Purified Reconstituted Inositol 1,4,5-Trisphosphate Receptors

THEJOURNAL OF BIOLWNCAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 46, Issue of November 18, pp, 28972-28978, 1994 Printed in U.S.A.

Purified Reconstituted Inositol 1,4,5-Trisphosphate Receptors THIOL REAGENTS ACT DIRECTLY ON RECEPTOR PROTEIN* (Received for publication, March 18, 1994, and in revised form, August 15, 1994)

Adam I. Kaplin, Christopher D. Ferris, Susan M. Voglmaier, and Solomon H.Snyder$ From the Departments of Neuroscience, Pharmacology a n d Molecular Sciences, Psychiatry a n d Behavioral Sciences, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205

Thimerosal, a sulfhydryl oxidizing reagent, has been shown to induce Ca2+ mobilizationin several cell types and to increase the sensitivity of intracellularCa2+ stores to inositol 1,4,54risphosphate (IP,). Using purified IP, receptor (IP,R) protein reconstitutedin vesicles, we demonstrate pronounced stimulationby thimerosal of its Ca2+channel activity. Effects of thimerosal aredependent onthe redox state of the receptor, implying an action of thimerosal on a critical sulfhydryl group(s)of IP,R. Thimerosal enhances the affinity of IP,R for IP, binding. Themanner in whichthimerosalmodulates IP,R responsiveness to IP, provides evidencefor receptor heterogeneity with implicationsfor mechanisms of quantal Ca2+ release. These results clarify regulationof IP,R activity by redox modulation.

We successfullyreconstituted IP, inducedCa2+fluxin vesicles containing only purified IP,R protein (17). IP,-mediated Ca2+flux in the reconstituted vesicles is regulated by phosphorylation (18, 19) and adenine nucleotides (20). Using these reconstituted vesicles, we have characterized effects of thiol reagents on Ca2+flux at t h e level of the IP,R protein. Intracellular release of CaZ+is a discontinuous, quantalprocess in which successive increments of IP, transiently release precise amounts of Ca2+(21, 22). Utilizing IP,R reconstituted that intoproteolipidvesicles (IP,RV) wepreviouslyshowed quantal fluxof Ca2+ elicited by IP, is a fundamental property of the IP,R, suggesting that the receptors purified from rat cerebellumconstitute a heterogeneouspopulationwithvarying sensitivity to IP, (23). The effects of thiol reagents on IP,mediated flux in IP,RV observed here provide additional evidence for functional receptor heterogeneity, which may help account for quantalCa2+release.

The dynamicsof intracellular Ca2+ provide crucial signaling EXPERIMENTALPROCEDURES information in many aspectsof cellular regulation. Intracellular Ca2+flux is regulated by numerous processes, especially the Materi~ls-[~H]IP,, 45Ca2+ , and formula 963 scintillation mixture release of Ca2+ fromintracellularstoresbyinositol 1,4,5- were obtained from DuPont NEN. ~-myo-Ins(l,4,5)P,,hexapotassium trisphosphate (IP,)' (1,2) and a calcium-induced calcium re- salt was obtained from LC Laboratories (Woburn, MA). Concanavalin lease system involving channels labeled by the alkaloid ryan- A-Sepharose and G-25, superfine, were obtained from Pharmacia LKB Biotechnology Inc. Phospholipids forreconstitution were obtained from odine (3,4).Physiologic and pathologic changesin the oxidative Avanti Polar Lipids (Birmingham, AL). All other reagents were from state of cells influence Ca2+ disposition. Low levels of oxidants Sigma. can stimulatecell proliferation and differentiation ( 5 , 61, while Purification and Reconstitution of IP$-IP,R was purifiedfrom increased oxidativestress may exertcytotoxic effects, including adult male Sprague-Dawleyrat cerebellum and reconstituted into lipid death by apoptosis (7). Perturbationsin the oxidative state of vesicles as described (17). Briefly, IP,R was purified using a two-step affinity chromatography procedure employing sequential heparin-agasulfhydrylgroupscaninfluenceCa2+ flwr (7). Thimerosal rose and concanavalinA-Sepharose columns. Followingpurification to (TMS), a sulfhydryloxidizingagent, has beenreportedto apparent homogeneity, detergent-solubilized receptor protein was stimulate (8-12) or inhibit (11,121 intracellularCa2+flux. TMS mixed with sonicated lipids and the mixture wasdialyzed at 4 "C increases the potencyof IP, in releasing Ca2+(9, 11, 13-16). In against buffer A (50 mM NaC1, 50 mM KC1,20 mM Tris-HC1, pH 7.41, supplemented with 2.5 mM B-mercaptoethanol(BME)and 2mM EDTA, some studies, TMS increased the affinity of IP, for receptor binding sites (13, 16), while otherstudies showed no effect (11, to effect detergent removal and vesicle formation. Thebuffer was changed every8 hfor 48h, and EDTA was omitted from the final buffer 12). These findings suggestthat IP,-induced Ca2+ fluxi s influchange. For experiments performed in the absence of reducing agent, 1 enced by the oxidativestate of sulfhydryl groups, perhaps those ml of IP,RV was passed over a 5-ml G-25desalting column equilibrated of the IP, receptor (IP,R) itself. However, TMS can influence with buffer A to remove BME. 45Ca2+ Flux-Reconstituted proteoliposomes were assayed for1P,related Ca2+regulating systems, such as the endoplasmic reflux as described (17). Following preincubation under ticular Ca2+pump (11,13), indirectly affectingIP, induced Ca2+ stimulated 45Ca2+ various conditions vesicles wereincubated (for either 10 or 15 s) in the release. with or without IP,. Under these conditions, presence of 2 pCi of 45Ca2+ tracer 45Ca2+ gained access to the lumen of vesicleswhen the IP,R * This work was supported by United States Public Health Service channels were opened byIP,. The flux reaction was stopped by the addition of excess buffercontaining unlabeled divalent cations and hepGrant MH-18501,Research Scientist AwardDA00074 (to S. H. S.), content was isolated by immediTraining Grant GM-07309 (to A. I. K. and S. M.V.), Predoctoral Fel- arin (200 pg/ml). Intravesicular 45Ca2+ lowship MH-10018(to C. D. F.), and grants from the International Life ately passing the vesiclebuffer mixture over a cation-exchange column Sciences Institute and the W. M. Keck Foundation. The costs of publi- (Dowex 50W,Sigma). The vesicles were collectedand their intravesicucation of this article were defrayed in part by the payment of page lar 45Ca2+ content was measured by scintillation spectrometry. charges. This article must therefore be hereby marked "aduertisement" PHIIP, Binding-Ligand binding was assayed by precipitation of in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. IP,RV with polyethylene glycol using y-globulin as carrier protein, as $To whom all correspondence should be addressed. Tel.:410-955described (20). 3024; Fax: 410-955-3623. The abbreviations used are: IP,, inositol 1,4,54risphosphate; TMS, RESULTS thimerosal; IP,R, IP, receptor; IP,RV, IP, reconstitutes vesicles; BME, TMS stimulation of Ca2+ fluxin Several workers have shown B-mercaptoethanol; DTT, dithiothreitol; IAA, iodoacetamide; NEM, intact cells and platelets (14, 15, 24-29) and enhancement of N-ethylmaleimide.

28972

nditions

Effects of Thimerosal on Purified IP, Receptor Activity TABLEI Effects of thiol modifjing reagents on the Ca2+flux in IP$ vesicles Reconstituted vesicles were preincubated for 2 min a t room temperature under the conditions shown. Control vesicles were preincubated with buffer A. IAA and NEM refer to iodoacetamide andN-ethylmaleimide, respectively. All reagents were prepared freshas stock solutions in buffer A. Following this preincubation, Ca2+ flux was allowed to proceed for 15 s in the presence of tracer 45Ca2+with or without 40nM IP,. Specific Ca2+flux under each condition was calculated as cpm in the cpm in theabsence of IP, (blank). None presence of IP, (total) minus the of the reagents shown altered the blank. Data is expressed as specific Ca2+flux under each condition as a percentage of specific Ca" flux under control conditions. Measurements in duplicate or triplicate varied less than 10%. This experiment has been repeated three times with the same results.

28973

180000

I

I

1

2.5 mM TMS 2.5 mM BME

I

300004 IA

,

0 , 0

1

2

3

4

Preincubation %

3 TMS 1 ~ TMS l l ~ 1 mM BME 1 mM D m 1mM GSH lmMTMS+lmBME 1 mM TMS + 1 mM DTT lmMTMS+lmGSH 1m" lmMIAA+ImMBME 1 mM NEM 1 mM NEM + 1 mM BME

PREINCUBATION TIME (min)

131 0 100 71 43 823 489 506

99

2.5 mM BME 4 min Preincubation

65

94 30000

IPS-induced release of Ca2+in permeabilized cells and microsomes (11-13, 15, 16). Pretreatment of IP3RV with low micro0 , , molar concentrations of TMS produces a 30% stimulation in 0 0.5 1 1.5 2 2.5 IP,-mediated Ca2+flux (Table I).Relatively high concentrations (1 mM) of TMS abolish flux activity. However, when combined [TMSI (mM) with BME, which by itself has no effect, TMS pretreatment FIG. 1. Time and concentration dependence of T M S stimulaTMS and results in an8-fold enhancement of Ca2+flux induced by 40 m tion. Reconstituted vesicles were preincubated for 2 min with IP,. This effect is rapid withmore than 50% maximal Ca2+ flux BME followed by 45Ca2+ withand without 25 nM IP, for 10 s. In A , were preincubated with 2.5 mM TMS and 2.5 mM BME for evident at 10 s pretreatment with TMS, reaching a plateau a t vesicles variable lengthsof time prior to the initiation of Ca" flux. In B , vesicles 1-2 min (Fig. lA).The enhancement of Ca2+flux elicited by were incubated with variable concentrationsof TMS and 2.5mM BME TMS in the presence of excess BME (2.5m) is concentration for 4 min prior to Ca2+flux measurements. Specific Ca2+flux was caldependent, with half-maximal augmentation evident at about culated as described in Table I. Measurements in duplicate varied less 10%.This experiment has been repeated two times with the same 0.2 mM TMS and a plateau between 1 and 2.5 mM TMS than results. (Fig. 1B). Like BME, two other reducing agents, dithiothreitol (Dm) TABLEI1 Effect of variations in the orderof addition of TMS a n d BME on and reduced glutathione (GSH), also markedly augment the Ca2+flux TMS-mediated increase of Ca2+flux, whereas D'IT and GSH by Reconstituted vesicles were processed as described inTable I, except themselves reduce IP,-induced Ca2+flux (Table I). DTT and that two successive 2-min incubations were performed and 100nM IP, GSH are lessefficient than BME in facilitating the stimulationwas used to assess CaZ+ flux. Control vesicles were incubated in buffer of Ca2+flux mediated by TMS. The greater augmentation of the A during both preincubations. As shown in Fig. 2, relative to control of Ca2+flux by TMS + BME seen with 100nM IP, effects of TMS by BME might reflect its smaller size, permit- values the stimulation ting access to the relevant sites on the IP,R. Thus, there is a is less than that achieved with 40 nM IP, (compare Tables I and 11). in duplicate or triplicate varied less than 10%. This nonspecific requirement for a reducing agent to achieve maxi- Measurements experiment has been repeated three times with the same results. mal stimulation by the oxidant TMS. Preincubation No. 1 Preincubation No. 2 Control The TMS effect appears specific in that otherthiol oxidizing reagents, iodoacetamide (IAA) and N-ethylmaleimide (NEM), % 1 mM TMS + 1 mM BME fail to augment Ca2+flux either by themselves or in the presBuffer A 541 1mM TMS Buffer A 0 ence of BME (Table I). TMS, BME, and all of the other thiol 1 mM BME Buffer A 103 modifying reagents tested, either alone or in combination, have 1mM TMS 1mM BME 37 no effect on Ca2+flux in the absence of IP, (data not shown), 1 mM BME 1 lll~TMS 500 demonstrating that these reagentsdo not exert nonspecific effects on Ca2+ fluxin reconstituted vesicles. To clarify interactions between BME and TMS, we evaluated agents might maintaincrucial sulfhydryl groups in a reduced of addition (Table form that can be modified by TMS. This is consistent with effects on Ca2+flux of variations in their order 11).The pronounced facilitation of Ca2+ fluxoccurs only if BME established actions of TMS involving alkylation only of reduced is applied before TMS or simultaneously with it. When BME is sulfhydryl groups (30). applied after TMS, we observe only inhibition of IP3-induced To investigate themechanisms wherebyTMS stimulates IP,Ca2' flux. This suggests that reducing agents, such as BME, induced Ca2+ flux, we evaluated concentration-response relaDTT, and GSH, cause alterations in IP,R that facilitate the tionships for IP, (Fig. 2). In the absence of TMS pretreatment TMS mediated augmentation of Ca2+ flux. Thesereducing (i.e.. 2.5 m BME alone), IP, enhances Ca2+flux in a concen-

IB

Thimerosal of Effects

28974 150000

on Purified IP, Receptor Activity

1

0

0.5

1

2

2.5

[3H]IP 3 BOUND (nM) FIG.3. TMS increases the binding affinity of the purifiedIP& Reconstituted IP,R (0.4pg) was preincubated in thepresence of 2.5 mM TMS and 2.5 mM BME (0) or 2.5 mM BME alone (0)for 2 min followed by incubation with 4 1 1 [,H]IP, ~ and various concentrations of unlabeled FIG.2. Effect of TMS on theIP,concentration-response curve. IP,. Total volume was 250 pl. Radioligand binding was determined by Reconstituted vesicles werepreincubated with 2.5 mM TMS and 2.5 mM polyethylene glycol precipitation. Nonspecific binding was measured in or 2.5 mM BME alone ( 0 )for 2 min followed by 45Ca2+with BME (0) the presence of 1 p~ IPa. Measurements in duplicate vaned less than variable amounts of IP, for 10 s. Specific Ca" flux was calculated as 15%. This experiment has been repeated two times with the same described in Table I. Measurements in duplicate varied less than 10% results. This experiment has been repeated three times with the same results.

TABLE I11 IP3 preincubation augments the stimulation by TMS at tration dependent fashion with half-maximal stimulation Preincubation conditions and Ca2+flux were performed as described 150 nM and maximal effects at 1 PM IP,, similar to our earlier Table 11, except that 40 nM IP, was used during the flux assay. The observations (17). TMS (2.5 mM) in the presence of BME (2.5 in successive preincubation protocol employed in these experiments inmM) produces a leftward shift of the concentration-response volved incubating IP,RV at room temperature for a total of 4 min and relationship with a 7-fold increase in the potency of IP,. Pre- vigorously pipetting the vesicles to achievecomplete mixing at the treatment with TMS and BME also produces a modest (12%) beginning of each preincubation step. This resulted in a slight decrease both the maximal Ca2+flux elicited by IP, and the maximal stimuincrease in the maximal amount of Ca2+ flux elicitedat satu- in lation produced by TMS + BME, when comparedto the single preincurating concentrations of IP,. The average calculated Hillcoef- bation conditionsemployed in Table I. However, all samples repreficient in control samples performed three times (1.59, 1.66, sented in this table underwent identical treatments. Measurements in and 1.77) was 1.7, similar to values for TMS-treated samples triplicate varied less than 10%.This experiment has been repeated two times with the same results. (1.61, 1.71, and 1.85). Preincubation No. 1 Preincubation No. 2 Flux assay Control To explore factors accounting for the increased responsiveness to IP, following TMS treatment, we monitored [,H]IP, % bindingto IP,RV (Fig. 3). In bothcontrol and TMS/BME40 na; IP,, 15 s 98 40 nM IP, Buffer A treated samples, a single component of IP, binding is evident. 40 nM IP,, 15 s 131 Buffer A 3 PMTMS 40 1 1 IP, ~ 3 PMTMS 40 nM IP,, 15 s 164 TMSBME treatment reduces the dissociation constant (K,) 2-fold from 13.6 t o 6.4n M with no effect on B,,,. 40 nM IP,, 15 s 104 Buffer A 40 m IP, We wondered whether TMS acts directly at the recognition IP,, 15 s 772 Buffer A 1 mM TMS + 1 mM BME 40 I ~ M site for IP, binding on the receptor. If so, the two substances 40 nM IP, 1 mM TMS + 1 mM BME 40 nM IP,, 15 s 1439 might compete with each other so that pre-exposure of the IP,R to IP, would sterically inhibit the effects of TMS. Accordingly, we compared effects of TMS on Ca2+ fluxin samples receiving vesicles incubated in the absence of TMS (Fig. 4A).The maxiIP, (40 nM) either before or after TMS exposure (Table 111). mal stimulation, seen with3 m TMS, is greater than 30%.At Because of t h e quantal nature of the response of IP,RV to IP, concentrations of TMS between3 and 10 v,there is a decline (231, preincubation with IP, does not augment the measured in the stimulation of Ca2+ flux seen with 80 nM IP,. At even greater TMS concentrations, flux activity decreases below conCa2+ flux relative to controls. Similar to earlier experiments, (1m each) prior to trol levels, with 100 m TMS being sufficient to abolish flux exposure to3 PM TMS alone or TMSBME IP, produces respective 30 a n d 670% increases in Ca2+ flux activity. Surprisingly, when vesicles are preincubated exactly as decompared to samples treated withIP, alone (Table 111).When these samples are exposed IP, to prior to and during treatmentscribed above with varying concentrations of TMS, and subsea maximal concentrationof IP, (1 PM), in Ca2+ quently stimulated with with TMS(3 J ~ M )or TMSlBME(1 mM each), the increase but rather only inhibits Ca" flux relaflux is twice as great as in samples in which IP, stimulation TMS does not stimulate 4A).When TMS concentrations are plotted follows TMS treatment (Table 111). The augmentation of the tive to controls (Fig. TMS responses elicited by pre-exposure to IP, indicates that on a logarithmic scale and Ca2+ fluxis normalized relative to the following can be seen: TMS and IP, act at different sites which presumably interact control values(as shown in Fig. a), allosterically. The greater effectof pretreatment with IP, sug- 1)the effects of TMS preincubation on Ca2+ flux elicited bya gests that IP, binding elicits a conformational alteration of submaximal concentration of IP, are biphasic, with TMSconstimulating flux and between 3 and 100 centrations up to3 IP,R in whichcriticalsulfhydrylgroupsaremademore 1.1~inhibiting flux;2) the effects of TMS on Ca2+ flux elicited by accessible to TMS. As described above, whereas millimolar concentrations of a maximal concentration of IP, are monophasic and inhibitory. Having explored the effects of millimolar concentrations of TMS in the absence of BME inhibit IP,-induced Ca" flux, low micromolar concentrations have a stimulatory effect (Table I). TMS with BME and micromolar concentrations of TMS alone, Preincubation with 0.1-3 p~ TMS leads to a progressively we examined the effect of variable amounts of BME added to several fixed concentrationsof TMS. The relationship between greaterCa2+fluxelicitedby80 nM IP,, relativetocontrol ~

~~

~~~~~~

*

Effects of Thimerosal on Purified IP, Receptor Activity 1500

28975

k

8woO

I

FIG.5. Biphasic effects of BME on TMS stimulation. Reconstituted vesicles were preincubated with0 V M ( 0 ) , 3p d (01, 100 p~ (m), or 1000 VM (0)TMS and various concentrationsof BME for 2 min followed by 45Ca2+with or without 40 nM IP, for 15 s. Specific CaZ+flux was calculated as described in Table I. Measurements in duplicate or triplicate vaned less than 10%.This experiment has been repeated three times with the same results.

preincubation with IP, preserves, in a concentration-dependent fashion, between one-halfto one-third of the flux activity of the samples following TMS and BME treatment, when compared to samples for which the TMS treatment isomitted (Fig.6B).This suggests that IP, elicits conformational changes in the IP,R which render it resistant to subsequent degradationby TMS. Dramatically, however, addition of greater amountsof IP, after treatment with TMS and BME produces no further augmentation of Ca2+flux, regardless of the amount of IP, employed FIG.4. Effects of low [TMS] on IP,-mediated CaZ+ flux in the during preincubation (Fig. 6 A ) . This suggests that this treatabsence of a reducing agent. Reconstituted vesicles were preincuIP, concentrationment regiment dramaticallyaltersthe bated with various concentrations of TMS for 2 min followed by &Ca2+ response relationship of the IP,R which retained activity. with and without either 80 nM (0) or 1000 nM (0)IP, for 15 s. In A , Cells contain substantial basallevels of IP,, often about 1VM specific Ca2+flux was calculated as described i n Table I. B represents the same data shown normalized to the control (ie. [TMS] = 0) and (311, which should suffice to open a substantial proportion of presented on a semilogrithmic graph. Measurements in duplicate or IP,R channels. Physiologic and pathologic stimulation of cells triplicate varied less than 10%. This experiment has been repeated alters theoxidative-reductive status of the intracellularmilieu three times with the same results. with associated influences on sulfhydryl groups of proteins. Such sulfhydryl alterations of IP,R may evoke Ca2+ releasein BME, TMS, and Ca2+f l u is complex (Fig. 5). Maximal stimu- the absence of receptor-mediated generation of intracellular lation of Ca2+flux occurs with a 2-fold excess of BME over TMS IP,. To approximate this situation, we initially exposed IP,RV in the case of 3 and 100 PM TMS. At higher BME concentra- to IP, and 45Ca2+and then added TMS (2.5 mM) in the contintions, the stimulationproduced by TMS decreases gradually to uous presence of BME (2.5 mM) (Fig. 7). In theabsence of TMS, control levels, presumably because excess BME covalently re- IP, stimulates Ca2+flux with effects first detectable at 10 nM IP,. TMS markedly enhances these effects and permits the acts with TMS reducing its availability to modify the IP,R. demonstration of significant Ca" flux stimulation with only 1 Experiments using DTT and TMS yield essentially the same results (data notshown) except that: 1)DTT alone inhibits flux nM IP,. At 10 and 20 nM IP,, TMS increases Ca2+flux about at higher concentrations (see Table I), and 2) maximal stimu- 6-fold compared to valuesobtainedwithout TMS addition. DTT and TMS. Thus, to These findings suggest that intracellular Ca2+ release can be lation is obtainedwithequimolar achieve maximal stimulation of IP,R activity thereisan regulated by changes in oxidative states of IP,R sulfhydryl optimalreducing agent concentration range for each TMS groups in theabsence of receptor-mediated signal transduction. concentration. DISCUSSION If IP, pretreatment alters the conformation of IP,R to effect Earlier studies reported that TMS, a sulfhydryl oxidizing TMS stimulation of Ca2+flux, as suggested in Table 111, perhaps IP, exposure would also influencethe inhibitory actionsof TMS reagent,enhances IP,-induced Ca2+ release (9, 11, 13-16). on Ca2+flux that are observed when high concentrations of These effects could involve numerous mechanisms, such as inTMS are applied to IP,RV in theabsence of BME. Accordingly, hibiting the Ca2' pump of the endoplasmic reticulum, inactiwe evaluated Ca2+ flux in IP,RV exposed to 0,40,200, or 500 nM vating a negative modulatorof IP,R activity, or augmenting the IP, followed bytreatment withTMS, and thenby application of effects of a protein that acts on the IP,R to promote its release BME (Fig. 6 A ) . In samples preincubated without IP,, no flux of Ca". By utilizing IP,R purified to apparent homogeneity and activity is detected after treatmentwith TMS followed by BME, reconstituted into proteolipid vesicles, this study establishes despite subsequent stimulation with up to 500 nM IP,. In the that the stimulatory effects of TMS take place at the level of the samples pretreated with200 nM IP,, Ca2+flux is 2.5 times that IP,R protein itself. of samples pre-exposed to 40 nM IP,. At 500 nM IP, pre-expoWhereas pretreatment of IP,RV with micromolar concentrasure, Ca2+flux is only modestly greater than at 200 nM. Thus, tions of TMS provides a modest stimulation of IP,-mediated

Effects of Thimerosal on Purified IPS

28976

A

1

I

B .+TMS n

6

8

100000-

v

Receptor Activity

FIG.7. TMS stimulation following the initiation of Ca2' flux. Ca2+flux was initiatedby incubating reconstituted vesicles with 4sCa2+, 2.5 mM BME, and 1,10,or 20 nM IP,. 20 s after the initiation ofCa2+ flux, 2.5 mM TMS, or buffer A was added and flux was allowed to proceed for another 130 s. Specific Ca2' flux was calculated as described in Table I. 10%. This experiment has Measurements in triplicate varied less than been repeated two times with the same results.

75000L4

+ N

1_1

substantial concentrations would likely have been present during the Ca2+ release experiments. The requirement for a reducing agent to achieve maximal 2 El stimulation by TMS is unexpected because IP,R used in these experiments were purified, reconstituted, and stored until use a v) inthe continualpresence of BME. Nevertheless, the pro0 nounced facilitation of IP,-induced Ca2+flux by TMS occurs only if BME is applied before TMS or simultaneously with it. IP,RV were desalted over a gel filtration column to remove BME for use in experiments to assess the effects of TMS on FIG.6. Effect ofsuccessive treatment withIps, TMS then BME IP,R activity in the absence of a reducing agent (see "Experion CaZ+ flux. Reconstituted vesicles were preincubated with 0 ( 0 ) 40 . mental Procedures"). One reason for the requirement to add (O),200 (W), or 500 (0) nM IP, for 20 s followed by a 2-min preincubation with 2 msr TMS, then a 2-min preincubation with 2 mM BME. After back BME to achieve maximal stimulation by TMS might be these three incubations,a s depicted inA, Ca" flux was initiated by the that IP,R is capable of autooxidation, such as through the foraddition of "Ca" with various concentrationsof IP,. In B , the Ca2' flux mation of disulfide bonds, which prevents TMS from gaining seen following these successive preincubations is shown in comparison access to the relevantsulfhydryl groups. Anotherpossibility is to the flux observed with vesicles preincubated with 0 nM IP,, 0 mM that the redox reaction between TMS and sulfhydryl group(s) TMS, then 2 mM BME followed by "Ca2+ with 40, 200, or 500 nM IP,. Specific Ca2' flux was calculateda s described in Table I. Measurements on the IP,R is complex, possibly involving intermediate steps, in duplicate varied less than10%. This experiment has been repeated and requires thepresence of BME a t specific stages toachieve four times with the same results. the proper products which confer enhanced receptor activity. The fact that both the concentration and orderof addition of Ca2+flux, and millimolar concentrations of TMS abolish flux TMS and BME to IP,RV can dramatically effect IP,R activity activity, pretreatment with TMS in the presence of BME en- suggests that the oxidative state of critical sulfhydryl resihances IP,-induced Ca2+flux up to 8-fold. Although the stimu- due(s) on the IP,R proteincan have a profound impact on and responsiveness to redox modulation. latory effects of TMS are specific, there appears to be a non- receptoractivity specific requirement for the presence of a reducing agent to Which sulfhydryl groupson IP,R might be uniquely responsive maximize the stimulationby TMS. TMS itself does not gate the to oxidative-reductive changes?IP,R is a large protein comprisIP,R, rather, TMS pretreatment enhances the response of IP,R ing 4 protomeric subunits (18, 32). Mutagenesis (33, 34) and to IP,. Pretreatment of IP,RV with TMS in thepresence of BME chemical modification (35) studies havelocalized the IP, recognition site to the first 400 amino acidsat theNH, terminus. The produces a leftward shift of the concentration-response relationship with a 7-fold increase in the potency of IP,. TMSBME Ca2+ion channel resides in thecarboxyl-terminal portion of the treatment of IP,RV reduces the dissociation constant (IC,) for protein, where there is high sequence homology to the Ca2+ ['HIIP, binding 2-fold without affecting the Bmm.Recently, two channel in theryanodine receptor (36, 37). The large cytoplasgroups (13, 16) also observed a TMS elicited increase in IP, mic portion of the receptor betweenthe NH,-terminal IP, bindbinding affinity to permeabilized cells and microsomes with no ing domain and the COOH-terminal transmembrane domain has been designated the coupling domain (2). Sinceour studies change in B,,,. The enhancement of IPS-induced Ca2+flux by TMS in the indicate that TMS does not act at the IP, recognition site (Table presence of BME in our experiments ismore pronounced than III), alternative possibilities are the carboxyl-terminal Ca2+ similar effects observed by others (9,11,13-16). However, none channel domain or the largecoupling domain in the interiorof of the earlier studies examined a broadrange of concentrations the molecule. Antibodies directed against the carboxyl-terminal cytoplasof reducing agents and TMS. In previous studies, IP,R activity was assessed either in intact cells, which retain theirreducing mic tail of the IP,R, like TMS treatment, both augment or cytoplasmic environments, or in microsomes derived from tis- inhibit Ca2+release (38, 39).This area of the type 1 IP,R consues homogenized in the presence of reducing agents, so that tains two cysteines at positions 2610 and 2613 which are rd

v M

5oo00-

Effects of Thimerosal on Purified IPS present in a TXCFICG sequence motif highly conserved in all subtypes ofIP,R and ryanodine receptors (1). Interestingly, TMS stimulates CaZ+flux through the ryanodine receptor in much the same manner that it stimulates flux through IP,R (27). Accordingly, these cysteines are strong candidates to be the targets for the regulation of IP,R by TMS and possibly by endogenous oxidative-reductive processes. What might be the endogenous regulator of the TMS-sensitive sulfhydryl groupson the IP,R? Besides TMS, oxidized glutathione (GSSG) can stimulateCa2+ fluxin permeabilized cells (13, 40). Presumably, the GSSG reacts with freesulfhydryl groups of IP,R to form a mixed disulfide bond (Protein-SSG). Glutathione is themost abundant reducing agent in cells, and the relative proportion of oxidized and reduced glutathione changes markedly in response to alterations in cellular physiology and pathology (41).Our findings with TMS (Fig. 7) suggest that even in the presence of low ambient levels of intracellular IP, and ina reducing environment, cellularalterations that increaselevels of GSSG could alter theconformation of the IP,R and thereby stimulateCa2+ release. One puzzling observation was the enhancement by low concentrations of TMS (in theabsence of reducing agents) of Ca2+ flux at 80 nM IP, and inhibition at 1p~ IP, (Fig. 4). Conceivably the higher IP, concentration exposes IP,R to an inhibitory action of TMS masking the stimulatory effect of TMS seen at lower (IP,). This seems unlikely, as IP, occupancy of IP,R appears to enhance rather than inhibit stimulation by TMS of receptor activity (Table 111 and Fig. 6). Another explanation is that TMS pretreatment might cause IP,R to lock in a subconductance state following the addition of IP,. The IP,R conductance in this state could be greater than the conductance normally elicited by 80 nM IP, but less than the conductance elicited by a saturating concentration of IP,. This would be similar to the effects of low concentrations of ryanodine on the ryanodine receptor (42).This seems unlikely because thespecific Ca2+flux elicited by 80 nM IP, is considerably less than the flux obtained by 1 p~ IP, following TMS treatment, suggesting that under these conditions IP,R are not in the same conductance state. A third explanation invokes heterogeneity of IP,R with only a fraction of receptors sensitive to 80 nM IP,. By increasing IP, binding affinity(Fig. 31, low micromolar concentrations of TMS lower the threshold for receptor responsiveness. Under these conditions TMS would stimulate Ca2+flux by recruiting receptors that normally respond only to higher IP, concentrations. With saturating concentrations of IP,, such that all receptors present are occupied by their ligand, a shift in the binding affinity of the IP, receptors present would be inconsequential and no stimulation would be observed. Perhaps the inhibitory effects of TMS are the result of modification of a different sulfhydryl group on the IP,R than the site which accounts for IP,R stimulation. Successive treatment of IP,RV with IP,, TMS, and BME provides another example of the way in which redox modulation of the IP,R can dramatically alter itsresponsiveness to IP, (Fig. 6). Treatment with 2 mM TMS followed by 2 mM BME abolishes IP,R flux activity. Incubation of IP,RV with increasing concentrations of IP, prior to treatmentwith TMS and BME preserves flux activity. Dramatically, no further increase is seen in flux activity following addition of increasing amountsof IP, during t h e f l wassay, regardless of the initial IP, concentration used during preincubation. It appears that thereceptors which remain active following the treatment protocol achieve a maximal response over a very narrow range of IP, concentrations. Unfortunately, we were not able to characterize this range directly because in samples pretreated with IP,, the levels of the

Receptor Activity

28977

ligand could only be partially diluted following incubations with TMS and BME. It was not possible, therefore, to assess Ca2+flux activity a t extremely low levels of IP,. An alternative explanation for these results is that successive treatment of IP,RV with IP,, TMS, and BME lockthe IP,R in subconductance states with Ca2+conductances proportional to the amount of IP, initially present during preincubation. This seems unlikely because it would necessitate a very large number of possible subconductance states, whereas no more than four have been identified (431, or an unusual coincidence that we happened to have selected concentrations of IP, that led to the induction of distinct states. Moreover, these states would have tobe at least partiallyirreversible, in that they are no longer responsiveto changes in theconcentration of IP,. The design of the assaywe employ to measureCa2+flux necessitates that the channelactivation we measure is reversible. Flux assays are terminated by the addition of heparin, a competitive IP, antagonist, and divalent cations to closeIP,R channels which have become activated (see "Experimental Procedures"). If the channelswere to remain open, all of the 45Ca2+ which had entered the vesicles during the flux assay wouldbe free to diffuse out and be lost during passage over the Dowex column we employ to chelate extravesicular Ca2+. It thus appears that successive treatment with IP,, TMS, and BME results in a population of functional receptors that are maximally activated by very low levels of IP,. How is it that increasing concentrations of IP, during preincubation result in the preservationof increasing maximalflux activity? Again, we believe that functional receptor heterogeneitycan best account for this result. Low concentrations of IP, would enhance Ca2+ flux selectively from more sensitive populations of receptors while progressively increased IP, concentrations would influence less sensitive receptors. With IP,RV, where each vesicle contains on average one or fewer receptors (23), we propose that 40 nM IP, protects amore sensitive populationof IP,R from the degradative effects of TMS while 200 and 500 nM protect less sensitivereceptor populations. Treatment with TMS in the presence of 40 n~ IP, would inactivatethe lesssensitive receptors enabling us to selectively examine IP,R of greater sensitivity. These findings provide insight into quantal aspects of IP,inducedCa2+release. In intact and permeabilized cells, IP, releases Ca" in a discontinuous quantal process in which successive increments of IP, transiently release precise amounts of Ca2+(21, 22). It has been well documented that this discrete release of Ca2+takes place in the absence of receptor desensitization (44).This novel phenomenon has so far only been seen in IP, and ryanodine receptors (44).Although there has been considerable speculation regarding the mechanisms underlying quantal Ca2+release, two alternative models are currently believed to best account for this phenomenon: the all-or-none model and the steady-state model (44).The central difference between these two models relates to their assumptions about the differences in the sensitivityof IP,R to IP,. The all-or-none model proposes that quantal Ca2+release is the result of complete emptying of distinct stores that are heterogeneous in their sensitivity to IP,. The steady-state model proposes that quantal release occurs because decreasing lumenal CaZ+allosterically attenuates the channel activity of IP,R that are homogenous in their sensitivity to IP,. There exists evidence in support of both models (44).Data have been reported supporting heterogeneity in the sensitivity of different intracellular but few studies have directly addressed stores to IP, (45, 46), heterogeneous sensitivity of IP,R themselves. We provided the novel evidence indicating that IP,R are intrinsically heterogeneous in their sensitivityto IP,, by showing

28978

Effects of Thimerosal on Purified IP, Receptor Activity

that purified and reconstituted receptors display quantal Ca2+ release (23). As discussed above, the effects of TMS on IP3R response to various concentrations of IP, can best be explained as a result of intrinsic variations in receptor sensitivity to its ligand. In order for heterogeneously sensitive receptors t o manifest quantal Ca2+release, they must respond maximally over a very narrow range of IP, concentrations. This is one of the predictions of the all-or-nonemodel, and it accounts for how a highly sensitive subpopulation of receptors can release a fraction of the total releasable Ca" in response to low levels of IP, without desensitizing (44). This behavior is implied by the response ofIP,R t o successive treatment with IP,,TMS, and BME (Fig. 6). The molecular basis of the functional receptor heterogeneity could arise from variations in receptor subtypes, splice forms, and phosphorylation or redox states. In summary, our findings document that TMS directly modifies critical sulfhydryl group(s) present on the IP3Rto markedly increase or abolish its activity. The stimulation of IP,R activity by TMS, which occurs at a location remote from the IP, binding domain, doubles the affinity of the receptor for its ligand without increasing the number of available sites. In thepresence of a fixed concentration of IP, and 45Ca2+,addition of TMS can stimulate IP,R activity, with implications for physiologic and pathophysiologic redox modulation of intracellular Ca". Finally, we have used TMS mediated perturbations of IP3R activity to provide evidence for functional IP,R heterogeneity. These findings are consistent with our previous observations (23) and with the original hypothesis regarding the mechanism of quantal Ca2+release by IP,R (21).

(1993)Pfulgers Arch. Eur: J . Physiol. 424, 516-522 13. Hilly, M., Pietri-Rouxel, F., Coquil, J. F., Guy, M., and Mauger, J. P. (1993) J. Biol. Chem. 268, 16488-16494 14. Bootman, M. D., Taylor, C.W., and Berridge, M. J. (1992)J. Biol. Chem. 267, 25113-25119 15. Bird, G. St. J., Burgess, G. M., and Putney, J. W., Jr. (1993)J . Biol. Chem. 268, 17917-17923 16. Poitras, M., Bernier, S., Servant, M., Richard, D. E., Boulay, G.,and Guillemette, G. (1993)J . Biol. Chem. 268, 24078-24082 17. Ferris, C. D., Huganir, R. L., Supattapone, S., and Snyder, S. H. (1989)Nature 342,8749 18. Supattapone, S., Danoff, S. K., Theibert, A,, Joseph, S. K., Steiner, J., and Snyder, S. H. (1988)Proc. Natl. Acad. Sei. U. S. A. 8 5 , 8747-8750 19. Ferris, C. D., Huganir, R. L., Bredt, D. S., Cameron, A. M., and Snyder, S. H. (1991)Proc. Natl. Acad. Sci. U. S . A . 8 8 , 2232-2235 20. Ferris, C. D., Huganir, R. L., and Snyder, S. H. (1990)Proc. Natl. Acad. Sci. U. S. A. 87,2147-2151 21. Muallem, S., Pandol, S. J., and Beeker, T. G. (1989)J. Biol. Chem. 264, 205-212 and Stryer, L. (1990)Proc. Natl. Acad. Sci. U. S. A. 87,3841-3845 22. Meyer, T., 23. Ferris, C. D., Cameron, A. M., Huganir, R. L., and Snyder, S. H. (1992)Nature 366,350-352 24. Hecker, M., Brune, B., Decker, K., and Ullrich, V. (1989)Biochem. Biophys. Res. Commun. 159,961-968 25. Swann, K. (1991)FEBS Lett. 278,175-178 26. Miyazaki, S., Shirakawa, H., Nakada, K., Honda, Y., Yuzaki, M., Nakade, S., and Mikoshiba, K. (1992)FEBS Lett. 309, 180-184 27. Galione, A,, McDougall,A,, Busa, W. B., Willmott, N., Gillot, I., and Whitaker, M. (1993)Science 261, 348-352 28. Thorn, P., Brady, P., Llopis, J., Gallacher, D.V., and Petersen, 0. H. (1992) Pflugers Arch. Eur: J . Physiol. 422, 173-178 29. Fissore, R. A,, and Robl, J. M. (1993)Deu. Biol. 159, 122-130 30. Grinnell, F. L., Stollar, B. D., and Nishlmura, J. S. (1969)Biochim. Biophys. Acta 185,471-474 31. Bredt, D. S., Mourey, R. J., and Snyder, S. H. (1989)Biochem. Biophys. Res. Commun. 159,976-982 32. Maeda, N., Kawasaki, T., Nakade, S., Yokata, N., Taguchi, T., Kasai, M., and Mikoshiba, K. (1991)J. Biol. Chem. 266, 1109-1116 33. Mignery, G. A., and Sudhof, T. C. (1990)EMBO J. 9,3893-3898 34. Miyawaki, A,, Furuichi, T., Ryou, Y., Yoshikawa, S., Nakagawa, T., Saitoh, T., and Mikoshiba, K. (1991)Proc. Natl. Acad. Sci. U. S . A. 88,49114915 Acknowledgments-We thank Dr. Paul Talalay and Dr. Tom Woolf for R. J., Estevez, V. A,, Maracek, J. F., Barrow, R. K., Prestwich, G . D., helpful discussions and Nancy Brucefor help preparing the manuscript. 35. Mourey, and Snyder, S. H. (1993)Biochemistry 32, 1719-1726 36. Furuichi, T., Yoshikawa, S., Miyawaki, A., Wada, K., Maeda,N.,and REFERENCES Mikoshiba, K. (1989)Nature 342,32-38 37. Mignery, G. A,, Newton, C. L., Archer, B. T., 111, and Sudof, T. C. (1990)J . Biol. 1. Bemdge, M. J. (1993)Nature 361,315-325 Chem. 266,12679-12685 2. Fems, C. D.,and Snyder, S. H. (1992)Annu. Reu. Physiol. 54,469488 38. Miyazaki, S., Yuzaki, M., Nakada, K., Shirakawa, H., Nakanishi, S., Nakade, 3. McPherson, P. S., and Campbell, K. P. (1993)J. Biol. Chem. 268,13765-13768 S., and Mikoshiba, K. (1992)Science 257,251-255 4. Sorrentino, V., and Volpe, P. (1993)Dends Pharmacol. Sci. 14, 98-103 5. Burdon, R. H., and Rice-Evans, C. (1989)Free Rad. Res. Commun. 6,345-358 39. Fadool, D. A,, and Ache, B. W. (1992)Neuron 9,907-918 40. Renard, D.C., Steitz,M. B., and Thomas,A. P. (1992)Biochem.J. 284,507-512 6. Allen, R. G. (19913Proc. SOC.Ezp. Biol. Med. 196, 117-129 41. Uhlig, S., and Wendel, A. (1992)Life Sei. 51, 1083-1094 7. Kass, G. E., Nicotera, P., Orrenius, S. (1992)Cellular and Molecular Mecha42. Ehrlich, B. E., Kaftan, E., Bezprozvannaya, S., and Bezprozvanny, I. (1994) nism of Inflammation, Vol. 4,pp. 133-152,Academic Press, Inc., New York Trends Pharmacol. sci. 15, 145-149 8. Hatzelmann, A., Haurand, M., and Ullrich, V. (1990)Biochem. Pharmacol. 39, 43. Watras, J.,Bezprozvanny, I., and Ehrlich, B. E. (1991)J. Neurosci. 11,3239559-567 3245 9. Missiaen, L., Taylor, C. W., and Berridge, M. J. (1991)Nature 352, 241-244 10. Missiaen, L., Taylor, C. W., and Berridge,M. J. (1992)J . Physiol. 455,623-640 44. Bootman, M. D. (1994)Mol. Cell. Endocr 98,157-166 11. Sayers, L. G., Brown, G.R., Michell, R. H., and Michelangeli, F. (1993) 45. Thorn, P., Lawrie, A. M., Smith, P. M., Gallacher, D. V., and Petersen, 0. H. (1993)Cell 74,661468 Biochem. J. 289,883487 46. Kasai, H., Li, Y.X., and Miyashita, Y. (1993)Cell 74, 669-677 12. Parys, J. B., Missiaen, L.,De Smedt, H., Droogmans, G., and Casteels, R.