|
Fig. 1. Structural and functional analysis of the Xenopus Xcad3. (A) The Xcad3 genomic clone. Three exons and two introns are depicted with their respective nucleotide length together with 5′ and 3′ flanking regions. The transcription and translation initiation sites are numbered as +1 and +175, respectively. (B) The regulatory function of the genomic Xcad3 sequences. The 5′ flanking (–7000 to +174) and intron 1 sequences were inserted into a firefly luciferase (LUC) reporter plasmid pGL3, as illustrated left of the histogram. These Xcad3/LUC constructs were injected into eight-cell stage embryos together with an internal standard plasmid pRL-CMV that contained Renilla luciferase coding sequence. Injection sites are depicted right of the histogram as PNT and AB. Regions around these sites give rise to the posterior neural tube and anterior brain, respectively. Luciferase activities were measured at stage 23 and Xcad3/LUC reporter activities normalized to the pRL-CMV internal standard activities were presented in the histogram with arbitrary units. (C) Transgenic embryo carrying Xcad3/GFP construct. The reporter construct used for transgenesis is indicated on top. The transgenic embryo shown was photographed at stage 37. (D) Involvement of Fgf signaling in enhancing function of the 5′ flanking and intron1 sequences. Xcad3/LUC reporter constructs indicated left of the histogram were injected into the PNT site together with mRNA (100 pg/blastomere) encodingδ XFgf-4a or d59 (a control inactive version of δXFgfR-4a). The d59 lacks a stretch of 59 amino acids that is required for dimerization with a wild-type receptor subunit (Amaya et al., 1991). Luciferase activities were analyzed and presented as in B.
|
|
Fig. 2. Analysis of Fgf response element (FRE) of Xcad3. (A) Experimental design for the embryonic cell culture assay used in B. Ectodermal tissues were isolated from stage 10 gastrula embryos. The dissociated cells were inoculated into microculture wells at 200 cells/well. After completion of reaggregation by brief centrifugation, cells were cultured in the presence of increasing concentrations of bFgf until control embryos reached stage 23. The transcriptional levels of two position-specific neural markers were analyzed by RT-PCR (Kengaku and Okamoto, 1995). (B) High-dose-dependent activation by Fgf of endogenous Xcad3. Autoradiographs are shown of RT-PCR products of the transcripts from En2, an anterior neural marker gene and Xcad3, both of which were co-amplified with EF1α transcript, an internal standard (upper panels). Each RT-PCR product was quantified by a laser image analyzer and values for En2 (▴) and Xcad3 (δ) transcripts normalized to EF1α transcript are presented as percentages of the respective maximum value and plotted against bFgf dose (graph). (C) Experimental design for the embryonic cell culture assay used in D. Xcad3/LUC reporter and pRL-CMV plasmids were coinjected into four animal blastomeres of eight-cell stage embryos. When they reached stage 10, ectodermal tissues were isolated and processed as in A. To compare directly Fgf dose-dependence of Xcad3/LUC reporter with that of endogenous Xcad3, two parallel sets of cultures were prepared; one was assayed for luciferase activity, while the other was assayed for transcriptional levels of endogenous Xcad3. (D) Comparison of the Fgf dose-dependence profiles for a Xcad3/LUC reporter and endogenous Xcad3. Eight-cell stage embryos were injected with a Xcad3/LUC reporter depicted below the graph and an internal standard plasmid pRL-CMV and processed as described in C. Normalized Xcad3/LUC reporter activities are presented as percentages of the maximum value and plotted against bFgf dose (δ). Transcript levels of endogenous Xcad3 was assayed and plotted as in B (○). (E,F) Presence of FRE in the intron 1. Chimeric constructs injected are indicated below each graph: they contained either Xcad3 intron1 (E) or SV40 enhancer sequence (F). Reporter activities of these constructs were analyzed as in D and presented in arbitrary units. Note that inclusion of intron1 sequence in reporter constructs is essential for dose-dependent response to Fgf.
|
|
Fig. 3. Identification of multiple FREs within the intron 1. (A) Deletion and sequence analysis of the intron 1. Reporter constructs examined contain intron 1 sequences that are deleted to various extents as depicted left of the histogram. Reporter activities of these constructs were analyzed as in Fig. 2, except that ectoderm cells were cultured in the absence or presence (6 ng/ml) of bFgf. The ratio of the luciferase activity obtained with bFgf to that obtained without bFgf is taken as a measure of Fgf responsiveness and presented as a value of fold induction in the histograms. Numbers on full-length intron 1 indicate nucleotide length of respective segments. Multiple FREs appear to be distributed within domain1 and domain 2 (light green) of intron 1. Domain 2*, as defined in the text, is also indicated close to domain 2. Sequence analysis of intron1 reveals that a number of Ets-binding motifs (EBM; A/CGGAA/T) and Tcf/Lef-binding motifs (TLBM; CTTTGA/TA/T) are present throughout intron1 as indicated by blue and red triangles, respectively. (B) Deletion analysis of a 209 bp fragment within domain 2*. Consecutive deletion of binding motifs TLBM3, EBM4 and TLBM2, as indicated left of the histogram results in a progressive reduction of Fgf responsiveness. Numbered red and blue triangles point to the position of TLBMs and EBMs in the 209 bp fragment, respectively (see Fig. 4A for precise location).
|
|
Fig. 4. EBMs and TLBMs as sub-elements of FREs. (A) Nucleotide sequence of the domain 2* region of Xcad3 intron 1. EBMs and TLBMs are marked and numbered. (B,C) Mutational analysis of EBMs (B) and TLBMs (C). Reporter constructs examined contain the domain 2* fragments that carry various combinations of mutations in EBMs or TLBMs, as indicated to the right of each graph (E1, 3 means that EBM1 and EBM3 are mutated). Mutations were introduced as follows: GAA to agA in E1, 2 and 4; TCC to ctC in E3 and 5; ATCCT to gTCga in E6; TTTGT to Tcgag in T1; TCAAA to TagAc in T2; TCAAAGG to Tgtcgac in T3; CTTTG to gaaTt in T4. Reporter activities of mutated constructs were analyzed as in Fig. 3, except that ectoderm cells were cultured in the absence or presence of increasing concentrations of bFgf.
|
|
Fig. 5. Involvement of Ets transcription factors as activators in the Fgf response of Xcad3. Structural features of Ets proteins and their derivatives examined are illustrated at the bottom right. Black boxes represent the DNA-binding ets domains. Synthetic mRNAs encoding these proteins were coinjected with –546/LUC/intron1 and pRL-CMV into two animal blastomeres (either left or right side) of eight-cell stage embryos, which were then processed as in Fig. 2C. Reporter activity was analyzed as in Fig. 2D and presented in arbitrary units. (A) Effects of dominant-negative Ets proteins on the Fgf response of Xcad3/LUC. The injected amount was 32 pg/blastomere for dnElk1 (▴), dnXEts1 (♦), dnXER81 (▪) and EnR (δ) mRNA (control). (B) Effects of wild type Ets proteins on the Fgf response of Xcad3/LUC. The injected amounts are 30 pg/blastomere for Elk1 (▴), XEts1 (♦), XER81 (▪) and EnR (δ) mRNA. Note that XEts1 and XER81 activate the reporter gene more efficiently than Elk1. (C) Reversal of dnElk1 induced suppression of Fgf response by XEts1 and XER81. The injected amounts were 24 pg, 48 pg, 72 pg and 96 pg/blastomere for dnElk1, XEts1, XER81 and EnR mRNA. The total amount of injected mRNA was adjusted by adding neutral EnR mRNA. Note that the suppression of reporter gene expression induced by dnElk1 (▴) is reversed by the addition of XEts1 (♦) or XER81 (▪).
|
|
Fig. 6. Involvement of XTcf3 as a repressor in the Fgf response of Xcad3. Structural features of XTcf3, its derivatives and XLef1 examined are illustrated at the bottom. β-catenin binding domains, VP16 activation domain and DNA-binding HMG domains are marked. Experimental procedures are as described in Fig. 5. (A) Effects of dominant-negative XTcf3 on the Fgf response of Xcad3/LUC. The injected amounts were 1.8 (♦), 5.3 (▪) or 16.0 (▴) pg dnXTcf3 mRNA/blastomere. (B) Effects of wild-type XTcf3 on the Fgf response of Xcad3/LUC. The injected amounts were 1.8 (♦), 5.3 (▪) or 16.0 (▴) pg XTcf3 mRNA/blastomere. (C) Effects of XLef1 and VP16-dXTcf3 on the Fgf response of Xcad3/LUC. The injected amount was 6 pg/blastomere for XLef1 (▪), VP16-dXTcf3 (▴) and EnR mRNA (δ). VP16-dXTcf3 enhances the Fgf response of the reporter construct (–546/LUC/intron1), but XLef1 does not. (D) Effects of β-catenin on the Fgf response of Xcad3/LUC. The injected amounts were 32 (▪) or 64 (▴) pg β-catenin mRNA/blastomere.
|
|
Fig. 7. Involvement of Sox2 as a co-activator in the Fgf response of Xcad3. Structural features of Sox2, its derivatives and SoxD BD(–) examined are illustrated at the bottom. DNA-binding HMG domains and EnR repressor domain are marked. Experimental procedures are as described in Fig. 5 except that the reporter construct (–546/LUC/domain2*) was used in C. (A) Effects of dominant negative Sox2 on the Fgf response of Xcad3/LUC. The injected amounts were 5.3 (▴), 16 (♦) or 48 (▪) pg Sox2-EnR mRNA/blastomere. Sox2-EnR suppresses the Fgf response of the reporter construct (–546/LUC/intron1) in a dose-dependent manner. (B) Effects of wild-type Sox2 on the Fgf response of Xcad3/LUC. The injected amounts were 5.3 (▴), 16 (▪) or 48 (♦) pg Sox2 mRNA/blastomere. (C) Reversal of Sox2-EnR induced suppression of Fgf response by wild-type Sox2. The injected amounts were 16 pg, 96 pg and 112 pg/blastomere for Sox2-EnR, Sox2 and EnR mRNA. Note that the suppression of reporter gene (–546/LUC/domain2*) expression induced by Sox2-EnR (▴) is reversed by the addition of Sox2 (▪). (D) Effects of Sox2 BD(–) (▴) and SoxD BD(–)(▪) on the Fgf response of Xcad3/LUC. The injected amounts were 160 pg/blastomere for both Sox2 BD(–) and SoxD BD(–) mRNA.
|
|
Fig. 8. Direct interaction of EBM and TLBM with XTcf3, Sox2 and XEts1 proteins. An end-labeled probe, either wild type or mutated, was incubated with XTcf3 or Sox2, or with a combination of both proteins in A, and with Sox2 or XEts1, or with a combination of both proteins in B, as indicated in diagrams above. All proteins were tagged with a V5 epitope and made by in vitro translation. Probe T2/E4/T3: an intronic DNA fragment containing TLBM2, 3 and EBM4 (overlined in Fig. 4A). Probe E4: TLBM2 and 3 mutated. Probe T2/E4: TLBM3 mutated. Probe E4/T3: TLBM2 mutated. Mutations were introduced as described in Fig. 4. Unlabeled probe (competitor) or antibody against V5 epitope was added as indicated in the diagrams.
|
|
Fig. 9. Intron 1 as an immediate early target of Fgf signaling (A) and a proposed model of integration of multiple signaling pathways on FREs of Xcad3 intron1 (B). In A, experimental design was essentially the same as in Fig. 2C, except for the following changes. A reporter construct (–546/LUC/intron1) was injected into animal blastomeres of eight-cell stage embryos. Ectodermal cells at stage 10 were cultured for 30 minutes with or without 10 μg/ml cycloheximide and then for an additional 1.5 hours with or without bFgf (10 ng/ml). The transcriptional levels of the reporter gene (luciferase-coding portion) were analyzed by RT-PCR.
|
