- Email: [email protected]
MgATP specifically controls in vitro self-assembly of vertebrate skeletal myosin in the physiological pH range
J. Mol. Bid. (1985) 182, 159-172
MgATP Specifically Controls in C/l’tro Self-assembly of Vertebrate Skeletal Myosin in the Physiological pH Range Ingrid Pinset-HPrstriim Service de Biophysique, Departement de Biologic Centre d’Etudes Nucle’aires de Saclay 91191 Gif sur Yvette, Cedex, France (Received 12 August
1983, and in revised form, 1X July
1984)
The appearances in the electron microscope of rat and rabbit skeletal muscle myosin filaments and rod aggregates, formed in the presence of variable amounts of MgATP. were compared at different pH values. It is shown that small amounts of MgATP, similar to t’hose sufficient to trigger the dissociation of the actomyosin complex, were able to modify the geometry of myosin filament’s profoundly in the physiological pH range, whereas the conformation of rod aggregates remained unchanged even in the presence of high concent’rat’ions of MgATP. Myosin filaments formed in the absence of MgATP displayed the classical spindle-shaped conformation and variable diameters at all pH values, whereas myosin filaments formed in the presence of MgATP in the physiological pH range ha,d constant diameters, similar to those of natural thick filaments. These filaments of constant diameter frayed, rapidly and reversibly, into two types of subfilaments with respective diameters of 4 to 5 nm and 9 to 10 nm. when the pH of the medium was raised above 7.2. Spindle-shaped myosin filaments and rod aggregates remained unchanged by such small changes in pH. It was possible to change the conformation of preformed spindle-shaped filaments by simply adding MgATP to the medium. but t’his reaction was slow and t’ook several hours to be completed. Relatively high concentrations of MgATP, similar to t,hose in the living cell, increased the solubility of both myosin filament’s and rod aggregates in the alkaline pH range (pH 2 7.0). Low pH values ( I 6.5) and excess free Mg’ + ( 2 6 to 7 mM) abolished both the specific effect of MgATP on myosin filament conformation and it,s solubilizing effect on both myosin filaments and rod aggregates. The degree of purity of the myosin preparations and the level of phosphorylation of the LC-2 light chains did not influence filament behaviour noticeably and rat and rabbit myosins behaved similarl,v.
1. Introduction Natural thick filaments of vertebrat’e skeletal muscle are remarkably conservative structures. They are invariably 1.6 pm long and 15 to 18 nm wide and display a bipolar conformation (Huxley, 1963; see Fig. l(a)). The myosin molecules, which are the major constituents, lie antiparallel in the middle of the filaments, forming a 150 nm wide central bare zone (Huxley, 1963: Craig & Offer. 1976; Craig, 1977). This invariance is in contrast to the wide spectrum of sizes and structures displayed by synthetic filament,s produced by spontaneous aggregation of myosin molecules in vitro. This is largely due to t’he fact that myosin self-assembly is extremely sensitive, not’ only to small changes in the pH and ionic strength of the medium (Kaminer & Bell, 1966), and to t’he nature of the ionic species (Harrison et al.. 1971; Reisler et al.. 1980), but also
t,o purely kinetic factors, such as the polymrrization rate (Huxley. 1963). In addition, with few exceptions, synthetic filament populations are very heterogeneous (Kaminer & Bell, 1966) or display obviously non-physiological sizes (Moos rt al., 1975; Reisler rt al., 1980). Finally, synthetic mposin filaments do not consistently display cenbral bare zones (Hinnsen et al., 1978), but when they do these are usually not’iceably wider than those of natural thick filaments (Huxley, 1963). It has been suggested that’ the structure of natural thick filaments might be determined, not only by the intrinsic characteristics of the myosin molecules themselves, but in addition by some non-myosin components (Huxley & Brown, 1967; Moos et al., 1975). Natural filaments indeed contain a great number of non-myosin proteins (Starr bz Offer, 197 1). some of which are located very precisely (Craig bz Offer, 1976: Starr & Offer, 1983). However,
it has not been Jtossible t#odemonstjratje whether any one of these has a conformat~ion-determining role. On thr other hand. we have shown (PinsetHiirstriim & Truff~. 1979) t.hat. synt,hetic filamentis displa>,ittg some of the key features of nat,ural thick filatnentjs. for example, constant physiological cliameters and 150 nm wide central bare zones. ma) form from solul)ilized rat skeletal muscle tnyosin. Itroviderl MgATJ’ is present (luring filament format’ion. Tt thus seemed possible that the particQatior1 of MgATP might be necessary for the cvrrect positioning of the myosin molecules in the filament backbone. This hypothesis seems to be supported by the parallel behaviour of smoot,h trtuwlt~ and non-muscle
myosins
in the presence
of
MgATP. The conformation of these latter rnyosins (*ittt indeed be changed by interaction with MgA4TP. attd i his conformat,ional change may modify filament fhrma,tion irt &YJ (Suziki rf (I/., 1978: Scholey rt nl.. 19X0: Hendrick-Jones rt al.. 1983). Skeletal tnyosin has at least two types of specific MgATP ittteravtion sites: first. the high-af?inity sites loc*ated in t,he heads. which are responsible for ATT h!~drolysis and trigger the dissociation of the it~~tOttl~VOSitl complex (Lymn & Taylor, lY71): swontl, several low-affinity sibes located alortp the rot1 segment. which might be implicated in filament c.ohesion (Harrington & Himmetfarb. 1971 ). (‘0~ secluetttl~\-. it swtnetl of irtt,ercst t.0 try to assess \vlticsh sites were responsible for the effect of MgATT’
OII the vonformation of synthetic skeletal myosin filaments. anti this was the aim of the present work. To this entl 1 cvtnpared the behaviour of myositt filatnc-tits and roti aggregates formed in the prrsrnce 01’ various amounts of MgATP at tlifferent pH salllrs. Since ltttosphor?;lat,ion of the IA’-2 li$tt c.ttain of’stnooth tnuscle rnyosin. \vhich is putativeI> Ioc*ated at the ,jnnc%ion between the heads and the rocl (Flicker rt trl.. 1981: t(endrick-Jones P/ ~1..
l!A%). niotlulates the depentlenee of in rsitro tilament fi)rtttation on MgATT’. the behaviour of variably l)hoSl)horvlated pa.re(l. 1,nstly. iti prwious studiw r;tt artd rabbit
under strictly
myosin
preparations
were conu Ot’li t 0 of skeletal ttt?-osin wlf-aswtttl)ly, skeletal myosins were c30tnparetJ
ortier to relate tliC> l)rrsrttt
itlentical 2. Materials
cxperimenial
c~otttiit~iotts.
and Methods
(‘rude myosin was prepared front the hind lep n~usc~lrs of rat. and the Irg and back muscles of rabbit. as described previously (Pinset-HgrstrBm & Truffy. l!C!)). The major pa,rt of the artin contamination of these preparations was removed eit’her by ult~racrntrif~tgatiotI at the maximum speed (about 130.000 g) in the Beckman Airfuge for 30 min in the presence of 04 M-KU at pH 6.8. adjusted with 20 mM-Tris-maleate or by means of a.mmonium sulphate precipitation. wtaining the 1.5 t,o .iP, cut. In some instancrs further purificaation was carried out by means of column chromatography on Sepharose 2B in the presence of XTP (M’halert rt al.. 1981). As judged by gel electrophoresis in the presence of 8 M-urea (Perrie & Perry. l!??(j). t,hr rat myosins prepared
by these twhniyues were ntainly nnphosphor?-l~lt~~~l. a11d attempts to carry out subsequent l)hosl)hor?-latiorr t, different published techniques w-ew unsucwssful in rn,v hands. The rabbit, myosinn prepared 1)~ the sanlv techniques. on the other hand. displayed variable phosphorylation. Essentially unphosphor?-lated at~tl up t-c, about 8Oo, phosphorylatetl preparations \vet’e obta,inrcl by the techniques described hg Kardatnt & (iratzrr (1982). .Ul preparations were c-arrird out at 4 (’ and all media cw~t.aitied 2 m.n-tlithiothrritol and (1.1 mwsotlium azide in order to try to avoid osidation of the sulphydryl groups and proteolqsis. ~everthelesn. sinw it has IMYII shown previously (Pmaet-HGstrtim 8 \Vhalen. I!)i!)) that the I,(‘-” light chain is particwlarl~sfbtlsitivc> to endogenous proteasrs. anti that t’s<‘11 \-csr,v limittscl havr far wac~hin~ proteolysis of this light c,hain ma!. effects on tilatnent a.ssetttt)l>-. its integrity \+as systrttrati cxlly Arcked by means of iso-rlec*tr.ic focwsirly (\Vl~;tlrtt r,/ 01.. 1WX). Rat and rabbit myostn rod subf’ra~tnents ww prepww~ from c~rudr m?osiits by c~hymotryl)tic~ diqMiot1 aworditip to the method of \Vretls & Pope (I 977).
(1,) Filrrnwnf ,fbrnlotion Fi1amrttt.s ww formed by drcwasin~ the KC0 concentration of the media from 0.5 \t t.0 0.12 11 I)) dilution at various rates under mapttetic stirring using constant-speed peristaltic l)umps, or hy dialysis ovflrnight. In the case of dilution. 1 ml of an initial solution was diluted t.o a final rolumt, of IO ml wit.lt a dilution solvent 048 M in KU This operation was wrrird out at room temperature. whereas the dialysis was c.arrietl out itt the cold room. \Vith thr rxccptio~t of thtl K(‘t wncetttration. the ionic, cwtttpositiorr of t hti initial solutions and the dilution solvents were itletttic*al. In particular. it should be noted that. the ATP ~ottc,rrtt.ratiott was the same in both media. C~onseyuetttly. filamtvit formation took place at cwnstant ATP c*otrc~entratioil. The pH of the medium was adjusted by means of 20 mwTrismaleate and all solvents c~oritained 2 rnht-tiithiothreitol and @I mwsodium azi&. Ot.htbr ionic cwmponrnts of thrl tnedia were variable and !+‘ill br described in tttv liesult,s. The protein concentration of the initial solutions varic~tl between I and 0.1 mg/rttl. The tinal prott~in c.otlc.f,tltratioit of the filament suspensiotw thus varicxd betv wti I ttig,“rnl (in some rxlwrimcnts c,ar,rircI out bs clialysis) anal 10 /4’gjral. In rlo (*use did T n0tic.e arty siqitieant itrfwric~t~ of t’he protein concentration on the results ohtainetl. It1 or&r to c,heck nhethrr it \ras possiblt~ to (.IIHI~~IC t,ltcs geomrtr~~ of preformecl myosin tilamrttts directly. without previous solubilization. follo~vt~d by tvag~re~ation in ii different tnrdium. t.htl ionic. c*omposition of tttv tilatnent suspensions uas sometitnrs c~hanged either I)y simpl> adding the required agent(s) rapidly under ma#nrtic, stirring or t)y dialysis orrrni~ht
The final filament suspensions were spread onto thin collodion-backed carbon films using eithri t1w cytochrome c monolayer sprradirq techniyur alread~~ described (Pinset-HLrstrStn 9r Truffy. 1979) or direct deposition onto carbon fillkts made hydrophilit~ by means of ultraviolet irradiation as tlesc*ribetl by Trirtick Ni Elliott (1’379). The results obtained using these 2 spreading bechniques wrc very sitnilar and no aignifieant difference in size or apparent struvtuw of the filament populations distributed over the qids ~wuld be idrntititd. X(,sttrthc,-
MgATP
Control of Myosin Assembly
less, as applied in this laboratory, the monolayer technique provided cleaner grids and more-even negative contrasting. It was consequently preferred in the majority of the experiments described here. After spreading the grids were rinsed with a small quantity of buffer and negatively contrasted using 2 to 3O/; freshly uranyl acetate. The collodion prepared. unbuffered backing was removed just before the introduction of the grids into the microscope by touching a filter paper soaked in isoamylacetate. The observations were carried out, by means of a Siemens Elmiskop 101. modified to allow operation at 120 kV under normal high-dose conditions at nominal magnifications of 10,000, 20,000 and 40,000 diameters. respectively. The real magnification was checked using negatively cvontrasted. unfixed catalase crystals. Filament sizes were measured either with a ruler on magnified prints or dire&l?, on the micrographs with a microcomparator.
3. Results (a) Experiments curried out in the r*bsence of MgATP These experiment’s were done in order to relate the present work to previous studies of skeletal myosin filament formation (Huxley, 1963; Kaminer & Bell, 1966; Hinssen et al., 1978) and to try to obtain a strict comparison between t’he behaviour of the different types of myosin preparation used in under classical experimental these experiments conditions. Filament formation was carried out at
161
pH 6.5 to 8.5 and either in the absence of MgCl, or in the presence of 10 m&i-MgCl,. Other medium conditions are described in Materials and Methods. The change in KC1 concentration w-as obtained either by dialysis or by dilution at, rat’es varying between 0.5 ml/min and 20 ml/min. The results obtained were entirely consistent’ with those of t)he authors mentioned above. and are illust’rated by Figures 2: 3, 4. 5 and 6 and Table 1. It follows from these results that, within the limits of the techniques applied, the origin of the myosin preparations, their degree of purity and the level of light chain phosphorylation had no significant influence on filament behaviour in the absence of MgATP. As shown in all previous studies. the short filaments formed at all pH values by rapid dilution. and at high pH values whatever the dilution rate. looked clearly bipolar, but displayed 1argelJ variable diameters. Filaments formed at lower pH values and at intermediate dilution rates were spindk-shaped. very inflated in the middle, with ends tapering over long distances. Filaments formed at’ these lower pH values, using sloa dilut’ion rates or, in particular, dialysis, displayed constant’ diameters, similar to t’hose of the inflated centres of spindle-shaped filaments; in other words. several-fold t’hicker t ban natural they were filaments. Rod preparations polymerized under t,he same experimental conditions as the myosin preparations. consistently gave rise to flat spindle-
(b)
Figure 1. Electron micrograph of (a) natural thick filaments prepared from rat leg muscle according t,o Trinick & Elliott (1979) and (b) synthetic filaments formed from crude rat skeletal myosin in the presence of 2 mM-ATP and 3 mM-M&l, at pH 7.0 at a dilution rate of 2 ml/min. The lengths of the 2 types of filament are clearly very different, yet their diameters and general morphological appearances are roughly the same, and strikingly dissimilar to those of synthetic filaments formed in the absence of MgATP. (Kaminer & Bell. 1966; Hinssen et al., 1978) (see also Fig. 13). Magnification: 35,000 x diameters. 6
I62
1. Pinset-Htirstrii,rn
20 ml/mln
t
5 ml/mln
Dialysis
IL
n
IL PH 7.5
I 0.1
DH 8.0
I 6.5
I 7.0
I 7.5
III 8.0
6.5
I 7.0
I 7.5
I 8.0
PH (a)
(b)
Figure 4. Mean lengths
Diameter
(nm)
Figure 2. Histograms
representSing the diamet’ers of filaments formed from crude rat myosin in the absence of MgATP at various pH values and dilut,ion rates. In order to try to measure the degree of spindle shape. the diameters were measured at 2 points: in the centre of the filament and midway between the centre and the tips. It c-an be seen that in the case of filaments formed at intermediate dilution rates the measured values cluster into 2 distinct blocks. corresponding roughly to the 2 measuring points. The short bipolar filaments formed at the highest pH value and the very long ones formed at slow filution rates (particularly by dialysis), on the ot’her hand. displayed approximately constant diameters. similar to those displayed by the spindle-shaped filaments in thp rentre
1
0.11
I
I
I
I
6.5
7.0
7.5
0.0
I
I
I
I
I
6.5
7.0
7.5
0.0
of filaments obtained from 3 distincat column-purified rat myosin preparations at various dilution rates and different pH values (a) in the absencr of MgATP and (b) in the presencr of 2 miv-ATP a’nd 3 mM-MgC1,. Symbols as in Fig. 3.
shaped aggregates of various sizes (Figs 7 and 8). The mean diameters of these aggregates changed with pH in a very similar way to those of myosin filaments (Fig. 8). but their overall sizes. and in particular their lengths, were much less influenced by the dilution rate than were myosin filaments. With the exception of the aggregates obtained at t’he most rapid dilution rates. they all displayed constant lengths of about I pm. Tn the ph,vsiologlcal pH range (6.5 to 7.5). howc,ver. t,hese aggregates showed a marked t’endency to cluster together on standing for longer periods (several hours or overnight), eventually forming “super-aggregat,es”. sometimes over 1 pm wide and several pm long. Both lateral and longitudinal growth thus seemed to be virtually unlimited. It’ should be noted, however, that even the longest) super-aggregates
-J
PH (a)
(b)
Figure 3. Mean lengths of filaments formed from 4 distinct crude rat myosin preparations at various dilution rates and different pH values (a) in the absence of MgATP and (b) in the presence of 2 mM-ATP and 3 mM-MgC1,. The dilution rates were as follows: (0, a), 20 ml/min; (0, n ) 10 ml/min; (a, A) 5 ml/min; (0. v) 2 ml/min; (0, 0) d’la ly sis. It can be seen that in every case the filament lengths depended strongly upon both the pH and the dilution rate, yet at each dilution rate the longest filaments were consistently obtained in the presence of MgATP and the physiological pH range.
I 6.5
I 7,O
I 7.5
I 0.0
I
I
I
I
6.5
7.0
7.5
0.0
PH (0)
Cb)
Figure 5. Mean lengths of filaments formed from 3 distinct non-phosphorylated, crude rabbit myosin preparations at various dilution rates and different pH values (a) in the absence of MgATP and (b) in the presence of 2 mM-ATP and 3 mM-MgCl,. Symbols as in Fig. 3.
MgATP
Control of Myosin Assembly -
-
163
Table 1 Mean diameters of jlaments formed at various dilution rates at different pH values from different types of myosin preparations in the absence of MgATP A. Crude, non-phosphorylated
rat myosin 5
ml/min Dialysis
20 ml/min ~ PH
6.5
7.0
7.5
0.0
I
I
I
6.5
7.0
7.5
I: 0-C )
.?b
Xa
Xl,1
i.0
24.5k4.0 21.1k2.7
35.2k10.5 44.8k4.0 32.0* 9.2 41.8k2.9
i.5 8.0
19.4k3.2 20.3+3.1
24.9+ 23.4*
6.5
I I I I
r
7.2 34.912.3 4.3
j:
24.6k4.0
44.Ok4.3
22.5k3.2 19.7k2.6
43.0f4.2 33.2k4.1 17.1+ 5.3
PH (a)
(b)
Figure 6. Mean lengths of filaments
formed from 3 distinct approx. 80% phosphorylated, crude rabbit myosin preparations at various dilution rates and different pH values (a) in the absence of MgATP and (b) in the presence of 2 mM-ATP and 3 mM-MgATP. Symbols as in Fig. 3.
never approached the lengths of the very long myosin filaments obtained by slow dilution or dialysis. Rat and rabbit rod preparations behaved very similarly and no influence of the presence of 10 mM-MgCl, on t*he aggregation was noticed.
carried out in the presence of 2 m&w-ATP and 3 WLM-A%fgC12
(b) Experiwbents
It has been shown (Pinset-HtirstrBm & Truffy, 1979) that the size and conformation of rat skeletal myosin filaments depend critically upon the relative concentration of ATP and MgCl,. Therefore, it was decided to carry out’ the first series of experiments in the presence of approximately physiological concentrations of these two agents, in order to explore the influence of pH on filament formation under these conditions, and to compare the behaviour of the different, myosin and myosin rod preparat,ions used in this work. To this end concentrations of 2 mM-ATP and 3 miw-MgCl, were chosen. Filaments were formed between pH 6.5 and 8.5, at dilution rates varying between 0.5 ml/min and 20 ml/min, as well as by dialysis. At pH 6.5 the filaments formed from rat and rabbit, purified and unpurified, phosphorylated and unphosphorylated myosins were virtually indistinguishable from those formed at corresponding dilution rates in the absence of MgATP (Figs 2, 3, 4, 5, 6 and 9; Table 2). From pH 6.8 and above, however, the situation changed very strikingly: all filaments formed at pH 6.8 and 7.0 displayed constant diameters of 15 to 18 nm, which were not modified by the dilution rate (Fig. 9; Table 2). With the noticeable exception of the lengths, the overall appearance of these filaments in the electron microscope was extremely similar to that of natural thick filaments (Fig. 1). These filaments were consistently longer than those formed at corre-
B. Column purijed,
6..5 7.0 7.5 84
23.6k4.2 20.4k4.2 21.3f5.0 19.5f4.5
non-phosphorylated
34.0+ 9.5 36.2k10.5 26.0& 9.5 21.5k10.5
C. (‘rude. non-phosphorylated
6.5 7.0 7.5
25.6f4.5 24.0+5.0 20.3f4.5
37.5+ 33.5+ 24++
8.0
18.9k4.5
20.6&
D. Prude, approximately
rat myosin
42.Ek4.3 45.Ok5.5 35.OL5.0
25.6f5.0 24.6f4.5 20.5f5.0
46.Ok4.3 44.5k4.6 32.6k4.7 245k5.0
22.7k5.5 25.2k4.7
47.5k5.5 44.5k4.5
rabbit myosin
10.5 5.5 4.8
44.5k4.5 42.5f5.0 33.7145
19.5Ifr4.5
4.8
31.5f5.0 20.514.7
SO0/Ophosphorylated rabbit myosin
5 mllmin 20 ml/min PH
s
%t
xa
2,
Dialysis r
6.5 7.0 7.5
22.6k4.3
34.8k4.6
43.7i4.6
23.7k4.6
44.8+5.8
19.5k4.6
36.6k4.3
46.054.8
22.9k4.7
24+i+5.0
34.055.1
225k4.8 19.5k4.6
8.0
19.0+4.4
20.5k4.3
43.9k4.8 33.7i4.6 18.7 & 5.0
In order to try to get an approximate evaluation of the degree of spindle shape, the diameters were measured at 2 points, at the centre of the filament, (&), and midway between the rentre and the tips (2,). The diameters are expressed in nanometers. Each mean value (i) and its corresponding variance correspond to a minimum of 400 individual measurements. A. Mean values corresponding to 6 distinct preparations. B. Mean values corresponding to 6 distinct preparations. C”. Mean values corresponding to 3 distinct preparations. I). .Ilean values corresponding to 3 distinct preparations.
sponding dilut’ion rates in the absence of MgATP (Figs 3, 4, 5 and 6). At pH 7.2 these filaments of constant diameter were replaced by markedly thinner ones, displaying various diameters, with coexisting some apparently unchanged filaments of constant diameter (Figs 9. 10 and 11). These thinner filaments were slightly shorter than t’he filaments of constant diameter formed at
164
I. Pinset-Hiirstriim
Figure 7. Elect,ron micrographs of aggregates pH 7-O at a dilution rate of 2 miimin. M&$&ation:
formed from rabbit, rod preparations in the absence (a) 35,000 x diameters(bj 140,000 x diameters. 20 ml/mln
corresponding dilution rates. Since the thinner filaments were easily breakable, this length difference might, however, be only apparent. The diameters of these thin filaments were rather variable, but the histograms of Figure 9 show that,, on the whole, they gather into two distinct groups, with apparent diameters of 4 to 5 nm and 9 t,o 10 nm, plus a supplementary group, presumably representing undissociated filaments of constant diameter. As can be seen from the micrographs of Figures 10, and 11, some of these latter, and even some of the thicker “sub-filaments”, split, up, mostly symmetrically about the centre of the filament’, whereas others only seemed to slim down gradually. At pH 7.5 the mean filament lengths reduced and the filament were substantially
5 ml/mln
of MgATP L
at
Dialysis
pH 6-5
pH 6.8
pH 7.0
>H 7-2
3H 7.5
0
20
40 Diameter
Figure
1
I 6.5
I 7.5
I 7-o
I 8.0
I
PH
Figure 8. Mean diameters
of aggregates formed from 2 distinct rabbit rod preparations at a dilution rate of 2 ml/min and various pH values in the absence of MgATP (O), and in the presence of 2 mM-ATP and 3 m&i-MgCI, (0).
9. Histograms
(nm)
representing the diameters of filaments formed from crude rat myosin in the presence of 2 mix-ATP at 3 mix-Mgcl, at various pH values and dilution rates. It can be seen that at pH 6.5 filament behaviour was similar to that displayed in the absence of MgATP (see Fig. 2). At pH 6.8 and 7.0, on the other hand, the diameters were uniform and independent of the dilution rate. At pH 7.2 and 7.5 distinctly thinner filaments with various diameters were formed, and above pH 8.0 no filaments at all could be distinguished.
MgATP
Control of Myosin Assembly
Table 2 Mean diameters of jlaments formed at various dilution rates and different pH values from different types of myosin preparations in the presence of 2 mM-ATP and 3 mM-MgCl, A. Crude, non-phosphorylated
rat myosin 5 mllmin
PH
20 ml/min ~~ -..-~~-s Xl01
6.5 6.8 7.0 7.2 7.5
P15k3.0 17,.5+_2.7 16.0&%2 lO.Sf6.0 9.0+3.2
B. C‘olumn-pwijed,
36.1klO.l 45.7f3.5 lS.l+_ 2.3 15.8_f 2.3 10.4k 5.6 9.5k 3.3
PH 6.5 6.8 7.0 7.2 7.5
2O.Qk3.0 18.3f2.8 15652.3 95k5.3 9.S + 3.0
36.9kQ.Q l&9+2.3 lSQf2.3 11.6+56 Q.Q_f3.1
C. Prude, non-phosphorylatrd
20.9*3.1 16Gjf2.9 16.4k2.2 11.3k6.2 9.it3.4
26.3k3.5
Dialysis f 44.Ok3.5 17.852.2 16.152.4 11.2k6.0 9.2k3.2
non-phosphorykxted rat myodin
5 ml/min 2() fnl/min ~~ -~ ~- - -~~.__~. s GJt ra
6.5 6.8 7.0 7.2 7-5
cl
%
44.8k3.8
CJ
Dialysis z
25.lk3.4
44.2k3.6 18.3f2.5 15.6f2.6 13CZ+6.6 10*3+3.5
25.7k3.3
44.Ok3.6 17.Ok2.4 lS.Ok 2.2 11.2k5.8 9.2k3.6
rabbit myosin
34.8T10.7 43.8*3,4 17.2+ 2.5 15.7* 2.4 lO.Q& 5.9 9.4k 3.6
1). Ckuie XOObphosphorylated rabbit myosin 5 ml/min PH
20 mlimin s
6.5 6.8 7.0 7.2 7.5
22.0 k 3.3 17.6k2.3 16.1k2.4 11.4k6.1 Q.Ok3.8
~J& 35.2 k lS.l+ 15.8+ 10.8+ 9,2*
X8
Xb
Dialysis f
10.4 46.0* 3.2 24.7 + 3.7 44.7 + 3.7 2.5 16.9* 2.3 2.2 15.9* 2.3 5.9 11.6k6.2 4.0 9.3k4.0
Measurements as indicated in A. Mean values corresponding IS. Mean values corresponding C. Mean values corresponding D. Mean values corresponding
Table 1. to 6 different preparations. to 6 different preparations. to 3 different preparations. to 3 different preparations.
diameters did not exceed 12 nm (Figs 2, 3, 4, 5 and 9). The micrograph of Figure 12 also shows that these short, thin filaments coexisted with substantial amounts of solubilized myosin molecules. At pH values of 8.0 and higher, only solubilized myosin molecules, and no filamentous structures, were distinguishable. Consequently, the presence of MgATP not only modified the filament conformation in the physiological pH range, but markedly
165
increased the actual solubility at pH values of 7.5 and above. By increasing the concentration of MgCl, from 3 InM to 10 mM, while maintaining the ATP constant, the specific effect of MgATP on myosin filament geometry was abolished (Fig. 13) and the filament solubility at high pH values decreased. The situation displayed in the absence of MgATP was thus restored. All the different myosin preparations behaved similarly, and I was unable to distinguish any influence of the origin, t’he degree of purity or light chain phosphorylation on myosin filament behaviour in the presence of 2 rnM-ATP and 3 m&r-MgCl,. As mentioned in Materials and Methods. in these experiments filament formation took place in the actual presence of MgATP. In order t,o see whether it was possible to change the filament geometry by simply adding MgATP to pre-formed filaments, 2 mM-ATP was added rapidly under magnetic stirring, or slowly by dialysis overnight, to filament suspensions formed in the presence of 3 mM-MgCl,. but in the absence of ATP. When this experiment was carried out by means of dialysis at pH 6.8 OI 7.0, the transformation of the classical spindleshaped filaments into substantially longer ones with constant diameters of 15 to 18 nm was complete. Adding ATP rapidly, on the other hand, did not have any immediate effect but, slowly. the transformation nevertheless took place, and was finally accomplished after four to five hours. Adding excess MgCl, to filaments pre-formed in t,he presence of 2 mM-ATP and 3 mM-MgCl,, did not have an immediate effect either but, gradually, the filaments became thicker and shorter, displaying the characteristic spindle-shaped conformaiion. Changing the pH from 6.8 or 7.0 to 7.2. on t,hr ot’her hand. by simply adding a small quantity of Trismaleate under magnetic stirring, immediately dissociated the constant diameter filaments formed in the presence of 2 mM-ATP and 3 mM-Mg(?l,. into subfilaments, and the reverse reaction was as rapid and apparent,ly complete. Raising the pH to 7.5 instantaneously partly dissolved the filaments formed at lower pH values in the presence of MgATP. Lowering the pH again, however, only produced relatively short filaments, which nevertheless displayed t*he same diameters as the initial ones. The overall conformation of the rod aggregates was not significantly influenced by the presence of 2 mM-ATP and 3 m&r-MgCl, at any pH value (Fig. 8). In the physiological pH range a,nd below the size of the aggregates was not significantly influenced either, but clustering together and fusing into super-aggregates did not occur even on standing for days. The mean diamet#er of the aggregates, however, decreased slightly at pH 7.5 and at values higher than 8.0 no aggregation was visible. The solubility of the rod aggregates was consequently increased at high pH values in the presence of 2 mM-ATP and 3 mM-MgClz. very much like that of myosin filaments. Increasing the concentration of MgCl, to 10 mM abolished this
166
I. Pinset-Hiirstriim
Figure 10. Micrographs of filaments formed at pH 7.2 at’ a dilution rate of 2 mljmin from column-purified, phosphorylated rat myo$in in the presence of 2 rnM-ATl’ and 3 mnl-MgCl,. Magnification: 35,000 x diameters.
solubilizing effect of MgATP. The presence of excess MgCl, also promoted fusing into super-aggregates in the physiological pH range. No difference was noticed between t,he behaviour of rat and rabbit, rod preparations. (c) Experiments carried out at various concentrations of MgA T P Since it’ has been shown (Pinset,-HSrst’rGm &, Truffy, 1979). t’hat free ATP may have a disorganizing effect on myosin filaments. it was decided to maintain the concentration of MgCI, at’ a consistently slightly higher value than that of the added ATP. Tn fact, a fixed ratio of t’wo thirds was maintained throughout these experiments. This means, for example. that’ when the ATP concentration was 0.1 m;M that of MgCl, was 0.15 mM. Nevertheless, for simplicity in t’he text. when referring tjo MgATP only the corresponding concentration of ATP is given. As described in Materials and Methods, the MgATP concentration was the same in the initial solutions and the dilution solvent; in ot’her words. ATP was continuously renewed during filament formation. Nevertheless, ATP hydrolysis inevitably occurred during the dilution and the ensuing spreading. To try to limit unknown effects of t,his. the experimental protocol was strictly standardized. The myosin concentration was maint’ained constant and low, at, 10 &ml, in the final filament suspension, and spreading was carried out according to a fixed time scheme in less than five minutes after dilution had finished. The indicated MgATP concentrations consequent’ly represent
non-
maximum values. which were presumably rather close t,o the real MgATP concentrations at high nominal values, but might represent substantial overest,imations at low nominal concentrations. These experiments were limited to crude. not column-purified. rabbit myosin. with either virtually or unphosphorylated Ll p t.0 80?,(, phosphorylated IL-2 light, chains. and to rabbit myosin rod preparations. Filament format,ion was obtained by dilution at a constant rate of 2 ml/min at pH 6.5 to 7.5. In other cvords. the KU concentration was changed from 0.5 M t,o 0.18 M in five minutes. The results from t,hese experiments are summarized in Figures 14, 15. 16. and 17. Tt is obvious that phosphorylated and unphosphorylated rabbit myosins behaved similarly, and t,hat no significant’ difference between t,he electron rnicroscopic appearance of the two types of filament was apparent at any pH value or MgATP ctoncentration. Another outstanding observation was t,hat the critical concentration of MgATP, necessary t)o modify the myosin filament conformation, was dramatica1l.y. pH-dependent. Indeed, at pH 6.5 no direct,ly rlslble effect on the overAl filament conformation was brought. about even by up t’o 10 mM-MgATP; only a very slight decrease in the mean filament diameter was noticed at ,MgATP concent,rations over 3 mM (Figs 14 and 15). At pH 6.8 a gradual decrease in the mean filament width, accompanied by a parallel increase in the mean lengths, occurred over a wide range of MgATP concentrations between 0.1 mM and 2 mM. At MgATP concentrations above 2 rnM the filaments displayed constant diameters of 15 to 18 nm and
MgATP
Control of Myosin Assembly
Figure 11. Micrograph of filaments formed at pH 7.2 at a dilution rate of 2 ml/min from column-purified phosphorylated rabbit myosin in the presence of 2 mM-ATP and 3 mM-MgCI,. Magnification: 35,000 x diameters.
obviously increased lengths. At, pH 7.0 a similar transition occurred in a narrower and at a much lower range of MgATP concentration. The midpoint of this transition was at about 0.02 mM-MgATP. As mentioned above, this value was presumably overestimated. This assumption was corroborated by the following observation: the preparations used in these experiments were very crude, and most of them contained noticeable amounts of actin, easily identified in the electron microscope images. At
167
809;)
nominal MgATP concentrations below 0.02 mM these actin filaments were clearly decorated with myosin: e.g. the actin-myosin complex was not dissociated. Whatever the actual MgATP caoncentration needed to trigger the substitution of spindleshaped filaments for filaments displaying constant physiological diameters, it thus seemed t,o be the same as the concentration sufficient to dissociate the actin-myosin complex. At’ pH 7.2 and 7-5 the spindle-shaped filaments were replaced by the t,hin
1fit3
1. Pinset-H&striim
Figure 12. Micrograph of filaments formed at pH 7.5 at a dilution rate of 2 ml/min from column-purified in the presence of 2 mM-ATP and 3 mM-MgCl,. Magnification: 35.000 x diameters.
subfilaments, low range
already described, in this same very
of MgATP
concentration. Whatever of MgATP, and whatever
actual concentration pH, these specific structural myosin filament conformation
the the
modificat,ions of t.he were abolished when
(0)
rat myosin
the concentration of MgCI, was raised to 10 rnM in excess over that of MgATP. As to the myosin rod aggregates, no significant modification of the overall conformation was noticed at any pH value and at. MgATP concen-
(b)
Figure 13. Micrographs of filaments formed at a dilution rate of 5 ml/min at. (a) pH 7.0 and (b) 8.0 from crude rat myosin in the presence of 2 mM-ATP and 10 mM-MgCl,. It can be seen that in both rases the filaments displayed t.he characteristic behaviour of filaments formed in the absence of MgATP. Nevertheless. MgATP is obviously present in the medium, since undecorated actin filaments are clearly distinguishable in the background. Magnification: 35,000 x diameters.
MgATP
169
Control of Myosin Assembly
r
IIO
0.01
0.1
I-0
IO.0
Mg ATP (RIM)
Figure 14. Mean diameters of filaments formed at
Mg ATP CrnM) a
dilution rat,e of 2 ml/min at different pH values, from 3 distinct non-phosphorylated crude rabbit myosin preparations as a function of the MgATP concentration. Symbols for Figs 14 to 18: (0, 0) PH 6.5; (0, n ) PH 6% (A, A) pH 7.0: (V. V) pH 7.2; (0, e, pH 7.5.
trations up to 15 mM. Only a slight gradual decrease in the mean diamet’er of the aggregates was noticed at and above pH 7.0 when the MgATP concentrat.ion exceeded 5 mM (Fig. 18). Nevertheless, the thinner aggregates formed at a given, lower pH (e.g. 7-O) in the presence of high amounts of MgATP (e.g. 10 mM MgATP) were indistinguishable from those formed at higher pH (e.g. 8.5) in the total absence of MgATP (Fig. 19). Raising the pH and adding thus had high c>oncentrations of MgATP qualitatively similar effects, as judged by electron
Figure 16. Mean lengths of filaments
formed at a dilution rate of 2 ml/min at different pH values, from 3 myosin rabbit non-phosphorylated crude distinct preparations as a function of the concenkation of MgATP. Symbols as in Fig. 14.
microscopy. Here again, the presence of high amounts of excess free MgC12 abolished the effect of MgATP. 4. Discussion (a) MgATP
and jilament formaticm
The results
show that low concentrations of to those that are sufficient to trigger the dissociation of the actomyosin complex. are able to transform profoundly the geometry of MgATP,
similar
40
r -5 5 E B
30
20
I( I-
-
Mg ATP CrnM)
Figure 15. Mean diameters of filaments formed at a dilution rate of 2 ml/min at different pH values, from 3 distinct, approx. SOS/, phosphorylated crude rabbit myosin preparations as a function of the concent>ration of MgATP. Symbols as in Fig. 14.
Mg ATP (ITIM)
Figure 17. Mean lengths of filaments formed at a dilution rate of 2 ml/min at different pH values, from 3 distinct approx. 800/b phosphorylated crude rabbit myosin preparations as a function of the concentration of MgATP, Symbols as in Fig. 14.
I. Pinset-Htirstr6m
170
160
0.01
0.1
I.0
IO.0
Mg ATP (mM)
Figure 18. Mean diameters of aggregates formed at a dilution rate of 2 ml/min at different pH values, from 2 distinct rabbit rod preparations as a function of the concentration of MgATP. Symbols as in Fig. 14.
synthetic filaments formed in t,he physiological pH range from all the different myosin preparat’ions st’udied here. whereas the conformation of rod aggregates was not noticeably influenced even by MgATP concentrations much greater than those of t,he living cell. This strongly suggests t’hat t’he ,l;lgATP interact’ion sites involved are the highaffinity sit,es located in the heads, and is to my knowledge the first indication that these sit’es are
(a I
implicated in skeletal muscle myosin filament assembly. In contrast to synthetic filaments formed in t)he absence of MgATP, which displayed the classical spindle-shaped conformation and various diameters (Kaminer & Bell, 1966; Hinssen et al.. 1978). those formed in its presence had constant diameters similar to those of natural filaments (Huxley, 1963) (Fig.1). Furthermore, t’he myosin filament,s formed in the presence of MgATP frayed, rapidly and reversibly, ink) two types of subfilaments wit’h diameters of I to 5 nm and 9 to 10 nm when the pH was raised t,o 7.2, whereas myosin filaments formed in the absence of MgATP. as well as rod aggregates, dissolved only gradually at) substantially higher pH values. This last observation strongly suggests that the assembly modes of myosin filaments formed in the presence a,nd in the absence of MgATP are fundamentally different. This conclusion is corroborated by t,he observntion that the direct transformation of one preformed filament t,ype int,o t,he other is a slow process. requiring several hours. The fact, that, this last phenomenon is so slow suggests that it may involve an actual reposit,ioning of the myosin molrculcs in the filament ba,ckbone. Such an event, cBertainly caannot take place during cross-bridge movement, but it might, represent a structural prerequisite for t’he proper functioning of the caontra&ile appa,ratus. Since the solubilities of bot’h myosin tilaments and rod aggregates were similarly influenced by relatively high concentrat,ions of MgATP. not very different from t.hose of the living cell. the rrsult>s are not cont’radictory to the
tb)
Figure 19. Micrographs representing rod aggregates formed at a dilution rate of 2 ml/min at pH 8.0 (a) in the absence of MgATP and (b) at pH 7.0 in the presence of 15 mM-MgATP. Magnification: 35,000 x diameters.
MgATP
Control of Myosin Assembly
suggestion of Harrington $ Himmelfarb (I971), that the low affinity MgATP interaction sites of the rod might modulate the overall cohesion of the myosin molecules in the filament backbone and hence, eventually, influence cross-bridge movement. The observation that the high-affinity MgATP interaction sites of the heads are implicated in filament assembly suggests that skeletal muscle myosins might be more analogous to smooth muscle and non-muscle myosins than previously realized. An important difference still persists between the two myosin types, since in the examples studied here LC-2 phosphorylation, which so dramatically modifies the behaviour of smooth muscle and nonmuscle myosin filaments (Kendrick-Jones et al., 1983), was without any noticeable effect on the conformation of skeletal muscle myosin filaments. However, it should be noted in this regard that electron microscopy, at least as it was applied here, is a very crude technique, which is certainly unable to reveal subtle structural differences at the level of the filament backbone or in the parts of the myosin molecules projecting from the filament surface. The fa,ct t’hat, slight conformational differences might exist between filaments formed from nonphosphorylated and 80% phosphorylated myosins, respectively, was indeed recently suggested by the work of Kardami & Gratzer (1982). These authors reported that suspensions of non-phosphorylated myosin filaments were substantially more turbid than suspensions of phosphorylated myosin filaments at a critical sodium chloride concentration of 0.16 M. The problem consequently deserves to be defined more precisely, using different approaches. The critical pH dependence of the MgATP effect on myosin filament conformation suggests that electrostatic interactions are of prime importance for filament assembly. This has already been predicted theoretically on t.he basis of sequence studies on nematode myosin (McLachlan & Karn, 1982, 1983) and demonstrated experimentally using selective proteolysis and chemical crosslinking st,udies on both natural and synthetic myosin filaments (Sutoh et al.. 1978; Chiao & Harringt,on, 1979; Ueno & Harrington, 1981) and rod aggregates (Reisler & Liu, 1982). Yet the present, results go beyond those sited above, since these latter studies consider only inberactions at the level of t,he rod, independently of any possible interference of MgATP. The present observation that lowering the pH and raising the concentration of free magnesium have qualitatively similar effects. e.g. that both abolished the specific MgATP effect on myosin filament conformation, might provide a plausible explanation to the mechanism of this effect. Indeed lowering the pH and raising the magnesium concentration are both liable to neutralize negative charges. which might be necessary for the correct positioning of the myosin molecules in the filament backbone. Consequently, it is possible that, interaction with MgATP at the level of the heads liberates negative charges at, some strategical point
171
on the surface of the myosin molecule, maybe at the junction between the heads and the rod, determining the specific conformation of natural thick filaments. (b) Nature of the subjilaments There is considerable evidence that natural thick filaments are built up from roughly parallel threadlike subunits, but there is no general agreement as to the nature of these subunits. Several authors have reported the existence of approximately 4 nm wide “microfilaments”, using electron microscopy (Katsura & Noda, 1973; Pepe & Dowben, 1977; Stewart et al.. 1981) and X-ray diffraction (Wray, 1979). Others have shown that when treated with distilled water isolated natural thick filaments may fray rapidly and reversibly into a maximum of three subfilaments displaying diameters at least double the size of the microfilaments (Maw & Rowe, 1980; Luther et al., 1981; Trinick, 1981). The present’ results show that. synthetic myosin filaments formed in the presence of MgATP can fray as rapidly and reversibly as natural ones, but that the subunits obtained display various diameters. Since, however, these subunits fall int,o two statistical blocks corresponding to mean diameters of 4 to 5 nm and 9 to 10 nm. respectively, it is tempting to associate the smaller subunits with the microfilaments of Katsura & Noda. and the larger subunits with the subfilaments of Maw & Rowe, which would mean that the t*wo are not incompatible. The observat,ion that fraying of synthetic filaments went far beyond t,hat of are natural filaments indicates that they substantially less stable. This suggests that some of the numerous non-mvosin proteins of nat.ural thick filaments have a stabilizing role. If this is the case it might, at first sight. seem surprising that t.hr stabilities of synthetic filaments obtained from crude and column-purified rnyosins wet-v the same. CI~UdP mposin approximatel? preparations mderd contain the majority of these proteins in physiological proportions (Starr & Offer, 1971). It should be not,ed, however. t ha.t t,hrse proteins are locat,ed very preciseI)- in nat)ural filaments (Craig & Offer, 1976; Starr & Offer. 1983), wherras incorporation into synthetic filaments in vitro presumably occurs at random. In synthetic. filaments the local concentrations of any one of t,hese proteins would. consequent,]\;, most probabl? be much lower t’han in natural filaments. Moreover. part,s of some proteins. which are undeniably natural filaments, are solubilized only with great. difficulty, and are not contained in classical crude myosin preparations like those used in this work. In particular, this is t,he case for t,he protein making up the “ctnd-filament,s” (Trinick, 1981). I an indebted t,o Elissavet Kardami for help with the preparation of phosphorylated rabbit mvosin; the rabbit and rat rod preparations were generous g’ifts from her and
172
I. Pinset-Hkrstriim
from Lawrence B. Bugaisky. Ketty Schwartz and Robert (:. Whalen made mang helpful comment,s about t’he and Alfred Rut,herford kindly tried to manuscript. improve the English.
References Chiao, E’. C. & Harrington. W. F. (1979). Biochemistry, 18, 959-963. Craig. R. (1977). J. MO& Biol. 109, 69-81. Craig, R. & Offer, G. (1976). Proc. Roy. Sot. ser. B, 192, 451466. Flicker, I’., Wallimann. T. & Vibert, P. (1981). Biophys. J. 33, 279a. Harrington, W. J. & Himmelfarb. S. (1972). Biochemistry, 11, 2945-2952. Harrison, R. G., Lowey. S. & Cohen, C. (1971). J. Mol. Biol. 59, 531-535. Hinssen, H., D’Haese, J.. Small, J. V. 8r Sobieszek. A. (1978). J. liltrastruct. Res. 64, 282-302. Huxley, H. E. (1963). J. Mol. Biol. 7. 281-308. Huxley. H. E. & Brown, W. (1967). J. Mol. Biol. 30. 383-434. Kaminer, B. & Bell; A. L. (1966). ,1. ,UoZ. Biol. 20. 391-402. Kardami, E. & Gratzer, W. B. (1982). RiochPmistry. 21, 1186.-1191. Katsura, T. & Xoda. H. (1973). A4;ldvclrr.Hiophys. 5. 177--202. Kendrick-Jones. J.. (‘ande, W. Z., Tooth, P. J.. Smith. R. S. & Scholev. J. M. (1983). ,I. Mol. Riol. 165. 139-162. Luther. P. K., Munro, Y. M. & Syuirv. J. M. (1981). .J. Mol. nior. 151, 703~-730. Lymn. R. \n;. & Taylor. E. W’. (197 I). Kiotltenristry. 9. 2975~2983.
Maw. M. (‘. & Row?. A. .I. (1980). .2;rctlrrv (Lmdm). -cl2 II-C.
286.
McLachlan,
A. D. & Karn. J. (1982). ,Vature (London),
299.412-414.
McLachlan,
A. D. & Karn, ,J. (1983). .I. Mol. Biol. 164,
605-626.
Moos. (‘.. Off’er. G., Starr, R. & Bennett. P. (1975). .I. Mol. Biol. 97. l-9. Pepe. F. .A. & Dowben. I-‘. (1977). ,i. Mol. Rinl. 113. 199-218. Perrie. W. T. & Perry. S. V. (1970). Biochenr. 1. 119, 31-38. Pinset-Hlrstrtim. T. &. Truf@. ?J. (1979). J. Mol. Biol. 134. 173 188. Pinset-HLrstrGm. I. 8 Whalen. R. G. (1979). 1. ;tfol. Hiol. 134. 189-197. Reisler. E. & Liu, J. (1982). J. Mol. Biol. 157, 659-669. Reisler. E.. Smith: C. & Segan. G. (1980). J. ~Ifol. Biol. 143. 129~-145. Scholeg. J. M.. Taylor, K. A. & Kendrick-*Jones, J. (1980). Nuture, (London), 287, 233-235. Starr. IA. $ Offer, Q. (1971). FEBB Letters, 15. 4W4. Starr, R. $ Offer, G. (1983). 3. Illol. Biol. 170. 67%698. Stewart. 31,. Ashton. F. T., Lieberson, R’. 8r Pepe. F. A. (1981). .J. Mol. Biol. 153, 381-392. Sutoh. K., (‘hiao. Y. C. & Harrington. W. F. (1978). Biochemistry, 17. 1234-1239. Suziki. H., Onishi. H.. Takahashi, K. CyrWatanabe, S. (1978). .J. Riochem. 84, 1529-1542. Trinick. *I. A. (1981). .I. Mol. Biol. 170. 675-698. Trinick, .I A. &r Elliott. .A. (1979). .I. Mol. Biol. 131. I 33 136. Steno. H. KL Harrington, \1:. F. (1981). J. Izlol. Biol. 149. 619 -640. Weeds. A. & Pope. K. (1977). J. Nol. Hiol. 111. 129-157. Whalen. R. (:., Butler-Browne. G. S. & Gras. F. (1978). .I. Mol. Hiol. 126. 415-431. Whalen. K. G.. Sell. S. 31.. Butler-Brownr, G. S.. Schwvart.z. K.. Bouveret. I’. $ Pinset-HlrstrCim, T. ( I981 ). Xatwe (Londonj, 292. 805-809. Wray ,I. S. (1979). Xatuw (London). 277. 37 -40.
Edited by H. E. Huxley
MgATP Specifically Controls in C/l’tro Self-assembly of Vertebrate Skeletal Myosin in the Physiological pH Range Ingrid Pinset-HPrstriim Service de Biophysique, Departement de Biologic Centre d’Etudes Nucle’aires de Saclay 91191 Gif sur Yvette, Cedex, France (Received 12 August
1983, and in revised form, 1X July
1984)
The appearances in the electron microscope of rat and rabbit skeletal muscle myosin filaments and rod aggregates, formed in the presence of variable amounts of MgATP. were compared at different pH values. It is shown that small amounts of MgATP, similar to t’hose sufficient to trigger the dissociation of the actomyosin complex, were able to modify the geometry of myosin filament’s profoundly in the physiological pH range, whereas the conformation of rod aggregates remained unchanged even in the presence of high concent’rat’ions of MgATP. Myosin filaments formed in the absence of MgATP displayed the classical spindle-shaped conformation and variable diameters at all pH values, whereas myosin filaments formed in the presence of MgATP in the physiological pH range ha,d constant diameters, similar to those of natural thick filaments. These filaments of constant diameter frayed, rapidly and reversibly, into two types of subfilaments with respective diameters of 4 to 5 nm and 9 to 10 nm. when the pH of the medium was raised above 7.2. Spindle-shaped myosin filaments and rod aggregates remained unchanged by such small changes in pH. It was possible to change the conformation of preformed spindle-shaped filaments by simply adding MgATP to the medium. but t’his reaction was slow and t’ook several hours to be completed. Relatively high concentrations of MgATP, similar to t,hose in the living cell, increased the solubility of both myosin filament’s and rod aggregates in the alkaline pH range (pH 2 7.0). Low pH values ( I 6.5) and excess free Mg’ + ( 2 6 to 7 mM) abolished both the specific effect of MgATP on myosin filament conformation and it,s solubilizing effect on both myosin filaments and rod aggregates. The degree of purity of the myosin preparations and the level of phosphorylation of the LC-2 light chains did not influence filament behaviour noticeably and rat and rabbit myosins behaved similarl,v.
1. Introduction Natural thick filaments of vertebrat’e skeletal muscle are remarkably conservative structures. They are invariably 1.6 pm long and 15 to 18 nm wide and display a bipolar conformation (Huxley, 1963; see Fig. l(a)). The myosin molecules, which are the major constituents, lie antiparallel in the middle of the filaments, forming a 150 nm wide central bare zone (Huxley, 1963: Craig & Offer. 1976; Craig, 1977). This invariance is in contrast to the wide spectrum of sizes and structures displayed by synthetic filament,s produced by spontaneous aggregation of myosin molecules in vitro. This is largely due to t’he fact that myosin self-assembly is extremely sensitive, not’ only to small changes in the pH and ionic strength of the medium (Kaminer & Bell, 1966), and to t’he nature of the ionic species (Harrison et al.. 1971; Reisler et al.. 1980), but also
t,o purely kinetic factors, such as the polymrrization rate (Huxley. 1963). In addition, with few exceptions, synthetic filament populations are very heterogeneous (Kaminer & Bell, 1966) or display obviously non-physiological sizes (Moos rt al., 1975; Reisler rt al., 1980). Finally, synthetic mposin filaments do not consistently display cenbral bare zones (Hinnsen et al., 1978), but when they do these are usually not’iceably wider than those of natural thick filaments (Huxley, 1963). It has been suggested that’ the structure of natural thick filaments might be determined, not only by the intrinsic characteristics of the myosin molecules themselves, but in addition by some non-myosin components (Huxley & Brown, 1967; Moos et al., 1975). Natural filaments indeed contain a great number of non-myosin proteins (Starr bz Offer, 197 1). some of which are located very precisely (Craig bz Offer, 1976: Starr & Offer, 1983). However,
it has not been Jtossible t#odemonstjratje whether any one of these has a conformat~ion-determining role. On thr other hand. we have shown (PinsetHiirstriim & Truff~. 1979) t.hat. synt,hetic filamentis displa>,ittg some of the key features of nat,ural thick filatnentjs. for example, constant physiological cliameters and 150 nm wide central bare zones. ma) form from solul)ilized rat skeletal muscle tnyosin. Itroviderl MgATJ’ is present (luring filament format’ion. Tt thus seemed possible that the particQatior1 of MgATP might be necessary for the cvrrect positioning of the myosin molecules in the filament backbone. This hypothesis seems to be supported by the parallel behaviour of smoot,h trtuwlt~ and non-muscle
myosins
in the presence
of
MgATP. The conformation of these latter rnyosins (*ittt indeed be changed by interaction with MgA4TP. attd i his conformat,ional change may modify filament fhrma,tion irt &YJ (Suziki rf (I/., 1978: Scholey rt nl.. 19X0: Hendrick-Jones rt al.. 1983). Skeletal tnyosin has at least two types of specific MgATP ittteravtion sites: first. the high-af?inity sites loc*ated in t,he heads. which are responsible for ATT h!~drolysis and trigger the dissociation of the it~~tOttl~VOSitl complex (Lymn & Taylor, lY71): swontl, several low-affinity sibes located alortp the rot1 segment. which might be implicated in filament c.ohesion (Harrington & Himmetfarb. 1971 ). (‘0~ secluetttl~\-. it swtnetl of irtt,ercst t.0 try to assess \vlticsh sites were responsible for the effect of MgATT’
OII the vonformation of synthetic skeletal myosin filaments. anti this was the aim of the present work. To this entl 1 cvtnpared the behaviour of myositt filatnc-tits and roti aggregates formed in the prrsrnce 01’ various amounts of MgATP at tlifferent pH salllrs. Since ltttosphor?;lat,ion of the IA’-2 li$tt c.ttain of’stnooth tnuscle rnyosin. \vhich is putativeI> Ioc*ated at the ,jnnc%ion between the heads and the rocl (Flicker rt trl.. 1981: t(endrick-Jones P/ ~1..
l!A%). niotlulates the depentlenee of in rsitro tilament fi)rtttation on MgATT’. the behaviour of variably l)hoSl)horvlated pa.re(l. 1,nstly. iti prwious studiw r;tt artd rabbit
under strictly
myosin
preparations
were conu Ot’li t 0 of skeletal ttt?-osin wlf-aswtttl)ly, skeletal myosins were c30tnparetJ
ortier to relate tliC> l)rrsrttt
itlentical 2. Materials
cxperimenial
c~otttiit~iotts.
and Methods
(‘rude myosin was prepared front the hind lep n~usc~lrs of rat. and the Irg and back muscles of rabbit. as described previously (Pinset-HgrstrBm & Truffy. l!C!)). The major pa,rt of the artin contamination of these preparations was removed eit’her by ult~racrntrif~tgatiotI at the maximum speed (about 130.000 g) in the Beckman Airfuge for 30 min in the presence of 04 M-KU at pH 6.8. adjusted with 20 mM-Tris-maleate or by means of a.mmonium sulphate precipitation. wtaining the 1.5 t,o .iP, cut. In some instancrs further purificaation was carried out by means of column chromatography on Sepharose 2B in the presence of XTP (M’halert rt al.. 1981). As judged by gel electrophoresis in the presence of 8 M-urea (Perrie & Perry. l!??(j). t,hr rat myosins prepared
by these twhniyues were ntainly nnphosphor?-l~lt~~~l. a11d attempts to carry out subsequent l)hosl)hor?-latiorr t, different published techniques w-ew unsucwssful in rn,v hands. The rabbit, myosinn prepared 1)~ the sanlv techniques. on the other hand. displayed variable phosphorylation. Essentially unphosphor?-lated at~tl up t-c, about 8Oo, phosphorylatetl preparations \vet’e obta,inrcl by the techniques described hg Kardatnt & (iratzrr (1982). .Ul preparations were c-arrird out at 4 (’ and all media cw~t.aitied 2 m.n-tlithiothrritol and (1.1 mwsotlium azide in order to try to avoid osidation of the sulphydryl groups and proteolqsis. ~everthelesn. sinw it has IMYII shown previously (Pmaet-HGstrtim 8 \Vhalen. I!)i!)) that the I,(‘-” light chain is particwlarl~sfbtlsitivc> to endogenous proteasrs. anti that t’s<‘11 \-csr,v limittscl havr far wac~hin~ proteolysis of this light c,hain ma!. effects on tilatnent a.ssetttt)l>-. its integrity \+as systrttrati cxlly Arcked by means of iso-rlec*tr.ic focwsirly (\Vl~;tlrtt r,/ 01.. 1WX). Rat and rabbit myostn rod subf’ra~tnents ww prepww~ from c~rudr m?osiits by c~hymotryl)tic~ diqMiot1 aworditip to the method of \Vretls & Pope (I 977).
(1,) Filrrnwnf ,fbrnlotion Fi1amrttt.s ww formed by drcwasin~ the KC0 concentration of the media from 0.5 \t t.0 0.12 11 I)) dilution at various rates under mapttetic stirring using constant-speed peristaltic l)umps, or hy dialysis ovflrnight. In the case of dilution. 1 ml of an initial solution was diluted t.o a final rolumt, of IO ml wit.lt a dilution solvent 048 M in KU This operation was wrrird out at room temperature. whereas the dialysis was c.arrietl out itt the cold room. \Vith thr rxccptio~t of thtl K(‘t wncetttration. the ionic, cwtttpositiorr of t hti initial solutions and the dilution solvents were itletttic*al. In particular. it should be noted that. the ATP ~ottc,rrtt.ratiott was the same in both media. C~onseyuetttly. filamtvit formation took place at cwnstant ATP c*otrc~entratioil. The pH of the medium was adjusted by means of 20 mwTrismaleate and all solvents c~oritained 2 rnht-tiithiothreitol and @I mwsodium azi&. Ot.htbr ionic cwmponrnts of thrl tnedia were variable and !+‘ill br described in tttv liesult,s. The protein concentration of the initial solutions varic~tl between I and 0.1 mg/rttl. The tinal prott~in c.otlc.f,tltratioit of the filament suspensiotw thus varicxd betv wti I ttig,“rnl (in some rxlwrimcnts c,ar,rircI out bs clialysis) anal 10 /4’gjral. In rlo (*use did T n0tic.e arty siqitieant itrfwric~t~ of t’he protein concentration on the results ohtainetl. It1 or&r to c,heck nhethrr it \ras possiblt~ to (.IIHI~~IC t,ltcs geomrtr~~ of preformecl myosin tilamrttts directly. without previous solubilization. follo~vt~d by tvag~re~ation in ii different tnrdium. t.htl ionic. c*omposition of tttv tilatnent suspensions uas sometitnrs c~hanged either I)y simpl> adding the required agent(s) rapidly under ma#nrtic, stirring or t)y dialysis orrrni~ht
The final filament suspensions were spread onto thin collodion-backed carbon films using eithri t1w cytochrome c monolayer sprradirq techniyur alread~~ described (Pinset-HLrstrStn 9r Truffy. 1979) or direct deposition onto carbon fillkts made hydrophilit~ by means of ultraviolet irradiation as tlesc*ribetl by Trirtick Ni Elliott (1’379). The results obtained using these 2 spreading bechniques wrc very sitnilar and no aignifieant difference in size or apparent struvtuw of the filament populations distributed over the qids ~wuld be idrntititd. X(,sttrthc,-
MgATP
Control of Myosin Assembly
less, as applied in this laboratory, the monolayer technique provided cleaner grids and more-even negative contrasting. It was consequently preferred in the majority of the experiments described here. After spreading the grids were rinsed with a small quantity of buffer and negatively contrasted using 2 to 3O/; freshly uranyl acetate. The collodion prepared. unbuffered backing was removed just before the introduction of the grids into the microscope by touching a filter paper soaked in isoamylacetate. The observations were carried out, by means of a Siemens Elmiskop 101. modified to allow operation at 120 kV under normal high-dose conditions at nominal magnifications of 10,000, 20,000 and 40,000 diameters. respectively. The real magnification was checked using negatively cvontrasted. unfixed catalase crystals. Filament sizes were measured either with a ruler on magnified prints or dire&l?, on the micrographs with a microcomparator.
3. Results (a) Experiments curried out in the r*bsence of MgATP These experiment’s were done in order to relate the present work to previous studies of skeletal myosin filament formation (Huxley, 1963; Kaminer & Bell, 1966; Hinssen et al., 1978) and to try to obtain a strict comparison between t’he behaviour of the different types of myosin preparation used in under classical experimental these experiments conditions. Filament formation was carried out at
161
pH 6.5 to 8.5 and either in the absence of MgCl, or in the presence of 10 m&i-MgCl,. Other medium conditions are described in Materials and Methods. The change in KC1 concentration w-as obtained either by dialysis or by dilution at, rat’es varying between 0.5 ml/min and 20 ml/min. The results obtained were entirely consistent’ with those of t)he authors mentioned above. and are illust’rated by Figures 2: 3, 4. 5 and 6 and Table 1. It follows from these results that, within the limits of the techniques applied, the origin of the myosin preparations, their degree of purity and the level of light chain phosphorylation had no significant influence on filament behaviour in the absence of MgATP. As shown in all previous studies. the short filaments formed at all pH values by rapid dilution. and at high pH values whatever the dilution rate. looked clearly bipolar, but displayed 1argelJ variable diameters. Filaments formed at lower pH values and at intermediate dilution rates were spindk-shaped. very inflated in the middle, with ends tapering over long distances. Filaments formed at’ these lower pH values, using sloa dilut’ion rates or, in particular, dialysis, displayed constant’ diameters, similar to t’hose of the inflated centres of spindle-shaped filaments; in other words. several-fold t’hicker t ban natural they were filaments. Rod preparations polymerized under t,he same experimental conditions as the myosin preparations. consistently gave rise to flat spindle-
(b)
Figure 1. Electron micrograph of (a) natural thick filaments prepared from rat leg muscle according t,o Trinick & Elliott (1979) and (b) synthetic filaments formed from crude rat skeletal myosin in the presence of 2 mM-ATP and 3 mM-M&l, at pH 7.0 at a dilution rate of 2 ml/min. The lengths of the 2 types of filament are clearly very different, yet their diameters and general morphological appearances are roughly the same, and strikingly dissimilar to those of synthetic filaments formed in the absence of MgATP. (Kaminer & Bell. 1966; Hinssen et al., 1978) (see also Fig. 13). Magnification: 35,000 x diameters. 6
I62
1. Pinset-Htirstrii,rn
20 ml/mln
t
5 ml/mln
Dialysis
IL
n
IL PH 7.5
I 0.1
DH 8.0
I 6.5
I 7.0
I 7.5
III 8.0
6.5
I 7.0
I 7.5
I 8.0
PH (a)
(b)
Figure 4. Mean lengths
Diameter
(nm)
Figure 2. Histograms
representSing the diamet’ers of filaments formed from crude rat myosin in the absence of MgATP at various pH values and dilut,ion rates. In order to try to measure the degree of spindle shape. the diameters were measured at 2 points: in the centre of the filament and midway between the centre and the tips. It c-an be seen that in the case of filaments formed at intermediate dilution rates the measured values cluster into 2 distinct blocks. corresponding roughly to the 2 measuring points. The short bipolar filaments formed at the highest pH value and the very long ones formed at slow filution rates (particularly by dialysis), on the ot’her hand. displayed approximately constant diameters. similar to those displayed by the spindle-shaped filaments in thp rentre
1
0.11
I
I
I
I
6.5
7.0
7.5
0.0
I
I
I
I
I
6.5
7.0
7.5
0.0
of filaments obtained from 3 distincat column-purified rat myosin preparations at various dilution rates and different pH values (a) in the absencr of MgATP and (b) in the presencr of 2 miv-ATP a’nd 3 mM-MgC1,. Symbols as in Fig. 3.
shaped aggregates of various sizes (Figs 7 and 8). The mean diameters of these aggregates changed with pH in a very similar way to those of myosin filaments (Fig. 8). but their overall sizes. and in particular their lengths, were much less influenced by the dilution rate than were myosin filaments. With the exception of the aggregates obtained at t’he most rapid dilution rates. they all displayed constant lengths of about I pm. Tn the ph,vsiologlcal pH range (6.5 to 7.5). howc,ver. t,hese aggregates showed a marked t’endency to cluster together on standing for longer periods (several hours or overnight), eventually forming “super-aggregat,es”. sometimes over 1 pm wide and several pm long. Both lateral and longitudinal growth thus seemed to be virtually unlimited. It’ should be noted, however, that even the longest) super-aggregates
-J
PH (a)
(b)
Figure 3. Mean lengths of filaments formed from 4 distinct crude rat myosin preparations at various dilution rates and different pH values (a) in the absence of MgATP and (b) in the presence of 2 mM-ATP and 3 mM-MgC1,. The dilution rates were as follows: (0, a), 20 ml/min; (0, n ) 10 ml/min; (a, A) 5 ml/min; (0. v) 2 ml/min; (0, 0) d’la ly sis. It can be seen that in every case the filament lengths depended strongly upon both the pH and the dilution rate, yet at each dilution rate the longest filaments were consistently obtained in the presence of MgATP and the physiological pH range.
I 6.5
I 7,O
I 7.5
I 0.0
I
I
I
I
6.5
7.0
7.5
0.0
PH (0)
Cb)
Figure 5. Mean lengths of filaments formed from 3 distinct non-phosphorylated, crude rabbit myosin preparations at various dilution rates and different pH values (a) in the absence of MgATP and (b) in the presence of 2 mM-ATP and 3 mM-MgCl,. Symbols as in Fig. 3.
MgATP
Control of Myosin Assembly -
-
163
Table 1 Mean diameters of jlaments formed at various dilution rates at different pH values from different types of myosin preparations in the absence of MgATP A. Crude, non-phosphorylated
rat myosin 5
ml/min Dialysis
20 ml/min ~ PH
6.5
7.0
7.5
0.0
I
I
I
6.5
7.0
7.5
I: 0-C )
.?b
Xa
Xl,1
i.0
24.5k4.0 21.1k2.7
35.2k10.5 44.8k4.0 32.0* 9.2 41.8k2.9
i.5 8.0
19.4k3.2 20.3+3.1
24.9+ 23.4*
6.5
I I I I
r
7.2 34.912.3 4.3
j:
24.6k4.0
44.Ok4.3
22.5k3.2 19.7k2.6
43.0f4.2 33.2k4.1 17.1+ 5.3
PH (a)
(b)
Figure 6. Mean lengths of filaments
formed from 3 distinct approx. 80% phosphorylated, crude rabbit myosin preparations at various dilution rates and different pH values (a) in the absence of MgATP and (b) in the presence of 2 mM-ATP and 3 mM-MgATP. Symbols as in Fig. 3.
never approached the lengths of the very long myosin filaments obtained by slow dilution or dialysis. Rat and rabbit rod preparations behaved very similarly and no influence of the presence of 10 mM-MgCl, on t*he aggregation was noticed.
carried out in the presence of 2 m&w-ATP and 3 WLM-A%fgC12
(b) Experiwbents
It has been shown (Pinset-HtirstrBm & Truffy, 1979) that the size and conformation of rat skeletal myosin filaments depend critically upon the relative concentration of ATP and MgCl,. Therefore, it was decided to carry out’ the first series of experiments in the presence of approximately physiological concentrations of these two agents, in order to explore the influence of pH on filament formation under these conditions, and to compare the behaviour of the different, myosin and myosin rod preparat,ions used in this work. To this end concentrations of 2 mM-ATP and 3 miw-MgCl, were chosen. Filaments were formed between pH 6.5 and 8.5, at dilution rates varying between 0.5 ml/min and 20 ml/min, as well as by dialysis. At pH 6.5 the filaments formed from rat and rabbit, purified and unpurified, phosphorylated and unphosphorylated myosins were virtually indistinguishable from those formed at corresponding dilution rates in the absence of MgATP (Figs 2, 3, 4, 5, 6 and 9; Table 2). From pH 6.8 and above, however, the situation changed very strikingly: all filaments formed at pH 6.8 and 7.0 displayed constant diameters of 15 to 18 nm, which were not modified by the dilution rate (Fig. 9; Table 2). With the noticeable exception of the lengths, the overall appearance of these filaments in the electron microscope was extremely similar to that of natural thick filaments (Fig. 1). These filaments were consistently longer than those formed at corre-
B. Column purijed,
6..5 7.0 7.5 84
23.6k4.2 20.4k4.2 21.3f5.0 19.5f4.5
non-phosphorylated
34.0+ 9.5 36.2k10.5 26.0& 9.5 21.5k10.5
C. (‘rude. non-phosphorylated
6.5 7.0 7.5
25.6f4.5 24.0+5.0 20.3f4.5
37.5+ 33.5+ 24++
8.0
18.9k4.5
20.6&
D. Prude, approximately
rat myosin
42.Ek4.3 45.Ok5.5 35.OL5.0
25.6f5.0 24.6f4.5 20.5f5.0
46.Ok4.3 44.5k4.6 32.6k4.7 245k5.0
22.7k5.5 25.2k4.7
47.5k5.5 44.5k4.5
rabbit myosin
10.5 5.5 4.8
44.5k4.5 42.5f5.0 33.7145
19.5Ifr4.5
4.8
31.5f5.0 20.514.7
SO0/Ophosphorylated rabbit myosin
5 mllmin 20 ml/min PH
s
%t
xa
2,
Dialysis r
6.5 7.0 7.5
22.6k4.3
34.8k4.6
43.7i4.6
23.7k4.6
44.8+5.8
19.5k4.6
36.6k4.3
46.054.8
22.9k4.7
24+i+5.0
34.055.1
225k4.8 19.5k4.6
8.0
19.0+4.4
20.5k4.3
43.9k4.8 33.7i4.6 18.7 & 5.0
In order to try to get an approximate evaluation of the degree of spindle shape, the diameters were measured at 2 points, at the centre of the filament, (&), and midway between the rentre and the tips (2,). The diameters are expressed in nanometers. Each mean value (i) and its corresponding variance correspond to a minimum of 400 individual measurements. A. Mean values corresponding to 6 distinct preparations. B. Mean values corresponding to 6 distinct preparations. C”. Mean values corresponding to 3 distinct preparations. I). .Ilean values corresponding to 3 distinct preparations.
sponding dilut’ion rates in the absence of MgATP (Figs 3, 4, 5 and 6). At pH 7.2 these filaments of constant diameter were replaced by markedly thinner ones, displaying various diameters, with coexisting some apparently unchanged filaments of constant diameter (Figs 9. 10 and 11). These thinner filaments were slightly shorter than t’he filaments of constant diameter formed at
164
I. Pinset-Hiirstriim
Figure 7. Elect,ron micrographs of aggregates pH 7-O at a dilution rate of 2 miimin. M&$&ation:
formed from rabbit, rod preparations in the absence (a) 35,000 x diameters(bj 140,000 x diameters. 20 ml/mln
corresponding dilution rates. Since the thinner filaments were easily breakable, this length difference might, however, be only apparent. The diameters of these thin filaments were rather variable, but the histograms of Figure 9 show that,, on the whole, they gather into two distinct groups, with apparent diameters of 4 to 5 nm and 9 t,o 10 nm, plus a supplementary group, presumably representing undissociated filaments of constant diameter. As can be seen from the micrographs of Figures 10, and 11, some of these latter, and even some of the thicker “sub-filaments”, split, up, mostly symmetrically about the centre of the filament’, whereas others only seemed to slim down gradually. At pH 7.5 the mean filament lengths reduced and the filament were substantially
5 ml/mln
of MgATP L
at
Dialysis
pH 6-5
pH 6.8
pH 7.0
>H 7-2
3H 7.5
0
20
40 Diameter
Figure
1
I 6.5
I 7.5
I 7-o
I 8.0
I
PH
Figure 8. Mean diameters
of aggregates formed from 2 distinct rabbit rod preparations at a dilution rate of 2 ml/min and various pH values in the absence of MgATP (O), and in the presence of 2 mM-ATP and 3 m&i-MgCI, (0).
9. Histograms
(nm)
representing the diameters of filaments formed from crude rat myosin in the presence of 2 mix-ATP at 3 mix-Mgcl, at various pH values and dilution rates. It can be seen that at pH 6.5 filament behaviour was similar to that displayed in the absence of MgATP (see Fig. 2). At pH 6.8 and 7.0, on the other hand, the diameters were uniform and independent of the dilution rate. At pH 7.2 and 7.5 distinctly thinner filaments with various diameters were formed, and above pH 8.0 no filaments at all could be distinguished.
MgATP
Control of Myosin Assembly
Table 2 Mean diameters of jlaments formed at various dilution rates and different pH values from different types of myosin preparations in the presence of 2 mM-ATP and 3 mM-MgCl, A. Crude, non-phosphorylated
rat myosin 5 mllmin
PH
20 ml/min ~~ -..-~~-s Xl01
6.5 6.8 7.0 7.2 7.5
P15k3.0 17,.5+_2.7 16.0&%2 lO.Sf6.0 9.0+3.2
B. C‘olumn-pwijed,
36.1klO.l 45.7f3.5 lS.l+_ 2.3 15.8_f 2.3 10.4k 5.6 9.5k 3.3
PH 6.5 6.8 7.0 7.2 7.5
2O.Qk3.0 18.3f2.8 15652.3 95k5.3 9.S + 3.0
36.9kQ.Q l&9+2.3 lSQf2.3 11.6+56 Q.Q_f3.1
C. Prude, non-phosphorylatrd
20.9*3.1 16Gjf2.9 16.4k2.2 11.3k6.2 9.it3.4
26.3k3.5
Dialysis f 44.Ok3.5 17.852.2 16.152.4 11.2k6.0 9.2k3.2
non-phosphorykxted rat myodin
5 ml/min 2() fnl/min ~~ -~ ~- - -~~.__~. s GJt ra
6.5 6.8 7.0 7.2 7-5
cl
%
44.8k3.8
CJ
Dialysis z
25.lk3.4
44.2k3.6 18.3f2.5 15.6f2.6 13CZ+6.6 10*3+3.5
25.7k3.3
44.Ok3.6 17.Ok2.4 lS.Ok 2.2 11.2k5.8 9.2k3.6
rabbit myosin
34.8T10.7 43.8*3,4 17.2+ 2.5 15.7* 2.4 lO.Q& 5.9 9.4k 3.6
1). Ckuie XOObphosphorylated rabbit myosin 5 ml/min PH
20 mlimin s
6.5 6.8 7.0 7.2 7.5
22.0 k 3.3 17.6k2.3 16.1k2.4 11.4k6.1 Q.Ok3.8
~J& 35.2 k lS.l+ 15.8+ 10.8+ 9,2*
X8
Xb
Dialysis f
10.4 46.0* 3.2 24.7 + 3.7 44.7 + 3.7 2.5 16.9* 2.3 2.2 15.9* 2.3 5.9 11.6k6.2 4.0 9.3k4.0
Measurements as indicated in A. Mean values corresponding IS. Mean values corresponding C. Mean values corresponding D. Mean values corresponding
Table 1. to 6 different preparations. to 6 different preparations. to 3 different preparations. to 3 different preparations.
diameters did not exceed 12 nm (Figs 2, 3, 4, 5 and 9). The micrograph of Figure 12 also shows that these short, thin filaments coexisted with substantial amounts of solubilized myosin molecules. At pH values of 8.0 and higher, only solubilized myosin molecules, and no filamentous structures, were distinguishable. Consequently, the presence of MgATP not only modified the filament conformation in the physiological pH range, but markedly
165
increased the actual solubility at pH values of 7.5 and above. By increasing the concentration of MgCl, from 3 InM to 10 mM, while maintaining the ATP constant, the specific effect of MgATP on myosin filament geometry was abolished (Fig. 13) and the filament solubility at high pH values decreased. The situation displayed in the absence of MgATP was thus restored. All the different myosin preparations behaved similarly, and I was unable to distinguish any influence of the origin, t’he degree of purity or light chain phosphorylation on myosin filament behaviour in the presence of 2 rnM-ATP and 3 m&r-MgCl,. As mentioned in Materials and Methods. in these experiments filament formation took place in the actual presence of MgATP. In order t,o see whether it was possible to change the filament geometry by simply adding MgATP to pre-formed filaments, 2 mM-ATP was added rapidly under magnetic stirring, or slowly by dialysis overnight, to filament suspensions formed in the presence of 3 mM-MgCl,. but in the absence of ATP. When this experiment was carried out by means of dialysis at pH 6.8 OI 7.0, the transformation of the classical spindleshaped filaments into substantially longer ones with constant diameters of 15 to 18 nm was complete. Adding ATP rapidly, on the other hand, did not have any immediate effect but, slowly. the transformation nevertheless took place, and was finally accomplished after four to five hours. Adding excess MgCl, to filaments pre-formed in t,he presence of 2 mM-ATP and 3 mM-MgCl,, did not have an immediate effect either but, gradually, the filaments became thicker and shorter, displaying the characteristic spindle-shaped conformaiion. Changing the pH from 6.8 or 7.0 to 7.2. on t,hr ot’her hand. by simply adding a small quantity of Trismaleate under magnetic stirring, immediately dissociated the constant diameter filaments formed in the presence of 2 mM-ATP and 3 mM-Mg(?l,. into subfilaments, and the reverse reaction was as rapid and apparent,ly complete. Raising the pH to 7.5 instantaneously partly dissolved the filaments formed at lower pH values in the presence of MgATP. Lowering the pH again, however, only produced relatively short filaments, which nevertheless displayed t*he same diameters as the initial ones. The overall conformation of the rod aggregates was not significantly influenced by the presence of 2 mM-ATP and 3 m&r-MgCl, at any pH value (Fig. 8). In the physiological pH range a,nd below the size of the aggregates was not significantly influenced either, but clustering together and fusing into super-aggregates did not occur even on standing for days. The mean diamet#er of the aggregates, however, decreased slightly at pH 7.5 and at values higher than 8.0 no aggregation was visible. The solubility of the rod aggregates was consequently increased at high pH values in the presence of 2 mM-ATP and 3 mM-MgClz. very much like that of myosin filaments. Increasing the concentration of MgCl, to 10 mM abolished this
166
I. Pinset-Hiirstriim
Figure 10. Micrographs of filaments formed at pH 7.2 at’ a dilution rate of 2 mljmin from column-purified, phosphorylated rat myo$in in the presence of 2 rnM-ATl’ and 3 mnl-MgCl,. Magnification: 35,000 x diameters.
solubilizing effect of MgATP. The presence of excess MgCl, also promoted fusing into super-aggregates in the physiological pH range. No difference was noticed between t,he behaviour of rat and rabbit, rod preparations. (c) Experiments carried out at various concentrations of MgA T P Since it’ has been shown (Pinset,-HSrst’rGm &, Truffy, 1979). t’hat free ATP may have a disorganizing effect on myosin filaments. it was decided to maintain the concentration of MgCI, at’ a consistently slightly higher value than that of the added ATP. Tn fact, a fixed ratio of t’wo thirds was maintained throughout these experiments. This means, for example. that’ when the ATP concentration was 0.1 m;M that of MgCl, was 0.15 mM. Nevertheless, for simplicity in t’he text. when referring tjo MgATP only the corresponding concentration of ATP is given. As described in Materials and Methods, the MgATP concentration was the same in the initial solutions and the dilution solvent; in ot’her words. ATP was continuously renewed during filament formation. Nevertheless, ATP hydrolysis inevitably occurred during the dilution and the ensuing spreading. To try to limit unknown effects of t,his. the experimental protocol was strictly standardized. The myosin concentration was maint’ained constant and low, at, 10 &ml, in the final filament suspension, and spreading was carried out according to a fixed time scheme in less than five minutes after dilution had finished. The indicated MgATP concentrations consequent’ly represent
non-
maximum values. which were presumably rather close t,o the real MgATP concentrations at high nominal values, but might represent substantial overest,imations at low nominal concentrations. These experiments were limited to crude. not column-purified. rabbit myosin. with either virtually or unphosphorylated Ll p t.0 80?,(, phosphorylated IL-2 light, chains. and to rabbit myosin rod preparations. Filament format,ion was obtained by dilution at a constant rate of 2 ml/min at pH 6.5 to 7.5. In other cvords. the KU concentration was changed from 0.5 M t,o 0.18 M in five minutes. The results from t,hese experiments are summarized in Figures 14, 15. 16. and 17. Tt is obvious that phosphorylated and unphosphorylated rabbit myosins behaved similarly, and t,hat no significant’ difference between t,he electron rnicroscopic appearance of the two types of filament was apparent at any pH value or MgATP ctoncentration. Another outstanding observation was t,hat the critical concentration of MgATP, necessary t)o modify the myosin filament conformation, was dramatica1l.y. pH-dependent. Indeed, at pH 6.5 no direct,ly rlslble effect on the overAl filament conformation was brought. about even by up t’o 10 mM-MgATP; only a very slight decrease in the mean filament diameter was noticed at ,MgATP concent,rations over 3 mM (Figs 14 and 15). At pH 6.8 a gradual decrease in the mean filament width, accompanied by a parallel increase in the mean lengths, occurred over a wide range of MgATP concentrations between 0.1 mM and 2 mM. At MgATP concentrations above 2 rnM the filaments displayed constant diameters of 15 to 18 nm and
MgATP
Control of Myosin Assembly
Figure 11. Micrograph of filaments formed at pH 7.2 at a dilution rate of 2 ml/min from column-purified phosphorylated rabbit myosin in the presence of 2 mM-ATP and 3 mM-MgCI,. Magnification: 35,000 x diameters.
obviously increased lengths. At, pH 7.0 a similar transition occurred in a narrower and at a much lower range of MgATP concentration. The midpoint of this transition was at about 0.02 mM-MgATP. As mentioned above, this value was presumably overestimated. This assumption was corroborated by the following observation: the preparations used in these experiments were very crude, and most of them contained noticeable amounts of actin, easily identified in the electron microscope images. At
167
809;)
nominal MgATP concentrations below 0.02 mM these actin filaments were clearly decorated with myosin: e.g. the actin-myosin complex was not dissociated. Whatever the actual MgATP caoncentration needed to trigger the substitution of spindleshaped filaments for filaments displaying constant physiological diameters, it thus seemed t,o be the same as the concentration sufficient to dissociate the actin-myosin complex. At’ pH 7.2 and 7-5 the spindle-shaped filaments were replaced by the t,hin
1fit3
1. Pinset-H&striim
Figure 12. Micrograph of filaments formed at pH 7.5 at a dilution rate of 2 ml/min from column-purified in the presence of 2 mM-ATP and 3 mM-MgCl,. Magnification: 35.000 x diameters.
subfilaments, low range
already described, in this same very
of MgATP
concentration. Whatever of MgATP, and whatever
actual concentration pH, these specific structural myosin filament conformation
the the
modificat,ions of t.he were abolished when
(0)
rat myosin
the concentration of MgCI, was raised to 10 rnM in excess over that of MgATP. As to the myosin rod aggregates, no significant modification of the overall conformation was noticed at any pH value and at. MgATP concen-
(b)
Figure 13. Micrographs of filaments formed at a dilution rate of 5 ml/min at. (a) pH 7.0 and (b) 8.0 from crude rat myosin in the presence of 2 mM-ATP and 10 mM-MgCl,. It can be seen that in both rases the filaments displayed t.he characteristic behaviour of filaments formed in the absence of MgATP. Nevertheless. MgATP is obviously present in the medium, since undecorated actin filaments are clearly distinguishable in the background. Magnification: 35,000 x diameters.
MgATP
169
Control of Myosin Assembly
r
IIO
0.01
0.1
I-0
IO.0
Mg ATP (RIM)
Figure 14. Mean diameters of filaments formed at
Mg ATP CrnM) a
dilution rat,e of 2 ml/min at different pH values, from 3 distinct non-phosphorylated crude rabbit myosin preparations as a function of the MgATP concentration. Symbols for Figs 14 to 18: (0, 0) PH 6.5; (0, n ) PH 6% (A, A) pH 7.0: (V. V) pH 7.2; (0, e, pH 7.5.
trations up to 15 mM. Only a slight gradual decrease in the mean diamet’er of the aggregates was noticed at and above pH 7.0 when the MgATP concentrat.ion exceeded 5 mM (Fig. 18). Nevertheless, the thinner aggregates formed at a given, lower pH (e.g. 7-O) in the presence of high amounts of MgATP (e.g. 10 mM MgATP) were indistinguishable from those formed at higher pH (e.g. 8.5) in the total absence of MgATP (Fig. 19). Raising the pH and adding thus had high c>oncentrations of MgATP qualitatively similar effects, as judged by electron
Figure 16. Mean lengths of filaments
formed at a dilution rate of 2 ml/min at different pH values, from 3 myosin rabbit non-phosphorylated crude distinct preparations as a function of the concenkation of MgATP. Symbols as in Fig. 14.
microscopy. Here again, the presence of high amounts of excess free MgC12 abolished the effect of MgATP. 4. Discussion (a) MgATP
and jilament formaticm
The results
show that low concentrations of to those that are sufficient to trigger the dissociation of the actomyosin complex. are able to transform profoundly the geometry of MgATP,
similar
40
r -5 5 E B
30
20
I( I-
-
Mg ATP CrnM)
Figure 15. Mean diameters of filaments formed at a dilution rate of 2 ml/min at different pH values, from 3 distinct, approx. SOS/, phosphorylated crude rabbit myosin preparations as a function of the concent>ration of MgATP. Symbols as in Fig. 14.
Mg ATP (ITIM)
Figure 17. Mean lengths of filaments formed at a dilution rate of 2 ml/min at different pH values, from 3 distinct approx. 800/b phosphorylated crude rabbit myosin preparations as a function of the concentration of MgATP, Symbols as in Fig. 14.
I. Pinset-Htirstr6m
170
160
0.01
0.1
I.0
IO.0
Mg ATP (mM)
Figure 18. Mean diameters of aggregates formed at a dilution rate of 2 ml/min at different pH values, from 2 distinct rabbit rod preparations as a function of the concentration of MgATP. Symbols as in Fig. 14.
synthetic filaments formed in t,he physiological pH range from all the different myosin preparat’ions st’udied here. whereas the conformation of rod aggregates was not noticeably influenced even by MgATP concentrations much greater than those of t,he living cell. This strongly suggests t’hat t’he ,l;lgATP interact’ion sites involved are the highaffinity sit,es located in the heads, and is to my knowledge the first indication that these sit’es are
(a I
implicated in skeletal muscle myosin filament assembly. In contrast to synthetic filaments formed in t)he absence of MgATP, which displayed the classical spindle-shaped conformation and various diameters (Kaminer & Bell, 1966; Hinssen et al.. 1978). those formed in its presence had constant diameters similar to those of natural filaments (Huxley, 1963) (Fig.1). Furthermore, t’he myosin filament,s formed in the presence of MgATP frayed, rapidly and reversibly, ink) two types of subfilaments wit’h diameters of I to 5 nm and 9 to 10 nm when the pH was raised t,o 7.2, whereas myosin filaments formed in the absence of MgATP. as well as rod aggregates, dissolved only gradually at) substantially higher pH values. This last observation strongly suggests that the assembly modes of myosin filaments formed in the presence a,nd in the absence of MgATP are fundamentally different. This conclusion is corroborated by t,he observntion that the direct transformation of one preformed filament t,ype int,o t,he other is a slow process. requiring several hours. The fact, that, this last phenomenon is so slow suggests that it may involve an actual reposit,ioning of the myosin molrculcs in the filament ba,ckbone. Such an event, cBertainly caannot take place during cross-bridge movement, but it might, represent a structural prerequisite for t’he proper functioning of the caontra&ile appa,ratus. Since the solubilities of bot’h myosin tilaments and rod aggregates were similarly influenced by relatively high concentrat,ions of MgATP. not very different from t.hose of the living cell. the rrsult>s are not cont’radictory to the
tb)
Figure 19. Micrographs representing rod aggregates formed at a dilution rate of 2 ml/min at pH 8.0 (a) in the absence of MgATP and (b) at pH 7.0 in the presence of 15 mM-MgATP. Magnification: 35,000 x diameters.
MgATP
Control of Myosin Assembly
suggestion of Harrington $ Himmelfarb (I971), that the low affinity MgATP interaction sites of the rod might modulate the overall cohesion of the myosin molecules in the filament backbone and hence, eventually, influence cross-bridge movement. The observation that the high-affinity MgATP interaction sites of the heads are implicated in filament assembly suggests that skeletal muscle myosins might be more analogous to smooth muscle and non-muscle myosins than previously realized. An important difference still persists between the two myosin types, since in the examples studied here LC-2 phosphorylation, which so dramatically modifies the behaviour of smooth muscle and nonmuscle myosin filaments (Kendrick-Jones et al., 1983), was without any noticeable effect on the conformation of skeletal muscle myosin filaments. However, it should be noted in this regard that electron microscopy, at least as it was applied here, is a very crude technique, which is certainly unable to reveal subtle structural differences at the level of the filament backbone or in the parts of the myosin molecules projecting from the filament surface. The fa,ct t’hat, slight conformational differences might exist between filaments formed from nonphosphorylated and 80% phosphorylated myosins, respectively, was indeed recently suggested by the work of Kardami & Gratzer (1982). These authors reported that suspensions of non-phosphorylated myosin filaments were substantially more turbid than suspensions of phosphorylated myosin filaments at a critical sodium chloride concentration of 0.16 M. The problem consequently deserves to be defined more precisely, using different approaches. The critical pH dependence of the MgATP effect on myosin filament conformation suggests that electrostatic interactions are of prime importance for filament assembly. This has already been predicted theoretically on t.he basis of sequence studies on nematode myosin (McLachlan & Karn, 1982, 1983) and demonstrated experimentally using selective proteolysis and chemical crosslinking st,udies on both natural and synthetic myosin filaments (Sutoh et al.. 1978; Chiao & Harringt,on, 1979; Ueno & Harrington, 1981) and rod aggregates (Reisler & Liu, 1982). Yet the present, results go beyond those sited above, since these latter studies consider only inberactions at the level of t,he rod, independently of any possible interference of MgATP. The present observation that lowering the pH and raising the concentration of free magnesium have qualitatively similar effects. e.g. that both abolished the specific MgATP effect on myosin filament conformation, might provide a plausible explanation to the mechanism of this effect. Indeed lowering the pH and raising the magnesium concentration are both liable to neutralize negative charges. which might be necessary for the correct positioning of the myosin molecules in the filament backbone. Consequently, it is possible that, interaction with MgATP at the level of the heads liberates negative charges at, some strategical point
171
on the surface of the myosin molecule, maybe at the junction between the heads and the rod, determining the specific conformation of natural thick filaments. (b) Nature of the subjilaments There is considerable evidence that natural thick filaments are built up from roughly parallel threadlike subunits, but there is no general agreement as to the nature of these subunits. Several authors have reported the existence of approximately 4 nm wide “microfilaments”, using electron microscopy (Katsura & Noda, 1973; Pepe & Dowben, 1977; Stewart et al.. 1981) and X-ray diffraction (Wray, 1979). Others have shown that when treated with distilled water isolated natural thick filaments may fray rapidly and reversibly into a maximum of three subfilaments displaying diameters at least double the size of the microfilaments (Maw & Rowe, 1980; Luther et al., 1981; Trinick, 1981). The present’ results show that. synthetic myosin filaments formed in the presence of MgATP can fray as rapidly and reversibly as natural ones, but that the subunits obtained display various diameters. Since, however, these subunits fall int,o two statistical blocks corresponding to mean diameters of 4 to 5 nm and 9 to 10 nm. respectively, it is tempting to associate the smaller subunits with the microfilaments of Katsura & Noda. and the larger subunits with the subfilaments of Maw & Rowe, which would mean that the t*wo are not incompatible. The observat,ion that fraying of synthetic filaments went far beyond t,hat of are natural filaments indicates that they substantially less stable. This suggests that some of the numerous non-mvosin proteins of nat.ural thick filaments have a stabilizing role. If this is the case it might, at first sight. seem surprising that t.hr stabilities of synthetic filaments obtained from crude and column-purified rnyosins wet-v the same. CI~UdP mposin approximatel? preparations mderd contain the majority of these proteins in physiological proportions (Starr & Offer, 1971). It should be not,ed, however. t ha.t t,hrse proteins are locat,ed very preciseI)- in nat)ural filaments (Craig & Offer, 1976; Starr & Offer. 1983), wherras incorporation into synthetic filaments in vitro presumably occurs at random. In synthetic. filaments the local concentrations of any one of t,hese proteins would. consequent,]\;, most probabl? be much lower t’han in natural filaments. Moreover. part,s of some proteins. which are undeniably natural filaments, are solubilized only with great. difficulty, and are not contained in classical crude myosin preparations like those used in this work. In particular, this is t,he case for t,he protein making up the “ctnd-filament,s” (Trinick, 1981). I an indebted t,o Elissavet Kardami for help with the preparation of phosphorylated rabbit mvosin; the rabbit and rat rod preparations were generous g’ifts from her and
172
I. Pinset-Hkrstriim
from Lawrence B. Bugaisky. Ketty Schwartz and Robert (:. Whalen made mang helpful comment,s about t’he and Alfred Rut,herford kindly tried to manuscript. improve the English.
References Chiao, E’. C. & Harrington. W. F. (1979). Biochemistry, 18, 959-963. Craig. R. (1977). J. MO& Biol. 109, 69-81. Craig, R. & Offer, G. (1976). Proc. Roy. Sot. ser. B, 192, 451466. Flicker, I’., Wallimann. T. & Vibert, P. (1981). Biophys. J. 33, 279a. Harrington, W. J. & Himmelfarb. S. (1972). Biochemistry, 11, 2945-2952. Harrison, R. G., Lowey. S. & Cohen, C. (1971). J. Mol. Biol. 59, 531-535. Hinssen, H., D’Haese, J.. Small, J. V. 8r Sobieszek. A. (1978). J. liltrastruct. Res. 64, 282-302. Huxley, H. E. (1963). J. Mol. Biol. 7. 281-308. Huxley. H. E. & Brown, W. (1967). J. Mol. Biol. 30. 383-434. Kaminer, B. & Bell; A. L. (1966). ,1. ,UoZ. Biol. 20. 391-402. Kardami, E. & Gratzer, W. B. (1982). RiochPmistry. 21, 1186.-1191. Katsura, T. & Xoda. H. (1973). A4;ldvclrr.Hiophys. 5. 177--202. Kendrick-Jones. J.. (‘ande, W. Z., Tooth, P. J.. Smith. R. S. & Scholev. J. M. (1983). ,I. Mol. Riol. 165. 139-162. Luther. P. K., Munro, Y. M. & Syuirv. J. M. (1981). .J. Mol. nior. 151, 703~-730. Lymn. R. \n;. & Taylor. E. W’. (197 I). Kiotltenristry. 9. 2975~2983.
Maw. M. (‘. & Row?. A. .I. (1980). .2;rctlrrv (Lmdm). -cl2 II-C.
286.
McLachlan,
A. D. & Karn. J. (1982). ,Vature (London),
299.412-414.
McLachlan,
A. D. & Karn, ,J. (1983). .I. Mol. Biol. 164,
605-626.
Moos. (‘.. Off’er. G., Starr, R. & Bennett. P. (1975). .I. Mol. Biol. 97. l-9. Pepe. F. .A. & Dowben. I-‘. (1977). ,i. Mol. Rinl. 113. 199-218. Perrie. W. T. & Perry. S. V. (1970). Biochenr. 1. 119, 31-38. Pinset-Hlrstrtim. T. &. Truf@. ?J. (1979). J. Mol. Biol. 134. 173 188. Pinset-HLrstrGm. I. 8 Whalen. R. G. (1979). 1. ;tfol. Hiol. 134. 189-197. Reisler. E. & Liu, J. (1982). J. Mol. Biol. 157, 659-669. Reisler. E.. Smith: C. & Segan. G. (1980). J. ~Ifol. Biol. 143. 129~-145. Scholeg. J. M.. Taylor, K. A. & Kendrick-*Jones, J. (1980). Nuture, (London), 287, 233-235. Starr. IA. $ Offer, Q. (1971). FEBB Letters, 15. 4W4. Starr, R. $ Offer, G. (1983). 3. Illol. Biol. 170. 67%698. Stewart. 31,. Ashton. F. T., Lieberson, R’. 8r Pepe. F. A. (1981). .J. Mol. Biol. 153, 381-392. Sutoh. K., (‘hiao. Y. C. & Harrington. W. F. (1978). Biochemistry, 17. 1234-1239. Suziki. H., Onishi. H.. Takahashi, K. CyrWatanabe, S. (1978). .J. Riochem. 84, 1529-1542. Trinick. *I. A. (1981). .I. Mol. Biol. 170. 675-698. Trinick, .I A. &r Elliott. .A. (1979). .I. Mol. Biol. 131. I 33 136. Steno. H. KL Harrington, \1:. F. (1981). J. Izlol. Biol. 149. 619 -640. Weeds. A. & Pope. K. (1977). J. Nol. Hiol. 111. 129-157. Whalen. R. (:., Butler-Browne. G. S. & Gras. F. (1978). .I. Mol. Hiol. 126. 415-431. Whalen. K. G.. Sell. S. 31.. Butler-Brownr, G. S.. Schwvart.z. K.. Bouveret. I’. $ Pinset-HlrstrCim, T. ( I981 ). Xatwe (Londonj, 292. 805-809. Wray ,I. S. (1979). Xatuw (London). 277. 37 -40.
Edited by H. E. Huxley