Several aspects of the behaviour of polyoma virus-transformed BHK cells in culture have suggested that they are more deformable than BHK cells. This possibility was tested by applying negative pressure at the cell surface by means of a micropipette. It was found that PyBHK cells in early mitosis are twice as deformable as BHK cells in the same stage. In addition, the taut, non-ruffling margins of both cell types when fully spread are much less deformable than the extending, ruffling leading lamella. The degree of deformability of these cells is correlated with the distribution and organization of microfilaments and, consistent with this, deform-ability increases greatly in the presence of cytochalasin B. The significance of deformability studies such as these is discussed.
Regional variation in the deformability or extensibility of the cell surface may be important in determining cell morphology and locomotory behaviour. Wolpert (1963) found that during sea-urchin cleavage the cleavage furrow becomes less deformable and Tickle & Trinkaus (1973) noted that an increase in deformability of Fundulus deep cells is correlated with the transformation of nonmotile early blastula cells into locomotory gastrula cells. Since baby hamster kidney fibroblasts (BHK/13) and polyoma virus-transformed BHK fibroblasts differ in important ways, both morphologically and in their contact and locomotory behaviour (Erickson, 1978), it seems possible that differences in deformability could in part be responsible. These differences include: (1) more frequent blebbing of PyBHK cells, especially when spreading on a plane substratum, (2) pulsating of PyBHK cells while moving, and (3) greater ease of nuclear underlapping (greater compressibility) of PyBHK cells (i.e., they appear to be able to squeeze under other cells more readily).
In addition, localized differences in deformability of these cells might also control the orientation and manner of cell locomotion. For example, localized variations in deformability could account for the fact that BHK cells are bipolar and only extend from either end, whereas PyBHK cells are characteristically multipolar and often spread in several directions simultaneously.
A number of lines of evidence suggest that the microfilamentous cortex of the cytoplasm is responsible for cellular resistance to deformation. Microfilaments exist both as a meshwork in the cortex of the cytoplasm, where they apparently insert into the plasma membrane (Pollard & Korn, 1973), and organized in long bundles, which course through the cytoplasm close to the cell surface apposed to the substratum (Buckley & Porter, 1967; McNutt, Culp & Black, 1973; Lazarides, 1976). It has been suggested that these cortical microfilaments might contribute to the rigidity of the cell periphery, in addition to their presumed contractile activity (Wolpert, 1963; Tickle & Trinkaus, 1977). As the cleavage furrow forms in dividing sea-urchin eggs, for example, and a thickened band of cortical microfilaments appears in the cortical cytoplasm of the furrow (Schroeder, 1972), this region becomes highly resistant to deformation (Wolpert, 1963). Also, Mazor & Williamson (1977) have shown that both deformability and phagocytosis in macrophages are dependent upon the actin-myosin interaction. And, finally, Pollard (1976) has found that when F-actin is complexed with myosin in vitro, it is more viscous than F-actin alone. Thus, we might expect that the thicker the microfilamentous cortex, the less deformable the cell will be.
In the present study, deformability was measured in culture in rounded, early mitotic cells and in spread interphase cells of the BHK and PyBHK fibroblastic cell lines, with a modification of the ‘elastimeter’ of Mitchison & Swann (1954).
MATERIALS AND METHODS
Baby hamster kidney fibroblasts (BHK/13) and polyoma virus-transformed baby hamster kidney fibroblast (PyBHK) cell lines, originally established by Stoker & Macpherson (1964), were used in these studies. Both cell lines were a gift of R. D. Goldman. The cells were grown in BHK21 medium (Gibco), supplemented with 10% calf serum, 10% tryptose phosphate broth, and 10% antibiotic-antimycotic solution (Gibco), and were maintained in a 5 % CO2-water-saturated atmosphere at 37 °C. Cultures were transferred every 4 days by removing the cells from the substratum with Viokase (Gibco, 10 ×; diluted 1 to 10 in distilled water).
Determination of deformability
To examine the relative deformability of BHK cells and PyBHK cells, a constant amount of negative pressure was exerted on rounded early mitotic cells by means of a micropipette touched to the cell surface and the length of the protrusion pulled was measured. The apparatus used was a modification of that described by Mitchison & Swann (1954) and consisted of a micropipette connected by way of tygon tubing to a reservoir of water which rested on a lab jack. The jack could be raised or lowered at controllable speeds by a motor to regulate the amount of negative pressure in the pipette. A reservoir of Brodie’s fluid was also placed on the lab jack and connected to a manometer which was calibrated in centimetres. The micropipettes used were pulled on a vertical pipette puller (Lab products) from acetone-washed 1-mm glass tubing (Corning) and broken with a pipette breaker to yield pipettes with internal diameters ranging from 5 to 10 μm. Care was taken to use only those pipettes with smooth tips. Each pipette was calibrated to zero pressure by finding the level of the fluid reservoir where small debris neither flowed into nor out of the pipette. Cells were deformed by lowering the lab jack until 2 cm of negative pressure were being exerted and then bringing the pipette to the surface of a rounded mitotic cell by means of a micromanipulator (Leitz) for one minute. The length of the protrusion pulled in this time interval was measured with a calibrated ocular micrometer. Each pipette was used for measurements on both normal and transformed cells, so that any variability between the 2 cell lines due to pipette shape and inner diameter was eliminated.
The deformation measurements were carried out in a shallow well formed by attaching a glass ring around the culture, which was growing on a coverslip. This well provided both ample space to move the micropipette and excellent optical conditions. A Zeiss inverted microscope with a 40 × planapochromat phase lens was used throughout the investigation.
Some cells were treated with 10 μl/ml cytochalasin B (Aldrich) in 01 % DMSO (dimethyl sulphoxide) for 24 h before testing their deformability.
Scanning electron microscopy
In order to examine the surface contour of the protrusions, cells were fixed for scanning electron microscopy after a protrusion was pulled by quickly removing the medium and adding warm 3 % glutaraldehyde. The cells were then processed for scanning electron microscopy (Porter, Kelley & Andrews, 1972) and relocated in an ETEC scanning electron microscope. Stage tilt was adjusted so that the protrusion was observed perpendicularly to its long axis. Photographs were taken with Polaroid PN55 film.
Measurements of cell surface area
Total surface area of each protrusion was estimated from the scanning electron micrographs, assuming their shapes to be that of a cylinder. From other studies (Erickson & Trinkaus, 1976) we know the surface area per μma of an early mitotic cell to be essentially doubled by microvilli. Thus the total surface area of the region where the protrusion was pulled is the area circum-scribed by the base of the protrusion plus the surface membrane incorporated in microvilli, or approximately twice the area of the base of the protrusion.
Transmission electron microscopy
Both cell types were also examined in the transmission electron microscope. Cells were grown on carbon-shadowed Formvar-coated 300-mesh Athene type grids and fixed and critical point dried according to the method of Buckley & Porter (1975). These cells could then be observed directly in the electron microscope as whole-mount preparations at 80-100 kV.
For preparation of thin sections, cells growing on Falcon plastic were fixed according to a modification of the procedure of McNutt et al. (1971), with 1% paraformaldehyde, 2% glutaraldehyde in 0·1 M sodium cacodylate buffer pH 7 2 with 0·075% CaCl2 for 15 min at 37 °C. After rinsing with cacodylate buffer at room temperature, specimens were osmicated for at least 1 h at 4 °C and finally stained with 0·5 % uranyl acetate in veronal acetate buffer for 1 h at 4 °C. The cells were dehydrated in ethanol at room temperature, embedded in Epon 812 and polymerized at 60 °C until hard.
To make cross-sections, specimens were cut perpendicularly through the Falcon dish according to the method of Eguchi & Okada (1971). To section parallel to the substratum, an area of interest was cut out and glued with epoxy cement to an Epon block, so that the bottom side of the cells adjacent to the dish plastic was closest to the knife, and therefore first to be sectioned. Silver-grey sections were picked up on a 300-mesh Athene type grid, stained with uranyl acetate and Reynold’s lead citrate and examined in a Philips 201 electron microscope.
Rounded early mitotic PyBHK cells are approximately twice as deformable as early mitotic BHK cells, as shown in Table 1. In addition, with relatively wide-bore micropipettes (> 10 μm), whole PyBHK cells can be sucked up into the pipette, whereas BHK cells are never seen to be this deformable, even though early mitotic cells of the 2 cell types have approximately the same total surface area (Erickson, unpublished results). Indeed, PyBHK cells are so deformable that many times a protrusion pulled from one of them did not remain intact, but instead broke up in the pipette into small globules approximately 5·7 μm in diameter. In contrast, BHK cells are much more resistant to negative pressure exerted at their surface. Often a protrusion from the surface of BHK cells can be extended to a length of 30 to 40 μm in a pipette with a 5-μm bore without ever breaking up into small globules, when the negative pressure is increased drastically to over − 10 cm.
When the surfaces of these protrusions are examined in the SEM, they are observed to be quite smooth. The rest of the cell surface, in contrast, is still covered with blebs and microvilli (Erickson & Trinkaus, 1976) (Fig. 1). This suggests that the surface of each protrusion is provided in part by the unfolding of blebs and micro-villi. However, the surfaces of these protrusions are approximately 41 % larger than would be expected if only the microvilli of that region of the cell surface had contributed to their formation (Table 2). This suggests that additional surface must have come from beyond the region subjected to negative pressure, either by surface flow, by the formation of new cell surface, or by membrane stretching.
Because the cells had a tendency to roll after manipulation, it was difficult to measure the speed of retraction of the protuberances. They were generally resorbed into the cell body within a few minutes, however.
Protrusions were also pulled from spread cells of both cell types, but meaningful comparisons between BHK cells and PyBHK cells were not possible since the two cell types are flattened to such different degrees (Erickson, 1978). It is clear, however, that the spreading lamellipodium of a BHK cell is much more deformable than its taut lateral margins. Protrusions up to 10 μm long could be pulled from the extending region of the cell with the standard pressure, in contrast to the side of the cell, where the length of the protrusions never exceeded 2 /tm. Also, if a BHK cell is spread and not blebbing, only a small protrusion 1–2 μm in length could be pulled from the lateral margin with a 5-μm pipette. In contrast, if a BHK cell is spread and blebbing, much longer protrusions (up to 5–10 μm) could be pulled from the lateral margin. Not surprisingly, whatever change in the cell provides for blebbing also makes the cell more deformable.
It was difficult to obtain accurate comparative measurements from spread PyBHK cells since such a large portion of their perimeter forms lamellipodia. In addition, these cells are much less spread than BHK cells, which might be expected to have substantial effects on their deformability. In general, I observed that protrusions between 8 and 10 μm could be pulled from almost any region of the cell, including the ruffling lamellipodia, the lateral margins and the tops of the cells.
When the cells are treated with cytochalasin B for 12 h, and then subjected to negative pressure through a 5μm bore pipette, protrusions up to 16 μm in length can be pulled from early mitotic cells of both types (Table 1). Sixteen micrometres is the maximum length that has ever been achieved with a 5-μm pipette; at that length the protrusion breaks up into globules or into strings of membrane and cytoplasm. In fact, at times, under the influence of CB, even protrusions much shorter than 16 /tm would subdivide into globules. These protrusions, when released, retract slightly, but are never fully resorbed, unlike protrusions of untreated cells which are resorbed in a few minutes. Since treatment with cytochalasin B so drastically affects deformability of both cell types, BHK and PyBHK cells were examined in the transmission electron microscope for amount and arrangement of microfilaments.
Ultrastructural observation of microfilaments
When rounded mitotic cells are examined in the TEM, no cables of microfilaments are observed. Rather, a meshwork of microfilaments is found beneath the cell membrane as observed by others (Goldman & Knipe, 1973; Sanger, 1975; Bragina, Vasiliev & Gelfand, 1976). Although it was not possible to measure accurately the relative thickness of the cortical meshwork of the 2 cell types, several differences were observed in the amount and arrangement of 6–8-nm and 10-nm filaments in various regions of each cell in the spread state, when in both thin section and whole mounts.
In negatively stained, spread BHK cells, many microfilaments (6-8 nm) are arranged in thick bundles or cables, as previously reported (Goldman & Follett, 1969; Goldman & Knipe, 1973) and these bundles are aligned with, and adjacent to, the lateral margins of the cell. Microfilaments, in the form of bundles, are also present in microspikes and extend from these microspikes into the cytoplasm of the lamella (Fig. 2 A). Microfilaments also fill the microvilli, as has been observed by others (e.g. Buckley & Porter, 1967; Follett & Goldman, 1970; Betchaku & Trinkaus, 1978). The ruffling lamellae are primarily filled with a meshwork of microfilaments, as seems to be generally true of spread cells (Goldman & Follett, 1969; Abercrombie, Heaysman & Pegrum, 1971; Yamada, Spooner & Wessells, 1971). More detail can be discerned in thin sections. In 35 thin sections of BHK cells, the bundles of filaments were found to be clearly apposed to the cell surface, especially adjacent to the lateral margins (Fig. 3 A). In addition, there are many 10-nm filaments, especially well seen in the sections cut parallel to the substratum. Therefore, the region of the BHK cell which is most deformable, the leading edge, is filled with a meshwork of microfilaments, while the lateral margins, which are quite resistant to deformation, are underlain by bundles of microfilaments.
Few bundles of microfilaments are observed in whole mounts of PyBHK cells, although they may be obscured in some instances by the thickness of the preparation. Moreover, in 40 cells examined with negative stain, the microfilaments near the margins are organized in a meshwork in all but two. These 2 cells exhibit bundles near the margins. As in BHK cells, the ruffling lamellipodia of PyBHK cells are filled with a meshwork of microfilaments (Fig. 2B). In thin sections PyBHK cells are also observed to have microfilaments in their cortical cytoplasm of their lateral margins, but they are usually arranged in a meshwork (Fig. 3B) and are not nearly as prominent as in BHK cells. Filaments of the 10-nm variety are rarely observed and, when present, are usually found in elongated processes in association with microtubules. The relative absence of bundles of microfilaments in the leading edge and lateral margins of PyBHK cells, in contrast to BHK cells, correlates with their greater deformability, compared to BHK cells, everywhere around the cell perimeter.
These studies show that early mitotic PyBHK cells are twice as deformable as early mitotic BHK cells and that their greater deformability correlates well with behaviour in culture. PyBHK cells bleb profusely especially during early stages of spreading. In consequence, one would predict that their surfaces are much less resistant to internal pressures than BHK cells. Also, PyBHK cells underlap each other easily and even the thick nucleus of one cell can deform and pass under the margin of another. BHK cells appear rigid, however, and nuclei are generally prevented from passing under each other by their very taut margins.
It is probable that the resistance of the cell to deformation is due to the cortical microfilaments associated with the plasma membrane (Mitchison & Swann, 1954; Wolpert, 1963, 1971; Tilney, 1976; Tickle & Trinkaus, 1977). The increase in deformability after CB treatment is consistent with this possibility, since CB either detaches microfilaments from the membrane (Yamada & Wessells, 1973; Miranda, Godman, Deitch & Tanenbaum, 1974a; Miranda, Godman & Tanenbaum, 1974) or, at the very least, causes a drastic change in their organization (Goldman, 1972; Spudich, 1973).
Rounded mitotic cells of both BHK and PyBHK fibroblasts possess only a meshwork of cortical microfilaments associated with the plasma membrane; bundles are not evident (Erickson, unpublished data; Goldman & Knipe, 1973; Sanger, 1975; see also Bragina et al. 1976). It appears, therefore, that the microfilaments in early mitotic cells of both cell types are organized in the same manner, at least as observed in the TEM, and the difference in deformability of these cells might well be due to differences in the thickness or strength of this meshwork. In view of the probability that microfilaments are responsible for the resistance of cells to deformation, it would be of interest to know whether the cortical microfilament meshwork of BHK cells differs from that of the much more deformable Py transformed counterpart. Perhaps their cortical mesh work is thicker or their microfilaments are less labile? Although such determinations have not yet been made for these cells, Wickus, Gruenstein, Robbins & Rich (1975) found that 50% of the actin associated with the membrane of rat embryo fibroblasts is lost upon viral transformation. Similarly, Heaysman & Pegrum (1973) found that chick heart fibroblasts have a thicker microfilamentous cortex than sarcoma cells. Differences in membrane composition and fluidity that might contribute to differences in deformability have not been ruled out, but such differences seem unlikely since both cell types are equally deformable after CB treatment.
Further evidence that differences in the microfilamentous cortex lie at the basis of differences in deformability comes from comparison of the deformability of various regions of the cell. Regional differences in deformability of spread cells clearly correlate with variations in the organization of the microfilamentous cytoskeleton. The lateral marginal regions, which are under tension and where the microfilaments are largely oriented in bundles, are very stiff and not easily deformed, whereas at the leading edge, where microfilaments are organized as meshworks, the cells are much more deformable. This striking regional difference in deformability also correlates with a number of other observations. (1) A cell protrudes forward during locomotion only in the very deformable areas. This evidence supports the hypothesis that cells move forward by forcing cytoplasm into those areas where the membrane and associated microfilamentous cortex are the most deformable. (2) PyBHK cells, which have fewer microfilaments organized into bundles, are less polar and spread in several directions simultaneously, whereas BHK fibroblasts, which have extensive arrays ofmicrofilament bundles aligned with the lateral margins, are highly polar and generally form protrusions only at the end where the microfilaments are organized as a meshwork. (3) A blebbing cell is also quite extensible, as would be expected, since a bleb has little cortex organized into microfilaments (Price, 1967; Trinkaus & Lentz, 1967; Godman, Miranda, Deitch & Tanenbaum, 1975; Fujinami, 1976; Erickson, unpublished).
The basis for the difference in deformability of early mitotic BHK and PyBHK cells is not yet understood, but a brief discussion of what these deformability studies are measuring is in order. Since the surface of a pulled protrusion is quite smooth, compared to the rest of the cell, it seems likely that the membrane in the protrusion comes predominantly from an unfolding of the surface microvilli and blebs which adorn the surface of these cells at this stage (Wolpert & Gingell, 1969; Follett & Goldman, 1970; Erickson & Trinkaus, 1976). However, since the protrusions are approximately 40 % larger than would be expected if only microvilli and blebs have unravelled, some membrane may flow in from the rest of the cell surface. In fact, Tickle & Trinkaus (1977) found when protrusions were pulled from Fundulus deep cells, carbon particles adhering to the cell surface moved toward the micropipette. Membrane stretching and formation of new membrane at the point of deformation have not been ruled out. However, it has been shown that the surface area of erythrocytes remains constant as the cells are swollen before haemolysis (Jay, 1973). Further, Rand & Burton (1964) found that red blood cell membranes are capable of undergoing large amounts of bending, but very little stretching. Thus, it seems likely that deformability studies of fibroblasts reported here and elsewhere probably reflect in large measure the plasticity (or unfolding) of the cell membrane and cytoplasmic cortex rather than ‘stretching’ of the cell membrane (see also Weiss & Clement, 1969).
These deformability studies also support the contention that the cell surface membrane contributes little to the tension at the cell surface. It has been established by Plateau that if a column of fluid is elongated to a length greater than the circumference of the capillary it is pulled through it will be pinched off (for reviews see Thompson, 1961; Danielli, 1945). Rand (1964) observed this same phenomenon in protrusions pulled from red blood cells, which are known to have little visible cortical structure. A protrusion longer than its circumference could not be extended from PyBHK cells without pinching off. In contrast, however, longer protrusions could be pulled from BHK cells, indicating that a force in addition to surface tension is maintaining the integrity of the BHK cell protrusion. When cells are treated with CB, however, protrusions break off at approximately 16 μm, almost precisely the circumference of a 5-μm pipette. Thus, when the microfilamentous cortex is somehow disrupted, the membrane and cytoplasm together act like a uniform column of liquid. It is inferred from these experiments that the integrity of these longer protrusions is largely due to the structure of the cortical cytoplasm and not to the plasma membrane. This is consistent with the observation of Cole (1932) and Harvey (1932) who showed that tension at the surface of a cell is quite low (see Danielli, 1945, for discussion).
I would like to thank my dissertation advisor, Dr J. P. Trinkaus, for his support while this work was in progress and his help in preparing this manuscript. I am indebted to Dr Cheryll Tickle for the discussions that prompted these experiments.
This work was supported by NSF grant BHS 70-00610 and NIH grant USP HS-HD 07137 to J. P. Trinkaus and by a fellowship from NIH training grant HD 00032-11.
This work represents part of a dissertation submitted to the Graduate School of Yale University in partial fulfillment of the requirements for the degree of Doctor of Philosophy.