Molecular definition and sequence motifs of the 52 kD component of human SS A/Ro autoantigen

Molecular definition and sequence motifs of the

52-kD component of human SS-A/Ro

autoantigen.

E K Chan, … , J P Buyon, E M Tan

J Clin Invest.

1991;

87(1)

:68-76.

https://doi.org/10.1172/JCI115003

.

Serum SS-A/Ro autoantibodies are commonly found in patients with Sjogren's syndrome,

systemic lupus erythematosus, neonatal lupus, and subacute cutaneous lupus. Two

proteins of 60 and 52 kD have been described as targets for these autoantibodies. To define

the 52-kD component unambiguously, cDNA clones were isolated from human HepG2 and

MOLT-4 cell cDNA libraries. The identity of cDNA was established by (a) the specificity of

the antibody affinity purified from the recombinant protein, (b) the reactivity of the purified

recombinant protein with prototype SS-A/Ro sera in immunoblot and ELISA, and (c)

two-dimensional gel comigration of MOLT-4 cell 52-kD protein and the recombinant protein. A

1.9-kb cDNA encoded the complete 52-kD protein containing 475 amino acids (Mr 54,082).

Putative zinc-finger domains and a leucine zipper motif were identified in the amino-terminal

half of the 52-kD protein, implicating its possible association with DNA/RNA. Sequence

homology detected between the 52-kD protein and human ret transforming protein, and

mouse T cell gene expression down-regulatory protein rpt-1, may provide leads to the

functional role of the 52-kD protein in addition to the possibility that these proteins might

constitute members of a subfamily of finger proteins.

Research Article

Molecular Definition

and

Sequence

Motifs

of the 52-kD Component of

Human

SS-A/Ro

Autoantigen

EdwardK. L.Chan, JohnC.Hamel, Jill P. Buyon,* and Eng M. Tan

The W. M. Keck AutoimmuneDisease Center, Department of Molecular andExperimental Medicine, Scripps Clinic andResearch

Foundation, La Jolla, California 92037; and *HospitalforJoint DiseasesOrthopedicInstitute, New York 10003

Abstract

SerumSS-A/Roautoantibodies are commonly found in pa-tients with Sjogren's syndrome, systemic lupus erythematosus, neonatal lupus, and subacute cutaneous lupus. Two proteins of

60and52 kD have beendescribedastargets for these autoanti-bodies.Todefinethe52-kDcomponentunambiguously,cDNA clones were isolated from buman HepG2 and MOLT4 cell cDNAlibraries.Theidentity of cDNA was established by(a)

thespecificity oftheantibody affinity purified fromthe

recom-binant protein, (b) the reactivity of thepurified recombinant

protein with prototype

SS-A/Ro

sera in immunoblot and

ELISA, and(c)two-dimensionalgelcomigration ofMOLT-4

cell 52-kD protein and the recombinant protein. A 1.9-kb

cDNAencoded thecomplete 52-kD protein containing 475 aminoacids

_(M,

54,082). Putativezinc-finger domainsand a leucine zippermotifwereidentifiedin the amino-terminalhalf ofthe52-kDprotein, implicatingitspossible associationwith

DNA/RNA. Sequence homology detected between the 52-kD

proteinand human rettransforming protein,and mouse T cell gene expression down-regulatory protein rpt-1, may provide

leads to thefunctionalroleofthe52-kDproteinin additionto

thepossibilitythat theseproteins might constitutemembersof

a subfamily offinger proteins. (J. Clin. Invest. 1991.

87:68-76.) Keywords:

autoantibody

*

Sjogren's syndrome

-zinc

finger

protein-systemic lupus

erythematosus

*

autoimmunity

Introduction

Themolecular

characteristics

of

SS-A/Ro'

antigensareof spe-cialclinical interestbecause of the relatively high frequency of

Preliminaryreportsinabstract formwerepresentedattheFirst Interna-tional WorkshopontheMolecular and CellBiologyofAutoantibodies andAutoimmunity, Heidelberg, Germany, 27-29July 1989,andat

theWorkshoponStructure and Function ofEukaryotic RNP,Delphi,

Greece,1-5April 1990(1990. Mol. Biol.Rep. 14:53[Abstr.]).

AddressreprintrequeststoEdwardK. L.Chan, Ph.D.,W.M.Keck Autoimmune Disease Center, ScrippsClinic and Research

Founda-tion, 10666N.Torrey PinesRoad,LaJolla,CA 92037.

Theoriginalnucleotide and amino acidsequencedatahave been submittedtoEMBL/Genbankunder accessionNo.M35041.

Receivedfor publication13 June 1990 andinrevisedformIAugust 1990.

1. Abbreviations used in thispaper:

CD4,

clusterdesignation4;IPTG,

isopropyl-,B-thiogalactopyranoside; PBS-T, phosphatebufferedsaline with 0.05% Tween-20; SS-A/Ro,nuclearantigenAofSjogren's

Syn-drome; SS-B/La,nuclearantigenBofSjogren's syndrome.

theSS-A/Ro autoantibodyinpatients withseveral rheumatic

diseasesincluding systemic lupuserythematosus, Sjogren's

syndrome, neonatal lupus syndrome with congenital heart

block,and subacute cutaneous lupus (for current reviews, see

references 1-3).Variousmolecularspecies rangingfrom 50 to 150 kD have been described for theantigenic component(s)of SS-A/Ro (summarized in reference4). In 1984 Wolin and

Steitz(5)showedthatautoantibodiestoSS-A/Ro immunopre-cipitated small ribonucleoprotein particles composed of hY-RNAs and a 60-kD protein component. The 60-kDprotein

wasdefinedas themajor autoimmune target because SS-A/Ro

autoimmune seradid not recognizehY-RNAs alone. Recent workofBoire and Craft (6) reported a subpopulation

ofautoan-tibodies recognizinganantigenic epitope restrictedtothe intact ribonucleoprotein particle composed ofhY5 RNA andthe 60-kDSS-A/Roprotein. Besides the60-kDprotein, Ben-Chetritet al. (4) found that most SS-A/Ro-positive sera as defined by immunoprecipitation in Ouchterlony double diffusion assay also recognized another protein of 52 kD. It was not clear whetherthe 52-kDprotein interacted with hY-RNAsdirectly

orvia theassociation withthe60-kD proteinthat was known to

have in vitro RNA-binding activity(5, 7).Indirect immunoflu-orescenceusing these specific anti-52-kD reagentsgave nuclear punctatestaining similartothoseobservedfor antibodies

spe-cifictothe60-kD SS-A/Rocomponent (4). It was also shown

by antibody affinity purification that patient sera contained distinct, noncross-reacting antibodies directed to the 60- and 52-kD components(4). Autoantibodiesto the 52-kD compo-nent wereinfactvery common in SS-A/Ro autoimmune sera

analyzed(>80%) and the reason the 52-kD component had notbeen detected earlier wasprobably a result of coexisting autoantibodiestoSS-B/La and to theincomplete separation of

the 47-kDSS-B/La antigen fromthe52-kDcomponent in the usual gel separation system(8). These twoSS-A/Ro compo-nents werefound in allhumancelllinestestedincludingHeLa,

MOLT-4, Raji, and Wil-2(4). Rader et al. (9)alsodescribed

twohumanSS-A/Ro species of60 and 52 kDfrom

lympho-cytesand, in

addition,

tworedcell-specific SS-A/Ro compo-nentsof60 and 54 kD that wereimmunologicallyrelated to the lymphocyte 60- and52-kDspecies, respectively.Todefinethe 52-kD componentof

SS-A/Ro

protein unambiguously, we haveclonedthefull-lengthcDNAencodingthe 52-kDprotein. This report describes the isolation and characterization of

cDNAclonesusinga human autoimmune serumwithspecific autoantibody to the 52-kD SS-A/Ro protein and thespecial features ofthe deducedpolypeptidethat

might

provide

insights

intothefunctionalroleoftheprotein. Methods

Cellextracts. MOLT-4 cells(humanT celllymphoblastic leukemia, ATCC CRL1582)werecultured in DMEMcontaining10% calfserum

J.Clin. Invest.

© The AmericanSocietyforClinicalInvestigation,Inc.

0021-9738/91/01/0068/09 $2.00

at370Cinan8%CO2 incubator. Cultureswere

supplementedwith

2.5

Ag/ml

gentamicin sulfate and maintainedat106 cells per ml. Cellswere

harvested andextracted in bufferA(150mMNaCl, 10mMTris-HCl,

pH7.2,0.5% NonidetP-40) and freed ofcell nucleibycentrifugationat 12,000 g for 15 min.

Immunoblotting. For high-resolutionseparation of the 60-kD, 52-kD,and47-kDSS-A/Ro-SS-B/Laproteins,cellextracts wereseparated

by PAGEasdescribed by Laemmli (10)using 15%gelslabs(20X 13 X0.1 cm) and theratioof acrylamide:bis-acrylamidewaschangedto 172.4:1 in theseparating gel (8).Itwasshownrecently that the standard gelcomposition used by many laboratorieswas notoptimal for the separation of the 52-kDSS-A/Ro and 47-kDSS-B/Laproteins (8).

Proteins weretransferred from gelontonitrocelluloseusinga constant voltageof 60Vfor 2h(11). The nitrocellulose sheetwasair-dried,cut intostrips, and incubated for 30 min in 3% nonfat milkdiluted in PBS toblocknonspecific binding sites. The nitrocellulosestripswere incu-bated with 1 mlofa 1:100dilutionofseraforI hand then washed in PBSwith 0.05%Tween20(PBS-T).'25I-proteinA(ICNBiochemicals,

Irvine, CA) was used to detect bound human antibodies. Thefollowing protein standards(Bio-Rad Laboratories,Richmond, CA)wereusedto determine molecularweights: phosphorylaseB,92,500; BSA, 66,200; ovalbumin,45,000; carbonic anhydrase,31,000; soybeantrypsin inhib-itor, 21,500; andlysozyme, 14,400.

Screeningofphage librariesfor SS-A/RocDNA.AHepG2 cellAZap

cDNAlibrarywas akindgiftfrom Dr. Frank R.Jirik, Universityof BritishColumbia, Vancouver, BC, Canada.Ahightiter SS-A/Ro hu-man serum Bowasusedfor immunoscreening of106recombinants.

'25I-proteinAwasusedtodetect bound humanantibodies. One

posi-tive clone Clwasidentified aftermultiplescreeningsandwassubcloned invivo into pBluescript plasmidpCIusing R408 helper phage (Strata-gene, LaJolla, CA)asrecommended in themanufacturers' instruc-tions.

HumanTcelllymphoblastoma(MOLT-4)Xgt-llcDNAlibrary was constructedby Dr. K. Ogata andDr. D.J.Noonan,ScrippsClinic and ResearchFoundation. The MOLT-4library was screened for full-length SS-A/RocDNA clonesby DNAhybridization. Two partially complementary synthetic oligonucleotides

(5'-TGTGCAGTGCAT-GGAGAGAGACTTCACCTG-3' and 5'-TCTTTCTCACAGAAC-AGGTGAAGTCTCTCT-3') were designed basedonthe 5' sequenceof

pClcDNAinsert.Theyweremixed and labeled with [a

32P]-ATP

using

the standardfill-in reaction of Klenow polymerase (12).Allscreenings

werecarried out with duplicate filters and positive phages were plaque

purified.Theresulting cDNAsweresubclonedintopBluescriptvectors (Stratagene) for further analysis.

Cloningofrfpprotein.Twopartiallycomplementarysynthetic oligo-nucleotides (5'-TGGCCTCCGGGAGTGTGGCCGAGTGCC-3' and

5'-GGTGGTCTCCTGCTGCAGGCACTCGGCC-3') were designed basedonthe 5' sequence of therfpcDNA reported(13).Theywere mixed and labeled with

[a32P]-ATP

using the standard fill-in reaction of Klenowpolymerase(12)and used to screen 106 recombinant from theMOLT-4 cellcDNAlibraryasdescribed above. Twoinitial posi-tivesweredetected andoneof these, 2rl,waspurified and the cDNA insertsubclonedtopBluescript SK plasmidfor further analysis. A rab-bitanti-rfpreferenceserumwaskindly provided by Dr. M. Takahashi, Aichi Cancer Center Research Institute, Nagoya, Japan.

Affinity-purified

antibodies. XZapclone Cl phage plaques were in-ducedtoproduce recombinant protein by overlaying with

isopropyl-fl-thiogalactopyranoside (IPTG, Sigma Chemical Co., St. Louis, MO) saturatednitrocellulose filters and grown overnightat37°C. These ni-trocellulose filterswerewashed with PBS-T, incubated with diluted serumfor 1 h, washed again with PBS-T before elution of bound anti-bodies with0.1%BSA in0.1Mphosphatebuffer, pH 2.2.

Affinity-puri-fiedantibodieswereimmediately neutralized by the addition of 1 M Tris-HCI, pH 8.8. The antibodies were concentrated with Centricon

100microconcentrators (Amicon Corp., Danvers, MA).

Purificationof recombinant protein. Plasmid pCI was transformed inEscherichia coli strain XL1-Blue (Stratagene). A 200-ml culture of

therecombinant cellswasgrownto

OD16w

=0.6at370C and IPTGwas

addedto afinal concentration of 10 mM. The culturewasallowedto

resumegrowthovernightbefore

harvesting by

centrifugation. Purifica-tionof recombinantproteinwas_performedasdescribedbyAdametal.

(14). Thefinalpelletwasextracted in 8 Murea for 15 min and the supernatantwasstoredat-70'C in_aliquots.

ELISA.Standardprotocolfor ELISAwas

_employed

asdescribedby

Rubin(15). Purified recombinant

_proteins

in 8 Mureawerediluted

1,000-64,000-foldin PBS forcoatingImmulon 2microtiter plates

(Dynatech

Laboratories,

Inc., Alexandria,

VA).

Peroxidase-conjugated

goat anti-humanIgG+Mreagents(CaltagLaboratories,So. San

Fran-cisco, CA)and the substrate 2,2'-azinobis(3-ethylbenzthiazoline sul-fonic

acid) (Boehringer

Mannheim

GmbH,

Mannheim, FRG)were

usedas

detecting

agents.

Southern

analysis.

For

determining

thecomplexityof thegene

en-codingthe52-kDSS-A/Ro

protein,

human

_genomic

DNAsamples (10

iig/lane) from HeLa cells and normal peripheral bloodweredigested with EcoRIorHindIIIrestriction enzymeto

_completion

and

analyzed

essentially

according

toSouthern(16)

using

labeled

pC1

cDNA insert. In vitro RNAtranscriptionand translation.TheEcoRI

Xgtll

cDNA insertof 52FLwassubclonedinto

pBluescript

SKvector.The

_resulting

plasmidp52FLwith the cDNA insert in thesameorientationas

fl-ga-lactosidasewaslinearizedbyrestriction enzymeClaIatthe3'endofthe insert.RNAwastranscribed in vitro from the linearized

plasmid

using

T3RNApolymeraseandwastranslated in vitro inarabbit

_reticulocyte

lysate(Promega,

Madison,

WI) in the presence of

[35]-methionine

(Tran35S-label,

70%methionine and15%

_cysteine,

ICN

Biochemicals)

asdescribed in the manufacturers' instructions. Products of in vitro translationwerestored in

_aliquots

for

immunoprecipitation

(17, 18)

and two-dimensional

gel

analysis ( 19).

DNAsequencingand

_analysis.

_{The sequencestrategy}wasto

com-bine data derived from restriction

fragments

subclonedin

_pBluescript

vectorand theuseof

synthetic

oligonucleotides

for

_primers.

Plasmids

werepurifiedbytheChen andSeeburgmethod

(20)

andDNA

sequenc-ingwasperformed bythe

dideoxy

chaintermination method of

_Sanger

etal.(21).DNAand

_protein

sequenceswere

_{analyzed by}

theGenetics

ComputerGroup Sequence

Analysis

Software

_Package

for VAX

com-puters(22).

Alignment

of

protein

sequenceswas

_initially

achieved with the GAP program that

employed

the

algorithm

of Needleman and Wunsch(23). Multiplesequence

alignments

were

performed

with

CLUSTAL programs

(24, 25).

Results

Cloning

andcharacterization

of52-kD

SS-A/Ro

cDNAs.

Inthe

screening

of106 recombinant

phages

from the

XZap

HepG2

cell cDNA

library,

cloneC1wasselected

by

humanserumBo

andwas

purified

withthreeroundsof

plaque

purification.

To show that Cl encoded the 52-kD

SS-A/Ro

protein,

purification

of

affinity antibody

was

performed

with Cl

phage

protein

anda

differenthumanautoimmuneserumCa that

_recognized

both

SS-A/Ro

and

SS-B/La

proteins

in MOLT-4 cellextracts

(Fig.

1,

lane

2).

Affinity purified antibody

fromtheCl

phage

protein

was

specific

in

recognizing

the 52-kD

SS-A/Ro

protein

alone

(lane

4).

This

specific

reactivity

was consistent with our

previous finding

that human

affinity purified

52-kD

anti-bodies didnot

recognize

either the 60-kD

SS-A/Ro

or

_SS-B/La

proteins (4).

The

XZap

cloneClwassubcloned invivo into

pBluescript

plasmid

pCl.

Recombinant

protein

derivedfrom

pCl

grown in XL1-Bluebacteria had amolecular mass ofn

46,000

as de-tected

by

immunoblotting

and was

purified

as described in Methods. The final

preparation

was a

highly

antigenic

Figure 1. Immunoblotting analysis of affinity purified antibodies from XZap Cl-derived fusion protein. Molt-4 cell extractswere separated onSDS-PAGE using a20-cm 15% separating gelasdescribed in Methods.Aftertransfer tonitrocellulose filter, strips were cut and probedwith different humanantibodies: lane 1, a normal human serum; lane 2, human

-60

serumCa with weak reactivity to the 60-kD SS-A/Ro protein and

-

2

strongreactivities to

both the 52-kDSS-A/

Ro and 47-kDSS-B/La proteins; lane 3, Ca

-47 serumabsorbedwith Cl phageproteins shows adrop in intensity of the 52-kDband but no effect on the reactivity toSS-B/La; lane 4,

antibody from Ca serumaffinity purified from Cl phage protein shows very strong reactivity with 52-kD

SS-A/Roalone; lane 5,serumBo, theoriginalserumused in the

cloningof Cl. The results show thataffinity purified antibodyfrom Cl recombinantproteinrecognized onlythe52-kDSS-A/Ro protein.

ELISA format (Fig. 2 A). Even with the highest dilution of

1:64,000forcoatingmicrotiterwells, an

OD410

reading of 2 was obtained 1 haftertheaddition ofdetecting reagents. Fig. 2 B

illustratesthespecificity ofthe ELISA when a1:30,000dilution of thepurifiedrecombinantantigenwas usedforthe detection ofSS-A/Ro antibodies. Recombinant60-kDSS-A/Roand SS-B/Laproteins derived fromcDNAclonespreviouslyreported

(17, 26)werepurifiedsimilarlyand areincludedinthis figure

forcomparison. High OD readingsforthe 52-kD SS-A/Ro

re-combinant proteinwereonlyseeninserawith detectable anti-bodies by

immunoblotting.

Our data indicated that in most

anti-SS-A/Ro seratested, thereactivities asdeterminedby ELISA werehigher forthe 52-kD than the 60-kDrecombinant antigens. Besidesthenegativecontrolsshownin Fig. 2 B, 20 normalhumanseraand standard prototypeautoimmunesera were also negative including antibodies to PCNA, Sm, U

1-RNP,ribosomalRNP, Jo- 1,Scl-70, SL/Ki,and Kuspecificities

(datanotshown).

The

immunological

dataabove suggested thatpCl encoded

atleastoneimmunoreactive region ofthe 52-kDSS-A/Ro

pro-teinandthenextstep was todetermineif the cDNAinsertof

pCl

encoded the entire 52-kD

protein.

Plasmid

pCl

was di-gested withEcoRIrestrictionenzymeandacDNAinsertof1.5 kb wasdetectedafterelectrophoresisin astandardagarosegel.

Northern blot analysis of the mRNA from MOLT-4 and HeLa cellsusing radiolabeled 1.5-kb fragment showed that therewas a single band of - 1.9 kb. Thisindicated thatpCl probably

contained onlyafractionof the complete cDNA for the 52-kD SS-A/Ro protein. To obtainafull-lengthcDNA, 106

recombi-nant phages from the MOLT-4 cell cDNA library were

screened witharadiolabeled DNA probe correspondingtothe 5' sequence ofpCIcDNA insert. Of the twoclones selected,

clone 52FL had the longest insert of - 1.9 kb andwas

sub-clonedinto pBluescriptplasmidp52FLfor furtheranalysis.

Todetermine if p52FL cDNA insert encoded thecomplete

52-kD protein, comparison was made between the in vitro

transcriptionand translationproductsofp52FLand the52-kD SS-A/Ro protein from MOLT-4 cells. Preliminary data showed that they comigrated in standard one-dimensional gel SDS-PAGE, and therefore two dimensional gels were used for morediscriminating analysis (Fig. 3). In vitro [S35J-methionine labeled translation products of p52FL RNA gaveamajor spot (arrow) and several minorspotsthatwereobservedonly inan

overexposed autoradiogram (Fig. 3 A). Immunoprecipitation of these translation products with human SS-A/Ro sera

showed that the major 52-kD product was immunoreactive. Immunoblotting of MOLT-4 cell proteins withserumCa de-tected multiple SS-B/La isospeciesofdifferentpImigratingat

OD A

14-12

-10

- 6- 4-2-t

04

1:2000 1:8000 1:32000

NHS Ge

OD E

6

NHS Ze Ge Bo Ca Wr Be We

M SS-A/Ro52kDa

Figure 2.ReactivityofpurifiedrecombinantproteininELISA.

Recombinant proteinderivedfrompCI plasmidwaspurifiedas describedin Methods.(A)Titrationcurveof the reactivities(OD410)

ofaprototypeSS-A/RoserumGe andanormal humanserumwith

recombinantprotein dilutedfrom 1:1,000to 1:64,000. (B) Reactivities of humanserawithrespecttodifferentpurified recombinantproteins.Recombinant 60-kDSS-A/Roand 47-kD

SS-B/Laproteinswerepurifiedfrom cDNA clonespreviously reported (17, 26).Serum Ze is the CDC prototypeanti-SS-B/La,seraGe and

Boaremostlyreactive with52-kDSS-A/Ro, seraCa, Wr,Be

recognizedbothSS-A/RoandSS-B/La,andserumWe is theCDC prototypeanti-Sm.

1

2345

- SS-A/RoBOkDa E SS-B/La47kDa

ZIM

a .

B

ay.'I.;i.

IEF > Figure 3. Autoradiogramsof

two-dimensional gels

C/) separating MOLT-4 cell

proteinsand invitro

C/3

| translation

products

of

_p52FL

-60

RNA transcript. (A) Separation of[3S1-methioninelabeled in vitrotranslationproductof

47

p52FL RNA. Anoverexposed autoradiogramof the driedgel

*

_a

is shown toindicate themajor

52-kD spot(arrowhead)and other minorspecies. (B)The

immunoblotof MOLT-4cell

proteins with serumCa,

B

hd showing strong reactivityis with

SS-B/Laisospecies, migrating at 47kD and isoforms of the 60-kDSS-AfRo proteins. In

4 shcontrastthe reactivity with 52

kD SS-A/Ro protein is restricted toa_major

_spot

(arrowhead)withapIof- 6.5.

Theplestimated bythe ISOELECTRIC program' (22)

basedontheproteinsequence

is 6.35.Arrowsmarked aandt

correspondtothepositions of major cellproteinsactin andtubulin, respectively.Todetermineunambiguouslythegelmigration positionsof the 52-kD SS-A/Roprotein and the recombinantprotein,amixtureof MOLT4 cellproteinsand the[35S1-methioninelabeled invitro translation productofp52FLRNAwas separatedonasingle gel and thentransferredtonitrocellulose filter. (C) 4-d exposure of the labeled[35S]-methionine

onthenitrocellulose filter. (D) 12-h exposure withintensifyingscreenfor thesamenitrocellulose afterimmunoblottingwith humanserumGe,

which is specificfor 52-kD SS-A/Roprotein,and["'I]-proteinA. Themajor spots(arrowhead)in C andDmatchtothesameposition showing thatp52FLencoded the entire 52-kDproteinofSS-A/Ro.

47 kD and several isoforms ofthe 60-kD SS-A/Ro proteins

(Fig.3 B). On the other hand, the 52-kD SS-A/Roprotein was

restrictedto amajorspotwithanestimated pI 6.5(Fig.3 B,

arrowhead). The colocalization of the p52FL encoded protein and the 52-kD SS-A/Ro proteinwas determined unambigu-ously byrunning bothsamplesinthe samegel(Fig. 3, C and

D). These independent results provided further confirma-tion thatp52FLcDNA encodedthe entire 52-kDprotein of

SS-A/Ro.

Southern analysis. Human genomic DNA samples from

HeLa cells and normal peripheral blood were digested with EcoRI orHindIII restrictionenzyme tocompletion.Either re-striction sitewas not found in the complete 52FL sequence. Results ofthe genomic blot probed with Cl cDNA insert showed a

major

bandof 5 kb andaminor band of 8 kb

(HindII)and = 20 kb (EcoRI)(Fig. 4). These dataindicated

that therewasprobably onlyoneandat mosttwo genesforthe

52-kD SS-A/Ro protein inthe human genome.

Sequence

analysis.

Thenucleotide sequences of the cDNA inserts of Cl and 52FL weredetermined bysequencing both DNAstrands asoutlined in Fig.5. CloneClrepresented the 3'

80% of clone 52FL and therewas nodifference detected be-tweenCl and 52FL clones in thecorrespondingregions. The DNA sequenceof52FL and the deduced amino acid sequence areshown inFig. 6.Thereisonly one large openreading frame composed of 475 amino acids correspondingto acalculated molecularmassof54,082. The sequenceflankingthefirstATG

codon is in close agreement with the eukaryotic translation

initiationconsensus sequence (27). In the 5'noncoding region,

there are three stop codons thatare in-frame with the ATG

startcodon(Fig. 6).Twopolyadenylation signalsequencesare

found at the endof the3'noncodingsequence. Theconsensus

sequence for nuclear localizationsignalhas not beendetected.

Specialfeatures of the 52-kD SS-A/Ro proteinare

summa-rizedinFig.7.Theinteresting featuresarethe amino-terminal

zincfinger domains, a centraldomain with a leucine zipper motif,and acarboxy-terminal domainthatishighly similarto

humanprotein

rpi.

Fig. 8showstheputative zinc-finger struc-tures inthe 52-kD

protein.

ThearrangementoftheCys resi-dues for the chelation of metal suchaszinchas been found in

many DNA/RNA-bindingproteins (28, 29) and thesezinc

fingerstructures aregenerally thoughttoberesponsible forthe interaction with either DNA or RNA. Theproposed finger structures inFig. 8 are atypicalascompared to most known

finger proteinsequences(28, 29).Thereisnopublisheddata at the present toindicate whether the

SS-A/Ro

proteinbinds zinc orother metal ions.However,asinglezincfingermotif has also been described in the 60-kDSS-A/Ro protein (7, 17).A leucine

zipper motif (residues 211-232)wasdetectedin the center

do-main ofthe 52-kD protein. Leucine zippers were originally

described inDNA-binding proteins (30)andnowthesemotifs

areknown toparticipate in protein-protein interaction or

dimer formation that isimportantforDNA-binding (31-33). This motifin52-kD

SS-A/Ro

may beimportantfor interaction

withotherproteinsorfor homo-dimerformation.

1

₂

₃

too

-23

-9.4

-6.6

_

-4.4

-2.3

-2.0

Figure 4. Southern blot analysis of human genomicDNA samples probed withClcDNA insert. Lane 1, HeLa cell

1.4

DNAdigestedwith EcoRI; lane 2, normalhumanperipheral blood DNA digested withEcoRI;lane3, human peripheral bloodDNAdigestedwith HindIII restrictionenzyme. The results suggestonlyone or two genes encoding the 52-kDSS-A/Ro protein.

the 52-kD protein, homology searchwith other known

pro-teinswasperformed withtheWORDSEARCHprogram avail-able from theGenetics Computer Group Sequence

Analysis

Software Package (22). Three

proteins

were identified with

high degree of homologytothe52-kD

protein

andtheywere

the mouserpt- 1

protein

(34),humanrettransformingprotein (35),and its relatedprotein

rfp

( 13). Fig.9 A shows the

align-mentsamong the three

protein

sequences. Note that the least numberofgaps has beenintroducedto

optimize

thealignment.

Thereisover30%identityand50-60%

similarity

insequences

atthe

corresponding

amino-terminalhalvesofthese

proteins.

Thecysteine residues ofthe zinc finger motifsare conserved

among all three sequences. The52-kD

protein

ismoresimilar

to either

rfp

or rpt- 1 than

rfp

istorpt- 1

(Fig.

9

A).

Thedata suggest that these three

protein

domains

might

be relatedas a

subfamily

of

finger

proteins.

In

addition,

the

carboxy-terminal

halvesofthe 52-kDSS-A/Roand

rfp proteins

are

highly

similar with> 50% ofidentityand 65%similarity

(Fig.

9B).

rfp

cDNAisolation and

analysisfor

antigenic reactivity.

Be-cause upto 50%

identity

in amino acid residues was found between the

carboxy-terminal

halves of the 52-kD

SS-A/Ro

and

rfp proteins,

it was ofinterest to determine whether

(a)

anti-52-kD SS-A/Ro autoantibodiescould

recognize

rfp

pro-tein, and

(b)

rfp

might bea memberof

_{SS-A/Ro proteins.}

A

cDNAclone 2rlwasisolated from thesameMOLT-4

library

and restriction enzyme

analysis

confirmed its

identity

to the

published

rfp

cDNAclone

(13).

Thiswas

supported

byDNA

sequencing

data which showedthat 2rl was an authentic

rfp

proteincDNA with 104extranucleotidesatthe 5' end as

com-paredtotheoriginalclonedescribed by Takahashietal.

(13).

The extra5' sequenceis

TCCGCTCGGACGCGGCCACGT- TGTCTTGCGCGCTTTGCCCGCCTGGCCCTGGGAC-

TCTGACCCTCGGCTACCCTTTCCTGCCCCACTAGCG-TGGCCGCGAGCCT. In vitro[S35]-methioninelabeled trans-lation

product

of2rl-encodedRNAwas a

_protein

of 58 kD. Noneofthe 10differenthumananti-SS-A/Ro sera recognized

this58-kDprotein in a standard immunoprecipitation assay. In immunoblotting, therabbit antiserum to recombinant

rfp

protein recognized a 58-kD MOLT-4 cell protein that was

dis-tinct fromthe 52-kD and 60-kD SS-A/Ro proteins. These data

indicatedthat,although there are significant sequence similari-ties between the 52-kD SS-A/Ro and

rfp

proteins, no

immuno-logicalcross-reaction could be observed between these proteins

using autoantibodies. It alsoreconfirms the general

observa-tion that human autoantibodies are highly specific for their

antigenic targets and are generally noncross-reactive with re-lated proteins (1).

Discussion

AcDNA encoding the complete 52-kD SS-A/Ro protein

de-scribed by Ben-Chetrit et al. (4) was cloned from a human T-cell (MOLT-4) cDNA library. A partial clone was also

ob-tained fromahumanliver (HepG2) cDNA library and there was nodifference detected between their overlapping regions. Thenucleotideand deduced amino acid sequences for the 52-kDprotein showed no homology with the corresponding se-quencesfortherelated60-kD

SS-A/Ro

(7, 17) and 47-kD SS-B/La antigens (26). DNA hybridization experiments using the cDNAsfromallthreeantigensabove also showed no

cross-hy-bridization (datanotshown). Human SS-A/Roautoantibodies recognized epitope(s) expressed on the recombinant proteins

derived fromtheinitialXZapphage and thesubcloned

pBlue-script plasmid.Thepartially purified recombinant proteinwas

aspecificsubstrateforthesensitivedetection of anti-SS-A/Ro

antibodiesinanELISAformat.Inaddition,invitrotranslation

productderived fromcDNA clone could be immunoprecipi-tatedby anti-SS-A/Ro sera and wasfound tocomigrate with the cellular52-kDprotein asdetermined byimmunoblotting

and two-dimensional gel electrophoresis. Because data from mRNA blotting indicated that there was only one bandof

1.9 kb in MOLT-4

cells,

the 1.9-kb cDNA insert of

p52FL

probablyrepresented the full-length mRNA. The presence of putative zinc fingersandaleucine zipperin the 52-kDprotein

0 1 1.9kb

R N S N R

I,

-

_{-~~~~~~~~~~~~~~~~~~~~~~~~F}

,~~~~~~~~~~~~~~

CL

52FL

- . - _

Figure5.Schematicrepresentationof the cDNA clonesCland52FL derivedfrom HepG2and MOLT-4 celllibraries, respectively. Coding regionsareindicatedbysolid bars and the 5' and 3' untranslated

regionsarerepresented by thin lines. Thinandthickarrowsrepresent DNA sequencederived from restrictionfragment subcloningand

-85 GGGAGCTGGCTACCAGCGTTGAGTTGCCCTCTAAAGCCAAACCCCCIAAAGGTCTCCACACTGCTGTTTAACGGCACACTTGACA 1 ATGGCTTCAGCAGCACGCTTGACAATGATGTGCGAGGAGGTCACATGCCCTATCTGCCTGGACCCCTTCGTGGAGCCTGTGAGC

1 M A S A A R L T M M W E E V T C P I C L D P F V E P V S

85 ATCGAGTGTGGCCACAGCTTCTGCCAGGAATGCATCTCTCAGGTTGGGAAAGGTGGGGGCAGCGTCTGTGCTGTGTGCCGGCAG 29 I E C G H S F C Q E C I S Q V G K G G G S V C A V C R Q

169 CGCTTTCTGCTCAAGAATCTCCGGCCCAATCGACAGCTAGCCAACATGGTGAACAACCTTAAAGAAATCAGCCAGGAGGCCAGA

57 R F L L K N L R P N R Q L A N M V N N L K E I S Q E A R

253 GAGGGCACACAGGGGGAACGGTGTGCAGTGCATGGAGAGAGACTTCACCTGTTCTGTGAGAAAGATGGGAAGGCCCTTTGCTGG

85 E G T Q G E R C A V H G E R L H L F C E K D G K A L C W 337

113

GTATGTGCCCAGTCTCGGAAACACCGTGACCACGCCATGGTCCCTCTTGAGGAGGCTGAGAG GTACQAGGAGAAEKL1QAG V C A Q S R K H R D H A M V P L E E A A Q E Y Q E K L Q 421 GTGGCATTAGGGGAACTGAGAAGAAAGCAGGAGTTGGCTGAGAAGTTGGAAGTGGAAATTGCAATAAAGAGAGCAGACTGGAAG

141 V A L G E L R R K Q E L A E K L E V E I A I K R A D W K

505 AAAACAGTGGAAACACAGAAATCTAGGATTCACGCAGAGTTTGTGCAGCAAAAAAACTTCCTGGTTGAAGAAGAACAGAGGCAG

169 K T V E T Q K S R I H A E F V Q Q K N F L V E E E Q R Q

589

197

CTGCAGGAGCTGGAGAAGGATGAGAGGGAGCAGCTGAGAATCCTGGGGGAGAAAGAGGCCAAGCTGGCCCAGCAGAGCCAGGCC

L Q E L E K D E R E Q L R I L G E K E A K L A Q Q S Q A

673 CTACAGGAGCTCATCTCAGAGCTAGATCGAAGGTGCCACAGCTCAGCACTGGAACTGCTGCAGGAGGTGATAATTGTCCTGGAA

225 L 0 E L I S E L D R R C H S S A L E L L Q E V I I V L E

A A

757 AGGAGTGAGTCCTGGAACCTGAAGGACCTGGATATTACCTCTCCAGAACTCAGGAGTGTGTGCCATGTGCCAGGGCTGAAGAAG

253 R S E S W N L K D L D I T S P E L R S V C H V P G L K K 841 ATGCTGAGGACATGTGCAGTCCACATCACTCTGGATCCAGACACAGCCAATCCGTGGCTGATACTTTCAGAAGATCGGAGACAA

281 M L R T C A V H I T L D P D T A N P W L I L S E D R R Q

925 GTGAGGCTTGGAGACACCCAGCAGAGCATACCTGGAAATGAAGAGAGATTTGATAGTTATCCTATGGTCCTGGGTGCCCAGCAC

309 V R L G D T Q Q S I P G N E E R F D S Y P M V L G A Q H

1009 TTTCACTCTGGAAAACATTACTGGGAGGTAGATGTGACAGGAAAGGAGGCCTGGGACCTGGGTGTCTGCAGAGACTCTGTGCGC

337 F H S G K H Y W E V D V T G K E A W D L G V C R D S V R

1093

365

AGGAAGGGGCACTTTTTGCTTAGTTCCAAGAGTGGCTTCTGGACAATTTGGTTGTGGAACAAACAAAAATATGAGGCT(GGCA(C

R K G H F L L S S K S G F W T I W L W N K Q K Y E A G T

1177 TACCCCCAGACTCCCCTCCACCTTCAGGTGCCTCCATGCCAAGTTGGGATTTTCCTGGACTATGAGGCTGGCATGGTCTCCTTC

393 Y P Q T P L H L Q V P P C Q V G I F L D Y E A G M V S F

1261 TACAACATCACTGACCATGGCTCCCTCATCTACTCCTTCTCTGAATGTGCCTTTACAGGACCTCTGCGGCCCTTCTTCAGTCCT

421 Y N I T D H G S L I Y S F S E C A F T G P L R P F F S P

1345 GGTTTCAATGATGGAGGAAAAAACACAGCCCCTCTAACCCTCTGTCCACTGAATATTGGATCACAAGGATCCACTGACTATTGA

449 G F N D G G K N T A P L T L C P L N I G S Q G S T D Y * 475

1429 TGGCTTTCTCTGGACACTGCCACTCTCCCCATTGGCACCGCTTCTCAGCCACAAACCCTGCCTCTTTTCCCCATGAACTCTGAA

1513 CCACCTTTGTCTCTGCAGAGGCATCCGGATCCCAGCAAGCGAGCTTTAGCAGGGAAGTCACTTCACCATCAACATTCCTGCCCC 1597 AGATGGCTTTGTGATTCCCTCCAGTGAAGCAGCCTCCTTATATTTGGCCCAAACTCATCTTGATCAACCAAAAACATGTTTCTG 1681 CCTTCTTTATGGGACTTAAGTTmTTmTTITCTCCTCTCCATCTCTAGGATGTCGTCTTTGGTGAGATCTCTATTATATCTTGTA

1765 TGGTTTGCAAAAGGGCTTCCT M AATTT AAAAAAAA 1819

Figure6. Nucleotidesequenceof 52FL cDNA insert and deduced amino acidsequenceof the encoded 52-kD SS-A/Roprotein.Three in-framestopcodonsatthe 5'-untranslatedregionandtwo

polyadenylation signalsequences areunderlined.Aleucinezipper

motif is also underlined with the leucine residues marked with arrowheads. Thesesequenceshave been submittedtoEMBL/Genbank

under accession No. M35041.

0

Zinc fi dom

Total se of homc

100 200 300 400 475 Amino A

_PV

SIE

' Acids E

V C

052-kD F G

Li

~~~~SS-A/Ro

H

ingers Leucine _rfp-like D S

Lain zipper domain L F

domain C C

s

~~~~I

% Q

quence showing high degree p Zn

Dlogy with human rfp protein _C C

1C16

39_

B

G K G

V G

Q G

S S

I V

C C

E% ff A

Q Zn+ v .._

IIC

C36

C54

_",%,

C

DGK

K A E L

LHLFC C

EG

GH zlZn++ lv

VA J* _fc

00,C2

114

t.

Sequence showing homology with

human ret and mouse rpt-1 proteins

Figure 7.Diagrammaticsummary of the features of the 52-kDSS-A/ Roprotein.Theentire 52-kD protein sequence shows high degree of homology withhumanrfpprotein. In addition, theamino-terminal half of the 52-kDprotein is similar to the corresponding halves of humanrettransformingprotein and mouse rpt- 1 gene regulatory

protein(see Fig. 9).

Figure8.Putative zinc fingers in the amino-terminal domain of 52-kDSS-A/Roprotein.Thesefingerstructures are arranged based on theproposedfingerstructuresinrpt-1 by Patarca et al.(34)and are further modifiedfor the 52-kD protein.Intheregion encompassing residues 16-54, eitheroneoftwopossiblezinc fingers can beformed in theconfigurations depicted inAand B. Anotherregion, residues 92-114(C)couldbe thesiteof another zinc finger configuration(also

rfp 1 MASGSVAECLQQETTCPVCLQYFAEPMMLDCGHNICCACIARCWGT-- -AETNVSCPQCRETFPQRHMRP SS-A/Ro 1 MASAARLTMMWEEVTCPICLDPFVEPVSIECGHSFCQECISQV-GK--- GGGSV-CAVCRQRFLLKNLRP rpt-1 1 MAS-SVLEMIKEEVTCPICLELLKEPVSADCNHSFCRACITLNYESNRNTDGKGNCPVCRVPYPFGNLRP

**... *.** .* ..**. . * .* ..* .*.. .... *. ** . ..* 68 NRHLANVTQLVKQLRTERPSGPGGEMGVCEKHREPLKLYCEEDQMPICVVCDRSREHRGHSVLPLEEAVE

66 NRQLANMVNNLKEISQEAREGTQGER--CAVHGERLHLFCEKDGKALCWVCAQSRKHRDHAMVPLEEAAQ

70 NLHVANIVERLKGFKSIPEEEQKVN--ICAQHGEKLRLFCRKDMMVICWLCERSQEHRGHQTALIEEVDQ

*.*.. *.. .. * .***.*.*.***. .**.*.*.**.*-**** .* *.

138 GFKEQIQNQLDHLKRVKDLKKRRRAQGEQARAELLSLTQMEREKIVWEFEQLYHSLKEHEYRLLARLEEL

134 EYQEKLQVALGELRRKQELAEKLEVEIAIKRADWKKTVETQKSRIHAEFVQQKNFLVEEEQRQLQELEKD 138 EYKEKLQGALWKUMKKAKICDEWQDDLQLQRVDWENQIQINVENVQRQFKGLRDLLDSKENEELQKLKKE

..*.* .* * . .*.*.... . * ..* *. *. .

208 DLAIYNSINGAITQFSCNISHLSSLIAQLEEKQQQPTRELLQDIGDTLSRAERIRIPEPWI 268 204 EREQLRILGEKEAKLAQQSQALQELISELDRRCHSSALELLQEVIIVLERSESWNLKDLDI 264

208 KKEVMEKLEESENELEDQTELVRDLISDVEHHLELSTLEMLQGANCVLRRSQSLSLQQPQT 268

SS-A/Ro 287 VHITLDPDTANPWLILSEDRRQVRLGDTQQSIPGNEERFDSYPMVLGAQHFHSGKHYWEVDVTGKEAWDL rfp 316 VDVTLDPDTAYPSLILSDNLRQVRYSYLQQDLPDNPERFNLFPCVLGSPCFIAGRHYWEVEVGDKAKWTI

* ..******* ***** ..**** .** ..* .**** ..* *** ..* .* .***** .* ..* .*..

357 GVCRDSVRRKGHFLLSSKSGFWTIWLWNKQKYEAGTYPQTPLHLQVPPCQVGIFLDYEAGMVSFYNITDH

386 GVCEDSVCRKGGVTSAPQNGFWAVSLWYGKEYWALTSPMTALPLRTPLQRVGIFLDYDAGEVSFYNVTER

*** *** *** .. ..***.. ** . .* * * * *.*.*. .* .*******.** *****.*..

427 GSLIYSFSECAFTGPLRPFFSPGFNDGGKNTAPLTLCPL--- NIGSQG--- STDY 475 456 -CHTFTFSHATFCGPVRPYFSLSYS-GGKSAAPLIICPMSGIDGFSGHVGNHGHSMETSP 513

.

..**. . * ** .**..** . .. *** . . *** . . ** . . . *. . * .* .

sequence mightsuggest DNA and/or RNA bindingactivities

although experimental evidence forthe direct association of the 52-kDprotein withDNA/RNAis still lacking.The molecu-larcharacteristics andcDNAsequencedescribed in thisreport

providesaprecise definition forthe 52-kDSS-A/Ro protein. It

isnecessary tomentionthatdueto asequencingerror, an

in-frame carboxy-terminal region rich in Ser/Thr/Pro amino acidswasreported to be present in the 52-kD clone (36). The error wasdetectedand the corrected sequenceis presentedin

thisreport.

Other SS-A/Ro protein sequences. There were three

previouscDNA sequences reported forproteins ofthe

SS-A/

Ro complex. Deutscheret al. (7)andBen-Chetrit et al. (17)

bothdescribedthecloningand sequenceofcDNAforthe

60-kDSS-A/Ro protein. Their corresponding deduced protein se-quences were identical except forshort regions in their

car-boxy-termini.

Bothrecombinant

proteins

were

recognized

by

human autoantibodies in severaldifferent immunoassay

for-mats (7, 17,unpublished data). Thecommonregion between

the twodeducedproteinsequencescontainedaputative zinc fingerand aRNA-bindingproteinconsensusmotifthat could accountforthedirectinteraction withhY-RNAs.Deutscheret al.(7)wereable toreconstitute in vitroribonucleoprotein com-plexescomposedofhY 1 RNAandrecombinant 60-kD

SS-A/

Roprotein.Becausethetwo cDNAs wereobtained from

differ-enthumancells,thedifferencesin sequence could be a resultof differential expression of separate

SS-A/Ro

genesor differen-tialmRNA

splicing.

Athirdsequence wasreported

initially

by Lieuetal.(37) describing the amino-terminal24residues de-termined by standard

protein sequencing

ofa

purified

60-kD protein fromWil-2 cells. Asynthetic peptide correspondingto

thisamino-terminal regionwasreported tobe

recognized by

SS-A/Ro

positive

autoimmunesera

(37);

itwasthus

proposed

67

Figure

9.

Comparison

of amino

65 acidsequences

generated

from 69 the CLUSTALprograms.(*)

identical match and (.) conservative substitution in 137 _{amino acid residue, respectively.}

137

(A)

Comparison

ofthe amino-terminal halvesof 52-kDSS-A/

Ro, human

rfp,

andmouse rpt- 1 207 proteins. The sequence shown

203 _for

_rfp

_{protein is identical in} 207

human ret transforming protein. The sequence for the 52-kD protein is more similar to rfp

(33%identity and 54%

A similarity)andto rpt-l (39%

identity and 61% similarity) than the

rfp

protein is to the rpt-I protein (29% identity and 52%

356 _{similarity). (B) Comparison of}

385 _{the carboxy-terminal halves of}

52-kD SS-A/Ro and human

rfp

426 proteins.>50%of amino acid

455 residues are identical, and there is a65%similarity between the twosequences. Percentidentity

B

and

similarity

areobtained from theGAP program(23).

that a major autoepitope resided within this amino-terminal region (37). McCauliffe et al. (38) from thesame laboratory

recentlydescribeda cDNAencoding thisprotein.Thededuced amino acidsequence appears to beahuman analogue of the rabbitand mouse calreticulin

(calregulin),

a high-affinity

cal-ciumbinding proteinpresent in the lumenof

endoplasmic

re-ticulum (39).Itisunclearwhetherthe

protein

of McCauliffeet

al.(38) is another

species

of

SS-A/Ro antigen

becausethe

reac-tivity

ofthe recombinant

protein

with human

anti-SS-A/Ro

sera was not

presented (38).

The52-kD

SS-A/Ro

protein

de-scribed inthisreportis distinct fromthe cDNA sequence

re-ported by McCauliffe and from other sequences

reported

to

date.

Homology with ret,

rfp,

and

rpt-J

proteins. In 1985, the

activation ofa

transforming

generet wasfirst detected

during

transfection ofNIH 3T3 cellswith DNA from a human T

lymphoma (40). The ret gene was clonedfrom transformed

NIHcellsbyhybridization withhuman sequence

probes

and was shown tobe a resultofDNA rearrangement

(40).

Later, it

wasfoundthat the ret gene was afusionbetweentwounlinked

segmentsofhuman DNA

encoding

foranamino-terminal finger domainand a

carboxy-terminal tyrosine

kinase

domain,

respectively (35). This kinase domainwaspreceded

by

a

hydro-phobictransmembranesequence and thus theret

protein

was

thoughttobeacellsurfacereceptor

(35).

Acellular

homologue

correspondingtotheamino-terminal half ofret wascloned and named

rfp

(ret

finger

protein)and we show thatthelatter hasa

high

degree of homology with the52-kD

SS-A/Ro protein

(Fig.

9).Anuclear localization

signal

sequencewasfoundinthe

rfp

sequence,

suggesting

thatit

might

bea nuclear

protein.

Its mRNA level was 20times

higher

in testis than in liverand

kidneyandthisobservationledto a

proposed

functional role in

mRNA level was highest in 11.5-d mouse embryos as com-pared to other stages ( 13). The fact that

rfp

ispart of a

trans-forminggene retraisesthequestion whetherthe52-kDprotein

might havesimilar transformingpotential. The secondprotein

that the 52-kD SS-A/Ro protein was found to share similarity with was rpt-l (for regulatory protein ofT-lymphocyte

[l]).

Rpt-l was first described by Patarca et al. (34) as a protein

selectivelyexpressed by restingbutnotactivated

CD4'

inducer Tcells. Rpt- 1 isa 4 l-kDnuclearproteinthatdown-regulates

geneexpressionof IL-2 receptor a-chain gene and human

im-munodeficiency virus type 1 genes. Sequence similarities

among the 52-kDSS-A/Ro,

_rjp,

and rpt-I proteinscanbe

use-ful when the function of an index protein such as the 52-kD

SS-A/Ro is unknown, but it is well

acknowledged

that such interpretations should be takenwith caution. Nevertheless, it

providesnewavenues for futureinvestigationsinto thepossible functionalrolesof the 52-kD

SS-A/Ro

protein.

Because the report of Ben-Chetrit et al. (4) showingthat

SS-A/Ro autoantibodytargetsconsisted ofatleasttwo compo-nentsof52and 60 kD,theresults ofmanystudies basedon the

assumptionthat the60-kD

protein

wastheonly

SS-A/Ro

au-toantigen would havetobereevaluated. Buyonet al. (41)

re-cently showed thatinneonatallupus syndrome with

congenital

heart block, all20 mothers ofpermanently affected infants had antibodiestoeither SS-A/RoorSS-B/La

antigens.

The predom-inant antibody response was to the 52-kD SS-A/Ro

antigen

(41). Inanotherreport byBen-Chetritetal.(42), itwasshown

that although most

SS-A/Ro-positive

autoimmune sera had reactivityto both the52-and60-kDantigens in immunoblot-ting, antibodytothe52-kD

antigen

withoutconcomitant anti-bodytothe60-kDantigenwasseenonlyinpatientswith

pri-marySjogren's syndromewhereasantibodytothe60-kD

anti-gen withoutconcomitant antibodyto the52-kDantigen was seenonlyin

patients

with

systemic

lupus erythematosus (42).

Thelatterobservationthat there was adissociation of immune responsesto the two SS-A/Ro componentsmightsuggest dif-ferenteventsstimulatingtheautoimmuneprocessin these

dis-eases. The currentreport onthesequenceand molecular char-acteristics ofthe52-kDprotein establishes its identity and mightbeusefulinelucidating its functionand thatofthe

multi-componentSS-A/Ro complex. Such information could

con-tribute to understanding some features of this autoimmune reaction.

Acknowledgments

Thisis publication6400-MEMfrom the Research Institute of Scripps Clinic. This work is supported by Grants AR32063 and AI10386 from theNational Institutes of Health. Dr. Chan isarecipient ofan Arthritis Foundation Investigator Award.

Noteadded in proof: Recently we have detected protein sequence simi-larity between the 52-kD SS-A/Ro protein and the 55-kD RAD18 gene product ofSaccharomycescerevisiae(Chanet et al. 1988. Gene(Amst).

74:543-547). Theregionof similarity is restricted toNH2-terminal54 aminoacidresidues, and the Cys residues of the zinc finger domain are conserved.

References

1.Tan, E. M. 1989. Antinuclear antibodies: diagnostic markers for autoim-mune diseases and probes for cell biology. Adv. Immunol. 44:93-151.

2.Buyon,J.P.,and R. Winchester. 1990.Congenital completeheart block. A human model ofpassively acquiredautoimmuneinjury.Arthritis Rheum. 33:609-614.

3. Provost, T. T., and M. Reichlin. 1988. Immunologic studies ofcutaneous

lupus erythematosus. J.Clin.Immunol. 8:223-233.

4. Ben-Chetrit,E., E. K. L. Chan, K. F.Sullivan,and E. M.Tan. 1988. A 52-kD protein isanovel component of theSS-A/Ro antigenic particle.J.Exp. Med.167:1560-1571.

5.Wolin, S. L., and J. A. Steitz. 1984.TheRo smallcytoplasmic ribonucleo-protein: identificationoftheantigenic proteinanditsbindingsiteonthe Ro RNAs.Proc.Nadl. Acad. Sci.USA.81:1996-2000.

6.Boire, G.,and J.Craft.1989.Biochemicalandimmunological

heterogene-ityofthe Roribonucleoprotein particles. Analysiswithseraspecificfor theRohYs particles. J. Clin.Invest. 84:270-279.

7.Deutscher,S. L., J. B.Harley, and J. D. Keene. 1988. Molecularanalysisof the60-kDahuman Roribonucleoprotein.Proc.Nadl.Acad. Sci. USA. 85:9479-9483.

8. Buyon, J.P.,S. G.Slade,E. KL.Chan,E. M.Tan,and R. Winchester. 1990. Effective separation ofthe 52kDSS-A/Ro polypeptide fromthe48kD

SS-B/Lapolypeptidebyaltering conditions ofpolyacrylamidegelelectrophoresis.

J.Immunol. Methods.129:207-2 10.

9. Rader, M. D., C. O'Brien, Y. Liu, J.B.Harley, andM.Reichlin. 1989. Heterogeneity of theSS-A/Ro antigen: differentmolecular forms inlymphocytes

and redblood cells. J.Clin.Invest.83:1293-1298.

10.Laemmli,U.K1970.Cleavage of structuralprotein duringtheassembly ofthe headofbacteriophageT4.Nature(Lond.).227:680-685.

11.Towbin,H., T.Staehelin,andJ. Gordon. 1979.Electrophoretictransferof proteinsfrompolyacrylamidegelstonitrocellulose sheets. Procedures and appli-cations.Proc.Natl.Acad.Sci. USA. 76:4350-4354.

12.Feinberg,A.P., and B.Vogelstein. 1983.Atechnique for radiolabeling

DNArestrictionendonucleasefragmentstohigh specific activity. Anal. Biochem.

132:6-13.

13.Takahashi,M., Y. Inaguma, H.Hiai,and F.Hirose.1988.

Developmen-tally regulated expressionofahuman"finger"-containinggene encodedbythe 5' half of the rettransforminggene.Mol.Cell. Biol.8:1853-1856.

14.Adam, S. A., T. Nakagawa,M. S.Swanson, T. K.Woodruff,andG.

Dreyfuss. 1986. mRNApolyadenylate-binding protein:geneisolationand

se-quencingandidentification ofaribonucleoproteinconsensussequence.Mol. Cell Biol.6:2932-2943.

15.Rubin, R. L. 1986.Enzyme-linkedimmunosorbent assay for anti-DNA andanti-histone antibodies.InManualof Clinical LaboratoryImmunology.3rd ed. N. R.Rose, H.Friedman,and J.L.Fahey,editors. AmericanSociety for Microbiology, Washington,DC. 744-748.

16.Southern,E. M. 1975.Detection ofspecificsequences among DNA

frag-mentsseparated by gelelectrophoresis.J.Mol. Biol.98:503-517.

17.Ben-Chetrit,E.,B.J.Gandy,E.M.Tan, and K. F. Sullivan. 1989. Isola-tion andcharacterization ofacDNA cloneencodingthe 60-kD componentofthe humanSS-A/Roribonucleoproteinautoantigen.J.Clin.Invest.83:1284-1292.

18.Chan, E. K. L.,A.M. Francoeur, and E. M. Tan. 1986.Epitopes, struc-turaldomainsand asymmetryofamino acid residues inSS-B/Lanuclearprotein.

J.Immunol. 136:3744-3749.

19.Garrels, J. I. 1983.Quantitative two-dimensional gel electrophoresis of

proteins. Methods Enzymol.100:411-423.

20.Chen,E. Y., and P. H. Seeburg. 1985.Supercoil sequencing:afastand

simplemethodforsequencing plasmidDNA. DNA.4:165-170.

21.Sanger, F. 1981. Determination of nucleotide sequence in DNA. Science

(Wash.DC).214:1205-1210.

22. Devereux, J., P. Haeberli, and0.Smithies. 1984. A comprehensive set of sequenceanalysisprograms for the VAX.Nucleic AcidsRes.12:387-395.

23. Needleman, S. B., and C. D. Wunsch. 1970.Ageneral method applicable

tothe searchforsimilarities in the amino acid sequence oftwoproteins.J.Mol. Biol.48:443-453.

24.Higgins,D.G., and P. M. Sharp. 1988. CLUSTAL: a package for

perform-ing multiplesequencealignmentson amicrocomputer. Gene(Amst.). 73:237-244.

25.Higgins,D.G.,andP. M.Sharp. 1989. Fast and sensitive multiple se-quencealignmentson amicrocomputer. Comp. Appl. Biosci. 5:151-153.

26.Chan, E. K. L., K. F.Sullivan,and E.M. Tan. 1989. Ribonucleoprotein

SS-B/La belongsto aprotein familywith consensus sequences for RNA-binding. Nucleic Acids Res. 17:2233-2244.

27. Kozak, M. 1989. The scanning model for translation: an update. J. Cell Biol. 108:229-241.

28. Berg,J.M.1986. Potentialmetal-binding domains in nucleic acid binding

proteins.Science(Wash.DC). 232:485-487.

29. Klug, A., andD. Rhodes. 1987. Zincfingers: a novel protein fold for nucleic acidrecognition. ColdSpring Harbor Symp. Quant. Biol. 52:473-82.

31.Kouzarides,T., and E.Ziff. 1988. The roleoftheleucine zipper inthe

fos-jun interaction. Nature(Lond.). 336:646-651.

32. Landschultz, W. H., P. F. Johnson,andS.J.McKnight. 1989. The DNA

binding domain oftheratliver nuclear protein C/EBP is bipartite.Science(Wash. DC). 243:1681-1688.

33.Sassone-Corsi, P., L.J.Ransone,W. W.Lamph, andI.M. Verma. 1988. Directinteraction between fos and jun nuclear oncoproteins: role ofthe "leucine zipper" domain. Nature(Lond.). 336:692-695.

34. Patarca, R., J.Schwartz,R.P.Singh, Q.-T.Kong, E. Murphy, Y. Ander-son,F.-Y. W. Sheng,P.Singh,K. A.Johnson, S. M. Guarnagia, T. Durfee, F. Blattner, and H.Cantor.1988.rpt- 1, anintracellular protein from helper/inducer

Tcells that regulatesgeneexpression of interleukin2 receptorand human immu-nodeficiency virustype 1.Proc.Natl. Acad.Sci. USA. 85:2733-2737.

35.Takahashi, M., and G.M.Cooper. 1987.rettransforminggeneencodesa

fusionprotein homologoustotyrosine kinases. Mol. Cell. Biol. 7:1378-1385. 36. Chan,E. K.L.,J.C. Hamel, C.L.Peebles, J. P. Buyon, and E. M.Tan.

1989. Protein domainofthe 52kDaSS-A/Roribonucleoprotein particle is

homol-ogous tothe N-terminalmetal-finger domain of human ret/rfp andmouserpt- 1

proteins. J. Cell Biol. 109:14a. (Abstr.)

37.Lieu, T.-S.,M. M.Newkirk, J.D.Capra, andR. D.Sontheimer. 1988.

Molecularcharacterization ofhuman Ro/SS-Aantigen. Amino terminal se-quence of theproteinmoiety of human Ro/SS-A antigen and immunological

activity ofacorresponding synthetic peptide. J. Clin. Invest. 82:96-101. 38. McCauliffe,D. P., F. A. Lux, T.-S.Lieu, I. Sanz,J. Hanke, M. M.

New-kirk,L. L.Bachinski, Y. Itoh,M. J.Siciliano,M.Reichlin,R. D.Sontheimer,and J. D.Capra. 1990.Molecular cloning,expression,and chromosome 19

localiza-tionof a humanRo/SS-A autoantigen. J.Clin. Invest.85:1379-1391. 39.Lieu, T. S., D. P. McCauliffe, P. Volpe, B. H. Alderson-Lang, M. Micha-lak, J. D.Capra, and R.D.Sontheimer. 1990. Structuralandfunctional homol-ogy between a humanRo/SS-A autoantigenand thecalcium-binding protein, calreticulin. Clin. Res.38:408. (Abstr.)

40.Takahashi, M., J. Ritz, and G.M.Cooper. 1985.Activation ofanovel humantransforminggene, ret,by DNArearrangement.Cell. 42:581-588.

41.Buyon,J. P., E.Ben-Chetrit,S. Karp, R. A.S.Roubey,L.Pompeo, W. H.

Reeves,E. M.Tan,and R.Winchester. 1989.Acquired congenital heart block. Patternof maternal antibodyresponse tobiochemicallydefined antigens of the SS-A/Ro-SS-A/Lasystemin neonatal lupus.J.Clin. Invest. 84:627-634.

42.Ben-Chetrit, E.,R. I.Fox,and E. M.Tan.1990.Dissociation of immune

responses to theSS-A (Ro) 52-kdand60-kdpolypeptides in systemic lupus