J. Mol. Bid. (1987) 195, 351-358

Bundling of Myosin Subfragment-l-decorated Actin Filaments

Toshio Andot

Cardiovascular Research Institute University of California

San Francisco, CA 94143, IJ.S.A.

(Received 29 July 1986, and in revised form 2 January 1987)

We have reported previously that rabbit skeletal myosin subfragment-l (S-l) assembles actin filaments into bundles. The rate of this reaction can be estimated roughly from the initial rate (V,) of the accompanying turbidity increase (“super-opalescence”) of the acto-S-l solution. V, is a function of the molar ratio (r) of S-l to actin, with a peak at r = l/S to l/7 and minimum around r = 1.

In the present paper we report a different type of opalescence (we call it “hyper- opalescence”) of a&o-S-l solutions, which also resulted from bundle formation. Adjacent filaments in the bundles had a distance of approximately 180 8. Hyper-opalescence occurred at r x 1 when KCOOCH, was used instead of KCl. By comparing the effects of ADP, .+ADP, tropomyosin or ionic strength upon the super- and hyper-opalescence, we concluded that the two types of S-l-induced actin bundling had different molecular mechanisms. The hyper-opalescence type of bundling seemed to be induced by S-l, which was not complexed with actin in the manner of conventional rigor binding. The presence of the regulatory light chain did not affect hyper-opalescence (or super-opalescence), since there were no significant differences between papain S-l and chymotryptic S-l with respect to these phenomena.

1. Introduction

Actin is ubiquitous in eukaryotic cells. In non- muscle cells, actin exists in a variety of forms (i.e. monomer, filament, bundle, and mesh-work struc- ture) and can change its form when required. Assembly of actin filaments into higher-order structures is assisted by a variety of actin-binding proteins such as fascin, fimbrin, villin, filamin, and a protein identified in lung macrophages (for reviews, see Weeds, 1982; Korn, 1982; Sheterline, 1983). Heavy meromyosin prepared from skeletal muscle myosin cross-links actin filaments in vitro by binding each S-l portion to a different actin filament to form a raft-like structure (Trinick & Offer, 1979). It has been found by Ando & Scales (1985) that even the single-headed subfragment of rabbit skeletal myosin, S-1$, can assemble actin filaments into bundles whose interfilament distance

t Present. address: Department of Physics, Faculty of Science, Kanazawa University, 1-l Marunouchi, Kanazawa, Ishikawa 920 Japan.

$ Abbreviations used: myosin S-l, subfragment- 1; TM. tropomyosin; TES, N-tris[hydroxymethyl]methyl- 2-aminoethane sulfonic acid; E-ADP, 1 ,N6- ethenoadenosine 5’-diphosphate.

is about 90 A. Initially, they observed that, when an appropriate amount of S-l was added to F-actin solutions, an instant enhancement of light- scattering (due to S-l binding to F-actin) was followed by a gradual increase of the scattering with time. This gradual increase was called “super- opalescence”. The initial rate of super-opalescence was not proportional to the amount of S-l added, but had a peak at an S-l to actin molar ratio (r) of l/S to l/7. The rate was nearly zero at r values above 0.5. In the course of the experiments, it’ was noticed that the rate of super-opalescence was enhanced drastically, simply by replacing KC1 with KCOOCH3. Not only was the magnitude changed, but so also was the behavior of the rate as a function of r. Once r approached 1, the light-scattering intensity of the acto-S-l solutions increased abruptly. Since this behavior is so peculiar, we can expect that some unknown interactions between the proteins must be involved. In order to investigate such interactions, we have characterized acto-S-l solutions that contain acetate as a predominant anion, by studying the effects of nucleotides, tropomyosin (TM), and ionic strength on the light- scattering behavior, and by observing acto-S-l with an electron microscope.

351 0 1987 Academic Press Inc. (London) Ltd. OO22-2836/87/100351-08 $03.00/O

352 T. Ando

I 20 40 60 80 I

KCI, KEr. KNO, hd

Figure 1. Effects of KCl, KBr and KNO, upon the initial rate (V,) of super-opalescence. F-actin (3 PM) and chymotryptic S-l (0.45pM) were mixed in a solution containing 20 mM-TES (pH 7.0), 1 ITIM-UU3gneSiUm

acetate, 0.2 m&r-calcium acetate, 0.1 mxl-dithiothreitol, 0.1 mM-KaKs and various amounts of one of the neutral salts to be tested (0, KCl; A, KBr; 0, KNO,) and potassium acetate at 20°C. The sum of potassium acetate concentration and the neutral salts concentration was kept at 0.1 M.

2. Materials and Methods

(a) Protein preparations

Myosin was prepared from rabbit skeletal back and leg muscles (Tonomura et al., 1966) and stored in 0.3 M-KCl,

50% (v/v) glycerol at -20°C. Papain S-l was obtained by digesting myosin with insoluble papain in a solution containing 0.1 M-ammonium acetate (pH 7.2), 2 mM- MgCl,, and 0.2 miw-dithiothreitol, and purified by the use of DEAE-Sepharose CL-6B. The purified S-l was concentrated using Sephadex G-25 powder. Chymotryptic S-l was obtained by the method of Weeds & Taylor (1975). S-l(A1) and Sl(A2) were separated using preparative DEAE-cellulose high-pressure liquid chromatography (h.p.1.c.). The eluted S-l was concen- trated by 60% saturation of ammonium sulfate. After removal of the ammonium sulfate, the concentrated S-l was lyophilized in 0.15 M-SUCrOSe, 50 tIIMn-a~IUOniUIn

acetate and 05 mw-dithiothreitol at pH 7.0. The lyophi- lized S-l was stored at -20°C. Just before use, the lyophilized S-l was dissolved with a small volume of the desired buffer solution, and clarified by centrifuging for 90 min at 47,000 revs/min using a 50 rotor. The h.p.1.c. and lyophilization steps were omitted in some cases. However, this omission did not significantly affect the results given below. Acetone powder of rabbit skeletal muscle was prepared as described (Ando & Asai, 1979). Actin was first purified by the method of Spudich & Watt (1971), and further purified by Sephacryl S-300 gel chromatography. A high concentration (15 mg/ml) of F-actin was stored in 0.1 mmATP, 20 mM-KCOOCH,, 20 miw-TES (pH 7.0). 2 mM-Mg(COOCH,),, 0.2 mM- Ca(COOCH,),, 0.2 m&r-dithiothreitol, 0.1 mmNaru’, at 6°C. Just before use, the stored F-actin was diluted to approx. 20 pM with an ATP-free buffer solution. TM and troponin were prepared from rabbit skeletal muscle

according to the methods of Bailey (1948) and Ebashi et al. (1971), respectively. The molar concentrations of papain S-l, chymotryptic S-l and F-actin were deter- mined by their ultraviolet absorption in 0.6 M-NaCl using E&& = 8.3, Ei& = 7.5 and Ei& = 65, respectively. and respective molecular weights of 1.33 x 105, 1.15 x lo5 and 4.2 x 104. Corrections for turbidity were made as described (Ando, 1984). The molar concentrations of TM and troponin were determined by the method of Lowry et al. (1951), taking values of 7.0 x lo4 and 7.5 x IO4 for their respective molecular weights.

(b) Light-scattering

As described by Ando & Scales (1985), the light- scattering intensity of acto-S-l solutions was measured at 400 nm using a Hitachi Perkin-Elmer MPF4 fluoro- meter. Since the units of light-scattering intensity are arbitrary, the intensity was normalized by assigning a value of 10 to the intensity of 5 p&r-F-actin in a solution containing 0.1 M-KCl, 20 miw-TES (pH 7.0), 2 mM- Mg(COOCH,),, 0.2 miw-Ca(CGGCH,),, 0.2 mm-dithio- threitol and 0.1 mM-NaN, at 15°C. A small volume (5 to 160 ~1) of S-l solution was added to 2 ml of F-actin solution in a thermostatted fluorescence cuvette. The solution was mixed by sucking and blowing several times with a l-ml Pipetman automatic pipet, and the time- course of changes in light-scattering intensity was immediately recorded on time-scanning chart paper. The initial rate (V,) of super-opalescence or hyper-opalescence was read from the chart paper.

(c) Electron microscopy

Droplets of solutions of 1 mw-F-actin plus 1.5 PM-S-~ were applied to freshly carbon-coated Formvar grids. After 30 s, a few drops of 1 ye (w/v) uranyl acetate were applied to the grid for another 30 s. The negative stain was then removed with torn filter paper. The grids were allowed to dry for several minutes before they were observed at 80 kV in a Philips EM 200 electron microscope.

3. Results

(a) Further characterization of super-opalescence

When super-opalescence of acto-S-l solutions was first studied (Ando & Scales, 1985), KC1 was used for setting an appropriate ionic strength. Here, how- ever, Cl- was found to decelerate super-opalescence (see Fig. 1). F-actin (3 ,UM) and chymotryptic S-l (0.45 PM) (i.e. r = 0.15) were used in this experi- ment. The sum of KCi and KCOOCH, concentra- tions was kept at 0.1 M. V, with 0.1 M-KCOOCH, was about 15-fold higher than that with 0.1 M-KCl. Therefore, unless otherwise mentioned, KCOOCH3 was used instead of KCl. The inhibitory effect of other anions (Br- and NO;) was also tested. As shown in Figure 1, these anions inhibited super- opalescence with the order of increasing effective- ness being Cl- < Br- < NO;.

V, decreased when ADP had been preincubated in F-a&in solutions before adding chymotryptic S-l. V, decreased to zero with increasing ADP concentration (Fig. 2). V,, decreased by half at approximately 20 PM-ADP. This value is not very

Acto-S-l Bundling 353

F

E

T .E .E 2 4

i

C

r- 0

14

0 j-

\ 0

\ 0

) \

\ ‘\, 0

‘0

\ \@,

,O\” I I I I I 0.2 0.4 0.6 C

ADP, E -ADP (mtd

Figure 2. Effect of preincubation of F-actin with ADP or E-ADP on the initial rate of super-opalescence. Chymotryptic S-l (0.45 PM) was added to 3 pni-F-actin preincubated in various amounts of ADP/e-ADP, 60 mM-potassium acetate, 20 mM-TES (PH 74L 2 mM-magnesium acetate, 0.2 mM-calcium acetate, 0.1 mw-dithiothreitol, and 0.1 mM-NaNu’,, at 20°C. 0. ADP effect; 0, E-ADP effect.

different from the reported values of the dissocia- tion constant of Mg 2 + -ADP for rabbit skeletal acto- S-l: 37 PM (Highsmith, 1976), 143~~ (Greene & Eisenberg, 1980). When E-ADP was used instead of ADP, V, decreased by half at approximately 270 PM-E-ADP. This value is also not far from the dissociation constant of Mg’+-&-ADP (350 /AM) for rabbit skeletal acto-myosin in fibers reported by Yanagida (1981). Therefore, the acto-S-l-bound nucleotides seem to be responsible for the reduction of V,. This conclusion is plausible, considering that V, is not, a linear function of the amount of nucleotide-free acto-S-l complex.

On the other hand, when ADP was added to a&o-S-l solutions that had been preincubated and were already showing super-opalescence, the light- scattering intensity stayed at a level slightly lower than just before ADP was added (Fig. 3). Since mechanical perturbation alone of acto-S-l solutions by mixing with a Pipetman reduced the super- opalescence level, the slight reduction observed here was not due to ADP addition. Therefore, we can conclude that ADP addition to actin bundles cannot disassemble the bundles, although ADP inhibits bundle formation.

TM had similar effects on the super-opalescence of acto-S-l solutions. TM addition to actin bundles cannot disassemble the bundles whereas TM inhibits bundle formation. A molar ratio of TM/a&in of as little as approximately 0.018 was sufficient to reduce V, by half (Fig. 4). Troponin had little influence on the TM effects on super-

I I I I I I I 0 2 4 6 8 IO I2 14

Time (mini

Figure 3. Effect on the super-opalescence of ADP addition to the acto-S-l solution that has shown super- opalescence. At arrow 1, 0.45 PM-chymotryptir S-l was added to 3 PM-F-actin in a solution containing 0.1 M-potassium acetate, 20 mM-TES (pH 7.0), 1 mM- magnesium acetate, 0.2 mM-calcium acetate, 0.1 mM- dithiothreitol, 0.1 mM-NaN,, at 20°C. At arrow 2. 0.2 mM-ADP was added.

opalescence, irrespective of the presence of Ca’+. To study the effect of the regulatory light chain

on super-opalescence, the previous experiments were repeated using papain-S-1 instead of chymo- tryptic S-l. It gave essentially the same results as those obtained with chymotryptic S-I.

(b) Hyper-opalescence

We have reported (Ando & Scales, 1985) that V, reaches a peak at r = l/S to l/7, and then steadily decreases to nearly zero when r is further increased (see Figs 2 and 3 of Ando & Scales, 1985). Though V, data at r above 0.7 were not shown in that paper, V, was zero at this range of r values (Fig. 3 of Ando & Scales (1985) implies this fact). Now, V, versus r was re-studied in solutions that did not contain Cl-, but contained CH,COO- instead. As shown in Figure 5 (chymotryptic S-l was used), V, reached a peak at r = l/6 to l/7 and then decreased to zero when r was increased (the first phase). However, as r was further increased, V, st#arted increasing very rapidly just’ before r reached the value 1.0, and reached a plateau at r z 2.5 (the second phase). The plateau value of V,, at r z 2.5 was about 14 times larger than the peak value of the first phase. The contrast between the present and previous observations of V, behavior around r = 1.0 resulted only from the difference of anion (i.e. Cl- or CH,COO-). Henceforth, in order to

354 T. Ando

I I I I I I I 0 0.04 0.08 0.12 0.16

TM Q.LM)

Figure 4. Effect of preincubation of F-actin with TM on the initial rate of super-opalescence. Chymotryptic S-l (0.45 PM) was added to 3 PM-F-a&in plus various amounts of TM in a solution containing 60 mm-potassium acetate, 20 mM-TES (pH 7.0), 2 m&r-magnesium acetate, 0.2 mm-calcium acetate, 0.1 m&r-dithiothreitol, O-1 mM- NaN,, at 20°C.

distinguish the second phase from the first phase, we call the second one “hyper-opalescence” and we use “super-opalescence” for the first one.

(c) Bundle formation

Does hyper-opalescence also result from forma- tion of bundles by S-l-decorated F-actin? In order to study this question acto-S-l was observed with an electron microscope. An acto-S-l sample (1.5 PM-

chymotryptic S-l plus 1 pi%-F-actin) was incubated at room temperature for one hour in a solution containing 0.1 M-KCOOCH,, 20 mM-TES (pH 7.0), 1 miw-Mg(COOCH,),, 0.2 miw-Ca(COOCH,),, 0.2 m&r-dithiothreitol, O-1 mM-NaN,, and then mounted on grids. As shown in Figure 6(a), well- ordered bundles containing a few S-l-decorated actin filaments were observed. The bundles possessed distinct transverse stripes with a periodi- city of about 350 A and an interfilament distance of about 180 A. It was difficult to see any arrowhead polarity of an S-l-decorated filament within the bundles. Fortunately, we had a few specimens of bundles in which two adjacent filaments branched at an end, so that we could clearly see the polarity (Fig. 6(b), (c), (d)). The arrowheads of both branched filaments pointed in the same direction (i.e. to the junction ((b), (c)), or to the splay direction (d)). That is, actin filaments within these particular specimens had parallel polarity.

(d) Further characterization of hyper-opalescence

How high a concentration of Cl- would be required in order to suppress hyper-opalescence

Figure 5. The initial rate of super- and hyper- opalescence of acto-S-l solutions as a function of the molar ratio of S-l to actin. Various amounts of chymotryptic S-l were added to 2 PM-F-a&in in a solution containing 60 mM-potassium acetate, 20 mM- TES, (pH 7.0), 2 mlcr-magnesium acetate, O-2 mm-calcium acetate, 0.1 mM-dithiothreitol, 0.1 mM-NaN,, at 20°C.

completely? As shown in Figure 7, V, of hyper- opalescence was sharply reduced by Cl-, while V. of super-opalescence was gradually reduced. About 26 mM of Cl- was sufficient for extinguishing hyper- opalescence. Other anions (Br- and NO,) were also tested. These anions inhibited hyper-opalescence with the order of increasing effectiveness of Cl- < Br- < NO;.

Why did hyper-opalescence occur around r = l.O? There seem to be two possible interpretations of this observation. One is that actin filaments may acquire an ability to form bundles once they are almost completely decorated with S-l. Another is that free S-l may possess an ability to assemble S-l-decorated actin filaments, for free S-l concen- tration starts increasing sharply around r = 1.0. Here, “free S-l” is not meant literally. A first interpretation might be that the agent responsible for the actin bundling is S-l that has formed a rigor complex with actin (there cannot be a direct interaction between actin molecules of neighboring filaments). I f ADP or .s-ADP inhibit hyper- opalescence as well, it may be easy to test which of the interpretations is right. Ranges of the ADP/e-ADP concentrations required for inhibition should differ between the two cases, since the

Acto-S-l Bundling 355

Figure 6. (a) to (d) Electron micrographs of hyper-opalescence type of acto-S-l bundles: 1.5 PM-chymotryptic S-l plus 1 PM-F-a&in was incubated at room temperature for 1 h in a solution containing 0.1 m-potassium acetate, 20 mM-TES <pH 7-O), 1 mM-magnesium acetate, O-2 mr+calcium acetate, 0.2 mm-dithiothreitol and 0.1 m&x-NaN,, and then mounted on grids. The bars represent 0.1 pm. Note that in (b) and (c) the arrowheads on both branched filaments point to the ,junction, and in (d) they point to the splay direction.

affinities of ADP/&-ADP for acto-S-l are much lower than those for free S-l. As shown in Figure 8, Lb of hyper-opalescence was much more steeply reduced by preincubating F-actin with ADP/&-ADP, compared to that of super-opalescence (for comparison the data for V, versus ADP/&-ADP concentrations in Fig. 2 are also put on Fig. 8). V, decreased by half at approximately 2 PM-ADP. When E-ADP was used, V, decreased by half at approximately 15 PM-E-ADP. These numbers are close to the respective dissociation constants for S-l ( 1.6 pM for ADP, Konrad & Goody, 1982; 9 PM for E-ADP, Ando, unpublished results). It is very unlikely that the observed suppressive effect of ADP/&-ADP on V, was made by the quite tiny amount of act&-l -bound ADP/&-ADP. Therefore, the second interpretation seems correct. In contrast, ADP addition disassembled the hyper- opalescence type of bundles.

Ionic strength effects on V,, of hyper-opalescence

also seem to favor the second interpretation mentioned above. The curve of V,, of hyper- opalescence versus the total S-l concentration shifted to the left as ionic strength of the acto-S-l solutions was enhanced, with no significant changes in the plateau value of V, (Fig. 9). As is well known, when ionic strength is enhanced, S-l affinity for actin is reduced. This results in the curve of free S-l concentration versus the total S-l concentration shifting to the left. This fact would be consistent with the shift of V,, if we suppose that free S-l possesses an ability to assemble S-l -decorated actin filaments.

In contrast to the effect of TM upon super- opalescence, preincubation of F-actin with TM did not significantly affect the V, of hyper-opalescence (Fig. IO). V,, was reduced by 30% by at, most a stoichiometJric amount of TM. This fact suggests also that the molecular process of hyper-opalescence differs from that of super-opalescence.

356 T. Ando

1

KCI, KBr, KNO, hwl ADP, r-ADP QLLM)

Figwe 7. Effects of KCl, KBr and KN03 upon the initial rate of hyper-opalescence. F-actin (1 PM) and 2 PM- chymotryptic S-l were mxied in a solution containing 20 mM-TES (pH 7.0), 1 mm-magnesium acetate, 0.2 mM-

calcium acetate, 0.1 mnn-dithiothreitol, 0.1 mM-NaN,, various amounts of one of the neutral salts to be tested (0, KCI; A, KBr; 0, KNO,), potassium acetate, at 20°C. The sum of potassium acetate concentration and the neutral salts concentration was kept at O-1 M.

Finally, the ability of papain S-l to induce hyper- opalescence was examined. Papain S-l also induced hyper-opalescence in essentially the same manner as did chymotryptic S-l (data not shown).

4. Discussion

Tt is known that neutral salts are potent structural destabilizers (at high concentrations) and inhibit the activity of several kinds of enzymes (von Hippel & Wong, 1964; Warren et al,, 1966). The order of increasing effectiveness for monovalent anions is as follows: CH3COO- < Cl- < Br- < NO, < ClO, < I- < SCN-. This order is known as Hofmeister’s series or the lyotropic series. The order of the first four anions coincides with that of their ability to inhibit super- and hyper-opales- cence. Stafford (1985) has reported a disruptive effect of Cl- on the myosin tail. Carboxyl ions that exist as organic ions generally comprise 30 to 40% of the intracellular anions. The rest is predo- minantly composed of phosphate ions. Chloride ion exists only at 1.5 mM in frog sartorius muscle cells. (In spite of these facts, many biochemists have long been using KC1 rather than KCOOCH, to set the ionic strength.) Taking the foregoing facts into consideration, it is natural to assume that Cl- inhibited an intrinsic ability of S-l to assemble actin filaments, rather than that CH,COO- affects S-l, so that S-l acquires an ability to assemble actin filaments.

4-

2-

0

Figure 8. Effect of preincubation of F-actin with ADP or E-ADP on the initial rate of hyper-opalescence and super-opalescence. The data for super-opalescence (0, 0) are the same as those in Fig. 2. (A) Hyper-opalescence rate with ADP; (A) hyper-opalescence rate with E-ADP. Chymotryptic S-l (2 p(M) was added to 1 PM-F-actin preincubated in various amounts of ADP/&-ADP, 0.1 M- potassium acetate, 20 mM-TES (pH 7*0), 1 mM-mag- nesium acetate, O-2 mM-calcium acetate, 0.1 mM-dithio- threitol, 0.1 m&i-NaN,, at 20°C.

In actin bundles formed in the super-opalescence phase, an S-l molecule seems to cross-link two adjacent actin filaments. Indeed, electron micro- graphs of the actin bundles show an interfilament distance of 90 A (Ando & Scales, 1985), which is larger than the interfilament distance expected when actin filaments bind directly to each other side by side. When the S-l particle cross-links actin filaments, it must bind tightly to an actin filament at its so-called rigor binding site and must bind weakly to another filament at its secondary binding site. Although ADP binding to acto-S-l weakens the rigor binding, it does not weaken it sufficiently to dissociate S-l from actin. Therefore, the inhibitory effect of ADP on super-opalescence must act by weakening the secondary binding. On the other hand, ADP addition to the super-opalescence type of actin bundles could not disassemble the bundles. There may be two possible interpretations for this. ADP may not be able to bind to S-l that has cross-linked actin filaments. This idea implies that the ADP binding site and the secondary actin binding site may overlap on S-l. Alternatively, there may exist a “locked state” of the bundles, following an intermediate bundle state that can revert to single filaments. The locked state would not revert to the intermediate state even when it attaches to ADP, and ADP-attached single fila- ments cannot go forward to the intermediate bundle state.

Acto-S-l Bundling 357

r(S-i/actin)

Figure 9. Ionic strength effect on the relation of V, versus r. Various amounts of chymotryptic S-l were added to 1 PM-F-a&in in a solution containing various amounts of potassium acetate, 20 mM-TES (pH 7.0), 1 man-magnesium acetate, 0.2 mm-calcium acetate, 0.1 m&r-dithiothreitol, 0.1 mm-NaN,, at 20°C. Potassium acetate was at (0) 60 mM; (A), 0.1 M; (O), 0.2 M.

Although the maximum initial rate of hyper- opalescence was 14 times greater than that of super- opalescence, this does not mean that the corres- ponding rates of the two types of bundling processes differ by this much; a unit change in light- scattering intensity accompanied by dimerization of fully S-l-decorated filaments must be much larger than that accompanied by dimerization of partially S- 1 -decorated filaments.

It should be mentioned here that a few preparations of chymotryptic S-l and papain S-l did not show hyper-opalescence. Even such S-l complexes, however, always showed super-opales- cence and normal Ca’+- and EDTA-ATPase activities. This subtle change in S-l quality occurred mainly in the lyophilization process. Aging of S-l solutions also reduced the rate of hyper- opalescence. Since further addition of fresh dithio- threitol solution somewhat restored S-l that had not shown normal hyper-opalescence, oxidation of S-l seems to be responsible for the subtle change in S-l behavior.

Seymour & O’Brien (1985) have observed, with an electron microscope, well-ordered acto-S-l bundles formed in a crack in the carbon film. They have mentioned that adjacent filaments in the bundles had antiparallel polarity and spacing of around 150 A. It is not clear whether the bundles they observed are the same as those observed here, because they could not know the solution condi-

12

IO

8

_c1 Ic

$

4

2

0 I I I

0.04 0.08 0.14

TM (prd

Figure 10. Effect of TM on the initial rate of hyper- opalescence. Chymotryptic S-l (2 FM) was added to 1 pm-

F-actin plus various amounts of TM in a solution

containing 0.1 M potassium acetate, 20 mM-TES (pH 7.0), 1 mM-magnesium acetate, 0.2 m&r-calcium acetate, 0.1 mm-dithiothreitol, 0.1 mM-NaN,, at 20°C.

tions (e.g. molar ratio of S-l to actin, ionic condition) in the crack of the carbon film where the acto-S-l bundles were formed. Although we saw a parallel polarity of adjacent filaments in a few specimens of bundles whose filaments branched at an end, it may be too early to generalize this observation to all hyper-opalescence types of bundle. This issue could be resolved in future structural studies on the complex.

Hyper-opalescence-type acto-S-l bundles have an interfilament distance of 180 A. This distance is twice that of the super-opalescence-type bundles. Therefore, it is unlikely that an S-l molecule fills this interfilament gap by itself. That is, the S-l molecule is not in contact with both of the neighboring actin filaments.

Hyper-opalescence scarcely occurred when r was less than unity. It occurred abruptly when S-l concentration exceeded the value of the actin concentration. Furthermore, the inhibitory effect of ADP on hyper-opalescence appeared in a range where the ADP concentration was much lower than the dissociation constant of ADP for acto-S-l. That is, in spite of the small amount of ADP bound to acto-S-l (at most only 2 to 3 o/o of actinbound S- 1 contains ADP), formation of the hyper-opalescence type of bundles was inhibited by ADP addition. Therefore, ADP acted on S-l that was not involved in rigor complexes with actin (we usually call such S-l “free S-l”), which resulted in inhibition of bundle formation. Therefore, data of the inhibitory effect of ADP suggested that “free S-l ” induced the hyper-opalescence type of acto-S- 1 bundling. This

358 T. Ando

idea can give a clear answer to the question of why hyper-opalescence occurred abruptly when S-l concentration exceeded actin concentration. The “free S-l” concentration is small before the S-l concentration reaches the same value as the actin concentration, but sharply increases after that. This idea is also consistent with the data on ionic strength effects on hyper-opalescence behavior as a function of S-l concentration. “Free S-l” can be obtained to some extent even before S-l concentra- tion reaches the level of the actin concentration; this is achieved by reducing S-l rigor affinity for actin by increasing the ionic strength of the acto- S-l solutions. Therefore, in higher ionic strength solutions, hyper-opalescence appeared before the S-l concentration reached the level of the actin concentration. Because these three facts concerning hyper-opalescence are consistent with each other, it is very probable that “free S-l” somehow bundles S- 1 -decorated actin filaments.

The next question will be how “free S-l ” bundles acto-S-l filaments. It is uncertain at present whether the “free S-l” is incorporated into the final structure of the acto-S-l bundles. It might be possible that the “free S-l” is incorporated into a transient filament-filament complex, from which it is eventually expelled. Whether this is the case or not, it is evident that the “free S-l” interacts with actin-bound S-l. Studies by Morel’s group (Morel & Garrigos, 1982; Bachouchi et al., 1985) have shown that S-l molecules can dimerize through end-to-end contact. According to these authors the dimer is predominant at low ionic strength, but, when the ionic strength is raised using KC1 (approx. 0.1 M),

the dimer is disassembled. It would be interesting to see whether such S-l dimers could be retained when KCOOCH, is used instead of KCl, because hyper-opalescence occurred with KCOOCH, but not with KCl.

I thank MS Mei Lee Wong for excellent technical assistance in taking electron micrographs. This work was supported by grants NSF PCM75-22698 and NHBLI- 16683.

References

Ando, T. (1984). Biochemistry, 23, 375-381. Ando, T. & Asai, H. (1979). J. Mol. Biol. 129, 265-277. Ando, T.&Scales, D. (1985). J. Biol. Chem. 260,2321-2327. Bachouchi, N., Gilik, A., Garrigos, M. & Morel. J. E.

(1985). Biochemistry, 24, 6305-6310. Bailey, K. (1948). Biochem. J. 43, 271-279. Ebashi, S., Wakabayashi, T. & Ebashi, F. (1971).

J. Biochem. 69, 441-445. Greene, L. E. & Eisenberg, E. (1980). J. Biol. Chem. 255,

5433548. Highsmith, S. (1976). J. BioE. Chem. 251, 6170-6172. Konrad, M. & Goody. R. S. (1982). Eur. J. Biochem. 128,

547-555. Korn. E. D. (1982). Physiol. Rev. 62, 672-737. Lowry, 0. H., Rosenbrough, N. J., Farr, A. L. & Randall,

R. J. (1951). J. Biol. Chem. 193, 265-275. Morel, J. E. & Garrigos, M. (1982). Biochemistry, 21,

2979-2986. Seymour, J. & O’Brien, E. J. (1985). J. Muscle Res. Cell

Motil. 6, 725-756. Sheterline, P. (1983). In Mechanism of Cell MotiZity:

Molecular Aspects of Contractility, pp. 104-196, Academic Press, New York.

Spudich, J. A. & Watt, S. (1971). J. BioZ. Chem. 246, 4866-487 1.

Stafford, W. F. (1985). Biochemistry, 24, 3314-3321. Tonomura, Y., Appel, P. & Morales, M. F. (1966).

Biochemistry, 5, 515-521. Trinick, J. & Offer, G. (1979). J. Mol. Biol. 133, 549-556. von Hippel, P. H. & Wong, K. Y. (1964). Science, 145,

577-580. Warren, J. C., Stowring, L. & Morales, M. F. (1966).

J. Biol. Chem. 241, 309-316. Weeds, A. G. (1982). Nature (London), 296, 811-816. Weeds, A. G. & Taylor, R. S. (1975). Nature (London),

257, 5456. Yanagida, T. (1981). J. Mol. Biol. 146, 539-560.

Edited by H. E. Huxley