INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

REGULAR CONTRACTILE ACTIVITY is a potentially potent stimulus for altering the composition, structure, and function of the muscle cell. Nowhere is this more evident than with the chronic low-frequency electrical stimulation (CLFS) model in which contractile activity is characteristically induced for 12-24 h/day over several weeks. Induced patterns of submaximal contractions of this nature typically applied to muscles composed of a predominance of fast-twitch (type II) fibers result in extensive reorganization of both the excitation and contraction processes and the energy metabolic pathways involved in energy supply. Although the magnitude of the adaptations vary, qualitatively similar adaptations appear to occur between species (dog, rat, mouse, rabbit) (29).

Despite the fact that the CLFS model has proved invaluable for studying the limits of phenotypic plasticity in skeletal muscle and the factors regulating the expression of a large number of proteins involved in fiber-type transformation, it is nonphysiological (30). In physiological models, voluntary activity is used to elicit training adaptations. Voluntary training typically involves a relatively brief session of exercise followed by a prolonged recovery period, which could last for 2-3 days depending on the training frequency. Moreover, it is not clear how the intracellular signals mediating altered protein expression in CLFS translate into different programs of voluntary training. Unlike CLFS, the properties that have been explored with voluntary training have been much more limited. To a large extent, the focus has been directed toward the energy metabolic pathways and more recently the contractile protein myosin. The processes involved in excitation-contraction coupling, namely those involved in the transmission of the neural signal to the interior of the fiber, and which result in an increase in the free cytosolic Ca²⁺ ([Ca²⁺]_f) in signal involved in myofibrillar activation and muscle contraction, has received scant attention.

One vital organelle that is the primary regulator of [Ca²⁺]_f levels in the skeletal muscle cell is the sarcoplasmic reticulum (SR). The SR is a membranous network that envelops the myofibrils and acts as a storage depot for Ca²⁺. Also embedded in the SR is a collection of special proteins that regulate the release of Ca²⁺ into the cytosol and the sequestration of Ca²⁺ back into the lumen of the SR. The ryanodine receptor (RyR) or Ca²⁺-release channel and the SR Ca²⁺-ATPase (SERCA) are the principal proteins involved in Ca²⁺ release and Ca²⁺ uptake, respectively. Both RyR and SERCA exist as several isoforms in mammalian tissue (1, 8).

As with a variety of other cellular proteins, the proteins involved in SR Ca²⁺ cycling can also undergo extensive replacement when CLFS is administered to fast-twitch (type II)-based muscles. Provided that CLFS is of sufficient daily volume and duration, RyR protein content decreases dramatically (9, 14, 26) without changes in RyR1 to RyR2, the isoform that predominates in heart muscles (8). Although not measured, the downregulation in RyR would be expected to be accompanied by a pronounced decrease in Ca²⁺-release kinetics (8). The CLFS model can also induce large reductions in SERCA content. However, unlike RyR, a shift from SERCA1, the predominant isoform observed in type II muscle fibers, to SERCA2a, the major isoform that exits in slow-twitch (type I) fibers, occurs (3, 14, 26, 27). Because the alteration in SERCA isoform content with CLFS might be expected to change, the Ca²⁺ affinity as measured by the Hill coefficient and Ca²⁺ concentration at half-maximal activity (pCa₅₀), it is surprising that no study has apparently addressed this issue (29). Similarly, the coupling ratio, defined as the ratio between Ca²⁺ uptake and Ca²⁺-ATPase activity, appears unexplored with CLFS (3). On the basis of the adaptations that occur with CLFS, it is reasonable to suggest that the functional alterations in SR Ca²⁺ cycling occur, in large part, as a result of a reduction in protein content both to the RyR and SERCA.

In contrast to CLFS, SR adaptations to voluntary training remain poorly characterized, particularly with regard to the relationship between the nature of the contractile stimulus and the resulting alterations in protein expression and function. At least in terms of the contractile intensity and the strain put on the excitation-contraction processes and metabolic flux, low-intensity aerobic running would appear to have a close parallel to CLFS. As a consequence, the adaptations that occur with a training program of this nature would be expected to resemble, at least qualitatively, those occurring with CLFS. There is some evidence that this occurs. In one of the earliest studies using treadmill running in rats, Kim et al. (17) reported decreases in maximal SR Ca²⁺ uptake (V_max) and in Ca²⁺ affinity but only in tissue composed of a predominance of fast-twitch, low-oxidative fibers and not in tissue of predominant fast-twitch, high-oxidative fibers or slow-twitch-based muscle. That the reduction in V_max is due at least in part to a reduction in SR Ca²⁺-ATPase protein is supported by the findings of Green et al. (12), who reported reductions in Ca²⁺-ATPase protein content after a similar but more extreme running program also using rats.

Additional evidence that the adaptations in the SR may be specific to the nature of the training has been recently published using sprint training in humans (28). In this study, large increases in both SERCA1 and SERCA2 protein content were observed in tissue obtained from the vastus lateralis. Surprisingly, however, the increases in the Ca²⁺-ATPase protein were not accompanied by increases in maximal Ca²⁺-ATPase activity or in Ca²⁺-uptake kinetics. These investigators also reported increases in SR Ca²⁺-release rates that were accompanied by increases in RyR protein content as measured by Western blot but not by [³H]ryanodine binding (28). These results suggest that modification of the protein levels of the SR may occur independently of changes in the functional properties of the SR. At present it is unclear whether this effect is species specific or due to the nature of the exercise used for training.

The purpose of this study was to investigate the alterations in the Ca²⁺-exchange properties of the SR in human muscle in response to a training program involving prolonged aerobic-based exercise. We have hypothesized that, similar to CLFS, reductions in Ca²⁺ uptake and Ca²⁺-ATPase would occur in conjunction with a shift toward SERCA2a, the predominant isoform of type 1 fibers, and that these adaptations would occur independently of changes in Ca²⁺ sensitivity or coupling ratios. Moreover, we have also postulated that the adaptations in Ca²⁺ sequestration would be accompanied by reductions in Ca²⁺-release kinetics.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Ten healthy men (age 21.4 ± 0.3 yr; body mass 76.8 ± 2.2 kg; means ± SE) volunteered for the study. All participants were active but not engaged in exercise, either high resistance or endurance, on a regular basis for at least 6 mo before the beginning of the study. Written consent was obtained from all volunteers as required after approval of the study by the Office of Human Research.

Experimental design. The training program consisted of 10 wk of single-leg submaximal cycle performed in the recumbent position. Participants cycled at 60 revolutions/min at a power output that initially corresponded to 75% of the pretraining peak power output. Training initially consisted of 3 × 30 min sessions/wk, gradually progressing to 5 × 60 min sessions/wk. The power output was adjusted throughout the training to maintain individual target heart rates in the range of 140 and 160 beats/min. The leg assigned for training, dominant vs. nondominant, was randomly assigned. The dominant leg was selected on the basis of the leg preferred in a vertical pump with a single foot takeoff. For each participant, the other leg served as the untrained control.

Measurements of SR function. Measurements of Ca²⁺-ATPase, Ca²⁺ uptake, and Ca²⁺ release were performed on stored homogenates previously made from tissue samples prepared immediately after extraction from the vastus lateralis and stored at 80°. Homogenates were prepared by use of 11:1 (vol/wt) dilution of buffer containing (in mM) 200 sucrose, 40 L-histidine, 1 EDTA, 10 NaN₃, and 1 DTT (pH 7.8) by using a hand-held glass homogenizer (Kontes, Duall 20) (24). The homogenate was separated into a number of aliquots, quick-frozen in liquid N₂, and stored at 80°C. Immediately before analysis, aliquots were thawed and processed for specific measurements.

Ca²+-ATPase activity. Total Ca²⁺-ATPase activity was measured by using the spectrophotometric assay developed by Simonides and van Hardeveld (37) with minor modifications. The reaction buffer contained (in mM) 200 KCl, 20 HEPES, 15 MgCl₂, 1 EGTA, 10 NaN₃, 5 ATP, and 10 phospho(enol)pyruvate (pH 7.0). The pH was adjusted at 37°C. Just before starting the reaction, 18 U/ml lactate dehydrogenase (LDH), 18 U/ml pyruvate kinase (PK), 33 µl homogenate, 1 µM Ca²⁺ ionophore A-23187 (Sigma C-7522), and 0.3 mM NADH were added to a cuvette containing 1 ml of reaction buffer. Assays were performed in duplicate at 37°C and 340 nm (Shimadzu UV 160) with 1 mg wet wt of tissue per assay. After the recording of baseline absorbance and fluorescence of NADH for ~1 min, the reaction was initiated by addition of 100 mM CaCl₂ and was monitored for 1-2 min. Maximal Ca²⁺-ATPase activity and Ca²⁺-dependency of Ca²⁺-ATPase activity were assessed by using CaCl₂ in 0.5-µl additions. Ca²⁺-ATPase activity increases with [Ca²⁺]_f, until a plateau occurs, once maximal activity is reached. The [Ca²⁺]_f corresponding to each CaCl₂ addition was assessed separately by using dual-wavelength spectrofluorometry and the Ca²⁺-fluorescent dye indo-1. The measurement of [Ca²⁺]_f using this procedure is based on the difference in maximal emission wavelengths between the Ca²⁺-bound-indo-1 complex and the Ca²⁺-free-indo-1 complex. An excitation wavelength of 355 nm was used, and the emission maxima were recorded at 405 and 485 nm for Ca²⁺-bound (G) and Ca²⁺-free (F) to indo-1. The ratio of F to G decreases during Ca²⁺ uptake and was used to calculate [Ca²⁺]_f. Felix software (Photon Technology International) was used to calculate the ionized Ca²⁺ concentration by use of Eq. 2.5 of Grynkiewicz et al. (13)
where K_d is the equilibrium constant for the interaction between Ca²⁺ and indo-1, R_min is the minimum value of R with the addition of 250 µM EGTA, G_max is the maximum value of G with the addition of 250 µM EGTA, G_min is the minimum value of G with the addition of 1 mM CaCl₂, and R_max is the maximum value of R with the addition of 1 mM CaCl₂. Simultaneous photon counts per second were recorded for both emission wavelengths. Before the experiment, Ca²⁺-independent background fluorescence was recorded in the reaction mediums in the absence of indo-1 and was used to correct each assay. The dissociation constant (K_d) that was used for the interaction between Ca²⁺ and indo-1 was 250 nM (13). It should be emphasized that in homogenates the actual K_d is significantly affected by the protein composition. Ca²⁺-ATPase activity was obtained by subtracting basal activity from total (Mg²⁺-activated) ATPase activity. Basal activity was measured in the presence of 40 µM cyclopiazonic acid, which completely inhibits SR Ca²⁺-ATPase activity (35). The validity of the SR Ca²⁺-ATPase assay using whole muscle homogenates has been previously demonstrated in both animals (37) and humans (31).

Ca²+ uptake. Oxalate-supported Ca²⁺ uptake was measured by using the Ca²⁺-fluorescent dye indo-1 according to the methods of O'Brien and colleagues (25) as modified by our group (40). Fluorescence measurements were made on a spectrofluorometer (RatioMaster system, Photon Technology International) equipped with dual-emission monochromators. The measurement of [Ca²⁺]_f using the indo-1 procedure is based on the difference in the maximal emission wavelengths between the Ca²⁺-bound form of indo-1 and the Ca²⁺-free form. The excitation wavelength was 355 nm, and the emission maxima were 485 and 405 nm for Ca²⁺-free (G) and Ca²⁺-bound (F) indo-1, respectively (24). Photon counts per second were recorded simultaneously for both emission wavelengths. The Ca²⁺-independent (background) fluorescence was measured in the reaction medium (without indo-1) at each emission wavelength before the experiment was started. Background fluorescence was automatically corrected with Felix software (Photon Technology International) before the start of each assay.

Ca²+ release. Ca²⁺ release was measured on muscle homogenates in conjunction with Ca²⁺ uptake according to the methods of Ruell et al. (31) with minor modifications (40). Ca²⁺ release was measured in a single assay, immediately after Ca²⁺ uptake. After active loading by the SR when the [Ca²⁺]_f declined to a plateau, 3 µl of AgNO₃ were added to give a final concentration of 141 µM. The reaction was then allowed to proceed for ~3 min. With the addition of AgNO₃, Ca²⁺ release generally proceeded in two phases as observed by our group (40) and previously by Ruell et al. (31). There was an initial rapid release (phase 1) followed by a slower, more prolonged rate of release (phase 2). As with the Ca²⁺ uptake, the curve representing [Ca²⁺]_f vs. time was smoothed over 21 points and differentiated. The maximal rate of Ca²⁺ release was calculated by taking the maximum positive derivative for the second phase (31). Our laboratory has previously shown that, with AgNO₃ as the releasing agent, phase 2 is regulated by the Ca²⁺-release channel (39).

Western blots. To assess the effects of training on Ca²⁺-ATPase isoform distribution (SERCA1 and SERCA2a), electrophoresis and Western blotting were performed. These analyses were conducted on the supernatant (postnuclear homogenate) obtained after a one-step centrifugation process prepared from frozen tissue. The homogenates were prepared in buffer consisting of 5 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (pH 7.5), 250 mM sucrose, 0.2% NaN₃, and 100 µM phenylmethylsulfonyl fluoride in a ratio of 15:1 (vol/wt). Extraction was accomplished by adding an equal volume of buffer containing 10 mM sodium phosphate (pH 7.4), 150 mM NaCl₂, 2% Triton X-100, 2% deoxycholate, 0.2% SDS, 200 µM phenylmethylsulfonyl fluoride, and 1,000 IU aprotinin. This mixture was centrifuged at 10,000 g for 15 min, and the supernatant was withdrawn and stored at 80°C until analyzed. Wu and Lytton (43) have reported that 95-99% of the SERCA protein remains in the supernatant after this procedure.

Statistical procedures. The data were analyzed with both one-way and two-way ANOVA procedures for repeated measures. One-way ANOVA procedures were used to examine the differences between the two legs (trained vs. untrained) on a single variable. When an additional variable was present (i.e., [Ca²⁺]_f vs. leg), a two-way ANOVA was employed. When significant differences were found, Newman-Keuls procedures as appropriate were used to locate differences between specific means. Significance was set at the 0.05 level.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

O_2 peak. In the trained leg, O_2 peak increased (P < 0.05) by 7% (2.98 ± 0.12 vs. 3.19 ± 0.13 l/min). No differences in O_2 peak were found in the untrained leg before and after training (3.0 ± 0.14 vs. 3.09 ± 0.14 l/min). Differences in O_2 peak between the control and experimental legs before training were not significant.

SR. V_max in the trained leg was 18.8% lower than in the untrained leg (Table 1). Of the two properties used to assess the kinetic characteristics of the enzyme, pCa₅₀ and the Hill coefficient, only the pCa₅₀ was altered with training (Table 1: Fig. 1). Training resulted in a small but significant (P < 0.05) increase in pCa₅₀. It should be noted that basal ATPase activity was not different between the control and trained legs (23.8 ± 4.0 vs. 18.6 ± 6.1 µmol · mg protein¹ · min¹).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Ca2+-dependent ATPase activity determined in vastus lateralis of control and trained legs. Values are means ± SE (n = 7). See Table 1 for description of the kinetic properties of the enzyme that were obtained from these curves.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Ca-dependent Ca2+ uptake determined in vastus lateralis of control and trained legs. Values are means ± SE (n = 10). Ca2+ uptake was determined at 5 different free cytosolic Ca2+ concentration ([Ca2+]f) levels (250, 500, 1,000, 1,500, and 1,800 nM). * Significantly different from control (P < 0.05).

1

1

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Sample tracing showing the [Ca2+]f vs. time. Both Ca2+ uptake and Ca2+ release for a representative volunteer are indicated. The values are for homogenates from control leg and trained leg. Note the initial additions of Ca2+. Silver nitrate (AgNO3) was used to initiate Ca2+ release. The maximum rate of Ca2+ uptake and Ca2+ release (phase 2) were determined by differentiating linear fit to the curve after ATP and AgNO3 additions.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   Relative protein contents of sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA1 and SERCA2a) isoforms in vastus lateralis of control and trained legs. Values are means ± SE (n = 10). % Control, percent change relative to the control leg. * Significantly different (P < 0.05) from control.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Western blot analyses of fast (SERCA1) and slow (SERCA2a) Ca2+-ATPase isoforms protein with use of monoclonal antibodies A52 and 7E6. Analysis was performed on postnuclear fractions (see METHODS). For postnuclear protein, 20 µg were loaded per lane. Blots represent the measurement of 3 subjects (1, 2, 3) for SERCA1 and SERCA2a. STD, standard; T, trained leg; C, control leg.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

As hypothesized, we have found that a 10-wk program of prolonged submaximal exercise results in adaptations to the SR in the vastus lateralis muscle. The adaptations consisted of reductions in Ca²⁺-ATPase activity, Ca²⁺ uptake, Ca²⁺ release, and SERCA1 content. Collectively, these adaptations are consistent with a shift in the properties of the SR toward those typically observed in slow-twitch or type 1 fibers.

In this study, all of the SR functional measurements were performed on whole muscle homogenates prepared from tissue samples obtained by needle biopsy. The measurement of SR Ca²⁺-ATPase activity can be problematic under such circumstances given the presence of other ATPases in the homogenate. However, the potential limitations of other contaminating ATPases have been circumvented by using a procedure development by Simonides and van Hardeveld (37) and subsequently modified by Reull et al. (31) for application to human tissue. This procedure provides for the selective exclusion of other interfering ATPases. The specificity of the procedure has been clearly indicated by demonstrating the complete loss of Ca²⁺-ATPase activity with cyclopiazonic acid, an inhibitor of the enzyme. Moreover, the assay is a regenerating assay in which ATP is continually produced by the addition of the glycolytic enzymes PK and LDH. The oxidation of NADH is measured fluorometrically. This procedure prevents the accumulation of ADP and P_i, which could inhibit the enzyme. Because the myokinase inhibitor AP5A had no effect on enzyme activity (unpublished observations), ADP loss via AMP formation would not appear important.

Although the plasticity of the SR has been demonstrated in animals to a variety of mechanically based chronic perturbations such as induced contractile activity (14), hindlimb unweighting (34), and overload induced by surgical removal of synergistic muscles (16), this appears to be one of the few longitudinal studies to examine SR changes in response to submaximal training in previously untrained humans. Adaptations at the level of the SR in response to regular exercise were expected given the extensive changes in Ca²⁺-regulatory membrane processes and proteins that are elicited by chronic electrical stimulation. These changes are typically induced in muscle of rats (36), rabbits (14), and dogs (26) consisting of a predominance of fast-twitch fibers. In the present study, the nature of the adaptations that occurred in the SR in specific fiber types is unknown, given the mixed distribution of fiber types in the vastus lateralis. It should be emphasized that not all studies report adaptations in the SR in human muscle with regular submaximal exercise. In preliminary work, our laboratory failed to detect changes in maximal Ca²⁺-ATPase activity (11), whereas others have failed to observe changes in Ca²⁺-ATPase concentration (as assessed by Ca²⁺-dependent steady-state phosphorylation) (21). Although the reasons underlying the contradictory findings remain unclear, particularly with regard to our laboratory's initial study in this area, the fact that previously trained volunteers were used by Madsen et al. (21) might explain the lack of an effect with additional training. In a recent study (19), selected SR properties in the vastus lateralis muscle were compared between untrained control subjects and endurance-trained and resistance-trained athletes. Compared with untrained control subjects, endurance-trained athletes demonstrated lower Ca²⁺ uptake and Ca²⁺ release. Given the lower type II fiber-type distribution observed in the endurance-trained athletes and the fact that endurance training does not appear to induce significant shifts in the major fiber types (32), these results suggest that the differences observed in SR functional measurements might, in large part, be explained by inherent differences in major fiber-type distribution.

The most inviting mechanism to explain the adaptations that resulted in the SR in response to training is a reduction in the content of specific proteins. A reduction in Ca²⁺-ATPase protein content is clearly indicated from the Western blot analyses and densitometric measurements in which a relative reduction in SERCA1, the isoform typically observed in fast-twitch muscle, occurs with training but not in SERCA2a, the predominant isoform in slow-twitch muscle. The reduction in Ca²⁺-ATPase protein level was accompanied by a reduction in V_max that was observed in the trained leg compared with the control leg. It is noteworthy that the decrease in V_max occurred in the absence of changes in the Hill coefficient, a measure of the Ca²⁺-binding sensitivity. In contrast to the Hill coefficient, a small but significant increase in pCa₅₀ was noted with training. This finding indicates that, after training, a lower [Ca²⁺]_f level can recruit 50% of V_max. Interesting, we have also found that the training program resulted in a significantly lower homogenate [Ca²⁺]_f in the trained leg compared with the control leg (unpublished observations).

Reductions in Ca²⁺-ATPase protein levels could also explain the rightward shift in [Ca²⁺]_f vs. Ca²⁺ uptake relationship after training, including the reduction in maximal Ca²⁺ uptake. Because no difference in the coupling ratio was observed between the trained and untrained legs, the apparent downregulation in Ca²⁺-ATPase did not compromise the efficiency of Ca²⁺ transport into the SR. Our findings would also suggest that training did not result in increased vesicle Ca²⁺ leakiness, abnormal enzyme expression, or changes in relative phospholipid content of the SR membrane, because all of these factors would be expected to alter coupling ratios (22).

A coordination in SR coupling adaptations to aerobic training is also suggested in Ca²⁺ uptake and Ca²⁺ release because the reductions observed in Ca²⁺ uptake were also accompanied by reductions in Ca²⁺ release. Functionally, cooperative changes in these processes could help maintain the [Ca²⁺]_f-time integral that occurs during activation because the reduced rate of Ca²⁺ release is also paralleled, at least qualitatively, by reductions in Ca²⁺ uptake kinetics. Reduced Ca²⁺ cycling, as suggested by the coupled changes in Ca²⁺ release and sequestration, could conceivably preserve [Ca²⁺]_f-time integral and activation of the myofibrillar complex and force generation (pCa-force relationship) but at lower energy cost (38).

It should be emphasized that the interpretation of the effects of training on Ca²⁺ release may be influenced by the type of releasing agent employed. We have followed the procedure used by Ruell et al. (31), in which AgNO₃ is used as the releasing agent. In recent work, we have found that, unlike 4-chloro-m-cresol, which is specific to the Ca²⁺-release channel, AgNO₃ can also act on the Ca²⁺-ATPase and can lead to reversal of the pump action (39). Moreover, because Ca²⁺ release occurs in two phases, training effects may be different between the phases. This study was completed before our examination of the Ca²⁺-release kinetics using different releasing agents (39). As such, we have followed the popular practice of using the second phase of Ca²⁺ release and the releasing agent AgNO₃. Additional study is warranted investigating the possibility that the effect of training may be modified depending on the releasing agent used and the Ca²⁺ release phase examined.

It should also be noted that limitations exist with regard to the assessment of Ca²⁺-release kinetics, particularly with regard to the role of different Ca²⁺-binding proteins. Ca²⁺-binding proteins such as troponin-C and Ca²⁺-ATPase would be expected to bind some of the Ca²⁺ and potentially affect the slope of [Ca²⁺]_f vs. time (41). As a consequence, even though our results have been normalized to total protein content, changes in specific Ca²⁺-binding proteins with training could potentially modify our conclusions regarding the effects of training on SR Ca²⁺-release function per se.

A particularly important issue is the mechanisms(s) underlying the reduction is maximal Ca²⁺-ATPase activity. Reductions in function could occur because of reductions in protein, reductions in protein function, or both. Previous studies have clearly demonstrated that prolonged induced contractile activity can result in initial reductions in maximal Ca²⁺-ATPase activity at least in fast muscle as a result of structural alterations in the region of the adenine nucleotide binding site, whereas protein content is unchanged (5). Although the effect of voluntary exercise on Ca²⁺-ATPase activity remains controversial, particularly in rats (7), studies have reported reductions in muscles composed of a predominance of fast-twitch fibers (4, 42), presumably as a result of inactivation of the nucleotide binding site (18). Such a mechanism also appears to exist during prolonged exercise in humans because several studies have reported disturbances in Ca²⁺-ATPase activity in the vastus lateralis (2, 10, 15). Reductions in Ca²⁺-ATPase activity also appear to be accompanied by reductions in Ca²⁺ uptake (2, 15). In addition, reductions in Ca²⁺ release also appear to accompany the exercise-induced reductions in Ca²⁺-ATPase activity (6, 15, 42). The length of time that the depressions in the Ca²⁺-ATPase activity, Ca² uptake, and Ca²⁺ release persist when contractile activity has stopped remains uncertain. In this study, a period of ~4 days was provided after the last training session before the posttraining tissue sample was obtained.

The adaptations that we have observed in the properties of the SR do not appear to be linked to modifications in the major fiber-type distribution in the vastus lateralis because we have found, using both myosin-based histochemistry and myosin heavy chain analyses by Western blot techniques, that no differences existed in these properties between the trained and untrained legs (unpublished observations). Previous studies using CLFS in animals to investigate SR plasticity have utilized muscles of a predominant type II composition. The vastus lateralis in untrained humans is typically composed of an approximately equal percentage of type I and type II fibers (32), and both fiber-type populations would be expected to be activated during the training sessions (32). On the basis of the dramatic changes in SR protein expression and function that have been observed in animal type II-based muscles to CLFS, it might be expected that a large percentage of the SR adaptations observed in our study occurred in the type II fibers of the vastus lateralis and particularly those with a low oxidative potential. This hypothesis is supported by the findings of Kim et al. (17), who found that reductions in SR Ca²⁺ uptake follow treadmill running in rats were restricted to the superficial portion of the vastus lateralis, a region composed primarily of fast, low-oxidative fibers.

With our experimental design, SR alterations with training were assessed by comparing the trained leg with the untrained or control leg only at the end of the 10-wk program. This one-leg design is the model employed with CLFS. Although changes in SR expression could conceivably occur in the inactive leg as a result of hormonal influences, this does not appear to occur, at least in CLFS, in which the induced-contractile activity represents the dominant signal regulating protein expression (30). The use of time-course measurements, as has been used in CLFS, would have provided important insights not only into the pattern of changes in specific SR proteins and processes but also into the cooperative nature of the expression of multiple proteins. As an example, it has been reported that cardiac-like replacement of proteins occurs early in CLFS (9). The reductions that were observed in Ca²⁺ release in this study could be due not only to a downregulation in total RyR1 protein but to an isoform shift. Increases in RyR2, which predominates in the heart, have been found to increase in fast-twitch-based muscle with CLFS (9). Unfortunately, because of tissue limitations, we could not measure RyR protein content. With respect to Ca²⁺-ATPase, our results indicate a reduction in protein levels with training that is due to a reduction in SERCA1 because no change was found for SERCA2a. These findings are consistent with the CLFS model, in which changes in SERCA1 precede changes in SERCA2a (27). The reduction in SERCA 2 protein occurred in the absence of changes in total muscle protein. This could occur by increased expression of other proteins. The mitochondrial protein, as an example, is known to be increased by endurance training (32). Because no changes were found in the Hill coefficient with training and only small changes in pCa₅₀ were observed, the isoform shift would appear to have minimal consequence on the Ca²⁺ sensitivity of the Ca²⁺-ATPase. This conclusion is consistent with previous work in which it has been reported that SERCA isoforms expressed in COS cells displayed similar Ca²⁺ affinities (20).

Because both SR Ca²⁺ uptake and Ca²⁺ release may be altered by a variety of SR proteins other than SERCA and RyR, it is possible that altered expression of these proteins could have influenced the SR Ca²⁺-handling changes that we observed with training. A number of proteins such as the dihydropyridine receptor in the t tubule and triadin in the SR couple with RyR, and these demonstrate a similar pattern of downregulation with long-term CLFS (9, 14). Similarly, changes in phospholamban in conjunction with SERCA isoform could alter Ca²-ATPase dependency. However, because the effect of phospholamban is restricted only to SERCA2a and because we found no change in SERCA2a with training, changes in phospholamban expression should be of little consequence on Ca²⁺-sequestration properties (9).

Perspectives. The results of this study have several potentially important functional implications. First and foremost, it is apparent that, like many other proteins of the muscle cell, regular voluntary contractile activity can also alter proteins involved in SR Ca²⁺ handling. Our results also suggest that when the training is prolonged and of submaximal intensity, such that aerobic metabolism represents the dominant source of ATP supply, the adaptations favor a reduction in the proteins and processes involved in Ca²⁺ handling. These changes observed in the SR are consistent with reduced Ca²⁺-cycling kinetics similar to what is observed in slow-twitch fibers. Reduced cycling kinetics would be expected to allow isometric or low-velocity activities to be performed with increased efficiency. Because the coupling ratio was not altered by the training, this could occur primarily as a product of the decreased cycling rate per se and not by altering the energy required per unit Ca²⁺ transport. In addition, because it is well documented that aerobic-based training leads to an upregulation in mitochondrial potential, the downregulation in SR Ca²-ATPase could lead to a more favorable balance between the utilization of the ATP by the SR Ca²-ATPase and aerobic delivery of ATP. As a consequence, cellular energy homeostasis may be better protected during prolonged exercise.