Frequency-encoding Thr17 phospholamban phosphorylation is independent of Ser16 phosphorylation in cardiac myocytes

Both Ser(16) and Thr(17) of phospholamban (PLB) are phosphorylated, respectively, by cAMP-dependent protein kinase (PKA) and Ca(2+)/calmodulin-dependent protein kinase II (CaMKII). PLB phosphorylation relieves cardiac sarcoplasmic reticulum Ca(2+) pump from inhibition by PLB. Previous studies have suggested that phosphorylation of Ser(16) by PKA is a prerequisite for Thr(17) phosphorylation by CaMKII and is essential to the relaxant effect of beta-adrenergic stimulation. To determine the role of Thr(17) PLB phosphorylation, we investigated the dual-site phosphorylation of PLB in isolated adult rat cardiac myocytes in response to beta(1)-adrenergic stimulation or electrical field stimulation (0. 1-3 Hz) or both. A beta(1)-adrenergic agonist, norepinephrine (10(-9)-10(-6) m), in the presence of an alpha(1)-adrenergic antagonist, prazosin (10(-6) m), selectively increases the PKA-dependent phosphorylation of PLB at Ser(16) in quiescent myocytes. In contrast, electrical pacing induces an opposite phosphorylation pattern, selectively enhancing the CaMKII-mediated Thr(17) PLB phosphorylation in a frequency-dependent manner. When combined, electric stimulation (2 Hz) and beta(1)-adrenergic stimulation lead to dual phosphorylation of PLB and exert a synergistic effect on phosphorylation of Thr(17) but not Ser(16). Frequency-dependent Thr(17) phosphorylation is closely correlated with a decrease in 50% relaxation time (t(50)) of cell contraction, which is independent of, but additive to, the relaxant effect of Ser(16) phosphorylation, resulting in hastened contractile relaxation at high stimulation frequencies. Thus, we conclude that in intact cardiac myocytes, phosphorylation of PLB at Thr(17) occurs in the absence of prior Ser(16) phosphorylation, and that frequencydependent Thr(17) PLB phosphorylation may provide an intrinsic mechanism for cardiac myocytes to adapt to a sudden change of heart rate.     

The alpha-helical propensity of the cytoplasmic domain of phospholamban: a molecular dynamics simulation of the effect of phosphorylation and mutation

We have used molecular dynamics simulations to investigate the effect of phosphorylation and mutation on the cytoplasmic domain of phospholamban (PLB), a 52-residue protein that regulates the calcium pump in cardiac muscle. Simulations were carried out in explicit water systems at 300 K for three peptides spanning the first 25 residues of PLB: wild-type (PLB(1-25)), PLB(1-25) phosphorylated at Ser16 and PLB(1-25) with the R9C mutation, which is known to cause human heart disease. The unphosphorylated peptide maintains a helical conformation from 3 to 15 throughout a 26-ns simulation, in agreement with spectroscopic data. Comparison with simulations of a fourth peptide truncated at Pro21 showed the importance of the region from 17 to 21 in preventing local unfolding of the helix. The results suggest that residues 11-16 are more likely to unfold when specific capping motifs are not present. It is proposed that protein kinase A exploits the intrinsic flexibility of the 11-21 region when binding PLB. In agreement with available CD and NMR data, the simulations show a decrease in the helical content upon phosphorylation. The phosphorylated peptide is characterized by helix spanning residues 3-11, followed by a turn that optimizes the salt-bridge interaction between the side chains of the phosphorylated Ser-16 and Arg-13. Replacing Arg-9 with Cys results in unfolding of the helix from C9 and an overall decrease of the helical conformation. The simulations show that initiation of unfolding is due to increased solvent accessibility of the backbone atoms near the smaller Cys. It is proposed that the loss of inhibitory potency upon Ser-16 phosphorylation or R9C mutation of PLB is due to a similar mechanism, in which the partial unfolding of the cytoplasmic helix of PLB results in a conformation that interacts with the cytoplasmic domain of the calcium pump to relieve its inhibition.     

Proteolytic cleavage of phospholamban purified from canine cardiac sarcoplasmic reticulum vesicles. Generation of a low resolution model of phospholamban structure

Purified phospholamban isolated from canine cardiac sarcoplasmic reticulum vesicles was subjected to proteolysis and peptide mapping to localize the different sites of phosphorylation on the protein and to gain further information on its subunit structure. Five different proteases (trypsin, papain, chymotrypsin, elastase, and Pronase) degraded the oligomeric 27-kDa phosphoprotein into a major 21-22-kDa protease-resistant fragment. No 32P was retained by this protease-resistant fragment, regardless of whether phospholamban had been phosphorylated by cAMP-dependent protein kinase, Ca2+/calmodulin-dependent protein kinase, or protein kinase C. Phosphoamino acid analysis and thin-layer electrophoresis of liberated phosphopeptides revealed that 1 threonine and 2 serine residues were phosphorylated in phospholamban and that 1 of these serine residues and the threonine residue were in close proximity. Only serine was phosphorylated by cAMP-dependent protein kinase, whereas Ca2+-calmodulin-dependent protein kinase phosphorylated exclusively threonine. The results demonstrate that phospholamban has a large protease-resistant domain and a smaller protease-sensitive domain, the latter of which contains all of the sites of phosphorylation. The 21-22-kDa protease-resistant domain, although devoid of incorporated 32P, was completely dissociated into identical lower molecular weight subunits by boiling in sodium dodecyl sulfate, suggesting that this region of the molecule promotes the relatively strong interactions that hold the subunits together. The data presented lend further support for a model of phospholamban structure in which several identical low molecular weight subunits are noncovalently bound to one another, each containing one site of phosphorylation for cAMP-dependent protein kinase and another site of phosphorylation for Ca2+/calmodulin-dependent protein kinase.     

Atrial natriuretic peptide-dependent phosphorylation of smooth muscle cell particulate fraction proteins is mediated by cGMP-dependent protein kinase

The atrial natriuretic peptide (ANP) stimulates cGMP production and protein phosphorylation in a particulate fraction of cultured rat aortic smooth muscle cells. Three proteins of 225, 132, and 11 kDa were specifically phosphorylated in response to ANP treatment, addition of cGMP (5 nM), or addition of purified cGMP-dependent protein kinase. The cAMP-dependent protein kinase inhibitor had no effect on the cGMP-stimulated phosphorylation of the three proteins but inhibited cAMP-dependent phosphorylation of a 17-kDa protein. These results demonstrate that the particulate cGMP-dependent protein kinase mediates the phosphorylation of the 225-, 132-, and 11-kDa proteins. The 11-kDa protein is phospholamban based on the characteristic shift in apparent Mr from 11,000 to 27,000 on heating at 37 degrees C rather than boiling prior to electrophoresis. ANP (1 microM) increased the cGMP concentration approximately 4-fold in the particulate fractions, from 4.3 to 17.7 nM, as well as the phosphorylation of the 225-, 132-, and 11-kDa proteins. In contrast, the biologically inactive form of ANP, carboxymethylated ANP (1 microM), did not stimulate phosphorylation of any proteins nor did the unrelated peptide hormone, angiotensin II (1 microM). These results demonstrate the presence of the cGMP-mediated ANP signal transduction pathway in a particulate fraction of smooth muscle cells and the specific phosphorylation of three proteins including phospholamban, which may be involved in ANP-dependent relaxation of smooth muscle.     

Role of conformational sampling of Ser16 and Thr17-phosphorylated phospholamban in interactions with SERCA

Phosphorylation of phospholamban (PLB) at Ser16 and/ or Thr17 is believed to release its inhibitory effect on sarcoplasmic reticulum calcium ATPase. Ser16 phosphorylation of PLB has been suggested to cause a conformational change that alters the interaction between the enzyme and protein. Using computer simulations, the conformational sampling of Ser16 phosphorylated PLB in implicit membrane environment is compared here with the unphosphorylated PLB system to investigate these conformational changes. The results suggest that conformational changes in the cytoplasmic domain of PLB upon phosphorylation at Ser16 increase the likelihood of unfavorable interactions with SERCA in the E2 state prompting a conformational switch of SERCA from E2 to E1. Phosphorylation of PLB at Thr17 on the other hand does not appear to affect interactions with SERCA significantly suggesting that the mechanism of releasing the inhibitory effect is different between Thr17 phosphorylated and Ser16 phosphorylated PLB.     

Phospholamban phosphorylation in intact ventricles. Phosphorylation of serine 16 and threonine 17 in response to beta-adrenergic stimulation

Phospholamban is the major membrane protein of the heart phosphorylated in response to beta-adrenergic stimulation. In cell-free systems, cAMP-dependent protein kinase catalyzes exclusive phosphorylation of serine 16 of phospholamban, whereas Ca2+/calmodulin-dependent protein kinase gives exclusive phosphorylation of threonine 17 (Simmerman, H. K. B., Collins, J. H., Theibert, J. L., Wegener, A. D., and Jones, L. R. (1986) J. Biol. Chem. 261, 13333-13341). In this work we have localized the sites of phospholamban phosphorylation in intact ventricles treated with the beta-adrenergic agonist isoproterenol. Isolation of phosphorylated phospholamban from 32P-perfused guinea pig ventricles, followed by partial acid hydrolysis and phosphoamino acid analysis, revealed phosphorylation of both serine and threonine residues. At steady state after isoproterenol exposure, phospholamban contained approximately equimolar amounts of these two phosphoamino acids. Two major tryptic phosphopeptides containing greater than 90% of the incorporated radioactivity were obtained from phospholamban labeled in intact ventricles. The amino acid sequences of these two tryptic peptides corresponded exactly to residues 14-25 and 15-25 of canine cardiac phospholamban, thus localizing the sites of in situ phosphorylation to serine 16 and threonine 17. Phosphorylation of phospholamban at two sites in heart perfused with isoproterenol was supported by detection of 11 distinct mobility forms of the pentameric protein by use of the Western blotting method, consistent with each phospholamban monomer containing two phosphorylation sites, and with each pentamer containing from 0 to 10 incorporated phosphates. Our results localize the sites of in situ phospholamban phosphorylation to serine 16 and threonine 17 and, furthermore, are consistent with the phosphorylations of these 2 residues being catalyzed by cAMP- and Ca2+/calmodulin-dependent protein kinases, respectively.     

Increased inhibition of SERCA2 by phospholamban in the type I diabetic heart

The sarcoplasmic reticulum (SR) plays a critical role in mediating cardiac contractility and its function is abnormal in the diabetic heart. However, the mechanisms underlying SR dysfunction in the diabetic heart are not clear. Because protein phosphorylation regulates SR function, this study examined the phosphorylation state of phospholamban, a key SR protein that regulates SR calcium (Ca2+) uptake in the heart. Diabetes was induced in male Sprague-Dawley rats by an injection of streptozotocin (STZ; 65 mg kg(-1) i.v.), and the animals were humanely killed after 6 weeks and cardiac SR function was examined. Depressed cardiac performance was associated with reduced SR Ca2+-uptake activity in diabetic animals. The reduction in SR Ca2+-uptake was consistent with a significant decrease in the level of SR Ca2+-pump ATPase (SERCA2a) protein. The level of phospholamban (PLB) protein was also decreased, however, the ratio of PLB to SERCA2a was increased in the diabetic heart. Depressed SR Ca2+-uptake was also due to a reduction in the phosphorylation of PLB by the Ca2+-calmodulin-dependent protein kinase (CaMK) and cAMP-dependent protein kinase (PKA). Although the activities of the SR-associated Ca2+-calmodulin-dependent protein kinase (CaMK), cAMP-dependent protein kinase (PKA) were increased in the diabetic heart, depressed phosphorylation of PLB could partly be attributed to an increase in the SR-associated protein phosphatase activities. These results suggest that there is increased inhibition of SERCA2a by PLB and this appears to be a major defect underlying SR dysfunction in the diabetic heart.     

Analysis of acetylcholine receptor phosphorylation sites using antibodies to synthetic peptides and monoclonal antibodies.

Three peptides corresponding to residues 354-367, 364-374, 373-387 of the acetylcholine receptor (AChR) delta subunit were synthesized. These peptides represent the proposed phosphorylation sites of the cAMP-dependent protein kinase, the tyrosine-specific protein kinase and the calcium/phospholipid-dependent protein kinase respectively. Using these peptides as substrates for phosphorylation by the catalytic subunit of cAMP-dependent protein kinase it was shown that only peptides 354-367 was phosphorylated whereas the other two were not. These results verify the location of the cAMP-dependent protein kinase phosphorylation site within the AChR delta subunit. Antibodies elicited against these peptides reacted with the delta subunit. The antipeptide antibodies and two monoclonal antibodies (7F2, 5.46) specific for the delta subunit were tested for their binding to non-phosphorylated receptor and to receptor phosphorylated by the catalytic subunit of cAMP-dependent protein kinase. Antibodies to peptide 354-367 were found to react preferentially with non-phosphorylated receptor whereas the two other anti-peptide antibodies bound equally to phosphorylated and non-phosphorylated receptors. Monoclonal antibody 7F2 reacted preferentially with the phosphorylated form of the receptor whereas monoclonal antibody 5.46 did not distinguish between the two forms.     

Phosphorylation of phospholamban by calcium-activated, phospholipid-dependent protein kinase. Stimulation of cardiac sarcoplasmic reticulum calcium uptake

Ca2+-activated, phospholipid-dependent protein kinase (protein kinase C) is able to catalyze the phosphorylation of phospholamban in a canine cardiac sarcoplasmic reticulum preparation. This phosphorylation is associated with a 2-fold stimulation of Ca2+ uptake by cardiac sarcoplasmic reticulum similar to that seen following phosphorylation of phospholamban by an endogenous calmodulin-dependent protein kinase or by the catalytic subunit of cAMP-dependent protein kinase. Two-dimensional peptide maps of the tryptic fragments of phospholamban indicate that the three protein kinases differ in their selectivity for sites of phosphorylation. However, one common peptide appears to be phosphorylated by all three protein kinases. These findings suggest that protein kinase C may play a role similar to those played by cAMP- and calmodulin-dependent protein kinases in the regulation of Ca2+ uptake by cardiac sarcoplasmic reticulum, and raise the possibility that the effects of all three protein kinases are mediated through phosphorylation of a common peptide in phospholamban.     

Concerted phosphorylation of the 26-kilodalton phospholamban oligomer and of the low molecular weight phospholamban subunits

Phospholamban (PLB) from cardiac sarcoplasmic reticulum (SR) was phosphorylated under various conditions by the adenosine cyclic 3',5'-phosphate (cAMP)-dependent and/or the calmodulin-dependent protein kinase. The small shifts in apparent molecular weight resulting from the incorporation of Pi groups in the PLB complexes were analyzed by high-resolution sodium dodecyl sulfate-polyacrylamide gel electrophoresis. In parallel experiments, PLB was dissociated into its subunits and analyzed by using a newly developed isoelectric focusing system. The pI values of the PLB subunits phosphorylated by the cAMP- or calmodulin-dependent kinase were 6.2 and 6.4, respectively. Double phosphorylation of the same subunit resulted in an acidic shift of the pI to 5.2. The combined analysis of the behavior of the PLB complex and of its subunits has greatly simplified the interpretation of the complex phosphorylation pattern and has led to the following conclusions: The PLB complex is composed of five probably identical subunits, each of them containing a distinct phosphorylation site for the calmodulin- and the cAMP-dependent kinase. The population of PLB interacting with the endogenous calmodulin-dependent kinase cannot be phosphorylated by the cAMP-dependent kinase unless previously phosphorylated in the presence of calmodulin. It was also observed that after maximal phosphorylation of PLB in the presence of very large amounts of the cAMP-dependent protein kinase, the Ca2+ pumping rate of the cardiac SR ATPase is stimulated up to 5-fold, i.e., a level of a stimulation which exceeds considerably the values so far reported in the literature.     

Subunit structure and multiple phosphorylation sites of phospholamban

The phosphorylation-induced mobility shift of the high molecular weight form of phospholamban (24,500 daltons) in the cardiac sarcoplasmic reticulum produced on 3',5'-cyclic AMP (cAMP)-dependent phosphorylation with 5 mM ATP was resolved into five clear steps on sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and on Ca2+-calmodulin-dependent phosphorylation into ten steps. The mobility shift of the low molecular weight form of phospholamban (less than 14,400 daltons) in these reactions occurred in one step and two steps, respectively. With the two protein kinase activities, the electrophoretic pattern of the mobility shifts of the high and low molecular weight forms of phospholamban was similar to that obtained with Ca2+-calmodulin-dependent protein kinase alone. The results of pulse-chase experiments involving the centrifuge column method suggested that the site(s) of phosphorylation by cAMP- and Ca2+-calmodulin-dependent protein kinase activities are on the same phospholamban molecule. Two-dimensional tryptic peptide maps of phosphorylated phospholamban indicated that cAMP-dependent protein kinase phosphorylates at a single site, A, and Ca2+-calmodulin-dependent protein kinase phosphorylates at sites C1 and C2 in the low molecular weight form, where A is different from C1 but may be the same as C2. The high molecular weight form of phospholamban is suggested to be a pentamer of identical monomers (low molecular weight form) having one phosphorylation site for cAMP-dependent protein kinase and two for Ca2+-calmodulin-dependent protein kinase.     

Phosphate groups as substrate determinants for casein kinase I action

Phosphorylation of rabbit muscle glycogen synthase by cyclic AMP-dependent protein kinase has been shown to enhance subsequent phosphorylation by casein kinase I (Flotow, H., and Roach, P. J. (1989) J. Biol. Chem. 264, 9126-9128). In the present study, synthetic peptides based on the sequences of the four phosphorylated regions in muscle glycogen synthase were used to probe the role of substrate phosphorylation in casein kinase I action. With all four peptides, prior phosphorylation significantly stimulated phosphorylation by casein kinase I. A series of peptides was synthesized based on the NH2-terminal glycogen synthase sequence PLSRTLS7VSS10LPGL, in which phosphorylation at Ser7 is required for modification of Ser10 by casein kinase I. The spacing between the P-Ser and the acceptor Ser was varied to have 1, 2, or 3 intervening residues. The peptide with a 2-residue spacing (-S(P)-X-X-S-) was by far the best casein kinase I substrate. When the P-Ser residue at Ser7 was replaced with P-Thr, the resulting peptide was still a casein kinase I substrate. However, substitution of Asp or Glu residues at Ser7 led to peptides that were not phosphorylated by casein kinase I. Phosphorylation of one of the other peptides showed that Thr could also be the phosphate acceptor. From these results, we propose that there are substrates for casein kinase I for which prior phosphorylation is a critical determinant of protein kinase action. In these instances, an important recognition motif for casein kinase I appears to be -S(P)/T(P)-Xn-S/T- with n = 2 much more effective than n = 1 or n = 3. Thus, casein kinase I may be involved in hierarchal substrate phosphorylation schemes in which its activity is controlled by the phosphorylation state of its substrates.     

Sites on the cytoplasmic region of phospholamban involved in interaction with the calcium-activated ATPase of the sarcoplasmic reticulum.

Proton NMR studies have shown that when a peptide corresponding to the N-terminal region of phospholamban, PLB(1-20), interacts with the Ca2+ATPase of the sarcoplasmic reticulum, SERCA1a, docking involves the whole length of the peptide. Phosphorylation of Ser16 reduced the affinity of the peptide for the pump by predominantly affecting the interaction with the C-terminal residues of PLB(1-20). In the phosphorylated peptide weakened interaction occurs with residues at the N-terminus of PLB(1-20). PLB(1-20) is shown to interact with a peptide corresponding to residues 378-405 located in the cytoplasmic region of SERCA2a and related isoforms. This interaction involves the C-terminal regions of both peptides and corresponds to that affected by phosphorylation. The data provide direct structural evidence for complex formation involving residues 1-20 of PLB. They also suggest that phospholamban residues 1-20 straddle separate segments of the cytoplasmic domain of SERCA with the N-terminus of PLB associated with a region other than that corresponding to SERCA2a(378-405).     

Phosphorylation and dephosphorylation of purified phospholamban and associated phosphatidylinositides

Phospholamban, the putative regulator for the calcium pump, was purified to apparent homogeneity and in high yields from canine cardiac sarcoplasmic reticulum membranes. Purified phospholamban migrated with an apparent Mr of 27,000 in alkaline sodium dodecyl sulfate-polyacrylamide gels, and upon boiling in 7.5% sodium dodecyl sulfate, it dissociated into a lower molecular weight component of 5500-6000. Purified phospholamban contained 0.62 +/- 0.09 mumol of lipid Pi/mg of protein, and the major phospholipids were phosphatidylserine (34%), phosphatidylcholine (22%), sphingomyelin (17%), phosphatidylinositol (13%), and phosphatidylethanolamine (9%). Phospholamban was phosphorylated by cAMP-dependent protein kinase to a level of 207 nmol of Pi/mg, and this would indicate an incorporation of 1 mol of phosphate/mol of protein, assuming a molecular weight of 5500 for phospholamban. Phosphorylation of phospholamban could be reversed by a "phospholamban phosphatase" isolated from canine cardiac cytosol. Phospholipids associated with the purified phospholamban were also phosphorylated in the presence of the catalytic subunit of cAMP-dependent protein kinase, and the maximal phosphate incorporation was 4 nmol/mg of protein. The main phospholipids phosphorylated were phosphatidylinositol 4-monophosphate and phosphatidylinositol 4,5-bisphosphate. Phosphorylation of phospholipids was inhibited by the heat-stable inhibitor protein of the cAMP-dependent protein kinase, and it could be also reversed by the phospholamban phosphatase.(ABSTRACT TRUNCATED AT 250 WORDS)     

Structural changes in the cytoplasmic domain of phospholamban by phosphorylation at Ser16: a molecular dynamics study

Phospholamban is a 52-residue integral membrane protein that regulates the activity of the sarcoplasmic reticulum calcium pump in cardiac muscle. Its inhibitory action is relieved when phospholamban is phosphorylated at Ser16 by cAMP-dependent protein kinase. To computationally explore all possible conformations of the phosphorylated form, and thereby to understand the structural effects of phosphorylation, replica-exchange molecular dynamics (REMD) was applied to the cytoplasmic domain that includes Ser16. The simulations showed that (i) without phosphorylation, the region from Lys3 to Ser16 takes all alpha-helical conformations; (ii) when phosphorylated, the alpha-helix is partially unwound in the C-terminal part (from Ser10 to Ala15) resulting in less extended conformations; (iii) the phosphate at Ser16 forms salt bridges with Arg9, Arg13, and/or Arg14; and (iv) the salt bridges with Arg13 and Arg14 distort the alpha-helix and induce unwinding of the C-terminal part. We then applied conventional all-atom molecular dynamics simulations to the full-length phospholamban in the phospholipid bilayer. The results were consistent with those obtained with REMD simulations, suggesting that the transmembrane part of phospholamban and the lipid bilayer itself have only minor effects on the conformational changes in the cytoplasmic domain. The distortions caused by the salt bridges involving the phosphate at Ser16 readily explain the relief of the inhibitory effect of phospholamban by phosphorylation, as they will substantially reduce the population of all helical conformations, which are presumably required for the binding to the calcium pump. This will also be the mechanism for releasing the phosphorylated phospholamban from kinase.     

A single site (Ser16) phosphorylation in phospholamban is sufficient in mediating its maximal cardiac responses to beta -agonists

Phospholamban (PLB) can be phosphorylated at Ser(16) by cyclic AMP-dependent protein kinase and at Thr(17) by Ca(2+)-calmodulin-dependent protein kinase during beta-agonist stimulation. A previous study indicated that mutation of S16A in PLB resulted in lack of Thr(17) phosphorylation and attenuation of the beta-agonist stimulatory effects in perfused mouse hearts. To further delineate the functional interplay between dual-site PLB phosphorylation, we generated transgenic mice expressing the T17A mutant PLB in the cardiac compartment of the null background. Lines expressing similar levels of T17A mutant, S16A mutant, or wild-type PLB in the null background were characterized in parallel. Cardiac myocyte basal mechanics and Ca(2+) kinetics were similar among the three groups. Isoproterenol stimulation was associated with phosphorylation of both Ser(16) and Thr(17) in wild-type PLB and Ser(16) phosphorylation in T17A mutant PLB, whereas there was no detectable phosphorylation of S16A mutant PLB. Phosphorylation of Ser(16) alone in T17A mutant PLB resulted in responses of the mechanical and Ca(2+) kinetic parameters to isoproterenol similar to those in wild-type myocytes, which exhibited dual-site PLB phosphorylation. However, those parameters were significantly attenuated in the S16A mutant myocytes. Thus, Ser(16) in PLB can be phosphorylated independently of Thr(17) in vivo, and phosphorylation of Ser(16) is sufficient for mediating the maximal cardiac responses to beta-adrenergic stimulation.     

Secondary structure of detergent-solubilized phospholamban, a phosphorylatable, oligomeric protein of cardiac sarcoplasmic reticulum

The structure of phospholamban, a 30-kDa oligomeric protein integral to cardiac sarcoplasmic reticulum, was probed using ultraviolet absorbance and circular dichroism spectroscopy. Purified phospholamban was examined in three detergents: octyl glucoside, n-dodecyloctaethylene glycol monoether (C12E8) and sodium dodecyl sulfate (SDS). Ultraviolet absorption spectra of phospholamban reflected its aromatic amino acid content: absorption peaks at 275-277 nm and 253, 259, 265 and 268 nm were attributed to phospholamban's one tyrosine and two phenylalanines, respectively. Phospholamban phosphorylated at serine 16 by the catalytic subunit of cAMP-dependent protein kinase exhibited no absorbance changes when examined in C12E8 or SDS. Circular dichroism spectroscopy at 250-190 nm demonstrated that phospholamban possesses a very high content of alpha-helix in all three detergents and is unusually resistant to denaturation. Dissociation of phospholamban subunits by boiling in SDS increased the helical content, suggesting that the highly ordered structure is not dependent upon oligomeric interactions. The purified COOH-terminal tryptic fragment of phospholamban, containing residues 26-52 and comprising the hydrophobic, putative membrane-spanning domain, also exhibited a circular dichroism spectrum characteristic of alpha-helix. Circular dichroism spectra of phosphorylated and dephosphorylated phospholamban were very similar, indicating that phosphorylation does not alter phospholamban secondary structure significantly. The results are consistent with a two-domain model of phospholamban in which each domain contains a helix and phosphorylation may act to rotate one domain relative to the other.     

Sequence analysis of phospholamban. Identification of phosphorylation sites and two major structural domains

Phospholamban is a regulatory protein in cardiac sarcoplasmic reticulum that is phosphorylated by cAMP- and Ca2+/calmodulin-dependent protein kinase activities. In this report, we present the partial amino acid sequence of canine cardiac phospholamban and the identification of the sites phosphorylated by these two protein kinases. Gas-phase protein sequencing was used to identify 20 NH2-terminal residues. Overlap peptides produced by trypsin or papain digestion extended the sequence 16 residues to give the following primary structure: Ser-Ala-Ile-Arg-Arg-Ala-Ser-Thr-Ile-Glu-Met-Pro-Gln-Gln-Ala- Arg-Gln-Asn-Leu-Gln-Asn-Leu-Phe-Ile-Asn-Phe-(Cys)-Leu-Ile-Leu-Ile-(Cys)- Leu-Leu-Leu-Ile-. Phospholamban phosphorylated by either cAMP-dependent or Ca2+/calmodulin-dependent protein kinase was cleaved with trypsin, and the major phosphorylated peptide (comprising greater than 70% of the incorporated 32P label) was purified by reverse-phase high performance liquid chromatography. The identical sequence was revealed for the radioactive peptide obtained from phospholamban phosphorylated by either kinase: Arg-Ala-Ser-Thr-Ile-Glu-Met-Pro-Gln-Gln-. The adjacent residues Ser7 and Thr8 of phospholamban were identified as the unique sites phosphorylated by cAMP- and Ca2+/calmodulin-dependent protein kinases, respectively. These results establish that phospholamban is an oligomer of small, identical polypeptide chains. A hydrophilic, cytoplasmically oriented NH2-terminal domain on each monomer contains the unique, adjacent residues phosphorylated by cAMP- and Ca2+/calmodulin-dependent protein kinase activities. Analysis by hydropathic profiling and secondary structure prediction suggests that phospholamban monomers also contain a hydrophobic domain, which could form amphipathic helices sufficiently long to traverse the sarcoplasmic reticulum membrane. A model of phospholamban as a pentamer is presented in which the amphipathic alpha-helix of each monomer is a subunit of the pentameric membrane-anchored domain, which is comprised of an exterior hydrophobic surface and an interior hydrophilic region containing polar side chains.     

Tyrosine-specific dephosphorylation-phosphorylation with alkaline phosphatases and epidermal growth factor receptor kinase as evidenced by 31P NMR spectroscopy

The dephosphorylation of phospho-amino acids with alkaline phosphatase (AlPase) from calf intestine or Escherichia coli and the phosphorylation of bovine serum albumin (BSA) with epidermal growth factor (EGF) receptor kinase from human A431 epidermoid carcinoma cells were investigated by 31P NMR spectroscopy. The initial rates of the dephosphorylation of phospho-tyrosine (P-Tyr) and phosphoserine (P-Ser) with AlPase were essentially the same in the one-substrate system. In the two-substrate system (P-Tyr plus P-Ser), however, the ratio of the initial rate for P-Tyr vs. P-Ser was 2.4 to 4.5 depending on the buffer and pH conditions employed. This substantiates for the first time the specificity of AlPases to P-Tyr over P-Ser at the free amino acid level. In the stationary phase of the overall process, the dephosphorylation of P-Ser became slow compared to that of P-Tyr in the one-substrate system. The decrease in the rate for P-Ser was further pronounced in the two-substrate system. For this remarkable effect, the rephosphorylation of serine was responsible, as demonstrated in the reaction mixture containing serine, Pi, and AlPase. BSA phosphorylated by EGF receptor kinase exhibited sharp 31P resonances around 0 ppm at neutral pH, far distant from the peak positions (4.9 ppm) of histone H1 phosphorylated by cAMP-dependent protein kinase. These NMR data are directed evidence that BSA was phosphorylated exclusively at the tyrosyl residues, whereas the phosphorylation of histone H1 was at the seryl residues.