A conserved maternal-specific repressive domain in Zelda revealed by Cas9-mediated mutagenesis in Drosophila melanogaster

Abstract

In nearly all metazoans, the earliest stages of development are controlled by maternally deposited mRNAs and proteins. The zygotic genome becomes transcriptionally active hours after fertilization. Transcriptional activation during this maternal-to-zygotic transition (MZT) is tightly coordinated with the degradation of maternally provided mRNAs. In Drosophila melanogaster, the transcription factor Zelda plays an essential role in widespread activation of the zygotic genome. While Zelda expression is required both maternally and zygotically, the mechanisms by which it functions to remodel the embryonic genome and prepare the embryo for development remain unclear. Using Cas9-mediated genome editing to generate targeted mutations in the endogenous zelda locus, we determined the functional relevance of protein domains conserved amongst Zelda orthologs. We showed that neither a conserved N-terminal zinc finger nor an acidic patch were required for activity. Similarly, a previously identified splice isoform of zelda is dispensable for viability. By contrast, we identified a highly conserved zinc-finger domain that is essential for the maternal, but not zygotic functions of Zelda. Animals homozygous for mutations in this domain survived to adulthood, but embryos inheriting these loss-of-function alleles from their mothers died late in embryogenesis. These mutations did not interfere with the capacity of Zelda to activate transcription in cell culture. Unexpectedly, these mutations generated a hyperactive form of the protein and enhanced Zelda-dependent gene expression. These data have defined a protein domain critical for controlling Zelda activity during the MZT, but dispensable for its roles later in development, for the first time separating the maternal and zygotic requirements for Zelda. This demonstrates that highly regulated levels of Zelda activity are required for establishing the developmental program during the MZT. We propose that tightly regulated gene expression is essential to navigate the MZT and that failure to precisely execute this developmental program leads to embryonic lethality.

Fig 1. ZLD-PB is the dominant isoform expressed both in the embryo and imaginal disc.

(A) Schematic of predicted transcripts from the zld locus (above). Boxes indicate exons with coding sequence (light gray) and untranslated regions (dark gray). The length of the resulting mRNA is given in nucleotides (nt). Two gRNAs flanking the downstream exon of zld-RD used to generate an isoform specific deletion are shown. Schematics of the predicted protein products for each splice variant (below) with amino acid numbers in D. melanogaster ZLD shown above. Above are the approximate locations of the transcriptional-activation and DNA-binding domains as demonstrated in Hamm et al (2015) [20]. (B) Confocal images of embryos homozygous for either an N-terminal mCherry-tagged ZLD or for ZLD-PD-mCherry at stages 5, 12–13, and 14–16. The first column shows the maximum projection images for mCherry-ZLD expressing embryos. All other images show a single confocal slice. ZLD-PD-mCherry (1) and (2) indicate embryos from two distinct editing events. (C) Confocal images of mCherry-ZLD isoforms demonstrating the endogenous ZLD and ZLD-PD specific expression in third instar larval wing discs. Outlines show the borders of the wing discs as determined by transmitted light images. All images shown are the maximum projection. (D) Immunoblot for ZLD on S2 extract from cells expressing either ZLD-PB or ZLD-PD or total lysate from third instar wing discs. (E) Sequence of zld-RD following Cas9-mutagenesis demonstrating removal of the splice acceptor and coding sequence. Small insertions (red sequence) and deletions (dashed lines) shown. Lower case letters indicate intron sequence. Capital letters indicate exonic sequence.

Fig 2. Cas-9 mediated mutation of highly conserved domains in ZLD.

(A-C) Alignment of amino acid sequences of ZLD protein domains from four insect species (Drosophila melanogaster, Drosophila virilis, Anopholes gambiae, Nasonia vitripennis). Red dots indicate conserved cysteine and histidine residues in the C₂H₂ zinc finger domains. Blue dots indicate the conserved aspartate and glutamate residues in the acidic domain. Point mutations generated in the endogenous zld locus by Cas9-mediated genome engineering are shown above. Numbers above indicate amino acids in D. melanogaster ZLD. (D) Schematic of the experimental design for homology directed repair from a single-stranded donor oligonucleotide (ssODN) following a Cas9-mediated, double-strand break. Silent restriction site (yellow), silent mutation in the gRNA target site (blue) and the desired point mutation (red) are shown. (E) Wild-type DNA sequence coding for the targeted ZLD domains (below) with mutated nucleotides generated by homology directed repair (above) are shown. Arrowheads indicate Cas9 cleavage site.

Fig 3. The second zinc finger in ZLD is required for maternal, but not zygotic function.

(A) Viability and fertility of heterozygous and homozygous animals carrying point mutations in zinc finger 1 (ZnF1), the acidic domain (EDD), or zinc finger 2 (ZnF2). ^aPercent recovered was calculated in comparison to heterozygous FM7 female siblings. Percent based on Mendelian inheritance. ^bViability was determined by crossing to a w¹¹¹⁸ mate. (B) Immunoblots with anti-ZLD antibodies demonstrate wild-type levels of protein expression from alleles harboring point mutations in conserved domains. Protein levels were assayed on total protein lysate from embryos 1–2 hours after egg laying (AEL) maternally inheriting both the mutated allele and a superfolder GFP-tagged allele that is fully functional, allowing internal normalization.

Fig 4. Mutation in the second C₂H₂ zinc finger of ZLD results in increased ability to activate transcription.

Fold activation of luciferase reporters driven by either a wild-type scute promoter (WT) or one with mutations in the ZLD-binding sites (MUT). Immunoblot with anti-ZLD antibodies demonstrates comparable levels of ZLD expression. n = 3, error bars indicate +/- standard deviation.

Fig 5. Conserved residues in the second zinc finger domain shared with JAZ-like zinc fingers are essential for wild-type ZLD activity.

(A) Alignment of the second zinc finger of ZLD with the consensus amino acids in JAZ domains. Bold indicates amino acids shared between the consensus and ZLD with red indicating the zinc-chelating cysteines and histidines and blue indicating additional shared residues. Underlined residues are those that contact double-stranded RNA. (B) Alignment of amino acid sequence of second zinc finger in ZLD with other insect species. Green bars below the sequence indicate residues conserved in all arthropods that contain the ZnF2 domain, as identified by Ribeiro et al. [25]. Red dots indicate conserved cysteine and histidine domains in the C₂H₂ zinc finger. Blue dots indicate the amino acids conserved in JAZ zinc fingers. Point mutations generated in the endogenous zld locus by Cas9-mediated genome engineering are shown above. (C) Wild-type DNA sequence coding for the targeted ZLD domains (below) with mutated nucleotides generated by homology directed repair (above) are shown. Blue arrowhead indicates Cas9 cleavage site. (D) Immunoblot with anti-ZLD antibodies demonstrates wild-type levels of protein expression from the allele harboring point mutations in the conserved JAZ-domain amino acids. Protein levels were assayed on total protein lysate from embryos 1–2 hours after egg laying (AEL) maternally inheriting both the mutated allele and a superfolder GFP-tagged allele that is fully functional, allowing internal normalization. (E) Viability and fertility of heterozygous and homozygous animals carrying point mutations in the JAZ domain amino acids. ^aPercent recovered was calculated in comparison to heterozygous FM7 female siblings. Percent based on Mendelian inheritance. ^bViablity was determined by crossing to a w¹¹¹⁸ mate. (F) Fold activation of luciferase reporters driven by either a wild-type scute promoter (WT) or one with mutations in the ZLD-binding sites (MUT). Immunoblot with anti-ZLD antibodies demonstrates comparable levels of ZLD expression. n = 3, error bars indicated +/- standard deviation.

Fig 6. Mutation of the second zinc finger results in both increased zygotic genome activation and maternal RNA degradation during the MZT.

(A) Volcano plot showing log2 fold change values (x-axis) by–log10 corrected p-values (y-axis) for all genes identified by RNA-seq. Data represent a comparison of gene expression in embryos laid by either mothers with C554S mutation in zinc finger 2 (zld^ZnF2) or by w¹¹¹⁸ mothers. Genes that are significantly altered in expression (p-value < 0.05, > 2 fold change) are indicated as red dots (up-regulated) or blue dots (down-regulated). (B) Numbers of up-regulated (red) and down-regulated genes that are maternally expressed (maternal), zygotically expressed (zygotic) or both maternally and zygotically expressed (mat-zyg). Classification of gene expression is as defined in Lott et al. 2011 [34]. (C) Venn diagram showing the overlap of up-regulated, zygotically expressed genes and genes bound by ZLD (as defined by ChIP-seq in Harrison et al. 2011 [13]). (D) GO term enrichment for up-regulated genes associated with ZLD-binding sites. (E) Venn diagram showing the overlap of down-regulated, maternally expressed genes with the sets of genes subject to maternal or zygotic degradation pathways as defined in Thomsen et al. 2010 [10]. p-values are calculated by Fisher’s exact test (S2 Table). (F) Model of the effects on maternal and zygotic gene expression over the MZT due to maternal inheritance of zld mutants.