Cloning of a complementary DNA coding for the 100 kD antigenic protein of the PM Scl autoantigen

Cloning of a complementary DNA coding for the

100-kD antigenic protein of the PM-Scl

autoantigen.

Q Ge, … , C O'Brien, I N Targoff

J Clin Invest.

1992;

90(2)

:559-570.

https://doi.org/10.1172/JCI115895

.

Anti-PM-Scl antibodies are associated with polymyositis-scleroderma overlap or either

disease alone. Among sera from 39 patients with anti-PM-Scl, 23 recognized the 100-kD

band in immunoblot against HeLa cell extract, 16 of which also stained the 70-kD band. A

human thymocyte lambda gt11 cDNA expression library was screened with anti-PM-Scl

serum, and two clones were identified whose products reacted with 33 and 37 of 39

anti-PM-Scl sera, respectively, but none of 26 negative control sera. Affinity-purified antibody

reacting specifically with plaques of the clone stained the 100-kD band on immunoblot,

reacted with nucleoli of HEp-2 cells, and immunoprecipitated the PM-Scl protein complex.

Partial sequences of both inserts were identical. One insert was fully sequenced, and

additional 5' and 3' sequence was obtained using a gene-specific primer to form a cDNA

with HeLa cell RNA as template followed by PCR. The complete nucleotide sequence

included 2,739-bp coding for a predicted full-length protein of 98,088 D. There was no

homology with the PM-Scl 75-kD protein and no significant homology with other proteins. A

mixed-charge cluster was identified, with 22 charged amino acids of 37. In conclusion, the

full-length cDNA sequence was determined coding for the PM-Scl 100-kD protein, the most

commonly antigenic protein of the PM-Scl complex.

Research Article

Find the latest version:

Cloning of

a

Complementary

DNA

Coding

for

the

100-kD

Antigenic Protein of the PM-Scl Autoantigen

QunGe,*Mark Barton Frank,*CharlesO'Bnen,*and IraN.Targoff,**

*Arthritis and ImmunologyProgram, Oklahoma Medical Research Foundation 73104; and $Department ofMedicine, Universityof Oklahoma Health Sciences Center 73190; and Veterans Affairs Medical Center, Oklahoma City, Oklahoma 73104

Abstract

Anti-PM-S& antibodies are associated with

polymyositis-scleroderma overlaporeither disease alone. Among sera from

39 patients with

anti-PM-S&,

23 recognized the 100-kD band

in immunoblot against HeLa cell extract, 16 of which also stained the 70-kD band.Ahumanthymocyte Xgtll cDNA

ex-pression library was screened with

anti-PM-S&

serum, and two

clones wereidentified whose products reacted with 33 and 37 of 39

anti-PM-Sd

sera,

respectively,

butnoneof 26negative con-trol sera.

Affinity-purified

antibody

reacting specifically

with plaques of the clone stained the 100-kD bandonimmunoblot, reacted with nucleoli of HEp-2 cells, and immunoprecipitated

the

PM-Sd

protein complex. Partial sequences of both inserts were identical. One insertwasfully sequenced, and additional 5'and3'sequencewasobtained usinga

gene-specific

primerto

form a cDNA with HeLa cell RNAas template followed by PCR. The complete nucleotide sequence included

2,739-bp

coding fora

predicted full-length protein

of

98,088

D. There wasnohomology with the PM-Scl 75-kD protein and no

signifi-canthomology with other

proteins.

Amixed-charge clusterwas

identified, with 22 charged

amino

acids of 37.In

conclusion,

the full-length cDNA sequencewasdeterminedcoding for the

PM-Scl 100-kD

protein, the

most

commonly antigenic protein of

the PM-Scl complex. (J.

Clin.

Invest. 1992.

90:559-570.)

Key words:

polymyositis

* scleroderma * autoantibodies * anti-PM-Scl * nucleolar

proteins

Introduction

Autoantibodies reacting

with the PM-Scl

antigen

have been

found

inserafrom

patients

with

polymyositis

orscleroderma

orfrom

patients

withoverlapof these two conditions

(1-5).

They

wereidentifiedin 8%ofa

large

groupof

myositis

pa-tients and - 3% ofagroup ofscleroderma patients ( 1,

4).

However, it isvery rare

for

this antibody

tobe found in

patients

with other conditions, making itavaluablediseasemarker, and

suggesting

that itmay havelinkstofundamental disease

mecha-nisms in those

patients

in whom it is

found.

Anti-PM-Sclisthe

unique specificity derived

fromanti-PM-l

( 1,

6).

The PM-Scl

antigen

is

found

inthe nucleolus

(7, 8),

and all

Portions of this workwerepresented atthe 54th and 55th Annual

Meetingsof the American College ofRheumatology,27 October-l

November 1990 and 17-21 November 1991.

Addresscorrespondence to Ira N.Targoff, M. D., Arthritis and

ImmunologyProgram, Oklahoma Medical ResearchFoundation,825 NE13th Street, OklahomaCity,OK 73104.

Receivedfor publication8August1991andinrevisedform2March

1992.

TheJournal of ClinicalInvestigation,Inc. Volume 90, August 1992, 559-570

serawithanti-PM-Scl stain the nucleolus in indirect

immuno-fluorescence.

Most sera with the antibody also stain the nu-cleus, but

antibody

could be affinity purified that reacted only with the nucleolus (7). The role of the antigen in cell physiol-ogyis not known, but it may be involved in preribosomal for-mation, a function of the nucleolus, since compounds that

in-terfere with

ribosomal

formation

also interfere with its produc-tion (8).

Immunoprecipitation

of

35S-labeled

HeLa cell extracts with anti-PM-Scl showsagroup of1

1-16

proteinsthatappearto be acomplex (8, 9). Aprominent band is seen at 100 kD, but bandsare present from 20 to

110

kD. No nucleic acids have been

associated

with the antigen by

immunoprecipitation.

Most sera with the antibody react with a

100-1 10-kD

protein in immunoblot studies(9), andsomealsoreactwith asecond

protein

migrating in the range of 70-75 kD (10). Occasional

serareactprimarily with this second protein.Thegene encod-ing the second protein has recently been cloned and sequenced and has been found to produce a protein of 39.2 kD that mi-grated aberrantly on PAGE, apparently because ofa region rich in

acidic

residues in the carboxyl halfofthe molecule(

10).

The

100-1

0-kD protein that is commonly antigenichas notbeen

analyzed

in

detail.

Tostudy

further

thelarger antigenic protein

of

the PM-Scl

antigen

ona molecular level,wehave isolated a complementaryDNA(cDNA)'that codes for this protein and have

determined

its sequence.

Methods

Sera. Sera were drawnfromclinicalscreeningattheOklahoma Medi-cal ResearchFoundationorfromserareferred for testing for myositis-associated autoantibodies. Sera from39patientswerefoundtocontain anti-PM-Sclby immunodiffusionandconfirmed by immunoprecipita-tion. 20 myositissera fromthiscollection foundto benegative for

anti-PM-Sclweretestedascontrols,someof whichhad other autoanti-bodies.Sixserafromhealthylaboratory workerswerealsotested.

Immunoblotting. Each serum was tested by

immunoblotting

againstwhole HeLa cell

lysate

asantigensource todetermineits compo-nentreactivity.Antibodies that wereaffinitypurifiedfromEscherichia coli-expressed

PM-Scd

fusion protein as below were tested against

whole HeLa cell

lysates

oragainst immunoprecipitates preparedwith

anti-PM-Sclserumfrom HeLa cell

lysates.

Thesizeofthe PM-Scl mole-culesexpressedin E. coliwerealsodeterminedwith immunoblot

as-saysagainstE.coli

lysate

containingtherecombinantprotein.

The HeLalysatewaspreparedby sonication in 0.15MNaCl in 0.05 MTris-HCI bufferatpH7.4 whereas E.coli

lysates

wereprepared by

lysis

with

Xgtl

1bacteriophagewithorwithout the cloned insertorwith thesamplebuffercontainingSDSforusein

electrophoresis

(

11).

The

lysates

and immunoprecipitates were electrophoresed in 10% SDS-PAGE andblottingwascarriedoutaspreviouslydescribed(

11),

ex-1. Abbreviations usedin this paper: cDNA, complementary DNA;

IPTG,

isopropylthio-jt-galactoside;

PCR, polymerasechain reaction;

TBST, 0.05 M Trisbuffer pH8.0 with 0.15MNaCl,0.05%Tween-20.

ceptthat8% nonfat milkwasusedforblocking.Serawerediluted1:100

for routine testing,andaffinity purified antibodywasusedundiluted. Isolation ofPM-Scl cDNA. It was previously known thatthe PM-Scl antigenwaspresent inthymus tissue byimmunodiffusion against calfthymusextract(I)and in HeLacellsby immunoprecipitation (8). Therefore,humanthymocytecDNAand HeLa cell cDNA Xgtl 1

ex-pressionlibrariesfrom Clontech Laboratories (Palo Alto, CA)were

screened.Serum frompatient J.H. with diffuse scleroderma, shownto

contain anti-PM-Scl byimmunodiffusion andtoreactwith both the 100- and 70-kDproteins by immunoblotting,wasusedforscreening, diluted 1/500in 0.05 M Tris bufferatpH8.0 with 0.15 M NaCland 0.05% Tween-20 (TBST) with 2% BSA and 10 mMsodium azide.

Recombinant phagewereplated withE.coliY1090onLB agar, incu-batedat42°C for3-4 h, overlayeredwith nitrocellulose membranes

soakedinisopropylthio-(3-galactoside (IPTG),and incubated for

an-other 3 hat370C.The membraneswerethenwashed, blocked, incu-batedwith dilutedpatientserum, anddevelopedasfor

immunoblot-ting, usingalkaline-phosphatase-coupledanti-humanIgG conjugate

and 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium substrate.Atotalof2X 106plaquesfrom each of the librarieswere

screened. Positive cloneswereplaquepurified.

Affinity

purification ofantibodies. Specific antibodieswereaffinity purified by elution fromproteinsderivedfromplaquesofthe

recombi-nantPM-Sclbacteriophage.To reduce thebackgroundattributableto

anti-E. coliantibodies,serawerefirstpreabsorbedwith nitrocellulose

membranescoatedwithproteins fromE.colithat had beenlysedwith

wild-type Xgtl1bacteriophageasdescribedabove. LBplateswere

pre-paredwithahigh concentrationof eitherplaque-purifiedrecombinant

PM-Scl bacteriophage or wild-type bacteriophage. The plates were

overlaid with IPTG-soaked nitrocellulose membranes, which were

then incubatedwith absorbedanti-PM-Sclorotherserumovernightat

40C,asdescribedaboveforscreening.AfterfiveTBSTwashes,

mem-brane-bound antibodieswereeluted with0.2 Mglycine-HClatpH2.5 for 1.5 min. Collected eluateswereneutralizedimmediatelybyadding

1 MTrisbase. Theactivity oftheaffinity-purifiedantibodywas

con-firmed by testing against recombinantand wild-typebacteriophage

proteinsonnitrocelluloseasabove.

Testing of patientsera. To determinewhether patientsera with anti-PM-Sclorcontrols reacted with theproteinproductof theclone, plaque-purified recombinant bacteriophageweremixed5:1 with

wild-type Xgtl1 andplatedwith E. coli Y1090,asforscreening.Piecesof nitrocellulosecarryingplaqueproteinswereincubated with5 mlofan

individualserumdiluted 1:600 anddeveloped. Reactionswere

consid-eredpositiveifsignificant staining of - 20%oftheplaqueswas seen.

Intensity ofstainingwasestimatedon a0(negative)to4+(strongest)

scale. Threeanti-PM-Sclandtwonormalserawerealso testedagainst

IPTG-inducedE.colilysed withrecombinantbacteriophagein

immu-noblotasdescribed above.

Immunoprecipitation. Immunoprecipitation from

"S-methio-nine-labeled HeLa celllysatewasperformedaspreviouslydescribed

(12). 1mgprotein A-Sepharose (PharmaciaLKB, Uppsala, Sweden)

( 10Mlof swelled bead suspension) was coated with 20Mlof serum or 500

_Al

ofaffinity-purified antibody. Immunoprecipitates were ana-lyzed on 10% SDS-PAGE. Preparation of unlabeled immunoprecipi-tatesfor immunoblot analysis of affinity-purifiedantibodies was simi-lar but the immunoprecipitates were treated as for protein analysis. Immunoprecipitation from unlabeled HeLa cell lysate for nucleic acid

analysiswasalsoperformedaspreviously described( 12). One 150-cm2

flask ofHeLa cells was usedforthree samples.4mgof

protein-A-Se-pharosewerecoatedwith20

_Al

ofserum.Phenol-extracted

immuno-precipitateswereanalyzed on 7 M urea, 10% PAGE, and developed with silver stain (Bio-Rad Laboratories, Richmond, CA).

Northern blot analysis. Total RNA was isolated from HeLa cells by

the procedureofChomczynski and Sacchi (13), andpoly-(A)'RNA wasisolated from HeLa cells using a commercial kit (MicroFastTrack; Invitrogen, San Diego, CA). Both were electrophoresed inadenaturing 1% agarose gelcontaining2 Mformaldehyde and were transferred onto Nytranmembrane(Schleicher & Schuell, Inc., Keene, NH) ( 14).

Pre-hybridizationwith salmon sperm DNA and hybridization with radiola-beled cDNA probe (15, 16) were performed in 2x prehybridization andhybridization buffersfrom 5 Prime-3Prime Inc. (Boulder, CO)

accordingtotheir instructions, with formamide at420C.Membranes were washed under high stringency conditions with0.1Xstandard sa-line citrate and 0.1% SDS at650Cfor 30 min and were developed with

autoradiography.

DNA sequencing. DNA from the plaque-purified recombinant Xgtl1bacteriophagewasprepared from liquidlysatesand digested with EcoRI toexcisethe cDNA insert. The insert was isolated by

fraction-ationinanagarosegel and subcloned into the plasmidpUCl8 (17).

Afterrestriction enzyme mapping of the cDNA, individual restriction

fragments wereisolated from agarose gels. These were ligated into M I3mpl8 andmpl9(17)andwereusedtotransformE.coli strain

XLlBlue. Thecomplete nucleotide sequencesofboth strandsofthe cDNAweredetermined by usingthedideoxynucleotide

chain-termina-tion method (18), using35S-labeled dATP and T7 DNApolymerase [UnitedStatesBiochemical, Cleveland,OH(SequenaseVersion2.0)

orPharmaciaLKB Biotechnology, Piscataway, NJ] (19).

Compres-sionregionswereresolvedby usingdeoxyinosine triphosphate. Determinationof the5'and3'endsofthecDNAsequence. Addi-tional cDNA sequence for the gene for thePM-Scl100-kD proteinnot

included in theoriginal bacteriophageinsertwasobtained byusingthe methodsofLohetal.(20)and Frohmanetal.(21 ).

Foramplification ofthe 5' endofthecDNA, 15

Mg

oftotal HeLa cell RNA and2MgofHeLa cellpoly-(A)'RNAwereseparatelyreverse

transcribedusingthecDNACyclekit(Invitrogen)withgene-specific primer1 (TableI). AftercDNAsynthesis,20

Mg

ofribonucleaseAwere

addedtothereversetranscribed cDNA pool andincubatedat37°Cfor 30 min. Excess primerand deoxynucleotide triphosphateswere

re-movedby three roundsofspinfiltration in a Centricon-30(Amicon Corp., Scientific Sys.Div., Danvers, MA),in 5 mMTris-HCl bufferat

pH8.0 with 0.5 mM EDTA. Thesamplewasconcentrated ina vacuum

centrifugeandadjustedto23MlwithH20.Aterminal

deoxynucleoti-TableI.Gene

Specific

PrimersUsed

for

cDNA Synthesis and Polymerase Chain Reaction

Primer Sequence Length Region Use

nt

1 5'-GGCTTCATCCAGTAAAATACC-3' 21 431-411* cDNAsynthesis for 5' extension 2 5'-CCAGATCTGGTACTGCATTACTCTGC-3' 26

310-292"4

5' End amplification

3 5'-CATCrAGAAGCTTCAGGTACAACTGGC-3' 27 2593-26101 3' End amplification

4

5Y-TAGGATCCGTCCTGTCGGCGACCAGC-3'

26

69-86"1

5' Primer for PCRfor expression 5

5'-CAAGATC1'GCCAGTTGTACCTGAAG-3'

25 2610-2594*t 3' Primer for PCRfor expression Bold sequences were added for restriction enzyme sites. Numbered regions correspond to sequence shown in Fig. 9. nt,=nucleotides; * Com-plementary to coding sequence. t Underlined sequence isBglIIrestrictionsite. § Underlined sequence isXbaIrestriction site. 11Underlined sequence is BamHI restriction site.

dyl-transferase tailing reactionwasthenperformedwith dATP accord-ing toinstructionsfrom Promega Corp.(Madison,WI).

The resultingextended,tailed 5' end wasamplifiedusingthe poly-merase chainreaction(PCR). The reaction wasperformedin 100

,l,

containing thefollowing in final concentrations: lx Taq DNA poly-merase buffer (Promega Corp.); 200 uM deoxynucleotide

triphos-phates;0.25

,M

gene-specificprimer 2; and 0.25 IMoligo-d(T)17with restriction-site adapter. The primer usedwasdifferent from that used forfirst-strand cDNA synthesis in ordertoincrease thespecificityfor this gene and avoid amplification of tailedfirst-strand primer. The cDNAtemplate was denaturedat950C for 5 min followedby chilling

onice. It wascentrifuged briefly,andthen 2.5 UofTaq DNA

polymer-ase(PromegaCorp.) andtwodrops of mineral oilwereadded.Atotal of 35 cycles were performed,consistingof 940C for 50 s, 540C for 1

min,and720Cfor 1min,with a finalelongationstepat720Cfor 5 min. The 3' end wasamplifiedby asimilarprocedure. Total RNA and

poly-(A)'RNA from HeLa cellswere reversetranscribed usingan

oligo-d(T) primer. After RNAse digestion,the deoxynucleotide tri-phosphates were removed byusingaG-50 spincolumn (5 Prime-3 Prime Inc.). The 3' end wasamplifiedfrom the cDNA pool inaPCR usinggene-specificprimer3(Table I)and theoligo-d(T) primer.

The 5' and 3' extended cDNAproductswereanalyzedby electro-phoresis in 8% PAGE. To confirm that the products werespecific,

Southernblottingwasperformed (22),inwhich thegel productswere

electroblottedontoHybond-N membrane (AmershamCorp.,

Arling-tonHeights,IL), andhybridizationwasperformedwitha32P-labeled cDNAinsertasprobe.Specific productswereisolatedfrom thegel by

diffusion ( 14).Aportionof theproductsweresubcloned into the

plas-mid vector pUC 18 afterdigestionofthe 5' products with PstI andBgl II,

ordigestionofthe 3'productswith XbaI and PstI. Other PCRproducts

weresubclonedinto theplasmid vector PCR1000(Invitrogen) accord-ing to manufacturer's instructions withoutprevioustreatment.All

ex-tendedproductsweresequenced fromadouble-strandedDNA

tem-plate withT7DNApolymerase(PharmaciaLKB)asabove. Expression of the recombinantprotein.Expression of theprotein

codedfor by the sequenceincludingthe extended5'endwasachieved from a new cDNA prepared by reversetranscriptionfrom total HeLa

_4 -|Anti-PM-Sci Sera

-Z Z A B C D E F G H I J K L M

f * 1S=- S

..,Al.s;,.J

oz W 'v

-195

-105

Table II. Comparison ofReaction inImmunoblotwith Reaction

withJH4Anti-PM-SclClones forAnti-PM-SclSera

Serum

no. 100 kD 70 kD Other JH4B, JH4C,

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

+ + + 4+

+ + + 3+

+ + - 3+

+ + - 3+

+ + - 4+

+ + + 2+

+ + + 3+

+ + - 4+

+ + + 4+

+ + + 4+

+ + + 3+

+ + + 3+

+ + + 2+

+ + + 2+

+ + + 2+

+ + - 4+

+ - - 1+

+ - - 4+

+ - - 4+

+~~~~~~4

+ - + 4+

+ - + 4+

+ - + 4+

_ + + 2+

_ _ ₊

₃₊

_ _ _{+ 1+}

_ _{- + 4+}

-

+~~~~~4

-+~~~~~~3

_ _{- + 3+}

4+ 3+ 4+ 3+ 3+ 3+ 3+ 3+ 4+ 4+ 2+ 2+ 3+ 3+ 2+ 3+ 3+ 4+ 4+ 1+ 4+ 3+ 4+ 1+ 2+ 1+ 3+ 4+

- - - 4+ 3+

_ _ - 1+ 2+

_ - - 1+ 2+

- - - 3+ 4+

_ _ _- -

₁₊

-

_ _- - ₁₊

- - - 3+ 3+

_ - - 1+ 1+

_ _ _- -

₂₊

-71

-44

-28

All 39 sera werepositive for anti-PM-Scl by immunodiffusion and weretestedfor reaction with the 100- and 70-kD bands on immuno-blotagainstwhole HeLa cellextract. Reaction in immunoblotwas

assessedaspositive(+)ornegative (-),whereaspositivereaction

againstthetwoJH4cloneswasestimatedas 1+ through4+.Other referstoreaction in immunoblot with bands other than the 100- and 70-kD bands.

Figure1. Immunoblot withanti-PM-Sdlpatientsera. Whole HeLa cellextract asantigen waselectrophoresedina 10%SDS-PAGE,

transferredtonitrocellulose,anddevelopedwithanti-PM-Sclsera

(A-M) or normal (NL) sera. Thepositionof molecular mass markers in kD is shown at the right. The spots are artifactsappearingafter

original gel development.Lanesdevelopedwith selectedsera are

shown,illustratingthe patterns observed:staining ofthe 100-kD band (A-E), stainingof both the 100- and 70-kD bands (F-M), and

stainingof other bands (A, C, F, H-M).

cellRNAusinganoligo-d(T) primer,followed by PCRamplification using gene-specific primer4(TableI)beginning 31 bpafter the starting ATGwith a BamHI site attached and gene-specific primer 5 beginning 8 bpupstream of the stop codon with aBglII site attached. Initial attemptsusingaprimerbeginning5'to thestartingsite were unsuc-cessful.

Constructs of the pQE expression plasmids prepared using the

QIAexpresskitfromQiagen,Inc.(Chatsworth, CA) were used as

vec-tor forprotein expression. Purified PCR product wasdigested with

BamHIandBglII togeneratecompatibleends and subclonedinto the type IIIpQEvector constructs.Thevector wastransformed into E.coli

strainM15(pREP4). Colonieswereselected byBamHI-BglII

diges-tionof plasmid DNAandwereanalyzed for optimal protein expres-sion.

Othermethods. Ouchterlonyimmunodiffusionwasperformed by

standard methodsusing concentrated calf thymusextract asantigen

compared withastandardanti-PM-Sclserum(1)ora serumverified againstthestandard.Indirect immunofluorescenceagainst HEp-2cell

substratewasperformed usingacommercial kit(INOVADiagnostics,

Inc., SanDiego, CA),followingtheinstructionsof the manufacturer. Computeranalysis of nucleotide and amino acidsequenceswas

per-formedwith theSequenceAnalysisSoftwarePackageof the Genetics

ComputerGroup oftheUniversityof Wisconsin(23), usingaVAX

8250 computer, and alsousingasupercomputerthroughthe Pittsburgh Supercomputing Facility as previously described (24). Tests for

chargedrunsand clusters inproteinsequenceswereperformed using

the PROPAT program(25),kindlyprovided byDr. V.Brendel,

De-partmentofMathematics,StanfordUniversity,Stanford,CA.Protein motifswereexaminedusingadatabaseobtained fromDr. A.Bairoch,

Centre MedicalUniversitaire, Geneva, Switzerland andadaptedfor

use on aVAX computerby M. BartonFrank.

Results

Immunoblotting. The 39 sera identified as positivefor

anti-PM-Scd

were tested by

immunoblotting against

whole HeLa

cellextracttodeterminethe

antigenic proteins

for eachserum

(shown for 13

patient

serain

Fig. 1).

23 sera

(59%)

staineda

l00-kD band, and 16 of these 23 sera (41% oftotal) also

staineda70-kD band

(Tables

II and

III). Only

1 serumstained

the 70-kD bandandnotthe100-kD

band, although

1 of the 23

that stained the l00-kD band stained the 70-kD band much morestrongly. The 70- and 100-kD bands

appeared

to

corre-spondto the 75- and1 0-kDbands

reported by

others(9, 10). The 100-kDbandappearedas adoublet insomeruns,

appar-ently

dueto

degradation

of the lOO-kD band since the second

band was not

consistently

present. 21 sera

recognized

bands

other than the 100-and 70-kD

bands,

6 of whichwerenegative

against

the 100- and 70-kD bands. Amongthe other

bands,

bands of 30- and 32-kDwere seenwith2sera

(K

and M in

Fig.

1);a50-kD band with3; anda

weak,

inconsistent 85-kD band with 10,

which

was

similar

to aband

reported by

others

(10).

Other bands were

unique

to

individual

sera. Nine sera had

negative

immunoblots.

Staining

of the 100- and 70-kD bands by all

positive

sera wasweak ineach of theseveral runs

per-formed,

evenbyserawith

high anti-PM-Scl titers,

butsomeof theother bandswerestained

strongly.

The

relationship

of the otherbands tothePM-Scl

antigen complex

remains unclear.

TableIII. SummaryofAnti-PM-SclSera Reactions

CLONEJH4C,

WB-100kD WB-70 kD +

+ + 16 (41) 0 (0)

+ ₇₍₁₈₎ 0

(0)

+ 1(2.6) 0 (0)

-- 13 (32.5) 2 (5.1)

Summary of reactions of all 39 sera found to be positive for

anti-'a

Figure2.Testingofpatientseraforimmunoreactivitywithexpressed protein. E. coliwasinfected witha5:1 mixture ofwild-typeand

JH4B, Xgtl andinducedwith IPTG. Plaqueproteinswere trans-ferred tonitrocellulose, whichwasdividedintosegments,and

devel-opedwithdifferentsera.Sera usedfordevelopment: and 2: normal sera; 3: noserum;4-7: myositisserawith autoantibodies other than

anti-PM-Sdl; 8-2 1:anti-PM-Sdlsera_{(identified by}

immunodiffu-sion). Insegmentsdevelopedwithanti-PM-Sdl, approximatelyone fifth oftheplaquespresentwerepositively stainedwhereassegments

developedwithotherserashowednopositive plaques.

Isolation ofPM-Sdl cDNA clone. Two Xgtl clones that

persistently produced immunoreactive plaques, designated

JH4B,

andJH4Cj, wereidentified from thethymocyte library,

andnonewereidentified fromthe HeLacelllibrary. Multiple

anti-PM-Sdlpositiveandnegativesera weretested for

reactiv-itywithplaques produced bytheseclonesasdescribed in Meth-ods

(Fig.

2).Serafrom 33of39patientspositivefor

anti-PM-Seby immunodiffusion reacted with cloneJH4Bf plaques

(84.6%)

whereas 37 of 39 (94.9%) reactedwith clone

JHa4C-plaques,

ofwhich 26 (66.7%) stained strongly at 3-4+. The

fourserathatdidnotstainJHe4Bf plaquesbutdidstainJH4Cn plaques generallystainedJHt4Ce weakly

(i+),

suggesting that

thedifferenceinreactivitybetween the cloneswasquantitative

rather than qualitative. As shown in Table III,9 22 of 23sera withanti-Pl00-kD

activity

in immunoblotting were positive

against plaquesofbothclones, and the other reacted faintly

withJH4Cw.Theserumwithanti-70-kD activitybutnot

anti-I00-kDalso reactedwith bothclones.Four ofsixserareacting onlywith otherbandsin

immunoblotting

(neither100nor70

kD) reacted with both cloneswhereas the othertwo reacted

withneitherclone,theonly2ofthe39testedthatwerenegative

against

JH4Cl.Sixofnineserathatwerenegativein

immuno-blot reacted with both clones, and the other three reacted

weaklywith

JH4Co.

Noneof the sera from 20 patients with otherautoantibodies,nor serafrom 6 normal controlsreacted

withplaquesof either clone(notshown).

Toconfirm that thisreactivitywaswith fusionprotein,

fur-ther testing in immunoblotting was performed. Reaction

againstalysate ofE.coli produced by the

JH4C,

Xgt

bacterio-562 Q.Ge,M.BartonFrank,C.O'Brien, andLN. Targoff

phage clone was compared with reaction with a

lysate

pro-duced bywild-type Xgtl 1 (Fig. 3).A

protein

of - 200 kDwas

recognized

inJH4C, lysate by all three anti-PM-Scl tested but was notfound in wild-type

lysate.

Thetwonormalseradidnot

recognize any

proteins

in

this

region from

either

lysate.

Thus,

proteins carrying antigenic determinants recognized

specifi-cally bymostanti-PM-Sclsera wereexpressed by these

clones,

suggesting

that the inserts coded for an

antigenic

protein

of

PM-SCd.

Confirmation

of

the

identification

of

the clones.

The immu-nological

specificity of

the

expressed

protein

of thetwoclones wasfurther

confirmed

by

testing affinity-purified antibody

pre-pared, as described in

Methods, by

elution from

plaque

pro-teins of the JH4 clone or

wild-type Xgtl

1 after

they

had been incubated with

anti-PM-Scl

serum.

Affinity-purified

antibody

fromanti-PM-Scl positive JH serum using either of the JH4 clones

showed

strong

activity

against plaques

of both

clones,

demonstrating

that the recombinant

proteins

of thetwoclones shared

antigenic

determinants with each

other,

and also

con-firming that the

affinity-purified antibody

remained immuno-logically active. Eluates from

wild-type plaques

after

incuba-m

be

qL

I

*_ 1

w

A.

CL

z~~~~

z I

2

21

4 5 6

l

ba

4..

I *

1 2 3

u

4..

I

I. I

4 5 6

V- V

m

U

T

PM-Scl

ad

3: n 3: --j (.) I -;

--200

-97

-44

-27

Figure 4. Immunoblot withaffinity-purifiedantibodies. Whole HeLa cellextract asantigenwaselectrophoresedina10%SDS-PAGE in alllanes,transferredtonitrocellulose,anddevelopedwith

affinity-purifiedantibodies fromplaquesofJH4B,clone,

JH4C,

clone,orwild-type (wt) X-gtl1,anti-PM-Sclsera(JH, CK, MJ, FL),

orcontrolserawith anti-Jo- l (Jo-I)oranti-EJ(EJ). Althoughwhole JH serumstainsmultiplebandsofHeLacell extract,affinity-purified

antibodiesfrom either clone stainonlythe I00-kDprotein.The

posi-tionof molecularmassmarkers inkilodaltons is shownattheright.

PM-SC

1 2 3

3 3 3M 3 a

j'

L 1

-:> .; --

I.

_? .FM _?

V

1

4 5

3. is3. 3: X- 3

-68

-44

-27

J H4C1 wild type

Figure 3. Immunoblot of E. colilysates.Lysates ofE.coliprepared either withJH4CIXgtl1(left)orwild-type(right)were

electrophor-esed ina 10%SDS-PAGE and developed with theseraindicated: 1,

mouseanti-13-galactosidase (developedwithanti-mouse conjugate);

2and3,normalhumansera;4-6, anti-PM-Sclsera. Theposition of

molecularmassmarkers inkilodaltons is shownatthe left.

Figure5. Immunoblotwithaffinity-purifiedantibodies against

anti-PM-Scl immunoprecipitates. Immunoprecipitates preparedfrom HeLa cellextractusingJHserumwereelectrophoresedinall lanes of a10% SDS-PAGE andtransferredtonitrocellulose.Thepositionsof molecularmassmarkersinkilodaltonsareshowntotheright.Each

stripwasdevelopedwitheither wholeserum( Wh)at 1:00 dilution

oraffinity-purifiedantibodiespreparedfrom thatserumbyelution fromplaquesofJH4B, clone(Elu)orwild-typeXgtl1 ( Wt). The

serumineachcaseisindicated intheupperlineas:NL,normal serum;Ro, anti-Ro/SSA; Jo-l, anti-Jo-l; and anti-PM-Sclsera,

la-beled 1-5. Forserum 1(JH),lane Wh*wasdevelopedwithserum

that had been incubatedwithplaquesof

JH4Bj,

showingdecreased

stainingof the l00-kDdoublet,withoutdecrease in thestainingof the 70-kD band.

ComplementaryDNA Cloning ofPM-Scl 100-kD Protein 563

2

12-

170-

116-

76-Jo-1

a m

m3: NL Ro

mu 3-zu 3:

11

~

ixA

*.;..1

.--1,|1

Ii

iL,.m

Figure6.IndirectimmunofluorescenceonHEp-2 cells. (A) Whole JHserumusedat1:00dilution. (B) Affinity-purified antibody (antibody from

JH4B,plaques) usedundiluted.Antibodyfromwild-typeplaqueswascompletelynegative (not shown).

:h.i

tion with JH serum were negative against theplaquesof both clones.

Affinity-purified antibody from JH serum using clone JH4B1 stained a 100-kD protein inimmunoblotting against whole HeLa cellextract, similarin size to the 100-kDprotein stained by whole JH serum(Fig. 4). However, otherproteins

stainedinimmunoblotting by wholeJH serum,includingthe 70-kD protein, were not seenwith

affinity-purified

antibody.

Wild-type eluates were negative in immunoblotting.

Further testingwas performed todetermine whether this

result wastypical ofotheranti-PM-Sclsera.Anti-PM-Scl

im-munoprecipitates from HeLa cell extractprepared using JH serum wereused asantigen ratherthanwhole HeLa cell ex-tract. Whole JH serumrecognizeda 100-kD doubletand the 70-kDband whereas

affinity-purified

antibodyfromJH serum

recognized onlythe100-kD doublet (Fig. 5). After incubation

with JH4B1 plaqueproteins, staining ofthe 100-kD doubletby JHserum wassubstantially decreasedbutstainingofthe 70-kD band was not. Antibody was

similarly

affinity purified from

four of the otheranti-PM-Sclserathathadreactedwithboth

the 100-and 70-kD bands, usingboth

_JH4B,

andwild-type XgtI1. In each case, the

affinity-purified

antibody fromtheJH4

clone reactedwith the 100-kD doublet of the

immunoprecipi-tate but not the 70-kD band. The wild-type eluates were all

negative (Fig. 5).

JH serumantibodies

affinity

purified fromthe

_JH4B,

clone

were alsofoundtostainthenucleoliofHEp-2cells, ina pattern similar to that seen with whole JH serum (Fig. 6) whereas

wild-type eluateswerenegative. Affinity-purified antibodywas

alsoactivein

35S-immunoprecipitation.

Inadditiontothe 100-kD band,

affinity-purified

antibody immunoprecipitated the smallerbands (under 40 kD) typical of the immunoprecipita-tionpatternof

anti-PM-Scl

sera,apparently duetotheir affin-ity for the 100-kD bandas partofthe PM-Scl complex (Fig.7).

U 4

M2I -J

-195

Figure

7.

Immunoprecipitation

with

affinity-purified antibody. Immuno-precipitatesfrom

35S-methionine-la-_ ₁₀₅ beled HeLa cellextracts

using

anti-PM-Sclserum(PM-Scl),

affinity-pu-rifiedantibodyelutedfromJH4B,

-71 (JHB clone), wild-type eluate (Wild

XgtJJ),

orcontrolserawith

_anti-Sm,

* anti-PL-12 withanti-Ro/SSAand ***_44

anti-La/SSB,

andanti-Jo-I,were

electrophoresed ina 10%SDS-PAGE

* andautoradiographed.Theposition of molecularmassmarkersin kD is

.*

_* shownattheright. Theentire PM-Scl

complexseenwith whole anti-PM-Scl _ -2 7

a_W serumisimmunoprecipitated bythe

JH4B,

eluates, exceptfor the

artifac--1

8 tual bands

_surrounding

the 44-kD marker. The PM-Scl pattern is very ^-15 faintlyseenwith control eluates but

W -15 wasmuch stronger withJH4B,.

A

i

I

U

i Ae

0.5

.1:0115

1

o

,

0 0.5 .0 1.5 2.0 2.5 kb

B TaqI/KpnI

C SRmT

K

I '%

2.5 kb

S %,S

|

l ~ -.- l.I -

%-6

D 5' Extension

0 Insert DNA

2.5 kb

3' Extension -2

2.5 kb

Figure8.Sequencing strategyforcDNAencodingthePM-SclIO0-kD

protein.Bar Ashows a restriction enzyme map ofthecleavagesites on theJH4B, cloned cDNA. In bars B andC,the arrows show the maximumsequencedlengthanddirection for eachfragment. Bar D showsthe extended ends.

Antibody elutedfromwild-type plaquesalso showed this pat-tern,apparently becauseofbackground bindingofpatient

anti-body

to theplaques, but itwas much weaker than thatseen

with antibody eluted from the clone. These studiesindicated that anti-100-kD anti-PM-Scl antibodiesreacted withthe pro-teinexpressedbythe JH4 clones.

Sequencing ofPM-Scl

cDNA. The DNAs of clones

JH4B1

and

_JH4C,

weredigestedwithEcoRI.Agarose gel electrophore-sisofthe products showedasimilarsized single insert of - 2.5

kb for each clone. Inserts from both clones were subcloned into theEcoRIsite ofpUC18.Partialsequencing from the

double-strandedplasmidDNA showed that both clones hadidentical

terminal sequences.

Arestriction map of the

_JH4B,

insert was obtained by

sin-gle and double digestions with selected restriction enzymes

(Fig. 8).Four fragments were obtained by digestion withKpnI

andTaqI, andthreefragments wereobtained followingSmaI digestion. All seven fragments ofJH4B1 were then isolated

from agarosegels, subclonedinto Ml3mp18,andsequenced. Subclonesfrom each of thesestrategieswereused to verify the positions of the five restriction sites shown in Fig. 8.

The sequence of both strands of the JH4B1 insert was thus

determined. IncludingEcoRI linkers (not shown), it contained 2,479 nucleotides, with a single long open-reading frame of 2,331 nucleotides from the first ATG codon atposition288 in Fig. 9 to a stop codon at position 2,619. Thepredictedprotein from the insert cDNA included 777 amino acids, with a pre-dicted molecular mass of 89 kD, smaller than theobserved

massof the antigenic protein in this study (100 kD) or that reported in other studies(110 kD). Nopoly-(A)tailwas de-tected in the cDNA insert.

Northern blot

analysis.

Northern blot

against

both total RNA and

poly-(A)'

RNA from HeLa cellsshowed hybridiza-tion of 32P-labeled

_JH4B,

cDNA insert with asingleband of2.7 kb(Fig. 10). Similar resultsshowing hybridization ofasingle 2.7-kb band were obtainedusingtotal RNA from an SKHepl cell line from human liveradenocarcinoma (kindlyprovided by Dr. P.Kincade,OklahomaMedicalResearchFoundation)

ComplementaryDNA Cloning ofPM-ScllOO-kDProtein 565

T T

11 4

GACAAGcTCTCGCGAGACGAGcCGTGC&GGCTGAAAAAATGGCGCC&CcCAGTAccCGGGAGcCC&GGGTCCTGTCGGCG

MetAlaProProSerThrArgGluProArgValLeuSerAla 14

81

ACCAGCGCAAC~CAA&TCCXACGGAGKGATGGTGCTGCC&GGcrnTccCGGACGCCGACGaccTTGTGA&G.L

=X rLjrrDGG

ThrSerAlaThrLysSerAspGlyGluMetValLeuProGlyPheProAspAlaAepSerPheValLysPheAlaLeuGly

41

162 CATCGGGnCARC&C&GTATGATTTTTACCGAAGTTTTCCTGGCA

SerValValAlaValThrLysAlaSerGlyGlyLeuProGlnPheGlyAspGluTyrAspPheTyrArgSerPheProGly

68

243 TTCCAAGCATTTTGCGAAACACAGGGAGACAGGTTGCTTCAGTGCATGAGCAGAGTAATGCAGTACCATGGGTGTCGCAGC

PheGlnAlaPheCysGluThrGlnGlyAspArgLeuLeuGlnCysMetSerArgValMetGlnTyrHisGlyCysArgSer

95

324 AACATT.'AGGATCGAAGTAAAGTGACTGAGCTGGAAGACAAGTTTGATTTACTAGTTGA.TGCCAATGATGTAATTCTGGAG

AsnIleLyBAspArgSerLysValThrGluLeuGluAspLysPheAspLeuLeuValAspAlaAsnAspValIleLeuGlu

122

405 AGAGTGGGTATTTTACTGGATGAAGCCTCAGGTGTAAACAAGAATCAACAGCCTGTCCTCCCTGCCGGCTTGCAGGTCCCC

ArgValGlyIleLeuLeuAspGluAlaSerGlyValAsnLysAsnGlnGlnProValLeuProAlaGlyLeuGlnValPro

149

486 AAAACGGTAGTGTCCAGCTGGAACCGTAAGGCAGCAGAATATGGCAAAAAAGCAAAATCTGAAACTTTCCGGCTGCTTCAT

LysThrValValSerSerTrpAsnArgLysAlaAlaGluTyrGlyLysLysAlaLysSerGluThrPheArgLeuLeuHis

176

567 GCAAAAAATATCATCCGACCTCAGCTCAAGTTTCGAGAGAAGATTGACAATTCCAACACACCATTTCTTCCTAAAATCTTC

AlaLysAsnIleIleArgProGlnLeuLysPheArgGluLysIleAspAsnSerAsnThrProPheLeuProLysIlePhe

203

648 ATCAAACCCAATGCTCAGAAACCTCTCCCTCAAGCTCTCTCTAAGGAAAGGCGGGAACGCCCACAGGATCGTCCTGAGGAC

IleLysProAsnAlaGlnLysProLeuProGlnAlaLeuSerLysGluArgArgGluArgProGlnAspArgProGluAsp

230

729 TTGGACGTCCCCCCTGCACTGGCTGATTTCATCCATCAGCAGAGAACCCAGCAGGTTGAGCAAGACATGTTTGCACATCCT

LeuAspValProProAlaLeuAlaAspPheIleHisGlnGlnArgThrGlnGlnValGluGlnAspMetPheAlaHisPro

257

810 TATCAATATGAACTAAATCACTTTACCCCAGCAGATGCAGTGCTTCAAAAGCCACAACCCCAGTTATACAGACCTATAGAA

TyrGlnTyrGluLeuAsnHisPheThrProAlaAspAlaValLeuGlnLysProGlnProGlnLeuTyrArgProIleGlu

284

891 GAGACACCATGCCATTTCATATCCTCCCTGGATGAACTCGTGGAACTCAACGAAAAGCTCTTGAATTGTCAGGAATTTGCA

GluThrProCysHisPheIleSerSerLeuAspGluLeuValGluLeuAsnGluLysLeuLeuAsnCysGlnGluPheAla

311

972 GTTGACTTGGAGCACCACTCTTACAGGAGCTTCCTGGGACTGACCTGCCTGATGCAAATTTCTACTCGGACGGAAGACTTC

ValAspLeuGluHisHisSerTyrArgSerPheLeuGlyLeuThrCysLeuMetGlnIleSerThrArgThrGluAspPhe

338

1053 ATCATTGACACCCTCGAGCTTCGAAGTGACATGTACATTCTCAATGAGAGCCTCACAGACCCAGCCATCGTTAAGGTCTTT

IleIleAspThrLeuGluLeuArgSerAspMetTyrIleLeuAsnGluSerLeuThrAspProAlaIleValLysValPhe

365

1134 CATGGTGCTGATTCAGACATAGAATGGCTACAGAAAGACTTTGGGTTGTATGTAGTAAACATGTTTGATACTCATCAGGCA

HisGlyAlaAspSerAspIleGluTrpLeuGlnLysAspPheGlyLeuTyrValValAsnMetPheAspThrHisGlnAla

392

1215 GCACGCCTTCTTAACCTGGGCAGGCACTCACTCGATCATCTCCTGAAACTCTACTGCAACGTGGACTCAAACAAGCAATAT

AlaArgLeuLeuAsnLeuGlyArgHisSerLeuAspHisLeuLeuLysLeuTyrCysAsnValAspSerAsnLysGlnTyr

419

1296 CAGCTGGCTGATTGGAGAATACGCCCTCTGCCCGAGGAGATGCTCAGCTACGCCCGGGATGACACCCATTACCTGCTATAT GlnLeuAlaAspTrpArgIleArgProLeuProGluGluMetLeuSerTyrAlaArgAspAspThrHisTyrLeuLeuTyr 446

1377 _{ATCTATGACAAAATGAGGCTGGAGATGTGGGAGCGCGGCAACGGGCAGCCCGTGCAGCTGCAGGTGGTGTGGCAACGGAGC}

IleTyrAspLysMetArgLeuGluMetTrpGluArgGlyAsnGlyGlnProValGlnLeuGlnValValTrpGlnArgser

473

Figure

9.Thecompletenucleotide andpredictedaminoacidsequenceof the PM-ScII00-kDprotein.The sequence in regular type was derived

fromthe

_JH4B,

cDNAwhereas thesequencesin boldwerederived from the extendedends. The asterisk indicates the stop codon. The

super-scripted

dots markevery 10thnucleotide. The underline indicates thepolyadenylationsignal. Nucleotideposition is shown at the left and amino

acid

position

is shownonthe

_right.

(not shown).

This indicated that the JH4 cDNA inserts, at Determinationofthe

'and3'ends

ofthefull-length cDNA.

<2.5

_kb,

included

most but not all of the full length ofthe The ends ofthe cDNA were extended and amplified as

de-corresponding

mRNA.

scribed

in Methods. The products of

5'

end amplification

566 Q.Ge,M.BartonFrank,C.O'Brien,andI. N.

Targoff

1458 AGGGACATCTGCCTCAAGAAATTCATCAAACCTATCTTCACGGATGAGTCCTACCTTGAACTCTATAGGAAGCAGAAGAAG

ArgAspIleCysLeuLysLysPheIleLysProIlePheThrAspGluSerTyrLeuGluLeuTyrArgLysGlnLysLys

1539 CACCTTAACACACAGCAGTTGACAGCCTTTCAGCTGCTGTTTGCCTGGAGGGATAAAACAGCTCGCAGGGAAGATGAAAGT

HisLeuAsnThrGlnGlnLeuThrAlaPheGlnLeuLeuPheAlaTrpArgAspLysThrAlaArgArgGluAspGluSer

1620 TACGGATATGTACTGCCAAACCACATGATGCTGAAAATAGCTGAAGAACTGCCTAAGGAACCTCAGGGCATCATAGCTTGC TyrGlyTyrValLeuProAsnHisMetMetLeuLysIleAlaGluGluLeuProLysGluProGlnGlyIleIleAlaCys

1701 TGCAACCCAGTACCGCCCCTTGTGCGGCAGCAGATCAACGAAATGCACCTTTTAATCCAGCAGGCCCGAGAGATGCCCCTG CysAsnProValProProLeuValArgGlnGlnIleAsnGluMetHisLeuLeuIleGlnGlnAlaArgGluMetProLeu

1782 CTCAAGTCTGAAGTTGCAGCCGGAGTGAAGAAGAGCGGACCGCTGCCCAGTGCTGAGAGATTGGAGAATGTTCTCTTTGGA

LeuT.vsSerGluValAlaAlaGlyValLysLysSerGlyProLeuProSerAlaGluArgLeuGluAsnValLeuPheGly

1863 CCTCACGACTGCTCCCATGCCCCTCCGGATGGCTATCCAATCATCCCAACCAGTGGATCTGTGCCAGTTCAGAAGCAGGCG

ProHisAspCysSerHisAlaProProAspGlyTyrProIleIleProThrSerGlySerValProValGlnLysGlnAla

1944 AGCCTCTTCCCTGATGAAAAAGAAGATAACTTGCTGGGTACCACATGCCTGATTGCCACAGCTGTCATCACGTTATTTAAT

SerLeuPheProAspGluLysGluAspAsnLeuLeuGlyThrThrCysLeuIleAlaThrAlaValIleThrLeuPheAsn

2025 GAACCTAGTGCTGAAGACAGTAAAAAGGGTCCATTGACAGTTGCACAGAAAAAAGCCCAGAACATCATGGAGTCCTTTGAA

GluProSerAlaGluAspSerLysLysGlyProLeuThrValAlaGlnLysLysAlaGlnAsnIleMetGluSerPheGlu

2106 AATCCATTTAGGATGATCAGCAACCGTTGGAAGCTGGCCCAGGTACAAGTACAAAAAGACTCTAAAGAAGCTGTCAAGAAG

AsnProPheArgMetIleSerAsnArgTrpLysLeuAlaGlnValGlnValGlnLysAspSerLysGluAlaValLysLys

2187 AAGGCAGCTGAGCAAACAGCTGCCCGGGAACAGGCAAAGGAGGCGTGCAAAGCTGCAGCAGAACAGGCCATCTCCGTCCGA

LysAlaAlaGluGlnThrAlaAlaArgGluGlnAlaLysGluAlaCysLysAlaAlaAlaGluGlnAlaIleSerValArg

2268 CAGCAGGTCGTGCTAGAAAATGCTGCAAAGAAGAGAGAGCGAGCAACAAGCGACCCAAGGACCACAGAACAGAAACAAGAG

GlnGlnValValLeuGluAsnAlaAlaLysLysArgGluArgAlaThrSerAspProArgThrThrGluGlnLysGlnGlu

2349 AAL ArAgAGeuyICTAAAATlTTCLCAAGAAs CAAAGGAsC GAGCAro AGlAuyAsGTLuACGherrTTrA TyAreAG AGln LysLysArgLeuLysIleSerLysLysProLysAspProGluProProGluLysGluPheThrProTyrAspTyrSerGln

2430 TCAGACTTCAAGGCTTTTGCTGGAAACAGCAAATCCAAAGTTTCTTCTCAGTTTGATCCAAATAAACAGACCCCGTCTGGC

SerAspPheLysAlaPheAlaGlyAsnSerLysSerLysValSerSerGlnPheAspProAsnLysGlnThrProSerGly

2511 AAGAAATGCATTGCAGCCAAAAAAATTAAACAGTCGGTGGGAAACAAAAGCATGTCCTTTCCAACTGGAAAGTCAGACAGA

LysLysCysIleAlaAlaLysLysIleLysGlnSerValGlyAsnLysSerMetSerPheProThrGlyLysSerAspArg

2592 GGCTTCAGGTACAACTGGCCACAGAGATAGTCCTGGAAGACACGTGGCGCCTGTGGACCGGAAGCACCAAATGCTGGTGCT GlyPheArgTyrAsnTrpProGlnArg *

2673 GCTTTTG1Ti A T r A A T L

Figure9(Continued)

showed a predominant band ranging from 320 to 340 bp, whether totalor

poly-(A)'

RNAwasused for thetemplate for

cDNAsynthesis. InsomePCRamplifications using thesame

cDNA astemplate, analysis of the products showed multiple bandsinadditiontothepredominant 340-bp band. However, afterelectroblotting of electrophoresed PCR products ontoa

nylon membrane, only the 340-bp band hybridized with 32P-la-beled

JH4B,

insert.

Toidentify the longest extensionspresent,arangeof

prod-uctswereisolated from the polyacrylamide gel bydiffusion.

Aliquots of PCR products were subcloned into pUC1 8 or

PCR 1000 (see Methods). 19 subcloneswereobtained and 11

wereselected and sequenced from plasmid mini-preps.These

11 werederived from 5 different PCRs thatwere performed

using cDNA pools from threereversetranscription reactions.

The sequencesshowed that all of the clones startedfrom the

gene-specific primer, contained the 88 bp of knownsequence

from the

_JH4B,

cDNA (discounting the EcoRI linker),

ex-tended 180-210nucleotidesbeyond the 5' end of the cDNA, and then showed thepoly-(T) expected from the tailing

reac-tion. The first initiation codon of the

maximum-length-ex-tended5'portionwasatposition 39 (Fig. 9), outsidethe origi-nal cDNA. All of the 5' extension subclones contained this codon. Therefore, all PCR-produced 5' ends gave the same

ComplementaryDNA Cloning ofPM-Scl 100-kDProtein 567 500

527

554

581

608

635

662

689

716

743

770

797

824

851

4

Figure10. Northern blotofHeLa cell RNA. 40

_Ag

(left)or 20

,g

(middle) of total HeLa

RNA, or8

,4g

ofisolated

poly-(A)'HeLaRNA, wereelectrophoresed in

a1% agarose gel, trans--2.7kb ferred to Nytran

mem-brane, andhybridized with32P-labeledPM-Scl

100-kD cDNA insert isolated from clone

JH4Bj.

Asingle band wasidentified in all threepreparations.The

positionof size markers is shown at the left.

putative translational starting site, which differed from that

predicted from the initial cDNA insert by 249nucleotides(83 aminoacids). None of these subclonescontainedanin-frame stopcodonupstreamof thefirst ATG.

Two slight discrepancies were noted among the 11

se-quenced 5' extensions: 1of11showedachangefromaTtoaC

at position 96of the 5' extension whereas another showeda

change fromaGtoanAatposition191.Althoughit ispossible that one ofthese represents a different allele, their low

fre-quency suggests that theymore likely represent mutant Taq polymerase products.

Although a stop codon was present in the open-reading frame ofthe cDNA insert (Fig. 9, position 2,619),thepoly-(A)

tailwasnotpresent,and thereforethe 3' endwasamplifiedas

well. PAGE of PCR amplification productsof the 3' extensions

showed a predominant band at 180 bp, but multiple other

bands were identified. Southern blotting and hybridization with a 32P-labeled cDNA insert, however, showed hybridiza-tion only with the 180-bpband. Analiquotof PCRproducts

was digested with XbaI and PstI and subcloned into pUC18

and PCR1000. Eightsubcloneswereselected for sequencing.

Thesequencesof all ofthe clones started from the gene-specific

primer 99 bpupstream from the 3'end oftheoriginal cDNA

insert, included the 90 bp of knownsequence uptothe EcoRI linker, extended 32 nucleotides further, and ended witha poly-(A) tail that usually contained - 20bp. Two of the eight

se-quenced 3' extensions showedanabsence ofthe three nucleo-tides 2,716-2,71 8 justupstreamofthepoly-(A)tail.Aputative polyadenylation signal, ATTAAA,waslocated 19 nucleotides from the poly-(A) tail within the extended region.

Expression of the recombinant protein. E. coli M 15-(pREP4)wastransformed withtypeIIIconstructsofpQE plas-mid containing PCR-amplified insert coding for the PM-Scl 100-kD protein asdescribed in Methods. Theexpressed

pro-tein lacked the amino-terminal 10 amino acids and the

car-boxyl-terminal 3 amino acids of the predicted protein, but the primers and the expressionsystemadded 15 unrelated amino acids,sothat the expressed proteinwaspredictedtobe2 amino acids longer and 381 Dgreater in molecular mass than the nativeprotein. Anti-PM-Sclserarecognizedaproteinin lysate

ofinduced E. colitransformed with vector containing this in-sert, butnot in controls without insert orwith recombinant plasmids without IPTG induction (Fig. 11). The mobility of the E. coli-expressed proteininPAGEwasslightly larger, by

- 2-4 kD, than the native HeLa cellimmunoprecipitated

100-kD protein.

Analysis ofthefull-length

sequence.Thefull-length

nucleo-tide sequence of thisPM-Scl gene including the extended 5' and 3' ends was2,739 bp long. This sequence consists of 38 nucleotides in the 5' untranslated region, 2,580 bp in the pre-dictedcoding regionfromthe first methionine codon, and100

nucleotides from the stop codon up to the 21 nucleotide poly-(A)tail(Fig. 9). Thededuced amino acid sequence was

com-posed of 860 residues with a predicted molecular mass of 98,088 D, very close in size to the observed relative molecular weight in PAGE of the PM-Scl 100-kDantigen.

Severalpossiblesites ofposttranslational modificationwere observed. Two potential N-glycosylation sites were identified at residues 353 and 839. A number of potential phosphorylation

siteswereidentified, includingthose forprotein kinase C (at residues 849, 823, and846),casein kinase II (at residues 18, 104, 292, 293, 333, 370, and665), and tyrosine kinase (at 523).There were also amidation sites identified at residues 163

and823.However,the presence ofthese sites does not

necessar-j4-HeIa-*I

[4 - E.coli - H

Insert Insert No Insert

IPTG NoIPTG IPTG

F I

_I~j

I

Zn 1 2 z ° 1 2 z- 1 2 z ° 1 2

200-97-

1uI,

68- g

44-

27-Figure 11. Immunoblot of expressed recombinant protein.

Immuno-precipitatefrom HeLalysateusingserumJH(Hela)orlysatesof recombinantE. coliwereelectrophoresedina 10%SDS-PAGE,

transferredtonitrocellulose, and developed with eithernormalserum

(NL),anti-Jo-I serum(Jo),or oneoftwoanti-PM-Sclsera(lanes

1and2). Insert/IPTG:E. coli transformed with plasmidvector

con-tainingthePCR-amplified PM-Scl 100-kD insert, induced with IPTG;Insert/no IPTG:E.coli transformed with plasmid containing

insertbutnotinduced;NoInsert/IPTG:E.coli transformedwith

plasmid without insert and induced.The 100-kD band recognized in

immunoprecipitates by anti-PM-Sclsera(near the97-kD marker)

wassimilarinsizetothefull-lengthexpressedprotein.Multiple

weakeradditional bands ofexpressedprotein reacting with

anti-PM-Sclappeartorepresentincomplete, partial-lengthexpressedprotein

ordegradation products.

568 Q.Ge,M.BartonFrank,C.O'Brien,andLN. Targoff

7.5kb-

4.4kb-

2.4kb-1

.4kb-z

cc j +

m _'R *%

4---6

ilyindicate that posttranslational modificationoccurs atthese sites, and theyareoften observed in proteins thatare not modi-fied. There has been noevidence of posttranslational

process-ing ofthe PM-Scl 100-kD protein reported.

One reportindicated that a number oflupus and

sclero-dermaautoantigens showedahigherfrequencythan other

pro-teins ofchargeruns (alongseries ofconsecutivecharged amino

acids) and charge clusters(ahighfrequency ofcharged amino

acids, not necessarily consecutive, within a given stretch of

aminoacids) (25). Althoughnocharge runs were observed in

this PM-Scl protein, a "mixed charge cluster" (consisting of

both positive and negative residues)was foundbetween resi-dues 753 and 789, with 14 positive and 8negativelycharged

residues among 37 amino acids, for a total of 59% charged

amino

acidswithin this region.

Therewas nosignificanthomology with the sequence ofthe

75-kD PM-Sclprotein (10), and therewere no sequences of

more thanfour consecutive amino acids shared between the two

proteins.

Independence ofthetwomajor PM-Scl antigenic proteinsontheamino acid level is consistent withthe

observa-tion noted above that they were antigenically independent,

since

affinity-purified

JH4B,

eluates stained the 100-kD

pro-tein butnotthe 70-kDprotein. Alderuccioetal.(10)had simi-larlynotedspecific reactivity withthe expressed 75-kDprotein without cross-reactionwith the 1I0-kD protein.Nosignificant homologywithothereukaryotic proteins inthe data banks was

identified.

Discussion

Inthis study, the

full-length

nucleotide and predicted amino acidsequence

for

the 100-kD

antigenic protein

of

thePM-Scl

antigen

wasdetermined. The

l00-kD protein

wasfoundtobe

antigenic

for thegreat

majority

of

patients

withanti-PM-Scl antibodies,evenifno

activity

was seen onroutine

immunoblot-ting.

This proteinappears tobe

independent

of the 70-kD

pro-tein

immunologically,

consistent with the lack of primary amino acidsequence

similarity

between their predicted

poly-peptides.

Strong

support was obtained for the conclusion that the clonedcDNA encodesa

protein

thatreactswith

anti-PM-Scl

antibodies.

First,

extensive serum

testing

demonstrated that reaction with recombinant

protein

was

specific

forsera with

anti-PM-Scl.

Noreactionwas seenwithsera

containing

other autoantibodies or normal human sera.

Second,

the

affinity-purified

antibodies that reacted with the fusion

protein

demon-strated similar

specificity

tothatof antibodiesto the 100-kD PM-Scl

protein.

Theelution from the recombinant

plaques

resulted in

affinity

purification

of antibodies reactive with the 100-kD

protein,

but nottheantibodiestothe 70-kD

protein

fromthe same sera,

confirming

the

specific reactivity

of the fusion

protein. Affinity-purified

antibodies alsoreacted in

im-munoflourescence

and

in

immunoprecipitation

as

expected

for

anti-PM-Scl.

Much

higher

activity

was seen with

_JH4B,

eluates than

wild-type

eluates in

immunoprecipitation,

and the background

reactivity

of

wild-type

eluateswas

probably

dueto

nonspecific

binding.

Because the

patient

serum contained a

high titer of

antibody,

specific anti-PM-Scl

would bea

signifi-cant

proportion

ofany

nonspecific

antibody

binding

that

oc-curred.

Evidence was also provided to support the conclusion that

thecDNA codes for the same l00-kD PM-Scl protein thatis

partof-theimmunoprecipitated PM-Scl antigen complex.The size oftheexpressedrecombinantproteinand the immunopre-cipitated antigenic proteinwereshown to be similar.

Affinity-purified antibodies recognizedthe l00-kD protein immuno-precipitated from HeLa cell extract by anti-PM-Scl sera.

How-ever,thepossibilitythat the cDNA codes for a separatePM-Scl 100-kD protein that cross-reacts with the 100-kD antigenic protein ofthe immunoprecipitated PM-Scl complex has not beenunequivocally excluded, since direct amino acid sequence

of theprotein was notavailable.Ofinterestwas thatinsome cases,anti-PM-Sclserastainedadoubletat100kDin immuno-blotting. The twoforms of thisprotein shared epitopes since both reacted equally with affinity-purified antibodies. This

may representadegradation productassuggested byits inter-mittentappearancebutotherpossibilities, suchas adifference inposttranslational modifications,or,lesslikely,twodifferent but closely related proteins,cannotbe excluded.

Thereisalso strong supportfortheconclusionthatthis is

thefulllengthof thecDNAsequence.The mRNA detected by

Northern blot wasthesamesizeastheoriginalcDNAplusthe extended ends. The predicted protein from the sequence is

within 2% oftheestimatedmolecularweight (100kD) of the native antigenic PM-Scl protein observedbySDS-PAGE. The recombinant protein appearedevenlarger than the immuno-precipitated protein probably because of

differences

in

process-ing or posttranslational modifications. Others have reported theproteinto be 1 10 kD. If 110kD isthetruephysical molecu-lar mass,the

difference

insize from the predicted proteinmay

resultfrom posttranslational modificationsatoneofthe

poten-tial sites identified above. The sequence preceding the first ATGis consistent with commonly observed starting sites in that an A is present in the third position before the ATG whereasit isusually notfound in the unusualeventofanATG codonupstreamof thestarting ATG (26). There isno Athree nucleotides upstream to the second ATG, which is found withintheoriginal insert, suggesting itwas not thestarting site. Supporting the identification ofthe 3' end,aclearstopcodon is

presentand the extended end containstheexpectedpoly-(A) tail.

Although

the most common

polyadenylation

signal is AATAAA,ATTAAAisarecognized naturally occurring vari-ant found in 12% ofmRNAs and provides almostas much

polyadenylation activity

astheAATAAAsignal (27). Although the 100-kD proteinwas the protein most

com-monly

recognized

by

anti-PM-Scl

positiveserain immunoblot-tingagainst whole HeLa cellextract,reactionwas not seenwith

serafrom41%of

patients

with anti-PM-Scl. Although

Gelpi

et

al.

(9)

found that all of 10

anti-PM-Scl

serareacted with the

1 I0-kD

protein,

Alderuccioetal.(10) found results similarto ours,with

only

halfof

anti-PM-Scl

sera

reacting

in immuno-blotagainstthe100-1 10-kD

protein.

However,wehave found thatserafromasmanyas94.9% of

patients

with anti-PM-Scl

reactedwith theproteinexpressed inIPTG-inducedplaquesof theclone.Themost

likely

explanation isthat some sera

recog-nized exclusively conformational

epitopes

that were lostby

de-naturation during SDS-PAGE butwere presentonthefusion

protein.

Although the recombinant

portion

offusion

proteins

isoften denaturedorotherwise presentedinanonnativestate,

this doesnotexclude the

possibility

ofconformation.

Alterna-tively,

lowabundanceofthe

100-ckD protein

inthewholeHeLa

extract comparedwiththe plaquesmay also haveaccounted

for thedifference.

Sera oftwo patients did not react with plaques of either

cloneor with the 100-kD proteininimmunoblot, suggesting that thosetwo sera did notrecognizetheI00-kD

protein

in any

form. Since theyshowed lines of identity withstandard

anti-PM-Sclserain immunodiffusionandshowedthetypical

PM-Scl complex inimmunoprecipitation, they clearly react with

PM-Scl. They may react with aconformational epitope not

presentonthefusion protein;they may react with the native

75-kD protein only;orthey may reactexclusivelywith other

componentsofthe complex. Both ofthe twoseradid react with

other bands in immunoblot, but it is not yet known if these

bandsarecomponents of the complex.

Anti-PM-Scl

isanexamplein which aprotein complexis

targeted

by autoantibodies. Asdemonstratedin this and other

studies, antibodyto the 100-kD protein is the most common

specificity

found.Thissuggeststhat the 100-kDband may be

the maintargetofthe anti-PM-Sclresponse. However, reaction

withatleastoneother complex componentiscommonlyseen,

that of 70-75 kD. In this study, specific antibodies reacting

with the 100-kD protein in immunoblot did not cross-react withthe 70-kDprotein. Othershave also found that antibodies tothetwoproteinsareindependent (10). Independent

reac-tion with multiple componentssuggests that thecomplexasa

whole hasbecome antigenic, which supports the

antigen-dri-ven nature of the autoantibody response. It is possible that

some

of

the other

protein

bandsrecognizedby

anti-PM-Scl

sera

in immunoblottingare also complexcomponents.

Preliminary

results of Bluthner et al. (28), presented in

abstract form,

described

cDNA cloningofasmall portion of the PM-Scl 100-kD proteinand indicated that allanti-PM-Scl

serareacted with that small portion, suggestingapredominant

epitope.

Thesequenceof thatfragment has not yet been

pub-lished, andwe aretherefore not able to compare it to our

se-quence. Identification of the location of the epitopes of the

PM-Scl 100-kD proteinwillbefacilitated by theavailabilityof

the full-length sequence determinedin thisstudy.Analysisof

the sequence didnotreveal motifs that mightsuggest a

specific

function,

but the availabilityof this recombinantproteinwill

assist in further studies of thecellular function and cellbiology of PM-Scl.

Acknowledgments

The authors thank Dr. Russell Rother, Dr. Kazuko Itoh, and Sally

Bleichingerforhelpful discussions;Drs. Morris Reichlin, Frank C. Ar-nett, Paul H. Plotz, and Frederick W. Miller for referral of sera; and Edward P. Trieu for experttechnicalassistance. The oligonucleotide primers were synthesized at the Molecular Biology Resource Facility of theSaint FrancisHospitalofTulsa Research Institute at the University ofOklahoma Health Sciences Center.

This work was supported by NIH grants AR-322 14,AI-27181, and

AI-21568;a grant from the Pittsburgh Supercomputer Center through

cooperativeagreementU41-RR04154from the National Institutes of Health Division of Research Resources; and by Department of

Vet-eransAffairs medical research funds.

References

1.Reichlin,M.,P. J.Maddison,I.N.Targoff,T.Bunch,F. C.Arnett,G. Sharp,E.Treadwell, and E.M.Tan. 1984. Antibodiesto anuclear/nucleolar

antigen inpatientswithpolymyositis-overlap syndrome. J. Clin. Immunol. 4:40-44.

2.Bernstein, R. M., C. C. Bunn, G. R. V. Hughes, A. M. Francoeur, and M. B.

Mathews. 1984. Cellular protein and RNA antigens inautoimmunedisease.Mol. Biol.Med. 2:105-120.

3. Treadwell, E. L., M. A. Alspaugh, J. F. Wolfe, and G. C. Sharp. 1984. Clinical relevance of PM-I antibody andphysiochemical characterizationof

PM-antigen. J.Rheumatol. 11:658-662.

4. Reimer, G., V. D. Steen, C. A. Penning, T. A. Medsger, Jr., and E. M. Tan. 1988. Correlates betweenautoantibodiesto nucleolar antigens and clinical fea-tures in patients withsystemic sclerosis (scleroderma).Arthritis Rheum. 31:525-532.

5. Genth, E., R. Mierau, P. Genetzky, C. A. Von Muhlen, S.Kaufmann,H. von Wilmowsky, M. Meurer, T. Krieg, H.-J. Pollmann, and P. W.Hartl.1990.

Immunogeneticassociations ofscleroderma-relatedantinuclear antibodies. Ar-thritis Rheum.33:657-665.

6. Wolfe, J. F., E. Adelstein, and G. C. Sharp. 1977. Antinuclear antibody with distinct specificity forpolymyositis. J. Clin. Invest.59:176-178.

7. Targoff,I.N., and M. Reichlin. 1985. NucleolarlocalizationofthePM-Scl

antigen. Arthritis Rheum.28:226-230.

8. Reimer, G., U. Scheer, J.-M. Peters, and E. M. Tan. 1986.

Immunolocaliza-tion and partialcharacterizationof anucleolar autoantigen (PM-Scl) associated

with polymyositis/scleroderma overlap syndromes. J. Immunol. 137:3802-3808.

9. Gelpi, C., A. Alguero, M. Angeles Martinez, S. Vidal, C. Juarez, and J. L.

Rodriguez-Sanchez. 1990. Identification of protein components reactive with

anti-PM-Scl autoantibodies.Clin. Exp. Immunol. 81:59-64.

10. Alderuccio, F., E. K. L. Chan, and E. M. Tan. 1991. Molecular character-ization of anautoantigen ofPM-Sclin thepolymyositis/scleroderma overlap syndrome. A unique and complete human cDNA encoding an apparent 75-kD acidic protein of the nucleolar complex. J. Exp. Med. 173:941-952.

11. Targoff,I.N., F. C. Arnett, L. Berman, C. O'Brien, and M. Reichlin. 1989.

Anti-KJ:a new antibody associated with themyositis/lungsyndrome that reacts with atranslation-relatedprotein. J. Clin. Invest.84:162-172.

12. Targoff,I.N. 1990.Autoantibodiestoaminoacyl-transferRNA synthe-tases for isoleucine and glycine: two additionalsynthetasesare antigenic in myosi-tis. J. Immunol. 144:1737-1743.

13.Chomczynski,P., and N. Sacchi. 1987. Single-step method of RNA isola-tion byacid-guanidiniumthiocyanate-phenol-chloroformextraction. Anal.

Bio-chem. 162:156-159.

14. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular Cloning: A laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 6.46-6.48.

15. Feinberg, A. P., and B. Vogelstein. 1983. A technique forradiolabeling DNA restrictionendonucleasefragments to high specific activity. Anal. Biochem. 132:6-13.

16. Feinberg, A. P., and B.Vogelstein. 1984. Addendum: a technique for

radiolabelingDNA restrictionendonucleasefragments to high specific activity. Anal. Biochem. 137:266-267.

17.Yanisch-Perron,C., J. Vieira, and J. Messing. 1985. Improved M 13 phage cloning vectors and host strains: nucleotide sequences of theM13mpl8 and

pUCl9vectors. Gene(Amst.).33:103-119.

18. Sanger, F., S.Nicklen,and A. R.Coulson. 1977. DNAsequencingwith

chain-terminatinginhibitors. Proc.Nat!.Acad. Sci. USA.74:5463-5467.

19. Tabor, S., and C. C. Richardson. 1987. DNA sequence analysis with a modified bacteriophage T7 DNA polymerase. Proc. Nat!. Acad. Sci. USA.

84:4767-4771.

20. Loh, E. Y., J. F. Elliott, S. Cwirla, L. L. Lanier, and M. M. Davis. 1989. Polymerase chain reaction with single-sided specificity: analysis of T cell receptor delta chain. Science (Wash. DC).243:217-220.

21. Frohman, M. A., M. K. Dush, and G. R. Martin. 1988. Rapid production offull-lengthcDNAs from raretranscripts: amplificationusing a single gene-spe-cificoligonucleotideprimer. Proc.Nat!.Acad. Sci. USA.85:8998-9002.

22. Southern, E. M. 1978. Detection of specific sequences among DNA frag-ments separated by gelelectrophoresis.J.Mol. Biol.98:503-517.

23. Devereux, J., P. Haeberli, and 0. Smithies. 1984. Acomprehensiveset of sequence analysis programs for the VAX. Nucleic Acids Res. 12:387-395.

24. Itoh, K., Y. Itoh, and M. B. Frank. 1991. Proteinheterogeneityin the human Ro/SSAribonucleoproteins.The 52- and 60-kD Ro/SSA autoantigens are encoded by separate genes. J. Clin. Invest.87:177-186.

25. Brendel, V., J. Dohlman, B. E. Blaisdell, and S. Karlin. 1991. Very long charged runs in systemic lupuserythematosus-associatedantigens. Proc.Nat!. Acad. Sci. USA.88:1536-1540.

26. Kozak, M. 1984. Compilation and analysis of sequences upstream from thetranslationalstart site in eukaryotic mRNAs.NucleicAcidsRes.12:857-872.

27. Wickens, M. 1990. How the messenger got its tail: addition of poly-(A) in the nucleus. Trends Biochem. Sci. 15:277-281.

28. Bluthner, M., E. Genth, and F. A. Bautz. 1989. Cloning of a cDNA-frag-ment coding for an epitope recognized byanti-PM-Scl-autoantibodies.In Molecu-lar and Cell Biology ofAutoantibodiesandAutoimmunity:Abstracts ofthe First

International Workshop.E. K. F. Bautz, J. R. Kalden, M.Homma,and E. M. Tan, editors. Springer-Verlag, Heidelberg, FRG. p. 2.