Visualization of a system of filaments 7-10 nm thick in cultured cells of an epithelioid line (Pt K2) by immunofluorescence microscopy

Directional change during active diffusion of viral ribonucleoprotein particles through cytoplasm

A mesh of cytoskeletal fibers, consisting of microtubules, intermediate filaments, and fibrous actin, prevents the Brownian diffusion of particles with a diameter larger than 0.10 μm, such as vesicular stomatitis virus ribonucleoprotein (RNP) particles, in mammalian cells. Nevertheless, RNP particles do move in random directions but at a lower rate than Brownian diffusion, which is thermally driven. This nonthermal biological transport process is called "active diffusion" because it is driven by ATP. The ATP powers motor proteins such as myosin II. The motor proteins bend and cross-link actin fibers, causing the mesh to jiggle. Until recently, little was known about how RNP particles get through the mesh. It has been customary to analyze the tracks of particles like RNPs by computing the slope of the ensemble-averaged mean-squared displacement of the particles as a signature of mechanism. Although widely used, this approach "loses information" about the timing of the switches between physical mechanisms. It has been recently shown that machine learning composed of variational Bayesian analysis, Gaussian mixture models, and hidden Markov models can use "all the information" in a single track to reveal that that the positions of RNP particles are spatially clustered. Machine learning assigns a number, called a state, to each cluster. RNP particles remain in one state for 0.2-1.0 s before switching (hopping) to a different state. This earlier work is here extended to analyze the movements of a particle within a state and to determine particle directionality within and between states.

Mechanisms of active diffusion of vesicular stomatitis virus inclusion bodies and cellular early endosomes in the cytoplasm of mammalian cells

Viral and cellular particles too large to freely diffuse have two different types of mobility in the eukaryotic cell cytoplasm: directed motion mediated by motor proteins moving along cytoskeletal elements with the particle as its load, and motion in random directions mediated by motor proteins interconnecting cytoskeletal elements. The latter motion is referred to as "active diffusion." Mechanisms of directed motion have been extensively studied compared to mechanisms of active diffusion, despite the observation that active diffusion is more common for many viral and cellular particles. Our previous research showed that active diffusion of vesicular stomatitis virus (VSV) ribonucleoproteins (RNPs) in the cytoplasm consists of hopping between traps and that actin filaments and myosin II motors are components of the hop-trap mechanism. This raises the question whether similar mechanisms mediate random motion of larger particles with different physical and biological properties. Live-cell fluorescence imaging and a variational Bayesian analysis used in pattern recognition and machine learning were used to determine the molecular mechanisms of random motion of VSV inclusion bodies and cellular early endosomes. VSV inclusion bodies are membraneless cellular compartments that are the major sites of viral RNA synthesis, and early endosomes are representative of cellular membrane-bound organelles. Like VSV RNPs, inclusion bodies and early endosomes moved from one trapped state to another, but the distance between states was inconsistent with hopping between traps, indicating that the apparent state-to-state movement is mediated by trap movement. Like VSV RNPs, treatment with the actin filament depolymerizing inhibitor latrunculin A increased VSV inclusion body mobility by increasing the size of the traps. In contrast neither treatment with latrunculin A nor depolymerization of microtubules by nocodazole treatment affected the size of traps that confine early endosome mobility, indicating that intermediate filaments are likely major trap components for these cellular organelles.

The Emerging Role of Cytoskeletal Proteins as Reliable Biomarkers

Cytoskeletal proteins are essential building blocks of cells. More than 100 cytoskeletal and cytoskeleton-associated proteins are known and for some, their function and regulation are understood in great detail. Apart from cell shape and support, they facilitate many processes such as intracellular signaling and transport, and cancer related processes such as proliferation, migration, and invasion. During the last decade, comparative proteomic studies have identified cytoskeletal proteins as in vitro markers for tumor progression and metastasis. Here, these results are summarized and a number of unrelated studies are highlighted, identifying the same cytoskeletal proteins as potential biomarkers. These findings might indicate that the abundance of these potential markers of tumor progression is associated with the biological outcome and are independent of the cancer origin. This correlates well with recently published results from the Cancer Genome Atlas, indicating that cancers show remarkable similarities in their analyzed molecular information, independent of their organ of origin. It is postulated that the quantification of cytoskeletal proteins in healthy tissues, tumors, in adjacent tissues, and in stroma, is a great source of molecular information, which might not only be used to classify tumors, but more importantly to predict patients' outcome or even best treatment choices.

The Molecular Revolution in Cutaneous Biology: Keratin Genes and their Associated Disease: Diversity, Opportunities, and Challenges

The abundance of keratin proteins and the filaments they form in surface epithelia has long been appreciated. This said, the remarkable diversity of keratin proteins and the notion that they are encoded by one of the largest gene families in the human genome has come to the fore relatively recently, coinciding with the sequencing of whole genomes. This complexity has generated some practical challenges, notably in terms of nomenclature and tractability. More importantly, however, studies of keratin have seeded the discovery of the genetic basis for a large number of genodermatoses and continue to provide a unique perspective on and insight into epithelial cells and tissues, whether normal or diseased.

Analysis of ER-mitochondria contacts using correlative fluorescence microscopy and soft X-ray tomography of mammalian cells

Mitochondrial fission is important for organelle transport, quality control and apoptosis. Changes to the fission process can result in a wide variety of neurological diseases. In mammals, mitochondrial fission is executed by the GTPase dynamin-related protein 1 (Drp1; encoded by DNM1L), which oligomerizes around mitochondria and constricts the organelle. The mitochondrial outer membrane proteins Mff, MiD49 (encoded by MIEF2) and MiD51 (encoded by MIEF1) are involved in mitochondrial fission by recruiting Drp1 from the cytosol to the organelle surface. In addition, endoplasmic reticulum (ER) tubules have been shown to wrap around and constrict mitochondria before a fission event. Up to now, the presence of MiD49 and MiD51 at ER-mitochondrial division foci has not been established. Here, we combine confocal live-cell imaging with correlative cryogenic fluorescence microscopy and soft x-ray tomography to link MiD49 and MiD51 to the involvement of the ER in mitochondrial fission. We gain further insight into this complex process and characterize the 3D structure of ER-mitochondria contact sites.

Immunolabeling artifacts and the need for live-cell imaging

Fluorescent fusion proteins have revolutionized examination of proteins in living cells. Still, studies using these proteins are met with criticism because proteins are modified and ectopically expressed, in contrast to immunofluorescence studies. However, introducing immunoreagents inside cells can cause protein extraction or relocalization, not reflecting the in vivo situation. Here we discuss pitfalls of immunofluorescence labeling that often receive little attention and argue that immunostaining experiments in dead, permeabilized cells should be complemented with live-cell imaging when scrutinizing protein localization.

Immunofluorescence detection of the cytoskeleton and extracellular matrix in tissue and cultured cells

"A picture is worth a thousand words" goes the proverb. A poor picture however can be worse than saying nothing at all. This is particularly true for immunofluorescence pictures that in addition to their informative character bear an esthetic component. We here provide a panel of straightforward methods to process tissue sections and cultured cells for immunostaining of cytoskeletal elements, primarily those associated with actin filaments. We want to emphasize to the reader the fact that the choice of the processing method will have an important influence on the outcome of the immunostaining and thus on the interpretation of the results. Fixation of cultured cells with cross-linking reagents such as paraformaldehyde efficiently preserves structural elements at the expense of reduced antigenicity. The degree and timing of cell permeabilization with detergents, along with chemical cross-linking, contributes to the clarity and resolution of distinct structures but can also lead to loss of information. Fixation with organic solvents like methanol will, in most cases, better preserve antigens but will produce a higher background and impact on structural integrity. Therefore, it is recommended to test different protocols for a "new" protein or epitope - the results will pay back your investment.

Fluorescence nanoscopy in whole cells by asynchronous localization of photoswitching emitters

We demonstrate nanoscale resolution in far-field fluorescence microscopy using reversible photoswitching and localization of individual fluorophores at comparatively fast recording speeds and from the interior of intact cells. These advancements have become possible by asynchronously recording the photon bursts of individual molecular switching cycles. We present images from the microtubular network of an intact mammalian cell with a resolution of 40 nm.

Les contrôles nécessaires en immunohistochimie : de la recherche au diagnostic

The specificity of an immunohistochemical reaction is guaranteed by two sets of controls. Positive controls verify the specificity of the primary antibody and demonstrate that it binds only to the protein which was used as an immunogen. Negative controls ensure that the labelling technique is specific and that the primary antibody is responsible for generation of the immunostaining. In fact, the production of a labelling may also be related to cross reactivity or to non-specific physical or chemical interactions. This paper reviews the characteristics of various epitopes and antibodies, describes different strategies which prove the specificity of the immunohistochemical reaction in research or diagnostic pathology and point towards the essential information which should be reported in a paper.

4Pi microscopy of quantum dot-labeled cellular structures

The most prominent restrictions of fluorescence microscopy are the limited resolution and the finite signal. Established conventional, confocal, and multiphoton microscopes resolve at best approximately 200nm in the focal plane and only 500nm in depth. Additionally, organic fluorophores and fluorescent proteins are bleached after 10(4)-10(5) excitation cycles. To overcome these restrictions, we synergistically combine the 3- to 7-fold improved axial resolution of 4Pi microscopy with the greatly enhanced photostability of semiconductor quantum dots. Co-localization studies of immunolabeled microtubules and mitochondria demonstrate the feasibility of this approach for routine biological measurements. In particular, we visualize the three-dimensional entanglement of the two networks with unprecedented detail.

2,2′-Thiodiethanol: A new water soluble mounting medium for high resolution optical microscopy

The use of high numerical aperture immersion lenses in optical microscopy is compromised by spherical aberrations induced by the refractive index mismatch between the immersion system and the embedding medium of the sample. Especially when imaging >10 micro m deep into the specimen, the refractive index mismatch results in a noticeable loss of image brightness and resolution. A solution to this problem is to adapt the index of the embedding medium to that of the immersion system. Unfortunately, not many mounting media are known that are both index tunable as well as compatible with fluorescence imaging. Here we introduce a nontoxic embedding medium, 2,2'-thiodiethanol (TDE), which, by being miscible with water at any ratio, allows fine adjustment of the average refractive index of the sample ranging from that of water (1.33) to that of immersion oil (1.52). TDE thus enables high resolution imaging deep inside fixed specimens with objective lenses of the highest available aperture angles and has the potential to render glycerol embedding redundant. The refractive index changes due to larger cellular structures, such as nuclei, are largely compensated. Additionally, as an antioxidant, TDE preserves the fluorescence quantum yield of most of the fluorophores. We present the optical and chemical properties of this new medium as well as its application to a variety of differently stained cells and cellular substructures.

Dissection of keratin dynamics: different contributions of the actin and microtubule systems

It has only recently been recognized that intermediate filaments (IFs) and their assembly intermediates are highly motile cytoskeletal components with cell-type- and isotype-specific characteristics. To elucidate the cell-type-independent contribution of actin filaments and microtubules to these motile properties, fluorescent epithelial IF keratin polypeptides were introduced into non-epithelial, adrenal cortex-derived SW13 cells. Time-lapse fluorescence microscopy of stably transfected SW13 cell lines synthesizing fluorescent human keratin 8 and 18 chimeras HK8-CFP and HK18-YFP revealed extended filament networks that are entirely composed of transgene products and exhibit the same dynamic features as keratin systems in epithelial cells. Detailed analyses identified two distinct types of keratin motility: (I) Slow (approximately 0.23 microm/min), inward-directed, continuous transport of keratin filament precursor particles from the plasma membrane towards the cell interior, which is most pronounced in lamellipodia. (II) Fast (approximately 17 microm/min), bidirectional and intermittent transport of keratin particles in axonal-type cell processes. Disruption of actin filaments inhibited type I motility while type II motility remained. Conversely, microtubule disruption inhibited transport mode II while mode I continued. Combining the two treatments resulted in a complete block of keratin motility. We therefore conclude that keratin motility relies both on intact actin filaments and microtubules and is not dependent on epithelium-specific cellular factors.

Insights into the dynamic properties of keratin intermediate filaments in living epithelial cells

The properties of keratin intermediate filaments (IFs) have been studied after transfection with green fluorescent protein (GFP)-tagged K18 and/or K8 (type I/II IF proteins). GFP-K8 and -K18 become incorporated into tonofibrils, which are comprised of bundles of keratin IFs. These tonofibrils exhibit a remarkably wide range of motile and dynamic activities. Fluorescence recovery after photobleaching (FRAP) analyses show that they recover their fluorescence slowly with a recovery t(1/2) of approximately 100 min. The movements of bleach zones during recovery show that closely spaced tonofibrils (<1 microm apart) often move at different rates and in different directions. Individual tonofibrils frequently change their shapes, and in some cases these changes appear as propagated waveforms along their long axes. In addition, short fibrils, termed keratin squiggles, are seen at the cell periphery where they move mainly towards the cell center. The motile properties of keratin IFs are also compared with those of type III IFs (vimentin) in PtK2 cells. Intriguingly, the dynamic properties of keratin tonofibrils and squiggles are dramatically different from those of vimentin fibrils and squiggles within the same cytoplasmic regions. This suggests that there are different factors regulating the dynamic properties of different types of IFs within the same cytoplasmic regions.

Detection of cytokeratin dynamics by time-lapse fluorescence microscopy in living cells

To monitor the desmosome-anchored cytokeratin network in living cells fusion protein HK13-EGFP consisting of human cytokeratin 13 and the enhanced green fluorescent protein was stably expressed in vulvar carcinoma-derived A-431 cells. It is shown for A-431 subclone AK13-1 that HK13-EGFP emits strong fluorescence in fixed and living cells, being part of an extended cytoplasmic intermediate filament network that is indistinguishable from that of parent A-431 cells. Biochemical, immunological and ultrastructural analyses demonstrate that HK13-EGFP behaves identically to the endogenous cytokeratin 13 and is therefore a reliable in vivo tag for this polypeptide and the structures formed by it. Time-lapse fluorescence microscopy reveals that the cytokeratin 13-containing network is in constant motion, resulting in continuous restructuring occurring in single and migratory cells, as well as in desmosome-anchored cells. Two major types of movement are distinguished: (i) oscillations of mostly long filaments, and (ii) an inward-directed flow of fluorescence originating as diffuse material at the cell periphery and moving in the form of dots and thin filaments toward the deeper cytoplasm where it coalesces with other filaments and filament bundles. Both movements are energy dependent and can be inhibited by nocodazole, but not by cytochalasin D. Finally, disassembly and reformation of cytokeratin filament networks are documented in dividing cells revealing distinct and rapidly occurring stages of cytokeratin organisation and distribution.

Induction of tyrosine aminotransferase of primary cultured rat hepatocytes depends on the organization of microtubules

We investigated the relationship between the expression of tyrosine aminotransferase (TAT) and cytoskeletal systems of cultured rat hepatocytes by using serum-free culture conditions and changing three factors: 1) the concentration of calcium, 2) the dish-coating material, and 3) the cell-plating density. In hepatocytes in low-calcium medium, induction of TAT by dexamethasone and glucagon was maintained, although cell-cell adhesion was lost. Hepatocytes on Matrigel formed a nonspreading, spherical shape that provided them with the full extent of TAT activity without cell-cell adhesion. Hepatocytes plated on collagen at low cell density spread and changed shape, and the induction of TAT activity was markedly reduced. By using confocal laser-scanning microscopy, we analyzed the three-dimensional organization of cytoplasmic microtubules of hepatocytes maintaining the ability of TAT induction. Hepatocytes plated on collagen at low cell density possessed the radial filamentous structure of cytoplasmic microtubules. When the spherical shape of hepatocytes was maintained by cultivating cells on Matrigel, a ring-like structure of cytoplasmic microtubules beneath the plasma membrane was dominant. Moreover, the induction of TAT activity of hepatocytes in a standard culture system was strongly inhibited by the addition of 1 microM colchicine. These studies suggest that the organization of cytoplasmic microtubules may participate in the shape-related regulation of cell function.

5-Lipoxygenase expression in Langerhans cells of normal human epidermis

We studied expression of the 5-lipoxygenase (5-LO) pathway in normal human skin. In situ hybridization revealed a 5-LO mRNA-containing epidermal cell (EC) population that was predominantly located in the midportion of the spinous layer, in outer hair root sheaths, and in the epithelial compartment of sebaceous glands. Examination of skin specimens by immunohistochemistry and of primary ECs by flow cytometry mapped the 5-LO protein exclusively to Langerhans cells (LCs). The LC 5-LO protein was largely found in the nuclear matrix, in nuclear envelopes, and perinuclear regions as indicated by in situ confocal laser scan microscopy. Reverse transcription-PCR and immunoblot analyses of purified primary EC populations further indicated that LCs are major 5-LO expressing cells. Enriched primary LCs were also found to contain 5-LO activating protein (FLAP), leukotriene (LT) C4 synthase, and LTA4 hydrolase. By contrast, 5-LO, FLAP, and LTC4 synthase were undetectable or largely reduced, but LTA4 hydrolase transcripts and protein were identified in ECs depleted of LCs. These data show that naive LCs are major, and possibly the sole, 5-LO pathway expressing cells in the normal human epidermis.

Neurofibromin colocalizes with mitochondria in cultured cells

Mutations in neurofibromatosis type 1 target the gene coding for neurofibromin. While neurofibromin is able to accelerate the rate of GTP hydrolysis by cellular Ras proteins, its biological function is not well understood. To gain information regarding its function, the intracellular localization of neurofibromin was analyzed in cultured cell lines using polyclonal antisera raised against four neurofibromin-specific peptides, three from the carboxyl terminus and one from the amino terminus. In methanol-fixed cells distinct rod-like structures distributed throughout the cytoplasm were recognized by the antisera. Similar structures were seen with each antiserum, including affinity-purified antibodies, and in each of the cultured cell lines tested. Similar structures were seen in paraformaldehyde-fixed cells. Double staining experiments showed that these structures colocalize with mitochondria, but not with actin, beta-tubulin, or endoplasmic reticulum. When actin or tubulin structures within the cell were disrupted by separate antimitotic drugs, these stained structures retained their shape. Neurofibromin association with mitochondria was confirmed biochemically when highly purified mitochondrial fractions from bovine heart tissue were shown in Western analysis to contain neurofibromin. This association might be helpful in predicting identification of some of the cellular proteins with which neurofibromin interacts.

Centripetal transport of microtubules in motile cells

The study of microtubule (MT) dynamics in cells has largely been restricted to events occurring over relatively short periods in nonmotile or stationary cells in culture. By using the antioxidant, Oxyrase, we have reduced the sensitivity of fluorescent MTs to photodamage and this has allowed us to image fluorescent MTs with good temporal resolution over much longer periods of time. We have used our enhanced imaging capabilities to examine MT dynamics in fibroblasts moving directionally into a wound. We found that MTs in these cells exhibited dynamic instability similar to that reported for other cells. More interestingly, we found a novel dynamic behavior of the MTs in which entire MTs were moved inward from the leading edge toward the cell nucleus. This centripetal transport (CT) of MTs only occurred to those MTs that were oriented with their long axis parallel to the leading edge; radially oriented MTs were not transported centripetally. Both small bundles of MTs and individual MTs were observed to undergo CT at a rate of 0.63 +/- 0.37 micron/min. This rate was similar to the rate of CT of latex beads applied to the cell surface and of endogenous pinocytotic vesicles in the cytoplasm. When we imaged both MTs and pinocytotic vesicles, we found that the pinocytotic vesicles were ensheathed by a small group of parallel MTs that moved centripetally in concert with the vesicles. Conversely, we found many instances of MTs moving centripetally without associated vesicles. When cells were treated with nocodazole to depolymerize MTs rapidly, the rate of pinocytotic vesicle CT was inhibited by 75%. This suggests that centripetal transport of MTs may be involved in the movement of pinocytotic vesicles in cells. In conclusion, our results show that MTs in motile cells are redistributed by a novel mechanism, CT, that does not require changes in polymer length. The centripetally transported MTs may play a role in transporting pinocytotic vesicles in the cell.

A confocal microscopy study of anticytoskeletal antibody activity in patients with connective tissue disease

The significance of the presence of antibodies to cytoskeleton proteins in patients with connective tissue diseases is not clear, as there is a high level of these antibodies in healthy controls. In an attempt to improve the visualization of the immunofluorescence binding pattern of autoantibodies to cytoskeletal structures in cultured fibroblasts, we have used confocal microscopy. Of the 256 serum samples tested, 155 (61%) WERE reactive with cytoplasmic structures. These reactive samples could be divided into seven patterns of binding, as determined by double-blind examination of single-section confocal images. While confirming the results of previous immunofluorescence studies which have shown that autoantibodies that bind to filamentous structures in the cytoplasm of cultured cells are common in patients with connective tissue diseases, we were able to identify three patterns of cytoskeletal binding which may be useful as an adjunct to other tests for the diagnosis of some connective tissue diseases, in particular systemic sclerosis (scleroderma) and rheumatoid arthritis/Sjogren's syndrome. None of the seven patterns was exclusive to a particular disease. We conclude that confocal microscopy may be of limited use as an adjunct to other serological assays in the diagnosis of some forms of connective tissue disease.

An immunofluorescence study of the effects of ultraviolet radiation on the organization of microfilaments, keratin intermediate filaments, and microtubules in human keratinocytes

Indirect immunofluorescence microscopy has been used to investigate the ultraviolet (UV) radiation induced disruption of the organization of microfilaments, keratin intermediate filaments, and microtubules in cultured human epidermal keratinocytes. Following irradiation, concurrent changes in the organization of the three major cytoskeletal components were observed in cells incubated under low Ca2+ (0.15 mM) conditions. UV irradiation induced a dose-dependent condensation of keratin filaments into the perinuclear region. This collapse of the keratin network was accompanied by the reorganization of microfilaments into rings and a restricted distribution of microtubules, responses normally elicited by exposure to high Ca2+ (1.05 mM) medium. The UV induced alteration of the keratin network appears to disrupt the interactions between keratin and actin, permitting the reorganization of actin filaments in the absence of Ca2+ stimulation. In addition to the perinuclear condensation of keratin filaments, UV irradiation inhibits the Ca2+ induced formation of keratin alignments at the membrane of apposed cells if UV treatment precedes exposure to high Ca2+ medium. Incubation of keratinocytes in high Ca2+ medium for 24 hours prior to irradiation results in the stabilization of membrane associated keratin alignments and a reduced susceptibility of cytoplasmic keratin filaments to UV induced disruption. Unlike results from investigations with isogenic skin fibroblasts, no UV induced disassembly of microtubules was discernible in irradiated human keratinocytes.

An immunofluorescence study of the calcium-induced coordinated reorganization of microfilaments, keratin intermediate filaments, and microtubules in cultured human epidermal keratinocytes

Indirect immunofluorescence microscopy has been used to investigate the coordinated reorganization of microtubules, microfilaments, and keratin intermediate filaments in cultured human epidermal keratinocytes following a switch from low-Ca++ (0.15 mM) medium to high-Ca++ (1.05 mM) medium. A dramatic reorganization occurs concurrently in the three major cytoskeletal components shortly after the calcium switch. The most prominent features are the alignment of keratin filaments at the plasma membranes of apposed cells, the induction of microfilament rings, the restriction of microtubules to the area within the boundaries of the microfilament rings, and the alignment of actin bundles at cell borders. Additional changes are observed in terminally differentiated cells. This is the first report that describes simultaneous changes in the organization of the three major cytoskeletal components of epidermal keratinocytes. Cytochalasin D and demecolcine (colcemid) studies were performed to determine whether the organization of microtubules, microfilaments, and keratin filaments, as well as the calcium-induced reorganization of these cytoskeletal elements, may be dependent on the existence of structural relationships between them. These studies demonstrate that the disruption of microfilaments results in the formation of a latticelike keratin network, with a close association of actin and keratin being maintained. The formation of keratin filament alignments occurs even in the absence of intact microfilaments. In addition, it was found that the Ca(++)-induced reorganization of microfilaments and keratin filaments is not dependent on an intact microtubule network. Furthermore, the reorganization of actin into concentric rings can be dissociated from changes in the organization of keratin filaments.

Regulation of the biosynthesis of subgroup C adenovirus protein IVa2

The IVa2 gene is located between 16 and 11.3 map units on the left strand of the adenovirus type 5 (Ad5) genome. The coded RNA contains an intron of 277 nucleotides. To determine whether protein IVa2 is synthetized during productive infection and to obtain an immunological reagent to study its function, we prepared antibodies directed to 414 amino acids of protein IVa2 fused to the N-terminal domain of Staphylococcus aureus protein A. Western immunoblot analysis of viral proteins demonstrates that protein IVa2 is a minor component of mature viral particles and that it is also present in assembly intermediates and young virions. Thus, contrary to a previous report (H. Persson, B. Mathisen, L. Philipson, and U. Pettersson, Virology 93:198-208, 1979), protein IVa2 is not related to the 50-kDa polypeptide, a scaffolding protein present in assembly intermediates. The biosynthesis of protein IVa2 during productive infection was examined. Time course studies using immunofluorescence analysis with polyclonal antibodies targeted to protein IVa2 revealed that this protein is first synthesized at 12 h in a few cells exhibiting very striking fluorescence. Synthesis continues until at least 24 h postinfection. When hydroxyurea is added, protein IVa2 is not detected. In cells infected with mutant H5 ts125, blocked at the nonpermissive temperature (40 degrees C) in viral DNA replication, protein IVa2 is overexpressed. These results suggest that protein IVa2 synthesis requires cellular rather than viral DNA replication. RNase protection assay results indicate that hydroxyurea inhibits protein IVa2 synthesis at the transcriptional level. Thus, overexpression of protein IVa2 in H5 ts125-infected cells may be regulated at the translational level.

Immunofluorescence localization of the epidermolytic toxin target in mouse epidermal cells and tissue

An epidermolytic toxin target was observed in keratohyalin granules of sectioned epidermis by a 'direct' fluorescence procedure using FTC-toxin, but not by an 'indirect' procedure using sequential reaction with toxin, anti-toxin and FTC-secondary antibody. The investigation of the two procedures was extended to keratinocytes. A dispase digestion procedure yielded three fractions which corresponded to basal, spinous and granular cells according to biochemical and morphological criteria. It was shown that the 'direct' and 'indirect' procedures both detected the toxin target in the keratohyalin granules of granular cells, but that the 'indirect' procedure was very insensitive. In control experiments, the profilaggrin of keratohyalin granules was detected readily in cells by a 'direct' procedure using FTC-antiprofilaggrin but only weakly by an 'indirect' double antibody procedure. Insensitivity to 'indirect' procedures thus appears to be a particular property of the keratohyalin granule site. It was shown that the toxin target was readily accessible in permeable (trypsin-isolated) granular cells but inaccessible in impermeable (dispase-isolated) cells.

Immunogold labelling of the intermediate filament-lamina-nuclear matrix system in HeLa and BHK-21 cells

Whole-mount, sequentially extracted cells combined with immunogold electron microscopy were developed to demonstrate the intermediate filaments, lamina, and nuclear matrix (IF-L-NM) and to identify their protein components. The IFs of HeLa cells were reacted both with keratin and vimentin monoclonal antibodies; meanwhile, the IF network of BHK-21 cell was reacted only with vimentin monoclonal antibody. The lamina and nuclear matrices of both HeLa and BHK-21 cell were labelled, respectively, with lamin monoclonal antibody-gold complex and 280 Kd nuclear matrix protein monoclonal antibody-gold complex. The monoclonal antibody to keratin could cross-react with the lamina both of HeLa and BHK-21 cells.

Immunolocalization of myosin in intact and wounded cells of the green alga Ernodesmis verticillata (Kützing) Borgesen

Myosin localization was examined in the coenocytic green alga E. verticillata using indirect immunofluorescence microscopy. A polyclonal antibody affinity-purified against the heavy chain of slime-mold myosin recognizes a 220000 to 230000 Mr protein that electrophoretically migrates slightly behind rabbit myosin. A second polypeptide of 85000 Mr is also consistently detected in immunoblots, indicating that two forms of myosin-like proteins may be present in these cells. In intact cells, myosin immunofluorescence is present on the chloroplast surfaces, in nuclei and in cytoplasmic strands between plastids. Myosin labeling also occurs in association with pyrenoids primarily in apical chloroplasts. During wound-induced cytoplasmic contractions, myosin is localized near the plasma membrane in longitudinal arrays superimposed over a reticulate pattern of fluorescence; both these patterns become apparent upon wounding. Double-label immunofluorescence of actin and myosin demonstrates that these arrays represent the longitudinal bundles of actin microfilaments and the actin-containing reticulum, the former being directly associated with contraction in these cells. These results indicate that both actin and myosin are associated with contractility in Ernodesmis, probably representing the apparatus and "molecular motor", respectively, which effect motility.

E-cadherin-mediated cell-cell adhesion prevents invasiveness of human carcinoma cells

The ability of carcinomas to invade and to metastasize largely depends on the degree of epithelial differentiation within the tumors, i.e., poorly differentiated being more invasive than well-differentiated carcinomas. Here we confirmed this correlation by examining various human cell lines derived from bladder, breast, lung, and pancreas carcinomas. We found that carcinoma cell lines with an epithelioid phenotype were noninvasive and expressed the epithelium-specific cell-cell adhesion molecule E-cadherin (also known as Arc-1, uvomorulin, and cell-CAM 120/80), as visualized by immunofluorescence microscopy and by Western and Northern blotting, whereas carcinoma cell lines with a fibroblastoid phenotype were invasive and had lost E-cadherin expression. Invasiveness of these latter cells could be prevented by transfection with E-cadherin cDNA and was again induced by treatment of the transfected cells with anti-E-cadherin mAbs. These findings indicate that the selective loss of E-cadherin expression can generate dedifferentiation and invasiveness of human carcinoma cells, and they suggest further that E-cadherin acts as an invasion suppressor.

The characterization of free, cytoskeletal and membrane-bound polysomes in Krebs II ascites and 3T3 cells

Polysomes from Krebs II ascites and 3T3 cells were separated into three populations by using a sequential extraction method. Free polysomes were released by using a combination of low salt (25 mM KCl) and NP-40 detergent in the lysis buffer. The cytoskeletal bound polysomes were subsequently released by raising the salt concentration to 130 mM and finally, polysomes bound to the membranes of the endoplasmic reticulum were extracted by the combined treatment with Triton X-100 and deoxycholate. The results presented here illustrate that the three polysome-containing fractions differ in many parameters such as polysome profiles, cytoskeletal components and phospholipid content. When polyA-containing mRNA was isolated from the three polysome fractions and translated in an in vitro system, some differences were observed in the patterns of proteins being synthesized.

Study of vimentin expression in non-Hodgkin's lymphoma using paraffin sections

Human non-Hodgkin's lymphomas were studied by means of an avidin-biotin complex immunoperoxidase method using several monoclonal antibodies against the intermediate filament protein, vimentin. The study cases were 61 B-cell lymphomas (including 2 plasmacytomas) and 30 T-cell lymphomas (including 8 cases of mycosis fungoides). Twelve of the 61 B-cell lymphomas were positive for vimentin, and were composed of extrafollicular-center cells such as immunoblastic and plasmacytoid cells. On the other hand, lymphomas of follicular center cell origin were negative for vimentin. All cases of T-cell lymphoma except for 14 (all of 9 AILD-type lymphomas, all of 4 lymphoblastic lymphomas and one diffuse mixed small/large lymphoma) were positive for vimentin. Although vimentin expression appeared to be influenced by various conditions such as the proportion of T- and B-cell subsets, or B-cell proliferation rate, follicular center cells were constantly negative for vimentin.

Reciprocal/dichotomic expression of vimentin and B cell differentiation antigens in Reed-Sternberg's cells

An immunohistochemical study of 63 cases of Hodgkin's disease was undertaken using formalin-fixed paraffin embedded tissue sections. The antibodies used were against L26, LN-1, LN-2, EMA (epithelial membrane antigen), Leu-M1, Vimentin, UCHL-1, S-100, and lysozyme. Hodgkin's disease could be divided into three groups: the first group was LN-1+/L26+/vimentin-, the second LN-1-/L26+/vimentin+, and the third LN-1-/L26-/vimentin+). Sixteen cases of follicular lymphomas were also examined and were all positive for LN-1 and L26 and negative for vimentin. Thus the vimentin negativity of the first group, including 7 nodular lymphocyte-predominant cases, gives further evidence of their germinal center B-cell origin. Since vimentin is expressed mainly in the immature stage of B-lymphocytes, the second group of Hodgkin's disease may represent immature B-cell Hodgkin's disease. In the third group, vimentin was present in Reed-Sternberg's (RS) and Hodgkin's (H) cells in 45 of the 48 cases (92.5%). In none of 48 cases were these cells positive for S-100 or lysozyme, but strong vimentin-positivity still suggested monocytic or histiocytic origin. The results of our study suggest, at least, divergent origin of RS's and H's cells.

Vitronectin shows complement-independent binding to isolated keratin filament aggregates

Keratinocyte cell death, whether produced by skin disease or by physiologic apoptosis in normal skin, may result in formation of dermal keratin bodies, consisting mainly of keratin intermediate filament aggregates. Vitronectin, a multifunctional plasma and tissue glycoprotein, which inhibits the complement membrane attack complex and promotes cell attachment and spreading, is, like amyloid P component, associated with keratin bodies in vivo. To investigate a potential role for vitronectin in the removal of keratin bodies, we studied the interaction of vitronectin with keratin intermediate filaments in normal human skin and in Hep-2 cells, as well as with isolated keratin intermediate filament aggregates in vitro. Following pre-incubation of skin sections and Hep-2 cells with normal human serum (as a source of vitronectin), cytoplasmic staining of keratinocytes and of cytoskeletal filaments in Hep-2 cells was observed by immuno-fluorescence staining with polyclonal and monoclonal anti-vitronectin antibodies. Vitronectin binding to keratin intermediate filament aggregates extracted from normal human epidermis was demonstrated by immunofluorescence and by immunoblotting, and was not dependent on complement activation, because it occurred even when heat-inactivated human serum or C4-deficient serum was used as a source of vitronectin. Amyloid P component shows Ca++- dependent binding to keratin intermediate filament aggregates. does not involve amyloid P component because it occurred when binding of the latter protein was inhibited by EDTA buffer. Moreover, purified vitronectin also bound to keratin intermediate filament aggregates in immunofluorescence studies. Vitronectin binding to keratin intermediate filaments may play a role both in limiting complement-mediated tissue damage (because keratin bodies may activate complement) and in promoting removal of keratin bodies by fibroblasts and/or macrophages.

Intermediate filament proteins immunologically related to desmin in astrocytes: a study of chicken spinal cord by two-dimensional gel electrophoresis and immunoblotting

Co-migration experiments by two-dimensional SDS-PAGE using chicken spinal cord extracts and desmin purified from chicken gizzard showed that desmin is not present in spinal cord. However, by the immunoblotting procedure, desmin antibodies recognized 3 spinal cord antigens with different molecular weights and isoelectric points than desmin and the glial fibrillary acidic (GFA) protein. These antigens which also reacted with GFA protein antibodies were not identified in chicken gizzard extracts. The reactivity of the antigens with a monoclonal antibody recognizing an epitope common to most intermediate filament proteins (1) suggests that immunostaining of astrocytes with desmin antibodies (2, 3) is due to the presence of new intermediate filament proteins immunologically related to desmin.

Differential binding of natural monoclonal antibodies to the surface of fixed or living cells

A few hundred monoclonal antibodies derived from normal mice were tested for binding to cell surface antigens in one T-cell hybridoma and one I-Ek-transfected fibroblast cell line. The assay, which is suitable for large screenings, used glutaraldehyde-fixed cells followed by immunoenzymatic detection of immunoglobulin. Of the 331 antibodies tested, 75 showed significant binding, not only on these cells, but also on a macrophage, fibroblast, thymoma, and pre-B cell line, and on normal syngeneic and allogeneic thymocytes. If the assay was modified so as to use live cells in a simplified ELISA on living cells, only 10 of 253 antibodies were found to be positive with the T-cell hybridoma line and 7 with the transfected fibroblast cell line. In both sets of conditions, about 75% of the positive antibodies were found to be 'multireactive' after being tested on a panel of antigens. In contrast, conventional 'immune antibodies' to cell surface antigens could be tested by routine methods in either type of assay. We conclude that, while glutaraldehyde fixation does not affect the reactivity of conventional antibodies, this technique is inappropriate for testing the binding of natural antibodies to cell surface antigens.

Analysis of cell division using fluorescently labeled actin and myosin in living PtK2 cells

Actin and the light chains of myosin were labeled with fluorescent dyes and injected into interphase PtK2 cells in order to study the changes in distribution of actin and myosin that occurred when the injected cells subsequently entered mitosis and divided. The first changes occurred when stress fibers in prophase cells began to disassemble. During this process, which began in the center of the cell, individual fibers shortened, and in a few fibers, adjacent bands of fluorescent myosin could be seen to move closer together. In most cells, stress fiber disassembly was complete by metaphase, resulting in a diffuse distribution of the fluorescent proteins throughout the cytoplasm with the greatest concentration present in the mitotic spindle. The first evidence of actin and myosin concentration in a cleavage ring occurred at late anaphase, just before furrowing could be detected. Initially, the intensity of fluorescence and the width of the fluorescent ring increased as the ring constricted. In cells with asymmetrically positioned mitotic spindles, both protein concentration and furrowing were first evident in the cortical regions closest to the equator of the mitotic spindle. As cytokinesis progressed in such asymmetrically dividing cells, fluorescent actin and myosin appeared at the opposite side of the cell just before furrowing activity could be seen there. At the end of cytokinesis, myosin and actin were concentrated beneath the membrane of the midbody and subsequently became organized in two rings at either end of the midbody.

Amyloid P component binds to keratin bodies in human skin and to isolated keratin filament aggregates in vitro

Dermal keratin bodies, consisting mainly of keratin intermediate filament aggregates (KIFA) coated with IgM anti-KIF autoantibodies, are present in normal human skin and occur in increased quantities in certain skin diseases. Keratin bodies are normally rapidly removed, but in primary localized cutaneous amyloidosis (PLCA) they are converted by an unknown mechanism to amyloid. Amyloid P component (AP), a glycoprotein identical to, and derived from, the normal plasma protein serum amyloid P component (SAP), is present in all forms of amyloid including PLCA. We investigated the interaction between SAP, keratin bodies, and KIFA. Immunofluorescence staining of normal skin using fluoresceinated anti-SAP and rhodamine-conjugated anti-IgM, or AE-1/AE-3 anti-keratin antibodies followed by Texas Red-conjugated anti-mouse immunoglobulin, showed that 52% +/- 4 (mean +/- sem, n = 6) of keratin bodies bound anti-SAP. Similar findings were present in a biopsy from a patient with lichen planus. Isolated KIFA, prepared by 8M urea extraction of normal human epidermis or cultured keratinocytes, were preincubated with normal human serum as a source of SAP and then stained with fluoresceinated anti-SAP. Bright fluorescence seen when the incubation medium contained Ca++ was absent in the presence of ethylenediamine tetraacetic acid. Specific Ca++-dependent binding of SAP to KIFA was confirmed using immunoblotting. Binding of SAP to KIFA did not prevent their degradation following exposure to trypsin or alpha-chymotrypsin. Similarly, partial enzymatic digestion of KIFA did not abrogate their ability to bind SAP. Our findings, that SAP is associated with keratin bodies in skin and exhibits Ca++-dependent binding to KIFA in vitro, identify keratin filaments as a newly recognized ligand for SAP.

Evaluation of 111In labelled white blood cells by in vitro functional tests and electron microscopy. Comparison of three labelling methods

We have studied the influence of granulocyte labelling with commercially available 111In-oxine, tropolone (trop) or home made 111In-Mercapto pyridine (Merc) prepared by the method of Thakur (1985) on the cell structure by electron microscopy and on the cell function by enzymatic tests, random migration, chemotaxis, phagocytosis and bactericidal activity. The granulocytes were labelled with 400 microCi 111In-oxine in saline or 111In-trop or Merc in plasma. The effect of the chelating agents with and without addition of the tracer was studied (n = 4) with varying concentrations: 5-10 micrograms/ml oxine, 10-160 micrograms/ml trop and 1-4 micrograms/ml Merc. Chemotaxis and random migration were not affected by 111In-trop and clearly suppressed by 111In-oxine and Merc; the other tests were normal. The cell structure was disturbed by Merc. The labelling efficiency was excellent with oxine (90%), acceptable with trop (30%-80%) and poor with Merc (10%-25%). Without 111In, chemotaxis and random migration were normal up to a concentration of 80 micrograms/ml trop, 8.5 micrograms/ml oxine and 1 microgram/ml Merc. With addition of 111In, chemotaxis and random migration were unaffected up to 80 micrograms/ml by trop and markedly suppressed by Merc and oxine. It is concluded that labelling with 111In-trop assures intact cells.

Centrosomal proteins and lactate dehydrogenase possess a common epitope in human cell lines

A spontaneously arising rabbit anti-centrosome serum with strong human specificity, used to identify specific antigens in isolated centrosomes, was shown to react with several noncentrosomal proteins including a 36-kDa protein that appeared to be the major cellular antigen. To explore the immunological relationship between noncentrosomal and centrosomal antigens, immunoglobulins were affinity purified using the individual noncentrosomal antigens (from lymphoblastoma KE37 cells) and were tested for their capacity to bind to human centrosomes in situ and to proteins from isolated centrosomes. In this way, the 36-kDa antigen, an abundant cytosolic protein, was shown to share at least one antigenic determinant with high molecular weight centrosomal proteins. This antigen was further identified by mild proteolysis as the glycolytic enzyme lactate dehydrogenase. In all the analyzed human cell lines, the centrosomal staining in situ was correlated with a strong labeling of purified lactate dehydrogenase in immunoblots. Conversely, the absence of centrosomal staining in rodent cells was always correlated with the absence of lactate dehydrogenase labeling. These data suggest an evolutionary relationship between centrosomal proteins and this "housekeeping" enzyme.

Natural IgM and IgG autoantibodies to epidermal keratins in normal human sera. I: ELISA-titration, immunofluorescence study

This paper presents a study of autoantibodies (autoAB) to keratins and to epidermis by a double approach associating a specific immunoenzymatic technique and immunofluorescence. The existence of natural autoAB to keratins in all normal human sera was asserted and the heterogeneity of natural autoAB to the epidermis explored. By a sensitive enzyme-linked immunosorbent assay we detected natural IgM and IgG autoAB to keratin polypeptides extracted from human plantar stratum corneum (SC) in 60 randomly selected normal human sera. The interindividual variation factors of their titers were about 100X in IgM and 50X in IgG antikeratin (AK) autoAB. The IgM and IgG AK autoAB titers varied independently. By a semiquantitative indirect immunofluorescence assay we detected in these sera IgM and IgG autoAB that labeled normal epidermis according to various morphologic patterns. The IgG autoAB labeled SC and suprabasal layers (SBL) in 57.4% of sera, SC in 20.4% and SBL in 7.4%. The IgM autoAB labeled SC and SBL in 52% of sera, SC in 24%, SC and SBL plus basal layer (BL) in 18%, and SBL in 2%. Like the titers, the patterns of IgM and IgG autoAB to epidermis were found to be unrelated. The IgG AK autoAB titers were found to significantly correlate only with the IgG autoAB directed to SC + SBL; the IgM AK autoAB titers only with the IgM autoAB directed to SC + SBL + BL. This showed that these patterns of labeling are typical for AK autoAB and that autoAB to SC, which could not be related to AK autoAB, exist in some normal sera. Antikeratin and antiepidermis IgM autoAB titers were found to be strongly correlated to total amounts of IgM assayed by radial immunodiffusion, indicating that the synthesis of these natural IgM autoAB vary in the same way as that of general IgM synthesis. For the AK and antiepidermis IgG autoAB, however, the same correlation to total serum IgG was found to be much weaker.

Uterine smooth muscle cells in primary culture. Alterations in fine structure, cytoskeletal organization and growth characteristics

Smooth muscle cells (SMC) were enzymatically isolated from the myometrium of adult rat and human uteri and grown in primary culture. Cell fine structure and cytoskeletal organization were followed by transmission electron microscopy and cytochemical demonstration of actin filaments, microtubules and intermediate filaments, and initiation of DNA synthesis was investigated by thymidine autoradiography. During the first few days in culture the cells spread out on the substrate and went through a morphological transformation including loss of myofilaments followed by formation of an extensive rough endoplasmic reticulum and a large Golgi complex. Actin filaments aggregated in stress fibers spanning the entire length of the cells and microtubules and intermediate filaments formed a radiating system originating in the juxtanuclear region. In vivo, the SMC contained intermediate filaments reactive for desmin, but as early as the first day of culture expressed vimentin as well. For five days at least, all cells remained positive for both proteins, but the staining for desmin decreased while that for vimentin increased. This structural modification was accompanied by initiation of DNA synthesis, with a peak on day 3 (45-55% labeled nuclei). Subconfluent, growth-arrested primary cultures responded weakly to purified platelet-derived growth factor and serum, and in secondary cultures no response to the mitogenic stimulation was obtained. The observations indicate that uterine SMC cultivated in vitro undergo a transformation from contractile to synthetic phenotype, similar to the transformation described previously for arterial SMC under the same conditions. The proliferative potential of the uterine cells is, however, markedly lower. The findings support the notions that the transition into synthetic phenotype is a necessary but not sufficient requirement for initiation of DNA synthesis in SMC and that visceral and vascular SMC represent separate differentiation pathways.

In vitro translation of rat liver and Novikoff hepatoma cytokeratin mRNAs

Cytoplasmic poly(A)+ RNA was isolated from normal rat liver and Novikoff ascites hepatoma cells, translated in vitro using rabbit reticulocyte lysate system and the translational products were assayed by immunoprecipitation with antibodies specific for Novikoff hepatoma principal cytokeratins p39, p49 (a group of hepatic cytokeratins C, D, and E) and p56. The identity of the precipitated antigens was further confirmed by two-dimensional polyacrylamide gel electrophoresis. Only the Novikoff hepatoma poly(A)+ RNA contained translatable mRNA coding for the p39 cytokeratin while the p49 and p56 cytokeratins were translated from both the normal rat liver and Novikoff hepatoma poly(A)+ RNAs. Immunoprecipitations employing monoclonal antibody specific for p39 also recovered significant quantities of p56 and 49K cytokeratins, presumably due to oligomeric associations of these proteins with p39 immediately after in vitro synthesis. Similar results were observed after experiments with anti-p56 monoclonal antibody in which p39, not reactive with this antibody, was recovered in immunoprecipitates. Overall, the two-dimensional gel fluorograms of cytokeratins synthesized in vitro from NAH or liver poly(A)+ RNA are quite similar to isolated antigenic and cytokeratin profiles reported previously. These results suggest that overt posttranslational processing is not likely responsible for the diversity of cytokeratins observed in the liver.

Use of monoclonal antibodies for analyzing the distribution of the intermediate filament protein vimentin in human non-Hodgkin's lymphomas

A series of human non-Hodgkin's lymphomas was examined for immunoreactivity with monoclonal antibodies to the intermediate filament protein vimentin with the use of an avidin-biotin immunoperoxidase method. The lymphoid cell nature of each tumor was established with the use of a panel of monoclonal antibodies to lymphoid cell differentiation antigens. There were 28 B-cell and 2 T-cell lymphomas in the series; of the 30 tumors, 11 (37%) were immunoreactive for vimentin. There was no correlation between vimentin immunoreactivity and the histopathologic type of lymphoma. In some tumors, there was nonspecific stromal immunoreactivity for vimentin, but the neoplastic lymphocytes were not immunoreactive. The selective expression of vimentin in non-Hodgkin's lymphomas may be due to masking of the appropriate epitopes or to selective expression of the vimentin gene in certain tumors. On the basis of these results, monoclonal antibodies to vimentin appear to be of limited usefulness in establishing the diagnosis of non-Hodgkin's lymphoma.