Human Ro ribonucleoprotein particles: characterization of native structure and stable association with the La polypeptide

Human Ro ribonucleoprotein particles:

characterization of native structure and stable

association with the La polypeptide.

G Boire, J Craft

J Clin Invest.

1990;

85(4)

:1182-1190.

https://doi.org/10.1172/JCI114551

.

Anti-Ro autoantibodies, found in sera of patients with systemic lupus erythematosus,

Sjogren's syndrome, and related diseases, target the Ro ribonucleoprotein particles

(RNPs). Although the polypeptide and RNA components of the Ro RNPs have been

characterized, much less is known about the native structure of these particles. We have

now characterized by biochemical techniques intact Ro ribonucleoprotein particles from

cultured HeLa cells. These particles segregated in three discrete subpopulations with

characteristic physicochemical properties: one containing hY5 RNA (RohY5 particles), one

containing only hY4 RNA (RohY4 particles) and one with hY1, hY3, and hY4 RNAs

(RohY1-hY4 particles). The RohY5 particles were purified free of contaminating

ribonucleoproteins; both the La and the 60-kD Ro polypeptides were stable components of

this portion of the Ro RNPs. The La RNPs co-purified with the RohY4 particles and

contaminated the RohY1-hY4 RNPs. The stable association between the La and the 60-kD

Ro polypeptides provides a potential macromolecular target for the linked set of anti-Ro and

anti-La antibodies, and suggests a possible functional association of these polypeptides.

Research Article

Human Ro Ribonucleoprotein

Particles:

Characterization

of

Native Structure

and

Stable

Association with the La Polypeptide

Gilles Boire and Joe Craft

Section ofRheumatology, Department ofMedicine, YaleUniversity, New Haven, Connecticut06510

Abstract

Anti-Ro autoantibodies, found in sera of patients with systemic lupus erythematosus,

Sjogren's

syndrome, and related dis-eases, target the Ro ribonucleoprotein particles (RNPs). Al-though the polypeptide and RNA components of the Ro RNPs havebeencharacterized,much less is known about the native structure of these particles. We have now characterized by

biochemical techniques intact Ro ribonucleoprotein particles fromcultured HeLa cells. These particles segregated in three

discrete subpopulations with characteristic physicochemical

properties: one containing hY5 RNA (RohYs particles), one

containingonly hY4 RNA(Ro Y4 particles) and one with hYl, hY3,and hY4 RNAs(RohYIhY4particles). TheRohYsparticles were purified free ofcontaminating ribonucleoproteins; both the La and the60-kDRopolypeptideswerestablecomponents ofthisportionof the Ro RNPs. The La RNPsco-purifiedwith the

RobY

particlesandcontaminatedthe

RoM`YIhY4

RNPs. The stableassociation between the La and the60-kDRo

polypep-tidesprovidesapotentialmacromolecular targetforthelinked

setof anti-Roandanti-Laantibodies,and suggests a possible

functionalassociation ofthese polypeptides. (J. Clin. Invest.

1990.

85:1182-1190.)

autoantibodies* anti-Ro antibodies *

ri-bonucleoproteins-Roprotein-Laprotein

Introduction

Autoimmune diseases such assystemic lupus erythematosus (SLE)arecharacterized

by

autoantibodiesin

patient

serathat bind

highly

conserved

autoantigens

(reviewed

in

1-3).

The spectrum of this autoimmune response is often restricted in any

given

patient

(3,4), with certain

autoantibody specificities

that correlate with clinical manifestations

(reviewed

in 1-3). Thusmanyeffortshave been made todefinethestructuralor

biological features of

autoantigens

that

might

nominate them astargets ofautoimmunity (for examples,see5-8).

Anti-Ro antibodiesare aprominent feature ofthehumoral autoimmune response in many

patients

with SLE and with

Sjogren's

syndrome(9, 10).Theseantibodiestargetsmall

ribo-Publishedin abstractform in 1989.(ArthritisRheum. 32:S1 18.) Address reprint requests to Dr. Craft, Department of Medicine,

YaleUniversity School of Medicine,Box3333, 609 LCI, New Haven, CT065 10.

Dr.Boire's permanent addressisDivisionof Rheumatology, De-partment of Medicine, Centre Hospitalier Universitaire de

Sher-brooke, 3001-12th Ave. North, Sherbrooke, Quebec, Canada J1H

5N4.

Receivedforpublication 18 July 1989 and in revisedform 27 No-vember 1989.

nucleoprotein particles (RNPs)' that consist ofa 60-kD Ro polypeptide associated with 2-4 small RNAs (1 1, 12). Recent evidence also indicates thata52-kD polypeptide, anda54-kD polypeptide restrictedtoerythrocytes, arecomplexed with the Ro RNAs (13, 14). The number of RNAs differs among mammalian cells: for example, human cells contain four Ro RNAs, called hY1, hY3, hY4, and hY5, and murine cells containtwosuchRNAs, mY 1 and mY2 (15). These RNAsare

RNApolymerase IIItranscription products, and as such, are at least transiently associated with the La polypeptide (1 1). The 60-kD Ro polypeptide binds its Ro RNA at the base ofa

double-stranded stem formed by base-pairing of 5' and 3' ends (12); the binding site of the La polypeptide is believed to be at leastpartiallyconstituted by a poly (U) stretch at the 3' end of the RNA that is processedtoformthe mature Ro RNA tran-script with release of this polypeptide (16-18). The RNA

and/

orproteinbinding sites for the 52- and 54-kD Ro polypeptides have yet to be defined.

Indirect evidence indicates that one 60-kD Ro polypeptide binds to one Ro RNA (12), implying the existence of distinct populations of Ro RNPs. We recently reported the frequent

occurrenceinpatient seraof antibodies that targetedan anti-genic epitope restricted to intact Ro particles containing the 60-kD Ro polypeptide and the hY5 RNA

(Ro'Ys

RNPs) (19). The production of antibodies to a specific Ro RNP suggested structuralheterogeneityof theseparticles;moreover, the pro-duction of antibodies that only recognized intact RohY5 parti-cles indicated that native Ro RNPs might serve as autoim-munogens, inadditionto(orratherthan) isolatedRo

polypep-tides.

Thus, to determine if Ro particles are indeed physically

heterogeneous, and to identifystructural features that could possibly designate a subset of thesenativeRoparticles as

tar-gets of aspecific autoimmune response,wehave biochemically purifiedintact human Ro particlescontainingthe 60-kD Ro

polypeptide and the Ro RNAs. These particles segregatedin

three subpopulations with characteristic physicochemical

properties.The Lapolypeptide, frequently targeted bythe

au-toimmune response inconjunctionwiththeRopolypeptides,

wasstablyassociatedwith aportionof thesepurified particles.

Methods

Cells andsera.Human HeLacells,initiallyobtained fromAmerican Type Culture Collection(Rockville, MD),weremaintainedat37°C

under 5% C02, in RPMI 1640(Gibco Laboratories, Grand Island,

NY)supplemented with 10% heat-inactivated fetal calfserum,60 ,g of penicillin/mland100ggofstreptomycin/ml.Serawereobtainedfrom healthy laboratory workersandfromAmerican and French Canadian

patientswithvarious connective tissue diseases. Control anti-Rosera

1.Abbreviations usedinthis paper: RNP,ribonucleoproteinparticles;

TBS,Tris-buffered saline;TSE,Tris-Cl,NaCl, andEDTA. J.Clin.Invest.

©)TheAmerican SocietyforClinicalInvestigation,Inc.

0021-9738/90/04/1182/09 $2.00

were defined as those which immunoprecipitated the 60-kD Ro

poly-peptide and all fourhY RNAsfromradiolabeled HeLa cell extracts

(19),whileanti-RohY5sera immunoprecipitated only the 60-kD

poly-peptide andthe hY5 RNAfromidentical extracts (19). The anti-Ro sera useddid notimmunoprecipitate the LaRNAs from HeLa cell extracts norrecognize the 50-kDLa polypeptide in immunoblots. As

previously reported(12), a protein that comigrated with the La

poly-peptidewasfaintly visible in[35S]methioninelabeledanti-Ro

immu-noprecipitates.Noneofthe sera used in these studies

immunoprecip-itatedthe 52-kD Ropolypeptide fromHeLacell extracts, nor

recog-nized this proteininimmunoblotsoftheseextracts(13).

Preparation and immunoprecipitation ofcell extracts. For analysis ofinvivolabeled RNAs, HeLa cells wereradiolabeledwith

[32P]or-thophosphate (10 uCi/mlcells; Amersham Corp., Arlington Heights, IL)for14 h aspreviously described(20). Cells were collected by

centrif-ugation, washed in TBS (10 mM Tris-Cl[7.5], 150mMNaCI), and

sonicated in NET-2 buffer (50 mM Tris-Cl[7.5], 150mMNaCI,0.05%

Nonidet P-40). Immunoprecipitation of radiolabeledand unlabeled

cell extracts was performed as described (19). Immunoprecipitated

32P-labelednucleic acidswerevisualized by autoradiography on XRP

film(Eastman Kodak, Rochester, NY) while the unlabeled RNAs were

visualizedbysilver staining (21). For analysis of proteins, HeLa cells were labeled with[35S]methionine(10 zCi/mlcells; Amersham Corp.)

andimmunoprecipitations performedasdescribed(19).

Inexperiments to assess thestability oftheassociationof hY RNAs

withthe60-kDRopolypeptide,4MNaClwas added to parallel sam-plesof theHeLa cellsonicatestoincrease theirsaltconcentrationto

0.25, 0.35, 0.50,0.75, and 1.00 M, respectively. After rotation for1 h at

4°C,thesesamples with high salt concentrationswereadded to anti-body-coated Sepharose beads resuspended in NET-2 bufferthathad beenpreviously modifiedtocontainthe sameNaCIconcentrationsas thesonicates.

Biochemical purification of theRoparticles.The first steps of the purification ofthe Roparticleswereperformedasdescribed ( 19), with modifications. Inbrief, 1-2X109HeLacellswerecollectedby centrif-ugation and washed twice in chilled TBS. Subsequentlyallprocedures

wereperformedat4°Cand allbuffersweresupplemented withI mM dithiothreitol and 1 mMphenylmethylsulfonylfluoride. A cytoplas-micfraction wasprepared without detergent: cellswereallowed to

swell in sixpellet-volumes of bufferA(10mMTris-CI[7.5], 1.5 mM

MgCl2,10mMKCI)for 10 minonice,spundown,resuspendedintwo

pellet-volumes of bufferAanddisruptedby 10 strokes ofanall-glass

Douncehomogenizer; 100 U of RNAsin (PromegaBiotec,Madison,

WI)were addedafterdisruption ofthecells.Extractswerethen clari-fied bycentrifugationat13,000gfor20 min.The salt concentration of thesupernatantwasincreasedto140mMby adding0.1 vol of bufferB

(0.3 MTris-Cl [7.5], 1.4 M NaCl, 1.5 mM MgCl2).Theextract was

thenlayeredover a7.5-mlsucrosecushion(800mM sucrose, 10mM

Tris-Cl[7.5],5mM MgCI2) in 30-mltubesandcentrifugedat26,000

rpminaSW28rotor(Beckman Instruments, Inc., PaloAlto,CA)for 90min. The supernatantwascollectedanddiluted 1:1 with TSE(50

mM Tris-CI[7.5], 150 mM NaCl, 1 mM EDTA)andapplied to a

40-mlcolumn of DE52(Whatman, Maidstone, Kent,England)

pre-equilibratedinTSE. After extensivewashingswithTSE,stepelutions ofthecolumnusingTSEcontaining 175, 190, 210, 250,and 300mM

NaClwereperformed.Afteraddition of 40UofRNAsin,each eluate

wasseparately concentrated byvacuumdialysis (Pro-Di-Con;

Bio-Mo-lecularDynamics, Beaverton,OR) againstTSE. Elutionprofilesof the

Ro RNPsin eachof the eluted fractionsweredeterminedby

immuno-precipitationwith control anti-Rosera.

Theconcentrated DE52eluates(0.5 ml)containingthedifferent Roparticleswereeachlayeredon a10-30%(wt/wt)sucrosegradient

(10.5 ml)andcentrifugedat35,000rpm inanSW41rotor(Beckman

Instruments, Inc.) for 18h.Aftercentrifugation, serial0.5-ml fractions

werecollected fromthe bottomof thetubeandassayed by immuno-precipitation forthe presenceofRoparticles.Thefractionscontaining

thepeaksoftheRo RNPswerefurther fractionatedbyHPLC inTSEat

roomtemperatureon a gel permeation column(1TKG3000 SW600

X7.5 mm; Phenomenex, Rancho Palos Verdes, CA) at a flow rate of 0.5ml/min (260psi).Molecularweight standards included

thyroglob-ulin(669kD), ferritin (440 kD),catalase(232 kD), aldolase (158 kD), bovineserum albumin (67 kD), chymotrypsinogen A (25 kD), and

uridine(0.244 kD). Theeluates werecollected in

250-jul

fractionsand

assayed byimmunoprecipitationfor thepresenceofRoparticles.

SDS-polyacrylamide gel electrophoresisandimmunoblotting. Pro-teins present in extracts ateachstepof the purification wereeither concentrated inCentricon-30cones(AmiconCorp., Danvers, MA),or

precipitatedin25% TCA,spun down bycentrifugationat 12,000 gfor

15 min, the pellet washed twiceincoldacetone, anddried. Proteins

werethendissolved inSDS-samplebuffer (63mMTris-Cl [6.8],2%

SDS,2%2-mercaptoethanol,20%glycerol, 0.004% bromophenol blue) before fractionation in SDS-7.5% polyacrylamidegels (22) (acrylam-ide/bis, 30:0.8). 4ggperlaneof 60-kDRopolypeptides affinity-puri-fied from human placenta (kindly provided byDr. M.Mamula,Yale

UniversitySchool ofMedicine)werefractionatedas apositivecontrol

oneach gel. After transferto nitrocellulose(SchleicherandSchuell, Keene,NH) in 25% methanol, Tris-glycine[8.3] (23),the nitrocellu-lose sheetswere blocked with 3% BSA in TBSovernightand

subse-quentlyincubated with human antisera diluted 1:50 in TBScontaining 0.1% Tween 20and 1.0% BSA. Bound antibodiesweredetectedwith

251-labeledproteinA(IX 105cpm/ml;ICN,Irvine,CA) followed by

autoradiographyonXRPfilm(Kodak).

Affinitypurification ofantibodies. Anti-La antibodieswere

affin-ity-purified(24) from aserumcontaining hightiters of La anti-bodies andapaucityof anti-Roantibodies,asdeterminedby

immu-noblottingassays.TheLapolypeptidefromasonicate of total HeLa cellswaslocalizedonimmunoblots;thecorrespondingarea wasthen

cut out, incubated withtheanti-La serum, andextensively washed. The bound antibodieswereeluted with 0.2Mglycine-HCI[2.5],

neu-tralized with2 MTris-Cl [7.5],andassayedbyimmunoprecipitation.

Controlantibodies eluted fromanunrelatedstripof the nitrocellulose

sheet lacked anti-La activity in the 32P immunoprecipitation assay (19, 20).

Results

TheRo RNAs dissociate

from

theRo

particles

at

high

ionic

strength.

Todefinetheconditions of

purification compatible

with isolation of nativeRo

particles,

weverifiedthe

stability

of

the Ro

particles

in buffers of

increasing

ionic

strength.

As shown in

Fig.

1

A,

relatively

low salt concentrations disso-ciatedthe Ro RNAsfromthe 60-kD Ro

polypeptides.

Anti-Ro

sera

specific

forthe 60-kD

polypeptide

couldnot

immunopre-cipitate

Ro RNAs from extracts

containing

500 mM NaCl

(Fig.

1_{A,compare lanes 3 and}

_6).

Thisdidnotresult fromthe lossof

binding

of anti-Ro antibodiestothe

polypeptide

itself,

since the

intensity

ofthe 60-kD band

immunoprecipitated

from

[35S]methionine

labeled cellextractsdidnotdecreasein buffers

containing

upto I MNaCI

(Fig.

1

B;

compare lane 2to

lane3). The

sensitivity

todissociationinbuffers of

high

ionic

strength

variedamong the Ro RNPs: hY3 and hY4 were

to-tally

dissociated fromRo

polypeptides

in350mM

_NaCl,

while

hY 1 and hY5were

only

partially

dissociated inthesamesalt

concentration

(Fig.

I

A,

lane

4).

The presenceoflow

millimo-larconcentrations of EDTAin thebuffers

appeared

to

_slightly

stabilizethe Ro

RNA-protein

bond

(data

not

shown),

thus all

buffers used in the

purification procedure

contained 1 mM EDTA.

Dissociation

inthese salt

concentrations

isa

characteristic

ofRo

particles,

since the

immunoprecipitation

ofthe La RNAsby anti-Lasera was notaffectedin upto750mMNaCl

0

'U '0

o 000000

z zzzzzz E E E E 222

0 OC) ;C) tO C:)D

0 0 3 0 0 0 0

a-,ulh ULO If)f U

'n, N-_ _

0

IC

0 C) Z

B

z

2 E

B

U)

69

-

46-olo

O.-A "LaRNAs

I~

.;_

'Pg44* 2 3 4

-C)

10

10

Figure1. Immunoprecipitationof Ro RNPsfrom HeLacellextracts inincreasingsaltconcentrations. (A) Immunoprecipitation of 32P-labeled extracts with

ananti-Roserum(lanes2-6)andwith aserum con-taining anti-Roandanti-Laantibodies(lanes 7and 8). Immunoprecipitationswereperformed inNET-2

buffermodified with4 MNaClinordertoobtain a R0 _{final concentration of 0.15M(lanes6and 8), 0.25 M}

(lane5), 0.35M(lane4), 0.50M(lane3)and 0.75M

Lo NaCl(lanes 2 and 7). Lane 1 represents RNAs pres-entin total cellularextracts.(B) Immunoprecipita-tion of[35S]methioninelabeledextractswitha

nor-mal humanserum(NHS;lane1), withananti-Ro

serum(lanes 2 and 3) and witha serumcontaining bothanti-Ro and anti-La antibodies (lane 4). The immunoprecipitationswereperformed inNET-2

buffercontaining 150 mM NaCI (lanes 1, 2, and 4) andcontainingI MNaCl(lane3). The positions of

the60-kDRo and the 50-kD Lapolypeptidesare in-dicated.Abandcorrespondingtothe 52-kDRo poly-peptide (13)was notimmunoprecipitated with these

sera.Thebandsat45and 65 kDare notspecific since theyappearin theimmunoprecipitate formed withthenormal control serum. Molecularweight markerswere'4C-methylatedbovineserumalbumin (69kD) and '4C-methylatedovalbumin (46 kD).

60-kDRopolypeptide,ananti-Laserumstill

immunoprecipi-tatedanRNAco-migrating withhY5; however,bands

corre-spondingtohY1and hY2(abreakdownproductof hY1) have

disappeared (Fig. I A;comparelane 7tolane8).

Three subpopulations ofRoparticlescanbebiochemically purifiedfromHeLacells. AsreportedbyKatoetal.(25),more than90% oftheRoparticleswererecovered in thecytoplasmic fraction of HeLa cells prepared by gentle cell disruption,even

without detergent (datanotshown). The cytoplasmic fraction

wasthen loaded on aDE52anionexchange column, washed

extensively, and the bound material eluted stepwise with buffers ofincreasingionicstrength.AsillustratedinFig. 2, and

as previously reported (19), the RohY5 particleswere present

inboththe 175mMand 190mMNaCl eluates (lanes3 and4, lanes 10and 11), a group ofparticles containing all fourRo

RNAs

(RohYIlhY5

particles) eluted between 210 mMand 250

mMNaCl (lanes 6 and 13), and Ro particles containing hY1,

hY3, and hY4RNAs (RohYIhY4_{particles), but nearly devoid}

ofany hY5 RNA, eluted between 250 and 300 mM NaCl

(lanes7 and14). The Ro RNAs in each of these fractionswere

immunoprecipitated with an anti-Ro serum that bound the 60-kDRo polypeptide (lanes 10-14), indicating that this poly-peptidewascomplexedwith its cognateRNAs. Contaminating

tRNA and 5S RNAwerepresentin eachofthethreefractions;

however, it was noteworthy that the hY5 RNA and the La

RNAsconstituted majorRNAspeciesinthe 175and 190 mM

and in the 210to 250mM NaCleluates, respectively (Fig. 2,

lanes 3 and4, and lane6).From these elutionexperiments,it was evident thatRo particles existedasatleasttwo

popula-tions: theRohYs RNPs,and the RoYl,hY3,hY4 RNPs. The

exis-tenceofathirdpopulation ofparticles,onethat containedall

four Ro RNPs (RohYl-hY5) and that eluted between 210 and 250 mM NaCl couldsimplyresult fromacontamination with each of the othertwopopulations; alternatively,itcould

repre-sentagroupof Ro particles containingthe RoRNAsin ap-proximately equimolarratios.

After concentration to - 500 ,ul, the three eluates _were

loaded separatelyonto 10-30%(wt:wt) sucrosegradientsand fractionated by rate zonal ultracentrifugation. The RohYM RNPs,elutedat 175 mM NaCl fromDE52,sedimentedasa

single peakwhere themajor contaminatingRNAwas5SRNA

(Fig. 3A, lane6); in thepeak fraction, 5SRNAwas not

im-munoprecipitablewith anti-La antibodies(datanotshown).In the 250 mM NaCl DE52 eluate, the Ro RNPs appeared to sediment slightly ahead ofthe La RNPs (Fig. 3 B, compare

lanes 5 and 6 to lanes 7and 8). This observation was con-firmed by immunoprecipitations ofindividualgradient frac-tions with an anti-Ro serum and with an anti-La serum as

A

:I

58S

---U I

-55- f hY

-hY2

--nY3

-hY4

5-u)0 z

4-0

F--5

8S

=

I

a)

CDNA

10

0

0

-J -1 10 r

+

0

10

LUJ

F~~~~~

E E EE E {

-aE

E EE E

LC)OD

0000-+-40-

uO 0000)

r-as) -O O >1 >- L O

--

C\J

C'4J

N U Q-

cmJ0

r'

a

! s I m

..:...

2

3

4

5

6

7

8

9

10

11

12

13

14

shown in Fig. 3 C (note thatthe peak of immunoprecipitable anti-Ro sera

Ro RNPs was in fractions 14 and 15, whereas the peak of plied that thi

immunoprecipitable La RNPs wasinfractions 16 and 17). A tated Ro RIN

population ofRohY4RNPsdid sediment with the peak of the investigate tI La RNPs (Fig. 3 C, lanes 4 and 9), however, indicatingthat anti-Ro, as i

RohY4particles couldrepresent athird discrete subpopulation eluates conts

ofRoRNPs,inadditiontoRohY5RNPsandapopulation that RNPs, respe

appeared to contain the hYl-hY4 RNAs

(RohYI-hY4

RNPs). top, lane 4) The displacement ofthe RohY4 particles from the other Ro bottom, lane RNPs was notsecondary to in vitrodissociation of the hY4 tainingthe F

RNA from the antigenic 60-kD Ro polypeptide since these parallelingtE

RNPscould beimmunoprecipitatedbyaserumthat targeted lanes5and;

the 60-kDprotein. contaminatit

The peaks ofRoparticles from thesucrosegradientswere was stably a

further fractionated by gel filtration HPLC.Again, RohY4par- both theLaa tides behaved differently from the other Ro RNPs (Fig. HPLC eluatc 4). While the Stokes' radius of the

RohY5

particles and the and the Rob

RohYIhY4particles appeared compatible with theelution pro- contained b( file ofaglobular protein of300-350 kD, the RohY4 particles Ofnote,I elutedat - 230kD(Fig. 4) with the bulk of theLaparticles dominantly

(data not shown), consistent with their cosedimentation in with minima

sucrosegradients. After HPLCpurification of theRohY5 RNP comparedto

fraction from sucrose gradients, hY5 RNA represented the immunoaffin

predominant RNA visible in silver-stained gels of

phenol/

bottom,lane

chloroformextractsof the peak eluates(Fig. 5A, lanes4and blotsof theI

5), although several polypeptideswere present in these frac- RohYI-hY4ani

tions (Fig. 5 B, lanes 5 and 6). The peak fractions of the polypeptide RohYIhY4particles and of theRohY4 particleswerestill heavily that the Ro

contaminated withLaRNPs; in thesefractions,the Ro and La substantially

RNAswere notdegraded (datanotshown). To inves

Characterization of the

purified

Roparticles. After each the HPLC-pi

purification

step,the Ro RNAswerS

immunoprecipitated

with itated the

pei

15S

-/hY

--Il

hY3

-hY:3

hY

4

' hY5

Figure 2. Ro particlesafter fraction-ation withDE52anion exchange chromatography. Total RNAs are shown inlanes 1-7. To help

iden-tifythefractionated RNAs, immu-noprecipitates formed with an anti-Roplus anti-La serum (lane 8) and

ananti-Ro serum (lanes 9-14) are also shown. Lane 1, RNAs in total cellular extract; lanes 2, 8, and 9, RNAs in cytoplasmic fraction; lanes 3 and 10, 175 mMeluate;lanes 4 and11, 190 mMeluate;lanes5and 12, 210mMeluate;lanes 6 and13, 250mMeluate;lanes 7 and 14, 300 mMeluate.

specific for the 60-kD Ropolypeptide. This im-isprotein remained boundtothe

immunoprecipi-NAsthrough the final stages ofpurification. To hispossibility, weperformed immunoblots using well as anti-La sera, of the peaks of the HPLC aining the RohY5, the

RohYI-hY4,

and the RohY4

ctively. Both the 60-kD Ro polypeptide (Fig. 6,

and the 50 kilodalton La polypeptide (Fig. 6, 4)were presentin the peak HPLCfractions

con-ZohY5 RNPs, with thequantitiesof these proteins hepeak ofthe RohY5 particles (compare Fig. 5A,

7 toFig.6, lanes4and5).Thesefractions lacked

ng La RNAs, suggestingthat the Lapolypeptide Associated with the RohY5 particles. As expected,

and the 60-kDRopolypeptideswerepresent inthe

escontaining thepeak fractionsof the RohYI-hY4 iY4_{RNPs (data notshown), since these fractions} )thRoand La RNAs.

theLapolypeptide in the HPLC eluateswas

pre-composed of the intact 50 kilodalton molecule,

.1_{degradation (Fig.}6, bottom, lanes 1,2,4,and5), )degradationofcontaminatingLaprotein in the

nity-purified Ro polypeptide preparation (Fig. 6,

-3). Similar resultswereobtained with

immuno-HPLC eluatescontaining the peak fractionsofthe Id theRohY4RNPs(datanotshown). Since theLa

ishighlysensitive toprotease(26), thisindicated particles purified underourconditions were not

degraded.

tigate

more precisely the protein constituents of

iurified,intactRohY5 particles, we

-u cE (5)

L.H

en _s?

uo

p

0

CU)₁

13 15 16 17 19 FRACTION -?4

-5S

-hY5

2 3 4 5 6 7 8 9

E c:} 0

c)

LZ

a)

-+

° 3 8 13 14 15 16 17 FRACTION #

5S -5S

hY

hY3 2

hY4, ~a~tRNA

1 2 3 4 56 7 8 9

_

_)

.2

Is

m

f

_W

f

6~

CM OD

X up

cuJ_C

_

DGDl

C

'1

Co

I-RohY4 + +

RohY1-hY4 - + +

hYS

++-Roh .

0 5 10 15 20 25 30 35 40 45

Fraction Number

Figure4.HPLC gelfiltration of the fractions fromthe sucrose

gra-dientscontaining the peaks of Ro particles. The elution profile of molecularweight standards is shown (BSA, bovineserumalbumin). Theabscissa shows fraction number (see Methods for details). The elution profiles oftheRohY5, RohY4, andtheRohYIhY4 particlesare

shownontheordinateas-to+++,where-is undetectable by im-munoprecipitation, +is barely detectable,++is clearly detectable, and+++is the maximumintensity. Blankspacesindicate thata

particular fractionwas nottestedbyimmunoprecipitation.

Ropolypeptideand the Lapolypeptide,respectively; antibod-ies that selectively immunoprecipitated the RohY5 particle

from crude cell extracts (19) werealso usedin these

experi-ments. As shown inFig. 7, lanes2, 3, and 5, these three

anti-bodies immunoprecipitated the RohY5 particle. Since the

RohYF-hY4and theRohY4 fractionswerecontaminated withLa

RNPs,asimilar

immunoprecipitation

ofthese

purified

parti-cles was notattempted.

0 C

C) C

(n 'O

z o o

aE 10 0

14 15 16 17 18 14 15 16 17 18 FRACTION#

9 LORNAs

1 2 3 4 5 6 7 8 9 10ti

Discussion

We havepartially purified Roparticles from HeLa

cells

that

appear

immunologically

and

biochemically

intact, without

significant degradation

ofthe RNA or

protein

components. These particles likely exist asdistinct populations of RNPs, since they

partitioned

into threesubgroups, RohYI-hY4, RohY4,

and RohY5, with atleast the latterfraction

being

stably

asso-ciated withthe La

polypeptide.

Several lines of evidence indicate that the Ro RNPs puri-fiedduring these studieswereintact. First, the RNA compo-nentsofthepurifiedRNPsdidnotdegrade

during

the

purifica-tion,

since their

electrophoretic mobility

remained constant

andidenticaltothatof

freshly

immunoprecipitated

Ro RNAs.

Second,immunoblotsofthefinal HPLCeluates indicated that the60-kDRo

polypeptide,

aswellasthe

protease-sensitive

La

polypeptide,

were largely intact.

Thirdly,

the

autoantigenic

Figure 3. Sucrosegradientratezonalultracentrifugation ofRo

RNPs.(A) Total RNAspresentinsuccessive fractions of the gradient

of the 175mMDE52 eluate. LaneI shows total RNAspresentin the 175 mMeluatefromDE52, and lane2showsan

immunoprecipi-tateofaHeLacytoplasmicextractusingacontrol anti-Roserumto

help identify the Ro RNAs. Successive fractions (1-22)were

col-lected fromthebottom of the tube; total RNAspresentin

represen-tativefractionsareshown.(B) TotalRNAspresentinsuccessive

frac-tions of thegradientof the 250 mMDE52eluate.Lane1shows the

immunoprecipitated controlLa RNAsand lane2shows

immuno-precipitated control Ro RNAs, both fromacytoplasmicextract. Suc-cessive fractionswerecollected from the bottom of the tube.(C)

Im-munoprecipitates of the fractions ofagradient of the 250mMDE52 eluateusingcontrol anti-Ro(lanes 2-6)and anti-La(lanes 7-11)

sera.The fractionscorrespondtothose in B.

cn

z

0 0

A

0 , r

l 3 8

hY \ hY3 hY4-- 4w

hY5T )iig

C

5.8S\

u 'I

5S-l

hY I

/-hY3 hY4

/-hY5T

ur)

I--J

M:

U)

z

z CC

00

o~6 8 10 12 14 16 18 10 FRACTION

-5.

w~~~~mi

BS

5S- -5S

hYI

hY3 l _

hY4T

-hY5

hRNA L

2 3 4 5 6 7 8 9 10

B

Mr

(x

103)

97

-

66-43

-

P.-,a) O -0

~0 E E.

rf- - L C'J (Q D

0

WL 0 D

c) a:1 c

-J

8 10 12 Fraction#

-Ro

2 3 4 5 6 7

Figure5.HPLCpurified RohY5 particles. (A) TotalRNAs presentin successive fractions from the HPLC gel filtration column containing thepurifiedRohYsparticles.Lane 1shows totalRNAs presentin

cy-toplasmicextractsand lane 2 showsthe controlRoRNAs immuno-precipitated from theseextracts. Thefractionswerenumberedfrom thefirst fraction eluted after the dead volume of the column. Lane 10 shows theRNAsimmunoprecipitated from fraction 10usinga

control anti-Roserum.(B) Silver-stainedSDS-polyacrylamidegel

showingtheproteinspresent inthe175-mMNaCl eluate fromDE52

column (lane2),the peakRohYM fraction inthesucrosegradient

(lane4),and thepeakRohYsfractionsinHPLC gel filtration (lanes

5-7). Extracts loadedineach lane werederived from approximately equalnumbersof cells.Lane3illustrates60 kD Ropolypeptides af-finitypurified fromhumanplacenta.

60-kD Ro polypeptide remained complexed with its RNAs

throughoutthe stepsof

purification,

sinceRo RNAs could be

immunoprecipitated with anti-Rosera;similarly,the La

poly-peptide

remained

stably

associated withthe RohY5

particle

since it was immunoprecipitable with an

affinity-purified

anti-Laserum.

Finally,

the

conformation-dependent

antigenic determinant specifictothe

RohYs

particlewasstill intactsince

these

particles

could be

immunoprecipitated

with an anti-RohYs serum.

CC

-1)

V

NQ o

ho 0

ICry L)C-)

0 Cl) 0

-Y

0

0

0

-4

ftO

14

FRACTION #

Ro-2 34 5

Ro-

La-1

2

3

4 5

Figure 6.ImmunoblotsofRo"Ysparticles after HPLC fractionation. Upperpanel.Thenitrocellulosesheetwasincubatedwithananti-Ro

serum(1:I 00).Lowerpanel.Thenitrocellulose sheet from panelA wasreprobedwithastronganti-Laserum(1:200).In eachpanel,

lane1represents the175mMNaCl eluatefromaDE52 column,

lane 2 represents thepeak ofthe sucrosegradient forthe eluatefrom the DE52column, lane3,60 kD Ropolypeptides affinity-purified

from human placenta, lane4shows thepeak of theRohYsparticles eluted from HPLCgelfiltration,andlane5showsafraction col-lected after theRohY5peak.

Previous studies have concentrated upon purification of theproteincomponent of the RoRNPs, eitherbiochemically (27-29), or by affinity purification (30); in both cases, the conditions ofpurification made itunlikelythatintact

ribonu-cleoprotein particlescould bepurified. Therecent identifica-tion of anti-Ro antibodies, that are apparently restricted to

intact RohY5 particles (19), aswell asthe discovery ofother

autoantibodies that target conformational

epitopes (31, 32),

indicatesthatisolated RNPs

containing

theircognate RNAs will beusefulin furtheranalysis ofthehumoralautoimmune

response.

Inextension ofourpreviousimmunologicaland

biochemi-calobservations suggesting that theRoparticlesare

heteroge-neous(19), sensitivity to salt disruption was uneven among these RNPs: the RohY4 and RohY3 RNAsbegantodissociate

fromtheirautoantigenic polypeptideat250 mMNaCl, with theRohY' and

RohYl

RNPs

being

somewhatmoreresistantto

disruption in high salt. Binding of all four Ro RNAstothe 60-kD Ropolypeptidewastotally

disrupted

inbuffers of

LO)

o 0

Er

cE

ICU HO

0

Er

Imt

Cn

cc

I

-z I1

hYl

hY 3_

j

hY

4

hY5'

1

2

3

4

5

Figure 7. Immunoprecipitation of HPLC-purifiedRohY" particles. TheRohY5particles from the peak of the HPLC gel filtration column were immunoprecipitated with 10 !d of sera specific for an epitope restricted to theRohY5 particle (19) (lane 2), or the 60-kD Ro

poly-peptide (lane 3); immunoprecipitates were also formed with a nor-mal human serum (lane 4) and with affinity-purified La

anti-bodies (lane5)(24). LaneIshows the control Ro RNAs immuno-precipitated from a HeLa cytoplasmic extract using the same anti-Ro serumshownin lane 3.

tively low ionic strength (500 mM NaCl), a characteristic

shared by many nucleic acid binding proteins (33).

Con-versely, the binding of the La RNAs to the La protein did not

appear to be affected by exposure to buffers containing up to 750 mM NaCl. Thus, interactions of these two proteins with their respective RNAs are likely to be quite different, even if they appear to be present simultaneously on the same ribonu-cleoprotein particles, as previously shown (28, 29, 34), and as confirmed in our studies.

Heterogeneity of Ro particles was further substantiated during purification, since three populations of these RNPs

were obtained that differed in discretebiochemical properties.

The first population containing the hY5 RNA (RohYs

parti-cles) eluted from an anion-exchanger at low salt, sedimented in

sucrose gradients somewhat faster than the La RNPs, and had a Stokes' radius compatible with a globular protein of 300 to

350 kD. A second one enriched in RohY4 RNPs eluted from

DE52 at higher salt, sedimented in sucrose gradients with the La RNPs, and had an apparent molecular

weight

of 230

kD. Finally, a group of particles containing the hY 1, hY3 and hY4 RNAs (RohYl-hY4 RNPs) eluted at an intermediate salt

concentration, sedimented in sucrose gradients ahead of the La RNPs, and had an apparent molecular weight of 300 to 350 kD. At present we are unsure if this latter group of particles consist of individual Ro RNPs with similar physical proper-ties, or if they exist asa complex.

Although othershave previously reported that Ro particles

fractionated with an apparent molecular weight of 100-150 kDin gel filtration (30, 35), the purification procedures used in theseearlier studies made it unlikely that the sized particles wereintact. In the current study, Ro particles had unexpect-edly high Stokes' radii corresponding to

globular

proteinsof 230 and 300-350 kD. Molecularweight determinations of RNPs by gelfiltration are imprecise, however, at least in part because these particlesin solution may not behave as globular proteins

(36).

For example, the 7S RNP, comprised ofone copy each of the 5S RNA and a 34-kD polypeptide with a

combinedmolecularweight of 73 kD, has a Stokes' radius by gelfiltrationHPLC corresponding to a globular protein of 120 kD. On the other hand, the actual molecular weight of one 60-kD Ro polypeptide plus one Ro RNA (27 kD forhY5,and 37 kD for hY

I)

is similar to that of the 7SRNP,yet theStokes'

radius is much greater for Ro RNPs; this observation at least suggests the possibility that Ro particles or their individual componentsexist as multimers, or that they contain polypep-tide components in addition to-the 60-kD protein.

Although we did not address all the possibilities thatcould

account for the observed size of Ro RNPs, we did showthat the

Roh'5

RNPscontained the La and the 60-kD Ro

polypep-tides since antibodies specific for these proteins

immunopre-cipitated the purified RohY5 particles. Moreover, in theHPLC

purified

RohY5

fractions, immunoblots demonstrated the pres-ence of both the La. and the 60-kD Ro polypeptides; since immunoprecipitates of these fractions lacked the La RNAs, it

did not seem likely that contamination with La RNPs

ac-counted for the presence of the Lapolypeptide. Further,since

the HPLC fraction that was enriched inRohY5 particles eluted

at 300-350kD, it is unlikely that free Lapolypeptide

specifi-cally contaminated this fraction and accounted for the im-munoblot results (Fig. 6,bottom). Thus, the

immunoprecipi-tation and immunoblot experiments indicate that both

poly-peptides are capable of associating with thehY5RNA and are likely stablycomplexedwiththisRNA, atleast under the con-ditions utilized in this study. To accountfor the 300-kD mo-lecular weight of the RohYl RNPs, however, these particles

may exist as dimers. At present, we do not know if thenewly

described 52 kD Ro polypeptide (13) is also a constituent of these purified RNPs, as the sera we used as probes in these studies only targeted the La andthe60-kD Ro proteins.

It is unclearif the other populations of Ro RNPs(RohY'-hY4

and

RohY4

RNPs) are stably associated with the La

polypep-tide. These fractions werecontaminatedwith the La RNPs and immunoprecipitation andimmunoblotting experiments with anti-La antibodieswereobscured bythis

contamination;

how-ever, the sizeof theseparticles (> 220 kD)suggestthat

they

are

composed of more than one copy of the60-kDRopolypeptide

and more than one 30-kD RNA.

Certain observations suggest that the La polypeptide may

have a unique relationship with RohY5 RNPs. For example,

under conditions where all Ro RNAs were dissociated from

the 60-kD Ropolypeptides(Fig. 1 A, lane 2),anti-La antibod-ies stillimmunoprecipitated afraction of thehY5 RNAs, but

none of the hYI andhY2RNAs (Fig. 1 A, lane 7).Similarly,

we also haveobservedthatimmunoprecipitates oftotal HeLa cell extracts using affinity-purified anti-La antibodies show a

It is significant that among all mature, processed RNA polymerase III transcripts, Ro RNAs (ora portion of these RNAs such as hY5) are the only ones stably associated with the Lapolypeptide, and that these RNAs are the RNA polymerase IIItranscripts still produced incells grown under adverse con-ditions ( 11). It thus follows that the function of Ro RNPs may be related to that of the La polypeptide, which is thought to

serveas aterminatorofRNApolymerase III transcription (17, 18). Ro RNPs could serve as a cofactor that modulates La activity, or as a "sink" to help recycle used La polypeptides for another round of transcription termination. Alternatively, Ro RNPsbearing the La polypeptide could contain incompletely processed Ro RNAs (17, 18); however, this latter possibility appears less likely since no difference in mobility of the hY5 RNA can be detected when RohY5 RNPs are immunoprecipi-tated with either anti-Ro or anti-La antibodies.

Antibodies specific forthe 60-kD Ro and the La polypep-tidesfrequentlyoccurintandem, or as a linked set (4, 10). The closeproximityandthe stable association of these two

poly-peptideson atleast a subset of the Ro RNAs provide a poten-tial macromolecular target for this linked autoantibody set, and thus favors the hypothesis that autoantigens drive im-mune responsesinSLE and related illnesses (4).

In conclusion, both protein and RNA components have been shown to be essential forcompletefunction ofcertain

other RNP autoantigens (37-39). Thus, intact Ro particles

such as those obtained in these experiments may represent reagents to studytheactivity of thisRNP, either aloneor in

conjunction withthe Lapolypeptide, perhapsin theregulation oftranscription byRNApolymeraseIII.

Acknowledgments

Weare verygratefulto Dr. JohnHardin (Yale University Schoolof Medicine) for invaluable discussions,support, andreview of the

man-uscript.We alsothankDr. H. A.Menard(Centre Hospitalier Univer-sitairedeSherbrooke) forsera,andDr. M.J.Mamula(Yale University

School ofMedicine) forhelpful discussions,review of thiswork,and

affinity-purifiedhuman60-kDRopolypeptides.

Thiswork wassupported bygrantsfrom the National Institutes of Health (AI-26853 to J. Craft, AR-32549 to Dr. John Hardin), the United Scleroderma Foundation(toJ.Craft),the Arthritis Foundation anditsConnecticut Chapter, andtheLupus Foundation of America.

Dr.Boire isarecipient of ArthritisSocietyPostdoctoralFellowshipFC 13-27-(87).Dr.Craft isaPewScholar in the Biomedical Sciences.

References

1. Tan,E. M. 1982. Autoantibodiesto nuclearantigens (ANA):

theirimmunobiologyand medicine. Adv. Immunol. 33:167-240. 2.Hardin,J. A.,andT.Mimori. 1985. Autoantibodiesto

ribonu-cleoproteins.Clin.Rheum.Dis. 11:485-505.

3.Tan, E. M., E. K. L.Chan,K.F.Sullivan,and R. L.Rubin. 1988. Antinuclear antibodies(ANAs): diagnostically specific immune markersand cluestoward theunderstandingofsystemic

autoimmu-nity.Clin. Immunol.Immunopathol.47:121-141.

4.Hardin,J. A. 1986. Thelupusautoantigensandthepathogenesis

ofsystemiclupus erythematosus.Arthritis Rheum. 29:457-460.

5. Sturgess, A. D., M. G. Peterson, L. J. McNeilage, S.

Whit-tingham,and R. L.Coppel. 1988. Characteristics andepitope mapping

ofacloned humanautoantigen La. J.Immunol. 140:3212-3218.

6. Vande Water, J.,M. E.Gershwin,P.Leung,A.Ansari,and R. L.

Coppel. 1988. The autoepitope of the74-kDmitochondrial

autoanti-genof primarybiliarycirrhosis correspondstothefunctional site of dihydrolipoamide acetyltransferase.J. Exp.Med. 167:1791-1799.

7.Deutscher, S. L.,J. B.Harley,andJ. D.Keene. 1988. Molecular

analysisof the60-kDahumanRoribonucleoprotein. Proc.Nati.Acad. Sci. USA. 85:9479-9483.

8. Query, C. C., andJ. D. Keene. 1987. Ahuman autoimmune

proteinassociated withU I RNAcontainsaregion ofhomology that is

cross-reactive withretroviralp3O0 antigen.Cell. 51:211-220. 9. Reichlin, M. 1986. Significance ofthe Ro antigensystem. J.

Clin.Inmunol. 6:339-348.

10. Harley,J. B., E. L. Alexander,W. B. Bias, 0.F. Fox, T. T.

Provost, M. Reichlin, H. Yamagata, and F. C. Arnett. 1986. Anti-Ro(SS-A) and anti-La(SS-B) in patients with Sjogren's syndrome.

Ar-thritisRheum. 29:196-206.

11. Hendrick,J.P.,S. L.Wolin,J.Rinke, M.R.Lerner, andJ. A.

Steitz. 1981. Ro smallcytoplasmic ribonucleoproteinsareasubclass of

Laribonucleoproteins: further characterization ofthe Ro and La small ribonucleoproteins from uninfected mammalian cells. Mol. Cell. Biol.

1:1138-1149.

12.Wolin,S. L., andJ. A.Steitz. 1984. TheRosmallcytoplasmic

ribonucleoproteins: identification oftheantigenic proteinandits bind-ing siteontheRoRNAs. Proc.Natl.Acad. Sci. USA.81:1996-2000.

13.Ben-Chetrit,E.,E. K. L.Chan,K. F.Sullivan,and E. M. Tan.

1988.A52-kDproteinisanovelcomponentoftheSS-A/Ro antigenic particle.J.Exp. Med. 167:1560-1571.

14. Rader,M. D., C.O'Brien,Y.Liu,J.Harley,and M.Reichlin. 1989.Heterogeneity ofthe Ro/SSAantigen:Different molecular forms in lymphocytes and red bloodcells. J.Clin. Invest. 83:1293-1298.

15.Wolin,S. L., andJ. A.Steitz. 1983. Genes fortwosmall

cyto-plasmic Ro RNAsareadjacent andappeartobesingle-copyin the humangenome.Cell. 32:735-744.

16. Stefano, J. E. 1984. Purifiedlupusantigen Larecognizesan oligouridylatestretchcommon tothe3'terminiofRNApolymeraseIII

transcripts.Cell. 36:145-154.

17.Gottlieb,E., andJ. A.Steitz. 1989. TheRNAbindingprotein

Lainfluences boththe accuracy and theefficiencyofRNApolymerase

IIItranscription in vitro. EMBO(Eur. Mol. Biol. Organ.)J. 8:841-850.

18.Gottlieb, E., andJ. A.Steitz. 1989. Function ofthe mammalian

La protein: evidence for its action in transcription termination by

RNApolymeraseIII. EMBO(Eur.Mol. Biol. Organ)J. 8:851-862. 19.Boire,G., andJ.Craft. 1989. Biochemical andimmunological

heterogeneityof theRoribonucleoprotein particles.Analysiswithsera specificfor the RohY5 particle. J. Clin.Invest.84:270-279.

20.Craft,J.,T.Mimori,T.L.Olsen,and J. A.Hardin. 1988. The U2smallnuclearribonucleoproteinparticleas anautoantigen: analy-sis with serafrom patientswith overlap syndromes. J. Clin. Invest. 81:1716-1724.

21.Forman,M.S.,M.Nakamura,T.Mimori,C.Gelpi,andJ. A.

Hardin. 1985. Detection of antibodiestosmall nuclear

ribonucleopro-teins and smallcytoplasmic ribonucleoproteins usingunlabeled cell

extracts.Arthritis Rheum.28:1356-1361.

22. Laemmli,U. K. 1970.Cleavageof structuralproteins during

the assembly of the head ofbacteriophage T4. Nature

(Lond.).

227:680-685.

23.Towbin, H.,T.Staehelin,andJ.Gordon. 1979.

Electrophoretic

transfer ofproteinsfrom polyacrylamide gelstonitrocellulose sheets:

procedure and some applications. Proc. Natl. Acad. Sci. USA. 76:4350-4354.

24.Smith,D. E.,andP. A.Fisher. 1984.

Identification,

develop-mental regulation, and response to heat shock oftwo

antigenically

related forms ofamajornuclearenvelopeproteinin

Drosophila

em-bryos.Applicationofanimprovedmethod for

affinity

purification

of antibodiesusingpolypeptidesimmobilizedonnitrocellulose sheets.J.

25. Kato, N., H. Hoshino, and F. Harada. 1982. Nucleotide se-quenceof4.5S RNA(C8orhY5) fromHeLacells.Biochem. Biophys.

Res. Commun. 108:363-370.

26. Habets, W. J., J. H. den Brok, A. M. Th. Boerbooms, L. B. A. van de Putte, and W. J. vanVenrooij. 1983. Characterization of the

SS-B(La)antigen inadenovirus-infected anduninfected HeLa cells.

EMBO (Eur.Mol. Biol. Organ.)J.2:1625-1631.

27.Lieu, T.S., M.Jiang, J. C. Steigerwald, andE. M. Tan. 1984. Identification oftheSS-A/Rointracellular antigen with autoimmune

sera.J.Immunol. Methods. 71:217-228.

28.Elkon,K.B.,andL.Culhane. 1984. Partial immunochemical characterization ofthe RoandLaproteins using antibodies from

pa-tients with the siccasyndrome and lupus erythematosus. J. Immunol. 132:2350-2356.

29.Deng, J.S.,R. D.Sontheimer,and J. N.Gilliam. 1985.

Molecu-larcharacteristicsofSS-B/La and SS-A/Ro cellularantigens.J. Invest.

Dermatol. 84:86-90.

30.Yamagata, H., J.B.Harley,and M.Reichlin. 1984.Molecular

properties oftheRo/SSAantigenandenzyme-linked immunosorbent

assayforquantitation of antibody.J.Clin.Invest. 74:625-633. 31. Elkon,K., S. Skelly,A. Parnassa, W. Moller, W. Danho,H. Wessbach, and N. Brot. 1986.Identificationandchemicalsynthesis of

aribosomal protein antigenic determinant in systemic lupus

erythema-tosus. Proc.Nati. Acad. Sci. USA.83:7419-7423.

32.Boire, G.,S.Lapointe,F. J. Lopez-Longo,andH. A. Menard.

1988. Naturally occurringautoantibodies recognize conformational determinantsonthe Roparticles. Arthritis Rheum. 31 :S67 (Abstr.).

33. Schleif,R. 1988. DNAbinding byproteins. Science (Wash.

DC).241:1182-1187.

34. Eisenberg, R. A. 1985. Association between the Ro and La

antigenic determinants: immunodiffusion analysis ofhuman spleen

extract.J.Immunol. 135:1707-1713.

35. Scopelitis,E., J. J. Biundo, andM. A.Alspaugh. 1980. Anti-SS-Aantibodyand otherantinuclearantibodies in systemiclupus

ery-thematosus. Arthritis Rheum. 23:287-293.

36.Monteilhet, C.,and L. P.Aggerbeck. 1986. High performance liquidchromatographic purification ofaribonucleoprotein complex

anditsproteincomponent.Anal.Biochem. 154:162-170.

37. Sharp, P. A. 1987. Splicing ofmessenger RNA precursors.

Science(Wash. DC).235:766-771.

38.Altman, S.A., H.Gold, andM.Bartkiewicz. 1988.

Ribonucle-ase P as an snRNP. InSmall NuclearRibonucleoproteins. M. Birn-stiel, editor. Springer Publishing Corp.,New York. 183-195.