Efficient targeted mutagenesis in the monarch butterfly using zinc-finger nucleases

Abstract

The development of reverse-genetic tools in "nonmodel" insect species with distinct biology is critical to establish them as viable model systems. The eastern North American monarch butterfly (Danaus plexippus), whose genome is sequenced, has emerged as a model to study animal clocks, navigational mechanisms, and the genetic basis of long-distance migration. Here, we developed a highly efficient gene-targeting approach in the monarch using zinc-finger nucleases (ZFNs), engineered nucleases that generate mutations at targeted genomic sequences. We focused our ZFN approach on targeting the type 2 vertebrate-like cryptochrome gene of the monarch (designated cry2), which encodes a putative transcriptional repressor of the monarch circadian clockwork. Co-injections of mRNAs encoding ZFNs targeting the second exon of monarch cry2 into "one nucleus" stage embryos led to high-frequency nonhomologous end-joining-mediated, mutagenic lesions in the germline (up to 50%). Heritable ZFN-induced lesions in two independent lines produced truncated, nonfunctional CRY2 proteins, resulting in the in vivo disruption of circadian behavior and the molecular clock mechanism. Our work genetically defines CRY2 as an essential transcriptional repressor of the monarch circadian clock and provides a proof of concept for the use of ZFNs for manipulating genes in the monarch butterfly genome. Importantly, this approach could be used in other lepidopterans and "nonmodel" insects, thus opening new avenues to decipher the molecular underpinnings of a variety of biological processes.

Figure 1.

Validation of ZFN activity to target cry2 in DpN1 cells. (A) Proposed core transcriptional feedback loop of the monarch circadian clockwork. CLOCK (CLK) and CYCLE (CYC) heterodimers drive the transcription of period (per), timeless (tim), and cryptochrome 2 (cry2), which upon translation form complexes, cycle back into the nucleus, where CRY2 inhibits CLK:CYC-mediated transcription on a 24-h basis. (B, top) Schematic of monarch cry2 gene and position of the ZFN-targeted site (red star). (Black boxes) Exons. (Bottom) Magnified view illustrating binding sites for the ZFN pair, each consisting of four zinc-finger modules linked to either DD or RR variants of the FokI endonuclease. (C) Wild-type CRY2 (amino acids 1–742) represses CLK/CYC-mediated transcription in S2 cells in a dose-dependent manner, while the truncated protein (amino acids 1–160) does not. The monarch per E-box enhancer luciferase reporter (dpPer4Ep; 10 ng) was used in the presence (+) or absence (−) of dpCLK/dpCYC expression plasmids (5 ng each) and either the full-length CRY2 (FL-CRY2; 1, 2, and 10 ng) or truncated CRY2 (Tr-CRY2; 1, 2, and 10 ng). Luciferase activity is relative to beta-galactosidase activity and normalized so that the relative activation by dpCLK/dpCYC alone is 100%. Each value is the mean ± SEM of three independent transfections. (D, top) Strategy used to detect ZFN-induced mutations in monarch cells by restriction endonuclease assay. (Red and blue lines) Wild-type genomic amplicon showing the presence of an EagI site. Genomic fragments with mutations induced by NHEJ are resistant to restriction enzyme digestion because of the loss of the EagI site and appear uncleaved (green line). (Bottom) Estimation of ZFN activity in DpN1 cells. Genomic amplicons from a pool of cells untreated or treated with 1 μg of ZFNs, subjected to restriction digest. The frequency of NHEJ in treated cells was estimated by quantification of ethidium bromide staining and densitometry of the resistant band (green arrow) relative to wild-type fragments (blue and red arrows). (E) ZFN-induced cry2 mutations in DpN1 cells. (Gray shaded boxes) The ZFN recognition sites on the wild-type sequence. (Blue letters) The EagI site. The positions of (red dashes) deletions and (red letters) insertions.

Figure 2.

Injection and screening strategy for recovering germline-transmitted ZFN-induced mutants. (1) To target the first nuclei during embryonic development, fertilized eggs are injected with ZFN-encoding mRNAs within 20 min after egg laying (AEL) at the micropyle, the region at which the sperm is transferred and fertilization occurs in holometabolous insects. (2) Surviving larvae are reared to the fourth instar larvae and then are subjected to noninvasive genotyping to screen for mosaics presenting a high frequency of targeting in somatic cells. The presence of lesions is tested by restriction endonuclease assay from fleshy filaments of a single animal (larval sensors at the anterior thorax) whose removal does not alter the butterfly's survival or fertility, and the mutated alleles are sequenced. Of note, for species that do not possess external structures, genotyping could alternatively be performed using hemolymph extracts. (3) To reduce breeding efforts, only mosaic larvae targeted with high frequency in somatic cells are reared to adulthood (mixed orange and gray) and backcrossed to wild-type butterflies (orange). (4) The G₁ progeny are screened for targeted mutations as in 2.

Figure 3.

ZFN-induced mutagenesis at the cry2 locus in somatic cells. (A) Effects of ZFN mRNA injections on monarch embryos. The bar graph depicts the rate of hatching larvae (gray bars) and of embryos presenting development but dying before hatching (white bars), associated with egg injection of each of the ZFNs mRNA alone (linked to either of the obligate heterodimeric versions of the FokI nuclease domain) or both ZFNs mRNA at 0.5 μg/μL. For comparison, control embryos injected with water in similar conditions hatch at a rate of 19.5% with an additional 3.6% that develop but die before hatching (n = 82). Uninjected embryos hatch at a rate of ∼75%. (B) Agarose gel example of lesions observed in somatic cells of hatched larvae and dead embryos injected with both ZFN mRNAs, showing a high degree of targeting (green arrow), compared with an uninjected control. (C) Mutations observed in somatic cells of dead embryos, and dead or live larvae, from eggs injected at different doses of ZFN mRNAs.

Figure 4.

Germline transmission of cry2 ZFN-targeted mutations. (A) Table of heritable NHEJ mutagenesis at the cry2 locus. Potential founders (sex and number) presenting a high level of targeting in somatic cells that produced fertile crosses and those that yielded mutant progeny are shown. The approximate level of mosaicism, the number of progeny screened, the number of mutants per parent, the percentage of mutants in the progeny of each founder, and the number and name of mutations recovered are reported. The level of mosaicism was estimated by quantification of ethidium bromide staining and densitometry of the resistant band relative to wild-type fragments. (B) Alleles carried by founders showing mutations at the targeted site, a 4-bp deletion (M1) and a 2-bp insertion (M2), respectively, transmitted to 50% and 39.2% of the offspring (see above). Both mutations cause frameshifts leading to truncated proteins. (Top) Sequences; (bottom) chromatogram profiles. (C) Western blot analyses of CRY2 in brains of sibling wild-types (+/+), and heterozygous (+/−) and homozygous (−/−) mutants for each mutation. (Left) Four-base-pair deletion (M1); (right) 2-bp insertion (M2). The specific band recognized by a monarch-specific anti-CRY2 antibody directed against the C terminus of the protein (lacking in homozygous mutants) is identified by the arrow on the left of each panel.

Figure 5.

CRY2 deficiency disrupts circadian behavior. (A) The act of monarch eclosion from chrysalis to butterfly. (B) Profiles of adult eclosion in constant darkness (DD) of wild-type (+/+, black), heterozygous (+/−, gray), and homozygous mutant (−/−, red) siblings of the 4-bp deletion (left) and the 2-bp insertion (right) cry2 mutant line, entrained to LD throughout their larval and pupal stages. Eclosion occurred on the first day and second day of DD. Data from both days are pooled and binned in 1-h intervals (for detailed data throughout the 2 d in DD) (see Supplemental Fig. S3). Effect of genotype on eclosion time, one-way ANOVA: P < 0.0001 for both lines; Tukey post-hoc test: +/+ versus +/−, P > 0.05; +/+ versus −/−, P < 0.01; +/− versus −/−, P < 0.01 for both lines. (Black horizontal bars) Night/subjective night; (gray horizontal bars) subjective day.

Figure 6.

CRY2 deficiency disrupts the molecular clockwork. (A) Circadian expression of per and tim in brains and antennae of adult wild-type in constant darkness (DD) of wild-type (black lines), heterozygous (gray lines), and homozygous mutant (red lines) siblings of the 4-bp deletion cry2 mutant line entrained to LD throughout their larval and pupal stages. Values are mean ± SEM of four animals. Interaction genotype × time, two-way ANOVA: per in brain, P < 0.0001; per in antenna, P < 0.0001; tim in brain, P < 0.005; tim in antenna, P < 0.0001. (Box shading: light gray) Subjective day; (dark gray) night. (B) Expression levels of per and tim in brains and antennae of adult from the three genotypes of the 2-bp insertion cry2 mutant line at circadian time (CT)12, which corresponds to the trough of per and tim rhythmic expression in wild type. One-way ANOVA: P < 0.0001 for per and tim in both tissues; post-hoc t-test: P < 0.05 for per and tim between wild-type and knockouts, and heterozygotes and knockouts in both tissues. Values are the mean ± SEM of six animals. (Black bars) Wild type (+/+); (gray bars) heterozygous mutants (+/−); (red bars) homozygous mutants (−/−).