RNAi-mediated gene silencing in zebrafish triggered by convergent transcription

Abstract

RNAi based strategies to induce gene silencing are commonly employed in numerous model organisms but have not been extensively used in zebrafish. We found that introduction of transgenes containing convergent transcription units in zebrafish embryos induced stable transcriptional gene silencing (TGS) in cis and trans for reporter (mCherry) and endogenous (One-Eyed Pinhead (OEP) and miR-27a/b) genes. Convergent transcription enabled detection of both sense and antisense transcripts and silencing was suppressed upon Dicer knockdown, indicating processing of double stranded RNA. By ChIP analyses, increased silencing was accompanied by enrichment of the heterochromatin mark H3K9me3 in the two convergently arranged promoters and in the intervening reading frame. Our work demonstrates that convergent transcription can induce gene silencing in zebrafish providing another tool to create specific temporal and spatial control of gene expression.

Figure 1. Convergent silencing of mCherry.

(A) Non-CT and CT constructs were designed using the Tol2 transgenesis system with the β-actin promoter driving sense transcription and an inverted CMV promoter driving antisense transcription of the mCherry open reading frame (see Supplement for maps and sequence information). All constructs harbored heart-specific GFP (cmlc2-GFP) which enabled identification of transgenic embryos. (B) Widespread expression of mCherry was observed in non-CT embryos (386 out of 386 transgenic embryos), whereas near complete loss of mCherry was observed in CT-mCherry embryos at 2 dpf (370 out of 402 transgenic embryos).

Figure 2. mCherry silencing is Dicer dependent and results in increased levels of H3K9me3 chromatin modification.

(A) Detection of sense and antisense mCherry transcripts via qPCR analysis in F2 embryos from 3 different CT-mCherry F1 lines. (B) Increased mCherry sense mRNA levels upon co-injection of a Dicer MO compared to control (ctrl) MO injected transient transgenics at 54 hpf. (C) Rescue of mCherry protein levels upon co-injection of Dicer MO. Protein lysates were prepared from embryos as in (B) and Western blots were prepared with antibodies against mCherry and α-tubulin. The full gel is shown in Supplemental Fig. 1. For comparison, protein levels in a non-CT embryo are as shown. (D) ChIP-qPCR showing significant enrichment of H3K9me3 levels on convergent chromatin from CT-mCherry F2 embryos (F1 line 9119). Enrichment was determined compared to negative IgG control at 5 dpf. ChIP values and standard deviations are shown from three independent biological experiments. ***p < .001 based on unpaired, two-tailed distribution Student's t-test.

Figure 3. Silencing of One Eyed Pinhead.

Convergent silencing of the zebrafish One Eyed Pinhead (OEP) gene with the β-actin promoter driving sense transcription and an inverted CMV promoter driving antisense transcription. The OEP open reading frame was directly cloned between the two promoters. (B–G) Compared to uninjected control embryos (UICs), CT-OEP injected embryos phenocopied OEP mutants with curved body axis, cyclopia (black arrow in C), and reduced notochord (red arrows in F compared to G). In 12% of the CT-OEP embryos, we observed a more severe phenotype with complete loss of eye development (red asterisk in E). (H) Graph showing % transient transgenics showing WT, classic OEP, or more severe phenotypes. n = 167. (I). Silencing of OEP is Dicer dependent. Embryos were injected as indicated in the absence or presence of Dicer MO or a control mismatch morpholino. **p < 0.005.

Figure 4. Silencing of miR-27a/b.

(A) Convergent silencing of the miR-27a/b genes with ubiquitin promoters driving both sense and antisense transcription. A synthetic DNA sequence encompassing precursor sequences for both miR-27a and miR-27b were directly cloned between the two promoters. (B) Compared to control embryos injected with dye only (DIC), lateral and ventral views show that CT-miR-27a/b injected embryos phenocopy miR-27a/b morphants (see Supplemental Figure 4) with defects in pharyngeal arch morphogenesis, craniofacial defects, and inhibition of pectoral fin outgrowth as shown by decreased staining of ECM (cartilage) with alcian blue. (C) Compared to DICs, craniofacial defects observed with alcian blue staining were observed in 8–9.5% of CT-miR-27a/b embryos using two different plasmid clones of the construct in (A). n = 191 for DIC, n = 458 for CT-miR-27a/b clone 1, and n = 272 for CT-miR-27a/b clone 2. All images are at 4 dpf.