Duplicate VegfA genes and orthologues of the KDR receptor tyrosine kinase family mediate vascular development in the zebrafish

Abstract

Vascular endothelial growth factor A (VEGFA) and the type III receptor tyrosine kinase receptors (RTKs) are both required for the differentiation of endothelial cells (vasculogenesis) and for the sprouting of new capillaries (angiogenesis). We have isolated a duplicated zebrafish VegfA locus, termed VegfAb, and a duplicate RTK locus with homology to KDR/FLK1 (named Kdrb). Morpholino-disrupted VegfAb embryos develop a normal circulatory system until approximately 2 to 3 days after fertilization (dpf), when defects in angiogenesis permit blood to extravasate into many tissues. Unlike the VegfAa(121) and VegfAa(165) isoforms, the VegfAb isoforms VegfAb(171) and VegfAb(210) are not normally secreted when expressed in mammalian tissue culture cells. The Kdrb locus encodes a 1361-amino acid transmembrane receptor with strong homology to mammalian KDR. Combined knockdown of both RTKs leads to defects in vascular development, suggesting that they cooperate in mediating the vascular effects of VegfA in zebrafish development. Both VegfAa and VegfAb can individually bind and promote phosphorylation of both Flk1 (Kdra) and Kdrb proteins in vitro. Taken together, our data support a model in the zebrafish, in which duplicated VegfA and multiple type III RTKs mediate vascular development.

Figure 1

Evolutionary relationship between human and zebrafish Vegfa isoforms and receptors. Protein sequences were aligned using ClustalW (European Bioinformatics Institute, Cambridge, United Kingdom) with the complete open reading frames. Species are indicated as follows: zf, zebrafish; m, mouse; and h, human. (A) Alignment of VegfA proteins shows a closer relationship of the newly isolated gene to VEGFA than VEGFB, VEGFC, VEGFD, and PLGF. (B) Alignment of RTK sequences. The zebrafish gene (Kdra) formerly referred to as flk1 shows a closer relationship to hFLT1 and mFlt1 than to hKDR (hFLK1) and mKDR (mFlk1). (C) Radiation hybrid mapping of zebrafish kdra, flt1, and kdrb and synteny comparison. Zebrafish kdrb is located on chromosome 20 in a region syntenic to human KDR, nearby the zebrafish orthologues of kit and pdgfra, which map near human KDR on human chromosome 4q12. Zebrafish kdra is on chromosome 14, and flt1 near cdk8 and ipf1 in a region on chromosome 24, which is syntenic to the human 13q12 region where FLT1, CDK8, and IPF1 are located.

Figure 2

Expression and structure of zebrafish VegfA genes and receptors. (A) Zebrafish vegfA exon/intron boundaries were predicted by blasting against the zebrafish genome. The structural similarities between vegfAb₁₇₁, vegfAa₁₆₅, and human VEGFA₁₆₅ suggest that vegfAb₁₇₁ is an orthologue of human VEGFA₁₆₅ and similarly, that vegfAb₂₁₀ is orthologous to human VEGFA₁₈₉ and/or VEGFA₂₀₆. (B) Alignment of the cytoplasmic domain of zebrafish Flk1 and human and mouse KDR. Conservation of critical tyrosine residues (arrows) shows that the flk1 RTK described here (Kdrb) is more closely related to KDR than the previously described putative flk1 gene (referred to now as kdra). See “Cloning and analysis of VegfAb and Flk1” for a full description of the critical residues and their function.

Figure 3

Expression of the zebrafish VegfA and Kdra genes in the developing embryo. (A) Whole-mount in situ hybridization with probes to vegfAa₁₂₁, vegfAa₁₆₅, or vegfAb. Both vegfAa₁₂₁ and vegfAa₁₆₅ are diffusely expressed at 12 somites, whereas vegfAb is not appreciably expressed at this time point. At 24 hours after fertilization (hpf), both isoforms of vegfAa were broadly expressed, although vegfAa₁₆₅ was more clearly seen in the somites. vegfAa₁₆₅ is more highly expressed in the developing CNS than vegfAb, whereas in the lens vegfAb is more highly expressed, and both are expressed in the developing somites (). At 2 dpf, vegfAa₁₂₁ is expressed in the developing heart vasculature and pectoral fins. The aortic vasculature is also identified in vegfAb in situs, but in contrast, only vegfAb is expressed in the developing pronephros (◀). At 4 dpf, significant expression is restricted to vegfAb in the vasculature surrounding the eye. (B) Wild-type embryos were analyzed by whole-mount in situ hybridization with a probe to either kdrb (top panels) or kdra (bottom panels) at 16 somites (left panels), 2 dpf (middle panels), and 4 dpf (right panels). At 16 somites, expression is limited to the inner cell mass. By 2 dpf, the developing intersomitic vasculature expresses both kdrb and kdra. At 4 dpf, both genes are expressed in the developing subintestinal veins (SIVs) and in the remaining vasculature. See “In situ hybridization and photography” for image acquisition information.

Figure 4

Vascular defects in VegfAb morphants. (A) Photomicrograph of a 5-bp mismatch vegfAb morpholino–injected (4.5 ng) fli1-gfp embryo at 2 dpf demonstrates no vascular defects (only one shown, although no phenotype was seen with either). (B) Defects in the formation of the ISVs anteriorly are clearly visible beginning at 2 dpf. (C) Photomicrograph of a 5-bp mismatch vegfAb control morpholino–injected (4.5 ng) fli1-gfp embryo at 4 dpf for comparison. (D-F) Beginning at 3 dpf, the number of circulating RBCs gradually decreased in 63% (162/255) of the embryos injected with 4.5 ng of the vegfAb ATG morpholino and in 47% (74/156) with 9 ng of the vegfAb-75 (5′UTR) morpholino (day-4 embryos are shown). Injection of an equivalent amount of 5-bp mismatch morpholinos did not affect vasculogenesis, and coinjection of the active vegfAb-75 (5′UTR) morpholino together with 50 pg each of vegfAb₁₇₁ and vegfAb₂₁₀ RNA reduced the number of embryos with defects in vasculogenesis by more than 50% (8/32 vs 22/40 abnormal embryo). VegfAb morphants showed decreased intersegmental vessel number and size (, ◀ in panels D and F) and aberrant head vascular development. SIVs were severely reduced in number, size, and branching that was more pronounced anteriorly. (E) At 4 dpf, blood is apparent in the head and anterior embryos (), which (F) corresponds to the areas with defects in angiogenesis. (G) Control morpholino–injected embryos demonstrate normal subintestinal vein (SIV) architecture by alkaline phosphatase staining, and (H) RBCs are shown by staining with o-dianisidine that stains hemoglobin reddish/brown. However, (I) injection of either morpholino targeting VegfAb leads to SIVs that are erratically placed and thin or nearly completely absent (only the start codon morpholino is shown) and (J) extravasation of RBCs in various structures, which is more pronounced anteriorly where the angiogenic defects are most visible (B-F). The results are combined from at least 3 separate experiments, and the photomicrographs are representative of the visible defects.

Figure 5

Kdra and Kdrb cooperate to mediate vasculogenesis during zebrafish embryogenesis. Fli-gfp transgenic embryos were injected with combinations of morpholinos against kdra and kdrb or 5-bp mismatch control morpholinos. (A) Twenty-eight hpf embryos injected with either of 2 5-bp mismatch control morpholinos (4.5 ng each) corresponding to separate 5′UTR kdrb morpholinos (m1-kdrb [n = 32] and m2-kdrb [n = 44]) or the previously published kdra morpholino (m-kdra [n = 38]) had no vascular abnormalities. (B) Injection of 4.5 ng kdra or kdrb morpholino has no demonstrable effect on vasculogenesis at 28 hpf (107 embryos). Vascular defects were seen at higher concentrations of kdrb morpholinos; however, similar defects, but at a reduced penetrance, were also seen with equivalent amounts of the mismatched controls, suggesting these effects were not directly related to Kdrb function. (C) Coinjection of either 4.5 ng m1-kdrb or m2-kdrb with 4.5ng kdra morpholino caused variable loss of intersegmental arteries in 42 (24%) of 178 kdra/m1-kdrb or 53 (38%) of 138 kdra/m2-kdrb morpholino–injected embryos. (D) RBCs are seen in the anterior axial vasculature but cannot circulate. (E,F) Kdra/kdrb double morphant defects are still seen at 2 dpf. (G-I) At 4 dpf, injection of 4.5 ng kdra morpholino led to defects in the angiogenesis (◀, ) in 34 (33%) of 101 injected embryos. Embryos coinjected with morpholinos targeting both kdra/b had severe axial vessel defects, although some subintestinal vasculature is apparent. The results are combined from at least 4 separate experiments and the photomicrographs are representative of the visible defects.

Figure 6

VegfAb secretion and binding. Full-length vegfAa₁₂₁, vegfAa₁₆₅, vegfAb₁₇₁, and vegfAb₂₁₀ tagged with a V5 epitope at the N-termini were transfected into COS7 cells. (A) Western blot with a V5 antibody showed that although all 4 proteins are detectable in cell lysates (B) only VegfAa₁₂₁ and VegfAa₁₆₅ were secreted. (C) Transfected full-length kdra and kdrb fused to a V5 epitope were detectable in cell lysates of COS7 cells. (D) Replacing the endogenous signal peptide on all isoforms with the Igκ signal peptide resulted in secretion of both VegfAa and VegfAb. HEK 293 cells were transfected individually with Igκ-HA-vegfAa₁₆₅-Myc, Igκ-HA-vegfAb₁₇₁-Myc, soluble kdra-V5, and soluble kdrb-V5. Transfected cells were grown for 4 days, and conditioned media were collected, mixed, and incubated overnight. Immunoprecipitation (IP) with a V5 antibody (RTKs), blotting with Myc antibody (VegfAs), demonstrates that soluble Kdra bound to both VegfAa₁₆₅ and VegfAb₁₇₁; however, Kdrb did not appear to bind VegfAb₁₇₁ as well as VegfAa₁₆₅ in this assay. (E) Igκ-HA-vegfA-V5 constructs and Kdra-V5/pcDNA3.1 or Kdrb-V5/pcDNA3.1 were cotransfected into COS7 cells blotted with antiphosphotyrosine antibody, indicating activation of both RTKs by both VegfAa and VegfAb or (F) anti-V5 antibody demonstrating expression of the Kdra or Kdrb proteins.