Neuroprotection by aspirin and sodium salicylate through blockade of NF-kappaB activation

I

>2 vt <I ....>:- >2.- *. . 24vt&ft, /i . *. .VKfft.*-<t... '.A.

remnant liverremoval. Bindingreactionswere

per-formedessentially asdescribed(3, 4) with nuclear

extractsfrommouseliver cells after hepatectomy.For

STAT binding, the probe usedwas apreviously

an-nealed high-performance liquid chromatography-pu-rified double-stranded oligonucleotidefrom the

sis-induciblefactor binding element in the c-fos promoter

(5'-GATCCTCCAGCATTTCCCGTAAATCCTCCAG-3') (22) and end-labeled with

[-y-32P]adenosine

triphosphate(ATP). Supershiftexperiments were

per-formed by incubating1,ulof primary antibodywith the nuclear extracts inbinding bufferfor 1 to 2 hours at

4°C before addition of the labeled oligonucleotide.

Antibody toSTAT3 (anti-STAT3)andanti-STAT5were

fromSanta Cruz Biotechnology.

13. S. Ruff-Jamison, K. Chen, S. Cohen, Proc. Natl. Acad.Sci. U.S.A. 92, 4215 (1995).

14. N.-O. Ku,S. Michie, R. G. Oshima,M. B.Omary, J. Cell Biol. 131, 1303(1995).

15. D. E.Cressmaneta/., unpublisheddata. 16. Hepatectomized animals were anesthetized and

ventral laparotomywasperformed. Normalliver was

prepared by subjecting animalsto laparotomy

fol-lowed by perfusionasdescribed(26). One hour be-fore theremnantliverwasharvestedandfixed, ani-mals wereinjected intraperitoneallywith BrdU (50 mg/kg) (0.2% solutioninPBS) [B. Schutte,M. M. J. Reynders, F. T. Bosman, G. H. Blijham, J. Histo-chem. Cytochem. 35, 1343 (1987)]. The portal vein

wascannulatedwith a22-gauge angiocatheter, the

liver wasflushed with PBS, and4% paraformalde-hyde (pH 7.2)(4°C)wasthen perfused for 10min at a rateof 6 ml/min. The fixed liverwasremovedand cutinto 5-mm slices with a razorblade and then fixed for1hourin4%paraformaldehydeat4°C. An auto-mated tissue processor wasused to embed liver

slices withparaffin. Tissuesections(5pm)were cut onamicrotomeandadheredto

poly-L-lysine-coat-edglass slides.Stainingof fixed tissuesampleswith anantibodytoBrdU (Boehringer Mannheim) allows

one todiscern proliferating cells(brown, stained

nu-clei) from quiescent ones(clear, unstained nuclei). The immunohistochemicalstudywasperformed

es-sentially as described [S. M. Hsu, L. Raine, H.

Fanger, Am. J. Clin. Pathol. 75, 734 (1981); L. E. Greenbaum,D. E.Cressman,B. A.Haber,R.Taub, J.Clin. Invest. 96(1995)]. StatWorks and Student's

ttest,respectively,wereused for statisticalanalyses

onanimal liverweights and DNA synthesis.

17. M. D.Dabeva andD. A.Shafritz,Am.J.Pathol. 143,

1606(1993).

18. H. M. Rabes, G. Iseler,S. Czichos, H. V.Tuczek, Cancer Res.37, 1105 (1977).

19. R.Taub,in LiverRegenerationandCarcinogenesis,

R. L.Jirtle, Ed. (Academic Press, San Diego, CA, 1995),p. 71.

20. ForRNApreparation, animalswerekilled atthe

indi-catedtimes, and total liverRNApreparation, North-ernblots, and hybridizationswereperformedas de-scribed [K.L. Mohnetal., Mol. Cell. Biol. 11,381

(1991)]. For immunoblots,20

pLg

of nuclear or whole-cell extract waselectrophoresedon10 to15% SDS-polyacrylamide gels, transferred to nitrocellulose,

and detected by chemiluminescence (Amersham) accordingtotheinstructions ofthe manufacturer as

described(3,4). Primary antibodies usedinprotein immunoblots, electrophoretic gel-mobility

super-shift, and immunohistochemistry studieswere

anti-Fos andanti-cyclinDl (Santa Cruz Biotechnology), anti-p50-NF-KB1(2), anti-JunB,and anti-LRF[J.-C. Hsu, R. Bravo, R.Taub, Mol. Cell. Biol. 12, 4654

(1992)]. TheAP-1probewas adouble-stranded

oli-gonucleotide containing the consensus AP-1 site

(3'-CGCTTGATGAGTCAGCCGGAA-5')(Promega). TheNF-KBprobewas apreviouslyannealed high-performance liquid chromatography-purified dou-ble-strandedoligonucleotidefromtheclass major histocompatibility complexenhancer elementH2-KB (5'-TCGAGGGCTGGGGATTCCCCATCTC-3')(2).

21. P. Cofferetal., Oncogene 10, 985(1995); L. M.

Robertsonetal.,Neuron14,241(1995).

22. B.J.Wagneretal., EMBOJ.9,4477(1990).

23. X.Oian,U. Samadani,A.Porcella,R. H.Costa, Mol.

Cell. Biol. 15,1364(1995).

24. R. H.Diamond, D. E.Cressman,T. M.Laz, C. S.

Abrams, R.Taub, ibid. 14, 3752(1994); F. Hilberg, A.Aguzzi, N.Howells, E. F. Wagner, Nature365,

179(1993); C. Schmidtetal., ibid.373, 699 (1995). 25. B.A.Haberetal., J. Clin. Invest. 95,832(1995).

26. J. Lee etal., Hepatology19,656(1994).

27. H.Baumann,K.K. Morella, S.P.Campos,Z.Cao, G.P.Jahreis, J. Biol. Chem.267,19744(1992).

28. J. I.Daksis,R. Y.Lu,L. M.Facchini,W. W.Marhin,L.

J.Z.Penn,Oncogene 9, 3635 (1994); J. Phuchareon andT.Tokuhisa,CancerLett.92,203(1995). 29. T. Hunter and J. Pines, Cell 79, 573 (1994); D.

Resnitzky,M. Gossen,H. Bujard, S. I. Reed, Mol. Cell. Biol.14,1669 (1994);J. H.Albrecht,M. Y.Hu, F.B.Cerra,Biochem.Biophys. Res. Commun. 209,

648(1995).

30. J. Deviere et al., Clin. Exp. Immunol. 77, 221

(1989);A. M.Gressner, KidneyInt.49, S-39 (1996); C.McClain,D.Hill, J.Schmidt,A. M.Diehl,Semin. Liver Dis. 13,170(1993);H.Tilgetal., Gastroen-terology 103,264(1992);P.Greenwel,J.Rubin,M.

Schwartz,E. L.Hertzberg,M.Rojkind,Lab. Invest.

69,210

(1993).

31. J.Baueret

al.,

Blood

72,

1134

(1988);

T.Kishimoto, S. Akira, T.

Taga,

Science 258, 593

(1992);

Y.

Yamada, I.

Kirillova,

J. J.

Peschor,

N.

Fausto,

in

preparation.

32. Wethank J.Darnell for the STAT3cDNA,U. Muller

Eberhard and S. Maeda for the

hemopexin

(HPX)

and serum

amyloid

P-component (SAP) probes,

andC. Steerfor the

gift

of the

cyclin

Dl cDNA and

helpful

discussions. We also thankC.

Deutsch-mann, R.Diamond,D.Tewari,P.Traber,M.Lazar, andF.Rauscherfor

helpful discussions;

S.

Hwang

andV. Miles for technicalassistance;and J. Mat-thewsfor assistancewith

manuscript

preparation.

Thiswork was in part

supported by

NIH

grants

DK44237,

DK49210,

andDK49629

(to R.T.);

NIH

grant K08

(to

L.E.G.);

the

University

of

Pennsylva-nia Genetics

Training

Grant

(to

R.A.D.);

andthe

Howard

Hughes

Medical Institute.

17July 1996;

accepted

1October1996

Neuroprotection by Aspirin

and

Sodium

Salicylate Through

Blockade of NF-KB Activation

Mariagrazia

Grilli,*

Marina

Pizzi,

Maurizio

Memo,

PierFranco Spano

Aspirin

(acetylsalicylic acid)

is

a

commonly prescribed drug

with

a

wide

pharmacological

spectrum.

At

concentrations

compatible

with amounts in

plasma during

chronic

anti-inflammatory

therapy, acetylsalicylic

acid and its metabolite sodium

salicylate

were

found to be

protective against neurotoxicity

elicited

by

the

excitatory

amino acid

glu-tamate in rat

primary neuronal cultures

and

hippocampal

slices. The site of action of the

drugs

appeared

to be downstream of

glutamate receptors

and to involve

specific

inhi-bition of

glutamate-mediated

induction of nuclear factor

kappa

B.

These results

may

contribute to

the

emerging

theme of

anti-inflammatory drugs

and

neurodegeneration.

Glutamate

is

the most abundant excitatory

neurotransmitter in

the

brain; however,

un-der

certain

conditions,

it

may

become

a

potent

excitotoxin

and contribute

to

neu-rodegeneration (1).

On

the other

hand,

an

accumulation

of

clinical and

experimental

evidence suggests that

neurodegeneration

is

often associated with

inflammation

(2).

We

tested

the

possibility that the

anti-inflam-matory

drugs aspirin

[acetylsalicylic acid

(ASA)]

and sodium

salicylate (NaSal),

be-cause

of

their wide spectrum of

pharmaco-logical

activities

and

multiple

sites of action

(3), may confer neuroprotective properties.

Several models

of neurons in

culture

have been

used to

unravel the

molecular

events

triggered by glutamate that lead

to

cell

death

as

well

as to

develop

pharmaco-logical compounds able

to counteract

exci-totoxicity.

Here we

used

primary cultures of

rat

cerebellar

granule cells, where

a brief

pulse of glutamate, through

activation

of

DivisionofPharmacology,DepartmentofBiomedical

Sci-encesandBiotechnologies, Universityof Brescia Medical

School,Brescia,1-25123Italy.

*To whomcorrespondenceshould be addressed. E-mail:

mpharm@master.cci.unibs.it

the

N-methyl-D-aspartate

(NMDA)

type

of

glutamate

receptor,

induces cell death

(4).

ASA

and NaSal

were

added

to

the culture

medium

5

min

before

and

during

a

15-min

application

of

50

pM

glutamate

(5),

a

con-centration

that reduced cell survival

by

70

to

80%.

The

range

of

concentrations

for

both

drugs

was

correlated with the

amounts

in

plasma (1

to

3

mM)

for

optimal

anti-inflammatory

effects

in

patients

with

rheu-matic

diseases

(3).

A

concentration-depen-dent

protection against

glutamate-induced

neurotoxicity

was

observed

in

the

presence

of

both

drugs (Fig. 1A).

For

ASA,

the

calculated median effective

concentration

(EC50)

was

1.7

mM, with maximal

effect

(83%

protection)

at

3 mM.

The

concentra-tion

of

NaSal

giving 50%

protection

was

5

mM,

and maximal

response

(87%

protec-tion)

was

observed

at

10 mM.

Unlike

sa-licylates,

at concentrations

compatible

with

the

plasma

levels

during

chronic

drug

treat-ment

(1

to

20

,uM) (3),

the

anti-inflamma-tory

drug

indomethacin

was

unable

to

pre-vent

glutamate-induced

cell

death

(6).

Neuroprotection

was

also

evaluated

in

hippocampal

slices of

8-day-old

rat

brain

SCIENCE * VOL.

274

* 22 NOVEMBER 1996 1383

on October 29, 2012

www.sciencemag.org

(7),

a system

that

more

closely

represents in

vivo

conditions.

In

the

hippocampal slices,

most

pyramidal

neurons

of

CAI

and

CA3

and granule cells of dentate

gyrus

(DG)

be-came

acutely

necrotic,

exhibiting swollen

cytoplasm with

large

vacuoles,

nuclear

shrinkage, and focal

clumping

of chromatin

(Fig.

IB).

Application

of

ASA

preserved

hippocampal cell viability

from

the

NMDA-mediated

injury

(Fig.

1, B

and

C).

ASA

did

not

modify cell viability

at

1

mM,

but

at

3

mM

it

specifically produced

significant

neu-roprotection in

the

CA3

region

(Fig. IC).

Higher

concentrations

of

ASA

completely

inhibited

the

NMDA

effect

in

CAI

and

DG

as

well

as

in

CA3

cells (Fig.

1, B

and

C).

Compared with

primary

cultures of

rat

cere-bellar

granule

cells,

2 mM

NaSal

efficiently

counteracted NMDA-mediated

toxicity

in

hippocampal slices (Fig. IC).

To

dissect the molecular mechanisms

by

which

salicylates preserved cell

viability

against

excitoxicity,

we

tested

whether these

drugs diminished

glutamate-mediated

calci-um

entry

(8).

In rat

cerebellar

granule cells,

application

of

glutamate

in

the

absence

of

external

Mg2+

caused

a

rapid

increase in

the

A

C .o 0 0

0.

0. z

C

intracellular Ca2+ concentration

([Ca2+]i)

followed by

a

sustained

plateau (Fig. 2A),

principally

because of the

NMDA

receptor

subtype

activation

(9). ASA, applied

at

neu-roprotective concentrations

(1 to 3

mM),

induced

a

very

low

and

short-lasting

[Ca2+]i

increase

and

did not

modify

glutamate-me-diated calcium

entry

(Fig. 2B). Similar

re-sults

were

obtained with

NaSal (9). Thus,

it

was

likely that

salicylates

were

acting on

intracellular

molecular

targets

further

down-stream

of

glutamate

receptor

activation, a

property

that makes them

distinguishable

from

most

neuroprotective

drugs.

It

also

ap-pears

that neuroprotection occurred

inde-pendently of mechanisms controlling

[Ca

2

1

homeostasis.

At

plasma

concentrations

maintained

dur-ing treatment

of

chronic

inflammatory

diseas-es,

ASA

and

NaSal, but

not

indomethacin,

inhibit the

activation

of nuclear factor

kappa

B

(NF-KB)/Rel

transcription

factors

in T

and

pre-B

lymphocytes (10). The NF-KB/Rel

family

is

implicated

in

controlling

expression

of

several

genes

crucially involved

in

immune

and

inflammatory function (11). NF-KB/Rel

proteins are present in primary neurons

and

in

B

CAl

several

brain areas

(12). Administration of

glutamate to primary cultures of rat cerebellar

granule cells also results

in

up-regulation

of

NF-KB

nuclear

activity

(13)

and of the

tran-scriptional complex AP-1 (14). Cells were

exposed

to

50

F.M glutamate

in

the absence

or

presence

of ASA (1 or 3

mM)

and NaSal

(3

or

10

mM),

and nuclear extracts

(15)

were

prepared 1 hour after

stimulation. Both

drugs

10C 8C _ 6C -I + 4C 2C o0

o0-)O

0* 10-0 II

10001

800-2 c 6C -T + 4C 2C 10-JU CA3 DG

A

Glutamate 1 2 3 4 5I6 1 2 3 4 5 6

B

ASA Glutamate 1 2 3 4 Time(min) 5 6

Fig.

2.

Original recording showing the

glutamate-induced

[Ca2-]j

increase

in rat

cerebellar granule

cells.

(A) Effect of 50 ,uM glutamate (n

=

95). (B)

Control Effect of 50

F.M glutamate

inneurons

pretreated

with 3 mM

ASA (n

=

98). Traces

are

from

repre-sentative cell

recordings.

Drug concentration(mM) l... 0 | CAI I0- A NMDA ASA(mM) - - 3 NaSal(mM) - 3 10 Glutamate I_ I_

_I_

(50 pM) C NMDA +ASA NMDA 30pM -ASA 1 mM-ASA 3mM-ASA 5mM ASA 10 mM-NaSal 2 mM

Fig.

1.

Neuroprotection

bysalicylates. (A)

Concentration-dependent

effectelicited by ASA and NaSal in ratcerebellargranule cells. A glutamate(50

F.M)

pulse was applied in the absence or presence of ASA

(-)

and

NaSal

(m).

Neuronal survival was expressed as percent of neuroprotection, with glutamate inducing 78 +

3%

ofcell loss. Thexaxis represents drug concentrations. Points represent the means+

SEM

of six

experiments,

runin

triplicate,

ondifferentculture preparations. (B) Prevention of excitotoxic effect of NMDA in rat

hippocampal

slicesby

ASA. Sections

wereexposed to vehicle (control), 30 ,uM NMDA

(NMDA),

or30

p.M

NMDA and 5 mMASA (NMDA +ASA). Cell viability was evaluated in CA1,

CA3,

and

DG.

Scale

bar,

10

pLm.

(C)

Effect of ASA and NaSal on NMDA-induced cell loss in rat

hippocampal

slices. Test drugs were added to the slices at the indicated concentrations and cell viability in

CA1,

CA3,

and

DG

wasanalyzed.

Columns

represent the means ± SEM of three experiments run on four slices each.

Differences

compared

with NMDAaloneweresignificant at P<0.01 asindicatedby anasterisk.

Fig.

3. Effect of

neuroprotective

concentrations of

ASA

andNaSalon

glutamate-induced

NF-KB and AP-1 DNA

binding

activities. Nuclear extracts fromratcerebellar

granule

cellswere

subjected

to an

electrophoretic mobility-shift

assay with

-y_32p-labeled

oligonucleotide probes containing

the

im-munoglobulin

KB

(lanes

1 to

6)

and the AP-1 DNA

binding

sites

(lanes

7 to

12) (15).

Cellswereeither unstimulated

(lanes

1and

7)

orstimulated with 50

,uM glutamate

(1

5-min

pulse)

in the absence

(lanes

2and

8)

orpresence

(lanes

3to6and 9to

12)

of the

drugs

asindicated. SCIENCE * VOL. 274 * 22 NOVEMBER 1996

rsn 1384 3 3 0 . . I I . A. I A kA 71 I I

on October 29, 2012

www.sciencemag.org

Downloaded from

inhibited the glutamate-induced

increase

of

NF-KB

activity in a

concentration-dependent

manner

(Fig.

3),

with calculated EC50 values

of 1.3

and

6 mM

for ASA

and NaSal,

respec-tively. Parallel

experiments

in

which cell

vi-ability

was

measured

24

hours

later revealed

a

strict

correlation between

neuroprotective

concentrations of

anti-inflammatory

drugs

and

blockade of

NF-KB

induction

(EC50

val-ues

of 1.5 mM for ASA

and

5.8 mM

for

NaSal). The

salicylate effect

on

NF-KB/Rel

proteins was

specific.

In

fact,

ASA

and NaSal

failed

to

modify the

glutamate-mediated

nu-clear induction of

the

transcriptional complex

AP-1

(Fig.

3).

Thus,

at concentrations

compatible with

amounts

in

plasma during

treatment

of

chronic inflammatory

states,

salicylates

pre-vented

glutamate-induced

neurotoxicity.

The neuroprotective

effect

correlated

nei-ther

with the anti-inflammatory

properties

of

these

compounds

nor

with

cyclooxygen-ase

inhibition.

In

fact, indomethacin

exert-ed anti-inflammatory but

not

neuroprotec-tive

properties,

and NaSal

was

neuroprotec-tive

but did

not

interfere

with

cyclooxygen-ase

activity

(3). The

common

molecular

target

for

ASA

and NaSal but

not

for

in-domethacin

(10, 16)

was

the blockade

of

NF-KB

induction,

suggesting a

link between

neuroprotection

and

the nuclear

event.

Here

we

provide evidence

for

an

unusual

pharmacological

effect of ASA

and

its

me-tabolite NaSal.

In view

of

their distinct

ability

to

act

not

merely

as

anti-inflamma-tory

compounds but also

as

neuroprotective

agents against excitotoxicity,

these

drugs

appear to

possess

a

wider

pharmacological

spectrum

than other nonsteroidal

anti-in-flammatory

drugs.

REFERENCES AND NOTES

1. S. A. LiptonandP. A.Rosenberg,N. Engl. J. Med.

330,613(1995);M.Memo, M.Pizzi,A.Valerio,M.

Grilli,P.F.Spano,Int.Rev.Psychiatry 7,339(1995).

2. P. L.McGeerandE. G.McGeer, BrainRes. Rev. 21, 195(1995).

3. P.Insel,inGoodman and Gilman's The Pharmaco-logicalBasisof TherapeuticsJ.G.Hardman, P. B.

Molinoff,R. W. Rudden,A.Goodman Gilman, Eds. (McGraw-Hill,NewYork, 1996),pp.617-657;K.D.

Rainsford, in Aspirin and the Salicylates (Butter-worths,London, 1984); G. Weismann,Sci.Am. 84,

264(January 1991); J.P.Famaey andH.E.Paulus, Eds., Therapeutic Applications ofNSAIDS

Subpopu-lations and NewFormulations (Dekker, NewYork, 1992).

4. V. Gallo, M. T.Ciotti, F. Coletti, F.Aloisi, G. Levi, Proc. Natl. Acad.Sci. U.S.A. 79,7919(1982); M.

Favaronetal., ibid. 85, 7351 (1988).

5. Primaryculturesof cerebellargranulecells were

pre-pared from cerebella of8-day-old rats (Sprague-Dawley)killedaccordingtothePolicyon theUseof Animals in Neuroscience Research. The cultures wereused at 10 to 12daysinvitro and contained morethan95%glutamatergicneurons.

Neurotoxic-itywasinduced asdescribed[M. Pizzi, C.Fallacara, V.Arrighi,M.Memo,P.F.Spano,J.Neurochem.61, 683(1993)]. Drugstested forneuroprotectionwere added 5 min before andduringtheglutamate pulse. Cellsurvival was evaluated 24 hoursafterthe

gluta-mate pulse asdescribed by K. H. Jonesand J. A. Senft [J. Histochem. Cytochem. 33,77(1985)].ASA, NaSal, and indomethacindid notaffectneuronal

vi-abilityper se.

6. M. Grilli,M. Pizzi, M.Memo,P. F.Spano, datanot

shown.

7. Hippocampal slices wereobtained from brains of

8-day-oldratsandpreparedasdescribed[G. Garth-waite and J. Garthwaite, Neurosci. Lett. 97, 316

(1989)].Sliceswerepreincubatedwith either the test drugs,ASAandNaSal, orvehiclefor 30min, and

then NMDA (30pM)wasadded foranadditional30 min.Sliceswerewashedandincubated in fresh

buff-erfor 90minandthen fixedand embeddedinepoxy

resin. Semithin (1 ,um) sectionswere cut, stained

withmethylene blueand azur11, andexamined by

lightmicroscopy. For quantitation ofcell loss,cells werecounted in cell layer fields takenfrom CA1, CA3,andthe dorsalbladeoftheDGineachslice.

The fields measured1.5x 104 mm2. The

percent-ageof cell survivalwascalculated bythe ratio ofliving

cells to thetotal number of cells.

8. D. W.Choi,J.Neurosci.7,369(1987).

9. The[Ca2

2]i

wasmeasuredbyFura-2 microfluorim-etry insinglecellsasdescribed byM. Pizzi et al.[Mol. Pharmacol. 49, 586(1996)]. Cellswereexposedto

glutamatefor 2 min inMg2+-free Krebs-Ringer

solu-tion.ASAorNaSalwasadded2 minbefore

gluta-mateexposure. Fluorescenceimageacquisitionand

analysiswereperformed by usingtheMIRAcal (mul-tiple image ratioing and analysiswithcalibration) sys-temfromApplied Imaging(England).LikeASA,

Na-Sal (2 to 10 mM) produced a low and transient [Ca2

]i

increasewithoutaffecting glutamate-mediat-ed calciumentry.

10. E. KoppandS. Ghosh,Science265, 956(1994).

11. H.C.Liou and D.Baltimore,Curr. Opin.CellBiol. 5,

477(1993); M. Grilli,J.-S.Chiu,M. J.Lenardo,Int. Rev.Cytol. 143,1 (1993).

12.C. Kaltschmidt, B. Kaltschmidt, P. A. Baeuerle,

Mech. Dev. 43, 135 (1993); C. Kaltschmidt, B.

Kaltschmidt,H. Neumann,H.Wekerle, P.A.

Bae-uerle,Mol. Cell. Biol. 14, 3981(1994);M.Grillietal.,

J.Biol.Chem. 270,26774(1995); S.W.Barger and

M. P.Mattson,Mol. Brain Res. 40, 116(1996).

13. C.Kaltschmidt,B.Kaltschmidt,P. A.Baeuerle,Proc.

Natl.Acad.Sci. U.S.A. 92,9618(1995);L.Guerrini,

F.Blasi,S. Denis-Donini, ibid.,p.9077;M.Grilli,F.

Goffi,M. Memo, P. F.Spano, J. Biol. Chem. 271,

15002(1996).

14. T.Curran and B. R.Franza, Cell55,395(1988).

15. Nuclear extracts were prepared according to a

small-scale protocol described by N. C.Andrews

andD. V.Faller[NucleicAcids Res.19, 2499(199

1)1.

DNA bindingreactionswereperformedasdescribed [S.-M. Kang, A.-C. Tran, M. Grilli, M. J. Lenardo, Science 256,1452(1992)]withthefollowing

modifi-cations: 4

pg

of nuclearextracts wascombinedwith

50,000cpm(0.2ng) ofy-32P-labeled oligonucleo-tidesin atotal volumeof 12

Rl.

Oligonucleotide

se-quences were as follows: KB oligonucleotide

se-quencefrom theimmunoglobulinKlight-chain gene, 5'-CAGAGGGGACTTTCCGAGAGGC-3'; AP-1

oli-gonucleotide sequence, 5'-CGCTTGATGAGTCA-GCCGGAA-3'.

16. Indomethacin(1to20

pLM)

wastestedfor theability

tointerfere withglutamate-induced NF-KB activation

in cerebellar granule cells. No inhibition was

ob-served.M. Grilli,M. Pizzi, M.Memo, P. F. Spano, unpublished material.

17. Thiswork waspartiallysupported byConsiglio Na-zionaledelleRicerche, Italy.

26 June1996;accepted25September1996

Tricorn

Protease-The

Core

of

a

Modular

Proteolytic System

Tomohiro Tamura, Noriko Tamura, Zdenka

Cejka,

Reiner Hegerl,

Friedrich

Lottspeich,

Wolfgang Baumeister*

Large macromolecular assemblies have evolved

as a means

of

compartmentalizing

reactions in

organisms lacking

membrane-bounded compartments. A tricorn-shaped

protease

was

isolated from the archaeon Thermoplasma and

was

shown

to

form

a

multisubunit

proteolytic complex. The 120-kilodalton

monomer

assembled

to

form

a

hexameric toroid that

could

assemble further into

a

capsid

structure.

Tricorn

protease

appeared

to act

as

the

core

of

a

proteolytic system; when it interacted with several

smaller

proteins,

it

displayed multicatalytic activities.

In

vivo

proteolysis

is an

essential

element

of

many

regulatory

processes. It must

be

subject

to

spatial and temporal control

in

order

to

prevent

damage

to

the

cell.

Pro-karyotic cells, which lack

membrane-bounded

compartments,

have developed

large macromolecular assemblies

or

"molec-ular

organelles"

so as to

confine

proteolysis

to an

inner

cavity to

which

only

proteins

targeted

for

degradation

have access.

The

paradigm

of

such

a

proteolytic

complex

is

the

proteasome

(1),

which

is

ubiquitous

across

the three

urkingdoms

archaea

(2),

bacteria

(3, 4)

and

eukarya (5).

In

the

Max-Planck-lnstitute

forBiochemistry,D-82152

Martins-ried,Germany.

*To whomcorrespondenceshould beaddressed.

course

of

searching

for

regulatory

compo-nents of

the

proteasome (6)

in

Thenro-plasma acidophilum,

we

discovered

a

proteo-lytic complex

of

high molecular

mass

that

is

not

related

to

the

proteasome.

This

com-plex

seems

to

be the

core

of

a

modular

proteolytic

system

generating

multicatalytic

activities.

We

purified

the

high-molecular-weight

(HMW)

protein

to

homogeneity by

a

se-quence of

chromatography

steps

(7).

The

purified

protein migrates at 720

kD in

gel

filtration

chromatography (versus

migration

at 680

kD

by

the 20S

proteasome), and

it

turned

out to

be composed

of

a

single

polypeptide

of 120 kD

when

subjected to

SDS-polyacrylamide

gel

electrophoresis

(SDS-PAGE) (Fig. 1).

The

purified

protein

SCIENCE * VOL.

274

* 22

NOVEMBER

1996 1385

on October 29, 2012

www.sciencemag.org