Developmental defects by antisense-mediated inactivation of micro-RNAs 2 and 13 in Drosophila and the identification of putative target genes

Abstract

Micro-RNAs are a class of small non-coding regulatory RNAs that impair translation by imperfect base pairing to mRNAs. For analysis of their cellular function we injected different miRNA-specific DNA antisense oligonucleotides in Drosophila embryos. In four cases we observed severe interference with normal development, one had a moderate impact and six oligonucleotides did not cause detectable phenotypes. We further used the miR-13a DNA antisense oligonucleotide as a PCR primer on a cDNA library template. In this experimental way we identified nine Drosophila genes, which are characterised by 3' untranslated region motifs that allow imperfect duplex formation with miR-13 or related miRNAs. These genes, which include Sos and Myd88, represent putative targets for miRNA regulation. Mutagenesis of the target motif of two genes followed by transfection in Drosophila Schneider 2 (S2) cells and subsequent reporter gene analysis confirmed the hypothesis that the binding potential of miR-13 is inversely correlated with gene expression.

Figure 1

Developmental defects due to injection of DNA oligonucleotides complementary to miRNAs. All embryos are oriented with anterior to the left and ventral down. (A) Wild-type embryo just before hatching (buffer- injected control). Note the wild-type mouth parts (bracket) that are missing in all other embryos. (B) Example of non-specific developmental defects in buffer-injected embryos (loss/deformation of mouth parts). (C–F) Anti-mi-DNA injected embryos. Embryos injected with anti-miDNA-2a (C and D) or anti-miDNA-13a (E and F) consistently produce abdominal defects, such as cuticle holes (arrows) and disorganised denticle belts (arrowheads), seen clearly in the higher magnification detail in (F).

Figure 2

Schematic overview of miRNAs specific for interaction with the K box and the DNA antisense oligonucleotides used for injection or target gene identification. (A) The sequences of miR-2a, miR-2b, miR-13a and miR-13b are given; identical sequence elements are color-coded. The nucleotide which discriminates variant a and b of each miRNA are indicated as well. The anti-miDNA oligonucleotides used for injection are given and the positions of the mutations introduced to anti-miDNA-13a-7/14 are indicated. (B) Schematic interaction of the miRNAs with the K box consensus sequence as described (18). (C) Examples of base pairing interaction of a K box motif. The K1 box of E (spl) mδ is able to form an interaction with miR-11 within the family and the first name motif, while miR-13 can interact only via the family motif. (D) Strategy to identify target genes that are negatively regulated by miRNAs. (Top) The schematic map of a mRNA and the binding of a miRNA at the 3′ UTR (similar to B and C). Many miRNAs may pair perfectly via the family domain at their 5′ terminus, while the base pairing in the 3′ part of the miRNA is not continuous. Thus, the antisense DNA oligonucleotide will base pair perfectly via its 3′ domain allowing the initiation of a PCR on a cloned cDNA template (bottom), supported by additional base pairing in the 5′ domain. The second primer is specific for a vector sequence.

Figure 2

Schematic overview of miRNAs specific for interaction with the K box and the DNA antisense oligonucleotides used for injection or target gene identification. (A) The sequences of miR-2a, miR-2b, miR-13a and miR-13b are given; identical sequence elements are color-coded. The nucleotide which discriminates variant a and b of each miRNA are indicated as well. The anti-miDNA oligonucleotides used for injection are given and the positions of the mutations introduced to anti-miDNA-13a-7/14 are indicated. (B) Schematic interaction of the miRNAs with the K box consensus sequence as described (18). (C) Examples of base pairing interaction of a K box motif. The K1 box of E (spl) mδ is able to form an interaction with miR-11 within the family and the first name motif, while miR-13 can interact only via the family motif. (D) Strategy to identify target genes that are negatively regulated by miRNAs. (Top) The schematic map of a mRNA and the binding of a miRNA at the 3′ UTR (similar to B and C). Many miRNAs may pair perfectly via the family domain at their 5′ terminus, while the base pairing in the 3′ part of the miRNA is not continuous. Thus, the antisense DNA oligonucleotide will base pair perfectly via its 3′ domain allowing the initiation of a PCR on a cloned cDNA template (bottom), supported by additional base pairing in the 5′ domain. The second primer is specific for a vector sequence.

Figure 3

Altering the miR-13 target site in the 3′ UTR of gene CG10222 and its influence on gene expression. (Left) The 3′ UTR of gene CG10222 and three mutant constructs with its predicted base pairing with miR-13b. The 3′ UTRs were fused to a pact-GL3 reporter construct and transfected into S2 cells. Transfection efficiency was first normalised to a reporter plasmid expressing LacZ. (Right) Relative expression level in percent with respect to the wild-type 3′ UTR of gene CG1022, which was set to 100%. Average expression levels and standard deviations obtained from six independent experiments with 1800 (left bars) or 450 ng (right bars) are given. The relative effect of the mutations is stronger with the lower amount of transfected plasmid.