Two kinesins from Arabidopsis, KatB and KatC, have a second microtubule-binding site in the tail domain

Kinesins, as a kind of microtubule-based motor proteins, have a conserved microtubule-binding site in their motor domain. Here we report that two homologous kinesins in Arabidopsis thaliana, KatB and KatC, contain a second microtubule-binding site in their tail domains. The prokaryotic-expressed N-terminal tail domain of the KatC heavy chain can bind to microtubules in an ATP-insensitive manner. To identify the precise region responsible for the binding, a serious of truncated KatC cDNAs encoding KatC N-terminal regions in different lengths, KatC1-128, KatC1-86, KatC1-73 and KatC1-63, fused to Histidine-tags, were expressed in E. coli and affinity-purified. Microtubule cosedimentation assays show that the site at amino acid residues 74-86 in KatC is important for microtubulebinding. By similarity, we obtained three different lengths of KatB N-terminal regions, KatB1-384, KatB1-77, and KatB1-63, and analyzed their microtubule-binding ability. Cosedimentation assays indicate that the KatB tail domain can also bind to microtubules at the same site as and in a similar manner to KatC. Fluorescence microscopic observations show that the microtubule-binding site at the tail domain of KatB or KatC can induce microtubules bundling only when the stalk domain is present. Through pull-down assays, we show that KatB1-385 and KatC1-394 are able to interact specifically with themselves and with each other in vitro. These findings are significant for identifying a previously uncharacterized microtubule-binding site in the two kinesin proteins, KatB and KatC, and the functional relations between them.     

Characterization of microtubule binding domains in the Arabidopsis kinesin-like calmodulin binding protein.

The kinesin-like calmodulin binding protein (KCBP) is a new member of the kinesin superfamily that appears to be present only in plants. The KCBP is unique in its ability to interact with calmodulin in a Ca2+-dependent manner. To study the interaction of the KCBP with microtubules, we expressed different regions of the Arabidopsis KCBP and used the purified proteins in cosedimentation assays with microtubules. The motor domain with or without the calmodulin binding domain bound to microtubules. The binding of the motor domain containing the calmodulin binding region to microtubules was inhibited by Ca2+-calmodulin. This Ca2+-calmodulin regulation of motor domain interactions with microtubules was abolished in the presence of antibodies specific to the calmodulin binding region. In addition, the binding of the motor domain lacking the calmodulin binding region to microtubules was not inhibited in the presence of Ca2+-calmodulin, suggesting an essential role for the calmodulin binding region in Ca2+-calmodulin modulation. Results of the cosedimentation assays with the N-terminal tail suggest the presence of a second microtubule binding site on the KCBP. However, the interaction of the N-terminal tail region of the KCBP with microtubules was insensitive to ATP. These data on the interaction of the KCBP with microtubules provide new insights into the functioning of the KCBP in plants.     

Microtubule-associated Protein-like Binding of the Kinesin-1 Tail to Microtubules

The kinesin-1 molecular motor contains an ATP-dependent microtubule-binding site in its N-terminal head domain and an ATP-independent microtubule-binding site in its C-terminal tail domain. Here we demonstrate that a kinesin-1 tail fragment associates with microtubules with submicromolar affinity. Binding is largely electrostatic in nature, and is facilitated by a region of basic amino acids in the tail and the acidic E-hook at the C terminus of tubulin. The tail binds to a site on tubulin that is independent of the head domain-binding site but overlaps with the binding site of the microtubule-associated protein Tau. Surprisingly, the kinesin tail domain stimulates microtubule assembly and stability in a manner similar to Tau. The biological function of this strong kinesin tail-microtubule interaction remains to be seen, but it is likely to play an important role in kinesin regulation due to the close proximity of the microtubule-binding region to the conserved regulatory and cargo-binding domains of the tail.     

p180 is involved in the interaction between the endoplasmic reticulum and microtubules through a novel microtubule-binding and bundling domain

p180 was originally reported as a ribosome-binding protein on the rough endoplasmic reticulum membrane, although its precise role in animal cells has not yet been elucidated. Here, we characterized a new function of human p180 as a microtubule-binding and -modulating protein. Overexpression of p180 in mammalian cells induced an elongated morphology and enhanced acetylated microtubules. Consistently, electron microscopic analysis clearly revealed microtubule bundles in p180-overexpressing cells. Targeted depletion of endogenous p180 by small interfering RNAs led to aberrant patterns of microtubules and endoplasmic reticulum in mammalian cells, suggesting a specific interaction between p180 and microtubules. In vitro sedimentation assays using recombinant polypeptides revealed that p180 bound to microtubules directly and possessed a novel microtubule-binding domain (designated MTB-1). MTB-1 consists of a predicted coiled-coil region and repeat domain, and strongly promoted bundle formation both in vitro and in vivo when expressed alone. Overexpression of p180 induced acetylated microtubules in cultured cells in an MTB-1-dependent manner. Thus, our data suggest that p180 mediates interactions between the endoplasmic reticulum and microtubules mainly through the novel microtubule-binding and -bundling domain MTB-1.     

Regulation of dynactin through the differential expression of p150Glued isoforms

Cytoplasmic dynein and dynactin interact to drive microtubule-based transport in the cell. The p150Glued subunit of dynactin binds to dynein, and directly to microtubules. We have identified alternatively spliced isoforms of p150Glued that are expressed in a tissue-specific manner and which differ significantly in their affinity for microtubules. Live cell assays indicate that these alternatively spliced isoforms also differ significantly in their microtubule plus end-tracking activity, suggesting a mechanism by which the cell may regulate the dynamic localization of dynactin. To test the function of the microtubule-binding domain of p150Glued, we used RNAi to deplete the endogenous polypeptide from HeLa cells, followed by rescue with constructs encoding either the full-length polypeptide or an isoform lacking the microtubule-binding domain. Both constructs fully rescued defects in Golgi morphology induced by depletion of p150Glued, indicating that an independent microtubule-binding site in dynactin may not be required for dynactin-mediated trafficking in some mammalian cell types. In neurons, however, a mutation within the microtubule-binding domain of p150Glued results in motor neuron disease; here we investigate the effects of four other mutations in highly conserved domains of the polypeptide (M571T, R785W, R1101K, and T1249I) associated in genetic studies with Amyotrophic Lateral Sclerosis. Both biochemical and cellular assays reveal that these amino acid substitutions do not result in functional differences, suggesting that these sequence changes are either allelic variants or contributory risk factors rather than causative for motor neuron disease. Together, these studies provide further insight into the regulation of dynein-dynactin function in the cell.     

Association between elongation factor-1alpha and microtubules in vivo is domain dependent and conditional

Although the precise definition for a microtubule-associated protein (MAP) has been the subject of debate, elongation factor-1alpha (EF-1alpha) fits the most basic criteria for a MAP [Durso and Cyr, 1994a]. It binds, bundles, stabilizes, and promotes the assembly of microtubules in vitro, and localizes to plant microtubule arrays in situ. In this study, the in vitro and in vivo association of EF-1alpha with microtubules was further investigated. Analysis of the in vitro binding data for EF-1alpha and microtubules indicates that EF-1alpha binds cooperatively to the microtubule lattice. In order to investigate the interaction of EF-1alpha with microtubules in vivo, GFP fusions to EF-1alpha or to EF-1alpha truncates were transiently expressed in living plant cells. Using this method, two putative microtubule-binding domains on EF-1alpha were identified: one in the N-terminal domain I and one in the C-terminal domain III. The binding of domain I to microtubules in vivo, like the binding of full-length EF-1alpha, is conditional, and requires incubation in weak, lipophilic organic acids. The binding of domain III to microtubules in vivo, however, is not conditional, and occurs under normal cellular regimes. Furthermore, domain III stabilizes cortical microtubules as determined by their resistance to the anti-microtubule herbicide, oryzalin. Because the accumulation of EF-1alpha onto microtubules is unconditional in the absence of domain I, we hypothesize that domain I negatively regulates the accumulation of EF-1alpha onto microtubules in vivo. This hypothesis is discussed in terms of possible regulatory mechanisms that could affect the accumulation of EF-1alpha onto microtubules within living cells.     

The chromokinesin Kid is required for maintenance of proper metaphase spindle size

The human chromokinesin Kid/kinesin-10, a plus end-directed microtubule (MT)-based motor with both microtubule- and DNA-binding domains, is required for proper chromosome alignment at the metaphase plate. Here, we performed RNA interference experiments to deplete endogenous Kid from HeLa cells and confirmed defects in metaphase chromosome arm alignment in Kid-depleted cells. In addition, we noted a shortening of the spindle length, resulting in a pole-to-pole distance only 80% of wild type. The spindle microtubule-bundles with which Kid normally colocalize became less robust. Rescue of the two Kid deficiency phenotypes-imprecise chromosome alignment at metaphase and shortened spindles- exhibited distinct requirements. Mutants lacking either the DNA-binding domain or the MT motor ATPase failed to rescue the former defect, whereas rescue of the shortened spindle phenotype required neither activity. Kid also exhibits microtubule bundling activity in vitro, and rescue of the shortened spindle phenotype and the bundling activity displayed similar domain requirements, except that rescue required a coiled-coil domain not needed for bundling. These results suggest that distinct from its role in chromosome movement, Kid contributes to spindle morphogenesis by mediating spindle microtubules stabilization.     

Microtubule-associated protein 4 binds to actin filaments and modulates their properties

We previously reported that an isoform of microtubule-associated protein 4 (MAP4) is localized to the distal area of developing neurites, where microtubules are relatively scarce, raising the possibility that MAP4 interacts with another major cytoskeletal component, actin filaments. In the present study, we examined the in vitro interaction between MAP4 and actin filaments, using bacterially expressed MAP4 and its truncated fragments. Sedimentation assays revealed that MAP4 and its microtubule-binding domain fragments bind to actin filaments under physiological conditions. The apparent dissociation constant and the binding stoichiometry of the fragments to actin were about 0.1 μm and 1 : 3 (MAP4/actin), respectively. Molecular dissection studies revealed that the actin-binding site on MAP4 is situated at the C-terminal part of the proline-rich region, where the microtubule-binding site is also located. Electron microscopy revealed that the MAP4-bound actin filaments become straighter and longer and that the number of actin bundles increases with greater concentrations of added MAP4 fragment, indicating that MAP4 binding alters the properties of the actin filaments. A multiple sequence alignment of the proline-rich regions of MAP4 and tau revealed two putative actin-binding consensus sequences.     

Motor protein independent binding of endocytic carrier vesicles to microtubules in vitro

We have established an in vitro assay to characterize the binding of endocytic carrier vesicles to microtubules. Magnetic beads coated with microtubules were used as an affinity matrix. A fraction from nocodazole-treated cells enriched in endocytic carrier vesicles, labeled with internalized horseradish peroxidase, was used in the binding experiments. Binding of the endocytic carrier vesicles to microtubules in vitro was cytosol-dependent. This activity of cytosolic factors was saturable, heat-sensitive, and insensitive to N-ethyl-maleimide. Binding was sensitive to GTP and ATP. Addition of neuronal microtubule-associated proteins completely abolished binding of the endocytic organelles to microtubules. This binding was independent of the cytosolic microtubule-based motor proteins kinesin and cytoplasmic dynein, since cytosol depleted of these proteins remained fully active. Microtubule-binding proteins from HeLa cells, however, stimulated the interaction of endocytic carrier vesicles with microtubules. Trypsinized vesicles could no longer bind to microtubules in the presence of cytosol. These results suggest that cytosolic microtubule-binding proteins, other than the known microtubule-based motor proteins, as well as membrane proteins are involved in the nucleotide-dependent interaction of endocytic carrier vesicles with microtubules.     

Mapping the microtubule binding regions of calponin

The smooth muscle basic calponin interacts with F-actin and inhibits the actomyosin ATPase in a calmodulin or phosphorylation modulated manner. It also binds in vitro to microtubules and its acidic isoform, present in nonmuscle cells, and co-localizes with microfilaments and microtubules in cultured neurons. To assess the physiological significance and the molecular basis of the calponin-microtubule interaction, we have first studied the solution binding of recombinant acidic calponin to microtubules using quantitative cosedimentation analyses. We have also characterized, for the first time, the ability of both calponin isoforms to induce the inhibition of the microtubule-stimulated ATPase activity of the cytoskeletal, kinesin-related nonclaret dysjunctional motor protein (ncd) and the abolition of this effect by calcium calmodulin. This property makes calponin a potent inhibitor of all filament-activated motor ATPases and, therefore, a potential regulatory factor of many motor-based biological events. By combining the enzymatic measurements of the ncd-microtubules system with various in vitro binding assays employing proteolytic, recombinant and synthetic fragments of basic calponin, we further unambiguously identified the interaction of microtubules at two distinct calponin sites. One is inhibitory and resides in the segment 145-182, which also binds F-actin and calmodulin. The other one is noninhibitory, specific for microtubules, and is located on the COOH-terminal repeat-containing region 183-292. Finally, quantitative fluorescence studies of the binding of basic calponin to the skeletal pyrenyl F-actin in the presence of microtubules did not reveal a noticeable competition between the two sets of filaments for calponin. This result implies that calponin undergoes a concomitant binding to both F-actin and microtubules by interaction at the former site with actin and at the second site with microtubules. Thus, in the living cells, calponin could potentially behave as a cross-linking protein between the two major cytoskeletal filaments.     

The Microtubule Binding Properties of CENP-E's C-Terminus and CENP-F

CENP-E (centromere protein E) and CENP-F (centromere protein F), also known as mitosin, are large, multi-functional proteins associated with the outer kinetochore. CENP-E features a well-characterized kinesin motor domain at its N-terminus and a second microtubule-binding domain at its C-terminus of unknown function. CENP-F is important for the formation of proper kinetochore–microtubule attachment and, similar to CENP-E, contains two microtubule-binding domains at its termini. While the importance of these proteins is known, the details of their interactions with microtubules have not yet been investigated. We have biochemically and structurally characterized the microtubule-binding properties of the amino- and carboxyl-terminal domains of CENP-F as well as the carboxyl-terminal (non-kinesin) domain of CENP-E. CENP-E's C-terminus and CENP-F's N-terminus bind microtubules with similar affinity to the well-characterized Ndc80 complex, while CENP-F's C-terminus shows much lower affinity. Electron microscopy analysis reveals that all of these domains engage the microtubule surface in a disordered manner, suggesting that these factors have no favored binding geometry and may allow for initial side-on attachments early in mitosis.     

Centromere function on minichromosomes isolated from budding yeast.

Centromeres are a complex of centromere DNA (CEN DNA) and specific factors that help mediate microtubule-dependent movement of chromosomes during mitosis. Minichromosomes can be isolated from budding yeast in a way that their centromeres retain the ability to bind microtubules in vitro. Here, we use the binding of these minichromosomes to microtubules to gain insight into the properties of centromeres assembled in vivo. Our results suggest that neither chromosomal DNA topology nor proximity of telomeres influence the cell's ability to assemble centromeres with microtubule-binding activity. The microtubule-binding activity of the minichromosome's centromere is stable in the presence of competitor CEN DNA, suggesting that the complex between the minichromosome CEN DNA and proteins directly bound to it is very stable. The efficiency of centromere binding to microtubules is dependent upon the concentration of microtubule polymer and is inhibited by ATP. These properties are similar to those exhibited by mechanochemical motors. The binding of minichromosomes to microtubules can be inactivated by the presence of 0.2 M NaCl and then reactivated by restoring NaCl to 0.1 M. In 0.2 M NaCl, some centromere factor(s) bind to microtubules, whereas other(s) apparently remain bound to the minichromosome's CEN DNA. Therefore, the yeast centromere appears to consist of two domains: the first consists of a stable core containing CEN DNA and CEN DNA-binding proteins; the second contains a microtubule-binding component(s). The molecular functions of this second domain are discussed.     

Long range allosteric control of cytoplasmic dynein ATPase activity by the stalk and C-terminal domains

The dynein motor domain consists of a ring of six AAA domains with a protruding microtubule-binding stalk and a C-terminal domain of unknown function. To understand how conformational information is communicated within this complex structure, we produced a series of recombinant and proteolytic rat motor domain fragments, which we analyzed enzymatically. A recombinant 210-kDa half-motor domain fragment surprisingly exhibited a 6-fold higher steady state ATPase activity than a 380-kDa complete motor domain fragment. The increased ATPase activity was associated with a complete loss of sensitivity to inhibition by vanadate and an approximately 100-fold increase in the rate of ADP release. The time course of product release was discovered to be biphasic, and each phase was stimulated approximately 1000-fold by microtubule binding to the 380-kDa motor domain. Both the half-motor and full motor domain fragments were remarkably resistant to tryptic proteolysis, exhibiting either two or three major cleavage sites. Cleavage near the C terminus of the 380-kDa motor domain released a 32-kDa fragment and abolished sensitivity to vanadate. Cleavage at this site was insensitive to ATP or 5'-adenylyl-beta,gamma-imidodiphosphate but was blocked by ADP-AlF3 or ADP-vanadate. Based on these data, we proposed a model for long range allosteric control of product release at AAA1 and AAA3 through the microtubule-binding stalk and the C-terminal domain, the latter of which may interact with AAA1 to close the motor domain ring in a cross-bridge cycle-dependent manner.     

Identification and molecular characterization of E-MAP-115, a novel microtubule-associated protein predominantly expressed in epithelial cells

A novel microtubule-associated protein (MAP) of M(r) 115,000 has been identified by screening of a HeLa cell cDNA expression library with an anti-serum raised against microtubule-binding proteins from HeLa cells. Monoclonal and affinity-purified polyclonal antibodies were generated for the further characterization of this MAP. It is different from the microtubule-binding proteins of similar molecular weights, characterized so far, by its nucleotide-insensitive binding to microtubules and different sedimentation behavior. Since it is predominantly expressed in cells of epithelial origin (Caco-2, HeLa, MDCK), and rare (human skin, A72) or not detectable (Vero) in fibroblastic cells, we name it E-MAP-115 (epithelial MAP of 115 kD). In HeLa cells, E-MAP-115 is preferentially associated with subdomains or subsets of perinuclear microtubules. In Caco-2 cells, labeling for E- MAP-115 increases when they polarize and form blisters. The molecular characterization of E-MAP-115 reveals that it is a novel protein with no significant homologies to other known proteins. The secondary structure predicted from its sequence indicates two domains connected by a putative hinge region rich in proline and alanine (PAPA region). E- MAP-115 has two highly charged regions with predicted alpha-helical structure, one basic with a pI of 10.9 in the NH2-terminal domain and one neutral with a pI of 7.6 immediately following the PAPA region in the acidic COOH-terminal half of the molecule. A novel microtubule- binding site has been localized to the basic alpha-helical region in the NH2-terminal domain using in vitro microtubule-binding assays and expression of mutant polypeptides in vivo. Overexpression of this domain of E-MAP-115 by transfection of fibroblasts lacking significant levels of this protein with its cDNA renders microtubules stable to nocodazole. We conclude that E-MAP-115 is a microtubule-stabilizing protein that may play an important role during reorganization of microtubules during polarization and differentiation of epithelial cells.     

Short Stop provides an essential link between F-actin and microtubules during axon extension

Coordination of F-actin and microtubule dynamics is important for cellular motility and morphogenesis, but little is known about underlying mechanisms. short stop (shot) encodes an evolutionarily conserved, neuronally expressed family of rod-like proteins required for sensory and motor axon extension in Drosophila melanogaster. We identify Shot isoforms that contain N-terminal F-actin and C-terminal microtubule-binding domains, and that crosslink F-actin and microtubules in cultured cells. The F-actin- and microtubule-binding domains of Shot are required in the same molecule for axon extension, though the length of the connecting rod domain can be dramatically reduced without affecting activity. Shot therefore functions as a cytoskeletal crosslinker in axon extension, rather than mediating independent interactions with F-actin and microtubules. A Ca(2+)-binding motif located adjacent to the microtubule-binding domain is also required for axon extension, suggesting that intracellular Ca(2+) release may regulate Shot activity. These results suggest that Shot coordinates regulated interactions between F-actin and microtubules that are crucial for neuronal morphogenesis.     

GMAP-210, A cis-Golgi network-associated protein, is a minus end microtubule-binding protein

We report that a peripheral Golgi protein with a molecular mass of 210 kD localized at the cis-Golgi network (Rios, R.M., A.M. Tassin, C. Celati, C. Antony, M.C. Boissier, J.C. Homberg, and M. Bornens. 1994. J. Cell Biol. 125:997-1013) is a microtubule-binding protein that associates in situ with a subpopulation of stable microtubules. Interaction of this protein, now called GMAP-210, for Golgi microtubule-associated protein 210, with microtubules in vitro is direct, tight and nucleotide-independent. Biochemical analysis further suggests that GMAP-210 specifically binds to microtubule ends. The full-length cDNA encoding GMAP-210 predicts a protein of 1, 979 amino acids with a very long central coiled-coil domain. Deletion analyses in vitro show that the COOH terminus of GMAP-210 binds to microtubules whereas the NH2 terminus binds to Golgi membranes. Overexpression of GMAP-210-encoding cDNA induced a dramatic enlargement of the Golgi apparatus and perturbations in the microtubule network. These effects did not occur when a mutant lacking the COOH-terminal domain was expressed. When transfected in fusion with the green fluorescent protein, the NH2-terminal domain associated with the cis-Golgi network whereas the COOH-terminal microtubule-binding domain localized at the centrosome. Altogether these data support the view that GMAP-210 serves to link the cis-Golgi network to the minus ends of centrosome-nucleated microtubules. In addition, this interaction appears essential for ensuring the proper morphology and size of the Golgi apparatus.     

A microtubule-binding domain in dynactin increases dynein processivity by skating along microtubules

Microtubule-associated proteins (MAPs) use particular microtubule-binding domains that allow them to interact with microtubules in a manner specific to their individual cellular functions. Here, we have identified a highly basic microtubule-binding domain in the p150 subunit of dynactin that is only present in the dynactin members of the CAP-Gly family of proteins. Using single-particle microtubule-binding assays, we found that the basic domain of dynactin moves progressively along microtubules in the absence of molecular motors - a process we term 'skating'. In contrast, the previously described CAP-Gly domain of dynactin remains firmly attached to a single point on microtubules. Further analyses showed that microtubule skating is a form of one-dimensional diffusion along the microtubule. To determine the cellular function of the skating phenomenon, dynein and the dynactin microtubule-binding domains were examined in single-molecule motility assays. We found that the basic domain increased dynein processivity fourfold whereas the CAP-Gly domain inhibited dynein motility. Our data show that the ability of the basic domain of dynactin to skate along microtubules is used by dynein to maintain longer interactions for each encounter with microtubules.     

15 A resolution model of the monomeric kinesin motor, KIF1A

A two-headed structure has been widely believed to be essential for the kinesin molecular motor to move processively on the track, microtubules. However, we have recently demonstrated that a monomeric motor domain construct of KIF1A (C351), a kinesin superfamily protein, moves processively, taking about 700 steps before being detached from microtubules. To elucidate the mechanism of its single-headed processivity, we examined the C351 -MT interaction by mutant analysis and high-resolution cryo-EM. Mutant analysis indicated the importance of a highly positively charged loop, the "K loop," for such processivity. A 15 A resolution structure unambiguously docked with the available atomic models revealed "K loop" as an extra microtubule-binding domain specific to KIF1A, and bound to the C terminus of tubulin. The site-specific cross-linking further confirmed this model.     

Multivariate analysis of conserved sequence-structure relationships in kinesins: coupling of the active site and a tubulin-binding sub-domain

An extensive computational analysis of available sequence and crystal structure data was used to identify functionally important residue interactions within the motor domain of the kinesin molecular motor. Principal component analysis revealed that all current kinesin crystal structures reside in one of two main conformations, which differ at the active site, and in the position of a microtubule-binding sub-domain relative to a rigid central core. This sub-domain consists of secondary structure elements alpha4-loop12-alpha5-loop13 and contains a conserved hydrophilic surface patch that may be involved in strong binding to microtubules. A hinge point for the sub-domain motion lies near a conserved glycine at position 292. Statistical coupling analysis revealed a network of co-evolving positions that link this region to the nucleotide-binding site, via a highly conserved histidine in the switch I loop. The data are consistent with a model in which the nucleotide status of the active site shifts kinesin between weak and strong binding conformations via reconfiguration of the identified sub-domain. Our data provide a statistically supported framework for further examination of this and other structure-function relationships in the kinesin family.     

Arabidopsis katanin binds microtubules using a multimeric microtubule-binding domain.

Katanin is a heterodimeric protein that mediates ATP-dependent destabilization of microtubules in animal cells. In plants, the catalytic subunit of Arabidopsis thaliana katanin (AtKSS, Arabidopsis thaliana Katanin Small Subunit) has been identified and its microtubule-severing activity has been demonstrated in vitro. In vivo, plant katanin plays a role in the organization of cortical microtubules, but the way it achieves this function is unknown. To go further in our understanding of the mechanisms by which katanin severs microtubules, we analyzed the functional domains of Arabidopsis katanin. We characterized the microtubule-binding domain of katanin both in vitro and in vivo. It corresponds to a poorly conserved sequence between plant and animal katanins that is located in the N-terminus of the protein. This domain interacts with cortical microtubules in vivo and has a low affinity for microtubules in vitro. We also observed that katanin microtubule-binding domain oligomerizes into trimers. These results show that, besides being involved in the interaction of katanin with microtubules, the microtubule-binding domain may also participate in the oligomerization of katanin. At the structural level, we observed that AtKSS forms ring-shaped oligomers.