STUDIES ON ISOLATED SMOOTH MUSCLE CELLS: THE … · Isolated smooth muscle cells were prepared from guinea-pig taenia coli muscle in a manner similar to that described previously

J. Cell Sci. M, 327-349 (i977) 327Printed in Great Britain

STUDIES ON ISOLATED SMOOTH MUSCLE

CELLS: THE CONTRACTILE APPARATUS

J. V. SMALL

Institute of Molecular Biology, Aarhus University, 8000 Aarhus C, Denmark

SUMMARY

Smooth muscle cells may be isolated from the taenia coli muscle of the guinea pig which,when made permeable by treatment with Triton X-100 (005 %) show a sensitivity to Ca forcontraction with MgATP. The rate of contraction, about io / ims" 1 , corresponds closely tothe maximum velocity of shortening of the intact muscle. Electron microscopy of such partiallydemembranated muscle cells shows that myosin filaments of about 16-nm diameter are presentin both the rigor and the relaxed states. In addition, the actin and myosin filaments are com-monly seen to be associated in groups corresponding approximately in size to the fibrilsrecognizable in cells in rigor in the light microscope. The dense bodies and the 10-nm filamentsare found located between the actin-myosin filament groups. The thick myosin filaments maybe isolated by fragmentation of the cells under relaxing conditions. These native filamentsrange up to about 8 /*m in length and show the same structural organization as filamentsassembled from purified smooth muscle myosin : there is no central bare zone and bare edges,about 0 2 /im long, occur at the filament ends. The lack of bipolarity on the native smoothmuscle myosin filaments and the absence, in the contractile apparatus, of actin-associatedstructures equivalent to Z-lines suggests that the amount of shearing that can occur betweenthe actin and myosin filaments is considerably greater than in skeletal muscle.

INTRODUCTION

From the recent and numerous structural studies of vertebrate smooth muscle (seereview by Shoenberg & Needham, 1976) one fact emerges on which there may be saidto be general agreement, namely that the myosin component is rather labile. Thus, inthe electron microscope, sections from whole muscle processed in any of various wayscan exhibit considerable differences, even from cell to cell, in the number, form anddistribution of the myosin-containing filaments. And in contrast to the situation instriated muscle, myosin filaments of a size which would exclude in vitro assembly havenot been isolated directly from smooth muscle homogenates. This situation may belargely attributed to the solubility of smooth muscle myosin under appropriateconditions at low ionic strength (see review by Hamoir, 1973). From the dependenceof myosin solubility on such factors as pH and the concentrations of ATP and ofdivalent cations (Sobieszek & Bremel, 1975 ; Sobieszek & Small, 1976) it is clear thatthe conditions for the formation and maintenance of filaments in smooth muscle arerather different from those obtaining in skeletal muscle.

That myosin does exist in filament form in vertebrate smooth muscle has beenshown by the detection of a i4-4-nm meridional reflexion in X-ray diffraction patternsfrom living muscle (Lowy, Poulsen & Vibert, 1970; Lowy, Vibert, Haselgrove &

328 J. V. Small

Poulsen, 1973). But owing to the difficulties encountered in obtaining consistentresults in the electron microscope the question of the actual geometry of the myosin-containing filaments and of the organization of the contractile apparatus has remainedunsettled.

In studies of striated muscle the glycerinated fibre and myofibril as ' model systems'have played a key role in the elucidation of the detailed organization of the contractileapparatus of this muscle type (see e.g. Bendall, 1969, and review by Huxley, 1972).Similar systems have been produced by total or partial removal of the cell membraneseither by skinning or detergent treatment. In all cases the structural integrity of thecontractile apparatus is maintained and the contractile process may be readily con-trolled through modifications of the bathing medium.

In recent investigations of vertebrate smooth muscle it has been possible to isolatesmooth muscle cells which, after detergent treatment, may be induced to contractwith MgATP at a rate comparable to that of the living muscle. Furthermore, as ischaracteristic of the living system (see e.g. Van Breeman et al. 1973) and of smoothmuscle actomyosin (Sobieszek & Small, 1976) this contraction is dependent on thepresence of Ca2+ ions. These cells may thus be considered as model systems com-parable to striated muscle myofibrils. In an earlier study such cells yielded informationabout the gross organization of the contractile apparatus and provided evidence for achange in orientation of the myofilaments on shortening (Small, 1974). The presentreport describes first an improved method for preparing isolated smooth muscle cellsfrom the taenia coli muscle of the guinea pig. It is then shown, by the use of electronmicroscopy, that myosin filaments are present in cells fixed both in rigor and undercontrolled conditions of relaxation. In addition information about the length anddetailed structure of the myosin filaments was provided by their direct isolation fromthe cells.

MATERIALS AND METHODS

The isolation of smooth muscle cells

Isolated smooth muscle cells were prepared from guinea-pig taenia coli muscle in a mannersimilar to that described previously (Small, 1974). It was found important to carry out digestionin collagenase at about pH 65 to obtain cells capable of contraction and to include MgCls inthe bathing medium before and during collagenase treatment in order to consistently obtaina high yield of cells showing Ca-sensitivity. In addition, Xylocaine (see Shoenberg & Hasel-grove, 1974) was found to be a useful relaxant for keeping cells at an extended length duringcollagenase treatment.

Freshly dissected muscle strips from adult guinea pigs were transferred to a modified Ca-free Hanks' solution, 'solution 1' (in mM : NaCl, 137 ; KC1, 5 ; NajHPO4, 11 ; KH2PO4, 0 4 ;NaHCO3, 4 ; glucose, 55 ; MgCl8, 2 ; EGTA, 2 and streptomycin, 10 mg/1.) buffered addition-ally with PIPES, 5 mM and with the pH adjusted to 60 . (As indicated previously (Small,1974) EGTA served both as a relaxant, necessary to obtain cells at an extended length, and alsoas an effective pretreatment of the tissue facilitating subsequent cells dispersion with collagen-ase.) Again, the use of low pH proved imperative for obtaining a high yield (up to 100 %) ofresponsive cells. When the muscles had relaxed, after about 30 min, they were tied at extendedlength on to plastic plates and stored in test tubes in solution 1 for 2-24 h at 4°C. They werethen digested in a solution containing 1 mg/ml collagenase (Worthington CLS) of the followingcomposition: in mM, NaCl, 137; KC1, 5 ; NaHCO3, 4 ; glucose, 55 ; MgClj, 2 ; PIPES, io,

Contractile apparatus of smooth muscle cells 329

pH 6-5 and with the addition of either ATP (0-75 HIM) or Xylocaine (2 HIM) or both to maintainrelaxation. The digestion was carried out at 37 °C for approximately 1 h after which time themuscle strips became rather loose but remained mainly intact. The strips, still attached tothe plates were then transferred to solution 1 but now without MgCl2 added and containing0-5 mM dithiothreitol (DTT) and left, after 2 changes for 1-6 h at 4°C. The fragile musclestrips were then cut from the plates and the cells released in the same solution by passing thestrips repeatedly through a wide-bore Pasteur pipette. The undigested residue was sedimentedat very low speed and discarded and the cells in the supernatant collected by centrifugation atabout 200-500 g for 5-10 min. The cells were then resuspended in fresh medium, containingboth MgCl2 and DTT, and stored on ice before use. Cells were also prepared in the same waybut with MgClj omitted throughout. Such cells, while not routinely Ca-sensitive, proved to bemore suitable for the isolation of the myofilaments.

Various other muscles were also investigated, namely, the guinea-pig vas deferens, rabbittaenia coli, hog stomach, hog carotid and chicken gizzard, but the guinea-pig taenia coli gaveby far the highest proportion of extended intact cells with the above procedure (or slightmodifications of it) and was thus the main source of smooth muscle. With the vas deferens thecells in the external longitudinal layer shortened on the muscle surface during digestion incollagenase and this shortening could not be prevented by the use of conventional relaxants.However, while these cells were unsuitable for contraction experiments they could be used forthe isolation of myofilaments.

Light microscopy

The isolated cells were mounted on slides and covered with a coverglass having a thin layerof silicone grease along the 2 opposite edges parallel to the long axis of the slide. In this way themedium bathing the cells could be readily changed by application of the desired solution to oneedge and absorption with tissue paper at the other. All observations were carried out at roomtemperature using a Zeiss photomicroscope with either Normarski interference or phase-contrast optics. For contraction experiments images were recorded at regular intervals on 35-mm film or with a 16-mm movie camera.

Contraction and relaxation conditions

The media used for contraction and relaxation of isolated cells were based on those foundsuitable for assaying the ATPase activity of smooth muscle actomyosin (Sobieszek & Bremel,J975 J Sobieszek & Small, 1976). However, while a pH around 7-0 was adopted in the latterstudies it was found necessary to operate below about pH 67 to achieve a relaxation-contrac-tion cycle in the isolated cells (see also Results). The compositions of the relaxation and con-traction media were as follows : relaxation medium, KC1 80 mM, MgCl2 3 mM, EGTA 2 mM,ATP(Na) 2 mM, cysteine 1 mM, PIPES 10 mM, pH 6-5 at 20°C ; contraction medium, as for therelaxing medium but with EGTA replaced by CaCl2, 0-5 mM.

Electron microscopy

Section material. For electron microscopy of isolated cells the loose, collagenase-treatedmuscle strips were retained on the plastic plates and the final pipetting step omitted. Theparallel alignment of the cells was thus maintained while at the same time allowing direct con-tact of the cells with the external bathing medium (see Fig. 10, p. 333). Processing of the cellsin this way greatly facilitated subsequent thin sectioning and observation in the electronmicroscope.

For observation of the cells in the rigor state, that is in the absence of ATP, the digestedmuscle strips were treated with 003-005 % Triton X-100 in solution 1 plus DTT for 12-24 hat 4°C. They were then fixed with 2-5 % glutaraldehyde in the same medium (without Triton)for 2-3 h at 4°C, washed overnight, postfixed in 1 % OsO4, dehydrated in ethanol and em-bedded in Araldite.

For electron microscopy of cells in the relaxed state, conditions were used as defined by theresults obtained with isolated cells under the light microscope (see Results). Thus, the loose

330 J. V. Small

muscle strips were exposed to the relaxing medium (see previous section) containing 0-05 %Triton X-100 for 2-5 min at room temperature and then fixed in the same medium with added2-5 % glutaraldehyde and further processed as above. The same experiment was carried out at4°C, the relaxation time in this case being somewhat longer and in the range of 5—10 min.

Thin sections were cut with a diamond knife, on an LKB Ultramicrotome III, mounted onuncoated 300-mesh grids and stained with uranyl acetate (1 % in methanol-15 min at roomtemperature) followed by lead citrate (3-5 min).

Isolated filament preparations. Isolated cells were first resuspended in solution 1 (plus DTT)with added 0-2 % Triton X-100 for 10 min to 1 h at 4°C. This procedure yielded clean, demem-branated cells free of mitochondria and other contaminating components. The cells were thenwashed in solution 1 to remove Triton and subsequently in the relaxing medium lacking ATP.Fragmentation of the cells was then carried out in one of two ways. In the first method the cellswere first broken into smaller fragments, recognizable in the light microscope, by just a fewseconds homogenization at low speed in a Sorvall Omnimixer fitted with a micro-attachment.The cell pieces were then pelleted and resuspended in ice-cold relaxing medium (see above) byhomogenization in a glass-glass homogenizer or by the use of a fine-bore Pasteur pipette. Inthe second method, the washed and pelleted isolated cells were resuspended directly in therelaxing medium and homogenized in a glass-glass homogenizer. For the quantity of cellsobtained from the taenia coli of one guinea pig a volume of about 1 ml of relaxing medium wasemployed.

The cell homogenate was applied to carbon-coated grids and negatively stained in the coldwith 1 % uranyl acetate with the use of a 'protective agent' according to Huxley (Moore,Huxley & De Rosier, 1970).

While thick filaments were always observed in the presence of MgCl2 they commonly occur-red in rather large aggregates, unsuitable for studies of fine structure. More satisfactory isolatedfilament preparations were obtained using isolated cells prepared without the use of MgClj andwith ATP (075 mM) employed as a relaxant during collagenase digestion (see Results).

RESULTS

Contraction, relaxation and rigor in Triton-extracted isolated cells

If isolated taenia coli cells were first demembranated with Triton X-100 (forexample with o-i % Triton in solution 1 for 15 min), mounted in solution 1 and thenexposed to the contraction medium containing MgATP and Ca they contracted ratherslowly. The total time of shortening from their extended to fully contracted length,about 70 /im, was then in the order of 2-5 min. In sharp contrast, the exposure ofunextracted cells to Triton X-100 (0-05%) added together with the contractionmedium resulted in a very rapid and several-fold faster contraction, occurring within

Figs. 1—9. Contraction and relaxation in smooth muscle cells at room temperature.Normarski interference optics.

Figs. 1-4. Cells from guinea-pig taenia coli muscle at the beginning (Fig. 1) andafter exposure (Figs. 2-4) to a medium containing MgATP, Ca and Triton X-100(see text). Time sequence of frames corresponds to : o, 15, 25 and 45 s respectively,x 96.

Figs. 5-9. Taenia coli cells (Fig. 5) exposed first (Fig. 6) to a relaxing medium con-taining MgATP, EGTA and Triton X-100 (see text) and then (Figs. 7-9) to a similarmedium (without Triton) but with EGTA replaced by Ca. Time sequence of framescorresponds to : Fig. 5 ,0s ; Fig. 6, 3 min after exposure to relaxing medium; Fig. 7,beginning of effect of added Ca (as for Fig. 6 + 2 min) ; Fig. 8, as for Fig. 7 + 40 s ;Fig. 9, as for Fig. 7 + 80 s. x 80.

Contractile apparatus of smooth muscle cells 33*

\

332 J. V. Small

20-40 s. Without Triton X-100 no response was obtained. A typical result of thislatter type of experiment is illustrated in the sequence of frames shown in Figs. 1-4.The uniformity of the cell preparations was shown by the full contraction of almost100% of the cells observed under the coverslip and the same response was observedafter several days of storage, provided DTT was included in the storage medium (i.e.solution 1).

From time sequences such as those shown in Figs. 1-4 the average shortening rateof the cells was measured as about 10 /im s"1, a rate which compares closely with themaximum velocity of shortening of the living muscle. (Expressed as cell lengths persecond and for a mean rest length /„ = 200 fim, the velocity of shortening correspondsto 5 % IQ S - 1 as compared to the extrapolated value of 10% /„ s"1 given by Lowy &Mulvany, 1973.)

Figs. 5-9 show a similar experiment but in this case the cells, mounted in solution 1were first exposed to a relaxing medium containing MgATP and EGTA (see Methods)together with Triton X-100 (0-05%). Under these conditions, for which the Ca2 +

concentration was reduced to below io~9M, the cells were seen to maintain theirextended length for several minutes (Fig. 6) and became slightly smooth in contour.The maximum time of exposure to the relaxing medium was in the order of 5-7 min.The subsequent addition of Ca-containing contraction medium (as above but withoutTriton) caused the cells to contract (Figs. 7-9). Thus, the cells exhibited the samesensitivity to Ca as shown by the intact muscle and by smooth muscle actomyosin(see Introduction) and could, in the presence of EGTA and MgATP, be consideredto be in a relaxed state.

Treatment of cells with Triton X-100 (0-05-0-2% in solution 1) in the absence ofATP produces cells that resist distortion by liquid flow and which may be consideredto be in a state of rigor, brought about by the cross-Unking of the myosin and actinfilaments. In this state a fibrillar substructure may be recognized in the cells underthe light microscope (see Small, 1974 and Fig. 21, p. 340). In an earlier study (Small,1974) the cytoplasmic fibrils were interpreted as groups of myofilaments and werefurther noted to become more oblique with respect to the cell axis at progressivelyshorter cell lengths. As shown in Fig. 22 labelling of the cells with smooth musclemyosin S-i acts to increase the density of the fibrils in the light microscope. As wellas demonstrating the presence of actin in the fibrils this treatment also enhances theircriss-cross organization.

As noted earlier (Small, 1974) relaxation and contraction are accompanied bydissolution of the fibrillar arrangement, indicating a reorganization of the myofilamentsfrom the rigor to the relaxed and contracting states. This is illustrated most directlyby electron microscopy of cells fixed in rigor and relaxation as described in thefollowing section.

The filaments of the contractile apparatus

Fig. 10 shows a cross-section of a muscle digested with collagenase but not mechani-cally dispersed into single cells. In such preparations the cells in the outer part of


V.' 'V

10Fig. io. Cross-section of the outer region of a muscle digested with collagenase and at a

stage prior to normal dispersal into isolated cella-Xsee.tex.t). x 12000.

334 J. V. Small

Fig. I I . Taenia coli cell in rigor, io-nm filaments (circled) lie between the actin-myosin filament groups. The myosin filaments exhibit an approximately squareprofile in cross-section, x 80000.


Fig. 12. Taenia coli cell, fixed in and after i-min exposure at 4°C to a relaxing mediumcontaining MgATP, EGTA and 0-05 % Triton X-ico, pH 6-5. x 98000. Incubationfor s min gave essentially the same result, a, plasmalemma dense areas ; d, cyto-plasmic dense bodies ; m, myosin. Circle marks a group of io-nm filaments.

J. V. Small

* i

-.***•- • •«

Fig. 13. Taenia coli cell fixed in and after i-min exposure at room temperature to a re-laxing medium containing MgATP, EGTA and 005% Triton X-ioo, pH6-S- x 72C3OO.The same result was obtained for relaxation under the same conditions for 5 minat room temperature. Circle marks a dense body with associated io-nm filaments.


the muscle were well separated and were assumed to be in virtually as direct contactwith the bathing medium as completely isolated cells.

We have considered 2 states of the cells for electron microscopy, namely those ofrigor and relaxation. Fig. 11 shows the typical appearance of cells fixed in the rigorstate. After treatment with 0-05 % Triton X-100 in the bathing medium (solution 1)for 12-24 h the c e u s m t n e entire muscle exhibit the same general appearance. Inaccordance with the light-microscope observations of cytoplasmic fibrils in cells inrigor a grouping of myofilaments can be recognized in thin sections. The size andshape of the filament groups is, however, rather variable but is consistent with thenoted variability in fibril width in the light microscope. Of particular interest in thesecells is the clear exclusion of the io-nm filaments from the actin-myosin filamentgroups and also the cross-sectional shape of the thick filaments (Fig. 11). Instead ofbeing approximately circular in profile the myosin filaments appear roughly squarein shape, the average filament width being measured as 16-0+ i-o nm. In addition,the thin filaments are so arranged that not more than one or two rows separateneighbouring thick filaments with the common occurrence of rosettes of between 7and 9 thin filaments around each of the thick filaments.

Cells fixed under conditions of relaxation, that is by the addition of concentrated,buffered glutaraldehyde after 1-5 min in the Triton-containing relaxing mediumappear as shown in Fig. 12 (relaxation at 4°C) and Fig. 13 (relaxation at room tem-perature). Thick filaments are again present under these conditions but are morediffuse around their perimeter than in cells in rigor. In addition, and particularly forcells relaxed at room temperature (Fig. 13) a rough grouping of the thick filaments iscommonly observed, although these groups are not as clearly delineated as they arein rigor. Dense areas with associated io-nm filaments occur intermingled between thefilament groups.

Counts of the total numbers of thick and thin filaments revealed ratios of myosinto actin of about 1:10 and 1:13 respectively in cells fixed in rigor and relaxation.Owing to the difficulties associated with such counts this difference may not beconsidered as significant.

The native myosin filaments: isolation and structure

We have already seen that a state corresponding to relaxation may be induced inisolated and Triton-extracted cells in the presence of EGTA, Mg and ATP at aroundpH 6-5. Under these conditions fragmentation of the cells by gentle homogenizationresulted in the release of the thick and thin filaments that we have already identifiedin section material (Fig. 14). The occurrence of both thick and thin filaments inhomogenates of isolated cells contrasts markedly with the result obtained for wholemuscle for which, under the same conditions, only thin filaments were released (seealso Sobieszek & Small, 1973). Not unexpectedly, the thick filaments are moreaccessible in the isolated and demembranated cell. It should be noted, however, thatsuitably dispersed preparations of thick filaments such as shown in Fig. 14 wereobtained only from cells prepared without the use of Mg (see Methods). Homo-

22 CEL 24

338 J. V. Small

Fig. 14. Thick and thin filaments separated from isolated taenia coli cells by gentlehomogenization under relaxing conditions. The thick myosin filaments are of variablelength and show a periodicity of 140 run. x 24000.


genization of cells prepared with Mg included in the various media resulted in therelease of myosin filaments but these generally aggregated together into large bundleswithin which the details of the individual filaments were obscured.

The isolated, native myosin filaments were found to be rather variable in length.The majority of filaments were less than about 3 /MTL long but extremely long filamentswere also observed and ranged in length up to more than 8 /tm (Figs. 15-17). Whilethis length distribution could reflect a real variability of the myosin filament lengthin vivo it is also evident that some fragmentation of the filaments probably occursduring the isolation procedure. It is therefore not possible from these data alone toestablish a reliable estimate of the length distribution of the filaments in the intactcell.

The most important feature of the isolated filament preparations was the detailedstructure of the myosin filaments (Figs. 26-28). Comparison of these native filamentswith filaments assembled from purified smooth muscle myosin (Figs. 23-25; see alsoSobieszek, 1972, 1976, and Sobieszek & Small, 1973) indicated that they had thesame, or a very similar structural organization. This was indicated by the presenceof cross-bridges along the entire length of the filaments giving rise to a continuousrepeat period of about i4-o-nm, together with bare, cross-bridge-free edges approxi-mately 220 nm long at the filament ends. Such a close structural similarity be-tween the native thick filaments and filaments formed from purified smooth musclemyosin suggested rather strongly that the native filaments were likewise composedonly of myosin. Furthermore, their structural organization was clearly differentfrom that deduced for striated muscle thick filaments (see Discussion).

Optical diffraction patterns from the native filaments (Figs. 29-32) showed thesame type of variability as found for synthetic filaments (Sobieszek, 1972, 1976)and which likewise probably arises from distortions in the cross-bridge lattice thatoccur during negative staining. Nevertheless, a close correspondence was foundbetween the spacings of the reflexions from the native and the synthetic filaments.The more common occurrence in native filaments of reflexions at about 48-0 nm and28-0 nm gave only a better fit, with the other reflexions to a repeat period of 144-0 nm,rather than 72-0 nm as deduced for the synthetic filaments. However, and as illustratedby Sobieszek (1976) this longer repeat may easily be generated by a very slightdistortion of the lattice derived for the synthetic filaments.

A striking additional feature of the isolated filament preparations was the presenceof myosin filaments aggregated side-by-side and in such a way that the cross-bridgerepeat period of about 14-0 nm was always in register (Figs. 18-20). Synthetic myosinfilaments also show a similar tendency (Sobieszek, private communication). Thisobservation provided direct evidence for the ribbons earlier observed in sectionmaterial and which showed a continuous 14-nm repeat across their width (Small &Squire, 1972) being side-by-side aggregates of the native filaments described here.This aggregation of the native filaments was very likely favoured by the somewhatsquare cross-sectional shape of the filaments noted in cells in rigor (Fig. 11).


DISCUSSION

The form of myosin in vivo

The inability to demonstrate thick filaments at all in the early electron-microscopestudies of vertebrate smooth muscle (see review by Shoenberg & Needham, 1976)serves as a clear demonstration of the lability of the myosin component of this muscletype under conventional conditions of chemical fixation. In later studies, considerableemphasis has often been placed on the need to prepare muscles under as close to'physiological conditions' as possible (Somlyo, Devine, Somlyo & Rice, 1973 ; Jones,Somlyo & Somlyo, 1973) to 'preserve' the contractile apparatus. But in this empiricalapproach the well established fact has been conveniently ignored, that 'physiologicalconditions' of the extracellular environment are not relevant to those of the intra-cellular milieu under the non-physiological conditions of chemical fixation. Underthese conditions uncontrolled changes in the intracellular environment occur whichpreclude a meaningful evaluation of the 'state' in which the cytoplasmic componentswere fixed (see also Shoenberg, Goodford, Wolowyk & Wooton, 1973).

In consequence it was of particular importance, for further ultrastructural studies ofvertebrate smooth muscle, to develop a model contractile system for which the state ofthe contractile apparatus could be directly controlled via the external medium. Theisolated and partially demembranated smooth muscle cells described in the presentstudy constitute such a model system.

By demonstrating a relaxation-contraction cycle in these cells under the lightmicroscope we were able to provide evidence, independent of that from X-raydiffraction (Lowy et al. 1973), for the presence of thick filaments in the relaxedstate in vertebrate smooth muscle. If myosin were in a soluble form in this state itwould be lost from the cell and subsequent contraction would be prevented. At thesame time it was possible to process cells for electron microscopy under controlledconditions of relaxation and rigor and thus obtain information about the form of themyosin filaments and the general organization of the contractile apparatus.

It is perhaps appropriate here to make further comment on the inability to obtaincontraction following relaxation at pH's above about 6-75. This result could becorrelated directly with the increase in solubility of smooth muscle myosin, under theconditions used, on approaching neutral pH. Thus, we have shown elsewhere(Sobieszek & Small, 1976) that for homogenized gizzard muscle very little myosin isextracted around pH 6-5 in the presence of MgATP (2 mM) and EGTA, while

Figs. 15-17. Long native myosin filaments obtained from isolated cells. Lengths arerespectively 8-3, 5 and 4-5/zm. Fig. 15, x 20000 ; Fig. 16, x 32000 ; Fig. 17, x 35500.Figs. 18-20. Side-by-side aggregates of the native myosin filaments. The i4-4-nmperiodicity is seen to be exactly in register. Figs. 18, 20, x 115000 ; Fig. 19, x 93000.Figs. 21, 22. Isolated taenia coli cell in rigor (extracted with 0-2% Triton X-100)before (Fig. 21) and after (Fig. 22) incubation with smooth muscle myosin subfrag-ment 1. Smooth muscle myosin S-i was kindly supplied by Dr A. Sobieszek.x 800.

342 J. V. Small


substantial amounts are extracted at pH 7-0 under the same conditions. This merelyserves to underline the careful balance of conditions that must exist in the smoothmuscle cell, between divalent cation and ATP concentration, ionic strength and pHto maintain myosin in filamentous form. Contraction without prior relaxationoccurred even at pH 7-0 and above, and this would be expected from the effect of thepresence of Ca on myosin solubility. It is not difficult from these considerations toexplain the difficulties encountered to consistently preserve myosin filaments inelectron-microscope studies of whole muscle or, on the other hand, the early observa-tion of thick filaments in muscles glycerinated at pH 6-o (Kelly & Rice, 1968).

In an earlier study (Small & Squire, 1972) we chose to use the X-ray diffractionmethod to provide some control on the changes accompanying chemical fixation of theliving muscle. By virtue of the difficulties associated with obtaining X-ray patternsshowing the 14-4-1^1 reflexion characteristic of myosin (Lowy et al. 1973 ; Shoenberg& Haselgrove, 1974) these studies were carried out with muscles cooled to lowtemperatures and in hypertonic media, under which conditions the myosin reflexionwas most clearly seen. And for muscles that showed the myosin 14-4-11111 reflexionafter chemical fixation, thick filaments were indeed recognized in the electron micro-scope. But these filaments were commonly ribbon-like in shape and we consideredsuch ribbons at that time to be the in vivo form of myosin. The ribbons were up toseveral microns in length and showed a regular i4'0-nm periodicity. The majorcriticism of this latter approach was that the conditions used for the processing ofmuscles were unphysiological to the extent that they could lead to modifications in thecontractile apparatus in the living muscle.

The results of the present study show that myosin does not occur in the form ofribbons under the conditions of relaxation defined here for the demembranated cells.From the general absence of ribbons in cells in rigor, in which there is optimalinteraction between actin and myosin, we conclude that they are likewise absentduring contraction. The demonstration of ' in-register' lateral aggregation of isolatedmyosin filaments further indicates rather clearly that the ribbons originate from aside-by-side aggregation of the native thick filaments. Presumably such aggregationin vivo is made possible under certain imposed conditions such as low temperatureand hypertonicity. It may be considered as somewhat fortuitous that similar con-

Figs. 23-28. Structural similarity of filaments assembled from purified smoothmuscle myosin (Figs. 23-25) and the native filaments released from the isolated cells(Figs. 26-28). A periodicity of about 14 nm, arising from the myosin cross-bridgeorganization may be recognized along the entire length of the filaments and bareedges (arrows) occur at the filament ends. Figs. 23 and 24 were generously supplied byDr A. Sobieszek. Figs. 23, 24, x 70000 ; Fig. 25, x 80000 ; Fig. 26, x 77000 ; Fig. 27,x 130000 ; Fig. 28, x 80000.

Figs. 29-32. Optical diffraction patterns obtained from native myosin filaments suchas shown in Figs. 26-28. All patterns show a strong meridional reflexion at about14 nm corresponding to the cross-bridge repeat and indicated by an arrow on Fig. 29.Apart from this reflexion the patterns show meridional or off-meridional reflexionsat about 28 nm and 36 nm (Figs. 29, 30, 32) as well as 48-50 nm (Figs. 29, 31). Seealso text.

344 J- V- Small

elusions were reached in earlier studies of smooth muscle (Somlyo, Somlyo, Devine &Rice, 1971 ; Somlyo et al. 1973) using preparations for which there were no validcontrols on the preparative procedures used.

From the isolation of the native thick filaments from smooth muscle cells it isapparent that the myosin filaments of vertebrate smooth muscle are closely similar,if not identical, to filaments formed from purified smooth muscle myosin. That thefilaments obtained from the cells were not formed spontaneously in the cell homo-genates could be excluded from the fact that smooth muscle myosin is essentiallyinsoluble under the conditions employed, as already discussed above.

The most significant feature of the native myosin filaments was the presence, as onfilaments from purified myosin (Sobieszek, 1972, 1976; Sobieszek & Small, 1973),of projecting cross-bridges along the entire filament length, giving rise to a continuousperiodicity of about 14 nm. The obvious lack of any central 'bare zone' as is charac-teristic of the thick filaments from cross-striated muscle (Huxley, 1957; Craig &Offer, 1976) shows that the organization of the myosin molecules and thus thestructural polarity of these filaments is basically different from that deduced forstriated muscle. We shall return again to this difference below when considering theorganization of the contractile apparatus.

The contractile apparatus

Some indications of the gross organization of the contractile apparatus in thesmooth muscle cell were provided by earlier observation of cytoplasmic fibrils in rigor(Small, 1974; see also Figs. 21-22). These fibrils which were up to about 0-3 /im inwidth gave rise to a criss-cross pattern in the cell and became progressively moreoblique with respect to the cell axis at shorter cell lengths. It was concluded that thefibrils were bundles of myofilaments and that they possibly represented individualcontractile units within the cell. In the electron microscope we have now shown thepresence of groups of myofilaments in cells in rigor which evidently correspond tocross-sections of the cytoplasmic fibrils recognized in the light microscope. Whilethere is some variability in size of the myofilament groups, and likewise in the widthof the cytoplasmic fibrils (Small, 1974) the existence of these myofibrillar units inrigor must reflect in some way the gross organization of the contractile apparatus.As we have seen, some grouping of the myofilaments is also suggested from the

Fig. 3 3. Schematic diagram indicating the general features of the proposed organizationof the contractile apparatus of the smooth muscle cell. For clarity, the cell propor-tions have been chosen for a cell at about its shortest length. One family of contractileunits is here envisioned as having membrane attachment sites located on a commonspiral along the cell surface. Several families of such units on similar spirals translatedalong the cell axis would be required to occupy the entire cell volume. The densebody- 10-nm filament network (d) forms a structural framework between the con-tractile units and is also attached to the cell surface. The number of filaments in across-section of an individual unit (t) corresponds approximately to the averagenumber noted in the myofilament groups in rigor, n corresponds to the cell nucleuswith the associated organelle-containing regions at the nuclear poles.


Fig. 33. For legend see opposite.

346 J. V. Small

electron micrographs of cells in the relaxed state (Fig. 13). In this state there is notthe same clear separation of filament groups as is seen in rigor but this is fully con-sistent with the earlier light-microscope observations (see Small, 1974) in which thetransition from rigor to relaxation was seen to be accompanied by a change from afibrillar to a homogenous cytoplasmic appearance. Although the existence of separatefibrillar units has yet to be shown directly, for example by their isolation from thecells, the evidence we have obtained is consistent with a gross organization of thecontractile apparatus as indicated very schematically in Fig. 33. In this model we haveassumed that the contractile units are attached at each end to the cell membrane. Inthe extended cell the units would be approximately parallel to the cell axis andwould become oriented at about 20 ° in the shortened cell (Small, 1974).

We should next consider the composition of the myofibrillar bundles and in thisrespect the relationship of the 10-nm filaments and cytoplasmic dense bodies to thesecontractile units. The notion has recently been revived (Ashton, Somlyo & Somlyo,1975) that the cytoplasmic dense bodies act as intracellular attachment sites for actinand thus are homologous to the dense bodies of certain invertebrate muscles (Rosen-bluth, 1972; Sobieszek, 1973). If so, the dense bodies would be present as integralcomponents of the contractile units. Several lines of evidence indicate, however,that this is not the case. First of all such an arrangement would give rise to somedegree of grouping of the actin and myosin filaments into pseudo I-bands or M-bandsthat would be most easily detectable in extended cells, but evidence for such grouping)

as would be indicated by areas of only actin or myosin filaments, at least as large asthe dense areas (which range from about 0-08 to o-i /tm in diameter) were not found.Instead, the myosin filaments were commonly separated by only one or two rows ofthe thin filaments in both the rigor and relaxed states. Second, since the dense bodiesare always found to be associated with, or surrounded by longitudinally oriented10-nm filaments (Fawcett, 1966; Lowy & Small, 1970; Cooke & Chase, 1971 ;Uehara, Campbell & Burnstock, 1971 ; Somlyo et al. 1973 ; Ashton et al. 1975) wewould expect to see 10-nm filaments intermingled between the actin and myosinfilaments. On the contrary, the 10-nm filaments lie in areas between the myofilamentgroups (see also Small & Sobieszek, 1976). Third, in a previous study (Small & Squire,1972) it was shown in longitudinal sections of well oriented muscles that the thinfilaments always bypassed the dense bodies and never abutted them and this has alsobeen confirmed by others (Cooke, 1976). And finally, in parallel studies of the 10-nmfilaments (Small & Sobieszek, 1976) we have obtained evidence which indicates thatthese latter filaments together with the dense bodies form a cytoskeletal frameworkwithin the smooth muscle cell (see also Cooke, 1976) that is distinct from the con-tractile apparatus. The data available thus indicate that the contractile apparatusproper is composed only of the thick and thin myofi laments and thus that the 3-dimensional organization is markedly different from that found in other muscle types.

In the taenia coli muscle we have found a ratio of actin to myosin filaments inextended muscles of between 10 and 13 to 1. For different smooth muscles the ratioof actin to myosin by weight is apparently quite variable, ranging from about 1 :o-6in hog carotid artery to 1: i-6 in guinea-pig vas deferens (Small & Sobieszek, 1976).


We would then expect a corresponding variability in the relative numbers of thick andthin filaments from muscle to muscle. Somlyo et al. (1973) obtained a value of 15:1for the numbers of actin to myosin filaments in vascular smooth muscle which wouldat least be consistent with the higher mass ratio of actin to myosin, 1 :o-6, as comparedto a value of 1:1 obtained for the guinea-pig taenia coli (Small & Sobieszek, 1976).The latter mass ratio for taenia coli differs from that obtained by Tregear & Squire(1973) and this difference is at present difficult to explain. It is interesting to note,however, that from the mass and filament ratios we have obtained for the taenia colithe relationship given by Tregear & Squire for the number of cross-bridges peri4-o-nm repeat on the myosin filaments also yields a value of about 4. Using ouractin to myosin mass ratios and the filament ratio of Somlyo et al. for vascular smoothmuscle we arrive again at about the same value. Further estimates of myosin contentand filament numbers in appropriately prepared material from different sources will benecessary to establish the significance of this value for smooth muscle myosin filaments.

The problem of how the shearing interaction between the thick and thin filamentstakes place in smooth muscle is an interesting one and the answer hinges mainly on theprecise molecular organization and polarity of the thick myosin filaments. From bothvisual and optical diffraction analyses these filaments appear to be identical to filamentsassembled from purified smooth muscle myosin and for which a model of the possiblemolecular organization has already been proposed (Sobieszek, 1972, 1976).

From the continuous distribution of cross-bridges along the myosin filaments it isreasonable to assume that the relative shearing between the thick and thin filamentsinvolves interaction of individual thin filaments with the entire length of a thickfilament. For a particular myosin filament the sliding direction would then be equallyshared among the surrounding thin filaments. How the polarity of the myosin mole-cule on the thick filaments and the polarity of the thin filaments fit into such a schemeis not possible to answer at the present time. Assuming a helical organization for themyosin filaments each helical strand may include molecules with a given polarity,resulting in 'strand polarity', and a given actin filament may then interact withmolecules in appropriate positions in strands of common polarity. Alternatively, thefilaments could be constructed in such a way that there would be some kind of ' rowpolarity' with a common polarity for a longitudinal row of cross-bridges along thefilament. Since it is now apparent that the form of myosin filaments in non-musclemotile systems is likely the same as found in vertebrate smooth muscle (Hinssen &D'Haese, 1974; D'Haese & Hinssen, 1974, and private communication) the solutionto these fundamental problems will be essential to our general understanding of thecontractile process.

I am grateful to Dr A. Sobieszek for many helpful discussions and to him and Dr S. Fabczakfor comments on the manuscript. I thank Ms Tove Wiegers for excellent technical assistance.This work was supported by generous grants from the Muscular Dystrophy Association Inc.and the Volkswagen Foundation.

348 J. V. Small

REFERENCES

ASHTON, F. T., SOMLYO, A. V. & SOMLYO, A. P. (1975). The contractile apparatus of vascularsmooth muscle : Intermediate high voltage stereo electron microscopy. J. molec. Biol. 98,17-29.

BENDALL, J. R. (1969). Muscles, Molecules and Movement. London: Heinemann.COOKE, P. (1976). A filamentous cytoskeleton in vertebrate smooth muscle fiber. J. Cell Biol.

68, 539-556.COOKE, P. H. & CHASE, R. H. (1971). Potassium chloride-insoluble myofilaments in vertebrate

smooth muscle cells. Expl Cell Res. 66, 417-425.CRAIG, R. & OFFER, G. (1976). Axial arrangement of crossbridges in thick filaments of verte-

brate skeletal muscle. J. molec. Biol. 102, 325-332.D'HAESE, J. & HINSSEN, H. (1974). Structure of synthetic and native myosin filaments from

Amoebaproteus. Cell Tiss. Res. 151, 323-335.FAWCETT, D. W. (1966). In An Atlas of Fine Structure. The Cell: Its Organelles and Inclusions,

p. 243. Philadelphia : Saunders.HAMOIR, G. (1973). Extractability and properties of the contractile proteins of vertebrate

smooth muscle. Phil. Trans. R. Soc. Ser. B 265, 169-181.HINSSEN, H. & D'HAESE, J. (1974). Filament formation by slime mould myosin isolated at low

ionic strength. J. Cell Sci. 15, 113-129.HUXLEY, H. E. (1957). The double array of filaments in cross-striated muscle. J. biophys.

biochem. Cytol. 3, 631-647.HUXLEY, H. E. (1972). Molecular basis of contraction in cross-striated muscles. In The Struc-

ture and Function of Muscle, vol. 1 (ed. G. H. Bourne), 2nd edn, pp. 301-387. New York andLondon : Academic Press.

JONES, A. W., SOMLYO, A. P. & SOMLYO, A. V. (1973). Potassium accumulation in smoothmuscle and associated ultrastructural changes. .7. Physiol., Lond. 232, 247-273.

KELLY, R. E. & RICE, R. V. (1968). Localisation of myosin filaments in smooth muscle.J. Cell Biol. 37, 105-116.

LOWY, J. & MULVANY, M. J. (1973). Mechanical properties of guinea pig taenia coli muscles.Acta physiol. scand. 88, 123-136.

LOWY, J., POULSEN, F. R. & VIBERT, P. J. (1970). Myosin filaments in vertebrate smoothmuscle. Nature, Lond. 235, 1053-1054.

LOWY, J. & SMALL, J. V. (1970). The organisation of myosin and actin in vertebrate smoothmuscle. Nature, Lond. 225, 46-51.

LOWY, J., VIBERT, P. J., HASELGROVE, J. C. & POULSEN, F. R. (1973). The structure of themyosin elements in vertebrate smooth muscle. Proc. R. Soc. B 265, 191-196.

MOORE, P. B., HUXLEY, H. E. & DE ROSIER, D. J. (1970). Three dimensional reconstructionof F-actin, thin filaments and decorated thin filaments. J. molec. Biol. 50, 279—295.

ROSENBLUTH, J. (1972). Obliquely striated muscle. In The Structure and Function of Muscle,vol. 1 (ed. G. H. Bourne), 2nd edn, pp. 389-42C. New York and London : Academic Press.

SHOENBERG, C. F., GOODFORD, P. J., WOLOWYK, M. W. & WOOTON, G. S. (1973). Ion changesduring smooth muscle fixation for electron microscopy. J. Mechanochem. Cell Mot. 2, 69-82.

SHOENBERG, C. F. & HASELGROVE, J. C. (1974). Filaments and ribbons in vertebrate smoothmuscle. Nature, Lotid. 249, 152-154.

SHOENBERG, C. F. & NEEDHAM, D. M. (1976). A study of the mechanism of contraction invertebrate smooth muscle. Biol. Rev. 51, 53-104.

SMALL, J. V. (1974). Contractile units in vertebrate smooth muscle cells. Native, Lond. 249,324-327.

SMALL, J. V. & SOBIESZEK, A. (1977). Studies on the function and composition of the io-nm(ioc-A) filaments of vertebrate smooth muscle. J. Cell Sci. 23, 243-268.

SMALL, J. V. & SQUIRE, J. M. (1972). Structural basis of contraction in vertebrate smoothmuscle. ,7. molec. Biol. 67, 117-149.

SOBIESZEK, A. (1972). Cross-bridges on self-assembled smooth muscle myosin filaments.J. molec. Biol. 70, 741-744.

SOBIESZEK, A. (1973). The fine structure of the contractile apparatus of the anterior byssusretractor muscle of Mytilus edulis. J. Ultrastruct. Res. 43, 313-343.


SOBLESZEK, A. (1976). Vertebrate smooth muscle myosin : enzymatic and structural properties.In Smooth Muscle Biochemistry (ed. N. L. Stephens). Baltimore: University Park Press.(In Press.)

SOBIESZEK, A. & BREMEL, R. D. (1975). Preparation and properties of vertebrate smooth musclemyofibril8 and actomyosin. Eur. J. Biochem. 55, 49—60.

SOBIESZEK, A. & SMALL, J. V. (1973). Filaments from purified smooth muscle myosin. ColdSpring Harb. Symp. quant. Biol. 37, 109-112.

SOBIESZEK, A. & SMALL, J. V. (1976). Myosin-linked calcium regulation in vertebrate smoothmuscle. ,7. molec. Biol. 102, 75-92.

SOMLYO, A. P., DEVINE, C. E., SOMLYO, A. V. & RICE, R. V. (1973). Filament organisation invertebrate smooth muscle. Phil. Trans. R. Soc. Ser. B 265, 223-229.

SOMLYO, A. P., SOMLYO, A. V., DEVINE, C. E. & RICE, R. V. (1971). Aggregation of thick fila-ments into ribbons in mammalian smooth muscle. Nature, New Biol. 231, 243-246.

TREGEAR, R. T. & SQUIRE, J. M. (1973). Myosin content and filament structure in smooth andstriated muscle. J. molec. Biol. 77, 279-290.

UEHARA, Y., CAMPBELL, G. R. & BURNSTOCK, G. (1971). Cytoplasmic filaments in developingand adult smooth muscle. J. Cell Biol. 50, 484-497.

VAN BREEMAN, C, FARINAS, B. R., CASTEELS, R., GERBA, P., WUYLACK, F. & DETH, R. (1973).Factors controlling cytoplasmic Caa+ concentration. Phil. Trans. R. Soc. Ser. B 265, 57-71.

{Received 18 June 1976)

Documents

STUDIES ON ISOLATED SMOOTH MUSCLE CELLS: THE … · Isolated smooth muscle cells were prepared from guinea-pig taenia coli muscle in a manner similar to that described previously