Forces on adhesive contacts affect cell function.

Cellular forces acting on the adhesive contacts made with the extracellular matrix (ECM) contribute significantly to cell shape, viability, signal transduction and motility. In the past two years, research has determined how cell spreading influences cell viability as well as cytoskeletal organization. The cytoskeleton generates a level of tension against the ECM that is proportional to ECM stiffness. The strength of this tension exerted against the ECM affects the migratory speed of the cell.

Selective regulation of integrin--cytoskeleton interactions by the tyrosine kinase Src.

Cell motility on extracellular-matrix (ECM) substrates depends on the regulated generation of force against the substrate through adhesion receptors known as integrins. Here we show that integrin-mediated traction forces can be selectively modulated by the tyrosine kinase Src. In Src-deficient fibroblasts, cell spreading on the ECM component vitronectin is inhibited, while the strengthening of linkages between integrin vitronectin receptors and the force-generating cytoskeleton in response to substrate rigidity is dramatically increased. In contrast, Src deficiency has no detectable effects on fibronectin-receptor function. Finally, truncated Src (lacking the kinase domain) co-localizes to focal-adhesion sites with alpha v but not with beta 1 integrins. These data are consistent with a selective, functional interaction between Src and the vitronectin receptor that acts at the integrin-cytoskeleton interface to regulate cell spreading and migration.

Mutational evidence for control of cell adhesion through integrin diffusion/clustering, independent of ligand binding.

Previous studies have shown that integrin alpha chain tails make strong positive contributions to integrin-mediated cell adhesion. We now show here that integrin alpha4 tail deletion markedly impairs static cell adhesion by a mechanism that does not involve altered binding of soluble vascular cell adhesion molecule 1 ligand. Instead, truncation of the alpha4 cytoplasmic domain caused a severe deficiency in integrin accumulation into cell surface clusters, as induced by ligand and/ or antibodies. Furthermore, alpha4 tail deletion also significantly decreased the membrane diffusivity of alpha4beta1, as determined by a single particle tracking technique. Notably, low doses of cytochalasin D partially restored the deficiency in cell adhesion seen upon alpha4 tail deletion. Together, these results suggest that alpha4 tail deletion exposes the beta1 cytoplasmic domain, leading to cytoskeletal associations that apparently restrict integrin lateral diffusion and accumulation into clusters, thus causing reduced static cell adhesion. Our demonstration of integrin adhesive activity regulated through receptor diffusion/clustering (rather than through altered ligand binding affinity) may be highly relevant towards the understanding of inside-out signaling mechanisms for beta1 integrins.

Cell spreading as a hydrodynamic process.

Many cell types have the ability to move themselves by crawling on extra-cellular matrices. Although cell motility is governed by actin and myosin filament assembly, the pattern of the movement follows the physical properties of the network ensemble average. The first step of motility, cell spreading on matrix substrates, involves a transition from round cells in suspension to polarized cells on substrates. Here we show that the spreading dynamics on 2D surfaces can be described as a hydrodynamic process. In particular, we show that the transition from isotropic spreading at early time to anisotropic spreading is reminiscent of the fingering instability observed in many spreading fluids. During cell spreading, the main driving force is the polymerization of actin filaments that push the membrane forward. From the equilibrium between the membrane force and the cytoskeleton, we derive a first order expression of the polymerization stress that reproduces the observed behavior. Our model also allows an interpretation of the effects of pharmacological agents altering the polymerization of actin. In particular we describe the influence of Cytochalasin D on the nucleation of the fingering instability.

Modulation of membrane dynamics and cell motility by membrane tension.

The plasma membrane of most cells is drawn tightly over the cytoskeleton of the cell, resulting in a significant tension being developed in the membrane. The tension in the membrane can be calculated from the force required to separate it from the cytoskeleton; and the force itself can be measured rapidly by using laser tweezers. Recent observations indicate that decreasing membrane tension stimulates endocytosis and increasing tension stimulates secretion. Thus, membrane tension provides a simple physical mechanism to control the area of the plasma membrane. Here, we speculate that tension is a global parameter that the cell uses to control physically plasma membrane dynamics, cell shape and cell motility.

Cell migration: regulation of force on extracellular-matrix-integrin complexes.

Cell migration relies upon forces generated by the cell. Recent studies have provided new insights into the processes by which cells generate and regulate the forces applied to extracellular matrix (ECM)-bound integrins and have led us to the working model described here. In this model, ECM binding to integrins in the front of lamellipodia causes those integrins to attach to the rearward-moving cytoskeleton. Integrin-cytoskeleton attachments in the front are strengthened as a result of ECM rigidity, enabling the cell to pull itself forward. The reduction in contact area at the rear compared with that at the lamellipodium concentrates the traction forces in the rear on fewer integrin-ECM bonds, facilitating release. In such a model, cell pathfinding and motility can be influenced by ECM rigidity.

DNase-I-dependent dissociation of erythrocyte cytoskeletons.

The human erythrocyte contains a complex of peripheral membrane proteins which forms an extensive network or cytoskeleton on the cytoplasmic membrane surface. When I treat erythrocyte cytoskeletons with deoxyribonuclease I (DNase I), the cytoskeletons dissociate and erythrocyte actin is solubilized. The dissociation of the cytoskeletons by DNase I parallels the disruption of actin filaments in vitro by DNase I and is blocked by the addition of action to the DNase I. Large protein complexes remain after DNase I disrupts the cytoskeletons, but these complexes are no longer visible in the light microscope nor sedimentable and are selectively depleted with respect to actin. From these studies, I suggest that DNase I binds to and solubilizes actin, which serves as a structural link between protein complexes in the erythrocyte cytoskeleton.

Cell spreading and lamellipodial extension rate is regulated by membrane tension.

Cell spreading and motility require the extension of the plasma membrane in association with the assembly of actin. In vitro, extension must overcome resistance from tension within the plasma membrane. We report here that the addition of either amphiphilic compounds or fluorescent lipids that expanded the plasma membrane increased the rate of cell spreading and lamellipodial extension, stimulated new lamellipodial extensions, and caused a decrease in the apparent membrane tension. Further, in PDGF-stimulated motility, the increase in the lamellipodial extension rate was associated with a decrease in the apparent membrane tension and decreased membrane-cytoskeleton adhesion through phosphatidylinositol diphosphate hydrolysis. Conversely, when membrane tension was increased by osmotically swelling cells, the extension rate decreased. Therefore, we suggest that the lamellipodial extension process can be activated by a physical signal (perhaps secondarily), and the rate of extension is directly dependent upon the tension in the plasma membrane. Quantitative analysis shows that the lamellipodial extension rate is inversely correlated with the apparent membrane tension. These studies describe a physical chemical mechanism involving changes in membrane-cytoskeleton adhesion through phosphatidylinositol 4,5-biphosphate-protein interactions for modulating and stimulating the biochemical processes that power lamellipodial extension.

Control of the erythrocyte membrane shape: recovery from the effect of crenating agents.

Intact erythrocytes become immediately crenated upon addition of 2,4-dinitrophenol (DNP) or pyrenebutyric acid (PBA). However, when cells are incubated at 37 degrees C in the presence of the crenating agents with glucose, they gradually (4--8 h) recover the normal biconcave disc form. The recovery process does not reflect a gradual inactivation of DNP or PBA since fresh cells are equally crenated by the supernatant from the recovered cells. Further, after recovery and removal of the crenating agents, cells are found to be desensitized to the readdition of DNP as well as to the addition of PBA, but they are more sensitive to cupping by chlorpromazine. This alteration in the cell membrane responsiveness was reversible upon further incubation in the absence of DNP. Recovery is dependent upon cellular metabolic state since an energy source is needed and incubation with guanosine but not adenosine will accelerate conversion to the disc shape. It is suggested that the conversion of cells from crenated to disc shape in the presence of the crenators, represents an alteration or rearrangement of membrane components rather than a redistribution of the crenators within the membrane. This shape recovery process may be important for erythrocyte shape preservation as well as shape control in other cells.

Membrane expansion increases endocytosis rate during mitosis.

Mitosis in mammalian cells is accompanied by a dramatic inhibition of endocytosis. We have found that the addition of amphyphilic compounds to metaphase cells increases the endocytosis rate even to interphase levels. Detergents and solvents all increased endocytosis rate, and the extent of increase was in direct proportion to the concentration added. Although the compounds could produce a variety of different effects, we have found a strong correlation with a physical alteration in the membrane tension as measured by the laser tweezers. Plasma membrane tethers formed by latex beads pull back on the beads with a force that was related to the in-plane bilayer tension and membrane- cytoskeletal adhesion. We found that as cells enter mitosis, the membrane tension rises as the endocytosis rate decreases; and as cells exited mitosis, the endocytosis rate increased as the membrane tension decreased. The addition of amphyphilic compounds decreased membrane tension and increased the endocytosis rate. With the detergent, deoxycholate, the endocytosis rate was restored to interphase levels when the membrane tension was restored to interphase levels. Although biochemical factors are clearly involved in the alterations in mitosis, we suggest that endocytosis is blocked primarily by the increase in apparent plasma membrane tension. Higher tensions inhibit both the binding of the endocytic complex to the membrane and mechanical deformation of the membrane during invagination. We suggest that membrane tension is an important regulator of the endocytosis rate and alteration of tension is sufficient to modify endocytosis rates during mitosis. Further, we postulate that the rise in membrane tension causes cell rounding and the inhibition of motility, characteristic of mitosis.

Two activators of microtubule-based vesicle transport.

Cytoplasmic dynein purified by nucleotide dependent microtubule affinity has significant minus end-directed vesicle motor activity that decreases with each further purification step. Highly purified dynein causes membrane vesicles to bind but not move on microtubules. We exploited these observations to develop an assay for factors that, in combination with dynein, would permit minus end-directed vesicle motility. At each step of the purification, non-dynein fractions were recombined with dynein and assayed for vesicle motility. Two activating fractions were identified by this method. One, called Activator I, copurified with 20S dynein by velocity sedimentation but could be separated from it by ion exchange chromatography. Activator I increased only the frequency of dynein-driven vesicle movements. Activator II, sedimenting at 9S, increased both the frequency and velocity of vesicle transport and also supported plus end movements. Our results suggest that dynein-based motility is controlled at multiple levels and provide a preliminary characterization of two regulatory factors.

Keratocytes pull with similar forces on their dorsal and ventral surfaces.

As cells move forward, they pull rearward against extracellular matrices (ECMs), exerting traction forces. However, no rearward forces have been seen in the fish keratocyte. To address this discrepancy, we have measured the propulsive forces generated by the keratocyte lamella on both the ventral and the dorsal surfaces. On the ventral surface, a micromachined device revealed that traction forces were small and rearward directed under the lamella, changed direction in front of the nucleus, and became larger under the cell body. On the dorsal surface of the lamella, an optical gradient trap measured rearward forces generated against fibronectin-coated beads. The retrograde force exerted by the cell on the bead increased in the thickened region of the lamella where myosin condensation has been observed (Svitkina, T.M., A.B. Verkhovsky, K.M. McQuade, and G. G. Borisy. 1997. J. Cell Biol. 139:397-415). Similar forces were generated on both the ventral (0.2 nN/microm(2)) and the dorsal (0.4 nN/microm(2)) surfaces of the lamella, suggesting that dorsal matrix contacts are as effectively linked to the force-generating cytoskeleton as ventral contacts. The correlation between the level of traction force and the density of myosin suggests a model for keratocyte movement in which myosin condensation in the perinuclear region generates rearward forces in the lamella and forward forces in the cell rear.

Equilibrium and kinetic effects of drugs on the shapes of human erythrocytes.

We have previously proposed that if the two half-layers of a membrane are different in their protein and lipid compositions, they may respond differently to some membrane perturbation (the bilayer couple hypothesis). This hypothesis has been applied to explain the changes in shape of human erythrocytes that are produced by a variety of amphipathic compounds. These compounds are presumed to intercalate by their hydrophobic ends into the lipid portions of the membrane; if the compounds are anions, the binding is preferentially to the outer half of the bilayer, if cations, to the inner half. It is proposed that such preferential binding causes an expansion of one half-layer relative to the other, with a corresponding change in cell shape. The predicted sidedness of these shape changes is now demonstrated in experiments with methochlorpromazine and 2,4,6-trinitrophenol. Under appropriate nonequilibrium or equilibrium or equilibrium conditions, both of these compounds are shown to be either crenators or cup-formers of the intact erythrocyte, depending upon which side of the membrane they are concentrated in. These results therefore strongly support the bilayer couple hypothesis.

On the mechanism of ATP-induced shape changes in human erythrocyte membranes. I. The role of the spectrin complex.

Human erythrocyte ghosts have been shown, by scanning electron microscopy, to undergo ATP-dependent shape changes. Under appropriate conditions the ghosts prepared from normal disk-shaped intact cells adopt a highly crenated shape, which in the presence of Mg-ATP at 37 degrees C is slowly converted to the disk shape and eventually to the cup shape. These changes are not observed with other nucleotides or with 5'-adenylyl imidodiphosphate. Anti-spectrin antibodies, incorporated along with the Mg-ATP into the ghosts in amounts less than equivalent to the spectrin, markedly accelerate the shape changes observed with the Mg-ATP alone. The Fab fragments of these antibodies, however, have no effect. The conclusion is that the structural effect produced by the ATP is promoted by the cross-linking of spectrin by its antibodies, and may therefore itself be some kind of polymerization or network formation involving the spectrin complex on the cytoplasmic face of the membrane. The factors that contribute to the shape of the ghost and of the intact erythrocyte are discussed in the light of these findings.

ATP-dependent movement of myosin in vitro: characterization of a quantitative assay.

Sheetz and Spudich (1983, Nature (Lond.), 303:31-35) showed that ATP-dependent movement of myosin along actin filaments can be measured in vitro using myosin-coated beads and oriented actin cables from Nitella. To establish this in vitro movement as a quantitative assay and to understand better the basis for the movement, we have defined the factors that affect the myosin-bead velocity. Beads coated with skeletal muscle myosin move at a rate of 2-6 micron/s, depending on the myosin preparation. This velocity is independent of myosin concentration on the bead surface for concentrations above a critical value (approximately 20 micrograms myosin/2.5 X 10(9) beads of 1 micron in diameter). Movement is optimal between pH 6.8 and 7.5, at KCl concentrations less than 70 mM, at ATP concentrations greater than 0.1 mM, and at Mg2+ concentrations between 2 and 6 mM. From the temperature dependence of bead velocity, we calculate activation energies of 90 kJ/mol below 22 degrees C and 40 kJ/mol above 22 degrees C. Different myosin species move at their own characteristic velocities, and these velocities are proportional to their actin-activated ATPase activities. Further, the velocities of beads coated with smooth or skeletal muscle myosin correlate well with the known in vivo rates of myosin movement along actin filaments in these muscles. This in vitro assay, therefore, provides a rapid, reproducible method for quantitating the ATP-dependent movement of myosin molecules on actin.

Cell migration does not produce membrane flow.

We have previously reported that rearward migration of surface particles on slowly moving cells is not driven by membrane flow (Sheetz, M. P., S. Turney, H. Qian, and E. L. Elson. 1989. Nature (Lond.). 340:284-288) and recent photobleaching measurements have ruled out any rapid rearward lipid flow (Lee, J., M. Gustafsson, D. E. Magnussen, and K. Jacobson. 1990. Science (Wash. DC.) 247:1229-1233). It was not possible, however, to conclude from those studies that a slower or tank-tread membrane lipid flow does not occur. Therefore, we have used the technology of single particle tracking to examine the movements of diffusing particles on rapidly locomoting fish keratocytes where the membrane current is likely to be greatest. The keratocytes had a smooth lamellipodial surface on which bound Con A-coated gold particles were observed either to track toward the nuclear region (velocity of 0.35 +/- 0.15 micron/s) or to diffuse randomly (apparent diffusion coefficient of [3.5 +/- 2.0] x 10(-10) cm2/s). We detected no systematic drift relative to the cell edge of particles undergoing random diffusion even after the cell had moved many micrometers. The average net particle displacement was 0.01 +/- 2.7% of the cell displacement. These results strongly suggest that neither the motions of membrane proteins driven by the cytoskeleton nor other possible factors produce a bulk flow of membrane lipid.

Light chain phosphorylation regulates the movement of smooth muscle myosin on actin filaments.

In smooth muscles there is no organized sarcomere structure wherein the relative movement of myosin filaments and actin filaments has been documented during contraction. Using the recently developed in vitro assay for myosin-coated bead movement (Sheetz, M.P., and J.A. Spudich, 1983, Nature (Lond.)., 303:31-35), we were able to quantitate the rate of movement of both phosphorylated and unphosphorylated smooth muscle myosin on ordered actin filaments derived from the giant alga, Nitella. We found that movement of turkey gizzard smooth muscle myosin on actin filaments depended upon the phosphorylation of the 20-kD myosin light chains. About 95% of the beads coated with phosphorylated myosin moved at velocities between 0.15 and 0.4 micron/s, depending upon the preparation. With unphosphorylated myosin, only 3% of the beads moved and then at a velocity of only approximately 0.01-0.04 micron/s. The effects of phosphorylation were fully reversible after dephosphorylation with a phosphatase prepared from smooth muscle. Analysis of the velocity of movement as a function of phosphorylation level indicated that phosphorylation of both heads of a myosin molecule was required for movement and that unphosphorylated myosin appears to decrease the rate of movement of phosphorylated myosin. Mixing of phosphorylated smooth muscle myosin with skeletal muscle myosin which moves at 2 microns/s resulted in a decreased rate of bead movement, suggesting that the more slowly cycling smooth muscle myosin is primarily determining the velocity of movement in such mixtures.

Kinectin, a major kinesin-binding protein on ER.

Previous studies have shown that microtubule-based organelle transport requires a membrane receptor but no kinesin-binding membrane proteins have been isolated. Chick embryo brain microsomes have kinesin bound to their surface, and after detergent solubilization, a matrix with an antibody to the kinesin head domain (SUK-4) (Ingold et al., 1988) bound the solubilized kinesin and retained an equal amount of a microsome protein of 160-kD. Similarly, velocity sedimentation of solubilized membranes showed that kinesin and the 160-kD polypeptide cosedimented at 13S. After alkaline treatment to remove kinesin from the microsomes, the same 160-kD polypeptide doublet bound to a kinesin affinity resin and not to other proteins tested. Biochemical characterization localized this protein to the cytoplasmic face of brain microsomes and indicated that it was an integral membrane protein since it was resistant to alkaline washing. mAbs raised to chick 160-kD protein demonstrated that it was absent in the supernatant and concentrated in the dense microsome fraction. The dense microsome fraction also had the greatest amount of microtubule-dependent motility. With immunofluorescence, the antibodies labeled the ER in chick embryo fibroblasts (similar to the pattern of bound kinesin staining in the same cells) (Hollenbeck, P. J. 1989. J. Cell Biol. 108:2335-2342), astroglia, Schwann cells and dorsal root ganglion cells but staining was much less in the Golgi regions of these cells. Because this protein is a major kinesin-binding protein of motile vesicles and would be expected to bind kinesin to the organelle membrane, we have chosen the name, kinectin, for this protein.

Chemical subdomains within the kinetochore domain of isolated CHO mitotic chromosomes.

We have used indirect immunofluorescence in combination with correlative EM to subdivide the mammalian kinetochore into two domains based on the localization of specific antigens. We demonstrate here that the fibrous corona on the distal face of the kinetochore plate contains tubulin (previously shown by Mitchison, T. J., and M. W. Kirschner. 1985. J. Cell Biol. 101:755-765) and the minus end-directed, ATP-dependent microtubule motor protein, dynein; whereas a 50-kD CREST antigen is located internal to these components in the kinetochore. Tubulin and dynein can be extracted from the kinetochore by 150 mM KI, leaving other, as yet uncharacterized, components of the kinetochore corona intact. Microtubules and tubulin subunits will associate with kinetochores in vitro after extraction with 150 mM KI, suggesting that other functionally significant, corona-associated molecules remain unextracted. Our results suggest that the corona region of the kinetochore contains the machinery for chromosome translocation along microtubules.

Nuclear spreads: I. Visualization of bipartite ribosomal RNA domains.

Visualization of nuclear architecture is key to the understanding of the association between RNA synthesis and processing. This architecture is obscured by the high density of components in most nuclei. We have developed a method of spreading nuclei and nucleoli that reduces overlap of weakly associated components. Strong interactions among nuclear components are not disrupted by this method. Spread nucleoli remained structurally distinct and functionally competent in ribosomal RNA synthesis. Nascent ribosomal RNA colocalized with RNA polymerase I and fibrillarin, a protein required for processing of ribosomal RNA. Colocalization of nascent transcripts and fibrillarin was seen in nucleoli spread over several microns, suggesting a strong interaction. These data suggest that nucleoli are superassemblies of bipartite domains, each composed of a ribosomal RNA synthesis center tightly associated with areas likely to be involved in ribosomal RNA processing.