The production of AMP by adenylate kinase (AK) and subsequent deamination by AMP deaminase limits ADP accumulation during conditions of high-energy demand in skeletal muscle. The goal of this study was to investigate the consequences of AK deficiency (−/−) on adenine nucleotide management and whole muscle function at high-energy demands. To do this, we examined isometric tetanic contractile performance of the gastrocnemius-plantaris-soleus (GPS) muscle group in situ in AK1−/− mice and wild-type (WT) controls over a range of contraction frequencies (30–120 tetani/min). We found that AK1−/− muscle exhibited a diminished inosine 5′-monophosphate formation rate (14% of WT) and an inordinate accumulation of ADP (∼1.5 mM) at the highest energy demands, compared with WT controls. AK-deficient muscle exhibited similar initial contractile performance (521 ± 9 and 521 ± 10 g tension in WT and AK1−/− muscle, respectively), followed by a significant slowing of relaxation kinetics at the highest energy demands relative to WT controls. This is consistent with a depressed capacity to sequester calcium in the presence of high ADP. However, the overall pattern of fatigue in AK1−/− mice was similar to WT control muscle. Our findings directly demonstrate the importance of AMP formation and subsequent deamination in limiting ADP accumulation. Whole muscle contractile performance was, however, remarkably tolerant of ADP accumulation markedly in excess of what normally occurs in skeletal muscle.
- AMP deaminase
- tetanic contraction
- muscle relaxation
- calcium handling
- cross-bridge cycling
skeletal muscle can maintain favorable energetic conditions in the face of great energetic demands. For example, the ATP consumption rate needed to support a tetanic contraction (∼10 μmol/g per second) is ∼200-fold above the resting rate in rat fast twitch fibers (28, 42). The metabolic challenge this represents is highlighted by the fact that the ATP content of mammalian skeletal muscle is only 6–7 μmol/g. Therefore, without compensation, the entire ATP pool would be completely depleted in <1 s at this rate. Obviously, this does not occur due to the ATP synthetic processes of oxidative phosphorylation, glycolysis, the creatine kinase (CK) reaction, and the adenylate kinase (AK) reaction. Moreover, the free energy available from ATP hydrolysis (ΔGATP) is a function of the ATP content relative to the products of ATP hydrolysis, namely ADP and inorganic phosphate (Pi) where ΔGATP is the conventional expression for the free engery of ATP at standard conditions, R is the ideal gas constant at 0.0083143 kJ/mol·K, and T is the temperature in Kelvin. Furthermore, the accumulation of ADP and Pi direct physiological consequences independent of an impaired ΔGATP. Thus, the metabolic challenge when ATP turnover is high, is to preserve ATP content and limit inordinate accumulation of ADP and Pi.
When the rate of ATP hydrolysis is out of balance with the rate of ATP synthesis, the need to limit the accumulation of ADP and Pi is the greatest. The reaction catalyzed by CK via the transfer of a phosphate from phosphocreatine (PCr) to ADP, serves as an ATP buffer, thereby maintaining ATP content utilizing ADP. Furthermore, the reaction catalyzed by AK functions to minimize ADP accumulation through the transfer of a phosphate from one ADP to another ADP forming ATP and AMP. In addition, the capacity of AK to limit ADP accumulation is magnified by the coupled reaction catalyzed by AMP deaminase (AMPD). The AMP formed by AK is rapidly deaminated to inosine 5′-monophosphate (IMP) by AMPD. This keeps the AK catalyzed reaction proceeding in the direction limiting ADP accumulation. The tight coupling of AK and AMPD is evident by the exceptional accumulation of IMP that can occur in fast twitch muscle [∼3.5 μmol/g wet wt (40)], which is orders of magnitude greater than the estimated free AMP accumulation (12) This rapid flux through AK and AMP establishes a stoichiometric increase in IMP with a corresponding reduction in ATP (39). If the coupled reaction through AK and AMPD did not occur when ATP turnover is out of balance with ATP supply, an inordinate accumulation of ADP and a decline in the energy from ATP, would then be expected. Therefore, AK and AMPD serve to limit ADP accumulation, preserving ΔGATP, at the cost of a decline in the total ATP content of the muscle.
The contractile ATPase and the sarco(endo)plasmic reticulum calcium ATPase (SERCA) account for the vast majority (≥95%) of the energetic cost of a muscle contraction (46, 47). The cost of maintaining a low cytosolic calcium concentration by SERCA, pumping calcium back in to the sarcoplasmic reticulum (SR) after a contraction, is ∼30% of the total cost of contraction. In addition, the cost to maintain the large calcium gradient between the cytosol and the SR (a ratio ∼10,000:1) is near the limit of how much ΔGATP is available (21). Therefore, ADP accumulation and the coordinate reduction in the ΔGATP, could lead to compromised SERCA function and prolonged elevation in cytosolic calcium. Consistent with this idea, a study by Dawson et al. (9) revealed a close relationship between small reductions in the free energy of ATP hydrolysis, and slowed relaxation in frog skeletal muscle (implying a prolonged calcium transient that led to delayed relaxation kinetics). Therefore, minimizing ADP accumulation likely preserves normal contraction kinetics.
The appropriate management of intracellular calcium in muscle cells is critical for contraction and relaxation. The rate calcium returns to resting concentrations after muscle activation influences the rate of relaxation. This has been demonstrated using models of altered expression of the calcium-binding protein parvalbumin (7, 48, 50). Whereas overexpression enhances the calcium-buffering capacity, it also results in faster relaxation kinetics (7). Conversely, the opposite pattern was observed in fast-twitch muscle from parvalbumin-deficient mice (48, 50). These results imply that the rate of calcium dissociation from troponin C has an impact on muscle relaxation. Furthermore, Luo et al. (34) demonstrated that as the affinity of troponin C for calcium is either increased or decreased, a corresponding increase or decrease in the time for relaxation is observed. Therefore, altering the rate with which calcium leaves troponin C can have an impact on muscle relaxation. Thus if ADP accumulation slows the rate that cytosolic calcium returns to the SR, calcium should remain longer on troponin C and the rate of relaxation should be slowed.
An inordinate accumulation of ADP may also impact myosin cross-bridge dynamics. Tension generated by muscle is ultimately a function of the balance between the cyclical attachment and detachment of actin and myosin. The steps leading to myosin detachment from actin involve the dissociation of ADP and subsequent binding of ATP to the nucleotide-binding site on myosin (see Ref. 17 for a review). The rate of ADP dissociation from myosin limits the rate of cross-bridge cycling at the fastest cross-bridge cycling rates, and therefore inhibits the unloaded shortening velocity (8, 51, 59). Furthermore, elevated concentrations of ADP compete with ATP for the nucleotide-binding site, thereby slowing the overall cycling rate because more cross-bridges populate an ADP-myosin-actin force-generating state. Chase and Kushmerick (5) examined whether the ADP accumulation, at values estimated from intact muscle, impacted contractile parameters, such as the unloaded shortening velocity and tension development, and found that the effect was minimal. Therefore, an ADP-dependent slowed cross-bridge cycling rate is not apparent in intact muscle because ADP accumulation is limited.
The rate of cross-bridge detachment is ultimately what causes tension to fall after a contraction. ADP has been shown to slow the rate of cross-bridge detachment after fully activated isometric contractions in skeletal (23, 30, 55) and cardiac (52) muscle fibers. This is thought to be the same reason that unloaded shortening velocity is slowed: more cross-bridges populate the strong bound ADP state longer with higher ADP concentrations. Thus, the consequences of ADP accumulation include impaired calcium uptake by the SR, and a slowed myosin cross-bridge detachment rate, both of which are factors involved in relaxation.
The purpose of this study was to examine the functional consequences of ADP accumulation on muscle in situ. To achieve this we have examined the metabolic and functional profile of AK-deficient (−/−) muscle in vivo at high-energy demands. We hypothesized that AK deficiency would result in ADP accumulation when energy demands were highest, and this would lead to the functional consequence of slowed relaxation consistent with impaired calcium uptake, and/or slowed cross-bridge kinetics. Our results demonstrate that in intact muscle, ADP can transiently increase to concentrations well in excess of 1 mM in AK−/− muscle. This degree of ADP accumulation did not modify the ability to develop force but did coincide with a substantial slowing of relaxation kinetics. These results illustrate the importance of AK and AMPD in minimizing inordinate ADP accumulation and identify the consequences of ADP accumulation on muscle function.
The design of this study was aimed at evaluating the metabolic and functional consequences of ADP accumulation in skeletal muscle. The AK1−/− mouse (25) was employed as a model capable of ADP accumulation at high-energy demands. The contractile function and energetic profile were determined at rest and during contractions within a range of isometric tetanic contraction frequencies. The frequencies of contractions chosen were based on the understanding that AK and AMPD play an important role in minimizing ADP accumulation during energetically demanding conditions. More specifically, conditions when the rate of ATP hydrolysis is out of balance with ATP supply processes. Furthermore, we also evaluated the potential consequences of ADP accumulation on cross-bridge dynamics by measuring the whole muscle unloaded shortening velocity.
AK1−/− transgenic mice and wild-type (WT) control mice were described in detail previously (25). Animals were kept in a temperature-controlled environment (22°C) with a 12:12-h light:dark cycle, and fed standard rodent chow ad libitum. The Animal Care and Use Committee at the University of Missouri-Columbia approved the animal use and care protocols.
To evaluate relevant phenotypic adaptations between AK1−/−, we performed histochemical analysis of myofibrillar ATPase and capillarity of the gastrocnemius-plantaris-soleus (GPS) muscle complex in both AK1−/− and WT muscle. For histochemical analysis, cross-sections were taken from the middle of the GPS muscle group. To determine qualitative differences in ATPase activity in the muscle groups, cross sections were stained for ATPase activity after acid preincubation (pH 4.6) according to the method described by Guth and Samaha (20). Three degrees of ATPase were apparent and the qualitative distribution of fiber types was evaluated for the soleus, plantaris, and the lateral and medial gastrocnemius muscles. Capillaries were visualized by alkaline phosphate stain as done previously (63, 64), which rendered the fiber yellow with the capillary appearing dark. Capillarity was expressed as capillary contacts per fiber. In addition, we examined two sections of the gastrocnemius muscle for citrate synthase activity, using the method described previously (53), as an index of mitochondrial content. Recent work (25, 26) has identified an increased abundance of intermyofibrillar mitochondrial content in AK1−/− gastrocnemius muscle compared with WT, as measured by electron microscopy. We sought to verify this observation in the superficial gastrocnemius (13 ± 1% of GPS mass, n = 12), which has relatively low oxidative capacity, and the remaining lateral gastrocnemius muscle (26 ± 2% of GPS mass, n = 12), a mixed oxidative capacity muscle (3).
Sixty isometric tetanic contractions (each 100 ms long) of the GPS muscle complex were elicited from an in situ muscle preparation via electrical stimulation (3- to 5-V stimulation, 0.05-ms square wave, at a frequency of 150 Hz with the use of a Grass S48 stimulator) of the sciatic nerve at contraction frequencies of 30, 60, 90, and 120 tetanic contractions/min. The in situ muscle preparation was similar to that used previously in the rat (38, 40). Mice were anesthetized with ∼70 mg/kg ip pentobarbital sodium injection. The hamstrings were cut away from the GPS and the femur was secured on the medial and lateral sides of the knee with two 16-gauge pins to prevent movement. The foot was also secured by clamping it to the platform to eliminate movement of the lower leg. The sciatic nerve was exposed, tied off, and cut to facilitate direct stimulation. The experiments were performed in a heated chamber with a temperature probe placed next to the GPS muscle complex to ensure a consistent temperature of 37°C. The GPS complex was isolated by securing the Achilles tendon to a Cambridge force transducer lever arm (model 305B, Aurora Scientific) with a shortened lever arm to permit a calibrated range of tension of 0–700 g. Supplemental oxygen (100%) was supplied across the nose of the animal throughout the procedure.
Muscle sections of the gastrocnemius muscle mentioned above were quick frozen within 3–5 s using aluminum tongs cooled in liquid nitrogen. In addition, the remainder of the GPS was taken for analysis of total muscle water weight. Metabolites relevant to high-energy phosphate metabolism (ATP, ADP, AMP, IMP, PCr, Cr, and lactate) were measured from the mixed lateral gastrocnemius muscle section mentioned above at three time points during the stimulation of the GPS corresponding to 20, 40, and 60 contractions. After the contracted muscle was frozen, corresponding muscle sections were taken from the resting contralateral leg. Muscle samples were stored at −80°C until use.
Metabolite analysis were made possible by homogenizing muscle in cold 3.5% perchloric acid, followed by centrifugation to remove protein and neutralization with tri-n-octylamine and 1,1,2-trichlorotrifluoroethane (6). Adenine nucleotides (ATP, ADP, and AMP) and IMP were quantified by reverse-phase HPLC, as described previously (56). PCr and creatine (Cr) concentrations were quantified by ion exchange HPLC (62). Metabolites are expressed as micromoles per gram wet weight and corrected to a constant total water content of 76%. Lactate content was determined by measuring NADH fluorescence using the lactate dehydrogenase reaction. The net accumulation of ADP during contractions was calculated by taking the difference of the contracted muscle measure from the resting average.
To assess the energy state of WT and AK1−/− muscles, the ΔGATP was calculated both from resting gastrocnemius, and from contracting muscles with the ΔG° of −32 kJ/mol (42). The free ADP was estimated either from the known stoichiometry with the CK reaction (57), or in contracting muscle the difference in total ADP from rested ADP was used. The concentration of Pi was estimated in resting muscle based on previously reported values in mouse skeletal muscle (29) and in contracting muscle, Pi was estimated by the addition of the resting content to the decline in PCr, which has been found to mirror the accumulation of Pi (41). An additional correction for the loss of PCr (increase in Pi) that occurs with muscle freezing was also used to calculate both free ADP at rest and Pi accumulation during contractions (1).
15 for review). Relaxation from tetanic contraction occurs in at least two different phases, an initial phase characteristic of uniform isometric sarcomeres, followed by an exponential fall in force characteristic of nonuniform sarcomeres (see Ref. 16 for review). Thus we measured an early relaxation time (time for force to fall 5% of force at the end of stimulation) and a late relaxation time (time for force to fall from 50 to 25% of force at the end of stimulation) to have an index of relaxation during both of these phases. The exponential portion of the relaxation data was fit best by a single exponential function with the use of SigmaPlot version 7.101. Calculations were made on all contractions with the aid of a script created to process the text output from the Chart software.n = 14) determined in preliminary work. The peak force of each contraction was evaluated as well as the peak rate of force development. The peak rate of force development was normalized to the tension developed for each given contraction. This was necessary because the fall in the rate of force development parallels the fall in tension that occurs with fatigue (see Ref.
Unloaded shortening velocity.
Unloaded shortening velocity of the GPS complex was examined using the slack test method described previously (13), as applied to whole muscle in situ (10). Force and lever arm position data were acquired at a frequency of 20 kHz. The lever arm position was controlled with the use of Labview software driving a PCI-MIO-16E-4 data-acquisition board (National Instruments). After a time for full development of maximal isometric force (100 ms in nonfatigued muscle and 70 ms in fatigued muscle), the lever arm was slacked by a specific distance at a rate in excess of 500 mm/s. The duration of the stimulus was always set to end ∼30–50 ms after the initiation of the slack. Force fell from full tetanic force to 0 within 8 ms. The slack time (the time from the beginning of the slack to when force development occurred) was plotted against the distance of the slack. Unloaded shortening velocity was calculated from the slope of linear least-squares regression of 4–6 points (Δlength/time to redevelop force).
Two-way (contraction number × intensity) ANOVAs were performed within each genotype for ADP accumulation. Student's t-tests were used for all other comparisons with a Bonferroni test for multiple comparisons.
The phenotype of the AK1−/− muscle compared with WT indicates no apparent differences in myofibrillar ATPase (Table 1) and oxidative capacity, with the exception of a small superficial section (13% of GPS) of the gastrocnemius muscle. In addition, no differences in capillary contacts per muscle fiber were seen in the soleus and plantaris. However, a significant increase in capillary contacts per fiber was apparent in the medial gastrocnemius (P < 0.05) of AK1−/− muscles compared with WT. The lateral gastrocnemius muscle tended to have more capillary contacts per fiber, but did not reach statistical significance. Citrate synthase activity, an index of muscle oxidative capacity, was greater in a superficial section of the gastrocnemius 32.8 ± 3.0 and 23.7 ± 1.3 μmol·g−1·min−1 (P < 0.05) in AK1−/− and WT muscle, respectively, consistent with previous reports of increased intermyofibrillar mitochondrial content (25, 26). In the remaining lateral mixed-fiber gastrocnemius muscle section, no difference was observed (42.3 ± 1.9 and 38.5 ± 1.9 μmol·g−1·min−1 in AK1−/− and WT, respectively). Thus the phenotype of the AK1−/− muscle compared with WT indicates an increase in oxidative capacity in a small superficial section (13% of GPS) of the gastrocnemius and an increase in muscle capillarity.
The resting muscle contents of ATP, ADP, PCr, and Cr were similar between AK1−/− and WT in mixed gastrocnemius muscle; however, the total content of AMP was greater (P < 0.05) in AK1−/− muscle than that found in WT muscle (Table 2). Therefore, apart from AMP, AK1 deletion does not appear to result in an altered resting metabolic phenotype.
Overall, the content of the adenine nucleotides plus IMP remained constant across stimulation conditions, consistent with previous reports in rat muscle (39) [7.36 ± 0.22, 7.15 ± 0.19, and 7.11 ± 0.23 (means ± SE) for 20, 40, and 60 contractions, respectively]. However, a marked redistribution of adenine nucleotides and IMP was most apparent at 120 tetani/min (Fig. 1). A clear deficiency in IMP formation is apparent in the AK1−/− muscles during the high-energy demands present in muscle contracting at 120 tetani/min (Fig. 1A). The deficiency in IMP formation in AK1−/− muscle was matched by a tempered decline in ATP and a marked increase in ADP compared with WT muscles (Fig. 1, A and B). Furthermore, the initial rate of IMP formation in AK1−/− muscle over the contraction frequencies 60, 90, and 120 contractions/min was only 10–15% of that observed in WT muscle (Fig. 2A). In contrast to IMP, lactate accumulation was not significantly different between AK1−/− and WT muscle during these stimulation conditions (Fig. 2B). Consequently, the lack of AK capacity clearly presented the muscle with a condition deficient in AMP deamination as measured by the lack of IMP accumulation. In addition, the similarity of the initial rates of lactate accumulation, particularly at the highest contraction frequency measured after ∼10 s of contractions, suggests that the glycolytic flux was not different between the AK1−/− and the WT muscles even in the presence of large differences in IMP formation rates. This depreciates the importance of IMP in controlling glycolytic flux, as has been hypothesized to occur (31).
Differences in ADP were also evident in WT vs. AK1−/− muscle over a range of contraction frequencies (Table 3). Most of the ADP in rested muscle is bound to actin. Thus, to get an estimate of free ADP, the resting content of 0.90 μmol/g wet wt (Table 2) was subtracted from the total ADP content measured in the stimulated muscle. The large changes in ADP above the total resting ADP content likely reflect free ADP and not a rapidly changing bound pool of ADP. After ∼40 contractions at the most demanding contraction frequency (120 tetani/min), the accumulation of ADP in WT muscle was not different from 0 (0.007 ± 0.094) compared with nearly 1 (0.90 ± 0.15) μmol/g wet wt in AK1−/− muscle. The differences in ADP accumulation in the AK1−/− group are even more evident when put in the context of the rate of ADP accumulation (Fig. 3). Significant differences in the initial ADP accumulation rates were apparent at the more demanding contraction frequencies of 90 and 120 tetani/min but were not apparent at 30 and 60 tetani/min (Fig. 3). ADP accumulation increased overall with increasing contraction frequency in AK1−/− muscle (Table 3). Table 3 presents ADP accumulation at different contraction frequencies. No significant effect of the number of contractions on ADP accumulation was found within each contraction frequency, thus mean ADP accumulation within contraction frequency is reported here. The highest ADP accumulation in Table 3 corresponds to an intracellular concentration of ADP accumulation in AK1−/− muscle that is ∼1.5 mM at the most intense contraction frequency [calculation based on ADP distribution in an intracellular water content of 0.607 ml/g muscle wet weight (22)].
As a quantitative assessment of the difference in energy state of muscle with and without the large ADP accumulation, we calculated the ΔGATP in both rested muscle and muscle contracting at 120 tetani/min. The ΔGATP at rest was not different between groups; however, during contractions at 120 tetani/min, it was significantly less favorable in the AK1−/− group compared with WT (Table 4). Therefore, the energetic state of the AK1−/− muscle at the highest frequencies of contraction is significantly more depressed from rest, than the WT muscles at the same energy demands.
The initial peak tetanic tension was not different between WT and AK1−/− muscles (Table 5). Whereas tetanic contractions at an increasing frequency induced progressively greater fatigue, the pattern of fatigue was similar between groups over the course of the 60 contractions; however, there was an apparent tendency for the AK1−/− muscle to sustain greater force as time proceeded (Fig. 4, A–D). The mean performance of the first and fortieth individual contractions of both groups illustrates a more complete picture of contractile function (Fig. 4, E and F). The shapes of the initial contractions were similar between groups. By contraction 40, a few differences became apparent at the most demanding contraction frequencies of 90 and 120 tetani/min: 1) force development in AK1−/− appeared to be slowed at 120 tetani/min, as evidenced by the pattern of the tension over the initial ∼40 ms of the contraction in both WT and AK1−/−; 2) the force of the contraction is sustained until the end of stimulation (100 ms) in AK1−/−, whereas in WT muscle tension is already starting to fall; and 3) the time of relaxation is considerably longer in AK1−/− (Fig. 4, E and F); for example, at 200 ms in Fig. 4E the net force remaining in the AK1−/− group is more than double the value in the WT group. Thus, whereas the peak force developed and the pattern of fatigue was not much different between the groups, a clear slowing of the relaxation was apparent at the highest contraction frequencies.
The fall in tension after the end of stimulation occurred in two distinct phases (Fig. 4, E and F): an early period of delay before the exponential decline in tension, and an exponential phase that was best fit by a single exponential function and characterized as partial (50–75%) relaxation, as done previously in whole muscle (10, 14, 18). The early relaxation times for the first contractions (nonfatigued muscle) were not different between groups; however, the AK−/− group late relaxation times were slightly faster than the WT group (see Table 5). At the most demanding contraction frequency of 120 tetani/min, the early relaxation phase appeared to slow after ∼20 contractions, and the difference between the two groups expanded until about contraction 35 and then subsided (Fig. 5A). This delay in this early phase was before the exponential decline in tension. A similar but less exaggerated pattern was observed at the contraction frequency of 90 tetani/min with the slowing of early relaxation time occurring after ∼30 contractions (Fig. 5B). No differences in the early relaxation times were apparent at contraction frequencies of 30 and 60 tetani/min (Fig. 5, C and D), illustrating that the slowing of early relaxation is dependent on having sufficiently high-energy demands. A transient slowing of the late relaxation phase, with the largest difference occurring at 40 contractions, was also apparent at the demanding contraction frequencies of 90 and 120 tetani/min (Fig. 5, E and F). Differences are also apparent when relaxation rates are expressed as exponential rate constants [0.025 ± 0.001 and 0.016 ± 0.002 ms−1 in WT and AK1−/− muscles, respectively, contracting 120 tetani/min at contraction 40 (n = 9–10 animals each)]. Compared with the early relaxation phase, the AK1−/− group displayed somewhat faster late relaxation kinetics compared with the WT group before 20 contractions at contraction frequencies of 90 and 120 tetani/min (Fig. 5, E and F), and almost throughout the contractions examined at the lower contraction frequencies of 30 and 60 tetani/min (Fig. 5, G and H). The differences apparent in the late relaxation phase of the AK1−/− group suggest a phenotypic adaptation for faster late relaxation kinetics that slows dramatically when energy demands are sufficiently high. Therefore, further inspection of the slowing of relaxation that was apparent in Fig. 4 revealed a transient slowing of relaxation in both the early phase and a late phase of relaxation at the highest contraction frequencies of stimulation.
Rate of force development.
As mentioned above, one of the differences apparent between groups was the pattern of force development that occurred, particularly at the two highest contraction frequencies (Fig. 4, E and F). This was apparent when the peak rate of force development is compared between the groups over the contractions where the largest functional differences were apparent between the groups (contractions 35–40). In fact, an inverse relationship with increasing contraction frequency was apparent in the AK1−/− group and not in the WT muscle (Fig. 6). This pattern of response was observed even when the rate values were normalized to forces at the same time point (50 ms) rather than the peak of the contraction, which occurred later in the AK−/− mice (not shown). Furthermore, this inverse relationship occurred over roughly the same contractions where the relaxation times were also further depressed with increasing contraction frequency (Fig. 5, A, B, E, and F). Such patterns with increasing contraction frequency were not seen in the WT muscle group. Thus at the same time relaxation times were slowing, a depression in the peak rate of force development was apparent.
Unloaded shortening velocity.
The unloaded shorting velocity was measured to determine whether the ADP accumulation observed in the AK1−/− group would be sufficient to slow actin myosin cross-bridge dynamics in intact muscle. On the basis of the ADP accumulation observed in the experiments described above, we examined the whole muscle unloaded shortening velocity of the GPS muscle complex during infrequent nonfatiguing contractions and at roughly 35–40 contractions at 120 tetani/min (Fig. 7 and Table 6). The unloaded shortening velocities were not different between groups in fresh or fatigued (35–40 contractions at 120 tetani/min) conditions (Table 6). These results suggest that the millimolar ADP accumulation seen in AK1−/− muscle is not sufficient to slow cross-bridge dynamics enough to show a difference in the unloaded shortening velocity.
Normally, during intense contractions when there is a decline in ATP, ADP is rapidly processed through AMP via adenylate kinase (AK), then on to IMP via the AMP deaminase (AMPD) reaction. In the absence of AK, this cannot occur, and an accumulation of ADP would be expected if energy expenditure remained high. An inordinate accumulation of ADP should depress the free energy available from ATP hydrolysis (ΔGATP), affect calcium removal from the cytosol, and thereby alter muscle relaxation. With the use of a knockout mouse model of AK1 deficiency, conditions of high-energy demands induced an ADP accumulation of ∼1.5 mM. This large increase in total ADP, which likely represents free ADP, permitted us to evaluate the functional consequences of such an inordinate ADP accumulation on whole muscle contracting in situ. We found that despite the presence of >1 mM differences in ADP in AK−/− and WT control muscle, the overall pattern of muscle performance was similar (Fig. 4, A–C). However, the most striking functional difference was a transient slowing of relaxation kinetics (Fig. 5, A and B), accompanied by an attenuation of the peak rate of force development. Thus the muscle function was remarkably tolerant of ADP accumulation in dramatic excess of normal limits, if time for relaxation between contractions is sufficient.
ADP and relaxation rate.
The slowed relaxation kinetics coincided with the time of high ADP accumulation in AK−/− muscle. Muscle relaxation is ultimately due to the detachment of myosin from actin. Increasing concentrations of ADP prolongs the rate of myosin detachment after full isometric activation (23, 30, 55). If this applies to our muscle with elevated ADP, there should be a prolonged maintenance of force after the end of stimulation, similar to the delay in early relaxation apparent in AK−/− muscle (Fig. 5A). However, our findings cannot be simply attributed to altered cross-bridge cycling rate, because in intact muscle, cross-bridge detachment does not occur independent of calcium dissociation from troponin C. In fact, Luo et al. (34) demonstrated that when the affinity of troponin C for calcium is either increased or decreased, there is a corresponding increase or decrease in time for relaxation. The rate with which calcium dissociates from troponin C is a function of the rate with which the cytosolic calcium concentration declines, which occurs by pumping calcium back into the sarcoplasmic reticulum (SR). SERCA functions to sequester calcium at a high-energetic cost due to the large calcium gradient between the cytosol and the SR (a ratio near 1:10,000) (21). The energy available from ATP hydrolysis (ΔGATP) declines with increasing concentrations of ADP and Pi (57). Dawson et al. (9) found that the decline in ΔGATP correlated with an increase in relaxation time in isolated frog muscle, and interpreted the slowed muscle relaxation to be due to the progressive impairment in calcium uptake. Indeed, the decline in free energy available estimated by the ΔGATP with an inordinate ADP accumulation in our study (Table 4) is consistent with this hypothesis and may explain the delay in relaxation that we observed at high-energy demands.
The accumulation of ADP may also affect calcium sequestration in a more direct manner, without a drop in energy state. Recent work by Macdonald and Stephenson (36) found that ADP accounted for as much as a 4.4-fold reduction in SR calcium loading ability, due in part (30%) to depressed SERCA pump rate and primarily (70%) to increased calcium leak over a concentration range of 0.01 to 1 mM. Furthermore, the ADP-dependent impairment of SR calcium uptake was observed under energetically favorable conditions (ATP and Pi were maintained and ADP was manipulated) demonstrating that ADP can inhibit SR calcium uptake independent of a large decline in ΔGATP (36). Therefore, we would predict from our evidence of an elevated ADP, coupled with significant delay in muscle relaxation kinetics, that calcium removal from the cytosol of the AK1−/− muscle is delayed. Wahr et al. (58) determined that, in frog skeletal muscle, the initial phase of relaxation (characteristic of relaxation while sarcomeres remain isometric) was more sensitive to changes in calcium than the subsequent exponential phase (characteristic of relaxation when sarcomeres are not uniform). The delay in relaxation in the AK−/− muscle was first observed in the early phase after ∼20 contractions at the highest contraction frequency (Fig. 5A), followed by a delay in the late phase after 30 contractions (Fig. 5E). In addition, SERCA function may be impaired due to the decline in pH that occurs with fatigue (24, 35). We do not think the exaggerated relaxation seen with high ADP can be explained by a pH effect on SR function because the predicted pH, based on lactate accumulation, was similar in both groups. However, whether the exaggerated relaxation requires a high [H+], cannot be answered from our experiments. Furthermore, large differences in ADP between AK1−/− and WT muscles were not always coupled to differences in relaxation rates, such as at 60 contractions at 120 tetani/min (Fig. 1B and Fig. 5, A and E), where force was substantially reduced (∼56% of the force at 40 contractions was developed at 60 contractions). It is unknown whether the putative effect of high ADP requires a relatively high rate of calcium turnover (high force development). However, it seems reasonable to interpret the slowed relaxation kinetics observed in our study to be due to a transient impairment of calcium sequestration by ADP.
Unloaded shortening velocity.
The rate of ADP dissociation from myosin limits unloaded shortening velocity, which is dependent on the rate of cross-bridge detachment (8, 17, 51, 59). We found that the decline in unloaded shortening velocity of the calf muscle group was similar with or without inordinate ADP accumulation. This suggests that the ADP accumulation in AK1−/− was insufficient to meaningfully slow cross-bridge detachment, in contrast to that predicted (8, 61) or observed at lower ADP concentrations (5). Alternatively, the slowed unloaded shortening velocity that occurs with fatigue at a low pH (for review, see Ref. 15) may have obscured any ADP effect. The decline in unloaded shortening velocity that we observed with fatigue (∼50%) is in the range of values reported previously (10, 60, 61). Furthermore, whereas ADP accumulation of this degree has been found to depress the unloaded shortening velocity in single fiber preparations, this effect has been investigated at both lower concentrations of ATP and at much lower temperatures than are found in muscle contracting in situ (8). Interestingly, we did observe a significant slowing of the peak rate of force development, coincident with the delayed relaxation discussed above (Fig. 6). It is possible that this apparent slowing of peak rate of force development, reflects a depressed cross-bridge cycling rate, given that others have observed a high correlation between the peak rate of force development and the extrapolated measure of maximum shortening velocity (54). Thus a minor degree of delayed cross-bridge detachment due to ADP cannot be ruled out. However, it is likely that a direct effect of ADP on cross-bridge cycling is not the primary factor responsible for the decline in relaxation seen in the presence of high ADP.
ADP and isometric force production.
In isolated fiber preparations ADP has been shown to cause increased muscle tension due to more myosin cross-bridges populating a strongly bound conformation with actin (27, 33). This effect of ADP on muscle tension occurs in competition with ATP binding the myosin head that results in detachment from actin (27). Because the concentration of ATP remains manyfold higher than the concentration of ADP, even under the most demanding conditions, the effect of ADP on isometric tension has been predicted to be at most an ∼8% increase in tension under conditions with 1 mM ADP accumulation (27). Interestingly, we found a modest effect of improved tension maintenance within a given contraction (contraction 40, Fig. 3, E and F) in muscle with high ADP at the two most demanding contraction frequencies. This is consistent with the prediction of a relatively small effect ADP has on isometric tension (8, 27, 33). We did not, however, observe any substantial increase in the peak tension developed in muscle with high ADP throughout the contractions examined. Therefore, an inordinate accumulation of ADP does not have a large effect on whole muscle force production.
Critical assessment of our model.
Our findings should be considered in the context of the model employed in this study; whole muscle stimulated in situ, supported by intact circulation. This model does not allow for the direct control of specific metabolites like models of isolated muscle fibers and must be understood in the context of the complex architecture and mixed fiber type inherent in the whole muscle group. For instance, while the unloaded shortening velocity value that we observed in nonfatigued muscle is comparable to that in whole rat gastrocnemius muscle in situ (10), it is probably determined by the proportion of fast fibers that make up muscle (4). We believe that it is unlikely this confounds our interpretation because the fastest fibers in our preparation should be the ones with the highest ADP accumulation, due to greater ATP turnover with a relatively low mitochondrial capacity. On the other hand, the in situ calf muscle preparation has advantages that include a more physiologically relevant experiment, which preserves the structural and enzymatic integrity to evaluate functional and metabolic changes. As with any genetic knockout model, adaptations might impact research findings. In the AK1−/− mouse, an increase in intermyofibrillar mitochondrial content was reported (25, 26) in the superficial gastrocnemius muscle. Consistent with this finding, we measured a significant increase in citrate synthase activity in a small superficial section of the gastrocnemius muscle (e.g., ∼13%); however, we did not measure any significant differences in the mixed gastrocnemius muscle. We also observed significant enhancement in the capillarity, in this case throughout both heads of the gastrocnemius (Table 1). Therefore, under steady-state energy demands, where the ATP demand is adequately met by aerobic ATP supply, we would expect an enhanced ability for nutrient exchange. This, however, does not likely impact our findings, given the very demanding short-term conditions that we examined. Furthermore, we did not see any evidence of large differences in fiber-type distribution in the AK1−/−, suggesting similar ATP demands compared with WT muscle. Therefore, it is not likely that our findings are mitigated by altered fiber-type expression or arrangement, given the very similar initial contractile characteristics and unloaded shortening velocity. The only significantly different initial contractile parameter we observed was a somewhat faster (∼12%) rate of relaxation in the late, or exponential phase of relaxation in AK−/− muscle. In fact, faster late relaxation times were apparent throughout the contractions observed at the lowest energy demands (Fig. 5H). This small difference might be attributable to a slightly different muscle fiber alignment, or, more likely a small increase in fast twitch fibers, but within the variability of our measures. Because no other contractile parameter examined reflects such a shift in fiber type (rate of force development, early relaxation time, and unloaded shortening velocity), we conclude that the difference in relaxation times in both the late and early phases, between AK1−/− and WT mice at high-energy demands, are meaningful.
Human AMPD deficiency.
We believe that our results provide insight into whether functional consequences should exist due to AMPD deficiency in human muscle. While the importance of AK and AMPD in limiting ADP accumulation has been understood for many years (31, 32, 39), this study expands our understanding by illustrating what consequences ensue when this path for ADP removal is deficient. A genetic mutation in the gene encoding AMPD1 results in AMPD enzyme deficiency and has been described as “the most common muscle enzyme defect in humans” (19). AMPD enzyme deficiency is found at a frequency of roughly 2% in muscle biopsies (see Refs. 19, 43, and 44 for reviews), and has been thought to cause AMP and ADP accumulation at high-energy demands, if the deficiency were great (49). The presence of AMPD deficiency is often associated with symptoms of exertional myalgia, such as muscle cramps and stiffness, as well as exaggerated fatigue and delayed muscle relaxation (11, 49); however, deficits in muscle function are not necessarily independent of other myopathies (37). In fact, a large percentage of individuals with AMPD deficiency are asymptomatic. Indeed, Norman et al. (45) examined muscle power output using a short explosive exercise bout in healthy asymptomatic AMPD-deficient subjects and found no difference in maximum power output compared with normal subjects (45). Our results could reconcile these disparate findings. Even if AMPD deficiency caused an ADP accumulation similar to what we have observed in this study, muscle force is expected to be normal, similar to what was reported previously (45). However, ADP accumulation could delay relaxation as tended to be evident in at least one study (11) due to an expected slowed rate of calcium clearance from the cytosol. While a deficient calcium uptake can lead to contracture and muscle soreness (2), it is unlikely that this would happen with AMPD deficiency. We did not see it here with AK deficiency, where the ADP accumulation is likely far greater than any expected with AMPD deficiency alone. Any putative effects of AMPD deficiency, however, would only be expected at an extreme energy imbalance as in the study by Norman et al. (45), and not during easier submaximal exercise, where steady-state aerobic metabolism is expected to be sufficient for ADP rephosphorylation. Therefore, our results would predict that muscle performance with AMPD deficiency would be normal under most energy demands, in the absence of other myopathies.
In summary, we have shown that in the presence of high-energy demands, AK deficiency results in minimal IMP formation and inordinate ADP accumulation. ADP accumulation of ∼1.5 mM in intact muscle did not impair, or significantly enhance the peak tension developed or the fall in tension that occurred with fatigue. Furthermore, the clearest functional consequence of the ADP accumulation was delayed early and late relaxation kinetics, consistent with impaired calcium uptake by the SR. The results from this study directly illustrate the effectiveness of AK and AMPD in the management of ADP during high-energy demands, and demonstrate the functional tolerance to ADP accumulation in intact muscle.
This work was supported by National Institutes of Health Grant AR-21617 and Netherlands Organization for Scientific Research, Council for Medical and Health Research Grant 901-01-095.
We thank Dr. Bé Wieringa for providing the AK1−/− mouse model used in this study. We thank Jerry Hancock, Anish Nayani, and Rob Humfield for help with Java and Labview programming, and Jane Chen, Hong Song, and Catharine Clark for technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society