Sparc protein is required for normal growth of zebrafish otoliths

Abstract

Otoliths and the homologous otoconia in the inner ear are essential for balance. Their morphogenesis is less understood than that of other biominerals, such as bone, and only a small number of their constituent proteins have been characterized. As a novel approach to identify unknown otolith proteins, we employed shotgun proteomics to analyze crude extracts from trout and catfish otoliths. We found three proteins that had not been associated previously with otolith or otoconia formation: 'Secreted acidic cysteine rich glycoprotein' (Sparc), an important bone protein that binds collagen and Ca(2+); precerebellin-like protein, which contains a C1q domain and may associate with the collagenous otolin-1 during its assembly into a framework; and neuroserpin, a serine protease inhibitor that may regulate local protease activity during framework assembly. We then used the zebrafish to investigate whether Sparc plays a role in otolith morphogenesis. Immunodetection demonstrated that Sparc is a true constituent of otoliths. Knockdown of Sparc expression in morphant zebrafish resulted in four principal types of defective otoliths: smaller, extra and ectopic, missing and fused, or completely absent. Smaller size was the predominant phenotype and independent of the severity of otic-vesicle defects. These results suggested that Sparc is directly required for normal otolith growth.

FIG. 1.

Similarity and moderate complexity of protein complements in actinopterygian otoliths. a Otolith proteins of commercial and laboratory freshwater fishes separated by gel electrophoresis and stained with silver. Aliquots corresponding to about 1.2 mg otoliths per lane; apparent molecular masses in kDa indicated on the left. b Phylogeny of fish species mentioned in this study: White sturgeon (Acipenser transmontanus), zebrafish (Danio rerio), bighead carp (Aristichthys nobilis), channel catfish (Ictalurus punctatus), black ghost (Apteronotus albifrons), rainbow trout (Oncorhynchus mykiss), chum salmon (Oncorhynchus keta), tilapia (Oreochromis sp.), japanese medaka (Oryzias latipes), torafugu (Takifugu rubripes), spotted green pufferfish (Tetraodon nigroviridis), and three-spined stickleback (Gasterosteus aculeatus). Dotted lines in the cladogram indicate five or six actinopterygian clades each that were not represented. At three branch points, the estimated time since the last common ancestor is shown (Kumar and Hedges ; Wittbrodt et al. 2002).

FIG. 2.

Localization of zebrafish Sparc mRNA in the ear and protein in otoliths. a The mRNAs of sparc and β-actin detected in whole embryos during early development by RT-PCR. b–d The mRNA of sparc detected in the developing ear by in situ hybridization. Dorsal views of hindbrain area, anterior at the top; outline in b, left otic placode; arrowheads in d, posterior maculae along the medial wall of the otic vesicle; arrows in d, semicircular canals; scale bar in b for all three panels, 50 μm. e Sparc protein detected in an adult-otolith extract by immunoblotting. About 0.6 μg total protein per membrane strip; Silver, silver-stained gel slice; Sparc, anti-Sparc serum UI-37; E. coli-adsorbed, UI-37 pre-adsorbed with a plain bacterial extract; Sparc-adsorbed, UI-37 pre-adsorbed with a bacterial extract containing recombinant zebrafish Sparc; dashes on the right, molecular-mass markers (kDa, from top) 250, 160, 105, 75, 50, 35, 30, 25, 15, 10. f Sparc protein detected in otic vesicles by whole-mount immunofluorescence. Green, Sparc labeled with antiserum UI-37; blue, nuclei labeled with DAPI; arrowheads, lapillus (left) and sagitta (right); lateral view of otic vesicle, dorsal at the top, anterior to the left; projections of 10-μm stacks of optical sections; scale bar for panels f and g, 25 μm. g Same as f, but with purified naive rabbit IgG instead of UI-37. h–i Sparc protein detected in a transverse section of an adult lapillus. h brightfield view of an otolith slice; i immunofluorescence with antiserum UI-37 at the surface of the slice’s cross-section; scale bar in h for panels h to m, 100 μm. (j–k) Same as h–i, but UI-37 pre-adsorbed with a plain bacterial extract. l–m Same as h–i, but UI-37 pre-adsorbed with a bacterial extract containing recombinant zebrafish Sparc. Because the brightfield images in h, j, and l were overexposed to visualize the growth rings, the visible otolith margins do not exactly match the fluorescent areas; i, k, and m are projections of 58 to 87-μm-thick stacks of optical sections that were acquired under matched settings.

FIG. 3.

Diminished Sparc expression in morphant zebrafish embryos. a Schematic of the ATG-MO morpholino binding to the start codon of sparc mRNA and the e5i5-MO morpholino binding to the junction of exon 5 and intron 5. GT_a–c splice donor sites; AG splice acceptor site; For and Rev PCR primers used in b. b Misspliced sparc mRNAs in e5i5-MO morphants at 28 hpf detected by RT-PCR. Injected reagents at the top; size markers in bp on the left; products derived from splicing at donor sites GT_a–c or the unspliced pre-mRNA indicated on the right. c Example of protein detection in microinjected embryos at 52 hpf by two-color immunofluorescence blotting with antibodies against Sparc (red) and β-tubulin (green). Total protein pooled from ten embryos each; equivalent of one embryo loaded per lane. d Summary of Sparc levels measured by immunofluorescence blotting. One set of injections each for ATG-MO and e5i5-MO; mean and range indicated by error bar for vehicle and Std-MO controls.

FIG. 4.

Otolith abnormalities in zebrafish Sparc morphants. a–d Phenotypes of otoliths in microinjected embryos under brightfield illumination. Lateral views of otic vesicles, dorsal at the top, anterior to the left; arrowheads, lapillus (left) and sagitta (right) in a and extra/ectopic otolith in c; scale bar for all four panels in a, 30 μm; some otoliths were imaged in a slightly different focal plane. e–i Scanning electron micrographs of dissected otoliths from injected embryos. Scale bar for all five panels in e, 10 μm. j–k Structure of wild-type and morphant otoliths visualized by staining with toluidine blue. Coronal semi-thin sections of anterior maculae and lapilli; scale bar for both panels in k, 10 μm. l Summary of otolith areas measured in bright field images. Two-way ANOVA of three independent experiments for each morpholino; N = 94–107 per age and treatment for vehicle and Std-MO, 49-57 otherwise; error bars, 95% confidence intervals; *, significantly different in Bonferroni post-hoc pairwise comparisons from all other groups of the same age at P ≤ 10⁻⁴ and 403–418 df. m–n Hair bundles in wild-type and morphant anterior maculae visualized by labeling with fluorescent phalloidin. Dorsal views; projections of 17-μm stacks of optical sections; scale bar for both panels in n, 10 μm; 27 bundles in m, 26 in n. o Smaller otoliths in morphants regardless of otic-vesicle size. Only ears with both lapillus and sagitta, but no extra otoliths; r² of least-squares-fitted lines was 22%, 36%, and 39%, respectively, for Vehicle, ATG-MO, and e5i5-MO.