|
|
||||||||
Department of Medicine, University of California, San Diego, La Jolla, California
Submitted 16 August 2006 ; accepted in final form 26 November 2006
| ABSTRACT |
|---|
|
|
|---|
mitochondrial calcium; pyruvate dehydrogenase; cardiac energetics; sarco(endo)plasmic reticulum calcium-adenosine 5'-triphosphatase
-adrenergic receptor-2 overexpressionall of which improve contractile characteristics and lead to an increase in force generation (15, 17, 22, 25, 28). Despite extensive examination of how these changes effect cardiac function, little is known about the corresponding affect these changes have on cardiac metabolism in these animals. Studies using PLB-KO mice, which have increased SERCA activity associated with an enhanced contractile phenotype, have shown an increase (60%) in oxygen consumption normalized for work (8). In addition, 31P-NMR spectroscopic analysis indicated an increase in AMP and ADP and a decrease in the phosphocreatine (PCr)-to-ATP ratio, indicative of an energy-starved state involving increased mobilization of high-energy phosphate from PCr. In addition, adenoviral-mediated SERCA expression, involved in gene therapy studies aimed at improving contractile function in failing hearts, improved the PCr-to-ATP ratio in failing hearts but resulted in a decreased PCr-to-ATP ratio in normal hearts (10). Given these observations, it is difficult to know how any modification aimed at increasing calcium handling and improving contractile function will affect energy metabolism. More specifically, the link between energy supply and demand is thought to be mediated by mitochondrial matrix calcium levels, with increased contractile performance leading to a rise in mitochondrial calcium levels (1, 16, 26). How these changes might affect energy substrate metabolism in the form of fuel selection (carbohydrates vs. fatty acids) is unknown since many studies often provide only carbohydrates as an energy source. In the present study we examine glucose and fatty acid metabolism in the isolated working heart from the SERCA-overexpressing mouse to determine what affect increased SERCA expression has on substrate metabolism. We also examine mitochondrial calcium levels in isolated myocytes using an adenovirus expressing a mitochondrially targeted ratiometric pericam to see whether changes in mitochondrial calcium handling occur in conjunction with changes in calcium handling and how this is related overall to energy substrate metabolism. We show that SERCA overexpression is associated with an increase in mitochondrial calcium and that these hearts show an increased reliance on glucose as a fuel source.
| METHODS |
|---|
|
|
|---|
All animal protocols were approved by the University of California, San Diego, Animal Subjects Committee and conform to the Guide for the Care and Use of Laboratory Animals as outlined by the National Institutes of Health. The mice used in this study were 46 mo of age, with wild-type (WT) littermates serving as controls for SERCA-overexpressing transgenic mice. The SERCA isoform 2a-overexpressing mice used in this study have been characterized previously (17). SERCA protein expression in this transgenic model is increased by
20% to 40% (17, 47), resulting in increased contractile performance.
Ex Vivo Heart Perfusions
Oxygen consumption studies.
Oxygen consumption in SERCA and WT hearts was determined by measuring the oxygen tension in the coronary effluent using a fiber-optic oxygen sensor (FOXY-AL300, Ocean Optics, Dunedin, FL) as described by How et al. (18). Isovolumic contractile parameters were measured in Langendorff-perfused hearts using an intraventricular balloon as described previously (47). Under the present conditions hearts were perfused with a modified Krebs-Henseleit bicarbonate buffer (KHB) containing 11 mM glucose and 0.4 mM octanoate as energy substrates. Oxygen consumption and function (pressure development and heart rate) were measured in unpaced hearts, hearts paced at 8 Hz, and unpaced hearts perfused with 10 µM isoproterenol added to the KHB to increase workload. Oxygen consumption was normalized to cardiac mass and plotted against rate pressure product (RPP) as shown in
Fig. 2.
|
|
In vivo viral expression of ratiometric mitochondrial pericam in heart. Mitochondrially targeted ratiometric pericam was packaged into adenoviral vectors for in vivo cardiac myocyte infection of transgenic SERCA and WT littermate mice. The gene therapy was performed as outlined previously (27). Mice were anesthetized with a ketamine (100 mg/kg)-xylazine (8 mg/kg) mixture, intubated, and ventilated with room air. Access to the thoracic cavity was obtained via a lateral sternotomy at the level of the second intercostal. The heart was lifted from the thoracic cavity, and a stitch was placed at the apex of the heart using 8-0 (Ethalon) suture to allow manipulation of the heart. Mitochondrially targeted pericam-expressing adenovirus was administered by direct injection into the LV free wall (5 sites and 10 µl/site) using an insulin syringe with a 29-gauge needle. After virus injection, a 22-gauge plastic cannula was inserted through the chest wall to evacuate residual air following closure of the chest cavity with 5-0 vicryl suture. After the evacuation of any trapped air by gentle suction, the 22-gauge cannula was removed and the mouse was taken off the ventilator and allowed to recover 4 to 5 days following gene therapy before myocyte isolation.
Myocyte isolation cytosolic and mitochondrial calcium measurements. Individual myocytes were isolated from the excised hearts by collagenase digestion according to the method outlined in Belke et al. (3). Cytosolic calcium transients were measured in myocytes affixed to laminin-coated coverslips and loaded with the fluorescent calcium indicator indo-1 (3 µM AM ester) according to the methods outlined in Suarez et al. (44). After the loading was completed, cytosolic calcium transients were measured (100 Hz) by using a Photon Technologies (Toronto, Canada) system configured for single excitation (365 nm) and dual emission (405 and 480 nm) and attached to a Nikon Diaphot epifluorescence microscope. During these measurements, myocytes were constantly perfused with a HEPES-buffered Tyrode solution (pH 7.4) containing 1 mM calcium and electrically paced at a rate of 0.3 Hz.
Mitochondrial calcium measurements were performed on a separate set of myocytes isolated from hearts of animals that had undergone gene therapy with adenovirus expressing a ratiometric pericam as described in In vivo viral expression of ratiometric mitochondrial pericam in heart. Myocytes were visually inspected under a 480-nm wavelength excitation for green fluorescence, indicative of adenoviral pericam infection (see
Fig. 4A). The same system used to measure cytosolic transients was employed to measure mitochondrial transients, except that the system was configured for dual excitation (410 nm for calcium-free and 480 nm for calcium-bound forms) and single emission (545 nm) as described by Nagai et al. (33). The slew rate between the two excitation wavelengths occurred at a frequency of 20 Hz. Resting calcium level was measured in quiescent myocytes followed by electrical pacing of myocytes at 1 Hz (myocytes were observed through the ocular objective to ensure capture during the pacing protocol). After the measurement of mitochondrial calcium in resting and stimulated myocytes, the cells were exposed to 10 µM ionomycin and Tyrode solution, containing zero calcium and 10 mM EGTA, to obtain the fluorescent signal in the calcium-free state [minimum ratio (Rmin)] and 30 µM digitonin in Tyrode solution containing 2 mM calcium to obtain the fluorescent signal in the calcium-bound state [maximum ratio (Rmax)]. This concentration of digitonin has been previously shown in isolated myocytes to remove the sarcolemma with little disturbance to mitochondrial integrity (24). Ratiometric pericam has a dissociation constant (Kd) of 1.7 µM (33), so that mitochondrial calcium levels can be calculated according to the formula [calcium] = (Kd)(R Rmin)/(Rmax R), where R is the ratio obtained from resting and paced myocytes. The ratiometric nature of the pericam construct allows us to measure changes in the calcium-bound and calcium-free forms simultaneously. As a result, the mitochondrial calcium signal is not dependent on the loading of the construct within the mitochondria and, hence, not dependant on the level of viral infection in vivo. Our preliminary studies indicate that pericam expression within the myocyte does not affect the cytosolic calcium transient (data not shown). Ratiometric pericam has a dissociation constant of 1.7 µM and a Hill coefficient of 1.1 (33). Previous studies have indicated a linear response for this construct to calcium over a range of 0.1 to 10 µM (33). Ratiometric pericam was only expressed in those studies involving individual myocytes.
|
|
Statistical Analysis
Values shown represent means ± SE. All comparisons were performed by using an unpaired Student's t-test, with the exception of mitochondrial calcium values where a paired t-test was used to compare stimulated and unstimulated myocytes within the same group. Significance was set at P < 0.05.
| RESULTS |
|---|
|
|
|---|
Contractile performance was measured in hearts perfused in working mode with glucose and palmitate as energy substrates. Cardiac power output from isolated working hearts over the 60-min perfusion period is shown in Fig. 1. Power output was similar in both groups, with peak output being observed between 10 and 20 min of perfusion followed by a slow decline. Functional parameters for the groups averaged over 60 min of perfusion are shown in Table 1. Cardiac output, systolic pressure (measured in the afterload line), and the calculated power output were virtually identical for both SERCA and WT control groups. Interestingly, the hearts from SERCA mice tended to be slightly smaller in mass than from their WT littermates, resulting in a slightly smaller stroke volume (27.2 ± 2.2 vs. 32.1 ± 3.5 µl, SERCA and WT, respectively); however, these differences were not significant. In contrast, heart rate was slightly higher in the SERCA groups (355 ± 15 vs. 323 ± 27 beats/min, SERCA and WT, respectively). As a result, overall cardiac output was similar between the two groups (9.73 ± 0.94 vs. 10.34 ± 1.40 ml/min, SERCA and WT, respectively). Coronary flow rate was identical for both groups. When examining the pressure tracings from the aortic afterload line, we did observe that the time spent in systole was slightly (though not significantly) faster in the SERCA group relative to the WT group (53.1 ± 1.3 vs. 58.4 ± 2.5 ms, P = 0.07), suggesting that the SERCA group moved its volume up the afterload line at a slightly faster rate.
|
Since previous studies have indicated the possibility of increased oxygen consumption relative to work levels in hearts with enhanced contractile performance, we wanted to compare oxygen utilization in SERCA and WT mice relative to different work levels. These hearts were perfused with glucose and octanoate to provide carbohydrates and fatty acids for oxidative metabolism. The plot of oxygen consumption against the RPP in Langendorff-perfused hearts fitted with an intraventricular balloon to measure pressure development is shown in Fig. 2. The results revealed a similar response in oxygen consumption for increasing workload (RPP) in both groups with the slopes of the two lines being virtually identical, i.e., 0.894 vs. 0.865 (µmol oxygen·min1·g1)/[(beats·min1·mmHg)/1,000] for WT and SERCA hearts, respectively). The only difference was the tendency of the SERCA mice to achieve a higher level of work when stimulated with 10 µM isoproterenol. In general, these results show that hearts isolated from SERCA-overexpressing mice do not show excessive oxygen consumption relative to work level.
Metabolic Profile
The rates of anaerobic glycolysis and mitochondrially mediated glucose oxidation and palmitate oxidation measured in isolated working hearts are shown in Table 2 and Fig. 1. In both groups, the rate of glycolysis was virtually identical, suggesting that ATP production from glycolysis was not preferentially enhanced in either group (Table 2). In contrast, oxidative metabolism differed remarkably between the two groups. The amount of glucose entering the oxidative pathway was significantly increased in SERCA mice. In contrast, fatty acid oxidation was diminished in these mice. If we assume an efficient conversion of substrate metabolism into ATP production for oxidative metabolism (see METHODS), then overall ATP production from glucose and palmitate oxidation between the groups was similar (166 ± 15 vs. 158 ± 18 µmol ATP·min1·g dry weight1 for WT and SERCA hearts, respectively). This observation reflects both the similar level of work performance and the similar level of oxygen consumption observed in these hearts. However, the contribution of ATP from glucose oxidation was 52.6% in SERCA mice but only 29.2% in WT mice, reflecting the differences in the source of energy substrates between the two groups.
|
PDH activity was measured in whole heart homogenates by following the production of 14CO2 from [1-14C]pyruvate. The results of the assay are shown in Table 3. The active form of PDH was significantly (P < 0.05) higher in the SERCA group compared with their WT littermates. In contrast, the total PDH activity did not differ between the groups. As a result, the active PDH complex represented 77.7% of total in the SERCA group but was only 52.7% in the WT group.
|
Cytosolic calcium transients measured using the indo-1 calcium indicator is shown in Fig. 3. The indo-1 ratio measured in resting myocytes tended to be lower in the SERCA group (1.53 ± 0.04 vs. 1.59 ± 0.03), whereas the peak ratio tended to be higher in the SERCA group (1.94 ± 0.05 vs. 1.84 ± 0.04) so that the resting and peak indo-1 values bracket the WT group; however, none of these differences were significant. A composite of the individual transients for the SERCA and WT groups is shown in Fig. 3A. The average transient height was significantly higher (64%) in the SERCA group relative to the WT group (Fig. 3B), indicating a larger calcium transient in the SERCA group. If we look at the rate of decline of the transient from peak values (Fig. 3, C and D), the SERCA group showed a significantly faster rate of decline compared with the WT group, a feature indicative of increased SERCA activity. These data show that the cytosolic calcium transients in the SERCA group were larger than those seen in the WT group.
To measure mitochondrial calcium levels, SERCA and WT mice underwent in vivo gene therapy to express a ratiometric pericam mitochondrial calcium reporter (see METHODS). Figure 4A shows a myocyte isolated from a heart expressing mitochondrial pericam. The fluorescent signal is observed in lines running the long axis of the myocyte, reminiscent of mitochondrial packing between sarcomeres. An example of the mitochondrial calcium transients obtained from the ratio pericam is shown in Fig. 4B. Resting calcium was measured in quiescent myocytes before electrical pacing of the myocyte at 1 Hz for 60 s to obtain a new calcium level within the mitochondria. Mitochondrial calcium increased significantly in WT cells upon pacing at 1 Hz (Fig. 4C), rising from 185 ± 22 to 263 ± 33 nM (P = 0.029, paired t-test). Mitochondrial calcium also increased significantly in SERCA myocytes mice upon pacing (Fig. 4D), rising from 163 ± 18 to 472 ± 97 nM (P = 0.003, paired t-test). The level of mitochondrial calcium in the SERCA group did not differ from the WT group under resting conditions; however, it was significantly higher in the SERCA group upon pacing. These results suggest that the increased cytosolic calcium transient in SERCA mice translates into an increased level of mitochondrial calcium upon pacing.
| DISCUSSION |
|---|
|
|
|---|
The increased contractile performance of the SERCA-overexpressing mouse has been reported previously (17). It is interesting to note that oxygen consumption as a function of cardiac work (RPP) in the Langendorff-perfused hearts did not differ greatly between WT and SERCA mice, suggesting that this model of enhancing contractile performance does not lead to oxygen wasting. The only difference between the groups was the tendency of the SERCA group to develop higher levels of work when stimulated with isoproterenol. This effect could be explained by increased loading of the sarcoplasmic reticulum (SR) with calcium due to the increased rate of calcium uptake as a result of SERCA overexpression. The increased rate of calcium uptake observed in isolated myocytes in the present study matches our previous observations (17). In contrast to the higher work levels obtained in the isovolumic Langendorff-perfused hearts, we did not see an increased pressure development in SERCA hearts in the Neely working heart model. This is due to the nature of the working heart model where the heart is capable of performing both pressure and volume work and the fixed preload and afterload pressure values. As a result, the physical amount of work being performed by the two groups was similar, where the cardiac output and systolic pressure were virtually identical. In vivo measurement of pressure using a Millar pressure transducer has indicted that while the overexpression of SERCA increases the rate of contraction and relaxation, pressure development is identical to that in WT mice (17), reflecting the more physiological nature of performing both pressure and volume work as opposed to pressure work alone (Langendorff isovolumic model). In contrast, the PLB-KO mouse shows a higher systolic pressure development in the isolated working heart (25). In vivo measurement of maximal LV power normalized to end-systolic volume was also significantly enhanced (6). In contrast to the observations reported in the current study, previous studies such as that presented by Muller et al. (31) using SERCA-overexpressing models have reported an increase in LV pressure development in the working heart model. The discrepancy between these observations may relate to the means behind which the different working heart models use to generate the afterload pressures against which the hearts are required to work. The classic Neeley working heart model (34) employed in the current study utilizes a hydrostatic column (roughly 32 in. in height) to generate afterload and maintain perfusion pressure on the coronaries. In this model the heart needs only to generate sufficient pressure to overcome the afterload and move fluid up the column without resticting flow. In contrast, the working heart model employed by Muller et al. (31) generates afterload without the need for a hydrostatic column by restricting aortic flow (Starling resistors), permitting the heart to generate higher pressures. By a restriction of aortic flow, this model combines elements of the working heart with that of the isovolumic contracting heart (where aortic outflow is completely restricted). We have previously observed an increase in LV pressure development in SERCA mice in the isovolumic contracting heart (47).
It should be noted that this study provided only glucose (carbohydrate) and palmitate (fatty acid) as substrates for the working heart model. When presented with other carbohydrates, such as lactate and pyruvate, the heart will preferentially metabolize those substrates (7). However, unlike previous studies examining metabolism in hearts with improved contractile function (8, 10), we provide substrates that can oppose each other's metabolism through the reciprocal actions of the Randle cycle (37). Previous studies have examined oxygen consumption and PCr-to-ATP ratios in hearts provided with carbohydrates, such as pyruvate and/or glucose alone, despite the preference of the heart for fatty acids as a substrate (8, 10, 36). The present study does not account for any potential use of endogenous energy sources such as glycogen or fatty acids in the form of triacylglycerol; however, under the conditions used in this study, mobilization of endogenous fatty acid is generally minor, whereas the dynamic nature of glycogen utilization indicates that exogenous radiolabeled glucose mixes readily with this pool (14, 39). It is interesting that calculated energy production from optimal glucose and fatty acid metabolism was similar in both WT and SERCA hearts. While this calculation assumes optimal conversion of substrate metabolism into ATP, we note that the two groups also showed a similar level of power output, and oxygen consumption studies could find no evidence of increased oxygen consumption relative to workload in either of the groups. It is interesting to note that the rate of glycolysis is similar in both WT and SERCA mice. Whereas previous studies have suggested that glycolytically derived ATP may be preferentially utilized by SERCA (49), the present study demonstrates that this metabolic pathway is not upregulated to provide ATP for SERCA and does not contribute to the improved contractile phenotype observed in these hearts.
In contrast to glycolytic activity, SERCA transgenics did show alterations in oxidative metabolism, resulting in an increased reliance on glucose, and a decreased reliance on fatty acids as a fuel source for oxidative metabolism. The most reasonable explanation for these results would be an increased calcium concentration in the mitochondrial matrix as the result of increased calcium handling within the myocyte due to increased SERCA activity. A rise in mitochondrial calcium leads to activation of the PDH complex that would preferentially increase carbohydrate metabolism (46). Recent studies have demonstrated a close physical approximation of the SR and the mitochondria that could account for increased mitochondrial calcium levels as a result of increased SR calcium handling (9; see Ref 38 for review). Previous studies in PLB-KO mice did show an increase in PDH activity that would match the enhanced SERCA activity and the increased metabolic demand by the PLB-KO hearts (8). That study also used a scanning transmission electron microscope and electron probe microanalysis to look at calcium levels in various compartments within the myocyte. Although they did see an increase in calcium within the SR compartment, they were unable to find any increase in calcium levels within the mitochondria. However, the method used in that study cannot distinguish free calcium from bound calcium (29), and the capacity of the mitochondria to store large amounts of calcium bound to phosphate could have prevented them from observing a rise in mitochondrial matrix calcium. In the present study we employ a mitochondrially targeted ratiometric pericam to examine mitochondrial calcium levels (33). The construct was delivered to myocytes of WT and SERCA mice in vivo via adenoviral gene therapy. In vivo transfection (as opposed to infection of isolated myocytes in cell culture) was employed to allow expression in a working heart in its physiological milieu to avoid myocyte remodeling and the potential decrease in SR activity that accompanies quiescent myocytes kept in culture (32, 48). Using the ratiometric pericam construct, we were able to demonstrate that basal levels of calcium were similar in both groups and that electrical stimulation of cardiac myocytes results in a greater increase of mitochondrial calcium in SERCA than in WT myocytes. This increase in mitochondrial calcium could lead to activation of the PDH complex, which would account for the increased glucose utilization observed in the SERCA group. We observed that the activity of PDH was higher in the SERCA group than in the WT group when measured in whole heart homogenates.
A rise in mitochondrial calcium levels would be expected to increase not only PDH activity but also the activity of Krebs cycle enzymes isocitrate dehydrogenase and
-keto glutarate dehydrogenase that speed up the rate at which acetyl-CoA is metabolized regardless of its original source (i.e., carbohydrates or fatty acids), resulting in an increased NADH production for the electron transport chain (16, 26). Several studies have shown an increased NADH production accompanied an increase in mitochondrial calcium levels during electrical pacing of isolated myocytes and cardiac trabeculae (4, 21). This effect is thought to provide a mechanism for the coupling of the expected increase in energy demand (from increased myocyte work) to increased ATP production to prevent a large increase in ADP levels to drive ATP production (26). With the transposition of our observations made in myocytes to the whole heart, the increase in mitochondrial calcium during an increase in workload would be expected to anticipate the increased need for ATP from oxidative metabolism and reduce the limit of energy debt resulting from increased ADP production and local mobilization of PCr. Interestingly, 31P-NMR spectroscopy on SERCA and WT mice has indicated higher levels of PCr, a lower concentration of ADP, and a higher free energy of ATP hydrolysis (
GATP) for a given level of wall stress in hearts from SERCA mice than those from corresponding WT littermates (Belke DD, Pinz I, Swanson E, Scott B, Tian R, Ingwall J, and Dillmann W; unpublished observations). Such observations might be indicative of increasing mitochondrial calcium load through enhanced calcium handling by SERCA, increasing energy production within the mitochondria. These observations may also explain why fatty acid oxidation is not also subsequently increased by a rise in mitochondrial calcium in our study. The activation of the Krebs cycle enzymes isocitrate dehydrogenase and alpha-keto glutarate dehydrogenase by calcium also requires an increase in ADP levels to increase enzyme activity (11). In contrast, the activation of the PDH complex by calcium is independent of ADP levels (11). Given the similar workload of the two groups in the present study (due to fixed preload and afterload perfusion pressures) and the similar efficiency in oxygen consumption per unit work measured in the two groups, it is unlikely that such a rise in ADP would occur in the SERCA group to sufficiently activate the Krebs cycle enzymes to increase fatty acid oxidation.
Whether the beneficial effect of increased SERCA activity on myocardial energetics can be maintained in pathological cardiomyopathies remains to be examined. We have previously shown that SERCA overexpression protects the heart against the decrease in contractile function which accompanies diabetes (47). Recent work by Sakata et al. (40), using adenovirally mediated SERCA expression in rat hearts, has shown an improvement in energy utilization and oxygen consumption in a rat model of Type 2 diabetes. It should also be noted that adenovirally mediated SERCA overexpression did improve the ATP-to-PCr ratio in hypertrophied rat hearts (10), and we have shown an improved function in a pressure overload model of hypertrophy involving aortic-banded transgenic mice with doxycycline-inducible SERCA expression (43). Recent studies have suggested that an increase in fatty acid oxidation leads to a decrease in cardiac efficiency (19). Conversely, an increase in glucose oxidation could lead to an improvement in cardiac efficiency by decreasing the amount of oxygen required to generate the same amount of ATP. This may explain, in part, the improvement in cardiac efficiency observed in hearts with SERCA overexpression. We note, however, that the rate of oxygen consumption relative to RPP as measured in our Langendorff-perfused hearts (using glucose and octanoate as energy substrates) was similar in both the WT and SERCA groups: 0.894 vs. 0.865 (µmol oxygen·min1·g1)/[(beats·min1·mmHg)/1,000] for WT and SERCA hearts, respectively.
In summary, increased expression of SERCA leading to an increase in calcium handling by the SR translates into an increase in mitochondrial calcium levels in the working heart. The increased mitochondrial calcium levels influence energy metabolism by increasing glucose oxidation and possibly making overall energy production more efficient. This appears to be a benefit that accompanies the improved contractile phenotype of the heart and may play a role in the beneficial effects which accompany increased SERCA expression under pathological conditions such as diabetes.
| GRANTS |
|---|
|
|
|---|
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
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.
| REFERENCES |
|---|
|
|
|---|
-agonist stimulation. Circ Res 75: 401409, 1994.This article has been cited by other articles:
![]() |
K. E. Broderick, L. Alvarez, M. Balasubramanian, D. D. Belke, A. Makino, A. Chan, V. L. Woods Jr., W. H. Dillmann, V. S. Sharma, R. B. Pilz, et al. Nitrosyl-Cobinamide, a New and Direct Nitric Oxide Releasing Drug Effective In Vivo Experimental Biology and Medicine, December 1, 2007; 232(11): 1432 - 1440. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |