Journal of Muscle Research and Cell Motility 13, 523-533 (1992)

Two different acto-S1 complexes O . A. A N D R E E V

and J. B O R E J D O *

Baylor Research Institute, Baylor University Medical Center, 3812 Elm Street, Dallas, TX 75226, USA Received 11 November 1991; revised 4 February 1992; accepted 3 March 1992

Summary Based on change in anisotropy of fluorescently labelled $1 and on increase in turbidity of acto-SI complex when $I bound to F-actin, we reported previously that depending on the molar ratio of $I to actin two different complexes of actin monomer (A) and myosin subffagment I ($1) could be formed: AI*S1 (one actin with one $I) and A2"$1 (two actins with one $1). Here we extend these findings to F-actin labelled with pyrene and cross-linked to $1 with I-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC). The fluorescence of pyrene F-actin decreased with increase in $1 concentration and reached saturation at a molar ratio of $1 to actin of either 0.5 or 1.0, depending on whether $1 was added slowly (5 min) or quickly (10-20 s between additions). Incubation of A2"$I complex in excess of $1 for > I h caused a shift in equilibrium towards the AI*S1 complex. The Az*S1 complexes were not formed at high $I to actin ratios (> 1.0) owing to competition between heads. Crosslinking experiments showed that the formation of EDC crosslinked products, 175-185 kDa doublet and 265 kDa band, depended on the ratio $I to actin. To assess the relative ratio of $I and actin in crosslinked products, we labelled $1 and F-actin with different fluorescent probes (5-IAF and IATR). The $1 to actin ratio was proportional to the ratio of intensities of fluorescence of labelled $1 and actin. The $1 to actin ratio in 265 kDa product was two times smaller than in 175-185 kDa doublet (which is believed to be A~*S1 complex) and therefore 265 kDa band corresponded to A2"$1. Transition between two types of binding may be important to understanding how muscle contracts.

Introduction Force is thought to be generated in muscle by cycling interaction of myosin head ($1) with F-actin. It is believed that $1 carrying a nucleotide first binds weakly to thin filament (Schoenberg, 1988) and that force is generated when weakly-bound $1 undergoes orientational change (Botts et at., 1989) triggered by a release of inorganic phosphate (Hibberd et al., 1985), to form a strong (rigor) complex with F-actin (for review see Cooke, 1986). This rigor bond is thought to play a fundamental role in the generation of force, and its properties have been extensively studied. Turbidimetric measurements (Finlayson et al., 1969; White & Taylor, 1976) provided evidence that stoichiometry of binding of myosin head with F-actin was 1:1. This has been confirmed by ultracentrifugation experiments (Eisenberg et al., 1972; Margossian & Lowey, 1973, 1978; Greene & Eisenberg, 1980) and by recent light scattering measurements (Lehrer & Ishii, 1988; Ishii & Lehrer, 1990). Titrations of native myofibrils with fluorescent $1 also gave 1:1 stoichiometry (Borejdo & Assulin, 1980). Moreover, the stoichiometry of binding can be estimated by crosslinking actin with $1 in rigor *To whom correspondence should be addressed. 0142-4319 9

1992 Chapman & Hall

with 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC). A major product of such crosslinking is a double band with an apparent molecular weight of 175-185 kDa (Morner et al., 1981). A number of investigators showed (by double fluorescent or radioactive labelling) that actin and $1 exist in this doublet in equimolar concentrations (Sutoh, 1983; Greene, 1984; Heaphy & Tregear, 1984; Chen et al., 1985). However, quite a different suggestion was made by X-ray diffraction data of muscle decorated with $1. These experiments suggested that $1 molecules interacted with two actin monomers (Amos et al., 1982). This conclusion was strengthened by a recently proposed model of the thin filaments (Holmes et al., 1990) and of thin filaments decorated with $1 (Milligan et al., 1990). Even though these models are consistent with $1 to actin binding stoichiometry of 1:1, they point out the possibility that each $1 may bind to two actin monomers. Furthermore, using the fact that anisotropy of fluorescently labelled $1 increases when it binds to F-actin we recently showed that depending on the relative amounts of actin and myosin present during titrations, $1 could bind either to one or to two actin monomers (Andreev & Borejdo, 1991a). Similar results were obtained with the aid of turbidimetric titrations (Andreev & Borejdo, 1991b). In

524 the present paper we use the fact that binding of $1 quenches fluorescence of F-actin labelled with pyrene to show that, if titration is made slowly, quenching saturates at the molar ratio of $I to actin of 0.5. We present evidence that double-headed heavy meromyosin (HMM) molecule can bind two or four actins, and show that the time course of formatioh of major crosslinked products is very different when actin is crosslinked with excess and substoichiometric amounts of $1. When actin is in excess over SI, apart from doublet band with an apparent molecular weight of 175-185 kDa (Mornet et al., 1981), a prominent crosslinking product migrating in SDS-PAGE at 265 kDa is formed. The ratio of actin to $ 1 in this product is double of that of a doublet. These results support our earlier claim that myosin can bind to one or to two actin monomers in a filament.

Materials and methods

Materials and solutions ATP, Silver Stain, phalloidin and chyrnotrypsin were from Sigma (St. Louis, MO). Low molecular weight markers, Sephadex G-25 and G-50 gels were purchased from Pharmacia (Pharmacia, Piscataway, N J). Iodoacetamido-tetramethylrhodamine (IATR), 5-iodoacetamido-fluorescein (5-IAF), 5-[2((iodoacetyl)-amino)-ethyl]-aminonaphtalene- 1-sulphonic acid (1,5-IAEDANS) and N-(1-Pyrenyl)-iodoacetamide (pyrene) were from Molecular Probes (Eugene, OR). F-buffer contained: 50 mM KC1, 2 mM MgSO4, 10 mM Tris HC1, pH 7.5, 0.2 mM ATP, 1.0 mM DTT. G buffer contained 2 mM Tris acetate, pH 8.0, 0.2 mM ATP, 1.0 mM DTT. Buffer A contained 50 mM KC1, 10 mM Tris HC1 buffer, pH 7.5.

Proteins Myosin was prepared from rabbit skeletal muscle by the method of Tonomura and colleagues (1966). HMM and $1 were obtained by chymotryptic digestion of myosin according to Weeds and Pope (1977) and Weeds and Taylor (1975), respectively. Actin was prepared according to Spudich and Watt (1971). For light scattering experiments F-actin was dialysed against a buffer containing 50 mM KC1, 2 mM MgSO4, 10 mM Tris HCI, pH 7.5 (to remove free ATP). The concentration of proteins and dyes were determined by absorbance using the following extinction coefficients: $1, A~~176 7.5; HMM; A~~176 = 6.47; actin, AI~176 = 6.3; F-actin, A1~/~ 5-IAF, 42000M-~cm -~ at 493nm; IATR, 62 000 M-~cm-~ at 555 nm; and pyrene, 22 000M-~cm -~ at 344 nm. The absorbances of labelled proteins were corrected by subtracting the absorbance of bound dye at 280 nm (at 290 nm for actin).

Labelling of H M M and S I Proteins were incubated with 5-IAF, IATR and 1,5-IAEDANS at 1.5-fold excess of the dye for 6 h in ice. To remove unbound or non-covalently bound dye, proteins were centrifuged at 500 r.p.m, for 2 rain in a column made by filling I ml syringe with Sephadex G-50. The proteins were then dialysed overnight against buffer A.

ANDREEV and BOREJDO

Labelling F-actin with 5-IAF and IA TR The 5-IAF, IATR and 1,5-IAEDANS were dissolved in dimethylformamide (DMF) at a stock concentration of 10 mM. F-actin was incubated with 1.5 M excess of the dye for 12 h at 0~ in the dark. After labelling, actin was precipitated, resuspended in G buffer, dialysed for 48 h to depolymerize it and passed through Sephadex G-50 column to remove unbound dye. It was then polymerized with 50 mM KCI, 2 mM MgSO4 for 2 h at room temperature (RT). The typical degree of labelling was 60-90%.

Labelling F-actin with pyrene The labelling of F-actin by pyrene and measurements of concentration of actin and of bound dye were made as described by Cooper and colleagues (1983). Pyrene was dissolved in DMF at a stock concentration of 10 mg rnl -I. F-actin at I m g m1-1 was incubated with 7.5 M excess of pyrene for 16 h at RT in the dark. After labelling, actin was precipitated, resuspended in G buffer, dialysed for 48 h to depolyrnerize it and passed through Sephadex G-50 column equilibrated by the same buffer except ATP to remove unbound pyrene and free nucleotides. It was then polymerized with 50 mM KC1, 2 mM MgSO 4 for 2 h at RT. To stabilize actin filament a phalloidin was added at 1:1 molar ratio to actin. Typical degree of labelling was 70--95%.

Light scattering and measurements of pyrene fluorescence All measurements were done in SLM 500C (Urbana, IL) spectrofluorometer. Light scattering intensity was measured at right "angle to the incident beam. Excitation and emission wavelengths were set at 490 nm and slits at 5 nm. Pyrene fluorescence was excited at 368 nm with 2.5 nm slit and detected at 409 nm with 5 nm slit.

Crosslinking of the actin-S I complex F-actin and $1 were mixed at different molar ratios and pre-incubated for 30 rain at 22~ Then 50 or 100 mM EDC was added and reaction was stopped by adding of equal volume of a solution containing 4% SDS, 10% fl-mercaptoethanol, 20% sucrose, 60 mM Tris HC1, pH 6.8. Reaction time was varied from 5 rain to 40 rain. The same procedure was used to crosslink fluorescent-labelled F-actin and $1 except that 50 mM EDC was added for 30 rain.

SDS-PAGE PAGE was run according to Laemmli (I970) in a buffer containing 25 mM Tris, 192 mM glycine, 0.1% SDS, in 12% polyacrylamide gels. Staining solution contained 50% methanol, 7.5% acetic acid and 0.25% Coomassie Blue. Fixing and destaining solutions contained 50% ethanol and 7.5% acetic acid. Gels were imaged by a video camera (Javelin Model JE2362A, Javelin Electronics, Torrance, CA) connected to a flame grabber (MetraByte Corp Model MV1, Taunton, MA) operated by an AT type computer. The intensity of various bands was measured from the captured image by Java image analysis program (Java 1.4, Jandel Scientific, Corte Madera, CA) and plotted on a Linotronix typesetter using PageMaker program (Seattle, WA).

525

Two different acto-S1 complexes

Determination of relative amounts of actin and S1 in crosslinked complex F-actin labelled by IATR, or 5-IAF, was crosslinked with $I labelled by 5-IAF, or IATR, respectively. The crosslinked sample was resolved in 7.5% SDS-PAGE in a glass tube (model 150A, BioRad, Richmond, CA). After electrophoresis the fluorescent bands were marked under illumination with UV lamp. The fluorescence spectra were recorded directly from gel in a glass tube mounted in SLM 500C spectrofluorimeter. The glass tube was oriented 50 degrees relative to the incident beam. Front-face illumination was used to avoid a concentration quenching of fluorescence. The fluorescence of 5-IAF and IATR were exited at wavelengths 480 nm and 530 nm, respectively, with 2 nm excitation and 10 nm emission slit. The Coming 3-66 and 3-70 filters were used to eliminate scattered light at 530 nm and 480 nm, respectively. Only IATR fluorescence is excited at wavelength ~,ex= 530 nm. At ),e, = 480 nm the contribution of IATR to fluorescence was negligible because spectra were similar to 5-IAF in solution. It was shown in the control experiments with mixtures of 5-IAF and IATR in solution in a square cuvette (front-face illumination, 2~x= 480 nm) that a second peak appeared in the fluorescence spectrum when the molar ratio of IATR to 5-IAF was > 15. Another control experiment was done to show that the molar ratio of the 5-IAFand IATR-labelled proteins in the electrophoretic band was proportional to the intensity of fluorescence of 5-IAF and IATR emanating from this band. $1 was labelled with 5-IAF or IATR and the two labelled species were mixed at different molar ratios. Samples containing different molar ratios were electrophoresed in I0% polyacrylamide in glass tubes. The fluorescence spectra of the band corresponding to the heavy chain of $1 were recorded through 5-IAF and IATR filters as described above. The ratio of fluorescence intensities emanating from 5-IAF- and IATR-labelled $I was measured and plotted against the molar ratio of the two Sis (see Results). The weight of proteins in each crosslinked product band was 0.5 h (solid line), even when concentration of actin was 5 l.tM. In agreement with Ando and Scales (1985), when

Time

Fig. 3. Increase in turbidity of acto-S1 complex when $1 is added quickly to pyrene-F-actin. Left panel: $1 was added to the cuvette containing 1 IXM of pyrene-F-actin. Arrows show when $1 was added (stock solution of I3 IXM)to make the following molar ratios of $1 to actin (from left to right): 0.25, 0.5, 0.75, 1.0, 1.5, 2.0. Centre panel: showing that turbidity of the complex did not increase in over 2 h. Right panel: addition of I mM ATP (at arrow). Solution 50 mM KC1, 10 mM Tris HC1 buffer, pH 7.5, at RT. The vertical scale is the same as in Fig. 1.

Two different acto-S1 complexes

527 '0d~,ec

10 min

50 sec L.--J

1.0

30 min 30 m p ~ . . ~ P " "

~+ ATP

30 rain p.... "g_ 3 min _....,..._~p..,-.._

l--

Fig. 4. Increase in turbidity of acto-S1 complex when $1 is added slowly to pyrene-F-actin. Left panel: $1 was added to the cuvette containing I ~tM of pyrene--F-actin. Arrows show when $1 was added (stock solution of 13 IXM) to make the following molar ratios of $1 to actin (from left to right): 0.25, 0.5, 0.75, 1.0, 1.5, 2.0. Centre panel: showing that turbidity of the complex did increase in over 1.5 h. Right panel: addition of I mM ATP (at arrow). Solution 50 mM KCL 10 mM Tris HC1 buffer, pH 7.5, at RT. The vertical scale is the same as in Fig. 1.

Fig. 6. Crosslinking of SI to F-actin using different molar ratios of actin and $1. Samples were run on 10% polyacrylamide gel in Laemmli's (1970) buffer, and were stained by Coommassie Blue. (A-~E)Excess $1; 3.75 molar ratio $1 to actin (I.8 mg ml -~ $1, 0.17 mg ml -~ F-actin). (F-J) Excess actin; 0.25 molar ratio $1 to actin (0.12 mg ml -I $1, 0.17 mg ml -~ F-actin). Crosslinking was done in solution containing 50 mM KCI, 10 mM Tris HC1 buffer, pH 7.5 at RT with 100 mM EDC. Times of crosslinking were 0 min (A and J), 5 min (B and F), 10 rain (C and G), 20 min (D and H) and 40 min (E and I). The positions of 265 kDa band, 175-185 kDa doublet, $1 and actin are marked by 265, Db, $1 and Ac, respectively.

level expected of binding of one $1 to one actin monomer (Fig. 4, centre panel). The significance of this finding is considered in the Discussion.

Addition of ATP to either slow or fast forming acto-S1 complex reversed the increase in turbidity (Figs 3 and 4, right panel).

Time

2.0

I

Titrafions of F-actin by H M M measured by a change of turbidity

I

I

9 slow 0 fast

o

0

In analogy to the case of SI, the turbidimetric titration of F-actin by H M M show that binding isotherm depended on the speed of addition of protein. Fast forming complex saturated at a molar ratio of 1.02 heads per actin monomer, and slow forming complex saturated at molar ratio of 0.67 heads per actin monomer (Fig. 5).

0

1.5

v

o

1.0

Crosslinking experiments

b I--

0.5

0.0

0.0

o

I

0.5

,.'

1.0

I

I

1.5

2.0

2.5

Molar ratio [ h e a d ] / [ A ]

Fig. 5. Turbidimetric titration of F-actin with HMM. HMM was added either slowly ( 0 ) approximately 600 s between additions, or quickly (O) approximately 60 s between additions, to a fixed (1.2 p,M) concentration of F-actin. HMM (0--2.4 I,tM heads) was in 50 mM KC1, 10 mM Tris HC1 buffer, pH 7.5. The quantity 0.02 was subtracted from the turbidity values for fast titration in order to superimpose the first (no HMM) point.

Our data points to the presence of two different conformations of acto-S1 complex. We will argue (see Discussion) that in the slow forming complex $1 binds to two actin monomers, and in the fast forming complex to one actin monomer. Providing the experiments are done slowly, the only factor determining how $1 will bind is the relative concentrations of actin and $1. When $1 is in molar excess over actin, it will bind one monomer, and when actin is in excess over SI, it will bind two monomers. It may be that different stoichiometries of binding reflect different configurations of the acto-S1 complex. Thus providing that experiments are done slowly enough, the acto-S1 complex could crosslink differently depending on the molar ratio of actin and $1 (one is tempted to speculate that the complex should be easier to crosslink when each $1 binds to two actin monomers). Figure 6 shows that this is indeed the case.

528

ANDREEV and BOREJDO 20 S1 IN EXCESS

9 in doublet O in 265 KDa 15 7O C 0 ..0 0 G

10

E O

0

--0 0

0

0

I

I

I

0 I

10

20

30

4o

50

time (min)

Fig. 7. Time course of creation of crosslinked products when $1 is in excess over actin. Intensity of each doublet band in Fig. 6 was determined by measuring average intensity within an area-of-interest (AOI) enclosing given band. Each measurement was corrected for the intensity of the background. (0), I75-185 kDa doublet; (O), 265 kDa band. It compares the time course of crosslinking when $1 (A-E) or actin (F-J) are in molar excess. We emphasize that the crosslinking was initiated after 30 min of incubation of F-actin and $1 (Materials and methods), i.e. the complex formation was allowed to reach equilibrium. The time course reflects only the kinetics of the chemical reaction of EDC, not $1 isomerization. The crosslinked products are doublet (indicated in Fig. 6 as Db) with apparent molecular weight of 175-185 kDa and a band with an apparent weight of 265 kDa (indicated in Fig. 6 as 265; Morner et al., 1981). The amount of actin was the same in the two experiments, Figure 7 compares the time course of production of 175-185kDa doublet and 265 kDa band when $1 is in excess. Only doublet was formed and no 265 kDa band was produced at all. Figure 8 compares the time course of production of both products when actin was in excess. Little 265 kDa band was produced at first, suggesting that EDC crosslinks one of the acto-S1 contacts faster than the other. After 20 min of crosslinking, however, more 265 kDa band was formed than the doublet. The time course of production of the crosslinked products suggests that 265 kDa band was produced at the expense of 175-185 kDa doublet. Similar observation was made by Huang and colleagues (1990). When formation of 265 kDa band is compared in Figs 7 and 8 it is seen that no 265 kDa band was formed when $1 was in excess, even though in this case there was four times more of the acto-S1 complex available for crosslinking.

The above results suggest that different protein interfaces are being crosslinked depending on the molar ratio of actin and $1. Specifically, 175-185 kDa doublet may correspond to the situation when $1 is crosslinked to one actin (because it is a predominant product early during crosslinking and is always predominant when $1 is in excess over actin) and 265 kDa band may correspond to the situation when $1 is crosslinked to two actins (because it is absent early during crosslinking and it is a predominant product when actin is in excess over $1). If this is true, then there should be twice as much actin in 265 kDa band as in 175-185 kDa doublet. To test this, a control experiment was done to check whether the relative amounts of two proteins labelled with either 5-IAF or IATR, which were present in the same electrophoretic band, could be determined from the fluorescence of 5-IAF or IATR present in this band. $1 was labelled with 5-IAF and with IATR and the samples were prepared consisting of a mixture of the two labelled species at different molar ratios. Each sample was then electrophoresed in 10% polyacrylamide in glass tubes. The fluorescence spectra of the band corresponding to the heavy chain of $1 was recorded as described in the Materials and methods. Figure 9 shows that the ratio of fluorescence intensities contributed by 5-IAF and IATR was proportional to the molar ratio of the two Sls. This demonstrates that, when both 5-IAF and IATR dyes are present within the same electrophoretic band, the fluorescent intensity of one dye is not influenced by the fluorescence of the other, at least not when the dyes are in the range of molar ratios 0.I-10. Any dye which remained non-covalently bound to $1 after purification 20 ACTIN IN EXCESS 9 in doublet O in 265 KDa 15 7D C O 0 c

10

c

E Q

/ o

--o ~/0 0

I

I

I

I

10

2o

30

40

5o

time (mln)

Fig. 8. Time course of creation of crosslinked products when actin is in excess over $I. Each point was obtained as described in the legend to Fig. 7. (0), 175-I85 kDa doublet; (C)), 265 kDa band.

529

Two different acto-SI complexes 60

I

I

I

was also close to 2 when 1,5-IAEDANS and IATR were used to label actin and $1, and when labels were reversed (not shown). We therefore conclude that if 175-185 kDa corresponds to complex 1:1 then 265 kDa corresponds to complex 2:1.

I

50

40 Discussion

E 3O

20

1o

0

--

o

I

I

I

I

2

4

6

8

1o

[S 1IAF]/[S 1IATR] Fig. 9. Proportionally between the relative amounts of SI labelled with either 5-IAF or IATR, present within the same electrophoretic band, and the fluorescence of 5-IAF or IATR present in this band. Vertical axis: ratio of the intensity of fluorescence of 5-IAF (measured at ~'ex= 480 nm, L~m= 532 rim) to the intensity of fluorescence of IATR (measured at K~x= 530 nm, ~,e~= 582 nm). Excitation and emission slits were 2 nm and 10 nm, respectively. Horizontal axis: molar ratio of $1 labelled with 5-IAF to $1 labelled with IATR. (0), 36 p,g S-1; (O), 4 gg of $1. by Sephadex G-50 column dissociated from SI during electrophoresis in the presence of SDS. To determine the relative amounts of actin and SI in crosslinked products, we compared the fluorescence emission spectra of both. Proteins were crosslinked and the fluorescence spectra of I 7 5 - I 8 5 k D a doublet and of 265kDa band were recorded directly from gel in a glass tube as described in the Materials and methods. Figure I0 shows emission spectra of 175-185 doublet (curves I and 3) and 265 kDa hand (curves 2 and 4). As fluorescence intensity is proportional to the concentration of labelled protein, we determine molar ratio of actin to SI in 175-185 kDa doublet (MRI) and in 265 kDa band (MR2) as MRI

= a

The results lead us to propose that $1 can bind actin filament in two different attitudes. One attitude is such that $1 binds only to one actin monomer. This is fast forming, complex, which we call AI*S1. In analogy with the model of Huxley (1969) we visualize it with $1 having an attitude 'perpendicular' to F-actin. The long axis of $1 in this conformation may be perpendicular to F-actin and allow neighbouring actin monomer to be occupied by another $1. The other attitude is such that $1 binds to two actin monomers. This is slow forming, complex which we call A2"$1. In analogy with Huxley's model we visualise it with $1 having an attitude 'parallel' to F-actin. The long axis of $1 in this conformation may be parallel to F-actin and not allow the neighbouring actin monomer to be occupied by another $1. $1 in AI*S1 complex can transform (we use the word 'transform' interchangeably with words reorient, relax or isomerize) to attitude characteristic of A2"$1 complex. The idea is schematically illustrated in Fig. 11. When actin is in excess (A), the initial binding is rapid and results in the formation of 1:1 complex (transition 1--,2). Formation of an equilibrium I:2 complex (transition 2--+3) requires several minutes, but it does not mean that the isomerization itself is slow. Actual isomerization may be fast, but quenching and turbidity may reflect the time necessary to reach the final equilibrium between AI*S1 and A2"$1. This time may include time taken by slow processes different from isomerization of S1 (e.g. redistribution of $1 on the 0.5 0.4

where cr is proportionality constant and I532(1), Is32(2), I5~z(3) and I58z(4) are fluorescence intensities at 532 nm and 582 nm calculated from curves 1, 2, 3 and 4, respectively. Taking MRI = 1 (Sutoh, 1983; Greene, 1984; Heaphy & Tregear, 1984; Chen et al., 1985) we calculated that molar ratio in 265 kDa as MR2 = [I532(1)I582(4)]/[I582(3)I532(2)] = 2. I When labelling was reversed (i.e. F-actin was labelled by 5-IAF and S1 by IATR), MR2 was 2.06 (not shown). MR2

2

8 C

0.2

I582(3)/I532(I)

MR2 = ~ I582(4)/I532(2)

FIATR 9 S-11AF

.~

,-v 0.1



Fig. 10. Emission spectra of 175-185 kDa doublet (curves I and 3) and of 265 kDa band (curves 2 and 4). The spectra I and 2 recorded at ~,ex= 480 nm, spectra 3 and 4 at ~'ex= 530 nm, excitation slit 2 nm, emission slit I0 nm. The Coming 3-66 and 3-70 filters were used to eliminate scattered light at 530 nm and 480 nm, respectively. SI was labelled with 5-IAF (0.9 tool dye per tool SI) and F-actin with IATR (0.62 mol dye per mol actin).

530

A

ANDREEV and BOREJDO 1

2

A+S-1

B

,,,, "

A1 9 S-1

1

A+S-1

3

N

\

A2" S-1

2

.,

\

A1 oS-1

Fig. 11. Schematic illustration of the two modes of binding of $1 to F-actin. A: actin in excess. $I initially binds 1 actin monomer in perpendicular attitude, which slowly undergoes isomerization to assume parallel attitude. B: $1 in excess. surface of F-actin). Our results do not allow us to state whether our binding process is the same as the one observed by Geeves and colleagues (Geeves et al., 1984; Coates et al., 1985). Preliminary evidence, obtained by measuring linear dichroism of fluorescent $1 diffused into muscle fibres, shows that $1 diffused into fibres at high concentrations assumes different orientation than $1 diffused into fibres at low concentrations. Pyrene quenching experiments are explained by above '2 complexes' hypothesis as follows: when quenching was measured by slowly adding increasing amounts of $I to a fixed concentration of pyrene actin (Fig. 1B), the quenching initially increased because it reflected the amount of acto-S1 complex. In this complex $1 binds to two actin monomers. When the molar ratio exceeded two actins per one $1 and further $1 was added, the fluorescence remained constant because all actin sites were already occupied, $1 was unable to find a free site on actin filament and remained free in solution. However, when $1 was added quickly (Fig. 1A), $1 did not have enough time to reorient on the surface of F-actin to bind to two monomers. It therefore bound only one monomer and binding isotherm was completely different. The same difference between fast and slow addition was true of turbidity measurements (compare Fig. 1A, open circles with Fig. 1B, open circles). The fact that the shape of binding isotherm could be approximated by a hyperbola when complexes were formed quickly and by two straight lines when complexes were formed slowly is consistent with the idea that two complexes represent $1 binding to one and two actins, respectively. Binding to one actin is expected to be only moderately strong, hence

classical hyperbolic shape of the isotherm (Fig. 1A). Two actin binding is expected to be very strong, hence linear increase of binding with increase of $1 concentration, followed by a region where amount of complex is independent of $I concentration (Fig. 1B). Our results suggest that the equilibrium between the two distinct acto-S1 complexes is dependent on a molar ratio of $1 to actin. We concluded that at low molar ratio of $1 to actin the A2"$1 complex is predominant and that at higher molar ratio the AI*S1 is predominant. These are two different systems because molar ratio of $1 to actin are different in each, therefore their equilibrium states are different. Each system can be transformed to the other by changing the molar ratio of $1 to actin. This transformation was monitored by following changes in turbidity. The turbidity measurements reflect the amount of acto-SI complexes but do not reflect changes in $1 orientation with respect to actin. The transition of A2"$1 to A~*S1 induced by the addition of excess $1 took about 2 h and was accompanied by doubling of turbidity (Fig. 4, centre panel). This increase in turbidity would be expected because the amount of acto-S1 complexes was increased two times. The reverse transformation, transformation of AI*S1 to A2"$1, could not be detected by a change in turbidity because it was not accompanied by a change in the amount of acto-S1 complexes. Our data does not allow us to say whether in 'parallel' attitude $1 binds to two different strands of actin or to the same strand (Milligan et aI., 1990). Careful measurements of turbidity when F-actin was titrated by $1 was made by White and Taylor (1976), Lehrer and Ishii (1988) and Ishii and Lehrer (1990). The fact that only one type of acto-S1 complex was found (AI*S1) suggests that these workers added $1 quickly. One instance when stoichiometry was one $1 to two actins was found when $1 was interacting with G-actin (Valentin-Ranc et al., 1991). Our results are not inconsistent with ultracentrifugation experiments which clearly showed 1:1 stoichiometry (Eisenberg et al., 1972; Margossian & Lowey, 1973, 1978; Greene & Eisenberg, 1980). In these experiments each molar ratio of actin and $1 is examined in a separate tube. In the tubes in which actin is in excess, $1 indeed assumes parallel attitude because enough time elapses between mixing and measurement. Once actin and $1 are mixed at 1:1 molar ratio, however, $1 will bind in perpendicular attitude and cannot reorient itself to bind to two monomers because there is no room on the surface of a filament to allow it to do so. It is therefore not surprising that in the ultracentrifugation experiments (and the experiments in which each molar ratio of actin and $1 was examined separately (Ando & Asai, 1979)) the stoichiometry was 1:1. Ultracentrifugation experiments question the maximum amount of $1 that can bind to actin, and the answer is clearly one mol $1 per mol actin. In contrast, to obtain slow forming complex, it was

531

Two different acto-SI complexes necessary and sufficient to have actin in excess and to add $1 slowly. After binding to one actin monomer, $1 has enough time to reorient itself on the surface of a filament to bind to two monomers. These experiments question the minimum amount of $1 that can saturate all actin binding sites, and the answer is 0.5 mol $1 per mol actin. The fact that HMM gave the same stoichiometry per head as $I (Fig. 5) suggests that each head binds to actin independently. This is not surprising because it is known that the rotational mobility of each head of HMM is similar to that of $1 (Mendelson et al., 1973) and that the rigor orientation of HMM with respect to the muscle fibre axis is the same as that of $1 (Borejdo et al., 1982). Our finding that the time course of creation of crosslinking products is different depending on whether actin or $1 are in molar excess (Figs 6-8) strongly suggests that depending on the molar ratio, different protein interfaces are being crosslinked. This implies that the conformation of acto-S1 is different depending on the molar ratio. This suggestion may explain why the maximal ATPase rate and the apparent activation constant of acto-S1 in solution is several fold different depending on whether the measurements are done keeping actin or $1 constant (Stein & Harwalkar, 1989). It also may explain the observations of Yarnamoto (1990) that the binding manner between actin and the lysine-rich sequence at the junction between 50 K and 20 K domains of $1 depends on the actin to SI molar ratio and the fact that actin is necessary at twofold molar excess over $1 for effective protection against the trypsin cleavage at the 50-20 kDa junction (Mornet et al., 1979). Crosslinking experiments showed that the biggest difference in the production of crosslinking products was in creation of the 265 kDa band. In spite of the fact that there were four less acto-S1 complexes when actin was in excess, much more 265 kDa band product was created when actin was in excess than when $1 was in" excess (Fig. 8). It is likely that 265 kDa complex was produced as a result of crosslinking of 175-185 kDa complex with second actin monomer. In the original paper Momet and colleagues (1981) suggested that 175-185 kDa doublet indicated two actins crosslinked to one $1 and that the 265 kDa band indicated actin quadruplet crosslinked to one $1. More recent measurements (Sutoh, 1983; Greene, 1984; Heaphy & Tregear, 1984; Chen et al., 1985) indicated that the 175-185 kDa doublet was really an anomalously migrating 1:1 complex of actin and $1. Our results indicated that the 265 kDa band is an anomalously migrating 2:1 complex of actin and $1. On the other hand Heaphy and Tregear (1984) estimated the molar ratio of actin to $1 in 265 kDa product at approximately 3.5 • 1 and Lu and Wong (1991) estimated it at 1.0 (the same as in doublet). The suggestion that there are two conformations of the acto-S1 complex is not new. Thus Shriver and Sykes (1981) presented NMR evidence that Sl-nucleotide complex existed in two temperature-dependent confor-

mations, and suggested that ternary acto-SI-nucleotide complex may also exist in two distinct conformations. Borejdo and colleagues (1982) and Ajtai and Burghardt (1986) showed that the conformation of SI in muscle fibre was different depending on the presence of MgADP. Arata (1984) suggested by crosslinking that the structures of binary complexes may be different from the ternary complexes, and Bhandari and colleagues (1985) directly showed that the energy transfer efficiency between actin and $1 in binary acto-S1 complex was different from the efficiency in an analogue of the ternary acto-SI-ATP complex. In this paper two different acto-SI complexes were observed in the absence of nucleotide. In these complexes one or two actins interacted with one $1. Electron microscopy (EM) showed that SI in acto-SI complex looked different depending on the molar ratio of SI to actin (Craig et al., 1980). However, no conclusions could be drawn from EM without doing three-dimensional reconstructions of images. Unfortunately, three-dimensional reconstructions of electron microscopic images of A2"$1 (i.e. when actin is in excess over SI) are nearly impossible (personal communication Drs R. Craig, K. Taylor, R. Milligan and E. Egelman). We can speculate whether the two complexes observed in this paper have any relation with intermediate states of acto-SI during ATP hydrolysis. One attractive possibility is that during each ATP hydrolysis cycle S1 first binds to one actin (corresponding to 'moderately strong' binding), and then to two (corresponding to 'very strong' binding) actin monomers, and that this transition is accompanied by force generation. This possibility supports the idea that force is generated by orientational transition of the crossbridge between two attached states (Huxley, 1969).

Acknowledgements We thank A. Andreeva for expert technical assistance, and R. Takashi and S. Burlacu for helpful discussions. Supported by NIH AR40095-02.

References AJTAI, K. & BURGHARDT, T. P. (1986) Observation of two orientations from rigor crossbridges in glycerinated muscle fibres. Biochemistry 25, 6203-7. ALBERTS,B., BRAY,D., LEWIS,J., RAFF,M., ROBERTS,K. & WATSON, J. D. (1983) Molecular Biology of ~he Ceil, pp. 138 and 227. New York & London: Garland Inc. AMOS,L. A., HUXLEY,H. E., HOLMES,K. C., GOODY,R. S. & TAYLOR, K. A. (I982) Structural evidence that myosin heads may interact with two sites on F-actin. Nature 299, 467-9. ANDO, T. & ASAI, H. (1979) Conformational change in actin filaments induced by the interaction with heavy meromyosin. J. Mol. Biol. 129, 265-77.

ANDREEV and BOREJDO

532 ANDO, T. & SCALES, D. (1985) Skeletal muscle myosin subfragment 1 induces bundle formation by actin filaments. J. Biol. Chem. 260, 2321-7. ANDREEV, O. A. & BOREJDO, J. (1991a) Stoichiometry of binding of myosin subfragment-1 to F-actin measured by fluorescence polarization. Biophys. ]. 59, 429a. ANDREEV,O. A. & BOREJDO,J. (1991b) The myosin head can bind two actin monomers. Biochem. Biophys. Res. Comm. 177, 350-6. ARATA, T. (1984) Chemical crosslinking of myosin subfragmentI to F-actin in the presence of nucleotide. J. Biochem. 96, 337-47. BHANDARI,D. G., TRAYER,H. R. & TRAYER, L P., (I985) Resonance energy transfer evidence for two attached states of the actomyosin complex. FEBS Left. 187, 160-6. BOREJDO, J. & ASSULIN,O. (1980) Binding of heavy meromyosin and subfragment-1 to thin filaments in myofibrils and muscle fibres. Biochemistry 21, 4913-21. BOREJDO, J., ASSULIN,O., ANDO, T. & PUTNAM, S. (1982) Crossbridge orientation in skeletal muscle measured by linear dichroism of an extrinsic chromophore. ]. Mol. Biol. 158, 391-414. BOTTS, J., THOMASON,J. F. & MORALES,M. F. (1989) On the origin and transmission of force in actomyosin subffagment 1. Proc. Natl. Acad. Sci. USA 86, 2204-8. CHEN, T., APPLEGATE,D. & REISLER,E. (1985) Crosslinking of actin to myosin subfragment 1 in the presence of nucleotides. Biochemistry 24, 137-44. COATES, J. H., CRIDDLE, A. H. & GEEVES, M. A. (1985) Pressurerelaxation studies of pyrene-labelled actin and myosin subfragment I from rabbit skeletal muscle. Evidence for two states of acto-subfragment 1. Biochem. ]. 232, 351-6. COOKE, R. (1986) The mechanism of muscle contraction. CRC Crit. Rev. Biochem. 21, 53-118. COOPER, J. A., WALKER, S. B. & POLLARD, T. D. (1983) Pyrene actin: documentation of the validity of a sensitive assay for actin polymerization. J. Muscle Res. Cell Motil. 4,

253-62. CRAIG, R., SZENT-GYORGYI,A. G., BEESE,L., FLICKER,P., VIBERT,P. & COHEN, C. (1980) Electron microscopy of thin filaments decorated with a Ca 2+ regulated myosin. ]. Mol. Biol. 140,

35-55. CRIDDLE, A. H., GEEVES, M. A. & JEFFRIES,T. (1985) The use of actin labelled with N-(1-pyrenal)iodoacetamide to study the interaction of actin with myosin subfragments and troponin/tropomyosin. Biochem. ]. 232, 343-9. EISENBERG, E., DOBKIN, L. & KIELLEY, W. W. (1972) Heavy meromyosin: evidence for a refractory state unable to bind to actin in the presence of ATP. Biochemistry 11, 4567-60. FINLAYSON, B., LYMN, R. W. & TAYLOR, E. W. (1969) Studies of kinetics of formation and dissociation of the actomyosin complex. Biochemistry 8, 811-9. GEEVES, M. A. & HALSALL,D. J. (1980) The dynamics of the interaction between myosin subfragment-1 and pyrenelabelled thin filaments from rabbit skeletal muscle. Proc. R.

Soc. Lond. B 229, 85-95. GEEVES,M. A., GOODY, R. S. & GUTFREUND,H. (1984) Kinetics of acto-S1 interaction as a guide to a model for the crossbridge cycle. 1. Muscle Res. Cell. Motil. 5, 351-61. GREENE, L. E. (1984) Stoichimetry of actin*S-1 crosslinked complex. ]. Biol. Chem. 259, 7363-6.

GREENE, L. & EISENBERG, E. (1980) The binding of heavy meromyosin to F-actin. J. Biol. Chem. 255, 543-8. HIBBERD,M. G., DANTZIG,J. A., TRENTHAM,D. R. & GOLDMAN,Y. E. (1985) Phosphate release and force generation in skeletal muscle. Science 228, 1317-9. HEAPHY, S. & TREGEAR, R. (1984) Stoichiometry of covalent actin-subfragment 1 complexes formed on reaction with zero-length crosslinking compound. Biochemistry 23, 2211-4. HOLMES, K. C., POPP, D., GEBHARD,W. & KABSCH,W. (1990) The

Structure of F-actin calculated from X-ray fibre diagrams and the 0.6 nm crystal structure. Molecular Basis of Muscular Contraction (edited by Squire, S. M.), pp. 65-75. Boca Raton, FL: CRC Press. HUANG, Y-P., KIMURA,M. & TAWADA,K. (1990) Covalent crosslinking of myosin subfragment 1 and heavy meromyosin to actin at various molar ratios: different correlations between ATPase activity and crosslinking extent. J. Muscle Res. Cell MotiL 11, 313-22. HUXLEY, H. E. (1969) The mechanism of muscular contraction. Science 164, 1356---66. ISHII, Y. & LEHRER, S. S. (1990) Excimer fluorescence of pyrenyliodoacetamide-labelled tropomyosin: a probe of the state of tropomyosin in reconstituted muscle thin filaments. Biochemistry 29, 1160-6. KOUYAMA, T. & MIHASHI, K. (1981)~Pulse-fluorometry study on actin and heavy meromyosin using F-actin labelled with N-(1-Pyrene) maleimide. Eur. ]. Biochem. 114, 33-8. LAEMMLI,U. K. (1970) Cleaveage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-5. LEHRER,S. S. & ISHII,Y. (1988) Fluorescence properties of acrylodan-labelled tropomyosin and tropomyosin actin: evidence for myosin subfragment-1 induced changes in geometry between tropomyosin and actin. Biochemistry 27, 5899-906. LU, R. C. & WONG, A. (1991) Evidence for the existence of multiple interfaces between actin and myosin subfragment-1. Biophys, ]. 59, 411A. MARGOSSIAN,S. S. & LOWEY,S. (1973) Substructure of the myosin molecule. IV. Interactions of myosin and its subfragments with adenosine triphosphate and F-actin. J. Mol. Biol. 74, 713-30. MARGOSSIAN, S. S. & LOWEY, S. (1978) Interaction of myosin subfragments with F-actin. Biochemistry 17, 5431-9. MENDELSON, R. A., MORALES,M. F. & BOTTS, J. (1973) Segmental flexibility of the S-1 moiety of myosin. Biochemistry 12,

2250. MILLIGAN, R. A., WHITTAKER,M. & SAFER, D. (1990) Molecular structure of F-actin and location of surface binding sites. Nature 348, 217-21. MORNET, D., PANTEL,M., AUDEMARD,E. & KASSAB,R. (1979) The

limited tryptic cleavage of chymotryptic S-I: an approach to the characterization of the actin site in myosine heads. Biochem. Biophys. Res. Comm. 89, 925-32. MORNET, D., BERTRAND, R., PANTEL, P., AUDEMARD, E. & KASSAB, R. (1981) Structure of the actin-myosin interface. Nature 292, 301-6. SCHOENBERG, M. (1988) In Molecular Mechanism of Muscle Contraction (edited by Sugi, H. & Pollack, G. H.) New York: Plenum Publishing.

T w o different acto-S1 complexes SHRIVER, J. W. & SYKES, B. D. (1981) Phosphorus-31 nuclear magnetic resonance evidence for two conformations of myosin subfragment-1 nucleotide complexes. Biochemistry 20, 2004-12. SPUDICH, J. & WATT, S. (1971) Regulation of rabbit muscle contraction. ]. Biol. Chem. 246, 4866--71. STEIN, L. & HARWALKAR,V. A. (1989) Biochemical kinetics of skeletal actoS1 at high $1 concentrations. Biophys. J. 56, 263-72. SUTOH, K. (1983) Mapping of actin-binding sites on the heavy chain of myosin subffagment-1. Biochemistry 22, 1579-85. TONOMURA, Y., APPEL, P. & MORALES, M. F. (1966) On the molecular weight of myosin. II. Biochemistry 5, 515-21.

533 VALENTIN-RANC,C., COMPEAU,C., CARLIER,M-F. & PANTALONI,D. (1991) Myosin subfragment 1 interacts with two G-actin molecules in the absence of ATP. J. Biol. Chem. 266, 17872-9. WEEDS,A. G. & TAYLOR,R. S. (1975) Separation of subfragment-1 isoenzymes from rabbit skeletal muscle myosin. Nature 257, 54-6. WEEDS, A. & POPE, B. (1977) Studies on the chymotryptic digestion of myosin. Effects of divalent cations on proteolytic susceptibility. J. Mol. Biol. 111, 129-57. WHITE, H. D. & TAYLOR,E. W. (1976) Energetics and mechanism of actomyosin adenosine triphosphatase. Biochemistry 15, 5818-26. YAMAMOTO, K. (1990) Shift of binding site at the interface between actin and myosin. Biochemistry 29, 844-8.