A wild-type, E7.5 mouse embryo (left), stained with an anti-Amn antibody. Amn is present only in the visceral endoderm (VE). The apical localization of Amn in VE cells (right). Courtesy of Sundeep Kalantry and Katia Manova, Memorial Sloan-Kettering Cancer Center, USA.

Unravelling the role of BMP antagonists in controlling and fine tuning embryogenesis is a hot topic in developmental biology at present (see accompanying Highlight). And who'd have predicted that getting to the bottom of a mouse with no trunk might add further to the twists in this field? For that's what Elizabeth Lacy and colleagues might have done by identifying a novel, BMP-antagonist-like, gene that when mutated abolishes trunk development in mouse embryos.

The recessively lethal amnionless (amn) mutation, generated by a transgene insertion, disrupts the middle region of the primitive streak that gives rise to trunk mesoderm. After many years of work on amn, Lacy and colleagues made their breakthrough when they found a BAC that rescued the amn phenotype. Their analysis showed that this BAC spanned the insertion site and contained three genes, but only one of them in its entirety. Lacy's team found deletions in this gene in amn mutants at the transgene integration site and confirmed its role by knocking it out to reproduce the amn mutant phenotype.

The amn gene encodes a novel type I transmembrane protein, the extracellular domain of which shows sequence similarity to the cysteine-rich (CR) domains of BMP inhibitors such as Sog, Tsg and chordin (see accompanying Highlight). These CR domains mediate the activity of BMP antagonists by binding BMPs and sequestering them away from their receptors. In a comparison of mouse amn to its fly and human homologues, the authors found it was this amino-terminal, CR-containing region that was most highly conserved.

Expression studies showed that amn is exclusively expressed in the extra-embryonic visceral endoderm (VE) tissue, but curiously throughout the VE and not just where it overlies the middle primitive streak. Furthermore, amn is expressed on the apical surface of the VE, which faces away from the embryonic epiblast (see picture). This raises the question of how Amn mediates its effects without directly interacting with this embryonic tissue. The authors tackle this by proposing that Amn modulates Bmp2 signalling within the VE itself, controlling targets that might themselves directly interact with embryonic tissue. Amn might also act in concert with a Bmp receptor, Alk2, which is present on the VE's apical surface, to modify the activity of other extra-embryonically expressed Bmp molecules, such as Bmp7. Future functional studies should tell whether Amn is to join this hotly pursued family of signalling antagonists.