by
|
by |
|
| This article will appear in a forthcoming issue of Trends in Plant Science. | |
Abstract
Plant cells construct four microtubule assemblies that direct key processes in plant cell morphogenesis (figure 1). There are two cortical arrays, the interphase cortical array, which is necessary for the formation of the cell wall and the direction of cell expansion and the preprophase band, which delineates the plane of cell division. The spindle separates daughter chromosomes and the phragmoplast array guides vesicles to the site where they join to form the new cell plate. The alternation of these arrays through the cycle must be coordinated and this is probably performed by accessory proteins. These accessory proteins must anchor or crosslink microtubules or otherwise regulate microtubule formation and destruction within the cell [1,2].
Classically, microtubule-associated proteins (MAPS) were defined as proteins that copurified with microtubules after several rounds of polymerization and depolymerization. This was based on studies using neuronal microtubules but applied to plant cells. These biochemical approaches have been troublesome because of the low yields of tubulin in plant microtubule preparations. However, in 1993, Chang-Jie Jiang and Seiji Sonobe purified the plant MAP-65 family of proteins by cycling taxol-stabilized tobacco microtubules in tobacco BY2 cell extracts through three rounds of polymerization-depolymerization. This family constitutes three or four electrophoretically separable proteins, all ∼65 kDa (reference 3). Subsequently, microtubule-affinity experiments, where neuronal microtubules were incubated with isolated carrot cytoskeletons, followed by the co-sedimentation of microtubules with binding proteins, revealed a similar family of MAP-65 proteins [4]. The identity of MAP-65 was determined and found to encode a novel protein [5]. At the same time, another structural plant MAP of 190 kDa (MAP-190) was identified, and this had the ability to bind microtubules and actin microfilaments. Again, MAP-190 was shown to be a novel protein [6]. These, or similar biochemical procedures, have identified several proteins that bind microtubules, most of which have alternative functions, for instance, EF1alpha (reference 7), eIF4F (reference 8) and heat shock protein 90 (reference 9). Therefore, the original criterion that a MAP copurifies with cycling microtubules has had to be moderated to include proteins that are capable of binding microtubules in vitro and in vivo [1].
With the completion of the Arabidopsis genome sequence, it has been possible to compare known animal and fungal MAPs on the available database. Surprisingly, few comparisons can be made, for example, only with kinesins, XMAP215 or katanin. Genetic approaches are now aiding in the identification and function in plant cells of known and possibly novel MAPs. The maize tangled mutant has defective phragmoplast alignment and the gene encodes an adenmatous polyposis coli-like protein, which is shown to bind microtubules [10]. The Arabidopsis FRA2 gene (also known as BOTERO1) [11] encodes the microtubule-severing protein katanin [12]. Several other Arabidopsis mutants are known and the corresponding genes might encode microtubule-associated proteins, such as the spiral mutants [13].
MOR1 is Similar to XMAP215-like Proteins
In a recent paper, Geoffrey Wasteneys and colleagues identified temperature-sensitive A. thaliana mutants defective in microtubule organization [14]. Two mutant alleles of the MICROTUBULE ORGANIZATION 1 (MOR1) gene were isolated. If germinated at the restrictive temperature, plants are severely stunted, show radially swollen and short organs and do not produce flowers. If shifted to the restrictive temperature during later development, flowering occurs but no seeds are produced. Moreover, closer examination revealed a left-handed twisting of organs, isotropic cell expansion and impaired root-hair polarity. The mor1 mutant expressing chimeric tubulin-green florescent protein (GFP) fusion protein was used to visualize microtubules in epidermal cells enroute from the permissive to the restrictive temperature. At the restrictive temperature, the microtubules became disorganized and shortened. The shortening of the microtubules appeared to be the result of dissociation of tubulin subunits rather than microtubule severing, suggested by the increased patchy diffuse tubulin-GFP fluorescence at the restrictive temperature. The MOR1 gene was identified and the deduced amino acid sequence shows significant identity to a class of microtubule-associated proteins conserved across phylogeny (figure 2), including Xenopus MAP215, XMAP215 (reference 15) and human TOGp (reference 16).
MOR1 has at least ten HEAT repeats within its sequence. The HEAT repeat is a tandemly repeated 37-47 amino acid-long unit occurring in several cytoplasmic proteins, including the four proteins that give the repeat its name: Huntingtin protein, Elongation factor 3, Alpha regulatory unit of protein phosphatase 2A, and the yeast PI3-kinase TOR1. Arrays of HEAT repeats usually consist of 3 to 36 units, which produce a rod-like helical structure and function as protein-protein interaction surfaces [17,18]. The two mor1 mutant alleles have single amino acid substitutions in the extreme N-terminal HEAT repeat: mor1-1 and mor1-2 have Leu174 to Phe174 and Glu195 to Lys195 substitutions, respectively. The essence of this discovery is that these mutations define a functional HEAT repeat that is conserved in MOR1's animal and yeast counterparts.
The question that arises is, how can these mutations have such a dramatic effect on microtubule organization? Some clues come from recent work on XMAP215. XMAP215 stabilizes microtubules and XKCM1, a Xenopus kinesin-related protein of the Kin1 class, is a microtubule destabilizing protein. These two proteins are capable of regulating microtubule dynamics in Xenopus egg extracts [19-22]. XMAP215 dramatically increases the growth rate of microtubules by seven- to tenfold, and if XMAP215 is depleted by 50-70%, this results in an increase in the catastrophe frequency and a reduction in microtubule length. Furthermore, the effects of this depletion can be rescued by inhibition of XKCM1. Therefore the steady-state length of microtubules appears to be determined by the relative activities of these two factors [15]. Experiments using different fragments of the XMAP215 protein containing the N-terminus (FrN), the middle (FrM) and the C-terminus (FrC) showed that the N-terminus was responsible for the catastrophe-suppressing activity of XMAP215 and that the C-terminus was responsible for microtubule binding [23]. Microtubule binding is believed to be through the Tau-MAP4-MAP2b-like microtubule binding domain, which is also present in the C-terminus of MOR1 (reference 23). In microtubule dynamics experiments, FrN induced microtubule polymerization whereas FrC caused an increase in microtubule catastrophy. Andrei Popov and colleagues proposed a model to explain these findings23. In this model, XMAP215 interacts with Kin1 kinesins (i,e. XKCM1) near growing microtubule ends. The C-terminus locates the XMAP215 protein to the microtubule whereas the N-terminus exerts the protein's catastrophe opposition and microtubule growth-promotion activities by interacting with Kin1. In such a model, FrC would increase catastrophes by displacing the endogenous XMAP215 and, thus, exposing the microtubule to the destabilizing activity of the catastrophic kinesins. FrN alone would promote microtubule polymerization by binding to Kin1 kinesins in a similar fashion to the full-length protein.
By contrast, Cynthia Spittle and colleagues [24], using co-pelleting assays, showed that the N-terminus of TOGp binds microtubules, whereas the C-terminus does so with a lesser affinity. However, the C-terminus was found to bind tubulin dimers. The microtubule binding ability is believed to be situated in a 600 amino acid stretch at the N-terminus that contains homology to the yeast homologue Stu2 at TOGp residues 493-596. Deletion of this region alone in TOGp had a minimal effect on microtubule binding but the deletion of flanking regions, of "approximately 200 amino acids" did significantly decrease binding to microtubules. Interestingly, the region of MOR1 that shares this homology to Stu2 does not contain the HEAT repeat or the sites of the mor1 mutations (although it is possible that these sites lie in the respective flanking regions).
Effects of the mor1 Mutations on Microtubule Dynamics
Bearing these data in mind, it is possible to suggest models to describe the effect of the mor1 mutations on microtubule dynamics (figure 3). At the permissive temperature a kin1-like protein (kin1-like proteins are present in the Arabidopsis genome database) can bind to MOR1 at the N-terminal HEAT repeat. At the restrictive temperature a conformational change in the protein caused by the mor1 mutations releases the kinesin, which is now capable of depolymerizing the microtubules. Alternatively, MOR1 decorates the microtubules at the permissive temperature, stabilizing the long microtubules in the cortex. A shift to the restrictive temperature causes a loss of the N-terminus castastrophe-suppressing activity (as in XMAP215) or the microtubule-binding capability (as in TOGp) of MOR1, enabling a kin1-like kinesin to destabilize the long cortical microtubules.
If one of these models is correct, why do we not see the effect of the mor1 mutations on mitotic and cytokinetic arrays? One possibility is that kin1-like activity predominates in the spindle and the cytokinetic phragmoplast to generate the short microtubules present in these arrays. This suggestion is based on ideas relating to XMAP215 activity. XMAP215 is probably regulated by phosphorylation, which could reduce its ability to bind microtubules because phosphorylation of the CDK1 site inhibits the ability of XMAP215 to promote polymerization of pure tubulin [20]. Such findings lead to a model where CDK1 is reduced during interphase and XMAP215 is unphosphorylated, and its activity predominates over that of XKCM1 leading to long microtubules. In mitosis, CDK1 activity is high and XMAP215 is hyperphosphorylated, its activity is reduced and the activity of XKCM1 predominates, leading to short microtubules [25].
It is possible that the ability of MOR1 to bind microtubules is controlled by its phosphorylation, and there are several candidate sites for such a role including two potential CDK1 sites (starting at position 1450 and 1841 are the sequences xRT*Px). This is not a direct match to the CDK1 consensus phosphorylation site (xKT*Px or xKS*Px) [26], but it remains a strong possibility because a similar basic residue precedes the two major CDK1 site determinant residues (T or S and P). Such a control on MOR1 might surpass any catastrophe-suppressing activity of the N terminus and more specifically, the N-terminal HEAT repeat. Under these circumstances the mor-1 mutations might have little to no effect on the short microtubules present in mitosis. However, what is intriguing is the fact that at the restrictive temperature, the cortical microtubules in the mor-1 mutant were short but never completely disassembled, which would imply a minimum-length control by the microtubule destabilizing factor.
Other possible reasons why the mor1 mutations have no observable effect in mitosis are that MOR1 is not present in the spindle or phragmoplast or that different catastrophic kinesins are present in the cortical and spindle or phragmoplast arrays.
Conclusion and Perspective
The study of the mor1 mutant has led to the identification of a new plant microtubule-associated protein. Moreover, the locations of the mutations in MOR1 and the resulting mor1 phenotypes have helped to clarify the proteins possible function in the plant cell. With consideration to the models presented here, this function could be similar and as important as those of its homologues, XMAP215 and TOGp.
It remains to be determined whether all the components of these models are active in plant cells, and whether the different components colocalize and interact in vivo.
