INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE ROLE OF CYTOSKELETAL FILAMENT networks in modulating nonmuscle contractility is unclear. Giuliano and colleagues (7) have proposed that nonmuscle contractility is regulated both by modulating the activity of molecular motors, such as myosin II, and by altering the cytomatrix in such a manner as to either resist or yield to the tension applied by the motors. This hypothesis is supported by the observation that tension increases in response to disruption of microtubules by nocodazole (5) in cultured fibroblasts (4) and reconstituted fibroblast fibers (21). These data are also consistent with the view that microtubules provide some form of internal resistance. In tensegrity models (6), a portion of a cell's contractile force is borne by rigid internal structures, reducing the force transmitted to external structures, as monitored by force transducers. Alternatively, other mechanisms may underlie the effects of nocodazole. For example, the increase in smooth muscle contractility elicited by nocodazole has been correlated with an increase in intracellular Ca²⁺ concentration ([Ca²⁺]_i) (17).

If microtubules constitute an internal resistance in parallel with the actin-myosin network and are capable of bearing compressive forces, depolymerization of microtubules should result in a greater measured force as the load shifts from internal microtubule structures to the external force transducer. One would also expect that other mechanical parameters, such as cell mechanical stiffness and shortening velocity, would be sensitive to the presence of an intracellular mechanical resistance as that postulated for microtubules. In fact, on theoretical grounds, one might anticipate that the maximum shortening velocity (V_max) would be particularly sensitive to intracellular resistances.

With the use of similar arguments, the effects of microfilament assembly in fibroblast contractility could also be deduced by studying the mechanical effects of cytochalasin D, known to disrupt microfilaments (3).

In this study, we developed a reconstituted fibroblast fiber, based on previous models (19, 21), that enables not only force to be quantitated but also velocity and stiffness in nonmuscle cells. We report that disruption of microfilaments has a profound effect on fibroblast contractility but not [Ca²⁺]_i. Our detailed mechanical analysis, in contrast to hypotheses based on measurement of isometric force alone, does not support a role for microtubule resistance in the mechanical properties of reconstituted fibroblast fibers.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell culture and reconstitution of fibroblast into fibers. Swiss 3T3 fibroblast cells (passages 15-30) were cultured in Dulbecco's modified Eagle's medium (DMEM) plus 10% fetal calf serum (CS) and antibiotic-antimycotic (Ab/Am; 100 U/ml penicillin, 100 µg/ml streptomycin, and 0.25 µg/ml amphotericin B) at 37°C in a humidified, 5% CO₂-95% O₂ atmosphere. Fibroblast fibers were formed by growing 3T3 fibroblasts in rat tail collagen (Upstate Biotechnology, Lake Placid, NY) gel matrix by a modification of the method of Kolodney and Wysolmerski (21). Dispersed cells were suspended in an ice-cold collagen solution that contained 2 × 10⁶ cells/ml and 0.5 mg/ml rat tail collagen in DMEM + 10% CS with added Ab/Am (Sigma). An aliquot (2 ml) of the collagen/cell suspension was poured into a well (0.8 × 5 cm × 0.5 cm deep) cut in a layer of silicone rubber in 100-mm glass petri dishes and placed in a CO₂ incubator at 37°C. After 8 h, an additional 1 ml of DMEM + 10% CS and Ab/Am were added to each well. The preparations were incubated for 3-5 days or until the cells shrank the gel and formed a fiber. Except where noted, cells were placed in serum-free media the night preceding the day on which experiments were conducted.

Mechanics measurements. The fibers made from 3T3 fibroblasts were cut into 5-mm pieces and mounted between glass posts with a cyanoacrylate glue. One post was fixed and the other connected to an AME 801 silicon strain gage (SensoNor) force transducer. The fibers were bathed in physiological salt solution (MOPS-PSS) that contained (in mM): 140 NaCl, 4.7 KCl, 1.2 NaH₂PO₄, 0.02 EDTA, 1.2 MgSO₄, 2.5 CaCl₂, 5.5 dextrose, and 20 MOPS, pH 7.4, at 37°C.

Measurement of the [Ca²⁺]_i. Measurements of [Ca²⁺]_i levels were obtained from cells loaded with the Ca²⁺-sensitive fluorescent dye, fura 2, based on the techniques of Grynkiewicz et al. (10). The fibroblast fibers were loaded with fura 2-AM. Fura 2-AM was prepared as a stock solution of 1 mM dye in DMSO. The fura 2 loading solution contained 3 µM fura 2-AM, 0.015% Pluronic F-127, and 0.5% DMSO in MOPS-PSS buffer. To aid in dispersion of the fura 2-AM, the loading solution was well sonicated. Fibroblast fibers were placed in this solution at room temperature for 3 h. After this loading period, fibroblasts were washed in MOPS-PSS buffer for 20 min to remove any unesterified dye.

Confocal fluorescence microscopy. For phalloidin staining, fibers were fixed with 4% paraformaldehyde and embedded in 17% gelatin. Sections were cut at 100 µm using a Leica VT 1000 vibrating blade microtome and stained with rhodamine-conjugated phalloidin (Sigma, St. Louis, MO). Images were constructed from optical sections acquired by scanning confocal microscopy.

Data analysis. All data are presented as means ± SE. Control and serum data were pooled and include some data previously reported (23). To assess the effects of cytochalasin D and nocodazole, a control contraction was elicited by 30% serum and the mechanical parameters were measured. After treatment with the drugs, the mechanical parameters were again measured. Paired comparisons, with each fiber serving as its own control, were made to control for variability among fibers. Differences were analyzed with paired t-tests; differences with a P value <0.05 were accepted as statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Structure of artificial fibers. The formation of a fiber generally takes 3 days after the cell/collagen suspension is poured into the mold. This transformation is shown in Fig. 1. The final fiber has dimensions of ~40 mm long and 1-2 mm diameter. For contractility measurements, each fiber was cut into smaller segments (~5 mm). These segments are about one-eighth of the whole fiber and would, if proportional, contain ~5 × 10⁵ of the original cells. Histological analysis of sections indicated that the cells were generally aligned along the long axis of the fiber. A more detailed structural analysis of these types of fibers at both aggregate and intracellular levels has been reported (21).

View larger version (175K): [in this window] [in a new window]   Fig. 1.   Artificial fibers reconstituted from Swiss 3T3 fibroblasts. Fibroblast cells reaggregate to form a fiberlike preparation. Fibroblast cells (2 × 106 cells/ml) were cultured in collagen (0.5 mg/ml) medium matrix, as described in MATERIALS AND METHODS. Photograph shows a 100-mm petri dish in which 3 wells (0.8 × 4 cm × 0.5 cm deep) were cut in a layer of silicone rubber, previously cast in the dish. At 24-h intervals, an aliquot (2 ml) of the collagen/cell suspension was poured into 1 well and placed in a CO2 incubator at 37°C. Photograph was taken at 48 h so that the total incubation time was 48 h for mold 1, 24 h for mold 2, and 30 min for mold 3.

Isometric contraction. Fibers were mounted under isometric conditions, then the length was increased to match the original fiber length in the mold. After stress relaxation, force attained a baseline level of 53.0 ± 0.003 µN (n = 30). Addition of serum increased isometric force that returned to prestimulus levels on washout (Fig. 2). These contraction/relaxation cycles could be repeated for at least 8 h, though over prolonged periods some stress relaxation in baseline force was noted. We could not determine a length dependence of force generation as classically done for striated muscle (9) because these preparations did not sustain stable forces when lengthened beyond the length formed in the mold. If untethered, the fibers shortened to approximately one-third the original formed length. Thus, at shorter lengths, the force would decrease. Serum induced dose-dependent contractions (Fig. 3), and a maximum isometric force of 106 ± 12 µN (n = 30) was developed in response to 30% serum. The times to half-maximal contraction and relaxation were 4.7 ± 0.3 and 3.1 ± 0.3 min (n = 22), respectively. Thrombin (2 U/ml) also elicited an increase in force, averaging 19.0 ± 2.0 µN (n = 13) above the prestimulus baseline.

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Serum-induced contraction of a reconstituted fibroblast fiber. A: After mounting in the isometric apparatus, fiber length was increased to its original length at formation in the mold (ST), left scale. After stress relaxation, the force scale was changed (right scale). Contraction was induced by 30% calf serum (CS). PSS indicates a return to the original physiological salt solution. Rapid spikes on solution changes are due to surface tension changes. B: a continuation of A. The serum-induced contraction was reversible, and the contraction-relaxation cycle was reproducible for at least 8 h.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Contractile response of a reconstructed fibroblast fiber to CS. Mean dose-response curve for 10 fibers. Isometric force was normalized to the maximum value for each fiber. Maximum force was 101.8 ± 15.1 µN (n = 10). Fibers were incubated for short (30-60 min, ; n = 5) or long terms (>15 h, ; n = 10) in serum-free medium before force measurements. Inset: isometric force developed in response to cumulative addition of CS to a reconstituted fiber. PSS indicates a return to the original physiological salt solution.

View larger version (9K): [in this window] [in a new window]   Fig. 4.   Effect of cytochalasin D (CytoD) on serum-induced contraction of a reconstituted fibroblast fiber. Contraction was induced by 30% CS and then 1 µM CytoD was added.

Force-velocity relations. These relations were studied by imposing a series of constant speed decreases in length and measurement of the consequent force responses. Figure 5A shows a typical experimental record for a single fiber composed of individual force responses to eight imposed shortenings of varying speeds. Force can be seen to continuously decline with the duration of shortening, similar to behavior reported for rat aorta (22) and hog coronary artery (17). Thus there is a family of force-velocity relations (Fig. 5B), each corresponding to the point in time at which the force values were measured. In the inset to Fig. 5B, V_max taken from each Hill equation was plotted as a function of the point in time at which the forces were measured. V_max decreases rapidly at first, then is relatively stable between 3 and 5 s. The velocities measured at early time points may in fact include discharge of series mechanical compliances, before a quasisteady state is achieved. In subsequent experiments, the force at 5 s was taken for all force-velocity relations. This point in time was chosen as it provided the widest range of measurable values over the differing experimental conditions. This provides a means for determining an operational, relative V_max for comparison of the different conditions.

View larger version (29K): [in this window] [in a new window]   Fig. 5.   Force-velocity relations of reconstituted fibroblast fibers. A: original recording of force in response to the imposition of constant shortening velocities in a serum-stimulated contracture. Length step, L/Lo. B: force-velocity curve. Force (P) at 1 (), 2 (), 3 (), 4 (), and 5 () s normalized to the isometric force (Po) was plotted against the imposed shortening velocity. Inset: time dependency of maximum shortening velocity (Vmax). Vmax taken from each Hill equation was plotted against time at which the forces were measured. Symbols in A have same meaning as in B.

Fiber stiffness. Stiffness was measured by imposition of rapid (<1 ms) step changes in length. Typical responses are shown in Fig. 6A. After imposition of the step, the force response exhibited a peak value followed by stress relaxation. A plot of the peak force responses against the imposed step length change is shown in Fig. 6B. Typical of muscle behavior, the fibroblast peak force responses were linear in the region where stretches were imposed, and this linear range extended to short step decreases. In the serum-stimulated fibers, the slope of the linear relation between force and length (stiffness) was 23.4 ± 0.92 mN/L_o (n = 30). Extrapolation of the linear portion gives an intercept on the length axis of 0.0076 ± 0.0006 L/L_o (n = 30); i.e., a step of 0.76% L_o is required to discharge the maximum isometric force. This value is lower than isolated smooth muscle cells (28) but comparable to that reported for skinned smooth muscle fibers (2).

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Stiffness measurements in reconstructed fibroblast fibers. A: time course of changes in length and force in nonstimulated fiber. B: the change in force plotted against the magnitude of the imposed length step (L/Lo) in nonstimulated () and CS-stimulated () fibers. The peak force reached during rapid length steps (completed within 1 ms) is plotted against the amplitude of the imposed length steps. Regression lines (dashed lines) for force change vs. length change for stretches (positive L/Lo values) are shown in the figure.

Effects of nocodazole and cytochalasin D on ultrastructure and mechanical parameters of reconstituted fibroblast fibers. We used fluorescence microscopy to verify the ability of cytochalasin D and nocodazole to disrupt microfilaments and microtubules in these preparations, respectively. We selected concentrations based on the reported disruption of microtubules and actin filaments in cultured fibroblasts (21). As illustrated in Fig. 7A, intact microtubules were clearly visible in cells populating control fibers, whereas cells from fibers treated with 10 µM nocodazole for 10 min (Fig. 7B) exhibited a diffuse distribution of tubulin, indicating depolymerization of microtubules.

View larger version (53K): [in this window] [in a new window]   Fig. 7.   Immunofluorescent micrographs of fibroblast fibers immunolabeled with anti-tubulin. Control (A) and fibers treated with nocodazole (B; 10 µM) were sectioned and labeled with a monoclonal antibody to -tubulin. Intact microtubules were present in cells from control fibers but not in cells from nocodazole-treated fibers.

View larger version (120K): [in this window] [in a new window]   Fig. 8.   Immunofluorescent micrographs of fibroblast fibers labeled with rhodamine. Polymerized actin was visualized with rhodamine-conjugated phalloidin control fibers (A), fibers treated with CS (B), and fibers treated with CS followed by cytochalasin D (C; 1 µM). Scale bar (A) applies to all panels. Actin polymerization was enhanced by CS but disrupted by cytochalasin D.

View larger version (26K): [in this window] [in a new window]   Fig. 9.   Distribution of individual values for force (A), Vmax (B), and stiffness (C) under control conditions and in response to serum, nocodazole (serum + Noco), and CytoD. The responses for each individual fiber are designated by the same symbol and connected by dotted lines.

View larger version (14K): [in this window] [in a new window]   Fig. 10.   Effects of CS, nocodazole, and cytochalasin D on the relation between maximum shortening velocity and force. Force and Vmax were normalized to those in the resting state. Nocodazole and cytochalasin D were applied after the serum-induced contraction reached a plateau. , CS (Ser, 30%); , nocodazole (Noco, 10 µM); , cytochalasin D (CytoD, 1 µM). Regression lines for CS (dashed line), nocodazole (broken line), and cytochalasin D (solid line) are shown. The level of isometric force, independent of how it is achieved, appears to be the major determinant of Vmax.

View larger version (15K): [in this window] [in a new window]   Fig. 11.   Effects of CS, nocodazole, and cytochalasin D on the relation between stiffness and force. Force and stiffness were normalized to those in the resting state. Nocodazole or cytochalasin D were applied after the serum-induced contraction reached a plateau. Symbols and regression lines have same meanings as in Fig. 10. Note that nocodazole has little effect on the relation between stiffness and force, whereas cytochalasin D elicits major changes.

Effects of nocodazole and cytochalasin D on [Ca²⁺]_i. To further delineate the mechanisms underlying the effects of microtubule or microfilament disruption on contractility, we measured [Ca²⁺]_i using ratiometric fluorescence spectroscopy and the Ca²⁺-sensitive dye fura 2. Nocodazole (10 µM) elicited a small increase in force with little effect on [Ca²⁺]_i (Fig. 12). Addition of serum produced a transient increase in [Ca²⁺]_i that was smaller (72.1% ± 3.8, n = 8, P = 0.002) than that seen in the absence of nocodazole. The addition of serum in the presence of nocodazole elicited a force response that was nearly identical (97.4% ± 0.9, n = 8, P = 0.02) to that in its absence.

View larger version (37K): [in this window] [in a new window]   Fig. 12.   Effects of nocodazole on force (A) and intracellular Ca2+ concentration ([Ca2+]i; B and C) in fibroblast fibers. A: the effects of nocodazole on basal force and on the force developed in response to CS. B: fiber [Ca2+]i measured with fura 2 fluorescence in 2 typical cells. C: the average value for 9 cells in the fiber.

View larger version (39K): [in this window] [in a new window]   Fig. 13.   Effects of cytochalasin D on force (A) and [Ca2+]i (B and C) in fibroblast fibers. A: the effects of cytochalasin D on basal force and on the force developed in response to CS. B: fiber [Ca2+]i measured with fura 2 fluorescence in 2 typical cells. C: the average value for 4 cells in the fiber.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

We have shown that artificial fibers composed of fibroblasts grown in a culture media including collagen provide a unique model for measurement of mechanical parameters in nonmuscle cells. Because cytochalasin D nearly abolishes mechanical responses, these mechanical parameters reflect aggregate cellular properties rather than those of the collagen matrix in which the cells are grown. In the absence of cells, collagen alone does not form fibers or support mechanical loads.

The aggregated cells generate force under unstimulated conditions. This is demonstrated by the ability of the fiber to shorten and by the inhibition of force by cytochalasin D. Both serum and thrombin elicit dose-dependent increases in isometric force. The response to serum of these Swiss fibroblasts was approximately fivefold greater than thrombin. The maximum isometric forces measured in this study were on the order of 150 µN. If one assumes that the original 4 × 10⁶ fibroblasts were evenly distributed throughout the reconstituted fiber, and were 100 µm long with a 10-µm diameter, then the cellular cross-sectional area of the fiber can be calculated as ~0.8 mm². The actual fiber cross-sectional area is difficult to determine exactly but is roughly between 1 and 4 mm². Thus this estimate is not unrealistic and is comparable to some vascular smooth muscle tissue in which smooth muscle cells account for 25% of the vessel wall. Thus force per cellular cross-sectional area would be 0.2 mN/mm². This is considerably less than that for smooth muscle, in which tissue values range from 10 to 200 mN/mm², and potentially larger when translated to smooth muscle cellular cross-sectional areas. However, our calculated value of 0.2 mN/mm² is of similar magnitude to those estimated for fibroblasts from wound contraction models in skin preparations (15).

Hyperbolic force-velocity relations, similar to tonic smooth muscle, were observed in these reconstituted fibers. As force continuously declined with time during the imposed constant shortening, assignment of a maximum velocity depended on the time of measurement. This is similar to that observed in studies of smooth muscle, in which, for example, unloaded velocities showed a similar slowing with time. Whether this is due to some internal resistance, as suggested for smooth muscle (12), or a possible dependence on length (11), is not known. However, the maximum velocities deduced from these force-velocity relations were similar to those visually measured in unloaded chicken embryo fibroblasts cells on cutting from the naturally occurring constraints in the forming molds (21). These maximum velocities are much slower than striated muscle but similar to those reported for the slower, tonic smooth muscles, such as the aorta (25). Velocity and myosin ATPase activities are generally correlated. The low velocity of these fibroblast fibers is consistent with the low-myosin ATPase rates reported for nonmuscle myosin. The actin-activated Mg²⁺-ATPase reported for clonal mouse fibroblast myosin is 90 nmol · min¹ · mg¹ (1). This can be compared with 89 nmol · min¹ · mg¹ for platelet myosin and 51 nmol · min¹ · mg¹ reported for gizzard smooth muscle measured under similar conditions (26).

In assessment of the data as a whole, the level of isometric force appeared to be the major determinant of the speed of shortening. In Fig. 10, the maximum velocities under the various conditions studied were plotted against the level of isometric force. Whether the variations in force were due to the natural variations in response to serum or modified by nocodazole or cytochalasin D, higher levels of force were strongly correlated with higher velocities.

Fibroblast fiber stiffness was also assessed in this model and was found to be similar to that reported for smooth muscle cells but considerably higher than that for isolated smooth muscle tissue. This may reflect the differences in structures through which force is transmitted. Stiffness was related to the level of isometric force (Fig. 11), though this dependence was not as highly correlated as V_max and was sharply reduced after cytochalasin D treatment. The relations between stiffness and force were similar for serum contractures in the presence or absence of nocodazole, which suggest that microtubules are not a major determinant of this parameter. The linear dependence of stiffness on force of these cell aggregates is similar to that observed for individual cells and predicted by tensegrity models (27).

We tested whether microtubules constituted significant mechanical resistance or internal load by measuring the effects of nocodazole, an agent known to disrupt microtubules (16), on shortening velocity and stiffness. Our immunomicroscopy confirmed that nocodazole at 10 µM also disrupted microtubule structures in fibroblasts in our reconstituted fibers (Fig. 7). Nocodazole has been shown in previous studies to increase isometric force in stimulated fibroblast fibers (21) that would be consistent with a reduction in internal resistance opposing the force generated by actin-myosin motors. In contrast to previous studies, we did not consistently observe an increase in force in response to nocodazole (Fig. 9A). However, in our studies, nocodazole was added to cells previously contracted with CS. Contraction with CS sharply decreased the additional response to nocodazole in chicken embryo fibroblasts (18). This finding suggests that, in contrast to the predictions of the tensegrity model, nocodazole and CS may cause contraction through a common signaling pathway that may become saturated by CS, thereby diminishing the nocodazole response.

V_max would be expected to be very sensitive to such an internal resistance. Based on the measured force-velocity relations (Fig. 5B), eliminating an internal load with a magnitude of only 20% of the maximum isometric force would increase fiber V_max by 40%. The effects of nocodazole on V_max depended on its effects on force, which were variable. However, there was no net effect of nocodazole on V_max of serum-stimulated fibers (Table 1).

The relation between the level of isometric force and V_max seen after nocodazole treatment was also similar to that observed for serum stimulation alone (Fig. 10). This provides evidence against the presence of a specific mechanical resistance that can be attributed to microtubules. This is further supported by the effects of nocodazole on fiber stiffness. When nocodazole treatment was associated with increased force, stiffness increased. This is the opposite to expectation if disruption of microtubules removed an internal resistance. The increase may be attributed to more actin-myosin cross bridges under these conditions. This is consistent with recent studies suggesting that nocodazole treatment increases myosin light chain phosphorylation in fibroblasts (18).

The effects of nocodazole on both stiffness and V_max appear to be mediated through its effects on force. Larger forces, independent of how they were generated, were associated with greater V_max. This would be consistent with some type of nonmicrotubule internal resistance in the aggregated cell fiber. If such a resistance was constant, higher forces would lead to relatively less loading and faster velocities. This would also be consistent with the decrease in shortening velocity with time, as has been proposed for smooth muscle cells (12).

In contrast to nocodazole, cytochalasin D, which disrupts actin filaments (3), had dramatic inhibitory effects on the mechanical process(es) underlying all measured parameters. The effects of cytochalasin D on these mechanical parameters was paralleled by its depolymerization of actin filaments in the fibroblast fiber (Fig. 8). Thus the ability to develop force and to shorten is highly dependent on the integrity of the actin filament network. The decreases in both stiffness and V_max by cytochalasin D were parallel to that of its inhibition of force and consistent with force being the primary determinant of these parameters. This would be consistent with actin-myosin interaction being the primary locus of force generation in these fibers.

It is also of interest to consider the possibility of linkage between actin microfilaments and microtubules, as suggested by actin binding properties of some microtubule-associated proteins (8, 14). If, by some manner, actin and microtubule networks are linked, and nocodazole destroys the linkage decreasing parallel cross bridges, then force, as well as stiffness, would be expected to decrease. This is what is observed. If these cross bridges are "lost" to force generation, then this may underlie the loss of force with nocodazole observed in some cases. On the other hand, if the loss of parallel cross bridges is translated into more series cross bridges, then one might anticipate V_max to increase, which was not observed. Further speculation in lieu of high-resolution ultrastructure studies is not warranted. However, these mechanical data provide the first insight into the relations between force, stiffness, and shortening velocity in a nonmuscle contractile system.

Because cytochalasin D and nocodazole appear to be mediated via their effect on force, it was of interest to assess their effects on [Ca²⁺]_i. The mechanism of activation of nonmuscle myosin is controversial (20, 23), but [Ca²⁺]_i is likely a key second messenger. Treatment with nocodazole reduced the increase in [Ca²⁺]_i in response to serum by ~30%, but had little effect on the developed force. Cytochalasin D, which dramatically reduced force, had an even lesser effect than nocodazole on the Ca²⁺ response to serum. Thus it is unlikely that the effects of either were mediated primarily through effects on [Ca²⁺]_i.

In summary, reconstituted fibers formed by fibroblasts cultured in a collagen gel can be used for precise mechanical measurements in aggregated nonmuscle cells. Although the isometric force generated by these reconstituted fibers is significantly lower, maximum velocity and compliance are similar to that of smooth muscle fibers. In terms of the contractile response to serum, actin microfilaments play a major role in the mechanical properties of the reconstituted fibers, as might be expected for a myosin motor-driven system. Microtubules do not make a major contribution to an internal mechanical resistance or load against which the myosin system has been proposed to operate (7). Because the composition of the major components in these cultured cells can be readily manipulated by genetic techniques (23), this model system may provide a unique approach to the understanding of contractile mechanisms in nonmuscle as well as cultured muscle cells.