| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* CNR Institute of Molecular Biology and Pathology and
Department of Biochemical Sciences, University "Sapienza", Rome, Italy; and
Department of Physiology, University of Wisconsin Medical School, Madison, Wisconsin, USA
1Correspondence: CNR Institute of Molecular Biology and Pathology, University Sapienza, P.le A.Moro 5, 00185 Rome, Italy. E-mail: gianni.colotti{at}uniroma1.it
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: EF hand • heart • PEF proteins • channels • excitation-contraction coupling • calcium-induced channel release
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4)
and excitation-contraction (e-c) coupling through the interaction with important cardiac transporters such as the ryanodine receptor (RyR2), the Na+-Ca2+ exchanger (NCX), the 
I subunit of the L-type Ca2+ channel and the sarcoplasmic reticulum Ca2+-ATPase (4
5
6
7
8)
. All these interactions appear to concur functionally and are presumed to help terminate CICR in ventricular myocytes.
Sorcin activation occurs at micromolar Ca2+ concentrations on saturation of the two high-affinity EF-hands, EF3 and EF2. The two functionally relevant calcium-binding sites do not form a structural pair as in most other PEF proteins. Given that EF2 and EF3 are connected by the long and rigid D helix (Fig. 1
A), sorcin activation requires that information of the Ca2+ binding event to the site with the highest affinity for the metal, EF3, be transmitted to EF2 through the D helix and from there to the rest of the molecule via the canonical EF2-EF1 coupling (9)
. In the information transfer process a key role is played by the extensive network of hydrophobic and hydrogen bonding interactions around the D helix (10
, 11)
. This model for sorcin activation, elaborated for the hamster protein, is applied in the present work to human sorcin in view of the extremely high sequence conservation (188 identical residues out of 198, Fig. 1A
).
|
The structural framework for sorcin activation in turn predicts that mutations at any of the strategic regions of the molecule, EF3, EF2, or the D helix, are likely to have profound effects on the protein functionality. In fact, F112L-sorcin (12
, 13)
, a natural variant of the F112 residue, which is located at the end of the D helix next to Asp113 (one of the acidic calcium ligands in the EF3 site), was discovered to be associated with hypertrophic cardiomyopathy and hypertension in two unrelated families (14)
. The F112L variant was found to lack the ability of wild-type sorcin to modulate RyR2 activity (14
, 15)
. Most recently, overexpression of F112L-sorcin in transgenic mice was observed to result in an impaired capacity of the variant to translocate to the membrane fraction in vitro and in an alteration of the e-c coupling processes due to the increase in Ca2+ spark duration, width, and frequency, the increase in NCX activity, and the slow inactivation of the L-type Ca2+ current (16)
.
The present characterization of F112L-sorcin aims at establishing whether the structural basis of the impaired function of the mutant lies in the decreased affinity for calcium and/or in the impairment of information transfer through the D helix. The X-ray crystal structure (2.5 Å resolution) shows that the F112L mutation has structural effects that do not involve solely the EF3 region but, quite unexpectedly, reach as far as the EF1 hand. The biochemical properties in turn point to the marked decrease in Ca2+ affinity as the major cause for the decreased interaction of F112L sorcin with its targets and hence for impaired activity of the protein.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Human wt-sorcin, hereafter called wt-sorcin, and the F112L mutant were expressed in Escherichia coli BL21(DE3) cells and were purified as described for Chinese hamster ovary recombinant sorcin (1)
.
Protein concentrations were determined spectrophotometrically at 280 nm using 29,400 as molar extinction coefficient on a monomer basis (1
, 10)
.
Crystallization
Crystallization of full-length F112L-sorcin was achieved with the hanging-drop vapor diffusion technique at 294 K. A 2 µl volume of a protein sample at 7 mg/ml was mixed with an equal amount of the reservoir solution containing (NH4)2SO4 (concentration range 0.8–1.5 M), 0.1 M MES-NaOH at pH values 5.5–6.5 and 12% dioxane. Crystals grew in 2–4 days to 
0.4 x 0.3 x 0.2 mm, in hexagonal space group P65 (a=b=64.385 Å and c=314.837 Å) and diffracted to beyond 2.5 Å resolution. The solvent content of the crystal was 50%.
Data collection, data analysis, and structure determination
For X-ray data collection, F112L-sorcin crystals were equilibrated in a cryo-solution made from the reservoir solution plus 25% (v/v) glycerol before flash-freezing in nitrogen gas at 100 K. Data were collected as 0.5 oscillation frames using the MAR CCD detector on the X-ray beamline ID14–2 at ESRF, Grenoble (France), at a wavelength of 1.0 Å.
Data analysis was performed using DENZO (17)
. Data were scaled using SCALEPACK (17)
and had an Rsym value of 8.2% and a 
2 value of 1.155 with a completeness of 96.8% extending to 2.5 Å resolution.
The structure was solved by molecular replacement using as search probe a polyalanine truncated model of the wt-sorcin monomer (RCSB entry 1JUO). The rotational and translational searches, performed with the program MOLREP (18)
in the resolution range of 10–3.0 Å, produced a solution corresponding to a tetramer (two sorcin dimers) in the asymmetric unit.
Refinement of the atomic coordinates and displacement parameters was carried out by means of Refmac5 (19)
. Model building was performed using the program COOT (20)
. Water molecules and sulfate ions were added to the model manually. The final model included 168 residues/monomer (residues 1–30 are not visible). The final Rfactor and Rfree at 2.5 Å resolution are 25.4% and 28.9%, respectively, with an rms of 0.013 Å on the bond lengths and of 1.6° on the bond angles.
The quality of the model was assessed by the program PROCHECK (21)
. The most favored regions of the Ramachandran plot contain 93.2% of nonglycine residues. The G-factor is used in the program Procheck to provide a measure of how "normal" or alternatively how "unusual" a given stereochemical property is. The G-factor is 0.3, which is better than –0.5, the minimal expected value for a geometrically correct structure. The structural and refinement statistics are listed in Table 1
.
|
Circular dichroism measurements
CD spectra were recorded at 20°C on a Jasco J-710 spectropolarimeter. In the near UV region (250–350 nm), the buffers used were 100 mM Tris-HCl buffer at pH 7.5 and 50 mM sodium phosphate buffer at pH 7.5 and 6.0. In the far UV region (200–250 nm) the same buffers were used but at a concentration of 5 mM. All buffers were pretreated with Chelex 100.
Analytical ultracentrifugation experiments
Analytical ultracentrifugation experiments were carried out on a Beckman Optima XL-I ultracentrifuge, on wt- and F112L-sorcin dialyzed vs. 0.1 M MES-NaOH at pH 6.0, 0.1 M Tris-HCl at pH 7.5, 50 mM sodium phosphate buffer at pH 6.0 and pH 7.0, all containing 1 mM EGTA. Sedimentation velocity experiments were performed at 40,000 rev/min and at 20°C. Data were collected at 280 nm, at a spacing of 0.003 cm with 3 averages in a continuous scan mode. Sedimentation coefficients were obtained with the software Sedfit (Dr. P. Schuck, National Institutes of Health, Bethesda, MD, USA). The values were reduced to water and 20°C (s20,w) using standard procedures. The buffer density and viscosity were calculated by the software Sednterp (John Philo, Thousand Oaks, CA, USA).
Sedimentation equilibrium experiments were performed at 23,000 and 28,000 rpm at 20°C. Data were collected every 3 h at a spacing of 0.001 cm with 10 averages in a step scan mode. Equilibrium was checked by comparing scans up to 24 h by the software Winmatch (program by David Yphantis). Data sets were edited by REEDIT (J. Lary, National Analytical Ultracentrifugation Center, Storrs, CT, USA) and analyzed by the software Sedphat (NIH). Data from different speeds were combined for global fitting. The goodness of the fit was judged using a combination of the fit statistics (data not shown) and the randomness of residuals.
Determination of Ca2+ affinity
Indirect titrations were carried out on a Fluoromax spectrofluorimeter in 0.1 M Tris-HCl buffer at pH 7.2 and 22°C in the presence of the fluorescent calcium chelator Fluo3, according to Harkins et al. (22)
and Mella et al. (10)
. The excitation wavelength was 488 nm (slit width, 2 nm); the increase in emission intensity due to calcium binding to Fluo3 was followed at 526 nm. Ca2+ contamination was reduced to <1.0 µM by treating the solutions and the glassware with Chelex 100. Two sets of independent experiments were carried out, which included control titrations of Fluo3 alone. The fluorescence intensity of Fluo3 at high calcium concentrations was taken as the higher asymptote. The fluorescence intensity corresponding to zero free Ca2+ concentration was determined at the end of each titration by the addition of 5 mM EGTA. Ca2+ affinity constants were estimated by fitting each experimental data set with the program CaLigator (23)
.
Surface plasmon resonance (SPR) measurements
The interaction of wt- and F112L-sorcin with the annexin VII N-terminal peptide was studied in SPR experiments performed on a BIACORE X system (Biacore AB, Uppsala, Sweden) according to Mella et al. (10)
. The changes in the observed SPR signal are expressed as resonance units (RU). The experiments were carried out at 25°C in 10 mM HEPES at pH 7.4, 0.15 M NaCl, and 0.005% surfactant P-20 (HBS-P buffer), treated with Chelex 100 and degassed. For the experiments as a function of Ca2+ concentration, CaCl2 or EGTA were added to the buffer. Measurements were performed at a flow rate of 20 µl/min with an immobilization level of the annexin VII N-terminal peptide corresponding to 400–1500 RU. Values of the plateau signal at steady state (Req) were calculated from kinetic evaluation of the sensorgrams using the BIAevaluation 3.0 software. A Scatchard analysis of the dependence of Req on the concentration of wt- and F112L-sorcin was also performed to assess the equilibrium dissociation constant at 20 µM Ca2+.
Cerebral artery myocytes isolation, permeabilization, and Ca2+ sparks recordings
Rat cerebral vascular smooth muscle cells (VSMCs) were isolated enzymatically from Sprague-Dawley rats (200–220 g) as described before (15)
. Saponin permeabilization and Ca2+ spark recordings were conducted as described in the same reference (15)
. Briefly, isolated VSMCs adhered to the bottom of a glass coverslip and were permeabilized with saponin (0.005% for 60–90 s) in solution containing (in mM): 110 Na-aspartate, 30 NaCl, 6 KCl, 0.5 EGTA, 8% dextran, 20 HEPES, and 0.05 Fluo-3 pentapotassium salt, pH 7.2. After permeabilization, cells were equilibrated for 1 min in internal solution (in mM: 110 K-aspartate, 10 NaCl, 30 KCl, 3 MgATP (free [Mg2+] adjusted to 1 mM), 10 phosphocreatine, 0.05 Fluo-3 pentapotassium salt, 5 U/ml creatine kinase, 8% dextran, pH 7.2), and EGTA (0.5 mM). CaCl2 was added to obtain a free [Ca2+] of 100 nM.
Spontaneous Ca2+ release events (Ca2+ sparks) were recorded at room temperature in internal solutions in the presence or absence of sorcin (2 µM) with a laser scanning confocal microscope (LSM510 META, Zeiss, New York, NY, USA) equipped with a X63 oil-immersion objective (N.A. 1.4) in the line scan mode. Fluo-3 was excited at 488 nm with an Argon laser, with emitted fluorescence measured at >510 nm. Ca2+ sparks were measured with a custom-made program running in IDL 5.5 software (Research Systems Inc., Boulder, CO, USA). Images were normalized by dividing the fluorescence intensity of each pixel (F) by the average resting fluorescence intensity (F0) of a confocal image to generate an F/F0 image.
Planar bilayer technique
Cardiac RyRs were reconstituted in planar lipid bilayers as described previously (3
, 4)
. Bilayers were composed of phosphatidylethanolamine and phosphatidylserine (1:1) dissolved in decane at 30 mg/ml. Cardiac SR microsomes (100–200 µg) were added to a 900 µl chamber (cis side), held at virtual ground, which corresponded to the cytosolic face of the channel. The recording solution in the cis chamber was 300 mM cesium methanesulfonate and 10 mM MOPS (pH 7.2). The trans solution was the same except that cesium methanesulfonate was 50 mM before fusion and 300 mM after fusion.
Channel activity in the presence or absence of sorcin was recorded at a sampling rate of 5 kHz using a 16-bit acquisition and storage system. Records were filtered with an 8-pole low pass Bessel filter at 1.5 kHz and digitized at 4 kHz using a Digidata 1200 AD/DA interface. Data acquisition and analysis were carried out with Axon Instruments hardware and software (pClamp 8.0).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(A and C monomers) and β (B and D monomers).
Monomer fold of F112L sorcin: the and β conformers
In the F112L monomer the Gly-rich, highly flexible, and disordered N-terminal domain (residues 1–30) is not visible, whereas the five EF-hands comprised in the C-terminal domain are clearly defined. The same features have been described for wt-sorcin (24)
and for most PEF proteins (25
26
27
28)
.
The two monomers forming the AB and CD asymmetric dimers assume two different conformations, and β, depicted in pale yellow and ruby, respectively, in Fig. 1B
. The structural differences between the and β conformers can be described by dividing the monomer into two subdomains connected by the long D helix: the first subdomain is formed by the EF1 and EF2 hands and the second by EF3, EF4, and EF5. The subdomains from each monomer superimpose quite well, but the relative orientations of each subdomain differ in the individual monomers. The major difference concerns EF3, which is enlarged in the β conformer due to a shift in the C-terminal part of the D helix. In contrast to F112L-sorcin, in the wt protein the largest difference between the two conformers concerns the EF2 loop. This is enlarged in the β conformer with respect to the one due to a shift in the N-terminal part of the D helix (data not shown).
Comparison of the conformers in human F112L, human wt, and hamster sorcin
A comparison between the wt and F112L monomers shows that the mutation causes a large conformational change in both the and β conformers. Unexpectedly, the spatial relationships between the two subdomains are altered dramatically with respect to wt-sorcin (24)
. Therefore, for the sake of comparison, the wt and F112L monomers have been superimposed by fixing EF4 and EF5, which are unaltered (Fig. 2
).
|
The most dramatic change in the F112L crystal structure with respect to the wt protein concerns the EF1 motif, which is rotated 180° around Tyr67, located in the loop between the B and C helices. As a result, a tetrameric assembly is stabilized that is characterized by an exchange of the EF1 hands between the A, C and B, D monomers. The changes in the conformer are depicted in Fig. 2A
. In F112L sorcin (pale yellow), the D helix is shifted significantly with respect to wt sorcin (blue). Moreover, the hydrogen bonds established in wt-sorcin by the Phe112 main chain carbonyl oxygen with the Glu124 carboxylate (Phe112 O-Glu124 OE1=3.62 Å, Phe112 O-Glu124 OE2=3.82 Å) are disrupted. Further, the absence of the bulky phenylalanine side chain allows a 90° rotation of the His108 side chain, which forms a hydrogen bond with Tyr159 (Tyr159 OH-His108 ND1=3.79 Å). The structural changes are even more evident in the β conformers (Fig. 2B
) where both EF2 and EF3 display a rotation of 30° around Gly104 placed in the middle of the D helix. This rotation leads to an enlargement of the EF3 loop with respect to wt-sorcin, as mentioned above. The chemical surrounding of the 112 residue changes as described above for the conformer.
Collectively, the structural changes that characterize F112L-sorcin alter the molecular surface in the vicinity of the mutation as indicated by an in vacuo calculation of the electrostatic potential. For example, in the β conformer the region surrounding the mutated residue appears less hydrophobic than in wt-sorcin and is characterized by two large negatively charged patches created by the displacement of the D helix and EF1 region (Fig. 2C, D
).
The structures of the F112L monomers were compared also with those of the hamster sorcin calcium binding domain (SCBD). As for the F112L variant and wt-sorcin comparison, the monomers were superimposed by fixing the EF4 and EF5 motifs, which are unchanged. The major difference between the monomers in both conformations concerns again the EF1 hand, which in F112L-sorcin is rotated with respect to hamster SCBD 180° around Tyr67 (Fig. 3
A, B). Other structural changes are apparent in the position of the D helix and of the EF3 loop, which is enlarged in F112L relative to hamster sorcin. The displacement of the D helix is more evident in the β conformers. Thus, in F112L-sorcin the helix is rotated 30° around Gly91 (placed at the beginning of the D helix) with respect to hamster sorcin.
|
Dimeric and tetrameric interfaces
The dimeric interface of the F112L variant is practically identical to that of wt-sorcin and is likewise characterized by an accessible surface area of 1909.74 Å2/monomer, which corresponds to 15.73% of the total solvent accessible surface area. The dimeric interaction surface appears to be smaller than in hamster SCBD where it amounts to 2820 Å2, i.e., to 20% of the entire monomer surface area.
F112L-sorcin dimerizes by pairing of the EF5 hands of two adjacent monomers through the G and H helices, which give rise to a four-helix bundle, while the loops between them form an antiparallel β-sheet. A hydrophilic interaction occurs between NH1 of Arg147 and carboxylate oxygen OT2 of Val198 that are placed at a distance of 2.7 Å. The dimer interface is stabilized also by hydrophobic interactions, which include those established by three highly conserved phenylalanine residues of both subunits, Phe173, Phe186, and Phe191 (Fig. 4
A).
|
The tetrameric interface in F112L-sorcin is quite extended (2298.59 Å2) and represents 19% of the entire protein accessible surface area. It is formed by residues placed on EF1 (35–59), by residues 62–68 on the loop connecting EF1 and EF2 and forming a short antiparallel β-sheet, and by residues lining the long D helix (Fig. 4B
). The interactions stabilizing the interface are both hydrophobic and hydrophilic. The most buried hydrophobic residues at the interface are Leu35 and Phe39 on the A helix, Tyr67 (placed on the loop connecting the EF1 and EF2 hands) and Phe92, Phe95, Trp99 placed on the D helix. The most buried residues involved in hydrophilic interactions are Gly62, Gly65, Gly66, Lys68, forming the intersubunit β-sheet, Arg106, and Arg116. In wt-sorcin there is no tetrameric interface, most likely due to the different disposition of the EF1 region.
In the hamster SCBD crystals a tetrameric interface is formed, albeit with features other than those described for F112L-sorcin (9)
. It consists of the very first amino acid residues of the domain (34–38), which form an additional helical turn that crosses the interface. Moreover, the side chains of Trp99 and Tyr38 at the interface of the interacting monomers B and D give rise to stacking interactions. These interactions are too weak to allow formation of stable hamster sorcin tetramers in solution.
Sulfate binding sites
Sulfate ions are very often bound to phosphorylation sites due to the similarity with phosphate ions in terms of size and charge distribution (29)
. The hamster sorcin sequence contains two putative phosphorylation sites by cAMP-dependent protein kinase (PKA), namely RYS (residues 147–149), located between the F helix and the EF4 loop, and RRRDS (RRRDT in human sorcin; residues 174–178) comprising the last three residues of the G helix and the first two of the EF5 loop (30)
. In the hamster SCBD crystals, sulfate is close to Ser-172, Arg175, and Arg176 (Fig. 5
A); accordingly, only the RRRDS site is susceptible to phosphorylation by PKA (9)
. No corresponding structural data are available for wt-sorcin that was crystallized without ammonium sulfate (24)
.
|
The electron density map of F112L-sorcin reveals the presence of three strong peaks per tetramer located close to the EF5 hand, in proximity of the canonical phosphorylation site RRRDT. These peaks were assigned to sulfate anions, since the average peak intensity is five times stronger than that of water and the crystals were grown from concentrated (NH4)2SO4 solutions.
As shown in Fig. 5A
, a sulfate anion occupies the position identified in hamster sorcin and is coordinated as in the latter protein by two positively charged residues, Arg175 (NH2-Arg175 O4=3.98 Å) and Arg176 (O1-Arg176 NH1=2.65 Å, O2-Arg176 NH2=3.01 Å). Arg84, which is at hydrogen bonding distance from the anion (3.0 Å) in hamster SCBD, cannot coordinate sulfate in F112L-sorcin because the side chain is shifted toward solvent. As for hamster sorcin, each anion is placed at 3.08 Å from the side chain oxygen of Ser-172; moreover, in F112L-sorcin, it is at hydrogen bonding distance (3.7–4.0 Å) also from the sulfur atom of Cys194.
The F112L-sorcin electron density map also shows four additional strong peaks per tetramer at the tetramer interface, indicating the presence of another sulfate binding site. Thus, two sulfate anions are placed between monomers A and C and two between monomers B and D in the pocket formed by the C-terminal part of the D helix and EF4 of one monomer and by the β-sheet (64–68) and EF3 of the symmetry related subunit. The ligand is stabilized by a hydrogen bond network that involves two water molecules (indicated as A and B in Fig. 5B
), Tyr67, and Ala64 of one subunit and Arg106 of the symmetry related one.
Analytical ultracentrifugation
The state of association of F112L-sorcin was assessed in sedimentation velocity and equilibrium experiments at pH values near neutrality and at pH 6.0, which corresponds to the crystallization pH. F112L-sorcin sediments as a single homogeneous peak of sedimentation coefficient, s20,w, 2.95 S in 0.1M Tris-HCl at pH 7.5 and in 50 mM BisTris at pH 6.0. In contrast, in 50 mM phosphate buffer at pH 6.0 the sedimentation coefficient drops to 2.45 S (Fig. 6
A). Sedimentation equilibrium experiments carried out under the latter experimental condition yielded Mr values of 41,400 ± 300 Da for wt-sorcin and 40,500 ± 200 Da for the F112L variant, which correspond closely to the expected value for a sorcin dimer, 43.1 kDa. Similar Mr values were obtained at pH 7.5 (Fig. 6B
).
|
Circular dichroism spectra
Analysis of F112L-sorcin in the UV region shows that the mutation does not produce significant spectral differences relative to the wt protein in Tris-HCl and phosphate buffers over the pH range 6.0 to 7.5 (data not shown).
Calcium affinity measurements
The Ca2+ affinities of wt- and F112L-sorcin were estimated in indirect fluorescence titration experiments carried out in parallel in the presence of Fluo3 as a Ca2+ chelator. The results obtained in 100 mM Tris-HCl at pH 7.5 and 25°C are presented in terms of degree of Fluo3 saturation as a function of total Ca2+ concentration (Fig. 6C
). The data can be fitted with one Ca2+ binding constant in the micromolar range, namely K = (0.7±0.2)x10–6 M for wt-sorcin, and K = (4.0±1.2)x10–6 M for F112L-sorcin, using a dissociation constant for the chelator of 325 nM. The F112L mutation therefore results in a 5- to 6-fold decrease in the affinity for Ca2+.
Interaction with the annexin VII N-terminal peptide
SPR was used to characterize the interaction between wt- and F112L-sorcin with the immobilized annexin VII N-terminal peptide. Sorcin and annexin VII interact with each other through their respective N-terminal regions. When the interaction is assessed at a constant protein concentration, the Req values increase markedly with increase in Ca2+ concentration (Fig. 7
A-C). The F112L mutant gives rise to significantly lower sensorgrams with respect to wt-sorcin at all the Ca2+ concentrations analyzed. At 60 µM Ca2+, for example, the Req values are 388 for wt-sorcin and 26 for F112L-sorcin. A further series of experiments was performed as a function of protein concentration at 20 µM Ca2+ (Fig. 7D-F
) to assess the apparent dissociation constant, KD, by Scatchard analysis of the ratio Req/C vs. Req (31)
. The analysis yields apparent KD values of 2.5 ± 0.3 µM for wt-sorcin, and 27 ± 9 µM for F112L-sorcin at 20 µM Ca2+. In this calculation, the intercept on the abscissa was considered to be the same for both proteins since the low Req values pertaining to F112L-sorcin do not warrant a precise Rmax determination. By assuming that the sorcin dimer contains a single effective annexin VII binding site (32)
, the analysis yields KD values of 1.25 µM for wt-sorcin and 13.5 µM for F112L-sorcin. The decrease in affinity for annexin VII due to the F112L mutation therefore is 10-fold.
|
Activity of cardiac ryanodine receptors reconstituted in planar lipid bilayers
Previous experiments have determined that sorcin binds cardiac ryanodine receptors (RyR2) with high affinity through its Ca2+ binding domain and may completely inhibit channel activity (3
, 4)
. To determine whether the F112L mutation modifies these properties, the blocking effect of wt- and F112L-sorcin on RyR2 reconstituted into planar lipid bilayers was compared (Fig. 8
). Under control conditions and at activating 10 µM free Ca2+ (pCa 5) the probability of the channel being open (Po) was 0.091 ± 0.012 and decreased to 0.010 ± 0.002, i.e., 9-fold, in the presence of 2 µM wt-sorcin, as described before (3
, 4)
. By contrast, a higher concentration of F112L-sorcin (3 µM) failed to decrease significantly Po of RyR2 (0.083±0.014 and 0.081±0.015 in the absence and the presence of F112L-sorcin, respectively). Thus, the F112L mutation dramatically impairs the ability of sorcin to interact with and/or block the activity of RyR2.
|
Ca2+ sparks in saponin-permeabilized vascular myocytes
Ca2+ sparks in resting cells represent the spontaneous and coordinated opening of a cluster of RyRs (33)
. Ca2+ sparks measurements were, therefore, used as markers of RyR activity to assess sorcin modulation of RyRs in situ (Fig. 9
). Rat VSMCs were permeabilized with saponin as in Rueda et al. (15)
, and Ca2+ sparks were recorded with a laser scanning confocal microscope in an internal solution containing 60 nM Ca2+ and 30 µM Fluo-4 salt (Fig. 9A
).
|
Perfusion of 2 µM sorcin onto cells resulted in a clear reduction of Ca2+ spark amplitude, half-duration, and half-width (Fig. 9B
) compared with control. In contrast, perfusion of 2 µM F112L-sorcin onto cells resulted in no reduction of Ca2+ spark frequency, amplitude, duration, and spatial spread compared with control in accordance with data in Rueda et al. (15)
. The effect on Ca2+ spark parameters shows that sorcin imposes impressive attenuation of RyR activity also in vascular myocytes, while the F112L mutant has practically no effect. These results, therefore, together with the bilayer experiments, suggest that the structural changes brought about by the F112L mutation result in functional alterations that impair the capacity of sorcin to interact with RyRs in both cardiac and vascular tissues.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The present characterization of the F112L structure, taken together with the marked decrease in the affinity for Ca2+ produced by the mutation, lays the molecular basis to explain the impaired function of the human variant.
A model for the Ca2+-dependent activation of sorcin was proposed by Ilari et al. (9)
based on the crystal structure of hamster SCBD and was validated by the characterization of site specific mutants of the EF hands (10)
. Binding of Ca2+ to the highest affinity site, EF3, triggers the activation process, which requires that this information be transferred to the EF2-hand. The long and rigid D helix that connects EF3 to EF2 plays a major role in this process: the reorganization of the interaction network around the D helix renders it an efficient amplifier of the initial conformational change. Further information transfer to EF1 and to the rest of the protein takes advantage of the canonical pairing of EF2 and EF1 (10
, 11)
. These events result in the exposure of hydrophobic surface(s) that permit interaction of Ca2+-bound sorcin with target proteins.
In F112L-sorcin, the mutation involves a residue located at the end of the D helix, close to the EF3 hand and next to Asp113, one of the acidic residues in EF3 that bind calcium with high affinity. The structural effects of the F112 substitution involve a large part of the molecule (Figs. 1
2
3)
and lead to a decrease in calcium affinity (Fig. 6)
that represents the major cause for the impaired interaction with target proteins (Figs. 7
8
9)
and hence for the impaired function of the mutant.
The occurrence of marked structural changes at the EF3 site and at the D helix was anticipated given the position of the F112 residue in the protein sequence. The amino acid substitution disrupts an important component of the hydrogen bonding network around the D helix, the interaction between the Phe112 carbonylic oxygen and the Glu124 carboxylate, and the bidentate Ca2+ ligand in the EF3 hand, thereby impairing the EF3 triggering function. In addition, the EF3-EF4-D helix region is altered by the shift of the C-terminal part of the D helix in the β conformer and by the 30° rotation of EF2 and EF3 around Gly104. This region is believed to contact the N-terminal part of the protein in the Ca2+-free form and to interact with protein targets in Ca2+-bound sorcin (9
, 28)
. In contrast to the structural alterations just described, the 180° tilt of EF1 around the Tyr67 residue was fully unexpected. It is indicative of a high flexibility of the EF1 region that manifests itself in the formation of hybrid tetramers in the crystal (Fig. 4)
, but not in solution (Fig. 6B
), where this part of the molecule is not constrained by the lattice forces.
The effect of phosphate on the conformation of both wt- and F112L-sorcin is likewise intriguing since phosphorylation by PKA alters Ca2+ sensitivity of sorcin, favors its translocation to the sarcoplasmic reticulum (SR) and diminishes the capacity to inhibit ryanodine binding to RyR2 (3
, 7)
. Morever, the phosphorylation levels of sorcin and other proteins involved in regulating the cardiac e-c cycle are altered under pathological conditions. Thus, in the failing heart, sorcin is hyperphosphorylated and translocation to the SR membrane is increased, possibly resulting in preservation of the SR Ca2+ content and in improvement of cardiac relaxation (7)
.
Sorcin phosphorylation takes place at the RRRDS(T) site (residues 174–178), which includes the last three residues of the G helix and the first two of the EF5 loop. The phosphorylation site is occupied by sulfate ions in the crystal structures of F112L-sorcin and of hamster SCBD grown in (NH4)2SO4 (Fig. 5
and ref. 9
). It is not known whether the other sulfate binding site located at the tetramer interface of the F112L crystal is of possible physiological relevance. In solution, phosphate induces a conformational change in both wt- and F112L-sorcin at pH 6.0 without affecting the state of association. Thus, the sedimentation coefficient, s20,w, decreases from 2.95 S at pH 7.5, to 2.45 S at pH 6.0, although sedimentation equilibrium experiments prove that the proteins are dimeric at both pH values. The observed change points to a departure from spherical shape at pH 6.0, although the nature of the conformational change remains unclear. It is conceivable that phosphate bound to the RRRDT site changes the microenvironment around the three consecutive arginine residues, thereby loosening the interactions that in Ca2+-free sorcin are believed to stabilize the N-terminal domain in the cleft formed by EF4 and EF5 (9)
.
In functional terms, the structural changes that take place in F112L-sorcin are reflected in a 6-fold decrease in calcium affinity (Fig. 6C
) and in a diminished capacity to interact with target proteins as exemplified by the 10-fold decrease in affinity for the annexin VII N-terminal peptide (Fig. 7)
, by the impaired translocation capacity of F112L-sorcin to the membrane fraction in transgenic mice (16)
and by the failure to affect the open probability of RyR2, either in artificial reconstitution systems (Fig. 8)
or in situ (Fig. 9)
. The impairment in the capacity to interact with annexin VII and RyR2 may differ in extent due to the likely contribution of specific topological factors, such as the flexibility of the EF1 region and the decreased hydrophobicity in the vicinity of the mutation apparent from the electrostatic potential surface (Fig. 2C, D
). It is worth recalling that substitution of Glu124 (which establishes hydrogen bonding interactions with Phe112) with Ala likewise produces a significant impairment in the ability of sorcin to interact with annexin VII and RyR in vitro (10
, 11)
.
In excitable tissues, sorcin interfaces Ca2+ signals with several effector proteins of e-c coupling and determines negative regulation of SR calcium release. In particular, sorcin increases NCX activity (6
, 16
, 34)
, increases the rate of L-type Ca2+ channel inactivation (5)
and increases Ca2+ affinity of SERCA2a (7)
.
The interaction of sorcin with RyR2 is one of the best studied so far. Several reports have shown that sorcin directly inhibits the channel activity, which in cellular settings would result in decreased localized Ca2+ release events (Ca2+ sparks) and reduced global Ca2+ transients. In kinetic terms, sorcin reduces Ca2+ flow from RyR2 by decreasing the mean open time and the frequency of open events (3
, 4)
. These effects may explain why sorcin decreases the duration, width, and amplitude of Ca2+ sparks (Fig. 8
and refs. 4
,6
,11
). F112L-sorcin fails to modify Ca2+ sparks (Fig. 8)
, mainly as a consequence of its decreased affinity for Ca2+ with respect to wt-sorcin (Fig. 6C
), although a direct effect on the surface interacting with RyR2 cannot be disregarded. A similar situation applies to the effect of wt- and F112L-sorcin in vascular myocytes. Whereas wt-sorcin decreases Ca2+ spark frequency, width, and duration, as expected since Ca2+ sparks in VSMCs arise mainly from RyR2 (34)
, F112L-sorcin fails to do so (Fig. 9)
. These differences are reflected also in the directionally opposite effects of wt- and F112L-sorcin overexpression in transgenic mice (16)
. Ca2+ sparks in mice overexpressing F112L-sorcin were more frequent, wider, and of longer duration than those in wt mice or transgenic mice overexpressing wt-sorcin. These differences in fundamental aspects of Ca2+ signaling in cardiac cells undoubtedly contribute to the altered phenotype of F112L-sorcin overexpressing hearts (16)
.
In summary, the present work provides a structural framework that helps explain the functional alterations brought about by the F112L mutation. The structural effects obviously concern the EF3 region, thus determining a marked decrease in Ca2+ affinity but involve the D-helix, the EF2, and the EF1 hand as well, and cause a decreased interaction of the sorcin variant with its targets, resulting in marked e-c coupling alteration in the heart and other tissues.
|
ACKNOWLEDGMENTS |
|---|
Received for publication May 22, 2007. Accepted for publication June 27, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and 7.5 (
). C) Fluorescence titration with Ca2+ in the presence of Fluo3. Wt- (
) and F112L-sorcin (•) (20 µM) were titrated with Ca2+ in the presence of Fluo3 (15 µM) in 100 mM Tris-HCl at pH 7.5 and 25°C. Control titration of Fluo3 alone with Ca2+ (
