Spatio-temporal orientation of microtubules controls conical cell shape in Arabidopsis thaliana petals

The physiological functions of epidermal cells are largely determined by their diverse morphologies. Most flowering plants have special conical-shaped petal epidermal cells that are thought to influence light capture and reflectance, and provide pollinator grips, but the molecular mechanisms controlling conical cell shape remain largely unknown. Here, we developed a live-confocal imaging approach to quantify geometric parameters of conical cells in Arabidopsis thaliana (A. thaliana). Through genetic screens, we identified katanin (KTN1) mutants showing a phenotype of decreased tip sharpening of conical cells. Furthermore, we demonstrated that SPIKE1 and Rho of Plants (ROP) GTPases were required for the final shape formation of conical cells, as KTN1 does. Live-cell imaging showed that wild-type cells exhibited random orientation of cortical microtubule arrays at early developmental stages but displayed a well-ordered circumferential orientation of microtubule arrays at later stages. By contrast, loss of KTN1 prevented random microtubule networks from shifting into well-ordered arrays. We further showed that the filamentous actin cap, which is a typical feature of several plant epidermal cell types including root hairs and leaf trichomes, was not observed in the growth apexes of conical cells during cell development. Moreover, our genetic and pharmacological data suggested that microtubules but not actin are required for conical cell shaping. Together, our results provide a novel imaging approach for studying petal conical cell morphogenesis and suggest that the spatio-temporal organization of microtubule arrays plays crucial roles in controlling conical cell shape.Author summary: How cells achieve their final shapes is a fundamental question in biology. Most flowering plants have special conical-shaped petal epidermal cells that are thought to attract pollinators, but the molecular and genetic mechanisms that control conical cell shape remain unknown. In this study, we developed a live-confocal imaging approach for the quantitative study of conical cell morphogenesis. Through genetic screens, we showed that A. thaliana KTN1, ROP GTPases, and SPIKE1 are required for conical cell shaping. Live-cell imaging showed that loss of KTN1 prevented random microtubule networks from shifting into well-ordered microtubule arrays at later developmental stages, which is correlated with the tip sharpening of conical cells. Moreover, genetic and pharmacological data suggested that microtubules but not actin are required for conical cell shaping. Together, our findings provide significant insights into the spatio-temporal organization of microtubules that controls conical cell development.