2tctsdri

Archives of Biochemistry and BiophysicsVol. 379, No. 2, July 15, pp. 337–343, 2000doi:10.1006/abbi.2000.1889, available online at http://www.idealibrary.com on

Caspase-Mediated Proteolytic Activation of Calcineurin inThapsigargin-Mediated Apoptosis in SH-SY5YNeuroblastoma Cells

Neeta Mukerjee,†,1 Kim M. McGinnis,*,1,2 Yang Hae Park,* Margaret E. Gnegy,*and Kevin K. W. Wang*,†,3

*Department of Pharmacology, University of Michigan Medical School, Ann Arbor, Michigan 48109; and †Laboratory ofNeuro-biochemistry, Department of Neuroscience Therapeutics, Parke-Davis Pharmaceutical Research Divisionof Warner-Lambert Company, 2800 Plymouth Road, Ann Arbor, Michigan 48105

Received February 18, 2000

We previously demonstrated a loss in calmodulin(CaM)-dependent protein kinase activity in SH-SY5Ycells undergoing thapsigargin-mediated apoptosis,(K. M. McGinnis et al., 1998, J. Biol. Chem. 273, 19993–0000). Here we demonstrate that the large subunit ofhe CaM-dependent protein phosphatase 2B (cal-ineurin) is fragmented during SH-SY5Y cell apoptosiso a major fragment of 45 kDa in a caspase inhibitor-ensitive manner. A 45-kDa fragment was also pro-uced when purified calcineurin was digested withecombinant caspase-3. The major cleavage site wasdentified to be DFGD* G386ATAA, which removes the

C-terminal CaM-binding and autoinhibitory regionsfrom the catalytic domain. Phosphatase activity in-creased progressively with caspase-3 digestion, cou-pled with the eventual loss of CaM-dependency. Cal-cineurin-mediated dephosphorylation of NFATc wasalso detected in thapsigargin-treated cells. Last, cal-cineurin inhibitors FK506 and cypermethrin providedpartial protection against thapsigargin-mediated apo-ptosis, suggesting that calcineurin overactivation con-tributes to thapsigargin-induced apoptosis. © 2000

Academic Press

Protein kinases and phosphatases have been im-plicated in the apoptotic signal transduction path-way in a variety of cell types. Treatment with the

1 These authors have contributed equally to this work.2 Present address: Molecular Neurogenetics Unit, Massachusetts

General Hospital, 13th Street, Bldg. 149, Charlestown, MA 02129.3 To whom correspondence should be addressed. Fax: (734) 622-

7178. E-mail: Kevin.wang@WL.com.

0003-9861/00 $35.00Copyright © 2000 by Academic PressAll rights of reproduction in any form reserved.

nonspecific protein kinase inhibitor staurosporine isa well-established apoptotic paradigm (1–3). Thespecific kinases involved in staurosporine-mediatedapoptosis have not been determined. Previously, wedemonstrated a loss in Ca21/calmodulin-dependentprotein kinase (CaMK)4 activity in SH-SY5Y humanneuroblastoma cells undergoing thapsigargin-medi-ated apoptosis. The loss in activity is accompanied byproteolytic cleavage of CaMK-IV and CaM kinase (4).Several groups have reported that inhibition ofCaMK activity is associated with the onset of apo-ptosis (5, 6).

Calcineurin is a Ca21/calmodulin (CaM)-dependentprotein phosphatase (PP2B) found in high concentra-tions in the central nervous system and is the majorphosphatase activated by CaM (for review, see 7). It ismade up of two subunits: CN-A (60 kDa) and CN-B (19kDa). These subunits are tightly bound and only dis-sociate under denaturing conditions. CN-A has somephosphatase activity on its own but requires CN-Bbinding for highest activity. CN-B shares 35% homol-ogy with CaM, and both are members of the EF-handfamily of Ca21-binding proteins. CN-A has four maindomains (from N- to C-terminal): catalytic, CN-B-bind-ing, CaM binding, and autoinhibitory domains. Metalions (Zn21 or Fe21/31) in the catalytic center enhancecalcineurin activity. Calcineurin is further activatedupon the binding of Ca21/CaM, which induces a confor-mational change that displaces the autoinhibitory do-main from the catalytic domain.

4 Abbreviations used: MTX, maitotoxin; MPT, mitochondrial per-meability transition; STS, staurosporine; BDP, breakdown product;

CaM, calmodulin; PNPP, p-nitrophenyl phosphate; CN, calcineurin.

337

m

apa

sp

ttpsmad4bo

w

dwc

rbjTmw

s

igaa

338 MUKERJEE ET AL.

Known substrates include nitric neuronal oxide syn-thase (nNOS), DARP32 and inhibitor-1 (both of theseinhibit protein phosphatase-1), the nuclear factor foractivated T-cell (NFAT), IP3 receptors, and BAD. Acti-vation of NFAT by calcineurin-mediated dephosphory-lation is required for T-cell activation and IL-2 produc-tion. The immunosuppressant drugs cyclosporin A andFK506 work by preventing T-cell activation via indi-rect calcineurin inhibition. Calcineurin specificity maybe determined by subcellular localization: it is targetedto specific subcellular sites through binding to the anti-apoptotic protein bcl-2 (8) or an protein kinase A an-chor protein, AKAP79 (9).

Calcineurin has been implicated in the signal trans-duction pathway leading to Ca21-dependent apoptosis.Inhibition of calcineurin with FK-506 or cyclosporin Aattenuated activation-induced apoptosis in T-cells (10,11) and glutamate-mediated apoptosis in cerebellargranule neurons (12). In BHK-21 cells, calcineurinoverexpression leads to apoptosis in a Ca21-dependent

anner (8).Here, we found that calcineurin A is proteolytically

ctivated by caspases in thapsigargin-mediated apo-tosis. Also, inhibition of calcineurin partially attenu-ted thapsigargin-mediated apoptosis.

MATERIALS AND METHODS

All chemicals, unless stated otherwise, were obtained from SigmaChemical Co. N-Acetyl-Leu-Leu-Met-CHO (Calpain Inhibitor II,CalpInh II), MDL28170 (carbobenzoxy-Val-Phe-CHO), thapsigargin,staurosporine, and carbobenzoxy-Asp-CH2OC(O)-2,6-dichloroben-zene (Z-D-DCB) were from Calbiochem. Anti-calcineurin A used werefrom Chemicon (AB1696), Pharmingen (65061A), and Transductionlaboratories (C26920). Purified bovine calcineurin was from Biomol.Purified, recombinant human caspase-3 was generously provided byDr. Robert Talanian (BASF). In addition, recombinant humancaspase-1, caspase-2 and caspase-7 were purchased from Pharmin-gen.

Cell culture and treatment. SH-SY5Y cells were grown on 12-wellplates to confluency (roughly two million cells/well) at 37°C, 5% CO2

in a humidified atmosphere with Dulbecco’s modified Eagle’s me-dium (DMEM) supplemented with 10% fetal bovine serum, 100unit/ml penicillin, and 100 mg/ml streptomycin. At the beginning ofthe experiment, cultures were washed three times with serum-freeDMEM. As indicated, cells were pretreated for 1 h. The cultures werethen challenged with 2 mM thapsigargin for 24 h or 0.5 mM stauro-porine for 6 h (13, 14) and maintained for indicated time, whenrotein was extracted.Protein extraction and Western blotting. Total protein was ex-

racted by lysing cells with 2% SDS/Tris buffer, precipitating pro-eins with trichloroacetic acid, and solubilizing with Tris base asreviously described (13) or by Triton X-100 extraction buffer de-cribed before (4). Protein concentration was determined with aodified Lowry method (Bio-Rad D-C protein assay kit). Equal

mounts of protein were loaded on each lane and subjected to sodiumodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE;–20% acrylamide gradient gel; Novex) with a Tris/glycine runninguffer. The separated proteins were transferred to a polyvinyl diflu-ride (PVDF) membrane (0.2 mm) by semidry electrotransfer (Bio-

Rad semidry transfer unit) for 2 h at 20 V. The blots were probed

ith primary antibody, a biotinylated secondary antibody, and avi-

in-conjugated alkaline phosphatase (Amersham). The immunoblotsere developed with nitroblue tetrazolium (NBT) and 5-bromo-4-

hloro-3-indolyl phosphate (BCIP).Biotinylated calmodulin (Biot-CaM) overlaying. Calmodulin, pu-

ified in our laboratory from bovine testis (15), was labeled withiotin (biot-CaM) as described (16). Calcineurin digests were sub-ected to SDS–PAGE (4–20%) and transferred to a PVDF membrane.he blot was blocked in Biot-CaM buffer (150 mM Tris, pH 7.4, 150M NaCl, 100 mM CaCl2, and 5% (w/v) non-fat dry milk). The blotsere incubated in the same buffer with 100 ml Biot-CaM for 1 h,

washed (3 3 10 min), and developed with BCIP/NBT. To demon-strate that the Biot-CaM would bind only in the presence of Ca21,ome blots were incubated with 1 mM EGTA instead of CaCl2.N-terminal sequencing of digested recombinant calcineurin. Pu-

rified bovine calcineurin A (5–20 mg) was digested with 2.5 mg ofmature purified recombinant caspase-3 in a mixture of 100 mMHepes buffer (pH 7.4), 10 mM dithiothreitol, 10% (v/v) glycerol, and1 mM EGTA for indicated times. The digestion was halted by theaddition of SDS-containing sample buffer for PAGE. The digestionwas halted with the addition of SDS–PAGE sample buffer. Sampleswere subjected to SDS–PAGE as above and transferred in a 3-[cy-clohexylamino]-1-propanesulfonic acid (CAPS)/methanol buffer toPVDF membrane as described (17). The membranes were stainedwith 0.1% Coomassie in 50% methanol until bands appeared. Thestained bands were excised and subjected to N-terminal sequencingby Edman degradation (in-house at Parke-Davis).

Calcineurin digestion and activity assay. Purified bovine braincalcineurin (15 mg) was digested with recombinant caspase-3 (2.5 mg)n the presence of 100 mM Tris–HCl, pH 7.5, 10 mM DTT, 10%lycerol, 100 mM EGTA in a final volume of 165 ml at room temper-ture for 2 and 5 h. Control incubation contained calcineurin in thebsence of caspase-3. Following incubation, 50 ml (0.45 mg cal-

cineurin) of the digestion mixture was assayed for calcineurin activ-ity by measuring the hydrolysis of p-nitrophenyl phosphate at 405nm on a Molecular Devices Thermomax plate reader as previouslydescribed (18). Briefly, the assay buffer consisted of 50 mM Tris–HCl,pH 7.5, 2.5 mM DTT, 20 mM p-nitrophenyl phosphate, and one of thefollowing: (i) 1 mM EDTA for measurement of background activity;(ii) 100 nM EDTA 1 1 mM Ca21 1 1 mM Mn21 for basal activity; or(iii) 100 nM EDTA 1 1 mM Ca21 1 1 mM Mn21 1 0.8 mM calmodulinfor calmodulin-stimulated activity. Since calcineurin activity is en-hanced in the presence of Mn21, therefore for ease of detection, Mn21

was included in the assay buffer for conditions ii and iii. For quan-titation, the background activity in the presence of 1 mM EDTA wassubtracted from the Ca21/Mn21 and Ca21/Mn21/CaM activity andthen plotted.

Cell death measurement. SH-SY5Y cell apoptosis (14) was as-sessed by measuring the release of the cytosolic enzyme, lactatedehydrogenase (LDH), into the culture medium (19). Quantificationof LDH release was done using the Cytotox 96 colorimetric LDHassay kit (Promega), following the manufacturer’s directions.

RESULTS

Calcineurin A Is a Caspase-3 Substrate in ApoptoticSH-SY5Y Cells

We investigated whether calcineurin, like severalother CaM-dependent enzymes (e.g., CaMKIV, CaMKkinase) (4, 20), was a caspase substrate. On Westernblot, untreated cells showed two tightly migratedbands for CN-A, probably representing the two majorCN-A isoforms (CN-Aa and CN-Ab). We also foundthat in SH-SY5Y cells undergoing thapsigargin-in-

duced apoptosis, the 60 kDa CN-A was cleaved to a

Ct

f

pi

p

339CASPASE-MEDIATED PROTEOLYSIS OF CALCINEURIN IN APOPTOSIS

roughly 45 kDa calcineurin fragment (using a C-termi-nal directed polyclonal antibody, Chemicon AB1696)(Fig. 1A). We observed that calcineurin A fragmenta-tion (16–24 h) appeared at the same time frame asPARP and CaMKIV fragmentation (see Ref. 4) andbegan significantly before cell death (36–48 h) (14)(data not shown). Similarly, in staurosporine-treated(0.5 mM) cells, there was a rapid and almost completeloss of intact calcineurin A, coupled with the formationof a 45 kDa fragment (Fig. 1B). With both apoptoticchallenges, calcineurin cleavage was completelyblocked when SH-SY5Y cells were pretreated with Z-D-DCB (Z-D) inhibitor (Figs. 1A and 1B). In addition,Z-DEVD-CHO also blocked CN-A fragmentation inthapsigargin-treated cells. Calcineurin A is also a po-tential calpain substrate (18). In cells pretreated withcalpain inhibitor (MDL28170 or calpain inhibitor II),there was no significant reduction of the 45 kDa frag-ment (Figs. 1A and 1B). Since calcineurin A degrada-tion was observed with two different apoptotic chal-lenges (thapsigargin and staurosporine), it suggeststhat the caspase-mediated calcineurin fragmentationis a general apoptotic phenomenon.

We attempted to identify the effector caspase(s) in-volved in cleaving calcineurin A in SH-SY5Y cells. Tri-ton X-100 solubilized control cell lysate was digestedwith the same amount of recombinant humancaspase-1, caspase-2, caspase-3, and caspase-7. Wefound that CN-A was not degraded significantly by

FIG. 1. Calcineurin A proteolysis during SH-SY5Y apoptosis. (A)ells were untreated or treated for 48 h with 2 mM thapsigargin in

he presence or absence of 50 mM Z-D-DCB (Z-D), 50 mM Z-DEVD-fmk, or 20 mM MDL28170 (MDL). (B) Cells were untreated or treatedor 6 h with 0.5 mM staurosporine (STS) in the presence or absence

of 50 mM Z-D-DCB (Z-D) or 20 mM calpain inhibitor II (CI). Wholecell lysate was subjected to SDS–PAGE (4–20%, 15 mg protein/lane),electrotransferred to PVDF, and subjected to western blot analysiswith anti-calcineurin A antibody (Transduction laboratories,C26920). Arrows indicate intact calcineurin (60 kDa), and the trian-gle indicates its fragment (45 kDa). Results represent at least threeexperiments.

caspase-1 or -2 (Fig. 2). In contrast, it was very sensi-

tive to caspase-3 digestion and to a lesser degree, tocaspase-7 digestion. The 45 kDa fragment of CN-Agenerated by caspase-3 was completely blocked by thepresence of 50 mM Z-D-DCB (Fig. 2).

Calcineurin A Is a Caspase-3 Substrate in Vitro andCleavage Site Identification

To confirm that calcineurin is a caspase-3 substrate,we digested purified bovine calcineurin (containingboth CN-A and CN-B) with recombinant caspase-3.Figure 3A shows the caspase-3-mediated proteolysis ofcalcineurin A (60 kDa), as characterized by three CN-A

FIG. 2. Digestion of calcineurin A in control SH-SY5Y cell lysate byvarious recombinant caspases. Control SH-SY5Y cell lysate (50 mg

rotein) was treated with 2 mg of recombinant caspase-1, -2, -3, or -7n the absence or presence of 50 mM Z-D-DCB for 4 h in 100 mM

HEPES, 10 mM DTT, 1 mM EDTA, and 10% (v/v) glycerol. Sampleswere subjected to SDS–PAGE and immunoblotted with anti-cal-cineurin A antibody (Transduction laboratories, 26920). Arrows in-dicate intact calcineurin (60 kDa), and the triangle indicates thefragment (45 kDa). Results represent at least three experiments.

FIG. 3. In vitro digestion of purified bovine calcineurin bycaspase-3. (A) Purified bovine calcineurin A (5 mg) was digested with

urified, recombinant caspase-3 (2.5 mg) for 3 h. Digests were sub-jected to Western blot analysis with three antibodies to calcineurin A(see Methods). (B) CN-A digest from (A) was subjected to SDS–PAGE, electrotransferred to PVDF, and overlaid with a biotinylatedCaM solution. The blot was then developed with avidin-conjugated

alkaline phosphatase.

340 MUKERJEE ET AL.

antibodies. We found that a 45 kDa fragment wasdetected with a N-terminal directed antibody (Chemi-con, AB1696) while a 12 kDa fragment was detected bya C-terminal-directed antibody (Pharmingen, 65061A;Fig. 3A). A third antibody (Transduction laboratories,C26920) which used the CN-B- and CaM-binding re-gions as antigen, detected both the 45 kDa and the 12kDa fragments (Fig. 3A). These data suggest that theCN-A is cleaved by caspase-3 into an N-terminal 45kDa fragment and a C-terminal 12 kDa fragment (Fig.4). We also noted that the small subunit of calcineurin(CN-B) was not sensitive to caspase-3 digestion (datanot shown). The 45 kDa fragment is identical in size tothe fragment of calcineurin formed in thapsigargin-treated cells (Fig. 1). We thus explored whether the 45kDa fragment contained the CaM binding domain byusing a biotinylated CaM overlay. The calcineurin di-gests were subjected to SDS–PAGE, transferred to aPVDF membrane, and overlaid with biotinylated CaM.This is a sensitive method for investigating the pres-ence of CaM binding proteins (16). We found that onlythe intact 60 kDa CN-A binds biotinylated CaM but notthe 45 kDa fragment (Fig. 3B). In the presence ofEGTA, no CaM binding was seen (data not shown).

We proceeded to determine the caspase-3-mediatedcalcineurin cleavage site by digesting larger quantitiesof purified calcineurin with caspase-3. N-terminal se-quencing on both the intact calcineurin A (60 kDa) andthe 45 kDa major fragment using Edman degradationyielded no signal, suggesting an N-terminal blockade(Table I). On the other hand, a 12 kDa fragmentyielded a sequence of GATAAARKEV (Table I). In fact,another smaller fragment (8–10 kDa) also yielded thisN-terminal sequence. A reference to the amino acidsequence of human calcineurin A (-a isoform) (21, 22)identified the caspase-3 cleavage site to be

FIG. 4. Linear schematic of calcineurin A. Domain organizationand calcineurin B (CN-B), calmodulin binding site (CaM), and auto-inhibitory domain (AI). The caspase-3-mediated cleavage site(DGFD*GATAA) is indicated. Caspase-3 generates two fragments ofapparent size 45 kDa and 12 kDa, from the N- and C-termini,respectively. The antigenic site for the three antibodies (Fig. 3) usedis also shown. Results are representative of at least three experi-ments.

DGFD*G386ATAA (Fig. 4). This cleavage site conforms

to the preferred DXXD caspase-3 consensus sequencefor the P4-P3-P2-P1 positions (23). Proteolysis at thissite is predicted to cleave off the regulatory CaM-bind-ing domain (residues 392–414) and autoinhibitory do-main (residues 465–487) from the catalytic domain-containing 45 kDa fragment (see schematic in Fig. 4).Consistent with this model, the intact CN-A binds bi-otinylated-CaM (biot-CaM) in the blot overlay experi-ments, and the caspase-3 generated a 45 kDa fragmentthat no longer binds biot-CaM (Fig. 3B).

Effect of Caspase-3 Digestion on Calcineurin Activity

Because the 45 kDa CN-A fragment produced bycaspase-3 contains the whole catalytic and CN-B-bind-ing domains but lacks the CaM-binding and the auto-inhibitory domains (Fig. 4), we suspected that cal-cineurin activity would increase upon caspase-3 prote-olysis. Therefore, we assayed the activity of purifiedcalcineurin in response to caspase-3 digestion. Cal-cineurin activity was measured using pNPP as thephosphatase substrate in the presence or absence ofCaM. In control conditions (with no added caspase-3),the addition of calmodulin increased calcineurin activ-ity 2-fold, as expected (Fig. 5). Upon caspase digestion(2 h), there was a significant increase in calcineurinactivity in the absence or presence of CaM. Also, thestimulatory effects by CaM was diminished (Fig. 5).After 5 h of incubation with caspase-3, both the basaland CaM-stimulated calcineurin activity were signifi-cantly elevated. The elevated phosphatase activity hasprogressively lost its sensitivity to CaM (Fig. 5). Thepresence of caspase inhibitor Z-D-DCB prevented theproteolytic activation of calcineurin (data not shown).

Calcineurin-Mediated NFATc Dephosphorylation inThapsigargin-Mediated Apoptosis

In immune cells, such as T-cells, NFATp, NFATc,and NFAT4 are the major targets of calcineurin duringT-cell activation (24); in the brain and neural cells,however, NFATp and NFATc have been shown to exist(25, 26). All major isoforms of NFAT are in a constitu-tively phosphorylated state and are endogenous sub-strates of calcineurin when the latter is activated. For

TABLE I

N-Terminal Sequence of Calcineurin Aand Its Major Fragments

Polypeptidemol. mass N-terminal sequence Origin

60 kDa Blocked Assumed N-terminal45 kDa Blocked Assumed N-terminal

12 kDa GATAAARKEV DGFD*G386ATAA

S

o

ewatP(

(o

341CASPASE-MEDIATED PROTEOLYSIS OF CALCINEURIN IN APOPTOSIS

example, when NFATc is dephosphorylated by cal-cineurin, it is then translocated into the nucleus andbinds to DNA to exert its transcriptional function. Wereasoned that if calcineurin is activated by caspase inthapsigargin-treated cells, NFATc should be dephos-phorylated. Indeed, using an NFATc isoform-specificantibody, we detected NFATc-dephosphorylation (asobserved with the increased mobility of NFATc) (Fig.6A). Importantly, the dephosphorylation of NFATc wasinhibited by Z-D-DCB (Fig. 6A). MDL28170, in con-

FIG. 5. Calcineurin activity after digestion with caspase-3. Puri-fied bovine calcineurin was digested with purified recombinantcaspase-3 for indicated time as described under Methods. Cal-cineurin activity was assayed in the presence (closed bars) andabsence (open bars) of CaM (see Methods). Results represent repli-cate experiments.

FIG. 6. NFATc dephosphorylation in thapsigargin-treated SH-Y5Y cells. Cells were untreated or treated for 48 h with 2 mM

thapsigargin in the presence or absence of (A) 50 mM Z-D-DCB (Z-D)r 20 mM MDL28170 (MDL) or (B) 100 nM cyclosporine A. Whole cell

lysate was subjected to SDS–PAGE (4–20%, 35 mg protein/lane),lectrotransferred to PVDF, and subjected to Western blot analysisith NFATc antibody. Arrows indicate phosphorylated (P-NFATc)nd dephosphorylated forms of NFATc (NFATc). We also confirmedhat this anti-NFATc could detect dephosphorylated NFATc and-NFATc in resting and PHA-activated Jurkat T-cells, respectively

data not shown). Results represent at least three experiments.

trast, had no effect (Fig. 6A). As control, we found thatCsA (1 mM) was effective in preventing dephosphory-lation of NFATc (Fig. 6B). We also noted that theoverall intensity of the NFATc band in “control” laneswas consistently higher than that in “thapsigargin”lanes. This is likely a result of the higher amount oftotal protein present in control samples (the intensityof all protein bands in control lanes was higher), de-spite our attempt to load the same amount of protein inall lanes. Another putative calcineurin substrate isBAD. Upon dephosphorylation, BAD then translocatesto mitochondria and facilitates cytochrome c release.Unfortunately, with the antibodies that we have testedthus far, we were unable to detect reliable signals ofeither BAD or phospho-BAD in SH-SY5Y cells (resultsnot shown).

Calcineurin Inhibition Protects against Thapsigargin-Mediated Apoptosis

SH-SY5Y cells were pretreated with 100 nM FK506,a potent inhibitor of the immunophilin–calcineurincomplex (10, 27), and then treated with vehicle or 2 mMthapsigargin. To rule out nonspecific immunophilin ef-fects, we also used another class of calcineurin inhibi-tor, cypermethrin, and its inactive analogue, per-methrin (28). Cypermethrin is an insecticide with di-rect anti-calcineurin activity (29). Cells were alsopretreated with the pan-caspase inhibitor Z-D-DCB.We monitored thapsigargin-induced apoptotic celldeath by assaying LDH release, as we have previouslyestablished (14). After 24 h of treatment, LDH releasewas significantly attenuated by pretreatment withboth FK506 and, to a lesser extent, cypermethrin (Fig.7). These compounds alone were not cytotoxic in SH-

FIG. 7. Effects of calcineurin inhibition on thapsigargin-mediatedapoptosis. SH-SY5Y cells were untreated (open bars) or pretreatedfor 1 h with 10 mM permethrin (hatched bars), 10 mM cypermethrindiagonally striped bars), 500 nM FK506 (horizontally striped bars),r 50 mM Z-D-DCB (solid bars) and then treated for 24 h with 2 mM

thapsigargin. Cell death was monitored by assaying LDH release.Data are mean 6 SE (*P , 0.05, **P , 0.01 vs thapsigargin-treated, Student’s t test, n 5 6).

SY5Y cells. As expected, pretreatment with Z-D-DCB,

dvLacc1

ctTdtwcdctco

ettoocwm

342 MUKERJEE ET AL.

the caspase inhibitor, reduced LDH release to controllevel.

DISCUSSION

Using the human neuroblastoma cell line, SH-SY5Y,we demonstrated that calcineurin A is activated bycaspase-3-mediated proteolysis (Figs. 1 and 6). With invitro digestion studies, we showed that caspase-3 pro-

uced a 45 kDa calcineurin A fragment that is acti-ated and is calmodulin-independent (Figs. 3 and 5).astly, the inhibition of calcineurin partially protectedgainst thapsigargin-mediated apoptosis in SH-SY5Yells (Fig. 7). This confirmed earlier reports that cal-ineurin inhibition protects against apoptosis (8, 10–2). Purified calcineurin, digested with caspase-3 in

vitro, lost its CaM-dependency and also became hyper-active (Fig. 5). Consistent with this, as previously re-ported (18), calpain-mediated proteolysis of calcineurinalso resulted in a 300% increase in p-nitrophenyl phos-phatase activity. In addition, the major calcineurinfragments produced by calpain-mediated hydrolysislost the ability to bind and be activated by CaM (18).Calcineurin A contains a number of DXXD sequences,each of which is a potential cleavage site for caspase-3(23), but we found that CN-A a was cleaved at a singlesite, DGFD*GATAAARKE (Fig. 4) (22, 30). Interest-ingly, the DXXD sequence is also conserved in CN-A b(DQFD*GSAAARKE) (31). The calcineurin regulatorydomains (CaM binding site and autoinhibitory domain)is cleaved off from the catalytic domain (Fig. 4). Theloss of CaM sensitivity in caspase-3-digested cal-cineurin is likely due to the loss of the regulatorydomains.

Recently, Shibasaki et al. (8) demonstrated that cal-ineurin is sequestered to mitochondrial membraneshrough the formation of a tight complex with BCL-2.his binding is disrupted by BAX, which translocatesuring apoptosis from the cytsolic to membrane frac-ion (32, 33), through hetero-dimerization with BCL-2ithin the cytoplasmic membrane fraction. Cal-

ineurin bound to BCL-2 is an active phosphatase butoes not have access to all of its substrates. Cal-ineurin bound to BCL-2 is unable to promote nuclearranslocation of NFAT3 (NFAT-c4). Changes in cal-ineurin localization may be an important factor in thenset of apoptosis.Another calcineurin substrate that has been report-

dly linked to apoptosis is neuronal nitric oxide syn-hase (nNOS). Calcineurin-mediated dephosphoryla-ion increases nNOS activity, and production of nitricxide (NO) has been implicated in apoptosis. The rolef NO in the apoptotic signal transduction pathway isontroversial: while NO is more commonly associatedith the onset of apoptosis (34–36), in some systems it

ay suppress apoptosis (37). Isolated mitochondria ex-

posed to NO donors release pro-apoptotic factors (34).It is conceivable in SH-SY5Y cells, increased NO pro-duction accelerates thapsigargin-induced apoptoticdeath. Most recently, two important pro-apoptotic pro-teins, BAD and caspase-9, were found to be negativelyregulated by Akt-mediated phosphorylation (38, 39).Wang et al. (40) further showed that calcineurin isresponsible for dephosphorylation of BAD during apo-ptosis. Dephosphorylated BAD is then capable ofdimerizing BCL-2 or BCL-XL, thus promoting the re-lease of cytochrome c during apoptosis. Thus, caspase-3-mediated calcineurin activation can be viewed as apositive feedback mechanism in facilitating the apopto-sis cascade.

Protein phosphorylation/dephosphorylation are im-portant regulation processes of cellular functions andhave long been implicated in apoptosis. This study,together with previous studies, suggests a complex bal-ance between kinase and phosphatase activity in neu-ronal cells undergoing apoptosis. CaMPK activity wasdecreased in SH-SY5Y cells treated with various apo-ptosis inducers (4). Furthermore, CaMK inhibition po-tentiates or promotes apoptosis (4, 41). Conversely, inagreement with previous reports (12, 42), inhibition ofthe CaM-dependent protein phosphatase calcineurinprotects SH-SY5Y cells from thapsigargin-mediatedapoptosis (Fig. 7). In contrast to CaMK activity, cal-cineurin activity increases in response to digestionwith caspase-3 (Fig. 5). The regulatory subunit of pro-tein phosphatase 2A (PP2A) is also degraded bycaspases resulting in PP2A activation during T-cellapoptosis (43). Interestingly, PP2A also negatively reg-ulates CaMK-IV activity (44). Because CaMK-IV isactivated by phosphorylation, the loss of CaMK activ-ity previously observed during apoptosis (4), may be atleast in part associated with a caspase-mediated in-crease in phosphatase activity. Thus, through proteo-lytic modifications of a number of protein kinases andphosphatases, caspases appear to play an active role inshifting the balance of protein phosphorylation statewithin a cell during the onset of apoptosis.

REFERENCES

1. Koh, J. Y., Wie, M. B., Gwag, B. J., Sensi, S. L., Canzoniero,L. M., Demaro, J., Csernansky, C., and Choi, D. W. (1995) Exp.Neurol. 135, 153–159.

2. Jacobsen, M. D., Weil, M., and Raff, M. C. (1996) J. Cell. Biol.133, 1041–1051.

3. Posmantur, R. M., McGinnis, K., Nadimpalli, R., Gilbertsen,R. B., and Wang, K. K. W. (1997) J. Neurochem. 68, 2328–2337.

4. McGinnis, K. M., Whitton, M. M., Gnegy, M. E., and Wang, K. K.(1998) J. Biol. Chem. 273, 19993–20000.

5. Tombes, R. M., Grant, S., Westin, E. H., and Krystal, G. (1995)Cell. Growth Differentiation 6, 1063–1070.

6. Kawamura, K., Grabowski, D., Krivacic, K., Hidaka, H., and

Ganapathi, R. (1996) Biochem. Pharmacol. 52, 1903–1909.

1

1

11

1

22

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

343CASPASE-MEDIATED PROTEOLYSIS OF CALCINEURIN IN APOPTOSIS

7. Klee, C. B., Ren, H., and Wang, X. (1998) J. Biol. Chem. 273,13367–13370.

8. Shibasaki, F., Kondo, E., Akagi, T., and McKeon, F. (1997) Na-ture 386, 728–731.

9. Coghlan, V. M., Perrino, B. A., Howard, M., Langeberg, L. K.,Hicks, J. B., Gallatin, W. M., and Scott, J. D. (1995) Science 267,108–111.

10. Liu, J., Farmer, J. D., Jr., Lane, W. S., Friedman, J., Weissman,I., and Schreiber, S. L. (1991) Cell 66, 807–815.

11. Fruman, D. A., Mather, P. E., Burakoff, S. J., and Bierer, B. E.(1992) Eur. J. Immunol. 22, 2513–2517.

2. Ankarcrona, M., Dypbukt, J. M., Orrenius, S., and Nicotera, P.(1996) FEBS Lett. 394, 321–324.

3. Nath, R., Raser, K. J., Stafford, D., Hajimohammadreza, I.,Posner, A., Allen, H., Talanian, R. V., Yuen, P., Gilbertsen, R. B.,and Wang, K. K. (1996) Biochem. J. 319, 683–690.

14. Nath, R., Hajimohammadreza, I., and Wang, K. K. W. (1997)Biochem. Mol. Biol. Int. 43, 197–205.

15. Dedman, J. R., Welsh, M. J., and Means, A. R. (1978) J. Biol.Chem. 253, 7515–7521.

16. Billingsley, M. L., Pennypacker, K. R., Hoover, C. G., Brigati,D. J., and Kincaid, R. L. (1985) Proc. Natl. Acad. Sci. U.S.A. 82,7585–7589.

7. Matsudaira, P. (1987) J. Biol. Chem. 262, 10035–10038.8. Wang, K. K. W., Roufogalis, B. D., and Villalobo, A. (1989)

Biochem. Cell Biol. 67, 703–711.9. Koh, J. Y., and Choi, D. W. (1987) J. Neurosci. Methods 20,

83–90.0. Wang, K. K. W. (2000) Trends Neurosci. 23, 20–26.1. Muramatsu, T., and Kincaid, R. L. (1993) Biochim. Biophys. Acta

1178, 117–120.2. Kincaid, R. L., Giri, P. R., Higuchi, S., Tamura, J., Dixon, S. C.,

Marietta, C. A., Amorese, D. A., and Martin, B. M. (1990) J. Biol.Chem. 265, 11312–11319.

3. Talanian, R. V., Quinlan, C., Trautz, S., Hackett, M. C., Mank-ovich, J. A., Banach, D., Ghayur, T., Brady, K. D., and Wong,W. W. (1997) J. Biol. Chem. 272, 9677–9682.

4. Rao, A., Luo, C., and Hogan, P. G. (1997) Annu. Rev. Immunol.15, 707–747.

5. Ho, A. M., Jain, J., Rao, A., and Hogan, P. G. (1994) J. Biol.Chem. 269, 28181–28186.

6. Ferrari, D., Stroh, C., and Schulze-Osthoff, K. (1999) J. Biol.

Chem. 274, 13205–13210.

7. Schreiber, S. L., and Crabtree, G. R. (1992) Immunol. Today 13,136–142.

8. Enan, E., and Matsumura, F. (1992) Biochem. Pharmacol. 43,1777–1784.

9. Enz, A., and Pombo-Villar, E. (1997) Biochem. Pharmacol. 54,321–323.

0. Ito, A., Hashimoto, T., Hirai, M., Takeda, T., Shuntoh, H., Kuno,T., and Tanaka, C. (1989) Biochem. Biophys. Res. Commun. 163,1492–1497.

1. Kuno, T., Takeda, T., Hirai, M., Ito, A., Mukai, H., and Tanaka,C. (1989) Biochem. Biophys. Res. Commun. 165, 1352–1358.

2. Hsu, Y. T., Wolter, K. G., and Youle, R. J. (1997) Proc. Natl.Acad. Sci. U.S.A. 94, 3668–3672.

3. McGinnis, K. M., Gnegy, M. E., and Wang, K. K. W. (1999)J. Neurochem. 72, 1899–1906.

4. Hortelano, S., Dallaporta, B., Zamzami, N., Hirsch, T., Susin,S. A., Marzo, I., Bosca, L., and Kroemer, G. (1997) FEBS Lett.410, 373–377.

5. Leist, M., Volbracht, C., Kuhnle, S., Fava, E., Ferrando-May, E.,and Nicotera, P. (1997) Mol. Med. 3, 750–764.

6. Estevez, A. G., Spear, N., Manuel, S. M., Radi, R., Henderson,C. E., Barbeito, L., and Beckman, J. S. (1998) J. Neurosci. 18,923–931.

7. Dimmeler, S., Haendeler, J., Sause, A., and Zeiher, A. M. (1998)Cell Growth Differentiation 9, 415–422.

8. Del Peso, L., Gonzalez-Garcia, M., Herrera, R., and Nunez, G.(1997) Science 278, 687–689.

9. Cardone, M. H., Roy, N., Stennicke, H. R., Salvesen, G. S.,Franke, T. F., Stanbridge, E., Frisch, S., and Reed, J. C. (1998)Science 282, 1318–1321.

0. Wang, H. G., Pathan, N., Ethell, I. M., Krajewski, S., Yamagu-chi, Y., Shibasaki, F., McKeon, F., Bobo, T., Franke, T. F., andReed, J. C. (1999) Science 284, 339–343.

1. Wright, S. C., Schellenberger, U., Ji, L., Wang, H., and Larrick,J. W. (1997) FASEB J. 11, 843–849.

2. Yardin, C., Terro, F., Lesort, M., Esclaire, F., and Hugon, J.(1998) Neuroreport 9, 2077–2080.

3. Santoro, M. F., Annand, R. R., Robertson, M. M., Peng, Y. W.,Brady, M. J., Mankovich, J. A., Hackett, M. C., Ghayur, T.,Walter, G., Wong, W. W., and Giegel, D. A. (1998) J. Biol. Chem.273, 13119–13128.

4. Westphal, R. S., Anderson, K. A., Means, A. R., and Wadzinski,

B. E. (1998) Science 280, 1258–1261.