The microtubule lattice--20 years on

In 1974, optical diffraction and image analysis indicated that tubulin dimers in the cylindrically complete A-tubule of flagellar doublet microtubules are arranged with helical symmetry, while those in the incomplete B-tubule associate differently. Recently, electron micrographs of reassembled brain microtubules decorated with kinesin heads have shown that the tubulin dimers there are arranged as in the B-tubule. The lack of symmetry of microtubules assembled in vitro prompts Linda Amos to speculate here that the assembly process in vitro may differ from that occurring in the cell.

Discodermolide interferes with the binding of tau protein to microtubules

We investigated whether discodermolide, a novel antimitotic agent, affects the binding to microtubules of tau protein repeat motifs. Like taxol, the new drug reduces the proportion of tau that pellets with microtubules. Despite their differing structures, discodermolide, taxol and tau repeats all bind to a site on beta-tubulin that lies within the microtubule lumen and is crucial in controlling microtubule assembly. Low concentrations of tau still bind strongly to the outer surfaces of preformed microtubules when the acidic C-terminal regions of at least six tubulin dimers are available for interaction with each tau molecule; otherwise binding is very weak.

Repeat motifs of tau bind to the insides of microtubules in the absence of taxol

The tau family of microtubule-associated proteins has a microtubule-binding domain which includes three or four conserved sequence repeats. Pelleting assays show that when tubulin and tau are co- assembled into microtubules, the presence of taxol reduces the amount of tau incorporated. In the absence of taxol, strong binding sites for tau are filled by one repeat motif per tubulin dimer; additional tau molecules bind more weakly. We have labelled a repeat motif with nanogold and used three-dimensional electron cryomicroscopy to compare images of microtubules assembled with labelled or unlabelled tau. With kinesin motor domains bound to the microtubule outer surface to distinguish between alpha- and beta-tubulin, we show that the gold label lies on the inner surface close to the taxol binding site on beta-tubulin. Loops within the repeat motifs of tau have sequence similarity to an extended loop which occupies a site in alpha-tubulin equivalent to the taxol-binding pocket in beta-tubulin. We propose that loops in bound tau stabilize microtubules in a similar way to taxol, although with lower affinity so that assembly is reversible.

F-actin-like filaments formed by plasmid segregation protein ParM

It was the general belief that DNA partitioning in prokaryotes is independent of a cytoskeletal structure, which in eukaryotic cells is indispensable for DNA segregation. Recently, however, immunofluorescence microscopy revealed highly dynamic, filamentous structures along the longitudinal axis of Escherichia coli formed by ParM, a plasmid-encoded protein required for accurate segregation of low-copy-number plasmid R1. We show here that ParM polymerizes into double helical protofilaments with a longitudinal repeat similar to filamentous actin (F-actin) and MreB filaments that maintain the cell shape of non-spherical bacteria. The crystal structure of ParM with and without ADP demonstrates that it is a member of the actin family of proteins and shows a domain movement of 25 degrees upon nucleotide binding. Furthermore, the crystal structure of ParM reveals major differences in the protofilament interface compared with F-actin, despite the similar arrangement of the subunits within the filaments. Thus, there is now evidence for cytoskeletal structures, formed by actin-like filaments that are involved in plasmid partitioning in E.coli.

Prokaryotic origin of the actin cytoskeleton

It was thought until recently that bacteria lack the actin or tubulin filament networks that organize eukaryotic cytoplasm. However, we show here that the bacterial MreB protein assembles into filaments with a subunit repeat similar to that of F-actin-the physiological polymer of eukaryotic actin. By elucidating the MreB crystal structure we demonstrate that MreB and actin are very similar in three dimensions. Moreover, the crystals contain protofilaments, allowing visualization of actin-like strands at atomic resolution. The structure of the MreB protofilament is in remarkably good agreement with the model for F-actin, showing that the proteins assemble in identical orientations. The actin-like properties of MreB explain the finding that MreB forms large fibrous spirals under the cell membrane of rod-shaped cells, where they are involved in cell-shape determination. Thus, prokaryotes are now known to possess homologues both of tubulin, namely FtsZ, and of actin.

Antibodies to cytoplasmic dynein heavy chain map the surface and inhibit motility

Polyclonal antibodies have been raised against four 16 residue peptides with sequences taken from the C-terminal quarter of the human cytoplasmic dynein heavy chain. The sites are downstream from a known microtubule-binding domain associated with the "stalk" that protrudes from the motor domain. The antisera were assayed using bacterially expressed proteins with amino acid sequences taken from the human cytoplasmic dynein heavy chain. Every antiserum reacted specifically with the appropriate expressed protein and with pig brain cytoplasmic dynein, whether the protein molecules were denatured on Western blots or were in a folded state. But, whereas three of the four antisera recognized freshly purified cytoplasmic dynein, the fourth reacted only with dynein that had been allowed to denature a little. After affinity purification against the expressed domains, whole IgG molecules and Fab fragments were assayed for their effect on dynein activity in in vitro microtubule-sliding assays. Of the three anti-peptides that reacted with fresh dynein, one inhibited motility but the others did not. The way these peptides are exposed on the surface is compatible with a model whereby the dynein motor domain is constructed from a ring of AAA protein modules, with the C-terminal module positioned on the surface that interacts with microtubules. We have tentatively identified an additional AAA module in the dynein heavy chain sequence, which would be consistent with a heptameric ring.

Structural comparison of dimeric Eg5, Neurospora kinesin (Nkin) and Ncd head-Nkin neck chimera with conventional kinesin

Cryo-electron microscopy and 3D image reconstruction of microtubules saturated with kinesin dimers has shown one head bound to tubulin, the other free. The free head of rat kinesin sits on the top right of the bound head (with the microtubule oriented plus-end upwards) in the presence of 5'-adenylylimido-diphosphate (AMPPNP) and on the top left in nucleotide-free solutions. To understand the relevance of this movement, we investigated other dimeric plus-end-directed motors: Neurospora kinesin (Nkin); Eg5, a slow non-processive kinesin; and a chimera of Ncd heads attached to Nkin necks. In the AMPPNP (ATP-like) state, all dimers have the free head to the top right. In the absence of nucleotide, the free head of an Nkin dimer appears to occupy alternative positions to either side of the bound head. Despite having the Nkin neck, the free head of the chimera was only seen to the top right of the bound head. Eg5 also has the free head mostly to the top right. We suggest that processive movement may require kinesins to move their heads in alternative ways.

Helical tubes of FtsZ from Methanococcus jannaschii

Bacterial cell division depends on the formation of a cytokinetic ring structure, the Z-ring. The bacterial tubulin homologue FtsZ is required for Z-ring formation. FtsZ assembles into various polymeric forms in vitro, indicating a structural role in the septum of bacteria. We have used recombinant FtsZ1 protein from M. jannaschii to produce helical tubes and sheets with high yield using the GTP analogue GMPCPP [guanylyl-(alpha,beta)-methylene-diphosphate]. The sheets appear identical to the previously reported Ca++-induced sheets of FtsZ from M. jannaschii that were shown to consist of 'thick'-filaments in which two protofilaments run in parallel. Tubes assembled either in Ca++ or in GMPCPP contain filaments whose dimensions indicate that they could be equivalent to the 'thick'-filaments in sheets. Some tubes are hollow but others are filled by additional protein density. Helical FtsZ tubes differ from eukaryotic microtubules in that the filaments curve around the filament axis with a pitch of approximately 430 A for Ca++-induced tubes or 590 - 620 A for GMPCPP. However, their assembly in vitro as well-ordered polymers over distances comparable to the inner circumference of a bacterium may indicate a role in vivo. Their size and stability make them suitable for use in motility assays.

The conformational cycle of kinesin

The stepping mechanism of kinesin can be thought of as a programme of conformational changes. We briefly review protein chemical, electron microscopic and transient kinetic evidence for conformational changes, and working from this evidence, outline a model for the mechanism. In the model, both kinesin heads initially trap Mg x ADP. Microtubule binding releases ADP from one head only (the trailing head). Subsequent ATP binding and hydrolysis by the trailing head progressively accelerate attachment of the leading head, by positioning it closer to its next site. Once attached, the leading head releases its ADP and exerts a sustained pull on the trailing head. The rate of closure of the molecular gate which traps ADP on the trailing head governs its detachment rate. A speculative but crucial coordinating feature is that this rate is strain sensitive, slowing down under negative strain and accelerating under positive strain.

Focusing-in on microtubules

A good approximation of the atomic structure of a microtubule has been derived from docking the high-resolution structure of tubulin, solved by electron crystallography, into lower resolution maps of whole microtubules. Some structural interactions with other molecules, including nucleotides, drugs, motor proteins and microtubule-associated proteins, can now be predicted.

Congruent docking of dimeric kinesin and ncd into three-dimensional electron cryomicroscopy maps of microtubule-motor ADP complexes

We present a new map showing dimeric kinesin bound to microtubules in the presence of ADP that was obtained by electron cryomicroscopy and image reconstruction. The directly bound monomer (first head) shows a different conformation from one in the more tightly bound empty state. This change in the first head is amplified as a movement of the second (tethered) head, which tilts upward. The atomic coordinates of kinesin.ADP dock into our map so that the tethered head associates with the bound head as in the kinesin dimer structure seen by x-ray crystallography. The new docking orientation avoids problems associated with previous predictions; it puts residues implicated by proteolysis-protection and mutagenesis studies near the microtubule but does not lead to steric interference between the coiled-coil tail and the microtubule surface. The observed conformational changes in the tightly bound states would probably bring some important residues closer to tubulin. As expected from the homology with kinesin, the atomic coordinates of nonclaret disjunctional protein (ncd).ADP dock in the same orientation into the attached head in a map of microtubules decorated with dimeric ncd.ADP. Our results support the idea that the observed direct interaction between the two heads is important at some stages of the mechanism by which kinesin moves processively along microtubules.

Tubulin-like protofilaments in Ca2+-induced FtsZ sheets

The 40 kDa protein FtsZ is a major septum-forming component of bacterial cell division. Early during cytokinesis at midcell, FtsZ forms a cytokinetic ring that constricts as septation progresses. FtsZ has a high propensity to polymerize in vitro into various structures, including sheets and filaments, in a GTP-dependent manner. Together with limited sequence homology, the occurrence of the tubulin signature motif in FtsZ and a similar three-dimensional structure, this leads to the conclusion that FtsZ is the bacterial tubulin homologue. We have polymerized FtsZ1 from Methanococcus jannaschii in the presence of millimolar concentrations of Ca2+ ions to produce two-dimensional crystals of plane group P2221. Most of the protein precipitates and forms filaments approximately 23.0 nm in diameter. A three-dimensional reconstruction of tilted micrographs of FtsZ sheets in negative stain between 0 and 60 degrees shows protofilaments of FtsZ running along the sheet axis. Pairs of parallel FtsZ protofilaments associate in an antiparallel fashion to form a two-dimensional sheet. The antiparallel arrangement is believed to generate flat sheets instead of the curved filaments seen in other FtsZ polymers. Together with the subunit spacing along the protofilament axis, a fitting of the FtsZ crystal structure into the reconstruction suggests a protofilamant structure very similar to that of tubulin protofilaments.

How Taxol stabilises microtubule structure

The structure of tubulin shows paclitaxel (Taxol(R)) binding to a pocket in beta tubulin on the microtubule's inner surface, which counteracts the effects of GTP hydrolysis occurring on the other side of the monomer.

Tubulin and FtsZ form a distinct family of GTPases

Tubulin and FtsZ share a common fold of two domains connected by a central helix. Structure-based sequence alignment shows that common residues localize in the nucleotide-binding site and a region that interacts with the nucleotide of the next tubulin subunit in the protofilament, suggesting that tubulin and FtsZ use similar contacts to form filaments. Surfaces that would make lateral interactions between protofilaments or interact with motor proteins are, however, different. The highly conserved nucleotide-binding sites of tubulin and FtsZ clearly differ from those of EF-Tu and other GTPases, while resembling the nucleotide site of glyceraldehyde-3-phosphate dehydrogenase. Thus, tubulin and FtsZ form a distinct family of GTP-hydrolyzing proteins.

Nucleotide-dependent structural changes in dimeric NCD molecules complexed to microtubules

Complexes consisting of motor domains of the kinesin-like protein ncd bound to reassembled brain microtubules were visualised using cryoelectron microscopy and helical image reconstruction. Different nucleotide-associated states of a dimeric construct (NDelta295-700) of ncd were analysed to reveal ADP-containing, AMP.PNP-containing and empty (rigor) conformations. In these three states, each thought to mimic a different stage in ATP turnover, the double-headed motors attach to the microtubules by one head only, with the free head tethered in relatively fixed positions. The three structures differ both in the way the attached heads interact with tubulin and in the position of the tethered heads. In the strongly binding rigor and AMP.PNP (ATP-like) states, the attached head makes close contact with both subunits of a tubulin heterodimer. In the weakly bound ADP state, the contact made by the attached head with the monomer closer to the plus end appears to be more loose. Also, in the ATP-like state, the free head tilts nearer to the plus end than in the other two states. The data argue against model mechanisms in which a conformational change in the bound head guides the free head closer to its next binding site; on the contrary, the transition from ADP-filled via rigor to the AMP.PNP (ATP-like) state of the bound head produces a small motion of the free head in the counter-productive direction. However, the observation that the tethered head points towards the minus end, in all three states, is consistent with the idea that the relative arrangement of the heads in a dimer is a major determinant of directionality.

Crystal structure of the bacterial cell-division protein FtsZ

Bacterial cell division ends with septation, the constriction of the cell wall and cell membranes that leads to the formation of two daughter cells. During septation, FtsZ, a protein of relative molecular mass 40,000 which is ubiquitous in eubacteria and is also found in archaea and chloroplasts, localizes early at the division site to form a ring-shaped septum. This septum is required for the mechanochemical process of membrane constriction. FtsZ is a GTPase with weak sequence homology to tubulins. The nature of FtsZ polymers in vivo is unknown, but FtsZ can form tubules, sheets and minirings in vitro. Here we report the crystal structure at 2.8 A resolution of recombinant FtsZ from the hyperthermophilic methanogen Methanococcus jannaschii. FtsZ has two domains, one of which is a GTPase domain with a fold related to one found in the proteins p21ras and elongation factor EF-Tu. The carboxy-terminal domain, whose function is unknown, is a four-stranded beta-sheet tilted by 90 degrees against the beta-sheet of the GTPase domain. The two domains are arranged around a central helix. GDP binding is different from that typically found in GTPases and involves four phosphate-binding loops and a sugar-binding loop in the first domain, with guanine being recognized by residues in the central connecting helix. The three-dimensional structure of FtsZ is similar to the structure of alpha- and beta-tubulin.

Structure and dynamics of molecular motors

The structures of the oppositely directed microtubule motors kinesin and ncd have been solved to atomic resolution. The two structures are very similar and are also homologous to myosin. Myosins and kinesins differ kinetically but, tantalizingly, cryoelectron microscopy has recently revealed that both structures may tilt during ADP release. Such evidence suggests that the two motor families use common structural mechanisms.

Three-dimensional cryoelectron microscopy of 16-protofilament microtubules: structure, polarity, and interaction with motor proteins

We present a three-dimensional (3D) map, reconstructed from electron microscope (EM) images of naturally occurring 16-protofilament (PF) microtubules (MTs) in ice. We compare it with the tubulin in six 3D maps of MTs decorated with motor domains, three from frozen MTs decorated with kinesin or ncd in the tightly bound AMP-PNP state, and three from negatively stained MTs decorated with kinesin in different nucleotide states. The comparison confirms that kinesin and ncd bind to identical sites and interact with both monomers of a tubulin dimer. Maps of specimens in negative stain and in ice are similar except that the protein in the top half of a motor domain appears denser in negative stain. The interactions have only a small effect on tubulin structure; the outward appearance is unchanged, but there seems to be a small internal rearrangement. The relative polarity of undecorated and decorated MTs is evident from their 3D structures. This agrees with the absolute polarities indicated by the orientations of motors in decorated specimens and by polar superposition patterns calculated for undecorated MTs. An image of tubulin PFs in zinc-induced sheets has been tentatively oriented by similar criteria.

The structure of microtubule-motor complexes

New images, calculated from electron micrographs, show the three-dimensional structures of microtubules and tubulin sheets decorated stoichiometrically with globular motor protein domains (heads). Single heads of kinesin and ncd, the kinesin-related protein that moves in the reverse direction to kinesin, bind in the same way to the same site on tubulin. Dimeric kinesin and dimeric ncd show an interesting difference in the positions of their second heads.

Three-dimensional cryoelectron microscopy of dimeric kinesin and ncd motor domains on microtubules

Kinesin and ncd motor proteins are homologous in sequence yet move in opposite directions along microtubules. We have previously shown that monomeric kinesin and ncd bind in the same orientation on equivalent sites relative to the ends of tubulin sheets of known polarity. We now report cryoelectron microscope images of 16-protofilament microtubules decorated with both single- and double-headed kinesin and double-headed ncd. Three-dimensional density maps and difference maps show that, in adenosine 5'-[beta,gamma-imido]triphosphate, both dimeric motors bind tightly to microtubules via one head, leaving the other free, though apparently in a fixed position. The attached heads of dimers bind to tubulin in the same way as single kinesin heads. The second heads are connected to the tops of the first but, whereas the second kinesin head is closely associated with the first, pairs of ncd heads are splayed apart. There is also a distinct difference in orientation: the second kinesin head is tilted toward the microtubule plus end, while the second head of ncd points toward the minus end.

Microtubule minus ends can be labelled with a phage display antibody specific to alpha-tubulin

To investigate the orientation of alpha- and beta-tubulin heterodimers within microtubules, we cloned a phage display antibody to alpha-tubulin. The N-terminal 100 residues of alpha-tubulin were bacterially expressed and used to select clones from a large repertoire of antibody-expressing phagemid particles. One clone reacted with the expressed alpha-tubulin N terminus and native tubulin dimer but not with the expressed beta-tubulin N terminus. Electron microscopy showed 30 nm gold beads coated with the antibody binding to one end of brain microtubules. The beads bound to the minus ends of axonemes but not to the brain tubulin extensions from their plus ends. In sliding motility assays with a plus end directed motor, beads were pushed ahead of the microtubules. Our results indicate that an N-terminal epitope of alpha-tubulin is exposed only at the minus ends of microtubules.

Structural comparison of tektins and evidence for their determination of complex spacings in flagellar microtubules

Recent structural studies indicate that a tektin heteropolymer forms a unique protofilament of flagellar microtubules. We report here the sequence of tektin C (approximately 47 kDa), predicted from its cDNA (GenBank U38523), compared to tektins A (approximately 53 kDa) and B (approximately 51 kDa) from sea urchin (Strongylocentrotus purpuratus) sperm flagellar microtubules, and compared to partial sequences reported from mouse and human. We are now able to make several observations concerning the tektin family: (1) their common structural features, (2) a comparison of their structure to intermediate filament proteins, and (3) their possible organization in the tektin filament polymer. The predicted amino acid sequence identities/similarities are: for tektins A and C, 42/54%, for tektins A and B, 34/51%; for tektins B and C, 29/42%; for tektin C and a partial cDNA clone from mouse testis, 55/65%; and for tektin B and a partial cDNA clone from the human brain, 45/47%. The three tektins (and the human clone) possess the exact sequence repeat RPNVELCRD. The structural pattern of all three tektin polypeptides is similar to intermediate filament proteins. Tektins are predicted to form extended rods composed of two alpha-helical segments (approximately 180 residues long) capable of forming coiled coils, which are interrupted by short non-helical linkers. The two segments are homologous in sequence and secondary structure, indicating a gene duplication event prior to the divergence of the three tektins. Along each tektin rod cysteine residues occur with a periodicity of approximately 8 nm, coincident with the axial repeat of tubulin dimers in microtubules. From EM data and calculations of secondary structure, the segment length of tektin AB heterodimers is likely to be 16 nm. Both segments of tektin C may be 24 nm long, but one may be 16 nm. On the basis of the available evidence, we propose that coassembly of tektin AB heterodimers with tektin C dimers produces filaments with overall repeats of 8, 16, 24, 32, 40, 48 and 96 nm, generating the basis for the complex spatial arrangements of axonemal components.

Re-examination of the polarity of microtubules and sheets decorated with kinesin motor domain

Electron microscope images of microtubules and tubulin sheets decorated with kinesin head domains have shown the main mass of the kinesin head domain to be superimposed on one subunit of each tubulin dimer. We have polymerized brain tubulin extensions on to the ends of flagellar axonemes under varied conditions, in order to check the polarity of the tubulin-kinesin head complex. Since the polarity of axonemes incubated with normal brain tubulin may be ambiguous, we also tried 50% N-ethylmaleimide-treated tubulin which specifically blocks minus ends. Our conclusion, which conflicts with recently published results, is that the main mass of the kinesin head is associated with the tubulin subunit closer to the plus end of a microtubule.

Nucleotide-dependent angular change in kinesin motor domain bound to tubulin

Kinesin is a 'motor' molecule, consisting of two head domains, an alpha-helical coiled coil rod, and a tail part that binds to its cargo. When expressed in a bacterial system, the head domain is functional, and can bind to microtubules with the stoichiometry of one head per tubulin dimer. Kinesin moves along microtubules by means of a cyclic process of nucleotide binding, hydrolysis and product release. We have used negative-stain electron microscopy and image analysis to study the structures of microtubules and tubulin sheets decorated with the motor domain (head) of kinesin in three states: in the presence of an unhydrolysable ATP analogue, 5'-adenylylimidodiphosphate (AMP-PNP); without nucleotides; and with adenosine 5'-diphosphate (ADP). A single kinesin head bound to a microtubule has a pear-shaped structure, with the broader end towards the 'plus' end of the microtubule under all conditions; the reverse motor, ncd, is similarly oriented. Three-dimensional maps reveal that kinesin heads have a spike that is assumed to form the attachment to the tail of a complete kinesin molecule. This spike is perpendicular to the microtubule axis in the presence of ADP, but points towards the plus end (approximately 45 degrees) in the presence of AMP-PNP or absence of nucleotides. Our results provide direct evidence for a conformational change of the kinesin motor domain during the ATPase cycle.