Uncoupling of Ca2+ Transport in Sarcoplasmic Reticulum as a Result of Labeling Lipid Amino Groups and Inhibition of Ca2+-ATPase Activity by Modification of Lysine Residues of the Ca2+-ATPase Polypeptide*

Limited labeling of amino groups with fluorescarnine in fragmented sarcoplasmic reticulum vesicles inhibits Ca2+-ATPase activity and Ca2+ transport. Under the labeling conditions used, 80% of the label reacts with phosphatidylethanolamine and 20% with the CaZ+-ATP- ase polypeptide. This degree of labeling does not result in vesicular disruption or in loss of vesicular proteins and does not increase the membrane permeability to Caz+. Fluorescamine labeling of a purified Ca2+-ATPase devoid of aminophospholipids also inhibits Ca2+-ATP- ase activity, suggesting that labeling of lysine residues of the enzyme polypeptide is responsible for the inhi- bition of Ca2+-ATPase activity in sarcoplasmic reticulum. Fluorescamine labeling interferes with phos- phoenzyme formation and decomposition in both the native vesicles and the purified enzyme; addition of ATP during labeling, and with less effectiveness ADP or AMP, protects both partial reaction steps. Addition of a nonhydrolyzable ATP analog protects phosphoenzyme formation but not decomposition. The inhibition of Ca2+ transport but not of Ca2+-ATPase occurs in sarcoplasmic reticulum vesicles labeled in the presence of ATP, indicating that the transport reaction is uncoupled from the Ca2+-ATPase reaction. The inhibition of Ca2+ transport but not of Ca2+-ATPase activity is also found in sarcoplasmic reticulum

Limited labeling of amino groups with fluorescarnine in fragmented sarcoplasmic reticulum vesicles inhibits Ca2+-ATPase activity and Ca2+ transport. Under the labeling conditions used, 80% of the label reacts with phosphatidylethanolamine and 20% with the CaZ+-ATPase polypeptide. This degree of labeling does not result in vesicular disruption or in loss of vesicular proteins and does not increase the membrane permeability to Caz+. Fluorescamine labeling of a purified Ca2+-ATPase devoid of aminophospholipids also inhibits Ca2+-ATPase activity, suggesting that labeling of lysine residues of the enzyme polypeptide is responsible for the inhibition of Ca2+-ATPase activity in sarcoplasmic reticulum. Fluorescamine labeling interferes with phosphoenzyme formation and decomposition in both the native vesicles and the purified enzyme; addition of ATP during labeling, and with less effectiveness ADP or AMP, protects both partial reaction steps. Addition of a nonhydrolyzable ATP analog protects phosphoenzyme formation but not decomposition. The inhibition of Ca2+ transport but not of Ca2+-ATPase occurs in sarcoplasmic reticulum vesicles labeled in the presence of ATP, indicating that the transport reaction is uncoupled from the Ca2+-ATPase reaction. The inhibition of Ca2+ transport but not of Ca2+-ATPase activity is also found in sarcoplasmic reticulum vesicles in which only phosphatidylethanolamine has reacted with fluorescamine. Furthermore, the extent of labeling of phosphatidylethanolamine is correlated with the inhibition of Ca2+ transport rates. The inhibition of Ca2+ transport is a reflection of the inhibition of Ca2+ translocation and is not due to an increase in Ca2+ efflux. We propose that labeling of phosphatidylethanolamine perturbs the lipid environment around the enzyme, producing a specific defect in the Ca2+ translocation reaction.
Fragmented sarcoplasmic reticulum vesicles transport Ca2+ coupled to the hydrolysis of ATP. The elementary steps of the Ca*+-ATPase reaction have been studied in detail (for recent reviews, see Refs. 1 and 2). It is generally agreed that the reaction is initiated by binding of Ca2+ to the high affinity sites of the enzyme, followed by enzyme phosphorylation. This results in Ca2+ translocation presumably triggered by a conformational change in the enzyme molecule. After CaZ+ is translocated, the enzyme undergoes a decrease in affinity for * This research was supported by Grant HL 23007 from the National Institutes of Health and a grant from the American Heart Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Ca2+, allowing Ca2' release to the vesicular interior. The dephosphorylation of the enzyme resulting in Pi liberation completes the reaction cycle.
The isolated SR' vesicles have a relatively simple protein and lipid composition (3). This, together with the considerable knowledge of the mechanism of the Ca2+-ATPase reaction, makes the SR system particularly suitable for studying the relationship between structure and function. With this aim in mind, a complex of fluorescamine with cycloheptaamylose was used in previous studies as a nonpermeant covalent label to analyze the disposition of proteins and aminophospholipids in SR. It was found that the Ca2+-ATPase was significantly labeled with CFC, indicating that a considerable fraction of the Ca2+-ATPase polypeptide is exposed to the outside of the SR vesicles; calsequestrin did not react with CFC, indicating that it is buried in the vesicular interior (4). In addition, the labeling pattern of PE, the major aminophospholipid present in purified SR, indicates that PE has a highly asymmetric disposition, with 70-80% of it located in the external side of the membrane bilayer (4). The same asymmetric disposition of PE has been observed after labeling SR with other aminolabeling reagents (5,6 ) .
Labeling SR with CFC results in marked inhibition of both Ca2+-ATPase and Ca2' transport, although the labeled vesicles retain their structural integrity (4). The inhibition of the Ca2+-ATPase reaction caused by CFC labeling is due to inhibition of the phosphorylation step and can be prevented by the presence of ATP during labeling ( 7 ) . In contrast, the presence of ATP during labeling does not prevent the inhibition of Ca2+ transport, which is due presumably to a specific defect in the translocation reaction (7). Since in previous experiments even limited labeling with CFC resulted in labeling of both protein and lipid amino groups, it was not possible to ascertain unambiguously whether the labeling of lysine residues in the Ca2+-ATPase polypeptide or the modification of the polar head group of PE produces the inhibition of Ca" transport.
In the present study, fluorescamine was chosen instead of CFC to label SR amino groups since labeling with fluorescamine allows more precise control of the extent of labeling and, as will be shown below, it produces inhibitory effects similar to those of CFC. However, labeling with fluorescamine has to be restricted to a few amino groups; addition of 10 or more mol of fluorescamine/105 g of SR protein results in loss of Uncoupling of Calcium Transport in Sarcoplasmic Reticulum 209 cdsequestrin (4). In experiments using a purified Ca2'-ATPase devoid of amino phospholipids, it was found that the inhibition of phosphoenzyme formation, and of the Ca'+-ATPase reaction, observed after labeling with fluorescamine is due to modification of a limited number of lysine residues of the Ca"-ATPase polypeptide chain. Furthermore, labeling of SR with fluorescarnine under conditions in which the label reacts only with the polar head group of PE results in inhibition of Ca2+ transport but not of Ca2+-ATPase activity. We propose that labeling of PE perturbs the interactions between the enzyme and its surrounding phospholipids resulting in the observed defect in the Caz+ translocation reaction, which is presumably more sensitive to perturbations of lipid-protein interactions than the phosphorylation step and the subsequent release of inorganic phosphate.

EXPERIMENTAL PROCEDURES
Preparation of SR, Purified Ca"-ATPase, and Phospholipidreplaced Ca2'-ATPase-Fragmented SR was prepared from rabbit skeletal white (fast) muscle as described in detail previously (8), except that muscle was homogenized in 0.1 M KC1, 20 mM Trismaleate, pH 7.0, instead of in 0.3 M sucrose, 20 mM Tris-maleate, pH 7.0. Also, care was taken to eliminate the very light density material that contains significant amounts of transverse tubule vesicles (9), by sedimenting the muscle microsomes at 50,OOO-75,OOO X g instead of 150,000 X g. The SR vesicles were resuspended in 0.3 M sucrose, 20 mM Hepes, pH 7.2, and stored frozen at -20 "C. For the preparation of a phospholipid-replaced ATPase (devoid of PE), the Ca"-ATPase enzyme, purified from SR by extraction with deoxycholate, was used as starting material. To purify the Ca"-ATPase, the procedure of Warren et al. (IO) was modified in that the solubilization with deoxycholate was carried out with a total of 600 mg of SR, and the resulting suspension was layered on top of discontinuous sucrose gradients (15 ml each of 50% sucrose (w/v) and 25% sucrose in 1 M KCI, 50 mM potassium phosphate buffer, pH 8.0,5 mM dithiothreitol).
The tubes were centrifuged for 17 h at 5 "C in a Beckman SW 27 rotor at 100,OOO X g. After collecting the Ca2'-ATPase from the interface of the two sucrose solutions, it was diluted 10-fold in 0.3 M sucrose, 20 mM Tris-maleate, pH 7.0, and was collected by centrifugation at 4 OC for 60 min at 100, OOO X g. The resulting pellets, resuspended in 1.0 M KCI, 0.3 M sucrose, 5 mM dithiothreitol, 50 mM potassium phosphate buffer, pH 8.0, at a protein concentration of 10 mg/ml, were incubated with slow stirring at 0 "C for 2 h with a sonicated mixture of deoxycholate and DOPC. A weight ratio of protein/deoxycholate/DOPC of 1:1:2 was used. The purified enzyme containing DOPC (DOPC-ATPase) was obtained by centrifugation in discontinuous sucrose gradients, as described previously (11). Since after one cycle of phospholipid replacement some PE still remained associated with the enzyme, a second DOPC replacement was carried out in the same way as described above. The DOPC-ATPase preparation obtained after two replacements contain over 99.5% phosphatidylcholine, as shown by two-dimensional thin layer chromatographic analysis of its lipid extract.* Deoxycholate was obtained from Sigma and was recrystallized prior to use (12). DOPC and A23187 were obtained from Calbiochem and were used without further purification.
Labeling with Fluorescarnine-To label the SR vesicles or the purified DOPC-ATPase preparation with fluorescamine, the vesicles were resuspended at 22 "C in 0.1 M KC1, 20 mM Hepes, pH 7.2, at a concentration of 2 mg of protein/mL An acetone solution of fluorescamine was rapidly added under continuous stirring, taking care to maintain the final acetone concentration at or below 1%. The labeling reaction proceeds to completion in a few seconds (13). The same concentration of acetone was added to the unlabeled vesicles as a control. The labeled vesicles were diluted and collected by sedimentation and resuspended in 0.3 M sucrose, 20 mM Tris-maleate, pH 7.0. Fluorescamine was purchased from Roche, Determination of Fluorescence of Protein and Lipid Components-Lyophilized samples containing 2 mg of protein were extracted with 2 X 2 ml of chloroform/methanol (2:1), followed by 2 ml of chloroform/methanol (1:2) as described previously (11). The com-Since DOPC-ATPase contains 0.77 pmol of phosphorus/mg of protein (77 mol/105 g), this indicates that the upper limit for the amount of PE still associated with the enzyme is 0.4 mol/lOs g. bined extracts were adjusted to 6.0 ml final volume. To determine lipid fluorescence, a 0.75-ml fraction of the lipid extract was evaporated to dryness and resuspended in 2.0 ml of 1% SDS. The remaining lipid extract was evaporated to dryness and resuspended in 0.2 ml of chloroform/methanol (2:l) to analyze phospholipid composition as described below. The protein residues obtained after the extraction of 2 mg of SR with chloroform/methanol were washed with 2 ml of 70% acetone, dried, and resuspended in 2 ml of 10% SDS by sonication in a bath-type sonicator. To determine protein fluorescence, a 0.05ml fraction of the solubilized protein residue was diluted to 2 ml with 1% SDS. Fluorescence intensity was determined with excitation and emission wavelengths of 390 and 465 nm, respectively (4) in a Perkin-Elmer spectrofluorometer, Model MPF-4, equipped with a temperature-controlling unit. All fluorescence measurements were carried out at 25 "C, in 1.0% SDS, and at a protein concentration of 0.015-0.025 mg/ml. Measurement of Fluorescence Incorporation into SR Lipids-A fraction of the lipid extracts containing 0.1-0.25 pmol of phosphorus was applied to Silica Gel G thin layer plates and the phospholipids were separated by running the plate in one dimension with chloroform/methanol/water (56:25:4). This solvent system allows complete separation of PE (RF = 0.34) from fluorescamine-labeled PE (RF = 0.23) and from the rest of the SR lipids (RF values lower than 0.12).
The amount of phosphorus present in the PE spot, the labeled PE spot, and in the combined spots for the other phospholipids were determined by perchloric acid hydrolysis as previously described (11). The only fluorescent spot containing measurable amounts of phosphate corresponds to fluorescamine-labeled PE.
Enzymatic Determinations-Ca2+-ATPase activities were either measured colorimetrically as described previously (11) or by means of a radiometer pH stat. All measurements were carried out at 22 "C. The reaction solution contained 0.1 M KCI, 5 mM MgC12, 0.1 mM CaC12, 4 mM ATP, 20 mM Tris-maleate, pH 7.0, and 0.05 mg of protein/ml. The buffer was omitted when the reaction was carried out in the pH stat using a pH of 7.0.
Steady state levels of phosphoenzyme as well as P, liberation rates at 0 "C were measured using [y-3'P]ATP as described previously (11). The reaction mixture contained 0.1 M KCI, 5 mM MgCL, 0.1 mhi CaC12, 0.01 mM [y-:'"P]ATP, 20 mM Tris-maleate, pH 7.0, and 0.1 mg of protein/ml. The reaction was stopped by adding 7% trichloroacetic acid. To measure phosphoenzyme, 0.5-ml fractions were filtered through HA Millipore filters, previously washed with 1 ml of 0.1 mM ATP. After extensive washing of the acid-denatured protein retained in the filter as described previously (ll), the filters were dried and counted in a liquid scintillation counter. The amount of P, liberated was determined by extraction of the phosphomolybdate complex into an organic phase. A 0.7-ml fraction of the acid-quenched sample was centrifuged to remove precipitated protein, and 0.1 ml of 3.758 ammonium molybdate in 3 N H2S04 and 1.0 ml of isobutanol/xylene (35:65, v/v) were added to 0.4 ml of the supernatant. The mixture was stirred for 45 s and centrifuged for 5 min to separate the phases, and 0.5 ml of the organic phase was counted in a liquid scintillation counter. The amount of liberated P, was calculated by using a calibration curve obtained by carrying out the same extraction procedure as above with standard phosphate solutions containing J2P. Initial rates of phosphoenzyme formation were measured at 4 "C by means of a three-syringe Durrum multimixer Model 133 as described in detail elsewhere (14). A solution containing 0.1 M KCI, 10 mM MgC12, 0.1 mM CaC12, 20 mM Tris-maleate, pH 7.0, and 2 mg of protein/ml was mixed with an equal volume of a solution containing 0.1 M KCI, 20 mM Tris-maleate, pH 7.0, and 0.02 mM [y-""PIATP. The reaction was quenched at different times by addition of 10% trichloroacetic acid. The amount of phosphoenzyme formed was measured as described above.
Ca" uptake in the presence of oxalate was measured at 22 "C in a solution containing 0.1 M KCI, 5 mM K-oxalate, 5 mM MgC12,O.l mM "CaCI,, 20 mM Tris-maleate, pH 7.0, and 0.025 mg of protein/ml. The reaction was started by addition of 5 mM ATP and was stopped at different times by filtering 1.0 ml of the reaction through HA Millipore filters. A 0.2-ml fraction of the filtrate was placed on filter paper strips, dried, and counted by liquid scintillation. To measure Ca2+-ATPase activities in the presence of oxalate at 22 "C, the same reaction conditions as in the Ca" uptake measurements were used except that the solution contained 5 mM [Y-~~P]ATP and 0.1 mM CaC12 instead of "CaCI2. The amount of P, liberated was determined by extracting a fraction of the filtrate as described above.
To measure Ca2+ efflux, 8 mM EGTA was added to the SR vesicles at different times after starting the Ca2' uptake reaction as described by guest on March 24, 2020 http://www.jbc.org/ Downloaded from above. Samples were collected at I-min intervals after addition of EGTA for a total period of 15 min. standard.

RESULTS
It has been shown previously (13) that fluorescamine reacts quantitatively with SR amino groups at pH 8.3, provided the protein concentration is 2 mg/ml or more and the reagent concentration is in the range of 0.04-0.2 mM.
Using this concentration of SR, we found that addition of 0.04-0.16 mM fluorescamine (2 to 8 mol of fluorescamine/105 g of SR protein) at pH 7.2 results in a linear increase in fluorescence intensity (Fig. 1); the slope decreased when the concentration of fluorescamine was increased to 0.2 mM (10 mo1/105 g). Reagent concentrations are expressed in terms of moles/105 g of SR protein in Fig. 1 due to the fact that the molecular weight of the Ca2'-ATPase, which represents 75-80% of the total SR protein, is about 1.1 x lo5, so that the unit of moles/105 g of SR protein is roughly equivalent to moles of reagent/mol of enzyme. Control experiments indicate that labeling at pH 7.2 results in the same fluorescence intensities as those obtained by labeling at pH 8.3.
After separating the SR lipid and protein components as described under "Experimental Procedures," 80% of the total fluorescence of SR is present in the lipids and 20% in the protein (Fig. 1). The fluorescence in the lipids is due to labeling of PE, whereas the fluorescence in the protein is due to labeling of the Ca2+-ATPase polypeptide, as evidenced by the fluorescence pattern of SDS-containing polyacrylamide gels of the protein extract (not shown).
It has been shown previously that 7540% of the total PE present in SR is labeled with CFC (4). Since the SR vesicles used in these studies contain on the average 0.65 pmol of phosphorus/mg of protein and 13-14% PE (65 mol of phospholipids and 9 mol of PE/105 g of protein, respectively), this corresponds to a total of 7 mol of PE/105 g of protein that are labeled with CFC. It is interesting to note that in the present work the fluorescence intensity of the SR vesicles increases linearly up to 8 mol of fluorescamine/105 g of protein.
Above 8 mol of fluorescamine, lipid labeling levels off, while protein labeling continues to increase (Fig. 1). Analysis of the lipid extract of SR vesicles labeled with 10 mol of fluorescambe/ lo5 g of protein indicates that 78% of the PE present in SR, corresponding to 7 mol of PE/105 g of protein, has reacted with the label (Fig. 1). This clearly shows that the reaction of PE with fluorescamine proceeds to the same extent as the reaction of PE with CFC (7). Furthermore, from the fluorescence intensity of the lipid extract and from the number of moles of PE labeled, it is possible to calculate the fluorescence intensity/mol of amino group labeled. Using this number and the fact that the total fluorescence of SR in SDS is the sum of the fluorescence of its lipid and protein components (Fig.  l), the total number of lysine residues labeled can be calculated. AS shown in Fig. 1, a maximum number of 7 mol of PE and 2 mol of lysine/105 g of protein is labeled when a concentration of 10 mol of fluorescamine/105 g of protein was used. This degree of lipid labeling is in agreement with previous reports by Hasselbach et al. (13), who also found a plateau of lipid labeling at a concentration of fluorescamine of 0.1 pmol/ mg of protein (10 mo1/105 g). Effect of Labeling on Ca2+-ATPase Activity-Unlabeled SR vesicles display at 22 O Ca2'-ATPase activities in the range of 0.6-0.7 pmol of Pi/mg/min. Increasing the Ca2+ permeability of the SR vesicles by addition of the ionophore A23187 results in 2.5-to 3-fold stimulation of Ca2+-ATPase activity. As observed previously with CFC-labeled SR (7), labeling of SR vesicles with fluorescamine results in inhibition of Ca2+-ATPase activity, measured either with or without A23187 in the reaction solution ( Fig. 2). At low concentrations of fluorescamine, the percentage of inhibition of Ca2'-ATPase activity in the absence of A23187 is comparable with the percentage of inhibition in the presence of the ionophore (Fig. 2). This suggests that the vesicles labeled with low reagent-toprotein ratios retain their permeability barrier to Ca2' after labeling, since the above results indicate that A23187 stimulates the Ca2+-ATPase activity in the labeled vesicles to the Same extent as the control. ATP added prior to labeling protects Ca*+-ATPase activity (Fig. 2). For protection of activity, it is required only that ATP be present during labeling, since no inhibition of Ca2'-ATPase was observed in vesicles labeled in the presence of ATP, then sedimented and resuspended in ATP-free solution prior to the Ca2+-ATPase assay. The same protective effect was observed by addition prior to

FIG. 2. Inhibition of Caa+-ATPase activity produced by la-
beling SR vesicles with fluorescamine. Ca2+-ATPase activities were measured as described in the text. Ca2+-ATPase activities measured with (0) or without (0) A23187 present in the reaction solution.
A, Ca2+-ATPase activity of SR vesicles labeled with fluorescamine in the presence of 5 mM ATP, measured either with or without A23187.
The moles of NH, labeled were calculated as described in the text. As seen previously with CFC ( 7 ) , labeling of SR with fluorescamine produces inhibition of phosphoenzyme formation (Table I). Addition of ATP prior to labeling prevents the inhibition in both cases (Table I, Ref. 7 ) .
Most likely, the inhibition of Ca*+-ATPase is due to labeling of lysine residues of the enzyme polypeptide. It has been shown recently that a large fraction of the total lysine residues present in the Ca2'-ATPase is contained in hydrophilic sequences (16), which makes it likely that they are exposed to the outside of the SR vesicles. To test whether labeling lysine residues with fluorescamine results in inhibition of Ca2+-ATPase activity, a purified ATPase preparation devoid of PE, DOPC-ATPase, was used. It was found that increasing label incorporation, as revealed by increasing fluorescence, results in progressive inhibition of both steady state phosphoenzyme levels and Pi liberation rates (Fig. 3). This clearly shows that labeling of lysines results in inhibition of Ca'+-ATPase. It is interesting to note that at the same extent of labeling the inhibition of the rate of P, liberation is more pronounced than the reduction of steady state phosphoenzyme levels (Fig. 3).
To further analyze this, the percentage of steady state phosphoenzyme levels and the percentage rate of P, liberation observed after labeling DOPC-ATPase with fluorescamine were plotted as a function of the number of lysine residues modified (Fig. 4). Although there is some deviation at the lower end due to the low levels of phosphoenzyme formed, the values for percentage of steady state phosphoenzyme fit a straight line that intercepts the abscissa (100% inhibition) at 7 mol of lysine labeled/1O5 g of enzyme. These results suggest that modification of 1 out of 7 lysine residues, all with equal reactivity to fluorescamine, results in 100% inhibition of phosphoenzyme formation. If we assume that labeling of a different lysine of the 7 with equal reactivity produces inhibition of phosphoenzyme decomposition but not of phosphoenzyme formation, at a given number of lysine residues modified, the rate of P, liberation should be inhibited more than the steady state phosphoenzyme levels (since phosphoenzyme decomposition cannot take place in the absence of phosphoenzyme formation, labeling of either the lysine required for phosphoenzyme formation or for decomposition would result in inhibition of Pi liberation). The predicted curves for the cases in which labeling of the lysine residue involved in phosphoen-   zyme decomposition causes either 50 or 100% inhibition of P, liberation rates are shown in Fig. 4. The experimental points fall between both curves, suggesting that blocking of this lysine residue results in more than 50% but less than 100% inhibition of P, liberation rates.
As shown above, addition of ATP prior to labeling SR with fluorescarnine protects CaZ+-ATPase activity and steady state phosphoenzyme levels ( Fig. 2 and Table I). ATP added prior to labeling DOPC-ATPase with fluorescamine also protects steady state phosphoenzyme levels and P, liberation rates, which reflect the Ca"-ATPase activity of the DOPC-ATPase preparation (Table 11). A concentration of ATP as low as 0.05 mM still affords partial protection.
ADP and AMP also protect the enzyme against inhibition of both activities although less efficiently than ATP does (Table 11). However, addition of AMP-PNP before labeling results in protection only of phosphoenzyme formation since the inhibition of Pi liberation rates is the same as that observed after labeling without AMP-PNP (Table  11). This finding supports the proposed model for there being two classes of lysine residues in the enzyme, so that modification of one class of lysine would result in inhibition of phosphoenzyme formation and of the other, in inhibition of P, liberation. It is important to note at this point that the inhibition of P, liberation does not result in a significant increase in the levels of the steady state phosphoenzyme intermediate. The reason for this is that at 0.01 mM ATP and at 0 "C, the apparent rate constant for phosphoenzyme formation (17 s", Table 111) is much higher than the rate of phosphoenzyme decomposition (0.09 s", calculated as the ratio between the rate of Pi liber-

TABLE I1
Labeling of DOPC-ATPase with fluorescamine: protective effect of nucleotides on steady state phosphoenzyme levels and P, liberation rates Samples were labeled as described in the text, diluted to remove the added nucleotides, and collected by centrifugation. The pellets were resuspended in 0.3 M sucrose, 20 mM Tris-maleate, pH 7.0, at the original volume. Phosphoenzyme and P, liberation rates were determined as described in Table I.

TABLE I11 Effect of labeling SR with fluorescamine on initial rate of phosphoenzyme formation and on steady state phosphoenzyme levels measured at 4 OC at three different substrate concentrations
For details, see text. Apparent rate constants of phosphoenzyme formation were calculated by assuming f i s t order reaction kinetics, using the final steady state level of phosphoenzyme (Columns 5 and I " 6) as the maximum value in each case. ation of 11.97 nmol mg" min" and the steady state phosphoenzyme value of 2.11 nmol mg", Table 11). From these values it follows that if the rate of phosphoenzyme decomposition would become zero, the steady state phosphoenzyme level would only increase by 0.5%. However, although the model proposed is the simplest that can account for the experimental observations, its validity can only be tested by direct identification in the primary sequence of the lysine residues labeled with fluorescarnine in the presence or in the absence of ATP or of AMP-PNP. Similar degrees of labeling were found in the presence or absence of ATP (not shown). This can be reconciled with the proposal that ATP prevents labeling of specific lysine residues (that once modified interfere with the Ca'+-ATPase reaction) by assuming that lysine residues that in the absence of ATP do not react with fluorescamine have now become labeled.
Compared with the considerable inhibition of steady state phosphoenzyme levels produced by labeling, the relative lack of inhibition of phosphoenzyme formation rates (Table 111) is compatible with the proposal that labeling of a specific lysine inhibits phosphoenzyme formation. Enzyme molecules in which the lysines involved in phosphorylation are labeled would not be phosphorylated by ATP, whereas enzyme molecules in which such lysines have not been labeled would carry out the phosphorylation reaction a t near normal rates.
Effect of Labeling on ea2+ Transport-The results obtained with DOPC-ATPase suggest strongly that labeling of lysine residues results in the inhibition of steady state phosphoenzyme levels and of Ca2'-ATPase activity in SR. However, even in the presence of ATP, labeling of SR with fluorescamine produces inhibition of Ca2+ uptake rates, although the labeled vesicles display normal phosphoenzyme levels and Ca"-ATPase activities (Table I, Fig. 2).
It is important to resolve whether the selective inhibition of Ca2+ uptake rates is due to labeling of amino groups of the protein, of the lipid, or of both. Although DOPC-ATPase would represent a suitable system to test the effect of protein labeling on Ca" uptake, under the conditions used in the present work to prepare DOPC-ATPase, no Ca2+ uptake can be measured, presumably due to leakiness of the replaced enzyme vesicles. Looking for conditions to label SR lipids only, SR was labeled in the presence of bovine serum albumin (BSA). The rationale behind this experiment is that if an exogenous protein such as BSA, which contains a large number of lysine residues, was added to SR prior to labeling, fluorescamine would react preferentially with the lysines of the exogenous protein and would not label the Ca"-ATPase polypeptide. Furthermore, since fluorescamine reacts preferentially with PE, it was likely that even if a large number of exogenous lysines were present some labeling of P E would still take place. That this is indeed the case is illustrated in the experiment depicted in Fig. 5.
When SR was labeled in the presence of BSA, and then separated from BSA by two cycles of sedimentation, it was found that all the label was covalently bound to PE and none to the Ca2'-ATPase polypeptide. No inhibition of phosphoenzyme formation rates (not shown), steady state phosphoenzyme levels (Table IV), or Ca'+-ATPase activity (Fig. 5) was observed in these conditions, conclusively showing that these are due to protein labeling. In contrast, the inhibition of Ca'+ uptake rates persisted (Table IV). The labeled vesicles accumulate as much Ca2+ as the unlabeled control (Table IV). This finding makes unlikely the possibility that the inhibition of Ca2+ uptake rates is due to the presence of vesicles that have become fully permeable to Cas+ as a result of labeling. However, experiments that conclusively rule out the possibility that labeling produces the inhibition of Ca" uptake rates  IV Inhibition of Ca" uptake rates in SR by labeling phosphatidylethanolamine with fluorescamine in the presence of BSA SR vesicles were labeled with fluorescamine in the presence of BSA as described in Fig. 5, so that only PE reacted with the label. The SR vesicles used in this experiment contained 20 pg of lipid phosphorus/mg of protein (0.65pmol/mg), of which 13% corresponded to PE. This gives a total of 8 mol of PE/105 g of protein. The number of moles of PE modified was calculated from this number and from the percentage of the total PE modified by fluorescamine. The Ca2+ uptake rates are expressed as mean f S.D.; the number of determination is in parentheses. For details on the determination of phosphoenzyme and Ca'+ uptake rates, see text. Ca2+ uptake was measured in the presence of 5 mm mM oxalate. by creating a leaky vesicular subfraction will be presented below.
To demonstrate unambiguously inhibition of Ca2+ uptake rates but not of Ca"-ATPase, both activities should be meas-ured under the same conditions (i.e. in the presence of oxalate). Measuring directly the amount of Pi liberated during the uptake reaction, as described under "Experimental Procedures," c o n f m s t h a t PE labeling results in inhibition only of Ca2+ uptake rates and not of Ca2+-ATPase activity (Table  V). Furthermore, a clear correlation between the extent of PE labeling and the degree of inhibition of Ca2+ uptake rates was found (Fig. 6), further indicating that t h e inhibition is due to the modification of PE.
The inhibition of the initial rates of Ca2+ uptake is a reflection of the inhibition of the ATP-dependent Ca2' influx and is not due to an increase in Ca2+ efflux (Table V). To measure the ATP-dependent Ca2+ influx, a tracer amount of 45Ca was added 30 s after starting the uptake reaction by addition of A T P t o vesicles equilibrated with 4"Ca. Samples were collected at 10-s intervals after addition of 4sCa. Since the efflux is initially composed only of 4"Ca, the '5Ca initially transported into the vesicles represents only the ATP-dependent influx component and not the net uptake rate. However, since in the presence of oxalate the efflux is such a small fraction of the influx (Table V), the initial rate of uptake is practically equal to the ATP-dependent Ca'+ influx. To measure Ca2+ efflux, EGTA was added 1 min after starting the and Ca2+ efflux, all measured in the presence of oxalate Ca2+ uptake and Ca2*-ATPase activity were measured at 22 "C in the presence of oxalate as described in detail in the text. Samples were collected at 15-s intervals after starting the reaction by addition of ATP to vesicles equilibrated for 2 min in the reaction solution containing '%a for uptake or ""Ca for ATPase determinations. To measure ATP-dependent Ca2+ influx, Ca2+ uptake was initiated in the presence of 40Ca by addition of ATP; after 30 s, a tracer amount of Ca was added and samples were collected at 10-s intervals for a total period of I min. Caz+ efflux was measured by addition of 8 mM EGTA 1 min after starting the uptake reaction by addition of ATP to vesicles preincubated for 2 min in the presence of 45Ca. The numbers represent mean -+ S.D.; numbers in parentheses indicate number of determinations.  (6) 2100 * 96 (6) 1192 f 103 (7) 1226 f 44 ( uptake reaction in the presence of oxalate. Since the observed low values for the efflux could conceivably be due to slow dissociation of the calcium oxalate precipitate inside the vesicles, and if so they would not reflect the true membrane permeability to Ca'+, a control experiment in which EGTA and the ionophore A23187 were added at early times during the uptake reaction was carried out. Simultaneous addition of A23187 and EGTA 1 min after starting the uptake reaction with ATP stimulates the efflux rate about 50-fold over that seen with EGTA alone, from 40 to 2000 nmol/mg/min, showing that at the early phase of the uptake reaction the dissociation of calcium oxalate is not the rate-limiting step in the efflux of Ca2+. This finding demonstrates conclusively that in the early phase of the uptake reaction the efflux measured after addition of EGTA is in fact controlled only by the membrane permeability to Ca2+. It has been shown recently (17) that during the early phase of Ca2+ uptake in the presence of oxalate, the crystallization of calcium oxalate does not reach equilibrium, so that addition of A23187 produces a marked increase in Ca2+ efflux by competing with oxalate for intravesicular Ca2+. In the present work, addition of A23187 in the absence of EGTA 1 min after the initiation of the reaction either stops Ca2+ uptake or produces a small increase in Ca2+ efflux (not shown), indicating that under these conditions the ATP-dependent Ca2+ influx is approximately equal to the ionophore-induced Ca2+ efflux.
As discussed above, after 10 min of starting the uptake TABLE VI Comparison between control SR vesicles and SR vesicles containing fluorescamine-labeled PE in terms of Ca2+ uptake rates, ea2+ efflux, and calcium oxalate loading SR vesicles were labeled with fluorescamine in the presence of BSA as describedin Fig. 5, using a concentration of fluorescamine of 10 mol/105 g of SR protein. The Ca" uptake reaction was carried out in a total volume of 225 ml, at a protein concentration of 0.01 mg/ml. To measure the Ca2+ uptake rate, I-ml samples were taken at 15-s intervals for the f i t 2 min after starting the reaction by addition of ATP. The amount of Ca2+ taken up by the vesicles was measured as described in the text. 20 min after starting the uptake reaction, 210 ml of the reaction solution were sedimented by centrifugation at 100,000 X g for 60 min at 4 "C. The resulting pellets, resuspended in 1.0 ml of a solution containing 0.1 M KC1, 5 mM MgCl,, 5 mM Koxalate, 0.1 mM 45CaC12, 20 mM Tris-maleate, pH 7.0, and 4 mM ATP (Ca" uptake solution), were layered on top of a discontinuous sucrose gradient composed of two equal layers of 35 and 50% sucrose (w/v) in Ca2+ uptake solution. After centrifugation for 75 min at 150,000 X g at 4 "C, most of the protein recovered from the gradient was found as a pellet, and the rest at the interface of the 35 and 50% sucrose layers. The pellets were resuspended in 0.3 M sucrose, 20 mM Tris-maleate, pH 7.0, concentrated by sedimentation at 100,000 X g, and resuspended once more in 0.3 M sucrose, 20 mM Tris-maleate, pH 7.0. The amount of calcium oxalate accumulated by the vesicles was determined by measuring the 45Ca content of a fraction of the resuspended pellets. The extent of PE modification with fluorescamine was determined by thin layer chromatographic analysis of lipid extracts of the initial as well as of the loaded vesicles. A detailed description of the lipid analysis procedure is given in the text. To measure Caz' efflux, the same conditions as described in Table V  reaction by addition of ATP, the vesicles containing fluorescamine-labeled P E contain the same amounts of calcium oxalate as the unlabeled vesicles (Table IV), indicating that although they take longer they eventually accumulate as much calcium oxalate as the unlabeled veiscles. To completely rule out the possibility that the labeled vesicles represent a leaky subfraction, both labeled and unlabeled vesicles were fractionated in sucrose gradients after loading with calcium oxalate as described in detail in Table VI. A relatively high concentration of fluorescamine that produced 55% inhibition of Ca2+ uptake rates but no increase in Caz+ efflux (Table VI) was chosen to label the vesicles to be loaded with calcium oxalate and fractionated in sucrose gradients. The same fraction of the initial vesicular population was found loaded with calcium oxalate for both control and labeled vesicles (Table  VI). If the vesicles had become progressively leaky as a function of the extent of labeling, a lower fraction of the initial vesicular population for the labeled vesicles should have been found loaded with calcium oxalate as compared to the unlabeled control vesicles. Furthermore, both the unfractionated labeled vesicles as well as the calcium oxalate-loaded subfraction had the same content of fluorescamine-labeled PE (Table  VI), showing conclusively that labeling does not produce a leaky membrane subfraction.
These results clearly indicate that the inhibition of Ca'+ uptake rates produced by labeling P E is due to a defect in the transport process itself and it is not caused by membrane leakiness or by inhibition of the Ca2+-ATPase reaction.

DISCUSSION
Fluorescamine reacts with primary amino groups and yields highly fluorescent products (18), which makes it suitable as an amino-labeling reagent. However, labeling of SR with fluorescamine at pH 8. 3 and at a reagent-to-protein ratio of 10 mol of fluorescamine/105 g of SR results in dissociation of calsequestrin from the SR membranes (4). This is presumably the cause of the significant extent of calsequestrin labeling observed under these conditions (4, 13). For these reasons, in previous experiments aimed at analyzing the disposition of proteins and lipids under conditions in which the vesicles retain their original structure even after extensive labeling, CFC, which is a water-soluble derivative of fluorescamine, was used (4, 7). Labeling of SR with CFC produces considerable inhibition of Ca" transport and Ca2+-ATPase activity (4, 7), indicating that there are some amino groups in SR crucial for these activities. Furthermore, under certain labeling conditions, such as limited labeling of SR with CFC in the presence of ATP, it was found that the labeled vesicles display normal Ca*+-ATPase activity but decreased Ca2+ transport rates (7), indicating that uncoupling of Ca"+ transport has taken place. Since even limited labeling with CFC in the presence of ATP results in covalent labeling of both the Ca2'-ATPase polypeptide chain and PE (7), it was not possible to ascertain unambiguously whether it is the modification of amino groups in the protein, or in the lipid, or maybe of both, which results in inhibition of Ca2+ transport. In the present work, fluorescamine was chosen instead of CFC since fluorescamine, as opposed t o CFC, reacts quantitatively and quickly with free amino groups (18), allowing a better control of the number of amino groups modified than CFC does. Also, since with CFC it was found that at the initial stages of the labeling reaction a larger fraction of the label was covalently bound to PE and much less to the Ca*+-ATPase, it was hoped that by limiting the amount of fluorescamine to 1 or 2 mo1/105 g of SR it would be possible to label only PE to analyze whether this had any effect on Ca2+ transport.
To avoid the problem of calsequestrin dissociation, the concentration of fluorescamine was kept below 10 mol/105 g of SR, and labeling was carried out at pH 7.2 instead of 8.3. No dissociation of calsequestrin was observed under these conditions, and the labeled vesicles retained their permeability barrier to Ca". However, although a t a concentration of 1 or 2 mol of fluorescamine/105 g of S R most of the label reacts with PE, still significant label incorporation in the Ca2+-ATPase polypeptide chain was observed. This was avoided by labeling S R in the presence of BSA, as will be discussed below.
Labeling of the Ca2+-ATPase Polypeptide Chain and Inhibition of the ca*+-ATPase Actiuity-The results obtained with DOPC-ATPase conclusively show that modification of lysine residues results in inhibition of Ca2'-ATPase activity. It is interesting to note in this regard that covalent labeling of SR amino groups with two other reagents also results in varying degrees of inhibition of Ca"-ATPase activity (19,20). Using pyridoxal-5'-phosphate followed by reduction with NaBH4, progressive loss of Ca2+-ATPase activity was observed (19). Furthermore, Ca-ATP and Mg-ATP offer complete or partial protection, respectively, against this inactivation (19). It was proposed that modification of a single lysine residue, located in the ATPase fragment of M, = 30,000, is responsible for the loss of ATPase activity (19). The results obtained in this work in addition to confirming that modification of lysine residues produces inhibition of Ca'+-ATPase, clearly indicate that the inhibition of the Ca'+-ATPase is due to inhibition of both phosphoenzyme formation and phosphoenzyme decomposition caused presumably by modification of two different lysine residues. Addition of ATP prior to labeling, and with less effectiveness ADP and AMP, preserves steady state phosphoenzyme levels and decomposition. It is likely that these nucleotides prevent the reaction of fluorescamine with lysine residues that are crucial for both phosphoenzyme formation and decomposition.
The results obtained with AMP-PNP indicate that there are lysine residues whose modification interferes only with phosphoenzyme decomposition and not with phosphoenzyme formation, since addition of AMP-PNP before labeling results in protection only of the lysine groups involved in phosphoenzyme formation.
The protective effect of ATP was interpreted by Murphy (19) as an indication of an active site location of one or more lysine residues, since lysines might bind to the anionic substrate Mg-ATP. Allen and Green (21) have reported most of the sequence of a 31-residue peptide containing the active site aspartyl residue that is phosphorylated by ATP. Of possible significance is the location of a lysyl residue adjacent to it, Although active site residues need not be near one another in the primary structure, it is likely that modification of this lysine residue could be blocked by ATP and by AMP-PNP, which would bind to the substrate site. Furthermore, this lysine residue could be involved in the phosphorylation reaction and a different lysine (one or more) in the phosphoenzyme decomposition steps. However, it is possible to visualize other protective mechanisms, such as allosteric effects, which might involve lysine residues other than those present in the active site.
As analyzed under "Results," the inhibition of steady state phosphoenzyme formation and of the rates of Pi liberation can be explained assuming that there are seven lysine residues in each enzyme molecule that have equal reactivity to fluorescamine. Labeling of one of them would result in total inhibition of phosphoenzyme formation (and hence of P, liberation rates) and of another, in inhibition of P, liberation rates without affecting steady state phosphoenzyme levels. However, it is important to note here that this model involving seven lysines with equal reactivity to fluorescarnine is valid for the purified Ca"-ATPase only, since it is conceivable that after purification more lysine residues than in the original SR vesicles can react with the label. In fact, this was shown to be the case for SR and purified enzyme labeled with CFC (4). However, the data shown in Table I and in Fig. 2 indicate that labeling of 2 mol of lysine/105 g of protein in SR produces 50% inhibition of phosphoenzyme formation and 70% inhibition of Ca2+-ATPase activity. Since the SR vesicles contain about 70% Ca"-ATPase and since this is the only protein that reacts with fluorescamine under the labeling conditions used in this work, this represents about 3 mol of lysine labeled/mol of enzyme. If SR had seven lysines of equal reactivity to fluorescamine, labeling of three should produce 43% inhibition of phosphoenzyme and 70% inhibition of Ca"-ATPase, which is within the range of the observed experimental values. So it is conceivable that both the purified enzyme system as well as the original vesicles have the same number of lysine residues of equal reactivity to fluorescamine and that blocking of one prevents phosphoenzyme formation and of the other, phosphoenzyme decomposition.
Uncoupling of Caz+ Transport-As shown under "Results," labeling of SR with fluorescamine in the presence of BSA produces sealed vesicles in which only PE has reacted with the label. The fact that labeling of P E produces inhibition of the ea2+ transport rates but does not affect the rate or the extent of phosphoenzyme formation, or the rate of phosphoenzyme decomposition, indicates that the transport reaction has been uncoupled from the hydrolysis of ATP. Although the intermediate reaction steps of the Ca2'-ATPase reaction have been clarified in considerable detail (1,2), very little is known concerning the actual molecular mechanisms involved in the translocation of Ca" from the outside to the inside of the SR vesicles. For this reason, it is difficult to visualize how the enzyme proceeds with the intermediate steps of the reaction a t normal rates but transports Ca2+ to the vesicular interior at a fraction of the rate of the unlabeled vesicles. Since leakiness has been ruled out, the present results suggest that there are membrane configurations in which the rate of ATP hydrolysis is not controlled by the rate at which CaL+ is transported into the vesicular lumen. The uncoupling of CaZ+ transport from the Ca2+-ATPase produced by limited tryptic digestion of SR (22) or by proton inactivation (23) has been explained by postulating selective defects in the enzyme polypeptide. The present findings would be the fmst example of uncoupling produced not by a change in the enzyme polypeptide itself but in the lipids around the enzyme.
As to the nature of the defect induced by P E modification, there is some indication that it might be due to a highly localized modification of the membrane in the immediate vicinity of the Ca2+-ATPase. Modifications of the lipid environment that produce changes in membrane fluidity result in significant changes in Ca2+-ATPase activity (11,(24)(25)(26). As expected from the lack of effect on Ca'+-ATPase activity, no changes in membrane fluidity are observed after PE modification, since the same EPR spectra were detected with fatty acid spin probes added to the SR membranes with or without modified PE.3 In support of the view that the labeled phospholipid is in the close vicinity of the enzyme is the finding that modification of as little as 1 mol of PE/mol of enzyme results in significant quenching of intrinsic tryptophan fluorescence due to fluorescence energy transfer between the enzyme tryptophan residues as donors and the fluorescaminelabeled PE as a~ceptor.~ Furthermore, another fluorescent P E analog, dansyl PE, when incorporated at low concentrations

Uncoupling of Calcium Transport in Sarcoplasmic Reticulum
into sealed SR vesicles also establishes fluorescence energy transfer with the CaZC-ATPase tryptophan residues (27); the interaction between these two chromophores persists even after the SR vesicles have been completely dissociated by SDS or deoxycholate. All these data suggest a strong and close range interaction between the enzyme and labeled PE analogs. It is conceivable that the modification of the polar head group of PE, a phospholipid which is externally located (4-6), causes a change in the immediate environment around the enzyme (such as a change in local charge in the external surface of the membrane) that uncouples the transport of ea2+ from the Ca2'-ATPase reaction.