Regulation of the interaction between smooth muscle myosin and actin

KATHLEEN M. TRYBUS

Rosenstiel Research Center, Brandéis University, Waltham, MA 02254, USA

and DAVID M. WARSHAWDepartment of Physiology and Biophysics, University of Vermont, College of Medicine, Burlington, VT 05405, USA

Summary

Phosphorylation of the regulatory light chain of smooth muscle myosin efficiently regulates the actin- activated ATPase activity of myosin filaments in solution and actin movement in an in vitro motility assay, independently of thin-filament regulatory pro­teins. Filaments containing both phosphorylated and dephosphorylated heads move actin at intermediate rates, depending on the relative proportions of the two myosin species. The decrease in velocity can be accounted for by mechanical interactions between phosphorylated heads and ‘weak-binding’ dephos­phorylated crossbridges. These results imply that shortening velocity could be modulated in any muscle by varying the relative proportions of two populations of crossbridges with different cycling rates.

Key words: smooth muscle myosin, light chain phosphorylation, motility assay.

Introduction

Phosphorylation of serine 19 of the 20x l0 s Mr regulatory light chain by myosin light chain kinase (MLCK) is a key event regulating smooth muscle actomyosin interactions and force development. The calcium sensitivity of smooth muscle contraction arises because MLCK alone is inactive, while the calcium-calmodulin-MLCK complex is active. Although regulation of smooth muscle myosin is con­sidered to be primarily myosin-based, thin-filament associated proteins such as caldesmon and calponin have recently been given considerable attention as putative regulatory proteins (Fig. 1). Caldesmon, an 87x l0 3Mr 74 nm long protein that binds actin, tropomyosin, cal­modulin and myosin, inhibits actin-activated ATPase activity. It has been suggested that caldesmon sterically blocks binding of the weak-binding M.ATP and M.ADP.Pi complexes to actin (Hemric and Chalovich, 1988). Cal­ponin, a 35 x 10s Mr calcium-binding, calmodulin-binding, troponin-T like protein also inhibits actin-activated ATPase activity (Takahashi et al. 1988). Recent evidence suggests that phosphorylation of calponin reverses the inhibition of activity (Winder and Walsh, 1990).

Phosphorylation regulates the activity of filamentous myosin

The presence of inhibitory actin-binding proteins raises the question of whether phosphorylation—dephosphoryl­ation alone can regulate the activity of myosin filaments to a high degree, or whether a second regulatory system isJournal of Cell Science, Supplement 14, 87-89 (1991)Printed in Great Britain © The Company of Biologists Limited 1991

required to depress the activity of dephosphorylated filaments. In solution, it has been difficult to determine the activity of dephosphorylated filaments because of the lability of this structure in the presence of MgATP. Dephosphorylated filaments readily dissociate to a mono­mer in which the tail is folded into thirds and the heads are bent toward the rod. Single-turnover measurements, in which each active site is given only one ATP molecule, established that this conformation traps the products of ATP hydrolysis, and releases them at the very low rate of 0 .0005 s-1 (Cross et al. 1986). Phosphorylation of the regulatory light chain causes reassembly into filaments and a several hundred-fold increase in actin-activated activity to 0.3 s-1 (Fig. 2). It was not clear from these observations, however, whether conformation or phos­phorylation was the major determinant of activity. By analogy with other myosins, the regulatory light chain is located near the head/rod junction, at least 10 nm from the active site; thus even a ‘direct’ effect of phosphorylation implies communication over distances equal to half the length of the myosin head. In smooth muscle, the 17xl03Mr light chain has been shown by photoaffinity labelling to be near the active site (Okamoto et al. 1986), and this subunit might be the link between the two regions.

The extent to which phosphorylation dependent as- sembly-disassembly occurs in vivo has not been well established, although in at least one type of smooth muscle a change in filament density between rest and contraction has been observed (Gillis et al. 1988). Nevertheless, it has been established that relaxed smooth muscle cells contain some dephosphorylated filaments (Somlyo et al. 1981), and a mechanism to inhibit their activity must exist. If dephosphorylation is not sufficient, then thin-filament proteins may be required to inhibit their activity.

In order to block disassembly of dephosphorylated filaments in the presence of MgATP, monoclonal anti­bodies with epitopes along the central portion of the myosin were used to stabilize the polymer, without affecting the activity of phosphorylated filaments. The actin-activated ATPase of antibody-stabilized, dephos­phorylated filaments, obtained by steady-state and single­turnover measurements, was approximately 0.002 s-1 , well below the rate obtained with phosphorylated fila­ments, and only 4 -5 fold higher than the value obtained with folded myosin (Trybus, 1989). These results suggest that three levels of activity can be distinguished: (1) enzymatically-incompetent folded myosin, (2) inactive dephosphorylated filaments and (3) active phosphorylated filaments. Interaction of the myosin rod with the neck in the folded conformation appears to stabilize ADP and phosphate at the active site to a degree that cannot quite be achieved in the filament.

87

Myosin heavy chains

ActinTropomyosin

Light chains

Fig. 1. Schematic diagram of smooth muscle myosin regulatory components.

0.002 s "1 Inactive

MLCK »0.3s“ 1Active

phosphataseDephosphorylated filaments

P^= __________

Phosphorylated filaments

I ’

fMLCK

Folded monomer Inactive

0.0005 s ' 1

Transient unfolded monomer

Fig. 2. Conformational states adopted by smooth muscle myosin, and their actin-activated ATPase rates.

Based on these experiments, a second regulatory system is not required to inhibit the enzymatic activity of dephosphorylated filaments. Phosphorylation, however, acts predominantly on a kinetic step affecting product release, and has only minor effects on binding of M.ATP and M.ADP.Pj to actin (Sellers, 1985). The role of caldesmon might then simply be to inhibit binding of myosin to actin in a relaxed muscle, although cycling between weak and strong binding states could be adequately regulated by phosphorylation. If filament

disassembly to the folded conformation occurred, it would only enhance an already efficient regulatory system. It would also produce a soluble species that does not bind actin, and which in principle could be recruited to different areas of the cell more readily than filamentous myosin.

Dephosphorylated myosin slows movement of actin by phosphorylated myosin

Actin-activated ATPases in solution showed a large difference in ATP turnover by phosphorylated and dephos­phorylated myosin, but is this factor enough to control movement of actin by myosin? The degree to which phosphorylation regulates the motion of single actin filaments by myosin was observed by an in vitro motility assay. Phosphorylated filaments moved actin at 0.2-0.4,ums_ (at 22°C), but dephosphorylated filaments held actin in an immobile, rigor-like conformation even in the presence of MgATP (Warshaw et al. 1990). The surprising aspect of this observation was that dephosphor­ylated myosin, which from solution studies is predomi­nantly in the ‘weak-binding’ AM.ADP.P; state, main­tained strong enough interactions with actin that the polymer did not diffuse away into solution (Warshaw et al. 1990).

During a contraction there are many times when myosin is not fully dephosphorylated or phosphorylated, but only partially phosphorylated. As calcium levels increase and myosin gets phosphorylated, the muscle begins to rapidly shorten and produce force (Fig. 3). With time, although force levels remain high, the muscle shortens more slowly, and levels of light chain phosphoryl­ation decrease as phosphatase activity predominates over kinase activity. This apparent modulation of shortening velocity by phosphorylation implies that phosphoryl­ation-déphosphorylation may be more than a simple ‘on -o ff switch. Shortening velocity is independent of the number of phosphorylated heads that move actin, so changes in velocity cannot be simply explained by changes in the number of active crossbridges.

Two ways that intermediate shortening velocities could be obtained are: (1) dephosphorylated bridges act as a load against which the phosphorylated bridges must work, thus slowing their velocity, or (2) the intrinsic rate of ATP

Force

------- L w i . _ Velocity..... myosin LC-P

Latchbridge (internal load)

Modulation

Fig. 3. Schematic diagram of the response of a smooth muscle to calcium. Dashed line indicates calcium concentration, degree of light chain phosphorylation and shortening velocity, while the solid line shows the level of force production.Curved arrows represent the crossbridge cycling rate.

K. M. Trybus and. D. M. Warshaw

F ig . 4. Effect of ‘weak-binding’ bridges on actin movement. Dephosphorylated smooth muscle myosin slows actin movement by phosphorylated smooth muscle myosin (filled circles) and skeletal muscle myosin (filled boxes). The effect of pPDM-skeletal myosin, a non-cycling analog of the ‘weak- binding’ conformation, on phosphorylated smooth (open circles) or skeletal myosin (open boxes) is very similar to that obtained with dephosphorylated smooth muscle myosin. Conditions:25 m M KC1, 4mM MgCl2, I m M EGTA, I m M DTT, p H 7.4, 22 C.

turnover by phosphorylated myosin is decreased as the number of neighboring dephosphorylated molecules in the filament increases. The motility assay is considered to be the in vitro correlate of unloaded shortening velocity, thus this technique in combination with solution studies should allow one to distinguish between the above possibilities.

The rate at which actin was moved by phosphorylated smooth muscle myosin slowed markedly as the proportion of dephosphorylated myosin in the filament increased to greater than 60% of the total (Fig. 4, 25 m M KC1). Intrinsic changes in ATPase activity were ruled out both by solution actin-activated ATPases, and by the obser­vation that mixtures of monomers resulted in changes in velocity similar to that observed for copolymers (unpub­lished data, D. M. Warshaw and K. M. Trybus). Surpris­ingly, dephosphorylated myosin had an even more pro­found effect on the movement of actin by skeletal muscle myosin (Fig. 4), suggesting that under these conditions skeletal myosin has relatively fewer crossbridges in a ‘strong-binding’ state compared to smooth muscle myosin.

Another way of observing the effect of ‘weak-binding’ bridges on movement is to use the well-characterized chemically modified analog, pPDM-skeletal myosin, which is thought to resemble a non-cycling M.ADP.P; conformation. This analog behaved mechanically in the same way as dephosphorylated myosin, based on its effect on actin movement by phosphorylated smooth and skeletal muscle myosin (Fig. 4). In contrast, skeletal myosin extensively modified with NEM, which is an analog of a ‘strong-binding’ conformation, completely abolished move­ment when present as only 1 % of the total myosin.

Modulation of actin velocity by active crossbridges with different cycling rates

Actin velocity was not only modulated when copolymers contained ‘weak-binding’ bridges, but could also be observed when filaments were formed from two popu­lations of active bridges that cycle at different rates, such as phosphorylated smooth muscle myosin and skeletal myosin. The observed velocity of actin movement by

smooth-skeletal copolymers was not a simple linear relationship that depended only on the proportion of fast and slow myosin. A nonlinear relationship, reflecting a mechanical interaction between the two myosin species, could arise if the crossbridge force-velocity relationships for the two species are assumed to be hyperbolic and different in curvature (Warshaw et al. 1990). Mixtures of smooth and skeletal myosin have no counterpart in the cell, but these results could be generally applied to muscles where multiple myosin isoforms with different turnover times are expressed, such as in developing muscles or in the heart.

Future prospects

The observations described here could explain why shortening velocity correlates with the degree of light chain phosphorylation, but they do not offer any infor­mation as to why a smooth muscle can maintain force at low phosphorylation levels (Fig. 3). The latchbridge hypothesis of Murphy and co-workers (Hai and Murphy, 1989) suggests that phosphorylation is required for the myosin head to attach to actin, but once attached, it can be dephosphorylated by phosphatases and become a ‘latch­bridge’. This slowly detaching, force-producing, dephos­phorylated latchbridge allows force to remain high while phosphorylation levels decrease. The experiments de­scribed here analyze the properties of dephosphorylated myosin that has not gone through the pathway required of a latchbridge (i.e phosphorylation, attachment to actin, and subsequent dephosphorylation), yet these dephosphor­ylated bridges are having profound effects on the observed movement of actin by phosphorylated myosin. Direct measurement of the force produced by dephosphorylated and phosphorylated crossbridges interacting with a single actin filament (Kishino and Yanagida, 1988) should help resolve some of these questions.

ReferencesC r o s s , R. A., C r o s s , K. E. a n d S o b ie s z e k , A. (1986). ATP-linked

monomer-polymer equilibrium of smooth muscle myosin: the free folded monomer traps ADP.Pj. EMBO J. 5, 2637-2641.

G i l l i s , J. M., C a o , M. L. a n d G o d f r a i n d - D e B e c k e r , A. (1988). Density of myosin filaments in the rat anococcygeus muscle, at rest and in contraction. II. J. Muscle Res. Cell Motil. 9, 18-28.

H a ï , C. a n d M u r p h y , R. A. (1989). Ca2+, crossbridge phosphorylation and contraction. A. Rev. Physiol. 51, 285-298.

H e m r ic , M. E. a n d C h a l o v i c h , J. M. (1988). Effect of caldesmon on the ATPase activity and the binding of smooth and skeletal myosin subfragments to actin. J. biol. Chem. 263, 1878-1885.

K is h in o , A. a n d Y a n a g i d a , T. (1988). Force measurements by micromanipulation of a single actin filament by glass needles. Nature 334, 74-76.

O k a m o t o , Y . , S e k i n e , T., G r a m m e r , J. a n d Y o u n t , R. G . (1986). The essential light chains constitute part of the active site of smooth muscle myosin. Nature 324, 78-80.

S e l l e r s , J. R. (1985). Mechanism of the phosphorylation dependent regulation of smooth muscle heavy meromyosin. J. biol. Chem. 260, 15815-15819.

S o m ly o , A. V., B u t l e r , T. M., B o n d , M. a n d S o m ly o , A. P. (1981). Myosin filaments have non-phosphorylated light chains in relaxed smooth muscle. Nature 294, 567-569.

T a k a h a s h i , K . , H i w a d a , K . a n d K o k u b u , T . (1988). Vascular smooth muscle calponin. A novel troponin T-like protein. Hypertension 11, 620-626.

T r y b u s , K. M. (1989). Filamentous smooth muscle myosin is regulated by phosphorylation. J. Cell Biol. 109, 2887-2894.

W a r s h a w , D . M., D e s r o s i e r s , J. M., W o r k , S. S. a n d T r y b u s , K. M. (1990). Smooth muscle myosin cross-bridge interactions modulate actin filament sliding velocity in vitro. J. Cell Biol. I l l , 453-463.

W i n d e r , S. J. a n d W a l s h , M. P. (1990). Smooth muscle calponin. Inhibition of actomyosin MgATPase and regulation by phosphorylation. J. biol. Chem. 265, 10 148-10 155.

Regulation o f myosin and actin interactions 89