Fig. 4. Analysis of endogenous BRF1 and TTP mRNA expression. Semi-quantitative RT–PCR was carried out with RNA from wt, slowA, slowB and slowC cells using primers specific for BRF1 (upper panel), TTP (middle panel) and, as internal control, glyceraldehyde-3-phosphate dehydrogenase (G3PDH, lower panel). To allow for comparison of the expression levels, the number of PCR cycles was kept below saturation of the amplification reaction (25 cycles for BRF1, 25 for TTP and 20 for GAPDH).

Fig. 7. In vitro binding of BRF1 to ARE-containing RNA. Gel mobility shift assays were performed with 5–50 ng of recombinant BRF1_wt protein (lanes 2–5 and 8) or 50 ng of BRF1_C120R protein (lanes 6 and 9). Radiolabelled ARE⁺ RNA (lanes 1–6) was synthesized from a 259 nt-long region of the IL-3 3′UTR containing the ARE. For ARE^– RNA (lanes 7–9), the 33 nt-long ARE has been deleted from the same region. Samples were resolved by non-denaturing 4% PAGE.

Fig. 8. Intracellular localization of BRF1 and TTP. (A) NIH 3T3 cells were transiently cotransfected with pTRE-BRF1 together with pTET-On. Expression of BRF1 was induced with 2 µg/ml doxycycline, and the myc-tagged protein was detected by immunofluorescence using 9E10 antibody at 60× magnification. (B) Localization of TTP was analyzed as in (A) by transient cotransfection of pTRE-TTP and pTET-On. (C) As a control for specificity of the 9E10 antibody, pcDNA-HuR-myc/His.B was transfected encoding myc-tagged HuR, a protein predominantly localized in the nucleus. The same area is shown by phase contrast in the right panel.

Fig. 3. In vitro transcription and translation of BRF1 cDNA amplified from slowA, slowB, slowC and wt cells. Use of a T7 promoter-linkered upstream primer allowed for in vitro RNA synthesis. Translation was carried out in reticulocyte lysate with 400 ng of RNA in the presence of [³⁵S]methionine. Protein products were resolved through 15% SDS–PAGE and visualized by direct autoradiography (upper panel). For control, an equivalent amount of input RNA was analysed by 1% agarose/formaldehyde gel electrophoresis, and stained with ethidium bromide (lower panel).

Fig. 1. Schematic representation of the strategy to clone a regulator of ARE-dependent mRNA degradation. (A) A GFP reporter gene linked to the 3′UTR of IL-3 was stably transfected into HT1080 cells, and a diploid subclone, referred to as wt, was chosen. GFPIL3 reporter mRNA decays rapidly in wt cells due to the ARE in the 3′UTR, and correspondingly, GFP expression is low. (B) Upon mutagenesis using frame-shift inducing compound ICR191, high GFP-expressing cells were isolated by FACS, subsequent cloning and direct fluorescence microscopy. Three mutant cell lines, in which GFP overexpression is due to a defect in degradation of GFPIL3 mRNA, were identified and termed slowA, slowB and slowC. (C) For rescue of mutant slowC, a cDNA library generated from wt cells was packaged into retrovirus and used for infection. Low GFP-expressing cells were isolated and screened for revertants with reconstituted decay of the GFPIL3 mRNA. (D) Provirally integrated cDNAs were recovered from the revertants by genomic PCR which allowed to clone the regulator.

Fig. 5. Inhibition of BRF1 expression by siRNA. (A) wt, slowB and slowC cells were transfected either with medium alone or with a 21-mer ds RNA BRF1-oligo. Cytoplasmic RNA was isolated 72 h later, and northern blotting analysis was performed to determine BRF1 mRNA expression levels. For quantification (bottom panel), signal intensities were measured by phosphoimager, normalized to actin and represented as a percentage of the control values from transfections with medium alone. (B) Upon transfection of wt cells with either medium alone, the BRF1-oligo or a control-oligo, stability of GFPIL3 reporter mRNA was monitored by actinomycin D treatment (5 µg/ml) for 0, 1 and 2 h. Northern blot analysis and quantification (right panel) was carried out as described for Figure 2. (C) GFP expression was measured by flow cytometry 72 h after transfection of medium alone or the BRF1-oligo into in wt cells, (D) slowB and (E) slowC.

Fig. 2. Analysis of GFPIL3 reporter gene expression in (A) unmutagenized wt cells; (B) slowC-R46, a clone derived from mutant slowC expressing the murine ecotropic retroviral receptor; and in the following clones selected for low GFP expression after cDNA library transfer: (C) CM81; (D) CM128; (E) CM1; (F) CM14. (G) slowC-BRF1-6 and (H) slowC-BRF1-13 are subclones of mutant slowC transfected with puCombi(SV40/BRF-1), a BRF1 cDNA expression vector. GFP levels were measured by flow cytometry (left panels), and stability of GFPIL3 mRNA was determined by actinomycin D chase experiments using northern blotting analysis (middle panels). Twenty micrograms of total RNA per lane was resolved by 1.1% agarose/formaldehyde gel electrophoresis. GFPIL3 mRNA was detected using a radiolabelled SP6-probe specific for the IL-3 3′UTR, and the membranes were rehybridized with a β-actin probe for loading control. GFPIL3 mRNA signal intensities were quantified by phosphoimager, normalized to the actin signal, and plotted as a percentage of the initial value against time (right panels).

Fig. 6. Influence of BRF1 on mRNA decay using a Tet-Off system. (A) Plasmid pSRL contains a β-globin reporter gene fused to the 3′UTR of IL-3 under transcriptional control of the tetracycline responsive element. NIH 3T3-B2A2-23 cells were transiently transfected with a constant amount of pSRL (12.5 µg) in combination with empty vector (7.5 µg, lanes 1–3), bsd-HisBRF1_wt (1.5 µg, lanes 4–6), rCD2-p110 (6.5 µg, lanes 7–9) or rCD2-p110 and bsd-HisBRF1_wt together (lanes 10–12). After 48 h, reporter gene transcription was blocked with tetracycline (2 µg/ml), and RNA was isolated at the time points indicated. Northern blotting analysis was performed as described in the legend of Figure 2 using a β-globin-specific probe for detection of the reporter mRNA. (B) The same co-transfections were carried out with 12.5 µg of reporter plasmid pSRL-ΔAU lacking the ARE. (C) Zinc-finger mutants of BRF1 were analyzed by cotransfection of pSRL together with bsd-HisBRF1_C120R (lanes 1–3) or bsd-HisBRF1_C158R (lanes 4–6). (D) For quantification of globin-IL3 mRNA decay, signal intensities from (A) and (C), as well as from independent repeat transfections were measured by phosphoimager, normalized to actin, and plotted as mean values ± SE against time. Number of repeat experiments: n = 12 for control, 8 for p110, 6 for p110+BRF1_wt, and 4 for p110+BRF1_C120R and p110+BRF1_C158R.