Involvement of myosin regulatory light chain diphosphorylation in sustained vasoconstriction under pathophysiological conditions.

Smooth muscle contraction is activated primarily by phosphorylation at Ser19 of the regulatory light chain subunits (LC20) of myosin II, catalysed by Ca(2+)/calmodulin-dependent myosin light chain kinase. Ca(2+)-independent contraction can be induced by inhibition of myosin light chain phosphatase, which correlates with diphosphorylation of LC20 at Ser19 and Thr18, catalysed by integrin-linked kinase (ILK) and zipper-interacting protein kinase (ZIPK). LC20 diphosphorylation at Ser19 and Thr18 has been detected in mammalian vascular smooth muscle tissues in response to specific contractile stimuli (e.g. endothelin-1 stimulation of rat renal afferent arterioles) and in pathophysiological situations associated with hypercontractility (e.g. cerebral vasospasm following subarachnoid hemorrhage). Comparison of the effects of LC 20 monophosphorylation at Ser19 and diphosphorylation at Ser19 and Thr18 on contraction and relaxation of Triton-skinned rat caudal arterial smooth muscle revealed that phosphorylation at Thr18 has no effect on steady-state force induced by Ser19 phosphorylation. On the other hand, the rates of dephosphorylation and relaxation are significantly slower following diphosphorylation at Thr18 and Ser19 compared to monophosphorylation at Ser19. We propose that this diphosphorylation mechanism underlies the prolonged contractile response of particular vascular smooth muscle tissues to specific stimuli, e.g. endothelin-1 stimulation of renal afferent arterioles, and the vasospastic behavior observed in pathological conditions such as cerebral vasospasm following subarachnoid hemorrhage and coronary arterial vasospasm. ILK and ZIPK may, therefore, be useful therapeutic targets for the treatment of such conditions.

The central role of myosin regulatory light chain phosphorylation in the activation of smooth muscle contraction

Vascular smooth muscle contraction is activated by an increase in cytosolic free Ca2+ concentration ([Ca2+]i) as a result of Ca2+ entry from the extracellular space and/or Ca2+ release from intracellular stores, primarily the sarcoplasmic reticulum (1). Ca2+ diffuses to the contractile machinery where it binds to calmodulin (CaM) (2). The (Ca2+)4-CaM complex induces a conformational change in myosin light chain kinase (MLCK), which involves removal of the autoinhibitory domain from the active site, thereby converting the kinase from an inactive to an active state (3). MLCK is physically bound through its N-terminus to actin filaments and, upon activation, phosphorylates nearby myosin molecules (4). Myosin II filaments are composed of hexameric myosin molecules, each consisting of two heavy chains and two pairs of light chains (17-kDa essential light chains (LC17) and 20-kDa regulatory light chains (LC20)) located in the neck region of the myosin molecule (Fig. 1). Activated MLCK phosphorylates Ser19 of LC20 and this simple post-translational modification induces a conformational change that is transmitted to the myosin heads, resulting in actin interaction and a marked increase in the actin-activated MgATPase activity of myosin (5). The energy derived from the hydrolysis of ATP then drives cross-bridge cycling and the development of force or contraction of the muscle. Relaxation follows the removal of Ca2+ from the cytosol, primarily by CaATPases, which pump Ca2+ out of the cell and back into the sarcoplasmic reticulum (6). MLCK is inactivated as Ca2+ dissociates from CaM and the autoinhibitory domain of MLCK blocks the active site. Phosphorylated myosin is then dephosphorylated by myosin light chain phosphatase (MLCP), a type 1 protein serine/threonine phosphatase (7).

Fig. 1.

Schematic representation of smooth muscle myosin II. Smooth muscle myosin II is a hexameric protein composed of two heavy chains (205 kDa each) and two pairs of light chains: the 17-kDa essential light chains (LC17) and the 20-kDa regulatory light chains (LC20). The N-termini of the heavy chains make up most of the head domains (pink) and the C-termini of the heavy chains account for the complete coiled-coil rod domain (which is responsible for assembly of myosin filaments) and the terminal unstructured tail (black). Each globular head includes an actin-binding site (blue) and an ATP-binding site (green). The light chains, LC17 (yellow) and LC20 (red), are associated with the neck region. "P" denotes phosphorylation of LC20. This figure was originally published in IUBMB Life. Walsh MP. Vascular smooth muscle myosin light chain diphosphorylation: mechanism, function and pathological implications. 2011; 63(11): 987–1000. © John Wiley & Sons, Inc.

Ca2+ sensitization

MLCP is a trimeric phosphatase with a 38-kDa catalytic subunit (PP1cδ), a 130-kDa regulatory subunit (MYPT1) and a 21-kDa subunit of uncertain function. MYPT1 targets the phosphatase holoenzyme to myosin and enhances its activity. MLCP is inhibited by both protein kinase C (PKC) and RhoA/Rho-associated kinase (ROK) pathways. Phosphorylation of the cytosolic phosphatase inhibitory protein of 17-kDa (CPI-17) at Thr38 by PKC converts CPI-17 to a potent inhibitor of MLCP, which is achieved by direct interaction of phosphorylated CPI-17 with PP1cδ (8). ROK is also capable of phosphorylating CPI-17 at Thr38 (9). In addition, and apparently of greater physiological significance, ROK phosphorylates MYPT1 to induce inhibition of MLCP activity. ROK phosphorylates MYPT1 at Thr697 and Thr855 (rat numbering) in vitro and both phosphorylation events result in phosphatase inhibition (10). However, it appears that ROK predominantly phosphorylates Thr855 in intact tissues (11). Activation of PKC and ROK pathways, therefore, can result in decreased MLCP activity, resulting in an increase in the ratio of MLCK: MLCP activity and an increase in force. Since ROK and novel PKC isoforms are Ca2+-independent, this results in Ca2+ sensitization of contraction, i.e. an increase in force without an increase in [Ca2+]i.

Myosin regulatory light chain diphosphorylation

The possibility that LC20 may also be phosphorylated at Thr18 was originally demonstrated in vitro when it was shown that high concentrations of MLCK phosphorylate Thr18 in addition to Ser19 (12). It is clear, however, that MLCK does not phosphorylate Thr18 of LC20 in intact smooth muscle tissues, and most contractile stimuli are associated with LC20 phosphorylation exclusively at Ser19, which can be attributed to MLCK (13). Interest in Thr18 phosphorylation was revived, however, when it was shown that treatment of smooth muscle tissues with membrane-permeant phosphatase inhibitors (such as calyculin-A) induced Ca2+-independent contractions that correlated with diphosphorylation of LC20 at Thr18 and Ser19 (14). Two Ca2+-independent kinases (integrin-linked kinase (ILK) and zipper-interacting protein kinase (ZIPK)) were shown to be the most likely kinases responsible for LC20 diphosphorylation (15,16,17).

Evidence of myosin regulatory light chain diphosphorylation in intact smooth muscle tissues

Phosphorylation of LC20 at Thr18 and Ser19 has been observed in several smooth muscle tissues, e.g. bovine tracheal smooth muscle in response to neural stimulation or carbachol (18, 19), rabbit thoracic aorta treated with prostaglandin-F2α (20, 21) and renal afferent arterioles stimulated with endothelin-1 (22). LC20 diphosphorylation has frequently been associated with pathophysiological conditions involving hypercontractility, including cerebral vasospasm (23), coronary (24, 25) and femoral arterial vasospasm (26), intimal hyperplasia (27) and hypertension (28).

The functional effects of diphosphorylation of myosin regulatory light chains

Early studies comparing the properties of smooth muscle myosin phosphorylated at Ser19 (by low concentrations of MLCK) and at both Thr18 and Ser19 (by high concentrations of MLCK) indicated that the additional phosphorylation at Thr18 increased the actomyosin MgATPase activity some two- to three-fold (12, 29,30,31). However, the velocity of movement of myosin-coated beads along actin cables (32) or of actin filaments over immobilized myosin in the in vitro motility assay (31) was similar whether LC20 was phosphorylated at Ser19 alone or at both Thr18 and Ser19. We recently addressed the hypothesis that phosphorylation at Thr18 may enhance the level of steady-state force achieved with Ser19 phosphorylation (13). To test this hypothesis, we used Triton-skinned rat caudal arterial smooth muscle strips. The objective was to achieve stoichiometric phosphorylation of LC20 at Ser19, then elicit phosphorylation at Thr18 and observe whether or not there was a further increase in steady-state force. The challenge was to achieve stoichiometric phosphorylation at Ser19 since this cannot be done in intact tissue due to the competing actions of MLCK and MLCP, which results in stable phosphorylation of LC20 at steady-state of ≤ 0.5 mol Pi/mol LC20. Furthermore, this problem cannot be overcome by inhibition of phosphatase activity since phosphatase inhibitors such as calyculin-A and okadaic acid unmask the basal activities of ILK and ZIPK, which phosphorylate both Thr18 and Ser19 of LC20. We took advantage of the fact that protein kinases generally, including MLCK, can utilize adenosine 5′-O-(3-thiotriphosphate) (ATPγS) as a substrate to thiophosphorylate their protein substrates (33), but the thiophosphorylated protein (LC20 in this case) is a very poor phosphatase substrate (34). It was necessary to use Triton-skinned tissue for this experiment since the plasma membrane is impermeant to ATPγS. Fig. 2 shows the experimental protocol and corresponding force measurements (Fig. 2A) and analysis of LC20 (thio)phosphorylation by Phos-tag SDS-PAGE (see below for a description of this technique, which enables the separation of unphosphorylated and phosphorylated forms of LC20) (Fig. 2B). At resting tension (pCa 9), LC20 is unphosphorylated (lanes 1 and 8, Fig. 2B). A control contraction was elicited by increasing [Ca2+] to pCa 4.5 in the presence of ATP and relaxation followed removal of Ca2+ (Fig. 2A). When basal force was restored, the tissue was washed several times in pCa 9 solution without ATP to remove all ATP, following which ATPγS was added at pCa 4.5 to elicit close-to-stoichiometric LC20 thiophosphorylation (Fig. 2B, lanes 2 and 3). Contraction did not occur under these conditions since ATPγS is not hydrolysed by the actin-activated myosin MgATPase and, therefore, ATPγS cannot support cross-bridge cycling (35,36,37). Following LC20 thiophosphorylation, ATPγS was washed out by several washes with pCa 9 solution (Fig. 2A). This was accompanied by very little dethiophosphorylation of LC20 (Fig. 2B, lane 4). ATP was then added at pCa 9, whereupon the tissue contracted rapidly due to the fact that LC20 was previously thiophosphorylated. Again there was very little dethiophosphorylation during this treatment (Fig. 2B, lane 5). The level of force achieved during this treatment was comparable to that elicited initially by addition of ATP at pCa 4.5 (Fig. 2A). Finally, the phosphatase inhibitor microcystin was added at pCa 9 in the presence of ATP, whereupon Ser19-thiophosphorylated LC20 was phosphorylated at Thr18 (1S1P in lanes 6 and 7 of Fig. 2B). Small amounts of monophosphorylated (1P) and diphosphorylated (2P) LC20 were also detected due to the presence (Fig. 2B, lane 5) of a small amount of unphosphorylated LC20 prior to the addition of microcystin. Fig. 2B, lane 9 shows a control Triton-skinned tissue treated with microcystin and ATP at pCa 9 to indicate the migration of unphosphorylated (0P), monophosphorylated (1P) and diphosphorylated (2P) LC20, as previously established (13). The key finding from this experiment was that phosphorylation of LC20 at Thr18 on top of close-to-stoichiometric thiophosphorylation of LC20 at Ser19 did not elicit an increase in force.

Fig. 2.

Stoichiometric thiophosphorylation of LC20 at Ser19 in Triton-skinned rat caudal arterial smooth muscle. (A) The viability of Triton-skinned rat caudal arterial smooth muscle strips was initially verified by transfer from relaxing solution (pCa 9) to high [Ca2+] (pCa 4.5) solution containing ATP and an ATP regenerating system (RS), which induced a contractile response. Tissues were then relaxed by 3 washes in pCa 9 solution containing ATP and RS. ATP was then removed by 6 washes in pCa 9 solution without ATP or RS. Tissues were then incubated in pCa 4.5 solution containing ATPγS (4 mM) in the absence of ATP and RS. Excess ATPγS was then removed by washing twice with pCa 9 solution without ATP or RS. Contraction was evoked by transfer to pCa 9 solution containing ATP and RS. Once steady-state force was established, microcystin (1 µM) was added in pCa 9 solution containing ATP and RS. Tissues were harvested at the indicated times during this protocol for Phos-tag SDS-PAGE and western blotting with anti-pan LC20 (B), as shown by the arrows in (A) (the numbers correspond to the lanes in (B)): (i) lanes 1 and 8, tissue incubated at pCa 9 showing exclusively unphosphorylated LC20; (ii) lanes 2 and 3, pCa 4.5 + ATPγS in the absence of ATP and RS; (iii) lane 4, pCa 9 in the absence of ATP and RS following thiophosphorylation; (iv) lane 5, at the plateau of force development following transfer to pCa 9 solution containing ATP and RS; (v) lanes 6 and 7, following treatment with microcystin at pCa 9 in the presence of ATP and RS. An additional control is included in lane 9: Triton-skinned tissue treated with microcystin at pCa 9 for 60 min to identify unphosphorylated (0P), monophosphorylated (1P), and diphosphorylated (2P) LC20 bands. Thiophosphorylated forms of LC20 are indicated as follows: 1SP, monothiophosphorylated LC20; 2SP, dithiophosphorylated LC20; 1SP1P, LC20 thiophosphorylated at one site and phosphorylated at the other. Data are representative of 8 independent experiments. This research was originally published in the Journal of Biological Chemistry. Sutherland C, Walsh MP. Myosin regulatory light chain diphosphorylation slows relaxation of arterial smooth muscle. 2012; 287(29): 24064–76. © the American Society for Biochemistry and Molecular Biology.

We next addressed the hypothesis that phosphorylation of LC20 at both Thr18 and Ser19 reduces the rate of dephosphorylation and relaxation compared to phosphorylation at Ser19 alone. To test this hypothesis, LC20 was phosphorylated at Ser19 only or at both Thr18 and Ser19 to comparable stoichiometry. This was achieved by treatment of Triton-skinned rat caudal arterial smooth muscle strips with ATP at pCa 4.5 (for monophosphorylation at Ser19) or with ATP and the phosphatase inhibitor okadaic acid at pCa 9 (for diphosphorylation at Thr18 and Ser19). The time-courses of dephosphorylation and relaxation were then followed to determine if there was a reduction in the rates of dephosphorylation and relaxation when LC20 was diphosphorylated at Thr18 and Ser19 compared to monophosphorylated at Ser19. As shown in Fig. 3A and C, both treatments induced LC20 phosphorylation to ∼0.5 mol Pi/mol LC20. At pCa 4.5, this was exclusively Ser19 phosphorylation, whereas in response to okadaic acid at pCa 9, phosphorylation occurred at both Thr18 and Ser19 (Fig. 3C). Comparison of the time-courses of dephosphorylation (Fig. 3A and C) and relaxation (Fig. 3B) clearly shows that both dephosphorylation and relaxation rates were significantly slower when LC20 was diphosphorylated at Thr18 and Ser19 compared to monophosphorylated at Ser19. This finding suggests a functional effect of LC20 diphosphorylation and raises the possibility that pathological situations of hypercontractility may result from impaired relaxation due to this dephosphorylation event. MLCP has been shown to dephosphorylate Thr18 in addition to Ser19 (38) suggesting that the same phosphatase dephosphorylates mono- and diphosphorylated LC20in situ.

Fig. 3.

Comparison of the time courses of relaxation and LC20 dephosphorylation in Triton-skinned rat caudal arterial smooth muscle following contraction with Ca2+ or okadaic acid in the absence of Ca2+. Triton-skinned tissues that had been contracted with Ca2+ (open circles) or okadaic acid (20 µM) at pCa 9 (closed circles) were transferred to pCa 9 solution and the time courses of dephosphorylation (A) and relaxation (B) were followed. Tissues were harvested at 10, 20, 30, 40, 50, 75, and 100% relaxation and LC20 phosphorylation levels were quantified by Phos-tag SDS-PAGE and western blotting with anti-pan LC20. Values represent the mean ± S.E. (n = 5). Representative western blots are shown in (C). This research was originally published in the Journal of Biological Chemistry. Sutherland C, Walsh MP. Myosin regulatory light chain diphosphorylation slows relaxation of arterial smooth muscle. 2012; 287(29): 24064–76. © the American Society for Biochemistry and Molecular Biology.

The renal microvasculature

Fig. 4 provides a schematic representation of the renal microcirculation, which consists of the afferent arteriole conveying blood from the interlobular artery to the glomerular capillaries where it is filtered prior to returning to the systemic circulation via the efferent arteriole. The afferent arteriole plays a key role in regulating glomerular inflow resistance and must be able to respond very rapidly to sudden changes in systemic blood pressure in order to protect the fragile glomeruli from pressure-induced damage (39). Angiotensin II (Ang II) is a renal-selective vasoconstrictor that contributes to renal vascular resistance under normal physiological conditions and thereby plays an important role in modulating renal hemodynamics (40). Endothelin-1 (ET-1), on the other hand, is a renal vasoconstrictor that does not contribute to renal vascular resistance under normal physiological conditions, but is implicated in abnormal renal vasoconstriction and reduced glomerular filtration in pathological states such as diabetes and chronic kidney disease (41,42,43,44,45,46,47).

Fig. 4.

Schematic diagram illustrating the anatomical relationship of the afferent arteriole, glomerulus, and the efferent arteriole. The afferent arteriole controls the glomerular inflow resistance. This vessel must constrict rapidly in response to fluctuations in blood pressure to prevent pressure elevations from being transmitted to the downstream glomerular capillaries. The efferent arteriole originates at the glomerular capillaries and regulates glomerular outflow resistance. When renal perfusion pressure is compromised, a sustained increase in efferent arteriolar tone maintains adequate filtration pressure within the upstream glomerular capillaries, thereby preserving renal function. This figure was reproduced from ref. (52) with permission.

Based on our studies of the effects of LC20 diphosphorylation described above, we developed a hypothesis that pathological situations of hypercontractility may result from impaired relaxation due to diphosphorylation of LC20 at Thr18 and Ser19. We have tested this hypothesis by studying the effects of two contractile stimuli, one physiological (Ang II) and one pathophysiological (ET-1), on renal afferent arteriolar constriction and LC20 phosphorylation. We compared the patterns of phosphorylation of LC20 in the afferent arteriole to determine whether or not LC20 diphosphorylation is associated with the pathophysiological stimulus ET-1 and not the physiological stimulus Ang II. This proved to be a challenging proposition due largely to the very small size of the afferent arteriole: a single afferent arteriole has a diameter of 15–20 µm, contains < 100 smooth muscle cells and is ∼1/10th the size of a human eyelash. This necessitated developing a technique to isolate individual afferent arterioles and enhance the sensitivity of detection and quantification of LC20 phosphorylation. Perfusion of the renal artery with molten agarose followed by cooling to solidify the agarose enabled dissection and recovery of intact afferent arterioles (48). Phos-tag SDS-PAGE (49) proved to be a suitable technique for rapid and efficient separation of unphosphorylated and phosphorylated forms of LC20 (50). In this technique, tissue proteins are separated in Laemmli SDS gels in which a phosphate-binding ligand (Phos-tag reagent) is immobilized in the running gel. In the presence of Mn2+ ions, the migration of proteins containing phosphorylated serine, threonine and/or tyrosine residues is retarded due to binding to the ligand. The higher the stoichiometry of phosphorylation, the slower the migration rate through the gel. The protein of interest (LC20 in this instance) is then detected by western blotting with an antibody that recognizes all forms of the protein, phosphorylated and unphosphorylated. The effectiveness of the separation of the various LC20 species by Phos-tag SDS-PAGE can be clearly seen in Figs. 2B and 3C. Quantification of the different LC20 species by densitometric scanning enables determination of the stoichiometry of phosphorylation, and the individual phosphorylation sites can be identified by using phosphorylation site-specific antibodies in parallel western blotting experiments. We were able to increase the sensitivity of detection of LC20 > 4,000-fold over existing methods (50). This was achieved by: (i) the use of biotinylated secondary antibodies in conjunction with streptavidin-conjugated horseradish peroxidase, combined with enhanced chemiluminescence detection of LC20 species, (ii) fixing the LC20 on the PVDF membrane by treatment with glutaraldehyde, and (iii) incorporating CanGetSignal® (Toyobo, Japan) into the protocol. Utilization of a minimum number of steps in the protocol with the least possible number of sample transfers maximized LC20 yield. The limit of detection of LC20 was thereby increased from ∼200 fmol (4 ng) to ∼0.05 fmol (1 pg); we estimate that a single afferent arteriole contains ∼2.5 fmol (50 pg) of LC20.

Using this approach, we succeeded in quantifying LC20 phosphorylation levels in single isolated afferent arterioles and observed that Ang II induced exclusively monophosphorylation of LC20 whereas ET-1 induced diphosphorylation, both in a time- and concentration-dependent manner (22). ET-1-induced diphosphorylation was confirmed to occur at Thr18 and Ser19 using phosphorylation site-specific antibodies. ET-1-induced LC20 diphosphorylation was confirmed by the proximity ligation assay (51). Furthermore, afferent arteriolar vasodilation (relaxation) occurred more slowly following washout of ET-1 than Ang II (22). These findings are, therefore, consistent with the hypothesis that pathophysiological signals such as ET-1 that are associated with prolonged vasoconstrictor responses involve LC20 diphosphorylation, whereas physiological signals such as Ang II induce LC20 phosphorylation exclusively at Ser19. The additional phosphorylation at Thr18 induced by ET-1 is, therefore, proposed to account, at least in part, for the sustained contractile response of the afferent arteriole to ET-1 compared to Ang II.

Conclusions

LC20 is phosphorylated at Thr18 and Ser19 in a Ca2+-independent manner by ILK and/or ZIPK, which are associated with the contractile machinery in vascular smooth muscle. This occurs in concert with inhibition of MLCP by ROK-catalysed phosphorylation of MYPT1, the regulatory and targeting subunit of the phosphatase. Diphosphorylation of LC20 occurs in response to ET-1 (pathological stimulus) but not Ang II (physiological stimulus) in renal afferent arterioles, and diphosphorylation of LC20 is associated with decreased rates of LC20 dephosphorylation and relaxation. ILK and ZIPK are, therefore, potential therapeutic targets for the treatment of diseases associated with hypercontractility, such as hypertension, cerebral vasospasm following subarachnoid hemorrhage, coronary arterial vasospasm, intimal hyperplasia, acute renal insufficiency and chronic kidney disease.

Conflict of interest

The authors declare that they have no conflict of interest.