Structure of the upstream promoter P1 of the ATF3 gene. (A) Primer extension analysis of mRNAs from the P1 promoter of the human ATF3 gene. A primer specific for the P1 transcript was annealed to 1 μg of polyA mRNAs isolated from the serum-treated HCT116 cells (lane 5), 1 μg yeast tRNA (lane 6) or none (lane 7), and the TSSs were determined as in Materials and methods section. Lanes 1–4 represent a sequence ladder. The arrows on the right indicate the TSSs, and the sequence is shown on the left. Bold letters represent multiple TSSs identified. (B) Alignment of the nucleotide sequence of the P1 promoter of the human and mouse ATF3 gene. Arrows indicate the multiple TSSs, and the numbers in parentheses represent the clones analyzed by 5′-RACE of the serum-stimulated cells. Putative TATA box is shown in the box. Multiple TSSs of the human and mouse genes were deposited in DDBJ with accession numbers AB291910 and AB291911, respectively.

Efficiency of translation of the ATF3 mRNA isoforms from the P1 and P2 promoters. (A) Expression reporter plasmids encoding the ATF3 transcripts with various 5′-UTRs from the P1 (upper panel) and P2 promoters (lower panel) are shown. Arrows a through e indicate the 5′-ends of mRNAs corresponding multiple TSS of the P1 promoter in Figure 2B, and arrows a and b indicate the 5′-ends the P2 transcripts obtained from the database and reference (27). (B) Each plasmid was assayed for the in vitro translation as described in the Materials and methods section. The relative expression of ATF3 to plasmid without UTR is shown. (C) 293 cells were transfected with each expression plasmid, cultured for 24 h, and the ATF3 mRNA and protein generated from the transgene was measured as in the Materials and methods section. Relative efficiency of translation is the ratio of the ATF3 protein normalized to mRNA compared to that without 5′-UTR. (D) 293 cells were transfected with each plasmid, starved with 0.5% serum, followed by stimulation with 20% serum for 24 h or 50 μg/ml MMS for 12 h. After measurement of both ATF3 protein and mRNA generated from the transgene, relative efficiency of translation is shown as in (C). All the data from (B) through (D) represent the means with standard error bar of three independent experiments.

Histone modifications across the human ATF3 gene locus in LNCaP and Hodgkin L428 cells. (A) ChIP assay was performed in LNCaP, DU145, L428, or DAUDI cells using anti-panacetyl H3, anti-trimethyl H3K4 or anti-trimethyl H3K9 antibodies as in Materials and methods section. The P1 and P2 promoter region of the ATF3 gene and the control GAPDH gene was amplified and analyzed by gel electrophoresis. In L428 and DAUDI cells, the promoter region of the human β-globin gene was also examined. (B) Immunoprecipitated DNA by anti-panacetyl H3 or anti-trimethyl H3K4 antibodies of LNCaP (filled circle, open circle), DU145 (filled traingle, open triangle) cells was measured by quantitative PCR throughout the human ATF3 gene locus, and expressed as percent of the input. Data represent the means of three independent experiments with standard error bars for control IgG (open) and specific antibodies (closed), respectively.

Conservation of the upstream alternate promoter of the ATF3 gene in human and mouse. (A) Schematic representation of the structure of the human and mouse ATF3 genes. Exon A is further devided into A-1 and A corresponding to the novel promoter recently identified (35) and that previously reported (27), respectively. (B) Total RNAs from serum-induced human HCT116 or mouse RAW264 cells were assayed for the expression of transcripts containing the exon A-1 or A by RT–PCR. (C) Specific bands in (B) were subcloned, sequenced and the nucleotide sequences of the junction between the exon A-1 and B of human and mouse are shown.

Differential response of the ATF3 alternate promoters to various stimuli. (A) HCT116 cells were serum-starved in the presence of 0.5% serum for 48 h, and then stimulated by 20% serum, or 100 μg/ml MMS. At each time indicated, the ATF3 mRNA isoforms from the P1 (filled circle) or P2 promoter (filled square) was measured. Relative expression of these transcripts was normalized to GAPDH as in Materials and methods section and Supplementary Figure S1. Lower panel shows a regular RT–PCR gel of the representative experiment. (B) HCT116 cells transfected with siRNA oligos control (siCtl), specific for the P1 (siP1) or P2 transcript (siP2) were stimulated with serum or MMS as in Materials and method section. At each time indicated, the expression of the P1, P2 transcript or ATF3 protein was measured. *P < 0.05 compared with control siCtl. (C) HCT116 cells were treated with 20 μg/ml etoposide (Etopo), 50 μM H₂O₂, 1 μM doxorubicin (Dox), 2 μM thapsigargin (Thap), 1 μM tunicamycin (Tuni) and HaCaT cells were treated with 100 pM TGF-β. At each time indicated, the expression of the P1, P2 transcript or ATF3 protein was measured. U2OS cells stably expressing the normal, oncogenic 12V or 61L mutants of HRAS were also assayed. (D) Cells transfected with pLuc-P1ATF3 (closed) or pLuc-P2ATF3 (open) were treated with various agents as in (C) for 16 h and assayed for luciferase activity. Data represent the fold induction compared to unstimulated cells. For U2OS cells expressing oncogenic HRAS, the fold induction of activity compared to the empty vector is shown. All the data from (A) through (D) represent means with standard error bars of three independent experiments.

Constitutive activation of the upstream P1 promoter in human prostate and Hodgkin RS cancer cells. (A) The expression level of ATF3 protein in DU145 (lane 1), PC-3 (lane 2), LNCaP (lane 3), DAUDI (lane 4), MOLT4 (lane 5), Hodgkin RS L428 (lane 6) or L540 (lane 7) cells was measured by western blot (upper panel). In the lower panel, total RNAs from these cells were assayed for the P1 or P2 transcripts as in Figure 3A. Relative expression of the ATF3 transcripts represents the ratio to that of GAPDH. (B) LNCaP, L428 or L540 cells were transfected with control siRNA oligos (siCtl, lane 1), or specific for the P1 (siP1, lane 2) or P2 transcript (siP2, lane 3), cultured for 48 h and assayed for the P1 and P2 transcripts or ATF3 protein. The relative expression of the P1 and P2 transcripts is shown as per cent of the control (siCtl). *P < 0.05 compared with control siCtl. (C) LNCaP (filled circle, open circle) or DU145 (filled triangle, open triangle) cells were fixed with formaldehyde and immunoprecipitated using anti-RNA polymerase II (upper panel) or anti-Leo1 antibody (lower panel) as in Materials and methods section. Immunoprecipitated DNA was measured by quantitative PCR throughout the human ATF3 gene locus and expressed as percent of the input DNA for the control IgG (open) and specific antibodies (closed), respectively. All the data from (A) through (C) represent the means of three independent experiments with standard error bars.