Probing the Structure of Cytoplasm

We have used size-fractionated, fluorescent dextrans to probe the structure of the cytoplasmic ground substance of living Swiss 3T3 cells by fluorescence recovery after photobleaching and video image processing. The data indicate that the cytoplasm of living cells has a fluid phase viscosity four times greater than water and contains structural barriers that restrict free diffusion of dissolved macromolecules in a size-dependent manner. Assuming these structural barriers comprise a filamentous meshwork, the combined fluorescence recovery after photobleaching and imaging data suggest that the average pore size of the meshwork is in the range of 300 to 400 ~,, but may be as small as 200 A in some cytoplasmic domains. ULK cytoplasm is a gel-like substance that exhibits viscoelastic and thixotropic behavior. That is, there is hysteresis in its recoil after stretching (or compression), and its viscosity is much higher at low, rather than at high, rates of shear. Cytoplasm appears to exist in three interconvertible states: solated, gelated, and contracted. Subcellular, localized transitions between these states are thought by some investigators to be intimately involved in cellular events such as cell division, phagocytosis, locomotion, and intracellnlar transport. (For reviews, see Porter, 1984; Pollard, 1984; Taylor and Condeelis, 1979; Stossel, 1982; for an overview of some current approaches to the study of cytoplasm, see The Cytoplasmic Matrix, 1984, J. Cell Biol. 9911, Pt. 2].) Among the primary candidates for the structural components of the cytoplasmic matrix are three types of filaments found abundantly in cytoplasm, F-actin, microtubules, and intermediate filaments, along with an assortment of accessory proteins that cross-link these filaments in vitro (e.g., Pollard, 1984; Lieska et al., 1985). Morphological studies have indicated that the filamentous components and their associated proteins in some combination may be interconnected into an isotropic network that pervades the cytoplasmic space, and confers on cytoplasm the properties of a viscoelastic, thixotropic, contractile gel (Wolosewick and Porter, 1979; Schliwa and Van Blerkom, 1981; Heuser and Kirschner, 1980; Penman et al., 1981; Stossel, 1982; Pollard, 1984; Taylor and Fechheimer, 1982; Ris, 1985). However, there is little direct evidence from intact, living cells as to precisely how the interaction of filaments and cross-linkers determines the properties of cytoplasm and what role their interaction plays in cellular events. The presence of a cross-linked network of filaments has not yet been demonstrated in living ceils, the fluid phase viscosity of the cytoplasmic gel in living cells is not well understood, and the average "pore size" (better described as the permeation size parameter, or mesh) of the cytoplasmic gel in living cells is completely unknown. Measurements of the fluid phase viscosity of cytoplasm in several types of cells, by electron spin resonance and by low temperature autoradiography of small radiolabeled dextrans have uniformly indicated a value 2-6 times that of water at 22"C (Mastro et al., 1984; Lepock et al., 1983; Paine and Horowitz, 1980). While several of these studies were done specifically to look for cytoplasmic structure, no evidence was found that the cytoplasm is anything other than a fairly concentrated solution of protein. However, as noted by these authors, probes of the relatively small sizes used for these studies would not necessarily have detected higher order structure (Paine and Horowitz, 1980; Mastro et al., 1984; Lepock et al., 1983). In contrast, the resistance to displacement of magnetic particles introduced into intact, living cells suggests that large particles the size of organelles may experience an effective viscosity orders of magnitude greater in cytoplasm than in water (Crick and Hughes, 1950; Sato et al., 1983; Valberg and Albertini, 1985). Fluorescence recovery after photobleaching (FRAP) ~ can be used both to measure the fluid phase viscosity of cytoplasm and to detect higher order structure (such as a cross-linked cytoplasmic gel) in living specimens, without the application of mechanical stress (shear) that damages or disrupts cytoplasmic structure. By using fluorescent molecules of a range of sizes as probes, the mesh of a cytoplasmic gel can be estimated from FRAP data as well. The ideal probes for such a study would be rigid compact spheres of known sizes that do not exhibit long-range interactions with or binding to intracellular components. Originally several laboratories attempted to use fluorescently labeled peptides and globular "control" proteins as probes (Wojcieszyn et al., 1981; Wang et al., 1982; Kreis et al., 1982; Jacobson and Wojcieszyn, 1984; Luby-Phelps et al., 1985). Unfortunately, the cyto1. Abbreviations used in this paper." AF-actin, iodoacetamidofluorescein-labeled actin; D~, aqueous diffusion coefficient; Dc.o, cyloplasmic diffusion coefficient; FRAP, fluorescence recovery after photobleaching; FTC, fluorescein thiocarbamyl; RG, radius of gyration; Rn, hydrodynamic radius; TRTC, tetramethylrhodamine thiocarbamyl. © The Rockefeller University Press, 0021-9525/86/06/2015/08 $1.00 The Journal of Cell Biology, Volume 102, June 1986 2015-2022 2015 on July 1, 2017 jcb.rress.org D ow nladed fom plasmic mobilities of these fluorescent peptides and proteins did not scale inversely with molecular size, suggesting to these investigators that their diffusion in cytoplasm is complicated by transient binding to some unidentified cytoplasmic component (Wojcieszyn et al., 1981; Wang et at., 1982; Gershon et al., 1985; Mastro et al., 1984; Jacobson and Wojcieszyn, 1984; Luby-Phelps et al., 1985). Size-graded fluorescein thiocarbamyl-dextrans (k-q'C-dextrans) appear to be better probes for FRAP studies of cytoplasmic structure than fluorescent proteins, since dextrans do not appear to interact significantly with intracellular components (Paine and Horowitz, 1980). This is supported by our preliminary evidence: FTC-dextrans in the cytoplasm of living 3T3 cells exhibited a single component recovery after photobleaching and recovered to -100% of'the fluorescence intensity before photobleaching. The smallest dextrans indicated a value for cytoplasmic viscosity comparable to previous reports (see above), and mobility in cytoplasm was inversely dependent on molecular size, as predicted by diffusion theory (Luby-Phelps et al., 1985). The dextran backbone is electrically neutral, so the FTC--dextrans have only a very low charge density, conferred on them by the fluorescein residues, thus minimizing the potential for screened electrostatic interactions. FTC-dextrans are not particularly hydrophobic molecules and are readily soluble in water. Dextran does not have a well-defined configuration because it is a flexible, branched random-coil polymer, rather than a rigid, compact sphere. However, its size can be defined statistically by the radius of gyration (Re). We have used fluoresceinand rhodaminelabeled dextrans to probe the fluid phase viscosity and higher order structure of cytoplasm in living 3T3 cells by FRAP and by video image analysis, as a first step toward understanding the structure and function of the cytoplasmic ground substance. Materials and Methods