Myosin Essential Light Chain Isoforms Modulate the Velocity of Shortening Propelled by Nonphosphorylated Cross-bridges*

The differential effects of essential light chain isoforms (LC17a and LC17b) on the mechanical properties of smooth muscle were determined by exchanging recombinant for endogenous LC17 in permeabilized smooth muscle treated with trifluoperazine (TFP). Co-precipitation with endogenous myosin heavy chain verified that 40–60% of endogenous LC17a could be exchanged for recombinant LC17aor LC17b. Upon addition of MgATP in Ca2+-free solution, recombinant LC17 exchange induced slow contractions unaccompanied by regulatory light chain (RLC) phosphorylation only in TFP-treated, but not in untreated, permeabilized smooth muscle; the shortening velocity and rate of force development were approximately 1.5 and 2 times faster, respectively, in response to LC17a than LC17b. Additional incubation with recombinant, thiophosphorylated RLC increased the shortening velocity, independent of the LC17 isoform exchanged. The LC17-induced contractions of TFP-treated muscles were abolished by prior addition of nonphosphorylated RLC. We suggest that LC17 stiffens the lever arm of myosin and, in the absence of regulation by RLC, permits cross-bridge cycling without requiring RLC phosphorylation. Our results are compatible with nonphosphorylated RLC acting as a repressor and with LC17isoforms modulating the MgADP affinity and, consequently, rate of cooperative cycling of nonphosphorylated cross-bridges.

The markedly different rates of contraction and relaxation in fast, phasic and slow, tonic smooth muscles reflect differences not only at the level of the membrane potential, signal transduction, rates of myosin light chain phosphorylation and dephosphorylation, but also in the kinetic properties of the actomyosin motor. The muscle-specific differences in actomyosin kinetics have been ascribed to the existence of specific myosin isoforms, but their relative contributions to the mechanical properties of smooth muscles have not been unequivocally demonstrated (reviewed in Ref. 1). The expression of the two myosin heavy chain (MHC) 1 isoforms, SM-1 and SM-2 (2) that differ by, respectively, the presence or absence of a 34-amino acid extension of the carboxyl terminus (3), does not correlate with phasic or tonic kinetics of contraction. On the other hand, motility assays show different rates of movement propelled by MHC isoforms that are the products of an alternative splicing mechanism resulting in the presence or absence of a 7-amino acid insert near the nucleotide-binding region of the myosin head (4,5). Disparate results have been obtained concerning the effects on contractile kinetics of the two LC 17 isoforms, acidic (LC 17a ) and basic (LC 17b ), that differ in 5 of the 9 COOHterminal amino acid residues and are products of a single gene generated by an alternative splicing mechanism (6). The faster kinetics of phasic, smooth muscle in situ (7)(8)(9) correlates with the expression of the LC 17a isoform (10,11), albeit the proportion of myosin heavy chain containing the insert was also variant in these muscles, whereas in vitro studies have produced conflicting results. Increasing the LC 17a content of isolated aortic myosin has been reported to increase ATPase activity (12), but motility assays failed to detect a difference in the velocity of actin movement as a function of LC 17 isoform, although they showed faster movement propelled by myosin containing the 7-amino acid insert in the motor domain (4,13). Thus, it is possible that the LC 17 isoform, the motor domain insert, or a combination of both (1) are responsible for the different kinetic properties of phasic and tonic smooth muscle actomyosin in situ. The aim of the present study was to determine whether a change in the ratio of LC 17 isoforms alone can alter the contractile kinetics of smooth muscle. To that end, we exchanged the basic isoform, LC 17b , into bladder and amnion smooth muscles, both of which have fast, phasic properties and express only the LC 17a isoform (Ref. 10 and the present study).
We used TFP, a hydrophobic cation that facilitates extraction and/or exchange of light chains from isolated myosin (12,14,15) and single cells (16), to promote light chain exchange in permeabilized smooth muscle. Our results show a significant LC 17 isoform-specific effect on the unloaded shortening velocity of nonphosphorylated, slowly cycling cross-bridges, but no detectable effect on muscles in which the RLCs are thiophosphorylated.

Tissue Preparation and Light Chain Exchange
Bundles of bladder detrusor muscle (3 mm ϫ 200 -400 m) from male, New Zealand rabbits weighing 2-3 kg or chicken amnion sheets (3 ϫ 3 mm) from 10 -13 day eggs were dissected and the ends tied with monofilament silk, and attached to a fixed hook and a force transducer (AE 801; AME, Horten, Norway) for recording isometric tension at room temperature. The chicken amnion also expresses 100% LC 17a and was used in preliminary studies to optimize conditions for LC 17 exchange in smooth muscle, because of its thinness and lack of connective tissue and * This work was supported by American Heart Association, Virginia Affiliate, Inc. Grant VA-96-F-11 (to J. D. M.) and National Institutes of Health Grants PO1-HL19242 and PO1-HL48807. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Expression, Purification, and Thiophosphorylation of Myosin Light Chains
Recombinant light chains, from a gizzard library, were expressed in Escherichia coli as described previously (18). In some experiments, thiophosphorylated recombinant RLC was added to muscle strips. These were prepared by incubating recombinant RLC (1.0 mg⅐ml Ϫ1 ) with 20 g⅐ml Ϫ1 myosin light chain kinase (MLCK; generous gift of Dr. D. J. Hartshorne), 5 M calmodulin (Calbiochem), 2 mM ATP␥S (Boehringer Mannheim, Indianapolis, IN), 2.4 mM MgCl 2 , and 0.6 mM CaCl 2 overnight at 4°C, followed by extensive dialysis against zero Ca 2ϩ -rigor solution. Greater than 90% thiophosphorylation of recombinant RLC was confirmed by SDS-PAGE using urea/glycerol gels (19).

Photolysis of Caged ATP
Muscle strips were treated according to the TFP protocol and incubated in zero Ca 2ϩ -rigor solution containing either LC 17a or LC 17b for 60 min. This was replaced by photolysis solution containing 5 mM caged ATP (1-(2-nitrophenyl)-ethyl-ATP; Molecular Probes, Eugene, OR), and 40 mM DTT, in zero Ca 2ϩ -rigor solution (pH 7.1, ionic strength 0.16) for 3 min. Caged ATP was photolyzed using a frequency-doubled ruby laser (Lumonics, Warwick, United Kingdom), delivering a 50-ns pulse of near-UV light at 347 nm, sufficient to release 1.2 Ϯ 0.1 mM ATP (n ϭ 5), as determined by high performance liquid chromatography of the trough solution at the completion of force measurements. Contaminant ATP and ADP were removed from the caged ATP stocks using 100 g⅐ml Ϫ1 apyrase (grade V; Ref. 10). The set-up used for laser flash photolysis has been previously described (20,21).

Slack-test Protocol
Unloaded shortening velocity (V us ) was determined using the slacktest method (22). Briefly, strips were attached to hooks via aluminum T-clips or, in some earlier experiments, with small loops. After reaching steady state isometric tension, length changes of varying amplitude, sufficient to just reduce tension to zero, were applied to one end of the muscle strip using an isotonic lever (Model 308, Cambridge Technology Inc., Watertown, MA, maximum step time 300 s). The time taken for the strip to shorten against zero load and take up the slack was recorded and plotted against the length change. The slope of this relationship was taken as unloaded shortening velocity. Linear regression analysis was used to obtain the best fit straight line. In order to compare the effects of LC 17 isoforms on V us , strips were treated according to the TFP protocol and incubated with the appropriate LC 17 isoform in zero Ca 2ϩ -rigor for 60 min before force activation in relaxing solution and determination of V us as detailed above. In some experiments, V us was also measured in LC 17 -exchanged muscles after an additional 45-min incubation in zero Ca 2ϩ -rigor containing recombinant thiophosphorylated RLC (0.7 mg⅐ml Ϫ1 ).

Analysis of Light Chain Exchange
Muscle strips were rapidly frozen and stored in liquid nitrogen for subsequent analysis of protein composition. To discriminate between true exchange and nonspecific binding to the tissue, the extent of myosin light chain isoform exchange was determined by either immunoprecipitation of myosin using anti-myosin heavy chain antibody (generous gift from Drs. C. Kelley and R. Adelstein) or by sedimentation of myosin in low salt buffer at high speed centrifugation, followed by SDS-PAGE and transfer to polyvinylidene difluoride membranes for subsequent Western blotting using anti-LC 17 antibody (clone 7B5.1) and anti-RLC antibody (generous gift from Dr. K. Kamm). The immunoprecipitated pellets were washed from three to six times and the supernatant of the final wash treated with trichloroacetic acid to precipitate remaining protein. The entire precipitate was loaded on the gel. No LCs were found in the final wash supernatants. Calponin was detected using an anti-calponin antibody generously provided by Dr. M. Walsh. Using one-dimensional gel electrophoresis (15% polyacrylamide; 16 ϫ 14 cm), the LC 17a and LC 17b isoforms could be readily distinguished according to their mobility. In experiments involving incubations with recombinant, thiophosphorylated RLC, its level of exchange with endogenous, nonphosphorylated RLC could not be accurately determined, since both species migrate to the same position on an isoelectric focusing gel, as a consequence of an NH 2 -terminal 4-amino acid tag on the recombinant RLC (18).
Immunoprecipitation of Myosin-Treated strips were homogenized in ice-cold lysis buffer (1% Nonidet P-40, 150 mM NaCl, 1 mM AEBSF, 0.1 mM leupeptin, and 10 mM Tris-HCl (pH 7.1)). Cellular debris was pelleted by centrifugation at 800 ϫ g (4°C) and 10 mM MgATP was added to the collected supernatant for 15 min before incubation with bovine polyclonal anti-myosin heavy chain antibody, which recognizes both SM-1 and SM-2 heavy chain isoforms, for 60 min at room temperature. The homogenate was incubated for a further 1 h at room temperature with protein A-10% agarose (v/v) (Santa Cruz, Santa Cruz, CA) and immune complexes were collected by centrifugation at 300 ϫ g (4°C) for 5 min and washed 4 times with lysis buffer. The final supernatant was run as a control after precipitation with 20% trichloroacetic acid in acetone and several washes with acetone. Both immune complex and final supernatant pellets were resuspended in 2 ϫ Laemmli sample buffer.
Sedimentation of Myosin-Treated strips were homogenized in icecold lysis buffer (1% Nonidet P-40, 300 mM NaCl, 1 mM AEBSF, 0.1 mM leupeptin, and 10 mM Tris-HCl (pH 7.1)). Cellular debris was pelleted by centrifugation at 300 ϫ g (4°C) and 10 mM MgATP added to the supernatant for 15 min before centrifugation at 80,000 ϫ g (4°C) for 20 min. The collected supernatant was diluted with 9 volumes of ice-cold buffer containing 1 mM MgCl 2 and 10 mM MOPS and allowed to stand for 30 min to precipitate myosin before centrifugation at 105,000 ϫ g (4°C) for 30 min. After removal of the supernatant, the pellet was resuspended in low salt buffer and the last centrifugation step repeated before resuspension of the pellet in 2 ϫ Laemmli sample buffer.

RLC Phosphorylation in Situ
Phosphorylation levels of endogenous RLC in Table I were determined by two-dimensional gel electrophoresis, transfer to nitrocellulose membranes, and staining with colloidal gold as described previously (23).

Effect of TFP on Light Chain Content, RLC Phosphorylation, and Circular Dichroism (CD) of LC 17 in Vitro
Turkey gizzard myosin was prepared as described previously (24). To determine the effects of TFP on isolated myosin using the same (1 mM) concentration of TFP as used in experiments on skinned fibers, myosin (12 g) was incubated in 500 mM KCl, 10 mM EDTA, 1 mM ATP, 1 mM DTT, and 20 mM PIPES (pH 6.5) for 15 min at 30°C, in the presence or absence of 1 mM TFP. The solution was diluted with 9 volumes of ice-cold buffer containing 1 mM MgCl 2 and 10 mM MOPS, and allowed to stand for 30 min to precipitate myosin before centrifugation at 105,000 ϫ g (4°C) for 30 min and the pellet resuspended in 2 ϫ Laemmli buffer. For comparison, myosin precipitates were also prepared following treatment with 5 mM TFP, a concentration known to extract RLC (14). Because TFP could inhibit RLC phosphorylation in skinned fibers (see "Results") through its known effects on calmodulin, we used a Ca 2ϩ /calmodulin-independent MLCK (generous gift from Dr. M. Ikebe) to determine the direct (CaM-independent) effect of TFP on RLC phosphorylation. Precipitates of TFP-treated and nontreated myosin were prepared as described, but resuspended in buffer containing 20 g ml Ϫ1 Ca 2ϩ /calmodulin-independent MLCK, 2 mM ATP␥S, 50 mM KCl, 1 mM MgCl 2 , 5 mM DTT, and 30 mM Tris HCl (pH 7.5), and incubated overnight at 4°C. Myosin was then precipitated by highspeed centrifugation and the pellet resuspended in 8 M urea buffer. The 2 A. V. Somlyo, unpublished observation. extent of RLC thiophosphorylation was determined using one-dimensional isoelectric focusing mini-gels (7.5% polyacrylamide, 2% pH ampholytes 4.0 -5.4 (Pharmalyte, Pharmacia, Piscataway, NJ)).
CD spectra of LC 17 were obtained to determine whether TFP had a direct effect on LC 17 structure. LC 17a (17 M), in the presence or absence of TFP (100 M), was incubated in 50 mM Tris buffer (pH 7.0) at 30°C for 15 min and kept at 20°C for approximately 30 min. CD spectra were then recorded in a 0.5-mm fused silica cuvette at 20°C using a J-600 spectropolarimeter (JASCO, Tokyo, Japan). Five scans were accumulated per sample and spectra are presented as difference spectrum, i.e. spectrum recorded in the presence or absence of TFP minus spectrum of TFP or Tris buffer alone. Measurements were repeated three times with different solutions and the results were reproducible. The ␣-helical content was estimated from the mean residue molar ellipticity at 220 nm using the DICROPROT v. 2.4 analysis program (available at http:/www.ibcp.fr/DIRCROPROT.html).

Fluorescent Labeling
In order to determine the intracellular distribution of LC 17 added to skinned muscle fibers, LC 17b was fluorescently labeled at an approximate molar stoichiometry with tetramethylrhodamine 5-iodoacetamide (Molecular Probes). Labeling was performed for 2 h at room temperature followed by extensive dialysis to remove free tetramethylrhodamine 5-iodoacetamide. Labeling of LC 17b was confirmed by SDS-PAGE. Muscles were carried through the same standard protocol for LC 17 exchange using the fluorescently labeled LC 17b , and examined using a confocal microscope (Bio-Rad MRC 1000). Approximately 1-m thick optical sections were sampled throughout the thickness of the muscle strips.

Electron Micoscopy
Strips of bladder muscle were carried through the standard permeabilization protocol, followed by treatment with TFP. Untreated permeabilized control strips were carried through the same solutions, minus TFP. Following TFP treatment the muscles were incubated in relaxing solution or zero calcium rigor solution for 15 min and subsequently fixed in 2% glutaraldehyde in 0.075 M sodium cacodylate buffers with 4% sucrose plus 0.2% tannic acid for 2 h, followed by fixation in 2% osmium tetroxide and en bloc staining with saturated uranyl acetate, dehydration in alcohol, and embedment in Spurr's resin. Thin longitudinal or transverse sections were stained with lead citrate and examined in a Philips CM 12 electron microscope at 80 KeV. Cells were examined across the entire strip in transverse sections and evaluated for filament content and differences in morphology between the control and TFP-treated specimens.

Statistics and Data Analysis
Data are presented as the mean Ϯ S.E., and n refers to the number of muscle strips. Statistical significance was determined (p Ͻ 0.05) with unpaired Student's t test. Unless otherwise stated, all gels presented are representative of at least three similar experiments. The data from the flash photolysis and slack test experiments were collected using LabView 3.1.1 data acquisition software (National Instruments, Austin, TX) and curve fitting of force transients was performed using Sigma Plot 2.0 software (Jandel Scientific, San Rafael, CA).

The Effect of TFP on the Light Chain Content and Properties of Isolated Myosin and Myosin in Smooth
Muscle-The effects of TFP treatment were explored on isolated myosin for comparison with the in situ effects. Unlike experiments on isolated myosin (14) or single cells (16), treatment of Triton-permeabilized rabbit bladder strips with TFP caused no detectable loss of the RLC, LC 17 , or calponin, as determined in whole homogenates (Fig. 1A) or MHC immunoprecipitates (Fig. 1B). To verify whether this apparent discrepancy was the result of method-based differences, we examined the effects of two different treatment protocols on LC removal from isolated gizzard myosin. A concentration of TFP (1 mM TFP, 15 min at 30°C) which did not extract light chains from the muscle strips also did not extract light chains from isolated myosin in solution (Fig. 2). Loss of both RLC and LC 17 , however, was observed with the method (5 mM TFP, 1 h at 4°C) used by Trybus et al. (14) to extract the RLC (Fig. 2).
To assess the plausibility of TFP having a direct effect on the properties of LC 17 , we compared CD spectra of TFP-treated with untreated LC 17 . The difference spectra of LC 17a (17 M) in the presence or absence of TFP (100 M) are shown in Fig. 3. TFP increased the estimated ␣-helical content of LC 17a from 31 to 47%. Following TFP treatment of rabbit bladder strips, the RLC, although still associated with the MHC, could no longer be thiophosphorylated (Fig. 4A). This could reflect a direct effect of TFP on the RLC or an effect on calmodulin leading to an inhibition of MLCK. To test this second possibility, we determined the effect of TFP treatment on thiophosphorylation of RLC by a Ca 2ϩ /CaM-independent MLCK in vitro. The TFP concentration (1 mM) that did not extract RLC or LC 17 from myosin in situ (see below) or in vitro, as described above, inhibited thiophosphorylation of RLC of isolated myosin by a Ca 2ϩ /CaM-independent MLCK (Fig. 4B), suggesting that TFP made threonine and serine unavailable for thiophosphorylation rather than through inhibition of CaM-MLCK. A similar result was obtained using full-length MLCK in the presence of Ca 2ϩ / CaM (data not shown). Paired controls of non-TFP-treated myosin showed the expected phosphorylation of RLC by MLCK and Ca 2ϩ /CaM-independent MLCK (Fig. 4B).
Efficiency of Exchange of Recombinant Light Chains-Experiments were carried out to exchange the endogenous LC 17a isoforms in bladder and chicken amnion cells with the recombinant LC 17b and to characterize the efficiency and extent of exchange. Although no selective extraction of either LC 17 or RLC was observed after 1 mM TFP treatment, subsequent incubation with recombinant LC 17 led to significant exchange of recombinant for endogenous LC 17 as determined by immunoprecipitation (Fig. 5A) or sedimentation of myosin (Fig. 5B). The amount of recombinant LC 17b associated with myosin heavy chain, expressed as percent of total LC 17 , was 50 Ϯ 1% (n ϭ 6) in the rabbit bladder and 44 Ϯ 5% (n ϭ 6) in the chicken amnion. No LC 17b was detectable in myosin precipitates (Fig.  5B) or in the final wash of immunoprecipitates if the TFP solution was omitted from the exchange protocol. Further evidence of TFP facilitating LC 17 exchange was obtained using confocal microscopy. Incubation with fluorescently labeled LC 17b according to the exchange protocol followed by extensive washing revealed a diffuse cytoplasmic signal, excluding nuclei and the extracellular matrix (Fig. 6A), whereas in paired experiments without TFP pretreatment the fluorescent signal obtained with identical detection parameters was markedly less intense (Fig. 6B). These findings indicate that added LC 17 can freely diffuse into and out of the filament lattice, but binds to the myosin heavy chain only after TFP treatment. Similar results were found for the RLC, where ϳ50% of total RLC was exchanged by recombinant RLC (data not shown). This was easily discernible, as the recombinant and endogenous nonphosphorylated RLCs ran separately on one-dimensional isoelectric focusing gels.
The Effects of Light Chain Exchange on Contractile Properties-Surprisingly, in all smooth muscles treated with TFP, incubation with either recombinant LC 17a or LC 17b in relaxing solution (5 mM EGTA, 2 mM MgATP) produced large (ϳ30% of initial high K ϩ -induced force), but slow, Ca 2ϩ -independent force generation (Fig. 7A). In approximately 50% of strips examined, transfer to relaxing solution without added LC 17 also induced a small contraction (9 Ϯ 1% of initial high K ϩ ; n ϭ 9/19 fibers), but subsequent addition of LC 17a or LC 17b caused force development irrespective of whether this smaller, initial con- traction was present (compare Fig. 7, A and B). Both contractions could be promptly relaxed by 3 mM vanadate (H 2 VO 4 Ϫ ; Fig.  7B), indicating that force was generated by cross-bridges. No significant increase in phosphorylation of endogenous RLC above basal levels was detectable at the peak of LC 17 -induced contraction (Table I), whereas in paired, untreated strips, Ca 2ϩ -induced contractions of a similar magnitude induced the expected increase in RLC phosphorylation. An inhibitor of myosin light chain kinase, ML-9 (300 M), had no effect on the contraction induced by exogenous LC 17 in TFP-treated muscles (n ϭ 2; data not shown), although it relaxed control fibers contracted with pCa 5.0. These findings imply that the function of the RLC has been altered, leading to a loss of its repressor role allowing the unregulated head with the functional, exchanged LC 17 to generate force upon addition of MgATP. Addition of nonphosphorylated RLC relaxed both the LC 17 -induced force (data not shown) and the initial, variant small contraction sometimes observed upon transfer to relaxing solution (Fig. 7C).
Muscles preincubated with nonphosphorylated RLC failed to develop force in response to LC 17 (data not shown), suggesting that the ability of nonphosphorylated RLC to switch off actomyosin ATPase activity had been regained. Recombinant thiophosphorylated RLC produced a qualitatively similar contraction (Fig. 7D) to that observed in response to LC 17 . None of the recombinant light chain species had any effect on force if TFP was omitted from the protocol detailed under "Experimental Procedures," consistent with the premise that TFP treatment enables subsequent exchange of recombinant light chains.
The Differential Effects of Exchanged Light Chain Isoforms on Unloaded Shortening Velocity (V us )-We next wished to determine whether the V us propelled by nonphosphorylated, cycling cross-bridges during force generation induced by recombinant LC 17 (see above) were affected by the type of LC 17 (a or b) isoform. To avoid time-dependent changes in V us (25,26), we determined it from the linear component during the initial 150 ms, and found it to be significantly (p Ͻ 0.05) faster after treatment with LC 17a (0.27 Ϯ 0.03 lengths s Ϫ1 ) than with LC 17b (0.17 Ϯ 0.02 lengths s Ϫ1 ; Fig. 8 and Table II, rows a and c). In LC 17 -exchanged fibers, additional incubation with recombi-nant, thiophosphorylated RLC significantly increased V us , indicating that the relatively slow V us following LC 17 exchange was not due to rundown of the preparation. However, V us under these conditions was independent of the LC 17 isoform exchanged (Table II,

rows b and d).
Effects of Light Chain Exchange on the Rate of Force Development-Photolysis of caged ATP was utilized to compare the relative rates of force development and circumvent diffusional delays. Following incubation of TFP-treated strips with either LC 17a or LC 17b in the absence of ATP, the initial rate of force development following photolytic release of ATP, determined by the time taken for 20% of the maximal force to develop (t 20% ), was significantly (p Ͻ 0.01) faster in LC 17a -treated muscles than in the LC 17b group (Table III). The maximal force generated after flash photolysis was not significantly different between the two groups.
Morphology-Electron microscopy of TFP-treated muscles viewed in transverse section showed a normal distribution of myosin filament arrays across the muscle bundle, consistent with the gels of intact and permeabilized tissues showing retention of contractile proteins with the permeabilization protocol utilized. Cell packing and outlines were normal. DISCUSSION The major findings of our study are that: 1) essential light chain exchange with the acidic isoform, LC 17a , results in a significantly faster V us and rate of force development by nonphosphorylated cross-bridges compared with exchange with the basic isoform, LC 17b ; this is the first demonstration of an effect solely attributable to LC 17 isoforms in situ; 2) in the presence of thiophosphorylated RLC, shortening velocities of TFP-treated fibers are restored and this effect is independent of the type of LC 17 isoform exchanged; and 3) TFP can directly affect the properties of LC 17 and RLCs in vitro. The major advantage of the current approach is that it allows a direct assessment of the mechanical effects of isoform exchange within the physiological context of organized, oriented, and strained cross-bridges capable of developing force and may be a useful model system for investigating cycling of nonphosphorylated cross-bridges in situ. Following exposure to 1 mM TFP, endogenous RLC and LC 17 remained stoichiometrically associated with myosin heavy chain both in situ (Fig. 1B) and in vitro (Fig. 2). However, both light chains appeared to be modified: RLC could no longer be thiophosphorylated in situ (Fig. 4A) and the endogenous LC 17 could now exchange with recombinant LC 17 (Fig. 5). The refractoriness of RLC to thiophosphorylation was not due to inhibition of calmodulin by TFP (remaining in skinned fibers), because TFP also inhibited phosphorylation of isolated myosin (Fig. 4B) by a Ca 2ϩ /calmodulin-independent myosin light chain kinase (42). Similarly, direct modification of LC 17 by TFP was demonstrated by its effect on the CD spectrum (Fig. 3). Our results suggest that lower (0.1-1 mM) concentrations of TFP directly interact with LC 17 and RLC, also increasing their exchangeability in situ. More prolonged treatment and/or high concentrations (e.g. 5 mM) of TFP remove light chains from the heavy chain (14,16,27).

FIG. 6. Confocal micrograph showing longitudinally oriented muscle cells in rabbit bladder strips after Triton permeabilization and incubation with tetramethylrhodamine B isothiocyanate (TRITC)-labeled LC 17b , with (Panel
We suggest that the exchange of recombinant with endogenous, TFP-modified LC 17 (present study) enabled the lever arm to convert unregulated actomyosin ATPase activity into active tension, possibly by stiffening the regulatory domain of S1. Transient electrical birefringence studies of skeletal S1 suggest that essential light chain and RLC stiffen S1 (28), and removal of light chains reduces the force developed by isolated myosin molecules (29). Thiophosphorylation of RLC increases rigor stiffness of permeabilized smooth muscle, consistent with a regulatory effect of light chains on the mechanical properties of the myosin heavy chain (43).
The effects of both thiophosphorylated and nonphosphory-   8. Plot of (l/l i ) versus time obtained by the slack test in rabbit bladder strips after Triton permeabilization, TFP treatment, and incubation with recombinant LC 17 isoforms. Following TFP treatment, muscles were transferred to zero Ca 2ϩ -rigor solution plus LC 17a or LC 17b and equilibrated for 60 min before transfer to relaxing solution. V us measurements were made once the contraction plateau was reached. LC 17a (⅜), V us ϭ 0.27 Ϯ 0.02 lengths s Ϫ1 , n ϭ 8; LC 17b (q), V us ϭ 0.17 Ϯ 0.02 lengths s Ϫ1 , n ϭ 8. See also Table II, rows a and c. Force, normalized to the initial high K ϩ contracture in the intact muscle, was 0.30 Ϯ 0.03 (LC 17a ) and 0.34 Ϯ 0.03 (LC 17b ); l i , initial length; l, applied length change. lated recombinant RLC on force can also be explained by a process of exchange of recombinant with endogenous, TFPmodified RLC. Exchange with the thiophosphorylated species would result in constitutively active actomyosin ATPase activity which correlates with the observed force generation, whereas the inhibitory effect of nonphosphorylated RLC on force is the expected result of restoration of regulation of contraction, consistent with the view that nonphosphorylated RLC is a repressor (Ref. 14 and reviewed in Ref. 30). The formation of nonphosphorylated, slowly cycling cross-bridges has also been demonstrated in permeabilized, single smooth muscle cells (16), from which calponin had been extracted; however, we did not find detectable loss of calponin from bundles of smooth muscle (present study).
The exchange of recombinant LC 17b for endogenous LC 17a was verified by co-immunoprecipitation and co-sedimentation with myosin heavy chains and ranged between 40 and 60%. In tonic smooth muscles, such as rabbit femoral artery and aorta and porcine aorta, expression of LC 17b ranges between 43 and 58% of total LC 17 content (10,11,31). Therefore, the level of exchange achieved in the present study was consistent with the LC 17a /LC 17b ratio that correlates with the lower ATPase activity and slower V us and rates of force development by tonic smooth muscle (10,11,31).
The V us is thought to be rate-limited by the dissociation of MgADP from cross-bridges (32). MgADP affinity for actomyosin cross-bridges is higher in tonic, than in phasic, smooth muscle (8,10), and slower cycling can be attributed to the slower off-rate of MgADP from cross-bridges. This greater sensitivity of tonic smooth muscle to MgADP correlates with a higher relative content of the LC 17b isoform. Therefore, we suggest that the approximately 1.5-fold difference in V us between LC 17a -and LC 17b -induced contractions is due to the relatively greater MgADP affinity conferred by the LC 17b iso-form upon these (nonphosphorylated) cross-bridges that also contain the heavy chain insert, but in which the LC 17 isoform was the only variable modified in our study. Although the velocities obtained under conditions of LC 17 exchange are relatively slow, this does not reflect a deterioration and rundown of the muscle fibers, since additional incubation with recombinant, thiophosphorylated RLC significantly increased the V us of LC 17 -exchanged fibers. Our results are in agreement with Lowey et al. (33), who showed that both essential light chain and RLC of skeletal myosin were required for propelling actin filaments at maximal sliding velocity. The absence of a difference in V us between LC 17a -and LC 17b -treated fibers containing thiophosphorylated recombinant RLC is consistent with the finding that complete replacement of LC 17a with the LC 17b isoform does not change the in vitro motility of thiophosphorylated heavy meromyosin containing the 7-amino acid insert (13). However, unlike these thiophosphorylated myosins, under physiological conditions, even at maximal steady-state force, RLC phosphorylation is in the range of only 20 -30% or even lower in smooth muscles.
The initial rate of force development (time to reach 20% of maximal force), thought to be determined by the rate of P i , but not ADP, release (reviewed in Ref. 34), was approximately two times faster in LC 17a -, than LC 17b -exchanged muscles. This may have been due to a difference in the rate of internal shortening affecting the rate of force development, or, because contractions were initiated from a rigor state, due to an effect of MgADP slowing the detachment of rigor bridges (20) prior to transition into force-generating states (20,35). Essential light chain-deficient myosins have a reduced unloaded duty cycle, reflecting altered rates of myosin attachment and detachment from actin (29). It is worth noting that the rate of force development by the nonphosphorylated cross-bridges (present study) was considerably slower than the rate of force developed by phosphorylated or thiophosphorylated cross-bridges in phasic smooth muscles (7). This finding is consistent with the view that: 1) nonphosphorylated cross-bridges can cycle, and 2) they do so at slower rates than phosphorylated cross-bridges.
Although results obtained from cycling cross-bridges which are either maximally thiophosphorylated or nonphosphorylated could be considered "nonphysiological," a physiological latch or catch-like state of tonic smooth muscles is thought to reflect slow cycling nonphosphorylated cross-bridges (Refs. 36 -38 and reviewed in Refs. 39 and 40). Since myosin in tonic smooth muscles contains a greater proportion of LC 17b , has a lower K D for MgADP, and lacks the heavy chain insert, the present finding that the LC 17 isoforms can modulate the V us of nonphosphorylated, cycling cross-bridges may have implications concerning the molecular origin(s) of the latch or catchlike state. Furthermore, under most physiological conditions RLC are only partially phosphorylated, and cooperative cycling of nonphosphorylated cross-bridges under these conditions (20,41) could also be modulated by the LC 17 isoform. We propose that the coordinate expression of a particular LC 17 isoform and the presence or absence of the 7-amino acid insert loop near the catalytic site of the motor domain, leads to the different affinities for MgADP and underlies the tonic (slow) or phasic (fast) contractile phenotypes of smooth muscle (1).

TABLE II
Unloaded shortening velocity in Triton-permeabilized, TFP-treated rabbit bladder strips following exchange of light chains After TFP treatment, muscles were treated with recombinant light chains followed by activation of force by transfer to zero Ca 2ϩ ATP containing solution and determination of V us , as described under "Experimental Procedures." Rows a and c, incubation in rigor solution containing LC 17a or LC 17b (0.75 mg⅐ml Ϫ1 ) for 60 min; rows b and d, incubation in LC 17a or LC 17b , respectively, followed by incubation in rigor solution containing recombinant thiophosphorylated RLC (0.7 mg⅐ml Ϫ1 ) for 45 min. Numbers in parentheses refer to number of data points . Rows a versus b, a versus c, and c