Reactive disulfide compounds induce Ca2+ release from cardiac sarcoplasmic reticulum

ARCHIVES

OF BIOCHEMISTRY

Vol. 282, No. 2, November

AND

BIOPHYSICS

1, pp. 275-283,199O

Reactive Disulfide Compounds Induce Ca2+ Release from Cardiac Sarcoplasmic Reticulum’ Sumanth

D. Prabhu’

University

of Pittsburgh,

and Guy Salama” School of Medicine,

Department

of Physiology, Pittsburgh,

Pennsylvania

15261

Received April 6,1990, and in revised form June 13,199O

Reactive disulfide compounds (RDSs) with a pyridyl ring adjacent to a disulfide bond, 2,2’dithiodipyridine (2,2’ DTDP) and 4,4’ dithiodipyridine (44’ DTDP), induce Ca2+ release from isolated canine cardiac sarcoplasmic reticulum (SR) vesicles. RDSs are absolutely specific to free sulfhydryl (SH) groups and oxidize SH sites of low pK,, via a thiol-disulfide exchange reaction, with the stoichiometric production of thiopyridone in the medium. As in skeletal SR, this reaction caused large increases in the Ca2+ permeability of cardiac SR and the number of SH sites oxidized by RDSs was kinetically and quantitatively measured through the absorption of thiopyridone. RDS-induced Ca2+ release from cardiac SR was characterized and compared to the action of RDSs on skeletal SR and to Ca2+-induced Ca2’ release. (i) RDS-induced Ca2+ release from cardiac SR was dependent on ionized Mg2+, with maximum rates of release occurring at 0.5 and 1 mM Mg& for 2,2’ DTDP and 4,4’ DTDP, respectively. (ii) In the presence of adenine nucleotides (0.1-l mM), the oxidation of SH sites in cardiac SR by exogenously added RDS was inhibited, which, in turn, inhibited Ca2+ release induced by RDSs. (iii) Conversely, when the oxidation reaction between RDSs and cardiac SR was completed and Ca2’ release pathways were opened, subsequent additions of adenine nucleotides stimulated Ca2+ efIlux induced by RDSs. (iv) Sulfhydryl reducing agents (e.g., dithiothreitol, DTT, l-5 mM) inhibited RDS-induced Ca2+ efflux in a concentration-dependent manner. (v) RDSs elicited Ca2’ efflux from passively loaded cardiac SR vesicles (i.e., with nonfunctional Ca2+ pumps in the absence of Mg-ATP) and stimulated Ca’+-dependent AT-

i This work was supported by grants from the Western Pennsylvania Affiliate of the American Heart Association, The American Heart Association (87-1065) and the National Science Foundation (DCB-89 18672) to G.S. * Supported by the Physician Investigator Training Grant from the National Institutes of Health (5T32-DK0745807). ” Recipient of Research Career Development Award from the National Institutes of Health (5K04 NS00909) and to whom correspondence should be addressed. 000:1-9861/90 $3.00 Copyright (ra1990 by Academic All rights of reproduction

Press,

Inc.

in any form reserved.

Pase activity, which indicated that RDS uncoupled Ca2+ uptake and did not act at the Ca2+, Mg2+-ATPase. These results indicate that RDSs selectively oxidize critical sulfhydryl site(s) on or adjacent to a Ca2+ release channel protein channel and thereby trigger Ca2+ release. Conversely, reduction of these sites reverses the effects of RDSs by closing Ca2+ release channels, which results in active Ca2+ reuptake by Ca2+, Mg2+ATPase. These compounds can thus provide a method to covalently label and identify the protein involved in Ca2+ release from cardiac SR. o lwo Academic PRESS, I~C.

In cardiac muscle, Ca2+ efflux from sarcoplasmic reticulum (SR)4 occurs via a specific release pathway distinct from the SR Ca2+ uptake process (l-5). This release process is more pronounced in junctional rather than longitudinal SR, is activated by micromolar Ca” and millimolar adenine nucleotides, and is inhibited by millimolar Mg2+ and micromolar ruthenium red (4,6,7). Using density gradient centrifugation and binding to [“HI ryanodine, the cardiac SR Ca2+ release channel has been identified as a 450-kDa “foot” protein (6, 7). Incorporation of either the isolated protein or the cardiac SR vesicles into planar lipid bilayers has revealed the presence of ligand-gated, high conductance, Ca2+ channels modulated by the same agents that alter Ca2+ release in cardiac and skeletal muscle (6,8-10). Despite substantial progress made in identifying the proteins involved in SR Ca”’ release, the exact physiologic mechanism underlying SR Ca2+ release remains unclear. Current theories include: (i) Ca”+-induced Ca2+ 4 Abbreviations used: SR, sarcoplasmic reticulum; SH sulfhydryl; Tris, tris(hydroxymethy1) aminomethane; AP III, antipyrylazo III, RR, ruthenium red; EGTA, ethylene glycol bis(fl-aminoethyl ether) N,N’-tetraacetic acid; CK, creatine kinase; CP phosphocreatine; DTT, dithiothreitol; 2,2’ DTDP, 2,2’ dithiodipyridine; 4,4’ DTDP, 4,4 dithiodipyridine; SPDP, N-succinimidyl 3-(2-pyridyl) dithiopropionate; RDSs, reactive disulfide compounds; PR, phenol red; CICR, Ca*+-induced Ca*+ release; BH, biotin hydrazide; TP, thiopyridone. 275

276

PRABHU

AND

release (11, 12), (ii) voltage-dependent change displacement of regularly spaced feet structures which span the gap across the triadic junction thus opening SR Ca” channels (13, 14), and (iii) inositol 1,4,5triphosphateinduced Ca2+ release (1516). Studies from this laboratory in skeletal muscle have demonstrated an alternative method to activate SR Ca2+ release via sulfhydryl (SH) oxidation (17-20). A variety of SH reagents including heavy metal ions (Ag+, Hg’+, Cu*+, Zn’+, Cd2’) (17,18), mercaptans (e.g., cysteine, homocysteine, cysteamine) in the presence of low (2 PM) Cu2+ (a catalyst of disulfide bond formation) (19), hypochlorous acid (19), plumbagin (19), and phthalocyanine dyes (2) all induce Ca*+ release from skeletal SR. Anthraquinones, including the antineoplastic drug doxorubicin, have also been shown to be potent stimulators of Ca2+ release in skeletal SR, acting via a direct oxidative coupling with the Ca2+ release channel (21). Recently, this laboratory has shown that a class of compounds called “reactive” disulfides (RDSs) such as 2,2’ dithiodipyridine (2,2’ DTDP), 4,4’ dithiodipyridine (4,4’ DTDP), and N-succinimidyl 3(2-pyridyl) dithiopropionate (SPDP) trigger Ca2+ release from actively and passively loaded skeletal SR vesicles (22). RDSs specifically oxidize free SH sites via a thiol-disulfide exchange reaction with the stoichiometric production of thiopyridone (23). As with the other SH reagents, these compounds induce Ca2’ release via specific interaction with the Ca2+ release channel but not the Ca2+, Mg2’ATPase pump. Synthesis of biotin-conjugated RDSs and covalent linkage of biotin to skeletal SR protein has been used to identify a 106-kDa protein (24). The 106-kDa protein was purified by biotin-avidin chromatography and revealed the existence of Ca2+ release channels when reconstituted in planar lipid bilayers. Immunological analysis indicated that this protein is not Ca2+, Mg2+ATPase, nor did it appear to be a proteolytic fragment of the high molecular weight ryanodine receptor complex. The relationship of 106-kDa proteins to the feet proteins and its role in physiological Ca” release remains to be elucidated (24). Although it is clear that SH reagents markedly increase skeletal SR Ca” permeability, this phenomenon has yet to be explored fully in cardiac muscle. Recently we have shown that the heavy metal ions Agi and Hg2+ are potent stimulators of Ca2+ release from cardiac SR (25). These ions most likely act at a site “on” or “adjacent” to the physiological Ca2+ release channel. In this report, the effects of RDSs on Ca2+ release from cardiac SR are examined in the presence and absence of known stimulators of physiological Ca2+ release. As in skeletal muscle, RDSs induce Ca2+ release from actively or passively loaded SR by specifically reacting with the Ca2+ release channel and not the ATPase pump. Biotin-conjugated RDSs retain their ability to induce Ca2+ release

SALAMA

and may be used as a tool to identify the sulf’hydryl-activated cardiac SR Ca2+ release channel. MATERIALS

AND METHODS

Preparation of cardiac SR. Cardiac SR vesicles were prepared from canine ventricular tissue as described by Chamberlain et al. (26). Hearts from mongrel dogs (used as donor animals in liver transplantation experiments) were provided by the Department of Surgery after the animals were sacrificed by lethal anesthetic injection. A protease inhibitor, phenylmethylsulfonyl fluoride (1 mM) and/or leupeptin (1 fig/ml), was added to all solutions used to isolate SR vesicles. These were added primarily to retard proteolytic breakdown of Ca*+, Mg*+ATPase pumps and thereby maintain active Ca*+ uptake (25). After the last centrifugation step, the SR vesicles (5-10 mg/ml) were suspended in a medium containing 0.29 M sucrose, 3 mM NaN,, 10 mM imidazole HCI, pH 6.9, and stored in liquid nitrogen until used. Protein concentrations were determined by the method of Lowry et al. (27). Measurements of Cd+ transport. Ca2+ uptake and efflux from cardiac SR vesicles were measured spectrophotometrically through the differential absorption changes of antipyrylazo III (AP III), a metallochromic Ca” indicator dye, at 720-790 nm (28). Absorption changes were quantitatively and kinetically measured with a time sharing, dual wavelength spectrophotometer (Biomedical Instrumentation, University of Pennsylvania, Johnson Research Foundation, Philadelphia, PA) (29). The specificity of AP III as an indicator of free [Ca’+] and its application for similar SR experiments have been demonstrated previously (30). The concentrations of RDSs and dithiothreitol (DTT) used in this experiment did not produce significant changes in the absorption of AP III and did not interfere with measurements of [Ca2’]rree. A typical transport experiment utilized a standard l-ml assay containing: 0.2 mM AP III, 0.5 mM MgCl,, 2.5 u creatine kinase (CK), 4 mM phosphocreatine (CP), and 0.5-1.0 mM MgATP in 120 mM K ‘, 75 mM Pi at pH 7.0. CK and CP provided an ATP regenerating system to maintain a constant concentration of ATP and of free Mg*+ during Ca2+ transport experiments. A single addition of cardiac SR (0.2-0.6 mg) was made to the l-ml cuvette, and uptake of background Ca”+ in the reaction medium was monitored. Two aliquots of CaCl, (525 pM) were then added to calibrate the absorption changes of AP III and to actively load the vesicles. To obtain rapid and reproducible levels of total Ca”’ uptake by cardiac SK, time was allowed for complete sequestration of each Ca” addition before adding more Ca*+. Once Cazt sequestration was completed, Ca”’ release was triggered by adding various concentrations of a RDS (either 22 DTDP, 4,4’ DTDP, or SPDP-biotin hydrazide (BH)) (lo-200 PM). Upon completion of RDS-induced Ca2+ release, the Ca*+ ionophere A23187 was added to the reaction mixture to abolish any existing Ca”’ gradients and to determine the amount of releasable intravesicular Ca*+. The effect of DTT (O-5 mM) on RDS-induced Cazf release was examined by adding it to the initial reaction mixture and measuring uptake and release, as described above. For each Ca2+ transport experiment, the rate of Ca*+ efflux was determined from the initial slope of Ca2+ release (averaged over the first 60 s) just after addition of a RDS and expressed in nanomoles of Ca’+/mg of SR protein/min. The percentage Ca” release was determined from the ratio of total Ca2+ released to total releasable Ca’+. The latter was determined after the addition of A23187 and was, wit,hin experimental error, equal to total Ca*+ accumulated by the SR vesicles. Cd’+ eflux from passively loaded SR vesicles. SR vesicles were incubated in 2 mM CaCl, at 4°C for 12 h to allow enough time for equilibration of intravesicular and extravesicular Ca2+ concentration. Several aliquots (100 ~1) of the passively Ca*’ -loaded SR stock solution were then preincubated with 100 pM 2,2’ DDP either in the presence or in the absence of 1 mM CAMP. Aliquots of control SR (not preincubated with 2,2’DTDP) or RDS-preincubated SR were then added manually but rapidly to 1 ml of Ca” release assay medium while measuring free extravesicular Ca2+ in the medium through the absorption changes of

SH OXIDATION

TRIGGERS

Ca *+ RELEASE

FROM

CANINE

AP III, at 720-790 nm. The release assay medium was free of added ATP or Ca2’ and consisted of 0.1 mM AP III, 0.5 mM MgCl,, 100 mM KCI, 20 mM Tris at pH 6.8 and either 0 or 1 mM CAMP. The addition of Ca” -loaded SR (25 ~1) (either control or preincubated with a RDS) to 1 ml of assay medium resulted in a rapid change in differential absorption due to the opacity of the vesicles and addition of extravesicular Ca’+. This rapid change was followed by a slower change in differential absorption of AP III which represented Ca” release from the lumen of the vesicles to the extravesicular space, and was dependent upon the passive permeability of the SR membrane. After completion of passive efflux, A23187 was added to collapse all existing Ca” gradients. The difIerentia1 absorption signals of AP III were then calibrated by an addition of CaCl,. Rates of Ca”+ efflux were determined from the slope of the passive Ca’+ efflux curves (during the first 60 s) and expressed in units of nanomoles of Ca”/mg of SR protein/min. Measurements of thiopyridone production during Ca” efflux from SR. RDSs exhibited large spectral changes following a reduction reaction initiated by the addition of a mercaptan (31). The decrease in absorption of the RDS occurred concomitantly with the loss of the parent RDS compound and the production of the corresponding thiopyridone (TP). Thus, the oxidation of a mercaptan such as cysteine bv 2,” DTDP or 4,4’ DTDP results in the stoichiometric production of 1 molecule of 2.TP, or 4.TP, respectively. By measuring the time course of thiopyridone production spectrophotometrically, the oxidation of SH groups on SR proteins was monitored quantitatively. Parallel experiments measuring the time course of TP production and SR Ca” release were performed. SR vesicles (247 pg) were suspended in 1 ml of 120 mM K’, 75 mM P,, 1 mM MgCl,, 2.5 u CK, 4 mM CP, 061 m.M MgATP. Two 5 fiM additions of CaCl, were made to load the vesicles and either 2,2’ DTDP (20 mM) or 4,4’ DTDP (100 PM) was added to trigger Ca’+ release. Differential absorption at 340-310 nm (for 2thiopyridone) and 324-350 nm (for 4-thiopyridone) measured TP production, and 10 nmol of cysteine was added to calibrate the signals. Ca” transport was measured under identical conditions except that 0.2 mM AP III was added to the reaction mixture and differential absorption changes were measured at 720-790 nm. Mrnsurement,s of Ca’+-dependent ATPaw activity. Ca’+-dependent ATPase activity was determined spectrophotometrically by monitoring proton production during ATP hydrolysis using the differential absorption changes ofphenol red (PR) at 5077540 nm. A typical experiment utilized an assay medium (2 ml) consisting of 0.1 mM PR, 710 fig SR, 1 mM MgCl,, in 5 mM Na oxalate, 120 mM KCI, 2 mM Hepes at pH 7.0. Under these conditions 0.5 mol of H’ are released per mole of ATP cleaved (X2). CaCl, (25 PM) aliquots were added to the medium, followed by 1 mM MgATP to initiate and measure ATP. After each experimental run, HCl was added to calibrate the absorption signals. The effects of 2,2’ DTDP (200 lg) and/or A23187 (2 PM) on ATPase activity were measured by adding one or both of these reagents to the initial reaction mixture and allowing 10 min of preincubation prior to adding CaCl, and MgATP. Basal ATPase (Ca”-independent ATPase) activity was determined by adding 7 mM EGTA then MgATP to the reaction medium. CaZ’ -dependent ATPase activity was determined by subtracting basal ATPase activity from total ATPase activity and was expressed in units of nanomoles P,/mg protein/min. Materials. All reagents were of analytical grade. All reagents used were purchased from Sigma except antipyrylazo III, which was from ICN Pharmaceuticals (Plainview, NY). A23187, 2,2’ DTDP, and 4,4’ DTDP were dissolved in ethanol stock solutions of 1 mg/ml, 10 mM, and 10 mM, respectively. PDP-BH was formed by dissolving equimolar amounts of SPDP and BH in dimethylsulfoxide and reacting at room temperature for 2 h.

RESULTS

RDSs induce SR Ca” release. RDS reagents are absolutely specific for thiol groups and preferentially oxidize SH groups of low pK,, at low substrate concentra-

Ca’CI,

CARDIAC

277

SR VESICLES

t 2.2’-DTDP

B.

550 v-9 SR I \ CaCI, 1 25 pM

7 4.4’-DTDP

c I 2 min

FIG. 1. RDSs induce Ca2+ release from actively loaded cardiac SR. CaZ+ transport across SR vesicles was determined by measuring exabsorption changes of travesicular [Ca2+]rree through the differential AP III. Cardiac SR vesicles (0.55 mg protein) were added to a cuvette (final volume, 1 ml) containing a reaction medium with an ATP-regenerating system. The vesicles actively pumped the background ionized Ca” in the medium, which was measured through slow changes in base line signal. An aliquot of Ca ” (25 fiM) was added to the reaction mixture and sufficient time was allowed for the vesicles to sequester the added Ca” before adding a second aliquot of Ca”‘. Various concentrations of 2,2’DTDP (A) or 4,4’DTDP (B) were then added to trigger Ca”’ release. At the end of each run, A23187 was added to measure the total releasable pool of Ca”.

tions and high reaction rates (23). Micromolar concentrations of these compounds interacted with SR protein(s) and thus triggered Ca2+ release from cardiac SR vesicles. In Fig. lA, 550 pugof SR was added to 1 ml of assay medium containing 0.2 mM AP III, 1 mM MgClz, 2.5 u CK, 4 mM CP, and 0.5 mM MgATP in 120 mM K’, 75 mM Pi at pH 7.0. After background Ca’+ and two subsequent additions of CaCl, (25 PM) were actively sequestered by the SR, the addition of 2,2’ DTDP (40-150 PM) induced Ca2+ efflux. After efIlux was complete, all existing Ca2+ gradients were collapsed by adding A23187 (2 pg), thus liberating the remaining available intravesicular Ca*+ not released by 2,2’ DTDP. Similar results were obtained using 4,4’ DTDP (lo-200 PM) instead of 2,2 DTDP (Fig. 1B). The total Ca2+ released by RDS plus A23187 was, within experimental error, equal to the total Cazt accumulated (Figs. 2A and 2B). The rate of Ca’+ release averaged over the first minute of Ca’+ efflux and the percentage Ca*’ release are plotted as a function of 2,2’ DTDP (Fig. 2A) and 4,4’-DTDP (Fig. 2B) concentrations. The rate of Ca*+ release increased from 2.27 to 27.75 nmol Ca*+/mg protein/min with 2,2’ DTDP concentra-

278

PRABHU

AND

SALAMA -

A

50 -

-

Rate oi Release

""""*--

Percent

- 100

B

- 90

.-F E

Rate of Release

-""''o-"'

50 -

Percent

Release

- 100

Release - 90

- 80

-80

[4,4’-DTDP]pM

[2,2’-DTDP]pM

FIG. 2. Rate and percentage of Ca2+ release as a function of RDS concentrations. Ca*’ release was induced by 2,2’ DTDP (A) or 4,4’ DTDP (B) as depicted in Fig. 2. Release rates were calculated from the first 60 s of the efflux curve and expressed as nmol Ca’+/mg SR protein/min. Percentages of Ca*+ release were determined from the ratio of Ca2+ released to total Ca*’ accumulated. The latter was equal to (within 10%) total Ca2+ released in the presence of the Ca*’ ionophore A23187.

tions increasing from 1 to 100 PM. Higher concentrations of 2,2’ DTDP (150 FM) resulted in a significant dropoff in the rate of Ca2+ efflux (15.2 nmol/mg/min). The percentage Ca2+ efllux reached a maximum of 78.9% at 40 PM 2,2’ DTDP and was followed by a plateau and a slight decline thereafter. The pattern of 4,4’DTDP-induced release was similar except that the maximum rates of Ca2+ efflux were higher (30.1 nmol/mg/min with 100 PM 4,4’ DTDP). A significant dropoff in efflux rates was again observed with higher concentrations of 4,4’ DTDP (16 nmol/mg/min with 200 PM 4,4’ DTDP). The percentage Ca”+ release induced by 4,4’DTDP achieved a near maximal level (75.4%) at 25 PM 4,4’ DTDP, followed by a plateau and slight increase thereafter. SR Ca2+ release and thiopyridone production. RDS are known to be absolutely specific for free sulfhydryls (23). As in skeletal muscle (22), RDSs like 2,2’DTDP or 4,4’ DTDP interacted with SR protein by a thiol-disulfide exchange reaction resulting in the formation of a mixed disulfide bond between the SH group on SR proteins and a pyridyl ring. The reaction produced 1 mol of thiopyridone (either 2-TP or 4-TP) per mole of oxidized sulfhydryl. The exchange reaction is effectively unidirectional, proceeding in the direction of thiopyridone production (23). The spectral characteristics of RDSs before and after the addition of mercaptans or SR vesicles have been previously reported (22,31). Reduction of a RDS and the concomitant production of its thiopyridone increased the peak absorption at 270 nm and 340 nm for 2-TP and at 230 nm and 324 nm for 4-TP production. In Fig. 3, parallel experiments were performed to measure Ca2+ release and 4-TP production induced by

the addition of 4,4’ DTDP SR vesicles loaded with Ca2+. In the top trace, SR Ca2+ uptake was initiated by the addition of 247 pg SR to a medium containing an ATP regenerating system. After two successive additions of 5 PM CaC12, release was induced with 100 PM 4,4’ DTDP. In the lower trace, 4-TP production was monitored at 324-350 nm under identical conditions, except for the

’ AP III Abs. Increase 720-790 nm Ca2+ Efflux

5;M CaCI, ’ Thiopyridone Production Abslncrease 324-350 nm

I

I

I

FIG. 3. Measurement of thiopyridone production. Parallel experiments were performed measuring Ca*+ transport and 4.TP production during 4,4’ DTDP-induced Ca” release from actively loaded SR. Ca*+ fluxes were measured with AP III at 720-790 nm. 4-TP production was measured at 324-350 nm in the absence of AP III. The time course of 4-TP production slightly preceded the time course of Ca” efflux. Approximately 29 nmol of SH sites were oxidized per milligram SR.

SH OXIDATION

I

AP III Abs. Increase 720 - 790 nm Ca*’ Efflux

t CaCI, 1opd

+

TRIGGERS

dr

RELEASE

f/

b 40’pA 2.2~DTDP

Ca”

2 min

I I

FIG. 4. DTT inhibits Ca*’ release induced hy RDSs. Varying concentrations (O-3 mM) of DTT were added to the initial reaction medium hefore active Ca2’ loading of the vesicles and release induced by 2,2’ DTDP (40 PM). DTT had no effect on uptake hut resulted in a concentration-dependent inhibition of RDS-induced Ca2+ efflux. Similar results were obtained with 4,4’ DTDP and SPDP (not shown).

absence of AP III. The addition of 4,4’ DTDP yielded a fast phase of 4-TP production which preceded Ca2+ efflux slightly, followed by a slower phase of 4-TP production. The production of thiopyridone indicated that 4,4’ DTDP oxidized a maximum of 20 + 4 nmol of SH sites per milligram SR (rz = 6 different SR preparations). Thus, 4,4’ DTDP interacted selectively with a minor subset ( 1 mM) in both skeletal (18) and cardiac SR (25). To measure the effect of free Mg2+ on RDS-induced Ca2+ release, cardiac SR vesicles were actively loaded with Ca2+ in the presence of varying concentrations of Mg2+ (0.55 InM). Once Ca” uptake was complete, RDSs were used to stimulate Ca2+ release. As shown in Fig. 5A, increasing concentrations of Mg2+ inhibited 2,2’ DTDP-induced Ca2+ release. Increasing free Mg2+ from 0.5 to 5 mM decreased Ca2+ efflux rates by 72% (31.1 nmol/mg/ min to 8.75 nmol/mg/min), and the percentage Ca2+ release by 66%. With 4,4’ DTDP (Fig. 5B), Ca2+ efflux

FROM

CANINE

CARDIAC

279

SR VESICLES

rates increased from 0.5 to 1 mM Mg2’, plateaued between 1 and 2 mM Mg2+, and decreased thereafter. Reduced rates of Ca2+ release were not caused by diminished Ca2’ gradients across the SR membrane as total Ca2+ uptake was the same (105 nmol Ca”‘/mg SR protein for 2,2’ DTDP experiments and 85 nmol Ca”+/mg SR protein for 4,4’ DTDP experiments) at all concentrations of Mg2+ tested. As with Ca2+-induced Ca2+ release Ca”’ re(CICR), ionized Mg2+ modifies RDS-induced lease by interacting with the SR Ca2+ release protein. Effect of adenine nucleotides on RDS-induced Ca”+ release. Adenine nucleotides enhance the open time probability of skeletal and cardiac SR Ca2’ release channel incorporated in planar bilayers (6,22,33) and stimulate CICR (13, 34). In skeletal SR vesicles, they have been shown to inhibit Ca2+ efflux induced by RDSs by interfering with the oxidation of SR SH sites (22). In Figs. 6A and 6B, high concentrations of MgATP (5 mM) inhibited both the rate of Ca2+ release and the percentage Ca2+ release induced by 2,2’ DTDP. The data are similar for 4,4’ DTDP (data not shown). Several lines of evidence indicated that, as in skeletal muscle, adenine nucleotides inhibited the thiol-disulfide exchange reaction rather than Ca2+ efflux from oxidized and activated SR Cazt release channels. (i) As shown in Figs. 7A and 7B, the production of thiopyridone from its respective

A.

APIII I Abs. Increase

B.

’ CaCI, + 15 pM

4:4’-DTDP

100 FM

FIG. 5. Mg *+ dependence of RDS-induced Ca’+ release. Ionized Mg2+ concentration was increased in the initial reaction mixture hefore active Ca2+ loading of the vesicles and release induced by RDSs. Increasing concentration of free Mg2+ did not alter uptake hut resulted in a stepwise reduction in the rates of release induced by 100 pM of 2,2’ DTDP (A) or 4,4’DTDP (B). Ca” efflux.

PRABHU

AND

40 0.5 mM MgATP --t

5mMMgATP

1 30 -

20 -

lo-

0

I

1

I

100

200

300

[2,2’-DTDP] pM

SALAMA

pending on whether it was added after or before the RDS) (i.e., after or before the oxidation of SH sites on the SR). Both the stimulation and the inhibition of release occurred in the absence of functional SR Ca2+ pumps, since vesicles were passively loaded with Ca’+, without ATP and CAMP is not a substrate for the Ca’+, Mg2+-ATPase. Second, MgATP (1-5 mM) inhibited Cazf release by blocking oxidation by RDSs rather than through a stimulation of Ca2+ pumps since rates of Ca2+ uptake were already maximum with 100 pM MgATP and the ATP regenerating system. Effect of RDSs on Ca’+-dependent ATPase activity. Studies of RDSs on skeletal SR have indicated that RDSs induced Ca2+ release by acting at the same site as the CICR mechanism rather than the Ca*+, Mg”+-AT-

A

100 90 -

2,2’-DTDP 20 FM

0.5 mM MgATP --a--

60 -

5mMMgATP

70 60 50 40 30 -

.;

3

10 0 0

I 100

I

200

I

300

[2,2’-DTDP] pM Adenine nucleotides inhibit of 2,2’ DTDP-induced Ca”’ release. An increase in MgATP from 0.5 to 5 mM in the initial reaction medium reduced both the rate (A) and the percentage of Ca” release (B) induced by 20 to 250 FM 22’ DTDP.

RDS gradually decreased in the presence of increasing MgATP concentrations, indicating an inhibition of the oxidation rate of SH sites. (ii) In SR vesicles passively loaded with Ca2+ (with no ATP), RDSs elicited Ca2+ release but the subsequent addition of CAMP (1 mM) stimulated passive Ca2+ efflux at rates over and above the efflux rate produced by RDS alone (Fig. 8, compare traces 4 and 5). However, when CAMP (1 mM) was added before to the RDS, Ca2+ etllux rates were lower than those produced by RDS alone and were similar to release rates achieved by CAMP alone (Fig. 8, compare traces 2, 3, and 4). The modulation of sulfhydryl-activated Ca2+ release by adenine nucleotides was attributed to the direct action of nucleotides on the release mechanism rather than on the Ca”, Mg2+-ATPase pumps. First, RDS-induced Ca2+ release was stimulated or inhibited by CAMP (de-

4

;I . , , , . , , . , , . , , . , , . , 0.0 0.2

0.4

0.6

0.8

1.0

1.2

[MgATP] mM

4,4’-DTDP 100 pM

[MgATP] mM FIG. 7. Thiopyridone production as a function of [MgATP]. Increases of MgATP in the initial reaction mixture resulted in a reduction of both 2-TP(A) and 4-TP(B) production when either 2,2’ DTDP (20 pM) or 4,4’ DTDP (100 &M) was used to trigger Ca”’ release. The results indicated that MgATP inhibited the oxidation reaction between RDSs and SR vesicles.

SH OXIDATION 1 2 3 4 5

A.

+ APill 720

Abs.

TRIGGERS

Ca”+ RELEASE

-Control - CAMP alone - CAMP followed by 2,2’-DTDP - 2.P’-DTDP alone - 2.2’-DTDP followed by CAMP

- 790 nm

.0004

FROM

CANINE

CARDIAC

281

SR VESICLES

cles by 2,2’ DTDP compared to more potent uncoupling by A23187. As with skeletal SR (22), when RDSs (20-50 PM) were added 10 to 30 min before the addition of ATP to drive active Ca”+ uptake there was an initial fast phase of Ca2+ uptake followed by a phase of Ca”+ release. The initial rate of Ca”+ uptake was similar to that measured without pretreatment with RDSs which strongly implies that RDSs did not interfere with Ca”+ translocation by the Ca2+, Mg2+-ATPase (not shown). Taken together, the data indicate that RDSs act at the Ca2+ release channel, a site distinct from the Ca”+, Mg2+-ATPase. DISCUSSION

13.

= iiz 5

25 20

1 -Control 9 - CAMP alone, 1 mM 3 -CAMP (1 mhf)followed by 2,2’-DTDP,100 t&i 4 - 2J’-DTDPalone, 100FM 5 - 2,2’-DTDP(100

1

2

z ---d

3

4

5

FIG. 8. In passively loaded SR vesicles cAMP stimulates RDS-induced Ca” efflux. SR vesicles were passively loaded with Ca2+ and passive Ca”+ efflux was induced as described under Materials and Methods. (A) Passive efflux traces measured under different conditions: (1) passive Ca*+ efflux from control SR vesicle; (2) passive Ca2+ efflux stimulated by the presence of 1 mM CAMP in the Ca’+ release medium; (3) release induced by 2,2’ DTDP, with the prior addition of CAMP; (4) passive efflux induced by 2,2’ DTDP alone, in the absence of CAMP; (5) 22’ DTDP-induced efflux in the presence of CAMP, added after the RDS-SR interaction was completed. (B) Bar graph of Ca”’ efflux rates, calculated from the steepest slopes of the efflux traces shown in (A).

The main results are that in cardiac SR, reactive disulfides induce Ca2+ release from Ca”+-loaded SR vesicles by oxidizing free sulfhydryls on an SR Ca2+ release protein(s). The chemistry of RDS reagents dictates an absolute specificity for free SH sites which are oxidized via a thiol-disulfide exchange reaction. Such a reaction between RDSs and cardiac SR was confirmed (a) through absorption changes caused by thiopyridone production which were associated with Ca2+ release induced by RDSs and (b) because the time course of thiopyridone production preceded slightly that of Ca”+ release (Fig. 3). RDS-induced Cazf release was dependent on free Mg”+ (Fig. 5) as for CICR and RDS-induced Ca’+ efflux in skeletal SR vesicles. Dithiothreitol inhibited RDS-induced release when added before the RDS (Fig. 4) and partly reversed the effect of RDSs by the closure of Ca”’ release pathways resulting in Ca”+ reuptake by the Ca”+, Mg’+-ATPase. Adenine nucleotides inhibited the thioldisulfide exchange reaction, hence reduced thiopyridone production and consequently inhibited RDS-induced

% .= .1 z

U- I=

Pase pump (22, 24). Possible interactions of RDSs with cardiac Ca2+ pump were examined by measuring Ca2+dependent ATPase activity in the presence and absence of 2,2’DTDP. Incubation of SR vesicles with 2,2’DTDP (20 PM) stimulated Ca’+-dependent ATPase activity by 17% (Fig. 9, compare bars 1 and 2). The enhanced ATPase activity was due to a 2,2’ DTDP-dependent increase in Ca2+ permeability, which uncoupled the SR vesicles. The maximum levels of uncoupling and ATPase activity were typically achieved with A23187 (bar 4, 39% stimulation over control). With 22’ DTDP plus A23187 in the medium, ATPase activity reached its maximum levels measured with the Ca2+ ionophore, alone (bar 3, 39% stimulation). Thus, the partial stimulation of ATPase activity by 2,2’ DTDP, compared to A23187 alone was due to a partial uncoupling of the vesi-

l-Control 400 - 2 - 2,2’-DTDPalone, 200 pM . 3 - 2,2’-DTDP(200 a) and 350 A23187 2 Kg . 4 - A23107 alone, 2 pg

1

2

3

4

FIG. 9. Stimulation of Ca’+-dependent ATPase activity by 22’ DTDP. Ca” -dependent ATPase activity was measured as described under Materials and Methods: (1) control SK vesicles; SR preincubated with (2) 200 flM 2,2’ DTDP; (4) 2 pg A23187; or (3) both. Partial uncoupling and ATPase stimulation occurred with 2.2’ DTDP preincubation. Maximal uncoupling occurred during preincubation with either A23187 alone or with both A23187 and 2,2’DTDP.

282

PRABHU

AND

Ca2+ release (Figs. 6 and 7). However, adenine nucleotides stimulated RDS-induced Ca2+ release when added after the oxidation reaction was completed, thereby demonstrating the standard stimulation of SR Ca2+ release (Fig. 8). Rates of passive Ca2’ release from passively loaded SR were enhanced by RDSs. Moreover, RDSs stimulated Ca2+-dependent ATPase activity because the oxidation reaction increased the Ca2’ permeability and uncoupled the SR vesicles (Figs. 8 and 9). Taken together: (a) the rates and amounts of Ca2+ release, (b) the changes in RDS-induced Ca2+ release by reagents known to affect physiologic Ca2+ release, (c) the stimulation of ATPase activity, and (d) the stimulation of passive Ca2+ release all indicate that the site of action of RDSs is at one or more sulfhydryl(s) on or near the physiologic cardiac SR Ca2+ release channel and not the Ca2+, Mg2+-ATPase. Similar results have been reported for skeletal SR (22). Sulfhydryl oxidation causes SR Ca”+ release. Our previous studies have shown that the heavy metals Ag+ and Hg2+ induce Ca2+ release from cardiac SR vesicles by binding to free sulfhydryls on a Ca2+ release protein (25). Similar results have been observed with skeletal SR in isolated vesicles (17, 18) and in skinned muscle fibers (35). In addition to heavy metals, mercaptans (in the presence of CL?), phthalocyanine dyes, and antharaquinone have all been shown to induce Ca2’ release from skeletal SR most likely by interaction with a free sulfhydry1 (19, 20, 21). All these SH reagents appear to act via a common SH-dependent mechanism. Yet, none of them prove that a “critical” SH modulates Cazf release because these reagents are not known to exclusively interact with thiols, nor do they discriminate among the different populations of thiols. The present study with RDSs is of critical importance because RDSs differ from all the SH reagents previously studied in connection to SR function and offer major advantages: (a) RDSs are absolutely specific for free sulfhydryls (22, 23), (b) the oxidation of free SHs can be quantitatively and kinetically measured spectrophotometrically through reaction of thiopyridone production (31, 32), (c) RDSs preferentially react with low pK, sulfhydryls, even in the presence of other types of SH groups, and (d) reaction rates are rapid and require low concentrations of substrate (RDS) (23). Thus, RDSs offer a powerful tool to covalently label the SR proteins responsible for the Ca2+ release mechanism by selectively oxidizing a particular group of sulfhydryls. As demonstrated in skeletal SR, a RDS conjugated with biotin hydrazide (PDP-BH) was used to covalently link biotin to the SR Ca2+ release protein and to identify a 106-kDa protein (24). Upon incorporation into planar lipid bilayers, highly purified 106kDa proteins exhibited the characteristics of Ca2+ release channels, which were activated by Ca2+, ATP, and sulfhydryls (24, 36). Its conductance and subconductance states were similar to those reported for the 450kDa foot protein (4, 33, 37). As shown in Fig. 10, the

SALAMA ‘AP III Abs. Increase 720 - 790 nm Ca*+ Efflux

f lOpM CaCI,

PbP-BH

FIG. 10. SPDP-biotin conjugates trigger Ca2+ release from cardiac SR. PDP-BH was synthetized as previously described (24) and used to trigger Ca*’ release from actively loaded cardiac SR vesicles. PDPBH produced a concentration-dependent stimulating of the rates and extent of Ca*+ release.

RDS conjugated with biotin, and PDP-BH also induced Ca2+ release from actively loaded cardiac SR vesicles. Maximum release occurred at 150 PM, and levelled off thereafter. Ca2+ release induced by PDP-BH implies that biotin was covalently linked to the cardiac SR Ca2+ release channel protein. Biotin-avidin chromatography may be useful to isolate cardiac SR Cazf release channel proteins. Skeletal versus cardiac SR. In skeletal SR, RDSs cause Ca2+ release by oxidizing a specific class of SH groups on the SR Ca2+ release channel (22). This occurs via the nucleophilic attack of the disulfide in the RDS by the mercaptide anion on SR proteins resulting in a mixed disulfide. Release was most potent with compounds bearing the most electron withdrawal groups (i.e., pyridine rings adjacent to the disulfide bond as in 2,2” DTDP and 4,4’ DTDP) while SPDP with one pyridine ring resulted in slower, less potent release. Disulfides not immediately adjacent to the pyridyl ring (e.g., cystine and pyrithioxin) failed to induce Ca2+ release. Similar results were seen in cardiac SR where the RDS had the following order of potency: 4,4’ DTDP > 2,2’ DTDP > PDP-BH. As in skeletal muscle, when the disulfide bond is at the 4 position (4,4’ DTDP), the rates of Ca2+ efflux were maximal. The percentage Ca2+ efflux achieved was comparable with either 2,2’ DTDP or 4,4’ DTDP. Both 2,2’ DTDP and 4,4’ DTDP with two pyridyl rings caused a more rapid of Ca2+ release than SPDP-BH (with a single pyridyl group), although the percentage release was comparable. In rabbit skeletal SR vesicles, free Mg2+ from 0.1 to 1.0 mM stimulated but from 1 to 5 mM inhibited the rate and percentage of Ca2+ release (22). Similar characteristics were observed with canine cardiac SR vesicles (Fig. 5), except that the stimulation and inhibition by free Mg2+ was less potent in cardiac compared to skeletal SR vesicles.

SH OXIDATION

TRIGGERS

Ca2+ RELEASE

Sulfhydryl-activated Ca2+ release may be important to understand the mechanism(s) responsible for cardiac SR Ca2+ release and for the identification and purification of the protein(s) involved in this release mechanism. The rates of Ca2+ release reported here are too low to account for physiological rates of Ca2+ release. However, these rates of release were obtained from vesicles that were partially loaded with Ca” in order to maintain a reproducible Ca2+ gradient between the intra- and extravesicular space. In other experiments, vesicles were “fully” loaded with Ca2* by addition of 10 to 25 PM Ca2+ and waiting for complete uptake before addition of another aliquot of Ca2+. Under the latter protocol, cardiac vesicles accumulated the “maximum” amount of Ca2+ (25) and rates of release induced by RDSs were five to eight times faster, still too low to account for physiological release rates. Nevertheless, one may argue that the rates of Ca”+ release obtained through an exogenously added sulfhydryl reagent are significant because SH sites on Ca2+ release channel proteins may be primed and adjacent to other SH sites (on the same or different proteins) to undergo more rapid oxidation-reduction reactions. In such a scheme, oxidation-reduction reactions could rapidly open and close Ca2+ release channels and rates. The present gate Ca2+ release at physiological study does not provide evidence for nor does it address the possible physiological role of sulfhydryl-gated Ca2’ release. However, it strongly suggests that the sulfhydryl redox state may be of critical importance in pathophysiologic conditions such as ischemic heart muscle where abnormal Ca2+ handling, loss of glutathione, and altered intracellular redox potentials have been demonstrated (38-40).

FROM

CANINE

283

SR VESICLES

10. Smith, J. S., Rousseau, E., and Meissner, G. (1989) Circ. Rex 64, 352-359.

11. Chamberlain,

B. K., Volpe, P., and Fleischer, Chem. 259,7540-7546.

S. (1984) J. Hiol.

12. Martonosi,

A. N. (1984) Physiol. Reu. 64, 1240-1320. G. (1984) J. Biol. Chem. 259,23652374. 14. Agnew, W. S. (1988) Nature (London) 334,299~300. F., and Pozzan, T. (1985) Nature 15. Volpe, P., Salviati, G., DiVirgilio, 13. Meissner,

(London)316,347-349. 16. Vergara,

J., Tsien, R. Y., and Delay, M. (1985) Proc. N&l. Acad. Sci. IJSA 82,6352%6356.

17. Abramson,

J. d., Trimm, J. L., Weden, L., and Salama, G. (1983) Proc. Natl. Acad. Sci. IJSA 80, 1526-1530.

18. Salama, G., and Abramson,

J. (1984) J. Biol. Chem. 259,

13,363-

13,369. 19. Trimm, J. L., Salama, G., and Abramson, Chem. 261,16,092216,098.

J. .J. (1986) J. Biol.

20. Abramson,

J. J., Cronin, ,J. R., and Salama, G. (1988) Arch. Biothem. Biophys. 263,245-255.

21. Abramson,

J., Buck, E., Salama, G., Casida, ,J., and Pessah, I. (1988) J. Biol. Chem. 263, 18,750m18,758.

22. Zaidi, N. F., Lagenaur,

C. F., Abramson, J. T., Pessah, I., and Salama, G. (1989) J. Biol. Chem. 264,21,725-21,736. K. (1979) Int. J. Biochem. 10,259-274.

23. Brockelhurst, 24. Zaidi,

N. F., Lagenaur, C., Hilkert, R. J., Xiong, H., Abramson, J. J., and Salama, G. (1989) J. Riol. Chem. 264, 21,737-21,747.

25. Prabhu, S. D., and Salama, G. (1990) Arch. Riochem. and Biophys. 277(1),47-55. 26. Chamhelain,

B. K., Levitsky, Biol. Chem. 258,6602-6609.

D. O., and Fleischer,

S. (1983) J.

27. Lowry,

0. H., Rosebrough, N. J., Farr, A. L., and Randall, (1951) J. Biol. Chem. 193,265&275.

R. J.

28. Scarpa, A. (1978) in Methods

in Enzymology (Fleischer, S., and L., Eds.), Vol. 56, pp. 301-338, Academic Press, San

Packer, Diego.

29. Salama, G., and Scarpa, A. (1980) J. Biol. Chem. 255,6525-6528. 30. Scarpa, A., Brinley,

ACKNOWLEDGMENTS

F. J., Jr., and Dubyak,

G. (1978) Biochemistr>

17, 1378.

Thanks are due to Thomas Brown for his preparation of canine cardiac SR vesicles and to Alice Diven for typing the manuscript.

31. Grassetti,

D. R., Murray,

Pharmacol. 32. Alberty,

J. F., and Ryan, H. T. (1969) B&hem.

18,603%611.

R. A., Smith, R. M., andBeck,

R. M. (1951) J. Biol. Chem.

193,425-434.

REFERENCES

33. Smith,

J. S., Coronado, don) 315,446-449.

1. Fozzard, H. A. (1977) Annu. Keu. Physiol. 39,201-220. 2. Fabiato, A. (1983) Amer. J. Physiol. 245 (Cell Physiol. C14. 3. Fabiato, A. (1985) Amer. J. Physiol. 85, 247-289. 4. Meissner, G., and Henderson, 3065-3073. 5. Winegrad,

CARDIAC

14), Cl-

34. Meissner, 236-244. 35.

J. S. (1987) J. Viol. Chem. 262,

Pike, G. K., Abramson, 5

6. Rousseau, E., Smith, J. S., Henderson, (1986) Rio&vs. J. 50, 100991014.

G.

7. Inui, N. M., Wang, S., Saiti, A., and Fleischer, S. (1988) J. Biol. Chem. 263,10,843~10,850. 8. Anderson, K., Lai, F. A., Liu, Q., Rousseau, E., Erickson, H. P., and Meissner, G. (1989) J. Biol. Chen. 264,1329-1335. G. (1989) Amer. J. Physiol.

E., and Eveleth, J. (1986) Biochemistry

256,

25,

,J. ,J., and Salama, (:. (1987) Biophys. J.

R., Zaidi, N. F., Lagenaur, phvs. J. 57, 275a.

J. S., and Meissner,

G. (1985) Nature (Lon-

1,607a.

36. Hilkert,

S. (1982) Annu. Keu. Physiol. 44,451-462.

9. Rousseau, E., and Meissner, H328pH333.

G., Darling,

R., and Meissner,

C., and Salama, G. (1990) Bio-

37. Lai, A. F., Erickson,

H. P., Rousseau, E., Liu, Q.-Y ., and Meissner, G. (1988) Nature (London) 331,315-319.

38. Curello,

Guarrenri,

S., Ceconi, C., Bigoli, C., Ferrari, R., Alhertini, C. (1985) Experientia 41,42-43.

A., and

39. Tosaki, A., and Hearse, D. ,J. (1987) Amer. J. Physiol. 253, H22540.

H233. Hearse, D. J., and Tosaki, A. (1988) J. Mol. Cell. Cardiol. 20, 21% 223.