INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

PHOSPHOLAMBAN (PLB) is a key regulator of the sarcoplasmic reticulum (SR) Ca-ATPase in ventricular myocytes. Dephosphorylated PLB is closely associated with the SR Ca-ATPase and acts as an inhibitor of the Ca pump (21, 22, 35, 36). Phosphorylation by cAMP-dependent protein kinase (PKA) and Ca/calmodulin-dependent protein kinase II (CaMKII) can relieve this inhibition (25, 41), allowing greater Ca transport at a given intracellular Ca concentration ([Ca]_i). The adjacent residues Ser-16 and Thr-17 of PLB were identified as the unique sites phosphorylated by PKA and CaMKII, respectively (38). However, most of the information about PLB has come from in vitro studies, since multiple proteins are involved in Ca handling in vivo, and it is difficult to isolate the effects of PLB. The recently generated PLB knockout (PLB-KO) mouse (27) allows unique insights into the functional consequences of PLB ablation in either the intact animal or in relatively intact preparations such as isolated myocytes.

In permeabilized ventricular myocyte, CaMKII inhibitors (KN-62 and specific peptides) decreased SR Ca pump function, and this effect could be prevented by an antibody that interferes with the PLB-Ca pump interaction (30). In intact rat myocytes, [Ca]_i decline is slower postrest (PR) than during steady-state (SS) twitches (6). The CaMKII inhibitor KN-62 prevents the frequency-dependent acceleration of [Ca]_i decline in SS (vs. PR), whereas phosphatase inhibitors prevent the slowing of [Ca]_i decline at the PR twitch (vs. SS; see Ref. 6). This suggests that the faster [Ca]_i decline during SS twitch is due to activation of CaMKII, and PLB was a possible target for such CaMKII regulation. This frequency-dependent acceleration of relaxation may be physiologically important during alteration in cardiac frequency, and the PLB-KO mouse provides a unique opportunity to directly test whether PLB is required for this effect.

Normal activation of cardiac cells involves Ca-induced Ca release, and the increased [Ca]_i and contraction are only transient because Ca is removed from the cytosol by Ca transport systems (7). A great deal is known about Ca transport systems in ventricular myocytes from several mammalian species, such as rat (1, 31), rabbit (1, 9) and ferret (2). In all of these species, the SR Ca-ATPase plays the dominant role in rapid removal of Ca from the cytosol during relaxation, although the percentage of contribution varies widely among species. The sarcolemmal Na/Ca exchange is the next most important Ca transport system during relaxation, with the sarcolemmal Ca-ATPase and mitochondrial Ca uniport generally playing a small role in removing Ca from the cytosol (1, 31).

This type of detailed quantitative information about the balance of Ca fluxes is not available for mouse ventricular myocytes, and we address this in the present study for both wild-type (WT) and PLB-KO mice. Given the emergence of transgenic and knockout mouse models in cardiovascular research (10), this fundamental quantitative information may be of broad importance and utility. Many different genetically altered mice have been generated, and the partially characterized PLB-KO mouse has been a valuable one (12, 15, 16, 27-29, 43).

In the present study, we address two major issues. First, we directly test the hypothesis that PLB is required for the frequency- and CaMKII-dependent regulation of twitch relaxation and [Ca]_i decline. Second, we evaluate for the first time the quantitative balance of Ca fluxes in WT mouse myocytes during relaxation and excitation-contraction (E-C) coupling and extend this to compare how they are altered in the PLB-KO model. A major finding is that the acceleration of SS twitch relaxation and [Ca]_i decline is still observed in the PLB-KO mouse and is still abolished by the CaMKII inhibitor KN-93. We conclude that the frequency-dependent acceleration of relaxation and [Ca]_i decline cannot be attributed solely to PLB phospholyration. We also find that Ca transport in the WT mouse is quantitatively similar to that reported in rats (1, 31), whereas in the PLB-KO mouse the SR Ca-ATPase is even more dominant over the Na/Ca exchange. Furthermore, in the PLB-KO, we find increased SR Ca content, reduced fractional SR Ca release during E-C coupling (at a given SR Ca load), and reduced ability of the Na/Ca exchange to extrude Ca (even without competition from SR transport).

	METHODS

Top Abstract Introduction Methods Results Discussion References

Cardiac myocyte preparation. Isolation of ventricular myocytes from WT and PLB-KO mice was carried out as previously described (18). Briefly, hearts were excised from adult male mice (species-matched WT and PLB-KO, 35-45 g) anesthetized with pentobarbital sodium (70 mg/kg ip). Hearts were mounted in a Langendorff perfusion apparatus and perfused with nominally Ca-free Tyrode solution for 6 min at 37°C. Perfusion was then switched to the same solution containing 0.8 mg/ml collagenase (type B; Boehringer-Mannheim, Indianapolis, IN) and 0.03 mg/ml pronase (Boehringer-Mannheim), with perfusion continuing until the heart became flaccid (~7-12 min). The ventricular tissue was then dispersed and filtered. The cell suspension was rinsed several times, with a gradual increase in the Ca concentration ([Ca]) to 1 mM. Before experimental use, the myocytes were plated onto Plexiglas superfusion chambers, with the glass bottoms of the chambers treated with laminin (GIBCO, Grand Island, NY) to increase cell adhesion. Although our yield of viable cells was lower with mouse (~40%) than with other species, no systematic difference was seen between PLB-KO and WT.

Measurement of cell shortening. Myocyte shortening was measured as previously described (5). The cells were superfused with normal Tyrode (NT) solution at room temperature (22-23°C) and field stimulated (square waves, 0.5 and 1 Hz). Cells were transilluminated by a red light source (to avoid interference with indo 1 epifluorescence measurement), and shortening was measured using a video-edge detection system (Crescent Electronics, Sandy, UT).

Measurement of intracellular Ca. To obtain [Ca]_i measurements while allowing myocytes to control their own intracellular environment (including Na concentration) and action potential, cells were loaded with indo 1 by incubation with the acetoxymethyl ester form of the dye (indo 1-AM, 10 µM; Molecular Probes, Eugene, OR) for 20 min at room temperature. After loading, the cells were superfused with NT solution for at least 30 min to wash out excess indicator and allow deesterification (5). The excitation source was a 150-W xenon lamp (Oriel, Stratford, CT) with a 355 ± 5 nm interference filter (Chroma Technology, Brattleboro, VT). Within the microscope, a 380-nm dichroic mirror reflected the ultraviolet light toward a fluorescence objective (Nikon CF Fluor ×40). The field illuminated was restricted to a circular spot of 100 µm in diameter. The fluorescence emitted by the cells was transmitted by a dichroic mirror (600 nm) onto another dichroic mirror (440 nm) such that 405- and 485-nm light could be selected by interference filters (40-nm bandwidth; Chroma Technology) placed in front of two photomultiplier tubes (Hamamatsu, Bridgewater, NJ). The microscope emission field was restricted to a single cell with the aid of an aperture in the emission path. The average background fluorescence recorded at both wavelengths from cells not loaded with indo 1 was subtracted before the fluorescence ratio was calculated. Cell fluorescence signals were digitized at 120-133 Hz (filtered at 60 Hz) and stored on a computer.

In mouse myocytes, we found that these indo 1 loading conditions resulted in substantial intracellular compartmentalization of indo 1, which contrasts with other mammalian ventricular myocytes (1, 2). Digitonin permeabilization indicated that ~40% of indo 1 was compartmentalized (and shorter loading time did not reduce this fraction). This greatly complicates calibration, and use of the traditional Grynkiewicz et al. (14) equation was not practical. As a compromise, fluorescence ratios (R = fluoresence at 405 nm/fluoresence at 485 nm) were converted to [Ca]_i by a modified "pseudo-ratio" method (11), that is, [Ca]_i = K_d(R'/R_rest)/[(K_d/[Ca]_{i rest}) + 1  R'/R_rest] where K_d is the dissociation constant for indo 1 (450 nM),  is the ratio of the free to bound indo 1 fluorescence at 485 nm (3.3), R' is the fluorescence ratio minus the minimum R (R_min), R_rest is the resting fluorescence ratio minus R_min, and [Ca]_{i rest} is [Ca]_i at rest. Because we did not observe significant differences of resting fluorescence ratio in WT and PLB-KO myocytes, it was assumed that [Ca]_{i rest} was 150 nM (3). R_min was determined in vivo in indo 1-AM-loaded cells superfused with solutions containing 5 mM EGTA/nominally zero Ca in the presence of the nonfluorescent Ca ionophore bromo-A-23187 (10 µM; Calbiochem, La Jolla, CA). Although the resulting [Ca]_i values seem reasonable in the context of our previous experience, the absolute values should be taken as practical approximations. Most data were calculated with both traditional and pseudo-ratio methods, and there were no differences in the qualitative conclusions. Thus we feel that our practical assumptions are justified.

Assessment of SR Ca load, Na/Ca exchange, and slow mechanisms. Rapid application of 10 mM caffeine was used to induce release of SR Ca and assess the SR Ca load as well as the participation of Na/Ca exchange and slow transport systems (mitochondrial Ca uniporter and sarcolemmal Ca-ATPase) during [Ca]_i decline. Cells were superfused with NT solution and stimulated at 0.5 Hz until twitch characteristics stabilized before each caffeine application. The time between the last stimulation and the start of caffeine was normally 2 s. The amplitude of the caffeine-induced Ca transient (CafC) can be used as an index of SR Ca content (5, 24, 39). Because the amplitude of the CafC in NT can be affected by Ca extrusion via Na/Ca exchange (5), CafCs were also evoked in Na- and Ca-free solution (0 Na, 0 Ca). In this case, [Ca]_i decline was attributable to the mitochondrial Ca uniport and sarcolemmal Ca-ATPase (referred to as slow mechanisms). Decline of [Ca]_i during a CafC in NT (140 mM Na) was attributable to Na/Ca exchange and slow mechanisms. Caffeine solution was introduced into the chamber via a quick-switching device (5) and was continued for 20 s to study the kinetics of cell relaxation and [Ca]_i decline.

SS vs. PR twitch contractions. Cells were superfused with NT solution and stimulated at 0.5 or 1 Hz until twitch stabilization (SS twitch). Electrical stimulation was then stopped for 15, 30, or 60 s before stimulation was resumed to assess the first PR twitch. In experiments using the CaMKII inhibitor KN-93, cells were exposed to 1 µM KN-93 for 5 min after the control series was completed. The stimulation was then resumed to reach a new SS.

Quantitative immunoblotting of the Na/Ca exchange. The protein levels of the Na/Ca exchange in the PLB-KO and WT mouse hearts were determined using quantitative immunoblotting in conjunction with the BioMax chemiluminescent detection system (Scientific Imaging Systems; Eastman Kodak, Rochester, NY). Mouse hearts were excised, rinsed with ice-cold phosphate-buffered saline, and frozen rapidly in liquid N₂. The hearts were then homogenized in buffer (pH 7.0) containing (in mM) 10 imidazole, 300 sucrose, 1 dithiothreitol, 10 sodium metabisulfite, and 0.3 phenylmethylsulfonyl fluoride. Cardiac homogenates from six PLB-KO mice or six WT mice were pooled together and subsequently used for Western blot analysis of the Na/Ca exchanger. Aliquots of the pooled homogenates of WT and PLB-KO mice were applied to parallel lanes at different concentrations on each of eight separate SDS gels and processed in parallel. The cardiac homogenates (5-20 mg) were separated on a 13% SDS-polyacrylamide gel and blotted onto a 0.22-mm nitrocellulose membrane (Schleicher & Schuell, Keene, NH) at 200 mA and 4°C for 3 h. After transfer, the membrane was incubated in blocking solution containing 5% nonfat dried milk at 4°C overnight, washed with Tris-buffered saline (TBS: 10 mM Tris · HCl, pH 7.8, and 150 mM NaCl), and then incubated with a mouse monoclonal anti-Na/Ca exchange antibody (Affinity Bioreagents, Golden, CO) diluted 1:1,000 in 0.5% milk/TBS for 30 min with gentle agitation at room temperature. After the membrane was thoroughly washed with TBS and reincubated with a goat anti-mouse antibody conjugate (1:25,000 dilution; Kodak Scientific Imaging Systems) in 1% milk/TBS for 30 min, the membrane was immersed in the Kodak BioMax 1×CDS buffer to ensure optimal pH for the chemiluminescent reaction. The wet membrane was then placed right side up in a reaction folder, and a piece of chemiluminescent matrix was overlaid on this membrane with the emulsion side down. The reaction folder was exposed to a Kodak X-OMAT AR film for 5-10 min when the light output remained stable. The signals were analyzed by laser densitometry using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). To verify the amount of protein loaded for SDS-PAGE and the efficiency of protein transfer onto the nitrocellulose membranes, three of the membranes were also probed with a mouse anti-actin monoclonal antibody (1:500 dilution; Accurate Chemical & Scientific), and these signals were also quantified using the chemiluminescent method.

Reagents and solutions. Unless otherwise stated, experimental reagents used were of analytical grade and were supplied by Sigma (St. Louis, MO). 2-[N-(2-hydroxyethyl)-N-(4-methoxybenzenesulfonyl)]-amino-N-(4-chlorocinnamyl)-N-methybenzylamine (KN-93) was from Seikagaku America (Rockville, MD). A 10 mM stock solution was made up in water and was stored at 4°C. Aliquots of the solution were added to the perfusate immediately before use. The NT solution contained (in mM) 140 NaCl, 6 KCl, 1 MgCl₂, 1 CaCl₂, 10 glucose, and 5 HEPES, with the pH adjusted to 7.4 with NaOH at 22°C. In 0 Na, 0 Ca solution, NaCl in NT was replaced by LiCl, CaCl₂ was omitted, and 1 mM EGTA was added.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Before addressing the issue of frequency-dependent effects on relaxation and [Ca]_i decline, we will first describe quantitative results concerning the fundamental properties of cellular Ca transport in ventricular myocytes isolated from WT and PLB-KO mice.

Ca transients and SR Ca load. Figure 1 shows examples of Ca transients during twitches and CafCs under SS conditions and compares PLB-KO with WT myocytes (with pooled data in Fig. 2). Figure 1A shows superimposed Ca transients during SS twitches (0.5 Hz) from representative WT and PLB-KO myocytes. Each trace is the average of three SS twitch Ca transients. Time to peak [Ca]_i is shorter in the PLB-KO mouse (142 ± 4 vs. 162 ± 6 ms, P < 0.05). The time constant () for [Ca]_i decline during the SS twitch is also significantly faster in the PLB-KO mouse ( = 188 ± 14 ms in WT vs. 112 ± 6 ms in PLB-KO; n = 17 and 14, respectively; P < 0.001, Fig. 2B). The faster [Ca]_i decline in PLB-KO is not surprising, since the absence of PLB is expected to relieve inhibition of the SR Ca-ATPase and increase the Ca transport rate at a given [Ca]_i. This acceleration of SR Ca-ATPase may also curtail the peak [Ca]_i in PLB-KO and thereby contribute to the shorter time to peak (also see DISCUSSION). The amplitude of the SS twitch Ca transient was slightly higher in the PLB-KO, although not significantly {change in [Ca]_i ([Ca]_i; in nM): 167 ± 21 for WT, n = 16, and 177 ± 19 for PLB-KO, n = 19, Fig. 2A}. Given the increase in apparent SR Ca content (Figs. 1B, 1C, and 2C), a larger increase in SS twitch Ca transient and contraction might have been expected (15, 43).

View larger version (15K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   View larger version (23K): [in this window] [in a new window]   Fig. 1.   Intracellular Ca concentration ([Ca]i) transients in wild-type (WT) and phospholamban (PLB) knockout (KO) mouse myocytes stimulated at 0.5 Hz. A: during steady-state (SS) twitch in normal Tyrode (NT); B: during caffeine (Caff) application in NT solution; and C: during application of caffeine in 0 Na, 0 Ca. Data shown were from different cells, and SS twitch was the average of three traces in both WT and PLB-KO mice. Time constant () was obtained by monoexponential curve fit to [Ca]i decline.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Amplitude and time course of twitch and caffeine (Caff)-induced Ca transients and contractions in myocytes from WT (filled bars) and PLB-KO (KO, open bars) mice. A: pooled data of 0.5-Hz SS twitch contraction amplitude (left) as percentage of resting cell length (RCL, n = 45 and 42) and change in [Ca]i ([Ca]i; right, n = 16 and 19 for WT and PLB-KO, respectively). Unpaired t-test showed no significant difference (ns) for either contraction or [Ca]i. B:  of cell relaxation (left) and [Ca]i decline (right) during SS twitch at 0.5 Hz were significantly longer in WT ( = 116 ± 9 ms for twitch, n = 45; and 188 ± 13.6 ms for [Ca]i decline, n = 17) than in PLB-KO (31.7 ± 1.3 ms for twitch, n = 42; and 112 ± 6.3 ms for [Ca]i decline, n = 14, P < 0.0001 for both). Contraction data in A and B were collected from cells without indo 1-acetoxymethyl ester (AM) loading. C: mean value of sarcoplasmic reticulum (SR) Ca load evaluated by caffeine application in NT (left) and 0 Na, 0 Ca solution (right). SR Ca load was significantly higher in both cases in PLB-KO mouse (n = 10-17, P < 0.01). D: mean  values of [Ca]i decline were slower in PLB-KO during both caffeine application in NT (left, n = 14, P < 0.01) and in 0 Na, 0 Ca (right, n = 12 and 8).

Figure 1, B and C, is representative of traces during CafC in NT and CafC in 0 Na, 0 Ca solution, respectively. The peak [Ca]_i during CafCs are higher in the PLB-KO mouse both in NT ([Ca]_i = 456 ± 58 nM in WT, n = 16, vs. 918 ± 89 nM in PLB-KO, n = 17, P < 0.01) and in 0 Na, 0 Ca solution ([Ca]_i = 565 ± 74 nM for WT, n = 11, vs. 1118 ± 133 nM for KO, n = 10, P < 0.01, Fig. 2C). These results indicate a higher SS SR Ca content in the PLB-KO compared with the WT myocytes. Translating these values to total SR Ca content using the passive intracellular Ca buffering measured in rabbit ventricular myocytes by Hove-Madsen and Bers (17) gives SR Ca contents of 102 and 140 µmol/l cytosol for WT and PLB-KO myocytes, respectively.

During CafCs, [Ca]_i declines more slowly in myocytes from the PLB-KO mice. The  of [Ca]_i decline for the CafC in NT (Caff NT) is 2.19 ± 0.24 s in WT vs. 3.22 ± 0.21 s in PLB-KO (P < 0.01, Fig. 2D). During CafC in 0 Na, 0 Ca (Caff 0 Na, 0 Ca), the [Ca]_i transient in the PLB-KO is almost flat, as shown in Fig. 1C. About one-half of the PLB-KO cells showed such a CafC in 0 Na, 0 Ca, making estimation of  impractical (i.e.,  approaches infinity). Thus, for CafC in 0 Na, 0 Ca, we use the rate constant () of [Ca]_i decline (1/) for initial pooled comparisons (so when [Ca]_i decline is extremely slow,  approached 0). The mean  for CafC in 0 Na, 0 Ca is lower in PLB-KO than in WT mice but is not statistically different (0.0106 ± 0.0088 vs. 0.0284 ± 0.0088 s¹, respectively). The mean  values are converted back to  values for comparison with other  values in Fig. 2D ( was 35.2 s in WT vs. 94.5 s in PLB-KO).

The  of [Ca]_i decline during CafC in 0 Na, 0 Ca is much slower than the value of 12 s reported for rat and rabbit ventricular myocytes by Bassani et al. (1), even for the WT mouse. Because we found significant indo 1 compartmentalization in mouse ventricular myocytes, the extremely slow decay of [Ca]_i could be an artifact due to this compartmentalization. Because ~50% of the Ca released during a CafC in 0 Na, 0 Ca is transported into mitochondria [based on results in rabbit by Bassani et al. (5)], the apparently slow [Ca]_i decline could be confounded by a gradually increasing mitochondrial [Ca]. This would cause  to be overestimated for CafC in 0 Na, 0 Ca. The impact of this potential problem will be significantly less for the cases of the twitch and the CafC in NT, since <7% of Ca is taken up by mitochondria in these situations (1). When cell contraction is simultaneously recorded in WT mouse myocytes during CafC in 0 Na, 0 Ca, the  of cell relaxation (8.19 ± 1.15 s, n = 5) is similar to that reported in both rabbit and rat myocytes (where relaxation and [Ca]_i decline are roughly parallel; see Ref. 1) and much faster than the  of [Ca]_i decline in mouse myocytes. This result further suggests that indo 1 mitochondrial compartmentalization complicates our measurement of  in CafC 0 Na, 0 Ca.

We also recorded cell contractions from cells without indo 1 loading, thus preventing any disturbance of intracellular Ca buffering (Fig. 2, A and B, left). Although SS twitch contractions are, on average, 21% larger in PLB-KO mice, the difference is not significant (as % of resting cell length: 5.96 ± 0.72 for WT, n = 45, and 7.22 ± 0.62 for PLB-KO, n = 42). However, Fig. 2B shows that relaxation  is much faster in the PLB-KO mouse (WT, 116 ± 9 ms, n = 45, KO, 31.7 ± 1.3 ms, n = 42, P < 0.0001). Both of these results are consistent with data for [Ca]_i decline. Cells loaded with indo 1 also showed  values of twitch relaxation comparable to those without indo 1 loading (106 ± 10 ms in WT and 34.6 ± 2.7 ms for PLB-KO, n = 7-10, P < 0.001). This indicates that the degree of indo 1 loading used did not appreciably alter contraction.

We also measured cell relaxation during CafC in NT using cells without indo 1-AM loading. These were done explicitly to evaluate whether the prolonged  of [Ca]_i decline during CafC in PLB-KO in Fig. 2D, left, might be complicated by mitochondrial compartmentalization of indo 1. As for the Ca transients, the  of relaxation of CafC in NT is significantly slower in the PLB-KO (2.12 ± 0.15 s in PLB-KO vs. 1.21 ± 0.28 s for WT; P = 0.019). The 75% slowing of CafC relaxation in the PLB-KO (without indo 1) is consistent with the 50% slowing of [Ca]_i decline during CafC in NT (Fig. 2D, left) and with slowing of Na/Ca exchange in PLB-KO.

Overall, the results in Figs. 1A and 2, A and B, are consistent with an increase in SR Ca pump function in the PLB-KO mouse. This may be the factor responsible for the shorter time to peak and faster relaxation of contraction and [Ca]_i, as well as the higher SR Ca load. In addition, Ca extrusion via Na/Ca exchange may also be slower in the PLB-KO mouse (based on slow decline of CafC in NT). Although there might also be slower Ca transport by the slow Ca removal mechanisms (sarcolemmal Ca-ATPase and mitochondrial Ca uniporter), this is complicated by dye compartmentalization.

SR Ca load is an important factor regulating SR Ca release during E-C coupling. Figure 3 shows the amplitude of the SS twitch [Ca]_i as a function of the SR Ca content (CafC in NT) for individual WT and PLB-KO myocytes. Linear regressions shown in Fig. 3 are significantly different between WT and PLB-KO myocytes (P < 0.05). It can be seen that, for a given SR Ca content, there is a relatively larger twitch [Ca]_i in the WT mouse. This could be due to either decreased fractional SR Ca release for a given SR Ca load or faster SR Ca uptake in the PLB-KO mouse (see DISCUSSION).

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Relationship of SR Ca load and SS twitch [Ca]i in WT and PLB-KO myocytes. For each cell, the amplitude of the SS twitch [Ca]i is plotted as a function of the SR Ca load (based on the [Ca]i induced by caff NT). For a given SR Ca load, the SS twitch [Ca]i in PLB-KO mouse myocytes is smaller (, n = 17) than that of WT mouse myocytes (, n = 17). Slopes of linear regressions (solid line, WT; broken line, KO) were significantly different (P < 0.05).

Analysis of Ca fluxes during relaxation. The relative contributions of the different Ca removal systems (SR Ca-ATPase, Na/Ca exchange, and the slow mechanisms) to [Ca]_i decline during the SS twitch in WT and PLB-KO mouse myocytes are calculated in a manner similar to that described by Bassani et al. (1) for rat and rabbit myocytes (Figs. 4 and 5). Mean values for the of [Ca]_i decline and Ca transient amplitudes for twitches and CafCs from Fig. 2 are used, creating noise-free curves that matched the mean experimental data. Free [Ca]_i is first converted to total [Ca]_i ([Ca]_t) using cellular passive cytosolic Ca-buffering measurements by Hove-Madsen and Bers (17). We assume that there are no major differences in the intracellular Ca buffering or indo 1 Ca affinity between PLB-KO and WT myocytes. Ca buffering differences are unknown, but if different they could complicate our comparisons. The rate of change in [Ca]_t (d[Ca]_t/dt) is then attributed to the sum of fluxes by the individual Ca transport systems (J_SR + J_Na/CaX + J_slow where J_SR is flux of the SR, J_Na/CaX is flux of Na/Ca exchange, and J_slow is flux through the slow mechanisms). During the CafC in 0 Na, 0 Ca, it is assumed that only the slow mechanisms were functional and could be lumped into a Hill relationship describing Ca flux as follows: J_slow = V_min + (V_max  V_min)/[1 + (K_m/[Ca]_i)ⁿ], where V_max is the maximum transport velocity, K_m is a Michaelis constant, and n is the Hill coefficient. V_min was introduced to allow the J value to approach zero at nonzero values of [Ca]_i.

View larger version (16K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   View larger version (19K): [in this window] [in a new window]   Fig. 4.   Rate of Ca removal by SR Ca-ATPase, Na/Ca exchange, and slow mechanisms in WT and PLB-KO mouse myocytes. We assumed that total myoplasmic Ca concentration ([Ca]t) is [Ca]i + 215/[1 + (420/[Ca]i)] + 702/[1 + (79,000/[Ca]i)] + [indo]i/[1 + (450/[Ca]i)]], where [indo]i is intracellular indo 1 concentration. The last term reflects Ca binding to intracellular indo 1. Two classes of binding sites with maximal binding (215 and 702 µmol/l cytosol) and dissociation constant (420 and 79,000 nM) were used to describe endogenous cytosolic Ca buffering (8, 17). [Ca]t is then differentiated with respect to time to obtain a rate of Ca transport: d[Ca]t/dt = JSR + JNa/CaX + Jslow, where J terms reflect flux through SR, Na/Ca exchange (Na/CaX), and slow mechanisms (sarcolemmal Ca-ATPase and mitochondria), respectively. Each J is described by Jx = Vmin + (Vmax  Vmin)/[1 + (Km/[Ca]i)n] (see text for definitions). In all three panels, the thicker traces are obtained from mean Ca transient data, the thinner curves are the fits of Jslow during CafC in 0 Na, 0 Ca (A), JNa/CaX + Jslow during CafC in NT (B), and JSR + JNa/CaX + Jslow during the SS twitch (C). Data and fits for WT (solid curves) and PLB-KO (broken curves) are shown. Derived values for Vmax in JSR, JNa/CaX, and Jslow were 338, 30.5, and 1.7 µM/s for WT and 599, 18.4, and 0.6 µM/s for PLB-KO. The [Ca]i to reach one-half of Vmax for JSR, JNa/CaX, and Jslow, respectively, were 309, 280, and 277 nM for WT and 303, 266, and 264 nM for PLB-KO.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Integrated Ca flux during a SS twitch in WT (A) and PLB-KO (B) mouse myocytes. With the use of the free [Ca]i during the SS twitch and the JSR + JNa/CaX + Jslow equations derived in Fig. 4, the integrated Ca flux through each system functioning simultaneously was calculated during [Ca]i decline. Thus, during the SS twitch in WT mouse, the proportions of Ca transported by the SR, Na/CaX, and the slow mechanisms are 90.3, 9.2, and 0.5%; in PLB-KO mouse, the proportions are 96.4, 3.4, and 0.1%, respectively.

Figure 4A shows that the transformed flux data for CafC in 0 Na, 0 Ca is well described by the J_slow equations over the relevant [Ca]_i range. It should be noted that this provides an adequate empirical description of the [Ca]_i dependence of Ca transport by the slow systems. It does not require that the equations have strict mechanistic interpretations, and the same is true for Na/Ca exchange and SR Ca-ATPase shown in Fig. 4, B and C. However, the J_slow transport appears to show a maximum of only 0.5-1.6 µM/s and is one-half of this value at 200-300 nM [Ca]_i.

Figure 4B is obtained by transforming the Ca flux data from the CafC in NT and is fit by the sum of J_slow at each [Ca]_i (taken exactly as obtained in Fig. 4A) plus a curve fit for J_Na/CaX. The Ca flux data then should be attributable to J_Na/CaX + J_slow. Figure 4B shows that the d[Ca]_t/dt data as a function of [Ca]_i are again well described by J_Na/CaX + J_slow and that Na/Ca exchange fluxes are about 10 times larger than those by the combined slow mechanisms (repeated from Fig. 4A). Next it is assumed that all Ca removal systems are functional during the twitch. Thus the twitch Ca flux in Fig. 4C was fit by J_SR + J_Na/CaX + J_slow, where J_Na/CaX and J_slow parameters are held exactly as obtained in Fig. 4, A and B, allowing a fit for J_SR parameters. The upper curve fits in Fig. 4C are the overall fits for J_SR + J_NaCaX + J_slow, and they describe the d[Ca]_t/dt flux data quite well. The data curves in Fig. 4C do not appear to approach a clear V_max value because [Ca]_i does not rise as high during the twitch as during a CafC. This limits our ability to make realistic predictions of V_max or K_m for the SR Ca-ATPase. Nevertheless, it is clear that the SR Ca-ATPase flux is much higher at all [Ca]_i in the PLB-KO myocytes. Furthermore, for WT and PLB-KO, respectively, the SR Ca-ATPase fluxes are ~10 and 30 times higher than the sum of J_NaCaX + J_slow (replotted from Fig. 4B).

On the basis of the [Ca]_i dependence determined in Fig. 4 for the different Ca transport systems, we then calculate the cumulative Ca flux through each transporter during a normal twitch in Fig. 5. In this case, we use twitch [Ca]_i as the driving function and the flux values for each system as derived in Fig. 4. The fraction of activating Ca transported out of the cytosol by the SR, Na/Ca exchange, and the slow mechanisms in the WT mouse are 90.3, 9.2, and 0.5%, respectively. This is similar to the values obtained in rat ventricular myocytes by Bassani et al. (1) and Negretti et al. (31). Thus, as in rat ventricular myocytes, the SR Ca-ATPase plays a dominant role in removal of Ca from the cytosol during the normal twitch. In the PLB-KO mouse, the SR Ca-ATPase is even more dominant such that the SR Ca-ATPase accounts for 96.4% of Ca removal during relaxation and the Na/Ca exchange only 3.4%. Although SR Ca uptake is about two times faster in the PLB-KO mouse (Figs. 1, 2, and 4), it will not increase the fractional SR contribution twofold, since it is already >90% in WT. Looking at it another way, ~10% of activating Ca is removed by mechanisms other than the SR Ca pump in WT, but this drops to 3.5% in the PLB-KO mouse.

As we mentioned above, indo 1 compartmentalization in mouse myocytes probably causes an underestimation of J_slow. To test if this underestimation will bias our conclusion about Ca fluxes, we also repeated the entire analysis above (Figs. 4 and 5) using the faster J_slow parameters previously derived for rat ventricular myocytes (1) and consistent with the mouse myocyte relaxation data, but we kept the  and peak data of CafC NT and SS twitch [Ca]_i measured from the WT mouse. This analysis makes only a modest difference in the overall flux contribution. For WT mouse myocytes, the integrated Ca flux attributed to the SR Ca-ATPase, Na/Ca exchange, and slow systems by this new analysis, compared with Fig. 5, was 90.3, 7.2, and 2.5%, respectively. Although Ca transport by the slow systems is higher, the overall impact on the balance between the SR Ca-ATPase and Na/Ca exchange is not altered much.

Frequency dependence and postrest cell contractions. Figure 6A shows superimposed SS twitch contractions in WT and PLB-KO myocytes stimulated at 0.5 and 1 Hz. The strikingly faster twitch kinetics in the PLB-KO myocyte are apparent. WT mouse myocytes exhibit a flat to modestly negative force-frequency relationship over this range of frequencies (Fig. 6B), similar to our observations with rat ventricular myocytes (not shown). The PLB-KO mouse myocytes show a significantly more negative force-frequency relationship. Figure 6C shows that both WT and PLB-KO myocytes exhibit marked and comparable PR potentiation. PR twitch potentiation is also a normal feature in rat ventricular myocytes (7).

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Frequency- and rest-dependent changes in twitch contractions. A: raw shortening traces (as % of RCL) in representative WT and PLB-KO (KO) mouse myocytes during SS twitches at 0.5 Hz (left) and 1 Hz (right, on a different time scale). Each curve is an average of 3 SS original traces. Time to peak contraction and relaxation is faster in the PLB-KO mouse. B: pooled data for amplitude of SS twitches at 0.5 and 1 Hz stimulation. In WT, the reduction in twitch amplitude from 0.5 Hz (n = 45) to 1 Hz (n = 24) was not significant, but in PLB-KO it was significant (* n = 26-42, P < 0.01). C: both WT () and PLB-KO () show potentiation of the first postrest (PR) contraction after 1 Hz stimulation. Data are normalized to the twitch at 2 s rest and are shown as means ± SE (n = 5-10). D: relaxation  of SS (at 1 Hz) and PR (60 s rest) twitches. In both WT and PLB-KO, SS twitches relaxed significantly faster than PR twitches (n = 6-8, P < 0.05).

Figure 6D shows that both WT and PLB-KO myocytes also show a slowing of relaxation at the first PR twitch after 60 s of rest compared with that at 1 Hz SS. In the WT relaxation,  increases from 117 ± 17 ms for the SS twitch to 151 ± 19 ms for the PR twitch (n = 8, P < 0.05). In the PLB-KO, twitch relaxation was overall much faster, and 60 of s rest increases the  of relaxation by almost a factor of two (26.9 ± 2 ms at SS vs. 47.4 ± 6.8 ms at the PR twitch, n = 6, P < 0.05). Such slowing down of relaxation at PR twitches has been previously described in rat ventricular myocytes (37) and is attributed to slow dissipation of CaMKII-dependent phosphorylation during rest (6). Because PLB phosphorylation is considered to be a likely target for this CaMKII effect, the PLB-KO mice provide a unique system to test this possibility. The fact that slowed relaxation at PR twitches is still observed in the PLB-KO mouse myocytes indicates that this effect cannot be solely due to PLB phosphorylation.

We further studied the frequency-dependent acceleration of [Ca]_i decline in PLB-KO myocytes loaded with indo 1-AM to test whether this acceleration of SS twitch relaxation in PLB-KO is still sensitive to CaMKII inhibition. Figure 7 shows that the slowing of the PR (vs. SS) Ca transient decline in these indo 1-loaded PLB-KO cells in Fig. 7C is not as great as observed for unloaded shortening in Fig. 6D. Nevertheless, there is still a significant difference such that  of [Ca]_i decline is 92.9 ± 4.7 ms for SS and 111 ± 5.5 ms for PR (n = 19, P < 0.01). Furthermore, Fig. 7C shows that this difference in  is completely abolished by pretreatment of the PLB-KO myocytes for 5 min with the CaMKII inhibitor KN-93 (1 µM). After KN-93,  for both SS and PR [Ca]_i decline is the same as for the slower control PR , that is, after KN-93,  is 120 ± 4.6 ms for SS and 115 ± 6.5 ms for PR (n = 19, not significant).

View larger version (24K): [in this window] [in a new window]   Fig. 7.   Effect of the Ca/calmodulin-dependent kinase II (CaMKII) inhibitor KN-93 (1 µM) on the  of SS and PR twitch [Ca]i decline in PLB-KO mouse myocytes. A: superimposed Ca transients of SS (average of 3 traces) vs. PR in NT solution (left) are normalized and expanded, showing [Ca]i decline (right). In this case,  of SS and PR were 83 and 119 ms, respectively. B: same as in A, except traces were recorded after 5 min perfusion with 1 µM KN-93 (in a different cell). SS [Ca]i decline is slowed down to about that of control PR. C: mean  values of [Ca]i decline before and after treatment with KN-93. Open bars, SS; filled bars, PR. In control, SS twitch [Ca]i decline was significantly faster than at the PR twitch (n = 19, P < 0.01 by paired t-test). After 5 min treatment with 1 µM KN-93, the difference was abolished (n = 19, not significant). D: pooled values of [Ca]i before and after treatment with KN-93: control SS, 88.9 ± 8.5 nM and PR, 130 ± 12 nM (n = 19, P < 0.001); KN-93 SS, 73.4 ± 4.6 nM and PR, 111 ± 9.8 nM (n = 19, P < 0.001).

The amplitude of the Ca transient can have intrinsic effects on the  of [Ca]_i decline (8). However, the larger PR twitch would be expected to decrease  rather than increase  so that the observed slowing of PR  of [Ca]_i decline might be an underestimate of the functional effect. We also measured the amplitude of the SS and PR twitch Ca transients before and after KN-93 exposure (Fig. 7D). In both cases, the PR twitch [Ca]_i is potentiated (as for contraction in Fig. 6C), and the degree of potentiation is comparable before and after KN-93 treatment. Thus there seems to be a CaMKII-dependent process that accelerates [Ca]_i decline during SS stimulation, but this process still occurs in the complete absence of PLB.

Immunoblots of the Na/Ca exchange. We have previously shown that the transcript levels of the sarcolemmal Na/Ca exchange mRNA were not altered in PLB-KO hearts compared with WT (12). Because the Ca transient and contraction data (e.g., Figs. 1B and 2D) imply slower Na/Ca exchange function in the PLB-KO myocytes, we also assessed the protein levels of the Na/Ca exchange in six PLB-KO and six WT mouse hearts using quantitative immunoblotting in conjunction with the chemiluminescent detection system (Fig. 8). On parallel lanes of each of eight polyacrylamide gels, various concentrations of homogenates from WT and PLB-KO hearts were loaded. The densitometric signals obtained were linear functions of protein concentration over the range of 10-20 µg protein loaded onto gel lanes for both actin and the Na/Ca exchanger. The signals from PLB-KO mouse hearts were compared with those of WT mouse hearts, which were set as 100%. The amount of protein loaded on polyacrylamide gels and the efficiency of protein transfer onto the membrane were verified by probing three of the same blots with a mouse monoclonal antibody to actin. Actin served as an internal control, since its protein levels were similar between WT and PLB-KO mouse hearts (12). Our results indicate no significant alteration in the expression levels of the Na/Ca exchange protein in the PLB-KO mouse hearts (97 ± 4 vs. 100% in WT).

View larger version (35K): [in this window] [in a new window]   Fig. 8.   Representative quantitative immunoblots of the Na/Ca exchange and actin using homogenates (10, 15, and 20 mg) of 6 pooled hearts from WT or PLB-KO mice. Increasing concentrations of cardiac homogenates were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and incubated with an anti-Na/Ca exchange monoclonal antibody. The same blot was also probed with an anti-actin monoclonal antibody. This served as an internal control for protein loading and the efficiency of protein transfer onto the membrane.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

In the present study, we have provided new information with respect to 1) Ca regulation and contraction in normal mouse ventricular myocytes, 2) how Ca regulation is altered in the PLB-KO mouse, and 3) the mechanism of frequency-dependent acceleration of relaxation. In addition to confirming the faster kinetics of relaxation and [Ca]_i decline in PLB-KO (15, 27, 28, 43), we show the first quantitative analysis of cellular Ca balance in WT (and PLB-KO) mouse myocytes. Competition between the SR Ca-ATPase and Na/Ca exchange in WT mouse myocytes is quantitatively similar to that reported in rats (1, 31), whereas in the PLB-KO myocytes the SR Ca-ATPase is even more dominant over the Na/Ca exchange. Additional new results show that, in the PLB-KO myocytes (vs. WT), there is increased SS SR Ca content, reduced fractional SR Ca release during E-C coupling (at a given SR Ca load), and reduced ability of the Na/Ca exchange to extrude Ca (even without competition by SR transport). A key finding here is also proof that PLB is not required for the physiologically relevant frequency- and CaMKII-dependent acceleration of SS twitch relaxation and [Ca]_i decline.

Altered E-C coupling. In the present study, SR Ca load during SS twitches is 37% higher in the PLB-KO vs. WT mouse myocytes (based on a 98% increase in [Ca]_i during CafC). This agrees reasonably with the 86% increase in SR Ca content measured by Chu et al. (12) using electron microprobe analysis (EPMA). If all other things were constant, the increased SR Ca load would be expected to produce a greater SR Ca release (4). Indeed, because increased SR Ca load was also reported to increase the fraction of SR Ca release, it would have been reasonable to expect that the SS twitch Ca transient in the PLB-KO mouse would be increased by >37% compared with WT. However, our results indicate that both SS cell contraction and Ca transient are increased only 21 and 6%, respectively, in PLB-KO myocytes (not significant). Indeed, the twitch Ca transient for a given SR Ca load was actually somewhat smaller in the PLB-KO vs. WT myocytes (see Fig. 3).

Masaki et al. (29) reported no alteration in L-type Ca current density in PLB-KO vs. WT myocytes, suggesting that the trigger density for E-C coupling is not changed. The apparently smaller fractional SR Ca release [for a given I_Ca (Ca current) trigger and SR Ca load] could be due in part to the 25% reduction in the number of ryanodine receptors in the PLB-KO mouse (12). However, the more rapid Ca resequestration by the faster SR Ca-ATPase in the PLB-KO mouse may also curtail the peak of the twitch Ca transient. This could also contribute to the shorter time to peak [Ca]_i (Fig. 1A). This effect of the SR Ca-ATPase to curtail the normal twitch Ca transient was reported in control rat and rabbit ventricular myocytes (1).

To assess how reduced fractional SR Ca release and accelerated reuptake might interact, we used a simple model like that developed in Fig. 5 (including a simple SR Ca release flux during the time to peak of the Ca transients; 160 and 140 ms in WT and PLB-KO myocytes, respectively). In WT myocytes, an SR Ca release of 56.3 µmol/l cytosol is sufficient to produce the measured peak free [Ca]_i of 167 nM using the Ca removal parameters from the Ca flux analysis in Figs. 4 and 5. If we increase SR Ca load by 37%, as measured in the PLB-KO mouse, and assume the same fractional release (55%), the SR Ca release would be 77 µmol/l cytosol, and SS twitch [Ca]_i would be expected to be 256 nM [even with the faster SR Ca pump flux in the PLB-KO mouse (from Figs. 4 and 5)]. This 53% increase in predicted [Ca]_i is much higher than the [Ca]_i that we measured in PLB-KO myocytes here (177 nM), suggesting that the increased SR Ca-ATPase activity alone is not sufficient to explain the lack of significant increase in twitch [Ca]_i and contraction in the PLB-KO mouse. To obtain the observed 177 nM [Ca]_i in PLB-KO myocytes with the 37% increase in SR Ca load, we have to reduce the fraction of SR Ca released during the twitch by 16% (from 55 to 46% of the SR Ca load). This possible reduction seems plausible, based on the 25% lower number of ryanodine receptors reported in the PLB-KO compared with the WT heart (12). Thus the reduced number of ryanodine receptors may explain the apparent decrease in fractional SR Ca release in the PLB-KO myocytes (at a given SR Ca load). Put another way, the expected increase in SS twitch Ca transients in PLB-KO myocytes (vs. WT) for a similar Ca current and larger SR Ca load may be limited by an offsetting reduction in the number of ryanodine receptors (which could reduce the fraction of SR Ca released in the PLB-KO myocytes).

Another possible explanation for the limited cellular inotropy is a spatial dropout or inhomogeneity of SR Ca release during the twitch observed by Hüser et al. (19) in the PLB-KO mouse. They suggested that local SR Ca release at certain sites might be prevented by the uniquely strong negative feedback created by the strong SR Ca pump in the PLB-KO mouse. This could limit the ability of SR Ca release at one locus to activate Ca release from a neighboring region.

Other reports have shown larger increases in SS twitch [Ca]_i and contraction in the PLB-KO vs. WT myocytes than the 6 and 21%, respectively, that we found here (15, 43). It is not clear why our results differ in this regard [particularly as conditions were similar to the study by Wolska et al. (43)]. It is quite possible that these studies, which showed larger increases in contraction and [Ca]_i, actually had a higher SR Ca load in the PLB-KO mouse [e.g., 86% higher than in WT (12) compared with the 37% we reported here]. Indeed, SS SR Ca load is very sensitive to numerous factors, such as frequency, temperature, mean [Ca]_i, intracellular Na concentration, extracellular K concentration, action potential configuration, and duration. Plugging the 86% increase in SR Ca load into our model above (with a 25% decrease in fractional release due to fewer ryanodine receptors) would give a 60% increase in the SS twitch [Ca]_i. Thus higher SR Ca load could better offset the reduced number of ryanodine receptors and produce the stronger inotropic effect seen by some others. Indeed, because increased SR Ca load strongly increases the fraction of SR Ca release (4), such an effect could easily explain the quantitative discrepancy. Given the dramatic hyperdynamic state of the PLB-KO heart in the intact animal (27), it seems possible that our results here with respect to SS twitch [Ca]_i and contraction underestimate the in vivo situation. On the other hand, having intracellular Ca transients of more comparable amplitude is functionally advantageous for our main goals here concerning comparison of kinetic parameters (8). In any event, the results here do agree with others with respect to the more rapid kinetics of twitches and Ca transient in the PLB-KO mouse (15, 27, 28, 43).

Balance of Ca fluxes during relaxation. It is particularly important to understand basic cellular Ca handling in mouse myocytes because of the increasing number of studies using mouse myocytes, especially from genetically altered mouse phenotypes. This is the first report analyzing Ca fluxes during relaxation in mouse myocytes. On the basis of the strategy of Bassani et al. (1), we obtain the relative contribution by SR Ca-ATPase, sarcolemmal Na/Ca exchange, and slow Ca removal mechanisms (sarcolemmal Ca-ATPase and mitochondrial Ca uniporter) during relaxation of a single twitch. The results in the WT mouse (90.3, 9.2, and 0.5%, respectively) are very similar to those in rat ventricular myocytes (92, 7, and 1%, respectively; see Ref. 1) but differ greatly from those in rabbit, ferret, and guinea pig ventricular myocytes (1, 2, 7) in which the Na/Ca exchange is a much stronger competitor with the SR Ca-ATPase. It is not surprising that in the PLB-KO mouse the contribution of SR Ca-ATPase is increased (96.4%), with a decreased contribution of Na/Ca exchange (3.4%) and slow mechanisms (0.1%). This is because dephosphorylated PLB normally works as an inhibitor of SR Ca-ATPase. Thus, without PLB, the knockout mouse SR Ca-ATPase becomes an even more dominant competitor with Na/Ca exchange and other Ca transport mechanisms. Because relaxation in other species (e.g., rabbit and guinea pig) normally depends less on SR Ca-ATPase and more on Na/Ca exchange (7), the effect of PLB ablation in these species would be expected to be greater with respect to the percentage of Ca flux via the SR Ca-ATPase during relaxation.

If the Na/Ca exchange in mouse myocytes removes only ~9% of the activating Ca during SS twitch relaxation, it may also be anticipated that only 9% of the activating Ca enters the cell via Ca current during the action potential. Voltage-clamp results in rat ventricular myocytes support this sort of SS conclusion (i.e., the amount of Ca influx during action potential is about the same as Ca efflux for each cardiac cycle; see Ref. 13), and the short action potential in rat limits the amount of Ca influx via I_Ca (45). Similar voltage-clamp studies have not been carried out in mouse myocytes, but the results might be expected to resemble those in rat. In the PLB-KO mouse, the Ca extrusion via Na/Ca exchange was even smaller, which might imply a smaller Ca influx during the action potential. Although Masaki et al. (29) found no difference in peak L-type Ca channel density in WT and PLB-KO mice, they did observe faster I_Ca inactivation in PLB-KO myocytes. This more rapid I_Ca inactivation may well be secondary to a faster rise in [Ca]_i in the PLB-KO (43) and would also result in lower net Ca influx during the action potential.

Changes in slow Ca transport mechanisms. When both the SR Ca-ATPase and Na/Ca exchange were inhibited during CafC 0 Na, 0 Ca, the decline of the Ca transient was very slow in both WT and PLB-KO myocytes ( ~34 and 90 s, respectively). This is much longer than previous observations in rat and rabbit ( ~12 s; see Ref. 1). The problem of mitochondrial indo 1 compartmentalization in mouse myocytes (see METHODS and RESULTS) makes it impossible to draw any clear conclusion from these results. However, given the energetic changes observed in the PLB-KO mouse myocytes, including an increase in the fraction of pyruvate dehydrogenase, which is in the active form (12), it may be of interest to determine if mitochondrial Ca transport is altered. Initial estimates of mitochondrial Ca content (12) showed no difference between WT and PLB-KO using EPMA, but it is difficult to detect small Ca changes with this approach.

Depressed Na/Ca exchange. The  of [Ca]_i decline and relaxation during a CafC in NT was faster in WT than in PLB-KO mice. With SR net Ca uptake blocked, the main mechanism for Ca removal from the myoplasm is Na/Ca exchange (5, 24). These results suggest that the Na/Ca exchange system is slower at extruding Ca in the PLB-KO mouse. Measurements of [Ca]_i decline during CafC in NT should be much less affected by mitochondrial compartmentalization of indo 1. This is because very little Ca is expected to enter mitochondria during relaxation (1). Furthermore, we found 75% slowing of CafC relaxation in NT for PLB-KO vs. WT cells not loaded with indo 1. Thus the 50% slowing of [Ca]_i decline in the PLB-KO mouse probably reflects a true decrease in the ability of Na/Ca exchange to remove Ca from the cytosol in PLB-KO mice.

Chu et al. (12) reported that, in PLB-KO mouse heart, there was no change in Na/Ca exchange mRNA levels, and here we extend this to show that there is no change in Na/Ca exchange protein levels. Thus the decreased Na/Ca exchange function in the PLB-KO mouse might be attributable to changes in exchanger regulation. One possibility would be elevated intracellular Na concentration in the PLB-KO mouse, but EPMA did not reveal any difference in Na content in the myocyte A-band (12).

Enhanced SR Ca load. We found a 37% increase in the SR Ca content in the PLB-KO mouse (based on a 98% increase in caffeine-induced Ca transient amplitude in 0 Na, 0 Ca). This agrees qualitatively with EPMA measurements of 86% increase in SR Ca content (12). Three factors in our experiments could contribute to the increased SR Ca load. First, the absence of PLB is expected to increase the affinity of the SR Ca pump for Ca in a similar manner to the physiological phosphorylation of PLB (23, 26, 41). Second, there is a functional decrease of Ca transport by Na/Ca exchange as discussed above. Any inhibition of the Na/Ca exchange will tend to increase cellular Ca load and result in an even larger fraction of the activating Ca being taken up by the SR. Third, the apparent reduction of fractional SR Ca release in the PLB-KO mouse (see above) would also tend to increase SR Ca load. Thus a combination of faster SR Ca-ATPase, depressed Na/Ca exchange, and reduced fractional SR Ca release may all contribute to the increased SR Ca load in the PLB-KO mouse.

No difference was found in calsequestrin mRNA or protein levels in the PLB-KO mouse (12). Thus the SR Ca buffer capacity may not be changed. Unless there was a large increase in SR volume or other SR Ca buffers, the increased SR Ca load must increase free intra-SR [Ca] and more fully saturate existing SR calsequestrin. This indicates that, in the WT mouse, there was still room to increase intra-SR [Ca] before reaching the thermodynamic limit that the SR Ca pump can generate.

CaMKII-dependent acceleration of relaxation. The frequency- and CaMKII-dependent acceleration of relaxation in cardiac muscle is likely to be an important physiological mechanism of autoregulation (6, 20, 37). The SR Ca-ATPase has been implicated in this autoregulation and is also the major Ca transport system involved in relaxation in mouse ventricular myocytes (Figs. 4 and 5) responsible for >90% of Ca removal during a SS twitch. As such, changes in SR Ca uptake rate can dramatically modify the rate of [Ca]_i decline during a twitch in these cells. Indeed, both SS twitch relaxation and [Ca]_i decline are much faster in PLB-KO mouse myocytes, consistent with stimulation of the SR Ca-ATPase.

In rat ventricle, relaxation of the first PR twitch after a 1-min rest is slower than during SS (6, 37). Because this rest-dependent effect was suppressed or reversed by thapsigargin, ryanodine, caffeine, or by replacement of Ca by Sr (6, 37), it seems clear that the effect depends on SR Ca uptake and on [Ca]_i. Schouten (37) hypothesized that the abbreviation of SS twitches was due to enhanced SR Ca uptake secondary to PLB phosphorylation by CaMKII, which was activated by the cyclic increase in [Ca]_i. Rest would allow PLB dephosphorylation, reversing the frequency-dependent stimulation of SR Ca-ATPase and slowing the time course of the SR Ca uptake. Mattiazzi et al. (30) found that inhibition of endogenous CaMKII-dependent PLB phosphorylation increased the SR Ca uptake rate in permeabilized rat myocytes. Furthermore, Bassani et al. (6) showed that phosphatase inhibition prevents the slowing of [Ca]_i decline at the PR twitch and that CaMKII inhibition can prevent the activation-dependent acceleration of [Ca]_i decline. Thus SR Ca transport and CaMKII are involved in the acceleration of relaxation during SS twitches. However, the involvement of PLB phosphorylation has not been proven.

If PLB was the only target of CaMKII and caused the acceleration of relaxation during SS twitch, this phenomenon should be abolished in the PLB-KO mouse. However, as shown in Figs. 6 and 7, the PR slowing of [Ca]_i decline is still present in the PLB-KO mouse. Indeed, relaxation of the PR twitch was slowed by 76% compared with the SS twitch. This proves that PLB is not required for this effect. Recently, Hussain et al. (20) also reported that the stimulation rate-dependent changes in Ca transient duration in rat are not associated with PLB phosphorylation. On the other hand, they suggested that SR Ca-ATPase might still be involved, since the rate-dependent abbreviation of the Ca transient depends on a functional SR.

The accelerating effect of SS stimulation of the  of [Ca]_i decline could also be abolished by the CaMKII inhibitor KN-93 at 1 µM (a concentration at which this agent is expected to be quite selective for CaMKII; see Ref. 40). Although this result strongly implicates CaMKII in mediation of the acceleration of [Ca]_i decline during SS, this conclusion does rely on the ability and selectivity of 1 µM KN-93 to inhibit CaMKII. It might be speculated that a CaMKII-dependent phosphorylation of the SR Ca-ATPase could be responsible, since Toyofuku et al. (42) and Xu and colleagues (44) have reported direct CaMKII-dependent phosphorylation of the cardiac SR Ca-ATPase, which results in increased V_max for Ca transport. However, this result has been challenged (32, 34), so the identity of the CaMKII target involved in accelerating SR Ca transport during SS twitches in rat and mouse myocytes is not yet clearly identified.