Search

Integrins are critical for hemostasis and thrombosis because they mediate both platelet adhesion and aggregation. Talin is an integrin-binding cytoplasmic adaptor that is a central organizer of focal adhesions, and loss of talin phenocopies integrin deletion in Drosophila. Here, we have examined the role of talin in mammalian integrin function in vivo by selectively disrupting the talin1 gene in mouse platelet precursor megakaryocytes. Talin null megakaryocytes produced circulating platelets that exhibited normal morphology yet manifested profoundly impaired hemostatic function. Specifically, platelet-specific deletion of talin1 led to spontaneous hemorrhage and pathological bleeding. Ex vivo and in vitro studies revealed that loss of talin1 resulted in dramatically impaired integrin alphaIIbbeta3-mediated platelet aggregation and beta1 integrin-mediated platelet adhesion. Furthermore, loss of talin1 strongly inhibited the activation of platelet beta1 and beta3 integrins in response to platelet agonists. These data establish that platelet talin plays a crucial role in hemostasis and provide the first proof that talin is required for the activation and function of mammalian alpha2beta1 and alphaIIbbeta3 integrins in vivo.

Cells rapidly transduce forces exerted on extracellular matrix contacts into tyrosine kinase activation and recruitment of cytoskeletal proteins to reinforce integrin-cytoskeleton connections and initiate adhesion site formation. The relationship between these two processes has not been defined, particularly at the submicrometer level. Using talin1-deficient cells, it appears that talin1 is critical for building early mechanical linkages. Deletion of talin1 blocked laser tweezers, force-dependent reinforcement of submicrometer fibronectin-coated beads and early formation of adhesion sites in response to force, even though Src family kinases, focal adhesion kinase, and spreading were activated normally. Recruitment of vinculin and paxillin to sites of force application also required talin1. FilaminA had a secondary role in strengthening fibronectin-integrin-cytoskeleton connections and no role in stretch-dependent adhesion site assembly. Thus, force-dependent activation of tyrosine kinases is independent of early force-dependent structural changes that require talin1 as part of a critical scaffold.

Focal adhesion assembly and actin stress fiber formation were studied in serum-starved Swiss 3T3 fibroblasts permeabilized with streptolysin-O. Permeabilization in the presence of GTPgammaS stimulated rho-dependent formation of stress fibers, and the redistribution of vinculin and paxillin from a perinuclear location to focal adhesions. Addition of GTPgammaS at 8 min after permeabilization still induced paxillin recruitment to focal adhesion-like structures at the ends of stress fibers, but vinculin remained in the perinuclear region, indicating that the distributions of these two proteins are regulated by different mechanisms. Paxillin recruitment was largely rho-independent, but could be evoked using constitutively active Q71L ADP-ribosylation factor (ARF1), and blocked by NH2-terminally truncated Delta17ARF1. Moreover, leakage of endogenous ARF from cells was coincident with loss of GTPgammaS- induced redistribution of paxillin to focal adhesions, and the response was recovered by addition of ARF1. The ability of ARF1 to regulate paxillin recruitment to focal adhesions was confirmed by microinjection of Q71LARF1 and Delta17ARF1 into intact cells. Interestingly, these experiments showed that V14RhoA- induced assembly of actin stress fibers was potentiated by Q71LARF1. We conclude that rho and ARF1 activate complimentary pathways that together lead to the formation of paxillin-rich focal adhesions at the ends of prominent actin stress fibers.

We have used gene disruption to isolate two talin (-/-) ES cell mutants that contain no intact talin. The undifferentiated cells (a) were unable to spread on gelatin or laminin and grew as rounded colonies, although they were able to spread on fibronectin (b) showed reduced adhesion to laminin, but not fibronectin (c) expressed much reduced levels of beta1 integrin, although levels of alpha5 and alphaV were wild-type (d) were less polarized with increased membrane protrusions compared with a vinculin (-/-) ES cell mutant (e) were unable to assemble vinculin or paxillin-containing focal adhesions or actin stress fibers on fibronectin, whereas vinculin (-/-) ES cells were able to assemble talin-containing focal adhesions. Both talin (-/-) ES cell mutants formed embryoid bodies, but differentiation was restricted to two morphologically distinct cell types. Interestingly, these differentiated talin (-/-) ES cells were able to spread and form focal adhesion-like structures containing vinculin and paxillin on fibronectin. Moreover, the levels of the beta1 integrin subunit were comparable to those in wild-type ES cells. We conclude that talin is essential for beta1 integrin expression and focal adhesion assembly in undifferentiated ES cells, but that a subset of differentiated cells are talin independent for both characteristics.

A mutant cell line, derived from the mouse embryonal carcinoma cell line F9, is defective in cell-cell adhesion (compaction) and in cell-substrate adhesion. We have previously shown that neither uvomorulin (E-cadherin) nor integrins are responsible for the mutant phenotype (Calogero, A., M. Samuels, T. Darland, S. A. Edwards, R. Kemler, and E. D. Adamson. 1991. Dev. Biol. 146:499-508). Several cytoskeleton proteins were assayed and only vinculin was found to be absent in mutant (5.51) cells. A chicken vinculin expression vector was transfected into the 5.51 cells together with a neomycin-resistance vector. Clones that were adherent to the substrate were selected in medium containing G418. Two clones, 5.51Vin3 and Vin4, were analyzed by Nomarski differential interference contrast and laser confocal microscopy as well as by biochemical and molecular biological techniques. Both clones adhered well to substrates and both exhibited F-actin stress fibers with vinculin localized at stress fiber tips in focal contacts. This was in marked contrast to 5.51 parental cells, which had no stress fibers and no vinculin. The mutant and complemented F9 cell lines will be useful models for examining the complex interactions between cytoskeletal and cell adhesion proteins.

To define the actin-binding site within the NH2-terminal domain (residues 1-245) of chick smooth muscle alpha-actinin, we expressed a series of alpha-actinin deletion mutants in monkey Cos cells. Mutant alpha-actinins in which residues 2-19, 217-242, and 196-242 were deleted still retained the ability to target to actin filaments and filament ends, suggesting that the actin-binding site is located within residues 20-195. When a truncated alpha-actinin (residues 1-290) was expressed in Cos cells, the protein localized exclusively to filament ends. This activity was retained by a deletion mutant lacking residues 196-242, confirming that these are not essential for actin binding. The actin-binding site in alpha-actinin was further defined by expressing both wild-type and mutant actin-binding domains as fusion proteins in E. coli. Analysis of the ability of such proteins to bind to F-actin in vitro showed that the binding site was located between residues 108 and 189. Using both in vivo and in vitro assays, we have also shown that the sequence KTFT, which is conserved in several members of the alpha-actinin family of actin-binding proteins (residues 36-39 in the chick smooth muscle protein) is not essential for actin binding. Finally, we have established that the NH2-terminal domain of dystrophin is functionally as well as structurally homologous to that in alpha-actinin. Thus, a chimeric protein containing the NH2-terminal region of dystrophin (residues 1-233) fused to alpha-actinin residues 244-888 localized to actin-containing structures when expressed in Cos cells. Furthermore, an E. coli-expressed fusion protein containing dystrophin residues 1-233 was able to bind to F-actin in vitro.

Evidence has accumulated that di- and trisialogangliosides are involved in the interaction of cells with fibronectin. We have therefore tested the ability of variants of BALB/c 3T3 deficient in such gangliosides to organize a fibronectin matrix and to spread on fibronectin-coated substrates. Whereas BALB/c 3T3 cells contained gangliosides GM3, GM1, and GD1a, direct chemical analysis showed that five out of six variants isolated contained no detectable GD1a. By the overlaying of thin layer chromatograms of cellular gangliosides with 125I-cholera toxin, these variants were also found to lack ganglioside GM1. In contrast, the sialogalactoprotein profile of these cells, analyzed using an 125I-ricin/SDS polyacrylamide gel overlay technique, was similar to that of the parent cell line. All variants organized an extensive fibronectin matrix comparable to that of BALB/c 3T3, as shown using either immunofluorescence or lactoperoxidase-catalyzed iodination. The variants could also spread on fibronectin-coated substrates and adopt a morphology similar to that of BALB/c 3T3 cells, with little or no difference in the concentration of fibronectin required for 50% cell spreading. Cell spreading of the variants was accompanied by the formation of focal contacts and microfilament bundles, in a manner closely resembling that seen with BALB/c 3T3 cells. Treatment of BALB/c 3T3 cells with neuraminidase, which converts much of the cellular GD1a to GM1, did not affect cell spreading on fibronectin. The results clearly demonstrate that complex gangliosides are not essential for retention of a fibronectin matrix or for spreading on fibronectin-coated substrates.

We examined the nature of the tetanus toxin receptor in primary cultures of mouse spinal cord by ligand blotting techniques. Membrane components were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose sheets, which were overlaid with 125I-labeled tetanus toxin. The toxin bound only to material at or near the dye front, which was lost when the cells were delipidated before electrophoresis. Gangliosides purified from the lipid extract were separated by thin-layer chromatography and the chromatogram was overlaid with 125I-toxin. The toxin bound to gangliosides corresponding to GD1b and GT1b. Similar results were obtained with brain membranes; thus, gangliosides rather than glycoproteins appear to be the toxin receptors both in vivo and in neuronal cell cultures. To follow the fate of tetanus toxin bound to cultured neurons, we developed an assay to measure cell-surface and internalized toxin. Cells were incubated with tetanus toxin at 0 degree C, washed, and sequentially exposed to antitoxin and 125I-labeled protein A. Using this assay, we found that much of the toxin initially bound to cell surface disappeared rapidly when the temperature was raised to 37 degrees C but not when the cells were kept at 0 degree C. Some of the toxin was internalized and could only be detected by our treating the cells with Triton X-100 before adding anti-toxin. Experiments with 125I-tetanus toxin showed that a substantial amount of the toxin bound at 0 degree C dissociated into the medium upon warming of the cells. Using immunofluorescence, we confirmed that some of the bound toxin was internalized within 15 min and accumulated in discrete structures. These structures did not appear to be lysosomes, as the cell-associated toxin had a long half-life and 90% of the radioactivity released into the medium was precipitated by trichloroacetic acid. The rapid internalization of tetanus toxin into a subcellular compartment where it escapes degradation may be important for its mechanism of action.

We used cholera toxin, which binds exclusively and with a high affinity to the ganglioside GM1, as a probe to investigate the distribution of this glycolipid on the surface of mouse lymphocytes. When lymphocytes are incubated with cholera toxin (or its B subunit) and then sequentially with horse anti-toxin and FITC-swine anti-horse Ig at 37 degrees C, the cholera toxin-ganglioside GM1 complex is redistributed to a cap at one pole of the cell. The capping of cholera toxin-GM1 complexes is slower than the capping of surface-Ig complexes, requires two antibodies, and is inhibited at high toxin concentrations. Cholera toxin-GM1, like surface-Ig capping, is an energy-dependent process and is inhibited by sodium azide, low temperatures, or cytochalasin B, but is unaffected by demecolcine. An affinity-purified antibody against alpha-actinin was used to examine the distribution of this cytoskeletal component during the capping process. 88% of the cells that had a surface Ig cap displayed a co-cap of alpha-actinin, and 57% of the cells that had a cholera toxin-GM1 cap displayed a co-cap of alpha-actinin. Time course studies revealed similar kinetics of external ligand cap formation and the formation of alpha-actinin co-caps. We conclude that capping of a cell-surface glycolipid is associated with a reorganization of the underlying cytoskeleton. The implications of such an association are discussed in the context of current models of the mechanism of capping.

Binding of the cytoskeletal protein vinculin to talin is one of a number of interactions involved in linking F-actin to cell-matrix junctions. To identify the talin binding domain in vinculin, we expressed the NH2-terminal region of the molecule encoded by two closely similar, but distinct vinculin cDNAs, using an in vitro transcription translation system. The 5' Eco RI-Bam HI fragment of a partial 2.89-kb vinculin cDNA encodes a 45-kD polypeptide containing the first 398 amino acids of the molecule. The equivalent restriction enzyme fragment of a second vinculin cDNA (cVin5) lacks nucleotides 746-867, and encodes a 41-kD polypeptide missing amino acids 167-207. The radiolabeled 45-kD vinculin polypeptide bound to microtiter wells coated with talin, but not BSA, and binding was inhibited by unlabeled vinculin. In contrast, the 41-kD vinculin polypeptide was devoid of talin binding activity. The role of residues 167-207 in talin binding was further analyzed by making a series of deletions spanning this region, each deletion of seven amino acids contiguous with the next. Loss of residues 167-173, 174-180, 181-187, 188-194, or 195-201 resulted in a marked reduction in talin binding activity, although loss of residues 202-208 had much less effect. When the 45-kD vinculin polypeptide was expressed in Cos cells, it localized to cell matrix junctions, whereas the 41-kD polypeptide, lacking residues 167-207, was unable to do so. Interestingly, some deletion mutants with reduced ability to bind talin in vitro, were still able to localize to cell matrix junctions.