Vertical Pt-C replication for TEM, a revolution in imaging non-periodic macromolecules, biological gels and low-density polymer networks.

Vertical replication for TEM is ideal for studying non-periodic specimens from 0.7 to 3 nm, a resolution mid-range difficult to attain by any other technique. This paper discusses the importance of vertical replication, its methods and hardware for high-resolution TEM. Evidence from diverse published research will demonstrate vertical replication's versatility in imaging the molecular level normally unattainable in freeze-dried polymers, polyethylene tribological wear on surfaces, low-density polymer networks or biological gels. Vertical platinum-carbon (Pt-C) replication minimizes the horizontal movement of Pt-C on a surface. Surface objects are symmetrically enlarged by a vertically deposited Pt-C film. To estimate real size in replicas, 16-25 particles or filaments need to be measured in calibrated transmission electron microscopy (TEM) images and reduced by a value less than the Pt-C film thickness measured with a quartz monitor. Continuous, vertically deposited Pt-C films are formed on mica at a deposition thickness of around 1.0 nm and on silver at a thickness of 0.4-0.5 nm. The distance between helical turns in poly(1-tetradecene sulfone) of 0.7 nm is the highest resolution achieved with vertical replication. Two polysulfones freeze-dried and vertically replicated on mica contained structures are predicted by indirect physical chemical methods to be present in solution. Polymer chains are fully Pt-C coated, with no uncoated gaps along chains. Some side-chains on the extended non-helical poly(1-tetradecene sulfone) are also detected. To estimate the real chain width, polymer chains measured in images are reduced by the Pt-C film thickness minus 0.5 nm. The polymer chain widths estimated from molecular models are in the same range of widths as those measured using the image size correction method. Also, it is possible to distinguish random coil proteins (chain width of around 0.5 nm) from an alpha-helix (chain diameter of about 1 nm) in vertically replicated samples on silver substrates. In the future, subnanometer resolutions below 0.7 nm should be possible. The resolution of vertical replication depends on the thickness of a continuous, amorphous Pt-C film. That thin, continuous 0.4-0.5 nm Pt-C films on silver substrates can be made suggests that a point-to-point resolution limit of around 0.28 nm in TEM may ultimately be approachable with replication.

DNA bound to polypyrrole films: high-resolution imaging, DNA binding kinetics and internal migration.

We have studied the time-dependent uptake of 35S radiolabeled DNA with electrochemically prepared polypyrrole films. The two distinct polypyrrole film surfaces, a rough (solution polymeric growth face, R) and a smooth surface (electrode face, S) were characterized by low-resolution AFM and high-resolution transmission electron microscopy (TEM). These studies showed the presence of steep contours and defects in the form of large and small surface holes and valleys on the rough surface of polypyrrole. The void dimensions ranged from the nanoscale to micron size. By contrast, the smooth surface was flatter and largely devoid of significant structural defects and exhibited closer packing of the polypyrrole chains over large areas. Both surfaces were comprised largely of chains whose average diameters were 1.0-1.2 +/- 0.3 nm. The surface characterization studies were complemented by time-dependent DNA uptake studies which showed a t1/2-dependent total uptake of 35S DNA at higher levels on the rough surface compared to the smooth surface. This is consistent with the apparent higher effective surface area of the rough surface compared to the smooth. Using a proportional counter the time-dependent ratio (R/S) of the 35S DNA detected from the rough surface of the polypyrrole disk to that detected from the smooth surface suggested that DNA was migrating into the disk interior from its uptake surface. The rough side defect dimensions measured by TEM were more than sufficient to allow for the penetration and migration of DNA into the disk interior. Both R/S ratios were extrapolated and found to intersect at an R/S value close to 1.0, suggesting a kinetic process leading ultimately towards a nearly uniform radiolabeled DNA distribution in the disk. These kinetic results were in agreement with the surface characterization studies and suggest a model in which sizeable internal pores exist throughout the electrochemically prepared polypyrrole, that could account for the DNA migration effect. This was confirmed by TEM of the interior of a polypyrrole disk produced by Argon ion milling.

Cell-free DNA in human blood plasma: length measurements in patients with pancreatic cancer and healthy controls.

The amount of non-cell-associated DNA free in blood plasma from pancreatic cancer patients usually exceeds that from healthy donors. We have evaluated the plasma DNA by gel electrophoresis and measured the variation in length of soluble DNA fragments by electron microscopy in plasma from three patients with pancreatic cancer and from three healthy controls. Whereas electrophoresis of nick-translated DNA isolated from plasma obtained from healthy controls showed autoradiographic bands at sizes equivalent to whole-number multiples (1-5x) of nucleosomal DNA (185-200 bp), in the samples obtained from pancreatic cancer patients, stronger ladder patterns appeared. Likewise, strand length distributions of DNA (DNA-SL) in the two groups differ. The DNA-SL distribution data include 2,752 measurements made from cancer patient plasma and 3,291 for control plasma. The shortest DNA-SL measured approximately 30 nm (approximately 88 bp calculated at 0.34 nm/bp) and the largest approximately 28,000 nm (>80,000 bp), with 50% of all lengths measuring between 100 and 900 nm long. The average plasma DNA-SL in controls (311 nm; median, 273 nm) exceeded that in cancer patients (231 nm; median, 185 nm). Small excesses of DNA at approximately 63, approximately 126, approximately 189, approximately 252, and approximately 315 nm, corresponding to small multiples of lengths associated with nucleosomes, were more prominent in the cancer patient plasma than in the healthy control plasma. This study provides evidence indicating differences in non-cell-associated DNA in plasma between cancer patients and healthy controls and indicates that a significant amount of this DNA is probably derived from apoptosis in neoplastic and/or normal cells.

Alzheimer disease hyperphosphorylated tau aggregates hydrophobically.

The chemical interaction that condenses the hyperphosphorylated protein tau in Alzheimer's disease (AD P-tau) into neurofibrillary tangles and cripples synaptic transmission remains unknown. Only beta-sheet, positive ion salt bridges between phosphates, and hydrophobic association can create tangles of just AD P-tau. We have correlated transmission electron microscope (TEM) images of tau aggregation with different percentages of beta-sheet in aqueous suspensions of tau while using buffers that block dispositive or tripositive ionic bridges between intermolecular phosphates. Circular dichroism (CD) studies were performed at different temperatures from 5-85 degrees C using AD P-tau, AD P-tau dephosphorylated with hydrofluoric acid (HF AD P-tau) or alkaline phosphatase (AP AD P-tau), and recombinant human tau with 3-repeats and two amino terminal inserts (R-39) and using bovine tau (B tau) isolated without heat or acid treatment. Secondary structure was estimated from CD spectra at 5 degrees C using the Lincomb algorithm. Each preparation except one demonstrated an inverse temperature transition, Ti, in the CD at 197 nm. No correlation was found between beta-sheet content and aggregation, leaving only hydrophobic interaction as the remaining possibility. Thirteen of 21 possible phosphorylation sites in AD P-tau lie adjacent to positive residues in tau's primary structure. Occupation of five to nine phosphate sites on AD P-tau appears sufficient to reduce or neutralize tau's basic character. AD P-tau's hydrophobic character is indicated by its low inverse temperature transition, Ti. The Ti for AD P-tau was 24.5 degrees C or 28 degrees C, whereas for B tau with three phosphates it was 32 degrees C, for unphosphorylated tau R-39 it was 38 degrees C, and for dephosphorylated HF AD P-tau it was 37.5 degrees C. The hydrophobic protein elastin and its analogs coalesce and precipitate at their Ti of 24-29 degrees C, well below body temperature. We hypothesize that AD P-tau causes tangle accumulation by this mechanism.

Conformation of Lac repressor tetramer in solution, bound and unbound to operator DNA.

We tested whether the Steitz et al. [(1974) Proc. Natl. Acad. Sci. U.S.A., 71:593-597] model of lactose repressor (LacR) (14 x 6.0 x 4.5 nm) represented the shape of free or operator-bound LacR in solution. The model predicts a 14 nm length for bound LacR. Direct measurement, using Pt-C shadow width standards, was 9.6 +/- 0.2 nm long. Using the Steitz model, we generated a distribution of measurements and converted them into a distribution of shadow widths using gold ball standards. Direct measurement of LacR produced a narrower shadow width distribution with a larger mean size than the Steitz model predicted. Measurement along two orthogonal axes of negatively stained LacR images generated a size distribution, also converted into a shadow width distribution using the gold ball standards. Since the experimental shadow width distribution exactly matched the shadow width distribution derived theoretically from negatively stained LacR, our negative-stained images are representative of LacR's conformation in solution. Approximately 56% of negatively stained LacR had a V-shaped fold around an axis orthogonal to its length, bringing the DNA binding domains of each dimer adjacent. This open end of the V binds single operator DNA. The other 44% of the LacR tetramer is in the extended form with its DNA binding sites at opposite ends. Although the V-shaped conformation has a closed hinge with the dimers associated along a side, the extended open-hinged state remains important since LacR must bind two distant operator sites for full repression. Our measurements predict the normal presence of both conformations in nearly equal amounts, suggesting that both are equally active in repressing the lac operon.

Quantification of particle sizes with metal replication under standard freeze-etching conditions: a gold ball standard for calibrating shadow widths was used to measure freeze-etched globular proteins.

The real size of platinum-carbon (Pt-C) replicated particles is not directly equivalent to either its metal-coated diameter or its shadow width. This paper describes two indirect methods, shadow widths and coated particle diameters, for determining a particle's actual size beneath a Pt-C replication film. Both produce equivalent measurements using the same standardized conditions: 2.3 nm Pt-C films deposited at a 45 degree angle on an approximately -100 degrees C surface in a 10(-6) torr vacuum. For the first method, gold balls nucleated in a partial pressure of helium and deposited on flat indirect carbon films (root mean square roughness of 0.8 nm) on 400 mesh grids were used as test particles for calibrating shadow widths as a function of particle size. The gold ball test specimens were replicated, and a distribution of Pt-C shadow widths orthogonal to the Pt-C deposition direction was measured and averaged for gold balls 1.5 +/- 0.25 nm, 2.0 +/- 0.25 nm, etc. The diameter of each gold ball was measured within the Pt-C film along with its shadow width because the Pt-C did not obscure or adhere well to the gold. The shadow width distributions for each gold size do not differ significantly from log normal. Two proteins, the lactose repressor and the mitochondrial ATPase, F1, were also used as replication test objects. Negative staining of both proteins was conducted to measure their average diameters. In the second method, a distribution of Pt-C-coated lac repressor diameters perpendicular to the shadow direction was measured. The Pt-C film thickness measured on the quartz crystal monitor was subtracted from the average metal-coated protein diameter to obtain the lac repressor's diameter. The Pt-C-coated particle diameter distributions also did not differ significantly from log normal. While doing this work it was discovered that outgassing the Pt-C electron gun greatly affected Pt-C film granularity: 19 sec produced a high contrast, granular Pt-C film, whereas 120 sec yielded a low contrast, less granular Pt-C film. Both gold balls and protein particles were subjected in separate experiments to either 19 or 120 sec of outgassing of the Pt-C gun prior to Pt-C replication. Outgassing had a profound effect on the average size of the Pt-C shadow widths on both gold and protein particles. The Pt-C gun outgassing procedure also determined the smallest replicated particle that could be resolved. The frequency of some smaller gold ball sizes detected after replication was reduced disproportionately with 19 sec vs. 120 sec outgassing. However, Pt-C gun outgassing did not affect the average measured diameter of the Pt-C-coated protein particles. The "geometric assumption" that each metal-coated particle creates a shadow width the same size as the metal-coated particle diameter was tested using a globular protein. Pt-C replication of protein particles at a 45 degree and 20 degree angle could not confirm the geometric assumption because an average shadow width was always significantly larger than its average Pt-C-coated particle diameter. A model for how the large shadow widths are formed is presented. Gold balls were also replicated at a 45 degree angle with current high resolution conditions at a substrate temperature of -185 degrees C, and the results of these replicas were compared to the results reported here at approximately -100 degrees C.

Alzheimer paired helical filaments, untreated and pronase digested, studied by vertical platinum-carbon replication and high resolution transmission electron microscopy.

Untreated paired helical filaments (PHF) and pronase treated PHF filaments have been stereoscopically imaged with a freeze-drying vertical platinum-carbon replication preparation method for TEM. The untreated PHF have an average wide region, W = 22.8 +/- 2.4 nm, a narrow region width, T = 10.6 +/- 1.7 nm, and a helical turn period, L = 78.6 +/- 13.4. The widths of the pronase treated PHF were significantly reduced and had average measurements of W = 14.8 +/- 1.2 nm, T = 5.7 +/- 1.0 nm, with the helical period unchanged, L = 75.4 +/- 17 nm. The surfaces of the untreated PHF contained approximately 1.0 and approximately 0.4 nm strands, the size of normal and denatured tau monomer. The pronase treated PHF contained approximately 1.0 and approximately 0.4 nm strands as well as approximately 2.0 nm strands. The stereoscopic images of the untreated and the pronase digested PHF do not support a double helical morphology for the PHF. The PHF appear to be long helical ribbons. The approximately 1.0 and approximately 2 nm substructure has been organized both parallel and orthogonal to the PHF-core axis for distances less than 80 nm. The most frequent structural appearance is of a disorganized PHF core. The surfaces of the untreated PHF also have a similar disorganized appearance.

High resolution platinum-carbon replication of freeze-dried basement membrane.

High angle platinum/carbon (Pt/C) replication has proved to be a valuable tool in analyzing basement membrane structure in human amnion, bovine lens capsule, and the Engelbreth-Holm-Swarm (EHS) tumor. High resolution replicas for transmission electron microscopy (TEM) have been achieved by depositing 1.0 +/- 0.1 nm thick Pt/C films backed with rotary deposited 12.5 +/- 2.5 nm thick carbon films. The basement membrane collagen IV network was observed to consist of fine branching filaments containing globular domains intrinsic to the filaments. A second quasi-regular network is formed by laminin. Unidirectional 45 degree angle Pt/C replication was used for most of this work. The merits and deficiencies of unidirectional vertical replication (80 degree angle), unidirectional 45 degree angle, and 20 degrees low angle rotary replication are discussed. Vertical replication produces the highest resolution replicas and has the potential for revealing the overall pattern of basement membrane structural assembly if basement membrane preparations freeze-dried in low salt can faithfully maintain their in vivo structure.

Secondary structure of proteins associated in thin films.

The solid state secondary structure of myoglobin, RNase A, concanavalin A (Con A), poly(L-lysine), and two linear heterooligomeric peptides were examined by both far-uv CD spectroscopy1 and by ir spectroscopy. The proteins associated from water solution on glass and mica surfaces into noncrystalline, amorphous films, as judged by transmission electron microscopy of carbon-platinum replicas of surface and cross-fractured layer. The association into the solid state induced insignificant changes in the amide CD spectra of all alpha-helical myoglobin, decreased the molar ellipticity of the alpha/beta RNase A, and increased the molar ellipticity of all-beta Con A with no change in the positions of the bands' maxima. High-temperature exposure of the films induced permanent changes in the conformation of all proteins, resulting in less alpha-helix and more beta-sheet structure. The results suggest that the protein alpha-helices are less stable in films and that the secondary structure may rearrange into beta-sheets at high temperature. Two heterooligomeric peptides and poly(L-lysine), all in solution at neutral pH with "random coil" conformation, formed films with variable degrees of their secondary structure in beta-sheets or beta-turns. The result corresponded to the protein-derived Chou-Fasman amino acid propensities, and depended on both temperature and solvent used. The ir and CD spectra correlations of the peptides in the solid state indicate that the CD spectrum of a "random" structure in films differs from random coil in solution. Formic acid treatment transformed the secondary structure of the protein and peptide films into a stable alpha-helix or beta-sheet conformations. The results indicate that the proteins aggregate into a noncrystalline, glass-like state with preserved secondary structure. The solid state secondary structure may undergo further irreversible transformations induced by heat or solvent.

Alzheimer neurofibrillary tangles contain 2.1 nm filaments structurally identical to the microtubule-associated protein tau: a high-resolution transmission electron microscope study of tangles and senile plaque core amyloid.

Alzheimer neurofibrillary tangles (NFT) and senile plaque core amyloid (SPCA) isolated from the brain of patients with Alzheimer's disease were freeze-dried and replicated with a new platinum-carbon (Pt-C) vertical deposition method for high-resolution transmission electron microscopy (TEM). The resolution of this vertical Pt-C replication method is superior to either of the more conventional 20 degrees rotary replication or 45 degrees unidirectional replication methods and is dependent on the Pt-C film thickness coating the specimen. The paired helical filaments (PHF) observed within the tangles were right-handed helices with a fairly regular twist period averaging 79.3 +/- 5.9 nm and a fairly regular maximum width averaging 14.9 +/- 1.0 nm. The PHF regions of minimum width were not regular and fell into three size categories: 2.4 +/- 0.3 nm, 4.9 +/- 0.6 nm and 9.6 +/- 1.4 nm. In addition to the PHF found in the tangles, a new filament was found within all the tangles. These 2.1 +/- 0.2 nm diameter filaments were triple-stranded left helices with 1.0 +/- 0.2 nm diameter strands with a structure identical to bovine tau. Like bovine tau polymer a number of filaments (130 nm to 238 nm) were longer than a fully stretched tau monomer of 96 nm. Images of neuritic senile plaque core amyloid (SPCA) showed that amyloid had a more solid appearance than the NFT and its branched filament structures were unlike the approximately 2.1 nm diameter filaments or the PHF found in NFT.

The organization of the microtubule associated protein tau in Alzheimer paired helical filaments.

The structural relationship of the microtubule associated protein tau to paired helical filaments (PHF) was examined by high resolution transmission electron microscopy (TEM) without treatment with any chemical fixatives. Neurofibrillary tangles (NFT) were isolated in the absence of detergent from Alzheimer diseased brains, were freeze-dried, and were vertically platinum-carbon replicated for TEM. The PHF we observed made one helical turn (L) in 74 +/- 8.5 nm and had a wide region (W) of 14.8 +/- 0.6 nm similar to PHF previously modeled with a periodic morphology. The PHF thin region (T) measured approximately 2.4 nm, approximately 4.9 nm, approximately 7.4 nm and approximately 9.7 nm and the most often observed width was approximately 2.4 nm. No surface features regularly divide the PHF into two filaments. Morphologically the PHF are thin helical ribbons with an often observed thickness of approximately 2.4 nm. At high magnification, approximately 1.0 nm and some approximately 0.4 nm strands identical to normal and denatured tau monomer covered PHF surfaces and were aggregated in non-periodic fashion. Bovine tau polymer approximately 2.1 nm diameter filaments, trapped on a filter, were partially heat denatured and showed some of the morphological features of PHF.

Evidence for helical structures in poly(1-olefin sulfones) by transmission electron microscopy.

Transmission electron microscope images were obtained of fractions of poly(1-tetradecene sulfone) and poly(cyclohexene sulfone) cast from very dilute solutions (0.007%, wt/vol) and rapidly freeze-dried on a mica surface. The samples were then vertically platinum-carbon (Pt-C) replicated with 9 +/- 0.3-A Pt-C and held together with 128 A of electron-transparent evaporated carbon. The Pt-C coating enlarges the molecular chain diameters by approximately 5 A, so that a single polysulfone chain has an apparent diameter of 9-12 A in the transmission electron microscope. Poly(1-tetradecene sulfone) forms short helical regions that show irregular helical turns of pitch 7-18 A, two to eight turns long with apparent helix diameters of 16-22 A. The n-C12H25 side chains were not obvious on the helices, but a few were visible on extended chain sections. In contrast, poly(cyclohexene sulfone) appears as a collection of Pt-C-coated single chains 10-12 A in diameter with no side-chain projections. The presence and absence of helical regions in these freeze-dried polysulfones thus reflects their solution conformations as long ago inferred on the basis of dielectric measurements.

The microtubule-associated protein tau forms a triple-stranded left-hand helical polymer.

High resolution transmission electron microscopy (TEM) has shown that bovine tau are 2.1 +/- 0.2-nm diameter filaments which are triple-stranded left-hand helical structures composed of three 1.0 +/- 0.2-nm strands. The reported amino acid sequence of human and bovine tau have been computer processed to predict secondary structure. Within the constraints imposed by the images, the secondary structure models and other structural information have been used to calculate tau's maximum and minimum length. The length calculations and secondary structure form the basis for image interpretation. This work indicates that each approximately 1.0-nm strand is a tau polypeptide chain and that the approximately 2.1-nm filament is composed of three separate tau chains (tau3). Bovine tau length measurements indicate that tau trimer filaments are generally longer than a fully extended tau monomer. These measurements indicate that each trimer, tau3, is joined with other trimers to form long tau polymers, (tau3)n. An inverse temperature transition has been found in the circular dichroism spectrum of tau indicating that its structure is less ordered below 20 degrees C and more ordered at 37 degrees C. The implications of this phenomenon with respect to tau's temperature-dependent ability to reconstitute microtubules is discussed and a mechanism for the possible abnormal aggregation of tau into neurofibrillary tangles in Alzheimer's disease is proposed.

Solution structure of bacteriophage T4D and icosahedral capsid geometry visualized in freeze-fractured, deep-etched replicas.

The prolate icosahedral capsid geometry of wild type bacteriophage T4D has been determined by direct visualization of the triangular faces in stereoimages of transmission electron micrographs of phage particles. Bacteriophage T4 was prepared for transmission electron microscopy (TEM) following a protocol of freeze-fracturing, deep-etching (FDET) and replication by vertical deposition (80 degrees angle) of a thin platinum-carbon (Pt-C) metal layer of 1.01 nm. From direct statistical measurements of the ratio of the head length to width and of stereometric angles on T4 heads, we have estimated a Q number of 21. This confirms previous indirect studies on T4 and agrees with determinations on bacteriophage T2. Many of the structural features of T4 observed in FDET preparations differ significantly from those observed by classical negative staining methods for TEM imaging. Most important among the differences are the conformation of the baseplate (a closed rosebud) and the positioning of the tail fibers (retracted). The retracted position of the tail fibers in the FDET preparations has been confirmed by negatively staining phage previously fixed suspended in solution with 2% glutaraldehyde. The FDET protocols appear to reveal important structural features not seen in negative stained preparations. These have implications for bacteriophage T4 conformation in solution, viral assembly and phage conformation states prior to tail contraction and DNA ejection.

Ultrathin (1 nm) vertically shadowed platinum-carbon replicas for imaging individual molecules in freeze-etched biological DNA and material science metal and plastic specimens.

Single molecule resolution in beam-sensitive, uncoated, noncrystalline materials has heretofore not been possible except in thin (less than or equal to 150 A) platinum-carbon (Pt-C) replicas, which are resistant to electron beam destruction. Previously, the granularity of metal film replicas limited their resolution to greater than or equal to 20 A. This paper demonstrates that Pt-C film granularity and resolution are a function of the method of replication and other controllable factors. Low-angle 20 degrees rotary, 45 degrees unidirectional, and vertical 9.7 +/- 1 A Pt-C films deposited on mica under the same conditions were compared. Vertical replication had a 5 A granularity, the highest resolution, and evenly coated the whole surface. A 45 degrees replication had a 9.5 A granularity, a slightly poorer resolution, and a discontinuous surface coating. The use of 20 degrees rotary replication proved to be unsuitable for high-resolution imaging, with 20-25 A granularity and resolution two to three times poorer. Vertical and 45 degrees Pt-C replicas can visualize the deep-etched DNA helix and the 13.3 A 3(2) helix of pectin in a gel. The DNA double helix, the complex structures of sol-gel glasses, Immobilon filters (polyvinylidene fluoride), a polymethacrylate plastic, the metal oxide surfaces of 440c stainless steel, and aluminum are illustrated. This high-resolution vertical Pt-C replica technique can image in the context of solutions, gels, or solids, single molecular chains 3-7 A wide, their associations, and their conformation. Included in the present article are first time descriptions for removing replicas from metals and plastics and for making high-magnification photographic prints of normal contrast using a reversal rephotographic process.

Immunogold-labeled microtubule crossbridges in platinum-carbon replicas of the marginal band of erythrocyte cytoskeletons.

Cytoskeletons of erythrocytes from the toad Bufo marinus are composed of a surface-associated cytoskeleton that encapsulates the annular bundle of microtubules known as the marginal band (MB) and the centrally located nucleus. As seen by phase-contrast microscopy, the microtubules (MTs) of the MB remain tightly bundled after cell lysis without the need for added stabilizing factors. The integrity of this structure suggested that in addition to MTs other components were present in the MB and were responsible for its stability. Thin (less than 18 nm) platinum-carbon (Pt-C) replicas of freeze-dried cytoskeletons prepared by using a modified Balzers 300 system provided a novel method of sample preparation for a high-resolution study of the ultrastructure of the MB. Electron micrographs of replicas revealed that, the MTs of the MB displayed numerous filamentous projections which, when viewed in stereo, appear as side-arm connections between adjacent MTs. Immunofluorescence data show that monospecific antibodies to tubulin and to MT-associated protein 2 (MAP2) from brain each detect cross-reactive material in the MB. The combination of immunogold cytochemistry with Pt-C replication provided the increased resolution required to identify the individual structures recognized by antibodies to tubulin and MAP2. As expected, antitubulin labeled the MTs of the MB. However, anti-MAP2 antibodies were localized specifically to the cross-bridging filaments between adjacent MTs. Thus, a MAP2-like protein was identified in situ as the crossbridging filament that bundles MTs to form a stable MB.

Type IV collagen lateral associations in the EHS tumor matrix. Comparison with amniotic and in vitro networks.

The macromolecular structural organization of the type IV collagen network in the extracellular matrix of the EHS tumor has been investigated using a stereoscopic freeze-dry Pt/C replication technique. This network, which can be specifically decorated with type IV collagen antibody, is formed in great part by the lateral joining of narrow filaments (2.7 nm average metal coated diameter) to form a complex three-dimensional irregular polygonal array of variable diameter branching strands. Globular domains, similar to the C-terminal globular domains of purified type IV dimers, can be identified in the network. In many regions of the network the filaments appear to twist around each other along the strand axis. The network is similar to that visualized in the human amnion as well as to a reconstituted network formed in vitro. These data strongly suggest that the laterally and end-domain-associated network is a widespread supramolecular architecture of type IV collagen in basement membranes.

Electron microscopic studies of free and proteinase-bound duck ovostatins (ovomacroglobulins). Model of ovostatin structure and its transformation upon proteolysis.

Conformational changes of duck ovostatin (ovomacroglobulin) upon complexing with thermolysin have been studied by electron microscopy. Both free and thermolysin-bound ovostatin preparations were negatively stained with uranyl acetate, a series of three pictures were taken at 10 degrees specimen tilt intervals (+10 degrees, 0 degrees, and -10 degrees), and images of the inhibitor molecules were observed in three dimensions. Four approximately cylindrical subunits were observed in free ovostatin. Two subunits associated approximately midway from both ends to form a dimer of four arms. Two dimers associated with each other at the midpoint to form a tetramer. The proteinase susceptible "bait" regions were located near the center of the molecule. Eight arms of the tetramer take various configurations. The orthogonal extent of free tetrameric ovostatin in a two-dimensional micrograph averages 26.0 +/- 4.7 x 34.0 +/- 5.0 nm. Upon complexing with thermolysin, all eight arms curl toward the center of the molecule, having four arms upward and the other four downward. Thus, proteinase-bound ovostatin has a uniform structure with a 2-fold axis of symmetry. The overall structure of the complex is more compact with average dimensions of 16.9 +/- 0.6 x 16.9 +/- 0.6 x 19.9 +/- 0.4 nm. From these electron microscopic studies we propose that a proteinase reaches to the center of the free ovostatin molecule and attacks the bait region. Subsequent to proteolysis the subunit arms curl and entrap the enzyme within the ovostatin molecule. The results support the unique mechanism of inhibition of proteinases by alpha 2-macroglobulin and ovostatin postulated from biochemical observations (Barrett, A. J., and Starkey, P. M. (1973) Biochem. J. 133, 709-724; Nagase, H., and Harris, E. D., Jr. (1983) J. Biol. Chem. 258, 7490-7498).

Basement membrane structure in situ: evidence for lateral associations in the type IV collagen network.

To determine molecular architecture of the type IV collagen network in situ, the human amniotic basement membrane has been studied en face in stereo relief by high resolution unidirectional metal shadow casting aided by antibody decoration and morphometry. The appearance of the intact basement membrane is that of a thin sheet in which there are regions of branching strands. Salt extraction further exposes these strands to reveal an extensive irregular polygonal network that can be specifically decorated with gold-conjugated anti-type IV collagen antibody. At high magnification one sees that the network, which contains integral (9-11 nm net diameter) globular domains, is formed in great part by lateral association of monomolecular filaments to form branching strands of variable but narrow diameters. Branch points are variably spaced apart by an average of 45 nm with 4.4 globular domains per micron of strand length. Monomolecular filaments (1.7-nm net diameter) often appear to twist around each other along the strand axis; we propose that super helix formation is an inherent characteristic of lateral assembly. A previous study (Yurchenco, P. D., and H. Furthmayr. 1984. Biochemistry. 23:1839) presented evidence that purified murine type IV collagen dimers polymerize to form polygonal arrays of laterally as well as end-domain-associated molecules. The architecture of this polymer is similar to the network seen in the amnion, with lateral binding a major contributor to each. Thus, to a first approximation, isolated type IV collagen can reconstitute in vitro the polymeric molecular architecture it assumes in vivo.