ABSTRACT
We have previously shown that the widely expressed human transcription factor TCF11/LCR-F1/Nrf1 interacts with small Maf proteins and binds to a subclass of AP1-sites. Such sites are required for [beta]-globin 5' DNase I hypersensitive site 2 enhancer activity, erythroid porphobilinogen deaminase inducibility, hemin responsiveness by heme-oxygenase 1 and expression of the gene NAD(P)H:quinone oxidoreductase1. Here we report the optimal DNA-binding sequences for TCF11/LCR-F1/Nrf1 alone and as a heterodimer with MafG, identified by using binding-site selection. The heterodimer recognises a 5'-TGCTgaGTCAT-3' binding-site that is identical to the established NF-E2-site, the antioxidant response element and the heme-responsive element while the binding specificity of the homomer is less stringent. To investigate the activity of TCF11 through this selected site, both alone and in the presence of MafG, we have used a transient transfection assay. TCF11 alone activates transcription while MafG alone acts as a repressor. When co-expressed, MafG interferes with TCF11 transactivation in a dose dependent manner. This indicates that MafG protein, which heterodimerises efficiently with TCF11 in vitro (the heterodimer having a higher affinity for DNA than TCF11 alone), does not co-operate with TCF11 in transactivating transcription. We propose that since both these factors are widely expressed, they may act together to contribute to the negative regulation of this specific target site. Efficient positive regulation by TCF11 may require alternative partners with perhaps more restricted expression patterns.
CNC-bZIP proteins are identified by an ~40 amino acid homology region immediately N-terminal to the basic region-leucine zipper (bZIP) domain. The Drosophila melanogaster homeotic selector gene cap `n' collar (1,2) encodes the first bZIP-factor identified that contained this region. Other family members include Skn1, a basic-region transcription factor required for correct specification of certain blastomere fates in early Caenorhabditis elegans embryos (3,4) and three human proteins; p45 NF-E2, an erythroid-specific activator proposed to regulate the [beta]-globins (5,6) and the more widely expressed TCF11/LCR-F1/Nrf1 (7-11) [hereafter referred to as TCF11 (transcription factor 11)] and Nrf2 (12). Homologous and related genes have also been cloned in other vertebrate species (13-15). The leucine zipper, which is responsible for homo- and heterodimerisation is not particularly conserved among CNC-bZIP family members (8). However, all CNC-bZIP family members tested so far preferentially form heterodimers with the same group of small Maf proteins (11,13,14,16). The Maf family of bZIP factors, the prototype of which (v-Maf) is responsible for the transforming activity present in an avian retrovirus (17), is subdivided into two groups based on primary sequence and ability to activate transcription. The large Maf family members are transactivators and regulate genes important in neuronal differentiation (18-20) whereas the small Maf proteins, MafF, MafG and MafK/p18, are widely expressed and transrepress transcription when bound to Maf responsive elements [MAREs; 5'-TGCTGAC(G)TCAGCA-3'], probably as homodimers (16,21-25).
We have previously shown that chicken Maf proteins MafF, -G and -K, but not the large activators v-Maf or MafB, specifically interact with several TCF11 protein isoforms in vitro (11). Both TCF11 and the Maf proteins -F, -G and -K/p18 are widely expressed, implying that these proteins may interact in numerous cell types (8,21,23,26,27). Heterodimers of small Maf proteins and CNC-bZIP domain family members or Fos, bind preferentially to a site containing consensus sequences for both Maf homodimer and AP1 binding called the AP1/MARE-site (5'-TGCTGAGTCAT/C-3') (11,13,14,16,21). This is in fact a classical AP1 site with a 5'-TGC extension, as is found in the NF-E2 site implicated in the regulation of erythroid specific gene expression. In murine erythroleukemia (MEL) cells, one of the binding complexes that specifically recognises this site was isolated and shown to consist of a heterodimer between the CNC-bZIP domain family member p45 NF-E2 and the small Maf homologue p18 (22). Several other AP1/MARE-sites require the 5'-TGC triplet for their correct function, and as yet unidentified binding-activity to these sites has been detected in a range of cell-types (6,22,27-30) including F9 cells (28) which do not express NF-E2 or AP1 factors.
Using DNA affinity chromatography we observed sequence-specific binding activity of the endogenous TCF11 isoforms p47/49 to the NF-E2-site in K562 cells (11). In vitro binding-site selection experiments presented here show that heterodimerisation dramatically increases the DNA-binding potential of TCF11. We observed a 5'-TGCTgaGTCAT-3' binding-site that is identical to the NF-E2-site (6), the antioxidant response element (ARE) (31) and the heme response element (HRE) (30). TCF11 alone shows only limited sequence-specificity. The Maf halfsite recognised by the heterodimer contains the previously described 5'-TGC AP1 site-extension. This TCF11:Maf-site represents the first selected CNC-bZIP heterodimer binding-site and can aid in identifying genes regulated by this presumably widespread heterodimer. The functional relevance of this binding has been tested in transfected COS 1 cells and it was found that while TCF11 can transactivate a reporter construct through the NF-E2 site when expressed alone, MafG co-expression interferes with this activation in a dose dependent manner. This is in contrast to the co-operative activation observed for NF-E2:small Maf heterodimers in NIH3T3 cells (16) and Ech:small Maf heterodimers in quail fibroblasts (21) demonstrating that different heterodimeric combinations of CNC-bZIP factors and small Mafs may have different activities through the same or similar target sequences.
Standard methods in molecular biology were used (32).
MBP-TCF11-A is a fusion protein between Escherichia coli maltose-binding protein (MBP) and the 300 C-terminal amino acids of human TCF11, including the CNC-bZIP region, constructed in the plasmid pMALc [New England Biolabs (NEB)] as previously described (11). MBP-MafG is an MBP fusion protein with chicken MafG (21), which has only one conserved amino acid substitution compared to human MafG in the bZIP domain. Production of fusion proteins and their isolation by amylose affinity chromatography were performed according to NEB protocols. Proteins were stored at -20°C in protein elution buffer/25% ethylene glycol (Pierce) (11).
The reporter constructs for cell transfections were produced by cloning; (i) the 900 bp EcoRI-BamHI fragment (PBGD1.5Luc), (ii) the 320 bp AccI-BamHI fragment (PBGD3.2Luc) or (iii) the 180 bp PvuII-BamHI fragment (PBGD5.1Luc), of the porphobilinogen deaminase (PBGD) gene erythroid-specific promoter (33) in front of the firefly luciferase gene in the pGL3 enhancer vector (Promega). Site directed mutagenesis of the NF-E2 site within construct PBGD3.2Luc was performed using the Stratagene `Quick change' mutagenesis kit. Expression constructs were produced by cloning the full length coding sequence of TCF11 (8) (5' to the EcoRV site at bp 3550) or MafG wild-type and mutant form ([Delta]L2PM4P, 21) into the expression vector pCDNA3 (Invitrogen).
The DNA library R76 (36) consists of 26 randomised nucleotides flanked by 25 bp constant regions used for PCR amplification and subcloning: 5'-CAGGTCAGTTCAGCGGATCCTGTCG(G/A/T/C)26GAGGCGAATTCAGTGCAACTGCAGC-3', synthesised by Dr Eshrat Babaie, The Biotechnology Centre of Oslo. It was rendered double-stranded using the Klenow fragment and primer F (5'-GCTGCAGTTGCACTGAATTCGCCTC-3'). Primer R is 5'-CAGGTCAGTTCAGCGGATCCTGTCG-3'. Double-stranded R76 was purified by polyacrylamide electrophoresis (32) and suspended in TE (10 mM Tris-HCl, 1 mM EDTA, pH 8.0). MBP-TCF11-A (1.3 pmol) (and MBP-MafG when heterodimer binding-sites were selected) was incubated in binding buffer [5 mM Tris-HCl, pH 8.0, 75 mM NaCl, 2.5 mM MgCl2, 0.5 mM EDTA, 5% glycerol, 2.5% ethylene glycol, 1% Tween-20, 1 mM DTT and 100 µg/ml poly(dIdC)] for 10 min at room temperature before the addition of 0.1 nmol dsR76 in a total volume of 20 µl. DNA-binding was allowed at room temperature for 20 min. Protein-DNA was electrophoresed through a 5% polyacrylamide gel (acrylamide:bisacrylamide 36:1) in 0.4* TBE (1* TBE is 89 m Tris-borate, pH 8.3, and 2 mM EDTA) at 10 V/cm for 10 min. The upper 0.5 cm of the gel, containing protein-DNA complexes, was excised. Protein-bound oligonucleotides were isolated by crushing the gel-slice and incubating in 400 µl NEMS solution (0.5 M NH4Ac, 10 mM MgAc, 1 mM EDTA and 0.1% SDS) for 12-15 h at 37°C with agitation. Oligonucleotides present in the supernatant were precipitated by the addition of 1 ml ice-cold ethanol (2.5 vol) and 20 µg dextran, as carrier. The DNA was collected by centrifugation (13 000 r.p.m. for 30 min). Pelleted DNA was washed briefly in cold 70% ethanol and suspended in 10 µl ddH2O and amplified by nine cycles of PCR using 0.1 nmol of each primer and Taq polymerase (Gibco) in a total volume of 20 µl. NEMS solution was added to the PCR-mix (final volume 200 µl), and DNA precipitated and washed as described. Binding-site selection was then repeated as described, including 1 µl (10 µCi/ml) [[alpha]-32P]dATP (Amersham) in the PCR reaction for radiolabelling of DNA. Electrophoresis time in the second and subsequent rounds was extended to 75 min to resolve MBP-TCF11-A:MBP-MafG heterodimers from MBP-MafG homodimers. Protein-bound DNA was identified by autoradiography. DNA from round three was digested with EcoRI and BamHI and ligated into the corresponding sites of pBluescript SKII+ (Stratagene). Plasmids were transformed into E.coli DH5[alpha] and the nucleotide sequences of individual clones were analysed (Tables 1 and 2).
The various cell lines (acquired from ATCC) were grown in the recommended media at 37°C in 5% CO2. At ~60-70% confluence the cells were transfected with plasmid DNA by a standard calcium phosphate precipitation method (34) using a total of 10-15 µg of DNA per 9 cm culture dish. The DNA mixture consisted typically of 2 µg of luciferase reporter construct, 1 µg of internal control plasmid (either pRSV-CAT or pEF[beta]-gal), various amounts of the TCF11 and/or MafG expression constructs and empty vector to the required total weight. The cultures were grown for 48 h after transfection before harvesting and enzyme assays were performed. Luciferase activity was measured on a Lumat LB 9507 luminometer using 5-50 µl of cell extract (from a total of 700 µl) brought to a total volume of 200 µl in buffer containing 10 mM Mg(OAc), 50 mM Tris-MES, pH 7.8, 2 mM ATP. Aliquots (100 µl) of 1 mM luciferin (Sigma L6882) were added for each measurement. The Luciferase activity for each culture was normalised to the activity of the internal control, either CAT or [beta]-galactosidase. The [beta]-galactosidase control was used in experiments with MafG expression vector since MafG was found to interfere with CAT expression in the pRSV-CAT construct. CAT activity was measured using a standard protocol (35) and [beta]-galactosidase activity was measured using ONPG as substrate and a colorimetric assay.
The approach used to select optimal binding sequences involved incubating the protein with a pool of degenerate oligonucleotides followed by isolation of the bound complexes on a polyacrylamide gel and PCR amplification of the retained oligonucleotide fraction. This binding, selection and amplification is repeated cyclically before selected oligonucleotides are cloned, sequenced and analysed. The procedure has been widely applied to identify transcription factor binding-sites (24,36,37). To retain maximum complexity of the selected fraction we have limited the numbers of selection-amplification cycles. The formation of protein-DNA complexes was monitored throughout the experiment. We observed a strong increase in binding from the second to the third round and after three rounds of selection and amplification, no further increase in DNA-binding was obtained (Fig. 1). The amplified binding-sites were therefore cloned after the third round.
The TCF11 homomer, bound to the library DNA, did not enter a polyacrylamide gel upon electrophoresis but remained in the well-region. This was consistent with previous observations using specific oligonucleotides and confirmed that it was not an artefact resulting from the nature of the synthetic oligonucleotide probes that were previously used (11). The limited mobility of the TCF11 binding complex suggests that the fusion protein MBP-TCF11-A binds to DNA in a multimeric form. The sequences of oligonucleotides within this complex showed that the AP1 halfsite 5'-GTCAT was represented four times more frequently than expected at random, indicating some specific interaction with DNA (Table 1), however palindromic or semi-palindromic sites were not abundant. It is interesting to note that similar halfsites have been selected using C.elegans Skn-1, a transcription factor with 65% similarity to TCF11 in the DNA-binding region that lacks a leucine zipper and does not dimerise (4). It appears, therefore, given the absence of palindromic sites, that MBP-TCF11-A in vitro does not interact with DNA as a bipartite structure formed by conventional dimerisation through the bZIP domain, but that a homomeric complex is capable of interacting with AP1 halfsites. The fact that we did not select any tandem repeated sites, suggests that at least in vitro, MBP-TCF11-A alone has no strong preference for repeated motifs.
The MBP-TCF11-A:MBP-MafG heterodimer migrates slightly slower than the MBP-MafG homodimer (Fig. 1B) on a polyacrylamide gel. It was, therefore, possible to isolate sequences bound to either of the complexes. When the two proteins are mixed they preferentially form heterodimers, the preference being obvious by the second round of selection (not shown). Three rounds of selection generated an oligonucleotide pool that bound strongly to the proteins (Fig. 1B). Of the 36 DNA sequences that were inspected (Table 2A), 31 sites (86%) contained a sequence consistent with the deduced TCF11:MafG core consensus sequence 5'-TGCTgaGTCAT-3' (Table 2B and D), showing that binding of the heterodimer is highly specific.
Of the TCF11:MafG-selected oligonucleotides, 55% comprised nucleotides in the constant (primer) region. The primer sequences were utilised only as Maf halfsites indicating that the MafG half of the heterodimer is more promiscuous than TCF11 in its sequence specificity (see Materials and Methods for further details). The 5'-TCA triplet at positions 5-7 was 100% conserved in all oligonucleotides. This shows that heterodimerising with a small Maf protein renders TCF11 more stringent in AP1-site binding preference. Maf binding sites have previously been divided into two groups; cyclic AMP responsive element (CRE)-type (TGACGTCA) and TPA-responsive element (TRE)-type (TGAGTCA) (24). In the current experiments all TCF11:MafG-selected sites showed TRE-like halfsite spacing. This is in contrast to the previously reported EMSA competition experiment in K562 cell nuclear extracts which showed that TCF11 p47/49 also bound specifically to an NF-E2-site with a CRE-like halfsite (11). The two methods used are likely to differ in sensitivity but the observation could also be explained by TCF11 forming heterodimers with bZIP-factors other than Maf proteins in these cell extracts, or alternatively, the endogenous TCF11:Maf heterodimer may display sequence-specificity different from that of the corresponding heterodimer formed in vitro.
The consensus heme-responsive element (30) contains a T at position +11 (Table 2C) which is only partially conserved and which can be substituted to an A without loss of heme responsivity in mouse L929 fibroblasts. The same position has 45% T and 35% A in the heterodimer selected TCF11 halfsites (Table 2C). This suggests that TCF11 prefers an A/T pair located at this position which was shown to be protected in an [alpha]-globin NF-E2-site in K562 cells (38). The CNC motif, being located immediately N-terminal to the basic DNA-binding domain, could therefore have a role in contacting DNA. Targeted point-mutations in this region of TCF11 may reveal its role in stabilising protein binding to the TCF11:Maf-site.
Our binding site-selected consensus site is also interesting in that 15 of the 20 positions (75%) deviating from the TCF11:MafG consensus nucleotides selected within the degenerate library are located at positions 1, 2 and 3 (Table 2). This `hot-spot' may therefore be a central sub-element inside the 11 bp TCF11:Maf-site which is used to discriminate it from a strong AP1 (Jun/Fos) element. This possibility is now being tested.
It was immediately apparent that the binding site selected for the TCF11:MafG heterodimer is identical to the NF-E2 site that mediates erythroid specific gene expression of, for example, the [beta]-globin gene cluster and the PBGD gene (5,6,39,40). We carried out a search of the sequence database to gain an idea of the overall distribution of the selected sites and found that a number of genes contain potential binding-sites for the heterodimer in their regulatory regions (Table 3). Many of these putative target genes can be classified into genes involved in haemoglobin and iron metabolism and genes important in cellular detoxification. Transcriptional responses to antioxidants and several xenobiotics act through AREs (5'-GCnnnGTCA-3') (31) and AREs from heme oxygenase 1, NAD(P)H:quinone oxidoreductase (NQO1), glutathione S-transferase and phenol sulfotransferase 1 and 2 show similarly positioned nucleotides that together define a consensus site identical to that of the TCF11:small Maf binding-site (Table 3). Indeed, TCF11 has been shown to positively regulate chloramphenicol acetyl-transferase (CAT) gene expression when linked to an ARE derived from the NQO1 gene (41).
Table 3. The occurrence of the selected TCF11:MafG binding-site in association with various genes. The TCF11:MafG consensus binding sequence (5'-TGCTgaGTCAT-3') was used in a computer search of the EMBL database and of the eukaryotic promoter database (EPD)
Based on the similarity between TCF11 and p45 NF-E2, it was previously suggested that TCF11 may act through this site and such binding has been demonstrated in vitro (9,11). The results of the binding-site selection assay further underlined this possibility. We analysed the activity of TCF11 in a transient transfection assay where the firefly luciferase gene, under the control of the PBGD erythroid-specific promoter (39,40) (chosen for the simple context in which a single NF-E2 site is presented), acts as a reporter driven by TCF11 produced from a full length TCF11 cDNA. Correct expression and nuclear localisation of the TCF11 protein were confirmed by immunofluoresence staining using a polyclonal TCF11 antibody (11, data not shown). The activity of TCF11 was assayed in a number of different cell lines and was found to transactivate expression in COS 1 and CV 1 monkey kidney cells, in human HeLa cells and endothelial cells (EA.hy 926 and ECV304), murine NIH3T3 fibroblasts and rat PC12 cells (Fig. 2). The level of transactivation was low and somewhat variable from experiment to experiment but consistently positive against a high background level of activity in the absence of transfected TCF11 (not shown). Murine F9 embryonal carcinoma cells showed the highest variability and in some experiments no transactivation was observed. To demonstrate that the NF-E2 site within the PBGD promoter mediated the transactivation, a number of deletions of the promoter region were analysed (Fig. 3A), this showed that transactivation was lost when the NF-E2 site was deleted (Fig. 3B). The background levels also dropped 3-10-fold (not shown). To demonstrate the role of the NF-E2 site more directly, its sequence was mutated within the context of the shortest active promoter (PBGD3.2Luc) (Fig. 3A). The two mutations that were assayed both reduced or abolished transactivation (Fig. 3C). This implies that TCF11 can bind to and transactivate expression through the NF-E2 site.
DNA binding assays showed that TCF11 preferentially forms heterodimers with small Maf proteins (including MafG) in vitro (11). The binding site selection studies reported here indicate that the preferred DNA binding sequence for the TCF11:MafG heterodimer is a perfect NF-E2 site. Furthermore, these experiments showed that TCF11 alone did not form simple homodimeric binding units and while it did bind DNA with a sequence specificity, this specificity was limited compared to the very clear preference shown by the heterodimer. We therefore wished to compare the activities of TCF11 alone and TCF11 co-expressed with MafG. This comparison was performed in COS 1 cells. As described above, TCF11 transactivated expression and the level of transactivation was found to be dependent on the amount of TCF11 transfected (Fig. 4B). In such a transfection assay, which involves transcription through an NF-E2 site that harbours a core AP1 site, background expression of the reporter gene is high. As has been found previously (16,21) the expression of MafG alone efficiently repressed this background level (Fig. 4). This is not surprising since small Maf proteins do not contain a known transactivation domain and so binding of the Maf homodimer may block access of endogenous factors (possibly AP1) responsible for the background activity. There is also evidence that small Maf proteins may block endogenous activation indirectly (29). Expression of a mutant form of MafG, which harbours a single amino acid change within the leucine zipper and so cannot dimerise (MafG[Delta] L2PM4P; 21) showed no such repression. Surprisingly, when TCF11 and MafG were co-expressed in the same cell, MafG blocked the weak transactivation observed with TCF11. Different relative amounts of the expression vectors for TCF11 and MafG were transfected in an effort to titrate the interaction. It was found that even at low relative amounts of MafG to TCF11 (2:7 µg of vector DNA) a significant drop in transactivation was already obvious (Fig. 4A). Even lower relative amounts of MafG (down to 10 ng of vector DNA) showed no evidence of co-operative transactivation with TCF11 (Fig. 4B). It is clear that the presence of MafG interferes with TCF11 mediated transactivation but it is not known whether this is due to the TCF11:MafG heterodimer lacking transactivation ability or whether MafG preferentially forms homodimers in this cellular context which compete for binding site access. The first possibility is suggested by the fact that TCF11 clearly and preferentially forms heterodimers with MafG in vitro (11; Fig. 1B).
In this series of experiments it was also observed that MafG can repress expression from the SV-40 promoter. This is important to note for the design of future transfection experiments involving small Maf proteins since vectors using this promoter cannot, therefore, be used as independent internal controls. In early experiments the plasmid pRSV-CAT, where CAT gene expression is driven from the constitutive SV-40 promoter, was used as an internal control to correct for differences in transfection efficiency. It was found that whenever the MafG expression vector was used in these experiments the level of CAT activity was greatly reduced (in the region of 10-fold, results not shown). Therefore, if the luciferase activity were corrected for CAT activity, the results were highly variable, inconsistent and uninterpretable. From these experiments the luciferase activity not corrected for CAT showed consistently the same trend observed in later experiments using an alternative internal control (pEF[beta]-gal). The internal control used in subsequent experiments expresses [beta]-galactosidase under the control of the elongation factor I[alpha] promoter and did not appear to be influenced by MafG (29,42).
The PCR-assisted approach of cloning transcription factor binding-sites following in vitro selection, is a powerful tool in the identification of regulatory sequences (24,36,37). We have shown that the TCF11:MafG heterodimer shows a clear preference for a site identical to a number of known regulatory elements including the NF-E2 site, the ARE and the HRE. The sequences of the selected oligonucleotides show a number of interesting features that are discussed in the Results section. It is interesting that TCF11 alone does not form simple homodimers in vitro but apparently binds as a multimeric complex to AP1 half sites, as indicated by its retarded electrophoretic mobility. The reduced specificity seen in the selection of TCF11 sites, together with the preference shown for the formation of heterodimers when TCF11 and small Maf proteins are co-expressed in vitro, indicate that a functional form of TCF11 is as a heterodimer with small Maf and/or, perhaps, with other unidentified bZIP partners. However, our transfection experiments show that TCF11 can transactivate expression when transfected alone and the fact that a dose dependent increase is observed would suggest that this effect is not dependent on TCF11 heterodimerisation with limiting amounts of endogenous factors.
The transactivation observed with TCF11 alone is inhibited by co-expression of MafG. Since these factors are both widely expressed and are likely to be co-expressed in a variety of cell types in vivo, and since they preferentially form heterodimers in vitro, it seems likely that the factors may commonly exist as an inactive or repressive heterodimeric form. Positive regulation of the target sites to which they bind may depend on heterodimerisation of these factors with alternative partners, perhaps with more tissue restricted expression. This study has provided us with a system in which to test alternative TCF11 partners. The other small Maf proteins and CNC-bZIP family members are candidates but the identification of new, tissue restricted partners for TCF11 is also pertinent.
The absence of co-activation by TCF11 and MafG shows that TCF11 acts differently to other CNC-bZIP family members in similar transfection assays. It has been observed that while all three small Maf proteins (-K, -F and -G) repress expression through the NF-E2 site in NIH3T3 cells or quail fibroblasts, co-expression with NF-E2 p45 interferes with Maf repression and leads to co-transactivation of the reporter (16,21,26). The NF-E2 related chicken gene, Ech (most similar to Nrf2), transactivates very efficiently in quail fibroblast cells in the presence of MafK, also overcoming repression by MafK alone (13). It is important to note that homology between the CNC-bZIP family members is largely restricted to the CNC-bZIP region involved in DNA-binding and heterodimerisation. Therefore functions mediated by other domains, hypothetically co-activator interactions and contact with the transcriptional complex, may differ between, for example, p45 NF-E2:small Maf and TCF11:small Maf heterodimers. This suggests that different CNC-bZIP factors, in partnership with different (or perhaps the same) small Maf proteins can act differentially through the same or similar regulatory elements indicating a complex network of competitive interactions when these factors are co-expressed.
Two recent articles report the cloning of the human homologue of MafG and, in contrast with our results, claim that a very small relative amount of MafG co-expressed with a TCF11 isoform (Nrf1) leads to co-operative transactivation (43,44). However, the data presented show that the slight increase in reporter activity in the presence of the two expression vectors falls well within (44), or just outside (43), the range of errors for the experiment. The slight increase reported in the latter case is detected in the range of 2 µg Nrf1 expression vector: 1-10 ng MafG expression vector, the effect being lost at 100 ng MafG. It is difficult to understand why such a low relative amount of MafG would have a positive effect that is lost so rapidly. We have attempted to repeat these observations in our assay system and see no such effect (Fig. 4), especially given the large variability inherent in this kind of experiment. It is possible that this discrepancy can be explained by differences in our assay systems although Toki et al., used the same cell line (COS 1) for their transfections. They have, however, expressed the human MafG cDNA whereas we used the chicken homologue (94% identity at the protein level).
TCF11 is a transcription factor that has been implicated in the regulation of erythroid-specific expression because of its ability to bind the NF-E2 site, but it is not erythroid specific, showing widespread expression. No target genes for TCF11 have as yet been identified. This study has shown that potential targets, based on the presence of optimal binding-sites, fall into a number of groups of genes (Table 3) which are co-regulated in response to a specific signal; antioxidant response, heme biosynthesis and erythroid differentiation, implicating TCF11 in specific biological processes. The role of TCF11 in these processes can now be tested. It is not known how many of these potential TCF11 binding sites represent real targets of TCF11 but it is interesting to speculate that the ubiquitously expressed proteins TCF11, Maf and AP1 all participate in gene regulation through these sites and thereby are connected in a network regulating a broad range of genes. The number of sites that represent real targets may be limited by competition from alternative factors or physical unavailability of sites due to protein binding to flanking sequences. In the search for TCF11 target genes, it has become important to also consider genes involved in early embryonic development, specifically in the gastrulation process, since the work of Farmer et al. (45) has shown that mice lacking TCF11 (LCR-F1) are blocked during the early steps of gastrulation. The identification of functional target genes for TCF11 is now of primary importance for the further understanding of TCF11 activity and function in vivo.
This work was supported by grants to A.-B. K. from the Research Council of Norway and the Norwegian Cancer Society and to P. M. from the Norwegian Cancer Society. Ø. J. is a Research Fellow of the Research Council of Norway and the University of Oslo. The MBP-MafG construct, MafG cDNA clones and pEF[beta]gal were kindly provided by Dr Makoto Nishizawa, University of California at San Diego.
Nucleic Acids Research
Pages
Introduction
Materials And Methods
Plasmid constructs and fusion proteins
Binding-site selection
Transient transfection assay
Results
Binding sites selected by TCF11
Binding sites selected by TCF11:MafG
Potential TCF11/Maf target genes
TCF11 activity through the NF-E2 site
The activity of TCF11 in the presence of MafG
MafG interferes with expression from the SV-40 promoter
Discussion
Acknowledgements
References
REFERENCES
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 6 Jan 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. T. Berg, A. Gupta, M. A. Richardson, L. A. O'Brien, D. Calnek, and B. W. Grinnell Negative Regulation of Inducible Nitric-oxide Synthase Expression Mediated through Transforming Growth Factor- -dependent Modulation of Transcription Factor TCF11 J. Biol. Chem., December 21, 2007; 282(51): 36837 - 36844. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Yamamoto, M. Kyo, T. Kamiya, T. Tanaka, J. D. Engel, H. Motohashi, and M. Yamamoto Predictive base substitution rules that determine the binding and transcriptional specificity of Maf recognition elements Genes Cells, June 1, 2006; 11(6): 575 - 591. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Katsuoka, H. Motohashi, T. Ishii, H. Aburatani, J. D. Engel, and M. Yamamoto Genetic Evidence that Small Maf Proteins Are Essential for the Activation of Antioxidant Response Element-Dependent Genes Mol. Cell. Biol., September 15, 2005; 25(18): 8044 - 8051. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Chen, M. Kwong, R. Lu, D. Ginzinger, C. Lee, L. Leung, and J. Y. Chan Nrf1 Is Critical for Redox Balance and Survival of Liver Cells during Development Mol. Cell. Biol., July 1, 2003; 23(13): 4673 - 4686. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Braun, C. Hanselmann, M. G. Gassmann, U. auf dem Keller, C. Born-Berclaz, K. Chan, Y. W. Kan, and S. Werner Nrf2 Transcription Factor, a Novel Target of Keratinocyte Growth Factor Action Which Regulates Gene Expression and Inflammation in the Healing Skin Wound Mol. Cell. Biol., August 1, 2002; 22(15): 5492 - 5505. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Schachat and M. M. Briggs Phylogenetic implications of the superfast myosin in extraocular muscles J. Exp. Biol., August 1, 2002; 205(15): 2189 - 2201. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C.-F. Huang, Y.-C. Wang, D.-A. Tsao, S.-F. Tung, Y.-S. Lin, and C.-W. Wu Antagonism between Members of the CNC-bZIP Family and the Immediate-Early Protein IE2 of Human Cytomegalovirus J. Biol. Chem., April 14, 2000; 275(16): 12313 - 12320. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A Veraksa, N McGinnis, X Li, J Mohler, and W McGinnis Cap 'n' collar B cooperates with a small Maf subunit to specify pharyngeal development and suppress deformed homeotic function in the Drosophila head Development, January 9, 2000; 127(18): 4023 - 4037. [Abstract] [PDF]
|
|
|
|
|
|
M. Kwong, Y. W. Kan, and J. Y. Chan The CNC Basic Leucine Zipper Factor, Nrf1, Is Essential for Cell Survival in Response to Oxidative Stress-inducing Agents. ROLE FOR Nrf1 IN gamma -gcsL AND gss EXPRESSION IN MOUSE FIBROBLASTS J. Biol. Chem., December 24, 1999; 274(52): 37491 - 37498. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Alam, D. Stewart, C. Touchard, S. Boinapally, A. M. K. Choi, and J. L. Cook Nrf2, a Cap'n'Collar Transcription Factor, Regulates Induction of the Heme Oxygenase-1 Gene J. Biol. Chem., September 10, 1999; 274(37): 26071 - 26078. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Kobayashi, E. Ito, T. Toki, K. Kogame, S. Takahashi, K. Igarashi, N. Hayashi, and M. Yamamoto Molecular Cloning and Functional Characterization of a New Cap'n' Collar Family Transcription Factor Nrf3 J. Biol. Chem., March 5, 1999; 274(10): 6443 - 6452. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Kanezaki, T. Toki, M. Yokoyama, K. Yomogida, K. Sugiyama, M. Yamamoto, K. Igarashi, and E. Ito Transcription Factor BACH1 Is Recruited to the Nucleus by Its Novel Alternative Spliced Isoform J. Biol. Chem., March 2, 2001; 276(10): 7278 - 7284. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Husberg, P. Murphy, E. Martin, and A.-B. Kolsto Two Domains of the Human bZIP Transcription Factor TCF11 Are Necessary for Transactivation J. Biol. Chem., May 18, 2001; 276(21): 17641 - 17652. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Godbout and E. A. Monckton Differential Regulation of the Aldehyde Dehydrogenase 1 Gene in Embryonic Chick Retina and Liver J. Biol. Chem., August 24, 2001; 276(35): 32896 - 32904. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||