Identification of

(1)

Identification of

an

H-2Kd

^gene

using

a

specific cDNA probe

J.-H. Xin1, S.Kvist, and B. Dobberstein*

EuropeanMolecular Biology Laboratory, Meyerhofstrasse 1, Postfach 102209, D-6900Heidelberg, FRG

Communicated byB. Dobberstein Receivedon 8April 1982

AcDNAclone known to code for a mouse histocompatibility (class I) antigen was found to contain a sequence specific for a subpopulation of H-2 genes. This unique sequence is located in the 3' non-coding region close to the stretch of poly(A) nucleotides. A subclone containing this fragment (pH-2d-5) has been used to select hybridizing mRNA.

Translation of the mRNA in vitro shows thatH-2Kd mRNA is selected. Southern blot analysis of DNA from congenic recombinant miceshowthat at least one genecontaining this sequenceislocated at the K locus (region) of the major histo- compatibilitycomplex. This gene contains a 3.7-kbBgml and a 13-kb EcoRI restriction endonuclease fragment. This gene has beenisolated from a genomic DNA library.

Key words: histocompatibility genes/hybrid-selected translation/genemapping/3' non-coding region/Southern blot analysis

Introduction

Genes located in the major histocompatibility complex (MHC) play a fundamental role in several aspects of the cellular immune response (Klein, 1975). They code forthree classes ofproteins:class I, the transplantation antigens H-2K, D, L,andR(herecalled H-2 antigens); class II, the immune- response-associated antigens, Ia; and class III, complement components (Nathenson and Cullen, 1974; Klein, 1979;

Ploeghetal., 1981).

Genes coding for lymphoid differentiation antigens, Qa and Tlaarelocatedatthe telomeric side of the MHC(Flaher- ty, 1980). They are structurally similar to the H-2 antigens whicharemembraneglycoproteinswithamol.wt. of 43-47 kd and areassociated with

32-microglobulin

(Coliganetal., 1981; Michaelsonetal., 1977; Soloskietal., 1981).

Adetailed molecularanalysisof genescodingforH-2, Qa, and Tla antigens has becomepossible with the isolation of cDNA probes in several laboratories (Ploegh et al., 1980;

Kvistetal., 1981; Soodetal., 1981).Their sequenceanalysis confirmed thehigh degree ofhomology postulated for H-2 antigens encoded by different loci (Bregegere et al., 1981;

Steinmetz et al., 1981a). Southern blot

analysis using

these cDNA probes revealed that genes

coding

for H-2

antigens

constitute alarge multigene family (Steinmetz et al., 1981a, 1981b; Camietal., 1981). One gene locatedintheQa

region

has been isolated and characterized by sequence

analysis

(Steinmetzetal., 1981b). Fromthis

analysis

itwasconcluded that genescodingforH-2andQa

antigens

showa

high degree

of homology, and cDNA

coding

for H-2

antigens

cross-

hybridizes efficientlywith that

coding

forQaorTla

antigens

'Present address: Shanghai Institute of Biochemistry, Academia Sinica, PeoplesRepublicofChina.

*Towhomreprintrequestsshould besent.

Ltd.,Oxford,England.02614189/82/0104-0467$2.00.0

(Steinmetzetal., 198lb). Therefore,todefieH-2 antigens it is necessary to obtain specific DNA probes which characterize asubgroup of genes in this multigene family.

Inthis paper we reporton such a DNA probe derivedfrom the extreme3' end of thenon-coding region of a cDNAclone coding for an H-2 antigen. Using this probe in a Southern blotanalysis, we mapped a gene located in the K locus ofthe MHC.

Results

Selectionofa cDNA clonehybridizing to mRNA codingfor anH-2Kd antigen

H-2-like antigens coded by genes located at different loci of the MHC show extensive homology; cDNAs containing coding regions cross-hybridize (Coligan et al., 1981; Kvist et al., 1981; Steinmetz et al., 1981a). In order to identify and characterize genes from a particular region or locus it is necessary toobtain DNA probes of higher specificity.

Wetested cDNA clones coding for various parts of H-2 an- tigensfor their ability to hybridize selectively to H-2 mRNAs.

Clonescontaining extended regions common to all H-2, and possibly Qa and Tla, antigens should hybridize to mRNA coding for all H-2-like antigens. Those containing unique se- quencesshould only hybridize to a subgroup of mRNA which possesses such sequences.

Plasmid DNA of clones pH-2d-1, pH-2d-4, pH-2d-5 (Bregegereetal., 1981; Lalanne et al., 1982) and pBR322 was fixed onto nitrocellulose fiters andtested by hybrid-selected translation using spleen cell mRNA from DBA/2 mice.

HybridizedmRNAwaseluted andtranslated in a reticulocyte

CM

I0Q ^IQ.

LO

I la

CM1

0.

COI

m

Ql

MW x 10-3

t _-43

Fig.1.Selectivehybridizationof H-2 mRNAtocDNA clonepH-2dm-5.

Plasmid DNA ofpH-2d4l,pH-2d-4, pH-2d-5,andpBR322wasimmobiliz- edonnitrocellulose filtersandhybridizedwith DBA/2spleenmRNA.

Afterelution, mRNAwastranslated inareticulocytecell-free systemsup- plementedwithdogpancreasmicrosomes. Antigensinserted into microsomes werecharacterizedbySDS-PAGE andautoradiography.Theposi- tions of theH-2Kd, Ld, andDdantigensweredeterminedbyco-runningof invivo10-min labelledantigensprecipitatedwithalloantisera (Robinson, 1982).

467

K L

^00.

D

.Aw

(2)

cell-free system in thepresenceofdog pancreas microsomes (Kvist et al., 1981). Clone pH-2d-1 hybridized to mRNA codingforH-2D, L,and Kantigens (Figure1). Separationof theseantigens bypolyacrylamide gelelectrophoresis (PAGE) has beendemonstrated by Robinson (1982). Theregion of the H-2Dd antigen was obscured by endogenous proteins of the cell-free system. That the clone pH-2d-1 also hybridizes to mRNAcoding for H-2Ddantigens has been shown previously (Kvist et al., 1981).ClonepH-2d-4 preferentially hybridizesto mRNAcoding for H-2Kdantigensand clonepH-2d-5 nearly exclusivelybindsmRNAcodingfor H-2Kdantigen (Figure1).

Weconcluded that mRNA coding for H-2Kd antigenscon- tains anunique regionpresentin clonespH-2d-4andpH-2d-5.

Rabbit antiH-2antibodies unfortunately couldnotbe used in this analysis asthey didnot reactefficiently with H-2Kd and Ld antigenssynthesized in the cell-free system (Dobberstein et al., 1979). We believe that this is because, in the cell-free system,

f2-microglobulin

is notbound totheheavy chainS.

CODINGREGION 3 NONCODINGREGION

I-I-I

I I I

bp la $ $ $

I 5aI> I

5a 5b5

Fig.2. Partial restrictionmapsof cDNAclonespH-2d_l andpH-2d-5. The restrictionmaps wereconstructedasdescribed (Kvistetal., 1981). Thever-

ticalarrows indicate PstI sites. The filled box shows the coding region and the hatched boxinthe 3' non-coding region indicates thenon-homologous portionof thetwoclones(see Figure 3). Subclonedfragmentsused for hybridization areoutlinedby the horizontalarrowsanddenoted la, 5a, and5b.

20 40

Partof the 3' non-coding regionshowshigh diversity The cDNA clone pH-2d-5 was further characterized by restriction map analysis and Southern blot hybridizationus- ing subcloned PstI fragments 5a and b(Figure 2) as probes.

Fragment 5b inefficiently hybridized to pH-2d-1 whereas 5a hybridized to itefficiently (datanotshown). Thus, fragment 5bwasidentifiedastheunique regionhybridizing specifically to mRNA coding for H-2Kd antigen. Sequence analysis of fragment 5b revealedastretch ofpoly(A) residuesat oneend, and 37 residues at the otherend,thatwerenearly identicalto nucleotides 278-317 of the 3' non-coding portion ofpH- 2d-1 (Figure 3). pH-2d-5 thus contained part of the 3' non- coding region ofanH-2antigen. The sequence of the 3' non- coding regioninpH-2d-1 wasfurthercomparedwiththoseof the corresponding regions in clone pH-21I and in Qa pseudogene 27.1 recently characterized by Steinmetz et al.

(1981a, 1981b). Extensivehomology was again found up to bp 317 (Figure3). Following bp 317, pH-2d1, pH-2d-5, and gene27.1 diverge greatlywhereas clonespH-2d_1 andpH-2II show nearidentity.

FragmentSbhybridizesexlusivelyto 17S mRNA

Portions of the 3' non-coding regionofH-2 mRNAsmight be sharedbyunrelated mRNAs of different sizes. Totestthis, we compared probe 5b and coding region la inaNorthern blot hybridization on total mRNA from DBA/2 mouse

spleen. Under the conditions used, both probes exclusively hybridizedto 17SmRNA(Figure 4).Thevalue obtained here for theH-2 mRNA agreeswellwith that obtained earlier by sucrosegradient centrifugation (Dobbersteinetal., 1979).

60 80 100

pH-22d TGAAGACAGCTGCCTGGTGTGGACTTGGTGACAGACAATGTCTTCACACATCTCCTGTGACATCCAGAGACCTCAGTTCTCTTTAGTCAA GTGTCTGATGTTCCCTGTG pH-2II ---AC--T---GA---G----G----GGTC.---C---CA---CAT--- 27.1 ---AC--C---GA---G----G---TGTC---AC---A--CA---C---

pH-2 -1

pH-2II

120 140 160 180 200 220

--CT- --G-T-GT---A---C-TG---

27.1 --C--AT--A---T----G---C-TG--- ---T---G--- -C---T---

240 PstI 260 280 300 320

pH-2 -I CCAAACATTGGTGGACATCTGCAGCCTGTCAGCTCCATGCTACCCTGACCTTCAACTCCTCACTTCCACACTGAGAATAATAAT TTGAATGTGGGTGGCTGGAGAGATG pH-2II T---C---G---T---A---G--G--G--- ---

27.1 ----C--C---G-T---A--T---G---C--G--C--G----AT----T--- GA---G--TGTGACC----T---TATCAT--T---CCTGA

pH-2 -5 --G--G---G---- ---AACCTTGATTGTTATCA

340 360 380 tPstI 400 420 PstI 440

pH-2 -1 GCTCAGCGCTGACTGCTCTTCCAAAGGTCCTGAGTTCAAATCCCAGCAACCACATGGTGGCTCACAACCAATCTGTAATGGGATCTAACACCCTCTTCTGCAGTGTCTGAA

pH-2II --- ---T--- 27.1 -GGTT-ATTGCTTGTTAA--T --T---T ---AAT-CT-AG--TTTTT-T-T-TT-TAAAA-A-TAA-TAAATGGCA-ATGCAG---TTT-AGGATCTGTGTCAC---G pH-2 -5 T--TGA-CTA-GGCTGAT---TTGTTAATTTC-TGCTT-GAGGTTTT GTTTG-TTGTTTGATTTGTTTT-T-T-TT--AAGAAAT-A-TGA-AGATGAATA GA

57 6

pH-2 -1

pH-2II

460 480

GACAGCATACAGTGTACTTACATATAATAATAAATAAGTCTTTA40

27.1 CTGT-TG-GTCTGT-GGGAGAGCT-G-GGC--TGAG--CTG-G

Fig.3.Comparison ofDNAsequencesin the non-codingportions ofcDNAclonespH-2d_1, pH-211(Steinmetzetal., 1981b),and genomic clone 27.1 bp

5023-5504 (seeSteinmetzetal., 1981b) andpH-2d_5b.The non-homologous regionbetween the clones is underlined.

pH-2d-1

pH-2d- 5 100bp

(3)

1 2

-28S p ~-~-18S

Fig. 4. Size of mRNA coding for H-2 like antigens. 10AgmRNA from DBA/2 mousespleens was separated on a 1% agarose gel, transferred to nitrocellulose filter, and hybridized to probe la,(hne1) or probe 5b, (lane 2). Mouse18S and 28S rRNA was run in parallel tracks. Hybridization was in 50%formamide, 5 x SSC, 1 x Denhardt's (Gergen et al., 1980) at 42°Cfor 12h.Thefinalwash was in0.1SET at 50°C.

A

I

m

B

0

LL

-

CC

T-

col Nr CN

m

^co ^co ^Ct^G ^o

^d

FragmentSb is a low copy H-2 gene probe

specif

icfor H-2Kd region gene

The number of H-2 genes in a genome can be roughly estimated by the number of hybridizing bands obtained ina Southern blot analysis usingan H-2cDNAprobecomprising the coding region. In such an experiment >15 bands had been obtained, suggesting a large gene family (Steinmetzet al., 1981b; Camietal., 1981 andFigure 5A). Using fragment 5b as ahybridization probe, we expected to select onlyasub- group of the H-2 gene family. When BglII and EcoRI digested DNAfrom DBA/2mice wasprobed with fragment 5b in a Southern blot analysis only two major bands were detected (Figures 5B and C). Thus, probe 5b is a low copy gene probe and could be used to map the corresponding gene(s) within the MHCcomplex. Forthis purposecongenic mousestrains of the d, b, and k haplotype and recombinants in different parts of the MHC complex were used (Klein et al., 1978). Congenic mouse strainsareidentical exceptfora limited segment on one chromosome carrying a different allele. Ashybrid-selectedtranslation hadsuggestedto usthat 5b-specific sequences are contained in the H-2Kd molecule, we selected recombinant mouse strains in which therecom-

BgITL C ^Eco ^RI

T-

N N CN

_ D)

<

°: °:

Go

m- m

0Go

m

0co

M

d

o

m

cr-ED

d

co

23.7- 9.5- 6.8-

p,,_

4.3-

*0

|

1013

---| :

3.7+

'

3 7.2

iC

2.3- 2.0-

d m -*M

-23.7 -9.5 -6.8 -4.3

-2.3 -2.0

Fig. 5. Southern blot ofmouseDNAusing fragment la (A)or5b(B and C)ashybridizationprobes. 10 Ag of liver DNA from mice of different inbredH-2 haplotypes d(DBA/2; BMO.D2),k(BIO,BR), b (B1O) and H-2 recombinantsd/b(BIO,D2[R1031)k/d(BIO.A),k/b(BIO.A[4R])wascleavedtocompletion with BamHI(A), EcoRI (A and C)^orBglII (B). After separation^{on an}0.5%agarosegel, DNAwastransferredtonitrocellulosefilters.Hybridiz&tion^was with 40ng/ml of fragmentla(A)for20 hat68°CinI xDenhardt,4xSET andthe final wash in0.1xSET, 0.1I% SDSat68°C. In BandChybridization

waswith60 ng/ml of fragment 5bat60°C and also the final washat60°C. The sizes of the markerDNAfragmentsaregiveninkilobases.Thegenomic fragmentsmappedtotheKlocusareindicated byarrowsand their estimatedsizesgivenin kilobases.

M

23.7- 9.5- 6.8- 4.3-

2.3- 2.0-

(4)

Table I. Mappingof the 3.7-kbBglIIand 13-kbEcoRI fragment

Cell line Locus Hybridizingfragment

H-2K H-2D BgIII EcoRI

DBA/2 d d 3.7 kb 13 kb

BlO.D2 d d 3.7 kb 13 kb

BIO.D2(R103) d b 3.7 kb 13 kb

BIO.A k d 3.2 kb 10.5 kb

BIO.A(4R) k b 3.2 kb 10.5 kb

BIO.BR k k 3.2 kb 10.5 kb

B1O b b 2.5 kb 10.5 kb

binationaleventhad occurred between the K andD locus. Us- ing thesemousestrains and restrictionsite

polymorphism

asa marker, the approximate (relative) location ofgenes within the MHC can be mapped. A similar approach had been successfully appliedtothemappingofgeneslocated in theQa and Tla region (Steinmetz et al., 1981b; Margulies et al., 1982).

As most congenic recombinant ^mouse linesare found on the BlO background, we determined whether DBA/2 and B10.D2mice(bothd

haplotype)

show identicalbands inBglII andEcoRIdigestsofgenomicDNAprobedwithfragment5b inaSouthern blotanalysis. Figures5Band C show thatthey indeed have both a major 3.7-kb BglII and a 13-kb EcoRI band. In DNA from B10.BR mice

(k haplotype)

the 3.7-kb BgtII fragment was replaced by one of 3.2

kb,

and in BlO mice (b haplotype)byoneof 2.5 kb.

Using

the recombinant mice B10.D2 (R103), B10.A, and BlO (4R) the size ofthe BglII orEcoRIbandfollowed the allelic form presentattheK locus andnotthatattheDlocus(Figure 5). This is summariz- ed in TableI. Astherecombinational eventin BIOA(4R)oc- curredbetween theAandBloci of theI

region (Klein

etal., 1978), the 3.7-kb BglII fragment of the d

haplotype,

the 3.2-kbfragment inthekhaplotype, and the2.5-kbfragment in the b haplotype were mapped to theK locus. The 13-kb EcoRI fragmentisreplaced inDNA from k and bhaplotype micebya 10.5-kb fragment.As recombination in B10.Aoc- curred between the E and C loci of the I region, the 13-kb fragmentwasalso mapped totheKlocus.

When fragment 5b was used to identify genes from a genomic phage library constructed from DBA/2 mouse DNA,atleast threepositiveclonesweredetected perhaploid genome. Allthree clones alsohybridizedtoprobe lb. One of these clones containeda3.7-kb BglII anda13-kbEcoRIfrag- menthybridizing to probe 5b (Kvist et al., in preparation).

Discussion

Using hybrid selected translation and Southern blot analysis, we identified a segment on a cDNAclone specific forH-2gene(s) at theKlocus. This segment is located in the 3' non-coding region and comprises the 120 nucleotide residues upstream from the stretch of poly(A) in clone

pH-2d-5.

Itcontains onepolyadenylation site. An identical se- quenceis present in clone

pH-2d-4

although this clone extends atits 3' endbyastretchof 20nucleotides containing a second polyadenylation site (Lalanne et al., 1982). It is thus likely thatcDNAclones

pH-2d-4

and-5 are coded by the same gene but usedifferent polyadenylation signals in their processing.

In aSouthern blot analysis with

BglII

DNA from mice of the d haplotype, probe 5b hybridizes primarily to a 3.7-kb fragment, with khaplotype DNA to a 3.2-kb fragment, and

with that of the b haplotype to a 2.5-kb fragment. Using recombinant mouse strains thesefragments weremapped to the K locus. It will thus bepossibletoisolate andcompareK locusgenes from these threehaplotypes.

With EcoRI-digested DNA of the d haplotype a 13-kb

fragment

wasdetected and with that of the k and bhaplotype

a10.5-kb fragment.As the3.7-kbBgllI and the13-kbEcoRI

fragments

canbe found inonegenomic clone, they

probably

characterize thesame

gene(s).

Theselectivityofprobe5b could beincreasedbyapplying

more stringent conditions [0.1 x SET(SET: 150 mM NaCl, 1 mMEDTA, 10 mM Tris-HCl, pH 7.8, 0.1Wo SDS), 65°C]

in the Southern blot analysis. Under these conditions ap- pearance of the minor bands was reduced, whereas no decreasewas found in the3.7-kb

BglII

and the 13-kbEcoRI fragment (Xin, unpublished data). This would indicate that onlythese two fragmentscontainsequencesidentical tothat contained in fragment Sb. Under stringent conditionsprobe Sbis,therefore, K locus-specific. The number ofgenes atthe Klocus and whether all contain thissegment in the 3' non-

coding region have stillto beestablished.

The selectivity of fragment Sb and its property as a low copyprobeis incontrast totheequivalentsegmentinclones pH-2d-1 and pH-2II. Here, the portion closeto thepoly(A) tail ishighly repetitivein themousegenome(Steinmetzet

al.,

198la). Non-homologous regions in the 3' non-codingregion have alsobeen described by Cosmanetal. (1982) forSWR/J mice(q haplotype). As theirobservationsareonly based on cross-hybridization, and sequence analysis is not yet available, it remains unclear whether fragment 5b is homo- logoustooneof the D fragments described by Cosmanetal.

(1982).

Takentogether, the data presented here suggest that the 3' non-coding region in mRNA coding for H-2 antigens hasa bipartite composition. The first portion comprises the stretch of - 300nucleotides following thestop codonTGA and the secondthe 120- 160 nucleotidesupstreamfrom thestretchof poly(A). Extensivehomologyof ^- 80% isgenerally found in the first withnosignificant homology being found in these- cond portion[cf.pH-2d-1,pH-2d-5, and gene 27.1 (Steinmetz etal., 1981a) in Figure 3]. Thedivergence in the secondpor- tion mightindicate that this region has been joined to anH-2 geneina recentrecombinational event, probably duringthe event of gene duplication. The functional relevance of this portion remains unclear. The finding, however, that genes fromdifferent loci bearnon-homologous portions in their3' non-coding region make these segments excellent tools in defininggenes of differentregions orloci.

Materials and methods

Restrictionenzymes, Kienow DNA polymerase, andpolynucleotide kinase werepurchased from Biolabs(Bishop's Stortford, UK) or Boehringer (Man- nheim,FRG).[a-32P]dNTPs (10 mCi/ml, 3000 Ci/mmol) and[-y-32P]ATP(10 mCi/ml, 5000 Ci/mmol) were from the Radiochemical Centre (Amersham, UK). Nitrocellulosefilters were obtained fromSchleicher and Schuell(Dassel,

FRG), Oligo(dT)-cellulose (T3) from Collaborative Research (Waltham,

MA). Rabbit reticulocyte lysate and[35S]methioninewerefrom New England Nuclear(Dreieich, FRG). Mice were obtained from J. Klein, Max Planck In- stitute(Tubingen,FRG), or from our own colony.Plasmid pUC8 was obtain- edfrom J. Messing,University of Minnesota.

Selectionandisolationof cDNA clonespH-2d-1to -5 have been described earlier(Kvistetal.,1981;Bregegere et al., 1981; Lalanne et al., 1982). Hybrid selectedtranslation was performed according to Ricciardi et al. (1979) and selectedmRNAtranslated in areticulocytelysateinthepresence of dog pan- creasmicrosomes(Kvist et al., 1981).

470

(5)

mRNAisolation andNorthern blotanalysis

RNAwasextracted fromDBA/2(H-2d) mouse spleens by the guanidinium chloride method (Cox, 1968) as described by Deeley et al. (1977). The chloroform-butanol extraction was replaced by a phenol-chloroform-isoamyl- alcohol extraction. Poly(A)+ mRNA was isolated on oligo(dT)-cellulose (Aviv andLeder, 1972)and separatedon a1% agarose gel containing 2.2 M formaldehyde (Rave et al., 1979) and blotted onto nitrocellulose filter (Thomas, 1980).

DNAsequenceanalysis

cDNAclonepH-24-1 wassubcloned by the method of Frischauf et al.

(1980) into pBR322 and sequenced by the method of Maxam and Gilbert (1980). cDNAclone pH-2d-5wasdigestedwith PstI, and fragments were clon- ed intopUC8. Inserts were characterized by hybridization to cDNA clones pH-2d-1and -5(Thomas, 1980). One subclone, containing a 250-bp insert, which did not hybridize topH-2d4-but topH-2d-5,was selected and named pH-2d'-5b. The plasmid was cut at the BamHI site, labelled, recut with a EcoRI,andsequenced(Maxamand Gilbert, 1980).

Isolationofgenomic DNAand blot hybridization

DNA wasisolated fromasinglemouseliver essentiallyasdescribedbyBlin andStafford (1976). Powdered tissuewasmixedwith 15 mlof the Sarkosyl- proteinase K solution and DNA extracted by gentle agitation. DNA was dialysed against waterand concentratedbyAquacide. RestrictedDNA was run on a0.5%agarose gel andtransferredto anitrocellulose filteraccording toSouthern(1975).DNAprobeswerelabelled by nick-translationto aspecific activity of 1-5 x 108 c.p.m./4g (Rigby et al., 1977). Hybridization and washings of the filters were essentially as described by Gergen et al. (1980) and asspecified in the figure legends.

Ploegh,H.L., Orr,H.T., and Strominger,J.L. (1980) Proc. Natl. Acad. Sci.

USA, 77, 6081-6085.

Ploegh,H.L., Orr,H.T., and Strominger,J.L. (1981) Cell, 24, 287-299.

Rave,N., Crkvenjakov,R., and Boedtker,H. (1979) Nucleic Acids Res., 6, 3559-3567.

Ricciardi,R.P., Miller,J.S., and Roberts,B.W.(1979) Proc.Natl.Acad. Sci.

USA, 76, 4927-4931.

Rigby,P.W.J., Dieckmann,M., Rhodes,C., and Berg,P. (1977) J. Mol. Biol., 113, 237-251.

Robinson,P.J.(1982) Immunogenetics, in press.

Soloski,M.J., Uhr,J.W., Flaherty,L., and Vitetta,E.S. (1981)J.Exp.Med., 153, 1080-1093.

Sood,A.K., Pereira,D., and Weissman,S.M. (1981)Proc. Natl.Acad. Sci.

USA, 78, 616-620.

Southern,E.M. (1975)J. Mol.Biol., 98,503-517.

Steinmetz,M., Frelinger,J.G., Fisher,D., Hunkapiller,T., Pereira,D., Weissman,S.M.,Uehara,H., Nathenson,S., andHood,L.(1981a)Cell,24,

125-134.

Steinmetz,M., Moore,K.W., Frelinger,J.G., Sher,B.T., Shen,F.W., Boyse, E.A.,and Hood,L. (1981b)Cell, 25,683-692.

Thomas,P.S. (1980)Proc.Natl.Acad. Sci. USA, 77,5201-5205.

Acknowledgements

We areespecially gratefulto J.Klein andP.Singer for adviceonselecting the appropriatemousestrains and forproviding them.WethankH.Lehrach for adviceon using pUC8 as acloning vehicle, Ms. W. Moses fortyping the manuscript, andD.Meyerforcritically reading it. This workwassupported byagrantfrom the DeutscheForschungsgesellschaftDo199/4-1.

References

Aviv,H., and Leder,P. (1972)Proc.Natl.Acad.Sci. USA, 69, 1408-1412.

Blin,H., and Stafford,D.(1976)Nucleic AcidsRes.,3, 2303-2308.

Bregegere,F., Abastado,J.P., Kvist,S., Rask,L., Lalanne,J.L., Garoff,H., Cami,B., Wiman,K., Larhammar,D., Peterson,P.A., Gachelin,G., Kourilsky,P., and Dobberstein,B. (1981) Nature, 292,78-81.

Cami,B., Bregegere,F., Abastado,J.P., and Kourilsky,P. (1981) Nature, 291, 673-675.

Coligan,J.E., Kindt,T.J., Uehara,H., Martinko,J., and Nathenson,S.G.

(1981)Nature, 291, 35-39.

Cosman,D., Khoury,G., and Jay,G.(1982) Nature, 295, 73-76.

Cox,R.A. (1968) Methods Enzymol., 12, 120-129.

Deeley,R.G., Gordon,J.I., Burns,A.T.H., Mullinix,K.P., Binastein,M., and Goldberger,R.F. (1977)J.Biol. Chem., 252, 8310-8319.

Dobberstein,B., Garoff,H., Warren,G., and Robinson,P.J.(1979) Cell, 17, 759-769.

Flaherty,L. (1980) in Dorf,M.E. (ed.), The Role of theMajor Histocompati- bility ComplexinImmunology, GarlandSTEMPress, NY,pp. 33-57.

Frischauf,A.-M., Garoff,H., andLehrach,H. (1980)Nucleic Acids Res., 8, 5541-5549.

Gergen,J.P., Stern,R.H., andWensink,P.C. (1980)Nucleic AcidsRes., 7, 2115-2136.

Klein,J. (1975) Biology of theMouseHistocompatibility Complex,published bySpringer Verlag, Berlin.

Klein,J., Flaherty,L.,VandeBerg,J.L., andShreffler,D.C. (1978)Immuno- genetics, 6, 489-512.

Klein,J. (1979)Science(Wash.), 203, 516-521.

Kvist,S., Bregegere,F.,Rask,L., Cami,B.,Garoff,H., Daniel,F.,Wiman,K., Larhammar,D., Abastado,J.P., Gachelein,G., Peterson,P.A., Dobber- stein,B., andKourilsky,P. (1981) Proc.Natl.Acad. Sci. USA, 78, 2772- 2776.

Lalanne,J.L., Bregegere,F., Delarbre,C.,Abastado,J.P., Gachelin,G., and Kourilsky,P. (1982) NucleicAcidsRes., 10,1039-1049.

Margulies,D.H.,Evans,G.A.,Flaherty,L.,andSeidman,J.G. (1982)Nature, 295, 168-170.

Maxam,A.M., andGilbert,W. (1980)MethodsEnzymol., 65,499-560.

Michaelson,J., Flaherty,L., Vitetta,E., and Poulik,M.D. (1977) J. E:xp.

Med., 145, 1066-1070.

Nathenson,S.G., andCullen,S.E. (1974)Biochim.Biophys.Acta,344, 1-25.

471

Identification of