Functional differences of tau isoforms containing 3 or 4 C-terminal repeat regions and the influence of oxidative stress.

We report functional differences between tau isoforms with 3 or 4 C-terminal repeats and a difference in susceptibility to oxidative conditions, with respect to the regulation of microtubule dynamics in vitro and tau-microtubule binding in cultured cells. In the presence of dithiothreitol in vitro, a 3-repeat tau isoform promotes microtubule nucleation, reduces the tubulin critical concentration for microtubule assembly, and suppresses dynamic instability. Under non-reducing conditions, threshold concentrations of 3-repeat tau and tubulin exist below which this isoform still promotes microtubule nucleation and assembly but fails to reduce the tubulin critical concentration or suppress dynamic instability; above these threshold concentrations, amorphous aggregates of 3-repeat tau and tubulin can be produced at the expense of microtubule formation. A 4-repeat tau isoform is less sensitive to the oxidative potential of the environment, behaving under oxidative conditions similarly to the 3-repeat isoform under reducing conditions. Under conditions of oxidative stress, in Chinese hamster ovary cells stably expressing either 3- or 4-repeat tau, 3-repeat tau disassociates from microtubules more readily than the 4-repeat isoform, and tau-containing high molecular weight aggregates are preferentially observed in lysates from the Chinese hamster ovary cells expressing 3-repeat tau, indicating greater susceptibility of 3-repeat tau to oxidative conditions, compared with 4-repeat tau in vivo.

NMR solution studies of hamster galectin-3 and electron microscopic visualization of surface-adsorbed complexes: evidence for interactions between the N- and C-terminal domains.

Galectin-3, a beta-galactoside binding protein, contains a C-terminal carbohydrate recognition domain (CRD) and an N-terminal domain that includes several repeats of a proline-tyrosine-glycine-rich motif. Earlier work based on a crystal structure of human galectin-3 CRD, and modeling and mutagenesis studies of the closely homologous hamster galectin-3, suggested that N-terminal tail residues immediately preceding the CRD might interfere with the canonical subunit interaction site of dimeric galectin-1 and -2, explaining the monomeric status of galectin-3 in solution. Here we describe high-resolution NMR studies of hamster galectin-3 (residues 1--245) and several of its fragments. The results indicate that the recombinant N-terminal fragment Delta 126--245 (residues 1--125) is an unfolded, extended structure. However, in the intact galectin-3 and fragment Delta 1--93 (residues 94--245), N-terminal domain residues lying between positions 94 and 113 have significantly reduced mobility values compared with those expected for bulk N-terminal tail residues, consistent with an interaction of this segment with the CRD domain. In contrast to the monomeric status of galectin-3 (and fragment Delta 1--93) in solution, electron microscopy of negatively stained and rotary shadowed samples of hamster galectin-3 as well as the CRD fragment Delta 1--103 (residues 104--245) show the presence of a significant proportion (up to 30%) of oligomers. Similar imaging of the N-terminal tail fragment Delta 126--245 reveals the presence of fibrils formed by intermolecular interactions between extended polypeptide subunits. Oligomerization of substratum-adsorbed galectin-3, through N- and C-terminal domain interactions, could be relevant to the positive cooperativity observed in binding of the lectin to immobilized multiglycosylated proteins such as laminin.

Extracellularly truncated desmoglein 1 compromises desmosomes in MDCK cells.

The formation and stability of epithelial tissue involves cell adhesion and the connection of the intermediate filaments of contiguous cells, mediated by desmosomes. The cadherin family members Desmocollins (Dsc) and Desmogleins (Dsg) mediate desmosome extracellular adhesion. The main intracellular molecules identified linking Dscs and Dsgs with the intermediate filament network are Plakoglobin (PG), Plakophilins (PPs) and Desmoplakin (DP). Previous studies on desmosome-mediated adhesion have focused on the intracellular domains of Dsc and Dsg because of their capacity to interact with PG, PPs and DP. This study examines the role of the extracellular domain of Dsg1 upon desmosome stability in MDCK cells. Dsg1 was constructed containing an extracellular deletion (Dsg delta 1EC) and was expressed in MDCK cells. A high expressor Dsg delta 1EC/MDCK clone was obtained and analysed for its capacity to form desmosomes in cell monolayers and when growing under mechanical stress in three-dimensional collagen cultures. Phenotypic changes associated with the ectopic expression of Dsg1 delta EC in MDCK cells were: disturbance of the cytokeratin network, a change in the quality and number of desmosomes and impairment of the formation of cysts in suspension cultures. Interestingly, Dsg1 delta EC was not localized in desmosomes, but was still able to maintain its intracytoplasmic interaction with PG, suggesting that the disruptive effects were largely due to PG and/or PP sequestration.

Aspects of the structure and assembly of desmosomes.

Desmosomes are found principally in epithelial cells and consist of disc-like plaques, the extracellular face of which is paired with that of a neighbouring cell. There is increasing evidence that desmosomes are adhesive structures, and that two types of desmosomal glycoproteins, the desmogleins (Dsg) and desmocollins (Dsc) both Ca(2+)-binding cadherin-like molecules, perform this role in adhesion through interaction of their extracellular domains. A number of isoforms of Dsg and Dsc are present in specific tissues. The cytoplasmic side of the plaque is attached to intermediate filaments through desmoplakin, a major plaque protein. Also associated with desmosomes are plakoglobin and beta-catenin, suggesting that the adhesive function of desmosomes might be mediated by signal transduction. Formation of desmosomes can be studied by growing epithelial cells in low-Ca2+ medium (LCM, < 0.1 mM), where desmosomal proteins are either synthesized but not assembled, or form partially assembled but unstable half-desmosomes. Addition of Ca2+ (to about 2mM) initiates cell contact and, in the case of half-desmosomes, leads to stabilization by incorporation into membranes and formation of typical paired structures. In cases where such pre-assembled structures are not formed, recruitment of desmosomal proteins appears to occur by vesicular transport of desmocollins and desmogleins to the cell surface, where association is made with plakoglobin and later, with desmoplakin. Although much remains to be learned of the assembly process, specific interacting domains of the molecular components are being recognized. Desmosome assembly is part of a coordinated pattern of junction formation which accompanies the establishment of cell polarity, resulting in differentiation of apical and basolateral cell surfaces. Desmosomes are now being regarded, not as static and inert structures, but as membrane specializations linked to systems involved in cell-cell communication as well as adhesion.

Antisense expression of a desmocollin gene in MDCK cells alters desmosome plaque assembly but does not affect desmoglein expression.

The desmocollins are one of two types of putative adhesive proteins present in the desmosome type of cell junctions, the other type being the desmogleins; both are members of the cadherin superfamily. Each type of desmosomal cadherin occurs as a number of isoforms which have differing tissue distribution; within stratifying epithelia some isoforms occur only suprabasally. We have sought to analyse desmocollin function by reducing the amount of protein using antisense gene expression in the widely studied Madin-Darby canine kidney (MDCK) cell line. Although this is a simple epithelial cell line, we show by Northern blot analysis that it expresses multiple isoforms of the desmosomal cadherins. Desmocollins DSC2 and DSC3 and desmogleins DSG2 and DSG3 (the pemphigus vulgaris antigen PVA) were detected, but DSC1 and DSG1, which are present exclusively in the suprabasal layers of the epidermis, were absent. The major desmocollin isoform was the type 2 (DSC2). A DSC2 clone isolated from a MDCK cDNA library had the same cell adhesion recognition sequence (Phe-Ala-Thr) as human, bovine and mouse type 2 isoforms. This sequence appears diagnostic for the three desmocollin isoforms. This cDNA clone was used to isolate a genomic DSC2 clone; antisense expression of this clone in MDCK cells resulted in a drastic reduction of desmocollin protein as judged by Western blots; Dsc3 was not upregulated to compensate for the loss of Dsc2. This antisense expression significantly altered desmosome assembly. There was a loss of punctate staining evident when using a desmosome plaque protein (desmoplakin) antibody. Electron microscopy revealed that there was a reduction in the number of desmosomes and a notable increase in the asymmetry of plaques between adjacent cells. Immunolabelling showed that similar levels of desmogleins and E-cadherin were present. Immunoelectron microscopy also showed that many vesicular structures were labelled, at intervals along the lateral membranes between cells. The distinctive loose organization of the remaining desmosomes may originate in modifications to the targeting and incorporation of proteins into fully assembled plaques. Other junctions were unaffected and the cells maintained their integrity as a confluent monolayer.

Coexpression of both types of desmosomal cadherin and plakoglobin confers strong intercellular adhesion.

Desmosomes are unique intercellular junctions in that they invariably contain two types of transmembrane cadherin molecule, desmocollins and desmogleins. In addition they possess a distinct cytoplasmic plaque structure containing a few major proteins including desmoplakins and the armadillo family member plakoglobin. Desmosomal cadherins are putative cell-cell adhesion molecules and we have tested their adhesive capacity using a transfection approach in mouse L cells. We find that L cells expressing either one or both of the desmosomal cadherins desmocollin 2a or desmoglein 1 display weak cell-cell adhesion activity that is Ca2+-dependent. Both homophilic and heterophilic adhesion could be detected. However, co-expression of plakoglobin with both desmosomal cadherins, but not with desmoglein 1 alone, resulted in a dramatic potentiation of cell-cell aggregation and the accumulation of detergent-insoluble desmosomal proteins at points of cell-cell contact. The effect of plakoglobin seems to be due directly to its interaction with the desmosomal cadherins rather than to its signalling function. The data suggest that the desmosome may obligatorily contain two cadherins and is consistent with a model in which desmocollins and desmogleins may form side by side heterodimers in contrast to the classical cadherins that are homodimeric. Plakoglobin may function by potentiating dimer formation, accretion of dimers to cell-cell contact sites or desmosomal cadherin stability.

Visualisation by electron microscopy of the unique part of the cytoplasmic domain of a desmoglein, a cadherin-like protein of the desmosome type of cell junction.

Part of the cytoplasmic domain of a human desmoglein, Dsg1, a cadherin-like protein found in desmosomes of epithelial cells, has been visualised by electron microscopy. The cloned fragment contains five repeats of a 29 +/- 4 residue sequence unique to desmogleins, followed by a glycine-rich region. In rotary shadowed preparations the molecule consists of a globular head attached to a thin tail, the latter perhaps corresponding to the glycine-rich region. This portion of the molecule is thought to span the width of the inner dense plaque. The structure and dimensions concur well to the configuration deduced from the protein sequence.

Internalisation of desmosomes and their entry into the endocytic pathway via late endosomes in MDCK cells. Possible mechanisms for the modulation of cell adhesion by desmosomes during development.

MDCK cells grown in media with normal levels of Ca2+ (approximately 2 mM) contain internalised desmosomes, referred to as desmosome-associated vacuoles (DAVs). The DAVs consist of one to three plaques retained in the plane of a surrounding vacuolar membrane, and their entry into the endocytic pathway has been investigated using HRP, cationized ferritin and BSA/gold in combination with electron microscopy and immunogold labelling of frozen sections. Endocytic tracers supplied from the apical and basolateral surfaces to filter-grown MDCK cells met in a common perinuclear compartment but DAVs were not labelled during short (5-30 minutes) pulses of marker, whether applied apically or basolaterally. Only when the tracers were taken up from the basolateral surface and then chased for periods of 2-18 hours, were DAVs labelled. It is proposed that entry of an endocytic tracer to DAVs occurs by the association of the desmosomal vacuole with late endosomes. Immunolabelling studies with antibodies to desmosomal components (to Dsg, DPI/II), to HRP and to the cation-independent mannose 6-phosphate receptor (MPR), confirmed that Dsg and DPI/II are located within DAVs and late endosomes, but not in early endosomes. Passage of Dsg, but to a lesser extent DPI/II, was detected in MPR- structures (lysosomes). DAV-like structures have also been observed in developing tissues such as mouse kidney. Such engulfment may provide a general mechanism for handling insoluble junctional proteins, particularly where rapid morphogenetic changes are occurring in the pattern of cell-cell adhesion.

Localization of two fibronectin-binding glycoproteins in rat liver and primary hepatocytes. Co-distribution in vitro of integrin (alpha 5 beta 1) and non-integrin (AGp110) receptors in cell-substratum adhesion sites.

We have compared the localization of integrin alpha 5 beta 1 and AGp110 (apical glycoprotein of Mr 110 x 10(3]] in rat liver parenchyma and in primary hepatocyte cultures. Integrin alpha 5 beta 1 is a heterodimeric fibronectin receptor. AGp110 is a newly described monomeric glycoprotein of the apical (bile canalicular) membrane domain of liver parenchyma that binds in an RGD-independent manner to fibronectin and mediates spreading of hepatocytes onto fibronectin-coated substrata. Using Western blotting of fractionated liver membranes and immunocytochemistry of liver sections at light- and electron-microscope levels, we have confirmed that AGp110 is a canalicular glycoprotein and have established that integrin is located in approximately equal proportions in the sinusoidal, lateral and canalicular membrane domains. In the canalicular surface domain both glycoproteins are associated with microvilli. Examination of immunolabelled primary hepatocytes spread on fibronectin-coated substrata by light and laser scanning confocal microscopy revealed colocalization of AGp110, integrin, actin and vinculin in substratum-attached microextensions at the periphery of the basal cell surface. Actin filaments that terminated at these cell processes originated from circular sub-cortical actin fibres. Interference reflection microscopy revealed focal adhesive contacts at the edge of the basal cell periphery at the same location where AGp110 and integrin were observed by immunofluorescence. In vitro, a proportion of the primary hepatocytes seeded onto fibronectin-coated substrata aggregated into colonies of several cells with intercellular contacts between neighbouring cells. Cell-substratum contacts containing integrin, AGp110, actin and vinculin followed the contours of these colonies in the same manner as they delineated the basal periphery of single, substratum-attached cells. We conclude that both integrin and AGp110 contribute to hepatocyte-fibronectin adhesive interactions and that intercellular adhesion and cooperation among hepatocytes in their response to fibronectin matrices leads to colony formation and morphological differentiation of parenchymal cell monolayers in vitro.

Estimating the growth pattern of micro-organisms in distinct stages of the cell cycle.

Knowledge of the growth patterns of micro-organisms is required to understand how cell growth and division are controlled and co-ordinated in relation to mechanisms of wall assembly and chromosome duplication. Direct observation, e.g. by time-lapse studies, is usually limited in accuracy by the small size of the cells. Indirect methods have therefore been developed which give estimates of the growth patterns of cells, based on the analysis of distributions of cell size in populations in balanced exponential growth. Previously, we have compared such methods (Burdett & Kirkwood, 1983) and concluded that the most powerful approach is that proposed by Collins & Richmond (1962), in which growth rate is calculated as a function of cell size using size distributions of extant, separating and new-born cells. A limitation of this method has been, however, that it gives only an estimate for the average growth rate of cells at a given size, irrespective of the state of progress of individual cells through the cell cycle. In this paper, we describe an extension to the standard Collins-Richmond procedure which provides separate estimates for the growth pattern of cells in distinct stages of the cell cycle, and we illustrate the method in relation to growth of mononucleate, binucleate and septate cells of Bacillus subtilis. It is demonstrated that this three-stage analysis is clearly superior to the standard method, in that it provides more detailed and probably more realistic information. We also demonstrate how to assess the precision and accuracy of the estimated growth pattern. Generalization of the method to any number of stages and to multiple as well as binary fission is described.

Structure and assembly of desmosome junctions: biosynthesis and turnover of the major desmosome components of Madin-Darby canine kidney cells in low calcium medium.

Neither stratifying (primary keratinocytes) nor simple (Madin-Darby canine kidney [MDCK] and Madin-Darby bovine kidney [MDBK]) epithelial cell types from desmosomes in low calcium medium (LCM; less than 0.1 mM), but they can be induced to do so by raising the calcium level to physiological concentrations (standard calcium medium [SCM], 2 mM). We have used polyclonal antisera to the major bovine epidermal desmosome components (greater than 100 kD) in a sensitive assay involving immunoprecipitation of the components from metabolically labeled MDCK cell monolayers to investigate the mechanism of calcium-induced desmosome formation. MDCK cells, whether cultured in LCM or SCM, were found to synthesize the desmosome protein, DPI and desmosome glycoproteins DGI and DGII/III with identical electrophoretic mobility, and also, where relevant, with similar carbohydrate addition/processing and proteolytic processing. The timings of these events and of transport of DGI to the cell surface were similar in low and high calcium. Although the rates of synthesis of the various desmosome components were also similar under both conditions, the glycoprotein turnover rates increased dramatically in cells cultured in LCM. The half-lives decreased by a factor of about 7 for DGI and 12 for DGII/III and, consistent with this, MDCK cells labeled for 48 h in SCM had three and six times the amount of DGI and DGII/III, respectively, as cells labeled for 48 h in LCM. The rate of turnover and the levels of DPI were changed in the same direction, but to much lesser extents. Possible mechanisms for the Ca2+-dependent control of desmosome formation are discussed in the light of this new evidence.

Normal pole formation during total inhibition of wall synthesis of Bacillus subtilis.

Previous work has shown that the side wall of a Gram-positive rod is initially laid down as a compact layer inside the older wall. It is then stretched as it comes to bear tension due to the osmotic pressure inside the cell. If the polar wall is likewise capable of a degree of expansion, then no new murein need be added while the planar cross-wall splits and converts into two poles. In Bacillus subtilis mutant strain FJ6, which is deficient in autolytic enzymes, pole formation can be caused by addition of exogenous muramidase (10 micrograms hen egg white lysozyme ml-1 for 10 min at 35 degrees C). This strain grows as long filaments with many completed cross-walls, but enzymic treatment caused the formation of many new poles of normal morphology as judged by thin section electron microscopy. Fully separated poles of normal appearance were also found when more than 100 times the MIC (1 microgram ml-1) of vancomycin was added to block wall growth totally and rapidly 10 min before the addition of lysozyme. We conclude, therefore, that no new murein is needed in the conversion of the flat septum into poles and that the unstressed cross-wall is capable of the necessary expansion.

Biophysics of pole formation of gram-positive rods.

During pole formation in Bacillus subtilis the inner and outer surfaces of the nascent pole are enlarged by almost exactly the same extent. This means that the stress is almost uniformly distributed throughout the polar wall. This differs from the situation in the cylindrical side wall, where most of the stress is exerted in the outer portions of the intact wall. Because the stress is shared more uniformly, the maximum strain in any part of the polar wall is reduced, compared with the maximum strain within the side wall. The lowered stress may account, in part, for the resistance of the polar wall to hydrolysis by autolytic enzymes under certain conditions. The shape of the newly completed pole is significantly different from the spherical shape that the hydrostatic pressure would tend to produce. It does, however, achieve the shape that maximizes the polar volume under the restrictions arising due to expansion along the circumference not being possible near the junction of cylindrical and polar wall.

Growth kinetics of individual Bacillus subtilis cells and correlation with nucleoid extension.

The growth rate of individual cells of Bacillus subtilis (doubling time, 120 min) has been calculated by using a modification of the Collins-Richmond principle which allows the growth rate of mononucleate, binucleate, and septate cells to be calculated separately. The standard Collins-Richmond equation represents a weighted average of the growth rate calculated from these three major classes. Both approaches strongly suggest that the rate of length extension is exponential. By preparing critical-point-dried cells, in which major features of the cell such as nucleoids and cross-walls can be seen, it has also been possible to examine whether nucleoid extension is coupled to length extension. Growth rates for nucleoid movement are parallel to those of total length extension, except possibly in the case of septate cells. Furthermore, by calculating the growth rate of various portions of the cell surface, it appears likely that the limits of the site of cylindrical envelope assembly lie between the distal tips of the nucleoid; the old poles show zero growth rate. Coupling of nucleoid extension with increase of cell length is envisaged as occurring through an exponentially increasing number of DNA-surface attachment sites occupying most of the available surface.

How does a Bacillus split its septum right down the middle?

This is a speculative paper which considers the possible ways that Gram-positive cells might employ to achieve an even thickness of the two daughter poles resulting from the fission of the cross-wall. The first is that the protonmotive force generated by the extrusion of protons at the cytoplasmic membrane acts to inhibit autolysins to a distance of about 25 nm. The second has to do with the stresses that develop as the poles form. On the tacit assumption that the autolysins will function faster when their substrates are under tension, it is shown how this, too, can lead to even bisection of the cross-wall. These possibilities are not alternative, both probably function.

The variable T model for gram-negative morphology.

Gram-negative micro-organisms possess only a very thin murein sacculus to resist the stress caused by the internal hydrostatic pressure. The sacculus consists of at most one molecular layer of peptidoglycan in an extended conformation. It must grow by the insertion and cross-linking of new murein to the old before the selective cleavages of the stress-bearing murein are made which allow wall enlargement. Since insertion of new murein occurs all over the surface of Escherichia coli (even in completed poles), the internal pressure would tend to force the cells into a spherical shape and prevent both cylindrical elongation and cell division. Of course, Gram-negative bacteria do achieve a variety of shapes and do divide. Because prokaryote cells, unlike eukaryotic cells, do not have cytoskeletons and contractile proteins to transduce biochemical free energy into the mechanical work needed to achieve aspherical shapes and to divide, this paradox seems to be resolvable only by postulating that the details of the biochemical mechanism for wall growth vary in different regions of the surface, affecting the work required to enlarge the wall locally. Depending on the degree and rate of change in the biochemical energetics, it is possible to account for rod and the other more complex shapes of Gram-negative bacteria. Division occurs in Gram-negative organisms by the development of constrictions that progressively invade the cytoplasm. The work to cause these morphological processes must ultimately derive from the biochemical process of the stress-bearing wall formation. A biophysical basis for cell division in these prokaryotic organisms is proposed.

How does a bacterium grow during its cell cycle?

Rod-shaped bacteria such as Escherichia coli and Bacillus subtilis appear to extend continuously in length between divisions. However, the kinetics of growth of the individual cell in the steady state is still unknown. A brief, critical account of the main approaches used to determine the pattern of surface extension is given. In general, these approaches are of three types. Firstly, attempts have been made to relate average cell size to growth rate of the culture and to determine possible stages in the cell cycle at which the rate of length extension might change. Secondly, comparisons have been made between the measured length distribution of cells and theoretical distributions, based on three primary hypotheses (linear, bilinear and exponential growth). Thirdly, the principle of Collins and Richmond, involving the calculation of growth rate from the length distributions of extant, separating and new-born cells, is described. It is emphasized that there is a strong element of variation in size at different stages of the cell cycle. This variation imposes severe limitations on models which utilize only average cellular dimensions. We conclude that the Collins-Richmond principle affords the most powerful approach to the analysis of bacterial growth kinetics. However, we propose that the method be modified to permit calculation of separate rates of growth of cells between discernible events in the cell cycle, as well as simply between birth and division.

Identification of specific restriction fragments associated with a membrane subparticle from Bacillus subtilis.

When lysates of Bacillus subtilis were treated with restriction endonucleases EcoRI or HindIII, almost all of the DNA was released from the major plasma membrane fraction that was sedimentable at low speed. However, a very small part of the released DNA, when centrifuged at high speed, appeared to be bound to small membrane fragments. On agarose gels, this material, prepared with either enzyme, contained only a small number of restriction fragments, and the DNA in the sample hybridized with 11 to 12 EcoRI or HindIII fragments of chromosomal DNA. This DNA was used after nick-translation to screen Charon 4A clone banks for phages containing membrane-bound fragments. One of these was studied in detail. Only a part (about 5 kilobases) of the region present in this clone is important in binding the DNA to the membrane subparticle.

The bactericidal action of beta-lactam antibiotics on an autolysin-deficient strain of Bacillus subtilis.

An autolysin-deficient mutant of Bacillus subtilis was completely tolerant to 5 h incubation with 50-100 micrograms cycloserine ml-1 whereas the wild-type was rapidly lysed and killed by 12 micrograms ml-1. Lysis also did not occur when low concentrations of beta-lactams were added to exponentially growing cultures of the mutant, but over 90% of the bacteria were killed within 90-120 min. Protein, lipid and peptidoglycan synthesis as well as growth were inhibited after about 60 min. At this time, but not earlier, small amounts of these three cell components appeared in culture supernatants. Earlier, at about 20-30 min, the intracellular pools of amino acids started to decline rapidly and there was a temporary apparent increase in the rate of lipid synthesis. Neither of the latter phenomena occurred with cycloserine, with which protein and lipid synthesis declined only slowly and the rate of peptidoglycan synthesis was 80% inhibited within 30 min. Only occasional cells with damaged walls were seen 30-90 min after addition of either beta-lactams or cycloserine to the cultures. It thus seems unlikely that wall hydrolysis or penetration by residual autolysins in the mutant are responsible for mass cell death caused by the beta-lactams.