Derivation and characterization of Dicer- and microRNA-deficient human cells

Abstract

We have used genome editing to generate inactivating deletion mutations in all three copies of the dicer (hdcr) gene present in the human cell line 293T. As previously shown in murine ES cells lacking Dicer function, hDcr-deficient 293T cells are severely impaired for the production of mature microRNAs (miRNAs). Nevertheless, RNA-induced silencing complexes (RISCs) present in these hDcr-deficient cells are readily programmed by transfected, synthetic miRNA duplexes to repress mRNAs bearing either fully or partially complementary targets, including targets bearing incomplete seed homology to the introduced miRNA. Using these hDcr-deficient 293T cells, we demonstrate that human pre-miRNA processing can be effectively rescued by ectopic expression of the Drosophila Dicer 1 protein, but only in the presence of the PB isoform of Loquacious (Loqs-PB), the fly homolog of the hDcr cofactor TRBP. In contrast, Drosophila Dicer 2, even in the presence of its cofactors Loqs-PD and R2D2, was unable to support human pre-miRNA processing. Interestingly, although ectopic Drosophila Dicer 1/Loqs-PB or hDcr both rescued pre-miRNA processing effectively in these hDcr-deficient cells, there were significant differences in the ratio of the miRNA isoforms that were produced, especially in the case of miR-30 family members, and we also noted differences in the relative expression level of miRNAs vs. passenger strands for a subset of human miRNAs. These data demonstrate that the mechanisms underlying the accurate processing of pre-miRNAs are largely, but not entirely, conserved between mammalian and insect cells.

Keywords: Dicer; RISC; RNA interference; microRNAs; post-transcriptional regulation.

FIGURE 1.

Characterization of the NoDice cells. (A) Western analysis of endogenous hDcr expression in WT 293T cells and the NoDice(2-20) and NoDice(4-25) cell lines. (B) Northern analysis of pre-miR-155 and mature miR-155 expression in WT 293T cells and NoDice(2-20) or NoDice(4-25) cells transfected with a pri-miR-155 expression vector. In lanes 4 and 6, the NoDice cells were also cotransfected with an hDcr expression plasmid. (C) Indicator assay measuring the ability of a pri-miR-155 expression plasmid to repress an indicator plasmid consisting of RLuc linked to a 3′ UTR containing two artificial miR-155 target sites. The cells analyzed include WT 293T cells and the NoDice(2-20) and NoDice(4-25) cell lines. Data are normalized to the level of RLuc activity seen in each cell type in the absence of pri-miR-155. Average of three experiments with SD indicated. (D) Growth curve of WT 293T cells, NoDice(2-20) cells, and NoDice(4-25) cells. These data indicate that the doubling time has increased from ∼1.3 d for WT 293T to ∼2.5 d for both NoDice cell types.

FIGURE 2.

Analysis of small RNA expression in the NoDice cells. (A) Size distribution of small RNAs (16–27 nt) that align to known cellular pri-miRNA hairpins determined by total small RNA deep sequencing of WT 293T cells and the NoDice cells. (B) Similar to A, except that this figure presents data generated using qRT-PCR for a set of seven individual miRNA species that are expressed at readily detectable levels in WT 293T cells.

FIGURE 3.

Analysis of RISC-loaded small RNAs in the NoDice cells. (A) Western analysis of endogenous hDcr and Ago2 expression levels in WT 293T and NoDice(4-25) cells. β-actin was used as a loading control. (B) Size distribution of total small RNAs obtained by deep sequencing of RISC-associated small RNAs isolated by immunoprecipitation of Ago2 from 293T or NoDice(4-25) cells. (C) Similar to B except that only the reads that align to known human pri-miRNA hairpins are shown.

FIGURE 4.

Transfected synthetic miRNA duplexes efficiently program RISC in the NoDice(4-25) cells. NoDice(4-25) cells were cotransfected with an RLuc-based indicator plasmid containing two target sites perfectly (p) complementary to miR-92a, miR-155, or miR-K11 together with a synthetic miRNA duplex encoding one of these three miRNAs. RLuc activity was determined at 24 h post-transfection and was normalized to a control transfection using an irrelevant miRNA duplex as well as to a cotransfected Fluc control.

FIGURE 5.

PAR-CLIP of NoDice(4-25) cells transfected with a synthetic miR-155 or miR-K11 miRNA duplex identifies miRNA-specific RNA targets. NoDice(4-25) cells were transfected with a synthetic miR-155 or miR-K11 miRNA duplex and RISC-binding clusters present in one of the transfected cultures, but not in the other or in control NoDice(4-25) cells, identified by PAR-CLIP. We then generated RLuc-based indicator constructs for a number of these potential cellular target sites that displayed both differential RISC binding, as determined by PAR-CLIP, and different degrees of homology to miR-155 and miR-K11, which have an identical 7-nt seed sequence (see Supplemental Fig. S3). These indicators were then cotransfected into WT 293T cells together with the miR-155 or miR-K11 synthetic miRNA mimic (A) or with a pri-miR-155 or pri-miR-K11 expression vector (B). RLuc activity was determined at 24 h post-transfection. These data were normalized to an internal FLuc-based control and are given relative to the level of RLuc activity seen in the absence of an ectopic miRNA.

FIGURE 6.

Drosophila Dicer 1 and loquacious PB rescue pre-miRNA processing in NoDice(4-25) cells. (A) Western analysis of NoDice(4-25) cells transfected with vectors encoding His-tagged dDcr1 or dDcr2 or Flag-tagged R2D2, Loqs-PD, or Loqs-PB. Although dDcr1 and dDcr2 are similar in molecular mass, it has previously been observed that dDcr2 migrates more rapidly upon gel electrophoresis (Miyoshi et al. 2010). (B) Northern analysis of pre-miR-155 and mature miR-155 expression in NoDice(4-25) cells cotransfected with a pri-miR-155 expression vector and the indicated Dicer or Dicer cofactor expression vectors. (C) An analysis of the level of endogenous, mature miR-92a-3p expression detected in the NoDice(4-25) cells, or NoDice(4-25) cells transfected with the indicated expression vectors, as determined by qRT-PCR. These data were normalized to the parental 293T cell line, which was set at 100%, and are similar to what was observed by deep sequencing (Fig. 7B).

FIGURE 7.

Rescue of endogenous miRNA expression in NoDice(4-25) cells by expression of dDcr1 and its cofactor Loqs-PB. (A) Size distribution of small RNA species that align to human pri-miRNA hairpins in WT 293T cells, the parental NoDice(4-25) cells, and in NoDice(4-25) cells transfected with plasmids encoding full-length hDcr or expressing dDcr1 plus Loqs-PB, as determined by deep sequencing at 72 h post-transfection. (B) Relative expression of the 10 most highly expressed endogenous mature miRNA species in WT 293T cells, the parental NoDice(4-25) cells, or NoDice(4-25) cells expressing ectopic hDcr or dDcr1 plus Loqs-PB. These data were derived by deep sequencing of total small RNAs at 72 h post-transfection.