Novel insights into the structure and evolution of the human SAGA complex by affinity-ligand purification

Human SAGA is a 20-subunit complex that stimulates transcription and is essential for development. The most prominent addition to SAGA in metazoans compared to yeast is a 150kDa splicing-factor module (SPL). SPL is also a part of the U2snRNP but its role in SAGA is elusive, partially due to absence of high-resolution structural information regarding its incorporation into the complex. In yeast, subunit TAF5 and TAF6 of SAGA are shared with the general transcription factor TFIID. In metazoans, gene duplication created proteins that occur only in SAGA (TAF5L and TAF6L) or in TFIID (TAF5 and TAF6). What functions of SAGA benefit from this protein specialization is unclear. Here we report the structure of endogenous human SAGA purified via an affinity-ligand from cells that were not disturbed by any genomic engineering tools such as CRISPR-Cas9. Our work reveals the high-resolution structure of SPL and of the TAF6L HEAT repeat domain that provides the SPL with a docking surface. We elucidate how SPL and the HEAT repeats are incorporated into SAGA. We find multiple major differences between TAF6L/TAF5L and the canonical paralogues that are directly implicated in structural re-arrangements required to accommodate SPL. Furthermore, SPL binding to SAGA is very different and occupying much less interaction surface than to U2snRNP. However, the two cases still share similar sequences in a helix that is deeply inserted into the SPL. The seemingly weaker interaction of SPL with SAGA raises the possibility that SAGA serves to relay this module to the splicing machinery. Our structure also suggests mutations that could uncouple SPL from SAGA to further interrogate the role of this module.