Nagashima and Asakura reported that single filaments of F-actin decorated with heavy meromyosin (HMM) or S-l are directly visible under an optical microscope with dark-field illumination, although the filaments without HMM or S-l decoration are not visible . Recently, S. Asakura (personal communication) found that single filaments of F-actin, without HMM or S-l decoration, are made visible under a fluorescence microscope by labelling with fluorescein isothyocyanate. Using the fluorescent dye, phalloidin-rhodamine, we could clearly and continuously observe the movement of single filaments of F-actin or F-actin-tropomyosin (TM)-troponin (TN) complex for more than 5 min without photobleaching. Phalloidin-rhodamine is most convenient for the present purpose, because it is specifically bound to F-actin without any influence on physiological functions of F-actin7'8 and its fluorescence intensity is enhanced 4.2-fold by the binding. When G-actin is polymerized in the presence of phalloidin-rhodamine, many filaments become visible under a fluorescence microscope. Figure 1 shows micrographs of phalloidin rhodamine-labelled F-actin and F-actin-TM-TN complex in various conditions. All fluorescent filaments have the same brightness and the brightness is uniform along the filament. The number of filaments visible in the field of microscope agrees with the number expected from the concentration of F-actin. The visible filaments are not bundles of F-actin but single filaments of F-actin.Table 1 Flexural rigidity (?) of F-actin and F-actin-TM-TN complexes SampleMedium e(l(T17dyncm2)FA +Ca0.032 6.5FA+HMM +Ca0.042 4.9+Ca+ATP 0.0593.5FA+TM+TO -Ca0.028 7.4+Ca 0.0405.2 FA+TM+TN+HMM-Ca 0.0258.3 -l-Ca0.045 4.6-Ca+ATP 0.0277.7 +Ca+ATP0.0533.9FA-PM +Ca0.033 6.4The concentration of HMM was 5 mg HMM mr1. +Ca, rigor buffer -1-0.2 mM CaCl2; -Ca, rigor buffer +1 mM EGTA; +Ca+ATP, activating buffer; -Ca+ ATP, relaxing buffer. The flexural rigidity, e, was determined by measuring the end-to-end distance, R. About 50 values of R were measured for each filament in a 2-10-s period and the mean square end-to-end distance, (R2), was related to a parameter of flexibility, ?, by the equation (R2} = [2AL-1 + exp(-2AL)]/2A2 (rf. 16). The most probable values of A were obtained from a least-square fit to the plots of L versus V1?2) of more than 10 different filaments 5-10 ??? in contour length, L. The flexural rigidity, e, was calculated from the value of A by the equation ekT/2\ (rf. 17). FA-PM, F-actin cross-linked by p-phenylene-N,N'-bis(maleimide) by the method of Knight and Offer14.In the activating buffer, the motion is not a thermal one, but the apparent values of A and ? determined by the above procedure are shown for convenience to compare the extent of bending (see text). The thermal bending motion of the filaments is observable under the microscope. Table 1 gives the flexural rigidity of F-actin and F-actin-TM-TN complex estimated by the statistical analysis of the end-to-end distance of a number of filaments. The values of the flexural rigidity are in good agreement with those of unlabelled F-actin and F-actin-TM-TN measured by different methods69"11. This indicates that the binding of phal-loidin does not affect the dynamic property of F-actin although it stabilizes F-actin against depolymerization at low ionic strength or high concentration of KI12. The binding of Ca2decreased the rigidity of the complex and the binding of HMM further decreased the rigidity in the presence of Ca. When both HMM and Mg2-ATP are added to the F-actin-TM-TN complex in the presence of Ca2, the bending motion of long filaments becomes faster and larger in amplitude, and the tumbling motion of shorter filaments becomes conspicuous. During such motions, long filaments are often broken into fragments of up to a few micrometres (Fig. Id). Figure 2 shows sequences of micrographs taken to compare the bending motion of the F-actin-TM-TN filaments in a solution containing HMM and Mg2+-ATP, in the absence (Fig. 2?) and presence (Fig. 2b) of Ca2. The filaments in Fig. 2? and b have same length, 10 ???. In the absence of Ca2(Fig. 2?), the bending motion of the first order mode has an apparent period of 3-6 s, and that of the second order mode has a period of about 0.3 s. Previously, by using dynamic light scattering, the relaxation time of the bending motion of F-actin of 2.5 ??? in length was determined to be about 10 ms9. The above value, 3-6 s, for the filaments of 10 ??? length is reasonable because the relaxation time, rn, of the bending motion is proportional to the fourth power of the length, as given by where ? is the mode number, L the contour length of the filament, ? the frictional constant for a structural unit and ? the flexural rigidity9.In the presence of Ca2(Fig. 2), the apparent period of the bending motion of the first mode decreased to 1-3 s. If the bending is due to thermal brownian motion, the flexural rigidity in the presence of Ca2must be about twice as large as that in its absence, according to equation (1). However, as found in Figure 2? and , the average amplitude of bending of the filaments is distinctly larger in the presence of Ca2H" than in its absence. Therefore, we conclude that more rigid filaments in the presence of Ca2"" were forcibly bent by interaction with HMM and ATP. Conversely, it is also possible to conclude that the bending motion of the filament which is more flexible in the presence of Ca2+ was accelerated by interaction with HMM and Mg2-ATP. The latter case is more likely because the F-actin-TM-TN complex is more flexible in the presence of Ca2+ as shown in Table 1. In the present conditions (0.1 M KC1, 20 nM actin and 15 ?? HMM), the rate of hydrolysis of ATP is estimated to be about 1 mole per mole actin per s (rf. 13). That is, each actin monomer is attacked by energized HMM (HMM-ADP-Pi) about once per second. The energy liberated by the hydrolysis may be (partially) transferred to F-actin to induce its active motion. Fig. 1 Micrographs of phalloidin-rhodamine-labelled F-actin: ? and b are for the F-actin in rigor buffer (0.1MKC1, 5mMMgCl2 and 10 mM MOPS pH 7.0) in theabsence and the presence of 10 mg HMM ml"1, respectively, c and d are for the F-actin-TM-TN complexes in relaxing buffer (rigor buffer +1 mM EGTA, 3 mM ATP, 5 mM creatine phosphate and 2 mg creatine kinase per ml) and in activating buffer (rigor buffer +0.2 mM CaCl2,3 mM ATP, 5 mM creatine phosphate and 2 mg creatine kinase per ml) containing 10 mg HMM ml"1, respectively. All proteins were obtained from rabbit skeletal muscle and purified according to the method described previously10. G-actin of 1 ?? was polymerized in rigor buffer containing 1 ?? phalloidin-rhodamine overnight at room temperature. TM and TN were added to the F-actin solution at a weight ratio of 1:4, respectively. For observation the solutions were diluted 50-fold in buffer and put in a gap between a glass slide and a coverslip of 50 ??? in thickness. Final concentration of actin was 20 nM. -Mercaptoethanol of 0.1-0.5% (v/v) which did not affect actomyosin ATPase activity or muscle contraction was added to the solutions to suppress photobleaching. Fluorescent F-actin or F-actin-TM-TN complexes were observed with a Nikon VFD-TR microscope equipped with epifluorescence optics, a Nikon UV-F x 100 objective (oil immersion, NA 1.3), a 100-200-W mercury arc lamp and a Nikon rhodamine filter set (AEX 540 nm, AEM> 580 nm). Fluorescent images were recorded on videotape with a high-sensitivity television camera (Ikegami CTC-900) and a video recorder (National Time Laps NV8030). Fig. 2 Sequential micrographs of phalloidin-rhodamine-labelled F-actin having bound TM and TN in relaxing buffer +5 mg HMM mlVa) and in activating buffer +5 mg HMM ml"1 (b, 1, 2). The intervals between successive images were 0.15 s (a) and 0.10 s (b, l, 2), respectively. The contour length of both filaments was 10 ???. Single arrows indicate first-order mode bending motion and double ones second order mode bending motion. As measured by a semiconductor thermometer, the temperature of the solution was 25 C during illumination. No increase in the temperature of the solution during ATP hydrolysis by HMM, S-l, acto-HMM or acto-S-1 was detectable. The resolution of TV camera (1/60 s) was not high enough to record the rapid motion of filaments, especially in activated state (b).Similar results have been obtained by using pure F-actin instead of the F-actin-TM-TN complex and we have preliminary evidence that S-l has the same effects as HMM on the motion of F-actin and the F-actin-TM-TN complex. The two-headed structure of HMM is not absolutely necessary to produce faster and larger bending motion of F-actin and the complex. The effect of cross-linking of actin subunits by p-phenylene N,N'-bis(maleimide) (PM) on the motion of F-actin was also examined. Previously, by applying this reagent, Knight and Offer showed that the cross-Unking did not change the ability of F-actin to activate myosin ATPase and induce superprecipita-tion, and suggested that relative motion of actin subunits was not required for the activation14. According to our direct observation, however, the cross-linking by this reagent does not restrict the motion of F-actin and the cross-linked F-actin has the same value of the flexural rigidity as before cross-linking (Table 1).The present finding suggests that in an activated state, myosin heads induce local conformational changes in F-actin such as distortion or rotation of actin monomers and loosening of actin-actin bonds8'15, which result in fast and large bending in F-actin filaments. Since F-actin has a structural polarity, the motion in F-actin could be asymmetric, if it occurred during interaction with HMM or S-l in the presence of Mg2+-ATP. However, the interaction between single F-actin filaments and soluble myosin fragments did not produce directional movement. When F-actin filaments function in living cells, they are usually found as assemblies with same polarity and often supported by some backbone such as membrane. Therefore, higher organizations of F-actin such as polar bundles may be indispensable to transformation of the activated motion in F-actin into an effective sliding force. We thank Dr Th. Wieland for supplying us with phalloidin-rhodamine, Drs K. Hotani, O. Ohara and E. Prchniewicz-Nakayama for many useful discussions and Dr T. J. Herbert for reading the manuscript.