Integrative sequence–structure analysis reveals hidden WD40 domains forming stable β-propeller folds with potential biological functions in plants

WD40 proteins form β-propeller structures that are essential for plant development and signaling; however, their complete domain architecture is often missed owing to high sequence divergence. Here, we reannotated 117 plant genomes and found that many WD40 proteins contain fewer than the canonical seven domains, raising questions about their functionality as complete WD40 proteins. Structure-based modeling of 17,369 WD40-only genes that lack other associated domains revealed 41,379 additional WD40 domains that were entirely missed by sequence-based annotation but are required for the formation of stable β-propeller structures. Despite being annotated as partial, two rice genes with sequence-invisible domains form complete β-propeller folds and are critical for pollen development. CRISPR–Cas9 knockouts showed that OsWD40-31 is required for pollen tube elongation and that OsWD40-169 is necessary for pollen germination. Disruption of these sequence-invisible domains reduced binding affinity to the key reproductive regulators PME1 and Lipase3, as confirmed by structural interaction modeling, yeast two-hybrid assays, and co-immunoprecipitation assays. Residue substitution revealed that the stability of the WD40 domain depends on hydrogen-bonding residues not captured by sequence conservation, explaining why many functional domains evade sequence-based detection. These findings highlight a fundamental disconnect between sequence conservation and structural integrity, establishing a structure-guided framework for uncovering hidden domain architectures in complex, repeat-rich gene families across plant genomes.