| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Division of Pediatric Cardiology and the Leon H. Charney Division of Cardiology, New York University School of Medicine, New York, New York, USA
2Correspondence: Leon H. Charney Division of Cardiology, New York University School of Medicine, 550 First Ave., OBV A-615, New York, NY 10016, USA. E-mail: glenn.fishman{at}med.nyu.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Key Words: ryanodine receptor • transgenic • mouse model
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
2
3
4
5
6)
. Sorcin, an EC coupling regulatory protein (7
8
9
10)
, participates dynamically in several aspects of CICR and is thought to modulate numerous processes including LTCC inactivation (7
, 11)
, SR Ca2+-ATPase function (12)
, and the open probability of cardiac RyR2s (8
, 9
, 13
14
15)
.
Human genomic studies have identified a nonsynonymous single nucleotide polymorphism (SNP) in the sorcin gene that results in substitution of the phenylalanine at residue 112 with leucine (sorcinF112L). Crystallographic analysis of wild-type (WT) sorcin suggests that F112 is within a D helix and participates in transmission of Ca2+-mediated conformational changes. Consequently, replacement of this residue with leucine may impinge on the dynamic, Ca2+-dependent translocation of sorcin from soluble to cellular membrane sites (14
, 16
, 17)
. Indeed, preliminary in vitro studies in planar lipid bilayers suggest that, unlike WT sorcin, the L112 variant does not diminish RyR2 open probability (18)
. Intriguingly, a preliminary report suggests that in two families the F112L sequence variant segregates with hypertension and hypertrophic heart disease (14)
.
The goals of the present study were 2-fold: first, we sought to gain further mechanistic insight into the normal function of sorcin in the heart, and second, we wanted to elucidate the relationship between the sorcin F112L polymorphism and its effects on cardiovascular physiology. We hypothesized that L112-sorcin, unlike the WT protein, might be defective in its hastening of LTCC inactivation as well as its inhibitory effect on RyR2 gating. The results presented, obtained in transgenic mice overexpressing L112-sorcin in the heart, support this hypothesis and are consistent with the idea that sorcin is an important regulator of Ca2+ cycling in cardiac myocytes.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
MHC) -sorcinWT construct (7
, 19)
and generated TG mice according to standard techniques in C57BL/6 x C3H (B6C3) F1 hybrids. TG lines were established by breeding founders with C57BL/6 mice and offspring were genotyped by polymerase chain reaction (PCR) assay. Both the MHC-sorcinF112L and MHC-sorcinWT TG lines were maintained in the C57BL/6 background, and nontransgenic (NTG) littermates or age-matched C57BL/6 were used as controls. Experiments were performed using mice that were 4–6 months of age. All experiments were performed in accordance with the regulations of our Institutional Animal Care and Use Committees.
Immunoblotting
Levels of proteins in homogenates of hearts were determined by immunoblot analysis, as described previously (11
, 20)
. NH2-terminal sorcin peptide antibodies were obtained from Zymed Laboratories (San Francisco, CA, USA). Antibodies to the cardiac ryanodine receptor (RyR2), pore-forming subunit of the LTCC, sarcoplasmic reticulum ATPase (SERCA2A), and phospholamban (PLB) were obtained from Affinity Bioreagents (Golden, CO, USA) and those for the Na+-Ca2+ exchanger (NCX) were obtained from Swant (Bellinzona, Switzerland) (21)
.
Echocardiography and hemodynamics
Transthoracic echocardiography was performed in standard fashion as described previously (22)
. Standard functional analyses included left ventricular wall thickness, fractional shortening, aortic outflow peak velocity, mitral E/A ratio, velocity of circumferential fiber shortening, and myocardial performance index (23
24
25)
. All measurements were performed in triplicate and averaged, and performed without knowledge of genotype. Systemic blood pressure was measured in Avertin (2,2,2-tribromoethanol; 20 µg/g i.p.) -anesthetized, ventilated mice as described previously (7)
.
Ca2+-mediated translocation
Ca2+-mediated translocation of sorcin was examined in heart lysates prepared from sorcinWT and sorcinF112L TG mice. Cardiac ventricles were lysed in 20 mmol/L Tris (pH 7.4), 0.15 mol/L NaCl, 5 mmol/L EGTA, and protease inhibitors by Dounce homogenization. After centrifugation at 8000 g for 2 min, 1 mol/L CaCl2 was added to a final concentration of 1 mmol/L to half of each supernatant sample. All crude supernatants were centrifuged at 100,000 g for 30 min, and Western blot analysis of sorcin using the Zymed antibody (Ab) was carried out on high-speed supernatants and microsomal pellets.
Isolation of ventricular cardiac myocytes
Left ventricular myocytes were isolated from adult C57BL/6 mice (age, 4–6 months; wt 25–35 g) using a standard enzymatic technique as described previously (26)
. The myocytes were plated in 50 mm nonstick Valmark Petri dishes (Midwest Scientific, St. Louis, MO, USA) and subjected to stepwise Ca2+ reintroduction. Myocytes were stored at room temperature (20–22°C) in perfusion buffer containing 1 mmol/L Ca2+ and used within 8 h postdigestion.
Fractional shortening
Cardiac myocytes were perfused with standard Tyrode’s saline solution containing (in mmol/L) NaCl 137, KCl 5.4, HEPES 10, MgCl2 1, NaH2PO4 0.33, CaCl2 1.8 and glucose 10 (pH adjusted to 7.4 with NaOH). Quiescent myocytes were field-stimulated with platinum wire electrodes using a stimulator (Grass Instruments, West Warwick, RI, USA) at frequencies between 0.1 and 3 Hz. Cell edges were detected with a video-edge detector (Crescent Electronics, Windsor, Ont., Canada); the differential output was displayed using pClamp 8.1 (Axon Instruments, Union City, CA, USA) after digitizing the signal (Digidata 1322A, Axon Instruments). Myocytes were field-stimulated for 10 cycles at each frequency to achieve steady state, then five additional cycles were recorded. Spontaneously contracting cells were not used. Cuboidal cells were selected for ease of edge detection. Length calibration was performed by detecting edges on a hemocytometer at lengths of 50, 100, 150 µm and measuring the corresponding output. A linear regression line was fit to the calibration points and the slope was used as the appropriate calibration factor. Resting length (Lmax) and peak length during shortening (Lmin) were converted to microns, and fractional shortening was calculated as (Lmax-Lmin)/Lmax x 100 at all pacing frequencies tested. Time-to-peak (TTP) (ms) was calculated from the initiation point of the pClamp recording, which coincided with the onset of the stimulus spike from the stimulator. A mono-exponential function was applied to the decay of shortening (90% Lmax to Lmin) to obtain a time constant (SigmaPlot; v.9.0; Systat Software, Inc., San Jose, CA, USA).
Measurement of Ca2+ transients with fluo-4 acetoxymethyl ester (AM)
Ca2+ transients were recorded using fluo-4 AM and confocal microscopy. Myocytes were incubated in perfusion buffer (as described above) containing 10 µmol/L fluo-4 AM (Molecular Probes, Invitrogen, Carlsbad, CA, USA) for 20 min at room temperature (20–22°C). Fluo-4 AM stock solution (1 mmol/L) was prepared with 50 µg fluo-4 AM in 45 µl DMSO and 5 µl pluronic acid (2.5 µM; Molecular Probes). After the myocytes were pelleted, the supernatant was replaced with standard Tyrode’s solution (see above). Myocytes were washed for 30 min to allow for sufficient deesterification. Myocytes were plated on laminin (1 mg/ml) -coated coverslips (#1.5), mounted in a perfusion chamber (Warner Instruments, Hamden, CT, USA), and field stimulated (1 Hz) as described above. Myocytes were perfused with standard Tyrode’s solution at 20–22°C containing 1.8 mmol/L Ca2+.
To assess the degree of SR Ca2+ loading, myocytes were conditioned with 1 Hz field stimulation for 10 s to load the SR with Ca2+. Caffeine (10 mmol/L) was rapidly applied (2 s) via a large bore patch pipette (<1M
; TW150; World Precision Instruments, Sarasota, FL, USA) using a pneumatic pressure injection system (Picospritzer II; General Valve Corp., Fairfield, NJ, USA) at 5 psi. Caffeine was in a solution with 0 mmol/L Na+ and 0 mmol/L Ca2+ to prevent Ca2+ efflux through the NCX (27)
. The solution contained (in mmol/L) LiCl 140, KCl 4, MgCl2 4, HEPES 5, lidocaine 0.3, 4-aminopyridine 10, and glucose 10 (pH adjusted to 7.4 with LiOH). Line scans were performed through medial sections of myocytes to limit movement artifacts during caffeine application. The fractional release of Ca2+ to the total SR Ca2+ load was calculated as the ratio of peak twitch F1/F0 to the peak F1/F0 of the caffeine transient. A linear regression line was fitted to the decay of the caffeine transient, during perfusion with caffeine and Na+/Ca2+ free solution, to assess 
0.5 of slow Ca2+ transporters.
INCX measurement
NCX currents (INCX) were measured essentially as described (28)
but under chloride-free conditions. Myocytes were patch-clamped in the whole-cell configuration at 20–22°C at a holding potential of –80 mV. Glass pipettes (1–2 M) were pulled from 1.5 mm (1.12 mm I.D.) borosilicate glass capillaries (TW150; World Precision Instruments) on a DMZ Universal Puller and fire-polished. Seal and cell rupture was achieved in Tyrode’s solution. Current was measured with a patch-clamp amplifier (Axoclamp 200B, Axon Instruments) and low-pass filtered (–3 dB at 1 kHz). The reference electrode consisted of a KCl (3 mol/L)/agar (4%) bridge connected to the amplifier via a Ag+/AgCl wire. The offset potential was corrected before touching the surface of the cell with the pipette tip. After 5 min dialysis, cell capacitance, access resistance, and membrane seal resistance were measured using a +5 mV step pulse (pClamp v.8.1, Axon Instruments) at a holding potential of –80 mV. Access resistances of 2–10 M were deemed adequate. Cell capacitance and series resistance were thus compensated (the latter to at least 70%). During perfusion with the modified bath solution, clamp potentials were offset to correct for a shift in junctional potential of 0.7 mV (pipette negative), which was calculated using Clampex software (v.8.1; Axon Instruments, CA, USA). The external solution was K+-free (to block the inward rectifier K+ current and the Na+-K+ pump) and Cl-free (to remove Cl– current contamination) and contained (in mmol/L) Na+-glutamate 137, MgSO4 1, CsOH 5.4, CaSO4 1, HEPES 10, nifedipine 0.01, Glc 10. Nifedipine stock solution (10 mmol/L) was made up in DMSO. The pipette solution contained (in mmol/L) CsOH 45, Cs-methanesulfonate 55, HEPES 20, ATP (Mg2+ salt) 10, GTP (Tris salt) 0.3, MgSO4 0.8, CaSO4 2.2, glutamate (Na+ salt) 14, BAPTA (tetracesium salt) 5, and 5,5'-dibromo BAPTA, AM (Di Br BAPTA) 5, buffered to pH 7.4 with CsOH. BAPTA and Di Br BAPTA were predissolved in 1N CsOH. With this internal solution, [Ca2+]i is buffered to 100 nmol/L (Maxchelator, ver.2.4). A descending voltage ramp protocol (every 15 s) was applied to elicit INCX. Membrane potential was changed to –40 mV to inactivate sodium current. The voltage was stepped up to +70 mV for 100 ms, then ramped down to –130 mV at the rate of 90 mV·s–1 to induce remaining currents. Once INCX had reached steady state, 5 mmol/L NiCl2 was superfused for 20 s. INCX was defined as the current component blocked by Ni2+ and current density was obtained by normalization to the cell capacitance. The INCX equilibrium potential (ENCX) was calculated as –59 mV (29)
.
Simultaneous recording of ICa and [Ca2+]i: "CICR gain"
The relation between Ca2+ influx and subsequent Ca2+ release from the SR (or CICR gain) was assessed using a protocol as described previously (30)
with slight modifications. Experiments were performed at 20–22°C. Whole-cell voltage clamp conditions were the same as described above for INCX measurements. However, the pipette solution contained (in mmol/L) CsCl 120, MgCl2 1.5, ATP (Mg2+ salt) 5, NaCl 10, TEA-Cl 10, HEPES 20, and fluo-4 (pentapotassium salt) 0.05, buffered to pH 7.2 with CsOH). The bath solution contained (in mmol/L) NaCl 137, CsCl 5.4, HEPES 20, MgCl2 1.2, NaH2PO4 1, CaCl2 2, and Glc 10, buffered to pH 7.4 with NaOH. After 10 min dialysis, the holding potential was set at –50 mV (to inactivate Na+ currents) and 200 ms voltage clamp steps were applied at 0.01 Hz (up to +50 mV in 10 mV increments) to record the L-type Ca2+ current (ICa,L). The ICa,L was defined as the current component blocked by nifedipine (10 µmol/L). Data points during the inactivation of ICa,L were subjected to curve fitting to a biexponential function (pClamp; v.8.1). The CICR gain parameter for intracellular Ca2+ release was calculated according to ref. 31
, where gain is the rate of Ca2+ release (F1/F0/TTP) divided by the integral of ICa,L (
ICa, L) during a 200 ms sweep. ICa, L was evaluated with pClamp (v.8.1) and normalized to cell capacitance.
Ca2+ signals recorded using confocal microscopy and fluo-4
Changes in [Ca2+]free were imaged and recorded by a laser scanning confocal microscope (Leica DM IRE2; Leica Microsystems Heidelberg GmbH) with a 63x oil immersion objective (numerical aperture=1.3). Fluo-4 was excited at the 488 nm line of an argon laser with emission collected through a 500 nm dichroic filter (Leica TCS SP2; Leica Microsystems GmbH, Wetzlar, Germany). Raster point size was 0.1 µm with an overall lateral resolution of 0.16 µm. For Ca2+ spark measurements, fluorescence images of xt line scan sections were recorded with 1024 pixels (400 Hz; 2.5 ms per line) in the time domain and with 512 pixels (0.155 µm/pixel) in the spatial domain. For spark measurements, five xt sections were taken along the longitudinal axis of the cell, with 5 µm spacing to limit photobleaching of the fluorophore. For Ca2+ transient experiments, 1500 pixels (3.75 s) were used in the time domain and line scans were taken along the transverse axis of the myocyte. CICR gain experiments used 512 pixels (1.25 s) in the time domain through the transverse axis of the cell.
Ca2+ sparks in permeabilized myocytes
We modified the cardiac myocyte permeabilization method adopted by Li et al. (32)
. Briefly, cardiac myocytes were permeabilized by incubating with streptolysin O (SLO), a cholesterol binding cytolysin. SLO was preactivated in 5 mmol/L dithiothreitol for 3 h at 4°C. During permeabilization, myocytes were stored in a solution containing (in mmol/L) EGTA 0.1, ATP 5, HEPES 10, K+-glutamate 150, MgCl2 0.25, glutathione (reduced form) 10, SLO (330 U/ml), and 2 µmol/L fluo-4 (pentapotassium salt; Molecular Probes, Invitrogen). The efficacy of SLO to permeabilize the cells was ascertained by imaging an aliquot of cells with confocal xyz scans and examining the increase in fluo-4 fluorescence with time. After 20 min, the myocytes were pelleted by gravity and the supernatant was replaced with a mock intracellular solution containing (in mmol/L, unless indicated) EGTA 10, Cs+-glutamate 200, HEPES 10, Mg-ATP 5, phosphocreatine di-Tris 5, 5 U/ml creatine phosphokinase, MgCl2 0.5, glutathione (reduced form) 10, 8% dextran (MW 40,000), and 50 µmol/L fluo-4 (pentapotassium salt); 20 nmol/L [Ca2+]free was used to elicit sparks. The volume of CaCl2 (100 mmol/L stock) that would achieve the desired free [Ca2+]i was calculated (Maxchelator, ver.2.4). Fluorescence was calibrated to [Ca2+]free by using a Ca2+ concentration series from 0 to 10 mmol/L Ca2+, buffered to 0 to 39 µmol/L [Ca2+]free with EGTA (Calcium Calibration Kit; Molecular Probes, Invitrogen). Fluo-4 (50 µmol/L) was used in each calibration solution with similar excitation and emission settings on the confocal microscope.
Image data analysis
Digital image analysis was performed using Leica Confocal Software (Version 2.5; Leica Microsystems GmbH) and ImageJ (NIH, Bethesda, MD, USA). Fluo-4 images were background subtracted and analyzed for peak (F1) and resting (F0) fluorescence. A mono-exponential function was applied to the decay of the Ca2+ transient to obtain a time constant (SigmaPlot; v.9.0; Systat Software). TTP of Ca2+ transients was calculated using a 1 ms light-emitting diode (LED) pulse applied to the visible light channel of the confocal microscope.
Sparks were automatically detected and spark parameters assessed from raw image data using a computer-based algorithm (IDL [ver.5.4], ITT Visual Information Solutions, Boulder, CO, USA), modified from Cheng et al. (33)
. Spark events were defined as areas where fluorescence was 3.8-fold the SD of background fluorescence (Fo). The software allowed incomplete spark events on the edge of the image to be manually removed from the final collated data. F1/F0, half-maximal width (W50) and half-maximal duration (D50) were calculated from spark events. The algorithm was modified to include TTP and frequency of spark events per line scan.
Computational modeling of cellular Ca2+ cycling
To integrate the changes in Ca2+ regulation observed in sorcinF112L mice, we performed computer simulations using a recently published model of electrical dynamics and calcium cycling in the mouse ventricular myocyte (34)
. The goal was to determine whether the altered LTCC inactivation and NCX activity in sorcinF112L mice were sufficient to explain the increases in Ca2+ transient amplitude and SR Ca2+ load that were observed. Because this model does not describe microscopic aspects of Ca2+ signaling, we did not attempt to simulate the measured changes in Ca2+ spark characteristics.
Bondarenko et al. (34)
presents separate sets of equations for apical and septal mouse ventricular myocytes. In the simulations presented (see Fig. 7
), the behavior of cells from NTG hearts was described by the equations for septal cells. Similar qualitative behavior was seen in separate simulations (not shown) using the equations for apical cells. To simulate effects of L112-sorcin overexpression, the model constant kNaCa was scaled by the factor 2.49 to increase NCX current. Slowed inactivation of ICa,L with no change in peak current was simulated by reducing the rates governing voltage-dependent and Ca2+-dependent inactivation (
and Kpcf, respectively) by 58% each and reducing peak LTCC conductance (GCaL) by 17%. In a separate set of simulations, the model constant 
3 was also scaled by 1.8 to simulate a potential effect of L112-sorcin on the activity of the SR Ca2+ ATPase (SERCA).
|
The objective of the computer simulations was to compare steady-state Ca2+ transient amplitudes (peak systolic minus diastolic [Ca2+]) between groups (i.e., NTG vs. sorcinF112L) over a range of pacing frequencies. A peculiarity of the model, however, made it difficult to compare these quantities at a true steady state. At relatively fast pacing rates (i.e., above 2 Hz), SR Ca2+ load and Ca2+ transient amplitude continued to increase slowly, even after 300 s of pacing, due to slow changes in intracellular [Na+] and [K+]. To overcome this problem but still account for changes in [Na+]i and [K+]i at different pacing rates, the following protocol was followed. The NTG model was paced at a given frequency for 120 s, after which the full set of state variables, including [Na+]i and [K+]i was saved. These values were used as initial conditions in a modified version of the model in which [Na+]i and [K+]i were fixed. The steady-state Ca2+ transient amplitudes obtained with this simplified model after 60–180 s of pacing (depending on rate) were used to generate the plots shown in Fig. 7
.
Statistical analysis
Data are expressed as mean ± SEM. An unpaired t test and/or single-factor ANOVA was used to compare current and fluorescence data between mouse groups. The nonparametric Mann-Whitney test appraised differences between means of spark parameters (StatistiXL). A P value of <0.05 was considered statistically significant in all cases.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
MHC promoter (19)
. Four independent founders were identified, and one line with levels of L112-sorcin overexpression (Fig. 1
A) comparable to those we previously described in mice overexpressing WT sorcin (
20-fold) was selected and expanded for more detailed characterization (7)
. SorcinF112L mice developed normally and showed no outward signs of morbidity or early mortality. Heart/body wt ratios in sorcinF112L TG mice (7.3±1.1 mg/g) were not significantly different from NTG controls (7.15±0.93; P=NS). Histological (Fig. 1B
) and echocardiographic analyses (Table 1
) revealed the gradual development of mild left ventricular chamber dilatation without evidence of hypertrophy or contractile dysfunction. Measurements of systemic systolic blood pressure in anesthetized, ventilated mice revealed no significant differences between sorcinF112L TG mice and NTG controls (72.8±7.1 mm Hg, n=6 vs. 78.5±12.7 mm Hg, n=6).
|
|
Biochemical properties of L112-sorcin
Sorcin undergoes Ca2+-mediated translocation from soluble to membranous fractions, which likely plays a role in its regulation of the L-type Ca2+ channel (LTCC) and cardiac ryanodine receptor (RyR2) function (17)
. To determine whether the F112L variant influenced the biochemical properties of sorcin, we prepared soluble and particulate fractions from the ventricles of sorcinWT and sorcinF112L TG mice in the presence and absence of Ca2+. The integrity of the fractions was confirmed by immunoblotting with antibodies against ß-tubulin, found exclusively in the Ca+2-soluble fraction, and Na/K ATPase, found only in the Ca2+-insoluble fraction. Whereas in the presence of Ca2+, WT sorcin was recovered primarily in the membranous fraction (80±11%, n=5), significantly less of the L112 variant (46±8%, n=5) translocated into this fraction (Fig. 1C
), suggesting decreased association with membrane-bound binding partners such as LTCC and RyR2.
Expression of calcium regulatory proteins
To examine whether overexpression of L112-sorcin in the heart influenced the abundance of other cardiac Ca2+-regulatory proteins, we performed a series of Western blots (Fig. 1D
). Compared with NTG hearts, there was no significant difference in expression of Cav1.2, the pore-forming subunit of the LTCC, RyR2, the SR Ca2+-ATPase or PLB in sorcinF112L TG hearts. There was, however, a statistically significant 2-fold increase in the abundance of NCX in the hearts of sorcinF112L TG mice, as summarized in Table 2
.
|
Contractions, Ca2+ transients and SR Ca2+ content
We next determined the physiological consequences resulting from expression of L112-sorcin in the heart, as assayed in dissociated adult cardiomyocytes. Myocytes from sorcinF112L TG mice were significantly shorter in maximum length (113.7±18.3 µm (n=15) vs. 130.7±25.5 µm (n=12), P<0.05) and maximum width (32.0±10.5 µm (n=15) vs. 39.7±10.7 µm (n=22), P<0.05) compared with those from NTG controls. SorcinF112L TG myocytes exhibited a 83% increase in mean peak fractional shortening at 1 Hz and an increase in shortening at all pacing frequencies above 0.1 Hz (P<0.05; n=9–15) (Fig. 2
A, B). The magnitude and TTP of shortening decreased with pacing frequency for both mouse groups (P<0.05) (Fig. 2C
), and the sorcinF112L myocytes had a shorter TTP at all frequencies tested (n=9–15) (Fig. 2A, C
). Expression of L112-sorcin in the heart also influenced lusitropic properties, with a significant decrease in the time constant of decay in mutant myocytes (60.0±7.2 ms (n=15) vs. 81.7±7.0 ms (n=12); P<0.05) compared with NTG controls.
|
Consistent with the increases observed in twitch amplitude, peak Ca2+ transient amplitude (F1/F0) was also elevated in sorcinF112L myocytes relative to those from NTG mice (Fig. 3
A, C; P<0.05). As with contractions, the Ca2+ transient TTP was also shortened in the sorcinF112L myocytes (44.4±1.6 ms [n=24] vs. 50.1±2.0 ms [NTG; n=36]; P<0.05). Consistent with the change in lusitropy, time to half-maximal decay of Ca2+ transients was significantly shorter in the sorcinF112L cells (126.8±3.2 ms (n=23) vs. 141±5.4 ms (n=31); P<0.05) compared with NTG controls. decay tended to be smaller in sorcinF112L cells (165.0±7.0 ms; n=25) than in NTG cells (174.0±7.0 ms; n=31), but this difference did not reach statistical significance.
|
We next investigated whether the increased Ca2+ transients and shortening were attributable to an increase in SR Ca2+ load. We assessed SR Ca2+ content with a rapid application of caffeine in conditions where the NCX was blocked. Caffeine-induced Ca2+ transients were also increased in the sorcinF112L myocytes (Fig. 3B, D
; P<0.05) whereas the fractional SR Ca2+ release during each contraction was not different in the sorcinF112L myocytes compared with the NTG control (Fig. 3E
). The time constant (0.5) of decline in caffeine transient was reduced by 46% in sorcinF112L myocytes (6.4±2.6 s vs. 3.4±2.9 s; NTG (n=22) vs. sorcinF112L (n=17); P<0.05), suggesting acceleration of slow Ca2+ transporters (Fig. 3B
).
Na+-Ca2+ exchange activity
We also assessed the activity of the NCX. INCX was measured under Cl–-free conditions as the current component blocked by 5 mmol/L Ni2+. Representative current traces obtained from a NTG myocyte are shown in Fig. 4
A. The Ni2+-sensitive current was larger in the sorcinF112L myocytes (Fig. 4B
[e.g., at +70 mV, 0.40±0.09 pA/pF (n=7) vs. 0.18±0.07 pA/pF (n=7) in NTG; P<0.05]. Reversal potentials for INCX were the same in both groups of mice (
–50 mV), close to the calculated equilibrium potential (ENCX = –59 mV).
|
ICa,L vs. [Ca2+]i or the CICR gain function
To further investigate the cellular basis for the observed increase in twitch amplitude and Ca2+ transient in the sorcinF112L myocytes, we analyzed the Ca2+ release gain function by simultaneously recording Ca2+ transients and ICa,L over a range of voltages (–50 to+50 mV). Consistent with direct measurements of cell dimensions, cell capacitance (Cm) was significantly lower in the sorcinF112L myocytes compared with NTG controls (Table 3
; P<0.05). Representative recordings at 0 mV are shown in Fig. 5
A, B. Although ICa,L inactivation was slowed in sorcinF112L myocytes (Fig. 5A
, Table 3
), the current density was not different. Between 0 and +30 mV, peak Ca2+ transients were modestly but significantly larger in the sorcinF112L myocytes (P<0.05) relative to those in the NTG myocytes (Fig. 5C
, Table 3
). To determine whether this increase in Ca2+ transient at 0 mV was due to augmented coupling of CICR, we calculated CICR gain as the rate of Ca2+ release divided by the integral of ICa,L (pC/pF) during a 200 ms sweep. However, whereas gain was voltage-dependent (P<0.05; ANOVA), it was not different for NTG and sorcinF112L myocytes (Fig. 5D
, Table 3
). To assess the respective roles for INCX and ICa,L, nifedipine (10 µM) was used to block ICa,L from –50 to +50 mV. Nifedipine blocked Ca2+ transients at all voltages tested for both NTG and sorcinF112L cells, suggesting INCX does not contribute to Ca2+ influx during EC coupling.
|
|
Ca2+ spark activity in quiescent myocytes
We next measured Ca2+ spark characteristics from sorcinF112L myocytes. We compared these data not only to NTG controls, but also to cells from mice overexpressing WT sorcin (sorcinWT). The sorcinF112L myocytes exhibited modest but significant elevations in half-maximal duration (D50) and half-maximal width (W50); (P<0.05; Fig. 6
A, Table 4
). Conversely, sorcinWT myocytes displayed significant decreases in D50 and W50 relative to NTG controls (P<0.05; Fig. 6A
, Table 4
). TTP and spark frequency were not significantly affected in any of the groups, although spark frequency within groups was highly variable.
|
|
Ca2+ spark activity in permeabilized myocytes
Since sorcin influences multiple components of Ca2+ regulation, we clamped [Ca2+]i by performing experiments in permeablized myocytes to examine the effects of sorcin on spark activity under more uniform conditions. Experiments were performed using 20 nmol/L [Ca2+]free. and under these experimental conditions there was a marked increase (243%) in spark frequency in sorcinF112L myocytes compared with NTG controls (Fig. 6B
, Table 4
). In addition, all permeabilized sorcinF112L myocytes that were tested exhibited sparks whereas only 70% of NTG myocytes exhibited sparks. Conversely, permeabilized sorcinWT myocytes exhibited a modest but significant decrease in spark frequency compared with NTG controls (P<0.05; Table 4
).
Modeling of intracellular Ca2+ dynamics
The two main changes in Ca2+ cycling observed in sorcinF112L mice, increased INCX and slowed inactivation of ICa,L, would be expected to lead to opposing effects on Ca2+ transient amplitude and SR Ca2+ load in the steady state. Greater Na+-Ca2+ exchange would promote extrusion of Ca2+ from the cell and lead to smaller Ca2+ transients, whereas slowed inactivation of ICa,L would be expected to cause increased cellular Ca2+ loading. To gain insight into the balance between these effects, we performed simulations with a computer model of the mouse ventricular myocyte (34)
.
Figure 7
A, B shows that simple modifications to the model equations for the septal mouse ventricular myocyte could reproduce the changes in INCX amplitude and ICa,L inactivation observed in sorcinF112L cells (see Materials and Methods for details). When these simulated myocytes were paced at different rates, control cells (black) displayed consistently larger Ca2+ transients than sorcinF112L cells (red), contrary to the experimental data. Sample Ca2+ transients are shown in Fig. 7C
and the steady-state 
[Ca2+]i vs. frequency relationship is displayed in Fig. 7D
. This result indicates that increased INCX causes greater effects on Ca2+ homeostasis than slowed inactivation of ICa,L, at least in this model. Due to the rapid repolarization of the action potential in the mouse, closing of LTCCs is primarily due to deactivation rather than inactivation, and slowed inactivation of ICa,L has only minimal effects on steady-state SR load during action potential simulations.
Because these simulations failed to recapitulate the larger Ca2+ transients seen in sorcinF112L mice, we performed additional simulations to gain insight into which additional Ca2+ transport pathways might have been affected. Lengthening the action potential by blocking repolarizing K+ currents caused an increase in Ca2+ transient amplitude at pacing frequencies above 2 Hz but a decrease at lower frequencies, inconsistent with experimental data (results not shown). In contrast, an increase in SERCA activity, when incorporated in the model along with the changes in INCX and ICa,L, resulted in larger Ca2+ transients and SR Ca2+ load at all frequencies, similar to what was observed (green traces and symbols in Fig. 7C, D
). This suggests that overexpression of sorcin may also promote Ca2+ uptake into the SR, as hypothesized by Matsumoto et al. (12)
.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
, 18)
. We found that overexpression of L112-sorcin in the heart resulted in significant effects on several aspects of cardiac ECcoupling, including 1) a slowing of ICa,L inactivation, 2) an increase in NCX activity, 3) an increase in Ca2+ spark duration and width, and 4) under conditions of clamped [Ca2+]i, a marked increase in spark frequency. These cellular changes were associated with a moderate increase in contraction amplitude at the single-cell level and the development of mild left ventricular dilatation with preserved overall ventricular function as measured in vivo.
Effects on ICa,L and NCX
Expression of L112-sorcin in the heart resulted in significant slowing of inactivation of ICa,L, including both the fast and slow time constants. Conversely, in a previous study we found that inactivation of LTCC was enhanced in TG mice overexpressing WT sorcin (7)
. Taken together, these results suggest a direct effect of sorcin on LTCC kinetics and are consistent with our previous observations indicating that sorcin biochemically associates with Cav1.2, the pore-forming subunit of the major cardiac Ca2+ channel (11)
. Whether or not sorcin acts independently of other regulatory mechanisms that influence LTCC inactivation such as that associated with calmodulin (CaM) and Ca2+/calmodulin-dependent protein kinase (CaMKII) is uncertain (35)
. However, it is intriguing that sorcin and CaM have similar binding sites within the Cav1.2 carboxy-terminus IQ domain (11
, 36)
.
We also measured a significant increase in Na+-Ca2+ exchange activity in sorcinF112L myocytes. Based on our Western blot analysis, this effect is most simply attributable to an increase in the abundance of the exchanger rather than a direct modulation of NCX function by sorcin (although the latter possibility is not excluded). An increased exchange activity has also been reported in myocytes transduced with an adenovirus encoding WT sorcin (15)
.
Effects on cell shortening and [Ca2+]i transients
At the single-cell level, sorcinF112L myocytes displayed enhanced shortening over a range of pacing frequencies and, consistent with this observation, increased Ca2+ transient amplitude. Caffeine-induced Ca2+ transients were also increased in myocytes from sorcinF112L mice, suggesting an increased SR Ca2+ load. Since neither the fractional release nor the CICR gain function measured in voltage-clamp experiments was increased, our data suggest that the larger Ca2+ transients result primarily from the elevated SR load. In addition, the bell-shaped relationship between Ca2+ transient amplitude and voltage as well as the sensitivity of Ca2+ transients to nifedipine suggests a negligible role for INCX during Ca2+ influx.
To synthesize these results with the changes in NCX activity and ICa,L observed, we performed simulations with a computer model of the mouse ventricular myocyte (cite). In isolation, increased NCX activity would be expected to lead to smaller Ca2+ transients and reduced SR Ca2+ load whereas slowed ICa,L inactivation would be expected to have opposite effects. When both changes were incorporated into the computer model, depletion of SR Ca2+ and smaller Ca2+ transients ensued (Fig. 7)
, indicating that increased INCX influences beat-to-beat Ca2+ cycling more than does slowed inactivation of ICa,L. The larger Ca2+ transients seen in experiments were reproduced in simulations in which SR Ca2+ uptake was also enhanced, suggesting that this Ca2+ flux may be augmented in sorcinF112L myocytes. This hypothesis would be consistent with previous studies that proposed an effect of sorcin on SERCA function (12
, 23)
. Consistent with this, sorcinF112L cells exhibited enhanced lusitropy.
Effects on Ca2+ sparks
Myocytes from the sorcinF112L TG mice exhibited modest but significant elevations in Ca2+ spark duration and width compared with NTG controls. Conversely, these parameters were decreased in sorcinWT myocytes. Because of the complex effects of sorcin on multiple aspects of cardiac EC coupling, as well as the potential for secondary adaptive/maladaptive electrical remodeling associated with long-term in vivo expression in our transgenic models, we also examined spark characteristics under highly controlled, albeit artificial, conditions (i.e., in Ca2+-buffered permeabilized myocytes). Under these conditions, the major finding was a marked increase in spark frequency in sorcinF112L myocytes compared with NTG controls. In contrast, overexpression of WT sorcin modestly but significantly depressed Ca2+ spark frequency.
Taken together, these data are consistent with and reinforce accumulating biochemical and in vitro functional data, suggesting that sorcin exerts a direct inhibitory influence on RyR2 gating (9
, 13)
. Conceivably, sorcin stabilizes the Ca2+-release channel in its closed state and may do so in a Ca+2-dependent manner. The inhibitory effects of the L112-sorcin protein appear to be defective, perhaps as a consequence of altered Ca2+-dependent translocation. It is worth noting, however, that although the inhibitory effects of overexpressed WT sorcin on spark characteristics were statistically significant, they were relatively modest. This suggests that levels of sorcin in normal myocytes may be nearly saturating, at least with respect to inhibitory effects on RyR2 gating. Moreover, our computational analysis of cellular Ca2+ cycling suggests that, when considered in the context of beat-to-beat Ca2+ regulation, the effects of sorcin on RyR2 gating may be relatively subtle compared with its modulation of other proteins such as the Na+-Ca2+ exchanger.
Mechanistically, whether the L112-sorcin protein simply competes with WT sorcin or instead exerts a dominant-negative effect is not yet clear. There is ample evidence that sorcin can homodimerize and exists within a multicomponent complex, which would facilitate a dominant-negative mechanism (37
, 38)
. For example, recent data suggest that sorcin and presenilin-2 interact and undergo Ca2+-dependent translocation to membrane-bound targets, including the cardiac RyR2 (39
, 40)
. Amplitudes of Ca2+ transients and peak developed tension are both increased in papillary muscles isolated from presenilin-2 knockout mice compared with WT controls (40)
, consistent with the hypothesis that sorcin normally promotes the closure of RyR2 and termination of CICR.
Another question that remains answered is why overexpression of the L112 variant or WT sorcin most prominently influences Ca2+ spark duration and width parameters when measured in quiescent myocytes whereas Ca2+ spark frequency is most dramatically affected when assayed in permeabilized myocytes under conditions of constant [Ca2+]i. Certainly the intracellular milieu is altered by the permeabilization process, and the abundance and/or functionality of additional cofactors that modulate RyR2 gating may very well be influenced. Moreover, relative differences in the affinity of WT and L112-sorcin proteins for Ca2+ may be greater when [Ca2+]i is held constant in the nanomolar range and differences in the extent of the subsequent inhibitory interaction with RyR2 may be maximized. In addition, it is conceivable that the SR Ca2+ load differs significantly in permeabilized vs. quiescent myocytes and this factor may secondarily influence spark frequency. Regardless, under both experimental conditions the presence of L112-sorcin enhances RyR2 activity whereas forced expression of WT sorcin does the opposite.
In vivo effects
The phenotypic consequences of forced expression of sorcin in isolated myocytes and intact hearts have been examined by several groups with rather divergent results. Using invasive hemodynamic assays, we previously reported that TG mice overexpressing WT sorcin had reduced contractile performance (7)
. However, in the present study using echocardiography, we did not observe this functional deficit. Conceivably this may reflect the lighter level of anesthesia required for echocardiographic studies and the preservation of compensatory autonomic reflexes that might tend to preserve systolic function. Seidler and colleagues also reported that contractile performance in isolated adult rabbit myocytes transduced with a sorcin-expressing adenoviral vector was depressed (15)
. Those investigators suggested that enhanced NCX activity and lower SR Ca2+ content might account for the reduced fractional shortening. However, other groups have found that transduction of mouse and rat hearts or isolated cells with sorcin-expressing adenoviral vectors significantly enhanced cardiac function and even rescued the abnormalities observed in a model of diabetic cardiomyopathy (23
, 41)
. At present, there is no satisfactory explanation for these divergent findings. It should be pointed out that discrepancies between in vivo and cellular measurements of contractile performance are not without precedent, and these discordant results have been ascribed to a variety of factors including loading conditions, neurohumoral factors, as well as altered cell-cell interactions and geometric properties that affect the intact heart but not isolated cells (42)
. Moreover, in this instance, the effects of sorcin on cardiac behavior may be time, species, and/or dose dependent, variables that differ significantly in TG vs. adenoviral model systems. In particular, longer term expression of high levels of sorcin may beget secondary electrical remodeling, thereby further influencing calcium handling, EC coupling, and contractile performance.
Relationship to human disease
TG mice overexpressing the variant L112-sorcin protein only in the heart develop mild left ventricular dilatation without evidence of hypertension and with no demonstrable hypertrophy at either the cellular or whole heart level. In fact, myocyte dimensions appear diminished, a finding corroborated by measurements of cellular capacitance. In contrast, patients heterozygous for the sorcinF112L variant reportedly develop hypertension and hypertrophic heart disease (18)
. Thus, it remains uncertain whether the nonsynonymous SNP in the sorcin gene is indeed responsible for the clinical manifestations seen in affected individuals. If additional human genetic data support the association, our data suggest that the hypertension may well be due to expression of L112-sorcin in other organ systems that influence blood pressure, such as the vasculature or kidney—both tissues in which sorcin is expressed (43
, 44)
. Moreover, whereas expression of the variant form of sorcin in the heart may be necessary or perhaps permissive for development of the hypertrophic phenotype, our data indicate that it is not sufficient. Additional human genetic analysis, in concert with murine transgenic or gene targeting studies, will be required to further elucidate the molecular pathogenesis of the human condition.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication April 12, 2006. Accepted for publication August 29, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
