Main

Myocardial contractility increases with development. Adult myocardium develops more force than fetal myocardium (1). Adult cardiac myocytes shorten further and faster than immature myocytes (2). Potential contributors to these differences include developmental increases in calcium ([Ca2+]i) transients and calcium sensitivity of myofilaments (1,3). Calcium transients increase postnatally in the rabbit (4). The calcium sensitivity of myofilaments, however, decreases with development in the rat, chick, and rabbit (5–7); this fall could result from a developmental switch in isoform expression of cTnT, a thin filament protein essential for Ca2+-regulated force development (8). Expression of cTnT1 and cTnT2 falls and that of cTnT3 and cTnT4 increases with development in the rabbit.

We expressed recombinant rabbit (r)cTnT1 and rcTnT4 in Sf9 cells and compared their effects on Ca2+-binding of isolated troponin in solution and in skinned muscle fibers in vitro. Further, we compared calcium transients and L-type calcium currents (ICa) in cardiac myocytes from 3-wk-old rabbits expressing cTnT1 and from adults that express predominantly cTnT4 and no cTnT1.

We found that (a) in solution, Tn comprising cTnT1 (TncTnT1) has higher calcium affinity than Tn comprising cTnT4 (TncTnT4), (b) fibers reconstituted with cTnT1 were more sensitive to calcium than those reconstituted with cTnT4, and (c) peak [Ca2+]i and ICa were significantly greater in adult than in 3-wk-old myocytes. Our data provide no evidence that incorporation of Tn into the myofilaments alters Tn-calcium binding affinity.

By imparting to the myofilaments a higher calcium sensitivity, the higher expression of cTnT1 and cTnT2 in immature myocytes would compensate partially for their smaller [Ca2+]i transients.

MATERIALS AND METHODS

Animals used in this study were handled according to the animal welfare regulations of Duke University (Durham, NC). The protocol was approved by the Institutional Animal Care and Use Committee. The rabbits were anesthetized (100 mg/kg ketamine and 10 mg/kg xylazine i.p.), and the heart was removed to obtain ventricular muscle fibers or isolated myocytes.

Preparation of cTnT.

cTnT isoforms were expressed in Sf9 cells using the baculovirus system (Invitrogen). The cDNA of rabbit cTnT1 (9) was cloned at the EcoR I restriction site into pVL1392 by releasing the cDNA from pBSII KS(+) with HindIII and KpnI and cloning it into pSP72 (Promega) where a second XbaI site was added. The cTnT1 cDNA was then excised with XbaI and cloned into pVL1392. Orientation of inserts and sequences encoding cTnT1 and cTnT4 were verified by polymerase chain reaction and DNA sequencing. SF9 cells were transfected with pVL1392-containing cTnT1 cDNA or pVL1393-containing cTnT4 cDNA and infected with wild-type AcMPEV (Invitrogen). Clones were identified using monoclonal antibody (MAb) 13-11, a cTnT-specific MAb, and Western blots (10).

Sf9 cells (3–4 ×106 cells/mL) were infected with cTnT1 or cTnT4 recombinant baculovirus. cTnT was purified according to Potter (11), and frozen at −70°C in 6 mol/L urea.

Circular dichroism.

CD spectra for cTnT1 and cTnT4 were obtained at 25°C (0.5-nm steps with 1-nm bandwidth) in a buffer containing 0.5 mol/L KCl and 10 mmol/L potassium phosphate, and 3 mg/mL protein, pH 7.2.

Preparation of cTnC and cTnI.

Human cTnC was expressed in BL21 (DE3) Escherichia coli using an expression vector provided by Merck and Co. Inc. and purified (12). Recombinant mouse cTnI was expressed using a pET-3d vector and purified using established methods (13). Stock protein solution concentrations were measured using quantitative amino acid analysis.

2-[4′-(Iodoacetamido)aniline]-naphthalene-6-sulfonate (IAANS) labeling of cTnC.

Purified cTnC was dissolved in buffer A (0.2 mol/L KCl, 30 mmol/L 3-(N-morpholino)propane-sulfonate (MOPS), pH 7.0) to a final concentration of 50 μmol/L, and dialyzed serially against buffer A containing 1 mmol/L DTT and buffer A containing 6 mol/L urea without DTT (1,4-Dithiothreitol). cTnC was labeled using a 15-fold molar excess IAANS (Molecular Probes, Eugene, OR). The labeling stoichiometry was determined by 326-nm UV absorbance (IAANS molar extinction coefficient 26,000). Absorbance of cTnC at 280 nm was corrected for fluorophore absorbance (14). Labeling reactions yielded 2 mol bound IAANS/mol cTnC.

Preparation of soluble troponin complexes.

Binary and ternary troponin complexes were formed at 20°C in titration buffer (in mmol/L: KCl 400, ethyleneglycol-bis-(β-aminoethylether)-N,N,N′,N′ (EGTA) 1, MOPS 25, DTT 1.5, pH 7.2). Molar ratios of IAANS-cTnC:cTnI:cTnT were 1:2:2 to avoid dimer formation. Complexes of IAANS-cTnC:cTnI, IAANS-cTnC:cTnT1, and IAANS-cTnC:cTnT4 were prepared similarly.

Calculation of [Ca2+]free.

[Ca2+]free was calculated using CATCONC (15) software and the logarithmic association constants: H+:EGTA4− 9.47, H+:HEGTA3− 8.85, H+:H2EGTA2− 2.66, H+:H3EGTA− 2.0; Ca2+:EGTA4− 10.97, Ca2+:HEGTA3− 5.29, Mg2+:EGTA4− 5.21, Mg2+:HEGTA3− 3.43.

Ca2+-troponin affinity in solution.

Fluorescence measurements were made using a SLM 8100 spectrofluorometer (SLM/Aminco). Titration buffer (mmol/L): KCl 0.4, EGTA 1, and MOPS 25, pH 7.2, 20°C. Fluorescence emission spectra were obtained using 335-nm excitation (excitation and emission bandwidths: 4 and 32 nm, respectively). Uncorrected fluorescence intensities between 400 and 500 nm were integrated and the fraction of Ca2+-bound IAANS-cTnC (Fb) calculated: MATH

where Iobsv is the integrated fluorescence intensity observed after each addition of Ca2+, IEGTA is the integrated fluorescence in the absence of Ca2+, and Isat is the fluorescence under saturating Ca2+). Data were fit to a form of the Hill equation:

where nH is the Hill coefficient and Kd is the dissociation constant).

Force-pCa relationship.

Right ventricular muscle fibers from adult rabbit hearts (16) were skinned (0.5% Triton X-100) in relaxing solution. Compositions of relaxing (pCa 9) and pCa 4.5 solutions (6) and their preparation were as described previously (17), except pMg was 3.0, and all solutions contained 0.1 mg/L pepstatin A. Apparent Ca-EGTA stability constant used was 3.702 × 106 L/mol.

Fibers were mounted on a force transducer (Güth OP1M (18)) at 22.0°C and sarcomere length of 2.3 μm (determined by light diffraction). Active force was determined versus pCa.

Dialysis of rcTnT.

rcTnT was dialyzed at 4°C for 24 h, first against 2 mol/L urea and 1 mol/L KCl, then sequentially against 1, 0.8, 0.6, and 0.4 mol/L KCl, all containing 20 mmol/L MOPS, 1 mmol/L DTT, and 0.1 mg/L pepstatin A, pH 7.0.

Solutions for skinned preparation experiments.

The washing solution used (in mmol/L) was KCl 100, MgCl2 5, MOPS 20, DTT 0.5, pepstatin A 0.1, pH 7.0. The extraction solution was KCl 250, EGTA 5, MgCl2 5, MOPS 20, DTT 0.5, pepstatin A 0.1 (pH 6.5), and 1–3 mg/mL rcTnT1 or rcTnT4. The reconstitution solution was 2.5 mg/mL cTnI-cTnC complex added to either (in mmol/L) KCl 100, EGTA 5, MgCl2 5, DTT 0.5, pepstatin A 0.1 mg/L, MOPS 20, or imidazole 120, KCl 29, MgCl2 16, EGTA 7, ATP 5, phosphocreatine 16 (pH 7).

Extraction-reconstitution assay.

We superfused the preparation in washing solution (10 min), then in extraction solution (70–90 min) to remove troponin (19). Loss of Ca2+-regulated force was verified using pCa 4.75. The fiber was then washed (10 min) and placed in reconstitution solution 75 min). In some preparations, extraction-replacement was performed in one step: preparations were exposed for 100 min to extraction solution containing TncTnT1 or TncTnT4 complex. The force-pCa relationship was measured, the fibers placed in sample buffer, and their proteins resolved using sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and silver stained. Western blots were probed with MAb 13-11 (10,20).

Cell isolation.

Myocytes were isolated following the procedure of Mitra and Morad (21). The digestion solution for 3-wk-old animals contained 0.5% collagenase (Worthington type II) and 0.006% protease (Sigma Chemical Co., type XIV) in 50 mL of physiologic solution (no added calcium); for adult animals, we used 1% collagenase and 0.006% protease.

Electrophysiology and calcium measurement.

Cells were studied under a Zeiss IM microscope using a Nikon 40x, 0.85-N.A. Fluor objective. Electrophysiologic data were acquired using Axopatch 200A, Digidata 1200, and pClamp 6.0 (Axon Instruments) (in early experiments, we used a Dagan Corporation Model 8900). FluoroPlex-III system (Tracor Northern) was used for fluorescence measurements (excitation wavelengths 350 and 380 nm, chopper rate, 60 s−1). Ventricular myocytes were whole-cell clamped (22) at room temperature (pipette resistance 1.8–2.5 MΩ). ICa was acquired at 10 or 20 kHz (2-kHz, four-pole, low-pass Bessel filter). Capacitance and 70–80% series-resistance compensation were applied. Holding potential (VH) was −80 mV; five conditioning 400-ms pulses to +10 mV at 4-s intervals standardized sarcoplasmic reticulum calcium load. VH from the end of the last conditioning pulse to the 400-ms test pulse was −40 mV (to inactivate sodium and T-type calcium channels). Resting potential (Em) was recorded immediately upon patch rupture and [Ca2+]i and ICa 5–10 min post-rupture.

Solutions for cardiac myocyte experiments.

The extracellular solution (in mmol/L) was NaCl 141.8, KCl 5.1, MgCl2 1.0, dextrose 12, MOPS (acid) 3.75, MOPS (Na salt) 6.25, and CaCl2 2.0, pH 7.4 and the pipette solution was KCl 143, Na2ATP 5.1, MgCl2 5.04, MOPS (acid) 10, EGTA 0.002, and K5fura-2 0.05 or 0.1, pH 7.2.

Calibration of fura-2 fluorescence.

We carried out in-cell calibration at room temperature as described by Li et al. (23).

Data analysis.

[Ca2+]i was calculated according to Grynkiewics et al. (24) using Ca-fura-2 Kd of 200 μmol/L. Background was subtracted before calculating fluorescence ratio. ICa was measured as negative peak minus steady-state current at 300 ms; this was equal, both in adult and 3-wk-old animals, to the (200 μmol/L) Cd2+-sensitive current. Cell capacitance was calculated from the time constant of response to small (±5 mV) test pulses from Vh = −80 mV sampled at 250 kHz (10-kHz filter) before compensation.

We used the notation pCa50 = −log10[Ca2+] for the pCa that elicited half-maximal tension, and pKd = −log10Kd for protein solutions.

Statistics.

Means or fitted values ± their standard error (SE) are given unless otherwise stated. We used an unpaired t test or t test for unequal variance. The paired t test was used to assess the difference in binding of Ca2+ to IAANS-TncTnT1 and IAANS-TncTnT4. ICa was corrected for cell capacitance by linear regression; ICa density was based on Cm of 80 pF. Statistical analysis was carried out using R (http://www.R-project.org)

RESULTS

Characterization of recombinant protein.

The purity and integrity of rcTnT1 and rcTnT4 were verified by comparison to native cTnT using SDS-PAGE and Western blots (Fig. 1A), and CD spectra, which exhibited two minima at 208 and 222 nm and were like the spectrum of native cTnT (25).

Figure 1
figure 1

Recombinant and native cardiac troponin T. A, Silver stains comparing electrophoretic mobilities of purified recombinant (r)cTnT1 (a) and rcTnT4 (b). Western blots, probed with MAb 13-11, which recognizes an epitope specific to cTnT, compare electrophoretic mobilities of purified rcTnT1 (d) and rcTnT4 (e) to the four native isoforms (cTnT1–4) from 2-d-old rabbit myocardium (c). cTnT isoform expression in neonatal heart (c) differs from that of adult rabbit heart (C) in which cTnT1 is absent and TnT4 is the major isoform. B, Proteins from a control (I) and an experimental (II) fiber (8.5% acrylamide gel). Endogenous cTnT isoforms in II were replaced with rcTnT1. The third lane contains rcTnT1. C, Western blots (probed with MAb 13-11) of a control adult fiber (a) and two experimental fibers (b, c), reconstituted, with rcTnT1 (b) and rcTnT4 (c). In the experimental fibers, only rcTnT1 (b) and rcTnT4 (c) are evident.

Full size image

Ca2+ concentration effects on fluorescence of IAANS-cTnC, IAANS-cTnC:cTnI, IAANS-cTnC:cTnT, and IAANS-cTnC:cTnI:cTnT containing rcTnT1 or rcTnT4.

We examined the effects of binding of Mg2+ or Ca2+ to cTnC high-affinity sites III and IV and of Ca2+ to the low-affinity site II on IAANS-cTnC fluorescence. In the absence of Mg2+, IAANS-cTnC fluorescence intensity decreased slightly between pCa 8.5 and 6.0, then increased as pCa was further reduced to 3.2, the increase being attributable to Ca2+ binding to site II (Fig. 2A). The apparent Kd of free IAANS-cTnC:Ca2+ was 0.104 μmol/L for the high-affinity sites and 9.7 μmol/L for the low-affinity site (data not shown).

Figure 2
figure 2

Effect of Ca2+ on fluorescence of IAANS-labeled cTnC (A) and IAANS-cTnC:cTnI complex (B) expressed as percentage of change relative to initial fluorescence. Data shown as mean ± standard deviation (five titrations).

Full size image

The IAANS-cTnC:cTnI complex was titrated with Ca2+ in the presence of 4 mmol/L Mg2+ to saturate both cTnC high-affinity sites III and IV (Fig. 2B). Like Putkey et al. (26), we found a minimal decrease in fluorescence with increasing [Ca2+]. Fluorescence of IAANS-cTnC:cTnT1 and IAANS-cTnC:cTnT4 complexes did not change with [Ca2+] (data not shown).

The fluorescence intensity of IAANS-TncTnT1 and IAANS-TncTnT4 decreased with [Ca2+]: Kd for IAANS-TncTnT1:Ca2+ was 0.52 ± 0.08 μmol/L and for IAANS-TncTnT4:Ca2+ 0.83 ± 0.09 μmol/L (p = 0.001; paired t test, five determinations); nH were 1.0 ± 0.1 and 1.3 ± 0.2, respectively (Fig. 3).

Figure 3
figure 3

Calcium binding to TnC. Fraction of Ca2+-bound IAANS-cTnC (Fb) as a function of log10[Ca2+] for TncTnT1 (thick line, filled symbols), and TncTnT4 (thin line, open symbols) illustrates the greater calcium affinity of TncTnT1.

Full size image

Extraction-reconstitution experiments.

Mean pCa50 of the fibers with endogenous cTnT was 5.99 ± 0.07 and nH was 2.3 ± 0.5 (n = 10). SDS-PAGE (Fig. 1B) and Western blots (Fig. 1 C) demonstrated replacement of endogenous cTnT by rcTnT1 or rcTnT4. pCa50 postreconstitution was 6.07 ± 0.04 (n = 6) for rcTnT1, and 5.75 ± 0.07 (n = 4) for rcTnT4, p = 0.003 (Fig. 4) and nH was 2.4 ± 0.4 and 1.9 ± 0.1, respectively (p = 0.34).

Figure 4
figure 4

Force vs pCa. Normalized force vs pCa data for fibers in which the endogenous cTnT was replaced with either rcTnT1 (+, solid curve) or rcTnT4 (○, dashed curve) illustrates the greater calcium sensitivity of force development in fibers containing rcTnT1. The curves correspond to mean values for pCa50 and nH.

Full size image

Comparison of Tn-Ca binding in solution and in myofilaments.

To assess whether incorporation of troponin into the myofilaments changed troponin:calcium affinity, we compared the differences (Δ) in Gibbs free energy (ΔG): ΔΔGsolution = ΔGTnT1 − ΔGTnT4 = 2.303RT·(pKd [TnT1] − pKd [TnT4]) for Tn in solution, and ΔΔGmuscle = 2.303RT·(pCa50[TnT1] − pCa50[TnT4]) in the myofilaments (R, universal gas constant; T, absolute temperature). This assumes that at pCa50, calcium binding sites of myofilament-bound cTnC are half-occupied. The two quantities were not significantly different: ΔΔGsolution was 0.29 ± 0.07 (95% confidence interval: 0.13–0.46) kcal/mol, and ΔΔGmuscle was 0.42 ± 0.10 (0.18–0.66) kcal/mol (Fig. 5).

Figure 5
figure 5

Difference in free energy (ΔΔG) of Ca2+ binding to cTnI:cTnC:cTnT1 and to cTnI:cTnC:cTnT4 in solution (Solution) and in fibers reconstituted with rcTnT1 and rcTnT4 (Muscle). Error bars: 95% confidence intervals.

Full size image

Ventricular myocytes.

Body weights of 11 adult and 16 3-wk-old rabbits from which myocytes were obtained were 2840 (standard deviation [SD] = 300) g and 300 (SD = 100) g. Cells from 13 animals yielded [Ca]i and electrophysiologic data; cells from 13 animals yielded [Ca]i data.

Peak [Ca2+]i in adult myocytes was twice that in 3-wk-old myocytes (Table 1, Fig. 6A). Diastolic [Ca2+]i and resting potentials in adult and 3-wk-old myocytes were not significantly different. Membrane capacitance (Cm) was significantly greater in adult myocytes due to their larger cell size (2). ICa and ICa density (Table 1, Fig. 6B, C) were both greater in adult myocytes. The peak of the ICa versus voltage curve in both age groups occurred at +10 mV (Fig. 6C).

Table 1 Comparison of [Ca] i , resting potential (E m ), C m , and I Ca between ventricular myocytes from adult and 3-wk-old rabbits
Full size table
Figure 6
figure 6

Comparison of calcium transients and L-type calcium currents from 3-wk-old and adult rabbit ventricular myocytes. A, Calcium transients from the 3-wk-old (upper trace) show much smaller transients than the adult. Diastolic levels were 155 and 103 nmol/L, respectively. Black traces: filtered. B, L-type calcium currents from the 3-wk-old (black trace) show a significantly smaller current than the adult (gray trace) myocyte. C, ICa-voltage relations from a 3-wk-old (capacitance 61 pF, closed symbols) and an adult (capacitance 109 pF, open symbols) myocyte.

Full size image

DISCUSSION

Myocardial contractility has been shown to increase with development.

The [Ca2+]i transient increases postnatally, but the calcium sensitivity of the myofilaments decreases. cTnT isoform expression also changes with development and is altered by heart disease (9,27,28). Although these changes, with opposite effects on contractility, appear to be at cross-purposes, the greater calcium sensitivity of tension would seem to partially compensate for the lower [Ca]i in the immature myocyte.

Four cTnT isoforms, cTnT1-4, expressed in the rabbit and human heart (9,27,29), are products of combinatorial alternative splicing of two exons (9). Expression of cTnT1 peaks in fetal human (36% of total cTnT) (27) and neonatal rabbit heart. In adult rabbit heart, cTnT4 is the dominantly expressed isoform, whereas in the normal adult human, only cTnT3 is expressed (9,27). These changes in isoform expression suggest that cTnT isoforms have functional consequence.

cTnT isoforms in solution.

We determined for the first time, in solution, that the affinity of TncTnT1 was greater than TncTnT4 and that this difference did not require incorporation of Tn into the myofilaments.

Calcium titration did not affect IAANS-cTnC:cTnT fluorescence, although it affected IAANS:Tn fluorescence, the effect depending on the presence of cTnT1 or cTnT4, suggesting that cTnT isoform amino-terminal regions, which are distant from known cTnT, cTnC, and cTnI interaction sites, affect the interactions among cTnC, cTnI, and cTnT, and this effect requires cTnI.

cTnT isoforms in the myofilament: the force-pCa relationship.

We found that the calcium sensitivity of active tension was higher in rabbit fibers reconstituted with rabbit rcTnT1 than with rabbit rcTnT4, in agreement with Gomes et al. (30), who examined the effects of human (h)cTnT isoforms on the force-pCa relationship of porcine fibers. There was an apparent rightward shift of the force-pCa relationship post-replacement. This may be due to our using mouse cTnI, which is 92% homologous with rabbit cTnI (we also used human cTnC, but that has 100% homology with rabbit cTnC). Also, extraction can affect other myofilament proteins, e.g. myosin binding protein C (31). Whatever the reason, both experimental groups would be affected similarly because the only difference between them was the cTnT isoform used. These results suggest that the developmental decrease in expression of cTnT1 and cTnT2 and increase in cTnT3 and cTnT4 contribute to the decrease in calcium sensitivity in rabbit and rat myocardium (7,32).

Developmental changes and species differences in isoform expression of slow skeletal (ss)TnI, cTnI, tropomyosin and myosin heavy chain (MHC) may contribute to differences in calcium sensitivity (3,33,34). The neonatal switch in expression from ssTnI to cTnI could lower calcium sensitivity (3,35,36) and modify cTnT isoform effects on tension (30). However, we found, in neonatal rabbit, a positive correlation between the sensitivity of myofilaments to calcium and cTnT2 expression but no evidence of a difference in relative cTnI to ssTnI expression (16).

Comparison of calcium binding of CTnT isoforms in solution and the myofilaments.

In the thin filament, the cTnT amino-terminus overlaps the head-to-tail binding of adjoining tropomyosin molecules, suggesting a structure-function relationship. We hypothesized that incorporation of troponin into the myofilaments would alter the calcium affinity induced by the troponin isoforms in solution. If so, ΔΔG between calcium binding to TncTnT1 and TncTnT4 in solution should differ from that in myofilaments: we found, however, that the difference in ΔΔG between fibers containing rcTnT1 and rcTnT4 resulted mainly from the intrinsic differences in TncTnT1 and TncTnT4 and not from incorporation of troponin into myofilaments (Fig. 5).

Systolic [Ca2+]i, Cm, and ICa density were larger in adult than in 3-wk-old rabbit ventricular myocytes, in agreement with previous reports (37–39). In an earlier study (2), we found larger and faster sarcomere shortening in the adult myocyte, consistent with its higher systolic [Ca2+]i. We found no developmental difference in diastolic [Ca2+]i, consistent with our previous finding of no difference in resting sarcomere length (2).

SUMMARY

In solution and in myofilaments, the affinity of TncTnT1 for calcium was higher than that of TncTnT4; our results provide no evidence that these affinities are modified upon incorporation of Tn into myofilaments. The higher calcium sensitivity of myofilaments from immature myocardium, due at least in part to developmental changes in cTnT isoform expression, may compensate for its smaller [Ca2+]i transient: for the same [Ca2+]i, a small shift of the force-pCa curve can result in a dramatic difference in generated force, given the steepness of the curve in the activation region (Fig. 4).