Quick-freeze Deep-etch Electron Microscopy Research Articles

Assembly, Storage, and Secretion of Von Willebrand Factor.

Von Willebrand factor (VWF) plays a central role in hemostasis, and dysregulation of VWF can cause either bleeding or thrombosis. Endothelial cells assemble VWF multimers in two stages that occur in distinct intracellular compartments: proVWF subunits dimerize in the endoplasmic reticulum through “tail-to-tail” disulfide bonds between C-terminal CK domains, and the proVWF dimers form enormous multimers in the Golgi through “head-to-head” disulfide bonds between N-terminal D3 domains. The finished multimers are packaged into ordered tubules within storage organelles called Weibel-Palade bodies (WPB), and tubular packing is necessary for the secretion of VWF filaments that have normal hemostatic function. We have recreated VWF tubule assembly in vitro, starting with pure VWF propeptide (domains D1D2) and disulfide-linked dimers of adjacent N-terminal D’D3 domains. No other cellular proteins or components of WPB are needed. Assembly requires low pH and calcium ions, similar to conditions in the Golgi. Quick-freeze deep-etch electron microscopy and three-dimensional reconstruction of negatively stained images show that tubules contain a repeating unit of one D’D3 dimer and two propeptides arranged in a right-handed helix with 4.2 units per turn. The symmetry and location of interdomain contacts suggest that decreasing pH along the secretory pathway coordinates the disulfide-linked assembly of VWF multimers with their tubular packaging. Secretion into the blood exposes VWF tubules to neutral pH conditions, releasing the constraints that maintain tubular packing and permitting the orderly unfurling of large VWF multimers. Some VWF multimers bind platelets and initiate the growth of platelet-rich thrombi. Under normal circumstances, these thrombi are limited in size by ADAMTS13, a metalloprotease that cleaves VWF multimers and releases the platelets. The absence of large VWF multimers causes bleeding that is typical of von Willebrand disease. Conversely, congenital or acquired deficiency of ADAMTS13 prevents the dissolution of VWF-platelet aggregates, which can cause the widespread microvascular thrombosis that characterizes thrombotic thrombocytopenic purpura (TTP). Interestingly, ADAMTS13 deficiency alone need not trigger TTP. Some patients persist for months or years without active disease but become ill whenever they suffer additional inflammatory stress associated with infection, surgery, or pregnancy. Thus, interactions between inflammatory mediators and endothelial cells can determine the course of VWF-dependent thrombosis. The mechanisms responsible for these phenomena remain poorly understood.

Assembly of Weibel-Palade body-like tubules from N-terminal domains of von Willebrand factor.

Endothelial cells assemble von Willebrand factor (VWF) multimers into ordered tubules within storage organelles called Weibel-Palade bodies, and tubular packing is necessary for the secretion of VWF filaments that can bind connective tissue and recruit platelets to sites of vascular injury. We now have recreated VWF tubule assembly in vitro, starting with only pure VWF propeptide (domains D1D2) and disulfide-linked dimers of adjacent N-terminal D'D3 domains. Assembly requires low pH and calcium ions and is reversed at neutral pH. Quick-freeze deep-etch electron microscopy and three-dimensional reconstruction of negatively stained images show that tubules contain a repeating unit of one D'D3 dimer and two propeptides arranged in a right-handed helix with 4.2 units per turn. The symmetry and location of interdomain contacts suggest that decreasing pH along the secretory pathway coordinates the disulfide-linked assembly of VWF multimers with their tubular packaging.

Structure of pericellular matrix around agarose-embedded chondrocytes

Determine whether the structure of the type VI collagen component of the chondrocyte pericellular matrix (PCM) generated by agarose-embedded chondrocytes in culture is similar to that found in native articular cartilage. Confocal microscopy, quick-freeze deep-etch electron microscopy, and real-time polymerase chain reaction (PCR) were used to investigate temporal and spatial patterns of type VI collagen protein deposition and gene expression by bovine chondrocytes during 4 weeks of culture within a 2% agarose hydrogel. Similar analyses were performed on chondrocytes within samples of intact cartilage obtained from the same joint surfaces as those used for cell isolation for comparison. Type VI collagen accumulated uniformly around cells embedded in agarose, with the rate of deposition slowing after the second week. After 1 week, PCM fibrils were observed to be oriented perpendicular to the cell surface, in contrast with the primarily tangential fibrillar arrangement observed in native articular cartilage. Expression of col6 in agarose-embedded cells was initially much higher ( approximately 400%) than that in chondrocytes within cartilage. Expression of col6 in the cultured chondrocytes declined by approximately 60% after 1 week, and remained stable thereafter. PCM structure and composition around cells in a hydrogel scaffold may be different than that in native cartilage, with potential implications for mass transport, mechanotransduction, and ultimately, the success of tissue engineering approaches.

A procedure to analyze surface profiles of the protein molecules visualized by quick-freeze deep-etch replica electron microscopy

Quick-freeze deep-etch replica electron microscopy gives high contrast snapshots of individual protein molecules under physiological conditions in vitro or in situ. The images show delicate internal pattern, possibly reflecting the rotary-shadowed surface profile of the molecule. As a step to build the new system for the “Structural analysis of single molecules”, we propose a procedure to quantitatively characterize the structural property of individual molecules; e.g. conformational type and precise view-angle of the molecules, if the crystallographic structure of the target molecule is available. This paper presents a framework to determine the observed face of the protein molecule by analyzing the surface profile of individual molecules visualized in freeze-replica specimens. A comprehensive set of rotary-shadowed views of the protein molecule was artificially generated from the available atomic coordinates using light-rendering software. Exploiting new mathematical morphology-based image filter, characteristic features were extracted from each image and stored as template. Similar features were extracted from the true replica image and the most likely projection angle and the conformation of the observed particle were determined by quantitative comparison with a set of archived images. The performance and the robustness of the procedure were examined with myosin head structure in defined configuration for actual application.

2P537 Mesoscopic model of the transient actin-sliding myosin crossbridges as captured by quick-freeze deep-etch replica electron microscopy(52. Bio-imaging,Poster Session,Abstract,Meeting Program of EABS & BSJ 2006)

2P537 Mesoscopic model of the transient actin-sliding myosin crossbridges as captured by quick-freeze deep-etch replica electron microscopy(52. Bio-imaging,Poster Session,Abstract,Meeting Program of EABS & BSJ 2006)

Visualization of the Nuclear Lamina in Mouse Anterior Pituitary Cells and Immunocytochemical Detection of Lamin A/C by Quick-freeze Freeze-substitution Electron Microscopy

We examined the nuclear lamina in the quickly frozen anterior pituitary cells by electron microscopic techniques combined with freeze substitution, deep etching, and immunocytochemistry and compared it with that in the chemically fixed cells. By quick-freeze freeze-substitution electron microscopy, an electron-lucent layer, as thick as 20 nm, was revealed just inside the inner nuclear membrane, whereas in the conventionally glutaraldehyde-fixed cells the layer was not seen. By quick-freeze deep-etch electron microscopy, we could not distinguish definitively the layer corresponding to the nuclear lamina in either fresh unfixed or glutaraldehyde-fixed cells. Immunofluorescence microscopy showed that lamin A/C in the nucleus was detected in the acetone-fixed cells and briefly in paraformaldehyde-fixed cells but not in the cells with prolonged paraformaldehyde fixation. Nuclear localization of lamin A/C was revealed by immunogold electron microscopy also in the quickly frozen and freeze-substituted cells, but not in the paraformaldehyde-fixed cells. Lamin A/C was localized mainly in the peripheral nucleoplasm within 60 nm from the inner nuclear membrane, which corresponded to the nuclear lamina. These results suggest that the nuclear lamina can be preserved both ultrastructurally and immunocytochemically by quick-freezing fixation, rather than by conventional chemical fixation.

Interaction of myosin.ADP.fluorometal complexes with fluorescent probes and direct observation using quick-freeze deep-etch electron microscopy.

Myosin forms stable ternary complexes with ADP and phosphate analogues of fluorometals that mimic different ATPase reaction intermediates corresponding to each step of the cross-bridge cycle. In the present study, we monitored the formation of ternary complexes of myosin.ADP.fluorometal using the fluorescence probe prodan. It has been reported that the fluorescence changes of the probe reflect the formation of intermediates in the ATPase reaction [Hiratsuka (1998) Biochemistry 37, 7167-7176]. Prodan bound to skeletal muscle heavy-mero-myosin (HMM).ADP.fluorometal, with each complex showing different fluorescence spectra. Prodan bound to the HMM.ADP.BeFn complex showed a slightly smaller red-shift than other complexes in the presence of ATP, suggesting a difference in the localized conformation or a difference in the population of BeFn species of global shape. We also examined directly the global structure of the HMM.ADP.fluorometal complexes using quick-freeze deep-etch replica electron microscopy. The HMM heads in the absence of nucleotides were mostly straight and elongated. In contrast, the HMM heads of ternary complexes showed sharply kinked or rounded configurations as seen in the presence of ATP. This is the first report of the direct observation of myosin-ADP-fluorometal ternary complexes, and the results suggest that these complexes indeed mimic the shape of the myosin head during ATP hydrolysis.

Periplasmic chaperone recognition motif of subunits mediates quaternary interactions in the pilus.

The class of proteins collectively known as periplasmic immunoglobulin-like chaperones play an essential role in the assembly of a diverse set of adhesive organelles used by pathogenic strains of Gram-negative bacteria. Herein, we present a combination of genetic and structural data that sheds new light on chaperone-subunit and subunit-subunit interactions in the prototypical P pilus system, and provides new insights into how PapD controls pilus biogenesis. New crystallographic data of PapD with the C-terminal fragment of a subunit suggest a mechanism for how periplasmic chaperones mediate the extraction of pilus subunits from the inner membrane, a prerequisite step for subunit folding. In addition, the conserved N- and C-terminal regions of pilus subunits are shown to participate in the quaternary interactions of the mature pilus following their uncapping by the chaperone. By coupling the folding of subunit proteins to the capping of their nascent assembly surfaces, periplasmic chaperones are thereby able to protect pilus subunits from premature oligomerization until their delivery to the outer membrane assembly site.

Intermembrane bridges within membrane organelles revealed by quick-freeze deep-etch electron microscopy.

Intracellular membrane-bounded organelles, such as the endoplasmic reticulum, Golgi apparatus, and mitochondrion, possess and maintain their shape and intrinsic relationship due to the nature of their membrane organization. To reveal the membranous attachments that support these shapes and relationships, we examined various kinds of cells by quick-freeze deep-etch electron microscopy. In the cisternae of the endoplasmic reticula, we found intermembrane bridges linking opposite membranes of the cisternae. Membranes of adjoining rough endoplasmic reticulum cisternae were linked by intermembrane bridges crossing a narrow cytoplasmic gap between cisternae. Intermembrane bridges were also found in and between the Golgi cisternae and in nuclear envelopes. Three kinds of intermembrane bridges were found within mitochondria: one linking between outer and inner mitochondrial membranes and the other two spanning the intracristal space and intercristal matrix space. The presence of intermembrane bridges within membrane organelles, except for those between rough endoplasmic reticulum cisternae, was seen in all cell types examined. Intermembrane bridges within membrane organelles provide a structural basis for the membrane organization of the organelles and thus may contribute to the functional integrity of the organelles.

Intermembrane bridges within membrane organelles revealed by quick‐freeze deep‐etch electron microscopy

Intracellular membrane–bounded organelles, such as the endoplasmic reticulum, Golgi apparatus, and mitochondrion, possess and maintain their shape and intrinsic relationship due to the nature of their membrane organization. To reveal the membranous attachments that support these shapes and relationships, we examined various kinds of cells by quick-freeze deep-etch electron microscopy. In the cisternae of the endoplasmic reticula, we found intermembrane bridges linking opposite membranes of the cisternae. Membranes of adjoining rough endoplasmic reticulum cisternae were linked by intermembrane bridges crossing a narrow cytoplasmic gap between cisternae. Intermembrane bridges were also found in and between the Golgi cisternae and in nuclear envelopes. Three kinds of intermembrane bridges were found within mitochondria: one linking between outer and inner mitochondrial membranes and the other two spanning the intracristal space and intercristal matrix space. The presence of intermembrane bridges within membrane organelles, except for those between rough endoplasmic reticulum cisternae, was seen in all cell types examined. Intermembrane bridges within membrane organelles provide a structural basis for the membrane organization of the organelles and thus may contribute to the functional integrity of the organelles. Anat. Rec. 251:339–345, 1998. © 1998 Wiley-Liss, Inc.

Cytoskeletal organization in the neurohypophysial axons of rats

Cytoskeletal organization in axons of the rat neurohypophysis was examined by quick-freeze deep-etch electron micrscopy and conventional thin-section electron microscopy. In the neurohypophysial axons, neurotubules and neurofilaments were arranged almost in parallel with each other, and interlinked by cross-bridges. Because the localization of MAP1A on the axonal cytoskeleton was shown by immunoelectron microscopy, MAP1A is a component of the cross-bridges between neurotubules. The neurotubules and neurofilaments occupied their respective domains in the axon. The neurotuble:neurofilament ratio was variable axon by axon. In some axons, neurotubules overwhelmed neurofilaments in number. In other axons, neurofilaments were predominant, and neurotubules were few. The neurosecretory granules were located in association with either the neurotubule or neurofilament domain. Fine fibrils 5–10 nm in diameter were associated with the inner surface of the axonal plasmalemma. In the Herring body, a few neurotubules occurred in the marginal region while neurofilaments were rarely seen, indicating that most of the neurosecretory granules have no relationship with the neurotubules in the body. These results suggest that the cytoskeleton in neurohypophysial axons may be involved in regulation of axonal transport and storage of neurosecretory granules.

Intergranular bridges in the anterior pituitary cell and their possible involvement in Ca2+-induced granule-granule fusion.

Quick-freeze deep-etch electron microscopy showed the presence of bridge-like structures between adjacent secretory granules in rat anterior pituitary secretory cells. These intergranular bridges were variable in length and thickness. The finest bridges were 7-8 nm in length, while the longest ones were as long as 80 nm. Annexin II, one of the Ca2+-dependent phospholipid-binding proteins, is known to interlink between two membranes and induce aggregation of liposomes and chromaffin granules under the presence of Ca2+. In anterior pituitary cells, annexin II was detected by immunoelectron microscopy at the contact sites of secretory granules with other granules. The anterior pituitary cells treated under the presence of extracellular Ca2+ with Clostridium perfringens enterotoxin which induces Ca2+ influx showed multigranular exocytosis, i.e., multiple fusions of secretory granules with each other and with the plasma membrane. The granule-granule fusion in progress could be captured by the quick-freeze deep-etch technique. The membranes of adjacent secretory granules were partially fused at their contact sites where intergranular strands were no longer seen, while there existed intergranular strands between unfused portions of the granule membranes. From these results, we consider that the intergranular bridges, some of which may be composed of annexin II, are involved in Ca2+-induced granule-granule fusion in anterior pituitary cells.

Quick-freeze deep-etch electron microscopy of the actin-heavy meromyosin complex during the in vitro motility assay

Since mica is a substitute for glass in the in vitro actin motility assay, I examined the structure of heavy meromyosin (HMM) crossbridges supporting actin filaments by quick-freeze deep-etch replica electron microscopy. This method was capable of resolving the inter-domain cleft of the monomeric actin molecule. HMM heads that are not bound to actin, when observed by this technique, were straight and elongated in the absence of ATP but strongly kinked upon addition of ATP or ADP·inorganic vanadate to produce the putative long-lived analog of HMM-ADP·inorganic phosphate. The low-magnification image of the ATP-containing acto-HMM preparation showed features characteristic of sliding actin filaments on glass coverslips. At high magnification, all the HMM molecules were found attached to actin by one head with the majority projecting perpendicular to the filament axis, whereas in the absence of ATP, HMM exhibited two-head binding with a preponderance of molecules tilted at 45°. Detailed examination of the shape of HMM heads involved in sliding showed a rounded, and flat appearance of the tip and comparatively thin neck portion as if the heads grasp actin filament, in contrast to rigor crossbridges which have a pear-shaped configuration with more gradual taper. Such configurations of HMM heads were essentially the same as I observed previously on acto-myosin subfragment-1 (S1) by the same technique, except for the presence of an additional neck portion of HMM which makes interpretaion of the images easier. Interestingly, under actively sliding conditions, very few heads were tilted in the rigor configuration. At first glance, the addition of ADP to the rigor-complex gave images rather like those obtained with ATP, but they turned out to be different. The contribution of the structural change of crossbridges to the force development is discussed.

Three-dimensional image analysis of myosin head in function as captured by quick-freeze deep-etch replica electron microscopy.

Quick-freeze deep-etch replica electron microscopy combined with mica-flake technique provides high contrast, high time- and spatial-resolution images of protein molecules in solution, whose three-dimensional structure is well preserved. Thus, it might be quite useful to obtain structural information of individual functioning molecules, such as myosin crossbridges under in vitro motility assay conditions. With that method, we could actually show that both heads of heavy meromyosin (HMM) crossbridges are mostly straight and bound to actin filaments with about 45 degree tilt-angle under rigor conditions, whereas they attached to actin through only one head with a wide variety of angles under in vitro sliding conditions. We also demonstrated that free HMM heads are strongly kinked in the presence of ATP or ADP/inorganic vanadate (Vi) in contrast to almost straight configuration in the absence of nucleotide. To examine more detailed structure of individual crossbridges, we tried to reconstruct the three-dimensional architecture of intramolecular subdomains of single HMM molecule. We took a series of tilted images of single HMM-ADP/Vi particle and successfully obtained its 3-D image by filtered back-projection, even with restricted range of tilt-angles. By comparison of the reconstruction with the atomic model of subfragment-1 (S1) without nucleotide, we found some great structural difference, which partly might be attributable to the conformational change by nucleotide binding.

Native structure and arrangement of inositol-1,4,5-trisphosphate receptor molecules in bovine cerebellar Purkinje cells as studied by quick-freeze deep-etch electron microscopy.

We used quick-freeze deep-etch replica electron microscopy to visualize the native structure of inositol-1,4,5-trisphosphate receptor (IP3R) in the cell. In the dendrites of Purkinje neurons of bovine cerebellum there were many vesicular organelles whose surfaces were covered with a two-dimensional crystalline array of molecules. Detailed examination of the cytoplasmic true surface of such vesicles in replica revealed that the structural unit, identified as IP3R by immunocytochemistry and subsequent Fourier analysis, is a square-shaped assembly and is aligned so that the side of the square is inclined by approximately 20 degrees from the row-line of the lattice. Comparison with the ryanodine receptor (RyaR), another intracellular Ca2+ channel on the endoplasmic reticulum, suggested that IP3R, unlike RyaR, has a very compact structure, potentially reflecting the crucial difference in the function of the cytoplasmic portion of the molecule.

Geometry of the Flagellar Motor in the Cytoplasmic Membrane of Salmonella typhimuriumas Determined by Stereo-photogrammetry of Quick-freeze Deep-etch Replica Images

The precise geometry of the flagellar basal structure anchored in the cytoplasmic membrane was determined by digital stereo-photogrammetry of the images captured by quick-freeze deep-etch replica electron microscopy. In order to examine the structure on the periplasmic side of the membrane, we analyzed the MS ring complexes of Salmonella typhimuriumoverproduced in the cytoplasmic membrane of Escherichia coli. The rod, the S ring, and the shoulder of the M ring were exposed to the periplasm. On the cytoplasmic side of the membrane, small bumps corresponding to the cytoplasmic rod were discernible. We also examined the intact inner surface of the cells of polyhook mutant which was prepared by a new protocol and found the bell-shaped structure extending from the membrane towards the cytoplasm. It was identified as the C ring, since it was located at the base of the polyhook. Various dimensions of the MS ring complex and the C ring projecting from the membrane were determined by digital stereo-photo grammetry, and a three-dimensional model of the total basal structure is presented.

FimH adhesin of type 1 pili is assembled into a fibrillar tip structure in the Enterobacteriaceae.

Type 1 pili are heteropolymeric mannosebinding fibers produced by all members of the Enterobacteriaceae family. The bulk of the fiber is composed of FimA. Two macromolecular complexes responsible for mediating an interaction with mannose-containing receptors were purified from fimA- Escherichia coli by mannose affinity chromatography and ion-exchange chromatography. One complex contained only the mannose-binding adhesin, FimH, associated with FimG, a minor component of the type 1 pilus. In the other complex the FimG-FimH moiety was loosely associated with a chaperone-minor subunit complex (FimC-FimF), possibly representing an intermediate in tip fibrilla assembly. The FimC chaperone has also been shown to form a preassembly complex with FimH that has been purified and characterized previously. Purified FimC did not bind to the FimG-FimH complex but did recognize FimH dissociated from the FimG-FimH complex. Quick-freeze deep-etch electron microscopy revealed that the FimG-FimH complex had a thin fibrillar architecture. High-resolution electron microscopy of type 1 pili revealed that a 16-nm fibrillar tip structure with an architecture identical to that of the FimG-FimH complex was joined end-to-end to the pilus rod. In a fimH- deletion mutant, the tip fibrillae joined to pilus rods were approximately 3 nm in length. The full-length tip fibrilla was restored by complementation with the fimH gene in trans. The bipartite nature of the type 1 pilus was also demonstrated on pili purified from clinical isolates of members of the Enterobacteriaceae family arguing that it is a conserved feature of the type 1 pilus.

Direct visualization of the microtubule lattice seam both in vitro and in vivo.

Microtubules are constructed from alpha- and beta-tubulin heterodimers that are arranged into protofilaments. Most commonly there are 13 or 14 protofilaments. A series of structural investigations using both electron microscopy and x-ray diffraction have indicated that there are two potential lattices (A and B) in which the tubulin subunits can be arranged. Electron microscopy has shown that kinesin heads, which bind only to beta-tubulin, follow a helical path with a 12-nm pitch in which subunits repeat every 8-nm axially, implying a primarily B-type lattice. However, these helical symmetry parameters are not consistent with a closed lattice and imply that there must be a discontinuity or "seam" along the microtubule. We have used quick-freeze deep-etch electron microscopy to obtain the first direct evidence for the presence of this seam in microtubules formed either in vivo or in vitro. In addition to a conventional single seam, we have also rarely found microtubules in which there is more than one seam. Overall our data indicates that microtubules have a predominantly B lattice, but that A lattice bonds between tubulin subunits are found at the seam. The cytoplasmic microtubules in mouse nerve cells also have predominantly B lattice structure and A lattice bonds at the seam. These observations have important implications for the interaction of microtubules with MAPs and with motor proteins, and for example, suggest that kinesin motors may follow a single protofilament track.

Quick-freeze, deep-etch visualization of exocytosis in anterior pituitary secretory cells: localization and possible roles of actin and annexin II.

The exocytotic process in the anterior pituitary secretory cells was studied using quick-freeze deep-etch electron microscopy, fluorescein-isothiocyanate-phalloidin staining, heavy meromyosin decoration, and immuno-electron microscopy. The subcortical actin filaments are distributed unevenly in the peripheral cytoplasm. Few secretory granules are seen beneath the plasma membrane in the region where the peripheral cytoplasm is occupied by numerous subcortical actin filaments. On the contrary, in the region free of the subcortical actin filaments, many secretory granules lie in contact with the plasma membrane. Thus, the subcortical actin filaments may control the approach of the secretory granules to the plasma membrane in these cells. The granule and plasma membranes that lie in close proximity are linked by intervening strands. Unfused portions of both membranes remain linked by these strands during membrane fusion and opening. These strands may be involved in membrane contact, fusion and opening during exocytosis. Annexin II (calpactin I) has been demonstrated immunocytochemically to be localized at the contact sites between the granule and plasma membranes, and is therefore a possible component of the intervening strands. Membrane fusion starts within focal regions of both membranes less than 50 nm in diameter. The plasma membrane shows inward depressions toward the underlying granules immediately before fusion. The disappearance of intramembranous particles from the exocytotic site of the membrane has not been observed.

Interaction of dynamin with microtubules: its structure and GTPase activity investigated by using highly purified dynamin.

We purified a large amount of dynamin with high enzymatical activity from rat brain tissue by a new procedure. Dynamin 0.48 mg was obtained from 20 g of rat brain. The purity of dynamin was almost 98%. Dynamin plays a role of GTPase rather than ATPase. In the absence of microtubules, Michaelis constant (Km) and maximum velocity (Vmax) for dynamin GTPase were 370 microM and 0.25 min-1, respectively, and in their presence, both were significantly accelerated up to 25 microM and 5.5 min-1. On the other hand, the ATPase activity was very low in the absence of microtubules, and even in their presence, Km and Vmax for dynamin ATPase were 0.2 mM and 0.91 min-1. Despite slow GTPase turnover rate in the absence of microtubules, binding of GTP and its nonhydrolizing analogues was very fast, indicating that GTP binding step is not rate limiting. Dynamin did not cause a one-directional consistent microtubule sliding movement just like kinesin or dynein in the presence of 2 mM ATP or 2 mM GTP. We observed the molecular structure of dynamin with low-angle rotary shadowing technique and revealed that the dynamin molecule is globular in shape. Gel filtration assay revealed that these globules were the oligomers of 100-kDa dynamin polypeptide. Dynamin bound to microtubules with a 1:1 approximately 1.2 molar ratio in the absence of GTP. Quick-freeze deep-etch electron microscopy of the dynamin-microtubule complex showed that dynamin decorates the surface of microtubules helically, like a screw bolt, very orderly and tightly with 11.4 +/- 0.9 (SD)nm period. Contrary to the previous report, microtubules make bundles by the attachment of the dynamin helixes around each adjacent microtubule, and no cross-bridge formation was observed.