in a noncoding regulatory region. Indeed, quantitative PCR analysis demonstrated that nsfb expression is reduced in nsfb^au13 embryos, supporting this possibility. Interestingly, a comparable reduction of nsfb expression was also observed in nsfb^au18 mutant embryos. It is likely that this occurs via nonsense-mediated mRNA decay, a molecular phenotype frequently observed in zebrafish for genes that possess nonsense mutations.^29,61,64,65 

nsfb mutants are oculocutaneous albinos and they are first identifiable at ∼2 dpf based on their hypopigmentation phenotype. nsfb is one of the two zebrafish nsf paralogs; Nsf is ubiquitously expressed in mammals,^66,67 although enriched in the nervous system.⁶⁸ nsfa expression is limited to the nervous system,²⁷ whereas nsfb has more widespread expression (our data), suggesting that it may be more orthologous in function to nsf than nsfa. Interestingly, however, our phylogenetic analysis indicates the protein sequence of nsfa is more closely related to human NSF than is nsfb (Fig. 4). Thus, although nsfb diverged evolutionarily from mammalian Nsf, it possibly retained its system-wide functions, whereas nsfa evolved to hold more limited or specialized functions. 

Unlike nsfa, virtually nothing is known about the in vivo function(s) of nsfb. Here we demonstrate a critical requirement for nsfb in pigmentation and in regulating the maturation of melanosomes in the RPE. Defects in melanosomal maturation are accompanied by reductions in melanosome number (Fig. 5). Although phenotypes in both nsfb alleles were fully penetrant, there were differences in the expressivity of the pigmentation phenotype, with mildly and severely affected mutants detected within the same clutch of homozygous embryos. These data suggest that, although melanogenesis is disrupted in both classes of mutants, it proceeds to a limited extent in mildly affected mutants, whereas it is significantly impaired in severely affected mutants. It is common in zebrafish to have variability of a phenotype,^69–71 and often this variability is ascribed to differing levels of maternal rescue in each embryo resulting from the deposition of mRNAs and proteins into the developing oocyte.⁷² More work is necessary to understand the molecular basis of this phenotypic variability in nsfb mutants, and, more generally, how maternal rescue contributes to phenotypic variability in recessive zebrafish mutants. 

Our quantitative TEM data suggest a model in which nsfb activity is required for the maturation of melanosomes in the RPE. Although we limited our characterization to the RPE hypopigmentation phenotype, nsfb mutants possess widespread CNS degeneration by 3 dpf, have reduced circulation, and typically die by 4 dpf. Given the expanded and systemic severity of the phenotypes over those found in ocular albinism or oculocutaneous albinism patients, the nsfb phenotypes are more similar to those pathologies observed in patients with syndromic albinism. Indeed, patients with Hermansky Pudlak Syndrome or Chediak-Higashi Syndrome often have additional, systemic defects such as pulmonary fibrosis and blood clotting defects, in addition to their albinism. Although mutations in NSF have not yet been linked to albinism in humans, misregulation of NSF is linked to psychiatric diseases.⁷³ It will be interesting to screen through existing albinism patients for whom a causative mutation has not yet been identified, particularly those with more syndromic pathologies, to determine whether mutations in NSF underlie their phenotypes. In summary, this study identifies a role for nsfb during melanogenesis in the RPE, establishing an additional and novel oculocutaneous albinism mutant in zebrafish through which the molecular and cellular underpinnings of melanogenesis can be further elucidated. 

Acknowledgments

The authors thank Dwight Romanowicz and Trent Reznor for invaluable assistance with TEM, Jacqueline Norrie for assistance with qPCR, Ryoko Minowa for fish care, Etta James for technical assistance, and members of the Gross Laboratory for helpful discussions throughout the course of this work. 

Supported by National Institutes of Health Grant R21-EY22700 and CORE Grant P30 EY008098, the Alcon Research Institute, Eye and Ear Foundation of Pittsburgh (Pittsburgh, PA, USA), and an Unrestricted Grant from Research to Prevent Blindness (New York, NY, USA). 

Disclosure: N.J. Hanovice, None; C.M.S. Daly, None; J.M. Gross, None 

Footnotes

 Open Access Article: no