Visualizing filamentous actin on lipid bilayers by atomic force microscopy in solution

The surface structure of actin filaments (F-actin) was visualized at high resolution, by atomic force microscopy (AFM) in aqueous solution, in large paracrystals prepared on positively charged lipid monolayers. The increased stability of these closely packed specimens allowed us to show that both the long pitch (38 nm) and the monomer (5.8 nm) can be directly resolved by AFM in the contact mode. The right-handed helical surface, distinguishable in high resolution images, was compared with reconstructed models based on electron microscopy. The height of the rafts, a measure of the actin filament diameter, was 10 ± 1 nm, whereas the smaller inter-filament distance, 8 ± 1 nm, was consistent with interdigitation of the filaments. The 10 ± 1 nm F-actin diameter is in good agreement with the results of fibre X-ray diffraction. As such specimens are relatively easy to prepare without specialized equipment, this method may allow the study of the thin filaments in which F-actin-associated proteins are also present.     

Form and distribution of actin and myosin in non-muscle cells: a study using cultured chick embryo fibroblasts.

Attempting to throw light on the mechanical basis of movement of non-muscle (cf. muscle) cells, the present work aims to determine the form and distribution of actin and myosin in chick embryo fibroblasts. These cells were cultured on formvar, fixed in glutaraldehyde then osmium tetroxide vapours, dehydrated, critical-point dried and examined, in toto, in the electron microscope (EM). Stereoscopic pairs of micrographs were studied to define more exactly the form and distribution of cytoplasmic filaments topographically associated with deformations of the cell surface and with organelle movements through the cytoplasm. Permeating the cytoplasm, interconnecting long and short filaments closely surrounded all organelles, linked with microtubules and polyribosomes and joined to the plasma membrane. These filaments, which varied greatly in width (2-13 nm) were closely associated with large numbers of 'comma-shaped' globoid bodies of approximately 15 nm diameter. Attempting to establish the identity, form and distribution of cytoplasmic myosin, cultured cells were extracted with a cold (4 degrees C) glycerol/pyrophosphate solution for 24 h before being fixed and critical-point dried. EM examination of these cells revealed a residual three-dimensional network of branching and anastomosing 4-13 nm diameter smooth filaments, devoid of fine (2 nm) filaments and globoid bodies. Examination of fixed, critical-point dried, skeletal muscle heavy meromyosin showed globoid structures similar in form and size to the globoid bodies found in cultures fibroblasts. Similarly fixed and critical-point dried paracrystals of actin, polymerized in the presence of Mg2+, appeared as branching interconnecting filaments which, in form and dimensions, resembled the network filaments observed in pyrophosphate-extracted cells. It is concluded that the pyrophosphate-extractable globoid bodies found in cultured fibroblasts represent monomers of myosin, that the broader filaments to which these attach represent actin in Mg2+ paracrystalline form and that the various subcellular movements are brought about by interactions between the two, analogous to those occurring in muscle cells.     

Improved transfer of two-dimensional crystals from the air/water interface to specimen support grids for high-resolution analysis by electron microscopy

Electron crystallographic analysis of two-dimensional crystals grown on lipid layers at the air/water interface has been limited by loss or damage during transfer of the crystals to an electron microscope support grid. Two methods of transfer are described which are applicable on a small scale (10 microliters of protein solution) and which give greatly improved results for streptavidin crystals on biotinylated lipid layers. In the first method, a hydrophobic grid surface was produced by coating a carbon support film with a thin layer of SiO2, followed by alkylation with dimethyloctadecylchlorosilane. The transfer efficiency of protein crystals approached 50% coverage of the alkylated grid surface. The degree of order of crystals transferred to the alkylated grid surface and preserved in negative stain was significantly improved over that of crystals transferred directly to a carbon support film. In the second method, crystals at the air/water interface were transferred to a holey carbon support film. The efficiency of transfer across the holes was virtually 100% as nearly every hole was completely covered with crystals. After preservation of the crystals in 1% glucose and cooling to liquid nitrogen temperature, electron diffraction was obtained that extended to 1/2.8 A-1 resolution. This demonstrates that two-dimensional crystals grown on lipid layers at the air/water interface can be sufficiently well-ordered, even after transfer to a support grid, to yield high-resolution structural information.     

The relative stability of a mercury stain at room temperature and liquid helium temperature

Observations of the localization of the mercurial stain TAMM [tetrakis (acetoxymercuri) methane] have been made at room temperature and liquid helium temperature. Mercurial stains do not maintain a fixed position at room temperature in the electron microscope. In order to record room temperature micrographs with the minimum possible dose, a special single-electron-counting electron microscope system was used. Micrographs were taken at liquid helium temperature using a prototype microscope with a superconducting electron objective lens. Tropomyosin paracrystals were chosen as a convenient test object. At room temperature the fine banding pattern of the paracrystals was barely resolvable or not resolvable at all. At liquid helium temperature the fine banding pattern of the paracrystals was easily observed and more than one micrograph could be taken before the banding pattern faded. These observations indicate that this mercurial stain is more stable in position at liquid helium temperature than at room temperature.     

Unravelling the structure of the lamellipodium

Summary Pushing at the cell front is the business of lamellipodia and understanding how lamellipodia function requires knowledge of their structural organization. Analysis of extracted, critical-point-dried cells by electron microscopy has led to a current dogma that the lamellipodium pushes as a branched array of actin filaments, with a branching angle of 70 degrees , defined by the Arp2/3 complex. Comparison of different preparative methods indicates that the critical-point-drying-replica technique introduces distortions into actin networks, such that crossing filaments may appear branched. After negative staining and from preliminary studies by cryo-electron tomography, no clear evidence could be found for actin filament branching in lamellipodia. From recent observations of a sub-class of actin speckles in lamellipodia that exhibit a dynamic behaviour similar to speckles in the lamella region behind, it has been proposed that the lamellipodium surfs on top of the lamella. Negative stain electron microscopy and cryo-electron microscopy of fixed cells, which reveal the entire complement of filaments in lamellipodia show, however, that there is no separate, second array of filaments beneath the lamellipodium network. From present data, we conclude that the lamellipodium is a distinct protrusive entity composed of a network of primarily unbranched actin filaments. Cryo-electron tomography of snap-frozen intact cells will be required to finally clarify the three-dimensional arrangement of actin filaments in lamellipodia in vivo.     

Filopodia formation induced by active mDia2/Drf3

Filopodia are rod-shaped cell surface protrusions composed of a parallel bundle of actin filaments. Since filopodia frequently emanate from lamellipodia, it has been proposed that they form exclusively by the convergence and elongation of actin filaments generated in lamellipodia networks. However, filopodia form without Arp2/3-complex, which is essential for lamellipodia formation, indicating that actin filaments in filopodia may be generated by other nucleators. Here we analyzed the effects of ectopic expression of GFP-tagged full length or a constitutively active variant of the human formin mDia2/Drf3. By contrast to the full-length molecule, which did not affect cell behaviour and was entirely cytosolic, active Drf3 lacking the C-terminal regulatory region (Drf3DeltaDAD) induced the formation of filopodia and accumulated at their tips. Low expression of Drf3DeltaDAD induced rod-shaped or tapered filopodia, whereas over-expression resulted in multiple, club-shaped filopodia. The clubs were filled with densely bundled actin filaments, whose number but not packing density decreased further away from the tip. Interestingly, clubs frequently increased in width after protrusion beyond the cell periphery, which correlated with increased amounts of Drf3DeltaDAD at their tips. These data suggest Drf3-induced filopodia form and extend by de novo nucleation of actin filaments instead of convergent elongation. Finally, Drf3DeltaDAD also induced the formation of unusual, lamellipodia-like structures, which contained both lamellipodial markers and the prominent filopodial protein fascin. Microarray analyses revealed highly variable Drf3 expression levels in different commonly used cell lines, reflecting the need for more detailed analyses of the functions of distinct formins in actin cytoskeleton turnover and different cell types.     

Membrane deformation of living glial cells using atomic force microscopy

Using atomic force microscopy (AFM) it has been possible to detect actin filaments that are beneath the cell membrane of living cells despite the fact that the AFM tip is applied to the surface of the cell. To determine whether the AFM tip actually penetrates or deforms the cell membrane we determined whether an intracellularly trapped fluorescent indicator was lost from cells during AFM. Using epi-fluorescence illumination to monitor the presence of fluo-3 in the cell, we found that AFM did not cause dye leakage from the cell. Further, force–distance curves indicated that standard tips did not penetrate the membrane while sharper Supertips^TM did. In addition, the physiology of cells was found to be unaffected by AFM with standard tips since volume regulatory signal transduction mechanisms were intact in such studies. Thus, traditional AFM tips deform the cell membrane in order to reveal the presence of subcellular structures.     

Paired helical filaments (PHFs) are a family of single filament structures with a common helical turn period: negatively stained PHF imaged by TEM and measured before and after sonication, deglycosylation, and dephosphorylation

Isolated paired helical filaments (PHFs) were visualized on glutaraldehyde vapor-treated thin approximately 10-nm thick indirect carbon films using high-resolution transmission electron microscopy (TEM) and the negative stain, phosphotungstate acid (PTA) at near neutral pH of 6.8. PHF preparations were prepared with and without 1 minute of sonication. These same PHF were also deglycosylated with endoglycosidase F/N-glycosidase F for 1 hour or the PHF were dephosphorylated with PP-2A for 1 hour. The negatively stained PHF filaments were quantitatively studied by measuring their wide regions (W) their thin regions (T) and their helical turn period (L) and these separate parameters were averaged for each filament. In the unsonicated PHF preparation there were PHF, cylindrical filaments with periodic thin regions (CF-PT), cylindrical filaments (CF), as well as 2.0-nm tau polymer-like filaments. The CF-PT were characterized by W, T, and L measurements and the CF were characterized by diameter measurements. The paired helical filament model proposed by Kidd (1963, Nature 197:192-193) of two approximately 10 nm filaments twisting around each other every approximately 80 nm with a thin region of 10 nm and a wide region of 25 nm does not correspond to the PHF structures found. None of the PHF we observed were composed of a pair of filaments and all of the PHF appear to be a single filament. The wide regions ranged from 12.5-27 nm and the thin regions ranged from 4.5-12.3 nm. The helical turn periods ranged from 76-85 nm and were generally about 80 nm. Only the helical turn period of approximately 80 nm was a common property of the whole family of PHF structures. The CF-PT appear to be a PHF precursor filament. Deglycosylation of the PHF and CF-PT reduced their sizes by 0.5-0.6 nm and 0.7-1.0 nm, respectively, and the right-hand helicity of the PHF was lost after deglycosylation. Dephosphorylation with PP-2A reduced the PHF wide regions by 6.0 nm and the thin regions by 2.6 nm.     

Mechanism of the process formation; podocytes vs. neurons

In this review article we discuss the common mechanism for cellular process formation. Besides the podocyte, the mechanism of process formation, including cytoskeletal organization and signal transduction, etc., has been studied using neurons and glias as model systems. There has been an accumulation of data showing common cell biological features of the podocyte and the neuron: 1) Both cells possess long and short cell processes equipped with highly organized cytoskeletal systems; 2) Both show cytoskeletal segregation; microtubules (MTs) and intermediate filaments (IFs) in podocyte primary processes and in neurites, while actin filaments (AFs) are abundant in podocyte foot processes in neuronal synaptic regions; 3) In both cells, process formation is mechanically dependent on MTs, whose assembly is regulated by various microtubule- associated proteins (MAPs); 4) In both cells, process formation is positively regulated by PP2A, a Ser/Thr protein phosphatase; 5) In both cells, process formation is accelerated by laminin, an extracellular matrix protein. In addition, recent data from our and other laboratories have shown that podocyte processes share many features with neuronal dendrites: 1) Podocyte processes and neuronal dendrites possess MTs with mixed polarity, namely, plus-end-distal and minus-end-distal MTs coexist in these processes; 2) To establish the mixed polarity of MTs, both express CHO1/MKLP1, a kinesin-related motor protein, and when its expression is inhibited formation of both podocyte processes and neuronal dendrites is abolished; 3) The elongation of both podocyte processes and neuronal dendrites is supported by rab8-regulated basolateral-type membrane transport; 4) Both podocyte processes and neuronal dendrites express synaptopodin, an actin-associated protein, in a development-dependent manner; interestingly, in both cells, synaptopodin is localized not in the main shaft of processes but in thin short projections from the main shaft. We propose that the podocyte process and the neuronal dendrite share many features, while the neuronal axon should be thought of as an exceptionally differentiated cellular process.     

Electron phase and amplitude images of stained biological thin sections

Very-high-resolution electron microscopy demonstrates similar phase image ‘particles’ in uncontaminated, unsupported thin sections of epoxy-embedded biological material and in simple thin carbon films, both of which show better than 5 Å resolution. Under such conditions it is possible to examine biological specimens for amplitude images well below the 20 Å limit commonly accepted for such preparations. With this very high resolution discrete amplitude images are found in sections after treatment with heavy-metal stain. These amplitude images are as small as 10 Å, a size equal to that of lead aggregates, which they may represent.     

Lipid losses during processing of cardiac muscle for electron microscopy

The lipid retaining properties of several methods of processing tissue for electron microscopy (EM) have been assessed quantitatively. Guinea-pig hearts were perfused in vitro at 37°C with ³H-oleic acid bound to albumin. The hearts were fixed by perfusion with 4% glutaraldehyde in 0.1 M cacodylate buffer. Pieces of left ventricle and interventricular septum were removed., weighed and processed for EM. The fluids used at each stage of processing were monitored for loss of radioactive lipid by scintillation counting. Lipids were extracted from the processed tissue immediately before the embedding stage using a mixture of chloroform: methanol (2:1 v/v). Counts from processed tissue were compared with counts from tissue extracted directly after perfusion fixation in order to monitor subsequent losses during processing. A modified version of Epon processing, omitting 100% ethanol, acetone or propylene oxide, gave a lipid retention of only 20.6%. The addition of para-phenylenediamine to the procedure did not improve the retention although this has been shown to be a useful stain for intracellular lipid. Water soluble Durcupan which does not involve ethanol or acetone dehydration has an average retention of 63% with 100% recovery while the lesser known polymer GACH, a mixture of glutaraldehyde and carbohydrazide used both for dehydration and embedding showed a lipid retention of 82% of the counts recovered although recovery was only 69%. An attempt was made to determine which classes of lipids were present in the tissue after perfusion fixation using thin layer chromatography. It was found that the presence of any of the processing fluids affected the polarity of the lipids and their rates of migration on thin layer plates.     

Reconstruction of fluid Langmuir monolayers under shear forces

Scanning probe microscopy observations of monolayers of a classical boundary lubricant, stearic acid (STA), reveal long-range dynamics of wear and reconstruction of monomolecular films under the shear forces caused by the sliding tip. In the range of sliding velocities studied, the friction forces behave non-monotonically with a maximum value around 0.2 m/s. The STA monolayer in a fluid state displays a flow of material from the worn area and its redistribution to form local heterogeneous multilayers. The range of influence of the material removal on the monolayer structure around the worn area is up to 80 m, or more than an order of magnitude larger than the actual worn area. Surface diffusion of mobile organic material is responsible for the observed long-range spreading of local shear stresses produced in the area of wear. It demonstrates the non-local nature of the stress-induced reconstruction in the fluid monolayers, in contrast with solid monolayers, which show very localized wearing processes.     

On the preparation of specimens from multifilamentary Nb3Sn superconducting wires for transmission electron microscopy

A set of techniques for preparing transmission electron microscope specimens from sections of thin (≤ 1 mm overall diam.) multifilamentary wires (filaments being approximately 5 μm diam.) are described. Results illustrating the successful uniform thinning of the several phases present in ‘bronze-route’ Nb₃Sn are shown.     

The projection of a negatively-stained filamentous object down its central axis as revealed by image reconstruction from tilt series

Helical reconstructions of negatively stained biological objects contain distortions arising from filament flattening and the non-cylindrical profile of the stain envelope. Current methods of estimating flattening do not make full use of the information available in electron micrographs. We have applied the more rigorous approach of reconstructing the density profile from tilt series using digital image processing techniques. Tilt series were collected for tobacco mosaic viruses (TMV) and seven independent reconstructions were calculated using equatorial data out as far as ～1/9·3 nm^—1. They indicated that the filaments were flattened with an axial ratio of about 2·4:1, which was probably closer to 3:1 in the original specimen, because the limited resolution caused flattening to be underestimated. The stain envelope around TMV and some indication of the underlying carbon substrate were also observed. This information could enable correction factors for flattening to be developed, which could be useful when calculating helical reconstructions or indexing helical diffraction patterns. This method could also be extended to non-equatorial layer planes, which would provide three-dimensional information on a wide range of macromolecules that possess a one-dimensional repeat.     

A technique for exposure of the glycoproteic matrix (zona pellucida and mucus) for scanning electron microscopy.

The fine structure of the zona pellucida (ZP) covering the oocyte and of the mucus covering the surface of the intestinal villi was investigated by using a new method employing ruthenium red (RR), saponin, and osmium-thiocarbohydrazide impregnation. The glycoproteic matrices both appeared constituted by thin filaments (ranging from 22 to 50 nm in thickness) anastomosed to form a very fine network. RR prevented the dissolution and/or alteration of glycoproteins and polyanionic carbohydrates induced by acqueous fixatives. Saponin was a detergent of the soluble proteins. Osmium-thiocarbohydrazide preserved the glycoproteic matrix filaments from the mechanical stress induced by dehydration and critical point drying and reduced filaments packing and shrinkage. The technical improvement was demonstrated by the following results: 1) a regular arrangement of the filaments network; 2) a thickness of mucus filaments smaller than that obtained with other methods of preparation; 3) a homogeneous thickness of ZP filaments. This method allowed a very detailed study of the fine structural organization of the ZP and intestinal mucus. Therefore, this technique can be useful for a better evaluation of the morphodynamic of these and other glycoproteic matrices.     

Intercellular junctions in embryonic chick cardiac muscle revealed by rapid freezing and freeze-substitution.

Using the method of rapid freezing and freeze-substitution, the embryonic chick cardiac muscle was investigated by transmission electron microscopy. Initially, the intercellular junctional complexes (fasciae adherentes and desmosomes) were formed in close proximity to each other along a nearly straight line. Subsequently, the separation of fasciae from desmosomes took place to form intercalated discs. The cell membranes of fasciae adherentes were reinforced with highly interwoven fine fibrils at which myofibrils terminated. The intercellular space of fasciae was bridged with fine fibrillar structures seemingly connected by a thin line at their middle portions. In the intercellular space of desmosomes, central lamina and traversing filaments were clearly observed. The outer and inner leaflets of the desmosomal plasmalemma were asymmetrically differentiated; the outer leaflet was thinner than the inner leaflet. On the inner side of the cell membrane, an electron-lucent layer and a dense desmosomal plaque were observed. The latter structure had protrusions with less electron density towards the cytoplasmic side. Further inside, a meshwork of fine fibrils was seen along and toward which bundles of intermediate filaments ran. The results obtained with freeze-substitution appeared to provide more information than those with thin sections after conventional fixation or with replicas of chemically fixed/glycerinated or physically fixed/deep-etched materials.     

Whole-mount preparations of mouse lens epithelium for the fluorescent cytological study of actin

A technique is given for the preparation of a sheet of epithelial cells from the capsule of the crystalline lens. A new method is described for fixation and staining with fluorescent phalloidin or actin antibody in order to localize the actin cytoskeleton in this tissue. Optical section of the preparation resolves such actin features as apical polygonal arrays, sequestered actin bundles, perinuclear actin aggregates, observed here for the first time, and filamentous networks in the basal region of the cell. This method is superior to previous ones in its ability to preserve actin-abundant sectors distinctively.     

Diffraction patterns from stained and unstained helices: Consistency or contradiction?

Addressing the issue of how best to integrate negative staining electron microscopy with X-ray fiber diffraction of macromolecular helices, we investigated staining properties of helical models using computational simulations. We compared diffraction patterns of stained and unstained representations of F-actin, myosin S1-decorated actin, and microtubules, to test the reliability of selecting well preserved specimens as those whose optical diffraction patterns most closely match the X-ray patterns of hydrated fibers. We conclude that if the stain layer is considerably thicker than the outer diameter of the specimen, this criterion is likely to be reliable. However, specimens in such thick stain layers may be resolution-limited by dynamic scattering effects such as beam-broadening. With relatively thin staining (the more practically relevant situation) "edge effects" in the stain distribution result in differences between diffraction patterns of stained and unstained specimens. Diffraction pattern changes due to molecular distortions such as shrinkage or flattening were similarly modeled. Since differences between diffraction patterns of stained and unstained helices may be due to either "edge effects" or molecular distortions, it does not appear possible a priori to distinguish between these effects. Comparison of experimental data for the (2.3 nm)-1 layer-line of tobacco mosaic virus (as model system) reveals major differences between the X-ray diffraction pattern of hydrated sols and the Fourier transform of HREM images of negatively stained specimens.     

Tau binds ATP and induces its aggregation

Tau is a microtubule‐associated protein mainly found in neurons. The protein is associated with process of microtubule assembly, which plays an important role in intracellular transport and cell structure of the neuron. Tauopathies are a group of neurodegenerative diseases specifically associated with tau abnormalities. While a well‐defined mechanism remains unknown, most facts point to tau as a prominent culprit in neurodegeneration. In most cases of Tauopathies, aggregates of hyperphosphorylated tau have been found. Two proposals are present when discussing tau toxicity, one being the aggregation of tau proteins and the other points toward a conformational change within the protein. Previous work we carried out showed tau hyperphosphorylation promotes tau to behave abnormally resulting in microtubule assembly disruption as well as a breakdown in tau self‐assembly. We found that tau's N‐terminal region has a putative site for ATP/GTP binding. In this paper we demonstrate that tau is able to bind ATP and not GTP, that this binding induces tau self‐assembly into filaments. At 1 mM ATP the filaments are 4–7 nm in width, whereas at 10 mM ATP the filaments appeared to establish lateral interaction, bundling and twisting, forming filaments that resembled the Paired Helical Filaments (PHF) isolated from Alzheimer disease brain. ATP‐induced self‐assembly is not energy dependent because the nonhydrolysable analogue of the ATP induces the same assembly. Microsc. Res. Tech. 77:133–137, 2014. © 2013 Wiley Periodicals, Inc.     

High‐resolution cryo‐SEM allows direct identification of F‐actin at the inner nuclear membrane of Xenopus oocytes by virtue of its structural features

The nuclear envelope of Xenopus laevis stage VI oocytes was studied in a high‐resolution field emission cryo‐scanning electron microscope to compare the level of structural preservation obtainable by different procedures of specimen preparation. All approaches generally allowed frequent detection of long filaments of about 10 nm in diameter that were attached to the nuclear envelope's inner membrane facing the nuclear interior. Structural details of these 10‐nm filaments, however, could not be unveiled by standard procedures of specimen preparation and analysis, including critical point drying and imaging at room temperature. In contrast, after freeze‐drying and imaging at ?100°C, the 10‐nm filament type was found to be composed of distinct globular subunits of approximately 5 nm in diameter that were arranged in a helical manner with right‐handed periodicity. Stereoscopic images showed that some of these filaments were lying directly on the membrane whereas others appeared to hover at a certain distance above the nuclear envelope. The appearance of these filaments was highly similar to that of in vitro polymerized F‐actin analysed in parallel, and closely resembled the structural characteristics of F‐actin filaments described earlier. By virtue of their structural features we therefore conclude that these filaments at the nuclear periphery represent F‐actin. The high level of structural resolution obtainable by field emission cryo‐SEM illustrates the potential of this method for studying details of biological structures in a subcellular context.