Autoantibody to RNA polymerase I in scleroderma sera

Autoantibody to RNA polymerase I in

scleroderma sera.

G Reimer, … , U Scheer, E M Tan

J Clin Invest.

1987;

79(1)

:65-72.

https://doi.org/10.1172/JCI112809

.

Autoantibodies to components of the nucleolus are a unique serological feature of patients

with scleroderma. There are autoantibodies of several specificities; one type produces a

speckled pattern of nucleolar staining in immunofluorescence. In actinomycin D and

5,6-dichloro-beta-D-ribofuranosylbenzimidazole-treated Vero cells, staining was restricted to

the fibrillar and not the granular regions. By double immunofluorescence, specific rabbit

anti-RNA polymerase I antibodies stained the same fibrillar structures in drug-segregated

nucleoli as scleroderma sera. Scleroderma sera immunoprecipitated 13 polypeptides from

[35S]methionine-labeled HeLa cell extract with molecular weights ranging from 210,000 to

14,000. Similar polypeptides were precipitated by rabbit anti-RNA polymerase I antibodies,

and their common identities were confirmed in immunoabsorption experiments.

Microinjection of purified IgG from a patient with speckled nucleolar staining effectively

inhibited ribosomal RNA transcription. Autoantibodies to RNA polymerase I were restricted

to certain patients with scleroderma and were not found in other autoimmune diseases.

Research Article

Autoantibody to RNA

Polymerase

I in

Scleroderma

Sera

Georg Reimer,* Kathleen M. Rose,*UlrichScheer,1and Eng M. Tan*

*W. M. Keck Autoimmune Disease Center, Department ofBasic and Clinical Research, Scripps Clinic and Research Foundation, LaJolla, California92037; tUniversity of Texas Medical School, Department ofPharmacology, Houston, Texas 77225;

OGerman

Cancer Research Center, Institute of Cell and TumorBiology,D-6900Heidelberg, Federal Republic of Germany

Abstract

Autoantibodies to components of the nucleolus are a unique se-rological feature of patients with scleroderma. There are auto-antibodies of severalspecificities; one type produces a speckled pattern of nucleolar staining in immunofluorescence. In actino-mycin D and 5,6-dichloro-,8-D-ribofuranosylbenzimidazole-treated Vero cells, staining was restricted to the fibrillar and not the granular regions. By double immunofluorescence, specific rabbit anti-RNA polymerase I antibodies stained the same fi-brillar structures in drug-segregated nucleoli as scleroderma sera. Scleroderma sera immunoprecipitated 13 polypeptides from [35Sjmethionine-labeled HeLa cell extract with molecular weights ranging from 210,000 to 14,000. Similar polypeptides were pre-cipitated by rabbit anti-RNA polymeraseIantibodies, and their commonidentities were confirmedinimmunoabsorption exper-iments. Microinjection of purified IgG from a patient with speckled nucleolar staining effectively inhibited ribosomal RNA transcription. Autoantibodies to RNA polymerase I were re-strictedtocertainpatients with scleroderma and were not found in otherautoimmune diseases.

Introduction

Patients withsystemic autoimmune diseases spontaneously pro-duce antibodies against a variety of cellular components (1).

DNAandhistones are among nuclear autoantigens with known structure andfunction.Other nuclear antigens are either nuclear

proteins or RNA-protein complexes that have been partially characterized and called Ul-RNP, Sm, SS-B/La, Jo-i, Scl-70,

andcentromere/kinetochore antigens (2-10). Thereare expec-tations that knowledge ofthe structure and function ofthese

autoantigenscould lead to abetterunderstanding of the

etio-pathogenesisof autoimmune diseases andprovidebetter reagents

toconstruct accuratediagnostic probes. Furthermore, some of these naturally occurringautoantibodies have been used suc-cessfully as probes in molecular and cellular biology. For

ex-ample, antibodiesagainstUl-RNPwereshowntoinhibitsplicing of heterogenous nuclear RNA, thusimplicating animportant

role of this ribonucleoprotein in the processing of precursor

Addressreprintrequests to Eng M. Tan, M.D., W. M.Keck Autoimmune DiseaseCenter, ScrippsClinicandResearchFoundation, 10666 North TorreyPinesRoad, La Jolla,CA92037. Dr. Reimer'scurrent address isDermatologische Universitatsklinik,Hartmannstr. 14,D-8520 Erlan-gen, FRG.

Receivedfor publication28May1986andinrevisedform 27August 1986.

mRNA(1 1). Autoantibodies against SS-B/La identified this nu-clearribonucleoprotein as possibly involved in the transport of RNA, particularly products of RNA polymerase III, including virus encoded RNAs EBER 1and 2 of Epstein-Barr virus (12). More recentevidence suggests that the SS-B/La particle is either

atermination factor for RNA polymerase III or is involved in

posttermination processing events (13). Recently, anticentro-mere/kinetochore antibodies from scleroderma patients have facilitated identification and molecular cloning of a common human kinetochore protein (14).

Antinucleolar antibodies are part of the spectrum of auto-antibodies present in scleroderma (15-17). This systemic

au-toimmune disease is characterizedbymicrovascular lesions and abnormalities of the immune system leadingtofibrosis of the

skinandcertaintarget organs. One of theantigensreactive with sclerodermaantibodies was recently shown to bea novel

nu-cleolar 7-2

ribonucleoprotein

(18-19). U-3

ribonucleoprotein

was alsoshown to be part ofacomplexprecipitatedbycertain

antinucleolar autoantibodies from scleroderma patients (18).

These autoantibodies were found to interact with a basic

nu-cleolarprotein ofmolwt34,000 (20),termed

fibrillarin,

asthis

proteinwasshowntobeassociated with the fibrillar component

of nucleoli(21).The

identity

of other nucleolar

autoantigens

in scleroderma, however, is unknown.

In this study, we present data that show that theantigens

recognized byasubgroupofantinucleolarautoantibodies consist ofRNApolymeraseIcomplex. The clinical association of this

antibodyspecificitywaswith the diffuse form of scleroderma.

Methods

Sera. Included in thisstudyare sera from 208 sclerodermapatients, kindly provided by 0. P. Hornstein (Erlangen, FRG), H. Mensing

(Hamburg, FRG),M. Meurer(Munich, FRG),D.S. Prince(Birmingham,

AL),R. A.Kaplan (SanDiego,CA),J.G. Curd andG. W.Williams(La

Jolla, CA). Sera frommorethan 100patientswithothersystemic

au-toimmune diseases, including systemic lupus erythematosus

(SLE),'

mixed connective tissue disease(MCTD), drug-induced lupus, derma-tomyositis, polymyositis,andrheumatoidarthritis, and20 normal in-dividuals servedascontrols. Allsclerodermapatientsfulfilledthe prelim-inary criteria forthediagnosisofprogressive systemicsclerosis(22).A

rabbit antiserumto RNApolymerase Iwas usedas areference. This antiserum wasproducedin thelaboratory ofDr. K. M. Roseby im-munization ofarabbit withpurified RNApolymerase I from Morris hepatoma 3924Aina mannerdescribedpreviously (23).Rabbit anti-RNApolymeraseIantibodies,asusedinthe presentstudy,hadpreviously been used inaseries ofotherstudiestolocalizeRNApolymeraseIin normal cells(24)and cellstreated withthe nucleolussegregating drug

5,6-dichloro-l-D-ribofuranosylbenzimidazole

(DRB) (25). These anti-bodieshad been usedfurthertoshowinduction of increasedamount

1. Abbreviations used in thispaper: DRB, 5,6-dichloro-ft-D-ribofura-nosylbenzimidazole; MCTD, mixedconnective tissuedisease; PMSF, phenylmethylsulfonylfluoride; SLE, systemic erythematosus.

J.Clin. Invest.

©TheAmericanSocietyfor ClinicalInvestigation,Inc.

0021-9738/87/01/0065/08 $1.00

andactivity ofRNApolymeraseIin mouse Blymphocytestreatedwith lipopolysaccharide(26).

Indirect immunofluorescence. Indirect immunofluorescence was performed with Hep-2 cells(Bion Enterprises,Ltd., ParkRidge,IL)as

substrate. Other substratesincludedHeLaandVerocells,grownon mi-croscopic slides and fixed in ice coldacetone/methanol 3:1 for 3 min, andratlivercryostatsections. Further tissues included chicken liver and different organ sections from Xenopus laevis. Seraweredilutedstarting

at1:40 inphosphate-bufferedsaline(PBS), pH7.4. Afluorescein(Behring

Diagnostic, La Jolla, CA) andarhodamine-labeled goatanti-human IgGconjugate (Tago, Inc.,Burlingame, CA)wereused.A fluorescein-labeledgoatanti-rabbit IgG conjugatewasobtained from Tago, Inc. Intensity ofimmunofluorescence stainingwasgraded weak (+), moderate (++),orstrong(+++). Resultswerereadon anOrtholuxIIfluorescence microscope withaPloempak vertical illuminator (E.Leitz,Inc., Rock-leigh, NJ) usingaNPL Fluortar100/1.32oil immersion objective. Double immunofluorescence stainingwasperformedasdescribed (27).

Drug studies.Segregation of nucleoli in tissue cultureVerocellsinto fibrillar and granularregionswasinduced by the addition of 0.2,gg/ml actinomycinDtothe culture medium for4 horof 50gg/mlDRB,a

halogenatedadenosine analogue, for 6 h. Bothdrugswere purchased fromBehringDiagnostics. After drugtreatment,cellswerefixedin ace-tone/methanol3:1 andprocessed for immunofluorescenceasdescribed above.

Electron microscopicimmunocytochemistry. Frozensections(5-gm thickness) ofnormal andregeneratingratliverwerefixedinacetone at

-20'Cfor 10 min, air dried,andincubated with sclerodermaserumS18

orrabbit anti-RNApolymeraseIantibodies diluted 1:100inPBSfor2 h.After severalwashesin PBS,goatanti-humanIgGorgoatanti-rabbit IgG coupled to colloidal gold (5-nmdiameter; Janssen Life Sciences Products, Beerse,Belgium)wasaddedat adilution of1:4.After overnight incubationat4°C,thespecimenswerefixedandprocessed for electron microscopyasdescribed in detail elsewhere (24).

Radiolabelingofcells.HeLa cellswerecultured in monolayers and radiolabeled for20 hwith[35S]methionineor[32P]orthophosphate(New England Nuclear, Boston, MA)at0.1mCi/mlinmethionine-or phos-phate-free medium,respectively,asdescribed(28). The cellswerethen washedonce inPBS and harvested witharubberpoliceman in icecold bufferAcontaining10mMTris-HCl (pH7.4), 150mMNaCl, 1.5 mM MgCl2,0.5%Nonidet P-40, supplemented with 23 U/ml kallikrein in-activator (BehringDiagnostics)and2mMphenylmethylsulfonylfluoride

(PMSF)tominimizeproteolytic degradation.The cellswerethen freeze-thawedtwiceandcentrifugedat10,000gfor15min,and the supernatant wasstoredat-70°C until used. - 10 mgproteinwereextractedfrom

HeLacellsgrownin 150X25-mmtissue culture dish.

Immunoprecipitation. Allexperiments were performed at 0-4°C as described (28) with modifications.Inbrief,serumaliquots of 10 glwere

incubated with 100 ul of proteinASepharose CL-4B(Pharmacia, Inc., Uppsala, Sweden)at10%(vol/vol)suspensioninbufferBsupplemented with2 mg/mlbovineserumalbumin. BufferBcontained 50mM Tris-HC1(pH 7.4), 150mMNaCl, 5 mM EDTA, 0.5% Nonidet P-40, 0.5% sodium deoxycholate, and0.1% sodium dodecyl sulfate (SDS). After washing five times in 1.5 mlbufferB, 100 ggradiolabeled HeLacell proteinextractwasaddedtothe beads andincubated for1 hwithconstant mixing.Theradiolabeledextracthad beenprecleared with 1/10 vol of proteinASepharose CL-4B in bufferBbefore beingused.After centrif-ugation for 3min,thepelletwaswashedfivetimes in 1.5 ml buffer B

containing 23 U/ml kallikrein inactivator and 2 mM PMSF. The

[35S]methionine-labeled

pelletwasthen boiledinLaemmli sample buffer, and theproteinswereanalyzed by SDS-polyacrylamide gel electropho-resis (PAGE) (29), followedbyfluorographyand autoradiography. Half of the 32P-labeled pelletwasdigested withDNase and RNase (Millipore

Corp., Bedford, MA)at0.5 mg/ml eachfor 30min at 37°C, and the proteinswereresolvedby SDS-PAGE.The otherhalfof the pellet was phenol extracted, andany RNA wasprecipitated inethanol. RNA analysis

wasperformed in7M urea,8% polyacrylamidegels, then autoradiography

asdescribedelsewhere(2).

Antigendepletion.

(35S]methionine-labeled

HeLacellextract

₍₁₀₀

_Ag

ofprotein)was_{incubated with 40 gg rabbit anti-RNApolymerase}IIgG boundto_proteinA_SepharoseCL-4B for 1hat40C in bufferBto_deplete

theextractofRNA_polymeraseI. The

depleted

supernatantwasthen used in

immunoprecipitation

asthe

_antigen

sourceforhuman antinu-cleolarsera.Inanother

experiment,

[35Slmethionine-labeled

HeLa cell

extract wasabsorbed with 40 gghumanantinucleolar

_IgG

asdescribed and thenprobedwith rabbit anti-RNA

polymerase

I.Thesameamounts

ofIgGfrom normalhumanserumandnormalrabbitserum,aswellas

fromaSLEserumwith Sm

specificity,

wereincluded in control

exper-imentsto ensurethatchangesobservedwerenotdueto

nonspecific

pro-tein-proteininteractions.

Microinjection

of

antibodiesinto X. Iaevisoocytes.

IgG

wasobtained fromarepresentativesclerodermaserum

(SI18) by

column

chromatog-raphyonDEAE-cellulose(DE-52;WhatmanInc., Clifton, NJ)andwas

further

purified

by

chromatography

on

protein A-Sepharose (Sigma

Chemical Co., St.

Louis, MO).

IgG

wasconcentratedto2

mg/ml by

vacuumdialysisagainstPBS. Antibodieswere

microinjected

intonuclei

offull-grownoocytes ofX.laevisasdescribed

(30).

Eachoocytenucleus received 10 nlantibodysolution.Insomeexperiments,thesolutionwas

injected into thecytoplasmof the animal

hemisphere

of theoocytes.

PurifiedIgGfrom normal humansera ormonoclonal antibodies

reacting

with nonnucleolar constituents of oocyteswereusedascontrols. Each

experimentalgroupconsisted of sixoocytes.2 hafter

antibody injection,

each oocyte received 0.1gCi

a-[32P]GTP (10 mCi/ml;

Amersham Buchler GmBH,

Braunschweig, FRG)

by

injection

intothe

cytoplasm.

Afteran

incubation time of6 hat 20°C,RNAwasextracted and

analyzed

by

electrophoresisin 1.5% agarosegelsasdescribed

(30).

The RNA

corre-spondingto twooocyteswas

_applied

toeach

gel

lane, air

dried,

and identifiedby

autoradiography.

rRNAfromX. Iaevisovaryribosomes andRNAfrom tobacco mosaic virus servedasmolecular

weight

markers.

Results

Antinucleolar antibodiesinsclerodermasera.

_IgG

antinucleolar antibodies were present in 31 of208 scleroderma sera

_(15%).

Thesesera wereselected forstrong

(+++)

nucleolar

_{staining by}

indirectimmunofluorescenceonsubstrate

Hep-2

cells.

Nucleo-plasmic

staining

wasalso present withsome _{sera, but the}

des-ignation

ofserawith antinucleolar antibodies in this

screening

assaywasbasedonthecriterionthatnucleolar

_staining

exceeded

nucleoplasmic

staining by

at least one

gradation

of

intensity.

Nineof theantinucleolar

positive

sera_gavea_punctateor

speck-lednucleolar

staining

pattern,

usually

uptoa serumdilution of

1:2,560

or

higher.

The other sera

produced homogeneous

or

clumpy

nucleolar

staining

andwere classified

according

to a

previous

report on the association ofantinuclear and antinu-cleolarantibodiesin scleroderma

_(17).

Inthe present

study,

the

antigen

reactivewithscleroderma

antibodiesthat

produce speckled

nucleolar

_staining

wasfurther characterized.

Localization

by light

and electron

microscopic

immunocy-tochemistry

of

scleroderma antibodieswith

_speckled

nucleolar

staining.

The

following

immunofluorescence results were

ob-tainedwith all nineserawith

speckled

nucleolar

staining.

Fig.

1 showsindirect immunofluorescence

staining

and

phase-con-trast

microscopy

asobtained with

_{representative}

serumS18

_using

HeLacells

(a

and

b),

Verocells

(c

and

d),

and concanavalin-A-stimulated human

peripheral

blood

lymphocytes (e

and

f)

as

substrates.Notethat in different

phases

of the cell

_cycle,

different

staining

patternswereobserved. Nucleolar

_staining

thatconsisted of_many fluorescent

_speckles

in

_interphase

cells

_changed

to a

few

bright

fluorescent spotsin theareaofcondensed

concanavalin-A-o

't.~.

KI

:"r

a' . *'

:. ,

a

'.

-PL.,

4fin

k. S

9

______________________________

f

stimulatedperipheral blood lymphocytes, increased intensity of nucleolar staining was observed in blast-transformed lympho-cytes(Fig. 1 e, large cells). Weak nucleoplasmic staining (+) of

interphasecells was alsoobserved with these sera, whereas

cy-toplasmicstaining was usually absent.

To further definethe nature of the nucleolar antigens, we exposed Vero cells to the nucleolus-segregating drugs

actino-mycinDandDRB.After drug treatment, cells were processed for indirect immunofluorescence and phase-contrast micros-copy. Actinomycin D-inactivated nucleoli show a typical

seg-regation ofthefibrillar component from the granular component

(Fig.2b), withthefibrillar component forming distinct nucleolar hemispheres (31-33). Antibodies demonstrating speckled nu-cleolarstainingin untreated cells were now distinctly localized in thefibrillarregions of segregated nucleoli (Fig. 2 a).

Immu-nofluorescencestaining was most prominent in small crescents

atthe outermost margins of the segregated fibrillar component.

Littleif any decoration of the granular structures that appeared dark in phase opticswasobserved. When DRB was added to Vero cellsgrowing in culture, the normally compact nucleoli unraveledintoextended beaded strands, segregating the nucleo-lusintogranular andfibrillarregions (34, 35).

Immunofluores-cence staining of these cells using the same scleroderma

anti-Figure 1. Immunofluorescence microscopy of HeLacells (a), Vero cells(c), and concanavalin Astimulatedperipheral human blood lympho-cytes(e) afterstaining with antinucleolar anti-bodies(S18) from a patient with scleroderma. Nucleoli demonstrateaspeckled immunofluo-rescence patternof interphase cells. In a mitotic Verocell, several bright fluorescent dots (ar-rows) are presentthat may represent nucleolar organizer regions. Weak nucleoplasmic staining is also present. Noteincreaseof nucleolar stain-ing in blast-transformed lymphocytes(large

cellsineandf). For comparison, the cells are also shownin phasecontrastoptics (b, d, f). Bar, 20AmX1,450.

bodies resultedinbindingwithonlyfibrillar structuresproducing

a"necklace" distribution of fluorescent dotsdispersed

through-outthenucleoplasm (Fig. 2,candd).

Electronmiscroscopywasperformed usingreference serum

S18 and rabbit anti-RNA polymerase Iantibodiesin the im-munogold label technique. In this experiment, gold-antibody complexes werelocalized in the fibrillar centers of nucleoli in hepatocytes from normal and regeneratingratliver.IgG from

sclerodermaserumS 18wasselectivelyenrichedoverthefibrillar centersandwas notobservedinthe dense fibrillarcomponent

orthe granularcomponent of nucleoli (Fig. 3, a andb). The same staining pattern was obtained with a rabbit anti-RNA

polymerase I serum(Fig. 3c)showingdistribution of

gold-an-tibodycomplexes in the fibrillar centers,ashad been reported previously (24).

Interaction

ofantinucleolar antibodies from scleroderma

sera

with

RNA

polymerase

I

Double immunofluorescence staining.

The

enzyme

RNA

poly-meraseIhas been shownbyotherinvestigatorstobelocalized

in nucleoli. The immunofluorescence patterns obtained with

thehuman scleroderma seradepictedinFigs. 1 and2,aswell asthe electronmicroscopicstudies inFig. 3,aremorphologically

ft

os

-f

f-.Ak.

b

.::S

'I:Ik Vf _{.~~~~~~~~~~~~~~} .1s$_U 4_a

..

Figure 2. Cultured Vero cells were treated withnucleolus-segregating drugsactinomycinD(a andb)or DRB(c andd)andprocessed for indirect immunofluorescence and phase-contrast microscopy. In acti-nomycin D segregated nucleoli, scleroderma antibodiesS18 selectively

stained the fibrillarregions,asshown in(a)of thiscomposite. Immu-nofluorescence staining is most prominent intwosmall crescents

(ar-similartothoseobtainedwith rabbit anti-RNA polymerase I antibodies(24,25).Itremainedtobedemonstrated thatidentical subcellularstructures werereactive with both antibodiesby si-multaneousstaining.Ahumanserumrepresentativeof speckled

nucleolar staining (Sl 8) and anti-RNA polymerase I rabbit

serum wereincubated simultaneously with Vero cells treated

withDRB. Theantigen-antibody reactionwasvisualizedonthe

samesubstratewitharhodamine anti-human IgG and fluores-cein anti-rabbitIgGconjugate.Inthisexperiment itwasobserved thatbothhumanantinucleolarandrabbitserumrecognizedthe

samenucleolar structures inlocationandconformation,so

pho-tomicrographs fromtheseexperimentswerevirtually superim-posable(data notshown).

Immunoprecipitation. We thenexamined the molecular

na-ture ofthe antigen reactive with human sera demonstrating speckled nucleolarstaining.

[35S]methionine-labeled

HeLa cell extract was used for immunoprecipitation. Nine sera with speckled nucleolarstaining pattern uniformly immunoprecipi-tated 13polypeptides(Pl-PI3)with mol wt of 210,000, 190,000,

155,000, 130,000, 120,000, 80,000, 64,000, 62,000, 42,000, 25,000, 18,000, 16,000, and 14,000. The results obtained with

tworepresentative sera (SI 8 and S 124) are shown in Fig. 4 a, lanes 2 and 3. Examination ofthe immunoprecipitates obtained

Ir'I

4,

4i

g ,' * f .sa

~ *W~ ~ A

V3,~~~~~~~

tt in .

phase contrast) granular component in the center of this compact nu-cleolus (b) is essentially unstained. In Vero cells treated withDRB,.

518 antibodies give a necklace type of staining dispersed throughout the nucleoplasm (c). The same cell is also shown by phase-contrast mi-croscopy (d). Bar, 20,umx 1,450.

by the rabbit anti-RNA polymerase I antibodies side by side with the human antinucleolarantibodies revealed the same 13 polypeptides by SDS-PAGE and autoradiography (Fig. 4, lane 4). Fig. 4 b shows the same polypeptides as Fig. 4 a, but better resolution of the high molecular weightpolypeptides Pl-P6 was

obtained in 5% SDS-polyacrylamide gel instead of 17.5% gel. Our control group, consisting of normal human sera and more

than 100 autoimmune sera from patients with systemic lupus erythematosus, MCTD,orrheumatoid arthritis did not contain antibodiesprecipitating thispolypeptide profile. Moreover, none of the sera from 120 SLE patients analyzedby

immunoprecip-itation in another study in this laboratory (manuscript in prep-aration), precipitated these 13polypeptides. On the other hand, this pattern of 13 polypeptides was consistently produced by nine sera with speckled nucleolar immunofluorescence. Sera with

other patterns of nucleolar reaction also did not produce this

finding.

:*A

lawaF ..(.

*-FC

8S

s<

*w

w

shows thataphosphoprotein ofmol wt35,000 (arrow), which was not labeled with [35S]methionine, was precipitated as a phosphoprotein by S18 antibodies. In this series ofexperiments,

alonger gel (15cm vs.8.5 cm) was used toidentify unambig-uously the32P-labeledproteins.Inmanystudies using acid-urea

gels(seeMethods),RNAwas notcoprecipitated with the RNA

polymeraseIcomplex, as opposed toanti-U1 RNPantibodies

from a patient with MCTD, which precipitated the UI-RNA

species(datanotshown).

Immunoabsorption. Todemonstratefurtherthatthe human antinucleolar and rabbit anti-RNA polymerase I antibodies

in-teractwith the same antigen,i.e.,RNA polymeraseI,weabsorbed

[35S~methionine-labeled

HeLacell extract with the IgGfraction of rabbitantiserum. We then used this RNA polymerase

I-de-pleted extract in animmunoprecipitationassay andshowed that thispreparationlackedtheantigen recognized by our prototype

antinucleolarserumS 18(Fig.6, lane 2). In areciprocalfashion,

when HeLa cell extract was absorbed with this human serum andthenprobed with rabbitantiserum,theRNApolymerase I

Figure 3. Bindingofscleroderma antinucleo-larIgGS18(aand

b)

andrabbit antibodiesto

RNApolymerase I(c)tosubnucleolar com-ponents in thehepatocytesofnormal(aand c)andregenerating (b)ratliver,asrevealed by the electronmicroscopic immunogold la-beltechnique.The 5-nmgold particles(small

blackdots)areselectivelyenrichedoverthe

fibrillarcenters(FC),roundishzonesof low contrast.The densefibrillarcomponents

ac

:. < (DFC) surroundingthefibrillarcentersas

*i + wellasthegranularcomponents(GC)of the

;e~ljy

nucleoliareessentiallyfree ofgoldparticles after incubation with both antibodies.(Ch),

chromatinsurroundingthe nucleoli.Bars,0.1

AmX 110,000(a,b),0.1 m X 135,000.

complex wasnotprecipitated (Fig. 6,lane 4). Incontrast, de-pletion ofradiolabeledHeLa cell extract byautoantibodies of

different immunological specificities,e.g., anti-Sm antibodies as in lane 5, did notaffect specificantigen recognition by antinu-cleolar scleroderma(Fig. 6, lane 6)orrabbit anti-RNA poly-merase IIgG(datanotshown). Thesefindingssupporttheearlier

observationsthat RNApolymerase Iistheantigen recognized bycertain antinucleolar antibodiesin sclerodermasera.

Microinjection

of

antibodies intoXenopus laevis oocytes.

Purified IgGfrom referenceserumS18 and controlantibodies weremicroinjected into X.

laevis

oocytenuclei. In these

exper-iments,aconsiderablereduction in the accumulation of newly

synthesized 28S and 18SrRNAwasobservedasshown inFig. 7, lane 4. This 1.5% agarose gel also shows that there was no

1 2 3 4 5 Figure4. (A) Autoradio-I P1 ] gramof

immunoprecipi-I si

P2 2 _tated

[35Sjmethionine-la-P3 97 4 beled HeLa cellproteins

en-1 \en5 N -69 resolved ina 17.5%

wm

do

P6 _ SDS-polyacrylamide

- = .= \P7 _ 46 gel. Normalhuman

P_

serum(lane1).

Repre-sentativeantinucleolar - -30 scleroderma sera(S18

-

-~~~P1

0 andS

₁₂₄₎

with

speckled

P11 staining pattern (lanes2

P12

and3,respectively). P13 Rabbitanti-RNA

poly--12.3 meraseIantibodies

A (lane 4).Normal rabbit

serum (lane 5). The rab-bitanti-RNA

polymer-P I

₂₀ aseIandantinucleolar

P2-

--

-200

scleroderma

antibodies

from two representative

p3 - _ patientsprecipitated the

P4 same 13polypeptides

P5- _ (Pl-P13)of 210,000 to 5

116.2

14,000molwt that were

distinct from

polypep-B

92.5

tides of 70,000, 46,000,

and30,000 molwt pre-cipitated by normal humanserum. Inaddition, control rabbit serum alsobrought downaproteinof80,000molwt.The46,000mol wt protein is actin, the otherpolypeptides precipitated by normal sera areunknown.(B)Autoradiogramofimmunoprecipitated

[35S~methionine-labeledHeLaproteinsresolved ina5%

SDS-poly-acrylamidegel. Antinucleolar sclerodermaserumS18(lane 1). Rabbit anti-RNA polymerase Iantibodies (lane2). Normal humanserum (lane 3).This low-percent gel systemwasusedtodemonstrate that dis-tincthighmolecularweight polypeptidesof molwt210,000 (PI),

190,000 (P2), 155,000(P3), 130,000 (P4), 120,000 (P5), and 80,000 (P6) wereimmunoprecipitated with the rabbit and the human antinu-cleolarserum.

However, theyinhibittranscriptionof thelampbrush

chromo-someloops(datanotshown).

Microinjection

of antibodies S18 into the cytoplasm (lane 3) also had no effect on rRNA

tran-scription since antibodies do notmigrate through the nuclear pores(36).

Several attemptswere made toidentifythe epitopeor epi-topesof the RNApolymerase Icomplex reactive with certain scleroderma autoantibodies. Extractable whole HeLa cell

pro-teins, isolated nucleoli aswell as purified RNA polymerase I

from Morris hepatoma 3942A, were electrophoretically

trans-ferredtonitrocellulose for immunoblotting.Atpresent, no con-clusive resultsconcerningtheantigenicsubunit were obtained, although a protein of mol wt 190,000 corresponding to one of thelarge subunits ofRNApolymerase I from Morris hepatoma 3924A has beenweakly reactive with some sera.

Discussion

Inthe present study, we have provided evidence by

immuno-cytochemical localization, immunoprecipitation, and

microin-jection of autoimmune antibodies intoliving cells that the RNA polymerase Icomplex was the antigen recognized by a subset ofantinucleolar antibodies from scleroderma sera. Anti-RNA

polymerase I autoantibodies were present in 4% of the 208

I

200-....

97.4-

-A

69-

30-

12.3-Figure 5.Autoradiogram of immunoprecipitated

2 3 4 [35S]methionine- (lanesIand 3) and

[32P]orthophosphate-la--P2 beled (lanes 2 and 4) HeLa cell *: proteins digestedwith RNase

andDNase. Normalhuman

-P6

serum (lanes

1

and 2). Antinu-sL

_.*

p7 cleolarserumSt18(lanes3 and *- 4). Thepolypeptides of molwt

190,000 (P2),80,000(P6), 64,000 (P7), and 18,000 (P 11) labeled with S-methionineare alsophosphoproteins.An addi-tionalphosphoproteinof mol wt35,000,notlabeledwith '[35Smethionine,isdetectedby

autoradiography(arrow). This

*_k-P

1 1 studyusedalongergel(15cm

'AW ' vs. 8.5 cm) toidentify unam-biguouslythe[32P]labeled pro-teins. ThePl-Pl 3 RNA poly-meraseIcomplex in thisand subsequentfiguresare identi-fied by black dots.

sclerodermapatientsexamined and were found to be restricted tothisautoimmunedisease.

RNApolymerase I is anenzymecomplextranscribingthe

genes that code for the precursor moleculesof ribosomal RNA (31,37).The distribution of thisenzymeininterphasecells and mitotic chromosomeswasrecentlystudiedbyScheer and Rose

(24)andScheeretal.(25).Theseinvestigatorsusedanantiserum against RNApolymerase Ifrom a rabbit immunized with the purifiedenzyme.Byimmunofluorescence microscopy,they

ob-tainedapunctate nucleolarstainingpattern ininterphasecells

fromdifferentspecies. Theysuggestedthat these punctateareas

representedtranscriptional complexesofRNApolymeraseIand

rRNAgenes. During mitosis,RNApolymeraseImoleculeswere

detectedatthechromosomal nucleolusorganizerregions.

Tran-1 2 3 4 5 6 Figure 6.Depletion of RNA

polymerase

I

from

if -200'

[35S]methionine-labeled

w

HeLacell extractby rabbit I'A-69 and humanantibodies.

RNApolymeraseI immu-...- noprecipitated byrabbit

an-tibodies(lane 1). Immuno-precipitate obtained with

MAd

-30

humanantinucleolar

anti-..

____

-30

bodies S18 (lane 2) with su-pernatant fromlane I

ex-periment. Immunoprecipi--12.3 tateobtained with :D_-1 2.3 antinucleolarantibodies

S18 with unabsorbed ex-tract (lane 3).

Immunopre-cipitateobtained with rabbit anti-RNA polymeraseIantibodies with supernatant fromlane 3

ex-periment.Anti-Smserumfromapatient with SLE showed a different immunoprecipitation profile (proteins A-G, lane 5). The supernatant from lane 5experimentstill contained the entire RNApolymerase I complex thatwasimmunoprecipitated with serumS18(lane 6).

1 2 3 4

Figure 7.Autoradiogram ofRNA fromX.laevisoocytesradiolabeled af-termicroinjection of various antibod-ies. Intranuclearinjection of mono-clonalIgM-antibodytoDNA(lane 2).

4 W -40 S

Injection

of_{purified IgG}from sclero-derma serumSI 8 intothecytoplasm (lane3) and nucleus (lane 4).

Intro-_ -28 S duction of the scleroderma antibodies S

₁₈

into theoocytenucleicauses a

~- *-. -18 S considerablereduction in the

accumu-lation of newlysynthesized 28S and 18SRNAs.40Sindicatesposition of the pre-rRNA.

scriptionally activerRNAgeneswerelocalizedbyelectron

mi-croscopic immunocytochemistry inthe fibrillar centers of

nu-cleoli.

Thespeckled nucleolarstainingpatternweobserved with a subset ofantinucleolarantibodies from sclerodermapatientswas

verysimilar to, if not identical with, the pattern described for therabbit anti-RNA polymerase I serum byScheerandRose

(24).Ananti-RNA polymeraseIserum fromarabbitwasalso used inaseries ofexperimentsinthisstudy. Double

immuno-fluorescence stainingshowed that humanspeckled antinucleolar IgG decorated the same nucleolar structuresastherabbit anti-bodies.Localization ofautoantibodies in thefibrillar regions of

drug-segregated nucleoliand, moreimportantly,their localiza-tionby electron microscopyinfibrillarcenters(24), whichare

thesites ofrRNAgenetranscription,gaveconfirmingevidence ofbinding with RNApolymerase I.Interestingly, these

sclero-derma antibodies produced nucleolar immunofluorescence stainingnotonly in human but also in rodent, avian,and

am-phibian cells, suggesting that the epitope is highly conserved

during evolution. Whenscleroderma antibodieswith speckled nucleolar staining in indirect immunofluorescence were

mi-croinjected intoX. Iaevis oocytes, a dramatic decrease in the

accumulation of newly transcribed ribosomal RNA was ob-served. The human antibodies thus appearto beeffective in

interferingwith the function ofRNApolymerase I.

Onamacromolecular level, bothsclerodermaantinucleolar

antibodies and rabbit anti-RNA polymerase I antibodies

im-munoprecipitated identical polypeptides rangingfrom mol wt 210,000 to 14,000. RNA polymerase I is a complex enzyme

previouslyreportedtobecomposed ofeightsubunits inMorris

hepatoma3942Aranging insize frommolwt190,000to17,500 (23). Byimmunoprecipitationwithspecific antibody,werevealed five additional [35S]methionine-labeled polypeptides present in HeLacell extract that had notbeen reported tobe part of the enzymepurified bycolumnchromatography fromrathepatoma.

Furtherinvestigationswilldetermine whether these additional

polypeptidesarepart oftheRNApolymeraseIcomplex reflecting

additional subunitsormerelyposttranslational modificationsof

previously described subunits. It is alsoconceivablethat some of theseadditionalpolypeptidesrepresenttightly bound factors

of thisenzymethatmaybeassociated withinitiation or

termi-nation ofrRNA genetranscription (38), Recently,

immunopre-cipitation of intrinsicallyradiolabeled DNA polymerase a with

amonoclonalantibody resulted in the identification of an

ad-ditional catalyticcomponentofthis enzyme that had been

pre-viously

unknown

(39).

It seems

_likely

that theuse of

specific

antibody for immunoprecipitation will also facilitate further characterization of RNApolymeraseI.

Ina previousstudy, all RNA polymerase I subunits from

rathepatoma could be phosphorylatedbyanendogenous protein

kinase associated with this enzyme (40-41). Autoradiography

ofin vivo 32P-labeled RNApolymerase I obtainedby immu-noprecipitationinthis study revealedonlyfourphosphoproteins (P2, P6,P7, andP11)aspart of thecomplex.Sincethedegree

ofphosphorylationof this enzymeparalleled transcriptional ac-tivityasshowninpreviousstudies(23, 40),thepresentfindings might reflect aless active form of RNA polymerase I readily

extractable fromHeLacells.Atpresent it is unknown howexactly

thephosphorylation isregulatedinvivo.

In previousreports byStetler and co-workers(40, 42), an-tibodies to RNA polymerase I were detected by radioimmu-noassay andbyimmunoblottingin 100%ofpatientswith SLE andMCTD, and in 78%ofpatientswith rheumatoid arthritis.

However, we were only able to immunoprecipitate the RNA polymerase I complex with scleroderma serathat produceda speckled nucleolarimmunofluorescence pattern and notwith any other sera,includingsclerodermaserawith differentstaining

patterns. Thediscrepanciesbetween the work of Stetler and

as-sociates and thisstudy mayberelatedtodifferences intechniques used. A critical assay used by the

previous

authors wasa ra-dioimmunoassayinwhichtheyemployeda

preparation

ofrodent

RNApolymerase I asthe antigen to coat the wells of

plastic

immunoassay plates.Inthe presentstudy,

immunocytochemical

localization ofantibody,

immunoprecipitation

of

intrinsically

labeledantigens,andmicroinjectionofantibodyinto

living

cells

werethe maintechniquesused. Theseconsiderations, however,

donotcompletely

clarify

the

discrepancies.

IfRNA

polymerase

I, however, is an enzymepredominantly localized in the nu-cleolus,ashasbeen

previously

shown

(24, 25, 31, 37),

it is difficult tounderstandwhyserafromSLE,

MCTD,

and rheumatoid

ar-thritispatients rarelydemonstratenucleolar

staining

inindirect immunofluorescence (15-17).Itwould bereasonabletoexpect that manySLE, MCTD, and rheumatoid arthritis sera would show nucleolar

staining

with the

high frequency

of anti-RNA

polymeraseIantibodies

reported

bythe

previous investigators.

Several

explanations

concerning these

discrepancies

are now

under

investigation

andtheyinclude the

possibility

thatepitopes recognizedbySLE, MCTD,and rheumatoidarthritisseramay

notbeaccessibletoantibodyin theindirect immunofluorescence or

immunoprecipitation

assays.

Theassociationofspecklednucleolar

staining

byserum

an-tibodies with the

diagnosis

ofsclerodermawasnoted inanearlier

study (17).Wehaveextended the studiestoshowthat the reactive

antigen

demonstrates

properties

very similartoRNA

polymerase

I. Ourknowledge of the function of this key enzyme in the nucleolus and the furthercharacterization ofthe

epitope

may lead to another step in

elucidating

the immune response in scleroderma.

Acknowledgments

WewishtothankThomasH.Hogerand Jan-M.Peters, ofthe German CancerResearchCenter,Institute of Celland TumorBiology, Heidelberg,

FRG, forskillfully performingthe electronmicroscopic

immunocyto-chemistry presentedinthisstudy.Wealso thank Dr.G.Rhodes and Dr. A.M.Francoeur forhelpful suggestions.

Rose). Dr. Reimer is supported by grant Re 585/1-1 from Deutsche Forschungsgemeinschaft. Dr. Rose is the recipient of a Research Career Development Award from the National Cancer Institute.

References

1. Tan, E. M.1982. Autoantibodiestonuclear antigens (ANA): their immunobiology andmedicine.Adv. Immunol. 33:167-240.

2.Lerner, M. R.,andJ. A.Steitz.1979.Antibodiestosmall nuclear RNAscomplexed withproteinsareproduced bypatients withsystemic lupus erythematosus. Proc.Nat!.Acad. Sci. USA. 76:5495-5499.

3.Francoeur, A.M., E. K. L.Chan, J.I.Garrels, andM. B.Mathews. 1985.Characterization andpurification of lupus antigenLa, an RNA binding protein. Mol. Cell. BioL.5:586-590.

4.Pizer, L. I.,J.-S. Deng, R. M.Sternberg, andE.M. Tan. 1983. Characterization ofaphosphoprotein associated with the SS-B/La nuclear antigeninadenovirus-infectedanduninfected KBcells. Mol. Cell. Biol. 3:1235-1245.

5.Mathews, M. B.,and R.M.Bernstein. 1983.Myositis autoantibody inhibits histidyl-tRNA synthetase: a modelforautoimmunity.Nature (Lond.). 304:177-179.

6. Douvas,A.S.,M.Achten, andE. M. Tan. 1979.Identification of anuclearprotein (Scl-70)as auniquetargetofhumanantinuclear

an-tibodies in scleroderma.J.BioL. Chem. 254:10514-10522.

7.Moroi,Y.,C.Peebles,M.J.Fritzler,J.C.Steigerwald,and E.M. Tan. 1980.Autoantibodytocentromere(kinetochore)inscleroderma sera. Proc.NatL. Acad. Sci. USA. 77:1627-1631.

8.Cox,J.V.,E. A.Schenk,and J. B.Olmsted. 1983.Human anti-centromere antibodies: distribution, characterization ofantigens,and effectonmicrotubuleorganization. Cell. 35:331-339.

9. Earnshaw,W.C.,N.Halligan, C. Cooke, andN.Rothfield. 1984. Thekinetochore ispartofthemetaphasechromosomescaffold.J. Cell Biol. 98:352-357.

10.Shero,J.H., B.Bordwell,N. F.Rothfield,and W.C. Earnshaw. 1986. High titers ofautoantibodies of topoisomerase I(Scl-70)insera

from scleroderma patients. Science(Wash. DC). 231:737-740. 11.Yang, V. W., M. R.Lerner,J. A.Steitz,andS.J.Flint. 1981.A small nuclearribonucleoprotein isrequired forsplicingof adenoviral earlyRNA sequences. Proc. Natl. Acad. Sci. USA. 78:1371-1375.

12. Lerner, M.R.,N.C.Andrews, G.Miller,and J. A.Steitz. 1981. Two small RNAsencodedbyEpstein-Barrvirusandcomplexed with proteinareprecipitated by antibodies frompatientswithsystemiclupus erythematosus. Proc.Natl. Acad.Sci. USA. 78:805-809.

13.Stefano,J. E. 1984.Purified lupusantigenLarecognizesan oli-gouridylatestretch common tothe3' termini ofRNA polymeraseIII transcript. Cell. 36:145-154.

14.Earnshaw,W.C.,K. F.Sullivan,W.Saunders,D.S.Cleveland, N. F.Rothfield. 1985. Identificationandmolecularcloning of CENP-B, a human centromereprotein recognized by allhumananticentromere antibodies (ACA). Clin. Res.33:59 la.(Abstr.)

15. Ritchie, R. F. 1970. Antinucleolar antibodies: their frequency anddiagnostic association.N.Engl.J.Med. 282:1174-1178.

16.Pinnas,J.L.,J. D.Northway, andE. M. Tan.1973. Antinucleolar antibodiesinhumansera.J.Immunol. 111:996-1004.

17.Bernstein,R.M., J.C. Steigerwald,and E. M. Tan. 1982. As-sociation of antinuclearandantinucleolar antibodiesinprogressive sys-temic sclerosis. Clin. Exp. Immunol. 48:43-51.

18.Reddy, R., E. M. Tan, D. Henning, K. Nogha, and H. Busch. 1983. Detectionof a nucleolar 7-2 ribonucleoprotein and a cytoplasmic 8-2 ribonucleoprotein with autoantibodies from patients with sclero-derma.J.Biol. Chem.258:1383-1386.

19.Hashimoto,C., and J. A. Steitz. 1983. Sequential association of nucleolar 7-2RNAwithtwodifferentautoantigens. J.Biol.Chem. 258:

1379-1382.

20. Lischwe,M. A., R. Reddy, R. L. Ochs, L. C. Yeoman, E. M. Tan,M.Reichlin,and H. Busch. 1985. Purification of a nucleolar sclero-dermaantigen(Mr=34,000;pI,8.5) rich inN0,

N0-dimethylarginine.

J.Biol. Chem. 260:14304-143 10.

21.Ochs,R.L.,M. A.Lischwe,W. H._Spohn,and H. Busch. 1985. Fibrillarin:a newproteinofthenucleolus identifiedby autoimmune

sera.Biol. Chem. 54:123-134.

22.Masi,A.T.,andG.P. Rodnan. 1981. Preliminarycriteria for theclassificationofsystemicsclerosis(scleroderma).Bull. Rheum. Dis. 31:1-6.

23.Rose,K.M.,D. A.Stetler,andS. T. Jacob. 1981. Protein kinase

activityofRNA polymerase I purified from rat _{hepatoma: probable}

functionof Mr42,000and46,600polypeptides.Proc.NatL.Acad. Sci. USA. 78:2833-2837.

24.Scheer, U., andK.M. Rose. 1984.Localization of RNA

poly-meraseIininterphasecells and mitotic chromosomes

by

light

and electron microscopicimmunocytochemistry.Proc.NatL.Acad. Sci.USA.

81:1431-1435.

25.Scheer,U.,B.Hugle,R.Hazan,and K. M. Rose. 1984. Drug-induceddispersaloftranscribedrRNA genesandtranscriptional products:

immunolocalizationandsilverstainingof different nucleolarcomponents in rat cellstreated with

5,6-dichloro-,B-D-ribofuranosylbenzimidazole.

J.Cell BioL. 99:672-679.

26. Liu,S. L., and K. M. Rose.

Lipopolysaccharide

mediated in-duction ofRNApolymeraseI_activityandamountin murine

B-lym-phocytes.J.Biol.Chem., in press.

27. Reimer, G., R.L. Rubin,B. L.

Kotzin,

andE. M. Tan. 1984. Anti-nativeDNAantibodies from autoimmuneseraalso bindtoDNA inmitochondria. J. ImmunoL. 133:2532-2536.

28.Francoeur,A.M.,and M. B.Mathews. 1982. Interaction between VARNAandthelupusantigenLa:formation ofa

ribonucleoprotein

particleinvitro.Proc.NatL.Acad.Sci. USA.79:6772-6776.

29.Laemmli,U. K. 1970.Cleavageof structural

proteins during

the

assemblyof the head ofbacteriophageT4. Nature

(Lond.).

227:680-685. 30. Scheer, U., H. Hinssen,W. W. Franke, andB. M. Jockusch. 1984. Microinjection of

actin-binding

proteins

and actin antibodies demonstrates involvement of nuclear actin in

transcription

of

lampbrush

chromosomes. Cell. 39:1431-1435.

31. Busch, H., andK. Smetana. 1970. The Nucleolus. Academic

Press, Inc.,NewYork. 626pp.

32.Bernhard,W. 1971.Adv.

Cytopharmacol.

1:49.

33.Simard, R.,Y.

_Langelier,

R.

Mandeville,

N.

Maestracci,

andA.

Royal. 1974. The Cell Nucleus. AcademicPress, Inc.,NewYork. 3:447. 34.Granick,D. 1975.Nucleolar necklaces in chick

embryo

fibroblast

cells.I.Formation of necklacesbydichlororibobenzimidazole and other adenosine analoguesthatdecrease RNA

_synthesis

and

degrade

preri-bosomes.J. Cell Biol. 65:398-417.

35.Granick,D. 1975. Nucleolarnecklacesinchick

embryo

fibroblast cells. II. _Microscopeobservations oftheeffectof adenosine

analogues

onnucleolarnecklaceformation.J. CellBiol.65:418-427.

36. Einck, L., and M. Bustin. 1984. Functional histone

_antibody

fragmentstraversethe nuclearenvelope.J. Cell Biol. 98:205-213.

37. Rose,K.M., D. A.Stetler,and S.T.Jacob. 1983.Enzymesof Nucleic AcidSynthesisand Modification. CRCPress,BocaRaton,FL. 2:43-74.

38.Wandelt,C.,andI.Grummt. 1983.Formation of stable

preini-tiationcomplexesisa

_prerequisite

for ribosomalDNA

transcription

in vitro.Nuc!.AcidsRes. 11:3795-3809.

39. Karawya, E.,J.Swack,W.Albert,J.Fedorko,J. D.Minna,and S. H. Wilson. 1984. Identification ofa

_higher

molecular

weight

DNA polymerasea

_{catalytic polypeptide}

inmonkeycellsbymonoclonal

an-tibody.Proc.Natl. Acad.Sci. USA. 81:7777-7781.

40.Stetler, D.A.,and S. T. Jacob. 1984.

Phosphorylation

ofRNA

polymeraseIaugments its interaction with autoantibodies ofsystemic

lupus erythematosuspatients.J.Biol. Chem.259:13629-13632.

41.Duceman,B._W.,K. M.Rose,andS.T.Jacob. 1981.Activation ofpurifiedhepatomaRNApolymeraseI_{by homologous}

_protein

kinase NII.J.Biol. Chem. 256: 10755-10758.