Structure-function relationships in the regulation of energy transfer between mitochondria and ATPases in cardiac cells

Abstract

The present study discusses the role of structural organization of cardiac cells in determining the mechanisms of regulation of oxidative phosphorylation and interaction between mitochondria and ATPases. In permeabilized adult cardiomyocytes, the apparent K_m (Michaelis-Menten constant) for ADP in the regulation of respiration is far higher than in mitochondria isolated from the myocardium. Respiration of mitochondria in permeabilized cardiomyocytes is effectively activated by endogenous ADP produced by ATPases from exogenous ATP, and the activation of respiration is associated with a decrease in the apparent K_m for ATP in the regulation of ATPase activity compared with this parameter in the absence of oxidative phosphorylation. It has also been shown that a large fraction of the endogenous ADP stimulating respiration remains inaccessible for the exogenous ADP trapping system, consisting of pyruvate kinase and phosphoenolpyruvate, unless the mitochondrial structures are modified by controlled proteolysis. These data point to the endogenous cycling of adenine nucleotides between mitochondria and ATPases. Accordingly, the current hypothesis is that in cardiac cells, mitochondria and ATPases are compartmentalized into functional complexes (ie, intracellular energetic units [ICEUs]), which appear to represent a basic pattern of organization of energy metabolism in these cells. Within the ICEUs, the mitochondria and ATPases interact via different routes: creatine kinase-mediated phosphoryltransfer; adenylate kinase-mediated phosphoryltransfer; and direct ATP and ADP channelling. The function of ICEUs changes not only after selective proteolysis, but also during contraction of cardiomyocytes caused by an increase in cytosolic Ca²⁺ concentration up to micromolar levels. In these conditions, the apparent K_m for exogenous ADP and ATP in the regulation of respiration markedly decreases, and more ADP becomes available for the exogenous pyruvate kinase-phosphoenolpyruvate system, which indicates altered barrier functions of the ICEUs. Thus, structural changes transmitted from the contractile apparatus to mitochondria clearly participate in the regulation of mitochondrial function due to alterations in localized restriction of the diffusion of adenine nucleotides. The importance of strict structural organization in cardiac cells emerged drastically from experiments in which the regulation of mitochondrial respiration was assessed in a novel cardiac cell line, that is, beating and nonbeating HL-1 cells. In these cells, the mitochondrial arrangement is irregular and dynamic, whereas the sarcomeric structures are either absent (in nonbeating HL-1 cells) or only rarely present (in beating HL-1 cells). In parallel, the apparent K_m for exogenous ADP in the regulation of respiration was much lower than that in permeabilized primary cardiomyocytes, and trypsin treatment exerted no impact on the low K_m value for ADP, in contrast to adult cardiomyocytes where it caused a marked decrease in this parameter. The HL-1 cells were also characterized by the absence of direct exchange of adenine nucleotides. The results further support the concept that the ICEUs in adult cardiomyocytes are products of complex structural organization developed to create the most optimal conditions for effective energy transfer and feedback between mitochondria and ATPases.

In saponin-treated permeabilized (skinned) cardiac cells, the apparent K_m (Michaelis-Menten constant) for ADP in the control of mitochondrial respiration is very high, in the range of 200 μM to 300 μM, in contrast to that in skinned glycolytic fibres or isolated mitochondria (10 μM to 20 μM) (1,2). It is also known that ADP produced in skinned cardiac fibres by creatine kinase (CK) and adenylate kinase (AK), as well as by ATPases, stimulates mitochondria much more effectively than ADP exogenously added to the medium (3,4), and that ATP generated by mitochondria is a preferable energy source for the sarcoplasmic reticulum Ca²⁺ pump compared with exogenous ATP (5). Such endogenous cycling of adenine nucleotides between mitochondria and ATPases cannot be abrogated by the pyruvate kinase (PK)-phosphoenolpyruvate (PEP) system, capable of eliminating the cytoplasmic ADP, unless the intracellular structures are disintegrated by mild trypsinization (3,4,6). Based on these results, we have hypothesized that in myocardial cells, mitochondria and ATPases function as tightly coupled complexes, termed intracellular energetic units (ICEUs) (3,4), which may represent a basic pattern of organization of energy metabolism in cardiac cells. It is very likely that the cytoskeletal proteins (eg, desmin, dystrophin and others) are the main factors responsible for organizing the ICEUs (3,4). By forming barriers for diffusion of adenine nucleotides, these proteins compartmentalize a fraction of cellular adenine nucleotides that does not easily equilibrate with the adenine nucleotide pool in the cytoplasmic bulk phase. This fraction of nucleotides is thought to participate in the mechanisms of intracellular energy transduction, such as CK- and AK-mediated phosphotransfer networks, and direct channelling of ATP and ADP between mitochondria and ATPases (3,4,6,7).

If the ICEUs indeed play a central role in the organization of energy metabolism in cardiac cells, alterations in their structure and function should interfere with the regulation of processes of energy production and utilization, and through these processes, with cardiac contractile activity. In support of this, it has been recently shown that the apparent affinity of mitochondrial respiration to exogenously added ADP increases with sarcomere shortening in response to an elevated cytoplasmic free Ca²⁺ concentration in skinned cardiac cells (8). An increase in the apparent affinity of mitochondria to ADP is also observed after mild treatment of cardiac cells with proteolytic enzymes (eg, trypsin) that disintegrate the cell structure, presumably affecting the cytoskeletal proteins (4). These observations directly indicate the existence of structural and functional links between mitochondria and sarcomeres, which are strong enough to modulate the energy transfer feedback signalling between ATPases and mitochondria.

The present study further addresses the role of ICEUs in the intracellular organization of energy metabolism in cardiac cells by comparing the regulation of the functions of mitochondria and ATPases in adult myocardial cells, and in subtypes of a novel cultured cell line (HL-1) characterized by a differentiated cardiac phenotype but distinct contractile activities depending on growing conditions. The subtype of beating (B) HL-1 cells was developed by Claycomb and coworkers (9,10). Another subtype of the cells (nonbeating [NB] HL-1 cells) was derived from B HL-1 cells by growing them with a different serum (Gibco fetal bovine serum, Gibco France) for five weeks, which led to cells devoid of beating properties, but capable of proliferating in the normal Claycomb medium (11). The absence of spontaneous beating in these cells is attributed to the lack of pacemaker current, intracellular Ca²⁺ oscillations and sarcomere structures (11–13). At the same time, the mitochondria are randomly arranged and they dynamically change their positions inside the cytoplasm (11–13). Thus, these cells represent valuable objects for studying the importance of the intracellular structural organization of energy metabolism for determining the functional properties of mitochondria and ATPases in cardiac cells.

METHODS

Reagents and solutions

The reagents were purchased from Sigma (USA), Serva (Germany), Roche (France) and Fluka (Switzerland). Solution B contained: CaK₂EGTA 2.77 mM, K₂EGTA 7.23 mM, MgCl₂ 1.38 mM, dithiothreitol (DTT) 0.5 mM, potassium 2-(N-morpholino)-ethanesulfonic acid 100 mM, imidazole 20 mM, taurine 20 mM, K₂HPO₄ 3 mM and bovine serum albumin (BSA) 5 mg/mL, pH 7.1 at 25°C, with 0.1 μM of free Ca²⁺. The free Ca²⁺ concentrations in this and other solutions were calculated as described previously (8). Solution C contained: potassium HEPES 50 mM, MgCl₂ 10 mM, ATP 10 mM, DTT 0.5 mM, taurine 20 mM, potassium 2-(N-morpholino)-ethanesulfonic acid 80 mM (pH 7.1). Solution D contained: KCl 800 mM, potassium HEPES 50 mM, MgCl₂ 10 mM, ATP 10 mM, DTT 0.5 mM and taurine 20 mM (pH 7.1). The stock solutions used to add ATP or ADP into solutions contained MgCl₂ (0.8 mol/mol ATP or 0.6 mol/mol ADP) to avoid large changes in intracellular free Mg²⁺ concentration. Mitomed solution contained: sucrose 110 mM, potassium lactobionate 60 mM, EGTA 0.5 mM, MgCl₂ 3 mM, DTT 0.5 mM, taurine 20 mM, KH₂PO₄ 3 mM, potassium HEPES 20 mM, pH 7.1, glutamate 5 mM, malate 2 mM, with essentially fatty acid-free BSA 2 mg/mL.

Animals

Adult male Wistar rats weighing 200 g to 350 g were used in the experiments. The investigation conformed to the Guide for the Care and Use of Laboratory Animals (14).

Preparation of skinned muscle fibres

The saponin-skinned fibres were prepared from endocardial parts of left ventricles according to the method described in detail previously (15).

Preparation of ghost fibres

The ghost fibres deficient in myosin were prepared according to Saks et al (16). Briefly, the saponin-skinned fibres were washed in solution C for 10 min. Thereafter, the fibres were incubated in solution D for 30 min to extract myosin, washed twice in solution B for 5 min to remove the excess KCl, and kept in this solution until assessed.

Cardiomyocyte isolation and cell culturing

Adult cardiomyocytes were isolated from rat heart with collagenase as described previously (17). The B HL-1 cell line was isolated from tumoural atrial cardiac myocytes from transgenic mice grown in a specific medium containing, in particular, fetal bovine serum issued from one provider (9) and transferred into the NB HL-1 cell line as described previously (13).

Trypsin treatment

Isolated cardiomyocytes were permeabilized by incubation with 20 μg/mL saponin for 10 min and washed four or five times with Mitomed solution. Permeabilized cardiomyocytes were incubated with different trypsin concentrations (titrated trypsin concentrations of 0.025 μM to 1.0 μM) for 5 min at 4°C. When appropriate, trypsin activity was inhibited by the addition of 2 mM trypsin inhibitor and 2 mg/mL BSA.

Respirometric investigations

The rates of oxygen uptake were recorded using a high-resolution respirometer (Oroboros Oxygraph, Paar KG, Austria) equipped with a Clark oxygen sensor, and analyzed with DATGRAPH Analysis software (OROBOROS, Austria). The analysis of respiration in skinned cardiac fibres was accomplished (as previously described [3,8,15,18]) in solution B (1.5 mL) at 25°C in the presence of 5 mM or 10 mM glutamate and 2 mM malate, and different concentrations of free Ca²⁺ and ADP or ATP. To analyze the respiration of mitochondria in B HL-1 and NB HL-1 cells, the cells were detached by trypsinization, and the cell suspension was washed and centrifuged three times at 4°C with Mitomed solution and 2 mg/mL BSA, and stocked in this solution at 4°C. To assess the kinetics of respiration regulation, the cells were incubated in the oxygraph chamber by the addition of 25 μg/mL to 50 μg/mL saponin for 15 min, for permeabilization, in Mitomed solution (15,19) for both cell types, and supplemented with 4 U/mL hexokinase and 12 mM glucose.

Preparation of skinned muscle cells for electron microscopy

After having registered the respiration rates, the skinned fibres were removed from the oxygraph, slightly blotted by filter paper, fixed in 0.25% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) and postfixed in 1% OsO₄ in the same buffer. After dehydration with ethanol and acetone, the specimens were embedded in Epon 812 (Shell Chemical Co, USA). The sections were stained with uranyl acetate followed by lead citrate, and examined using a Tecnai 10 electron microscope (FEI Company, Netherlands) at 100 kV. For ultrastructure investigations, the longitudinally oriented cardiomyocytes were selected and, from five to seven separate images corresponding to the same experimental conditions, the lengths of 92 to 390 sarcomeres were measured.

Confocal imaging of mitochondria in living cells

To detect the mitochondrial functional state at the level of the single mitochondrion, images were acquired and analyzed by fluorescent confocal microscopy. For mitochondrial imaging (localization) studies, cells were loaded with MitoTracker Green (Molecular Probes, USA) (0.2 μM) or MitoTracker Deep Red (Invitrogen, France) (85 nM). Cells were placed in Lab-Tek chambered coverglass (Nalge Nunc, USA) and incubated for at least 2 h at 4°C for cardiomyocytes and 15 min to 2 h at 36°C for HL-1 cells before analysis (chamber volume 0.3 mL; 10×10³ to 20×10³ cells per chamber). The digital images of MitoTracker Green fluorescence were acquired with an inverted confocal microscope (Leica DM IRE2, Leica Microsystems, Germany) with a 63× magnfication water immersion lens. The MitoTracker Green fluorescence was excited with the 488 nm line of an argon laser, using 510 nm to 550 nm for emission. Tetramethylrhodamine methylester fluorescence was measured using 543 nm for excitation (helium-neon laser) and greater than 580 nm for emission.

Statistical analysis

Student’s t test was used to compare data in different muscle groups. The means ± SEM or SD are presented where indicated.

RESULTS AND DISCUSSION

Figure 1 demonstrates the remarkably regular, crystal-like (20) arrangement of mitochondria in adult cardiomyocytes (Figure 1A), because the mitochondrial positions adjacent to neighbouring sarcomeres are rather precisely fixed, which seems to be specific for intermyofibrillar mitochondria in cardiac cells (Figure 1B) (21). Nozaki et al (21) have demonstrated a linear relationship between the length changes of sarcomeres and adjacent mitochondria caused by stretching and increases in free Ca²⁺ concentration under normoxic and hypoxic conditions. Their findings indicate that in normal cardiomyocytes, strong structural links exist between mitochondria and sarcomeric proteins that persist even in hypoxic cells. It is very likely that the perfect arrangement of mitochondria is achieved due to specific organization of the cytoskeletal microtubular network, particularly due to desmin filaments that connect the mitochondria with the Z-disk of the sarcomere. It appears that the cytoskeletal network not only fixes the position of mitochondria but also transfers the forces onto mitochondrial membranes, thereby causing their compression and stretching, in correspondence to the altered length of sarcomeres such as during the contraction-relaxation cycle (8,21–23). Most interestingly, tight structural organization of mitochondria and contractile proteins within the cardiac cells directly determines the functional properties of mitochondria. Indeed, numerous studies have shown that the value of the apparent K_m for exogenous ADP in the regulation of mitochondrial respiration in permeabilized cardiomyocytes is much higher than that for isolated mitochondria (1,2), but it decreases to the levels characteristic of isolated mitochondria after partial disintegration of the cardiomyocyte’s structure either by trypsin or knockout of specific cytoskeletal proteins (4,24). These phenomena have been quantitatively explained by the existence of structural and functional complexes between mitochondria and other ATPase-carrying structures, referred to as the ICEUs, which create local restrictions of the diffusion of ADP and ATP at the level of the mitochondrial outer membrane and/or cytoskeletal barriers (3,4,25,26). Owing to these diffusion limitations, the ICEUs isolate a fraction of adenine nucleotides from their bulk phase in the rest of the cytoplasm, to be used for energy transfer via different main routes: CK-mediated phosphoryltransfer; AK-mediated phosphoryltransfer; and direct channelling of ATP from mitochondria to ATPases, and ADP from ATPases to mitochondria (3–5,7).

Figure 1).

Confocal (A, C and D) and electron microscopical (B) visualization of mitochondria in adult rat cardiomyocytes (A and B), and cultured nonbeating HL-1 cells (C) and beating HL-1 cells (D). A Mitochondria were visualized using MitoTracker Red CMXRos (Molecular Probes, USA) (0.1 μM) in the presence of 5 mM glutamate and 2 mM malate. B Mitochondrial arrangement in permeabilized rat cardiomyocyte within the skinned fibre preparation in solution B, in the presence of 20 mM ATP and the pyruvate kinase-phosphoenolpyruvate system (20 U/mL and 5 mM, respectively), without respiratory substrates and Ca²⁺. C Mitochondrial arrangement and dynamics in nonbeating HL-1 cells in the presence of MitoTracker Deep Red (Invitrogen, France) in Tyrode’s medium. D Mitochondrial arrangement and dynamics in normally cultured beating HL-1 cells visualized using MitoTracker Deep Red. Bars represent a distance of 10 μm

Characterization of direct channelling by means of measuring the fluxes of ADP produced by CaMgATPases in permeabilized cardiomyocytes with the coupled PK-PEP-lactate dehydrogenase (LDH) system led to the conclusion that this process markedly enhances the apparent affinity of CaMgATPases for its substrate – ATP (8). For example, at a diastolic free Ca²⁺ concentration of 0.1 μM, the apparent K_m for ATP was 2.33±0.41 mM in conditions absent of oxidative phosphorylation, whereas switching on this process by the addition of respiratory substrates (glutamate and malate) resulted in a K_m of 1.15±0.26 mM (P<0.05). In parallel to that change, initiation of oxidative phosphorylation diminished the ADP flux (by more than twofold) from the ATPases available to the PK-PEP system as the mitochondria took over phosphorylation of ADP. Thus, the organization of intracellular energy metabolism into ICEUs confers the effective regulative mechanisms of ATPases to cardiac cells, via direct channelling of ATP from ATPases to mitochondria and, vice versa, by removal of ADP produced by ATPases back to the mitochondria. As a result, the local concentrations of ATP increased and those of ADP decreased near ATPases, which explains the decreased K_m for ATP (8).

Another set of data indicates that mitochondrial oxidative phosphorylation and direct ATP/ADP channelling are directly controlled by structural modifications of ICEUs, in relation to changes in sarcomere length imposed by increasing cytosolic Ca²⁺ concentrations. First, it was found that increasing the cytosolic free Ca²⁺ concentration from its diastolic level (0.1 μM) up to 4 μM was associated with a decrease in the apparent K_m for ADP from 320±20 μM to 17±3 μM in the regulation of respiration, and a similar decrease was also observed for the apparent K_m for ATP in the same process (8). These changes were related to the activation of the contractile system rather than to the upregulation of intramitochondrial Ca²⁺-sensitive enzymes, because excess Ca²⁺ exerted no effect on the K_m after the extraction of myosin from the cells (8). Second, an increase in free cytosolic Ca²⁺ concentration was shown to attenuate the direct channelling of ATP/ADP between mitochondria and ATPases (8). Figure 2 shows that this suppressive effect of Ca²⁺ was underlaid by marked shortening of the sarcomeres at a free Ca²⁺ concentration of 2 μM. It is likely that strong contraction caused such structural alterations in the diffusion barriers within the ICEUs, which allowed ADP to escape these units and thereby limited the extent of mitochondrial phosphorylation of ADP.

Figure 2).

Relationships between the mitochondrial ADP flux (means ± SEM) and mean sarcomere length altered by changing the free Ca²⁺ concentration (2 μM at point A, 0.1 μM at point B and 0 μM at point C) in permeabilized rat cardiac cells incubated in solution B, with 20 mM ATP and the pyruvate kinase-phosphoenolpyruvate system (20 U/mL pyruvate kinase and 5 mM phosphoenolpyruvate), and in the presence of respiratory substrates (10 mM glutamate and 2 mM malate). ww Wet weight

Compared with adult rat cardiomyocytes (Figures 1A and 1B), the structural organization of the HL-1 cells is very different (Figures 1C and 1D). In contrast to the perfect intracellular organization of mitochondria in adult cardiomyocytes, these organelles are rather chaotically arranged in HL-1 cells, forming dynamically changing filaments due to ongoing fission and fusion. Whereas the B HL-1 cells possess some residual sarcomeres (9,10), the NB HL-1 cells are devoid of these structures. These properties suggest that the ICEUs may not exist in HL-1 cells, and hence, the regulation of mitochondrial respiration in these cells must occur differently from that in adult cardiomyocytes. Indeed, analysis of ADP concentration versus oxygen uptake relationships revealed that while the apparent K_m for ADP for adult cardiomyocytes was 360±51 μM, its mean value for NB HL-1 and B HL-1 cells was much less, varying between 20 μM and 45 μM, irrespective of the type of respiratory substrate (complex I- or complex II-dependent) used. The treatment of adult rat cardiomyocytes with trypsin led to decreased values of the apparent K_m for ADP (Figure 3), but exerted no effect on that parameter in either HL-1 cell subtype (results not shown). These findings clearly show that mitochondria in HL-1 cells have an easy access to exogenously added ADP, most likely due to the absence of marked intracellular diffusion barriers. In other words, mitochondria appear to function in these cells similarly to how they function in in vitro conditions, where the local concentrations of ADP near mitochondrial adenine nucleotide translocase do not significantly differ from those in the surrounding medium. Interestingly, as in HL-1 cells, the K_m for ADP in the regulation of respiration is very low in one-day-old postnatal rat cardiomyocytes, when the mitochondria exhibit random localization in the cytoplasm and the contractile apparatus is not yet developed (27). In light of these findings, the immortal HL-1 cells appear to be cells in which the intracellular energy metabolism has regressed from the ICEU type of organization toward a less developed system characterized by a lack of control over mitochondrial function by intracellular structures (cytoskeleton) in a course of cellular remodelling under conditions of prolonged culturing.

Figure 3).

The decrease in the apparent K_m (Michaelis-Menten constant) for exogenous ADP in respiration regulation of permeabilized cardiomyocytes after treatment with trypsin (TR). Before kinetic experiments, permeabilized cardiomycytes were incubated with TR in the indicated concentrations for 5 min at 4°C. The means ± SD are given

Further support for this assumption was obtained when a direct transfer of phosphoryl groups between mitochondria and ATPases was studied using the exogenous ADP trapping system PK-PEP (3,18). As Figure 4 shows, after the addition of ATP to the permeabilized adult cardiomyocytes, the coupled PEP-PK-LDH system measured the total ADP flux produced by ATPases. However, the activation of mitochondrial respiration by the addition of respiratory substrates significantly attenuated the ADP flux for the PK-PEP system, and only after inhibition of oxidative phosphorylation by atractyloside could the flux through the PK-PEP system be restored. Thus, a part of the total ADP flux (approximately 30%, as a difference between the ADP fluxes before and after the addition of respiratory substrates [Figure 4B]) was directly channelled from ATPases to mitochondria, without reaching a space controlled by the PK-PEP system. In contrast, mitochondria were unable to capture ADP in NB HL-1 cells, which indicates the absence of structural and functional interactions between mitochondria and ATPases in these cells (Figures 4C and 4D).

Figure 4).

Absence of channelling of endogenous ADP from MgATPases to mitochondria in nonbeating HL-1 cells (C and D) in contrast to primary adult cardiomyocytes (CM) (A and B). ADP fluxes were measured at 340 nm (Gary 50 spectrophotometer, Varian, USA) by the pyruvate kinase-phosphoenolpyruvate-NADH-lactate dehydrogenase coupled enzyme system at 25°C. A and C Representative traces of the ADP fluxes in the presence of 50 μg/mL saponin (Sap), MgATP (3 mM), 5 mM glutamate plus 2 mM malate (Glu+Mal) and 30 μM atractyloside (ATR). B and D Bar plots showing a summary of the measurements shown in the left panel. *Significantly different from ADP flux in the presence of 2 mM ATP. Error bars show ± SD. *Statistically significant (P<0.05; n= 4 to 7) change in ADP flux for cardiomyocytes (B) showing channelling of endogenous ADP to mitochondria; °Not statistically significant change in ADP flux for HL-1 cells (D) according to Student’s t test. +S Measurements in the presence of substrates (Glu+Mal); Abs Absorption; V_ATPase Rate of ADP synthesis

CONCLUSIONS

These results further support the concept that the ICEUs in adult cardiomyocytes are products of the complex structural organization developed to create the most optimal conditions for effective energy transfer and feedback between mitochondria and ATPases. This organization and related regulatory mechanisms explain the metabolic basis of the basic Frank-Starling law in the heart (28).