GREBP, a cGMP-response element-binding protein repressing the transcription of natriuretic peptide receptor 1 (NPR1/GCA)

Abstract

NPR1/GCA (natriuretic peptide receptor 1/guanylyl cyclase A) expression is controlled by several agents, including ANP (atrial natriuretic peptide). After ANP stimulation, NPR1/GCA down-regulates the transcriptional activity of its gene via a cGMP-dependent mechanism. Because we previously identified a cis-acting element responsible for this cGMP sensitivity, we proceed here to explore novel putative protein binding to cGMP-response element (cGMP-RE). Using the yeast one-hybrid technique with a human kidney cDNA library, we identified a strong positive clone able to bind cGMP-RE. The clone was derived from 1083-bp-long cDNA of a gene of yet unknown function localized on human chromosome 1 (1p33.36). We named this new protein GREBP (for cGMP-response element-binding protein). DNA binding assays showed 18-fold higher cGMP-RE binding capacity than the controls, whereas an electromobility shift assay indicated a specific binding for the cGMP-RE, and chromatin immunoprecipitation confirmed the binding of GREBP to the element under physiological conditions. By acting on cGMP-RE, GREBP inhibited the expression of a luciferase-coupled NPR1 promoter construct. In H295R cells, ANP heightened GREBP expression by 60% after just 3 h of treatment while inhibiting NPR1/GCA expression by 30%. Silencing GREBP with specific small interfering RNA increased the activity of the luciferase-coupled NPR1 promoter and GCA/NPR1 mRNA levels. GREBP is a nuclear protein mainly expressed in the heart. We report here the existence of a human-specific gene that acts as a transcriptional repressor of the NPR1/GCA gene.

FIGURE 1.

Identification of a cGMP-response element-binding protein. A, yeast one-hybrid screening of a human kidney cDNA library; the putative binding protein interacts with cGMP-RE (present in three copies) to activate the reporter gene via GAL4AD. B, full-length mRNA determination was by primer extension. HEK293 total RNA was retrotranscribed using a radiolabeled primer targeting GREBP mRNA and run beside pACT2-GREBP plasmid sequencing with the same radioactive primer. C, full-length GREBP mRNA was confirmed by RT-PCR with a specific primer targeting upstream and downstream regions of the transcription start site. The positive control was HEK293 DNA, whereas retrotranscribed HEK293 RNA was studied by −25/−5 and +1/+20 RT-PCR. D, nucleotide and amino acid sequences of GREBP. The mRNA destabilization motif is double underlined. Basic residues (framed) constitute 20% of GREBP (23 of 115) and are distributed along the whole sequence. Boldface residues are those with high possibility of serine phosphorylation (>0.800:1.00), and threonine could be strongly phosphorylated (>0.900:1.00) by serine/threonine kinase, including PKG.

FIGURE 2.

DNA binding assays. A, GST or GST-GREBP fusion proteins were immobilized on a glutathione-Sepharose matrix and incubated with ³²P-labeled, double-stranded cGMP-RE DNA probe. After draining and washing, a 250-fold excess of cold probe was applied to the matrix to compete for hot probe binding. B, results of three different experiments. Radioactive counts after elution of GST-GREBP, GST, and bead control columns are corrected for protein content (GST and GST-GREBP) and compared. The presence of GST and GST-GREBP proteins on the resin was confirmed by Western blotting (WB) with a specific antibody directed against GST. *, p = 0.012. N/A, not applicable; error bars, S.E.

FIGURE 3.

EMSA of GREBP interacting with cGMP-RE from human, rat, and mouse species. A, 5 μg of nuclear extracts from GREBP or Neo stably transfected HEK293 cells were incubated with 18-bp-long ³²P-labeled double-stranded consensus cGMP-RE oligonucleotides corresponding to human, rat, and mouse sequences (39) and were run on non-denaturing PAGE. The figure shows representative EMSA in which the three probes could bind to nuclear proteins (all lanes). The arrow indicates the presence of a protein band visible only with the human and rat RE probes in control (Neo) nuclear extracts (lane 2), which increased in intensity with nuclear extracts from GREBP-transfected cells (lanes 3 and 8). This binding is specific because it could be competed for by 50- and 100-fold excess cold probes (lanes 4, 5, 9, and 10). B, nuclear extracts of GREBP-expressing cells were used to investigate binding to 24-bp-long cGMP-RE oligonucleotides. Binding was visible only with the human 24-bp-long ³²P-labeled probe (H24; lane 3) that was more intense after GREBP compared with Neo overexpression (lane 2). The other 24-bp-long probes (rat, R24; mouse, M24; rat/mouse probe with the nucleotide A at position 11, P1; and rat/mouse probe with the nucleotide G at position 11, P2) showed only nonspecific faint bands (lanes 4–7). C, supershift experiments using nuclear extracts of GREBP-StrepTagII-expressing cells. The probe corresponding to the human cGMP-RE sequence was used here. Overexpression of GREBP-StrepTagII protein (lane 3) led to an increased binding over Neo (lane 2). The addition of 1 μg of StrepTagII antibody prior to the ³²P-labeled cGMP-RE inhibited DNA-protein interaction (lane 5), whereas the same amount of antibody added after the radiolabeled probe had almost no effect on the binding (lane 7). Bovine serum albumin used as a control added before the probe (lane 4) or after it (lane 6) did not alter the binding compared with GREBP-StrepTagII alone (lane 3). D, overexpression of GREBP-HisTag protein (lane 3) led to an increased binding of the human cGMP-RE probe over Neo (lane 2). The addition of 0.5 μg of HisTag antibody prior to the ³²P-labeled probe inhibited DNA-protein interaction (lane 5), whereas the same amount of antibody added after the probe had almost no effect on the binding (lane 7). Normal rabbit IgG added before (lane 4) or after (lane 6) had no effect on probe binding compared with GREBP-HisTag alone (lane 3). N.E., nuclear extracts; Comp., cold competitor; Anti-ST, antibody against StrepTagII; Anti-HT, antibody against polyhistidine tag.

FIGURE 4.

A, ChIP assays for GREBP·cGMP-RE complex. HEK293 cells were transfected with peGFP or peGFP-GREBP for 12 h and fixed with formaldehyde as described under “Experimental Procedures.” GFP-overexpressing cells were used as negative controls. Fixed DNA·protein complexes were immunoprecipitated with an anti-GFP antibody, and normal rabbit IgG served as (no antibody) negative control. Input represented one-tenth of cleared unprecipitated supernatant. B, cellular localization of GREBP in HEK239 cells. HEK293 cells were transfected with peGFP or peGFP-GREBP for 12 h. They were then fixed and stained with phalloidin and 4′,6-diamidino-2-phenylindole. The upper panel shows the whole cell distribution of GFP as control. The lower panels present nuclear localization of the chimeric protein GFP-GREBP. DAPI, 4′,6-diamidino-2-phenylindole.

FIGURE 5.

NPR1/GCA promoter activity is inhibited by GREBP overexpression, and this inhibition is dependent on cGMP-RE. A, HEK293 cells were co-transfected with 10 ng of pCMV-βgal, 200 ng of hGCAp-pGL3b, and a combination of varying amounts of pCDNA1-Neo and/or pCDNA1-Neo-GREBP for a total plasmid amount of 500 ng/test. GREBP overexpression was confirmed by sqRT-PCR with specific primers targeting GREBP expressed by pCDNA1-Neo-GREBP. B, HEK293 cells were co-transfected with 10 ng of pCMVβ, 400 ng of pCDNA1-Neo or pCDNA1-Neo-GREBP, and 200 ng of hGCAp-pGL3b or the deleted cGMP-RE form of the human NPR1/GCA promoter hGCAp(ΔcGMP-RE)-pGL3b. All values are expressed as means ± S.E. (error bars) of relative luciferase activity from four different experiments performed in triplicate (n = 12). *, p < 0.05; **, p < 0.0001 versus Neo alone.

FIGURE 6.

NPR1/GCA transcriptional regulation requires GREBP expression. HEK239 cells were transfected with the pSilencer plasmid producing an NT hairpin sequence (pSilencer 2.0-U6 NT) or shRNA directed against GREBP (pSilencer 2.0-U6-shRNA-GREBP). mRNA levels were analyzed for GREBP expression in A and for NPR1/GCA in B. C, HEK293 cells were co-transfected with 10 ng of pCMVβ, 200 ng of hGCAp-pGL3b, and 300 ng of pSilencer 2.0-U6 NT or of pSilencer 2.0-U6-shRNA-GREBP. All values are expressed as means ± S.E. of relative mRNA levels (A and B) or relative luciferase activity (C) from experiments performed in triplicate. *, p < 0.05; **, p < 0.01; ***, p < 0.001 versus NT. Error bars, S.E.

FIGURE 7.

ANP stimulation influences GREBP and NPR1/GCA expression patterns. A, HeLa cells were treated with 100 nm ANP for the indicated time, and GREBP expression was measured by sqRT-PCR corrected for the 18 S internal standard (from four different experiments). NCI-H295R cells were treated with 100 nm ANP, and gene expression was measured after 1, 3, 6, and 14 h and summarized in B for sqRT-PCR results and compiled in C for graphic representation. All data are from three different experiments performed in duplicate. Values are expressed as means ± S.E. (error bars) relative to the basal state after ANP stimulation. *, p < 0.05; **, p < 0.01; ***, p < 0.005.

FIGURE 8.

GREBP expression profile in human tissue samples. The human MTE array from Clontech is composed of tissues obtained from normal subjects (male/female) from 15 to 66 years old after sudden and/or accidental death. The fetal samples combine 30 embryos, average age 18–30 weeks. GREBP expression is ubiquitous, with the strongest signal in the heart and adult aorta.