Talin-Null Cells of Dictyostelium Are Strongly Defective in Adhesion to Particle and Substrate Surfaces and Slightly Impaired in Cytokinesis

For axenic growth, D. discoideum wild-type strain AX2 and talin-null mutants were cultivated in shaken suspension at 150 rpm in liquid nutrient medium at 23°C as described by Claviez et al. (1982). For starvation, cells were washed twice in 17 mM K-Na phosphate buffer, pH 6.0 (non-nutrient buffer) and were shaken at a density of 10⁷ cells per ml in the buffer. To achieve aggregation competence, AX2 cells were starved for 6 h, and talin-null cells were starved for 8 h.

For growth in bacterial suspensions, Escherichia coli B/r or Salmonella minnesota R595 were cultivated overnight in standard I nutrient broth (Merck, Darmstadt, Germany), washed, and adjusted to a density corresponding to 10¹⁰ E. coli B/r cells per ml. The suspensions of washed bacteria were inoculated with 10⁴–10⁵ cells of D. discoideum per ml. For cultivation on agar surfaces, cells of E. coli B/r were spread on agar plates containing 0.1% peptone and 0.1% glucose in non-nutrient buffer. AX2 or talin-null cells were picked onto the surface for measuring radial growth.

Experiments shown in Fig. 3 were performed with the talin-null mutant strain HG1664 and repeated with HG1666. Experiments in Figs. 2, 4, 5, and 6 were performed with the talin-null strain HG1666 and repeated with HG1663 (see Fig. 2), HG1664 (see Fig. 4, B/r suspension), or HG1665 (see Figs. 5 and 6). The micrographs of cytokinesis and immunofluorescence images of Fig. 7 were obtained with mutant HG1665.

The targeting vector was constructed by inserting the Bsr cassette from plasmid pBsr2 (Sutoh, 1993) into the single blunted PstI site of a genomic DNA fragment comprising base pairs 58–3,031 of the talin coding region. This fragment was obtained by PCR using primers 5′-GCGGATCCTTTGCACCAGATATGTGTATTC and 3′-CGGCAATTCAACTTAGC, and was cloned into pUC19. The construct was excised using BamHI and HindIII, and the linearized and dephosphorylated fragment was used to transfect AX2 wild-type cells of D. discoideum by electroporation. Talin-null mutants were selected with 7.5 μg/ml blasticidin S (ICN Biomedicals, Inc., Costa Mesa, CA) in nutrient medium.

For Southern blotting, genomic DNA was prepared according to Noegel et al. (1985). 15 μg of either AX2 or mutant DNA was run on 1% agarose gels and blotted onto Hybond N nylon membrane (Amersham Intl., Little Chalfont, UK). The blots were hybridized under high stringency for 15 h at 65°C in RapidHyb buffer (Amersham Intl.) with a PCR-generated probe comprising base pairs 3,086–3,442 of the talin coding region.

mAb 477 against D. discoideum talin was produced by Kreitmeier et al. (1995). The antibody recognized an epitope between amino acid residues 2,053 and 2,290, as deduced from bacterially expressed talin polypeptide fragments. mAb 227-341-4 (here designated as mAb 341) was obtained from a Balb/c mouse immunized with a His-tagged NH₂-terminal fragment comprising residues 20–124 of the talin sequence. For the production of this fragment, a genomic DNA fragment was generated by PCR and cloned into pQE30 (QIAGEN Inc., Chatsworth, CA). The sequence was verified and the fragment was expressed in E. coli M15. The protein was purified on a Ni²⁺-agarose column under denaturing conditions using 6 M guanidinium chloride followed by 8 M urea. The fragment was injected together with Freund's adjuvant, and spleen cells were fused with PAIB₃Ag81 myeloma cells.

SDS-PAGE and immunoblotting with anti-talin antibodies was performed in 3–20% gradient polyacrylamide gels.  mAb 477 was directly ¹²⁵I-labeled, and mAb 341 was detected with iodinated sheep anti–mouse IgG (Amersham Intl.). For detection of the csA glycoprotein on Western blots, mAb 294 was used (Bertholdt et al., 1985).

Aggregation-competent Dictyostelium cells were monitored while migrating on the surface of a glass coverslip preincubated with 0.2% BSA solution (Serva, Heidelberg, Germany). The topography of the ventral cell surface, including cell-to-substratum contacts, was imaged by reflection interference contrast microscopy (RICM)¹ using a double-view microscope (Weber et al., 1995). The wavelength of the green light used for RICM was λ = 546 nm. Contours of two-dimensional projections of the cell body, as seen in bright-field images, were extracted by an image processing routine (Weber and Albrecht, 1997) and superimposed onto the RICM images. Chemotactic responses were recorded while the cells were stimulated from a micropipette filled with a solution of 10⁻⁴ M cAMP and placed with its orifice at a distance of 10–20 μm from the cell surface (Gerisch and Keller, 1981).

Agglutination of wild-type and talin-null cells was assayed by measuring light scattering according to Beug and Gerisch (1972) in a microprocessor-controlled agglutinometer (Bozzaro et al., 1987). Phagocytosis assays using TRITC-labeled heat-killed yeast in shaken suspension were carried out essentially as described by Maniak et al. (1995), and quantitative and microscopic assays of fluid-phase uptake were performed using TRITC-labeled dextran according to Hacker et al. (1997). For the quantitative assays of Fig. 3, A–C, and Fig. 5, A and C, suspensions of wild-type and talin-null cells were adjusted to the same total cell volume by determining the volume of densely packed cells (Maniak et al., 1995). The volume of the mutant cells was, on the average, 1.5-fold the volume of wild-type cells.

For synchronization, cells from suspension cultures were washed and shaken in non-nutrient buffer for 3 h, before they were allowed to adhere for 30 min on glass coverslips. Subsequently, the buffer was replaced by nutrient medium and the culture continued for 4 h (Neujahr et al., 1997). Cytokinesis in the adherent cells was monitored by video recording. For fluorescence labeling, the cells were fixed for 15 min in a solution of 15% saturated picric acid, 2% paraformaldehyde, pH 6.0, postfixed with 70% ethanol, and processed according to Humbel and Biegelmann (1992). α-Tubulin was labeled with rat mAb YL 1/2 (Kilmartin et al., 1982) and TRITC-conjugated goat anti–rat antibodies (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). DNA was stained with 4′,6-diamidino-2-phenylindole (DAPI) (Sigma, Deisenhofen, Germany). Micrographs were taken with a Phaco ×100 oil Neofluar objective using a Zeiss Axiophot microscope (Oberkochen, Germany).

To inactivate the gene that encodes the Dictyostelium talin previously described by Kreitmeier et al. (1995), cells of D. discoideum were transfected with a vector causing gene replacement by homologous recombination. The vector was designed to interrupt the gene by a blasticidin resistance (bsr) cassette at the codon for T516 of the talin sequence (Fig. 1 A). Probing 40 blasticidin-resistant clones by Southern blotting revealed eight independent transformants in which the gene was disrupted. These mutant lines were characterized by two phenotypic alterations: (a) weak adhesion to the polystyrene surfaces of petri dishes, and (b) the tendency to form larger cells in suspension. All transformants completed development on agar with the formation of normal fruiting bodies.

In all of the eight mutants obtained, the vector was selectively inserted into the talin gene; no second insertion into other regions of the genome had occurred. In Fig. 1, B and C, data for four of these mutants are shown. EcoRV and BclI cut the gene at both sides of the inserted bsr cassette. When genomic DNA of the mutants was digested with either one of these two enzymes and probed with a sequence stretch of the talin gene downstream of the vector sequence, single bands were labeled. These bands indicated an increase in size by 1.4 kb relative to the wild-type fragments, an upshift that corresponds to the size of the bsr cassette (Fig. 1,B, upper panels). Using as a probe the bsr coding region of the vector, the same single bands showed up in the mutants. As expected, no hybridization signal was seen with wild-type DNA (Fig. 1 B, lower panels).

Western blots of total cellular proteins separated by SDS-PAGE were probed with two antibodies, one binding COOH-terminally to the predicted site of disruption of the talin polypeptide, the other recognizing an epitope at the NH₂-terminal side of the disruption (Fig. 1,A). The two antibodies no longer detected the talin band in the eight independent mutants, nor did the latter antibody recognize in the mutants any NH₂-terminal fragment of this protein that might originate from a truncated message (Fig. 1 C).

A phenotypic alteration immediately obvious in the mutant cells was their weak adhesion to the polystyrol surfaces of plastic petri dishes, as revealed by their detachment upon gentle pipetting. To study this adhesion defect, alterations in the interaction of talin-null cells with a substrate surface were analyzed by RICM. Areas of close cell-to-substrate contact are visualized by RICM as dark areas (Gingell and Todd, 1979), which means that the reflecting cell surface approaches the glass surface with a distance <50 nm close to the zero order intensity minimum of the interference pattern (Rädler and Sackmann, 1993). The first interference maximum corresponding to a distance of λ/4 appears then as a bright area.

Glass coated with BSA proved to be of optimal adhesiveness for the motility of D. discoideum cells. This substrate allows wild-type cells to attach and detach in a regular pattern (Weber et al., 1995) and to move persistently with negligible loss of material retained on the substrate surface (Schindl et al., 1995). In the aggregation-competent stage, which is reached after several hours of starvation, wild-type cells adhere to the substrate only at portions of their basal surface, and the areas of contact rapidly change in size and shape with movement of the cells (Weber et al., 1995). An example is given in Fig. 2 A.

Talin-null cells had almost completely lost adhesiveness to BSA-coated glass. This is seen in RICM images by the absence of dark areas of contact in the center of interference patterns (Fig. 2 B). The finding that the area of closest approximation between the cell and substrate surfaces was regularly bright means that the distance was in the order of the first maximum of λ/4 = 136 nm for the green light used. This distance made firm attachment of the talin-null cells impossible, so that they easily drifted with slight convections in the fluid.

Aggregation-competent wild-type cells are highly responsive to cAMP as a chemoattractant, and cell-to-substrate interactions are highly dynamic during reorientation. The cells turn into a new direction by bending their established front or by extending a new front into the direction of the gradient. At the beginning of a chemotactic response, the front does not need to be in contact with the substrate, and sometimes two fronts compete with each other for becoming the leading edge (Weber et al., 1995). In the example shown in Fig. 2 C, the wild-type cell initially moved toward the top of the frames, then turned with its front toward the right, and shortly later produced a second front, which subsequently established itself as the leading edge. New sites of attachment at the fronts increased in size and fused with each other, followed by detachment at the tail of the cell.

Talin-null cells efficiently responded to cAMP by turning toward the source. Stimulation with the chemoattractant also slightly enhanced interaction with the substrate, so that grayish foci of attachment were seen in RICM images. However, the mutant cells were impaired in spreading and fusion of the contact areas (Fig. 2 D). Foci of attachment increased only marginally in size; often they were fixed in a position essentially stationary on the substrate, with the cells gliding ahead of these local sites of loose attachment. Most of the foci of cell-to-substrate contact in talin-null cells vanished when they had passed the middle of a cell. Some of the patches existed for up to 2 min, whereas others were short-lived and never reached the middle region of the cell.

Consequences of the adhesion defect in talin-null cells for particle uptake were assayed using heat-killed, fluorescently labeled yeast particles in shaken suspension. At a shaking frequency of 150 rpm, the rate of phagocytosis in talin-null cells was ∼10% the rate in wild-type AX2 cells (Fig. 3,A). Assuming that this reduction in the rate of particle uptake is a consequence of impaired adhesion, one would predict a higher rate of uptake at conditions of low shear. Fig. 3 B shows that reducing the frequency of shaking from 150 to 100 rpm increased substantially the rate of particle uptake in talin-null cells and only moderately increased it in wild-type cells.

A particle adhering to a Dictyostelium cell triggers an extension of the cell surface that spreads as a cup-shaped lamella over the particle and finally engulfs it (Maniak et al., 1995). Uptake of liquid medium occurs in D. discoideum by macropinocytosis which, like phagocytosis, requires a response in the actin system to internalize a vesicle (Hacker et al., 1997). Macropinocytosis differs, however, from phagocytosis in being independent of the stimulus of a particle that impinges on the cell surface. To distinguish between a defect in adhesion and the process of uptake, internalization of TRITC-dextran, a fluid-phase marker, was measured. The rate of fluid-phase uptake proved to be indistinguishable in talin-null cells from that in wild-type cells (Fig. 3,C). Microscopic observations revealed that the mechanism of this uptake in talin-null cells is similar as recently described for wild-type cells (Hacker et al., 1997). Protrusions extending ⩾2 μm from the cell body eventually closed at their rim, thus entrapping an aliquot of medium (Fig. 3 D).

Bacteria that differ in the carbohydrate moieties of their surface lipopolysaccharides are convenient probes to evaluate the role of adhesion in phagocytosis. When fixed in a lawn on an agar plate, a large variety of E. coli or Salmonella strains can support growth of Dictyostelium cells, independently of the chemical nature or physical properties of the bacterial surfaces. If applied in shaken suspension, adhesion of the bacteria to the phagocyte surface becomes crucial for uptake, since a bacterium needs to adhere until the phagocyte has produced a cup to envelop and eventually to engulf the bacterium. As far as wild-type cells of D. discoideum are concerned, the uptake of bacteria in suspension is primarily determined by surface lipopolysaccharides. Long polysaccharide chains prevent uptake by the lack of adhesion. It suffices to coat these carbohydrate chains with antibodies to render the bacteria appropriate for uptake in suspension (Gerisch et al., 1967).

E. coli B/r is commonly used for the efficient growth of D. discoideum wild-type cells in shaken suspension. The carbohydrate chains on the surface of this E. coli strain are reduced to a glucose-containing core oligosaccharide (Malchow et al., 1967). On a lawn of E. coli B/r on an agar plate, talin-null cells grew as fast as wild-type cells, which confirmed that engulfment of the bacteria is not impaired in the mutant cells (Fig. 4,A). In a suspension, however, the growth of talin-null cells on E. coli B/r was dramatically slowed down (Fig. 4 B).

To provide evidence that the core oligosaccharide present on the surface of E. coli B/r interfered with adhesion to talin-null cells, we used Salmonella minnesota R595, a strain appropriate for D. discoideum suspension cultures (Malchow et al., 1967; Gerisch et al., 1985; Cohen et al., 1994), which lacks the entire carbohydrate moiety apart from two residues of 3-deoxy-d-manno-octulosonic acid linked to lipid A (Lüderitz et al., 1966). In suspensions of this bacterial strain, the generation times of wild-type and talin-null mutant cells differed only slightly, indicating that bacteria of proper adhesiveness can be taken up by the mutant cells even under the restrictive conditions of a shaken suspension (Fig. 4 C).

In suspension, undeveloped wild-type cells of D. discoideum aggregate into loose clusters. The cohesion of these cells is distinguished from that of aggregating cells by its EDTA sensitivity (Beug et al., 1973). Cell cohesion can be quantitatively assayed by an automated light scattering assay. Suspended cells are rotated in an agglutinometer in cuvettes that are designed to apply standardized shear to dissociate the cells (Beug and Gerisch, 1972). The strength of cell cohesion is then reflected in the size of aggregates formed against the dissociating forces, and the decrease of particle number accompanying aggregation results in reduced light scattering.

Fig. 5,A shows the behavior of undeveloped cells: when an equilibrium was reached after <1 h, light scattering in suspensions of wild-type cells was low in the absence and high in the presence of EDTA, whereas in suspensions of talin-null cells it was high both in the presence and absence of EDTA. The photographs of Fig. 5 B illustrate these quantitative data, indicating that the EDTA-sensitive cohesion of undeveloped cells is talin dependent.

In the aggregation-competent stage, talin-null cells adhered to each other in an EDTA-stable manner (Fig. 5,C), and the aggregates formed in suspension were only slightly smaller than those formed by wild-type cells (Fig. 5 D).

The initial observation that talin-null cells become larger than wild-type cells when cultivated in suspension suggested a deficiency in cytokinesis. To establish this by quantitative data, numbers of nuclei were counted in cells that were cultivated either in suspension or on a glass surface. In suspension cultures, the majority of talin-null cells contained two to four nuclei, whereas wild-type cells were primarily mononucleated (Fig. 6, top). Under the same conditions, the average volume of talin-null cells was 1.5-fold that of wild-type cells. These results indicate that, in talin-null cells undergoing mitosis in suspension, cytokinesis fails to be reliably coupled to nuclear division. This deficiency was less distinct when cells grew on a solid surface (Fig. 6, middle), and it was not at all evident when the liquid medium was replaced as a source of nutrients by a lawn of bacteria on an agar surface (Fig. 6, bottom).

The finding that only a few of the talin-null cells acquired more than four nuclei in suspension culture with nutrient medium suggests a process that counterbalances the increase in nuclear number caused by the slight impairment of cytokinesis. To characterize this process, cytokinesis was recorded in bi- or multinucleated talin-null cells. Fig. 7,A illustrates mitosis in a binucleated cell that divided symmetrically into four daughter cells. In fixed preparations, mitotic stages of binucleated cells were characterized by the presence of two spindles and four asters of microtubules, which held variable positions relative to each other in the three-dimensional space of the cell body. In the three cells shown in Fig. 7,B, the two mitotic complexes, each consisting of a pair of asters connected to a spindle, were arranged in the same plane. These images illustrate most clearly that only two of the four cleavage furrows that are indistinguishable in Fig. 7 A are directed toward the middle of the spindles, the other pair incising the cell body in between the mitotic complexes.

In addition to binucleated cells, cells that contained three or four nuclei were found to form multiple spindles. It appears therefore that the increase in nuclear number is limited in talin-null cells by mitotic division cycles that give rise to more than two daughter cells. This type of cleavage is based on the shaping of cleavage furrows around the asters of microtubules.

Dictyostelium has been the first developmental system in which two types of cell-to-cell adhesion, an EDTA-stable and a sensitive one, could be distinguished (Gerisch, 1961) and shown to be blocked separately by antibody Fab (Beug et al., 1973). According to the data presented here, the EDTA-sensitive “type B contacts” and also cell-to-substrate contacts represent talin-dependent types of cell adhesion (Fig. 5, A and B). Quantitative measurements confirmed that the energy of cell-to-substrate adhesion is strongly reduced in the talin-null cells of Dictyostelium (Simson, R., personal communication). A glycoprotein anchored to the membrane by a ceramide-based lipid (Stadler et al., 1989) is the major adhesion protein involved in the EDTA-stable cohesion of aggregating cells (Müller and Gerisch, 1978). Cell-to-cell adhesion mediated by this “contact site A” glycoprotein turned out to be rather independent of talin (Fig. 5, C and D).

In vertebrate cells talin interacts with integrins, which link the cell surface to receptors on the extracellular matrix (Beckerle and Yeh, 1990). The distinction of two adhesion systems in Dictyostelium on the basis of their linkage to talin suggests that also in this microorganism talin interacts as part of a specific complex with one type of adhesion protein in the plasma membrane. Integrins have not been identified in Dictyostelium; a candidate for a talin-regulated adhesion protein is a 126-kD (Chadwick et al., 1984) or 130-kD glycoprotein (Chia, 1996).

Whether or not Dictyostelium has a “protointegrin” on its surface is not only of phylogenetic interest but is also relevant for functional reasons. Dictyostelium cells manage in their natural habitat to move on soil particles of varying chemical nature and physical surface properties. The question is whether Dictyostelium cells have a chance in their natural environment to use specific molecular interactions in regulating their adhesiveness to soil particles or to a variety of bacterial surfaces. There are several possibilities of how coupling of talin to a membrane protein regulates cell adhesion. (a) Adhesion might be regulated by clustering of the membrane protein, as it is known for integrins in other cells. (b) Some selectivity might be involved in the interaction of the putative adhesion protein with the surface of soil particles, analogous to the stereospecificity that determines attachment of Xenopus kidney cells to the surfaces of (R, R) and (S, S) enantiomorphous tartrate crystals (Hanein et al., 1994). (c) Thermally induced out-of-plane fluctuations of soft membranes lead to a repulsive force, which has the same distance dependence as the Van der Waals attractive forces (Sackmann, 1994). Coupling of the cell membrane to the actin cortex via talin will suppress the undulations, thereby increasing cell-to-substrate adhesion (Zeman et al., 1990). (d) Single cells of Dictyostelium might coat a surface with a secreted protein and then adhere on this layer in an analogous fashion to an integrin–fibronectin type of cell interaction with an extracellular matrix. In fact, a matrix is produced in the multicellular slug stage of Dictyostelium and is used as a surface for migration (Freeze and Loomis, 1977; Abe et al., 1994).

The motility of fibroblasts or keratocytes depends on traction forces that these cells exert on their substrate (Oliver et al., 1994; Lauffenburger and Horwitz, 1996). The traction is thought to convert intracellularly generated contractile forces into net movement. Stress fibers and focal contacts appear to be optimized for efficient and directed pulling on the surrounding extracellular matrix. In contrast, aggregating cells of D. discoideum efficiently move while adhering only with part of their basal surface to the substrate. These cells lack static structures like stress fibers or focal adhesion complexes and are fast moving with speeds of up to 30 μm/min. In talin-null cells, substrate adhesion is even further reduced to an absolute minimum (Fig. 2). These almost nonadhesive cells exemplify a case of cell motility largely uncoupled from traction forces.

For the movement of talin-null cells on small patches of contact with a substrate, the following possibilities can be considered. First, loose adhesion might induce a local response, e.g., coupling of a membrane patch immobilized on the substrate to one of the myosins, so that this patch is driven in the membrane in a rearward direction (Jay and Elson, 1992). As a result, the cell will move ahead while the patch remains stationary on the substrate. Second, cell locomotion might be effected by a continuous membrane flux, either in the form of a rolling movement (Anderson et al., 1996) or a lipid flow generated by exocytosis at the front and by endocytosis along the lateral cell surface (Bretscher, 1996). Talin-null cells offer themselves for a detailed analysis of membrane events essential for cell locomotion because of the absence of spreading, which in wild-type cells masks the local interactions at small areas of cell-to-substrate contact.

Phagocytosis mutants may be classified into four major groups: (a) those deficient in the adhesion of particles to the phagocyte surface, (b) mutants altered in signal transduction from the cell surface to the cortical cytoskeleton, (c) cytoskeletal mutants impaired in the mechanism of particle uptake, and (d) mutants in proteins that play a role in later steps of the phagocytic pathway. To begin with the third category, a series of gene disruption mutants has been generated in Dictyostelium to eliminate proteins associated with the actin system. These proteins include coronin, which accumulates together with actin early at the phagocytic cup (Maniak et al., 1995) and later reassociates with postlysosomal endosomes (Rauchenberger et al., 1997), and two F-actin cross-linking proteins, 120-kD gelation factor (ABP120) and α-actinin (Cox et al., 1996; Rivero et al., 1996). Since integrity of the actin system is essential for phagocytosis to occur (Maniak et al., 1995), the lack of proteins that cross-link actin filaments or otherwise contribute to the viscoelastic properties of the actin cortex most likely affects particle uptake directly.

On the basis of the following results, we propose that the talin-null cells are impaired in the initial adhesion to a particle and in subsequent spreading on its surface. The rate at which yeast particles are taken up by talin-null cells in shaken suspension can be markedly increased by reducing the frequency of shaking (Fig. 3,B). This inverse dependence of uptake on the strength of shear indicates that the average persistence time of a particle at the phagocyte surface is shorter than the time a cell would need to irreversibly entrap the particle. Extreme differences in the uptake rate of bacteria that differ in their surface properties provide a second argument for an impairment of adhesiveness in talin-null cells (Fig. 4). Fluid-phase uptake shows that talin-null cells can form surface extensions and convert them into macropinosomes large enough to include a bacterium (Fig. 3 D). In suspensions of 10¹⁰ bacteria per ml, as used in our experiments, accidental uptake will supply the cells with not more than one bacterium per 100 macropinosomes filled only with buffer. This explains why talin-null cells are almost unable to grow on E. coli B/r in suspension, whereas they are perfectly capable of growing in a dense lawn of the same bacteria on an agar surface. By the same argument, it is the strong adhesiveness of the carbohydrate-deficient R595 strain of Salmonella minnesota that allows the talin-null cells to recruit these bacteria efficiently from a suspension.

Wild-type cells of D. discoideum extend a lamella with a speed of ∼10 μm/min around a particle. This process is controlled by a zipper mechanism; it can be interrupted at any stage before closure of a phagocytic vesicle (Maniak et al., 1995). If the assumption is correct that the initial stages of phagocytosis resemble the spreading of a motile cell on a planar substrate (Grinnell, 1984), one can deduce from the dotlike contacts that talin-null cells form on a substrate that phagocytosis is similarly altered by weak initial adhesion and insufficient spreading of the mutant cells on a particle surface. Along the same line, it has been inferred from the enrichment of talin in phagocytic cups of mouse macrophages that talin plays a role during F_c receptor– mediated phagocytosis in linking transmembrane receptors to the force-generating cytoskeleton (Greenberg et al., 1990), similar to its implication in focal adhesion assembly.

In several of the previously isolated phagocytosis mutants of Dictyostelium, the protein affected is unknown, and inactivation of talin appears to be one possibility. Mutants characterized by Cohen et al. (1994) do not grow in suspensions of S. minnesota R595, which distinguishes the defect in these mutants from the impaired adhesion of talin-null cells. Mutants selected earlier by Vogel et al. (1980) differ from the talin-null mutants by their ability to grow in suspensions of E. coli B/r. Two separate adhesion mechanisms are involved in the binding of these bacteria to Dictyostelium cells, one unspecific and a sugar-specific one. If the unspecific type of adhesion is eliminated by mutagenesis of D. discoideum cells, bacteria of the E. coli B/r type can still stick to the mutant cells by virtue of terminal glucose residues on their lipopolysaccharides (Vogel et al., 1980). These and other sugar residues are recognized by lectinlike receptors exposed on Dictyostelium cell surfaces (Bozzaro and Roseman, 1983). The fact that talin-null cells are almost incapable of growing on E. coli B/r in suspension implies that talin is linked to both adhesion systems, the sugar-specific and the unspecific one.

As in other eukaryotic cells, cytokinesis in D. discoideum involves the formation of a cleavage furrow that initiates separation of the daughter cells. Our data indicate that talin, as an actin-associated protein, supports formation of the cleavage furrow. Impairment of cytokinesis in talin-null cells that grow in shaken suspension indicates a role for talin in cytoskeletal functions distinct from its involvement in making the cells adhesive. In the suspension cultures, a high proportion of bi- to tetranucleated cells is observed (Fig. 6).

Changes in cell shape during cytokinesis are brought about by the actin-rich cell cortex in conjunction with the microtubule-based spindle and asters. The synchronous division of multiple nuclei in a talin-null cell clearly reveals that a cleavage furrow is induced wherever two asters of microtubules are placed adjacent to each other (Fig. 7). This finding extends observations made on a Dictyostelium mutant lacking myosin II (Neujahr et al., 1997). Mitosis in multinucleated cells of this mutant is accompanied by folding of the cell surface around each aster of microtubules.

The cleavage stages in talin-null cells resemble experimentally disordered stages in dividing eggs of the sand dollar, Arbacia lixula (Rappaport, 1986). If in these eggs the formation of a cleavage furrow is prevented during the first nuclear division by dislocating the mitotic apparatus, four furrows are simultaneously induced during the second division at spaces between the asters, giving rise to four blastomers. In Arbacia blastomers as well as Dictyostelium cells, cruciform stages are diagnostic of single cells dividing into four. The spindle is dispensable in these cases for a cleavage furrow to be formed, and it is the apposition of two asters closely associated with the cell cortex that induces a furrow at the midline between them. Since this effect is easily observed in talin-null cells, these cells may provide a tool to identify the proteins or molecular configurations that specify a cleavage furrow.

We thank Alicija Baskaya for mAb production, Gerhard Rahn for iodination, and Jean-Marc Schwartz for digitalization of the video images.

DAPI

4′,6-diamidino-2-phenylindole

RICM

reflection interference contrast microscopy