ALU repeats in promoters are position-dependent co-response elements (coRE) that enhance or repress transcription by dimeric and monomeric progesterone receptors

Abstract

We have conducted an in silico analysis of progesterone response elements (PRE) in progesterone receptor (PR) up-regulated promoters. Imperfect inverted repeats, direct repeats, and half-site PRE are widespread, not only in PR-regulated, but also in non-PR-regulated and random promoters. Few resemble the commonly used palindromic PRE with three nucleotide (nt) spacers. We speculated that PRE may be necessary but insufficient to control endogenous PR-dependent transcription. A search for PRE partners identified a highly conserved 234-nt sequence invariably located within 1-2 kb of transcription start sites. It resembles ALU repeats and contains binding sites for 11 transcription factors. The 234-nt sequence of the PR-regulated 8-oxoguanine DNA glycosylase promoter was cloned in the forward or reverse orientation in front of zero, one, or two inverted repeat PRE, and one or tandem PRE half-sites, driving luciferase. Under these conditions the 234-nt sequence functions as a co-response element (coRE). From the PRE or tandem half-sites, the reverse coRE is a strong activator of PR and glucocorticoid receptor-dependent transcription. The forward coRE is a powerful repressor. The prevalence of PRE half-sites in natural promoters suggested that PR monomers regulate transcription. Indeed, dimerization-domain mutant PR monomers were stronger transactivators than wild-type PR on PRE or tandem half-sites. This was repressed by the forward coRE. We propose that in natural promoters the coRE functions as a composite response element with imperfect PRE and half-sites to present variable, orientation-dependent transcription factors for interaction with nearby PR.

Figure 1

Position and spacing of putative PRE in 2000-nt proximal promoters of PR-up-regulated or random genes. A, 12 PR up-regulated (20,21,22) and B) 12 random genes were searched for 2000-nt promoter sequences using the EZ retrieve database (23). SRMS was used to define 16-nt conserved motifs and patterns within them. The MEME database was used for large-scale sequence alignment (27). The TRANSFAC database was specifically searched for consensus PRE using two matrices: V$GRE_C (red boxes) and V$GR_Q6 (yellow boxes). Note: the nucleotide numbering on the abscissa is the reverse of conventional promoter positions due to the default numbering in SRMS with 2000 at the putative transcription start site, and ″0″ designates -2000.

Figure 2

Composition of PRE in PR-regulated and random promoters, the TAT promoter and MMTV-LTR, and four experimentally confirmed PR-regulated promoters. A, SRMS was used to search 12 PR-up-regulated, 12 non-PR-regulated, and 12 random promoters for the presence of PRE using the TRANSFAC database and the 16-nt consensus PRE/GRE 5′-GGTACAANNTGTYCTK-3′, where Y = C or T; K = G or T. A total of 37 PR-regulated, 32 non-PR-regulated, and 23 random PRE were used to construct a logo illustrating sequence conservation at each of the 16 positions. Letter heights represent the degree of base conservation at that position. Four promoters containing five experimentally confirmed PRE (16,24,25,26) and the three PRE of MMTV-LTR and TAT were also analyzed. B, The sequences of 37 PRE in 12 PR-up-regulated promoters from the TRANSFAC database 16-nt V$GRE_C matrix are shown in the pileup above the derived logo. The statistical significance (P values) for each sequence is compared with the matrix sequence. The nucleotide location of each PRE and its presence on the sense (+) or antisense (−) strand is indicated on the left. Examples of PRE sequences include 1) imperfect indirect repeats (KCNG1, site 765–781); 2) direct repeats (SLC31A2, site 827–843); 3) perfect half-sites (vasodilator phosphoprotein, site 147–163); and 4) imperfect half-sites (P2RX4, site 52–68). C, In the same genes, the more stringent 19-nt V$GR_Q6 matrix identified seven PRE the pileup and summary logo of which are shown together with their statistical significance (P values) compared with the matrix sequence.

Figure 2

Composition of PRE in PR-regulated and random promoters, the TAT promoter and MMTV-LTR, and four experimentally confirmed PR-regulated promoters. A, SRMS was used to search 12 PR-up-regulated, 12 non-PR-regulated, and 12 random promoters for the presence of PRE using the TRANSFAC database and the 16-nt consensus PRE/GRE 5′-GGTACAANNTGTYCTK-3′, where Y = C or T; K = G or T. A total of 37 PR-regulated, 32 non-PR-regulated, and 23 random PRE were used to construct a logo illustrating sequence conservation at each of the 16 positions. Letter heights represent the degree of base conservation at that position. Four promoters containing five experimentally confirmed PRE (16,24,25,26) and the three PRE of MMTV-LTR and TAT were also analyzed. B, The sequences of 37 PRE in 12 PR-up-regulated promoters from the TRANSFAC database 16-nt V$GRE_C matrix are shown in the pileup above the derived logo. The statistical significance (P values) for each sequence is compared with the matrix sequence. The nucleotide location of each PRE and its presence on the sense (+) or antisense (−) strand is indicated on the left. Examples of PRE sequences include 1) imperfect indirect repeats (KCNG1, site 765–781); 2) direct repeats (SLC31A2, site 827–843); 3) perfect half-sites (vasodilator phosphoprotein, site 147–163); and 4) imperfect half-sites (P2RX4, site 52–68). C, In the same genes, the more stringent 19-nt V$GR_Q6 matrix identified seven PRE the pileup and summary logo of which are shown together with their statistical significance (P values) compared with the matrix sequence.

Figure 3

Location of PRE and coRE on 12 PR-up-regulated promoters and other promoter constructs. A, SRMS was used to locate 2000-nt proximal promoters of 12 PR-regulated genes, which were searched for PRE using the TRANSFAC database 16-nt consensus PRE V$GRE_C matrix (open red boxes) or the 19-nt experimental GRE V$Gr_Q6 matrix (open yellow boxes). A highly conserved pattern of seven motifs, each approximately 16 nt in length (each colored box), spanning about 234 bp total, was found on 11 of the 12 promoters. The 234-nt sequence is termed a “coRE”. Location of each coRE on the sense (+) and antisense (−) strand is indicated. The numbering on the abscissa is the reverse of standard promoter nomenclature with 2000 at the transcription start site, and 0 at −2000 bp. B, Luciferase reporters driven by promoters containing PRE and the coRE of the OGG1 promoter, cloned in the forward or reverse orientation as it aligns in OGG1, were made as described in Materials and Methods. Open red boxes denote ½, one, or two palindromic PRE derived from the TAT promoter with each half-site separated by a dashed line. Mutated sites are designated by X. Also shown are constructs containing the three half-sites of the MMTV-LTR. Open black box is the TATA box of the minimal tk promoter.

Figure 4

Enhancement or suppression of PRB-mediated transcription is dependent on the orientation of the coRE. A, The coRE of the OGG1 promoter in the fwd or rev orientation was cloned upstream of PRE2-tk-LUC, PRE1-tk-LUC, or tk-LUC. HeLa cells were transiently transfected with the reporters, an expression vector encoding PRB, and an simian virus 40-driven Renilla luciferase control vector. Cells were treated with 10 nm R5020 or ethanol for 24 h. Relative firefly luciferase units were corrected for transfection efficiency based on Renilla activity, and fold changes between control and hormone-driven activity were calculated. Error bars represent the sem of three independent studies. B, The coRE in the fwd or rev orientation was cloned upstream of one or two PRE half-sites or the three half-sites of the MMTV-LTR in front of tk-LUC. HeLa cells were cotransfected with the reporters, an expression vector encoding PRB, and a Renilla-LUC control, and treated with R5020 or ethanol. Fold transcription was calculated as above.

Figure 5

Effects of the coRE on transcription by other nuclear receptors. A, The PRE2-tk-LUC reporter without or with the coRE in the forward or reverse orientation was transiently transfected into HeLa cells together with a Renilla-LUC control, and cells were treated 24 h with 10 nm dexamethasone. B, The same reporters were transiently transfected into HeLa cells together with an expression plasmid encoding AR and the Renilla-LUC control. Cells were treated 24 h with 10 nm of the androgen R1881. C, The coRE in either orientation was cloned upstream of an ERE2-tk-LUC reporter, HeLa cells were transiently transfected with the reporter, an ERα expression vector, and Renilla-LUC and treated 24 h with 10 nm 17β-estradiol. Fold changes are shown for hormone treated vs. untreated transcription, corrected for the Renilla control. Error bars represent the sem of three independent studies.

Figure 6

A PRB dimerization mutant monomer up-regulates transcription of endogenous genes. T47D breast cancer cell lines that stably express wild-type PRB (YB) or PRB dimerization mutants (DBX136; DBX141), were treated 6 h without (−) or with (+) 10 nm progesterone and harvested and RNA was isolated. RT-PCR was performed with primers specific for tissue factor (F3), PNMT, or SGK. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is shown as a loading control. Prog, Progesterone.

Figure 7

Compared with wild-type PRB, DBX monomers strongly stimulate PRE-dependent transcription, which is modified by the coRE. A, HeLa cells were transiently transfected with the single or tandem PRE half-site tk-LUC, or single or tandem palindromic PRE tk-LUC, plus wild-type PRB or DBX dimerization mutants, and a Renilla-LUC vector. Cells were treated 24 h without or with 10 nm R5020. Luciferase activity was corrected for the Renilla control, and fold regulation by liganded vs. unliganded receptors was quantified. B, HeLa cells were transiently transfected with the single or tandem palindromic PRE tk-LUC without or with the coRE in the fwd or rev orientation. They were cotransfected with wild-type PRB or DBX expression vectors, and a Renilla-LUC vector, treated 24 h without or with 10 nm R5020, and quantified as above. C, HeLa cells were transiently transfected with the single or tandem PRE half-site tk-LUC without or with the coRE in the fwd or rev orientation. They were cotransfected with PRB or DBX expression vectors and a Renilla-LUC vector and treated 24 h without or with 10 nm R5020. Transcription was quantified as above. Error bars represent the sem of three independent studies.

Figure 8

The coRE of the OGG1 promoter contains 11 putative transcription factor-binding sites. Factor binding sites shown are variable in size and were identified using the TRANSFAC database.