HDAC3 is a histone deacetylase that modulates gene expression across a range of developmental processes, and its catalytic activity is dependent on association with the NCoR/SMRT co-repressor complex. In mice, HDAC3 also exerts some of its effects on gene expression via non-enzymatic means. Here, Mattias Mannervik and colleagues explore three different mutations (Y303F, K26A and HEBI) that reduce HDAC3's catalytic activity in Drosophila via different mechanisms. They show that HDAC3 depletion in fly embryos impairs hatching, embryonic patterning and segmentation. The hatching defects can be rescued by the Y303F and HEBI mutations, but not by K26A. By contrast, the segmentation defects can be rescued by HEBI, but defect frequency increases in Y303F-rescued embryos. This suggests that the regions targeted by these different mutations are associated with unique HDAC3 functions. The authors use protein-DNA interaction data and RNA sequencing to identify direct targets of HDAC3 that are dysregulated upon HDAC3 knockdown. They find that downregulated gene expression is more efficiently rescued by wild-type HDAC3 than by the mutant constructs, whereas upregulated gene expression is most efficiently rescued by the HEBI mutation. Overall, this work suggests that HDAC3's functions can be decoupled, with catalytic activity required for embryonic gene activation but not for repression.