Nucleic Acids Res
January 1, 2005;
An Oct-1 binding site mediates activation of the gata2 promoter by BMP signaling.
gene encodes a transcription factor implicated in regulating early patterning of ectoderm
, and later in numerous cell-specific gene expression programs. Activation of the gata2
gene during embryogenesis is dependent on the bone
morphogenetic protein (BMP) signaling pathway, but the mechanism for how signaling controls gene activity has not been defined. We developed an assay in Xenopus embryos to analyze regulatory sequences of the zebrafish gata2
promoter that are necessary to mediate the response to BMP signaling during embryogenesis. We show that activation is Smad dependent, since it is blocked by expression of the inhibitory Smad6
. Deletion analysis identified an octamer binding site that is necessary for BMP-mediated induction, and that interacts with the POU homeodomain protein Oct-1
. However, this element is not sufficient to transfer a BMP response to a heterologous promoter, requiring an additional more proximal
cooperating element. Based on recent studies with other BMP-dependent promoters (Drosophila vestigial and Xenopus Xvent-2
), our studies of the gata2
gene suggest that POU-domain proteins comprise a common component of the BMP signaling pathway, cooperating with Smad proteins and other transcriptional activators.
Nucleic Acids Res
[+] show captions
Figure 1. Structure of the zebrafish gata2 gene. (A) The diagram illustrates the exon/intron structure at the 5′ end of the zebrafish gata2 gene, determined by comparing the 5′-RACE product with the genomic sequence. Exons are shown as boxes and +1 indicates the transcription start site. (B) The position of the first intron relative to the genomic sequence is shown, with intronic sequences in lower case, and conserved intronic splice junction sequences in bold. (C) To generate a reporter gene, the sequences from the initiation ATG at +595 (relative to the transcription start site) to −6520 were placed upstream of the coding sequences for the firefly luciferase gene. This construct is analogous to a transgenic reporter shown to recapitulate the early gata2 expression pattern in zebrafish embryos (38).
Figure 2. The gata2 promoter directs reporter gene activity that is induced by BMP4 and dependent on Smad signaling. (A) The luciferase reporter was injected at the four-cell stage into blastomeres that contribute preferentially to ventral or dorsal regions of the embryo, as indicated. The reporter is more active in the ventral-posterior derivatives, but this is abrogated by co-injection of RNA encoding the inhibitory Smad6 (the error bar for this sample is present, but it is so small that it is not evident). The results indicate that the reporter responds as predicted to the endogenous Smad-dependent BMP signaling pathway. (B) The gata2 promoter is induced ∼5-fold by co-injection of RNA encoding BMP4. This activation is also blocked by co-expression of Smad6.
Figure 3. Deletion analysis maps a BRE to a 68 bp region upstream of the gata2 promoter. A progressive series of 5′ deletions was used to measure the relative activity of the promoter when injected with RNA encoding BMP4, compared to the same reporter when injected with non-coding control RNA, as described in Materials and Methods. For each construct, the site of truncation is indicated, the +1 indicates the transcriptional start site, and the box represents the firefly luciferase reporter. The fold difference (induced/uninduced) is plotted and shows a significant drop when sequences are deleted between −819 and −751. This 68 bp sequence is designated BRE1.
Figure 4. BRE1 mediates activation by BMP4 or Smads. The inducible reporter (−819) or the reporter lacking BRE1 (−751), as diagrammed in Figure 3, was co-injected into Xenopus embryos with RNA encoding Smad1, BMP4, Smad5 or Smad8, as indicated. Smad1 activates the reporter equivalent to BMP4, and in both cases this is dependent on the presence of the BRE1 sequences. Smad5 is only slightly less active, while Smad8 fails to induce reporter activity.
Figure 5. Mutation of an octamer element located at the 3′ end of the BRE1 is sufficient to block BMP-induced activation of the promoter. (A) A series of specific mutations were introduced into the BRE1 by site-directed mutagenesis for potential Smad-binding sites (mutations #1–5) or for an A/T-rich sequence (mutation #6), or a putative octamer binding site (mutation #7). (B) When compared to the BMP-inducible reporter containing the intact BRE1 (−819), none of the first six mutations had a significant effect on induction. In contrast, mutation of the putative octamer binding site (mutation #7) blocked BMP-mediated induction, similar to deletion of the BRE1 (−751).
Figure 6. The octamer site of BRE1 binds Oct-1 in Xenopus nuclear lysates. (A) Gel mobility shift assays were performed using a labeled probe containing the BRE1 sequences including the putative octamer binding site. Lane 0 represents probe alone and the position of the free probe is indicated (P). In lanes marked with a ‘+’, a specific complex forms using nuclear extracts derived from stage 13 embryos, as indicated by the arrow in the second lane. The complex is competed specifically by the addition in the reaction of excess (10×, 25× or 100×) unlabeled probe DNA (self) or DNA containing the octamer consensus (octamer), but not by DNA containing the same mutation of the octamer that blocks the induction by BMP4 (mutant). (B) Similar gel mobility-shift assays were performed. In this case, the lanes include probe alone (0), nuclear extract (1), nuclear extract and antibody to Xenopus Oct-1 (2) or a control isotype-matched antibody (3). The free probe is indicated (P) as are the positions of a non-specific complex (NS), the specific complex (SP) and the complex that is super-shifted by addition of the Oct-1 antibody (SS).
Figure 7. The BRE1 cooperates with a distinct BRE2 element located at ∼100 bp proximal to the Oct-1 binding site. (A) The SV40 minimal promoter was used to test if the BRE activity could be transferred to a heterologous promoter. The SV40 promoter itself is not induced by BMP4. Shown below this construct are six additional reporters containing various regions of the gata2 upstream region as indicated on the map below, placed upstream of the SV40 promoter. Sequences from −819 to −685 containing just the BRE1 (construct #1, 134 bp) does not function as a BRE, whereas the entire region from −819 to −443 (construct #2, 377 bp) is sufficient to mediate ∼3-fold activation of the SV40 promoter by BMP-4. Subsequent constructs delineated a 20 bp sequence (between −666 and −646) that is necessary for BRE activity (designated BRE2). Dark grey boxes indicate gata2 genomic sequences, black boxes indicate the SV40 promoter, and the light grey box represents the luciferase reporter gene. Black boxes on the map below represent the BRE sequences. (B) BRE2 is not active in the absence of BRE1. This is shown by transferring sequences from −751 to −443 upstream of the SV40 promoter, which fails to support BMP-mediated induction. Therefore, neither BRE1 nor BRE2 has activity on its own, but the two regions cooperate to mediate the response. Boxes represent sequences as in (A). (C) Sequences around the BRE2, defined so far by the 20 bp region between −646 and −666. Potential Smad (AGAC) or Vent-2 (ATTA) binding sites are indicated, although we have so far been unable to confirm that they are functional, either for binding or activity (data not shown).