International Journal of Molecular Sciences

Article Atomistic Structure and Dynamics of the Ca2+-ATPase Bound to Phosphorylated

Rodrigo Aguayo-Ortiz 1,2 and L. Michel Espinoza-Fonseca 1,* 1 Center for Arrhythmia Research, Department of Internal Medicine, Division of Cardiovascular Medicine, University of Michigan, Ann Arbor, MI 48109, USA; [email protected] 2 Departamento de Fisicoquímica, Universidad Nacional Autónoma de México, Mexico City 04510, Mexico * Correspondence: [email protected]; Tel.: +1-734-998-7500

 Received: 14 September 2020; Accepted: 29 September 2020; Published: 1 October 2020 

Abstract: Ca2+-ATPase (SERCA) and phospholamban (PLB) are essential components of the cardiac Ca2+ transport machinery. PLB phosphorylation at residue Ser16 (pSer16) enhances SERCA activity in the heart via an unknown structural mechanism. Here, we report a fully atomistic model of SERCA bound to phosphorylated PLB and study its structural dynamics on the microsecond time scale using all-atom molecular dynamics simulations in an explicit lipid bilayer and water environment. The unstructured N-terminal phosphorylation domain of PLB samples different orientations and covers a broad area of the cytosolic domain of SERCA but forms a stable complex mediated by pSer16 interactions with a binding site formed by SERCA residues Arg324/Lys328. PLB phosphorylation does not affect the interaction between the transmembrane regions of the two ; however, pSer16 stabilizes a disordered structure of the N-terminal phosphorylation domain that releases key inhibitory contacts between SERCA and PLB. We found that PLB phosphorylation is sufficient to guide the structural transitions of the cytosolic headpiece that are required to produce a competent structure of SERCA. We conclude that PLB phosphorylation serves as an allosteric molecular switch that releases inhibitory contacts and strings together the catalytic elements required for SERCA activation. This atomistic model represents a vivid atomic-resolution visualization of SERCA bound to phosphorylated PLB and provides previously inaccessible insights into the structural mechanism by which PLB phosphorylation releases SERCA inhibition in the heart.

Keywords: calcium ATPase; calcium; sarcoplasmic reticulum; membrane transport; molecular dynamics; ; inhibition mechanism; phospholamban; phosphorylation; structural disorder

1. Introduction The sarcoplasmic reticulum (SR) Ca2+-ATPase (SERCA) is a critical molecular component of cardiac muscle cells, regulating the rate of cardiac muscle relaxation and restoring the SR Ca2+ load necessary for in subsequent beats [1]. SERCA utilizes the energy originated from hydrolysis of one molecule of ATP to transport two Ca2+ ions into the lumen of the SR [2]. SERCA activity is regulated in the heart by phospholamban (PLB), a 52-residue membrane that resides in the SR [3]. PLB inhibits SERCA activity by decreasing Ca2+ affinity, and inhibition is relieved by phosphorylation, thereby regulating SERCA activity and muscle contractility [4]. PLB phosphorylation at residue Ser16 (pSer16) induces an order-to-disorder structural transition of the cytosolic domain of PLB, destabilizing a inhibitory ordered T state in favor of the disordered, non-inhibitory R state [5,6]. Two mechanisms have been proposed for the relief of SERCA inhibition by PLB phosphorylation: dissociation model and subunit model. The dissociation model proposes that upon the formation of the R state, PLB physically separates from SERCA to relieve inhibition [7,8] whereas the subunit model hypothesizes that PLB acts as a functional subunit of SERCA, and that the

Int. J. Mol. Sci. 2020, 21, 7261; doi:10.3390/ijms21197261 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2020, 21, 7261 2 of 15

T-to-R state structural shift of PLB induced by phosphorylation relieves inhibition while preserving the SERCA–PLB heterodimer [9,10]. The two models have been supported independently by several groups, but there is solid experimental evidence that supports the subunit model because they have shown a direct interaction of PLB and SERCA upon PLB phosphorylation [11–14]. Furthermore, it is known that the cytosolic domain of PLB becomes structurally disordered upon phosphorylation, yet there has been speculation that SERCA has a well-defined pocket that binds this unstructured domain, and that this interaction is directly responsible for relief of SERCA inhibition [15]. More recent studies have shown that phosphorylation of PLB does not induce substantial changes in the interaction of the inhibitory transmembrane (TM) domain of PLB [16], thus suggesting that SERCA regulation by PLB phosphorylation is likely due to an allosteric effect exerted by the disordered cytosolic domain of PLB. Despite these advances in the field, there are major gaps in our understanding of SERCA regulation by phosphorylation of PLB: (i) What is the atomistic structure of the regulatory complex between phosphorylated PLB and SERCA? (ii) Does SERCA have a well-defined binding site for the disordered cytosolic domain of phosphorylated PLB? (iii) What is the mechanism for control of SERCA structural dynamics by PLB phosphorylation? Answering these questions in high spatiotemporal resolution is much needed in the field to understand SERCA regulation in terms of intermolecular interactions and activities. The structure of SERCA bound to the TM domain of PLB was solved recently [17], but the cytosolic domain in the phosphorylated R state is intrinsically disordered, so the high-resolution structure of the of phosphorylated PLB bound to SERCA has remained inaccessible to experiments. Based on structures of the inhibitory SERCA–PLB complex and the solution structure of PLB, here we report a fully atomistic model of SERCA bound to PLB phosphorylated at residue Ser16 (SERCA–pPLB) and study the structural dynamics using microsecond-long all-atom molecular dynamics (MD) simulations in an explicit lipid bilayer and water environment. Here, we studied the complex under Ca2+-free conditions to study the direct effects of PLB phosphorylation alone on the functional dynamics of SERCA. The result is a realistic atomic-resolution model of the molecular interactions and mechanism by which PLB phosphorylation releases SERCA inhibition in the heart.

2. Results

2.1. Structure of the SERCA–pPLB Complex We performed eight independent MD simulations for a total simulation time of 22.7 µs to determine the structure and stability of the SERCA–pPLB complex in the microsecond timescale. The structural ensemble of phosphorylated PLB in the MD trajectories shown in Figure1a provides greater detail about the conformations populated in the bound complex. The structural ensemble of phosphorylated PLB in the complex is characterized by a largely unstructured N-terminal phosphorylation domain (residues 1–26) and a structurally ordered TM helix (residues 27–52). Analysis of the time-dependent RMSD of the phosphorylated PLB showed the N-terminal phosphorylation domain is flexible in the complex, with changes in the RMSD of up to 25 å compared to the initial structure of the complex (Figure1b). Root-mean-square fluctuation (RMSF) of the main chain Cα atoms of the phosphorylation domain showed that this region exhibits high flexibility in the bound complex (Figure1c). Time-dependent changes in the secondary structure showed that the N-terminal phosphorylation is largely unstructured, although the formation of transient local α-helix structure is observed in most of the trajectories (Figure1d). Despite the disordered nature of this domain, we found that in all eight trajectories this domain settles a plateau, indicating that the structure of the cytosolic domain converges in the microsecond timescale. Conversely, RMSF values of the main chain atoms indicate that the TM helix of phosphorylated PLB has very low mobility in the complex (Figure1c). Complementary RMSD plots showed that the TM helix rapidly settles a plateau around 1 Å in all eight trajectories, indicating that the TM domain converged to a position that is essentially identical to that of the initial structure used in this study (Figure1b). An opposing behavior was found for the phosphorylation domain, where the TM domain has low mobility in the complex (i.e., RMSF < 2 Å, Figure1c) and also populates a stable Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 3 of 15

Int. J. Mol. Sci. 2020, 21, 7261 3 of 15 that is essentially identical to that of the initial structure used in this study (Figure 1b). An opposing behavior was found for the phosphorylation domain, where the TM domain has low mobility in the αcomplex-helix structure (i.e., RMSF in the< 2Å microsecond-long , Figure 1c) and also trajectories. populates These a stable findings α-helix indicate structure that in the phosphorylation microsecond- oflong PLB trajectories. does not These produce findings a change indicate in the that orientation phosphorylation of the TMof PLB domain does innot all produ trajectoriesce a change of the in SERCA–pPLBthe orientation complex. of the TM domain in all trajectories of the SERCA–pPLB complex.

Figure 1. 1. ((aa)) Three Three-dimensional-dimensional structure structure of phosphorylated phospholamban ((PLB)PLB) bound to to sarcoplasmicsarcoplasmic reticulum reticulum (SR) (SR) Ca Ca22++--ATPaseATPase ( (SERCA).SERCA). Gray Gray ribbons represenrepresentt the structures structures of of phosphorylatedphosphorylated PLB PLB sampled sampled on theon SERCA–pPLB the SERCA– complex.pPLB complex SERCA. isSERCA shown as is a surface shown representation as a surface andrepresentation colored according and colored to its according functional to domains: its functional A-domain, domains: red; A P-domain,-domain, red; blue; P- N-domain,domain, blue; green; N- transmembranedomain, green; transmembrane (TM) domain, yellow. (TM) domain, (b) RMSD yellow. of the (b TM) RMSD domain of the (red TM trace) domain and the(red N-terminal trace) and phosphorylationthe N-terminal phosphorylation domain (NT, black domain trace) of(NT, PLB black calculated trace) fromof PLB each calculated individual from molecular each individual dynamics (MD)molecular trajectory. dynamics RMSD (MD was) trajectory. obtained byRMSD aligning was theobtained backbone by aligning of the TM the domain backbone with of the the TM structure domain at thewith beginning the structure of each at MDthe replicate.beginning ( cof) RMSFs each MD of main replicate. chain ( atomsc) RMSFs for each of main residue chain of phosphorylatedatoms for each PLB.residue RMSFs of phosphorylated were calculated PLB. from RMSF eachs were independent calculated MD from trajectory each independent using the backboneMD trajectory of the using TM domainthe backbone as a reference. of the TM (d )domain Secondary as a structure reference. evolution (d) Secondary of phosphorylated structure evolution PLB indicating of phosph the presenceorylated ofPLBα-helical indicating structure the presence (violet) andof α unstructured-helical structure regions (violet) of the and protein unstructured (white). regions of the protein (white). We also calculated the backbone RMSD for the TM and cytosolic domains of SERCA to determine the structuralWe also dynamics calculated of the the backbonepump in the RMSD trajectories. for the In TM all trajectories,and cytosolic the structuredomains ofof the SERCA 10-helix to TMdetermine domain the of structural SERCA undergoes dynamics a of rapid the (tpump< 200 in ns) the drift trajectories. in the RMSD In all (Figure trajectories,2); this the small structure change of (RMSDthe 10-helix< 2.5 TM Å) in domain the structure of SERCA can be undergoes attributed a to rapid the relaxation (t < 200 ns) of thedrift TM in domainthe RMSD in the (Figure lipid bilayer.2); this Followingsmall change this (RMSD rapid relaxation < 2.5Å ) in of the the structure structure, can the be RMSD attributed values to remained the relaxation fundamentally of the TM unchanged domain in forthe thelipid remainder bilayer. of Following the simulation this rapid time. These relaxation findings of the indicate structure, that the the structure RMSD of values the TM remained domain

Int.Int. J. J. Mol. Mol. Sci. Sci. 20202020, ,2211, ,x 7261 FOR PEER REVIEW 44 of of 15 15 fundamentally unchanged for the remainder of the simulation time. These findings indicate that the structureobserved of in the MDTM domain trajectories observed is similar in the to that MD of trajectories the initial structureis similar built to that based of the on structuresinitial structure of the builtinhibitory based SERCA–PLB on structures complex. of the inhibitory This suggests SERCA phosphorylated–PLB complex PLB. This does suggests not induce phosphorylated sizeable changes PLB doesin the not overall induce conformation sizeable changes of the in TM the domain overall of conformation SERCA. RMSD of the plots TM of domain the cytosolic of SERCA headpiece. RMSD of plotsSERCA of the indicate cytosolic that headpiece this domain of SERCA undergoes indicate substantial that this structural domain undergoes changes in substantial the trajectories structural of the changesSERCA–pPLB in the complextrajectories (Figure of the2). SERCA Previous–pPLB studies complex have shown(Figure that 2). P therevious cytosolic studies headpiece have shown of SERCA that theis inherently cytosolic headpiece flexible under of SERCA physiological is inherently conditions flexible [18 under–22]; but,physiological the RMSD conditions plots indicate [18– that22]; PLBbut, thephosphorylation RMSD plots indicate may induce thatchanges PLB phosphorylation in the structural may dynamics induce of changes the headpiece. in the structural These effects dynamics will be ofdiscussed the headpiece. in Section The 2.4se .effects will be discussed in Section 2.4.

FigureFigure 2. 2. TimeTime-dependent-dependent RMSD RMSD evolution evolution of of SERCA SERCA domains domains in in the the SERCA SERCA–pPLB–pPLB complex. RMSD RMSD ofof the the TM domain of SERCA (red trace) was calculated using backbone backbone alignment alignment for for TM TM helices helices of of thethe pump; pump; RMSD RMSD of of the the cytosolic cytosolic domain domain (black (black trace) trace) was was calculated calculated by by aligning aligning the the backbone backbone of of the the cytosoliccytosolic headpiece headpiece with with the the structure structure at the beginning of each trajectory.

2.2.2.2. Mapping Mapping of of Interactions Interactions between between the the Disordered Disordered Cytosolic Cytosolic Domain Domain of Phosphorylated PLB and SERCA TThehe structure structure of of the the SERCA SERCA–pPLB–pPLB complex complex shows shows that that the the disordered disordered phosphorylation phosphorylation domain domain ofof PLB PLB interacts interacts with with the the flexible flexible cytosolic cytosolic headpiece headpiece of of SERCA. SERCA. However, However, the the simulations revealed revealed thatthat the the N-terminal N-terminal phosphorylation phosphorylation domain domain of PLB of binds PLB unspecifically binds unspecifically to SERCA, thusto SERCA contradicting, thus contradictingprevious studies previous suggesting studies that PLB suggesting phosphorylation that PLB is phosphorylation required to produce is requireda relatively to well-defined produce a relativelystructure of well the-defined complex structure that is non-inhibitory of the complex [15]. Althoughthat is non a visual-inhibitory inspection [15]. ofAlthough the complex a visual could inspectionbe sufficient of to the determine complex whether could be the sufficient phosphorylation to determine domain whether of PLB the binds phosphory non-specificallylation domain to SERCA, of PLBthis approach binds non may-specifically provide incomplete, to SERCA, thisif not approach inaccurate, may information. provide incomplete, Consequently, if not we inacc calculatedurate, information.occupancy maps Consequently averaged,over we calculated all trajectories occupancy to search maps for intermolecularaveraged over interactionsall trajectories between to search the forphosphorylation intermolecular domain interactions of PLB between and SERCA. the The phosphorylation phosphorylation domain domain of of PLB PLB and interacts SERCA. with The all phosphorylationfour functional domains domain of of SERCA: PLB interacts namely, nucleotide-binding with all four functiona (N), actuatorl domains (A), of phosphorylation SERCA: namely (P),, nucleotideand TM domains-binding (Figure (N), actuator3a). We found (A), phosphorylation that most of the interactions(P), and TM are domains formed ( transientlyFigure 3a). as We reflected found thatby occupancy most of the factor interactions values lessare formed than 0.3; transiently transient as interactions reflected by are occupancy observed primarilyfactor values between less than the 0.3;phosphorylation transient interactions domain ofare PLB observed and the primarily N-, P-, and between A-domains the phosphorylation (Figure3b). The domain residues of of PLB SERCA and thethat N interact-, P-, and with A- thedomains N-terminal (Figure phosphorylation 3b). The residues domain of SERCA of PLB that include interact Trp107, with Asn111, the N-terminal Asn114, phosphorylationTyr122, Arg324, and domain Lys328 of (Figure PLB include3a), whereas Trp107, PLB Asn111,residues Met1,Asn114 Lys3,, Tyr122, Tyr6, pSer16, Arg324 Met20,, and Lys328 Gln22, (Fiandgure Arg25 3a),interact whereas with PLB SERCA residues in Met1, the microsecond Lys3, Tyr6, time pSer16, scale Met20, (Figure Gln223c,d)., and Arg25 interact with SERCA in the microsecond time scale (Figure 3c,d).

Int. J. Mol. Sci. 2020, 21, 7261 5 of 15 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 5 of 15

FigureFigure 3. 3. (a(a) )Depiction Depiction of of residues residues involved involved in in the the interaction interaction of of the the NT NT domain domain of of p phosphorylatedhosphorylated PLBPLB with with SERCA. SERCA. The The color color scale scale shows shows the the residues residues located located within within 3.0 3.0Å of Å p ofhosphorylated phosphorylated PLB PLB N- terminalN-terminal domain domain ranging ranging from from high high(red) (red)to low to (blue) low (blue) occupancy occupancy fraction fraction (OF) values (OF). values. (b) Per amino (b) Per acidamino occupancy acid occupancy contacts contacts of pPLB of pPLB N-terminal N-terminal domain domain with with each each SERCA SERCA domain domain in in the the simulated simulated systems.systems. (c (c) )P Phosphorylatedhosphorylated PLB PLB NT NT residues residues involved involved in in the the interaction interaction with with SERCA. SERCA. The The color color scale scale showsshows the the residues residues located located within within 3.0 3.0 Å Å of of SERCA SERCA ranging ranging from from high high (red) (red) to to low low (blue) (blue) OF OF values. values. ((dd)) Per Per amino amino aci acidd occupancy occupancy contacts contacts of of SERCA SERCA with with the the NT NT domain domain of of p phosphorylatedhosphorylated PLB PLB in in the the simulatedsimulated systems. systems. Residue Residue pS16 pS16 is is highlighted highlighted in in orange. orange.

TheThe analysis analysis showed showed that that SERCA SERCA residues residues Arg324 Arg324 and and Lys328, Lys328, as as well well as as PLB PLB residue residue p pSer16Ser16 havehave substantially substantially high higherer occup occupancyancy values values (OF (OF >> 0.750.75)) compared compared to to other other residues residues located located at at the the SERCASERCA–pPLB–pPLB interface interface (Figure (Figure 3 3)).. Based Based on on the complementary electrostatic properties of thesethese residues,residues, their their accessibility accessibility to to the the cytosol,, and and their their relative relative position position in the initial structure (R = 25Å25 Å), ), itit is is likely likely that that these these residues residues engage engage in inhighly highly specific specific interactions interactions in the in theSERCA SERCA–pPLB–pPLB complex. complex. To testTo test this this hypothesis hypothesis,, we we calculated calculated inter inter-residue-residue distance distancess between between pSer16 pSer16 and and Arg324 Arg324/Lys/Lys328328 of of SERCASERCA.. We We found found that that except except for for trajector trajectoryy MD2, MD2, residue residue pSer16 pSer16 of PLB of PLB is recruited is recruited by a bycationic a cationic site ofsite SERCA of SERCA formed formed by residues by residues Arg324 Arg324 and Lys328. and Lys328. Residue Residue Arg324 Arg324 forms forms salt bridge salt bridge interactions interactions with thewith phosphate the phosphate group group of Ser19 of Ser19 in five in fiveof the of theMD MD simulations simulations of ofthe the complex complex (Figure (Figure 44,, trajectories trajectories MD4MD4-MD8,-MD8, red redline). line). These These residues residues interact interact either either through through monodentate- monodentate or bidentate-like- or bidentate geometries-like geometrin our simulations.ies in our Nonetheless, simulations. the Nonetheless, bidentate geometrythe bidentate is preferred geometry over theis preferred monodentate over one the in monodentatethe trajectories one (Figure in the4 trajectories), in agreement (Figure with 4), studies in agreement showing with that studies the former showing is energetically that the former more is energeticallyfavorable in more solution favorable [23]. We in solution argue that [23] a. mixtureWe argue of that monodentate a mixture of and monodentate bidentate geometries and bidentate can geometriesfacilitate the can reversible facilitate interaction the reversible between interaction Arg324 between and the phosphorylationArg324 and the sitephosphorylation of PLB. For example, site of PLB.reversibility For example, of this interactionreversibility is ofobserved this interaction in the trajectory is observed MD5, where in the the trajectory interaction MD5, between wherethere the µ interactionresidues is between disrupted there at t =residues2.4 s (Figure is disrupted4). The at new t = 2.4 salt µs bridge (Figure between 4). The Lys328 new salt and bridge the phosphate between Lys328group ofand pSer16 the phosphate is formed more group frequently of pSer16 in is the formed complex, more and frequently it was observed in the in complex, trajectories and MD1 it was and observedMD3–MD8 in (Figure trajectories4). We MD1 also found and MD3 that– thisMD8 interaction (Figure 4 is). relativelyWe also more found stable that in this the interaction microsecond is µ relativelytime scale more than stable that formedin the microsecond between Arg324 time scale and pSer16,than that lasting formed for between up to 3.6 Arg324s (e.g., and trajectory pSer16, lastingMD8, Figurefor up4 to). 3.6 Notably, µs (e.g., we trajectory found that MD8, the Figure interaction 4). Notably, between we pSer16 found and that Arg324 the interaction/Lys328 stabilizesbetween pSa disordereder16 and Arg324/Lys328 structure of PLBstabilizes residues a disordered Thr17-Asn27 structure (Figure of 4PLB). This residues finding Thr is17 important-Asn27 (Fig becauseure 4). Thisformation finding of isα important-helical structure because of formation residues aroundof α-helical residue structure Asn27 of (Lys27 residues in human around PLB) residue is required Asn27 (Lys27 in human PLB) is required for SERCA inhibition [24–26]; thus, stabilization of an unfolded structure by pSer16-Arg324/Lys328 illustrates the functional importance of this interaction in the SERCA–pPLB complex.

Int. J. Mol. Sci. 2020, 21, 7261 6 of 15 for SERCA inhibition [24–26]; thus, stabilization of an unfolded structure by pSer16-Arg324/Lys328 illustratesInt. J. Mol. Sci. the 20 functional20, 21, x FOR importancePEER REVIEW of this interaction in the SERCA–pPLB complex. 6 of 15

Figure 4. Distance evolution is shown between pSer16–Arg324 (black trace) and pSer16–Lys328 (red Figure 4. Distance evolution is shown between pSer16–Arg324 (black trace) and pSer16–Lys328 (red trace). Distances between were calculated between the phosphorous atom of pSer16 and the Cζ and trace). Distances between were calculated between the phosphorous atom of pSer16 and the Cζ and Nζ atoms of Arg324 and Lys328, respectively. Here, we consider pSer16–Arg324 and pSer16–Lys328 Nζ atoms of Arg324 and Lys328, respectively. Here, we consider pSer16–Arg324 and pSer16–Lys328 pairs to form favorable interactions at distance < 6 Å [27]; the panels on the right show the structure pairs to form favorable interactions at distance < 6 Å [27]; the panels on the right show the structure of of pSer16, Arg324, and Lys328 at the end of each MD trajectory of the SERCA–pPLB complex. SERCA pSer16, Arg324, and Lys328 at the end of each MD trajectory of the SERCA–pPLB complex. SERCA and phosphorylated PLB are shown as white and purple ribbons, respectively; pSer16, Arg324, and and phosphorylated PLB are shown as white and purple ribbons, respectively; pSer16, Arg324, and Lys328 are shown as spheres. Lys328 are shown as spheres.

2.3.2.3. EEffectsffects ofof PLBPLB PhosphorylationPhosphorylation onon TMTMDomain Domain InteractionsInteractions TheThe crystalcrystal structurestructure ofof thethe inhibitoryinhibitory SERCA–PLBSERCA–PLB complex,complex, thethe TMTM domaindomain ofof PLBPLB bindsbinds toto aa canonicalcanonical sitesite formedformed betweenbetween TMTM heliceshelices M2,M2, M6,M6, andand M9M9 [17[17]],,and andprevious previousstudies studies havehave shownshown thatthat reliefrelief of PLB PLB-induced-induced SERCA SERCA inhibition inhibition by by Ca Ca2+ may2+ may occur occur without without changes changes in the in binding the binding mode modeof PLB of PLBin the in canonicalthe canonical site site[28] [28. To]. To determine determine if if phosphorylated phosphorylated PLB PLB induces induces changes in in the the interactioninteraction between between the the TM TM domain domain of of the the two two proteins, proteins we, we plotted plotted occupancy occupancy maps maps for for all all residues residues at theat theSERCA–pPLB SERCA–pPLB interface interface over the over combined the combined 22.7 µ s 22.7 of simulation µs of simulation time. We time found. We that found the canonical that the pocketcanonical of SERCA pocket is of predominantly SERCA is predominantly occupied (OF occupied> 0.8) by (OF phosphorylated > 0.8) by phosphorylated PLB in the microsecond PLB in the timescalemicrosecond (Figure timescale5a). SERCA (Figure residues5a). SERCA with residues the highest with the occupancy highest occupancy include Val89, include Leu96, Val89 Ile103,, Leu96, andIle103 Val104, and ofVal104 TM helixof TM M2, helix Leu802, M2, Leu802 and Thr805, and Thr805 of TM of helix TM M6,helix and M6, Trp932 and Trp932 and Leu939and Leu939 of TM of helixTM helix M9 (FigureM9 (Figure5b). 5b Similarly,). Similarly PLB, PLB residues residues that that interact interact directly directly with with the the TM TM domain domain of of SERCA, SERCA, includingincluding Leu31,Leu31, Asn34, Ile38 Ile38,, Leu42, Leu42, Ile48 Ile48,, and and Val49 Val49,, are are predicted predicted to tooccupy occupy the the canonical canonical site site(OF (OF> 0.8)> 0.8) upon upon PLB PLB phosphorylation phosphorylation (Figure (Figure 5b,c5b,c).). T Thesehese residues residues establish importantimportant intermolecularintermolecular interactionsinteractions thatthat areare necessarynecessary forfor thethe stabilitystability of of the the complex complex between between unphosphorylated unphosphorylated PLBPLB andand SERCASERCA [[17]17].. Therefore Therefore,, the the high high occupancy occupancy of of these these sites sites in inthe the SERCA SERCA–pPLB–pPLB complex complex suggests suggests that thatPLB PLBphosphorylation phosphorylation does does not induce not induce changes changes in the in interaction the interaction between between the TM the domains TM domains of both of bothproteins. proteins.

Int. J. Mol. Sci. 2020, 21, 7261 7 of 15 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 7 of 15

FigureFigure 5. 5. (a,b) Depiction of residues involved in the interaction interaction of the the TM TM domain domain of of pPLB pPLB with with SERCA.SERCA. TheThe colorcolor scale shows the residues located within 3.0 Å of pPLB NT ranging from high (red) toto lowlow (blue)(blue) occupancyoccupancy factorfactor values;values; ((cc,,dd)) perper aminoamino acidacid occupancyoccupancy contactscontacts ofof thethe TMTM domaindomain ofof phosphorylatedphosphorylated PLBPLB withwith SERCA;SERCA; ((ee)) time-dependenttime-dependent evolutionevolution ofof thethe hydrogenhydrogen bondbond betweenbetween PLBPLB residueresidue Asn34Asn34 andand SERCASERCA residueresidue Gly801Gly801 calculatedcalculated fromfrom eacheach independentindependent MDMD trajectory.trajectory.

Site-directedSite-directed mutagenesis,mutagenesis, X-rayX-ray crystallography,crystallography, andand molecularmolecular simulationssimulations havehave shownshown thatthat residueresidue Asn34Asn34 ofof PLBPLB playsplays aa centralcentral rolerole inin SERCASERCA inhibitioninhibition [17[17,29,30],29,30].. OurOur occupancyoccupancy analysisanalysis showedshowed thatthat Asn34Asn34 occupiesoccupies the the SERCA–PLB SERCA–PLB interface interface throughout throughout the the entire entire simulation simulation time time (OF (OF= 1,= Figure1, Figure5c,d), 5c,d and), and this this might might suggest suggest that that the the structural structural model model of of SERCA–pPLB SERCA–pPLB reported reported here here does does notnot represent the the non non-inhibited-inhibited state state of ofthe the complex. complex. We have We have recently recently shown shown that PLB that residue PLB residue Asn34 Asn34forms formsa stable a stablehydrogen hydrogen bond bondwith the with backbone the backbone oxygen oxygen of SERCA of SERCA residue residue Gly801, Gly801, and and that thatformation formation and and disruption disruption of this of thisinteraction interaction is a is hallmark a hallmark of PLB of PLB inhibition inhibition and and reactivation reactivation of ofSERCA SERCA–PLB–PLB [28,30,31] [28,30,.31 Therefore,]. Therefore, we analyzed we analyzed time-dependent time-dependent hydrogen hydrogen bonding bonding between between Asn34– Asn34–Gly801Gly801 for each for independenteach independent MD MD trajectory trajectory of the of the SERCA SERCA–pPLB–pPLB complex. complex. T Thehe hydro hydrogengen bond bond betweenbetween Asn34–Gly801Asn34–Gly801 is stable for more than 90% of the simulation time in only two trajectories,trajectories, MD1MD1 andand MD6MD6 (Figure(Figure5 e),5e) but, but it it is is largely largely disrupted disrupted in in most most of of the the trajectories. trajectories. This This hydrogen hydrogen bond bond is formedis formed for for about about 58% 58% of theof the time time in MD7,in MD7, but but it is it only is only transiently transiently present present (i.e., (i.e., less less than than 15% 15% of the of simulationthe simulation time) time) in trajectories in trajectories MD2 MD2 through through MD5 MD5 (Figure (Figure5e). 5e).

2.4.2.4. Phosphorylated PLB Induces Functional Transitions inin SERCASERCA TheThe changes in the the hydrogen hydrogen bonding bonding pattern pattern of of Asn34 Asn34–Gly801–Gly801 are are consistent consistent with with release release of ofSERCA SERCA inhibition inhibition and and suggest suggest that that the the structural structural ensemble ensemble of SERCA of SERCA–pPLB–pPLB represents represents the non the- non-inhibitoryinhibitory state stateof theof complex. the complex. Therefore Therefore,, we asked we whether asked disruption whether disruption of the inhibitory of the contacts inhibitory by contactsPLB phosphorylation by PLB phosphorylation induces functional induces functional structural structural changes changes in the in cytosolic the cytosolic headpiece headpiece that that are α α areassociated associated with with SERCA SERCA activation. activation. To To address address this this question, question, we we first first measured Cα-C-Cα distancesdistances β β betweenbetween SERCASERCA residuesresidues Arg139Arg139 (A-domain)(A-domain) andand Asp426Asp426 (loop(loop NNβ55--β6)6) (Figure(Figure6 6aa).). WeWe chose chose these these residuesresidues becausebecause previousprevious studiesstudies havehave shownshown thatthat theirtheir interactioninteraction correlatescorrelates withwith SERCASERCA activityactivity µ inin vitrovitro [[32,33]32,33].. WeWe excludedexcluded thethe firstfirst 0.20.2 µss fromfrom eacheach MDMD trajectorytrajectory toto avoidavoid biasesbiases associatedassociated withwith thethe initialinitial configurationconfiguration ofof thethe SERCA–pPLBSERCA–pPLB complex.complex. We also generated anan MD ensembleensemble ofof thethe unphosphorylatedunphosphorylated PLBPLB boundbound toto SERCA by combining fivefive MD trajectories generated previously byby µ ourour groupgroup intointo aa singlesingle trajectorytrajectory (total (total simulation simulation time time of of 17.8 17.8 sµs [34 [34]]).) These. These trajectories trajectories are are used used as as a controla control here here to to resolve resolve precise precise structural structural changes changes induced induced by by PLB PLB phosphorylation. phosphorylation. InIn thethe absenceabsence ofof PLBPLB phosphorylation,phosphorylation, thethe interresidueinterresidue distancedistance Arg139–Asp426Arg139–Asp426 fitsfits aa two-Gaussiantwo-Gaussian distribution,distribution, withwith a a structural structural population population with with means means at 21 Å, at and 21 Å a ,broader and a distribution broader distribution representing representing a structurally a dynamicstructurally population dynamic with population means atwith 27Å means (Figure at6 b).27 Å The (Figure absence 6b) of. The direct absence interaction of direct between interaction these residuesbetween isthese consistent residues with is consistent a catalytically with incompetenta catalytically structure incompetent of SERCA structure [32 ,of33 ].SERCA In the [32,33] presence. In ofthe bound presence phosphorylated of bound phosphorylated PLB, we resolved PLB, five we subpopulations resolved five subpopulationsof Arg139–Asp426: of Arg139 three structural–Asp426: β β statesthree withstructural means states at 20, with 27, andmeans 33 Å,at which20, 27, are and consistent 33 Å , which the A-domainare consistent and thethe loopA-domain N 5- and6 being the loop Nβ5-β6 being spatially separated, and two populations with means at 7.5 and 11 Å , which represent the binding of the A-domain and the loop Nβ5-β6 (Figure 6c). These results from our

Int. J. Mol. Sci. 2020, 21, 7261 8 of 15

spatiallyInt. J. Mol. Sci. separated, 2020, 21, x andFOR PEER two populationsREVIEW with means at 7.5 and 11 Å, which represent the binding8 of 15 of the A-domain and the loop Nβ5-β6 (Figure6c). These results from our unbiased MD simulations indicateunbiased that MD upon simulations PLB phosphorylation indicate that upon and in PLB the phosphorylation absence of Ca2+, theand headpiece in the absence in SERCA–pPLB of Ca2+, the populatesheadpiece structures in SERCA that–pPLB are distinctpopulates from str thoseuctures populated that are by distinct the unphosphorylated from those populated SERCA–PLB by the complex.unphosphorylated More importantly, SERCA– thisPLB structural complex. shiftMore leads importantly, to a reduction this structural in the distance shift leads between to aA-domain reduction andin the the distance loop Nβ 5- betweenβ6, in agreement A-domain with and the the notion loop that Nβ5 PLB-β6, phosphorylation in agreement with stabilizes the notion a more that compact PLB structurephosphorylati of SERCA’son stabilizes headpiece a more [22 compact]. structure of SERCA’s headpiece [22].

FigureFigure 6. 6.( a(a)) Structure Structure of of the the SERCA SERCA cytosolic cytosolic headpiece headpiece showing showing each functional each functional N-, A-, N and-, A P-domain-, and P- indomain green, in red, green, and blue, red, respectively.and blue, respectively. We also show We the also location show ofthe the location functional of the loop functional Nβ5-β6 (magenta).loop Nβ5- Yellowβ6 (magenta) spheres. Yellow show thespheres position show of the residues position used of residues to calculate used interdomain to calculate distanceinterdomain between distance the A-domainbetween the and A loop-domain Nβ5- andβ6. C loopα-C α Nβ5distance-β6. C distributionsα-Cα distance between distributions Arg139–Asp426 between obtained Arg139–Asp426 for the (obtainedb) SERCA–PLB for the ( andb) SERCA (c) SERCA–pPLB–PLB and (c) complexes. SERCA–pPLB Discrete complexes. distances Discrete were distances calculated were from calculated crystal structuresfrom crystal of structures the inhibitory of the SERCA–PLB inhibitory SERCA complex–PLB (4kyt complex, orange (4kyt dashed, orange line) dashed and SERCA line) and bound SERCA to Cabound2+/AMP-PCP to Ca2+/A (MP1vfp-PCP, green (1vfp dashed, green line). dashed line).

WeWe charted charted changes changes in in the the rotation rotation and and displacement displacement of the of N-, the P-, N and-, P- A-domains, and A-domains to gain furtherto gain insightfurther into insight the functional into the transitions functional of transitions SERCA that of occur SERCA in response that occur to PLB in phosphorylation. response to PLB We foundphosphorylation that both N-. We and found P-domains, that both the N- catalytic and P-domains elements, the of SERCA,catalytic occupyelements the of structuralSERCA, occupy space 2+ definedthe structural by the Ca space/nucleotide-bound defined by the Ca state2+/nucleoti of the pumpde-bound (Figure state7a,b). of This the pump structural (Figure space 7a,b includes). This thestructural crystal space structure includes of the the pre-hydrolysis crystal structure (cluster of the C2, pre blue-hydrolysis circles) (cluster and post-hydrolysis C2, blue circles) (cluster and post C3,- greenhydrolysis circles) (cluster E1 states C3, green of the circles) pump. E1 Conversely, states of the the pump. structural Conversely, map revealed the structural that the map A-domain revealed 2+ adoptsthat the a A structural-domain arrangementadopts a structural similar arrangement to that of SERCA similar in to the that absence of SERCA of Ca in the(Figure absence7c), whichof Ca2+ 2+ is(Figure defined 7c primarily), which is by defined crystal primarily structures by of crystal SERCA structures bound to of Mg SERCA, PLB, bound and theto Mg regulatory2+, PLB, proteinand the 2+ sarcolipinregulatory (cluster protein C2, sarcolipin blue circles). (cluster The C2, absence blue ofcircles) a Ca . -bound-likeThe absence orientation of a Ca2+-bound of the A-domain-like orientation in the 2+ trajectoriesof the A-domain of SERCA–pPLB in the trajectories is consistent of SERCA with– bindingpPLB is ofconsistent Ca ions with and binding correct charge of Ca2+ neutralization ions and correct of thecharge transport neutralization sites is required of the transport for the formation sites is required of a competent for the formation orientation of of a this competent domain orientation [35]. Overall, of thethis structural domain [35] shifts. Overall, in the t catalytiche structural N- and shifts P- in domains the catalytic in the N SERCA–pPLB- and P- domains complex in the SERCA are generally–pPLB similarcomplex to are those generally required similar for activation to those of required SERCA, for thus activation indicating of thatSERCA PLB, phosphorylationthus indicating that alone PLB is suphosphorylationfficient to populate alone a non-inhibitedis sufficient to statepopulate of the a pumpnon-inhibi (Figureted7 stated). of the pump (Figure 7d).

Int. J. Mol. Sci. 2020, 21, 7261 9 of 15

Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 9 of 15

FigureFigure 7. Rotation 7. Rotation/displacement/displacement map map ofof the ( (aa)) N N-domain,-domain, (b () bP)-domain P-domain,, and and(c) A (-cdomain) A-domain of SERCA of SERCA obtained from all trajectories combined of the SERCA–pPLB complex; the map is built based on the obtained from all trajectories combined of the SERCA–pPLB complex; the map is built based on crystal structures of SERCA, as described in reference [36]; (d) schematic representation of the Post- the crystal structures of SERCA, as described in reference [36]; (d) schematic representation of the Albers pumping cycle of SERCA linking biochemical and structural intermediates of the pump. C1– Post-Albers pumping cycle of SERCA linking biochemical and structural intermediates of the pump. C6 represent specific biochemical intermediates obtained through clustering analysis of all available C1–C6crystal represent structures specific of the biochemical pump. intermediates obtained through clustering analysis of all available crystal structures of the pump. 3. Discussion 3. Discussion We present an atomistic structure of phosphorylated PLB bound to SERCA and its structural dynamicsWe present under an Ca atomistic2+-free conditions structure, thus of providing phosphorylated new insight PLBs into bound the mechanism to SERCA for and regulation its structural 2+ dynamicsof SERCA under by CaPLB phosphorylation-free conditions,. In thus agreement providing with new EPR insights and NMR into spectroscopy the mechanism studies for [9,15] regulation, of SERCAMD simulations by PLB phosphorylation.showed that the N-terminal In agreement domain with of phosphorylated EPR and NMR PLB spectroscopy is largely unstructured studies [9,15], MDin simulations the complex, showed but it that forms the aN-terminal stable heterodimeric domain of complex phosphorylated with SERCA PLB is in largely the microsecond unstructured in the complex,timescale. butThe interaction it forms a stableof the disordered heterodimeric N-terminal complex phosphorylation with SERCA domain in the of microsecond PLB with SERCA timescale. Theis interaction enabled primarily of the disordered by the phosphate N-terminal group phosphorylationof pSer16. This interaction domain al ofso PLB facilit withates SERCAburial of isPLB enabled 13 primarilyresidues by Lys3 the and phosphate Tyr6, in groupagreement of pSer16. with NMR This spectroscopy interaction studies also facilitates showing that burial C ofresonances PLB residues of these residues are almost completely quenched upon PLB phosphorylation by spin-labeled SERCA Lys3 and Tyr6, in agreement with NMR spectroscopy studies showing that 13C resonances of these at position Cys674 [15]. Furthermore, we found that PLB phosphorylation does not disrupt the residuesheterodimeric are almost SERCA completely–pPLB quenched complex uponin the PLB microsecond phosphorylation timescale, by and spin-labeled also does SERCA not induce at position Cys674measurable [15]. Furthermore, changes in the we orientation found that of PLB PLB’s phosphorylation TM helix in the does complex not disrupt. This observation the heterodimeric is SERCA–pPLBimportant because complex we in used the microsecond the crystal structure timescale, of the and inhibited also does SERCA not– inducePLB complex, measurable and our changes in thesimulations orientation validate of PLB’s previous TM helix EPR in the spectroscopy complex. This studies observation suggesting is important that the TM because domain we used the crystalphosphorylated structure PLB of the binds inhibited to SERCA SERCA–PLB in an orientation complex, and similar our to simulations that of the validate inhibitory previous EPRunphosphorylated spectroscopy studies PLB [16] suggesting. The agreement that the between TM domain experiments phosphorylated and simulations PLB binds provide to solidSERCA in an orientationevidence that similar the unbiased to that of simulations the inhibitory reported unphosphorylated in this study recapitulate PLB [16 the]. The intrinsic agreement structural between experimentsfeatures of and the SERCA simulations–pPLB providecomplex solidin the evidencecell. that the unbiased simulations reported in this NMR spectroscopy studies have suggested that upon phosphorylation, the N-terminal study recapitulate the intrinsic structural features of the SERCA–pPLB complex in the cell. phosphorylation domain of PLB interacts with a preferred binding pocket located between the P- and NNMR-domains spectroscopy of SERCA [15] studies. Contrary have to conclusions suggested in these that studies upon, phosphorylation,we found that the disordered the N-terminal N- phosphorylationterminal domain domain does not of bind PLB to a interacts well-defined with region a preferred of the SERCA binding headpiece, pocket and located instead samples between the P- anddifferent N-domains orientations of SERCA and covers [15 ].a broad Contrary area of to the conclusions cytosolic domain in these of SERCA studies, that we includes found all that the disorderedthree N- N-terminal, P-, and A- domains. domain Nonetheless does not bind, our to analysis a well-defined revealed the region formation of the of SERCA stable electrostatic headpiece, and insteadinteractions samples involving different pSer16 orientations of PLB and and residues covers a Arg324/Lys328 broad area of of the SERCA. cytosolic We found domain that of the SERCA thatinteraction includes allbetween three Arg324/Lys328 N-, P-, and A- and domains. pSer16 is characterized Nonetheless, by ourdiverse analysis salt-bridge revealed configurations the formation. of stable electrostatic interactions involving pSer16 of PLB and residues Arg324/Lys328 of SERCA. We found that the interaction between Arg324/Lys328 and pSer16 is characterized by diverse salt-bridge configurations. This structural heterogeneity of this interaction is likely a functional feature of this Int. J. Mol. Sci. 2020, 21, 7261 10 of 15 interaction and likely contributes to both the adaptability of the complex as well as to the reversibility of the regulatory SERCA–pPLB interaction at physiological conditions [37]. We also found that this pSer16-mediated interaction stabilizes an unfolded structure of residues Thr17-Asn27, thus increasing the structural disorder of this region and favoring the disruption of key intermolecular contacts between SERCA and PLB [24–26]. Structural order in this region has been proposed to participate in important SERCA–PLB inhibitory contacts [38], thus further supporting the functional role of Arg324/Lys328 in PLB phosphorylation-mediated release of SERCA inhibition. The functional interaction between pSer16 and Arg324/Lys328 also helps explain previous findings showing that phosphorylation facilitates binding of PLB’s N-terminal phosphorylation domain to SERCA [15,39]. Both Arg324 and Lys328 are highly conserved in mammalian species and have been associated with important regulatory functions that include and coupling of SERCA [40,41]. Owing to the functional importance of these residues and their intrinsic electrostatic properties at physiological conditions, we propose that Arg324/Lys328 likely constitutes the canonical binding motif required to stabilize the heterodimeric SERCA–pPLB complex. Atomistic simulations showed that the SERCA–pPLB complex is stable in the microsecond time scale, so we asked whether PLB phosphorylation disrupts key inhibitory interactions between the two proteins. We focused on an intermolecular hydrogen bond between the side chain of PLB residue Asn34 and the backbone oxygen of SERCA residue Gly801 because this interaction is required for SERCA inhibition. In two of the MD trajectories of SERCA–pPLB, the complex forms a hydrogen bond between PLB residue Asn34 and SERCA residue Gly801. These hydrogen bond, which is an essential intermolecular contact responsible for PLB inhibition of SERCA [25,28,30,31,38], remains intact for most of the simulation time in these trajectories. Nevertheless, we found that this inhibitory interaction is disrupted in most of the MD trajectories of the SERCA–pPLB complex. In these cases, the initially formed hydrogen bond formed between Asn43 and Gly801 is present transiently in the trajectories, and several hydrogen bond formation and disruption events are observed in the time scales used here. Our dynamic structural model also showed that PLB phosphorylation acts upon Asn34–Gly801 and renders the hydrogen bond interaction between these two residues unstable. Although these structural changes have not been directly observed in experimental studies, our findings are consistent with previous studies showing disruption of this interaction is a hallmark of Ca2+-mediated relief of SERCA inhibition by PLB [28]. It is also worth noting that previous studies showed that Ca2+ binding to SERCA–PLB completely breaks the inhibitory interaction Asn34–Gly801 in the microsecond timescale, but the effect of PLB phosphorylation is not as robust as that induced by Ca2+ binding [28]. This observation suggests that while PLB phosphorylation alone disrupts the inhibitory contact between Asn34 and Gly801, Ca2+ binding in the transport sites is still required to induce complete disruption of this interaction. Nevertheless, destabilization of this interaction in the absence of Ca2+ agrees with the notion that PLB phosphorylation acts as a molecular switch that releases SERCA inhibition and further demonstrates that the structural model obtained in this study corresponds to that of the non-inhibited state of the regulatory complex. Finally, we ask whether PLB phosphorylation induces specific structural changes that are associated with the formation of a competent structure of SERCA. We found that in the SERCA–pPLB complex and upon release of the inhibitory contacts between Asn34/Gly801, SERCA residues Arg139 of the A-domain, and Asp426 of loop Nβ5-β6 come close together. This structural shift is similar to that found in the crystal structure of the activated Ca2+/ATP-bound complex [42], and it is important for early formation of the catalytically active structure of SERCA [32]. Moreover, we found that even in the absence of bound Ca2+, PLB phosphorylation is sufficient to guide the structural transitions of the N- and P-domains that are required to produce a competent structure of SERCA. These findings are in agreement with live-cell FRET experiments showing that PLB phosphorylation induces a redistribution of SERCA structural states to produce a more compact and catalytic competent structure of the pump [22]. It is reasonable to argue that pSer16 serves as a molecular anchor or tether that brings together the A-domain and loop Nβ5-β6 and stabilizes this bound conformation via favorable electrostatic interactions. Int. J. Mol. Sci. 2020, 21, 7261 11 of 15

However, we found that the phosphorylation site of PLB interacts primarily with a Arg324/Lys328, and that this site is located ~30 Å away from the A-domain–loop Nβ5-β6 interface. Based on this evidence, we propose a mechanistic model in which PLB phosphorylation relieves interdomain inhibitory contacts between SERCA and PLB and also serves as a mediator of allosteric signaling that favorably orients and strings together the catalytic elements required for ATP utilization, thus dictating the structural transitions required for regulation of Ca2+ transport in the heart [20,32,33,42]. In summary, we obtained an atomistic model of the SERCA–pPLB complex, mapped the interactions between the two proteins, and determined the effects of PLB phosphorylation on the functional dynamics of SERCA. We found that the unstructured N-terminal phosphorylation domain samples different orientations and covers a broad area of the cytosolic domain of SERCA, but it forms a stable complex via pSer16 interactions with a binding site formed by SERCA residues Arg324/Lys328. The dynamic nature of the interaction helps explain the adaptability and reversibility of SERCA–PLB inhibition, thus constituting a functional feature intrinsic of the regulatory mechanism. PLB phosphorylation does not affect the interaction between TM domains of SERCA and PLB. However, PLB phosphorylation releases inhibitory contacts within the complex and guides the structural transitions that are required to produce a competent structure of SERCA. We conclude that PLB phosphorylation serves as an allosteric molecular switch that releases inhibitory contacts and strings together the catalytic elements required for SERCA activation. Based on these findings, we provide novel hypotheses that can be tested by functional mutagenesis studies to validate the putative SERCA site Arg324/Lys328, and also by complementary spectroscopy and simulation studies [34] to establish the structure-function relationships of SERCA regulation by PLB. The mechanistic insights from this study will likely have important implications for our understanding of the molecular switches that underly regulation of Ca2+ transport, and will also help understand the structural features required for therapeutic modulation of SERCA in the heart.

4. Materials and Methods

4.1. Construction of the Initial Structure of SERCA Bound to Phosphorylated PLB We used the structure of the full-length structure of the R state of monomeric AFA-PLB (Cys36Ala, Cys41Phe, Cys46Ala) obtained using solution NMR (PDB: 2lpf,[43]) as the initial model to build the SERCA–pPLB complex. First, we aligned residues 28-51 of the solution NMR structure onto residues 27–50 of PLB4 in the crystal structure of SERCA–PLB (PDB: 4kyt,[17]). Upon alignment, we combined the NMR structure of the R state of PLB with the crystal structure of SERCA–PLB to produce an initial model of the complex. Finally, residues Ala36, Phe41, and Ala46 of AFA-PLB were changed to alanine to produce a model of native PLB. We performed initial 100-ns MD simulations of the complex in a 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) bilayer. The stereochemical quality of the initial models was verified with PROCHECK [44,45], showing no major aberrations in the geometry of the complexes. In this initial model, the cytosolic domain of PLB (Met1-Gln26) is intrinsically unstructured, whereas the TM domain (Asn27-Leu52) populate an α-helical structure. Furthermore, we found that cytosolic residues Met1-Leu7 and Gln22-Gln26 of PLB directly interact with SERCA, whereas residues Thr8-Pro21 (which includes pSer16) are primarily exposed to the solvent. This model was used as the starting conformation to simulate the dynamics of SERCA–pPLB.

4.2. Preparation of the SERCA–pPLB Complex We used the atomistic model of the full-length PLB bound to SERCA generated here study to simulate the SERCA-pPLB complex under Ca2+-free conditions. We modeled residue pSer16 of PLB, and SERCA residues Glu309 and Asp800 as unprotonated and residues Glu771 and Glu908 as protonated. In addition, we adjusted the pKa of other ionizable residues to a pH value of approximately 7.2 using PROPKA [46,47]. The complex was inserted in a pre-equilibrated 130 130 Å bilayer of × POPC lipids. We used the first layer phospholipids that surround SERCA in the E1 state [48] as a Int. J. Mol. Sci. 2020, 21, 7261 12 of 15 reference to insert the complex in the lipid bilayer. This initial system was solvated using TIP3P water molecules with a minimum margin of 30 and 25 Å in the z-axis between the edges of the periodic box and the cytosolic and luminal domains of SERCA, respectively. K+ and Cl- ions were added to neutralize the system and to produce a KCl concentration of ~100 mM.

4.3. Molecular Dynamics Simulations MD simulations of all systems were performed with NAMD [49] using periodic boundary conditions [50], particle mesh Ewald [51,52], a non-bonded cutoff of 12 Å, and a 2-fs time step. CHARMM36 force field topologies and parameters were used for the proteins [53], lipid [54], water and ions. The NPT ensemble was maintained with a Langevin thermostat (310 K) and an anisotropic Langevin piston barostat (1 atm). Fully solvated systems were first subjected to energy minimization, followed by gradually warming up of the systems for 200 ps. This procedure was followed by 10 ns of 1 2 equilibration with backbone atoms harmonically restrained using a force constant of 10 kcal mol− Å− . We performed 8 independent MD simulations of SERCA–pPLB for a total simulation time of 22.7 µs.

4.4. Structural Analysis and Visualization We calculated the backbone root mean square deviation (RMSD) for SERCA, cytosolic, and TM domains, and PLB, N-terminal phosphorylation, and TM domains. Changes in the secondary structure of PLB was computed employing the program dssp [55] included in the GROMACS package [56]. We also calculated the occupancy fraction (OF) of SERCA residues located within 3.0 Å of PLB and PLB residues located within 3.0 Å of SERCA using the select built-in tool implemented in GROMACS [56]. The presence of a hydrogen bond between the amide group of Asn34 and the backbone oxygen of Gly801 was calculated using HydrogenBondAnalysis library of MDAnalysis. We mapped MD structures onto the rotation-displacement map created using 74 crystal structures of SERCA to determine the effects of PLB phosphorylation on the structural dynamics of the cytosolic domain [36]. All SERCA structures were aligned to the Cα atoms of crystal structures 2c9m [57] using the following TM domain helices as reference: M7 (residues 831–855), M8 (residues 895–915), M9 (residues 933–948), and M10 (residues 966–994). Translation distances and rotation angles of the N-, P-, and A-domains were computed for the three cytosolic domains using 2c9m [57] as reference with the draw_rotation_axis.py script running on PyMOL (http://pymolwiki.org/index.php/RotationAxis). The programs PyMOL and VMD [58] were used for visualization and rendering of the structures.

Author Contributions: Conceptualization, L.M.E.-F.; methodology, R.A.-O. and L.M.E.-F.; formal analysis, R.A.-O.; writing—original draft preparation, L.M.E.-F.; writing—review and editing, R.A.-O.; funding acquisition, L.M.E.-F. All authors have read and agreed to the published version of the manuscript. Funding: This study was funded by the National Institutes of Health grant R01GM120142 to L.M.E.-F., and in part through computational resources and services provided by Advanced Research Computing at the University of Michigan, Ann Arbor. Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Abbreviations

A-domain Actuator domain EPR Electronic Paramagnetic Resonance FRET Fluorescence Resonance Energy Transfer MD Molecular dynamics N-domain Nucleotide-binding domain NMR Nuclear Magnetic Resonance P-domain Phosphorylation domain PDB Protein Data Bank PLB Phospholamban Int. J. Mol. Sci. 2020, 21, 7261 13 of 15 pPLB Phosphorylated phospholamban POPC 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine RMSD Root-mean-square deviation RMSF Root-mean-square fluctuation SERCA Sarcoendoplasmic reticulum Ca2+-ATPase SR Sarcoplasmic reticulum TM Transmembrane

References

1. MacLennan, D.H.; Kranias, E.G. Phospholamban: A crucial regulator of cardiac contractility. Nat. Rev. Mol. Cell Biol. 2003, 4, 566–577. [CrossRef][PubMed] 2. Moller, J.V.; Olesen, C.; Winther, A.M.; Nissen, P. The sarcoplasmic Ca2+-ATPase: Design of a perfect chemi-osmotic pump. Q. Rev. Biophys. 2010, 43, 501–566. [CrossRef][PubMed] 3. Koss, K.L.; Kranias, E.G. Phospholamban: A prominent regulator of myocardial contractility. Circ. Res. 1996, 79, 1059–1063. [CrossRef][PubMed] 4. Kranias, E.G.; Hajjar, R.J. Modulation of Cardiac Contractility by the Phopholamban/SERCA2a Regulatome. Circ. Res. 2012, 110, 1646–1660. [CrossRef][PubMed] 5. Metcalfe, E.E.; Traaseth, N.J.; Veglia, G. Serine 16 phosphorylation induces an order-to-disorder transition in monomeric phospholamban. Biochemistry 2005, 44, 4386–4396. [CrossRef] 6. Paterlini, M.G.; Thomas, D.D. The alpha-helical propensity of the cytoplasmic domain of phospholamban: A molecular dynamics simulation of the effect of phosphorylation and mutation. Biophys. J. 2005, 88, 3243–3251. [CrossRef] 7. Chen, Z.; Akin, B.L.; Jones, L.R. Mechanism of reversal of phospholamban inhibition of the cardiac Ca2+-ATPase by protein a and by anti-phospholamban monoclonal antibody 2D12. J. Biol. Chem. 2007, 282, 20968–20976. [CrossRef] 8. Toyoshima, C.; Asahi, M.; Sugita, Y.; Khanna, R.; Tsuda, T.; MacLennan, D.H. Modeling of the inhibitory interaction of phospholamban with the Ca2+ ATPase. Proc. Natl. Acad. Sci. USA 2003, 100, 467–472. [CrossRef] 9. Karim, C.B.; Zhang, Z.; Howard, E.C.; Torgersen, K.D.; Thomas, D.D. Phosphorylation-dependent Conformational Switch in Spin-labeled Phospholamban Bound to SERCA. J. Mol. Biol. 2006, 358, 1032–1040. [CrossRef] 10. Negash, S.; Yao, Q.; Sun, H.Y.; Li, J.H.; Bigelow, D.J.; Squier, T.C. Phospholamban remains associated with the Ca2+- and Mg2+-dependent ATPase following phosphorylation by cAMP-dependent . Biochem. J. 2000, 351, 195–205. [CrossRef] 11. Bidwell, P.; Blackwell, D.J.; Hou, Z.; Zima, A.V.; Robia, S.L. Phospholamban binds with differential affinity to calcium pump conformers. J. Biol. Chem. 2011, 286, 35044–35050. [CrossRef][PubMed] 12. James, Z.M.; McCaffrey, J.E.; Torgersen, K.D.; Karim, C.B.; Thomas, D.D. Protein-Protein Interactions in Calcium Transport Regulation Probed by Saturation Transfer Electron Paramagnetic Resonance. Biophys. J. 2012, 103, 1370–1378. [CrossRef][PubMed] 13. Mueller, B.; Karim, C.B.; Negrashov, I.V.; Kutchai, H.; Thomas, D.D. Direct detection of phospholamban and sarcoplasmic reticulum Ca-ATPase interaction in membranes using fluorescence resonance energy transfer. Biochemistry 2004, 43, 8754–8765. [CrossRef][PubMed] 14. Dong, X.; Thomas, D.D. Time-resolved FRET reveals the structural mechanism of SERCA-PLB regulation. BioChem. Biophys. Res. Commun. 2014, 449, 196–201. [CrossRef][PubMed] 15. Gustavsson, M.; Verardi, R.; Mullen, D.G.; Mote, K.R.; Traaseth, N.J.; Gopinath, T.; Veglia, G. of SERCA by phosphorylation-mediated conformational shift of phospholamban. Proc. Natl. Acad. Sci. USA 2013, 110, 17338–17343. [CrossRef][PubMed] 16. Martin, P.D.; James, Z.M.; Thomas, D.D. Effect of Phosphorylation on Interactions between Transmembrane Domains of SERCA and Phospholamban. Biophys. J. 2018, 114, 2573–2583. [CrossRef][PubMed] 17. Akin, B.L.; Hurley, T.D.; Chen, Z.; Jones, L.R. The structural basis for phospholamban inhibition of the calcium pump in sarcoplasmic reticulum. J. Biol. Chem. 2013, 288, 30181–30191. [CrossRef] 18. Espinoza-Fonseca, L.M.; Autry, J.M.; Thomas, D.D. Microsecond molecular dynamics simulations of Mg(2)(+)- and K(+)-bound E1 intermediate states of the calcium pump. PLoS ONE 2014, 9, e95979. [CrossRef] Int. J. Mol. Sci. 2020, 21, 7261 14 of 15

19. Das, A.; Gur, M.; Cheng, M.H.; Jo, S.; Bahar, I.; Roux, B. Exploring the conformational transitions of biomolecular systems using a simple two-state anisotropic network model. PLoS Comput. Biol. 2014, 10, e1003521. [CrossRef] 20. Espinoza-Fonseca, L.M.; Thomas, D.D. Atomic-level characterization of the activation mechanism of SERCA by calcium. PLoS ONE 2011, 6, e26936. [CrossRef] 21. Winters, D.L.; Autry, J.M.; Svensson, B.; Thomas, D.D. Interdomain fluorescence resonance energy transfer in SERCA probed by cyan-fluorescent protein fused to the actuator domain. Biochemistry 2008, 47, 4246–4256. [CrossRef][PubMed] 22. Pallikkuth, S.; Blackwell, D.J.; Hu, Z.; Hou, Z.; Zieman, D.T.; Svensson, B.; Thomas, D.D.; Robia, S.L. Phosphorylated Phospholamban Stabilizes a Compact Conformation of the Cardiac Calcium-ATPase. Biophys. J. 2013, 105, 1812–1821. [CrossRef][PubMed] 23. Mandell, D.J.; Chorny, I.; Groban, E.S.; Wong, S.E.; Levine, E.; Rapp, C.S.; Jacobson, M.P. Strengths of hydrogen bonds involving phosphorylated amino acid side chains. J. Am. Chem. Soc. 2007, 129, 820–827. [CrossRef][PubMed] 24. Glaves, J.P.; Trieber, C.A.; Ceholski, D.K.; Stokes, D.L.; Young, H.S. Phosphorylation and mutation of phospholamban alter physical interactions with the sarcoplasmic reticulum calcium pump. J. Mol. Biol. 2011, 405, 707–723. [CrossRef] 25. Kimura, Y.; Asahi, M.; Kurzydlowski, K.; Tada, M.; MacLennan, D.H. Phospholamban domain Ib mutations influence functional interactions with the Ca2+-ATPase isoform of cardiac sarcoplasmic reticulum. J. Biol. Chem. 1998, 273, 14238–14241. [CrossRef] 26. Young, H.S.; Jones, L.R.; Stokes, D.L. Locating phospholamban in co-crystals with Ca(2+)-ATPase by cryoelectron microscopy. Biophys. J. 2001, 81, 884–894. [CrossRef] 27. Donald, J.E.; Kulp, D.W.; DeGrado, W.F. Salt bridges: Geometrically specific, designable interactions. Proteins 2011, 79, 898–915. [CrossRef] 28. Fernandez-de Gortari, E.; Espinoza-Fonseca, L.M. Structural basis for relief of phospholamban-mediated inhibition of the sarcoplasmic reticulum Ca2+-ATPase at saturating Ca2+ conditions. J. Biol. Chem. 2018, 293, 12405–12414. [CrossRef] 29. Kimura, Y.; Kurzydlowski, K.; Tada, M.; MacLennan, D.H. Phospholamban inhibitory function is activated by depolymerization. J. Biol. Chem. 1997, 272, 15061–15064. [CrossRef] 30. Espinoza-Fonseca, L.M.; Autry, J.M.; Ramirez-Salinas, G.L.; Thomas, D.D. Atomic-level mechanisms for phospholamban regulation of the calcium pump. Biophys. J. 2015, 108, 1697–1708. [CrossRef] 31. Espinoza-Fonseca, L.M.; Autry, J.M.; Thomas, D.D. Sarcolipin and phospholamban inhibit the calcium pump by populating a similar metal ion-free intermediate state. BioChem. Biophys. Res. Commun. 2015, 463, 37–41. [CrossRef][PubMed] 32. Smolin, N.; Robia, S.L. A structural mechanism for calcium transporter headpiece closure. J. Phys. Chem. B 2015, 119, 1407–1415. [CrossRef][PubMed] 33. Raguimova, O.N.; Smolin, N.; Bovo, E.; Bhayani, S.; Autry, J.M.; Zima, A.V.; Robia, S.L. Redistribution of SERCA calcium pump conformers during intracellular calcium signaling. J. Biol. Chem. 2018, 293, 10843–10856. [CrossRef][PubMed] 34. Raguimova, O.N.; Aguayo-Ortiz, R.; Robia, S.L.; Espinoza-Fonseca, L.M. Dynamics-driven allostery underlies Ca2+-mediated release of SERCA inhibition by phospholamban. Biophys. J. 2020.[CrossRef] 35. Clausen, J.D.; Bublitz, M.; Arnou, B.; Montigny, C.; Jaxel, C.; Moller, J.V.; Nissen, P.; Andersen, J.P.; le Maire, M. SERCA mutant E309Q binds two Ca2+ ions but adopts a catalytically incompetent conformation. EMBO J. 2013, 32, 3231–3243. [CrossRef][PubMed] 36. Aguayo-Ortiz, R.; Espinoza-Fonseca, L.M. Linking Biochemical and Structural States of SERCA: Achievements, Challenges, and New Opportunities. Int. J. Mol. Sci. 2020, 21, 4146. [CrossRef] 37. Wright, P.E.; Dyson, H.J. Intrinsically disordered proteins in cellular signalling and regulation. Nat. Rev. Mol. Cell Biol. 2015, 16, 18–29. [CrossRef] 38. Trieber, C.A.; Douglas, J.L.; Afara, M.; Young, H.S. The effects of mutation on the regulatory properties of phospholamban in co-reconstituted membranes. Biochemistry 2005, 44, 3289–3297. [CrossRef] 39. Negash, S.; Sun, H.Y.; Yao, Q.; Goh, S.Y.; Bigelow, D.J.; Squier, T.C. Cytosolic domain of phospholamban remains associated with the Ca-ATPase following phosphorylation by cAMP-dependent protein kinase. Card. Sarcoplasmic Reticulum Funct. Regul. Contractility 1998, 853, 288–291. Int. J. Mol. Sci. 2020, 21, 7261 15 of 15

40. Autry, J.M.; Thomas, D.D.; Espinoza-Fonseca, L.M. Sarcolipin Promotes Uncoupling of the SERCA Ca(2+) Pump by Inducing a Structural Rearrangement in the Energy-Transduction Domain. Biochemistry 2016, 55, 6083–6086. [CrossRef] 41. Yamasaki, K.; Daiho, T.; Danko, S.; Suzuki, H. Multiple and distinct effects of mutations of Tyr(122), Glu(123), Arg(324), and Arg(334) involved in interactions between the top part of second and fourth transmembrane helices in sarcoplasmic reticulum Ca2+-ATPase—Changes in cytoplasmic domain organization during isometric transition of phosphoenzyme intermediate and subsequent Ca2+ release. J. Biol. Chem. 2004, 279, 2202–2210. 42. Toyoshima, C.; Mizutani, T. Crystal structure of the calcium pump with a bound ATP analogue. Nature 2004, 430, 529–535. [CrossRef] 43. De Simone, A.; Gustavsson, M.; Montalvao, R.W.; Shi, L.; Veglia, G.; Vendruscoo, M. Structures of the Excited States of Phospholamban and Shifts in Their Populations upon Phosphorylation. Biochemistry 2013, 52, 6684–6694. [CrossRef][PubMed] 44. Laskowski, R.A.; MacArthur, M.W.; Moss, D.S.; Thornton, J.M. PROCHECK—A program to check the stereochemical quality of protein structures. J. App. Cryst. 1993, 26, 283–291. [CrossRef] 45. Laskowski, R.A.; Rullmannn, J.A.; MacArthur, M.W.; Kaptein, R.; Thornton, J.M. AQUA and PROCHECK-NMR: Programs for checking the quality of protein structures solved by NMR. J. BioMol. NMR 1996, 8, 477–486. [CrossRef][PubMed] 46. Olsson, M.H.M.; Sondergaard, C.R.; Rostkowski, M.; Jensen, J.H. PROPKA3: Consistent Treatment of Internal and Surface Residues in Empirical pK(a) Predictions. J. Chem. Theory Comput. 2011, 7, 525–537. [CrossRef] [PubMed] 47. Sondergaard, C.R.; Olsson, M.H.M.; Rostkowski, M.; Jensen, J.H. Improved Treatment of Ligands and Coupling Effects in Empirical Calculation and Rationalization of pK(a) Values. J. Chem. Theory Comput. 2011, 7, 2284–2295. [CrossRef] 48. Norimatsu, Y.; Hasegawa, K.; Shimizu, N.; Toyoshima, C. Protein-phospholipid interplay revealed with crystals of a calcium pump. Nature 2017, 545, 193–198. [CrossRef] 49. Phillips, J.C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R.D.; Kale, L.; Schulten, K. Scalable molecular dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781–1802. [CrossRef] 50. Weber, W.; Hünenberger, P.H.; McCammon, J.A. Molecular Dynamics Simulations of a Polyalanine Octapeptide under Ewald Boundary Conditions: Influence of Artificial Periodicity on Peptide Conformation. J. Phys. Chem. B 2000, 104, 3668–3675. [CrossRef] 51. Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: An N log(N) method for Ewald sums in large systems. · J. Chem. Phys. 1993, 98, 10089–10092. [CrossRef] 52. Essmann, U.; Perera, L.; Berkowitz, M.L. A smooth particle mesh Ewald method. J. Chem. Phys. 1995, 103, 8577–8593. [CrossRef] 53. Best, R.B.; Zhu, X.; Shim, J.; Lopes, P.E.; Mittal, J.; Feig, M.; Mackerell, A.D., Jr. Optimization of the additive CHARMM all-atom protein force field targeting improved sampling of the backbone phi, psi and side-chain chi(1) and chi(2) dihedral angles. J. Chem. Theory Comput. 2012, 8, 3257–3273. [CrossRef][PubMed] 54. Klauda, J.B.; Venable, R.M.; Freites, J.A.; O’Connor, J.W.; Tobias, D.J.; Mondragon-Ramirez, C.; Vorobyov, I.; MacKerell, A.D., Jr.; Pastor, R.W. Update of the CHARMM all-atom additive force field for lipids: Validation on six lipid types. J. Phys. Chem. B 2010, 114, 7830–7843. [CrossRef][PubMed] 55. Kabsch, W.; Sander, C. Dictionary of protein secondary structure: Pattern recognition of hydrogen-bonded and geometrical features. Biopolymers 1983, 22, 2577–2637. [CrossRef][PubMed] 56. Pronk, S.; Pall, S.; Schulz, R.; Larsson, P.; Bjelkmar, P.; Apostolov, R.; Shirts, M.R.; Smith, J.C.; Kasson, P.M.; van der Spoel, D.; et al. GROMACS 4.5: A high-throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 2013, 29, 845–854. [CrossRef] 57. Jensen, A.M.L.; Sorensen, T.L.M.; Olesen, C.; Moller, J.V.; Nissen, P. Modulatory and catalytic modes of ATP binding by the calcium pump. EMBO J. 2006, 25, 2305–2314. [CrossRef] 58. Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual molecular dynamics. J. Mol. Graph. 1996, 14, 33–38. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).