Crosstalk between transcription factors in regulation of the human glutathione S-transferase P 1 gene expression in Me 45 melanoma cells

Aim. The human GSTP1 is a major enzyme of phase II detoxification in the most cell types. Aberrant expression of GSTP1 is associated with carcinogenesis and development of multidrug resistance. The GSTP1 gene expression is regulated at multiple levels including transcriptional, post-transcriptional and post-translational. We concentrated our attention on the transcriptional level of regulation. Methods. Transient transfection of Me45 melanoma cells with constructs containing the luciferase gene under the control of complete and truncated GSTP1 promoter was utilized to identify a role of different promoter regions in regulation of the gene transcription in Me45 cells. To identify the transcription factors, interacting with the GSTP1 promoter sites, the competitive EMSA and super shift assay were applied. Results. GSTP1 transcription in Me45 cells is positively regulated by binding NF-kB to the cognate site and ERb in complex with unknown protein to the ARE site; the complex of ERb with c-Fos negatively regulates the gene expression via CRE site. The interaction of c-Fos/ERb with GSTP1 CRE site and indirect interaction of ERb with GSTP1 ARE were identified. Conclusions. The positive regulation of the human GSTP1 gene in Me45 melanoma cells is exerted via NF-kB and ARE sites and the negative one via CRE site of the promoter. ERb is indirectly involved in the regulation of GSTP1 transcription. It is bound via c-Fos with CRE site and via unknown protein with ARE site.

Introduction.Glutathione S-transferases comprise a multigene superfamily of enzymes that catalyze the conjugation of electrophilic toxic compounds with glutathione, playing a key role in phase II of detoxification [1].The human GSTP1 isoform is a major GST isoenzyme in most cell types, except hepatocytes [2].Besides its typical role in detoxification it possesses other functions, including a ligandin function [3], modulation of signaling pathways [4], conjugation and transport of steroid hormones [5], and participates in dinitrosyl-diglutathionyl-iron complex storage and metabolism [6].Aberrant expression of GSTP1 is associated with carcinogenesis and development of multidrug resistance (MDR).
The regulation of the GSTP1 gene expression is in the focus of researchers and clinicians interests because the stimulation of GSTP1 expression is expected to be used as a preventive approach against cancer while its down-regulation is in need to counteract the development of MDR.The GSTP1 gene expression is regulated at multiple levels including transcriptional, post-transcriptional and post-translational [7].We concentrate our attention on the transcriptional level of regulation.Despite the vast literature devoted to GSTP1 enzyme the functional characteristics of responsive elements in gene promoter and tissue-specific peculiarities of their regulation are poorly understood.Moreover the previous investigations of molecular mechanisms involved in the GSTP1 regulation were focused mainly on breast cancer, leukemia and prostate cancer cells.In present research we performed a functional analysis of GSTP1 promoter in human melanoma cells Me45.We utilized truncated promoter constructions to compare the functional role of different cis-acting promoter elements and identified transcription factors binding the responsive elements by competitive EMSA (electrophoretic mobility shift assay) and supershift assay.
Transient transfection assay.Me45 cells were grown in 24-well plates to 60 % confluence and transfected with 500 ng of pGSTP together with 25 ng of pRL-TK (plasmid with Renilla Luciferase and Thómidine kinase Promoter) plasmid («Promega», USA) per well using Lipofectamine TM LTX and PLUS TM reagents («Invitrogen», USA).After 20 h the firefly and renilla luciferase activities were assessed using Dual Luciferase ® Reporter Assay System («Promega», USA).
The study of ARE, NF-kB-like, NF-kB, CRE and GATA binding sites interactions with nuclear proteins from Me45 cells.For identification of the transcription factors interacting with the GSTP1 promoter the electrophoretic mobility shift assay (EMSA) was applied.The ability of 20 bp promoter fragments, containing ARE, NF-kB-like, NF-kB, CRE and GATA sites to bind nuclear proteins from Me45 cells was examined in this experiment.Fig. 3 shows that all oligonucleotides form complexes with nuclear proteins.Specificity of the protein binding was assessed in a competition experiment, in which nuclear proteins were preincubated in 50-and 100-fold molar excess of unlabeled probe.In this experiment we determined, that ARE, NF-kB and CRE sites specifically bind nuclear proteins while NF-kB-like and GATA sites do not.One band observed in all elrctrophoregrams was non-specific because it was not eliminated in competitive experiments (fig.3, A, B, C).
Surprisingly, we did not find any proteins interacting with NF-kB-like element which was identified as a negative regulator of GSTP1 transcription in the transient transfection experiment.We suppose that the «negative» role of the NF-B-like element in GSTP1 transcription may be connected with the presence of palindrome GGGACCCtc in the region that may hinder an enchanceosome formation.
The region spanning nucleotides from -85 to +35 which is shown to be able to support the transcription of the reporter gene in Me45 cell at the level even higher than the full-length promoter is known to be a minimal promoter essential for the GSTP1 gene expression.This minimal promoter region contains ARE site which interacts with different transcription factors -AP-1 [15], Nrf2 [16], ERb [17] and RARa [18], depending on cell type.To identify the transcription factors acting on this site in Me45 cells we performed competitive EMSA with consensus oligos for AP-1, Maf (the DNA-binding component of Nrf2), ERb and RARa and supershift assay with antibodies for these transcription factors.Consistent with results shown in fig.4, A, a 50-and 100-fold molar excess of unlabeled consensus oligonucleotides for AP-1, Maf, estrogen receptor beta (ERb) and retinoic acid receptor (RAR) were not able to compete for the nuclear proteins binding to the ARE site.It means that AP-1, Maf, ERb and RARa do not interact with ARE site through their DNA-binding domains.To clarify these results the supershift experiment with polyclonal antibodies to c-Jun (cross-reactive to JunB and JunD), c-Fos (cross-reactive to FosB, Fra1 and Fra2), MafF/G/K, ER b and Nrf3 (the placenta-specific homolog of Nrf2) was performed.As indicated in fig.4, A, neither transcription factors Jun, Fos nor Maf and Nrf3 prevent the formation of specific complex of ARE site with a nuclear protein.Only ERb antibody prevents the whole complex formation resulting in appearance of a new complex with higher electrophoretic mobility.This result clearly indicates that in Me45 nuclear extracts ER b binds to the GSTP1 ARE site through another yet unknown protein and DNA-binding domain of ERb is not involved in these interactions.
The region of GSTP1 promoter from -405 to -105 contains NF-kB site and positively regulates the reporter gene transcription in Me45 cells.This site binds NF-kB in K562 leukemia cells and mediates the gene induction by TNFa [11].The results of the GSTP1 promoter NF-kB site binding assay are Fig. 3.In vitro binding of Me45 nuclear proteins to GSTP1 promoter sites: A -electrophoretic mobility shift assay, demonstrating the ability of Me45 nuclear proteins to form complexes with ARE, NF-kB, NF-kB-like, CRE and GATA sites; B -results of competitive EMSA demonstrating, that protein binding to NF-kB-like site is nonspecific; C -results of competitive EMSA demonstrating, that protein binding to GATA site is nonspecific; S -specific complex; NS -nonspecific complex summarized in fig.4, B. Two specific bands were observed in the reaction of genuine NF-kB site containing oligo with nuclear extract.The unlabeled NF-kB consensus was able to efficiently compete for the nuclear proteins from both specific complexes leading to suggestion that NF-kB binds to GSTP1 NF-kB site in this cell line.To clarify the matter, nuclear extract was incubated with polyclonal antibodies to p50 and p65 subunits of NF-kB before the probe was added to the EMSA reaction.In the supershift assay of NF-kB site two new bands were observed after the incubation with p50 antibody -one originated from the lower and one from the upper complex, providing the evidence that both complexes contain p50.The upper complex of nuclear proteins and NF-kB site contains the p50/p65 heterodimer, while the lower complex observed in binding reaction is the p50/p50.These data together with the results of transient transfection assay strongly suggest that NF-kB interacts with the human GSTP1 NF-kB site and up-regulates gene transcription in Me45 cells.
The negative regulatory element -1162 … -404 contains a CRE site and ATAAA-repeated sequence.It was previously reported that CRE site of GSTP1 mediates gene response to cAMP by interacting with CREB in Calu-6 lung cancer cells [19].Competitive EMSA was also conducted to determine which protein is a part of the DNA-protein complex formed by CRE Fig. 4. Analysis of the complexes formed by ARE, NF-B and CRE sites from the human GSTP1 promoter: A -ARE-protein complex formation was inhibited by unlabeled ARE site (cold probe) and by ERb antibody; B -NF-kB site forms two complexes with the nuclear proteins from Me45 cells; both complexes were disrupted by the cold probe and NF-kB consensus and supershifted by p50 antibody; p65 antibody disrupted only the upper complex; C -CRE site from the human GSTP1 promoter interacts with Me45 nuclear proteins and the complex formation can be inhibited by the cold probe; AP-1, but not CRE consensus compete with the CRE for the nuclear proteins and antibodies to Fos and ERb supershift the complexes site in Me45.Regarding the ability of CRE sites in different genes to interact with CREB [19] and AP-1 [20] proteins, consensus oligonucleotides for both transcription factors were utilized in the competitive EMSA.A representative autoradiograph in fig.4, C, shows, that CREB consensus oligonucleotide could not compete with GSTP1 promoter CRE for protein binding, however genuine oligonucleotide CRE and AP-1 consensus competed successfully.This suggests that CRE site forms the complex with AP-1 in Me45 cells.The supershift experiment with antibodies against the transcription factors known to interact directly or indirectly with CREs of other genes was performed to verify the results.Antibodies to c-Jun (cross-reactive to JunB and JunD), c-Fos (cross-reactive to FosB, Fra1 and Fra2), MafF/G/K, ERb and Nrf3 were utilized in this assay.The supershifted bands were observed after the incubation of nuclear extracts with Fos and ERb antibodies.The supershift analysis indicates that ERb together with Fos protein interacts with the human GSTP1 CRE in Me45 cells and this interaction has a negative regulatory effect.
The phenomenon that protein binding sites can be shared between different transcription factors is called a transcription factor crosstalk [20].It can be realized by interaction of a «noncanonical» transcription factor directly with a DNA sequence which has a partial homology to the binding sites of this and another transcription factors [20] or by protein-protein interactions of «noncanonical» transcription factor with a «genuine» protein bound to its recognition site.In case of the human GSTP1 promoter both types of crosstalk are present -noncanonical c-Fos together with ERb crosstalks with CREB at CRE site and ERb together with an unknown protein crosstalks with AP-1 at ARE site.In both cases CREB and Fos/Jun has an opposite effect on gene transcription.In case of the GSTP1 promoter this negative effect is seems to be potentiated by ERb binding which is known to repress Fos-driven transcription [21].In the present finding we identified ERb indirectly interacting with two promoter elements -CRE and ARE sites.It evidences for the importance of this protein for the formation of the enchanceosome on GSTP1 promoter.
The ER signaling mechanisms discussed until now provide an explanation for the regulation of genes lacking estrogen response element and requiring a second DNA-binding transcription factor to mediate ER association with the DNA.ERa and ERb have been shown to act in opposite ways at Fos/Jun-binding sites.In the presence of E2 ERa activates transcription via its AF-1 and AF-2 transactivating domains while ERb-E2 which lacks a functional AF-1inhibits the Fos/Jundependent transcription [22].We suggest that ERb exerts the similar inhibitory effect at CRE site of GSTP1 promoter.The role of ERb associated with an unknown protein at ARE site is different and may activate transcription.The dual function of ERb in regulation of different promoter elements may be considered in context of enchanceosome formation.
Conclusions.In the present research the transcriptional mechanisms controlling the basal level of GSTP1 expression in Me45 cells have been analyzed for the first time.The obtained data indicate that the GSTP1 transcription in this cell type is positively regulated by binding of NF-kB to -323 site and ERb in complex with unknown protein binding to the ARE site; the complex of ERb with c-Fos at CRE site negatively regulates the gene expression.The interaction of c-Fos/ERb with GSTP1 CRE site and indirect interaction of ER with GSTP1 ARE site have been discovered.
Results and discussion.Functional analysis of the GSTP1 promoter regions in Me45 cells.The structure of GSTP1 promoter is summarized in fig. 1.To identify the role of GSTP1 promoter regions in regulation of GSTP1 transcription in Me45 cells we utilized transient transfection assay with reporter constructs containing complete or truncated GSTP1 promoter fused to the firefly luciferase gene.For this purpose we designed the reporter constructs each lacking the DNA fragment with one transcription factor binding site (fig.2).The diagram in fig. 2 represents relative firefly luciferase activities in lysates of Me45 cells transfected with reporter constructs.Each bar in the graph represents the average of 3 independent experiments with triplications in each.Transfection of the largest vector (pGSTP1415) containing the GSTP1 promoter fragment from -1379 to +35 resulted in relatively high level of f-luc gene expression in Me45 cells.Deletion of the GSTP1-flanking region between -1379 and -1162, containing GATA-binding site, did not influence significantly the expression of the reporter gene.Deletion of the region from -1162 to -405, which contains CRE and ATA-AA-repeat, resulted in increase of f-luc expression approximately 1.8-fold in comparison with previous construct.Further deletion of the region from -405 to -105, containing NF-kB site, reduced the reporter gene expression 1.6-fold.Deletion of the region from -105 to -85, known as an NF-kB-like element, resulted in 1.5-fold increase of f-luc expression.Thereby, the results of the transient transfection experiments suggest the presence of the negative regulatory elements located in the regions from -1162 to -405 and from -105 to -85.Also it provided the evidence for the presence of the strong positive regulatory element located from -405 to -105.The