Caveolin-1alpha and -1beta perform nonredundant roles in early vertebrate development

Abstract

Caveolins are integral membrane proteins that localize predominantly to lipid rafts, where they oligomerize to form flask-shaped organelles termed caveolae and play important roles in membrane trafficking, signal transduction, and other cellular processes. To investigate potential roles for caveolin-1 (cav-1) in development, cav-1alpha and -1beta cDNAs were functionally characterized in the zebrafish. Cav-1alpha and -1beta mRNAs exhibited overlapping but distinct expression patterns throughout embryogenesis. Targeted depletion of either Cav-1 isoform, using antisense morpholino oligomers, resulted in a substantial loss of caveolae and dramatic neural, eye, and somite defects by 12 hours after fertilization, the time at which mRNA levels of both isoforms substantially increased in wild-type animals. Morphant phenotypes were rescued by injection of homotypic (cav-1alpha/cav-1alpha) but not heterotypic (cav-1alpha/cav-1beta) zebrafish and human cav-1 cRNAs, revealing nonredundant and evolutionarily conserved functions for the individual Cav-1 isoforms. Mutation of a known Cav-1 phosphorylation site unique to Cav-1alpha (Y14-->F) resulted in a failure to rescue the cav-1alpha morphant phenotype, verifying an essential role for Cav-1alpha specifically and implicating this residue in early developmental functions. Cav-1alpha and -1beta morphants also exhibited disruption in the actin cytoskeleton. These results support important and previously unanticipated roles for the Caveolin-1 isoforms in vertebrate organogenesis.

Figure 1

Cav-1α and -1β genomic organization, predicted amino acid sequences, and expression in cultured cells. A: Alignment of zebrafish cav-1α and -1β cDNAs with the zebrafish cav-1 genomic clone 66I19. Base pairs 244 to 1748 of the cav-1α cDNA, 76 to 1580 of the cav-1β cDNA, and 31,441 to 31,612 and 41,369 to 42,703 of the zebrafish genomic cav-1 clone exhibit 100% nucleotide sequence identity (triple-line boxes). Base pairs 1 to 243 of cav-1α and 1 to 75 of cav-1β are distinct from each other but are identical to cav-1 genomic clone bp 30,576 to 30,819 and 31,336 to 31,441, respectively. A single line indicates cav-1 intronic sequence; open boxes indicate unique sequences of individual cav-1 isoforms that are identical to the genomic cav-1 clone 66I19 nucleotide sequence; triple line boxed regions and arrows indicate identical sequences. The coding region of cav-1α consists of three exons, the boundaries of which lie between bp 243/244 and 412/413, whereas that of cav-1β is composed of two exons, the boundary of which falls between bp 244/245. ATG indicates the predicted initiation codon, and TAA indicates the predicted termination codon. B: Comparison of the predicted amino acid sequence of zebrafish Cav-1 to mouse and human Cav-1. The cav-1α cDNA encodes a protein of 181 amino acids. The cav-1β cDNA encodes a protein of 148 amino acids. Asterisks indicate amino acid identity, and hyphens indicate absence of amino acids. The arrowhead present at amino acid 34 indicates the amino terminus of the zebrafish Cav-1β isoform. Zebrafish Cav-1α exhibits 72% amino acid identity to human Cav-1α and 71% identity to mouse Cav-1α. Zebrafish Cav-1β exhibits 77% amino acid identity to human Cav-1β and 76% identity to mouse Cav-1β. C: Expression of zebrafish Cav-1α and -1β in human cells. Western analysis of LNCaP and HEK 293 cell lines transfected with vector alone (VO) or zebrafish cav-1α and -1β expression constructs. Zebrafish Cav-1α and -1β exhibited molecular masses of 20.6 and 16.8 kd, respectively, similar to mammalian Cav-1 isoforms.

Figure 2

Developmental expression of cav-1 isoform mRNAs and gene products. A: Developmental expression of cav-1 isoform mRNAs. RT-PCR indicated that both cav-1 mRNAs were weakly expressed as maternal mRNAs at 0.7 hpf, exhibited an increase in zygotic expression between 6 and 12 hpf, and were maintained at these levels throughout adulthood. NTC, no cDNA template control. B: Developmental expression of zebrafish Cav-1 protein isoforms. Protein extracts from developmentally staged and adult zebrafish were used to examine expression of endogenous Cav-1α and -1β isoforms by Western blot. Cav-1α was first detected as a faint band at 18 hpf and increased in intensity between 24 and 48 hpf. Cav-1β was first detectable as a faint band at 20 hpf and increased dramatically between 28 and 48 hpf. Both isoforms were highly expressed in the adult (adult protein extracts were diluted 1:20, relative to other extracts). Extracts from wt embryos injected with zebrafish cav-1α and -1β cRNAs served as positive controls. Ponceau S staining of the blot shows relative protein loading.

Figure 3

Whole mount in situ hybridization analysis of zebrafish cav-1α and -1β mRNAs in the embryo. At 24 hpf, a probe recognizing both cav-1 isoform mRNAs revealed ubiquitous cav-1 expression throughout the embryo, with distinct notochord and epidermal expression domains (A–C). At 48 hpf, cav-1 mRNAs were detected in skin (C) and in cranial and trunk neuromasts (E). At 90 hpf, cav-1 transcripts were detected in pharyngeal vasculature, intestinal epithelium, and skin (G and J, arrows). A probe specific for cav-1β mRNAs shows expression of this isoform in pharyngeal vasculature, heart, and skin but not in intestinal epithelium (H). Sense controls were negative (D, F, and I). h, heart; ie, intestinal epithelium; nc, notochord; nm, neuromasts; pv, pharyngeal vasculature.

Figure 4

Distinct phenotypes are exhibited by cav-1α and -1β morphant embryos. A–D: Mismatch control MO-injected embryo; E–H: Cav-1α morphant embryo; I–L: Cav-1β morphant embryo. A, inset: Western blot of 28-hpf Cav-1α and -1β depleted and wt control embryos. C, G, K are dorsal views, and the remaining panels present lateral views, anterior to the left, dorsal up. Cav-1α morphant embryos exhibited curved bodies and kinked tails, lacked distinct mhb, and exhibited RPE defects (E–G) compared with control embryos (A–C). F: The heart was enlarged in cav-1α morphants, and neural tissues appeared vacuolated. G: Dorsal view shows disorganized RPE and abnormal neural folds. H: Lateral tail view reveals disorganized somites and shortened tails. I: Cav-1β morphant embryos exhibited slightly curved bodies. J: Lateral head view reveals heart edema (arrow), eye pigmentation defects, and reduced neural tissues. K: Dorsal head view shows pigmentation defects and aberrant neural folding. L: Lateral tail view reveals notochord and somite defects, which appeared less severe than those of cav-1α morphants. e, eye; hb, hindbrain; mhb, midbrain hindbrain; s, somite; y, yolk. Scale bars = 200 μm.

Figure 5

Molecular characterization of 24-hpf cav-1α and -1β morphant phenotypes. The neural markers otx2 and pax2.1 and the muscle marker myoD were used to characterize cav-1α and -1β morphant phenotypes. The expression of neural markers was reduced and MyoD expression revealed aberrant somite patterning in Cav-1α and -1β morphants.

Figure 6

Rescue of morphant phenotypes by co-injection of isoform specific human cav-1 cRNAs. Human wild-type cav-1 isoform cRNAs and Y14F cav-1α mutant cRNAs were tested for their ability to rescue zebrafish cav-1 morphant embryo phenotypes. Representative rescue of cav-1β morphant phenotype by homotypic and not heterotypic human cav-1β cRNAs are shown (top), along with all of the experimental rescue results. Zebrafish cav-1α morphants were almost fully rescued by homotypic human cav-1α (98.6%) but not by hcav-1α-Y14F or heterotypic hcav-1β cRNAs. Likewise, zebrafish cav-1β morphants were rescued by homotypic human cav-1β cRNAs (96.6%) but not by heterotypic hcav-1α or hcav-1α Y14F cRNAs. Embryos injected with hcav-1α, hcav-1α-Y14F, or hcav-1β cRNAs alone appeared normal.

Figure 7

TEM analysis of caveolae in 24-hpf Cav-1α/-1β-depleted embryos. Numerous caveolae were detected in notochord tissues of wt zebrafish (A, arrows), whereas very few were observed in cav-1α/1β morphant notochord (B, arrow). C: Wt somite tissues exhibit organized myofibrils surrounded by distinct SR (arrows). D: Cav-1α/1β morphant somite tissues exhibited few myofibrils, and those were not attended by SR. Statistical analyses, using one-tailed (P = 0.005) and two-tailed (P = 0.01) Student’s t-test, confirmed that the differences in the frequency of caveolae were significant. Cropped images were generated from original magnifications of 54,000 (A), 52,000 (B), 73,500 (C), and 73,500 (D).

Figure 8

Disrupted actin cytoskeletal organization in 12- and 24-hpf cav-1α and -1β morphants. Confocal images of phalloidin-stained actin cytoskeletal filaments in head and tail epidermis of 12- and 24-hpf cav-1α and -1β morphants and mismatch control-injected embryos. At 12 hpf, actin filaments of head and tail epidermis of cav-1α and -1β morphants appeared disorganized, particularly in cav-1α morphants, in which extensive depolymerized and disrupted cytoskeletal architecture was evident by punctate staining at cell-cell junctions. Cell size varied extensively in cav-1α and -1β morphants, consistent with cytokinesis defects. At 24 hpf, distinct defects were evident in actin cytoskeletal organization in somite myofibrils. cav-1α morphants exhibited little to no identifiable myofibrils, whereas myofibrils in cav-1β morphant somite tissue appeared dysmorphic and disorganized, lacked midline myoseptum and chevron shape, and were shortened along the AP axis.

Figure 9

Cav-1α and -1β are required for proper patterning of the vascular endothelium. Lateral views of control (A, A′, B, and B′), cav-1α morphant (C, C′, D, and D′), and cav-1β morphant (E, E′, F, and F′) transgenic fli-EGFP embryos. Anterior is to the left, and dorsal is up. Control embryos exhibited green fluorescent protein expression in well-patterned head and tail vascular endothelial cells. In contrast, Cav-1α-depleted head and tail vascular endothelium appeared mispatterned and did not form patent vessels (C, C′, D, and D′). Head vasculature of Cav-1β-depleted embryos appeared disorganized, whereas tail vasculature was less affected (E, E′, F, and F′).