A BMP gradient is essential for patterning the dorsal-ventral axis of invertebrate and vertebrate embryos. The extracellular BMP binding protein Short Gastrulation (Sog) in Drosophila plays a key role in BMP gradient formation. In this study, we combine genome editing, structural and developmental approaches to study Sog function in Drosophila. We generate a sog knockout fly stock, which allows simple reintegration of altered versions of the sog coding sequence. As proof-of-principle, we test the requirement for two cysteine residues that were previously identified as targets for palmitoylation, which has been proposed to enhance Sog secretion. However, we show that the SogC27,28S mutant is viable with only very mild phenotypes, indicating that these residues and their potential modification are not critical for Sog secretion in vivo. Additionally, we use experimental negative stain EM imaging and hydrodynamic data to validate the AlphaFold structure prediction for Sog. The model suggests a more compact shape than the vertebrate ortholog Chordin and conformational flexibility between the C-terminal von Willebrand C domains. We discuss how this altered compactness may contribute to mechanistic differences in Sog and Chordin function during BMP gradient formation.

A BMP gradient is essential for patterning the dorsal-ventral axis of invertebrate and vertebrate embryos. The extracellular BMP binding protein Short Gastrulation (Sog) in Drosophila plays a key role in BMP gradient formation. In this study, we combine genome editing, structural and developmental approaches to study Sog function in Drosophila. We generate a sog knockout fly stock, which allows simple reintegration of altered versions of the sog coding sequence. As proof-of-principle, we test the requirement for two cysteine residues that were previously identified as targets for palmitoylation, which has been proposed to enhance Sog secretion. However, we show that the SogC27,28S mutant is viable with only very mild phenotypes, indicating that these residues and their potential modification are not critical for Sog secretion in vivo. Additionally, we use experimental negative stain EM imaging and hydrodynamic data to validate the AlphaFold structure prediction for Sog. The model suggests a more compact shape than the vertebrate ortholog Chordin and conformational flexibility between the C-terminal von Willebrand C domains. We discuss how this altered compactness may contribute to mechanistic differences in Sog and Chordin function during BMP gradient formation.Summary statementThe authors model the structure of Sog and establish a new sog knockout fly stock that they validate for the testing of specific sog mutations.