DIRECT LOADING OF PURIFIED ENDOGENOUS INHIBITOR INTO THE CYTOPLASM OF PATCHED CARDIOMYOCYTES BLOCKS THE ION-CURRENTS AND CALCIUM TRANSPORT THROUGH THE NCX1 PROTEIN

Abstract

The Na⁺/Ca²⁺ exchanger in mammalian heart muscle (NCX1) is the central transporter protein that regulates Ca²⁺ extrusion from the heart cell. However, the functional biochemistry and physiology of NCX1 has been severely hampered by the absence of any specific high-affinity inhibitor. Here we describe advanced procedures for purifying a candidate inhibitor, previously called endogenous inhibitor factor (NCX_IF) and demonstrate its direct actions on NCX1 activities in the single cell system. A combination of advanced HILIC (Hydrophilic Interaction Liquid Chromatography) procedures with analytical tests suggests that NCX_IF resembles the properties of small (disaccharide size) polar molecule lacking any aromatic rings, conjugated bonds, or primary amino group. The effects of NCX_IF on the NCX1-mediated ion-currents (I_NCX) and cytosolic Ca²⁺-extrusion were detected by a combination of patch-clamp and confocal microscopy under conditions at which the purified NCX_IF was directly loaded into the cytoplasm of patched cardiomyocytes. It was demonstrated that cytosolic NCX_IF blocks the Ca²⁺-activated NCX1 inward current and the accompanying Ca²⁺-extrusion from the cell with high efficacy. A constant fraction of NCX1 inhibition was observed under conditions at which the cytosolic [Ca²⁺]_i was varied at fixed doses of NCX_IF, suggesting that the degree of inhibition is controlled by NCX_IF dose and not by cytosolic Ca²⁺ concentrations. NCX_IF blocks equally well both the Ca²⁺ extrusion and Ca²⁺ entry modes of NCX1, consistent with thermodynamic principles expected for the functioning of a bidirectional “carrier-type” transport system. We concluded that NCX_IF interacts with a putative regulatory domain from the cytosolic side and thus, may play an important regulatory role in controlling Ca²⁺ signaling in the heart. This may represent a new potential tool for developing novel treatments for cardiac Ca²⁺ signaling dysfunction.

Proteins of the NCX gene family contribute to Ca²⁺ regulation in many cell types (1–4), with three genes responsible for expression, namely, NCX1, NCX2, and NCX3 with multiple splice variants (2,5–7). In heart, the electrogenic Na⁺/Ca²⁺ exchange (8) is due to the gene product of NCX1.1 and, although the molecular and biophysical properties have been broadly studied (2–7), endogenous regulation is incompletely understood (9–13). Nevertheless, it is clear that changes in NCX1 protein expression accompany the development of diverse diseases such as heart failure and arrhythmia (2–4,9). There is, however, considerable uncertainty regarding the role played by NCX1 in these diseases and, moreover, to date, there is neither evidence that the NCX1 activity is the primary cause of such diseases nor are mutations of NCX1 linked to any specific disease. Nevertheless, NCX1 protein levels may change and certainly contribute to Ca²⁺ transport and signaling dysfunction, although the details remain poorly understood (2–4,9).

The Na⁺/Ca²⁺ exchanger turnover rate is clearly affected by the intracellular Ca²⁺, Na⁺, and H⁺ ions, which interact with the regulatory cytosolic f-loop of NCX proteins (3,4,10–13). Other cellular factors (ATP, PIP₂, and lipids) also can modulate NCX1, but their physiological relevance is still unclear (11,12). No phosphorylation of the cardiac NCX1 protein has been demonstrated either in vitro or in vivo. Recent breakthroughs in elucidating the 3D structure revealed that the f-loop contains two Ca²⁺-binding domains, CBD1 and CBD2 (14–16), which may modulate the NCX1 activity by sensing the cytosolic Ca²⁺ concentrations. If specific endogenous factors exist, affecting the Ca²⁺ interactions with the cytosolic regulatory domains, they may be quite important.

We isolated (from calf ventricle extracts) and purified a low molecular substance that inhibits NCX1 in sarcolemma vesicles (17,18). Mass-spectroscopic and gel-filtration analyses of active substances revealed a small (350–550 Da) polar molecule that is not retained on the regular RP-HPLC columns (17,18). Chemical tests and spectroscopic analyses showed no content for conjugated bonds, aromatic rings, primary amines, or aldehyde groups, thereby resembling the properties of non-reducing “sugar-like” molecules (17,18). This low-abundance compound inhibits NCX1 activites in the range of 10⁻⁶–10⁻⁷ M (18), but it has no effect on the Na⁺/K⁺-ATPase or SR Ca²⁺-ATPase activities (19). This compound was termed “the NCX inhibitory factor” or NCX_IF.

Application of NCX_IF to organ bath solutions elevates the contractility of ventricle strips (17,20), meaning that NCX_IF could be a physiologically active substance. Importantly, the β₁-blocker (deraline) does not affect these actions (20), suggesting that NCX_IF may act through a mechanism that is independent of the β-adrenergic pathway. Moreover, NCX_IF could suppress ouabain-induced arrhythmias, thereby resembling the properties of anti-arrhythmic agents (23). This suppression of arrhythmia may represent the NCX_IF-induced block of Ca²⁺ activated inward I_NCX (triggered during Ca²⁺ overload).

Extracellular perfusion of electrically paced cardiomyocytes with NCX_IF resulted in enhanced amplitudes of cell contractions and of [Ca²⁺]_i transients (22). Although these observations can be rationally explained by primary inhibition of NCX1 in the cell, no evidence existed for this claim. The reason for this lack of information was that no patch-clamp techniques were applied at this time for testing the NCX_IF effects in a single-cell system. Therefore, in the present work the effects of NCX_IF on I_NCX currents and Ca²⁺-extrusion were tested under conditions at which NCX_IF was directly loaded into the cytoplasm of patched cardiomyocytes. Here, the cytosolic [Ca²⁺]_i was transiently elevated (by exposing cardionyocytes to caffeine for short periods), whereas the I_NCX and [Ca²⁺]_i signals were subsequently recorded with patch-clamp and confocal microscopy.

If NCX1 is affected by NCX_IF, it is also important to know where on the protein NCX_IF acts. Kinetic studies with sarcolemma vesicles suggested that NCX_IF acts as a non-competitive inhibitor with respect to extra-vesicular (cytosolic) Ca²⁺ (21). This was taken to mean that NCX_IF interacts with a regulatory (but not transport) site on NCX1 protein and may thereby alter the rate-limiting step of the ion-transport cycle. In intact cells, the time of action of extracellularly applied NCX_IF on contractility and [Ca²⁺]_i transients was quite slow (5–15 min), suggesting that NCX_IF must enter the cell to act (20,22). To test the sidedness of the action of NCX_IF, we directly loaded purified NCX_IF into the cytoplasm through the patch clamp electrode and examined the action of NCX_IF on I_NCX and Ca²⁺-extrusion. We provide strong evidence that NCX_IF blocks NCX1 from the inside of the cell with great efficacy, meaning that NCX_IF may be an endogenous regulator of Ca²⁺ transport through the regulation of NCX1 in intact cells.

Materials and Methods

Materials and reagents

Protease inhibitors (PMSF, pepstatin, leupeptin, and aprotenin), Deoxyribonuclease I (type DN-25), fluorescamine, Dowex-50W 1×8, Dowex-AG 1×4, and laminin were obtained from Sigma (St. Louis, MO). Sephadex G-10 (fine) was from Pharmacia (Uppsala, Sweden). The glass microfiber filters (GF/C Whatman) were purchased from Tamar (Jerusalem, Israel). The Chelex 100 was obtained from Bio-Rad (USA). ⁴⁵CaCl₂ (10–30 mCi/mg) was purchased from DuPont NEN (Boston, MA) or Perkin Elmer (Monza, Italy). Fluo-4 (K₅) and fluo-3 AM were from Teflabs (Austin, TX, USA), Biotium (Toronto, Canada) or Molecular Probes (Eugene, OR, USA). The following reagents were used for preparing cardiomyocytes: Penicillin streptomycin (Biological Industries, Haemek, Israel), Bovine calf serum (Hyclone, USA), collagenase II (Worthington, NG, USA), trypsin, minimal essential medium-1 and L-glutamine (Gibco Invitrogen, Scotland). The scintillation cocktail Opti-Fluor was from Packard (Groningen, The Netherlands). All other reagents were analytical or HPLC grade. Deionized water (18 MegaOhms/cm, Millipore system) was used for preparing solutions.

Effect of NCX_IF on Na_i-dependent ⁴⁵Ca uptake in sarcolemma vesicles

The sarcolemma vesicles were obtained according to established procedures (27,28). The Na_i-dependent ⁴⁵Ca uptake was initiated by rapid dilution of Na-loaded vesicles in an assay medium containing ⁴⁵Ca with lipid/protein-free aliquots of NCX_IF (17–21). The inhibitory capacity of NCX_IF preparations was quantitatively evaluated in arbitrary inhibitory units (one inhibitory unit represents the amount of NCX_IF that inhibits the Na_i-dependent ⁴⁵Ca-uptake by 1% under fixed experimental conditions) (17–23). The samples can be stored at −70°C without any loss of inhibitory activity of NCX_IF at least for 8–12 months.

Extraction and pre-column purification of NCX_IF

NCX_IF was extracted from 6–8 kg calf ventricle muscle with 95% ethanol, as previously described (17–23). Briefly, after ethanol was evaporated, the remaining aqueous phase was reduced to 0.3–0.5 L and insoluble material was removed by centrifugation (20,000 × g for 30 min). From the supernatant, lipids were removed by chloroform, and cold ethanol (−20°C) was added to the aqueous phase (to yield ~ 80%). White precipitate was removed by centrifugation and after ethanol was evaporated, cold acetone was added to yield ~ 90%. White precipitate was removed by centrifugation and concentrated material (100–150 ml) was passed through equal volumes of Dowex 2 × 8 (acetate form) and Dowex 50W (H⁺ form) slurry. The collected effluents were reduced to 50–80 ml (yellow-brown color) and then passed through the CarbPacking SPE (60 ml/10gr) cartridge (Supelco, Bellefonte, USA). The nearly colorless effluent (30–50 ml) was loaded onto the Sephadex G-10 column (5.5 × 70 cm) and flashed with water (2 ml/min). The fractions containing the NCX_IF activity were concentrated, filtrated, and stored at −70°C until purified further (17,18).

Chromatography

Fully computerized HPLC systems were used for NCX_IF purification, as described in previous publications (18–23). A combination of RP- and HILIC HPLC columns was used in the present work for NCX_IF purification: The Synergi Polar (Phenomenex, USA, 250 × 21 mm, 4 μm), TSK-gel Amide-80 (TosoHaas, Ltd. Japan, 21 × 300 mm, 7 μm), Krumasil Silica 100A (Phenomenex, USA, 250 × 21.2 mm, 5 μm), apHera-NH₂ (Astec, USA, 250 × 4.6 mm, 5 μm), Cyclobond I 2000, β-cyclodextran (Astec, USA, 250 × 4.6 mm, 5 μm), ZIC-pHILIC (SeQuant, Sweden, 150 × 4.6 mm, 5 μm) and Aminex HPX-87c (250 × 4 mm, 9μm). For HILIC columns, the injecting samples were dissolved in 80–95% methanol. All HPLC columns were run at 22–25°C, besides the Aminex column, which was explored at 60–80°C.

Preparation of cardiomyocytes

Isolated ventricular myocytes were obtained from adult female Sprague-Dawley rats (225–250 gr). Briefly, rats were deeply anesthetized with sodium pentobarbital (0.1 mg/gr ip) and heparinized (1.25 U/gr ip). Fifteen min after heparin was injected, the heart was rapidly excised and rinsed with cold 250 μM EGTA isolation buffer containing (in mM): 130 NaCl, 1 lactic Acid, 5.4 KCl, 3 Na-pyruvate, 25 HEPES, 0.5 MgCl₂, 0.33 NaH₂PO₄, 22 D-Glucose, 0.01 u/ml Insulin, at pH 7.4 (adjusted with NaOH). The aorta was quickly cannulated for Langendorff perfusion. The heart coronary arteries were perfused at 37° C for 2 min with EGTA isolation buffer, and then perfused for 8 minutes with the same buffer supplemented with 50 μM CaCl₂, 1.6 mg/ml Collagenase (type II; Worthington Biochemical, Lakewood, NJ), and 0.04 mg/ml Protease (XIV). The ventricles were cut down, minced, and then gently agitated for 4 min in isolation buffer containing 83 μM CaCl₂, 1 mg/ml collagenase, 0.066 mg/ml Protease, and 1.66% BSA. The cells were filtered through nylon mesh (300 μm), and after having been washed, were resuspended, in succession, in enzyme-free isolation buffer (1% BSA) containing 250 μM and 500 μM CaCl₂. After a final wash, the cardiomyocytes were resuspended at room temperature in the cell media (17.4 mg/ml HEPES-buffered DMEM) supplemented with 10% fetal calf serum (Gibco), 0.01 u/ml insulin, and 44 mM NaCl, at pH 7.4 (adjusted with NaOH).

Simultaneous measurements of I_NCX and [Ca]_i by patch-clamp and confocal microscopy

A patch-clamp method was combined with confocal Ca²⁺ imaging microscopy to enable simultaneous measurement of I_NCX and [Ca]_i recordings in single cardiomyocytes exposed to a sub-maximal but high level (5 mM) of caffeine (24–26) To this end, freshly isolated cardiomyocytes were attached to laminin-coated coverslips and used within 4–5 h. Only quiescent, rod-shaped myocytes with clear striations were used. Experiments were performed at room temperature (20–23°C). The Giga-Ohm seal was attained while superfusing the recording chamber with normal Tyrode containing (in mM): 140 NaCl, 10 D-glucose, 2 CaCl₂, 10 HEPES, 1 MgCl_2, 4 KCl, at pH 7.4 (adjusted with NaOH). The patch pipette (1.8–2.2-M′Ω) was filled with a solution that contained (in mM): 2.5 Na₂ATP, 2.5 MgATP, 2.5 MgCl₂, 30 KCl, 110 K-aspartic acid, 10 HEPES, 0.05 Fluo-4 K₅ (molecular probes), at pH 7.2 (adjusted with KOH). After the whole-cell patch clamp was established, the cells were superfused with normal Tyrode supplemented with 5mM 4-aminopyridine, and 0.1mM BaCl₂. Cells were loaded for 8–10 minutes with fluo-4 through the patch pipette before the experiments were started. Voltage control and recordings were performed with an Axopatch 200A amplifier and Digidata 1322A (Axon Instruments). Currents were low-pass filtered at 1 kHz and sampled at 2 kHz. All currents were normalized to cell capacitance to account for differences in cell size, and data are reported as current densities (pA/pF). Cell capacitance was calculated from the capacity transient elicited by 10-ms hyperpolarization pulses from −90 to −95mV. To elicit I_NCX, the patched cardiomyocyte was held at −80 mV and 5-mm Caffeine (5 mM) was locally applied for 5 sec by a Pneumatic Pico Pump (PV830, World Precision Instruments, Inc.). [Ca]_i imaging was performed with a Zeiss (LSM510) laser scanning confocal microscope equipped with an argon laser and a 63X 1.4 NA oil immersion objective (Zeiss). The scanned line was positioned along the longitudinal axis of the cardiomyocyte, fluo-4 was excited at 488 nm, and the fluorescence above 505 nm was recorded. Line-scan images (time vs. x) were acquired at a sampling rate of 1.92 ms per line. Caffeine application was repeated every 7 min for as long as the cardiomyocytes remained quiescent; the leak currents were stable and under 200 pA. In order to keep the cardiomyocytes undamaged due to the caffeine exposure, 5 mM caffeine concentrations were used in all experiments.

Detection of the Ca²⁺-entry and Ca²⁺-exit modes of NCX1 in single cardiomyocytes

The forward and reverse modes of NCX1 were measured in fluo-3-loaded rat cardiomyocytes, according to established protocols (33,34). In these experiments, the SR function was completely blocked by pretreatment of cardiomyocytes with 10 μM ryanodine and 1 μM thapsigargin. To assess the Ca²⁺ entry via NCX1, we abruptly replaced the extracellular NaCl (137 mM) by the same concentration of TMA for 20 sec. The Ca²⁺ exit mode was induced by replacing 137 mM TMA with 137 mM NaCl. The [Ca²⁺]_i in cardiomyocytes was detected with fluo-3 by using wide-field fluorescence microscopy (22). The prepared cover slips, containing fluo-3-loaded cardiomyocytes, were placed in a recording chamber on the stage of an inverted microscope (Olympus IX-71, Japan). The chamber was perfused with the following (in mM): NaCl 137, KCl 4, MgSO₄ 1.2, Glucose 10, HEPES 10, and CaCl₂ 1.5 (pH 7.4) at 27°C with a temperature controller system (Automatic Temperature Controller, TC-324B, Warner). The excitation light (mercury burner, 100W, Osram) was passed through a set of neutral density and excitation filters (HQ480, Chroma) and focused by a 40X objective (LCPlanF1, Olympus) and field diaphragm (allowing full illumination of single cardiomyocytes). Next, the emitted fluorescence was filtered through the dichroic mirror (Q505LP, Chroma, USA) and emission filter (HQ535, Chroma, USA). The fluorescence signal was recorded with a photon counting device (C&L Instruments, USA), and digitized by a PCI-DAQ card. Each experiment was recorded and stored as a separate file by using Chart v.4.12 software (ADInstruments, Australia). After subtraction of auto fluorescence, the fluo-3-recorded signal was calibrated to [Ca]_i values according to a pseudo-ratio equation. It was assumed that the K_d value for fluo-3 is 650 nM at 27°C and that resting [Ca]_i is 100 nM, before the cells were treated with thapsigargin and ryanodine.

Fluo-4 calibration

In the confocal microscopy experiments, the average fluorescence per pixel in a scanned line was taken as the observed fluorescence. Before a cardiomyocyte was patched, its auto-fluorescence was measured under the same scanning conditions as in the following experiment. The F value was taken after subtracting the auto fluorescence. Since the emission of fluo-4 is negligible in the absence of Ca (F_min ~ 0), the [Ca²⁺]_i was calculated as [Ca²⁺]_i =K_d·(F)/(F_max−F). F_max was measured directly from each cardiomyocyte at the end of the experiment, as previously outlined (24,25).

Miscellaneous assays

Protein was measured by using a modified method of Lowry (35). Fluorescamine was used for detecting the α-amines and peptides (36). Reducing sugars were detected with phenol-sulfuric acid reagent (37). All analytical assays were downscaled for microplate measurements.

Statistics

All data are presented as mean ± SEM. Statistical analysis was done by using the Two Sample t-Test (Origin 7.0, Northampton MA, USA). P values less than 0.05 were considered to be significant.

Results

HPLC procedures for NCX_IF purification

The previous purification procedures of NCX_IF suffered from inadequate pure separation quality, low loading capacity, and unsatisfactory levels of final yield (17,18). A new set of advanced HILIC columns was introduced to overcome these problems. The new purification scheme is based on combining new preparative (Krumasil-Silica) and analytic (ZIC-HILIC and Hypera) HILIC columns with previously explored preparative columns (RP Synergi-Polar and TSK-Amide 80) (Fig. 1). The new purification scheme provides a final yield of ~0.03 mg NCX_IF per kg of starting material, which is at least 3–4 times higher than the highest yields obtained in previous preparations (17–23). A major reason for this improvement is the efficacy of the new HILIC columns, Krumasil-Silica (Fig. 1B), ZIC-HILIC (Fig. 1D), and Hypera (Fig. 1E). Despite this progress, currently available quantities of purified NCX_IF are still not enough for structural studies by 2D NMR.

Fig. 1. HPLC procedures for extensive purification of NCX_IF.

NCX_IF was extracted and partially purified by solvent precipitation, gel-filtration, and ion-exchange procedures, as described in Materials and Methods. Partially purified NCX_IF was further loaded on the HPLC columns in the following order: A. Synergi-Polar (10 ml/min). B. Krumasil-Silica (5 ml/min). C. TSK-gel Amide-80 (5 ml/min). D. ZIC-HILIC (0.5 ml/min). E. Hypera (1.0 ml/min). F. Cyclobond (1 ml/min). All other chromatographic conditions are described in Materials and Methods. Aliquots were removed from collected fractions, lyophilized, and then assayed for inhibition of the Na⁺/Ca²⁺ exchanger by using the standard assay of Na_i-dependent ⁴⁵Ca-uptake in sarcolemma vesicles (see Materials and Methods). The active fractions obtained from 20–50 injections were pooled, lyophilized, and then the concentrated fractions were injected into the next column. Bars indicate the fractions containing the inhibitory activity of NCX_IF.

The Aminex column is an excellent matrix for separating disaccharide, hexose, and pentose structures. Therefore, an attempt was made to compare the retention time of purified NCX_IF to known molecular “witnesses” in order find any possible correlation between the chromatographic behavior of NCX_IF and its chemical structure. The inhibitory activity of NCX_IF is eluted as a single peak, while showing a complete overlap (in time) with optical signals recorded at short UV (Fig. 2). In general, this kind of overlap of optical signals with inhibitory activity refers to the high quality of the purified substance. Eluted fractions of active NCX_IF (Fig. 2) exhibit no chemical reactivity with fluorescamine or phenol-sulfur reagent (see Materials and Methods), thereby suggesting no content of the primary amine or aldehyde groups in the purified preparations of NCX_IF. The chromatographic properties of NCX_IF were compared with “witness” molecules (mono-, di-, tri-, and tetra-saccharides) by using the advanced HILIC-HPLC columns. The retention time of NCX_IF on the ZIC-HILIC (Fig. 1D) and Aminex (Fig. 2) columns is characteristic of disaccharide-sized molecules. Attempts at structural analysis by LC/MS were not successful due to the extremely weak ionization ability of NCX_IF.

Fig. 2. Comparison of the HPLC retention time of purified NCX_IF with the mono- and disaccharide witnesses.

The active fractions of NCX_IF, obtained from the final step of purification (see Fig. 1), were loaded on the Aminex HPX-87c (250 × 4 mm, 9μm) and flashed with water (0.3 ml/min) at 60°C. Next, the fractions were tested for inhibitory activity of the Na⁺/Ca²⁺ exchanger by using the standard assay of Na⁺_i-dependent ⁴⁵Ca-uptake in isolated sarcolemma vesicles (see Materials and Methods). In control samples, the Na⁺/Ca²⁺ exchanger activity was assayed under the same experimental conditions with column effluent. Bars indicate the inhibition of Na⁺/Ca²⁺ exchanger activity. The straight line indicates continuous detection of effluent from the column at 191 nm. The a and b arrows indicate the retention times of glucose and sucrose, respectively (measured on the same column under the identical chromatographic conditions).

Effect of NCX_IF on the Ca²⁺-entry and Ca²⁺-exit modes of the Na⁺/Ca²⁺ exchanger in intact cardiomyocytes

The Ca²⁺-entry and Ca²⁺-exit modes of NCX1 were monitored in intact cardiomyocytes under conditions at which the SR Ca²⁺ fluxes were completely blocked by pretreatment of cardiomyocytes with ryanodine and thapsigargin (33,34). To assess the Ca²⁺-entry via NCX1, extracellular Na⁺ was abruptly removed (replaced by TMA) for 20 sec and Ca²⁺ entry was monitored by measuring changes in [Ca²⁺]_i using fluo-3. Figure 3A shows that Na⁺ removal results in the elevation of [Ca²⁺]_i (net Ca²⁺-entry). Following the removal of extracellular Na⁺, the return of extracellular [Na]_o to normal levels resulted in decreased [Ca²⁺]_i and reflects the net exit of Ca²⁺ via the Na⁺/Ca²⁺ exchanger. Two consecutive series of Na⁺-abrupt episodes in the same cardiomyocyte (separated by 5 minutes) produced comparable changes in the amplitude and rates of the [Ca²⁺]_i (Fig. 3A). Therefore, the inhibitory effects of NCX_IF can be reliably evaluated in the same cardiomyocyte during the second Na⁺-abrupt episode (the first Na⁺-abrupt can be used as a control). The bidirectional Ca²⁺-fluxes through NCX1 were examined by transient superfusion of cadiomyocytes with extracellular NCX_IF (50 units/ml for 5 min). NCX_IF significantly reduced the rate and magnitude of the changes in [Ca²⁺]_i (Fig. 3B). NCX_IF results in a slowing of the rate of Ca²⁺-entry and of the Ca²⁺-exit (Fig. 3C). Moreover, NCX_IF reduces the amplitude of [Ca²⁺]_i elevation (to 15.0 ± 4.8% of the control, p < 0.001) and decreases the rate of Ca²⁺-entry (11.9 ± 7.2% of the control, p < 0.001) and of Ca²⁺-exit (5.6 ± 2.2% of the control, p < 0.0001) (Fig. 3D). Therefore, NCX_IF can effectively inhibit both the Ca²⁺-entry and Ca²⁺-exit modes of NCX1 in intact cardiomyocytes, as expected, based on transport thermodynamics (50,51). These properties of NCX_IF, which can comparably alter both the Ca²⁺-entry and Ca2+-exit modes of NCX1, seem to be very different from the properties of the synthetic NCX-blocker, KB-R7943, which exhibits a preferential selectivity for inhibiting the Ca²⁺-entry mode in a single cardiomyocyte system (33,38).

Fig. 3. NCX_IF–dependent inhibition of Ca²⁺-entry and Ca²⁺-exit modes of the Na⁺/Ca²⁺ exchanger in intact cardiomyocytes.

The Ca²⁺-entry and Ca²⁺-exit modes of NCX were monitored in intact cardiomyocytes by perfusing the recording chamber with 0 mM Na⁺ medium for 20 seconds (abrupt-0Na). Then the chamber was perfused with normal medium, as indicated, and subsequently, the second session of the Ca²⁺-entry and Ca²⁺-exit reaction was generated in the same cardiomyocyte by switching to the 0 mM Na⁺ perfusion medium for 20 sec. A. Two consecutive sections of Na⁺-abrupt reaction in the absence of NCX_IF. B. After the first abrupt-0 Na⁺, the recording chamber was perfused with NCX_IF (50 units/ml) for 5 min and the Ca²⁺-entry and Ca²⁺-exit reactions were monitored by performing a second abrupt-0Na in the same cardiomyocyte. C. The data representing the typical effects of NCX_IF on Ca-entry and Ca²⁺-exit modes (see panel B) were normalized. The trace-a represents the control before exposure to NCX_IF, and the trace-b depicts the recordings obtained after exposure of the same cardiomyocyte to NCX_IF (50 units/ml) for 5 min. The Ca²⁺-exit rate constant (k) was calculated from the signal obtained by abrupt Na⁺ removal and is presented as the percentage of the control rate constant. The control rate constant (100%) represents k = 0.24 ± 0.047 s⁻¹. D. Maximal rates (first derivative) of each reaction are presented as the percentage of control. The control rate (100%) represents 157.1 ± 37.2 nM [Ca²⁺]_i·s⁻¹ for the Ca²⁺-entry phase and 230.7 ± 54.8 nM [Ca²⁺]_i·s⁻¹ for the Ca²⁺-exit phase. The amplitude of the Ca²⁺-entry mode was measured in the presence and absence of NCX_IF under the conditions described in panel B. The control amplitude (100%) represents the Δ[Ca²⁺]_i value of 865.2 ± 200.9 nM measured in the absence of NCX_IF. Results are represented as the mean ± SEM. The n values represent the data collected from cardiomyocytes isolated from different animals.

Although the Na⁺-abrupt protocol enables to assess the Ca²⁺-entry and Ca²⁺-exit modes of NCX1 in intact single cell, this method cannot resolve whether NCX_IF interacts with the extracellular or intracellular domain. Therefore, the application of complementary approaches was necessary to resolve the side-ness of NCX_IF action in intact cardiomyocytes (see below).

Direct loading of NCX_IF into the cytoplasm inhibits the I_NCX and Ca²⁺-extrusion in intact cardiomyocytes

To further investigate the sidedness of NCX_IF interactions with NCX1, we carried out patch clamp experiments. Briefly, NCX_IF was applied to the interior of intact cardiomyocytes through the patch pipette. In these experiments, I_NCX was measured simultaneously with [Ca²⁺]_i using a combination of patch clamp and confocal microscopy. Under control conditions, application of caffeine (for 5sec) activates the release of SR Ca²⁺ and elevates [Ca²⁺]_i, consequently enhancing Ca²⁺-activated inward I_NCX current (24–26). In patched cardiomyocytes, the cytoplasm was equilibrated with pipette solution containing either no NCX_IF (gthe control group) or 50 units/ml NCX_IF (the NCX_IF group). Figure 4A shows a typical recording obtained from either a control or from a cardiomyocyte loaded with 50 units/ml NCX_IF. The [Ca²⁺]_i transient, induced by caffeine, was significantly higher in the NCX_IF-loaded myocytes (1.66 ± 0.35 μM, n=16) than in the control group without NCX_IF (0.34 ± 0.1 μM, n=10; p < 0.01), as shown in Fig. 4. The peak I_NCX value was insignificantly lower in NCX_IF-loaded cardiomyocytes, 0.61 ± 0.05 pA/pF than in the control group, 0.79 ± 0.20 pA/pF (data not shown). As shown in Fig. 4B, the I_NCX per Δ[Ca²⁺]_i is much greater in controls than in the presence of NCX_IF. Therefore, a given concentration of cytosolic NCX_IF significantly blocks NCX1 in a single cardiomyocyte system.

Fig. 4. Simultaneous recordings of I_NCX and [Ca²⁺]_i in patched cardiomyocytes after transient exposure to caffeine.

Adult rat cardiomyocytes were held at −80 mV and loaded with the Ca²⁺ fluorescent dye fluo-4 through the patch pipette. Ca²⁺ transients were induced by application of 5 mM caffeine (Caff) on a quiescent cell (see Materials and Methods). A. Typical recordings of I_NCX and [Ca²⁺]_i from a control (left) or NCX_IF-loaded cell (right). Representative line scan image (upper panel), corresponding [Ca²⁺]_i values (middle trace), and the simultaneous recorded I_NCX (lower trace). B. Upper chart, mean peak I_NCX normalized to Δ[Ca²⁺]_i in control (n = 10) and NCX_IF-loaded cells (n = 16, *P < 0.01). Lower chart, the mean values of Ca2+ transient amplitudes in control (n = 10) and NCX_IF-loaded cells (n = 16, *P < 0.01). The n represents the number of independent experiments.

Effect of NCX_IF on the I_NCX/[Ca]_i relationship

The dependence of I_NCX on [Ca²⁺]_i can be measured dynamically during the time when the application of caffeine increases [Ca²⁺]_i and its removal leads to the fall of [Ca²⁺]_i. The data analysis was established by plotting I_NCX during the elevation of [Ca²⁺]_i, as shown in Fig. 5 and its fall, as shown in Fig. 6. Intracellular NCX_IF reduces I_NCX at any [Ca²⁺]_i value, consistent with the NCX1 block. The data fitted by linear regression clearly indicates that the slope of I_NCX vs. [Ca²⁺]_i is less steep (~6-fold) in NCX_IF-loaded cardiomyocytes, than in control cardiomyocytes in the absence of NCX_IF. The cytosolic NCX_IF has very similar effects on the I_NCX/[Ca²⁺]_i slope during the raising (6.0 ± 0.02-fold change in the slope, Fig. 5B) or the decay phase (5.94 ± 0.02-fold change in the slope, Fig. 6B) of caffeine-induced [Ca²⁺]-transient. The present data clearly indicate that NCX_IF inhibits both the NCX-mediated inward current and [Ca²⁺]-extrusion from the cell in a comparable way.

Fig. 5. Effect of directly loading NCX_IF into the cytoplasm on I_NCX during [Ca²⁺]_i rise.

Cardiomyocytes were transiently exposed to 5 mM caffeine as described in Materials and Methods. The caffeine-evoked I_NCX and [Ca²⁺]_i values were measured simultaneously and then plotted as indicated. The results represent the rising phase of the caffeine [Ca²⁺]_i transient (starting from the beginning of [Ca²⁺]_i upstroke until it peaked) A. The average values of I_NCX, obtained at a given [Ca]_i value, were plotted for NCX_IF-loaded or control (-NCX_IF) cardiomyocytes as indicated. B. The slopes of I_NCX vs. [Ca²⁺]_i were calculated by fitting the data for each individual experiment (n = 10 for control cells and n = 15 for NCX_IF-loaded cells, P < 0.001). C. Different ranges of [Ca²⁺]_i values (observed in each specific experiment) were plotted vs. appropriate levels of I_NCX, as described in panel A.

Fig. 6. Effect of directly loading NCX_IF into the cytoplasm on I_NCX during the [Ca²⁺]_i decline.

The caffeine-evoked I_NCX and [Ca²⁺]_i were simultaneously recorded and plotted as described in Figs. 4 and 5. The data were obtained by taking the decline phase of the caffeine-induced Ca²⁺ transient (starting from the I_NCX peak until the baseline was reached). A. The average I_NCX values (for given [Ca²⁺]_i levels) were plotted for NCX_IF-loaded or control cardiomyocytes (-NCX_IF). B. The I_NCX/[Ca²⁺]_i curves were fitted by linear regression and the obtained slopes were averaged as indicated (n = 9 for control cells and n = 12 NCX_IF-loaded cells, P < 0.001). The n values represent the number of independent experiments. C. The observed ranges of [Ca²⁺]_i values (obtained in each specific experiment) were plotted vs. I_NCX, as described in Fig. 5.

At fixed doses of NCX_IF, the fraction of I_NCX inhibition is independent of [Ca²⁺]_i, at least in the tested range of [Ca²⁺]_i = 30–2500 μM (Figs. 5 and 6). This observation provides independent evidence that the degree of inhibition is determined by NCX_IF dose and not by [Ca²⁺]_i. As a precaution, one has to take into account that under the experimental conditions tested here, the regulatory sites of NCX1 are fully saturated by cytosolic Ca²⁺. Therefore, a more extensive investigation is required for testing the effects of NCX_IF on Ca²⁺ interaction with the CBD1 and CBD2 domains.

Discussion

Purification procedures developed in our laboratory (17,18) permitted primary characterization of NCX_IF properties (17–23), although these preparations still suffer from weak chromatographic resolution, inadequate pure loading capacity, and low yield. Since milligram quantities of purified NCX_IF are required for structural studies, here we describe further improvements in the purification procedures. Although our previous work provided accumulating evidence for the NCX_IF-induced inhibition of NCX1 activities in vitro (17–19) and for the modulation of muscle contractility (17,20–23), there was no evidence for inhibition of NCX1 in the single-cell system. Therefore, the present work represents the first attempt at measuring the direct actions of NCX_IF on the Na⁺/Ca²⁺ exchanger activities in a single-cell system by performing simultaneous measurements of I_NCX and cytosolic [Ca²⁺]_i. As predicted from thermodynamic considerations, NCX_IF blocks equally well both Ca²⁺ extrusion and Ca²⁺ entry modes of NCX1. It appears that NCX_IF acts effectively from the cytosolic side, exhibiting an inhibitory potency that is comparable with the inhibitory potency observed previously with in vitro systems of isolated sarcolemma vesicles (17–20). Some issues related to the characteristics of NCX_IF action are discussed below.

Characteristic properties of purified NCX_IF preparations

A several-fold increase in the yield of purified NCX_IF preparations were obtained by introducing new HILIC technologies at various stages of the purification scheme (Fig. 1). Despite a significant improvement in preparative procedures, the available amounts of purified NCX_IF are still not enough to elucidate the NCX_IF structure by 2D NMR. To obtain sufficient (milligram) quantities of NCX_IF for structural analysis by NMR, one has to scale-up the existing purification procedures at least 5–10-fold (with starting material of 40–50-kg ventricles). Unfortunately, these near-industrial procedures for purification could not be handled in the present work due to the lack of available resources.

Another principal problem in molecular identification of NCX_IF is that the de novo identification of unknown substances by MS-techniques depends on the chemical nature of the investigated substance (39,40). For example, small charged molecules (e.g., such as peptides) can be successfully analyzed even at 10⁻⁹ M, whereas the analysis of hydrophilic, uncharged molecules (e.g., such as small polyols) could be a challenge even at 10⁻³ M (39–41). If NCX_IF is a hydroxyl-rich compound (as we suspect) this would explain why it is so difficult to resolve its structure by MS techniques. Consistent with our previous findings (18), NCX_IF exhibits very weak absorption only in the short UV (190–200 nm) range (Fig. 2), thereby suggesting that neither aromatic rings (nucleotides, phenols, indols, etc.) nor conjugated chemical groups are present. In agreement with previous analyses (17,18), purified NCX_IF (Fig. 2) does not react with fluorescamine or phenol-sulfur reagents. Therefore, there is no evidence that NCX_IF contains primary amine or aldehide groups. In potential, the covalent modification of putative hydroxyl groups of NCX_IF may enhance the MS and/or NMR signals by an order of magnitude but unfortunately, due to the extremely low solubility of NCX_IF in organic solvents (even in pyridine), the chemical modification of hydroxyl group(s) is also not a trivial task.

NCX_IF inhibits both the Ca²⁺-entry and Ca²⁺-exit modes of NCX1 in intact cardiomyocytes

Although the synthetic NCX blocker KB-R7943 has been reported to preferentially block the Ca²⁺ entry mode of NCX1 in isolated cardiac myocytes (33,38), this explanation is unsatisfactory (42), given the requirements of thermodynamics (50,51). This may be explained by the fact that experimental conditions are not sufficiently comparable in the KB-R7943 studies. Here, the conditions for examining both modes of NCX1 action were nearly identical when NCX_IF was used to block NCX1. NCX_IF blocks equally well for both the Ca²⁺ entry and Ca²⁺ exit modes of NCX1, as shown in Figure 3. Moreover, NCX_IF alters not only the rates of Ca²⁺ entry/exit modes but also the rate constants of both modes (Fig. 3B). These data suggest that NCX_IF may alter the turnover rate constant of the exchange cycle by affecting the rate-limiting step of net ion transport. This is in a good agreement with previous results (21), suggesting that NCX_IF affects the rate-limiting step of ion-translocation without interfering with the ion-binding at the transport site(s) of NCX1.

Cytosolic NCX_IF inhibits the NCX1-mediated ion currents at any given [Ca²⁺]_i

NCX_IF was directly loaded into the cytoplasm of patched cardiomyoctes in order to investigate the sidedness of NCX_IF action. Inward I_NCX was activated by elevation of [Ca²⁺]_i following transient exposure to caffeine. Simultaneous monitoring of I_NCX and [Ca²⁺]_i has shown that intracellular NCX_IF reduces I_NCX at any given [Ca²⁺]_i (Figs. 5 and 6). At fixed doses of NCX_IF, the fraction of I_NCX inhibition is independent of [Ca²⁺]_i (Figs. 5 and 6), suggesting that the extent of inhibition is controlled by NCX_IF dose and not by cytosolic Ca²⁺ level. The significance of the present findings is that NCX_IF may interact with some specific cytosolic domain to inhibit NCX1. Notably, the regulatory sites of NCX1 are fully saturated by cytosolic Ca²⁺ under the experimental conditions tested here (Figs. 5 and 6). Therefore, the future testing of NCX_IF for its capacity to affect the Ca²⁺ interactions with the CBD1 and CBD2 domains is highly encouraging.

The present experiments clearly demonstrate that NCX_IF can be effectively “entrapped” inside the cell by direct loading of NCX_IF into the cytoplasm of patched cardiomyocyte (Figs. 4–6). Although unproven, it seems plausible that an extracellular site of action is unlikely since the extracellular concentration would be significantly reduced by dilution following intracellular administration. In contrast, the extracellular application of NCX_IF would tend to load the cell with the inhibitor if it should leak (transport) in, as we suspect it does. This would then explain the efficacy of extracellular application of NCX_IF in the present experiments describing the bidirectional inhibition of NCX1 (Figure 3). In the absence of NCX_IF breakdown, extracellular application of NCX_IF should lead to steady-state levels of intracellular NCX_IF approaching the extracellular concentration.

Brief comparison of NCX_IF with existing NCX-blockers

A considerable amount of work has been devoted to the development of peptide (27–32, 42–44) and organic (38,45–47) NCX-blockers; however, all of them have some specific, but principal drawbacks. For example, the synthetic blockers (KB-R7943, SEA-0400, and SN-6) inhibit not only NCX, but also the other ion-transport systems (38,48). Another principal drawback is the enigmatic putative preferential inhibition of the Ca²⁺-entry mode, which would appear to limit its pharmacological significance and utility (33,49). In contrast to the organic NCX-blockers, XIP and FRCRCFa can effectively inhibit both the inward and outward currents of NCX by interacting with the cytosolic domain (42–44), but these blockers are inactive when added extracellularly due to their impermeability to the cell membrane.

In contrast to the synthetic blocker, KB-R7943, NCX_IF results in a comparable inhibition of both the Ca²⁺-entry and Ca²⁺-exit modes in intact cardiomyocytes (Fig. 3). Although KB-R7943 and NCX_IF are both active when added outside the cell, the striking difference is that KB-R7943 interacts with the extracellular domain, whereas NCX_IF might approach the cytosolic site in order to inhibit NCX1 (Figs 5 and 6). More extensive research is required for comparing NCX_IF with the more recently developed NCX-inhibitors, SEA0400 and SN-6.

Abbreviations

Arsenazo III

2,2′-(1,8-dihydroxy-3,6-disulfonaphthylene-2,7-bisazo)bisbenzenearsonic acid

Fluo-3

N-[2-[2[bis(carboxymethyl)amino]-5-(2,7dichloro-6-hydroxy-3-oxy-3H-xanthen-9-yl) henoxy]ethoxy]-4 methylphenyl]-N-(carboxymethyl)glycine

Fluorescamine

4-Phenylspiro[furan-2(3H),1′-phthalan] 3,3-dione

FRCRCFa

the S-S bond cyclic hexapeptide Phe-Arg-Cys-Arg-Cys-Phe-NH₂

XIP

the inhibitory peptide containing twenty amino acids

LC/MS

Liquid chromatography coupled with mass-spectroscopy

LC/MS/MS

Liquid chromatography coupled with MS/MS

Mops

3-(N-morpholino)propanesulfonic acid

Tris

tris(hydroxymethyl)aminomethane

PMSF

Phenylmethanesulfonyl fluoride

TMA

tetramethylammonium