Research
With few exceptions, plant tissues are composed of non-motile cells whose shapes and positions are fixed by cell walls. Consequently, the cellular organization of plant tissues directly reflects the orientation of new cell walls during cell division. Achievement of appropriate organ and cell shapes depends critically on the orientation of cell expansion, and to some extent also on patterns of cell division. We are studying cytoskeleton-dependent mechanisms governing the spatial control of cytokinesis, asymmetric cell division, and the spatial regulation of cell expansion. Recent and current work in each of these areas is outlined below.
Schematic of cytokinesis in a somatic plant cell illustrating the preprophase band (PPB), spindle, phragmoplast and other features. TANGLED protein co-localizes with the PPB and remains at the cortical division site throughout mitosis and cytokinesis. The prerophase/prophase stage is shown in 3-D but other stages are shown in cross section for clarity. See paragraphs below for more details.
Spatial Control of Cytokinesis
Plant cells achieve cytokinesis by building a new cell wall between the daughter cells. Unlike animal cells, division planes of plant cells are established prior to mitosis and are marked by a cytoskeletal structure unique to plant cells called the preprophase band (PPB), which appears during the G2 phase of the cell cycle and appears to be critical for the establishment of a cortical division site where the new cell wall will later become attached. However, the PPB breaks down upon formation of the mitotic spindle. After mitosis, cytokinesis is initiated by a plant-specific cytoskeletal structure called a phragmoplast, which directs the formation of a new cell wall by transporting vesicles containing the raw materials for cell wall formation to its midline. The phragmoplast forms between the daughter nuclei and subsequently expands laterally as wall formation proceeds until the new cell wall attaches to the mother cell wall at the former PPB site. For recent reviews on this topic, see Wright and Smith, 2008; Müller et al., 2009. Our research addresses questions such as how cells choose a division plane that is appropriate for their shape and developmental context, how that choice is translated into the establishment of a division plane marked by the PPB, how cytoskeletal filaments within the PPB participate in establishment of the cortical division site, and how this division site interacts with the expanding phragmoplast to position the new cell wall.
Arabidopsis TAN::YFP localizes as a ring at the division site throughout mitosis and cytokinesis. A-P, Dual localization of TAN::YFP (monochrome in second column, green in third and fourth column) and CFP::TUA1 (monochrome in first column, red in third and fourth columns) in cells at the indicated cell cycle stages. Brackets in A indicate a pair of adjacent PPBs. Arrowheads point to a metaphase spindle in E, and to phragmoplasts in I and M. Arrow in G points to the junction between a spindle-associated TAN::YFP ring and the adjacent PPB. 45o rotations in D, H, and L show that TAN::YFP forms a complete ring encircling the cell. Scale bar = 10 µm. Reproduced from Walker et al., 2007.
Analysis of a mutant of maize called tangled showed that TANGLED (TAN) protein is required for expanding phragmoplasts to find their way back to the former PPB site in dividing cells of the maize leaf. tan encodes a highly basic protein that is not closely related to other proteins of known function, but can bind to microtubules in vitro (Smith et al., 2001) Subsequent analysis of Arabidopsis TAN identified this protein as the first known component of the cortical division site to persist after PPB disassembly (Walker et al., 2007). Arabidopsis TAN::YFP is recruited in a microtubule- and kinesin-dependent manner to the cortical division site, where it co-localizes with PPBs and is subsequently maintained after PPB disassembly, marking the division site throughout mitosis and cytokinesis. As in maize, Arabidopsis TAN is required for guidance of phragmoplasts to former PPB sites during cytokinesis and thus is implicated as a functional component of the cortical division site. This exciting discovery leaves us with many unanswered questions, such as how TAN is maintained at the cortical division site after PPB disassembly and how it acts during cytokinesis to influence phragmoplast orientation. Our current work investigates these questions with studies in Arabidopsis, maize, and tobacco BY-2 cells.
Stomatal development in maize. All stages are illustrated schematically except the mature stage, which is illustrated with an image from a Toluidine Blue O-stained epidermal peel. See paragraphs below for details.
Asymmetric Cell Division
Asymmetric cell divisions are those that produce daughters of different sizes, shapes and/or developmental fates. In plants, as in other eukaryotes, asymmetric divisions are associated with pattern formation during embryogenesis, establishment of new cell lineages, and the formation of specialized cell types. In all of these processes, developmental asymmetry is closely tied to division polarity, which is oriented by either intrinsic or extrinsic cues. However, very little is known about such cues in plants or how cells perceive and respond to them. Stomatal development in maize provides an excellent model for studies of asymmetric cell division. Stomatal complexes in grasses such as maize consist of pair of guard cells, which delimit a pore that opens and closes to permit gas exchange for photosynthesis, flanked by a pair of subsidiary cells that participate in regulation of pore opening and closing. These stomatal complexes form through an invariant sequence of coordinated asymmetric cell divisions (see illustration above). Following an asymmetric division that forms a guard mother cell (GMC), the lateral neighbors of the GMC called subsidiary mother cells (SMCs) polarize towards the GMC in preparation for their asymmetric division. SMC polarization is marked by the formation of dense patches of cortical F-actin at the GMC contact site and nuclear migration to that site. After mitosis, asymmetric cytokinesis in SMCs forms the subsidiary cells, and then the GMC divides longitudinally to form a guard cell pair.
Analysis of stomatal complex formation in grasses led almost 50 years ago to the proposal that SMC divisions are polarized by a signal emanating from GMCs, but the mechanism of interaction between the GMC and SMC has remained unknown. We have isolated mutations in two maize genes, pan1 and pan2, that are required for proper execution of SMC divisions (see examples of aberrant subsidiaries indicated by arrowheads in pan1 mutant leaf epidermis, right). Analysis of pan phenotypes shows that aberrant SMC divisions can be traced back to defects in SMC polarization including failure of nuclear migration to the GMC contact site, and lack of actin patch formation or formation of delocalized patches. pan1 encodes a leucine-rich repeat receptor-like protein that localizes within SMCs to GMC contact sites prior to nuclear migration and actin patch formation, suggesting that it functions in the perception of GMC-derived polarizing cues (Cartwright et al., 2009). Ongoing efforts to elucidate the function of PAN1 and the pathway in which it functions include analysis of the role of Rho-type GTPases (ROPs) in PAN1-dependent SMC polarization, identification of the pan2 gene, and identification of candidate PAN1 ligands and putative downstream pathway components. These studies in maize are complemented by studies of the closest relatives of PAN1 in Arabidopsis.
Mutations in three maize discordia genes (dcd1, dcd2 and dcd3) disrupt SMC divisions in a different way. In dcd mutants, SMC polarity is established normally but PPBs form aberrantly, leading to abnormally oriented SMC divisions (see examples of aberrant subsidiaries indicated by black arrow, right). dcd1 encodes a putative PP2A phosphatase regulatory B’’ subunit homologous to Arabidopsis FASS/TON2, which is required for PPB formation in all cells, not just asymmetrically dividing ones (Wright et al., 2009). We identified a close relative of dcd1 in maize, alternative discordia1 (add1), which encodes a protein that is 96% identical to DCD1. Knocking down dcd1 and add1 function simultaneously via RNAi or in dcd1;add1 double mutants causes PPB formation to fail in both symmetrically and asymmetrically dividing cells. Consistent with these findings, DCD1/ADD1 proteins co-localize with PPBs in both symmetrically and asymmetrically dividing cells. Unexpectedly, the DCD1/ADD1 ring persists at the cortical division site after PPB disassembly, but unlike TAN does not appear to be maintained there through the completion of cytokinesis. Thus, DCD1 and ADD1 are not implicated in marking of the division site for recognition by the expanding phragmoplast, but may play a role in division site establishment or maintenance above and beyond their role in supporting PPB formation (Wright et al., 2009). Future work will be directed at identifying dcd2 and dcd3 and identifying putative downstream targets of the DCD1/ADD1-dependent PP2A phosphatase complex.
Spatial Regulation of Cell Expansion
Proper patterning of plant cell expansion is critical for the generation of organ shape, and also of diverse cell shapes, which are often important for cell function. Proper orientation of cell expansion requires microtubules, which influence the pattern of cellulose deposition into the cell wall, and actin filaments, which are thought to guide secretion and may have other roles as well (for review see Smith and Oppenheimer, 2005). Mechanisms governing aspects of cytoskeletal organization and dynamics that are relevant to plant cell growth are poorly understood, as are the mechanisms by which cytoskeletal filaments influence cell growth patterns.
We initiated work on these problems with the isolation and analysis of brick mutants of maize, which define three genes required for the formation of lobes on the margins of leaf blade epidermal cells (compare wild type, left, to brk1 mutant, right). SMC polarization defects are also sometimes observed in brk mutants, leading to formation of aberrant stomatal subsidiary cells (arrowheads, right). Lobe outgrowth in wild type cells is associated with localized accumulations of cortical F-actin that fail to form in brk mutants, and SMC polarization defects in these mutants are associated with loss of cortical F-actin patches at GMC contact sites (Frank and Smith, 2002). Consistent with this finding, brk1 encodes a subunit of a multiprotein complex first identified in mammalian cells that also contains SCAR/WAVE, an activator of the actin-nucleating Arp2/3 complex (Frank and Smith, 2002). Initial analysis of plant genomes led to the conclusion that plants lack ARP2/3 activators of the WAVE/SCAR/WASP family. However, we and others identified genes encoding four Arabidopsis proteins (SCAR1-4) distantly related at their N- and C-termini to SCAR/WAVE (Frank et al., 2004). The N-terminal domains of these proteins interact with BRK1 in vitro and their C-terminal domains can activate the bovine ARP2/3 complex in vitro, so they most likely interact with BRK1 in vivo and function as bona fide SCAR proteins to activate ARP2/3 complex-dependent actin nucleation. Supporting this conclusion, studies with Arabidopsis brk1 mutants showed that BRK1 functions in a pathway with the ARP2/3 complex to direct the morphogenesis of epidermal pavement cells and trichomes (Djakovic et al., 2006). Interestingly, Arabidopsis SCAR1 is depleted in brk1 mutants, indicating that BRK1 functions to protect SCAR from degradation in vivo. To address the question of where SCAR- and ARP2/3 complex-dependent actin nucleation occurs in plant cells and how it influences cell growth, we examined Arabidopsis BRK1 and SCAR1 localization. We found that both are plasma membrane-associated proteins localized at sites of cell growth and wall deposition, suggesting a role for SCAR-ARP2/3-dependent actin polymerization in membrane dynamics at the cell surface (Dyachok et al., 2008).