p45 NF-E2 regulates expression of thromboxane synthase in megakaryocytes
1997; Springer Nature; Volume: 16; Issue: 18 Linguagem: Inglês
10.1093/emboj/16.18.5654
ISSN1460-2075
Autores Tópico(s)Platelet Disorders and Treatments
ResumoArticle15 September 1997free access p45 NF-E2 regulates expression of thromboxane synthase in megakaryocytes Sophie Deveaux Sophie Deveaux INSERM U.91, Hôpital Henri Mondor, 51 Avenue du Maréchal de Lattre, 94010 Créteil, France Search for more papers by this author Sylvia Cohen-Kaminsky Sylvia Cohen-KaminskyS.Cohen-Kaminsky and R.A.Shivdasani contributed equally to this work Search for more papers by this author Ramesh A. Shivdasani Ramesh A. Shivdasani Dana-Farber Cancer Institute, Boston, MA, 02115 USAS.Cohen-Kaminsky and R.A.Shivdasani contributed equally to this work Search for more papers by this author Nancy C. Andrews Nancy C. Andrews The Children's Hospital, Harvard Medical School and the Howard Hughes Medical Institute, Boston, MA, 02115 USA Search for more papers by this author Anne Filipe Anne Filipe INSERM U.91, Hôpital Henri Mondor, 51 Avenue du Maréchal de Lattre, 94010 Créteil, France Search for more papers by this author Isabelle Kuzniak Isabelle Kuzniak INSERM U.91, Hôpital Henri Mondor, 51 Avenue du Maréchal de Lattre, 94010 Créteil, France Search for more papers by this author Stuart H. Orkin Stuart H. Orkin The Children's Hospital, Harvard Medical School and the Howard Hughes Medical Institute, Boston, MA, 02115 USA Search for more papers by this author Paul-Henri Roméo Paul-Henri Roméo INSERM U.91, Hôpital Henri Mondor, 51 Avenue du Maréchal de Lattre, 94010 Créteil, France Search for more papers by this author Vincent Mignotte Corresponding Author Vincent Mignotte INSERM U.91, Hôpital Henri Mondor, 51 Avenue du Maréchal de Lattre, 94010 Créteil, France Search for more papers by this author Sophie Deveaux Sophie Deveaux INSERM U.91, Hôpital Henri Mondor, 51 Avenue du Maréchal de Lattre, 94010 Créteil, France Search for more papers by this author Sylvia Cohen-Kaminsky Sylvia Cohen-KaminskyS.Cohen-Kaminsky and R.A.Shivdasani contributed equally to this work Search for more papers by this author Ramesh A. Shivdasani Ramesh A. Shivdasani Dana-Farber Cancer Institute, Boston, MA, 02115 USAS.Cohen-Kaminsky and R.A.Shivdasani contributed equally to this work Search for more papers by this author Nancy C. Andrews Nancy C. Andrews The Children's Hospital, Harvard Medical School and the Howard Hughes Medical Institute, Boston, MA, 02115 USA Search for more papers by this author Anne Filipe Anne Filipe INSERM U.91, Hôpital Henri Mondor, 51 Avenue du Maréchal de Lattre, 94010 Créteil, France Search for more papers by this author Isabelle Kuzniak Isabelle Kuzniak INSERM U.91, Hôpital Henri Mondor, 51 Avenue du Maréchal de Lattre, 94010 Créteil, France Search for more papers by this author Stuart H. Orkin Stuart H. Orkin The Children's Hospital, Harvard Medical School and the Howard Hughes Medical Institute, Boston, MA, 02115 USA Search for more papers by this author Paul-Henri Roméo Paul-Henri Roméo INSERM U.91, Hôpital Henri Mondor, 51 Avenue du Maréchal de Lattre, 94010 Créteil, France Search for more papers by this author Vincent Mignotte Corresponding Author Vincent Mignotte INSERM U.91, Hôpital Henri Mondor, 51 Avenue du Maréchal de Lattre, 94010 Créteil, France Search for more papers by this author Author Information Sophie Deveaux1, Sylvia Cohen-Kaminsky, Ramesh A. Shivdasani2, Nancy C. Andrews3, Anne Filipe1, Isabelle Kuzniak1, Stuart H. Orkin3, Paul-Henri Roméo1 and Vincent Mignotte 1 1INSERM U.91, Hôpital Henri Mondor, 51 Avenue du Maréchal de Lattre, 94010 Créteil, France 2Dana-Farber Cancer Institute, Boston, MA, 02115 USA 3The Children's Hospital, Harvard Medical School and the Howard Hughes Medical Institute, Boston, MA, 02115 USA The EMBO Journal (1997)16:5654-5661https://doi.org/10.1093/emboj/16.18.5654 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Transcription factor p45 NF-E2 is highly expressed in the erythroid and megakaryocytic lineages. Although p45 recognizes regulatory regions of several erythroid genes, mice deficient for this protein display only mild dyserythropoiesis but have abnormal megakaryocytes and lack circulating platelets. A number of megakaryocytic marker genes have been extensively studied, but none of them is regulated by NF-E2. To find target genes for p45 NF-E2 in megakaryopoiesis, we used an in vivo immunoselection assay: genomic fragments bound to p45 NF-E2 in the chromatin of a megakaryocytic cell line were immunoprecipitated with an anti-p45 antiserum and cloned. One of these fragments belongs to the second intron of the thromboxane synthase gene (TXS). We demonstrate that the TXS gene, which is mainly expressed in megakaryocytes, is indeed directly regulated by p45 NF-E2. First, its promoter contains a functional NF-E2 binding site; second, the intronic NF-E2 binding site is located within a chromatin-dependent enhancer element; third, p45-null murine megakaryocytes do not express detectable TXS mRNA, although TXS expression can be detected in other cells. These data, and the structure of the TXS promoter and enhancer, suggest that TXS belongs to a distinct subgroup of genes involved in platelet formation and function. Introduction Transcription factor NF-E2 binds to extended Ap1 (Jun/Fos) motifs in the erythroid-specific promoter of the human porphobilinogen deaminase gene (PBGD) (Mignotte et al., 1989b), the β-globin locus control region (Ney et al., 1990; Talbot et al., 1990) and the ferrochelatase promoter (Tugores et al., 1994). NF-E2 originally appeared to be required for full expression of the PBGD (Mignotte et al., 1989a) and globin (Ney et al., 1990; Talbot and Grosveld, 1991; Caterina et al., 1994a; Lu et al., 1994; Kotkow and Orkin, 1995) genes. Purification of NF-E2 revealed that it is a heterodimeric member of the basic region–leucine zipper (bZip) class of transcription factors. Its tissue-specific subunit, p45, belongs to the cap‘n'collar (CNC) family (Andrews et al., 1993a) and it associates with a ubiquitous partner, either p18/MafK or MafG, which is one of the small Maf proteins (Andrews et al., 1993b; Igarashi et al., 1994, 1995; Blank et al., 1997). The consensus binding site for the p45/p18 NF-E2 heterodimer is an asymmetrical TGAC/GTCAGCA sequence, where p45 recognizes the TGA moiety and p18 the longer TCAGCA motif (Igarashi et al., 1994). In addition to erythroid cells, NF-E2 activity is present in megakaryocytes and mast cells (Roméo et al., 1990; Andrews et al., 1993a). p45-deficient mice, generated by gene targeting in embryonic stem (ES) cells, displayed only mild dyserythropoiesis. Instead, they exhibited high mortality from hemorrhage secondary to the absence of circulating platelets. Megakaryocytes were present in the fetal liver and bone marrow of mutant mice, but ultrastructural studies revealed an aberrant distribution of demarcation membranes, as well as a reduced number of granules, in their cytoplasm; these cells were able to proliferate in response to thrombopoietin, but were unable to form platelets (Shivdasani et al., 1995). Therefore, it is likely that p45 NF-E2 is essential for expression of a subset of specific genes in the late stages of megakaryopoiesis, through binding to its cognate sequence. Alternatively, NF-E2 could act on specific genes through protein–protein interactions not involving direct binding to DNA. The latter situation has been described for several transcription factors, including the glucocorticoid receptor (Konig et al., 1992). A number of marker genes of the megakaryocytic lineage have been studied recently. These include the genes for glycoprotein IIb, thrombopoietin receptor (MPL), platelet factor 4, glycoproteins Ibα, V and IX and P-selectin (Ravid et al., 1991; Prandini et al., 1992; Hickey and Roth, 1993; Lanza et al., 1993; Lemarchandel et al., 1993; Pan and McEver, 1993; Block et al., 1994; Hashimoto and Ware, 1995; Deveaux et al., 1996). The regulatory regions of these genes are characterized by the presence of closely spaced binding sites for GATA-1 and proteins of the Ets family, but they do not contain NF-E2 binding sites and their expression in the megakaryocytes of p45-deficient mice is normal (Shivdasani et al., 1995; Shivdasani, 1996). The finding that p45-deficient megakaryocytes could not undergo terminal differentiation favors the hypothesis that p45 regulates a distinct set of genes which are crucial for platelet formation and function and may be expressed late in differentiation. To test this hypothesis, we decided to isolate sequences that are bound by p45 NF-E2 in megakaryocytic cells, using an immunological selection procedure in active nuclei. Genomic DNA fragments co-immunoprecipitated with an anti-p45 antiserum were cloned and one of these fragments was assigned to the thromboxane synthase gene (TBXAS1, here abbreviated to TXS). We present data showing that TXS expression is indeed controlled by NF-E2 in megakaryocytes: first, NF-E2 binding sites were found in the promoter and in an intronic enhancer of TXS; second, p45-deficient mice show a deep reduction in megakaryocytic TXS expression; third, NF-E2 acts by direct binding to the TXS promoter. These data demonstrate the existence of at least one direct target of p45 NF-E2 in megakaryocytes and suggest that this factor co-regulates a group of genes during platelet formation. Results Isolation of a fragment of the TXS gene by chromatin immunoprecipitation In an attempt to identify in vivo target regulatory regions for p45, we developed an immunopurification protocol that enriches for short DNA fragments that are bound to p45 in chromatin (see Materials and methods). This approach has led to the identification of target genes for the ultrabithorax protein in Drosophila (Graba et al., 1992) and Hox-C8 (Tomotsune et al., 1993), the thyroid hormone receptor (Bigler and Eisenman, 1994) and c-Myc (Grandori et al., 1996) in mammals. It has been shown that NF-E2 is able to disrupt nucleosomes in vitro (Armstrong and Emerson, 1996). In addition, chromatin decondensation often precedes the onset of gene expression (Dillon and Grosveld, 1993). Therefore, we assumed that target genes for p45 NF-E2 in terminal megakaryocyte maturation would already be expressed, or at least be in an open chromatin configuration, in a megakaryoblastic cell line expressing p45. We chose the HEL cell line, because it expresses high levels of mRNA for the thrombopoietin receptor (Mignotte et al., 1994), glycoprotein IIb (Poncz et al., 1987) and p45 NF-E2 (Chan et al., 1993b). Immunoselected DNA fragments from HEL nuclei were cloned into a plasmid vector. A thousand recombinant clones were ordered, among which 55 were sequenced. Thirteen of these contained motifs with 70% homology to the NF-E2 consensus sequence TGACTCAGCA (data not shown). Comparison with the GenBank database showed that one fragment (clone so13) corresponded to a portion of the second intron of human TXS (Figure 1A). This region contains two motifs, TGACCTCAGGT (nfe2-1) and TGATCTCAGCT (nfe2-2), that show homology with the NF-E2 binding site. However, both of them are more homologous to the cAMP-response element than to the Ap1 motif (see Discussion). Figure 1.(A) Sequence of a 870 bp fragment of the second intron of the human TXS gene. This fragment was cloned by PCR (see text). It contains the sequence isolated by immunoprecipitation (clone so13, underlined). The two putative NF-E2 binding sites are indicated (nfe2-1 and nfe2-2), as well as putative GATA and Ets binding sites in the whole fragment. This sequence corresponds to GenBank entry D34615. (B) Sequence of the TXS promoter fragment cloned from K562 DNA and used for transfection assays. This sequence corresponds to GenBank entry D34613 (Miyata et al., 1994) except for a T→G nucleotide change indicated in lower case. The GATA and Ets motifs and the NF-E2 binding site (nfe2-3) are indicated. Numbering is from Miyata et al. (1994). Download figure Download PowerPoint Both motifs were tested for their ability to bind NF-E2 in HEL nuclear extracts. Figure 2 shows that an oligonucleotide containing the nfe2-2 motif is able to compete binding of NF-E2 to the canonical PBGD NF-E2 binding site (lane 4). When the nfe2-2 oligonucleotide was used as a probe it was able to form several complexes, one of which was disrupted by anti-p45 antibodies (lanes 9 and 10). An anti-MafG antiserum, which is also able to recognize MafK and probably other small Maf proteins (Blank et al., 1997), was also able to disrupt this complex, in contrast to an anti-GATA-1 antiserum (data not shown). The nfe2-1 oligonucleotide is bound by a protein unrelated to NF-E2 (not shown) and is unable to compete for NF-E2 binding (Figure 2, lane 3). We conclude that only the nfe2-2 motif can bind p45 NF-E2 in vitro. In addition, a TGATCTCAGCA motif found in another immunoprecipitated sequence (clone so7; data not shown) is also able to compete for NF-E2 binding (lane 5). Figure 2.Gel retardation assay in HEL nuclear extract with the NF-E2 binding site of the PBGD promoter (lanes 1–8) and the nfe2-2 oligonucleotide (lanes 9 and 10). Cold competitor oligonucleotides were added in 100-fold molar excess: lane 1, no competitor; lane 2, PBGD NF-E2 site; lane 3, nfe2-1 oligonucleotide; lane 4, nfe2-2; lane 5, clone so7 (a sequence isolated in the same immunoprecipitation experiment; data not shown). The NF-E2 complex could be disrupted (and partially supershifted) with anti-p45 antibodies (lane 6) and anti–MafG/K antibodies (lane 7), but not with anti-GATA-1 antibodies (lane 8). One of the complexes formed on nfe2-2 (lane 9, arrowhead) could be disrupted with anti-p45 antibodies (lane 10). Download figure Download PowerPoint Thromboxane synthase is highly expressed in mature megakaryocytes Thromboxane synthase (TXS) controls biosynthesis of thromboxane A2, an inducer of platelet aggregation and vasoconstriction. TXS mRNA is most abundant in platelets, but is also found in blood monocytes, spleen, lung and liver (Miyata et al., 1994). In order to measure TXS expression in various human hematopoietic cells, we performed reverse transcription–polymerase chain reaction (RT–PCR) assays on HEL cells, CD34+CD38− and CD34+CD38+ progenitors, erythroblasts and megakaryocytes. Figure 3 shows that TXS expression is significantly higher in megakaryocytes than in HEL cells or in immature progenitors. This expression profile differs from that of another megakaryocytic marker, the thrombopoietin receptor (MPL), which is also highly expressed in HEL cells and progenitors. As expected, TXS expression was only weakly detected in erythroblasts. Figure 3.RT–PCR on the thromboxane synthase (TXS), thrombopoietin receptor (MPL) and ribosomal protein S14 (S 14) mRNAs. The latter was used as an internal standard. The PCR reaction was run for 26 (TXS), 30 (MPL) or 25 cycles (S 14). The cells used for total RNA preparation are indicated. Download figure Download PowerPoint The TXS promoter contains a high-affinity NF-E2 binding site The expression profile of the TXS gene prompted us to examine the structure of its promoter (Figure 1B). Eighty base pairs upstream of the major transcription initiation site (Miyata et al., 1994) lies a TGAATCAGCA motif (nfe2-3) that shows extensive (9/10) homology to the NF-E2 motif. Gel retardation experiments in the presence of specific competitors and antibodies demonstrated that it has high affinity for NF-E2 (Figure 4, lanes 1–8), although this sequence differs from the NF-E2 consensus binding site at the fourth nucleotide (C→A). In order to test the effect of this change in the context of a well-described NF-E2 binding site, we introduced a C→A mutation at the fourth position in the PBGD promoter NF-E2 binding site (Mignotte et al., 1989b). This resulted in a slightly lower affinity for NF-E2, but a strong inhibition of Ap1 binding (Figure 4, lanes 9 and 10). We conclude that the sequence found in the TXS promoter exhibits a marked preference for binding of NF-E2 over Ap1 complexes. Figure 4.The TGAATCAGCA sequence binds NF-E2 with high affinity. Lanes 1–5, competition of protein binding to the nfe2-3 oligonucleotide in HEL nuclear extract. Lane 1, no competitor; lane 2, nfe2-3; lane 3, PBGD NF-E2 site; lane 4, mutant PBGD NF-E2 site; lane 5, nfe2-2. Lanes 6–8, the NF-E2 complex could be disrupted or supershifted with anti-p45 antibodies (lane 6) and anti-MafG/K antibodies (lane 7) but not with anti-GATA-1 antibodies (lane 8). Lanes 9 and 10, gel retardation profile of the PBGD NF-E2 site (lane 9) and the mutant PBGD NF-E2 site (lane 10) in mouse fetal liver nuclear extract. The bands corresponding to Ap1 and NF-E2 complexes are indicated. Download figure Download PowerPoint TXS mRNA is undetectable in megakaryocytes from p45-deficient mice One prediction from our results is that levels of TXS mRNA would be significantly reduced in megakaryocytes derived from mice lacking NF-E2 function. To test this prediction, we used RT–PCR to compare TXS mRNA levels in fetal livers of wild-type, heterozygous and homozygous null p45 NF-E2 knockout mice (Shivdasani et al., 1995). With hypoxanthine phosphoribosyl transferase (HPRT) transcripts as a control for input RNA, the levels of TXS transcripts are not appreciably decreased in heterozygous or homozygous fetal livers (Figure 5A). However, this tissue also contains other cell lineages known to express TXS mRNA, including monocytes, and only a very small proportion of megakaryocytes; therefore, the contribution of these other lineages to total TXS mRNA during fetal hematopoiesis might conceal an absence of TXS expression in megakaryocytes. Figure 5.RT–PCR analysis of fetal livers (A), Tpo-treated adult bone marrow (B) and cultured megakaryocyte colonies (C), indicating abrogation of TXS mRNA levels in megakaryocytes in the absence of NF-E2 function in vivo. Input RNA levels were normalized using HPRT transcripts as the control. PCR reactions were performed for the indicated numbers of cycles. The bands corresponding to the TXS and HPRT fragments are indicated. The intermediate band is a minor TXS PCR product. Download figure Download PowerPoint Analysis of hematopoietic tissues from adult animals is complicated by the abundance of megakaryocytes found in mice lacking p45 NF-E2 relative to normal mice, making direct comparison of megakaryocyte RNAs difficult. In contrast, both wild-type and p45 NF-E2-null adult mice treated with recombinant MPL ligand (Tpo) have excessive, and roughly equivalent, numbers of megakaryocytes (Shivdasani et al., 1995). RT–PCR analysis of bone marrow from Tpo-treated mice again reveals significant TXS expression in p45-null bone marrow (Figure 5B). However, this result does not address whether this expression stems from megakaryocytes or other cell types. We therefore compared TXS mRNA levels between megakaryocyte colonies cultured in vitro from the livers of p45 NF-E2-null and control fetuses. As shown in Figure 5C, TXS mRNA is undetectable in megakaryocytes cultured from mice lacking p45 NF-E2. Taken together, these data indicate that NF-E2 is a major transcriptional regulator of TXS expression within megakaryocytes, while other proteins likely provide such regulation in other cells. The TXS promoter is directly regulated by p45 NF-E2 These results could not exclude the possibility that p45 NF-E2 has only an indirect influence on TXS expression. To test this hypothesis, we destroyed the NF-E2 binding site and compared the activity of the wild-type and mutant promoters [cloned into the chloramphenicol acetyltransferase (CAT) expression vector pBL-CAT3] in stably transfected HEL cells. Abolition of NF-E2 binding to the TXS promoter decreases its activity 6-fold under these conditions (Figure 6). Therefore, we conclude that p45 NF-E2 is a direct regulator of TXS expression. Figure 6.Regulatory elements of the TXS gene. The above constructs were stably transfected into HEL cells and their respective CAT activities are shown (mean results of three transfections, SD 8%). The 870 bp fragment from intron 2 (long box; see Figure 1A) was able to stimulate the TXS promoter, but mutation of its NF-E2 motifs (crosses) failed to reduce this effect. A mutant promoter that cannot bind NF-E2 displayed a 6-fold reduction in expression level compared with the wild-type promoter. A 190 bp fragment containing the nfe2-1 and nfe2-2 motifs (short box) was able to partially rescue the mutated promoter. When these two motifs were destroyed (crosses), no activation could be observed. Download figure Download PowerPoint A fragment of intron 2 has enhancer activity As the original fragment we isolated by immunoprecipitation belongs to the second intron of the TXS gene, we wanted to determine the role of this region in TXS gene expression. The nucleotide sequence of intron 2 reveals four potential GATA-1 binding sites 100–230 bp downstream of the nfe2-2 motif, as well as three Ets motifs (Figure 1A). Therefore, we speculated that the whole region may have enhancer activity. We synthesized by PCR a 870 bp fragment that contains the 3′ part of TXS intron 2 (see Materials and methods) and subcloned it upstream of the TXS promoter in the reporter vector. Transient transfections in HEL cells failed to reveal any effect of the intron fragment (data not shown). In contrast, when stable transfections were performed with these constructs a 9-fold stimulation of CAT activity by the intron 2 fragment was observed (Figure 6; mean result of three transfections). We conclude that the second intron of the TXS gene is an enhancer element whose activity in HEL cells is only apparent in the context of chromatin. The CD34 gene (May and Enver, 1995) and the MPL gene (S.Deveaux, A.Filipe and V.Mignotte, unpublished results) contain similar control regions. The intronic NF-E2 motifs are able to rescue a mutated TXS promoter To measure the contribution of the nfe2-1 and nfe2-2 motifs to the activity of the TXS enhancer, we disrupted these motifs in the context of an enhancer–promoter construct, but this failed to reveal any significant decrease in expression (Figure 6). In addition, a 190 bp subfragment of the intronic enhancer spanning the two NF-E2 motifs was unable to stimulate the TXS promoter. One explanation could be that most of the enhancer activity resides in the GATA binding sites or other sequences and that the high affinity NF-E2 site in the promoter may functionally compensate for mutation of the intronic motifs. Therefore, we designed a series of constructs to test the ability of the nfe2-1 and nfe2-2 motifs to rescue a mutated TXS promoter (Figure 6). In stable transfections these motifs were indeed able to in part rescue the effect of loss of the nfe2-3 motif. We conclude that these motifs are by themselves able to mediate transcriptional activation in HEL cells, although the promoter NF-E2 binding site appears to be most important under these experimental conditions. Discussion Cloning of p45 binding sites by chromatin immunoprecipitation A number of approaches have been used in the quest for transcription factor target genes. These include differential cDNA library construction (Cayrol and Flemington, 1995), differential hybridization (Feinstein et al., 1995; Hartl and Bister, 1995) and differential display (Okamoto and Beach, 1994). Identification of a target gene is then based on its differential mRNA expression. These three methods require the availability of two cell populations whose only difference (ideally) is expression of the regulatory protein. Other strategies exploit the DNA binding ability of the transcription factor. Genomic binding site cloning (Inoue et al., 1993) and whole genome PCR (Kinzler and Vogelstein, 1990) rely on repetitive selection of genomic fragments that are bound by the purified factor in vitro, via immunoprecipitation and PCR steps. The starting DNA population can be enriched in active genomic sequences, for example by isolation of CpG-rich sequences (Shago and Giguère, 1996). The main drawback of these methods is preferential isolation of high affinity binding sites that may not reflect the in vivo situation. We chose to set up in vivo immunoprecipitation of p45 NF-E2-bound DNA, using an erythro-megakaryoblastic cell line that expresses high levels of megakaryocytic markers, as well as p45. This method has been used to identify homeotic target genes (Gould et al., 1990; Graba et al., 1992; Tomotsune et al., 1993). It does not select against low affinity binding sites. However, it has been shown that <10% of the fragments that are isolated in this manner contain consensus motifs (Gould et al., 1990; White et al., 1992) or are efficient in vitro binding sites (Bigler and Eisenman, 1994, 1995; Phelps and Dressler, 1996). It is therefore necessary to screen high numbers of clones in sequence analysis and gel retardation assays. Another difficulty is to identify the genomic locus to which a given binding site belongs. CRE-type sequences may be targets for p45 NF-E2 in vivo Using HEL cells and an anti-p45 antiserum we have cloned 1000 candidate fragments and sequenced ∼5% of the clones (unpublished data). As expected, only a minority of these contained NF-E2-related motifs. A number of the candidate sequences were found to be ‘CRE-type’ motifs, i.e. their sequence is more related to the cAMP-response element (CRE) TGACGTCA than to the Ap1 motif (or ‘TRE-type’) TGACTCA. This is the case of the nfe2-2 motif and the motif contained in clone so7 (Figure 2). Although all the NF-E2 binding sites described to date are Ap1 motifs, this result suggests that CREs can be recognized by NF-E2 in vivo. This is in accordance with a previous study that showed that Ech–Maf heterodimers, which have similar site specificity, can bind a consensus CRE with high affinity (Kataoka et al., 1995). The genomic location and function of clone so7 are under study. Neither the TXS promoter nor the erythroid PBGD promoter were found in our cloned fragments (not shown). This suggests that the population of p45-binding clones is incomplete, as expected from the low enrichment. In the case of the PBGD promoter it could also mean that this erythroid promoter is bound by a different CNC family protein in vivo, such as Nrf1 (Chan et al., 1993a), Nrf2 (Moi et al., 1994), LCR-F1 (Caterina et al., 1994b) or a Bach protein (Oyake et al., 1996). Differential protein–protein interactions might result in binding of two different CNC proteins on the PBGD and TXS promoters in vivo. In chicken erythroid cells the CNC protein Ech is more abundant than p45 and it is able to heterodimerize with small Maf proteins and activate transcription through NF-E2 binding sites (Itoh et al., 1995). Thromboxane synthase is essential for platelet function Thromboxane A2 (TXA2) is an important messenger molecule for platelet aggregation (Hamberg et al., 1975). It is synthesized in platelets when they are stimulated by collagen, thrombin, adenosine diphosphate or epinephrine (Bhagwat et al., 1985). Its biosynthesis is catalyzed in succession by phospholipase A2, prostaglandin H synthase (cyclooxygenase) and thromboxane A2 synthase (TXS). Phospholipase A2 catalyzes deacylation of arachidonic acid from the sn-2 position of membrane phospholipids. Prostaglandin H synthase catalyzes conversion of free arachidonic acid to prostaglandin G2 and then to prostaglandin H2 (PGH2). In platelets TXS catalyzes specific conversion of PGH2 to TXA2 (Smith et al., 1991). TXA2 is released into the extracellular milieu and acts via a specific receptor on adjacent platelets to reinforce and amplify platelet aggregation and secretion. Platelet TXS deficiency has been shown to be associated with bleeding disorders in patients (Weiss and Lages, 1977; Mestel et al., 1980; Wu et al., 1981; Rao et al., 1985; Sinzinger et al., 1985). In these patients platelet counts are normal but platelet aggregation is impaired. From the results presented in this paper it is clear that TXS expression is abrogated in p45-null murine megakaryocytes. Although the present data do not directly address the role of TXS in thrombopoiesis, they certainly raise the possibility that activity of this enzyme plays a role not only in platelet activation, but also in platelet development. Architecture of the TXS regulatory regions A number of megakaryocyte-specific regulatory regions have been studied recently, including the glycoprotein IIb promoter (Lemarchandel et al., 1993; Block et al., 1994) and enhancer (Prandini et al., 1992), the thrombopoietin receptor promoter (Deveaux et al., 1996) and enhancer (Mignotte et al., 1996) and the GpIbα promoter (Hashimoto and Ware, 1995). All these contain binding sites for GATA-1 and proteins of the Ets family, usually 15–35 bp apart (Lemarchandel, 1993), which are crucial for expression. Two of these genes at least (MPL and GpIIb) are expressed normally in p45-null megakaryocytes (Shivdasani et al., 1995). TXS is predominantly (although not exclusively) expressed in megakaryocytes, but its promoter (Figure 1B) and the 870 bp intronic enhancer (Figure 1A) display a very different architecture. In addition to the presence of the NF-E2 site, the TXS promoter contains nine GATA motifs, including two clusters of three motifs, around −280 and −250. Although some of these motifs are imperfect GATA-1 binding sites (Merika and Orkin, 1993), the −280 and −250 regions constitute very strong GATA-1 binding elements, comparable with the high affinity motif located at −70 in the PBGD erythroid promoter (data not s
Referência(s)