Proc. Nati. Acad. Sci. USA Vol. 88, pp. 537-541, January 1991 Immunology

Evolution of the major histocompatibility complex: Molecular cloning of major histocompatibility complex class I from the amphibian Xenopus (major histocompatibility complex class ll/immunoglobulin superfamily/CD1/expression library)

MARTIN F. FLAJNIK, CAMILO CANEL, JACK KRAMER, AND MASANORI KASAHARA Department of Microbiology and Immunology, P.O. Box 016960 (R-138), University of Miami, Miami, FL 33101

Communicated by Max D. Cooper, October 15, 1990

Class I major histocompatibility complex ABSTRACT (MHC) cDNA clones have been isolated from an expression library derived from mRNA of an MHC homozygous Xenopus laevis. The nucleotide and predicted amino acid sequences show definite similarity to MHC class I molecules of higher vertebrates. The immunoglobulin-like a-3 domain is more similar to the immunoglobulin-like domains of mammalian class II (3 chains than to those of mammalian class I molecules, and a tree based on nucleotide sequences of representative MHC genes is presented.

Major histocompatibility complex (MHC)-encoded class I and class II molecules allow the immune system of vertebrates to distinguish self from nonself (1). MHC molecules are made up of four external domains, each composed of =90 amino acids. The two membrane-proximal domains are members of the C-1 subtype of the immunoglobulin superfamily (2), shared only with immunoglobulins and T-cell receptors. The two membrane-distal domains of both MHC classes are not immunoglobulin-like and combine to form a peptidebinding cleft composed of a floor of eight /3-strands topped by two a-helices (3). This structure evidently binds peptides derived from proteins either endogenously synthesized (for MHC class I) or exogenously acquired (for MHC class II) (4). T-cell receptors either recognize peptides and MHC molecules simultaneously or the peptide alone in "the context" of the MHC molecules (5, 6). To confront basic questions concerning MHC evolution, such as the derivation of the peptide-binding domains and which MHC class emerged first, it is necessary to examine the MHC of primitive vertebrates (7, 8). The isolation of MHC genes from nonmammalian vertebrates, however, has proceeded very slowly. The apparently rapid divergence of MHC sequences has not often allowed crosshybridization with mammalian gene probes. Crosshybridization with MHC genes of other vertebrate classes has only been successful in chickens with a mammalian class 11 -chain probe (9)perhaps due to a "slower divergence" of class II /3-chain sequences (8). Another approach, the polymerase chain reaction (PCR) with degenerate oligonucleotides from conserved MHC regions, has recently succeeded in the isolation of fish (carp) class I and class II genes (10). Antibody screening of expression libraries led to the cloning of several MHC-linked genes of the chicken (11, 12). In total, molecular information of the evolution of MHC is scanty. Xenopus offers an opportunity to study MHC regulation in two unique ways (for review, see ref. 13). (i) Tadpoles, although immunocompetent, do not display cell-surface expression of MHC class I until metamorphosis; class II molecules, in contrast, are expressed at all stages of life albeit

with a different tissue distribution (14, 15). (ii) Xenopus speciates by allopolyploidization, and there exist species with different numbers of chromosomes (e.g., the common Xenopus laevis with 36 chromosomes and Xenopus ruwensoriensis with 108 chromosomes). Species with many chromosomes (and the potential to express many MHC haplotypes), nevertheless, usually express only two MHC haplotypes (16). This is in contrast to many other genes that are expressed according to the number of available gene copies. As a first step in addressing all of the aforementioned problems, we have now cloned a X. laevis MHC class I cDNA from a liver expression library. Here we report a comparison of its sequence* with all known MHC genes.

MATERIALS AND METHODS Preparation of the cDNA Library. RNA was isolated by the guanidinium method (Stratagene) from the liver, thymus, and spleen of a froglet X. laevis homozygous for the f MHC haplotype (17). cDNA was prepared with the Uni-Zap system (Stratagene), which allows for unidirectional cloning of the resultant cDNAs. Six million primary recombinants were obtained and comprised this library. Screening of the Library. Antisera specific for frog class I were produced in mice. Two monoclonal antibodies (mAbs), TB1 and TB17, specific for both the native and denatured forms of class I (18) were covalently coupled to a protein G-Sepharose (Pharmacia) column following a procedure described by Schneider et al. (19). Erythrocyte membrane proteins (Xenopus erythrocytes bear large amounts of class I molecules on the cell surface) in detergent lysates were passed over the column, and the class I proteins were purified as described (18). Mice were either immunized directly to the column eluates or the proteins were further purified by gel electrophoresis and electroelution before the injections. Antisera derived from these mice were diluted at 1:500 in mAb supernatants (mAb immunoglobulin at 10ug/ml) to screen nitrocellulose filters bound with fusion proteins. The unamplified library of =600,000 total clones was screened. Gene and Protein Comparisons. The Intelligenetics FASTDB program (20) was used to search the GenBank (21), EMBL (22), National Biomedical Research Foundation (23), and Swiss-Prot (21) data bases for entries similar to the Xenopus class I sequences. Multiple sequence alignments were performed with the Intelligenetics GENALIGN program (developed by H. Martinez, University of California at San Francisco). The timeconsuming but more accurate Needleman-Wunsch alignment method (24) was chosen for some comparisons. The Abbreviations: MHC, major histocompatibility complex; mAb, monoclonal antibody. *The sequence reported in this paper has been deposited in the GenBank data base (accession no. M58019).

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

537

538

Immunology: Flajnik et al.

Proc. Natl. Acad. Sci. USA 88 (1991)

GENALIGN multiple sequence alignments were used as input to the Felsenstein PHYLIP [version 3.3 (1990), University of California at Berkeley] phylogenetic inference programs to obtain both maximum parsimony and distance matrix trees. Only the number of sequences and length of sequence parameters were adjusted. PROTOPARS and DNAPARS produced the maximum parsimony trees, and DNADIST followed by the FITCH distance matrix program produced the tree shown in Fig. 3.

TB1, which is specific for the cytoplasmic and denatured forms of Xenopus class I (18). Cross-hybridization analysis showed that the three clones isolated with the antiserum, CL10, CL12, and CL13, were progressively longer related cDNA clones (Fig. 1). The other clones isolated with the mAb are not bona fide MHC class I cDNA clones because they do not hybridize with the other three and require further characterization. Clone CL13 was sequenced in its entirety and found to encode all of the amino acids in the mature class I protein (Fig. 1). The entire sequence was then compared with others, and all of the best matches from DNA and protein data bank searches by using any parameter were mouse, human, or chicken MHC class I followed by class II data base entries. The greatest nucleotide similarity (50%o) was with the mouse class I gene H-2Kd. The entire Xenopus protein was most similar (27% similarity) to the mouse H-2Kb-encoded class I

RESULTS AND DISCUSSION Cloning of Xenopus Class I cDNAs. Seven clones were obtained from the cDNA library after screening with a combination of class I-specific polyclonal antisera and mAbs (25). Fusion proteins from three of the clones reacted only with the antiserum; the other four reacted only with mAb Leader 3eqeoe

al

*

*

dmm

begin

*

50 / GCG TTT TTA TTT CTG TTA TTC AGC CTC GAA ACC TCA GCA GTA TAT TGT CGA AGT CAC JATG Met Ala Phe Lou Phe Leu Leu Phe Ser Leu Glu There Ala Val Tyr CyslArg Ser His .

100

ACT CTA CAC TTT CAT ATC ACA TTG GCA TCA GCT CCG ATC CCT GGG CTT TCC CAG TAT GCC Thr Leu His Phe His Ile Thr Leu Ala Ser Ala Pro Ile Pro Gly Lou Ser Gln Tyr Ala

150

ATT ATT GTG TAC ATG GAT GGA GTG CAG TAT GGA AGG TAC AAC AGT GAC ATA CGT Ile Ile Val Tyr Met Asp Gly Val Gln Tyr Gly Arg Tyr Asn Ser Asp Ile Arg clam 12 starts . 200 * / CAG TTC TAC TCT GCA TCA TTG AAC CCA CTC TCC GTG TAC TTG GACIATG CAG ACC Gln Phe Tyr Ser Ala Ser Leu Asn Pro Leu Ser Val Tyr Leu Asp Met Gln Thr 250

.

.

.

CGT GCT Arg Ala

MAA TTT Lys Phe 300

.

GCT CAG ACA TTT GAG GTG TGG CAA AGG CAC AGG CTG AAC TTT CTA ATG GGA GTT TTC AAT Ala Gln Thr Phe Glu Val Trp Gln Arg His Arg Leu Asn Phe Leu Met Gly Val Phe Asn a2 domin begins clam 10 stazts .** 350 . / *1 ACA ACA AC GGTIAAT GGA CAT ATC CAT GTC TAC CAG AGG AAA TCT GCA TGT CAT Thr Thr Asn Gly Asn Gly His Ile His Val Tyr Gin Arg Lys Ser Ala Cys Glu Leu His

GAGICTG

400

GAT GAT GGC ACT ATT GGT GGA TAT CAG GAT ATT GCC TTT GAT GGG AAa GAG TTT TTA GTA Asp Asp Gly Thr Ile Gly Gly Tyr Gln Asp Ile Ala Phe Asp Gly Lys Glu Phe Leu Val

450

TTT GAT AAG GAG aGA GCG GTG TAC ATA ACT TCA ACA CAA GAG GCT GTG ATG GTG TCT CAT Ph. Asp Lys Glu Arg Ala Val Tyr Ile Thr Ser Thr Gln Glu Ala Val Met Val Ser His

500

CTA TGG MAT MAG CGT TAT GAT TCA ACT GAT AGT AAG ATG TTC TTG GAA ATT GAC TGC ATC

Lou Trp Asn Lys Arg Tyr Asp Ser Thr Asp Ser Lys Met Phe Lou Glu Ile Asp Cys Ile a3 domain begins 600 . . . . / 550

aAG CAC ATG AAG ATG TAC CTC CCT TAC ATA TCA AAT GAC CTG GAG MAG AAA| GTT CCT CCA Lys His Met Lys Met Tyr Leu Pro Tyr Ile Ser Asn Asp Leu Glu Lys Lys Val Pro Pro 650

AMT GTG AAG ATT TCC AGT TCA GAG TCA GAG AGT GGT ATC AAA CTG CAC TGC TGG GTG TAC Asn Val Lys Ile Ser Ser Ser Glu Ser Glu Ser Gly Ile Lys Lou His Cys Trp Val Tyr 700

AGG TTC TAC CCC CGA GAT GTG GAA GTG AAG TGG ATA AaG AAT GGG ACA GAT GaG ATT TAC Arg Phe Tyr Pro Arg Asp Val Glu Val Lys Trp Ile Lys Asn Gly Thr Asp Glu Ile Tyr

750 TCC GAG GAA TCA GCA GAG ATC CTG CCA AAC CCT GAT GGC ACC TAT CAG ATC AGA GTC AGT Ser Glu Glu Ser Ala Glu Ile Leu Pro Asn Pro Asp Gly Thr Tyr Gln Ile Arg Val Ser 800

.

GTG GAG GTA ACA CCA GAG GAG GGC GCC ACT TAC TCC TGC CAT GTG GAT CAC AGC Val Glu Val Thr Pro Glu Glu Gly Ala Thr Tyr Ser Cys His Val Asp His Ser trnans brane/cytoiasmi domain beoin . / . . . 850 GAG AAT CCC CTG GTT GTT CCC TTTIGTT CCC GTT GAA TCC AAC AAA AAA AAT ATT Giu Amn Pro Lou Vai Vii Pro Phe Vii Pro Vii Giu Ser Amn Lys Lys ASn Ile

AGT TTG Ser Lou

900 TTG TAT

Lou Tyr

950

ATC ATC ATC CCT GTT GGT TCA GGA TTA CTC GTC CTC TTG ATT GTC CTG GGT GTG TTT ATA Ie Ioe Ile Pro Vai Gly Ser Gly Leu Leu Vii Leu Lou Ile Vai Leu Gly Vai Ph. Ie t e 3' . / 1000 TAC AGA AAA GTT CCC AGT GGC CTT CAG AGT GTG AGC ACA GAG GAC GGA GAC TGA AAT TCC Tyr Arg Lys Vai Pro Ser Gly Leu Gin Ser Vai Ser Thr G1u Asp Gly Asp * z1ion begins 1050 AAA TAC ACT CCT CTA TGA TTT GAT GTA CAG TAA TAA ATG CCT TTG GAT (A)n

FIG. 1. Nucleotide and deduced amino acid sequence of the Xenopus class I cDNA clone CL13. The presumed boundaries of the domains and the start sites of the two shorter cDNA clones, CL10 and CL12, are indicated with arrows. (A)n denotes the poly(A) tail. A putative polyadenylylation signal is underlined.

Immunology: Flajnik

Proc. Natl. Acad. Sci. USA 88 (1991)

et al.

protein (26). The regions of the other two Xenopus clones (CL10 and CL12) that were sequenced were identical to the corresponding regions of clone CL13 (Fig. 1). Thus, only one class I-like cDNA sequence was selected from the expression library with the polyclonal antiserum; this is consistent with biochemical studies that suggested only one polymorphic class I gene/haplotype is expressed by Xenopus adult frogs (14, 27, 28). Although the leader sequence of clone CL13 begins with a methionine, whether this amino acid is truly at the beginning of the coding sequence is not known because only one more base at the 5' end is present in clone CL13. The 3'-untranslated region is much smaller than those of mammalian or avian class I mRNAs (1, 29). Amino Acid Comparisons with Other MHC Molecules. The similarities of the Xenopus class I amino acid sequence to the MHC molecules of other vertebrates are conspicuous (Fig. 2). The following features are of note: canonical glycosylation site at Asn-83 of the a-1 domain (all numbering is derived from the Xenopus sequence); conserved cysteines presumed to form intrachain disulfide bonds within the a-2 and a-3 domains; salt bonds between His-3 and Asp-29, Gln-21 and Asp-39, and Lys-155 and Glu-159. The similarities of many other amino acids throughout the molecule as compared with the other MHC sequences, especially in the a-3 domain, unambiguously demonstrate that the isolated cDNA clone could encode a Xenopus class I molecule.

1I.0( _ _

atl

201

130

_ - YNS _ t S

40

It has been pointed out that amino acids associated with the peptide-binding cleft (Fig. 2) are conserved between chickens and mammals (29). These amino acids, especially those with side chains that point away from the cleft (dots in Fig. 2), and some that point into the cleft (asterisks) are also conserved or conservatively changed in the frog sequence (3, 34). In contrast, those conserved amino acids with side chains that point up and could possibly interact with the T-cell receptor (small letter t in Fig. 2) are divergent in the frog sequence. The immunoglobulin-like a-3 domain has a unique second potential glycosylation site at the exposed region between the C and D }-strands, a loop connecting the two sheets (2, 3). The recently described stretch of amino acids that apparently makes up the CD8-binding site (residues 221-226) is not conserved in the Xenopus class I molecule (35); this region, however, like the human sequence, is rather acidic and may serve the same function. The glutamine at position 113, found in the chicken and human sequences to interact with 82microglobulin (3), is not conserved in the frog sequence; however, a glutamine at position 112 is shared with CD1 and may serve the same purpose. The Xenopus class I is the only MHC molecule in which the a-3 domain ends with a phenylalanine instead of a tryptophan. It is not yet known whether clone CL13 encodes the polymorphic class I molecule or a "nonclassical" class I protein (1, 36). Preliminary analysis of families by Southern blotting, the fact that only one class I molecule/haplotype has

60

50

ICA-Con

B-F12

Carp CDla HLA-DOQA

80 -L~~~

70 -.2-

-

Xen.

-tHRYFY

-

EI

AE-ASYGVNLY -YGSG-OYSHE-EYVDL -------TVL-P-VIK-

----,w - - ---

t* .t*.t**.t*.t**.t*

Xen.

HIA-Con B-F12

Carp CDla HLA-DQB

cccc

c

xen.

CDla

c

c

**t

.

c

c

c

180 + 170 150 140 130 120 110 t ! RtIGK$ l MMI Nt* A'aI -E GS(--To*-$Dk->LL GLRY>NTD-AAD I%2-RLAEO ELY- -2IqC EXAMS!NLiPVDEY;EDC9DT--L~iDK-SPETE4-SDPH n D CYQ -----YPF--E [email protected] - - --RLVS6S I VDIM at% EDF- i vxqGEFVFR-AG -t2M tNt. Itt.* tt* tt t* tt*.t* At .t*.* I ~~~~~~* * * * * p2m *

280 HLA-Con B-F12

c

100

90

c cCc cc c

TM!/CYT

ox I

AsQYSASL-jLS_ r PRAPWIE-QEGP e $ Sq PWVTVd-LFH J. TA-VPRTENI am RM SS v -PVFSAVEI -& ----- IEEQV-SILTEDD1APPPGTRLDLIK2W--ADGL-KEPLSFHVIWIASFYNHSNKQNLV -LSDLQTHTWDSNSSTIVFLWPN-SRGFSNEEKEL IIRSFEGIRYAHELOFE SQYAII

H

c c

ot2

539

VESNKKNILI

290

300

310 - --

cc ccCCcc cCc

c c

320

QS~YQDGD

S VVTGAAVVAA A SDSAQGSDVSLTACKV -AS iBSQPT-IPIV -@-- -* KGYNIAPGSNPAI IVAIAIHV -QPNLVPIVAGV ---L c FI EHHS ----SVG

FIG. 2. Amino acid sequence comparisons of Xenopus class I (Xen.), consensus human (primate) class I (HLA-Con; ref. 30), chicken class I (B-F12; ref. 29), carp class I (10), the non-MHC-encoded human class I molecule CD1a (31), human class II a chain HLA-DQA (32), and human class II ,8-chain HLA-DQB (33). Potential glycosylation sites, putative salt bonds deduced from the crystal structure (3), potential intrachain disulfide bonds, and /-strands in the immunoglobulin (Ig)-like a-3 domain are presented. Amino acids in the Xenopus sequence that are the same as at least one of the other sequences are boxed. Notations below the a-1 and a-2 domains signify characteristics of amino acids whose side chains are associated with the antigen-binding cleft; , pointing away from the cleft; *, pointing into the cleft; t, pointing up and potentially interacting with the T-cell receptor; c, amino acids conserved in mammalian sequences. 62m, an amino acid in mammalian and avian sequences presumed to interact with p2-microglobulin; CD8, the potential CD8-binding site of human class I molecules. Generally, gaps were placed following the predictions already made from reported chicken (29) and carp (10) sequences. Stars in the HLA sequence indicate that no consensus amino acid could be determined due to the extensive polymorphisms at these sites. All numbering is based on the Xenopus sequence. TM/CYT,

transmembrane/cytoplasmic.

Immunology: Flajnik et al.

540

Proc. Natl. Acad. Sci. USA 88 (1991)

Indeed, the similarities of the Xenopus a-3 domain to CD1 and class 1113 immunoglobulin-like domains do not extend to the peptide-binding domains (apparent in Fig. 2), suggesting that these two major regions of MHC molecules evolved independently. The class I a-3 and class II 8-2 MHC sequences from the lower vertebrates were sandwiched between non-MHC-encoded CD1 class I genes (arbitrarily chosen as the root) and mammalian class II 1-chain genes. Consistent with our findings, previous data suggested that class II,8 chains diverge more slowly than other MHC genes; hybridization with a mammalian 1-chain probe was the only case of successful cross-hybridization with MHC genes of other vertebrate classes. In addition, the class I gene family has evolved differently than class 11 (53): (i) in contrast to the well-known class II isotypes, there are no families of class I isotypes found between mammalian species; (ii) unlike class II, a variety of functions besides antigen presentation has been postulated for class I; and (iii) antibodies (mAbs and antisera) specific for human class II molecules have been found to react with native amphibian class II proteins, but human class I-specific antibodies rarely crossreact (8). Such class I "plasticity" of mammalian class I genes may also result in (or be the result of) a faster sequence divergence than for class II. The tree presented here is obviously preliminary due to the paucity of sequences from the lower vertebrates and the enormous distances between the sequences near the root. However, the tree is consistent with the sequence similarities previously mentioned, and it will serve as a starting point for future analyses. Polymerase chain reaction, performed with degenerate oligonucleotides, has recently allowed cloning of carp (10) and reptile (D. Grossberger, personal communication) MHC genes. Thus, soon it will be known whether trees with more sequences conform to the one presented in Fig. 3. Milstein and coworkers (31, 54) have already proposed, from sequence comparisons, that the non-MHC-encoded CD1 may be the extant molecule most like the common ancestor of class I and class II. The clustering of the sequences from the lower vertebrates with CD1 is not inconsistent with this hypothesis. If heat shock cognate 70 is subsequently defin-

been detected biochemically (14, 27, 28), and the cDNA sequence itself suggest that clone CL13, indeed, encodes the polymorphic, "classical" class I molecule. In addition, we have cloned a second allele by PCR that is almost entirely sequenced (data not shown); the second allele differs by many amino acids from the one described here, and, as expected, some of the polymorphic residues are associated with the cleft (data not shown). Comparisons of the Individual Domains. When the peptidebinding domains (a-i and a-2) of the Xenopus class I were compared, by themselves, to the protein banks, only poor similarity was detected with any described proteins, including MHC proteins (24% at best, with gaps). One of the best matches-with heat shock cognate 70 (HSC70)-is ofinterest because this protein has also been shown to bind peptides (37). Preliminary evidence indicates that the peptide-binding site of heat shock cognate 70 may adopt a structure similar to that of MHC class I (55). From gene and protein data bank searches, it was found that the immunoglobulin-like domain (a-3) of the Xenopus class I was more similar to immunoglobulin-like domains of mammalian class II 3 chains and their genes than to those of mammalian class I genes (best match of 41% amino acid identity with rat A,3). The first nine nucleotide sequences selected to be most similar to Xenopus class I were chicken, mouse, or rat class II 18-chain genes. The first 35 protein sequences most similar to the Xenopus a-3 domain were all human, rat, or mouse class II 1 chains, predominantly HLA-DQ/H-2A or HLA-DO/H-2A,12 sequences. The most similar class I molecule detected was human CD1 (37% amino acid identity, Fig. 2), a class I gene encoded outside the MHC but related to bona fide MHC genes (31). Phylogenetic trees, using mammalian, avian, and fish representative class I and class II a- and 13-chain nucleotide se6uences, were then assembled (Fig. 3). The immunoglobulin-like domains were chosen to formulate the trees because they are believed to serve a structural role in MHC proteins; thus, these domains, in contrast to the peptide-binding domains that are apparently under strong positive selection (52) may best conform to a "mainline" evolutionary analysis. 0

CD1 (30) class

0.5

1.0

1.5

2.0

2.5

3.0

l

1

3.5 Genetic distance

Fish dass 11 (10) Fish class 1(10)

Xenopus class Chicken class 1 (29) Chicken class 11 (38)

H-2 Ab (41) H-2 Eb (42)

Cat class 1 (43) HLA -A (44) H-2 K

(45)

class

Rat class 1 (46) ) H-2 07 (36) H-2 Aa (47) HLA - DOA (32) HLA - DNA (48) HLA - DPA (49)

H-2 Ea (50)

class 11 alpha chains

_v,

human P2M (51)

FIG. 3. Phylogenetic tree of the nucleotide sequences of the immunoglobulin-like domains including all known MHC sequences of lower vertebrates, CD1, mammalian class II , chains, mammalian class I a chains, and mammalian class II a chains. The FITCH distance matrix method was used. References are in parentheses after each sequence.

Immunology: Flajnik et A itively shown to have a similar peptide-binding site as MHC class I, then such a structure could have been transferred en masse to an immunoglobulin-like domain and given rise to a primordial class I molecule. We thank the following colleagues: Donna Sarapata and Eleanor Taylor for excellent technical help; Donna Sarapata for preparation of the figures; Ronda and Gary Litman for the first screenings of the gene and protein banks; Andrew Greenberg, Donna Sarapata, J. Wayne Streilein, Bonnie Blomberg, Churchill McKinney, Ramakrishna Kattepogu, and Richard Voellmy for critical reading ofthe manuscript. This work was supported by National Institutes of Health Grant A127877 and National Science Foundation Grant 8819366.

1. Klein, J. (1986) Natural History of the Major Histocompatibility Complex (Wiley, New York). 2. Williams, A. F. & Barclay, N. A. (1988) Annu. Rev. Immunol. 6, 381-405. 3. Bjorkman, P. J., Saper, M. A., Samaouri, B., Bennett, W. S., Strominger, J. L. & Wiley, D. C. (1987) Nature (London) 329, 512-518. 4. Germain, R. N. (1988) Cell 54, 441-444. 5. Nagy, Z., Lehmann, P. V., Falcioni, F., Muller, S. & Adorini, L. (1989) Immunol. Today 10, 132-138. 6. Zinkernagel, R. M. & Doherty, P. C. (1974) Nature (London) 248, 701-702. 7. Kaufman, J. F., Skoedt, K. & Salomonsen, J. (1990) Immunol. Rev. 113, 83-117. 8. Kaufman, J. F., Flajnik, M. F. & Du Pasquier, L. (1991) in Phylogenesis ofImmune Function, eds. Cohen, N. & Warr, G. (CRC, Boca Raton, FL), in press. 9. Bourlet, Y., Behar, G., Guillemot, F., Frechin, N., Billaut, A., Chausse, A. M. & Auffray, C. (1988) EMBO J. 7, 1031-1039. 10. Hashimoto, K., Nakanishi, T. & Kurosawa, Y. (1990) Proc. Natl. Acad. Sci. USA 87, 6863-6867. 11. Goto, R., Miyada, C. G., Young, S., Wallace, R. B., Alplanalp, H., Bloom, S. E., Briles, W. E. & Miller, M. (1988) Immunogenetics 27, 102-109. 12. Kaufman, J. F., Salomonsen, J. & Skjodt, K. (1990) Immunogenetics 30, 440-451. 13. Flajnik, M. F. & Du Pasquier, L. (1990) Immunol. Rev. 113, 47-63. 14. Flajnik, M. F., Kaufman, J. F., Hsu, E., Manes, M., Parisot, R. & Du Pasquier, L. (1986) J. Immunol. 137, 3891-3899. 15. Du Pasquier, L. & Flajnik, M. F. (1991) Dev. Immunol., in press. 16. Du Pasquier, L., Miggiano, V. C., Kobel, H. R. & Fischberg, M. (1977) Immunogenetics 5, 129-138. 17. Du Pasquier, L., Chardonnens, X. & Miggiano, V. C. (1975) Immunogenetics 20, 433-442. 18. Flajnik, M. F., Taylor, E., Canel, C., Grossberger, D. & Du Pasquier, L. (1991) Am. Zool., in press. 19. Schneider, C., Newman, R. A., Sutherland, E. R., Asser, U. & Greaves, M. F. (1982) J. Biol. Chem. 257, 10766-10769. 20. Brutlag, D. L., Dautricourt, J. P., Maulik, S. & Ralph, J. (1990) CABIOS 6, 237-245. 21. Bilofsky, H. S. & Burks, C. (1988) Nucleic Acids Res. 16, 1861-1863. 22. Cameron, G. N. (1988) Nucleic Acids Res. 16, 1865-1867. 23. George, D. G., Barker, W. C. & Hunt, L. T. (1986) Nucleic Acids Res. 14, 11-15. 24. Needleman, S. B. & Wunsch, C. D. (1970) J. Mol. Biol. 48, 443-453. 25. Landau, N. R., St. John, T. P., Weissman, I. L., Wolf, S. C.,

Proc. Natl. Acad. Sci. USA 88 (1991)

26. 27.

28. 29.

30. 31. 32. 33. 34.

35.

36.

37.

38. 39. 40.

41. 42.

541

Silverstone, A. E. & Baltimore, D. (1984) Proc. Natl. Acad. Sci. USA 81, 5836-5840. Reyes, A. A., Johnson, M. J., Schold, M., Ito, H., Ike, Y., Itakura, K. & Wallace, R. B. (1981) Immunogenetics 14, 383392. Flajnik, M. F., Kaufman, J. F., Riegert, P. & Du Pasquier, L. (1984) Immunogenetics 20, 433-442. Flajnik, M. F., Kaufman, J. F. & Du Pasquier, L. (1985) Dev. Comp. Immunol. 9, 777-781. Guillemot, F., Billaut, O., Porquie, A. M., Chausse, A. M., Zoorob, R., Kreiblich, G. & Auffray, C. (1988) EMBO J. 7, 2775-2785. Lawlor, D. A., Warren, E., Ward, F. E. & Parham, P. (1990) Immunol. Rev. 113, 147-185. Calabi, F. & Milstein, C. (1986) Nature (London) 323, 540-543. Moriuchi, J., Moriuchi, T. & Silver, J. (1985) Proc. Natl. Acad. Sci. USA 82, 3420-3424. Boss, J. M. & Strominger, J. L. (1984) Proc. Natl. Acad. Sci. USA 81, 5199-5203. Brown, J. H., Jardetzky, T., Saper, M. A., Samraoui, B., Bjorkman, P. J. & Wiley, D. C. (1988) Nature (London) 332, 845-850. Salter, R. D., Benjamin, R. J., Wesley, P. K., Buxton, S. E., Garrett, T. P. J., Clayberger, C., Krensky, A. M., Norment, A. M., Littman, D. R. & Parham, P. (1990) Nature (London) 345, 41-46. Waneck, G. L., Sherman, D. H., Calvin, S., Allen, H. & Flavell, R. A. (1987) J. Exp. Med. 165, 1358-1370. Flynn, G. C., Chappell, T. G. & Rothman, J. E. (1989) Science 245, 385-390. Zoorob, R., Behar, G., Kroemer, G. & Auffray, C. (1990) Immunogenetics 31, 179-187. Tonnelle, C., DeMars, R. & Long, E. (1985) EMBO J. 4, 2839-2847. Kappes, D. J., Arnot, D., Okada, K. & Strominger, J. L. (1984) EMBO J. 3, 2985-2993. Larhammar, D., Hammerling, U., Denaro, M., Lund, T., Flavell, R. A., Rask, L. & Peterson, P. (1983) Cell 32, 179-188. Saito, H., Maki, R. A., Clayton, L. K. & Tonegawa, S. (1983)

Proc. Natl. Acad. Sci. USA 80, 5520-5524. 43. van Hoof, M., de Geus, J. P., Roos, M., Brown, C., Jacobs, H. & Ploegh, H. (1989) Immunogenetics 30, 330-334. 44. Strachan, T., Sodoyer, R., Damotte, M. & Jordan, B. R. (1984) EMBO J. 3, 887-894. 45. Lalanne, J.-L., Delarbe, C., Gachelin, G. & Kourilsky, P. (1983) Nucleic Acids Res. 11, 1567-1577. 46. Radojcic, A., Stranick, K. S., Locker, J., Kunz, H. W. & Gill, T. J., III (1989) Immunogenetics 29, 134-137. 47. Benoist, C. O., Mathis, D. J., Kanter, M. R., Williams, V. E., II, & McDevitt, H. 0. (1983) Cell 34, 169-177. 48. Trowsdale, J. & Kelly, A. (1985) EMBO J. 4, 2231-2237. 49. Gustafsson, K., Widmark, E., Jonsson, A.-K., Servenius, B., Sachs, D. H., Larhammar, D., Rask, L. & Peterson, P. (1987) J. Biol. Chem. 262, 8778-8786. 50. Mathis, D. J., Benoist, C. O., Williams, V. E., II, Kanter, M. R. & McDevitt, H. 0. (1983) Cell 32, 745-754. 51. Gussow, D., Rein, R., Ginjaar, I., Hochstenbach, F., Seeman, G., Kottman, A. & Ploegh, H. L. (1987) J. Immunol. 139, 3132-3138. 52. Hughes, A. L. & Nei, M. (1988) Nature (London) 335, 167-169. 53. Klein, J. & Figueroa, F. (1986) Crit. Rev. Immunol. 6, 295-386. 54. Martin, L. H., Calabi, F. & Milstein, C. (1986) Proc. Natl. Acad. Sci. USA 83, 9154-9158. 55. Flajnik, M. F., Canel, C., Kramer, J. & Kasahara, M. (1991) Immunogenetics, in press.