INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

CREATINE KINASE (CK) is an enzyme thought to be critically involved in intracellular energy homeostasis. It catalyzes the reversible transfer of a high-energy phosphoryl group between ATP and creatine (Cr): phosphocreatine (PCr)² + MgADP + H⁺  Cr + MgATP².

The primary and well-recognized role of this reaction is to maintain a constant level of cytosolic ATP and a low cytosolic free ADP level, particularly when ATP hydrolysis temporarily exceeds ATP synthesis. Biochemical studies revealed the existence of different mammalian CK isoforms, which are distributed in a tissue-dependent manner and are located in different cellular compartments (see Ref. 38 for review). Invariably, at least one type of cytosolic CK is coexpressed with a mitochondrially localized CK isoform (Mi-CK) (39). In cardiac muscle at least 70% of CK activity resides in the cytosol, primarily as the homodimeric MM-CK species and partly as the heterodimeric MB-CK isoform. The remainder is represented by a specific mitochondrial type of CK. Mi-CK primarily forms octamers in vivo and is localized along the outer leaflet of the mitochondrial inner membrane and at sites where the outer and inner mitochondrial membranes are in close vicinity (39). The various CK isoforms seem to be functionally associated with free-energy producing and delivering processes in the cell (38). A fraction of MM-CK activity is specifically localized to subcellular structures such as the myofibrillar M band (38), glycolytic enzyme complexes, and the sarcoplasmic reticulum (38). Mi-CK preferentially uses ATP produced via mitochondrial oxidative phosphorylation because of functional coupling with the adenine nucleotide translocase (39). Hence, PCr leaving the mitochondrion and Cr entering it are thought to be the principal metabolites connecting sites of ATP delivery and ATP utilization through diffusion (38). This so-called "CK/PCr shuttle" concept emphasizes not only temporal energy buffering but also energy transport via PCr. The CK/PCr shuttle concept relies mostly on data obtained in vitro, whereas its critical role in energy metabolism in vivo has not yet been convincingly demonstrated.

Noninvasive ³¹P NMR methods have been used extensively to study the importance of the CK system for energy metabolism in intact tissue. Notably, ³¹P magnetization-transfer techniques have proven useful to assess CK-mediated flux in muscle at rest and in response to changes in workload (6). Whether a direct correlation between CK flux and ATP synthesis rate is crucial for muscle function is still a matter of debate. In the perfused heart CK flux has been shown to increase with elevated workload (4, 5, 22, 26, 28, 29). By contrast, in the in vivo paced mammalian heart (2) and in skeletal muscle (8, 31) CK flux remains unchanged or, in the latter, even slightly decreases with elevated ATPase activity and oxidative ATP synthesis rate.

Thus far, how individual CK isoenzymes contribute to the total CK-related flux measured by NMR magnetization-transfer techniques has remained obscure. Mathematical analysis of transient ³¹P NMR saturation-transfer data of neonatal rabbit hearts by Zahler and Ingwall (40) provided estimates for the Mi-CK flux that predicted that this would increase linearly with ATP synthesis rate. Intuitively, this would seem to agree with the CK/PCr shuttle model, which suggests intimate coupling of Mi-CK activity with mitochondrial ATP synthesis. However, so far, direct experimental evidence to substantiate this point is not available.

Interesting progress in CK research has recently evolved from novel advances in gene technology (21, 33, 34). Wieringa and co-workers (33, 34) succeeded in creating mouse strains that are homozygous for the targeted disruption of the gene encoding for a specific CK isoenzyme. These knockout models permit exciting new studies into the functional importance of the individual CK isoenzymes. In M-CK-deficient hindlimb skeletal muscle, in which 2% of the wild-type CK activity remains represented by Mi-CK, Van Deursen et al. (33) recently did not detect (Mi-)CK flux by inversion-transfer ³¹P NMR.

In the present study we measured the kinetics of CK in isolated, perfused hearts of M-CK-deficient mice. The high capacity for oxidative ATP synthesis and the relatively high percentage of Mi-CK isozyme suggest a prominent role for the CK/PCr shuttle in the wild-type heart (38). In the M-CK-deficient heart such a CK/PCr shuttle can no longer exist. By applying ³¹P NMR and saturation transfer to investigate (Mi-)CK flux and metabolic changes in response to altered cardiac performance, we particularly aimed at gathering information about the kinetics of Mi-CK and its function in cardiac energy metabolism.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Langendorff perfusion. Hearts were isolated from male wild-type C57Bl and M-CK-deficient mice (33) weighing ~30 g. The mice were anesthetized through inhalation of diethyl ether. After heparinization, the heart was rapidly excised and rinsed in ice-cold buffer. The heart was subsequently perfused through an aortic cannula. The prefiltered perfusate, a modified Krebs-Henseleit medium, contained (in mM) 124 NaCl, 4.7 KCl, 1.3 CaCl₂, 1.0 MgCl₂, and 24 NaHCO₃, as well as 10 glucose and 5 Na-pyruvate as substrates, and was continuously bubbled with 95% O₂-5% CO₂ gas, resulting in a pH of 7.35 at 37°C. The heart was perfused at a constant temperature of 37°C, maintained through a circulating water bath, and at constant hydrostatic pressure of 76 mmHg. A fluid-filled catheter was inserted via the mitral valve, passed through the apex, and connected to a Gould-Statham P23 Db pressure transducer for monitoring left ventricular developed pressure (LVDP) and heart rate. Two small copper pace electrodes were attached to the right ventricular outflow tract (10). The heart was then carefully placed into an NMR tube of 10-mm diameter together with a small glass capillary containing methylene diphosphonate (MDP) for spectral reference and calibration. The perfusate effluent was removed from the NMR tube using a peristaltic pump, keeping the liquid level well above the heart and the NMR detection coils. The perfusate flow rate was measured at regular intervals.

³¹P NMR spectroscopy. ³¹P NMR spectra were recorded at 202.5 MHz on a Bruker AM500 spectrometer interfaced to an Aspect 3000 computer and a vertical 11.7-T superconducting magnet. All spectra were obtained using quadrature detection, a 90° (35 µs) flip angle, a sweep width of 11,905 Hz, and 2,048 data points, without proton decoupling. The magnetic field was homogenized by shimming on the ¹H signal of water, resulting in typical line widths of 8-12 Hz. First, a fully relaxed ³¹P NMR spectrum was recorded from 64 scans with an interpulse delay of 10 s. Next, a transient ³¹P NMR saturation-transfer protocol was employed by selectively irradiating the -ATP resonance for specific durations and observing the PCr signal. Analysis of the exponential decrease of the PCr peak area with saturation time gave an estimate for the pseudo first-order rate constant (k_for) of PCr-to-ATP exchange (see Kinetic analysis). Selective saturation was achieved by applying a delays alternated with nutations for tailored excitation (DANTE) pulse sequence with 0.8-µs pulses and a 125-µs interpulse delay. The durations of the selective irradiation at the -ATP resonance were 0, 0.3, 0.6, 1.5, 3.0, and 6.0 s, and it was applied randomly to minimize any effects of aging of the preparation, which might lead to overestimation of the calculated flux. The total repetition time remained constant at 6.1 s. Finally, a control experiment was performed in which the selective irradiation (6 s) was placed at a frequency () downfield from the PCr resonance [(PCr)  (-ATP)]. Direct radio frequency spillover was generally <10%. The full characterization of a heart at one workload level typically lasted 52 min. Each heart was studied successively at 400 and 600 beats/min and during cardiac arrest. Full diastolic arrest was induced by switching to a perfusate in which the KCl concentration had been increased to 20 mM with an equivalent decrease in the NaCl content. Before NMR data collection was started, the heart was allowed to adapt to each new performance level for ~10 min. Workload is expressed as rate-pressure product (RPP), i.e., heart rate × LVDP. The intracellular pH was assessed from the chemical shift difference between intracellular P_i and PCr, as described previously (13). The intracellular free Mg²⁺ concentration ([Mg²⁺]) was estimated from the chemical shift difference between the - and -ATP resonances as described before (20).

Kinetic analysis. Time-dependent ³¹P NMR magnetization transfer was used to determine the pseudo first-order unidirectional rate constants for the forward CK reaction, as described previously (5, 11). The relationship between the PCr signal [M_t(PCr)] and the time of -ATP saturation was analyzed by a monoexponential fit, according to the equation

(1)

where M₀(PCr) is the magnetization of PCr in the absence of -ATP saturation, k_for is the forward CK rate constant (in s¹), and T1_app is the apparent longitudinal relaxation time (in s), measured in the presence of -ATP saturation. The intrinsic longitudinal relaxation time (T1_intr), which would be T1 in the absence of phosphate exchange, was calculated from the relation T1¹_app = T1¹_intr + k_for.

Spectral analysis. ³¹P NMR signals were quantified by iterative fitting of the time-domain signal with the method of variable projection (VARPRO), using prior knowledge (32).

Enzyme activities and metabolite determinations. After the NMR experiment each heart was blotted and weighed to obtain wet weight and then dried in an oven at 80°C for at least 24 h and weighed again to obtain dry weight. Hearts from a different set of animals, but from the same strain, were used for biochemical assays. For that purpose, hearts were rapidly excised and immediately freeze-clamped in liquid N₂-cooled tongs and stored in liquid N₂ until use. For measurement of CK activity, hearts were allowed to warm up to 4°C in homogenization buffer (10% wt/vol), containing (in mM) 200 mannitol, 30 sucrose, 25 HEPES (pH 7.1), 1 EDTA, and 10 Mg-acetate with 0.1% Triton X-100. The tissue was homogenized at 650 rpm in a Potter-Elvehjem homogenizer and then centrifuged for 5 min at 550 g. The remaining supernatant was assayed spectrophotometrically for CK activity, using a coupled enzyme assay at pH 7.1 and 30°C. Details of the assay have been described previously (35). For metabolite assays, the frozen hearts were pulverized in a dry ice-cooled porcelain mortar and extracted with ice-cold perchloric acid (5% wt/vol). The extract was centrifuged for 2 min at 12,000 g, neutralized by addition of 2 M KOH in 100 mM Tris, and stored in liquid N₂. ATP and PCr were determined spectrophotometrically, using standard coupled enzyme assays at pH 6.7 (24). The Cr concentration was determined using a spectrophotometric assay (3). Total Cr (TCr) levels were determined by summation of the PCr and Cr concentrations. Metabolite concentrations were calculated on the assumption that the intracellular water volume is 0.47 ml/g tissue wet wt (1).

Statistics. Differences between parameters of wild-type and mutant animals were analyzed using an unpaired Student's t-test. Multiple comparisons were made by the multiple-range test of Bonferroni if a significant effect was found by one-way or two-way analysis of variance. Differences were considered significant if P < 0.05. Data are presented as means ± SE.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Tissue mass. After perfusion the mouse hearts were blotted dry and weighed. Transgenic hearts had a wet weight of 180 ± 10 mg (n = 10), and control hearts were 175 ± 15 mg (n = 16). Wild-type and M-CK-knockout hearts were found to have equal wet weight-to-dry weight ratios, namely 5.24 ± 0.14 (n = 16) and 5.13 ± 0.16 (n = 10), respectively. Recently, equal heart weight-to-body weight ratios have been reported for these mice (36, 37).

CK activities and metabolites. The hearts of wild-type and M-CK-deficient mice were analyzed biochemically to assess the activity of CK. The CK activity at 37°C in wild-type hearts was 39.4 ± 4.1 mM/s (n = 7) compared with 10.4 ± 0.8 mM/s (n = 7) in M-CK-deficient hearts. Thus the CK activity in M-CK-deficient hearts was reduced to ~26% of the wild-type level. These activities are somewhat higher for both types of hearts than those recently reported (36), which is possibly caused by the fact that we included 0.1% Triton X-100 in our medium for more efficient homogenization. It was previously reported that knocking out the gene for M-CK does not cause elevated expression of Mi-CK in either heart ventricles or skeletal muscles (36). In some M-CK-deficient ventricles a small increase in cytosolic brain CK activity has been observed (33, 36), which was on average 10% of the total CK activity in these hearts, so that 90% of the CK activity resides in the mitochondrion.

Physiological parameters. Left ventricular diastolic and systolic pressures were measured continuously. No overt discrepancies between wild-type and M-CK-deficient hearts in either of these parameters were observed. At 400 beats/min, the average LVDP was 70 ± 6 mmHg (n = 7) in wild-type hearts and 74 ± 7 mmHg (n = 6) in transgenic hearts. At 600 beats/min, the average LVDP was 55 ± 4 mmHg (n = 6) in control hearts and 58 ± 6 mmHg (n = 6) in M-CK-deficient hearts.

Response of high-energy phosphate metabolism to workload changes. Hearts were studied by ³¹P NMR during 400 and 600 beats/min pacing and KCl-induced cardiac arrest, successively. At the highest heart rate, the RPP was ~34,000 mmHg/min in normal as well as in transgenic hearts. Previous studies on perfused rat and rabbit hearts have indicated that RPP is linearly correlated with myocardial oxygen consumption rate (12). Figure 1 shows typical ³¹P NMR spectra of wild-type and M-CK-deficient hearts at different workloads. The spectra demonstrate that workload-related alterations in the levels of phosphate metabolites were small in both wild-type and M-CK-deficient perfused hearts. The NMR-observed PCr-to-ATP ratio (PCr/ATP) in transgenic hearts was similar to that of wild-type hearts at all workload levels. At 400 beats/min, PCr/ATP was 1.55 ± 0.13 in M-CK-knockout and 1.65 ± 0.13 in wild-type hearts. In both normal and transgenic hearts PCr/ATP was significantly higher during arrest than during pacing at 400 beats/min, i.e., 2.86 ± 0.31 and 2.79 ± 0.07, respectively. PCr/ATP was somewhat higher during 600 beats/min than during 400 beats/min, i.e., 2.52 ± 0.20 in wild-type and 2.00 ± 0.18 in transgenic hearts. The observed differences in PCr/ATP at different work states could be partially explained by somewhat diminished ATP levels in both transgenic hearts and controls. However, this occurred without any concomitant meaningful decline of PCr levels or elevation of P_i levels.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   31P NMR spectra showing effect of workload on phosphate metabolites in perfused wild-type and cytosolic muscle creatine kinase (M-CK)-deficient mouse hearts. Fully relaxed 31P NMR spectra were recorded at 202 MHz (with 64 scans and 10-s repetition time) and 37°C. Hearts were studied at 3 different performance levels, successively: 1) paced at 400 beats/min, 2) paced at 600 beats/min, and 3) during KCl-induced cardiac arrest. A: wild-type mouse hearts. B: M-CK-deficient hearts. Perfusion medium was a modified Krebs-Henseleit buffer with 10 mM glucose and 5 mM pyruvate. Peak assignments are phosphomonoesters (PME), Pi, phosphocreatine (PCr), and -, -, and -phosphates of ATP. , Chemical shift; ppm, parts per million.

1

³¹P NMR saturation-transfer measurements of CK flux. Using ³¹P NMR saturation transfer, we determined the velocity of CK-mediated exchange between PCr and ATP in the isolated, perfused heart at the aforementioned three levels of cardiac performance. Figure 2 shows typical saturation-transfer spectra of a control heart (A and B) and a M-CK-deficient heart (C and D), both beating at 400 beats/min; spectra C and D and the difference spectrum (C  D) clearly indicate that Mi-CK, which is the predominant CK isozyme remaining in the transgenic heart, catalyzes an appreciable exchange flux. By contrast, Van Deursen et al. (33) recently did not detect (Mi-)CK flux in the hindlimb skeletal muscle from these mice. Data from a transient saturation-transfer protocol were analyzed according to Eq. 1. Figure 3, A-C, shows the exponential decay of the average PCr peak area at progressively longer saturation times of the -ATP resonance, demonstrating magnetization transfer between PCr and ATP in the M-CK-deficient heart at all work levels examined. Figure 3, D-F, shows similar curves for the control heart. A two-parameter exponential fit to the data according to Eq. 1 yielded values for the pseudo first-order rate constant for the CK reaction toward ATP synthesis and of the apparent spin-lattice relaxation time (T1_app) as summarized in Table 2. The intrinsic T1 values of PCr (T1_intr), i.e., T1 in the absence of PCr-ATP exchange, were calculated and appeared similar in mutants and controls. The unusually long T1_intr in the arrested control heart is probably largely caused by the significant error in the determination of rate constant k and T1_app and is expected to be similar to the T1_intr in beating control hearts. In beating hearts of control mice, the apparent rate constant, k_for, was almost twice as high as in arrested control hearts (Table 2). The k_for was not significantly lower in arrested than in beating hearts of M-CK-deficient mice. In parallel, in control hearts the total CK-catalyzed flux increased significantly with cardiac performance, expressed as RPP (Fig. 4). A similar correlation between CK flux and workload has been observed in perfused hearts of various other animal species (4, 5, 22, 26, 28, 29). CK flux did not increase significantly in M-CK-deficient mouse hearts. The PCr-ATP exchange flux in transgenic hearts was lower than in wild-type hearts by ~40% at 400 beats/min.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   31P NMR saturation-transfer spectra of hearts from M-CK-deficient and control mice. A and B: control hearts. C and D: transgenic hearts. Spectra are from 64 scans with 6.1-s repetition time. Arrows denote frequency of selective irradiation (6 s). Spectra denoted A  B and C  D are difference spectra for control heart and mutant heart, respectively. For details see MATERIALS AND METHODS.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Transient saturation-transfer data for wild-type and M-CK-deficient hearts at different performance levels. A-C: wild-type hearts. D-F: M-CK-deficient hearts. Average PCr amplitudes ([PCr]; means from 6-8 animals) from a 31P NMR saturation-transfer series are shown as a function of duration (t) of saturation of -ATP resonance. Data are normalized individually to [PCr] in absence of -ATP irradiation (t = 0). Error bars represent SE for PCr signal at each saturation time. Solid lines are 2-parameter exponential curve fits to average data according to Eq. 1.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Flux mediated by CK in normal and M-CK-deficient hearts as a function of rate-pressure product (RPP). , Wild-type hearts; , mutant hearts. Error bars give SE in both parameters. Values are also given in Table 2.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This study has given important clues about the kinetic behavior of the Mi-CK isozyme through the analysis of its kinetics in the M-CK-deficient mouse heart. In normal tissue the in vivo kinetics of Mi-CK cannot be distinguished from that of the more abundant cytosolic isozyme. However, transgenic hearts that lack M-CK are uniquely suited for study of the kinetic behavior of Mi-CK. Isoenzyme electrophoresis has shown that Mi-CK activity constitutes at least 90% of the total CK activity in the M-CK-knockout heart (33, 36). The remainder is represented by the brain-type homodimeric B-CK (BB-CK) isoform, which is possibly a replacement for the MB isoform found in wild-type hearts. Mi-CK levels in cardiac tissue are not significantly enhanced by the loss of M-CK (33, 36).

To date, the consequences of a knockout of the M-CK gene for myocardial metabolism and function have not been characterized in full detail. The levels of pyruvate kinase and glyceraldehyde-3-phosphate dehydrogenase are significantly increased in M-CK-deficient ventricles, suggesting somewhat enhanced glycolytic activity, although no changes in the activities of adenylate kinase, fructose-6-phosphokinase, or lactate dehydrogenase have been found (37). Ultrastructural data on the transgenic ventricular tissue are not available as yet. Although the M-CK knockout in fast-twitch skeletal muscle led to conspicuous abnormalities, particularly pertaining to the mitochondrial ultrastructure, M-CK-deficient soleus muscle, which like cardiac muscle consists mainly of slow, oxidative fibers, showed neither histochemical nor ultrastructural abnormalities.

M-CK deficiency may be expected to have consequences for heart performance. A fraction of the MM-CK activity is attached to the myofibrils, where it is coupled to the actomyosin ATPase activity, thereby maintaining high and potentially critical (local) ATP-to-ADP ratios during muscle contraction. Our experimental data did not reveal any obvious differences in the performance of M-CK-deficient hearts compared with controls, i.e., LVDP were similar at each pace rate. Thus M-CK appears not to be critical for cardiac function under our experimental conditions. On a different time scale, however, Ventura-Clapier et al. (37) showed that the rate constant for force development in calcium-activated M-CK-deficient ventricular fibers was markedly smaller than in controls, although the developed force by each cross bridge was not altered.

In normal hearts, MM-CK activity constitutes ~75% of the total CK activity. Therefore, PCr/ATP are affected primarily by the reaction of cytosolic MM-CK, operating near equilibrium. Interestingly, in the M-CK-deficient heart PCr/ATP as well as absolute ATP and PCr concentrations were similar to those in wild-type hearts, whereas they are affected almost exclusively through equilibration via Mi-CK. Thus, even without the majority of cytosolic CK, the transgenic hearts are able to maintain their cytosolic phosphorylation potential. Similar observations have been made in the hindlimb skeletal muscle of these mice (33).

The time- and workload-dependent changes in PCr were small, and no consistent increase in intracellular P_i levels was observed. Because the K⁺-arrested heart was studied at the final stage of the protocol, i.e., ~2 h after the start of perfusion, PCr and ATP levels as well as the NMR-detected CK flux may have become slightly reduced by gradual deterioration of cardiac energy metabolism. Such aging effects were minimal during execution of the saturation-transfer protocol at one workload level. Because of the relative insensitivity of NMR, even at high field strength, and the low tissue mass of mouse hearts (typically 0.175 g wet tissue), rather long measuring times were required that may have affected the stability of the heart preparation. However, the rationale for our protocol was that potential metabolic changes with different workloads could be observed in a single heart preparation.

Although a number of studies have addressed the kinetic behavior of CK in different muscle types, including heart and skeletal muscle, controversy remains about the importance of the enzyme for muscle function and energetics. Part of the controversy may have arisen from conflicting results obtained with different muscle types. On stimulation of skeletal muscle, PCr/ATP gradually declines, while P_i levels rise concomitantly as does the estimated cytosolic free [ADP]. Hence, in skeletal muscle the cytosolic free [ADP] has been implied as an efficient feedback signal to the mitochondria to regulate oxidative ATP synthesis (23). By contrast, in perfused cardiac muscle, workload-related changes in PCr/ATP are often less pronounced and may depend on experimental conditions such as the nature of carbon sources or inotropic agents used (10, 12, 16, 17, 19, 41). In the working dog heart in vivo, PCr/ATP and free ADP levels apparently barely change, at least over the work range studied, suggesting that factors other than free ADP may contribute to the control of oxidative phosphorylation in the heart in vivo (2, 18).

In this study, free [ADP] in wild-type and mutant working hearts were estimated to be on the order of 30-50 µM. These values are near the Michaelis-Menten constant of ADP [K_m(ADP)] for mitochondrial oxidative phosphorylation of 25-30 µM, as observed in isolated mitochondria (15), perfused hearts (12), and in vivo skeletal muscle (8, 9). We found that free ADP levels in M-CK-deficient and control animals were similar at each workload level. Free [ADP] tended to be higher in beating than in arrested hearts from both mutants and controls. In transgenic hearts, a significant decrease in free [ADP] between pacing at 600 beats/min and K⁺ arrest was observed. Unfortunately, the present data lack sufficient accuracy to allow strict conclusions on the extent of workload-related changes in free [ADP] in the M-CK-deficient heart.

Veksler et al. (36) recently observed that in skinned ventricular fibers of M-CK-knockout mice the K_m(ADP) of oxidative phosphorylation in the absence of creatine was significantly lower than in wild-type fibers. The addition of creatine markedly increased the rate of oxygen consumption through ADP production by Mi-CK in normal fibers but not in mutant fibers. From these data the authors concluded that the mitochondrial outer membrane in M-CK-deficient ventricles may have an increased permeability for ADP (36). Our calculations of free ADP levels in the heart are largely in agreement with this hypothesis. In transgenic hearts Mi-CK may be expected to equilibrate with the ADP in the mitochondrial intermembrane space, whereas in wild-type mice the calculated ADP largely represents the cytosolic pool. Our finding that free ADP levels and their changes between the various workloads were similar in wild-type and M-CK-knockout hearts suggests equilibration of the ADP pool in the interspace with that in the cytosol in the transgenic heart.

In wild-type hearts, the NMR-observed CK flux will be a summation of the fluxes through the various CK isoenzymes. As a consequence, the kinetic response of individual CK isoenzymes to changes in workload, particularly that of Mi-CK, is still unknown. In this study, we were able to obtain a unique estimate of this Mi-CK-catalyzed flux from a ³¹P NMR magnetization-transfer examination of M-CK-deficient mouse hearts. We demonstrate for the first time that the steady-state Mi-CK flux is readily observed by NMR and that the apparent rate constant for the reaction in both the beating and the arrested intact heart can be obtained. As mentioned before, at least 90% of the CK activity in the transgenic tissue arises from Mi-CK (36) and the small fraction of BB-CK is not expected to contribute >10% to the CK exchange flux. Recently, quantification of Mi-CK flux in the hindlimb skeletal muscle of these mice did not appear possible (33). At least partially, this can be explained by the higher Mi-CK content of heart compared with skeletal muscle tissue. By comparison, the chemically determined Mi-CK activity in the M-CK-deficient fast-twitch skeletal muscle is only 1-2% of the total CK activity found in homogenates of wild-type muscle.

It has previously been postulated that Mi-CK and myofibril-bound MM-CK operate far from equilibrium and in opposite directions, enabling transport of high-energy phosphates via a CK/PCr shuttle (39). It is commonly assumed that the flux through these compartmentalized CK isozymes should increase with elevated ATP hydrolysis and synthesis rates during work. It should be noted that in transgenic heart a net flux through the Mi-CK reaction cannot exist because of the absence of cytosolic CK. Unfortunately, we cannot quantitatively compare Mi-CK flux with ATP synthesis rate, because we were not able to measure oxygen consumption. However, the observed Mi-CK flux in transgenic mouse heart is several times higher than previously reported rates of mitochondrial ATP synthesis in the heart, which did not exceed 1.3 mM/s at high workload (26). Thus in the transgenic heart Mi-CK-catalyzed exchange should be sufficiently rapid to maintain the reactants near equilibrium. This validates our calculations of free ADP levels in the transgenic heart from CK equilibrium. Interestingly, it may be inferred that because Mi-CK activity is apparently similar in transgenic and normal hearts (33, 36), the Mi-CK reaction should also be sufficiently rapid to maintain near equilibrium under normal steady-state conditions in the wild-type heart, as is generally assumed to be the case for the bulk cytosolic MM-CK isozyme (27). It cannot, however, be ruled out that in wild-type hearts the CK reactions localized at the myofibrils and mitochondria may become (temporarily) displaced from equilibrium, especially during peak workloads or rapid workload transitions.

In Fig. 4 we show that the total CK flux was significantly lower in K⁺-arrested than in beating wild-type mouse hearts. Remarkably, in transgenic hearts the difference in (Mi-)CK flux between arrest and beating was much less pronounced, with approximately equal changes in RPP and in estimated free [ADP]. Thus in the M-CK-knockout mouse heart an increased rate of mitochondrial oxidative phosphorylation with work is not accompanied by a substantially higher flux through Mi-CK. If we tentatively assume that we can extrapolate these data from the transgenic to the wild-type situation, MM-CK flux may increase significantly with elevated free ADP levels, whereas Mi-CK flux may increase to a much lesser extent. This would imply that in arrested and beating mouse hearts the free [ADP] may approach saturating levels for Mi-CK but may be near the K_m(ADP) for MM-CK. This is supported by previously reported values for K_m(ADP), which were in the physiological range and were generally higher for the MM-CK than for the Mi-CK reaction (30, 35, 39).

In the arrested heart the apparent k_for and CK flux are approximately equal in wild-type and M-CK-knockout hearts. Remarkably, the Mi-CK-mediated flux in beating mutant hearts was only lower by ~40%, whereas the total CK activity, measured biochemically, had decreased by 74% in M-CK-deficient cardiac tissue. This suggests that the ratio of the steady-state NMR-detected CK flux to the unidirectional CK activity (V_max) is remarkably higher for Mi-CK than for MM-CK. In other words, relative to its in vitro activity, Mi-CK may contribute much more to the NMR-detected flux than MM-CK. In agreement with this observation, recent studies on purified Mi-CK and MM-CK in solution (35) have shown that at identical V_max, Mi-CK flux measured by ³¹P NMR saturation transfer was about twofold higher than MM-CK flux. In that study, the observed fluxes could be explained quantitatively, using kinetic properties of the enzymes in solution (35). The differential contributions of Mi-CK and MM-CK to the total NMR-detected CK flux in vivo may have profound consequences for the interpretation of NMR studies, which aim to probe CK function during muscle contraction.