Nature, 410, 475 - 478 (2001) .

Figures 2 and 3g contained a number of errors. The correct figures are reproduced below.

Figure 2: Tsg alters cleavage of chordin by BMP1 and mTll1, and directly interacts with chordin and BMP4 to enhance chordin/BMP4 complex formation.
figure 1

a, Immunoblot of Flag-tagged chordin incubated with BMP1 or mTll1 in the presence (+)/absence (-) of Tsg. Bands were detected using anti-Flag antibodies. b, Electrophoretic patterns, visualized by zinc staining6, for chordin incubated with BMP1 in the presence/absence of Tsg. c, Positions of major chordin cleavage sites (arrows) utilized in the presence/absence of Tsg. d,e, Immunoblots of proteins precipitated with anti-protein C antibody (anti-PC) after incubation of (d) Flag-tagged BMP1, mTll1, or (e) chordin, alone or with protein C-tagged Tsg (Tsg-pc). f, Immunoblot of proteins precipitated with anti-BMP4 antibodies (anti-BMP4), after incubation of BMP4 with chordin-Flag with/without Tsg. g,h, Immunoblots of proteins immunoprecipitated with (g) anti-PC or (h) anti-BMP4 antibodies, after incubation of BMP4 and Tsg separately or together.

Full size image
Figure 3g: Tsg potentiates secondary axis induction by chordin and chordin fragments and increases their binding of BMP4.
figure 2

a, Effects of Tsg overexpression in Xenopus embryos. b, Whereas 1 pg chordin RNA induced no axes (0/23), co-injection of 1 pg chordin and 45 pg Tsg RNA induced axes in 83% of embryos (n = 49). Tsg RNA alone (50 pg) induced no axes (n = 24) (not shown). c, Whereas 300 fg chordin or 45 pg Tsg RNA alone induced no axes (n = 25 each), co-injection of the two induced axes in 80% (n = 25) and strong Xenopus Vent2 inhibition in 75% (n = 25) of embryos (arrows show areas of strong or weak Vent2 inhibition in ventral marginal zones). d, Equimolar amounts of CR1 (100 pg), 65K internal ‘Int.’ (410 pg), CR2/CR3 (150 pg) or CR4 (69 pg) fragment RNA were injected with/without 100 pg Tsg RNA. Without Tsg RNA, no axes were induced (CR1, 0/43; Int., 0/24; CR2/CR3, 1/26; CR4, 0/28) or embryos dorsalized, (CR1, 0/43; Int., 0/24; CR2/CR3, 0/26; CR4, 0/28). In the presence of Tsg, CR2/CR3 efficiently induced axes (84%, n = 25), CR1 induced a lesser number (21%, n = 24) of small axes, and other fragments inefficiently induced axes (Int., 1/27; CR4, 0/23). CR1+Tsg dorsalized embryos (67%, n = 24), including those with secondary axes. e, Results of d (in a separate experiment, CR1+Tsg induced a higher proportion of secondary axes and lower proportion of dorsalized embryos, relative to CR2/CR3, than in the experiment shown). f, Co-injection of 100 pg each Tsg and CR2/CR3 RNA efficiently induced axes (93%, n = 30), whereas 100 pg Tsg RNA alone did not induce axes (0%, n = 25). 100 pg CR2/CR3 alone inefficiently induced axes (4%, n = 25). In all experiments shown in panels a–f, untagged versions of CR and Tsg constructs were used. In control experiments, similar to d–f, enhancement of axis formation by CR1 in the presence of Tsg is still efficient when RNAs for the two are co-injected with BMP1 RNA (not shown). This excludes the possibility that Tsg enhancement of axis induction in such experiments is through interaction with endogenous full-length chordin, induced via a positive feedback loop10 by overexpressed chordin fragments. g, Immunoblot of proteins precipitated with anti-PC antibodies after incubation of each CR-containing fragment alone or with Tsg-pc. h, Immunoblot of proteins precipitated by antibodies (anti-N-chordin) to sequences in the CR1 fragment after incubation of CR1 fragment (1.7 nM) with BMP4 (1.4 nM) with/without Tsg (1.3 nM).

Full size image