Dicer-dependent turnover of intergenic transcripts from the human beta-globin gene cluster

Abstract

The widespread occurrence of intergenic transcription in eukaryotes is increasingly evident. Intergenic transcription in the beta-globin gene cluster has been described in murine and human cells, and models for a role in gene and chromatin activation have been proposed. In this study, we analyze intergenic transcription and the chromatin state throughout the human beta-globin gene cluster and find that the data are not consistent with such activation-linked models. Thus, intergenic transcript levels correlate with neither chromatin activation nor globin gene expression. Instead, we find that intergenic transcripts of the beta-globin gene cluster are specifically upregulated in Dicer-deficient cells. This is accompanied by a shift towards more activated chromatin as indicated by changes in histone tail modifications. Our results strongly implicate RNA interference (RNAi)-related mechanisms in regulating intergenic transcription in the human beta-globin gene cluster and further suggest that RNAi-dependent chromatin silencing in vertebrates is not restricted to the centromeres.

FIG. 1.

β-Globin cluster transcript and chromatin profiles. (A) Adult erythroid cell (Fibach culture) transcript profile (real-time RT-PCR). Both sense (dark bars) and antisense (light bars) intergenic transcripts are shown on a logarithmic scale. (B) K562 transcript profile (real-time RT-PCR). Most of the data shown in panels A and B have been reproduced at least once to verify the reproducibility of the profiles (data not shown). (C) Histone H4 acetylation of adult erythroid cells (ChIP by real-time PCR). (D) K562 histone H4 acetylation (ChIP by real-time PCR). A schematic of the β-globin gene cluster is aligned below panel B to indicate the relative location of an amplicon. Suffixes “up” and “down” indicate a position 5′ and 3′, respectively, to the named element; the absence of a bar for “antisense” indicates that “antisense” was not measured at this element.

FIG. 2.

K562 β-globin cluster transcript profile following hemin (A) and butyrate (B) treatment (real-time RT-PCR). Due to significantly higher spliced ɛ- and γ-globin steady-state levels, these values are juxtaposed in separate graphs to the right. Only the sense strand profile is shown for clarity.

FIG. 3.

TSA induction of HeLa cells. (A) Both sense and antisense intergenic transcripts are induced as illustrated for “ERV down” located just upstream of the LCR. As the emphasis was on induction efficiency, “ERV down” sense and antisense intergenic transcripts are shown at different cycle numbers. (B) TSA also induces genic transcription as illustrated by spliced ɛ-globin transcripts. For actin normalization, see panel A.

FIG. 4.

Dicer knockdown considerably increases TSA-induced β-globin cluster intergenic and unspliced genic transcript levels in HeLa cells. (A) Dicer knockdown (RT-PCR analysis). (B) Dicer knockdown (Western analysis). (C) Effect of Dicer knockdown on the level of sense and antisense intergenic transcripts at “ERV down.” Appropriate actin controls are shown in panel A. (D) Dicer knockdown effect on TSA-induced transcript levels in the β-globin gene cluster (real-time RT-PCR). A star indicates that unspliced actin and tubulin levels are not comparable with those in the β-globin gene cluster. Only sense transcripts are shown for clarity; ▵ marks genic elements; respective primer designs to detect spliced and unspliced isoforms are indicated in the inset. Only the spliced isoform is amplified under conditions used with primers separated by the ∼850-bp intron 2 (see also Fig. 5). Suffixes “up” and “down” indicate positions 5′ and 3′, respectively, to the named element.

FIG. 5.

Upregulation of β-globin gene cluster transcripts is specific for TSA-induced unspliced genic and intergenic transcripts. (A) Effect of Dicer knockdown (kd) on actin and β-like globin intron 2 spliced/unspliced transcript levels. (B) Effect of Dicer knockdown on β-like globin intron 1 spliced/unspliced transcript levels. The upper and bottom arrows indicate unspliced and spliced transcripts, respectively.

FIG. 6.

Nuclear RNAi. (A) Diagram showing β-globin reporter construct with inserted PTB siRNA target sequence as well as experimental scheme for PTB siRNA-mediated knockdown experiments and analysis. (B) Fractionation (ethidium bromide-stained agarose gel) of nuclear and cytoplasmic RNAs following siRNA treatment. About five times more nuclear than cytoplasmic cell equivalents are shown. DNA size markers (values in nucleotides) were loaded in the first well. con, control. (C) Mitochondrially encoded cytochrome oxidase II message is largely recovered in the cytoplasmic fraction. (D) Dicer siRNA treatment effect on nuclear and cytoplasmic Dicer transcript levels (spliced). (E) Spliced PTB is equally knocked down in both fractions. (F) Unspliced PTB transcript levels are unchanged following PTB siRNA treatment. (G) A smaller but highly significant PTB siRNA-dependent decrease of PTB readthrough transcripts is observed in the same fractions. (H) Analysis of the same fractions as shown in panel B and employed in panels E to H shows that a plasmid-derived spliced message is only knocked down in the cytoplasmic fraction. (I) Readthrough transcripts derived from the same plasmid are reduced in an siRNA-dependent manner. (J) ɛ-Globin siRNA treatment causes ɛ-globin transcript knockdown in both nuclear and cytoplasmic fractions. (C to J) Real-time RT-PCR analyses. Values on top of the bars indicate severalfold knockdown of analyzed RNAs; RNA was harvested 16 h after siRNA transfection. (B to I) HeLa. (J) K562. *P < 0.001 (Student's t test, paired).

FIG. 7.

Chromatin activation of the β-globin gene cluster after Dicer knockdown in HeLa cells (ChIP analysis by real-time PCR). (A) Time course analysis for “ERV down” shows increased histone H4 acetylation in Dicer knockdown (kd) cells before, during, and especially after treatment with TSA. (B) Widespread increase in histone H4 acetylation in Dicer knockdown cells in the absence of TSA. ORF, open reading frame. (C) Histone H3K9 acetylation remains unchanged in Dicer knockdown cells except for a decrease for GAPDH. DHFR, dihydrofolate reductase. (D) Histone H3K4 dimethylation is generally increased in Dicer knockdown cells. Values on the top of bars indicate the ratio of real-time PCR signal for Dicer knockdown to control siRNA-treated cells.