Plant Cytokinesis. Exploring the Links

Cytokinesis is the partitioning of the cytoplasm following nuclear division. This process presents a number of challenges for the plant cell: first, to avoid losing or bisecting the nucleus, this event needs to be carefully coordinated with respect to the nuclear cycle in space and in time. Second, a structure as complex as the plant cell wall needs to be laid down during the brief period of time between anaphase and telophase. The spatial and temporal regulation of cytokinesis requires a series of links between the nuclear cycle, the cell cortex, the Golgi apparatus, and the membrane trafficking apparatus. A number of genes have recently been identified that could enable us to probe such links. Cytokinesis in higher plants may be considered as a specialized form of secretion. At the end of anaphase, Golgi-derived secretory vesicles carrying cell wall materials are transported to the equator of a dividing cell. Fusion of these vesicles gives rise to a membrane-bound compartment, the cell plate. The cell plate expands from the middle out (centrifugally) until it reaches the “zone of attachment” or division site on the mother cell wall. Once this attachment has taken place, the cell plate undergoes a complex process of maturation during which callose is replaced by cellulose and pectin (Samuels et al., 1995, and references therein). Two cytoskeletal arrays, the preprophase band (PPB) and the phragmoplast, play central roles in cytokinesis in the somatic cells of higher plants. The PPB, a transient ring of cortical microtubules (MTs) and actin filaments, appears in late S phase, narrows throughout G2, and disappears during prophase when the nuclear envelope (NE) breaks down. The phragmoplast is an array of MTs and actin filaments present at the equator of a dividing cell during the anaphase to telophase transition (for review, see Assaad et al., 1997). This review focuses on a small number of genes identified by mutation and whose molecular identities have recently been determined. These include TITAN5, TAN1, KEULE,KORRIGAN, and CYT1, required for cytokinesis in the somatic cells of higher plants (McElver et al., 2000; Zuo et al., 2000; Assaad et al., 2001; Lukowitz et al., 2001; Smith et al., 2001). In addition, reverse genetics has implicated the mitogen-activated protein (MAP) kinase kinase kinase NPK1 in cytokinesis (Nishihama et al., 2001). A number of important insights into plant cytokinesis have been gained from localization, drug, inhibitor, or antibody injection studies, reviewed elsewhere (Sylvester, 2000). Variant forms of plant cytokinesis specific to certain cell types are the subject of a recent review (Otegui and Staehelin, 2000). Plant cytokinesis: exploring the links. A, The division site is established during preprophase. The nuclear surface organizes the PPB (1), which may establish spatial cues (orange ovals; 2). B, Cell wall polymers such as xyloglucans are synthesized at the Golgi, packaged into vesicles, and transported to the equator of a dividing cell. C, The cell plate arises via fusion of Golgi-derived vesicles at the equator. The plate expands through the addition of secretory vesicles at the periphery, the deposition of callose in the lumena of the vesicular network, and via the concentric displacement of MTs toward the cell cortex (compare MTs in B, C, and D). Callose synthesis occurs within the cell plate, and this may be triggered by high concentrations of membrane-associated calcium. D, The phragmoplast is guided toward the spatial cue established during preprophase. MTs are now reoriented, with their minus ends at the equator and their plus ends toward the cell cortex. E, The cell plate fuses with the parental membranes and cell wall at the division site. Cellulose biosynthesis occurs at the plasma membrane; pectin biosynthesis takes place within the Golgi. F, The biosynthesis and/or packaging of cell wall polymers, membrane traffic, a MAP kinase cascade, and cytoskeletal organization are presumably coordinated with respect to cell cycle progression. How this relates to anaphase-promoting complex (APC) activation is unclear. Also, the targets of the MAP kinase remain to be determined. PPB and phragmoplast MTs are shown in red. Black arrows highlight established links and purple arrows potential links. Genes required for plant cytokinesis are shown in blue. The NE breaks down during mitosis and reassembles during telophase. The immature cell plate, rich in callose and also containing xyloglucans, is colored green, and the mature walls containing cellulose and pectin are colored in brown. Modeled according to Samuels et al. (1995) and Moore and Staehelin (1988); also see text. Elegant experiments involving the displacement of nuclei or immature cell plates by centrifugation in moss protonemata or stamen hair cells have clearly demonstrated that the position of the interphase nucleus determines the division plane (Mineyuki and Gunning, 1990; Murata and Wada, 1991). Within a narrow window of time during the cell cycle, the nucleus is competent to dictate the position of the PPB that, in turn, marks the future division site (Murata and Wada, 1991). In stamen hair cells, immature cell plates displaced by centrifugation curve toward the site formerly occupied by the PPB (Mineyuki and Gunning, 1990). Thus, a spatial cue is laid down early during the cell cycle (Fig. 1A) and, later during the cell cycle, the nascent cell plate is guided toward this cue (Fig. 1D). A relay of links between the nucleus, the PPB, and the cell cortex/plasma membrane/cell wall may be implicated. First, the surfaces of plant nuclei have been shown to possess all the properties of MT organizing centers (MTOCs). These properties include: (a) the ability to nucleate and organize MTs, (b) the capacity to establish and/or anchor the minus ends of MTs, and (c) the presence at the nuclear surface of gamma tubulin and other proteins characteristic of animal MTOCs (Stoppin et al., 1994). Thus, it is likely that perinuclear MTOCs, which are associated with the NE, organize the PPB during late S-phase. Cell wall stubs and mutlinucleate cells in cytokinesis-defective embryos. A and B, Light micrographs of dermatogen stage embryos. A, keule mutant. B, Wild type. C, Electron micrograph of keule mutant embryo at dermatogen stage showing a non-vacuolated binuleate cell. The centripetal cell wall stub supports the model that the division site consolidates the immature cell plate (see text). N, Nucleus; P, protoderm. A star marks the uppermost cell of the suspensor. Arrows point to cell wall stubs and arrowheads to multinucleate cells. Reproduced, with permission, from Assaad et al. (1996; Figs. 4E, 4F, and 6E, © Springer-Verlag, Berlin). Bar = 1 μm in C. The above model is strengthened by the analysis of plant mutants impaired in their ability to orient cell walls. The Arabidopsisfas and tonneau mutants, as well as thetangled 1 mutants of maize (Zea mays), are characterized by misoriented cell walls, especially during asymmetric or longitudinal divisions; whereas fas andtonneau mutants altogether lack PPBs, these rings of cortical MTs are often misoriented in tangled mutants (for review, see Nacry et al., 2000). In tonneau mutants, cortical MTs in general, including the PPB, are perturbed, yet the spindle and phragmoplast appear normal (Traas et al., 1995). It is possible that the FAS/TONNEAU gene products function at the nuclear surface to organize cortical MT arrays (Fig.1A). The TANGLED1 gene has been cloned and characterized (Smith et al., 2001). The gene encodes a highly basic protein bearing little sequence similarity to other proteins, yet possessing domains weakly homologous to the MT binding domain of vertebrate APC. TAN1 binds to MTs in vitro, possibly in a cell cycle-dependent manner, and proteins recognized by anti-TAN1 antibodies localize to the PPB, spindle, and phragmoplast in dividing cells, providing evidence thatTANGLED1 may encode an MT-binding protein. It is interesting that in tangled mutants the leading edges of phragmoplasts are not guided to sites formerly occupied by PPBs. Mutant cell plates consistently do not undergo the flattening that accompanies cell plate maturation, but remain wrinkled (Smith et al., 2001, and references therein). In addition to orienting the PPB, TANGLED may be implicated in the establishment of the division site during preprophase and/or may guide the leading edges of the phragmoplast to this site during cytokinesis (Fig. 1, A and D). The molecular identity of the spatial cue that determines the division site, as well as its localization to the cortex, plasma membrane, and/or cell wall, remain to be determined. The onset of cytokinesis is concomitant with exit from mitosis. In plants, progression through mitosis relies on the activity of the cyclinB-cdc2 complex (M-CDK) active during M phase. Shortly before anaphase, M-CDKs are thought to activate the APC, a ubiquitin ligase that in turn destroys the M-CDKs. Thus, cyclin-dependent kinases and the APC regulate each other to ensure timely progression through mitosis (Meijer and Murray, 2001; Nigg, 2001). A number of studies in budding and fission yeast (Saccharomyces cerevisiae and Saccharomyces pombe) support the notion that mitotic exit does not alone suffice for the initiation of cytokinesis. In fact, an additional kinase cascade triggers a cytokinetic pathway (Nigg, 2001). In budding yeast, the polo kinase CDC5 participates in APC activation and, in addition, appears to regulate the cytokinetic pathway by interacting with septins (Song and Lee, 2001). Because plants lack septins, it is not clear how these findings in yeast relate to plant cytokinesis. A large number of kinases have been found at the phragmoplast (for list, see Nacry et al., 2000) and these are good candidates for orchestrating the onset and execution of cytokinesis. Compelling evidence that a MAP kinase cascade is required for plant cytokinesis comes from a recent study showing that kinase negative mutations inNPK1, a MAP kinase kinase kinase, disrupt cytokinesis in tobacco (Nicotiana tabacum) cells (Nishihama et al., 2001). NPK1 activity is up-regulated during late M phase. The protein is present in the nucleus during interphase, and at the equatorial zone of the phragmoplast where it may be required for phragmoplast expansion toward the cell cortex (Nishihama et al., 2001). Thus, NPK1 may provide continuity in space and in time between the interphase nucleus and the cell equator. A MAP kinase has been detected at the phragmoplast in alfalfa (Medicago sativa) cells (Bögre et al., 1999) and its tobacco orthologue might be a target of NPK1. Two important questions remain unanswered: First, what signals activate NPK1? And, second, what are the targets of the putative MAP kinase cascade downstream of NPK1? The multinucleate phenotype of cytokinesis-defective mutants suggest that nuclear division can be initiated and completed even if cytokinesis is incomplete. In contrast, cell wall stubs in such mutants are only observed in multinucleate cells, which suggests that cytokinesis can only be initiated once the nuclear cycle is complete.titan and pilz mutants, characterized by giant nuclei, consistently show marked cytokinesis defects (Liu and Meinke, 1998; for review, see Nacry et al., 2000). A simple hypothesis is that the cytokinesis defects observed in titan andpilz mutants are an indirect consequence of cell cycle arrest due to a primary defect in nuclear division. This is supported by the observation, based on tubulin stains, that not only cytokinesis but cell cycle progression in general is affected in pilzmutants (for review, see Nacry et al., 2000). A number of conserved cell cycle checkpoints known to monitor nuclear division and spindle assembly or orientation readily account for these observations (Nigg, 2001). In this context, it is interesting to note that cytokinesis defective mutants such as keule have enlarged nuclei (Assaad et al., 1996), suggesting that incomplete cytokinesis may also impact the nuclear cycle, though not in an absolute way. In general, however, the nucleus appears to play a dominant role in dictating the onset of cytokinesis. Drug studies have highlighted the importance of the Golgi apparatus in cytokinesis. The biosynthesis and assembly of numerous cell wall polysaccharides take place in the Golgi (Moore and Staehelin, 1988). In contrast, callose is synthesized within the cell plate and cellulose microfibrils are synthesized by cellulose synthases embedded in the plasma membrane, which explains why cell plate flattening and maturation only occur after fusion with the parental membrane and wall (Samuels et al., 1995, and references therein). Immature cell plates are rich in xyloglucans and devoid of pectins, whereas mature cross walls are rich in pectins and have low xyloglucan content (Moore and Staehelin, 1988; His et al., 2001). Because pectins and xyloglucans are synthesized in the Golgi and targeted to the cell plate, there appears to be a tight cell cycle regulation of Golgi activity and secretion (Moore and Staehelin, 1988; see Fig. 1). A, Vesicle trafficking can be broken down into four steps: formation, transport, tethering/docking, and fusion. Some of the key players required at each step are indicated in italics, and lines denote their range of action. ADP-ribosylation factor (ARF) function is best documented for vesicle formation, and Rabs regulate vesicle docking. Through a variety of effectors and through potential cross talk, both ARF and Rab GTPases in principle could regulate all four steps, as is thought to be the case for Rabs (Zerial and McBride, 2001). B, Vesicle docking is regulated by Rabs and Sec1s through the activity of rab effectors. These are large complexes or long molecules capable of multiple interactions with the Rab on the vesicle and the Sec1 and syntaxin on target membranes. These interactions may be regulated by phosphatidylinositol (PIP) signaling (Zerial and McBride, 2001). C, Vesicle fusion requires an interaction between vesicle-associated membrane proteins (VAMPs) on vesicles and syntaxins on target membranes; based on studies of neural cells and synapses, it is thought that syntaxins need to be primed or opened for this interaction to take place. Sec1 proteins bind the closed syntaxin (1) and induce a conformational change (2). The Sec1 is then actively removed (3) with the help of CDK5 and tomosyn. 4, With the help of SNAP25-like adapters, a core complex is subsequently formed between the VAMP on the vesicle and the syntaxin on the target membrane. This pulls the membranes close together, overcoming the barriers to fusion. For the sake of simplicity, we present models based on neural systems (Chen and Scheller, 2001) for which the vesicle and target membranes are distinct. As regards the initial rounds of fusion at the cell plate, however, these membranes are in all likelihood identical. An interesting recent finding is that one of the TITANgenes, TITAN5, encodes an ARF GTPase that may be implicated in vesicle formation during cell division (McElver et al., 2000, and references therein). This ARF may, for example, be required for NE breakdown into vesicles, and this could explain the giant nuclei observed in titan5 mutant embryos and endosperm. In addition to a role in regulating nuclear division, TITAN5 could also be required for vesicle formation during cytokinesis (McElver et al., 2000). With 45 members, ARFs represent a large gene family in Arabidopsis (www.Arabidopsis.org/Blast), and therefore it is likely that individual ARFs have specialized as opposed to multiple functions. Because plants lack orthologues of nuclear lamins (www.Arabidopsis.org/Blast), the identification of TITAN 5 provides a highly important and novel handle on the poorly understood process of NE dynamics and function. At anaphase, vesicles are transported to the equator of a dividing cell. Because the plus ends of phragmoplast MTs overlap at the equator, a plus end-directed motor such as kinesin would transport Golgi-derived vesicles to the equatorial region during cytokinesis. In addition, a minus end-directed motor such as may play a role in vesicle to the kinesin and have been found at the phragmoplast (Sylvester, 2000). to vesicles are found at their target membranes. is by a number of protein interactions that Rab GTPases on vesicles with syntaxins on target membranes (Fig. and McBride, 2001). Arabidopsis a large family of Rab proteins, has been implicated in cytokinesis. Vesicle fusion requires a specific and interaction between syntaxins on target membranes and or on vesicle membranes. of the Sec1 of proteins appear capable of conformational in or these for interactions with other (Fig. for review, see and Scheller, 2001). have identified genes whose and membrane fusion during cytokinesis. At a and keule mutants are characterized by multinucleate cells with or incomplete cross walls (see Fig. et al., Lukowitz et al., encodes a syntaxin in during M phase, and at the phragmoplast in dividing cells et al., et al., 1997). encodes a Sec1 protein that has been shown to bind in in (Assaad et al., 2001). The that vesicles but do not at the equator of dividing cells in keule embryos et al., 2000) that, the syntaxin is required for vesicle fusion. The of keule mutants has evidence that the genes in cytokinesis is impaired but not in keule mutants, which the it is in keule mutants that as multinucleate embryos et al., 2000). is the function of the Sec1 proteins are large and capable of multiple and may multiple In yeast and animal cells, Sec1s the membrane fusion on target membranes with the Rab cycle on vesicle membranes (for review, see and McBride, 2001). In a manner, Sec1 proteins a key between and the and of neural et al., an is that may cell cycle signals and to the cytokinetic vesicle fusion by of an interaction with the syntaxin this a MAP kinase has been found to regulate the activity of a complex in animal cells, regulating et al., 2001). by CDK5 the interaction in neural cells et al., Thus, the vesicle trafficking for be a target of the numerous kinases (see Nacry et al., including MAP present at the The initial rounds of fusion at the cell equator likely can be considered as in that occur between that is to between Golgi-derived The membrane by vesicle fusion undergoes a series of including the of a membrane and these could be by a change in the identity of the cell plate are as membrane from the cell plate. In contrast, fusion of the cell plate with the parental wall requires a fusion between cell plate membranes and the plasma membrane. this the membranes that the cell plate are with the plasma membrane, yet of Golgi-derived In fact, during surface expansion and cell are also to the plasma membrane. As of when does the novel membrane formed at the cell equator a plasma membrane and what is its analysis the syntaxin in a novel of as close to plasma membrane syntaxins as it is to and in the family as which has been to the plasma membrane; are animal Sec1s required for et al., et al., protein of the are targeted to both the plasma membrane and cell plate et al., 2000). that the cell plate membranes in from the plasma membrane is by the that the plasma membrane is from the cell plate et al., 1997). These observations support the that the cell plate is a novel and membrane from a form of Cell plate maturation requires callose as well as cellulose and pectin Two genes that cell wall and appear to be required for cell plate and with the cell wall stubs observed as of the and mutants, the cell wall stubs observed in and mutants appear later in and have only been documented in cells (Assaad et al., Lukowitz et al., and Meinke, 1998; Zuo et al., 2000). This the as to these cell wall stubs during cell expansion cell division. It clear that the of cell plates is in both mutant encodes a required for The in cellulose content observed in readily for its cell wall which include a high callose content and a of pectins and Meinke, 1998; Lukowitz et al., 2001). of these findings are that appears to be required for cellulose and that other polysaccharides such as callose and pectins may for cellulose cellulose and pectin content have also been observed in mutants et al., 2001). As an is likely to the assembly or of microfibrils et al., and is to be implicated in pectin Thus, the point to the pectin of the cell wall. The above a of molecular which may enable us to probe the links between the nuclear cycle, the cell cortex, membrane traffic, and cell wall As an protein required for the spatial regulation of TAN1 may provide a between the nucleus and/or phragmoplast and the cell cortex. The MAP kinase cascade by NPK1 may regulate cytoskeletal and/or membrane fusion during cytokinesis. The complex may vesicle fusion with the nuclear mutants on the complex process of cell plate the of plant however, it is clear that only of the players have been for are These play important roles as throughout the cell cycle in yeast and animal cells, and are key of membrane traffic, yet the only GTPase identified by mutation in plant cell division is the ARF by need to be in before an of links can be the of cytokinesis genes identified by mutation (for review, see Nacry et al., a of and reverse genetics likely be to and for and/or on the to the numerous whose could not be due to space