p53 and cell cycle independent dysregulation of autophagy in chronic lymphocytic leukaemia.

BACKGROUND: Activation of wild-type p53 with the small molecule sirtuin inhibitor Tenovin-6 (Tnv-6) induces p53-dependent apoptosis in many malignant cells. In contrast, Tnv-6 reduces chronic lymphocytic leukaemia (CLL) cell viability with dysregulation of autophagy, without increasing p53-pathway activity. METHODS: Here, we have investigated whether a quiescent phenotype (unique to CLL) determines the Tnv-6 response, by comparing the effects of Tnv-6 on activated and proliferating CLL. We further studied if these responses are p53-dependent. RESULTS: Unlike quiescent cells, cell death in activated cultures treated with Tnv-6 was consistently associated with p53 upregulation. However, p53 acetylation remained unchanged, without caspase-3 cleavage or apoptosis on electron microscopy. Instead, cellular ultrastructure and protein profiles indicated autophagy inhibition, with reduced ubiquitin-proteasome activity. In specimens with mutant TP53 cultured with Tnv-6, changes in the autophagy-associated protein LC3 occurred independently of p53. Cells treated with Tnv-6 analogues lacking sirtuin inhibitory activity had attenuated LC3 lipidation compared with Tnv-6 (P0.01), suggesting that autophagy dysregulation occurs predominantly through an effect on sirtuins. CONCLUSION: These cell cycle and p53-independent anti-leukaemic mechanisms potentially offer novel therapeutic approaches to target leukaemia-sustaining cells in CLL, including in disease with p53-pathway dysfunction. Whether targets in addition to sirtuins contribute to autophagy dysregulation by Tnv-6, requires further investigation.

The beaded filament of the eye lens: an unexpected key to intermediate filament structure and function.

In 1959, an unusual filamentous polymer, now called the beaded filament, was described in the lens of the eye. The constituent proteins, assembly properties and functions of the beaded filament have been elusive. The recent publication of the sequences for two major lens filament proteins (CP49 and filensin) and the reconstitution in vitro of structures closely resembling beaded filaments, suggests that the beaded filament is related structurally to intermediate filaments (IFs). The association of the lenticular chaperones, the alpha-crystallins, with the filament contributes to the characteristic beaded morphology, as well as giving important clues to the function of this unusual filament in the lens. These recent results have several implications for IF function and assembly.

The coated pit and macropinocytic pathways serve distinct endosome populations.

Clathrin-coated vesicle endocytosis and macropinocytosis are distinct endocytic pathways demonstrable in several cell types including human epidermoid A431 cells (West, M.A., M.S. Bretscher, and C. Watts. 1989. J. Cell Biol. 109:2731-2739). Here we analyze the extent of mixing of macropinocytic endosome (macropinosome) content with that of conventional endosomes served by coated vesicle endocytosis. Using laser scanning confocal fluorescence microscopy we detected very little delivery of macropinosome content to either early or late endosomes-lysosomes as defined by labeling with transferrin or with LDL. Mixing of the contents of the macropinosomes and conventional endosomes was not induced by the addition of brefeldin A. Moreover, the morphology of macropinosomes was not grossly altered in the presence of brefeldin A, whilst in the same cells there were dramatic tubulation effects on conventional endosomes as reported by others. Although refractory to fusion with conventional endosomes, macropinosomes were nonetheless dynamic structures which sometimes exhibited vesiculo-tubular morphology in living cells and were capable of fusing with each other. We suggest that different endocytic mechanisms can give rise to distinct endosome populations.

The prohibitin family of mitochondrial proteins regulate replicative lifespan.

Cellular senescence is determined by multiple factors, including the genetic regulation of metabolism and responses to endogenous and exogenous stresses [1-4]. Recent studies implicate a limited number of gene products in elongating lifespan in yeast and Caenorhabditis elegans [2-4]; these include the C, elegans gene cik-1, a central regulator of metabolism [5], and yeast RAS2, which controls the response to ultraviolet irradiation and other stresses [3]. Another gene postulated to effect senescence is PHB1, the yeast homologue of prohibitin [3], a rodent gene initially identified as a potential regulator of growth arrest and tumour suppressor [6-8]. Highly conserved prohibitin homologues have been identified in mammals [9], Drosophila [10], C. elegans [9], plants [11] and yeast. A second mammalian gene, encoding BAP37, a protein with sequence similarity to prohibitin, is thought to be involved in lymphocyte function [9]. Here, we show that the nuclear-encoded mammalian prohibitin and BAP37 proteins are present in mitochondria, are co-expressed, and interact physically with each other. Deletion of the Saccharomyces cerevisiae homologues, PHB1 and PHB2, results in a decreased replicative lifespan and a defect in mitochondrial membrane potential. Our observations highlight the relationship between the metabolic efficiency of cells and the ageing process, and provide evidence for its evolutionary conservation.

Rac is required for constitutive macropinocytosis by dendritic cells but does not control its downregulation.

BACKGROUND: Dendritic cells use constitutive macropinocytosis to capture exogenous antigens for presentation on MHC molecules. Upon exposure to inflammatory stimuli or bacterial products such as lipopolysaccharide (LPS), macropinocytosis is dramatically downregulated as part of a developmental programme leading to dendritic cell maturation, migration and activation of T cells. It is not known, however, how macropinocytosis is sustained in dendritic cells in the absence of exogenous stimuli, nor how it is downregulated upon maturation. We have tested the possibility that one or more members of the Rho family of GTPases are involved in and control pinocytosis in dendritic cells. RESULTS: We established dendritic cell populations that show constitutive macropinocytosis that was downregulated by LPS treatment. Microinjection of immature cells with dominant-negative Rac (N17Rac1) or treatment with Clostridium difficile toxin B, the phosphoinositide 3-kinase (PI3-K) inhibitor wortmannin, or LPS all inhibited the formation of macropinosomes but, surprisingly, did not eliminate membrane ruffling. Microinjection of N17Cdc42 or the Rho inhibitor C3 transferase eliminated actin plaques/podosomes and actin cables, respectively, but had little effect on the formation of macropinosomes. Surprisingly, dendritic cells matured with LPS had equivalent or even somewhat higher levels of active Rac than immature cells. Moreover, microinjection of a constitutively active form of Rac (V12Rac1) into mature dendritic cells did not reactivate macropinocytosis. CONCLUSIONS: Rac has an important role in the constitutive formation of macropinosomes in dendritic cells but may be required downstream of membrane ruffling. Furthermore, regulation of Rac activity does not appear to be the control point in the physiological downregulation of dendritic cell pinocytosis. Instead, one or more downstream effectors may be modulated to allow Rac to continue to regulate other cellular functions.

Gadd45 is a nuclear cell cycle regulated protein which interacts with p21Cip1.

GADD45 was originally identified as a cDNA clone induced by growth arrest and DNA damage. We show that Gadd45 is a nuclear protein, widely expressed in normal tissues, particularly in quiescent cellular populations. Using cell synchronisation methods we show that Gadd45 levels are highest in the G1 phase of the cell cycle, and are greatly reduced during S phase. Immunoprecipitation of Gadd45 from mammalian cells reveals that it is tightly associated with a protein which reacts with antibodies to the cyclin dependent kinase inhibitor p21Cip1. Binding of recombinant Gadd45 protein to overlapping p21Cip1 peptides in ELISA assays and use of the yeast two hybrid assay show that Gadd45 directly interacts with this cell cycle inhibitor. These data suggest that Gadd45 may act in the regulation of the cell cycle. It is postulated that the interactions of Gadd45 with both p21Cip1 and PCNA are important for the modulation of cell cycles, and for the inhibition of DNA replication.

The trigger to cell death determines the efficiency with which dying cells are cleared by neighbours.

Phagocytosis of apoptotic cells is required to prevent tissue injury. Professional phagocytes, such as monocyte-derived macrophages, are highly efficient scavengers of apoptotic cells but their presence cannot always be relied on; in that case, removal of effete cells is accomplished by helpful neighbours. This study describes differences in the efficiency with which apoptotic cells of the same type, but dying in response to different triggers, are engulfed; this varies from engulfment that is so proficient few or no unengulfed apoptotic cells are found, to engulfment that is so delayed apoptotic cells have become secondarily necrotic at the point of engulfment. In all cases the efficiency of engulfment is determined at least in part by the dying cells themselves. p53- and Bax-transfected kidney epithelial (293) cells (transiently transfected using a non-toxic method) were engulfed so proficiently by homotypic neighbours that cells did not show evidence of engagement of the apoptotic programme (chromatin condensation and TUNEL positivity) until engulfment had taken place. Engulfment nonetheless required activation of at least initiator caspases. 293 cells induced to apoptose by other means (etoposide and staurosporine treatment) were not so efficiently ingested: unengulfed apoptotic cells were consistently revealed at all doses and time points, even when treated cells were mixed with healthy, non-treated 293 cells. These data make it extremely unlikely that the fraction of viable, unaffected neighbours determines the efficiency with which engulfment proceeds. Furthermore, 293 cells treated with etoposide or staurosporine were differentially appealing both to homotypic neighbours and to cells in the professional phagocyte lineage (THP-1 cells). If different apoptotic stimuli programme cells to be recognised with different efficiencies, pathways to apoptosis may be injury limiting to greater or lesser degrees.

Structural characterisation of two forms of procyclic acidic repetitive protein expressed by procyclic forms of Trypanosoma brucei.

A procyclic acidic repetitive protein (PARP) fraction was purified from long-term cultures of Trypanosoma brucei procyclic forms by a solvent-extraction and reverse phase chromatography procedure. The PARP fraction yielded small quantities of a single N-linked oligosaccharide with the structure Man alpha1-6(Man alpha1-3)Man alpha1-6(Man alpha1-3)Manbeta1-4GlcNAcbeta1-4GlcNAc (Man5GlcNAc2). Fractionation of PARP on Con A-Sepharose revealed that the majority (80 to 90%) of the PARP fraction did not bind to Con A and was composed of the parpA alpha gene product that contains repeats of -Glu-Pro-Pro-Thr- (GPEET-PARP) and that lacks an N-glycosylation site. This form of PARP has not been previously identified at the protein-level. The minor Con-A-binding fraction was shown to be rich in the previously described form of PARP, encoded by the parpAbeta and/or parpB alpha genes, that contains a -Glu-Pro- repeat domain (EP-PARP) and an N-glycosylation site. Analysis of longer and shorter-term cultures suggested that procyclic cells initially express predominantly EP-PARP that is gradually replaced by GPEET-PARP. Both forms of PARP were shown to contain indistinguishable glycosylphosphatidylinositol (GPI) membrane anchors, where the conserved GPI core structure is substituted by heterogeneous sialylated branched polylactosamine-like structures that are predicted to form a dense surface glycocalyx above which the polyanionic -Glu-Pro-Pro-Thr- and -Glu-Pro- repeat domains are displayed. The phosphatidylinositol (PI) component of the GPI anchor was shown to be a mixture of 2-O-acyl-myo-inositol-1-HPO4-(sn-1-stearoyl-2-lyso-glycerol) and 2-O-acyl-myo-inositol-1-HPO4-(sn-1-octadecyl-2-lyso-glycerol), where the acyl chain substituting the inositol ring showed considerable heterogeneity. Mass spectrometric and light scattering experiments both suggested an average mass of approximately 15 kDa for GPEET-PARP, with individual glycoforms ranging from about 12 kDa to 20 kDa, that is consistent with its amino acid and carbohydrate composition. A measured translational diffusion coefficient of 3.9 x 10(7) cm2 s(-1) indicates that this molecule has a highly elongated shape. The possible functions of these unusual glycoproteins are discussed.

Changes in the nucleolar and coiled body compartments precede lamina and chromatin reorganization during fibre cell denucleation in the bovine lens.

Nuclear elimination accompanies differentiation in such specialized cell types such as erthyrocytes and lens fibre cells. It also accompanies apoptosis which has suggested that similar processes could operate in both. Denucleation occurs in the lens in order to reduce light scatter and this process is often disrupted in cataract. Using the adult bovine lens as a model system, nuclear changes accompanying denucleation are described with particular emphasis on the lamina, nucleolar and coiled body compartments in lens nuclei. Nuclear shape, chromatin reorganization and chromatin breakdown were also monitored to correlate the timing of events. Rearrangement of both A- and B-type nuclear lamins occurred in parallel with chromatin condensation and preceded changes in nuclear shape. The earliest changes detected in this study occurred in the coiled body and nucleolar compartments using coilin and fibrillarin antibodies respectively, suggesting that a shutdown in transcription is an early event in denucleation. Fibrillarin redistributed from an open floret pattern to several condensed spots which gradually decreased in intensity and eventually disappeared. Coilin, however, was localized in several microfoci prior to being reorganized into fewer larger foci. Prior to chromatin condensation, coilin redistributed to the nucleolar compartment and was absent from nuclei where chromatin had begun to condense. Such nuclei were positive by TUNEL staining. In contrast to the nucleus, mitochondrial degradation in lens fibre cells was a rapid process and involved a relatively sharp transition between positive and negative fibre cells for two mitochondrial specific markers, BAP 37 and prohibitin. A link between the changes in the nuclear lamina and chromatin with the initiation of mitochondrial fragmentation was also observed. Therefore, it is possible that the signal for the initiation of denucleation could originate from the mitochondria as proposed for apoptosis. Differences between apoptosis and lens fibre cell denucleation were noted and included the timescale of nuclear changes as well as the persistence of a nuclear remnant. These studies suggest that transcriptional shutdown precedes lamina reorganization and chromatin breakdown during lens fibre cell denucleation.

Distinct compartmentalization of TGN46 and beta 1,4-galactosyltransferase in HeLa cells.

The Golgi proteins, TGN46 and GalT, were characterized in human HeLa cells using specific polyclonal and monoclonal antibodies. A bacterially expressed soluble recombinant TGN46 protein was used to raise rabbit polyclonal antibodies and used to probe HeLa cell extracts. Human TGN46 had an apparent molecular mass of 100 to 120 kDa which reflects extensive glycosylation. Epifluorescence light microscopy indicated substantial colocalization of TGN46 and GalT. However, confocal laser microscopy and three-dimensional reconstruction of double-labeled HeLa cells revealed large areas of colocalization but also specific differences in the distribution of these two proteins within the Golgi apparatus. Importantly, quantitative immunoelectron microscopy showed that there was little overlap between the distribution of GalT and TGN46. Approximately 75% of GalT was in the Golgi stack, whereas 80% of TGN46 was detected in tubules. Distinct GalT-positive regions within the Golgi cisternal stack were not labeled for TGN46.

Filensin is proteolytically processed during lens fiber cell differentiation by multiple independent pathways.

Filensin is a lens-specific intermediate filament protein, expressed in the lens fiber cells but not the lens epithelium. Using antibodies to filensin and the other lens intermediate filament proteins, vimentin and CP49, the codistribution of filensin with CP49 and independence of this network from the vimentin network was confirmed. Monoclonal and polyclonal antibodies to peptides and specific subdomains of filensin were used to follow changes in the subcellular distribution of filensin during bovine lens fiber cell differentiation. Filensin is shown to be extensively processed during lens fiber cell differentiation to give protein fragments derived from distinct protein domains, one corresponding to the N-terminal non-alpha-helical/and rod domain and the other to the C-terminal non-alpha-helical tail domain. Immunoblotting analysis using anti-filensin peptide polyclonal antibodies suggested that the two fragment sets arose separately. Residues 331 to 430 in filensin have been identified as an important region in the processing pathway(s). Our results clarify previous confusion in the literature regarding the processing of filensin which arose because of the similar relative electrophoretic mobilities by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) of the different fragment sets. The predicted secondary structure characteristics of the different domains of filensin suggests different functions for the two fragment sets to give filensin a dual role in the lens. This suggestion is supported by the subtly different subcellular distributions in the peripheral and mature fiber cells of the two filensin fragment sets.

Nuclear localization of protein phosphatase 5 is dependent on the carboxy-terminal region.

Endogenous and overexpressed protein phosphatase 5 (PP5) localizes to the nucleus and cytoplasm of HeLa cells, while the overexpressed TPR domain of PP5 is restricted to the cytoplasm. Deletion and mutational analysis of human PP5 demonstrates that the C-terminal amino acids 420-499 are essential for nuclear localization and PP5 activity is not required. Since the phosphatase domain terminates at 473, these studies suggest that the highly conserved section (476-491) with the eukaryotic consensus FXAVPHPXPhiXPMAYAN is required for nuclear localization of PP5. Bacterially expressed PP5 is inhibited by several tumor promoters but not by the anticancer drug fostriecin.

Transformation of monomorphic Trypanosoma brucei bloodstream form trypomastigotes into procyclic forms at 37 degrees C by removing glucose from the culture medium.

African trypanosomes have been shown previously to undergo efficient transformation from bloodstream forms to procyclic (insect dwelling) forms in vitro by adding citrate and/or cis-aconitate to the culture medium and lowering incubation temperature to 27 degrees C. In this paper, it is shown that strain 427 monomorphic bloodstream forms of Trypanosoma brucei grown in axenic culture at 37 degrees C can be transformed to procyclic forms by simply replacing the glucose carbon source in the culture medium with glycerol. The removal of glucose from the medium results in the loss of the variant surface glycoprotein, the acquisition of cell surface procyclic acidic repetitive protein, the synthesis of procyclic-specific glycosylphosphatidylinositol precursors and the acquisition of substantial resistance to salicyl hydroxamic acid and glycerol within 72 h. A procyclic-specific cytoskeletal protein, known to be a marker of the late stage of transformation, is fully expressed by 96 h but full trans-sialidase activity appears only after 18-30 days. The transformation process described here is slower and less efficient than that previously described for monomorphic trypanosomes, using citrate and/or cis-aconitate and temperature shift as triggers. However, the separation of the transformation process from these stimuli is significant and the effects of glucose deprivation described here may reflect some of the events that occur in vivo in the tsetse fly midgut, where glucose levels are known to be very low.

Molecular characterisation of mitochondrial and cytosolic trypanothione-dependent tryparedoxin peroxidases in Trypanosoma brucei.

In trypanosomatids, removal of hydrogen peroxide and other aryl and alkyl peroxides is achieved by the NADPH-dependent trypanothione peroxidase system, whose components are trypanothione reductase (TRYR), trypanothione, tryparedoxin (TRYX) and tryparedoxin peroxidase (TRYP). Here, we report the cloning of a multi-copy tryparedoxin peroxidase gene (TRYP1) from Trypanosoma brucei brucei encoding a protein with two catalytic VCP motifs similar to the cytosolic TRYP from Crithidia fasciculata. In addition, we characterise a novel single copy gene encoding a second tryparedoxin peroxidase (TRYP2). TRYP2 shows 51% similarity to TRYP1, possesses a putative mitochondrial import sequence at its N-terminus and has a variant IPC motif replacing the second VCP motif implicated in catalysis in other 2-Cys peroxiredoxins. TRYP1 and TRYP2 were expressed in Escherichia coli, and the purified recombinant proteins shown to utilise hydrogen peroxide in the presence of NADPH, trypanothione, TRYR and TRYX from T. brucei, similar to the C. fasciculata cytoplasmic system. Western blots showed that TRYX, TRYP1 and TRYP2 are expressed in both bloodstream and procyclic forms of the life cycle. To determine the precise localisation of TRYX, TRYP1 and TRYP2 in the parasite, polyclonal antibodies to purified recombinant TRYX and TRYP1 and monoclonal antibody to TRYP2 were generated in mice. In-situ immunofluorescence and immunoelectron microscopy revealed a colocalisation of TRYX and TRYP1 in the cytosol, whereas TRYP2 was principally localised in the mitochondrion.

Rhabdomyosarcoma rho(0) cells: isolation and characterization of a mitochondrial DNA depleted cell line with 'muscle-like' properties.

Mutations of mitochondrial DNA are a significant cause of neuromuscular disease. Pathological mutant mitochondrial DNA has been studied in control nuclear backgrounds. These experiments entailed transfer of patient-derived mitochondria to rho(0) cells that lack mtDNA. A limitation of these studies has been the fact that the control nuclear backgrounds were unrelated to the affected tissues of patients. Therefore a rhabdomyosarcoma cell line that has 'muscle-like' properties was tested to determine whether it could be depleted of mtDNA. A human rhabdomyosarcoma cell line was treated with the DNA intercalating dye ethidium bromide (3, 8-diamino-5-ethyl-6-phenylphenanthridinium bromide) for 45 days. The treatment induced complete and permanent loss of mitochondrial DNA (rho(0)) in the rhabdomyosarcoma cells, as mtDNA remained undetectable after 8 months of growth in medium without drug. Crucially, the rhabdomyosarcoma rho(0) cells retained the ability to differentiate into myotubes with expression of muscle specific isoenzymes. The rhabdomyosarcoma rho(0) cell line provides a model system for studying pathological mutant mtDNA in cells that more closely resemble human muscle than the hitherto available human rho(0) cell lines.

Membrane and communication properties of tissue cultured human lens epithelial cells.

Explant cultures were established from capsule/epithelium preparations from both normal and cataractous lenses to investigate properties of human lens epithelial cells. The cultured cells were found to have similar membrane potentials to whole human lenses and isolated epithelia, and similar free ionic concentrations of potassium, sodium, and calcium (131 mM, 17 mM and 0.8 microM respectively) to whole human lenses. The free ionic concentrations were measured in both cases using neutral resin-filled electrodes. Cellular communication was investigated using electrical (two internal microelectrodes) and dye injection techniques. The electrical resistance of a confluent cell monolayer was approximately 4 M omega when the voltage measuring and current passing microelectrodes were in neighbouring cells or several cell diameters apart. Additionally, Lucifer Yellow dye injected into one cell spread rapidly over a wide area of cells. The cells thus appear to be extremely well coupled. Electrical communication could be disrupted by internal acidification (following exposure to 100% CO2), exposure to 1 mM octanol and by membrane depolarization. In the latter case the blockade was only partial. All uncoupling methods proved to be reversible. The diffusion of Lucifer Yellow dye was also inhibited by internal acidification and exposure to octanol. The sensitivity of both dye and electrical coupling to internal acidification and exposure to octanol is similar to that observed in hepatocytes and other tissues, whereas the effect on cell communication induced by changing the resting potential appears to occur only in a few cell types such as those of embryonic origin.

Identification of a novel intercellular structure in late-stage differentiating lens cells.

This study describes a novel intercellular structure in the adult bovine lens. In differential interference contrast images, the structure has the shape of a thickened torus or 'bagel' of 3-9 micrometer diameter and is contributed equally by 2 adjacent fibre cells. Due to its shape and location reaching into 2 neighbouring cells, the novel structure was termed 'intercellular torus' or 'bagel'. Intercellular bagels are present in a subset of late-stage lens fibre cells of the intermediate cortex, a considerable time after the cytoplasmic organelles have been broken down and the pyknotic nuclear remnants have disappeared. They are not present in deeper fibres. Our experiments show that intercellular bagels do not stain positive for DNA or RNAs, but are rich in lipids. Preliminary data indicate that the intercellular bagels contain calcium, suggesting that they might act as a place of transient Ca(2+) storage or sequestration after the intracellular organelles, such as the endoplasmatic reticulum, nuclear envelope, Golgi apparatus and mitochondria have been eliminated from the lens fibres during terminal differentiation.

Microtubules rich in modified alpha-tubulin characterize the tail processes of motile fibroblasts.

The organisation of microtubules rich in post-translationally modified alpha-tubulin has been investigated in a fibroblast cell line (NIH-3T3-T15) that can be reversibly transformed. An immunofluorescence microscopy study of the static non-transformed cells has revealed a central distribution of wavy microtubules showing post-translational modifications. When transformed there is a marked increase in cell motility and the appearance of long thin cytoplasmic 'tails'. These tails have been found to contain conspicuous bundles of post-translationally modified microtubules that run down the length of the processes and terminate close to the plasmalemma. Both detyrosinated and acetylated alpha-tubulin are present as major species in these modified microtubules. Such a pattern of modified microtubules is only occasionally seen in the untransformed NIH-3T3-T15 cells. We have also found them to be present in other transformed fibroblast lines. The presence of bundles of microtubules rich in modified alpha-tubulin in the cell tails is correlated with a marked reduction in the numbers of F-actin stress fibres. The possible role of these modified stable microtubules in cell motility is discussed.

Scatter factor affects major changes in the cytoskeletal organization of epithelial cells.

The effects of scatter factor on the cytoskeleton of MDCK and PtK2 cells are described. During the first 6 h after the addition of scatter factor, MDCK cells were found to increase their projected areas twofold, as well as the number and size of their F-actin stress fibers. In contrast PtK2 cells showed no change in their projected areas or in their stress fiber content. However, when both MDCK and PtK2 cells began to separate and scatter after approximately 6 h, the size and number of stress fibers was found to decrease considerably. Unscattered PtK2 cells and cells treated with scatter factor which had yet to scatter showed focal contacts present over the whole ventral surface, as judged by staining for both vinculin and talin. After treated cells separated, both vinculin and talin staining were mainly present in focal contacts on the ventral surfaces of the cell bodies and the distal ends of the processes. However, the cell processes showed few focal contacts along their lengths. The distribution of microtubules and vimentin and keratin intermediate filaments also did not change significantly until scattering had occurred. After cell separation, the processes were always packed with microtubules which were often, but not always, rich in detyrosinated alpha-tubulin and often, but not always, packed with intermediate filaments. All these changes in cytoskeletal organization are consistent with the adoption of a much more motile phenotype. The changes found are compared with those brought about by transformation.

Stable and slow-turning-over microtubules characterize the processes of motile epithelial cells treated with scatter factor.

The turnover of microtubules was studied in the processes of PtK2 cells, after treatment with the cytokine scatter factor (SF), using micro-injected biotin-tubulin as a reporter of new microtubule growth. Cells treated with SF became dispersed and fibroblast-like in morphology, showing one or more elongated processes. These processes contained bundles of microtubules, a significant proportion of which did not turn over during incubation times of up to an hour. Short broken pieces of microtubule were frequently found in all parts of the cell, particularly after longer incubation times, suggesting that more-stable microtubules were cut into pieces, which were subsequently degraded. From about half an hour after injection small tangles of stable microtubules were found. Some of these were clearly within the cell bodies. Others were usually larger in size and seemingly located outside the injected cells. These were considered to have formed part of small 'feet' presumed to be broken off during the retraction of trailing processes. The microtubules within the processes were resistant to the effects of both microtubule-depolymerizing drugs and cold under conditions where the processes were maintained. When these microtubules disappeared as the result of longer drug treatment the processes were also lost although, rarely, short processes lacking microtubules were found. It is concluded that the stable microtubules have a major role in process maintenance, although one that is indirect rather than a structural relationship.