Conserved major histocompatibility complex class II boxes--X and Y--are transcriptional control elements and specifically bind nuclear proteins.

HAL Id: hal-02561163

https://hal.archives-ouvertes.fr/hal-02561163

Submitted on 24 Nov 2020

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Conserved major histocompatibility complex class II boxes–X and Y–are transcriptional control elements and

specifically bind nuclear proteins.

A. Dorn, Beatrice Durand, C. Marfing, M. Le Meur, C. Benoist, D. Mathis

To cite this version:

A. Dorn, Beatrice Durand, C. Marfing, M. Le Meur, C. Benoist, et al.. Conserved major histocom- patibility complex class II boxes–X and Y–are transcriptional control elements and specifically bind nuclear proteins.. Proceedings of the National Academy of Sciences of the United States of America , National Academy of Sciences, 1987, 84 (17), pp.6249-6253. �10.1073/pnas.84.17.6249�. �hal-02561163�

Immunology

Conserved major histocompatibility complex class II boxes-X and Y-are transcriptional control elements and specifically bind

nuclear proteins

(immune responsegenes/transgenic mice/DNA-binding proteins/generegulation)

ARNULF DORN, BEATRICE DURAND, CORINNE MARFING,

MARIANNE LE MEUR,

CHRISTOPHE BENOIST,

AND

DIANE MATHIS

Laboratoire deGdndtique Moldculaire des Eucaryotes, Centre National de la Recherche Scientifique, Unite 184 de Biologie Moldculaire et de Genie Gdndtique, Institut National de la Santeetde la Recherche Mddicale, Institut de Chimie Biologique,Facultd deMWdecine, 11, rue Humann, 67085 Strasbourg Cedex, France

CommunicatedbyPierreChambon, May 20, 1987

ABSTRACT A conserved sequence motif exists at the 5' end of allmajor histocompatibility complex class II genes. This motif consists of the 14-base X and Y boxes separated by a short stretch of variable sequence. In this report, we provide evidence that the X and Y boxes play an important role in controlling transcription of the murine class IIgene

El.

We have developed transgenic mouse lines that carry Ea genes cleanly deleted for either the X or Y box and have compared the expression of these mutant transgenes with that of a nondeleted control. Both the X and the Y segments appear critical for accurate and efficient transcription of

Ek.

The most drastic effect is seen with y-interferon-treated macrophages, where deletion of the Y box completely abrogates transcription initiated by the normal promoter. In addition, we identify proteins from nuclear extracts that bind specifically to the X or Ybox.

The murineIaantigensareencoded byafamily of

genes-A, A/3, Eao

and

Etr-that

^{resides in}themajor histocompatibility complex (MHC) (1). Thesegenes,knownasclassIIloci,are expressed coordinately and are subject to a medley of regulatory influences (2). Accordingly, mouse class II mol- ecules are limited primarily to immunocompetent cells- including B cells, antigen-presenting cells, and thymic epi- thelial and dendritic cells. Withinagiven lineage, expression varies with thestate of differentiation; in theB lineage, for example, Ia antigensare notpresent onpre-Bcells,occur at high levelsonBcells, andareagain absent from plasma cells.

Finally, classIIMHCgeneexpressioncanbe modulated by avariety of effectors: yinterferon (IFN-y) isanup-regulator for macrophages, as is interleukin 4 for B cells, whereas glucocorticoids anda-fetoproteincanbe potentdown-regu- lators.

It is important to understand the basis of this elaborate control because aberrant

expression

of class II MHC mole- cules may engender or

amplify

certain ofthe autoimmune diseases (3). Some hints as to the location of crucial gene promoterelementsderive fromDNA sequencecomparisons.

Most suggestive has been the identification of a highly conserved sequence motif in the

5'-flanking region

ofall MHCclass II genes, whether human ormurine (4-8).This motifconsists oftwo14-base conservedsegments-theXand Y boxes-spaced by a 17- or 18-base stretch of variable sequence. Although there has been much speculation about how the conservedmotifmight influence MHC classIIgene

expression,

asyet,no

experimental

data existtosupport any theory.

Here,weprovideevidence that the X and Y boxes play an importantrole in regulating transcription of a murine class II MHC gene,

Egka.

Mutant genes bearing a clean deletion of either theXor Yboxcanbe transcribed in transgenicmice, but neitherefficientlynoraccurately. Theeffect ofthe Y-box deletion on transcription in IFN-y-induced macrophages is particularly drastic.Inaddition,we have beenabletodetect specificbinding ofnuclearproteinto the XandYsegments.

MATERIALS AND METHODS

Mice.

C57BL/6J

and

SJL/J

mice were

purchased

from Iffa-credo(LesOncins, France)and Ollac(Bicester, U.K.), respectively. Transgenic micewereproduced byinjectionof

E4,'

DNA into (C57BL/6 x SJL)F2 embryo pronuclei (9).

Lines carrying

Ek,

were maintained as heterozygotes by back-crossing positive animals toC57BL/6 mice.

Injected DNA. The injected fragments all derivefrom the

A/J

strainEa gene carried inthe cloneXA/J p34.13 (4). The 9-kilobase XbaIfragment (position-2170 toposition +6800) was subcloned into a pBR322-derived plasmid, and slight modificationswereintroduced toyieldthe plasmid pWE32.

(i) An Xho Ilinkerwasinserted into theAcc Isiteatposition -215. (ii) The Pvu Isite at position +12 was replaced bya BamHI site. The sequence of plasmid pWE32 reads as follows: ... AGTCTGCAGGATCCGCTTC ... instead of ... AGTCTGCGATCGCTTC ... (iii) The BamHI site at position +1581 was eliminated by partial digestion, repair, andligation.

To construct the AX and AY mutations, a Kpn I-Sac I fragment from pWE32 (positions -1320 to +101) was sub- cloned intoM13. X and Y boxeswere deleted(individually) byoligonucleotide-directed mutagenesis(10). Mutant phage DNAs weresequenced, and the AX21andAY301 mutations were grafted onto pWE32 by substitution of the Xho I- BamHIfragment (positions -215 to +12) to yield the plas- mids pWEAX21and pWEAY301.

Sucrose-gradient-purified

Bgl I fragments (8.2

kb)

were injected intomouseembryos.

Si

Analysis. Thepreparationof RNA and itsanalysis by S1 mappinghave beendescribed (11).In most

experiments,

the probewasasingle-stranded Sac I-Xho I

fragment (positions

+101 to-215) derived from

pWEAX21,

5'-end labeledatthe Sac I extremity.

Gel Retardation Assays. Various

Eck

promoter

fragments,

usually the residue -215 to residue +12 Xho I-BamHI segment, were

purified by gel electrophoresis

followed

by

electroelution.The39-base

double-stranded oligonucleotides

Abbreviations: MHC, major histocompatibility complex; IFN-y, y

interferon.

6249 Thepublicationcostsof this articleweredefrayedin partbypagecharge payment.This article mustthereforebeherebymarked "advertisement"

in accordance with 18U.S.C. §1734 solelytoindicate this fact.

Proc. Natl. Acad. Sci. USA 84 (1987) were prepared as follows: complementary strands were

synthesized on a Gene Assembler (Pharmacia) using the phosphoramidite method (12); the two strands were 5' end- labeled using [_y-32P]ATP, combined, denatured for 2 min at 80TC, andannealed 30min at 20'C;finally, double-stranded oligonucleotideswerepurifiedon a15% polyacrylamidegel.

The "Y" oligonucleotide spans positions -74 to -36; i.e.,it consists ofthe Yboxplus13 bases on the 5' side and 12 bases on the 3' side. "Y-Rd" includes thesame 5'-and3'-flanking sequences, but the Y box (TTCTGATTGGTTAA) has been replaced by random sequence(ACGTTCTGGCTACA),cho- sen by flipping two coins. The "X" oligonucleotide spans positions-105 to-68;i.e., it includesthe Xboxand 12 bases on the 5' side and on the3' side. "X-Rd"has the same 5'-and 3'-flanking sequences, but the Xbox (CCTAGCAACAGA- TG) has been replaced by random sequence (CATCTGT- TGCTAGG).

Nuclear extracts werepreparedaccordingtoDignamet al.

(13), except we routinely used 1 M NaCl to extract nuclear protein.Gelretardationassays wereconductedaspublished

(14).

RESULTS

Whenthe immediate 5'-flanking regions of human andmouse

class II genes are aligned, a conserved sequence motif becomes apparent (4-8). Theconsensus isasfollows:

CCTAGCAAG

GATG-*-TCTGATTGG

TTAA

Box X Box Y

The distance(denoted by the *) between theXand Yboxes (17or18bases) is also strongly conserved. The placement of X and Y at the 5' end of

Ek,

^a murine class II gene, is diagrammed in Fig. 1 A andB; the actual sequence of this region is presented in Fig. 1C.

X and Y Boxes Are Required for Efficient and Accurate Transcription of

El.

To determine whether the X and Yboxes have any regulatory role, we have chosen to study their influence on

Ek

expression in transgenic mice. In previous experiments, we injected this gene into B6 x SJL mouse

embryos and developed transgenic lines that harbor the injected DNA stably integrated into thegenome (11). (B6 x

SJL mice do not transcribe their endogenous E1a genes

because ofadeletion thatremovesthepromoter region and first exon) (15, 16). The

Ek

transgene was expressed effi- ciently, accurately, and with tissue and celltypespecificity.

Fortheexperimentsdescribedhere,wecreatedmutantEka

genesthatcarryaclean deletion of theXortheYbox. Asa

firststep,aslightly modified

Eka,

calledWE32,wasdesigned;

thisconstruct differs from the wild-type geneat only three positions,butgreatly facilitates the construction ofmutants.

Using WE32 as a base, we created the mutants WEAX21, which bearsacleandeletion of the X box (fromposition -93 toposition -80), and WEAY301, harboringaprecise deletion of the Y box (from position -61 toposition -48).

The three modified

Ek

genes-WE32, WEAX21, and WEAY301-were individually injected into (B6 x SJL)F2 embryos, alwaysas an8.2-kilobase Bgl I fragment. Multiple transgenic lines were developed for each construct. The number of integrated copies of the injected gene varied

among the different lines, fluctuating from 1 or 2copies to

>50.

Expression oftheEkTransgene in the Spleen.Asafirst test

oftransgenefunction, we measured Ea RNAexpression in thespleen by S1 mapping (Fig. 2). This allowedanestimate ofEaRNA levels^aswellas anassessmentof theaccuracyof initiation. To facilitate quantitation inthis particularexper-

5 Transcription 3,

A. Bgl

RV

XhoBam

RV

Bgl

I 9 _ I I

-kb i' L "- __ 1 2 TC 3UT

/ _

B. ,/ 5' end mRNA

". X YTATA F * WE-32

_~fITI

20bp Bam

WEAX-21 WE AY-301 C.

5'... ACAGCCTTTATTATTTTTTTGTTAATAAGTGGMAAATTTCTTCTTGAGGAAA-160

ATTATTTCTTGAAATGTTAAGTGGAAACTCGGATACTAATAGGACCTG6TTGCMGGAA-100

X Y

x...-.40

CCCTTT3CTAGCAACAGATGJGTCAGTCTGAAACATT TCTGTG6TTAGTTGAG TGCTTTGGATTTAATCCCTTTTAGTTCTiGTTAATTCTGCCTCAGTCTGCGA+1 ..3'

FIG. 1. The Ea gene and mutants thereof. (A) Diagram of the injected BgI I fragment. Black boxes represent the Ek exons as follows: L, leader peptide; 1 and 2, extracellular domains; TC, transmembranesegment andcytoplasmic tail; 3UT, 3'-untranslated region. The open circle denotes a mutated BamHl site; the under- lined XhoIandBamHIsiteswereintroducedartificially. Bgl,Bgl l;

RV, EcoRV. (B) The5' end. The mRNAcoding region appearsas a bold line.(C)Sequence of the Ek gene 5' end. ThemajormRNA start site isatposition +1.

iment, we used a single probe derived from WEAX21 that extendsfromposition +101 toposition -215.

The BxA lane illustrates the pattern of protected frag- ments for a nontransgenic, positive control: the bands at

positions

+1 and +3 reflect the normal Ek RNA initiation

sites;

the lower bands (arrow) also signify that RNA is initiatedatpositions+1and+3,butthefragmentsprotected by these transcripts have been digested internally by S1 at position +12,where thewild-type

Ek

sequence differs from theWEAX21probe bya4-base-pair insertion. TheB xSlane confirmsthat nontransgenic B6 x SJL mice donotexpress

Ek

RNA.

Twotransgenic lines bearingWE32 werestudiedtoverify that the seemingly minor alterations introduced tofacilitate the construction of

Ek

mutants were indeed innocuous. As Fig.2attests, theslightly modified

Eg

transgeneis

expressed

efficientlyandaccurately inthespleen. Inthis

instance,

we do not observe the bands at position +12 (cf. BxA lane) becauseWE32 and the WEAX21 probe are notmismatched at thisposition.

EaRNAisexpressed in the spleensof all micetransmitting WEAX21 with themutantgenebearing a clean deletion of the X box. However, most ofthis RNA is abnormal. A small percentageofthetranscriptsareinitiatedatpositions+1and +3, the normal sites, but by far the majority start at alternative sitesscatteredthroughoutthe 5' end. The scatter is even moreextensive than isapparent in Fig. 2 sincethe band labeled Up marks the end ofthe probe. With other probes,it ispossibletodetect transcriptsthatinitiateasfar upstream as the -1.0- to -1.5-kbregion, and thisfinding is consistentwith resultsofRNAgelblotting experiments (data notshown).There appears tobearoughcorrelationbetween levelsofmRNAand the number ofintegrated genecopies.

Mice transmitting

AY301,

the mutant

Ek

that bears a

precise

deletion ofthe Ybox, exhibitoneoftwophenotypes.

Some-lines22, 45,and 59-havelittleor noEaRNA in the spleen; others-lines54and 58-havenormalorgreater than normal levels of Ea RNA, but again initiation of these 6250

Immunology:

Dornet al.

Proc. Natl.Acad. Sci. USA 84 (1987) 6251

A.

DNA: WEZY301 WEAX21 WE32

Line: 12245 54 58591 5 16 23241 r2529 BxSBxA

Copy# 5 15 60 20 15 10 20 6 12 4 4 0 2

-220- _

I

-60-

-48- up-.I

-25- -19-

--220

40 4

_w ^_

_a

_ _

_D _

U0 U0

m

--60 --48

--25

--19

_0 --1

_ _

ft B.

BxS |I BxA ', AX21.16 I AY301.54

I1 ~ I'w

FIG. 2. Transgene expression in the spleen. (A) S1 nuclease mapping ofEatranscripts in spleen RNA. The probe was a single- stranded Sac l-Xho I fragment (residues +101 to -215) from pWEAX21. Bandslabeled Up denote forWEAX21all initiation on the 5'side ofresidue -215 (the end of the probe) and forWEAY301 alltranscriptsinitiated on the 5' side of residue -48 (the point of the AY deletion). BxS, (C57BL/6 x SJL) F2; BxA, (C57BL/10 x

A/J)F1.Copy numberwasestimatedby Southernblotting. (B) Flow cytometry of splenic B cells, achieved by gatingon surface IgM+

cells afterdouble-staining withafluorescein-isothiocyanate-labeled anti-mouse

1gM

antibodyandaphycoerythrin-labeled anti-E reagent (14-4-4). The dotted line in each panel shows cells stained with an irrelevant antibody. WEAX21-16 and WEAX301-54 are second generation.

transcripts is aberrant. The spread of transcripts is greater than whatisapparentinFig.2since the Up band reflects the

position

(residue -48) at which WEAY301 and the probe from WEAX21 are mismatched. In fact, S1 mapping with other probes and RNA gel blotting provides evidence for transcripts

initiating

as far on the 5' side as the -1.0- to -1.5-kbregion (datanotshown). Note that thetwolinesthat expresssignificant levels of

EB

RNAcarry manycopies(>20

copies)

ofthe transgene.

It was of interest to determine whether the aberrantly initiated transcripts from WEAX21 and WEAY301 spleens can be translated to make

EB

surface protein. Thus, we isolated splenocytes from various animals, and quantitated E-complex levels on B cells by two-dimensional cytofluo-

rimetry,

usingaphycoerythrin-labeled anti-E antibody anda fluorescein

isothiocyanate-labeled anti-IgM

antibody. Fig.

2B shows the profiles for control, nontransgenic animals

(BxS

and BxA). As anticipated, none of the BxS B cells expresstheEcomplex, whereas almostallof theB x AIgM+

cells are also

EB.

In other experiments, WE32 transgenics exhibit essentially the same staining pattern as the BxA controls

(data

notshown). Staining profiles forB cellsfrom

WEAX21.16 spleens

show that almost all are

EB, although

thereseemstobesometrailing of cells into the dull-staining

region.

In contrast, only about half the B cells from the

WEAY301.54

linearesurface

EB,

althoughmostof the cells that do express the E complex at the cell surface display normalamountscompared with the BxA controls. Qualita-

tively similar profiles have been observed with other WEAY301 and WEAX21 lines.

IFN-y

Inducibiity of the

Ek

Transgene. To determine whether themutated

Ek,

transgenes wereinduciblebyIFN-y, we isolated thioglycollate-primed peritoneal macrophages from various animals, cultured them in the presence or absence ofIFN-y, isolated RNA, and assayed

E. transcripts

byS1 mapping (Fig. 3A). TheBxApanel showsthe

degree

ofstimulation fora nontransgenic control: IFN-yincreases theamountofEatranscriptin RNAfrom peritonealmacro- phages 10- to 20-fold, andthis level reaches30-60%of that observed in spleenRNAfrom the same animal. Ea RNA is induced to a similar degree in the WE32.25 transgenic line (data not shown). In WEAX21animals, IFN-y increases the amount of Ea transcript, whether initiated at promoter- proximal orpromoter-distal sites,andtheamountinmacro- phage RNA is again about halfthat seen for spleen.

Note,

however, that thereare qualitative differences between the spleen and macrophage patterns: for WEAX21 a strong band at position -55 with spleen RNA is virtually absent with macrophage RNA.

IFN-y induction of Ea RNA synthesisin WEAY301

peri-

toneal macrophages is more complex. In line 58 mice, there is nodetectableinduction ofEa mRNA.Note(asabove)that normal induction would result in Ea RNA levels reaching 30-60% those observed in the spleen. We know that the interferon treatment was effective because the endogenous Aa gene wasinduced to normal levels (data not shown). In line 54 mice, there is no induction of the EB transcripts initiated atpromoter-proximal sites, but there is an increase in thenumber oftranscripts initiated at promoter-distalsites, although the levels remain well below spleen levels. (Here again, recall that the Up band at position -48 reflects initiationatall sites upstream from that point; this position

A. BxA AY30154

M -: § S~p M - | Sp

II

U-.-- Uw

+*-4

-0

a

AY30158 AX21t6

'S-p^I -

~~~~~

l^Upp-wSp^. - +^00-~amM,^*+1I8

4+1

FIG. 3. Transgeneexpression in macrophages. (A) S1 nuclease mappingof Ea transcripts in RNA from spleen (sp)or peritoneal macrophages cultured with IFN-y (+)oralone(-). Theright-most andleft-most panels show results from independent experiments.

Fortheformer, theprobe was a residues +72 to -215 fragment, 5'-end-labeledat residue +72, and, for thelatter, itwas aresidues +101to-215fragment,5'-end-labeledatresidue+101.Bothprobes derive from pWEA21. Transgenic mice are secondgeneration off- spring.(B) Flow cytometry ofperitonealmacrophagesculturedwith (+)orwithout(-)IFN-yand then stained withananti-Eantibody, 14-4-4. Dotted lines show staining profiles for IFN-y-stimulated macrophages stained withananti-Aantibody, 3G2.

Immunology:

Dometal.

-'O

-z- -19' _or, _

Proc. Natl. Acad. Sci. USA 84(1987) marks the point of mismatch between WEAY301 and the

WEAX21probe.)Whether or not thisdistinctionbetween the twoWEAY301lines is a true difference is not certainsincefor line58mice thereare sofewtranscripts initiatedatpromoter- distal sites. In any case, we have seen no induction of the promoter-proximaltranscripts around positions-25and-19 for either line inmultiple experiments.

Thus, theY boxseemsparticularly critical forexpression of

E,

transcriptsinIFN-y-inducedmacrophages.Thisfinding has been confirmed by cytofluorimetric quantitation of E complexonthesurface of peritoneal macrophages. As seen inFig.3B,there isnomeasurable

EaEs

proteinonstimulated WEAY301 macrophages,thoughaclear induction isdiscern- ibleoncellsfromBxAmice(andfrom WE32animals,data not shown). WEAX21 mice do have E complex on IFN-y- induced macrophages, but the level is somewhat reduced, probably because manyof theRNAs initiate farupstream.

Nuclear Proteins Bind Specifically to the X and Y Boxes.

Specific binding of nuclearproteinstothe5'-flanking region ofEa hasbeen studied with the gel retardation assay. This technique (17) depends on the relative

binding

affinity ofa given protein for a 32P-labeled DNA fragment versus unlabelednonspecific competitor DNA. If the protein binds specifically to the labeled DNA, migration of the resulting DNA-protein complex inaneutralpolyacrylamide gel will be retarded with respect to migration of the naked DNAfrag- ment.

NF-Y, a Y box-binding protein. Fig. 4A shows a gel retardation assay using the

32P-labeled

Xho I-BamHI frag- ment from the 5'-flanking region of

Ega

(positions -215 to +12, seeFig. 1), various quantities ofpoly[d(I-C)] compet- itor, anda 1 M NaCl nuclearextractfromthe B lymphoma M12. Astheamountof unlabeled

competitor

DNA

increases,

adistinct band becomesapparent(arrow). This band reflects thebinding ofaprotein (hereafter referredtoasNF-Y)tothe Y box. The evidence for this conclusion is 3-fold. (i) Gel retardation assays using a panel of mutant Ea promoter fragments showaperfect correlation between thepresenceof theY boxon agiven fragment and its

ability

tobindNF-Y (Fig. 4B). In particular, a clean deletion of this 14-base segmentcompletely abrogates

binding

(compareWTand 301 panels inFig. 4A). Ontheother

hand,

removalofthe TATA box

(A26),

Xbox

(A21),

orGTGGAAA motifs

(A6, BR)

does not reduce binding of NF-Y. (ii) DNase I

footprinting,

methylation interference

mapping,

and

point-mutational

analyses have delineatedcontactsites forNF-Ywithin theY box (27). (iii) NF-Y can be bound by a 39-mer (the Y

oligonucleotide)

centered on the Y box. The

prominent

retarded band in

Fig.

5doesreflect

binding

of whatwehave termedNF-Ybecause this band isnotobserved inassays that employ the Y-Rd

oligonucleotide,

which contains thesame

flanking

nucleotides buthasrandomsequence

replacing

the Y box. Furthermore unlabeled Y

oligonucleotide,

but not unlabeled Y-Rdoligonucleotide,competesforthebinding of NF-Y to the Xho I-BamHI

Eka

promoter fragment and unlabeled wild-type

Eka

promoterfragment, butnot the same fragment fromA301,competes for specificproteinbinding to the Y

oligonucleotide

(data not shown).

X

box-binding

protein. Gel retardation experiments with the variousmutantfragments depictedin Fig.4B(in partic- ularwiththeA21fragment)provided noevidencefor proteins that bind specificallyto the Xbox. But the better visualiza- tion of NF-Y afforded by using a short 39-base fragment prompted us to search for X-binding proteins using an

oligonucleotide.

Thus, we made the38-base X

oligonucleo- tide,

composed of theXbox and12basesoneachside,and the X-Rd

oligonucleotide,

which has the same 5'- and

3'-flanking

bases but random sequence replacingthe Xbox.

Asseenin

Fig.

5,wedetect threefaint bands(arrowed)that are

specific

tothe Xbox. All the other bandsarepresent in

A. WT A301

1 2 3 4 5 6 7 1 2 3 4 5 6 7

B.

Xho X Y

TATA Barm

WT

A 21 I

A301 I

A 26 A101

A103 I - t-4

A 6 'E -

'BR'

Binding

+

4.

+ FIG. 4. Gel retardation assaysonwild-typeandmutantpromoter fragments.(A)A32P-labeledfragment(residues -215to+12)from wild-typeEt(WT)orfromaY-box deletionmutant(A301,seebelow)

weremixed withincreasingamountsofpoly[d(I-C)]and 1-2 ,ugof nuclearextractfrom theBlymphomaM12.Thepositionofastable DNA-protein complexisindicated withanarrow; thestrong bottom band representsfree DNA. Amounts ofpoly[d(I-C)]for lanes1-7 were 0, 0.3, 0.6, 0.9, 1.2, 1.5, and 2 ug. (B) Effectof promoter deletions.Symbolsaredefined in thelegendtoFig.1.Deletionswere madeby BAL-31orrestrictionenzymedigestionorbyoligonucle- otide-directedorspontaneousmutagenesis.Theirpositionsare:

A21,

residues -93to-80;A301,residues -61to-48;

A26,

residues -40 to +1;

MO,

residues -93 to -48;

A103,

residues -43 to -80and residues -61to-48;A6,residues -186to-142; E,residues -129 to+12; BR,residues +215to -113.

the X-Rd control and thus reflect

nonspecific binding

or attachmentof

protein

tothe

flanking

sequences.

Also,

most

of theseextraneousbands

disappear

at

quite

low

poly[d(I-C)]

concentrations.

Unfortunately,

the bands that correlate with

X X-Rd Y Y-Rd

12345

12345I

¹ 2 3 4 5111 2 3 4 5

FIG. 5. Specific bindingtooligonucleotides.The labeled DNAs

were the double-strandedoligonucleotides Y, Y-Rd, X, and X-Rd.

Poly[d(I-C)]

concentrationwas0,50, 100, 200,or300 ng perlane for Yand Y-Rdin lanes1-5,respectively;andwas50, 100, 300, 600,and 800 ng per lane for X and X-Rd in lanes 1-5,respectively.

6252

Immunology:

^{Dornet al.}

the X boxare soweak thatwehave been unabletoverify the protein binding site by methylation interference or DNase footprinting.

DISCUSSION

It has been noted repeatedly that the MHC class II genes

harbor^aconservedsequencemotif in the 5'-flanking region- the X and Y boxes (4-8). Here, we provide evidence that these elements play a role in regulating transcription by assessing the effect of their deletiononEa RNA synthesis in transgenic mice. Deletion of either the X orY box hastwo profound effects: first, the efficiency of transcription is reduced; second, RNA initiation isno longeraccurate. The first point isbasedonthe observation that wild-type levels of Ea RNA only occurwhenmanycopies of thetransgeneare present. Thus, we surmise that high copy numbercompen- satesfor weakpromoters.Some of the variabilitymaybe due to position effects, butwedo notthinkthis isanimportant influence because the wild-type Eagene seems tobe rather oblivioustoposition effects in the four independent lineswe

have tested.

The AX and AYmutantshavesuperficially similar pheno- types: both show reduced efficiencyand abnormal initiation of Ea transcription; furthermore, the tissue specificity of Ea RNAsynthesis is maintained, in thesensethat therearehigh levels of RNA in the spleen butnodetectable transcripts in heart and brain(datanotshown). One mightargue,then, that the altered phenotype could result completely or partially fromaperturbation ofpromoter structureduetotheloss of 14nucleotidesrather than from theparticulardeletion itself.

While this work was in progress, Takahashi et al. (18) demonstrated that the distance between promoterelements

can affect their function, probably due to stereospecific alignment of proteins along the DNA. Thus, our results should be confirmed with X and Y box point mutations, rather than deletions. In the meantime, we emphasize that important distinctions do exist between the AX and AY phenotypes: the distribution of aberrantstartsitesdiffers, the sitesfor WEAY301 being much moreclustered; in addition, the Y box, butnot the Xbox, deletion essentially abolishes Ear transcription in IFN-y-treated macrophages. This char- acteristic also extendstoEaexpression inmature,uninduced macrophages, which appearlargely E- in spleenorthymus sections from AY301-54 mice (data not shown; W. Van Ewyk, personal communication).

As we discuss at length elsewhere (27), the Y box

TGCTGATTGGC TA actually

contains a CCAAT se- quence in reverse. CCAAT has been shown to be an

important component of severalpromotersandcanoperate in reverse orientation asATTGG (for review, see ref. 19).

CCAAT deletions ormutations have been shown to affect basal levelsof transcription in tissue culturecells(ref.20and references therein) andto becapable of provoking aberrant initiation of mRNA (21). But the CCAAT sequence and adjacent nucleotides may also have a more complicated function, modulating transcript levels duringdifferentiation (22, 23) ormediating induction by certain effectors (21, 24, 25). Notsurprisingly, then, theCCAAT/ATTGG sequence

specifically binds a nuclear protein or proteins of broad cell-type distribution (for references, see ref. 19). It is of interesttoreflectonthe Y box in thiscontext.Our datashow thatthe Y box influences basal levels oftranscriptionaswell

as theaccuracy ofmRNAinitiation. Perhaps ofmore con- sequence, ourresultsprovidesupport for the contention that the CCAAT box and adjacent sequences can serve a more

sophisticated role: the Y box, which harbors a CCAAT

sequence, is indispensable for IFN-y induction ofEa tran-

scripts

initiated in its

vicinity.

We have detected a

protein

thatbinds

specifically

totheYbox, butwedonotknowhow, or even

if,

this

protein

mediates Ea

expression

in macro- phages. SinceNF-YisaCCAATbox-binding

protein

andhas a

ubiquitous

tissue

distribution,

it ismost

likely

thatatleast oneother

protein

isinvolved.

Our

evidence for the

importance

of the MHC class II conserved motifrelies

largely

on

studies

of

transcription

in

transgenic

mice.Wechose

transgenic

mice, rather than cells in

culture,

as anassay system because MHC

class

II genes are

subject

toaninteractive networkof

regulatory influences,

and it would be

difficult,

ifnot

impossible,

to

reproduce

this

complexity

with cultured cells (a problem

already

docu- mented

for

other genes, see ref. 26). In

addition,

we had

hoped

to, and

apparently

have been

able

to, create new mouse strains that harbor class II genes

expressed

in

only

certain compartmentsof the immunesystem.

We are grateful to C. Waltzinger, A. Staub, F. Ruffenach, P.

Gerber,S. Phayboune,F. Mackay-Landau,and C. Repisforassis- tancewith various aspects of this work. RecombinantIFN-y (pro- duced byGenentech)wassupplied by Boehringer

Ingelheim.

This work wassupportedby grants from the Association de laRecherche contre le Cancer (6472). A.D. received a fellowship from the DeutscheForschungsgemeinschaft.

1. Mengkc-Gaw,L.&McDevitt,H.0.(1985)Annu.Rev.Immunol.3, 367-3%.

2. Flavell,R.A.,Allen, H., Huber, B.,Wake,C. &Widera,G,(1985) Immunol.Rev.84,29-50.

3. Bottazzo, G., Pujol-Borrell, R. &Hanafusa, T. (1983)Lancetii, 1115-1119.

4. Mathis,D.J.,Benoist,C. O.,Williams,V. E.,II,Kanter,M.R.&

McDevitt,H. 0.(1983)Cell32,745-754.

5. Saito, H., Maki, R. A., Clayton, L. K. & Tonegawa, S. (1983)

Proc.Natl.Acad. Sci. USA80,5520-5524.

6. Kelly,A.&Trowsdale,J.(1985) NucleicAcidsRes.13,1607-1619.

7. Auffray, C.&Strominger,J.(1986) Adv. Hum.Genet. 15, 197-247.

8. O'Sullivan,D. M., Larhammar,D.,Wilson, M., Peterson,P. A.&

Quaranta,V.(1986)Proc.Nati.Acad.Sci. USA83,4484-4488.

9. Brinster,R. L.,Chen,H. Y.,Trumbauer,M. E.,Yagle, M.K.&

Palmiter,R.D.(1985)Proc.Natl.Acad.Sci. USA82,4438-4442.

10. Grundstrom, T., Zenke, W.M., Wintzerith, M., Matthes, H. W.D.,Staub,A.&Chambon,P.(1985)NucleicAcidsRes.13, 3305-3316.

11. Le Meur, M., Gerlinger, P., Benoist, C. & Mathis, D. (1985) Nature(London) 316,38-42.

12. Matthes, H. W. D., Staub, A. & Chambon, P. (1986) Methods Enzymol., in press.

13. Dignam, J., Lebovitz,R. &Roeder,R.(1983)NucleicAcidsRes.

11,1475-1489.

14. Carthew, R.W., Chodosh,L. A. & Sharp, P.A. (1985)Cell43, 439-448.

15. Mathis, D.J., Benoist, C., Williams, V. E., II, Kanter, M. &

McDevitt, H.0.(1983)Proc. Natl. Acad.Sci. USA80,273-277.

16. Dembic,Z., Ayane, M.,Klein, J.,Steinmetz,M.,Benoist,C.0,&

Mathis,D.(1985)EMBOJ.4,127-131.

17. Fried, M. & Crothers, D. M. (1981) Nucleic Acids Res. 9, 6505-6525.

18. Takahashi, K., Vigneron, M., Matthes, M., Wildeman, A.,Zenke, M.&Chambon,P.(1986)Nature(London) 319,121-126.

19. McKnight,S. &Tjian,R. (1986)Cell46,795-805.

20. Myers, R. M., Tilly, K. & Maniatis, T. (1986) Science 232,

613-618.

21. Bienz,M. &Pelham, H.R. B.(1986) Cell 45,753-760.

22. Collins, F. S., Metherall, J.E., Yamakawa, M., Pan, J., Weis- smann, S. M. & Forget, B. G. (1985) Nature (London) 313, 325-326.

23. Gelinas, R., Endlich, B., Pfeiffer, C., Yagi, M. &Stamatoyan- nopoulos,G.(1985)Nature(London)313,323-324.

24. Coen,D.M.,Weinheimer,S. P.&McKnight,S. L.(1986)Science 234,53-59.

25. Graves, B.J., Eisenman, R. N. & McKnight, S. L. (1985) Mol.

Cell. Biol.5,1959-1968.

26. Hammer,R. E., Krumlauf, R.,Camper, S.A., Brinster, R. L. &

Tilghman,S. M. (1987)Science235,53-58.

27. Dorn,A.,Bollokens,J.,Staub,A.,Benoist, C. &Mathis,D.(1987) Cell,inpress.