Choudury, S. G. | bioRxiv Science

bioRxiv2025

A cryptic START domain regulates deeply conserved transcription factors

Transcription factors (TFs) integrate a diverse array of inputs to achieve the exquisite control of gene expression necessary for life. In plants, this is exemplified by the deeply conserved CLASS III HOMEODOMAIN LEUCINE ZIPPER (HD-ZIPIII) family of TFs. HD-ZIPIII activity is controlled by inputs at transcriptional, post-transcriptional, and post-translational levels. As part of their multidomain architecture, HD-ZIPIII TFs contain a StAR-related lipid transfer (START) domain, a ubiquitously distributed evolutionary module that binds various types of lipophilic ligands. Here, we show that HD-ZIPIII and HD-ZIPIV proteins contain a cryptic, deeply conserved START domain which we term the disorder-containing START domain (dSTART). The dSTART domain is required for HD-ZIPIII developmental function, controlling their subcellular localization and DNA-binding properties. The dSTART domain also helps discriminate responsive from non-responsive binding sites across the HD-ZIPIII shared genetic network. Finally, we identify candidate ligands of the dSTART domain including several species of phosphatidylglycerol and phosphatidic acid. The identification and functional characterization of a cryptic START domain provides new mechanistic insights into a deeply conserved family of TFs with roles in nearly all aspects of plant development.

plant biology↗

bioRxiv2024

START domains generate paralog-specific regulons from a single network architecture

Functional divergence of transcription factors (TFs) has driven cellular and organismal complexity throughout evolution, but its mechanistic drivers remain poorly understood. Here we test for new mechanisms using CORONA (CNA) and PHABULOSA (PHB), two functionally diverged paralogs in the CLASS III HOMEODOMAIN LEUCINE ZIPPER (HD-ZIPIII) family of TFs. We show that virtually all genes bound by PHB ([~]99%) are also bound by CNA, ruling out occupation of distinct sets of genes as a mechanism of functional divergence. Further, genes bound and regulated by both paralogs are almost always regulated in the same direction, ruling out opposite regulation of shared targets as a mechanistic driver. Functional divergence of CNA and PHB instead results from differential usage of shared binding sites, with hundreds of uniquely regulated genes emerging from a commonly bound genetic network. Regulation of a given gene by CNA or PHB is thus a function of whether a bound site is considered responsive versus non-responsive by each paralog. Discrimination between responsive and non-responsive sites is controlled, at least in part, by their lipid binding START domain. This suggests a model in which HD-ZIPIII TFs use information integrated by their START domain to generate paralog-specific transcriptional outcomes from a shared network architecture. Taken together, our study identifies a new mechanism of HD-ZIPIII TF paralog divergence and proposes the ubiquitously distributed START evolutionary module as a driver of functional divergence.

genomics↗