INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

CONTRACTION OF SMOOTH MUSCLE in response to excitatory neurotransmitters, hormones, or depolarization follows a widely recognized pathway: increased myoplasmic Ca²⁺ concentration ([Ca²⁺]_i), formation of a Ca²⁺-calmodulin complex, activation of myosin light chain kinase, and phosphorylation of myosin regulatory light chains (MRLC) on serine (Ser)-19 (Ser¹⁹) (12). Phosphorylation of MRLC enables cross-bridge attachment to the thin filament, thereby allowing cross-bridge cycling and force generation (10). In many cases, smooth muscle relaxation proceeds via a reversal of this contraction process: reduction of [Ca²⁺]_i, inactivation of myosin light chain kinase, and dephosphorylation of MRLC (9, 22, 29).

Smooth muscle relaxation induced by increases in cAMP or cGMP appears to be more complex. In submaximally contracted swine carotid media, addition of forskolin (to increase cAMP concentration) (20) or nitroglycerin (to increase cGMP concentration) (18) induced relaxations that were associated with both initial and sustained reductions in [Ca²⁺]_i and MRLC phosphorylation without a significant alteration in the Ca²⁺ sensitivity of MRLC phosphorylation. These findings suggest that cAMP or cGMP induced relaxation of submaximally stimulated tissues primarily by reductions in [Ca²⁺]_i-dependent MRLC phosphorylation.

When maximally contracted swine carotid media was treated with nitroglycerin, there was a rapid relaxation transient that was associated with a decrease in [Ca²⁺]_i and MRLC phosphorylation (18). However, low force persisted despite a return of [Ca²⁺]_i and MRLC phosphorylation to high levels indistinguishable from the levels observed in maximally contracted tissue (18). This phenomenon has been termed "relaxation without MRLC dephosphorylation," "the sustained inhibitory response," and "uncoupling of force from MRLC phosphorylation." It is best characterized as a rightward shift in the MRLC phosphorylation-force relation (18, 19, 24). This response is observed with activators of guanylyl cyclase, such as nitric oxide, and with phosphodiesterase inhibitors that increase intracellular cGMP concentration (6). Other treatments that induce relaxation without MRLC dephosphorylation include okadaic acid (27), some Ca²⁺ depletion protocols (8), Ca²⁺ channel blockers (14), high extracellular Mg²⁺ concentration (7), and other combinations of excitatory and inhibitory stimuli (2). Although this phenomenon was studied primarily in vascular smooth muscle, it is most prominent in tissues such as the corpus cavernosum, where inhibitory innervation is a key physiological control mechanism (6).

One potential explanation for sustained low tone without MRLC dephosphorylation could be phosphorylation of MRLC on an amino acid residue that does not activate cross-bridge cycling. However, we found that nitroglycerin phosphorylated MRLC entirely on Ser¹⁹ during nitroglycerin-induced relaxation (18) and high extracellular Mg²⁺ concentration-induced relaxation (7). Ser¹⁹ is the phosphorylation site that regulates myosin ATPase activity (21). This result indicates that the phosphorylation of MRLC at sites other than Ser¹⁹ cannot explain relaxation without proportional MRLC dephosphorylation. Indeed, such data clearly point to the existence of mechanisms that can block force generation by phosphorylated cross bridges in smooth muscle.

Recently, the sustained phase of cAMP- and cGMP-dependent relaxation was associated with phosphorylation of heat shock protein 20 (HSP20) on Ser¹⁶ (3, 4, 24). Flow-dependent vasodilation, a process mediated by nitric oxide, was also associated with HSP20 phosphorylation (13). We found that a peptide from HSP20 had a sequence homology with troponin I and that this peptide bound to thin filaments, reduced actin-activated myosin S1 ATPase activity, and relaxed skinned smooth muscle (24). We hypothesized that binding of Ser¹⁶-phosphorylated HSP20 to the thin filament turned "off" thin filaments so that phosphorylated myosin was unable to interact with the thin filament (i.e., a model similar to skeletal muscle troponin I). This would explain low force with elevated MRLC phosphorylation. Two predictions of this hypothesis were tested in this study.

1) If Ser¹⁶ phosphorylation of HSP20 is important in relaxation without MRLC dephosphorylation, then Ser¹⁶-HSP20 phosphorylation should correlate with the sustained inhibitory phase of relaxation rather than the initial relaxation transient. Furthermore, when the relaxing agent is removed, Ser¹⁶-HSP20 dephosphorylation should precede the force redevelopment.

2) If Ser¹⁶ phosphorylation of HSP20 is the only factor involved in relaxation without MRLC dephosphorylation, there should be an unique relation between Ser¹⁶-HSP20 phosphorylation and the sustained phase of relaxation regardless of whether cAMP or cGMP concentration is increased.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Tissues. Swine common carotid arteries were obtained from a slaughterhouse and transported at 0°C in physiological salt solution. Physiological salt solution contained (in mM) 140 NaCl, 4.7 KCl, 5 MOPS, 1.2 Na₂HPO₄, 1.6 CaCl₂, 1.2 MgSO₄, and 5.6 D-glucose, pH adjusted to 7.4 at 37°C. Dissection of medial strips, mounting, and determination of the optimum length for stress development at 37°C were performed as previously described (25). The intimal surface was mechanically rubbed to remove the endothelium.

Antibodies. Rabbit anti-HSP20 antibody was made commercially via repeated injection of gel-purified recombinant HSP20 (sequence confirmed by mass spectroscopy). After confirmation of an antigenic response, serum was collected and frozen for future use. Figure 1 demonstrates the specificity of the HSP20 antibody. Preincubation of the HSP20 antibody with recombinant HSP20 abolished immunostaining of a blot containing swine carotid HSP20.

View larger version (51K): [in this window] [in a new window]   Fig. 1.   Specificity of the anti-heat shock protein 20 (HSP20) antibody. Homogenates of a swine carotid media (stimulated with 10 µM histamine for 10 min and then relaxed by addition of 10 µM nitroglycerin for 20 min) were separated by isoelectric focusing and blotted to nitrocellulose paper. A: incubated in 1:10,000 anti-HSP20 antibody and developed with enhanced chemiluminescence (ECL). A shows an upper HSP20 band not phosphorylated [isoelectric point (pI) 6.3], a central band phosphorylated on a protein kinase C (PKC) site but not phosphorylated on Ser16 (pI 6.0), and a lower HSP20 band phosphorylated on both Ser16 and the PKC site (pI 5.7). On some blots, a minor fourth band was seen just below the central band at pI 6.0, which may represent phosphorylation only on Ser16. B: incubated in 1:10,000 anti-HSP20 antibody that was preincubated with 10 µg of recombinant human HSP20 (equivalent to 0.6 µM HSP20) for 30 min at 22°C and then developed with ECL. B shows lack of HSP20 immunoreactivity when antibody was preincubated with HSP20 antigen, demonstrating antibody specificity for HSP20.

Measurement of HSP20 and MRLC phosphorylation. Swine carotid arteries were pharmacologically treated and frozen in an acetone-dry ice slurry (25). After air drying, the tissues were homogenized in a buffer containing 1% SDS, 10% glycerol, and 20 mM dithiothreitol (20 mg wet wt/ml buffer). Full-strength, half-strength, and quarter-strength dilutions of samples were then separated on one-dimensional isoelectric focusing gels [ampholytes were a 50:50 mixture of isoelectric point (pI) 5-8 and pI 4-6.5 for HSP20 and a 50:50 mixture of pI 4.5-5.4 and pI 4.0-6.5 for MRLC], blotted to nitrocellulose, immunostained with our rabbit polyclonal anti-HSP20 antibody (1:5,000) or rabbit polyclonal anti-MRLC antibody (1:4,000 in 1% bovine serum albumin and 0.01% sodium azide), and detected with enhanced chemiluminescence (26). The dilutions ensured that the enhanced chemiluminescence detection system was in the linear range (26). Immunoblots were scanned on an Hewlett Packard flatbed scanner and quantitated with UNSCANIT software.

HSP20 has at least two phosphorylation sites. In the unstimulated swine carotid, our laboratory (24) found that over 90% of immunoreactive HSP20 was present in a band at pI 6.0 and that nitroglycerin-induced relaxation was associated with migration of some of the HSP20 immunostaining to a band at pI 5.7 [pI identification as in (4)]. Our laboratory interpreted this pI shift as a phosphorylation reaction. A less prominent band at pI 6.3 was also seen (Fig. 1). It was also previously reported (24) that the HSP20 band at pI 5.7 was phosphorylated at Ser¹⁶ based on mass spectroscopy sequencing (sequencing did not rule out additional phosphorylation at another site).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   HSP20 has multiple phosphorylation sites. Swine carotid artery tissues were stimulated with 10 µM histamine for 10 min and then relaxed by addition of 10 µM forskolin for 30 min. After being frozen in an acetone-dry ice slurry and thawed, tissues were homogenized in 100 mM Tris · HCl, 150 mM NaCl, 0.25% sodium deoxycholate, and 1.0% Tergitol (type NP-40). The homogenates were then treated with calf alkaline phosphatase (38°C for 30 min; A), vehicle (38°C for 30 min; B), or vehicle on ice (C). Proteins were then separated by isoelectric focusing, transferred to nitrocellulose, incubated in 1:5,000 anti-HSP20 antibody, and developed with amplified Opti-4CN (Bio-Rad). Treatment with calf alkaline phosphatase collapsed immunostaining to a single basic band at pI 6.3.

Statistics. Phosphorylation and force were compared with an ANOVA with Newman-Keuls tests for individual comparisons. Significance was defined as P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Time course of nitroglycerin-induced relaxation and HSP20 phosphorylation. If Ser¹⁶-HSP20 phosphorylation regulates the sustained inhibitory phase of relaxation, then the time course of Ser¹⁶-HSP20 phosphorylation should correlate with the sustained inhibitory phase. Maximal stimulation of swine carotid media with 10 µM histamine increased suprabasal MRLC phosphorylation and contractile stress without significantly increasing Ser¹⁶-HSP20 phosphorylation (Fig. 3). Force and MRLC phosphorylation decreased rapidly after addition of 10 µM nitroglycerin. One minute after addition of nitroglycerin, Ser¹⁶-HSP20 phosphorylation was not significantly increased. These data demonstrate that Ser¹⁶-HSP20 phosphorylation cannot account for the initial nitroglycerin-induced relaxation transient. Previous work in our laboratory (18) demonstrated that the initial phase of nitroglycerin-induced relaxation was associated with significant reductions in [Ca²⁺]_i and MRLC phosphorylation. Therefore, the initial phase of relaxation appears to be caused by a reduction in [Ca²⁺]_i-dependent MRLC phosphorylation.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   HSP20 phosphorylation (Phosphor) (shown in B) correlated with the sustained inhibitory phase of the nitroglycerin (NTG) response. Seven sets of swine carotid artery tissues were stimulated with 10 µM histamine alone for 30 min. Nitroglycerin (10 µM) was then added for 15 min and then washed out in the continued presence of histamine. A tissue from each set was frozen at 0 min for the unstimulated control, at 30 min with 10 µM histamine alone, and at 31, 32, and 45 min with 10 µM histamine plus nitroglycerin for 1, 2, and 15 min, respectively. After washout of nitroglycerin, the tissues were monitored for force redevelopment and were frozen when force redeveloped by 33% (actual time = 55.3 ± 1.6 min, 10.3 min after nitroglycerin washout), when force redeveloped by 67% (actual time = 60.7 ± 2.0 min, 15.7 min after nitroglycerin washout), and when force was stable (actual time = 74.8 ± 2.6 min, 24.8 min after nitroglycerin washout). After freezing, tissue homogenates were assayed for myosin regulatory light chains (MRLC; A) Ser16-HSP20 (B) phosphorylation. For clarity, the first 30 min of histamine stimulation are abridged. The lines connecting control and 30 min histamine are typical for swine carotid artery, based on previous data (25). C: force was normalized to that elicited by 109 mM extracellular K+ (K109) in the tissues. Results are means ± SE (n = 7 sets of tissues). Symbols are without error bars when SE was smaller than the symbol size.

Comparison of nitroglycerin- and forskolin-induced relaxation. Ser¹⁶-HSP20 phosphorylation was minimal at 0.006 ± 0.004 mol P_i/mol HSP20 in unstimulated tissues (Fig. 4). Stimulation with a maximal concentration of histamine (10 µM) for 30 min contracted the tissues, increased suprabasal MRLC phosphorylation by 0.24 ± 0.04 mol P_i/mol MRLC, and did not significantly elevate Ser¹⁶-HSP20 phosphorylation. Addition of 10 µM nitroglycerin for the last 20 min of the histamine treatment relaxed the tissues and increased Ser¹⁶-HSP20 phosphorylation to 0.19 ± 0.02 mol P_i/mol HSP20, without significantly changing suprabasal MRLC phosphorylation.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Forskolin induces an inhibitory response that correlates with Ser16-HSP20 phosphorylation but not MRLC phosphorylation. Five sets of swine carotid artery tissues were either 1) not stimulated (Cont), 2) stimulated with 10 µM histamine alone for 30 min (Hist), 3) stimulated with 10 µM histamine for 10 min and then relaxed by addition of 10 µM nitroglycerin for 20 min (H+NTG), or 4-7) stimulated with 10 µM histamine for 10 min and then relaxed by addition of 0.1, 0.3, 1.0 or 1.0 µM forskolin for 20 min. Contractile force was then measured, after which tissues were frozen for MRLC (A) and Ser16-HSP20 (B) phosphorylation measurements. C: force was normalized to that elicited with 109 mM extracellular K+ depolarization. Results are means ± SE (n = 5 set of tissues).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Dependence of contractile force on Ser16-HSP20 phosphorylation in tissues maximally stimulated with histamine. Individual tissue data from the experiments shown in Fig. 4 demonstrate the relation between Ser16-HSP20 phosphorylation (plotted on a log scale) and contractile stress in tissues that were 1) stimulated with 10 µM histamine alone (), 2) stimulated with 10 µM histamine for 10 min and then relaxed by addition of 10 µM nitroglycerin for 20 min (), or 3) stimulated with 10 µM histamine for 10 min and then relaxed by addition of 0.1, 0.3, 1.0, or 1.0 µM forskolin for 20 min (). Overall, there was a good inverse correlation between Ser16-HSP20 phosphorylation and contractile force with both forskolin- and nitroglycerin-induced relaxation (R2 = 0.81).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Previously, smooth muscle relaxation was hypothesized to be a single process involving reversal of the contractile process. This interpretation was based on a paradigm derived from skeletal muscle in which relaxation is the reversal of excitation. This paradigm appears to explain smooth muscle relaxation resulting from withdrawal of excitatory stimuli. In these cases, reduction in [Ca²⁺]_i-dependent MRLC phosphorylation precedes the reduction in force (9, 22, 29).

An important finding of this study is that "relaxation" of smooth muscle can involve two distinct phenomena with different regulatory mechanisms. Specifically, relaxation induced by agents that elevate cyclic nucleotides had two temporally distinguishable phases. The first is a rapid transient reduction in force characterized by reductions in [Ca²⁺]_i-dependent MRLC phosphorylation (Fig. 3 and Ref. 18). The second element is manifested as a sustained inhibition of force that occurs despite return of MRLC phosphorylation to preinhibition levels. This is the "uncoupling of force from MRLC phosphorylation" noted in a previous study (1). It is this second element, sustained inhibition of force redevelopment, that may be mechanistically associated with Ser¹⁶-HSP20 phosphorylation. We tested two predictions of this hypothesis.

First, if Ser¹⁶ phosphorylation of HSP20 inhibits force during the sustained phase of cyclic nucleotide-induced relaxation, then 1) HSP20 should be phosphorylated during the sustained inhibition of force when MRLC phosphorylation is increasing and 2) Ser¹⁶-HSP20 dephosphorylation should precede a recontraction induced by removal of the relaxing agent. Such a temporal correlation is one of the criteria described by Krebs and Beavo (15) in demonstrating the physiological relevance of a phosphorylation reaction. We found 1) that HSP20 became phosphorylated between 1 and 2 min after addition of nitroglycerin when MRLC phosphorylation began to increase and 2) that Ser¹⁶ dephosphorylation of HSP20 preceded force redevelopment induced by washing out nitroglycerin in the continued presence of histamine (Fig. 3). Suprabasal MRLC phosphorylation values did not significantly change during the force redevelopment, suggesting that changes in MRLC phosphorylation were not responsible for the change in force. These data are consistent with the hypothesis that Ser¹⁶-HSP20 phosphorylation is responsible for the nitroglycerin-induced sustained inhibition of force that occurs without MRLC dephosphorylation.

Importantly, Ser¹⁶-HSP20 phosphorylation did not correlate with the initial relaxation transient during which MRLC phosphorylation decreased significantly (Fig. 3). Previously, our laboratory (18) showed that nitroglycerin initially reduced [Ca²⁺]_i and MRLC phosphorylation in the swine carotid. These data suggest that nitroglycerin induced an initial relaxation transient that was mechanistically the reversal of excitation.

Second, if Ser¹⁶ phosphorylation of HSP20 regulates the sustained phase of force inhibition, there should be a unique relation between steady-state Ser¹⁶-HSP20 phosphorylation and reduction in force regardless of whether either cAMP or cGMP concentration is increased. We found that forskolin- and nitroglycerin-dependent increases in Ser¹⁶-HSP20 phosphorylation correlated with the sustained phase of force inhibition (Fig. 5). This is consistent with the hypothesis that Ser¹⁶-HSP20 phosphorylation regulates the sustained phase of both cAMP- and cGMP-dependent relaxation without MRLC dephosphorylation. Importantly, Figs. 3 and 5 demonstrate a temporal and quantitative relationship between Ser¹⁶-HSP20 phosphorylation and the sustained phase of force inhibition.

The results shown in Fig. 5 were plotted on a semilogarithmic scale for clarity. Further work is required to define the quantitative relation between Ser¹⁶-HSP20 phosphorylation and force inhibition. There was a suggestion that nitroglycerin relaxed the tissues slightly more for a given level of Ser¹⁶-HSP20 phosphorylation than that observed with forskolin (as shown in Fig. 5). However, other preliminary experiments did not support this finding (O'Connor and Rembold, unpublished observations). Forskolin also relaxed tissues more than nitroglycerin. Higher concentrations of nitroglycerin did not induce complete relaxation (data not shown), probably because the enzymes that metabolize nitroglycerin to nitric oxide were saturated.

Some other stimuli, such as elevated extracellular Mg²⁺ concentration, induce smooth muscle relaxation without MRLC dephosphorylation (7). In the case of elevated Mg²⁺ concentration, the relaxation was not associated with Ser¹⁶-HSP20 phosphorylation (26). These data suggest that that Ser¹⁶-HSP20 phosphorylation cannot fully explain the phenomenon of relaxation without MRLC dephosphorylation and that Ser¹⁶-HSP20 phosphorylation is only one of perhaps multiple mechanisms that cause relaxation without MRLC dephosphorylation.

It is well established, both by our work (20) and others (16, 17), that cyclic nucleotides can induce relaxation by reducing [Ca²⁺]_i. The response depends on the relative concentrations of the excitatory stimulus and the inhibitory agent (18). It appears that reduction of [Ca²⁺]_i and MRLC phosphorylation suffice to the response of submaximal activated carotid media to forskolin or nitrovasodilators. Thus the precise response may be variable, depending on the tissue preparation, the type and concentration of the excitatory agonist, the type and concentration of inhibitory agent, and the time of measurement. We hypothesize that the mechanisms that maintain elevated [Ca²⁺]_i are more susceptible to inhibition when smooth muscle is submaximally activated with submaximal increases in [Ca²⁺]_i. Maximal activation may more profoundly stimulate Ca²⁺ influx mechanisms so that increases in cAMP and/or cGMP concentration cannot substantially reduce [Ca²⁺]_i. It is possible that the "relaxation without MRLC dephosphorylation" mechanism exists so that force can be reduced even if cyclic nucleotides cannot reduce [Ca²⁺]_i and MRLC phosphorylation. On the basis of data present in this and previous studies (4, 24, 26), we propose that Ser¹⁶-HSP20 phosphorylation may be such a mechanism.

Correlation of Ser¹⁶-HSP20 phosphorylation and inhibition of force does not definitively establish cause and effect. Therefore, we must ask whether there is a plausible mechanism for HSP20 interfering with force generation. There are data suggesting that HSP20 binds to actin filaments (5, 24). We noted that a region of HSP20, contained in the peptide HSP20_110-121, had a sequence homology similar to troponin I, the major thin filament regulatory protein in skeletal and cardiac muscle. HSP20_110-121 both reduced actin-activated myosin S1 ATPase activity and prevented contraction of skinned swine carotid media, suggesting that HSP20 amino acid residues 110-121 may be important in force regulation (24). These data suggest that HSP20 may regulate force via an interaction with thin filaments.

Our working hypothesis, shown in Fig. 6, states that HSP20, phosphorylated on Ser¹⁶, binds to and/or alters the conformation of thin filaments to inhibit attachment of phosphorylated cross bridges. If this were the case, increasing Ser¹⁶-HSP20 phosphorylation should reduce the amount of active force that can be generated at a given level of MRLC phosphorylation. This action could be cooperative such that a modest elevation in Ser¹⁶-HSP20 phosphorylation could maintain low tone. We evaluated force in response to 10 µM histamine, a concentration that induces a maximal contraction. We found that the amount of Ser¹⁶-HSP20 phosphorylation correlated with the maximal sustained contraction observed in the presence of 10 µM histamine (Fig. 5). These data support the hypothesis that Ser¹⁶-HSP20 phosphorylation is a mechanism responsible for relaxation without MRLC dephosphorylation. Figure 5 can also be interpreted to suggest that the level of Ser¹⁶-HSP20 phosphorylation determined the maximal level of force generated with contractile stimuli. This suggestion leads to the hypothesis that Ser¹⁶-HSP20 phosphorylation reduces force by inactivating thin filaments so as to remove parallel contractile units.

View larger version (34K): [in this window] [in a new window]   Fig. 6.   Working hypothesis for inhibition of smooth muscle tone by Ser16-HSP20 phosphorylation. Left: currently accepted hypothesis for thick filament (i.e., myosin)-based regulation of smooth muscle contraction. An agonist, such as histamine, binds to a receptor and by various mechanisms (23) increases intracellular Ca2+ concentration. Ca2+ binds to calmodulin (CaM), activating myosin light chain kinase (MLCK), which phosphorylates myosin (M) on Ser19. Phosphorylated myosin (Mp) has a higher actin-activated ATPase activity, which allows attachment to the actin (A) containing thin filaments. Attached phosphorylated cross bridges (AMp) can either 1) rapidly cycle (reforming Mp) and cause muscle shortening or 2) be dephosphorylated while attached, forming force-maintaining latch bridges (AM). Latch bridges can either 1) slowly detach (forming M) or 2) be rephosphorylated to AMp and reenter the rapidly cycling pool. Right: our hypothesis for thin filament (i.e., actin)-based inhibition of smooth muscle force. Nitric oxide (NO) or nitrovasodilators activate soluble guanylyl cyclase producing cGMP. Hormones via receptors or forskolin activate adenyl cyclase producing cAMP. Both cAMP and cGMP appear to activate protein kinase G (PKG) in smooth muscle. PKG phosphorylates HSP20 on Ser16. We hypothesize that Ser16-phosphorylated HSP20 inactivates thin filaments. The result is fewer thin filaments available to cycle with phosphorylated myosin (Mp). Because fewer attached phosphorylated cross bridges (AMp) are formed, we also hypothesize that fewer latch bridges (AM) would be produced. Therefore, force (the sum of AMp and AM) is lower despite high MRLC phosphorylation. This model would predict that unloaded shortening velocity would not be affected by thin filament inactivation, since those phosphorylated cross bridges that attach to active thin filaments would produce normal unloaded shortening. Some regulatory pathways are oversimplified here for reasons of clarity. Myosin dephosphorylation is accomplished by myosin light chain phosphatase (MLCP). There are multiple regulatory systems for MLCP (11) that alter the Ca2+ sensitivity of MRLC phosphorylation. The respective importance of PKG vs. protein kinase A in the phosphorylation of HSP20 has not yet been determined. The phosphorylation-specific response of HSP20 at the thin filament level is not characterized: Ser16-HSP20 phosphorylation may cause a thin filament conformational change of previously bound HSP20 or may cause HSP20 association with (or even dissociation from) the thin filament. Finally, the identity of HSP20 phosphatase has not been identified.