XB-ART-4966Development. September 1, 2003; 130 (17): 4047-56.
The pro-BMP activity of Twisted gastrulation is independent of BMP binding.
The determination of the vertebrate dorsoventral body axis is regulated in the extracellular space by a system of interacting secreted molecules consisting of BMP, Chordin, Tolloid and Twisted Gastrulation (Tsg). Tsg is a BMP-binding protein that forms ternary complexes with BMP and Chordin. We investigated the function of Tsg in embryonic patterning by generating point mutations in its two conserved cysteine-rich domains. Surprisingly, Tsg proteins with mutations in the N-terminal domain were unable to bind BMP, yet ventralized the embryo very effectively, indicating strong pro-BMP activity. This hyperventralizing Tsg activity required an intact C-terminal domain and could block the anti-BMP activity of isolated BMP-binding modules of Chordin (CRs) in embryonic assays. This activity was specific for CR-containing proteins as it did not affect the dorsalizing effects of Noggin or dominant-negative BMP receptor. The ventralizing effects of the xTsg mutants were stronger than the effect of Chordin loss-of-function in Xenopus or zebrafish. The results suggest that xTsg interacts with additional CR-containing proteins that regulate dorsoventral development in embryos.
PubMed ID: 12874126
PMC ID: PMC2277362
Article link: Development.
Grant support: GM 56326 NIGMS NIH HHS , HD 21502-17 NICHD NIH HHS , R37 HD021502-17 NICHD NIH HHS
Genes referenced: bmp1 bmp4 chrd myod1 nog pclo pmp22 sox2 tdgf1.3 tll2 tubb2b twsg1
Article Images: [+] show captions
|Fig. 2. Effects of xTsg and xTsgW67G on neural patterning and dorsal development. (A-D) Expression of Sox-2 at stage 12 in embryos either uninjected (A), or microinjected in the animal pole at the four-cell stage with a total of 1 ng xTsg mRNA (B), 4 ng xTsg mRNA (C) or 1 ng xTsgW67G mRNA (D). (E,F) Expression of N-tubulin at stage 14 in wild-type embryos (E), or embryos microinjected at the two to four cell stage with 1 ng of xTsg (F), Xolloid (G) or xTsgW67G (H) mRNA. At least 10 embryos per sample were stained in each of two independent experiments. (I) Embryo injected four times into the marginal region at the two to four cell stage with 250 pg xTsgW67G mRNA. (J) Embryo co-injected with 250 pg TsgW67G and 250 pg dominant-negative BMP receptor (tBR) mRNA. (K,L) Sox-2 in situ hybridization of stage 42 embryos. (M-O) Histological sections of control, xTsg and xTsgW67G injected embryos. sc, spinal cord; so, somites; no, notochord; pn, pronephros.|
|Fig. 1. The ventralizing activity of Tsg resides in its C-terminal cysteine-rich domain. (A) The Tsg mutations presented here. (B) Alignment of the N-terminal cysteine rich domain of Tsg proteins from Xenopus (xTsg), mouse (mTsg), human (hTsg), Drosophila (dTsg, dTsg2) and Chordin CR1. (C-J) Stage 42 wild-type control embryo (C) and embryos radially microinjected at the four-cell stage with a total of 2 ng xTsg mRNA (D), xTsgS36A (E), xTsgS54A (F), xTsgC59A (G), xTsgW67G (H), xTsgC198A (I), xTsgW67G+C198A (J). Each construct was analyzed in at least three independent experiments with n>30. (K) xTsg and mutant xTsg proteins secreted by dissociated animal cap cells after microinjection of 2 ng of the indicated mRNAs.|
|Fig. 3. Wild-type xTsg and the hyperventralizing mutant xTsgW67G antagonize the dorsalization of mesoderm by Activin. (A-D) Animal cap explants were isolated at stage 8 and treated with 2 ng/ml Activin protein. Activin induces elongation (B). Microinjection of a total of 2 ng xTsg (C) or xTsgW67G (D) mRNA at the eight-cell stage into the animal blastomeres prevented elongation. (E) The accumulation of Activin-induced endogenous Chordin protein (lane 2) was reduced by wild-type xTsg mRNA injection (lane 3) and increased by xTsgW67G (lane 4). (F) RT-PCR analysis of animal cap explants treated with Activin after microinjection of 2 ng xTsg or xTsgW67G mRNA; both mRNAs inhibited the induction of dorsal mesodermal (MyoD, alpha-Actin) or neural (NCAM) marker genes by Activin. EF1α served as a loading control.|
|Fig. 4. Biochemical analyses of xTsg mutant proteins. (A) RT-PCR of animal cap explants injected with 2 ng xTsg, xTsgW67G or xTsgC198A mRNA. (B) Immunoprecipitation of BMP4 bound to xTsg. Recombinant BMP4 (5 nM) was preincubated with either 20 nM xTsgW67G, xTsgC59A or wild-type xTsg protein (lanes 2, 3, 4). Immunoprecipitation was performed with a polyclonal antibody recognizing an N-terminal peptide present in the Tsg expression vector (Piccolo et al., 1999) and the immunoblot probed with anti-BMP4 or anti-Flag antibodies. (C) Tsg-Chordin complexes formed after crosslinking of 20 nM of the indicated affinity-purified xTsg proteins to 5 nM xChordin protein with DSS (disuccinimidyl suberate). xTsg proteins were visualized via the FLAG epitope. Note that xTsgW67G binds to Chordin (lane 4) and that xTsgC198A does not (lane 6). (D) Chemical crosslinking of BMP-Chordin-Tsg ternary complexes after incubation of BMP4 (5 nM, lane 1) with Chordin (5 nM, lane 2) and xTsg (20 nM, lane 3) or xTsgW67G (20 nM, lane 4) proteins. The western blot was stained with anti-BMP4 monoclonal antibody. (E) Anti-FLAG immunoblot of the same experiment shown in D, after stripping of the BMP4 antibody and incubation with an anti-FLAG antibody. (F) Crosslinking with DSS of 293T supernatants from cell cultures transfected separately with the indicated murine expression vectors and then co-cultured. Formation of a ternary BMP4-Chordin-Tsg complex led to a significant increase in BMP avidity by the monoclonal antibody (lane 4). The amounts of BMP4 in the supernatants vary, due to binding to the extracellular matrix of the cultured cells.|
|Fig. 5. xTsgW67G inactivates an isolated CR module but not Noggin or a dominant-negative BMP receptor. (A) RT-PCR analysis of animal caps injected with the indicated mRNAs. Xenopus CR1 mRNA (80 pg total) was injected alone or together with 1 ng xTsg or xTsgW67G mRNA. (B) Microinjection of 20 pg CR1 mRNA induces secondary axes at high frequency (79%, n=79). (C) Formation of secondary axes was completely blocked by co-injection of 250 pg xTsgW67G mRNA (0%, n=39). (D) Secondary axes were efficiently induced by microinjection of 1 ng mRNA encoding a dominant-negative BMP receptor (tBR; 67%, n=49). (E) Axis-inducing activity of tBR was not affected by co-injection of 250 pg xTsgW67G (65%, n=57). (F) Injection of 0.1 pg Noggin mRNA induced secondary axes (68%, n=40). (G) The activity of Noggin was not inhibited by co-injection of 250 pg xTsgW67G mRNA (67%, n=42).|
|Fig. 6. xTsgW67G overexpression prevents dorsalization by LiCl. (A) Untreated stage 30 embryo. (B) Radially dorsalized embryo obtained after LiCl treatment at the 32-cell stage (DAI=9.25, n=16). (C) Embryo microinjected into the marginal zone of each blastomere at the four cell stage with 500 pg xTsgW67G mRNA (DAI=1.1, n=11) or (D) microinjected with a total of 8 ng anti-Chordin morpholino oligos at the two-cell stage (DAI=6.4, n=36) prior to LiCl treatment at 32-cell stage. Note that xTsgW67G causes complete ventralization, which cannot be reversed by LiCl treatment.|
|Fig. 7. Overexpression of hyperventralizing xTsg constructs in zebrafish embryos inhibits dorsoanterior development. (A) Uninjected sibling. Wild-type zebrafish embryos were injected at the one-cell stage with 400 pg wild-type xTsg mRNA (B-D) or with 330 pg xTsgW67G mRNA (E-G). Embryos injected with wild-type xTsg mRNA in two independent experiments (n=144) displayed a range of dorsalization (as classified by Mullins et al., 1996): 13% Class 1 (not shown), 44% Class 2 (B), 17% Class 3 (C) and 24% Class 4 (D). Of the embryos displaying Class 1 or 2 dorsalizations, 51% also exhibited duplication of the terminal ventral fin (inset in B, indicated by arrowheads), suggestive of a tail ventralization in zebrafish. Of 199 embryos injected with 330 pg xTsgW67G mRNA in two independent experiments, 73% were moderately ventralized to levels comparable with chordino (E). 14% showed a phenotype more severe than chordino (F) and 3% displayed an even more ventralized phenotype (G). At higher doses (800 pg), xTsgW67G mRNA caused phenotypes stronger than chordino in 86.8% of embryos (n=91), with 18.7% of the type shown in G and 67% of the type shown in F. Note the almost complete absence of brain and trunk somites in G. Injection of 200 pg xTsgC59A mRNA into 60 embryos (not shown) also caused ventralization, with 82% appearing similar to E, and 5% resembling F.|