Efficient site-specific transgenesis and enhancer activity tests in medaka using PhiC31 integrase

Abstract

Established transgenesis methods for fish model systems allow efficient genomic integration of transgenes. However, thus far a way of controlling copy number and integration sites has not been available, leading to variable transgene expression caused by position effects. The integration of transgenes at predefined genomic positions enables the direct comparison of different transgenes, thereby improving time and cost efficiency. Here, we report an efficient PhiC31-based site-specific transgenesis system for medaka. This system includes features that allow the pre-selection of successfully targeted integrations early on in the injected generation. Pre-selected embryos transmit the correctly integrated transgene through the germline with high efficiency. The landing site design enables a variety of applications, such as reporter and enhancer switch, in addition to the integration of any insert. Importantly, this allows assaying of enhancer activity in a site-specific manner without requiring germline transmission, thus speeding up large-scale analyses of regulatory elements.

Keywords: Fish; In vivo analysis; PhiC31 integrase; Position effects; Regulatory DNA.

Fig. 1.

Design of the PhiC31 transgenesis system. (A) Landing site with cmlc2 promoter (Auman et al., 2007; Huang et al., 2003) driving EGFP in the myocardium. The EGFP open reading frame (ORF) contains an in-frame attP site. This cassette is flanked by heterotypic lox sites (loxP and lox2272) as well as Sleeping Beauty transposon IR/DRs (SB). (B) The targeting vector contains an attB site and heterotypic lox sites corresponding to the respective sites of the landing site. Additionally, the vector contains a TagRFP-T coding sequence lacking both promoter and start codon. This coding sequence is preceded by the attB site. To prevent unscheduled expression of TagRFP-T by protein trapping, the attB site is furthermore preceded by stop codons in all forward reading frames. (C) Site-specific recombination between attP and attB (colored triangles/squares) is initiated upon microinjection of targeting vector and PhiC31o integrase mRNA into a transgenic line containing a genomic landing site. Specific targeting of the landing site is reported by targeting sensors that are activated through molecular complementation of targeting vector and landing site. Site-specific integration disrupts the attP-EGFP ORF of the landing site by replacing the EGFP with the TagRFP-T of the integration vector. This displacement occurs in-frame, thereby generating functional cmlc2::attL-TagRFP-T resulting in a fluorescence color switch in the myocardium. (D) The targeted genomic landing site initially contains the entire targeting vector plasmid (C). The complementation of heterotypic lox sites of targeting vector and landing site allows the removal of these undesired sequences through Cre-mediated locus cleanup.

Fig. 2.

Two genomic landing site lines. (A) Genomic locations of the landing sites of lines 1 (top) and 2 (bottom). Both sites are located in gene deserts. The closest neighboring genes for line 1 are a distance of 78 kb and 89 kb (including gaps in the assembly). No genes are annotated on scaffold 1518 that landing site 2 maps to (size of the scaffold is 29 kb). (B) Representative embryos stably targeted with hsp70::EGFP. Strong EGFP expression in lens (white arrowheads) reflects the autonomous activity of this promoter. EGFP expression in the myocardium (red arrowheads) occurs as the result of an interaction of hsp70 with the cmlc2 promoter in the targeting vector. No other EGFP fluorescence is detectable. Pigment cells are autofluorescent in both red and green color spectra (open cyan arrowheads; see also supplementary material Fig. S5). Scale bars: 50 μm. (C) Targeting efficiency and germline transmission of targeting vector. The targeting efficiency was calculated as the weighted average of three independent experiments, and expressed as percentage of the maximum 50% positive embryos in a heterozygous outcross. Using targeting sensors, candidate fish were preselected to establish stable transgenic lines. Four out of five (line 1) and six out of eight (line 2) of the candidates transmitted the targeting vector through the germline. (D) Southern blot analysis to confirm the presence of single copy genomic landing site integration in both lines. Genomic DNA was digested with an enzyme cutting once in the landing site (NcoI). An EGFP probe was used for hybridization. The ladder is the GeneRuler DNA Ladder Mix (Thermo Fisher).

Fig. 3.

Transcriptional accessibility of the landing site during development. (A-C) Landing site line 2, stably transgenic with hsp70::EGFP was heatshocked at different stages of development. At all tested stages, the heatshock resulted in ubiquitous EGFP expression (for pre-heatshock expression compare with Fig. 2B), highlighting full transcriptional accessibility of the locus during development. Identical results were obtained for line 1 (data not shown). Live embryos, imaged through the chorion (A,B) are shown; dorsal views, anterior to the left. Scale bars: 50 μm.

Fig. 4.

Cre-mediated locus cleanup. (A) The alg2 locus. Two landing site integrations are contained within the locus, one of them within exon 3. Note the high number of conserved non-coding sequences around the locus. (B) Genomic locus after targeting, before and after Cre-mediated removal of vector sequences. The Cre-cleanup removes the cmlc2 promoter from the locus, thereby disrupting its interaction with the hsp70 promoter. (C) Landing site line 3 stably transgenic for hsp70::EGFP; dorsal view (left) and time-averaged lateral view (right). In addition to the hsp70-autonomous lens expression (white arrowheads), and the heart expression resulting from its interaction with cmlc2 (red arrowheads), an enhancer trap in the brain is apparent (cyan arrowheads). (D) Landing site line 3 stable transgenic for hsp70::EGFP after Cre-mediated locus cleanup. This results in the loss of EGFP fluorescence in the heart (red arrowheads) but does not affect the brain-specific enhancer trap (cyan arrowheads), demonstrating that it is the result of the regulatory landscape of the alg2 locus. The hsp70-autonomous EGFP expression in the lens is retained (white arrowheads). (E) Southern blot analysis showing multiple landing site integrations in the line. Scale bars: 50 μm.

Fig. 5.

Testing targeting efficiency by locus-specific analysis of promoters in the injected generation. (A) The genomic landing site after targeting and molecular complementation is shown. The promoter sequence provided with the targeting vector can drive the expression of an attR-EGFP open reading frame (ORF) only upon proper integration of the targeting vector. (B-D) For the determination of targeting efficiency after injection, the hsp70 promoter was chosen. (B) Heatshock at ≤50% epiboly resulted in reporter gene expression patterns that fall into three groups with high, medium and low uniformity (quantification is shown in C). (D) A heatshock at 4 dpf triggers EGFP expression in the vast majority of cells, highlighting the high degree of uniformity achieved already in the injected generation. (E,F) Two other promoters, namely the lens crystallin (E, red arrowheads) (Emelyanov and Parinov, 2008) and the sonic hedgehog promoter (F, red arrowheads) (Neumann and Nuesslein-Volhard, 2000) were tested and showed reproducibly the expected expression pattern. Scale bars: 50 μm.

Fig. 6.

Locus-specific analysis of enhancer sequences in the injected generation. (A) The targeting vector with the enhancer test setup after targeting and molecular complementation is depicted. Variable enhancer sequences are cloned upstream to the hsp70 promoter and analyzed in the injected generation by imaging live embryos at 4 dpf (B,C). (B) The activity of the enhancer HMMA_4 is depicted (frontal view, ventral to the left). Reporter gene expression is detected in blood cells (red arrowheads). In line 3, additional activity from the characterized enhancer trap was detected in the brain (cyan arrowheads). (C) A mouse p300-bound element identified from mouse forebrain and the orthologous medaka sequence were tested. Both show specific reporter gene expression in the forebrain highly reminiscent to the stable lines (red arrowhead, compare with D). Dorsal view, anterior to the left. (D) Reporter gene expression patterns detected in a transgenic line generated with the same regulatory elements by a meganuclease-based approach (red arrowheads, compare with C). Additional expression in the hindbrain (cyan arrowhead) occurred at later stages (6 dpf, M. Eichenlaub, personal communication). Figures are taken from Eichenlaub and Ettwiller (Eichenlaub and Ettwiller, 2011). Scale bars: 50 μm.

Fig. 7.

Concatemeric assembly of the Fam44b enhancer increases reporter gene expression levels. (A) A single copy of the Fam44b enhancer element (Eichenlaub and Ettwiller, 2011) shows no detectable reporter gene expression except for hsp70-dependent expression in lens and heart. A heatshock confirms uniform targeting and integration. (B) Expected expression pattern from a transgenic line established using the meganuclease-based approach at 4 dpf. Figure is taken from Eichenlaub and Ettwiller (Eichenlaub and Ettwiller, 2011). (C,D) Seven copies of the Fam44b enhancer element cloned upstream of a single hsp70 promoter lead to prominent reporter gene expression. Time course to show expression at different stages of development in injected fish. Note the high degree of similarity of the expression pattern at 4 dpf between the injected fish and the transgenic line (B). Scale bars: 50 μm.