Novel intronic microRNA represses zebrafish myf5 promoter activity through silencing dickkopf-3 gene

Abstract

A strong, negative cis-element located at the first intron +502/+835 (I300) of zebrafish myf5 has been reported. To elucidate the molecular mechanism underlying this repression network, we microinjected zebrafish single-cell embryos with I300 RNA, resulting in the dramatic reduction of luciferase activity driven by the myf5 promoter. Within this I300 segment, we identified an intronic microRNA (miR-In300) located at +609/+632 and found that it was more highly expressed in the older mature somites than those newly formed, which negatively correlated with the distribution of zebrafish myf5 transcripts. We proved that miR-In300 suppressed the transcription of myf5 through abolishing myf5 promoter activity, and we subsequently identified the long isoform of the Dickkopf-3 gene (dkk3) as the target gene of miR-In300. We further found that injection of the dkk3-morpholinos (MOs) resulted in downregulation of myf5 transcripts in somites, whereas co-injection of myf5 mRNA with dkk3-MO1 enabled rescue of the defects induced by dkk3-MO1 alone. Finally, injection of miR-In300-MO enhanced both myf5 transcripts in somites and the level of Dkk3 protein in zebrafish embryos. Based on these findings, we concluded that miR-In300 binds to its target gene dkk3, which inhibits the translation of dkk3 mRNA and, in turn, suppresses zebrafish myf5 promoter activity.

Figure 1.

Intron 1 RNA is a repressive element for myf5-specific expression. (A) Plasmids used for the transient luciferase assay were microinjected into the one-cell stage of fertilized embryos. The luciferase gene was under-controlled in different promoters, including the myf5 promoter (pZmyf5 6.3R), myod promoter (pZmyoD 1.0R), CMV promoter (pCMV 0.7R) and TK promoter (pTK 0.7R). RNAs used for microinjection into zebrafish embryos were mRNA encoding red fluorescence protein (DsRed), sense strand of the first intron +502/+835 (I300) of zebrafish myf5 followed by DsRed mRNA (DsRed +502-> +835), and antisense strand of I300 followed by DsRed mRNA (DsRed +835-> +502). Luciferase activity was reduced only in the embryos co-injected with DsRed fused with sense I300 and the plasmid containing myf5 promoter. (B) Detection of the existence of primary transcript of I300 RNA by northern blot analysis. Using I300 RNA as the positive control, a positive signal (333 nt) was shown in the 16-hpf embryos, but not in the one-cell stage (0 hpf) embryos. The 5S rRNA (5srRNA) was served as a loading control.

Figure 2.

Expression patterns of zebrafish miR-In300. (A) miR-In300 is generated from intron 1 (+502/+2502) of the zebrafish myf5 gene. Pre-miR-In300 (+546/+644) and mature microRNA sequences (indicated in red; +609/+632) are presented. The predicted secondary structure of pre-miR-In300 is also illustrated. (B) Detection of the existence of miR-In300 transcript by northern blot analysis. The total RNAs extracted from the various stages of zebrafish embryos. The RNA level of miR-In300 was gradually increased in the embryos from 16 hpf to 20 hpf. The 5S rRNA (5srRNA) was served as a loading control. (C–F) Expression patterns of myf5, miR-In300 and miR-206 in somites at 20 hpf were detected by whole-mount in situ hybridization. myf5 was expressed only in the newly formed somites and in the presomitic mesoderm (PSM) at 20 hpf (C), whereas miR-In300 was predominant in the older formed somites, but only mildly present in the newly formed somites at 20 hpf (D). A muscle-specific microRNA in zebrafish, miR-206, was used as a positive control and was detected in mature muscle, but not in PSM at 20 hpf (E). In contrast, the antisense strand of myf5 intron 1 (A, red line) served as negative control and did not present any signal (F).

Figure 3.

Injection of the sense strand of I300 RNA containing miR-In300 repressed the target mRNA expression in embryos. (A) Two plasmids containing a luciferase reporter gene (hRluc) fused with five copies (5 X) of either a perfectly matched target sequence (PT) or a mutated-mismatched target sequence (mT) for miR-In300 were constructed. (B) Relative luciferase activities of zebrafish embryos microinjected with the materials as indicated. Co-injection of the sense strand of I300 RNA (DR_I+) with a vector DNA containing hRluc gene served as a standard for comparison to the luciferase activities driven by co-injection of DR_I+ with a plasmid DNA containing either 5 X PT or 5 X mT. In vivo transgenesis enabled the sense strand of I300 RNA to repress the gene expression of luciferase in the 5 X PT construct, but not in the 5 X mT construct.

Figure 4.

Change of miR-In300 level resulted in abnormal expression patterns of myf5. Expression patterns of myf5 (A and B) in somites at the 12–14 somite stage, as indicated. However, when miR-In300 dsRNA was injected, the expression of myf5 was greatly reduced in the newly formed somites and in the presomitic mesoderm (PSM), which became smaller (C and D). In wild-type embryos, myf5 was expressed in S-II to S1 somites (E and F), while in the miR-In300-MO-injected embryos, myf5 was expressed in S-II to S3 somites (G and H). Detection of the existence of miR-In300 transcripts by northern blot analysis. Results showed that miR-In300 was decreased in the miR-In300-MO-injected embryos at 16 hpf (I). The 5S rRNA (5srRNA) was served as a loading control.

Figure 5.

miR-In300 represses the translation of dkk3 mRNA. Schematic illustration of plasmid constructs used for microinjection. (A) Plasmid Tk-Rluc-dkk3 3′ UTR, in which the dkk3 3′ UTR was fused with the 3′ UTR of the luciferase gene and driven by TK promoter. Three putative miR-In300 binding sites, which are located at dkk3 3′ UTR, are indicated by empty boxes, and their possible targeted sequences are also presented. (B) Plasmid TK-Rluc-mdkk3 3′ UTR, which was the same as plasmid Tk-Rluc-dkk3 3′ UTR, except that the mutated target sequences (M1, M2 and M3) of dkk3 3′ UTR were included, are indicated by crossed boxes and the lowercase alphabet. (C) Luciferase activities in the microinjected embryos. TK-Rluc was constructed without ligating any sequence into the 3′ UTR region of the luciferase gene, which served as basal control. DsRed-I300 (+) was a sense strand RNA of I300. Compared to the embryos injected with TK-Rluc and I300 RNA, the luciferase activity of embryos injected with Tk-Rluc-dkk3 3′ UTR and I300 RNA decreased greatly, which was only 1/4th of the luciferase activity induced by injection of TK-Rluc and I300 RNA. However, the luciferase activities of embryos injected with I300 RNA and TK-Rluc-mdkk3 3′ UTR, which contained mutated sequences, remained unchanged, compared to the luciferase activity driven by I300 RNA and TK-Rluc. Co-injection of miR-In300-MO with Tk-Rluc-dkk3 3′ UTR showed that the luciferase activity was increased 3-fold over that of embryos injected with Tk-Rluc-dkk3 3′ UTR alone, without co-injection of miR-In300-MO.

Figure 6.

The myf5 promoter activity is controlled by the dkk3 gene. (A) Plasmid and MO, as indicated by +, were co-microinjected into the one-celled stage of fertilized embryos to carry out the transient luciferase assay. The luciferase activity driven by the upstream 6.3 kb of zebrafish myf5 promoter (pZmyf5 6.3R) and the co-injected dkk3-control-MO was measured in average (n = 6) and served as 100%. Compared to the embryos injected with pZmyf5 6.3R and dkk3-control-MO, the luciferase activity was greatly reduced in the embryos injected with pZmyf5 6.3R and dkk3-MO1. Meanwhile, the average of luciferase activity driven by pZmyf5 6.3R was also measured and served as 100%. Compared to the embryos injected with pZmyf5 6.3R, the luciferase activity was dramatically increased for the embryos in which the endogenous miR-In300 production had been inhibited by injection of miR-In300-MO. Zebrafish embryos derived from the transgenic line Tg (myf5 (80K): GFP), whose somites display GFP reporter, were used. (B) When dkk3-control-MO was injected, the GFP expression in somites remained unchanged at 16 hpf, whereas (C) the GFP was greatly reduced in somites when dkk3-MO1 was injected. (D–H) Whole-mount in situ hybridization of myf5 transcripts in zebrafish embryos at 16 hpf. Compared to the control (D), the myf5 signal in the somites of either dkk3-MO1-injected embryos (E) or miR-In300-dsRNA-injected embryos (G) was decreased. Co-injection of either dkk3-MO1 with dkk3 mRNA (F) or excess miR-In300-dsRNA with dkk3 mRNA (H) enabled embryos to rescue the defects induced by injecting either dkk3-MO1 alone (E) or miR-In300-dsRNA alone (G). The numbers shown in the lower-right corner of panels B, C, E, F, G and H indicate the number of phenotypes out of the number of embryos examined.

Figure 7.

Western blot analysis of Dkk3 protein during embryogenesis. (A) Total protein lysates were extracted from zebrafish embryos at 16, 18, 20, 24 or 30 hpf, as indicated. The molecular weight of protein markers and the positions of positive reactive bands for antiserum against Dkk3 and alpha-tubulin (A-tubulin) are also indicated. The protein level of Dkk3 was greatly reduced in the protein lysates extracted from embryos at 18 hpf. The Dkk3 protein became undetectable at 20, 24 and 30 hpf. The Dkk3 protein was reduced in the lysates extracted from the dkk3-MO1-injected embryos (lane 7), whereas the Dkk3 protein was greatly enhanced in the lysates extracted from the miR-In300-MO-injected embryos at 20 hpf (lane 8). The intensity of positive bands for antiserum against A-tubulin served as a loading control. (B) Possible model that illustrates the modulation of myf5 expression through miR-In300 mediation during somitogenesis in zebrafish embryos. The miR-In300 is an intronic microRNA within the first intron of zebrafish myf5, and dkk3 is the target gene of miR-In300. The miR-In300 blocks the target sequences located at the 3′ UTR of dkk3 mRNA, which results in repressing the translation of dkk3 mRNA. The absence of Dkk3 protein in zebrafish embryos then causes the downregulation of myf5 promoter activity through an unknown signal transduction pathway. Consequently, myf5 mRNA expression is gradually decreased at the later stages of zebrafish embryogenesis.