",,',

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICSVol. 278, No.2, May 1, pp. 000-000, 1990

Persistent Ca2+-lnduced Activation of Erythrocyte

Membrane Ca2+-A TPase Unrelated toCalpain Proteolysis 1

Basil D. Roufogalis,2 Izabela Brzuszczak, Arthur D. Conigrave, You-Han Xu,Cheryl L. Machan,:J and Kevin K. W. Wang3.4Department of Biochemistry, University of Sydney, Sydney. NS W 2006 Australia

Received September 29, 1989, and in revised form December 27, 1989

Ca2+ activation of the Ca2+-ATPase. On the other hand,Ca2+-pretreatment of the membranes modified the sus-ceptibility of the Ca2+-ATPase to both cleavage and ac-tivation by exogenously added calpain I. We concludethat pretreatment of Ca2+-A TPase in erythrocyte mem...branes with millimolar Ca2+ activates the enzyme by in-ducing a persistent conformational change of the en-zyme which is, however, subsequently reversed bydetergent solubilization. @ 1990 Acudemic Presa,lnc.

Preincubation of human erythrocyte membraneswith calcium in the submillimolar to millimolar concen-tration range resulted in an increase of the Ca2+ affinityand apparent maximum velocity of the Ca2+ -stimulatedMg2+-dependent ATPase (Ca2+-ATPase). The activa-tion was persistent, as it was not reversed when theCa2+-preincubated membranes were washed with eth-ylene glycol bis(,B-aminoethyl ether) N,N'-tetraaceticacid-containing buffers. Magnesium was not requiredfor the activation, whereas greater than 2 mM Mg2+partially antagonized the activation by Ca2+. In somemembrane preparations A TP was required in additionto Ca2+ for activation of the Ca2+-ATPase, but nonhy-drolyzable analogs of A TP had the same effect. Calmod-ulin prevented the activation by Ca2+ over the sameconcentration range in which it interacts with the Ca2+-A TPase. Taken together the results obtained providedstrong evidence that the Ca2i activation of the enzymewas not due to proteolytic cleavage by endogenous cal-pain. Thus, activation by Ca2+ was not blocked by leu-peptin (100-200 IJ.M), did not require dithiothreitol,and occurred at Ca2+ concentrations greater than thoserequired for activation of calpain I. Furthermore, Ca2+activation did not result in change in the mobility thenative 136-kDa species of the Ca2+-A TPase on SDS-gelelectrophoresis. Moreover, solubilization of the Ca2r-pretreated membranes with Triton X-100 reversed the

The calcium-stimulated, magnesium-dependentATPase (Ca2+-ATPase) in the erythrocyte plasma mem-brane maintains the intracellular level of free calcium inthe submicromolar concentration range (1,2). A numberof factors have been shown to activate the Ca2+-ATPase,including the ubiquitous calcium-binding protein cal-modulin (3, 4), limited proteolysis (5-9), interactionwith acidic phospholipids (10,11), including phosphati-dylinositoI4,5-bisphosphate (PIPJ5 (12, 13), and a vari-ety of organic anions (14). Partial activation has alsobeen obtained with cAMP-dependent protein kinase(15, 27) and protein kinase C (16). Of these mechanisms,only the activation by calmodulin has been clearly dem-onstrated under physiological conditions (17) .The Ca2+ -

., Abbreviations used: AMP-PNP, 5'-adenylimido-diphosphate:AMP-PCP, {:1-'Y-methylene adenosine 5'-triphosphate: ATP-'Y-S, ad-enosine-5'-0-(3-thiotriphosphate); BSA, bovine serum albumin:EGTA, ethylene glycol bis({:1-aminoethyl ether)N,N'-tetraacetic acid:Hepes, N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid; IP",inositol 1,4,5-trisphosphate; Mops, 3-(N-morpholino) propanesul-tonic acid: PAGE, polyacrylamide gel electrophoresis; PI, phosphati-dylinositol; PIP, phosphatidylinositol-4-phosphate; PIP2, phosphati-dylinositol-4,.5-bisphosphate; PLC, phospholipase C; SDS, sodiumdodecyl sulfate; Tris, tris(hydroxymethyl) aminomethane; PMSF,phenylmethylsulfonyl fluoride; CaM, calmodulin; PLA2, phospholi-pase A2°

I Part of the data was presented in ahstract form in Proceedings of

the Australian Biochemical Society Vol. 20, Ahst. P:\l ( 1988) and Vol.

21. C 1R (19R9)." To whom correspondence should he addressed at the Department

of Pharmacy, University of Sydney, NSW, 2006 Australia.1 From the Lahoratory of Molecular Pharmacology, Faculty of

Pharmaceutical Sciences, University of British Columhia, Vancouver,

B.C. V6T 1 Wfi, Canada.4 Present address: Department of Patholol..'Y, School of Medicine,

Wayne State University, Detroit, Michil(an 4R201.

001OOO:I-\JHI) I/!JO $:\.00Copyril(ht ", 1!J9f1 I)y Aclldemic PreR". Inc.All ril(ht" of reproduction in uny form re"erved.

002 ROUFOGALIS ET AL.

30A TPase is also activated by pretreatment of erythrocytemembranes with millimolar concentrations of calcium(18), an effect which was recently confirmed by Au et al.(19). The mechanism of this activation remains unclear.Whereas Klinger et al. (18) found no evidence for prote-olysis of the enzyme during activation, Au et al. ( 19) sug-gested that the activation may be due to limited proteol-ysis by endogenous calpain. In this study we presentevidence inconsistent with calpain proteolysis as thepredominant mechanism of Ca2+-induced activation ofCa2+-ATPase activity. A number of alternative mecha-nisms have been considered.

>....

">:::().,

".,.,

A.

e.;.I

+

C\I.,U

.1:;.-

"1;:-*

~

-0.0-0-.L~-r- ..

0

J.l 1.0 10.0

Free Ca2+ (jJ.M)

FIG. I. Effect of preincubation ofTris ghosts with Ca2+ on the Ca2+-ATPase activity. The preincubation medium contained 100 JIM EDTA(no added MgCI2) and either 0.1 mM EGTA (0) or 3 mM CaCl2 plus0.5 mM ATP (.). For comparison, membranes preincubated with 0.1mM EGTA were also assayed in the presence of 150 ~M calmodulin(~). The results are the means :t SD of four preparatitns.

a

-,I\.-.

described by Goldstein (25) using a Ca2+: EGTA.- association con-stant of 101°.97 and other constants from Martell and Smith (26).

Proteolysis of membrane-bound Ca2+ -A TPase by calpain I. CalpainI was purified from human erythrocyte cytosol as described previously(9, 28). Control or Ca2+-preincubated ghosts were incubated with cal.pain I ( llLg) at 25°C for 15 min, in the presence or absence ofcalmodu-lin (see figure legends), as previously described (9, 28). Proteolysis wasstopped by the addition of 200 ILM leupeptin. Samples were chilled onice and either phosphoenzyme formation was determined (see above)or the enzyme was solubilized (see figure legends for details).

Formation of the phosphorylated enzyme intermediate by the Ca"'".A TPase. The phosphoenzyme intermediate of the Ca"+ -A TPase wasdetermined essentially as described by Wang et al. (28). Control orCa2+-preincubated ghosts (50-100 ILg protein) were incubated at 4°Cfor 15 s in 40 mM Hepes, 2.5 mM CaCI2' 300 nM calmodulin, and 2ILMATP containing 6 mCi.mmol-' [-y-:1:!P)ATP (pH 7.2). The reactionwas initiated by the addition of ATP and stopped by the addition of10% (w/v) trichloracetic acid. The precipitated protein was processedfor acid gel electrophoresis at pH 6.6 as described before (9).

RESULTS

Characteristics of Ca2+-Induced Activationof Ca2+-A TPase

Preincubation of human erythrocyte ghosts with cal-cium activated their Ca2+-ATPase activity, which wasassayed in the same membranes subsequently washed toremove the added calcium. From preliminary studies wedetermined that activation was both time and tempera-ture dependent, and incubation for 60 min at 37°C wassufficient to ensure a maximum activation. Preincuba-tion with 3 mM CaC12 (2.4 mM free calcium) in the pres-ence of 0.5 mM A TP enhanced both the Ca2+ affinity andthe apparent maximum velocity of the Ca2+-ATPase rel-ative to a control incubated with EGTA in the absenceof calcium (Fig. 1). The activation appeared similar tothat obtained when the EGTA-preincubated mem-branes were assayed in the presence of calmodulin. Theincrease in maximum velocity after Ca2+ preincubation

MATERIALS AND METHODS

MateriaL'I. Leupeptin, histamine dihydrochloride, fatty acid freeBSA, PMSF, ATP (vanadium free, disodium salt), EOTA, Tris-base,Tris-phosphate, Mops, SOS, neomycin sulfate, AMP-PNP, andAMP-PCP were from Sigma. Adenosine, calmodulin (bovine brain),Coomassie brilliant blue R-250 and Hepes were from Calbiochem.Malachite green was from Hoskins and Williams and ammonium mo-lybdate was from BOH. 1"Y-31P)ATP (300 Ci/mmol) was from Amer-sham. Oithiothreitol and ATP-"y-S were from Boehringer. All otherreagents used were of analytical grade.

Preparation of calmodulin-depleted erythrocyte membranes. Cal-modulin-depleted human erythrocyte membranes were prepared ei-ther by saponin lysis in 0.2 mM EGTA-containing isotonic buffer (sa-ponin ghosts) ( 18, 20) or by hypotonic lysis in 1 mM EOT A-containingTris buffers (Tris ghosts) ( 10). All procedures were carried out at 4 oC.For saponin ghosts, 1 vol of washed red blood cells were hemolyzed in10 vol of 100 mM KCI, 50 mM Tris-HCI and 0.2 mM EGTA, pH 7.2/22°C, in the presence of 0.01-0.015% saponin for 10 min at room tem-perature, and white ghosts were isolated following washing of themembranes three times in the same buffer in the absence of saponin.Ghosts were frozen in liquid nitrogen and stored at -76°C before use.Tris ghosts were prepared by centrifugation as described previously(21). In some cases the procedure was performed by filtration of thelysate across 0.2-l'm pore size membranes similar to the method de-scribed in (21), except that the memhrane was a hollow fiber Plas-maftow Hi-05 unit (22). The white ghosts were stored in buffer con-taining 130 mM KCI, 2 mM dithiothreitol, 0.5 mM MgCI2, and 20 mMpotassium-Hepes, pH 7.4 (21). Protein, determined by the method ofLowry et a/. (23), was typically 2-5 mg/ml.

Preincubation with calcium. Ghosts (0.5 ml) were incubated for 60min at :17"C in 1 ml ofa medium containing (final concentration) Tris-HCI (5,,) mM), NaCI (66 mM), ouabain (0.1 mM), ROTA (100 I'M), andas required, either 0.1 mM EGTA or CaCI:! in the presence or absenceof ATP or calmodulin at pH 7.2 (see figure lel{ends for the particularcomposition of the preincuhation medium). The tubes were immersedin ice and after the addition of 0.5 ml of 40 mM Hepes-sodium (pH7.6) and sufficient EGTA to chelate the calcium (2 mM or more) thetubes were spun in an Eppendorf-type microcentrifuge for 15 min andthe membranes washed twice more with 1.5 ml of cold 40 mM Hepes-sodium buffer. The membranes were resuspended in this buffer totheir original volume.

Determination (if membrane-brJund Ca2'-A TPa.'ie activity. Pre-incubated washed membranes (20-50 I'g protein) were incubated at37"C for 30 min in a total volume of 0.4 mi containing .')5 mM Tris-HCI (pH 7.2), 66 mM NaCI, I mM MgCI:!, 0.1 mM ouabain, 2 mM ATP(disodium salt), 1fiO nM calmodulin (when added), 0.1 mM EGTA, andvarious concentrations of CaCI:! to achieve the desired level of freeCa:!'. The inorganic phosphate liberated to the medium was deter-mined colorimetrically usinl{ a sensitive malachite I{reen assay (24).Free Ca:!' concentrations were determined by the computer prol{ram

/m6390$5937 02-08-9008:34:13Archives arc

.::- 25

r~~ 20

00 15""'.~ae.5.

Ca2+.INDUCED ACTIVATION OF Ca2+.ATPase 003

25

-.A -- .~ 20...ItI 15

.~...I..oSo 5.~

qae

.E.

.

0/0

.

lU

~

01--0

16r

14'

1Z

10

e

6

5 62 3 4

Free Ca2+ (mY)B

..\\

,".,c

~

E,<I

+NIl 2U

01 I I I I0 5 10 15 20

Free Mg2+ (mIl)

FIG.2. Ca2+ and Mg:!t dependence of the Ca2+ activation oferythro-cyte membrane Ca:!+-ATPase. Membranes were preincubated at 3rcfor 60 min in a medium containing 100 I'M EDTA, 0.5 mM ATP atvarious CaCI2 concentrations (A) or at a fixed CaCI2 concentration of0.5 mM (8). The Ca:!t-ATPase activity of the subsequently washedmembranes was determined at 3.41'M free Ca:!t .A, tris ghosts (0) (rep-resentative experiment) or saponin ghosts (.) (mean of two experi-ments). 8, tris ghosts (0) and saponin ghosts (.) (mean(oftwo exper-iments). Similar results were- obtained on at least two other ghost

preparations.

0-0 0

the procedure used to wash the membranes prior to as-say, even when EGTA was present in the three washsteps (results not shown). This was further confirmedin a study in which further incubation of Ca2+ -activatedenzyme for 30 min at 37°C in fresh medium withoutadded Ca2+ and A TP did not decrease the activation ob-tained (results not shown).

It was found that in many membrane preparations ex-amined the presence of A TP enhanced the activation bycalcium, as was also reported by Klinger et at. (18). How-ever, the extent of this effect varied among particularmembrane batches (although it did not depend on thetype of membrane preparation used). Thus, the effect ofpreincubation with 3 mM CaCl2 alone varied from no ac-tivation to almost maximum activation between mem-brane batches. Nevertheless, enhancement of the Ca2+activation by A TP was a common finding, as illustratedin typical experiments shown in Fig. 3. Experiments todetermine the concentration dependence of the A TP re-quirement showed that maximum activation occurred at0.5 mM ATP (results not shown), suggesting that ATPwas bound at a low affinity site rather than at the cata-lytic site of the Ca2+-ATPase. This was further sup;;ported by results obtained with nonhydrolyzable A TPanalogs, where the enhancement of the activation byCa2+ was similar with 0.5 mM ATP, AMP-PNP, ATP-'Y-S (Fig. 3), or AMP-PCP (not shown).

As the activation by Ca2+ and A TP was kineticallysimilar to that by calmodulin, we examined the effect ofadding calmodulin together with Ca2+ and A TP in thepreincubation medium. Calmodulin protected againstthe Ca2+-activation of the Ca2+-ATPase (Fig. 4), as wasalso reported by Klinger et at. (18). The protection bycalmodulin was at a high affinity site, occurring over thesame concentration range of calmodulin required to acti-vate the Ca2+-ATPase activity.

was about 70% of that achieved with calmodulin (Fig. 1).However, the extent of Ca2+ activation varied betweenmembrane preparations, and over the course of thesestudies the Ca2+ -activated activity was between 50 and100% of the maximum obtainable with calmodulin.

The concentration dependence of the Ca2+ activationwas examined (Fig. 2). Half-maximal activation oc-curred at a free Ca2+ concentration around 0.5 mM (Fig.2A) and it saturated at around 2 mM free Ca2+, whilefurther increasing Ca2+ to greater than 5 mM tended todecrease the activation obtained. The concentration de-pendence of Ca2+ activation was similar in both saponinghosts and Tris ghosts (Fig. 2A). Magnesium was notrequired for the activation of the Ca2+-ATPase activityby Ca2+; in fact, the presence of magnesium at concen-trations above 2 mM in the preincubation medium inhib-ited the activation by CaCl2 (0.5 mM). The inhibitionwas only partial, as it plateaued between 6 and 20 mMMg2+ (Fig. 28). The small initial activation obtained atlow Mg2+ concentrations is probably due to an increasein free Ca2+ (from 0.2 to 0.3 mM) produced by Mg2+ inthis concentration range. In contrast to the activation ofthe enzyme by calmodulin, which is reversed when Ca2+or calmodulin is removed (3, 4), the activation by Ca2+preincubation was persistent, as it was not reversed by

20 ~

151

10

5

-:::;a>.

0>:;:()«1.,.,«1

11.

e.;cI

+

"'csU

6-~~.--L.--/67 0-0/

~~,...:::::::: e ::::;

011J~U 0.1 1.0 10.0

Free Ca2+ (JlM)

FIG. 3. Influence of ATP and ATP analogs on the Ca2+ activationof Ca2+-ATPase. Tris ghosts were incubated for 30 min at 3rc in thepresence of 0.1 mM EGTA (0), 3 mM CaCI2 alone (.) or together with0.5 mM ATP (6), AMP-PNP (A) or ATP-"y-S (0).

~

II:

.~

I...O..Po~

~'0

S

5

ROUFOGALIS ET AL004

B

3 4 5 6A

1 2

,', .

136-~

125124.)= ,..-127

82-;;= ,- --85

80

Ca -+ Ca -+ +

FIG. 6. Effect of calcium preincubation of human erythrocyteghosts on the M, of the phosphorylated enzyme intermediate of Ca:!+-ATPase and its calpain l-induced fragments. A, ghosts (100 Ilg pro-tein) were incubated for 30 min at 37"C in 55 mM Tris-HCI, 66 mMNaCI, 0.1 mM ouabain, 100 IlM EDTA, and 200 IlM leupeptin in theabsence of CaCI2 (lane 1) or in the presence of 2.5 mM CaCI:! (lane 2).B, a batch of control (lanes 3 and 5) or Ca2+-pretreated membranes(lanes 4 and 6) from A were treated with calpain I (lllg) in the absence(lanes 3 and 4) or in the presence (lanes 5 and 6) of 1.2 IlM calmodulinfor 15 min at 25°C. Proteolysis was stopped by the addition of leupep-tin (200 IlM). Samples from A and B \vere chilled on ice, and subjectedto phosphoenzyme formation, acidic gel electrophoresis, and autoradi-ography, as described under Materials and Methods. The relative M,(in kDa) of phosphoenzyme bands are indicated on the sides of the

autoradiograms.Effects of Detergent Solubilization

The effect of solubilizing Ca2+-ATPase, from controlor Ca2+ -pretreated membranes, was examined. The ac-tivity in the Triton X-I00-solubilized membranes wasthe same whether membranes were untreated or prein-cubated with EGTA or CaC12 (Fig. 5), despite the factthat the activity had been typically activated in themembranes preincubated with Ca2+ (results not shown).The calmodulin-activated Ca2+ -A TPase activity of thesesolubilized preparations was also unchanged (Fig. 5).

These results indicate that the Ca2+-activation was re-versed by detergent treatment and that the solubilizedenzyme from the Ca2+-pretreated membranes remainedfully calmodulin sensitive.

60

40

20

.

..'

!.

,~/

>.

~u<":J0-..<I

+

010U

/1I /@/~--~I I

-6 -7 -6 -5 -42+Log tree Ca (M)

FIG. 5. The effect of detergent ~olubilization on the Ca:!+-ATPa~eactivity of Ca:!i -preincubated erythrocyte membranes. Membraneswere preincubated for 60 min at :J7°C with either no additions (0, .),

)2 mML E(;TA (6, A), or 2 mM CaCI:! (0, .) and membranes washedfree of"11dded Ca:!1 or EGTA two times, a~ descrilJed in Fig. 1. Thepellet wa~ ~olubilized in the pre~ence of 0.1>% (w/v) Triton X-lOO, aspreviously de~crihed (21 ). The activity of the solubilized enzyme wasmea~ured at variou~ free Ca:!1 concentration~, in the ab~ence (opensymbols) or pre~ence (clo~ed ~ymhol~) ofcalmodulin ( 11>0 nM).

a I I

Susceptibility of Ca2+ -A TPase to Proteolysis

To examine whether Ca2+ preincubation caused prote-olysis of the Ca2+ -A TPase, the phosphorylated-enzymeintermediate of the enzyme was visualized by autoradi-ography on SDS-polyacrylamide gels in acidic condi-tions (Fig. 6A). There was no shift in the Mr of the Ca2+-A TPase following incubation of ghosts with 2.5 mMCaCl2 relative to a control incubated in the same condi-tions with EDT A (Fig. 6A, lanes 1 and 2). In these exper-iments the protease inhibitors leupeptin (200 JlM) andPMSF (200 JlM) were included, as the activation by Ca2+preincubation was not inhibited by the presence of theseprotease inhibitors.

Sensitivity to exogenous calpain. The Ca2+-ATPasewas, as expected, sensitive to calpain I added exoge-nously (Fig. 6B). The pattern of proteolysis before andafter preincubation with 2.5 mM CaCI2 was examined.We confirmed our previous finding that in the absenceof calmodulin the 136-kDa intact enzyme was cleavedpredominantly to a 124-lilla fragment in control mem-

02-08-9008:34:13/m6~90$59~7ArchiveR

'I:

~

'~

~~eaeE.

005Ca:!+-INDUCED ACTIVATION OF Ca2+-ATPase

branes. In membranes pretreated with 2.5 mM CaC12 thefragment formed by calpain was of slightly lower mobil-ity, which is likely closely related to the 125-kDa frag-ment which we identified in an earlier study (28) (Fig.6B, lanes 3 and 4). Consistent with this pattern, lessprominept fragments of 80 (with trace 82 kDa) and 82kDa were produced in the absence and presence of Ca2+preincubation, respectively. By contrast, in the presenceof calmodulin, identical fragments of 127 kDa (major)and 85 kDa (minor) were obtained for both control andCaC12-pretreated membranes (lanes 5 and 6).

A>-

-

~uo

QJVI0

a.1-~

I+N

OU

100.

80.

60

40

20

~~-~-~~.';//

t!/~ /0-0

.;~ /0/ 0. 0 o~o~-- -

-5

(M)

-6 -6

Log free Co2+

B

Influence of Calcium Pretreatment on the Activation ofthe Ca2+ -A TPase Activity by Calpain

The effect of Ca2+ pretreatment on the ability of cal-pain to activate the Ca2+-ATPase activity in the mem-branes by proteolysis (28) was examined. In these exper-iments it was necessary to solubilize the Ca2+-ATPaseactivity in ghosts in order to differentiate between acti-vation due to Ca2+ pretreatment (reversible by detergentsolubilization) and calpain fragmentation (irreversibleon detergent solubilization). Ghosts pretreated eitherwith 3 mM EDT A alone or 3 mM CaCl2 plus 1 mM EDT A(in the presence of PMSF (500 ILM) and leupeptin (200ILM) ) were thoroughly washed, treated with calpain I, sol-ubilized with Triton X-lOO and then assayed for Ca2+-A TPase (see Fig. 7 legend for details). A number of find-ings emerged from these experiments. First, the basalCa2+-ATPase activity i~olubilized preparations was es-sentially unchanged w~ether membranes had been pre-treated with EGTA (Fig. 7A) or CaCl2 (Fig. 7B), con-firming the finding in Fig. 5. Second, calpain fullyactivated the control Ca2+-ATPase activity (Fig. 7A),but only partially activated Ca2+ -pretreated membranes(Fig. 7B). Third, calmodulin fully activated the solubi-lized Ca2+-ATPase from either the control or CaCI2-pre-treated ghosts (Figs. 7A and B). Calmodulin also fullyactivated the partially activated enzyme from calpain-treated, Ca2+ -preincubated ghosts, showing that the lat-ter retained a substantial portion of its calmodulin-bind-ing domain (Fig. 7B).

/. 6~.7" ---6""

~-::;-"' /6 /0-0

.1!

.~.,6/

.?6-6-0/-O~O

>,

luOQJVIO

0-1-~I

+N

OU

.~

~0 .0

!

ol-8 -5

Log free Co2+ (M)

FIG. 7. Effect of Ca2+ preincubation of human erythrocyte ghostson the activation of Ca2+-ATPase by calpain I. Ghosts (2 ml, 4 mgprot/ml) were preincubated for 30 min at 37"C with 3 mM EGTA (A)or 3 mM CaCI2 and I mM EGTA (B) in the medium described in Fig. Iplus 500 I'M PMSF and 200 I'M leupeptin. Subsequently, the mixtureswere chilled on ice and 3 mM CaCI2 was added briefly to sample (A)and 2 mM EGTA was added to sample (B). The ghosts were thenwashed three times with a medium of 150 mM KCI, 20 mM Hepes (pH7.4) and 0.1 mM EDTA (Buffer 1) and resuspended to the originalghost volume. The ghosts were then treated with either none (0, .) or10 I'g calpain 1 (b., .) in the presence of 10 mM dithiothreitol, 300 I'MCaCI2' and 100 I'M EDTA for 60 min at 25°C. Proteolysis was arrestedby the addition of 200 I'M leupeptin. The ghosts were washed twicewith Buffer 1, as described above. The ghosts were then solubilizedwith Triton X-100, as described under Materials and Methods. Aftercentrifugation the supernatant was assayed for Ca2+-ATPase activityat various concentrations of free Ca2+ in the absence (0, b.) or in thepresence of 300 nM calmodulin (., .) (see Materials and Methods).Values shown are the means of two separate experiments.

-6

Other Po.'!sible Mechani.'!ms of Ca2+ -Activation

We examined the possibility that Ca2+ pretreatmentenhanced the Ca2+-ATPase activity by activating PI-specific phospholipase C, as it is known that in the sub-millimolar to millimolar Ca2+ concentration range phos-phatidylinositol-4,5-bisphosphate (PIP2) is hydrolyzedand that this phosphoinositide enhances the Ca2+affinity of the Ca2+-ATPase when incorporated togetherwith the pure protein in proteoliposomes (12). Neomy-cin 10 mM (which almost totally inhibits millimolarCa2+-induced IP:, formation in Tris membranes (29, 30)and B. D. Roufogalis, unpublished results) did not affectCa2+ plus ATP-induced activation in saponin ghosts

(Table I) while it blocked only 35% of the maximum acti-vation (without altering the increase of Ca2+ affinity) ofthe Ca2+-ATPase in Tris ghosts (results not shown).Furthermore, Ca2+ activation was not antagonized by 10mM adenosine, which inhibits PI and PIP phosphoryla-tion by ATP (31) (Table 0. Therefore, consistent withthe results showing that nonhydrolyzable A TP analogswere as effective as ATP in enhancing the Ca2+ activa-tion, phosphorylation of phosphoinositides was not as-sociated with the Ca2+ activation.

Another possible mechanism of the Ca2+ activationconsidered was disulfide crosslinking due to Ca2+ expo-

-E:JE

.xoE

~'--"

1

E:)

.~x0E

~

100

80

60.I

40.

20-

006 ROUFOGALIS ET AL.

TABLE I

Effect of Potential Antagonists on the Activation of Ca2+ -A TPase by Preincubation with Ca2+ Plus A TP

Ca2+-ATPase Activity"{nmol.mg prot-l. min-I)MechanismCompound

20.615.524.821.828.622.1

Calpain antagonist-y-Glutamyltransferase antagonistPI-kinase antagonistPIP2-specific PLC antagonistRemoval of fatty acids following PLA2 activation

Control

Leupeptin (100JlM)Histamine (5 mM)Adenosine (10 mM)Neomycin (10 mM)Fatty acid free BSA (in wash steps)

a Mean :t standard deviation of three determinations on saponin ghosts, preincubated with 3 mM CaCI2 and 0.5 mM ATP in the absence orthe presence of the compounds, as described under Materials and Methods.

b Activity without prior incubation with Ca2+ plus ATP was 2.0 :t 0.1 in the absence of calmodulin and 28.8 :t 2.5 nmol. mg prot-1. min-I inthe presence of 150 nM calmodulin. None of the treatments were significantly different from the control by a t test analysis at the 95% confidence

level.

sure of reactive sulfydryl groups. Intermolecular or in-tramolecular disulfide crosslinking is unlikely, however,as there was no evidence that Ca2+ induced higher mo-lecular weight aggregates of Ca2+-ATPase from SDS-P AG E of the Ca2+ -activated enzyme (Fig. 6) and the ac-tivation by Ca2+ was not reversed when the Ca2+ activa-tion was carried out in the presence of up to 50 mM di-thiothreitol (results not shown). We also considered thepossibility that Ca2+ caused polymerization of the Ca2+ -ATPase in the membrane, since dimerization of the pu-rified erythrocyte Ca2+-ATPase increases its activity(45) and miltnolar Ca2+ promotes the formation of sta-ble Ca2+-AT\:>ase oligomers in sarcoplasmic reticulum(46). However, radiation inactivation analysis of frozenmembranes, according to (47), showed no change in thefunctional molecular weight of the enzyme activity incontrol versus Ca2+-treated membranes (unpublishedresults). Finally, the possibility that Ca2+-sensitivephospholipase A2 may have activated the Ca2+-ATPaseby generation of free fatty acids (5) or lysophospholipidswas unlikely since washing Ca2+-pretreated membranesthree times in the presence of 0.1% fatty acid-free BSA( to remove any fatty acids that might have been formedduring the preincubation) or incubation with the phos-pholipase A2 inhibitor mepacrine (0.3 mM) did not pre-vent the enzyme activation (results not shown).

required to activate calpain I (2 to 100 ,uM) (34, 35), whilecalpain II (which is activated by millimolar Ca2+) cannotbe detected in erythrocytes (32, 33). The Ca2+ activationdid not require the addition of dithiothreitol (whereascalpain activation does) (9, 28). It is noteworthy that~dithiothreitol was included in the study of Au et at. (19),which concluded that the Ca2+ activation they observedwas due to proteolytic activation by endogenous calpainin their membranes.

The claim by Au et at. (19) that the activation of theCa2+ -A TPase by added calpain reported by Wang et at.(9) could have been accounted for by endogenous calpainin the membrane is also unlikely, as it was shown in (9)that no activation occurred when membranes were incu-bated with 100 ,uM Ca2+ alone (i.e. in the absence ofadded calpain). The significance of the small pool of cal-pain reported to be associated with isolated erythrocytemembranes (36, 41) therefore remains unclear. Finally,Ca2+ activation reported here was consistently foundeven in the presence of leupeptin (100-200 ,uM) (Table Iand K. K. W. Wang and B. D. Roufogalis, unpublishedresults). Although we did observe a small but not statis-tically significant inhibition of the Ca2+ -induced activa-tion by leupeptin (25%, Table I) and an equally smallamount of proteolysis of the Ca2+-ATPase incubatedwith Ca2+ in the absence of leupeptin and PMSF in thepresence of dithiothreitol (unpublished results) we donot consider that these are sufficient to explain most ofthe Ca2+-activation effect we observe. Leupeptin at theconcentrations used abolishes the activation of mem-brane Ca2+-ATPase by exogenous calpain I (9, 28). Al-though it was reported by Au et at. (19) that leupeptinblocked the activation of Ca2+-ATPase by Ca2+ preincu-bation, it should be noted again that their Ca2+ preincu-bation was in the presence of dithiothreitol (whichwould favor the activation of any calpain remaining as-sociated with the membrane) and, even then, the maxi-

DISCUSSION

Calpain proteolysis is an obvious mechanism by whichthe Ca2+-ATPase may be activated by pretreatment ofhuman erythrocyte membranes with millimolar calcium.Although a small amount of calpain is now recognizedto associate with the isolated membranes (19, 41), thismechanism has been eliminated as the predominantmode of Ca2+ activation in our studies for the followingreasons. The Ca2+ activation reported here occurs atCa:l+ concentrations (0.2 to 3 mM) well in excess of those

:t3.9b:t1.8:t1A:t2.1:to. 7:to.7

Ca2+.INDUCED ACTivATION OF Ca2+.ATPase 007

ghosts, via activation of PI-specific phospholipase C(30). However, the activation also occurred in saponin-lysed ghosts, which do not contain Ca2+-activated PI-specific phospholipase C activity (30). In addition 10 mMneomycin, an inhibitor of PIP2 breakdown in Tris ghosts(29 and results not shown) also failed to prevent the acti-vation by Ca2+ .

Another mechanism excluded was the liberation of ac-tivating free fatty acids or lysophosphatidylcholine byendogenous Ca2+ -dependent phospholipase A2 activa-tion, as Ca2+ activation was not abolished by inclusionof fatty acid-free BSA either during the preincubationor during the subsequent washing steps, or by addition ofthe phospholipase A2 inhibitor mepacrine. Furthermore,there was no evidence for activation by intramolecularor intermolecular cross linking of the enzyme by disulfidebond formation or by Ca2+-dependent activation of 'Y-glutamyltransferase (37) (Table I). Ca2+-induced dityro-sine formation (38) is also unlikely, as this reaction islight-dependent, whereas the Ca2+-activation occurredwhen membranes were preincubated with Ca2+ in thedark (results not shown). Finally, we found no evidencefor Ca2+-induced polymerization of the Ca2+-ATPase by;radiation inactivation analysis. Failure of millimolarCa2+ to activate the Ca2+-ATPase activity in Triton X-100-solubilized prep..?J:ations or when the enzyme was re-constituted into a~c or neutral phospholipid vesicles /aljJ;.~(unpublished results) also tends not to support an aggre-gation mechanism.

Evidence obtained from the calpain susceptibility ofthe enzyme suggests that activation by Ca2+ pretreat-ment involved a conformational change in the Ca2+-A TPase, rather than the proteolytic release of the inhib-itory and calmodulin-binding domains of the enzyme.Thus, calpain I cleaved the phosphoenzyme intermedi-ate in control membranes to a 124-kDa species, whichwe have shown previously to have lost the ability to bindcalmodulin and to be a high Ca2+ -affinity and high activ-ity enzyme fragment (28). However, after Ca2+ preincu-bation the Ca2+-ATPase was cleaved to a form interme-diate between the 124- and 127-kDa forms (Fig. 6B),probably the 125-kDa form which we have shown pre-viously to retain calmodulin-binding capacity and to beactivated (28). That the calmodulin-binding site is pre-served in the Ca2+ -activated state is further confirmedby the finding that calmodulin protected both the con-trol and the Ca2+ pre incubated Ca2+-A TPase against cal-pain cleavage of the enzyme to the 124- to 125-kDa frag-ments, giving rise to the 127-kDa fragment, which wehave shown (28) to retain its calmodulin-binding site(Fig. 6B). Furthermore, solubilization of the Ca2+-pre-treated enzyme by Triton X-100 restored its low basalactivity and its sensitivity to activation by calmodulin.Further evidence for a structural change in the Ca2+ -A TPase molecule on Ca2+ preincubation was that thecontrol enzyme was fully activated by calpain I incuba-

mum inhibition by leupeptin which they reported wasonly about 40%. While membrane-associated calpainhas been reported to be insensitive to its natural inhibi-tor, calpastatin (36,41), a comparable lack of sensitivityto leupeptin has not been reported.

Direct evidence that calpain-like proteolysis of theCa2+-ATPase did not occur during Ca2+-induced activa-tion of the Ca2+-ATPase in this study was obtained bySDS-gel electrophoresis. While we readily detectedfragmentation of the 136-kDa native Ca2+-ATPase byexogenously added calpain I (Fig. 6B and Ref. (28», nochange in relative molecular weight of the enzyme wasobserved after Ca2+ pretreatment under conditionswhich activated the Ca2+ -A TPase (Fig. 6A). By contrast,the virtually complete loss of recovery of protein follow-ing Ca2+-preincubation reported by Au et at. (19) mayhave been the result of Ca2+ -induced protein aggregationrather than complete proteolytic degradation as theysuggested.

Other experiments were also inconsistent with cal-pain-induced proteolysis. While the Ca2+-induced acti-vation was persistent, being retained following removalof Ca2+ with EGT A and EDT A washes and after dilutionof the membranes in the assay medium, it was, neverthe-less, reversed following solubilization of the Ca2+-ATP-ase by Triton X-100 (Figs. 5 and 7). This would not beexpected, of course, if endogenous calpain had inducedproteolysis, which removes a 9- to 12-kDa fragment andconverts the enzyme to a calmodulin-independent acti-vated state (28). Klinger et at. (18) also did not find anyevidence for inhibition of the Ca2+-induced activation byvarious protease inhibitors they examined, includingchymostatin, which should inhibit activation due to car-boxy-terminal cleavage by chymotrypsin-Iike activity(44). Therefore, unlike Au et at. (19), we conclude thatCa2+-activation mediated by calpain cleavage is not thepredominant mechanism of the Ca2+-activation ob-served in our studies.

The question then remains as to what other mecha-nism(s} may account for Ca2+ activation of the Ca2+-A TPase. The enhancement of Ca2+ activation by A TP insome membrane preparations may have been associatedwith A TP-dependent enzymatic activities, such as phos-phorylation by protein kinases of the Ca2+-ATPase (15,16) or of phospholipids (31). However, the effect doesnot appear to require A TP hydrolysis, as a range of thenonhydrolyzable analogs could substitute for A TP (Fig.3) and it did not require Mg2+ (Fig. 2). Furthermore, thepossibility that increased synthesis of PIP2 (a known ac-tivator of purified Ca2+-ATPase (31,12» from the phos-phorylation of PI and PIP in erythrocyte membranesmay indirectly activate the Ca2t-ATPase (31, 12) is ex-cluded, as adenosine, which blocks PI and PIP phos-phorylation (31), did not inhibit Ca2+ activation in thepresence of A TP (Table I). In fact, it is more likely thatthe PIP,J levels fell during Ca2+ preincubation of Tris

ROUFOGALIS ET AL.008

protects the Ca2+-ATPase against Ca2+-induced activa-tion by antagonizing the proposed lipid interactions ei-ther sterically or by a conformational change. The fail-ure of Ca2+ to activate the enzyme in the detergentor phospholipid-reconstituted state remains unclear.Clearly, further work is required to establish the mecha-nism of activation of the Ca2+ -A TPase by millimolar cal-cium concentrations.

~ out

acidicphospholipids

-: =~ ~ ~ ~ ~- in-Ca2+C;2+ -

+ ~A + -[124kDa -

-CaM-:, -

127kDa;. ~ ,

toC-term.

FIG. 8. Proposed model for the Ca2+ ..jnduced activation of Ca2+-A TPase based on the sites of action of acidic phospholipid, calmodu-lin, and calpain. The region labeled ( + ) is the calmodulin-binding do-main, while the putative Ca2+ /acidic phospholipid-interacting domainis labeled ( -)- Calmodulin ( CaM) interacts with the positively chargedcalmodulin-binding domain, where three calpain cleavage sites arealso located- These cleavages produce the 127, 125, and 124-kDa frag-ments, respectively. After pretreatment in the presence of millimolarCa"+, acidic lipids in the membrane may interact reversibly via "Ca2+bridges" with the negatively charged sequence. The enzyme remainspersistently activated even upon removal of the high concentrationsof Ca2+ initially used to activate the enzyme.

-~..."

10 N-term,

calpain

cleavagesACKNOWLEDGMENTS

We thank Dr. Michel Potier and his assistant, Louise Thausette,for carrying out the radiation inactivation analysis in Montreal. Thiswork was supported by a grant from the National Health and MedicalResearch Council of Australia. Some of the experiments were also sup-ported by the Canadian Heart Foundation (B.C.) and the Medical Re-search Council of Canada. We thank Sandy Butler for expert secre-tarial assistance.

tion whereas the Ca2+-pretreated Ca2+-ATPase could beonly partially activated by calpain (Fig. 7). Thus, Ca2+pretreatment protected the Ca2+-ATPase against cal-pain l-induced cleavage to its maximally active form.

On the basis of the predicted sequence of the plasmamembrane Ca2+-ATPase (39, 40) we propose a tentativemodel for the mechanism of Ca2+ activation of the mem -brane-associated Ca2+-ATPase (Fig. 8). Ca2+ pretreat-ment alters the normal interaction of membrane phos-pholipids at the negatively charged region immediatelyN-terminal to the calmodulin-binding domain (residues1079-1b94) (43) (see Fig. 8). Ca2+ might achieve this ei-ther by altering the lipid profile in the environmentaround this region and/or it might participate directlyin the binding of acidic phospholipids (which are knownto activate the purified Ca2+-ATPase (44» to the nega-tively charged region by Ca2+ bridges (see Fig. 8). Thisleads to a persistent activation which may be reversedonly by detergent treatment (Fig. 5). A role of lipids inthe mechanism of Ca2+-induced activation is supportedby experiments in which Ca2+ activation was largelyabolished by hydrolysis of membrane lipid head groupsby exogenously added phospholipase C (y -H. Xu and B.D. Roufogalis, unpublished results). The proposed con-formational change following Ca2+-induced activationalso alters the susceptibility of the protein to initialcleavage by calpain (Figs. 6 and 7), which has beenshown to occur at multiple sites within the calmodulin-binding domain (28, 42) (see Fig. 8). In turn, calmodulin

REFERENCES

1. Schatzmann, H. J. (1982) in Membrane Transport of Calcium(Carafoli, E., Ed.), pp. 41-107, Academic Press, London.

2. Lew, V. L., Tsien, R. Y., and Miner, C. (1982) Nature (London)298.478-481.

3. Gopinath, R. M., and Vincenzi, F. F. (1977) Biochem. Biophys.Res. Commun. 77.1203-1209.

4. Jarrett, H. W ., and Penniston, J. T. ( 1977) Biochem. Biophy.'!. Re.'!.Commun.77.1210-1216.

5. Taverna, R. D., and Hanahan, D. J. ( 1980) Biochem. Biophys. Res.Commun. 94. 652-659.

6. Sarkadi, B., Enyedi, A., and Gardos, G. (1980) Cell Calcium I.287-297.

7. Carafoli, E., Zurini, M., Niggli, V ., and Krebs, J. (1982) Ann. N. Y.Acad. Sci. 402.305-326.

8. Au, K. S. (1987) Biochim. Biophys. Acta 905.273-278.9. Wang, K. K. W., Villalobo, A., and Roufogalis, B. D. (1988) Arch.

Biochem. Biophys. 260. 696-704.10. Carafoli, E., Niggli, V., and Penniston, J. T. (1980) Ann. N.Y.

Acad. Sci. 358.159-168.11. Al-Jobore, A., and Roufogalis, B. D. (1981) Canad. J. Biochem.

59.880-888.12. Choquette, D., Hakim, G., Filoteo, A. G., Plishker, G. A., Bost-

wick, J. R., and Penniston, J. T. (1984) Biochem. Biophy,,!. Res.Commun. 125,908-915.

13. Enyedi, A., Flura, M., Sarkadi, B., Gardos, G., and Carafoli, E.(1987) J. Bioi. Chem. 262.6425-6430.

14. Minocherhomjee, A-e- V ., Al-Jobore, A., and Roufogalis, B. D.( 1982) Biochim. Biophys. Acta 690.8-14.

15. Neyses, L., Reinlib, L., and Carafoli, E. (1985) J. Bioi. Chem. 260,

10,283-10,287.16. Sma11wood, J. I., Gugi, B., and Rasmussen, H. (1988) J. Bioi.

Chem. 263.2195-2202.17. Mua1\em, S., and Karlish, S. J. D. (1982) Biochim. Biophy,,!. Acta

687, 329-:J32.18. Klinger, R., Wetzker, R., Fleischer, I., and Frunder, H. (1981) Cell

Calcium 2.553-560.19. Au, K. S., Lee, M. F., and Siu, Y. L. (1989) Biochim. Biophy,'I. Acta

978. 197-202.20. Hanahan, D. J., and Ekholm, J. E. (1978) Arch. Biochem. Biophys.

187.170-.179.

009Ca2+-INDUCED ACTIVATION OF Ca2+-ATPase

36. Hatanaka, M., Yoshimura. N.. Murakami, T., Kannagi. R.. andMurachi, T. (1984) Biochemistry 23, 3272-3276.

37. Anderson, D. R., Davis, J. L., and Carraway, K. L. (1977) J. Biol.Chem. 19.6617-6623.

38. Malencik, D. A., and Anderson, S. R. (1987) Biochemistry 26.695-704.

39. Verma, A. K., Filoteo. A. G., Standord, D. R.. Wieben, E. D.. Pen-niston. J. T., Strehler. E. E.. Fischer. R.. Heim, R.. Vogel. G.. Ma-thews, S., Strehler-Page. M.-A., James, P., Vorherr, T., Krebs, J.,and Carafoli, E. (1988) J. Biol. Chem. 263.14,152-14.159.

40. Shull, G. E., and Greeb, J. ( 1988) J. Bioi. Chem. 263.8646-8657 .

41. Gopalakrishna, R.. and Barsky. S. H. (1986) J. Biol. Chem. 261,

13.936-13,942.42. James, P., Vorherr, T., Krebs. J., Morelli, A., Castello, G., McCor-

mick, D. J., Penniston, J. T.. De Flora, A., and Carafoli, E. (1989)J. Biol. Chem. 264. 8289-8296.

43. Carafoli, E., Fischer. R., James. P., Krebs. J.. Maeda, M., Enydei,A., Morelli. A., and De Flora, A. (1987) in Calcium-Binding Pro-teins in Health and Disease (Means, A. N., Ed.). pp. 78-79, Aca-demic Press, New York.

44. Papp, B., Sarkadi, B., Enyedi. A., Caride, A. J.. Penniston, J. T.,and Gardos, G. (1989) J. Biol. Chem. 264.4577-4582.

45. Kosk-Kosicka, D.. and Bzdega, T. (1988) J. Biol. Chem. 263.

18,184-18,189.46. Dux, L., and Martonosi, A. (1983) J. Biol. Chem. 258.2599-2603.47. Minocherhomjee, A. M.. Beauregard, G., Potier, M., and Roufo:

galis, B. D. (1983) Biochem. Biophy.'I. Re.'I. Commun. 116,895-900.

21. Villalobo, A., Brown, L., and Roufogalis, B. D. (1986) Biochim.Biophys. Acta 854, 9-20.

22. Price, W. S., Roufogalis, B. D., and Kuchel, P. W. (1989) Anal.Biochem. 179, 190-193.

23. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J.(1951) J. Bioi. Chem. 193, 265-275.

24. Lanzetta, P. A., Alvarez, L. J., Reinach, P. S., and Candia, 0. A.(1979) Anal. Biochem. 100, 95-97.

25. Goldstein, D. (1979) Biophys. J. 26, 235-242.26. Martell, A., and Smith, R. (1982) in Critical Stability Constants,

Vol. 5, Plenum Press, New York.27. James, P. H., Pruschy, M., Vorherr, T. E., Penniston, J. T., and

Carafoli, E. ( 1989) Biochemistry 28, 4253-4258.28. Wang, K. K. W., Roufogalis, B. D., and Villalobo, A. (1988) Arch.

Biochem. Biophys. 267, 317-327.29. Allan, D., and Michell, R. H. (1981) Biochim. Biophys. Acta 508,

277-286.30. Hegewald, H., Miller, E., Klinger, R., Wetzker, R., and Frunder,

H. (1987) Biochem. J. 244, 183-190.31. Buckley, J. T. (1977) Biochim. Biophys. Acta 498, 1-9.

32. Murachi, T. (1983) Trend.'I Biochem. Sci. 8, 167-169.33. Takano, E., Park, Y -H., Kitahara, A., Yamagata, Y ., Kannagi, R.,

and Murachi, T. (1988) Biochem. lnt. 16, 391-395.34. Pontremoli, S., and Melloni, E. (1986) in Calcium and Cell Func-

tion (Cheung, W. Y., Ed.), Vol. vrpp. 159-183, Academic Press,New York.

35. Kuboki, M., lshii, H., and Kazama, M. (1987) Biochim. Biophy.'I.

Acta929,164-172.