Fig 1.
Efficient integration of loxP sites into the first and second introns of tbx20.
a. Diagram of the tbx20 locus. Exons are drawn to scale, introns not to scale. The number below each intron-exon junction indicates reading frame phase. b. Target sites of two highly active intronic tbx20 sgRNAs, tbx20 sgRNA9 and tbx20 sgRNA10, flanking exon 2. Sequence corresponding to single guide RNA is shown in blue, PAM motif is in bold, and expected Cas9 cut site is indicated by a red x. c-f. Integration of loxP site into intron 2 of tbx20. c. The single stranded oligonucleotide used for homology-directed repair was in sense strand with regard to the PAM and had loxP (aqua) site flanked by 3 nucleotide spacer sequences (grey), and 21 nucleotide long homology arms. d. Representative image of the 5’ nested PCR reaction used to screen pools of F1 embryos. L, 100 bp DNA ladder (ThermoFisher Scientific). 1–13, PCR on pools of embryos from 5 different F1 in-crosses (pools from the same F0 pair are grouped, loxP-positive cross is underlined in red). C1 control DNA from embryos injected with tbx20 sgRNA9 and tbx20sg9-loxP HDR oligonucleotide. C2, control DNA from embryos injected with tbx20 sgRNA10 and tbx20sg10-loxP HDR oligonucleotide. e. PCR genotyping of tail clips of adult F1 fish from the F0 in-cross underlined in red in d. Red arrows indicate fish heterozygous for the loxP-containing allele. Yellow arrows indicate fish which likely contain one loxP-containing allele, but presumably lack the wild type allele due to a deletion inherited from the other parent. f. Sequence of the recovered tpl135 allele containing integration of the full-length loxP site with 10-nucleotide partial target site duplication at the 5’ of the HDR oligonucleotide. g, h. Experimental design for integration of loxP site into intron 1 of tbx20. g. The single stranded oligonucleotide used for homology-directed repair was antisense to the PAM-containing strand and had loxP site flanked by 3-nucleotide spacer sequences, 62-nucleotide 5’ homology arm and 18-nucleotide 3’ homology arm. h. Sequence of the recovered tpl136 allele containing perfect integration of the loxP site.
Fig 2.
Generation and testing of a conditional (“floxed”) tbx20 allele.
a. Experimental design. b. Genotyping of adult fish for intron 2 loxP site. c. Results of nested PCR screening for loxP integration into intron 1. d. Sequence of intron 1 loxP integration in recovered tpl145 floxed tbx20 allele. e. Genotyping of “F1” adults. Primer binding sites are shown as black arrows, loxP sites as red or blue triangles, exon as an open box. f, g. Induction of tbx20 loss of function by injection of Cre mRNA. f. One quarter of embryos obtained by in-crossing tbx20tpl145 heterozygotes and injected with Cre mRNA display a consistent, severe heart development defect. g. Genotyping of embryos with severe heart defects (lanes 1–3) and phenotypically normal siblings (lanes 4–8). L, GeneRuler DNA Ladder (ThermoFisher Scientific). Genotyping lanes 1 and 5 correspond to images in f. h, i. Induction of tbx20 deletion by 4-hydroxytamoxifen. tbx20tpl145 heterozygote was crossed to Tg(ubi:CreERT2) line. GFP-positive embryos were collected at 2dpf and incubated with 5μM 4-HT for 24 hours. j. Adults raised from Cre-injected tbx20tpl145/+ embryos were incrossed, resulting in approximately 1/4 of embryos (36/128, 28%) with severe heart defects. k. Confirmation of Cre-mediated excision of the second intron of tbx20 by sequence analysis. l, m. Analysis of excision efficiency by qPCR. l. Excision efficiency was assessed by qPCR in embryos treated with 4-HT at various concentrations at different time points and in two different ubb:CreERT2 driver lines. The y-axis indicates un-excised tpl145 normalized to the untreated control. m. Images of tbx20 phenotypes shown after treatment with 4-HT. Individuals 1, 2, and 3 from l correspond to images labeled 1, 2, and 3.
Fig 3.
Conversion of a Cre-revertible Gene Breaking Transposon (GBT) allele to a conditional allele.
a. Diagram of the fleer locus. Exons drawn to scale, introns not to scale. Below each intron-exon junction reading frame phase is indicated. b. Diagram of the fleer gene trap allele flrtpl19. c. Diagram flrtpl19R locus reverted by Cre-mediated excision of the gene trap cassette. d. Diagram of flr sgRNA3 target site and antisense oligonucleotide HDR template. e. Sequence of loxP integration into intron 7, resulting in floxed fleer allele tpl141. f, g, h. Cre-mediated excision of exons 2–7 of fleer. f. Induction of fleer deletion by Cre mRNA. g. One quarter of embryos obtained by in-crossing fleertpl141 heterozygotes and injected with Cre mRNA display a phenotype consistent with fleertpl19 homozygotes in h, including formation of kidney cysts shown in g’ and h’. i. Sequence of the excision amplicon.
Fig 4.
a. Diagram of the aldh1a2 locus. Exons drawn to scale, introns not to scale. Below each intron-exon junction reading frame phase is indicated. b. Two highly active intronic aldh1a2 sgRNAs, aldh1a2 sgRNA1 and aldh1a2 sgRNA4, flank exon 8. Sequence corresponding to single guide RNA is shown in blue, PAM motif is in bold, and expected Cas9 cut site is indicated by a red x. c. Two independent aldh1a2 exon 8 deletion alleles recovered after co-injection of aldh1a2 sgRNA1 and aldh1a2 sgRNA4 along with nCas9n mRNA. For Sanger sequencing of the alleles, see S3 Fig. d, e. Deletion of exon 8 of aldh1a2 results in expected loss-of-function phenotype. F1 fish heterozygous for aldh1a2tpl137 and aldh1a2tpl138 deletions were crossed to each other. d. Images of 3 dpf embryos displaying wild type (top) and the expected aldh1a2 loss of function phenotypes: lack of pectoral fins, shortened hindbrain brain and cardiac edema. In addition, most of the 3 dpf embryos displaying these phenotypes had curved tails (d, bottom) consistent with uneven left/right somite numbers. e. aldh1a2tpl137/tpl138 trans-heterozygotes lack pectoral fin buds as revealed by loss of tbx18 expression at 32 hpf. f. Expression of tcf21 persists in the first and second branchial arches of aldh1a2tpl137/tpl138 trans-heterozygotes at 32 hpf. g. Diagram of the HDR template oligonucleotide used to knock in the loxP into aldh1a2 sgRNA1 target site. h. Sequencing of the precise loxP knockin allele. Additional knock-in alleles recovered from other F0 families are shown in S5 Fig. i. Genotyping of 16 phenotypically normal 3 dpf embryos from a cross between F1 fish heterozygous for aldh1a2tpl139 loxP knock-in results in Mendelian ratios of aldh1a2tpl139/tpl139 (red arrows), aldh1a2tpl139/wt (yellow arrows) and aldh1a2wt/wt embryos. PCR fragments from two of the embryos were sequenced to further confirm homozygosity for the loxP site.
Fig 5.
Conditional mutagenesis pipeline.
Upon deciding which exon to flox, we recommend sequencing the target sites to identify polymorphisms compared to reference genome. Next, sgRNAs should be designed and tested by either direct sequencing of PCR fragments, T7 endonuclease assay or loss of a restriction enzyme site on bulk DNA from pooled embryos. Once active sgRNAs have been identified, experiments integrating the first loxP site should be performed. In the absence of conclusive data that certain HDR template performs significantly better than others (such experiments are not practical at the only level that matters–germline transmission), we recommend using the design we successfully used to integrate loxP into fleer, aldh1a2 and tcf21: antisense to PAM, with 49-base 5’ homology arm and 21-base 3’ homology arm, with 3-nucleotide spacers flanking loxP site. As injected embryos are being raised, we then recommend to optimize nested PCR screening conditions DNA from pools of injected embryos. We found “plain” Taq polymerases (NEB #M0270, Thermofisher Scientific 2x PCR Master Mix Cat# AB-0575/DC and #EP0402, or similar) to be most suitable for nested PCR. In contrast, high-performance mixes such as Platinum Taq (Thermofisher #10966026) or Kapa 2G Fast ReadyMix + dye (Kapa Biosystems- KM5101) yield very high background and may only be used for the second (nested) reaction. It is also very helpful if primers for one end of the nested PCR are anchored within an exon. We recommend generating a deletion allele in parallel with integration of the first loxP site. Once highly active sgRNAs are identified, we recommend injecting a pair of sgRNAs flanking the exon to be floxed in order to confirm that removal of selected exon will yield an overt phenotype. We have been able to very efficiently delete exon 8 of aldh1a2 using sgRNAs spaced just over 450 base pairs, but larger deletions are certainly feasible too (1, 2). An additional benefit of a deletion allele is that it can be crossed to Cre drivers of interest, eliminating the need to back-cross floxed allele to obtain homozygotes. Screening for germline transmission should be performed by nested PCR on pools of embryos obtained from incross. Positive crosses should be analyzed by performing short flanking PCR (ideally under 400 base pairs) on DNA from individual embryos. Bands corresponding to loxP-containing allele should be extracted from gel and sequenced to ensure presence of intact loxP site. Siblings of screened embryos should be raised to adulthood and loxP-positive F1s should be identified by flanking PCR as well. Two strategies can be used for integration of the second loxP site. If speed is the main priority, loxP-positive F1s can be in-crossed and second sgRNA/HDR oligonucleotide can be injected. The main drawback of this strategy that there is only 50% likelihood that the second loxP site will integrate into a chromosome already containing the first loxP. It is therefore necessary to genotype adults for presence of the first loxP site before out-crossing. Even though we successfully used this strategy to engineer a floxed allele of tbx20, we consider it impractical and would generally recommend to first generate adults homozygous for the first loxP site, incross them and then inject the second sgRNA/HDR oligonucleotide.