Sensing microtubule states through the mitogen-activated protein kinase pathway during mitosis and morphogenesis


Protein phosphorylation provides a broadly used reversible switch to regulate cellular activities in time and space, and localized phosphorylation–dephosphorylation can also create gradients of activities that provide spatial information for building, remodelling and translocating cellular structures. In plants, building of the cell wall is inside-out, starting from the centre and proceeding towards the periphery by expansion of a disk under the guidance of phragmoplast – short and densely arranged microtubules (MTs) interdigitated at their plus ends (Barr & Gruneberg, 2007). What determines the timing and location of phragmoplast formation, the precisely set length of MTs, the size of overlap in this structure, and the dynamic displacement of MTs from the middle towards the periphery (cell wall expansion) are central questions that researchers studying cytokinesis are keen to answer. The article by Beck et al. (2010b) in this issue of New Phytologist (pp. 1069–1083) and recently published works from the laboratory of Y. Machida (Kosetsu et al., 2010; Takahashi et al., 2010) add a long-sought-after Arabidopsis mitogen-activated protein kinase (MAPK) component to a signalling cascade that regulates cytokinesis in higher plants.

‘The role of MPK4 in cytokinesis is surprising…’

MPK4: one MAPK, many functions

Mitogen-activated protein kinase signalling components with a function in cytokinesis have been largely discovered in yeast complementation screens, composed of HINKEL, a kinesin motor protein, ANP1-3 MAPKK kinase, and ANQ/MKK6 MAPK kinase. However, the identity of the Arabidopsis MAPK in this pathway remained elusive (Sasabe & Machida, 2006). To identify the downstream Arabidopsis MAPK component in the cytokinesis pathway, the Machida group performed a biochemical screen to phosphorylate and activate MAPKs by MKK6/ANQ, mixing purified components using in vitro kinase assays, and surprisingly they identified MPK4, rather than MPK13, as a primary MKK6/ANQ downstream target (Takahashi et al., 2010). For a long time the most plausible candidate has been MPK13, based on its highest sequence similarity to the previously identified tobacco and Medicago MAPKs in the homologous cytokinesis pathway, its cell proliferation-dependent expression and coexpression with Arabidopsis homologues of the phragmoplast-localised protein kinase 1 (ANPs) and the Arabidopsis homologue of the next to NPK1, NQK1 (ANQ). Further supporting evidence that MPK13 is downstream of ANQ is their functional interaction in yeast, and the ability of ANQ to phosphorylate and activate MPK13 (Melikant et al., 2004). The Machida group, however, found by pairwise yeast two-hybrid interaction between MKK6/ANQ and a panel of MAPKs that only MPK4, MPK12 and MPK1 interact with MKK6/ANQ. Furthermore, in a pull-down assay, only MPK4 showed interaction with MKK6/ANQ. MKK6 also activates MPK4 when coexpressed in protoplasts (Kosetsu et al., 2010). Most importantly, MPK4 was shown to localize to the midzone of phragmoplast by indirect immunofluorescence (Beck et al., 2010b), and to be dynamically recruited to this location during cytokinesis using a complementing MPK4–green fluorescent protein (GFP) construct (Kosetsu et al., 2010). The Machida group looked for a role in cytokinesis in mpk4 based on defects in cross-wall deposition and presence of binucleate cells in mature tissues in the mpk4 mutant (Kosetsu et al., 2010). In this issue, Beck et al. (2010b) show that both mpk4 and anp2anp3 mutants exhibit cytokinesis defects in dividing cells of the root mersitem. To monitor the dynamics of mitosis and cytokinesis, they introduced a microtubule GFP marker (35S:GFP:MBD) into the Columbia wild-type, mpk4 and anp2anp3 backgrounds. Images taken from these lines indicated a prophase–metaphase arrest for anp2anp3, and a block in cytokinesis phragmoplast expansion for mpk4. Time-lapse imaging, however, indicated that both mpk4 and anp2anp3 mutants are delayed in metaphase and anaphase progression, and both show similar mitotic abnormalities. It is known that overexpression of the GFP fused to microtubule binding domain (GFP:MBD) microtubule marker can overstabilize MTs and lead to abnormalities. However, in an experiment to confirm that the differences seen among the lines are not the result of different levels of GFP:MBD expression, Beck et al. (2010b) also examined the effect of a MEK inhibitor, PD98059, on mitotic progression and showed a similar delay in mitosis after drug treatment to that which they had observed in the mutants.

The role of MPK4 in cytokinesis is surprising, as MPK4 was originally discovered as a negative regulator for pathogenesis response upstream of salicylic acid (SA), while it is positively required for jasmonate and ethylene (ET)-responsive genes. SA is known to be required for defence against bacterial microbes, while jasmonic acid (JA) and ET are required for defence against fungal pathogens, and thus MPK4 is at the crossroads balancing these resistance mechanisms through hormonal pathways (Brodersen et al., 2006). Molecular and genetic research has placed MPK4 in a stress pathway regulated by reactive oxygen species (ROS) and involving MEKK1–MKK1/MKK2–MPK4 (Ichimura et al., 2006). We also know some of the downstream substrate targets of this stress pathway, for example, MKS1, an adaptor protein that brings this pathway to the WRKY33 transcription factor by regulating its nuclear localization (Qiu et al., 2008). How, then, can MPK4 function under ANP and ANQ/MKK6 to regulate a completely different biological process, cytokinesis? The answer to this question could lie in the flexibility of how MAPK signalling components are used in different cellular contexts. One example is the recent demonstration of how the Pseudomonas effector protein, AvrB, can divert the MPK4 pathway towards activating ET, jasmonate-responsive genes rather than repressing SA genes by recruiting the MPK4 complex to the HSP90 chaperone, which appears to regulate the activity of the pathway, and to a scaffold protein, RIN4, that is likely to take the MPK4 signalling pathway to genes in JA, ET response, perturbing hormone signalling and thereby enhancing plant susceptibility to bacterial pathogens (Cui et al., 2010). Furthermore, MEKK1 has a structural scaffold role to recruit MKK1/2 and MPK4, and apparently the MEKK1 protein kinase activity is not required for its function (Suarez-Rodriguez et al., 2007). This was shown by the complementation of a mekk1 mutant with a kinase-inactive mutant form of MEKK1. Thus, both the recruitment of MEK-MAPK modules to upsteam activators and scaffolds, or to downstream substrate can determine the composition and biological function of MAPK pathways.

MAPK pathways regulate microtubule cross-bridges through the phosphorylation of MAP65

MAP65 has an activity to crosslink MTs with an exact 25 nm spacing (Walczak & Shaw, 2010), and PLEIADE/MAP65-3 have been shown to be a regulator of cytokinesis (Muller et al., 2004). Independently, two collaborative works – between the laboratories of S. Sonobe and Y. Machida, and those of P. J. Hussey and L. Bogre – have shown that one of the MAP65 family members, MAP65-1, is phosphorylated by multiple MAPKs and cyclin-dependent kinases (CDKs); they have also mapped the MAPK phosphorylation sites and shown that phosphorylation regulates the ability of MAP65-1 to crosslink MTs (Sasabe et al., 2006; Smertenko et al., 2006). MAP65-1 phosphorylation was shown to be compromised in mpk6 and in anp2anp3 and mpk4 mutants, leading to cortical MT overstabilization through crosslinking (Beck et al., 2010a; Muller et al., 2010). Thus, both ANP2 and 3 MPK6 and MPK4 also regulate cortical MT functions and, thereby, cell expansion. These pioneering works on a plant MAP have been followed by identification of the MAP65 orthologues in yeast and animal systems, called PRC1, and established the structural mechanisms of how PRC1 together with a kinesin motor are responsible for determining the region of MT plus end overlaps (Walczak & Shaw, 2010).

How does the NACK1-NPK1-NQK1-NRK-MAP65 switch work?

Rather than sensing extracellular signals, the MAPK signalling pathway is triggered by a microtubule motor protein, NACK1. NACK1 activates NPK1 by direct interaction between their noncatalytic C-terminal surfaces. Both NACK1 and NPK1 contain MAPK and CDK phosphorylation sites at, and in the vicinity of, their interaction surfaces, suggesting that this interaction is regulated by phosphorylation. NPK1 becomes dephosphorylated as cells exit metaphase, and in parallel CDK becomes inactivated (Nishihama et al., 2001). This might be required for NACK1 interaction and activation of the pathway. Phosphorylation at the putative MAPK sites might provide a positive feedback loop, or input from other MAPK pathways. NACK1 binding possibly removes the C-terminal inhibitory domain of NPK1, leading to the activation of NQK1 and subsequently NRK1, and to the phosphorylation of MAP65. The phosphorylation of MAP65 at the midline of phragmoplast could destabilize the crosslink between overlapping MTs, allowing dynamic expansion of the cell plate (Sasabe & Machida, 2006; Barr & Gruneberg, 2007; Walczak & Shaw, 2010). On cortical MTs the same pathway might act to sense MT states, such as microtubule tension, and through the activation of the MAPK pathway and phosphorylation of MAP65, it could allow the reorganization of MTs and dynamic shaping of cells and meristem domains. Interestingly, the auxin efflux carrier PIN-FORMED (PIN) proteins also have MAPK phosphorylation sites at their regulatory insertion loop, and have been found to be phosphorylated at this site (Benschop et al., 2007). Thus, speculatively, the same MT sensing pathway could be linked to the dynamic patterning of the direction of auxin flow (Hamant et al., 2008; Heisler et al., 2010).