ARTICLES PUBLISHED: 2 MARCH 2015 | ARTICLE NUMBER: 15014 | DOI: 10.1038/NPLANTS.2015.14 Patterning mechanisms of cytoskeletal and cell wall systems during leaf trichome morphogenesis Makoto Yanagisawa1, Anastasia S. Desyatova2, Samuel A. Belteton1, Eileen L. Mallery1, Joseph A. Turner2 and Daniel B. Szymanski1,3* The plant actin cytoskeleton is an unstable network of filaments that influences polarized growth through poorly understood mechanisms. Here, we used a combination of live cell imaging and finite element computational modelling of Arabidopsis trichome morphogenesis to determine how the actin and microtubule cytoskeletons cooperate to pattern the cell wall and growth. The actin-related protein (ARP)2/3 complex generates an actin meshwork that operates within a tip- localized, microtubule-depleted zone to modulate cell wall anisotropy locally. The actin meshwork also positions an actin bundle network that organizes organelle flow patterns. This activity is required to maintain cell wall thickness gradients that enable tip-biased diffuse growth. These newly discovered couplings between cytoskeletal patterns and wall textures provide important insights into the cellular mechanism of growth control in plants. he microtubule and actin cytoskeletons cooperate to influence Fig. 1 and Supplementary Movie 1). Branch diameter at the base 1 2 shape change in both metazoan and plant cells. In plants, was nearly constant during growth, as would be expected for tip T 4 these cytoskeletal systems locally modulate the mechanical growth . Unlike the tip growth of pollen tubes and root hairs in properties of the cell wall3,4, and in the epidermis, heterogeneities which the tip radius is quite constant, the radius of curvature at in wall texture guide the complex patterns of irreversible cell expan- the tip decreased from ∼4µmto∼0.5 µm as the branch elongated sion5. In leaves, these growth patterns are important because they (Fig. 1a,b). Branch growth rates varied considerably during mor- can persist for days6,7 and influence organ size and shape8. phogenesis (Supplementary Fig. 1b), but there was a general trend Microtubule-dependent cellulose microfibril patterning is one towards higher elongation rates as the branches increased in central feature of polarized diffuse growth9,10; however, the means length. For example, short branches between 20 and 30 µm – by which the actin and microtubule cytoskeletons cooperate elongated at a rate of 3.1 ± 1.0 µm h 1, which was significantly less during cell morphogenesis is poorly understood. than the growth rate of branches in the 50–60 µm range, which – Leaf trichomes, specialized leaf epidermal cells that protect the was 5.8 ± 1.5 µm h 1 (P = 0.003). On the basis of the shape change growing organ against insect attack11, like single-celled cotton that occurs and on direct measurements of wall strain using cell fibres12, are model cell types that can be used to analyse how the wall bound pariticles13,22, it is known that trichome morphogenesis microtubule and actin cytoskeletons control polarized growth13,14. includes a diffuse growth component. To measure wall strain The distorted Arabidopsis leaf trichome mutants identified an evolu- patterns directly at high spatial resolution, externally applied tionarily conserved actin filament nucleation pathway, in which the fluorescent beads were used as fiducial marks along the branch WAVE/SCAR regulatory complex (W/SRC)15 positively regulates axis (Fig. 1c). Particle tracking of bead pairs detected an axial the actin filament nucleator ARP2/316–18. Although it is known strain gradient, in which distal regions of the branch grew at that W/SRC–ARP2/3 is required to maintain organized actin significantly faster rates than the branch base (Fig. 1d). Therefore, bundle networks and polarized cell elongation13,17,19–21, the mechan- the early stages of branch elongation include highly polarized isms by which it influences the cell wall properties that determine anisotropic diffuse growth and tip refinement to a final radius of the growth behaviours of the cell are not known. Here we integrate 0.61 ± 0.12 µm (n = 20). Our analysis captured the significant live cell imaging and computational analyses of the cell wall to shape change during branch development because mature branches determine how an ARP2/3-generated apical actin meshwork influ- that exceed 200 µm in length have similar length to width aspect ences cell wall properties that define the geometry of shape change. ratios and tip geometry13. Organized microtubules and cellulose microfibrils are required Creation of a plausible finite element model for polarized branch growth10,13,14. To quantify the degree of poten- Branch elongation and tip refinement are defining features of tial microtubule-based microfibril ordering, we measured the trichome morphogenesis13, but at present there are no clear hypoth- microtubule angle distribution relative to the longitudinal axis of eses about what combinations of cell wall mechanical properties the cell (Fig. 1f). As expected, microtubules and the cellulose would be needed to generate this type of shape change. Finite synthase (CESA) complex had a similar transverse alignment element computational modelling is a useful framework to simulate (Fig. 1e,g). However, the branch apex always contained an the complex relationships between turgor pressure, cell wall mech- obvious microtubule-depleted zone (MDZ) that lacked patterned anical properties and growth4. To parameterize a finite element CESA localization. An MDZ has been previously reported in the model of the cell wall and branch elongation, we conducted a series literature using different microtubule markers21,23, indicating that of reflected light time-lapse imaging experiments (Supplementary the branch apex wall has a more isotropic character than the flank. 1Department of Botany and Plant Pathology, Purdue University, West Lafayette, Indiana 47907, USA. 2Department of Mechanical and Materials Engineering, University of Nebraska-Lincoln, Lincoln, Nebraska 68588, USA. 3 Biological Sciences, Purdue University, West Lafayette, Indiana 47907, USA. *e-mail: [email protected] NATURE PLANTS | VOL 1 | MARCH 2015 | www.nature.com/natureplants 1 © 2015 Macmillan Publishers Limited. All rights reserved. ARTICLES NATURE PLANTS DOI: 10.1038/NPLANTS.2015.14 a b 4.0 Base widthwidth 60 BranchBranch lengthlength 3.5 TipTip radiusradius of curvaturecurvature m) Tip radius of curvature ( of curvature radius Tip µ 50 3.0 00:00 2.5 40 2.0 30 1.5 µ 20 m) 1.0 05:00 Branch length and base width ( Branch 10 0.5 0 0.0 0:00 2:00 4:00 6:00 8:00 10:00 Time (h) 10:00 c 00:00 04:00 d ** 20 16 0 1 15 2 3 4 1 16 6 10 0 5 8 0.8 1 7 2 3 4 (%/h) rate Strain 9 10 0.6 5 5 6 8 7 11 0.4 9 10 0 11 0 0.2 0.4 0.6 0.8 1.0 0.2 12 1312 13 15 Initial mean relative position 15 14 0 14 e g MBD–GFP f 60 40 20 Percentage of pixel pairs (%) pairs of pixel Percentage 0 0° (180°) 45° 90° 135° YFP-CESA6 MT angle relative to growth axis Figure 1 | The geometry of trichome shape change and cell wall strain. a,Reflected light time-lapse imaging of trichome morphogenesis. b,Branchlength (red), base width (blue), and tip radius of curvature (green) plotted as a function of time for a representative wild-type branch. c, Local wall strain analysis using fluorescent beads. Projected image of the tracked bead movement over a 4 h time course. Yellow dashed lines indicate cell outlines. d, Strain rates as a function of normalized branch length with 0 corresponding to the base and 1 the branch tip. Values are mean ± s.d. (n =43),**P < 0.01 (Student t-test). e, Microtubule localization visualized with the MBD–GFP reporter. Inset is an enlarged view of the microtubule-binding domain (MDZ) at the branch apex. Arrows mark branch apex. f, Microtubule angle distributions with 90° defined as transverse to the growth axis. Values are mean ± s.d. (n = 14 cells). g, Yellow fluorescent protein (YFP)-CESA6 expressing trichome branch showing linearly aligned CESA6 marked with red arrowheads. Time stamps in a and c indicate hours:minutes. Scale bars, 10 µmina and e;5µminc and g. Finite element modelling has been used to analyse the shape degree of fibre orientation strongly affected anisotropic expansion change of thin-walled pressurized plant cells4, and can incorporate (Fig. 2b). The measured microtubule angle distributions (Fig. 1f) realistic material models of fibres reinforcing a matrix24,25. Spatial provided experimental support for highly ordered fibres with heterogeneities in the model were introduced by subdividing the κ = 0.141 ± 0.029, which could generate branches with the observed shell from the tip to the base into multiple sections with varying high length to width aspect ratios (Fig. 2b). Anisotropic branch properties (Fig. 2a and see Methods). Finite element simulations expansion was also sensitive to the ratio of fibre to matrix elastic fi identi ed wall parameters that affected the rate of cell expansion moduli (k1/E0) of the wall (Fig. 2c). A sensitivity analysis of the κ and growth anisotropy (Supplementary Table 1). For example, the branch growth geometry to combinations of k1/E0 and parameters 2 NATURE PLANTS | VOL 1 | MARCH 2015 | www.nature.com/natureplants © 2015 Macmillan Publishers Limited. All rights reserved. NATURE PLANTS DOI: 10.1038/NPLANTS.2015.14 ARTICLES 18 Intersection points a b 90 κ c 19.0 d 16 Isotropic region = 0.05 k1/E0 = 10 Fit 80 κ = 0.1 18.5 14 Experimental κ 0 m) k E m) / = 100 12 Anisotropic region κ E µ 1 0 / = 0.2 µ 0.8 18.0 1 10 70 κ = 0.25 k E – ‘matrix’ modulus 8