A versatile gene trap to visualize and interrogate the function of the vertebrate proteome

Abstract

We report a multifunctional gene-trapping approach, which generates full-length Citrine fusions with endogenous proteins and conditional mutants from a single integration event of the FlipTrap vector. We identified 170 FlipTrap zebrafish lines with diverse tissue-specific expression patterns and distinct subcellular localizations of fusion proteins generated by the integration of an internal citrine exon. Cre-mediated conditional mutagenesis is enabled by heterotypic lox sites that delete Citrine and "flip" in its place mCherry with a polyadenylation signal, resulting in a truncated fusion protein. Inducing recombination with Cerulean-Cre results in fusion proteins that often mislocalize, exhibit mutant phenotypes, and dramatically knock down wild-type transcript levels. FRT sites in the vector enable targeted genetic manipulation of the trapped loci in the presence of Flp recombinase. Thus, the FlipTrap captures the functional proteome, enabling the visualization of full-length fluorescent fusion proteins and interrogation of function by conditional mutagenesis and targeted genetic manipulation.

Figure 1.

Visualizing the proteome by fluorescent tagging of endogenous full-length proteins. (A) The FlipTrap vector consists of a citrine coding sequence (green) flanked by intronic sequences that contain a splice acceptor (SA, beige) and donor (SD, beige). In the reverse orientation are mCherry (red) and polyA (red) sequences. Heterotypic lox sites are shown for loxP (light gray) and loxPV (dark gray). FRT sites (blue) are positioned internal to the TEs (outlined boxes). (B) Insertion of the FlipTrap by transposition into the intron of an actively expressed gene leads to splicing of the citrine exon, allowing the formation of an endogenously expressed full-length fluorescent fusion protein when the insertion is in-frame with the trapped gene. (C–K) Confocal images of Citrine fusion protein expression (green) in living FlipTrap embryos counterstained with the vital stain bodipy TR methyl ester (red). (C–E) Examples of tissue-specific expression. (C) Gt(magk-citrine)^ct30c expression in the nerve bundle of the spinal cord at 82 hpf. (D) Gt(mhc7bb-citrine)^ct42a is expressed in dorsal muscle fibers in a banding pattern of sacromeres at 78 hpf. (E) Gt(dmd-citrine)^ct90a is expressed and localized to the boundary of the myotome at 32 hpf. (F–H) Examples of ubiquitously expressed FlipTraps that exhibit heightened expression in specific tissues. (F) Gt(hmga2-citrine)^ct29a is expressed ubiquitously with heightened levels of expression in the pronephric ducts (arrowhead) at 23 hpf. (G) Gt(ppp1r14bb-citrine)^ct50a exhibits heightened expression in the head vasculature at 26 hpf. (H) Gt(mapkapk2a-citrine)^ct52b is ubiquitously expressed but only exhibits localized expression within the head vasculature at 26 hpf. (I–K) Examples of ubiquitously expressed FlipTraps with distinct subcellular localization. (I) Gt(sf3a3-citrine)^ct83a is localized to the nucleus of all cells. (J) Gt(znf-citrine)^ct65a is localized to the cell membrane as puncta. (K) Gt(magt1-citrine)^ct104a is expressed in all cells but localizes to the nuclear envelope of the lateral line cells. The inset in the top right corner of K shows magnified image of four lateral line cells. C, G, and H are projections of confocal Z-stacks. Lateral view, anterior to left, dorsal top. Bar, 20 μm. According to zebrafish nomenclature guidelines, Gt indicates gene trap. See also Supplemental Figure S2.

Figure 2.

Classification of expression, genes, and distribution of FlipTrap insertion identified by screening. (A) Pie chart of the expression classes identified from screening of FlipTrap-injected F₁ embryos (N = 170). (B) Distribution of the classes of genes trapped by the FlipTrap vector (N = 158). (C) The distribution of the frequency of FlipTrap insertions mapped against gene length. (N = 128). Only lines isolated from the Tol2-containing vector were assessed due to the low number of lines isolated from the Ds-containing vector. Only lines with full in silico annotation by Ensembl were used for this analysis.

Figure 3.

Web-based interface for storage and retrieval of FlipTrap expression and molecular data. The Web-based interface database stores the FlipTrap expression and molecular data, which can be queried by four parameters: gene name, anatomy, developmental stage, and expression pattern. An example of a gene expression report for Gt(rbms3-citrine)^ct30a displays associated molecular data consisting of allele designation, gene name and alias, and links to NCBI, Zfin, Ensembl, and the University of California at Santa Cruz Genome Browser. The FlipTrap integration site is mapped to genomic annotation of the trapped locus based on Ensembl data. Wide-field and confocal fluorescent images of a protein expression pattern are displayed based on developmental stage and are anatomically annotated.

Figure 4.

Cre-lox-mediated recombination of FlipTrap lead to knockdown of trap genes. (A) Schematic of Cre-mediated recombination. Cre recombination of the lox sites lead to two intermediates: (1) Recombination of the loxP sites lead to flipping of the mCherry and polyA sequences into the forward orientation and citrine and the splice donor into the reverse orientation. (2) Recombination of the loxPV sites lead to flipping of the mCherry and polyA sequences into the forward orientation. Further recombination of either intermediate lead to the excision of the citrine and splice donor, resulting in a mutant gene trap allele that contains a splice acceptor, followed by a 3′ exon encoded by mCherry. Expression of the Cre-induced mutant allele lead to the production of a truncated protein fused to mCherry. (B–E) Confocal images of the otic vesicle of a Gt(hmga2-citrine)^ct29a embryo injected with cre and membrane-cerulean mRNA. (F–I) Confocal images of the otic vesicle in embryos from a cross between Gt(hmga2-citrine)^ct29a and Tg(bactin2:cerulean-cre)^ct5000 adults. (B,D) Citrine expression after Cre-lox recombination. (C,G) Expression of mCherry upon Cre-lox recombination. (D) Membrane-cerulean counterstain. (H) Expression of Cerulean-cre localized to the nucleus Tg(bactin2:cerulean-cre)^ct5000 in embryos. (E,I) Merges of B–D and F–H, respectively, demonstrate that expression of either Cre (D) or Cerulean-Cre fusion protein (I) leads to the recombination and conversion of the Citrine to mCherry fusion protein. The arrowhead points to Citrine-positive nuclei that have not undergone Cre-lox recombination in the somatic tissue of progenies from a Gt(hmga2-citrine)^ct29a adult crossed to a Tg(bactin2:cerulean-cre)^ct5000 adult. Bar, 20μm. (J) Quantification of relative wild-type transcripts in wild-type (blue) and homozygous Cre-induced mutant embryos as determined by RT-qPCR. Relative levels of transcripts in wild-type siblings and homozygous mutant embryos were normalized to glyceraldehyde 3-phosphate dehydrogenase (gapdh) transcripts. Transcript levels of the trap gene in wild-type siblings are expressed as 100%, while transcript levels in homozygous mutant embryos are represent as a percentage relative to wild-type siblings. Error bars represent standard deviations from triplicate RT-qPCR experiments performed on five to 10 embryos per line. The trapped gene and designated alleles are listed in the X-axis. (K) Bright-field image of wild-type sibling (top of image) and homozygous mutant embryos (bottom of image) for Gt(tpm4a-mCherry)^ct31aR at 48 hpf. Homozygous mutant embryos exhibit defects in cardiac contraction and pericardial edema (arrow). See also Supplemental Figures S5 and S6.

Figure 5.

Cre-induced mutant phenotype in FlipTrap of cltca. (A,B) Schematic of cltca with FlipTrap insertion before (A) and after (B) Cre recombination. Predicted protein domains in the cltca gene (below) as determined by Simple Modular Architecture Research Tool (SMART) with Citrine (A, green bar) and mCherry (B, red bar) protein fusion. CLH domains are shown as blue ovals. In the cre mutant allele Gt(cltca-mCherry)^ct116aR, the terminal CLH domain is deleted. (C,D) Wide-field fluorescence image of Gt(cltca-citrine)^ct116a (C) and Gt(cltca-mCherry)^ct116aR (D) embryos taken with YFP and RFP filters, respectively. Cltca-Citrine is expressed ubiquitously but is heightened in the tail vasculature (arrow), while in the cre mutant allele Gt(cltca-mCherry)^ct116aR, mCherry is expressed uniformly. (E,F) Confocal image of the tail vasculature in Gt(cltca-citrine)^ct116a (E) and Gt(cltca-mCherry)^ct116aR (F) in which the cell membranes are labeled with a membrane-localized Cerulean protein (blue). In the cre mutant allele, mCherry expression in the tail vasculature is at levels similar to surrounding cell types (arrowhead). (G,H) Bright-field image of Gt(cltca-citrine)^ct116a (G) and homozygous Gt(cltca-mCherry)^ct116aR (H) embryos at 32 hpf. Homozygous Gt(cltca-mCherry)^ct116aR embryos have a smaller forebrain and midbrain, body curvature, and dilated caudal vein. Bars: E,F, 20 μm; C,D,G,H, 50 μm. See also Supplemental Figure S7.

Figure 6.

Mutant allele disrupts localization of novel collagen protein. (A,D) Schematic of col1al with FlipTrap insertion before (A) and after (D) Cre recombination. Predicted protein domains in genes as determined by SMART analysis with Citrine (A, green bar) and mCherry (D, red bar) protein fusion are shown as follows: The thrombospondin N-terminal-like domain is a blue hexagon, and the collagen domain is a blue-outlined rectangle. In the Cre-induced mutant allele, the collagen domain is deleted from the Col1al-mCherry protein. (B,E) Bright-field image of tail region of Gt(col1al-citrine)^ct98a (B) and homozygous Gt(col1al-mCherry)^ct98aR (E) embryos at 28 hpf. Holes are present in the tail mesenchyme of homozygous Gt(col1al-mCherry)^ct98aR embryos (arrowhead). (C,F) Projection of confocal Z-stacks of the tail region of Gt(col1al-citrine)^ct98a (C) and homozygous Gt(col1al-mCherry)^ct98aR (F) counterstained with phalloidin conjugated to Alexa-633 (blue). (C) Col1al-citrine fusion proteins (green) are localized as striated bands in the extracellular matrix of the tail mesenchyme. (F) The truncated Col1al-mCherry fusion protein (red) localizes as puncta within the cytoplasm of the tail mesenchyme. (G–I) Projection of confocal Z-stacks of col1al transcript (G, blue) and Col1al-citrine protein (H, green) expression in the tail region with antibody staining to β-catenin (white) as counterstain. col1al is expressed in the tail mesenchyme cells, while Col1al-Citrine protein localizes to the extracellular matrix. Bar, 20 μm.

Figure 7.

FRT-Flp-mediated cassette exchange in FlipTrap. (A) Schematic of cassette exchange system applied to a FlipTrap of the desma gene Gt(desma-citrine)^ct122a. The exogenous DNA cassette contained FRT sites (dark-blue triangles), a splice acceptor (SA, peach), and a 3′ exon encoding mCherry (red). In the presence of Flp recombinase and exogenous DNA, the FlipTrap cassette is replaced and converted from a Citrine fusion to a mCherry fusion protein. (B–D) Confocal image of a Gt(desma-citrine)^ct122a embryo injected with Flp recombinase and the conversion of Desma-Citrine expression to Desma-mCherry after Flp-FRT recombination. (B) Confocal image of the heart tube in a Gt(desma-citrine)^ct122a embryo after FRT-Flp-mediated cassette exchange showing a portion of the myocardial cells have exchanged the citrine exon for the mCherry cassette (red). (C) Projection of confocal Z-stack of the entire heart of the embryo in B. (D) Confocal image of trunk muscle cells in Gt(desma-citrine)^ct122a embryos after cassette exchange in a few muscle fibers as indicated by expression of Desma-mCherry (red). Confocal images were overlaid with a Nomarski image. Bar, 50 μm.