is expressed as at least three polypeptide chains (Id, sd, and ssd chains) specified by a ..... Hoffman, S., Sorkin, B. C., White, P. C., Brackenbury, R.,. Mailhammer, R. ... Murray, B. A., Hemperly, J. J., Gallin, W. J., MacGregor,. J. S., Edelman ...
Proc. Nati. Acad. Sci. USA Vol. 84, pp. 294-298, January 1987 Neurobiology
Organization of the neural cell adhesion molecule (N-CAM) gene: Alternative exon usage as the basis for different membrane-associated domains (exon/intron junctions/alternative splicing and polyadenylylation/functional domains/immunoglobulin gene superfamily)
GEOFFREY C. OWENS, GERALD M. EDELMAN, AND BRUCE A. CUNNINGHAM The Rockefeller University, 1230 York Avenue, New York, NY 10021
Contributed by Gerald M. Edelman, September 11, 1986
The neural cell adhesion molecule, N-CAM, ABSTRACT is expressed as at least three polypeptide chains (Id, sd, and ssd chains) specified by a single gene and derived by alternative splicing and polyadenylylation-site selection during RNA processing. We describe here the characterization of seven overlapping genomic phage clones reactive with N-CAM cDNA, indicating that the chicken N-CAM gene is more than 50 kilobases long. Analysis of the gene shows that there are at least 19 exons and that the coding sequences for the Id, sd, and ssd chains are assembled from 18, 17, and 15 exons, respectively. The first 14 exons appear to be common to all three chains and encode the amino-terminal portion of N-CAM, which contains five tandem homologous repeats resembling those seen in the immunoglobulin gene superfamily. In contrast to other genes containing such domains, each of these segments in N-CAM is specified by two exons. The carboxyl-terminal portion of each N-CAM chain is different as a result of the alternative use of exons. A single exon encodes the carboxylterminal 26 amino acids of the ssd chain and the 3' untranslated region of its mRNA, ending with a poly(A)-addition site. Two exons encode the transmembrane and cytoplasmic sequences common to the Id and sd chains, and another exon encodes the additional 261 amino acids found in the cytoplasmic domain of the Id chain. The carboxyl-terminal 21 amino acids common to the Id and sd chains and the 3' untranslated region common to their mRNAs are encoded by a single large exon of 3475 base pairs that ends with a second poly(A)-addition site. Sequences from the 13-kilobase intron that separates the exons encoding the amino-terminal and carboxyl-terminal regions of the molecule hybridize to a 2-kilobase poly(A)+ RNA transcript of unknown identity. This description of the chicken N-CAM gene provides a basis for determining the mechanisms that regulate the differential expression of the N-CAM polypeptide chains during development.
Specific adhesion between cells is an essential process in morphogenesis and histogenesis. Several cell adhesion molecules (CAMs) have been identified by functional assays and have been purified (1). From the distribution and properties of these CAMs, it has been proposed that modulation of CAM expression could generate the dynamic selectivity of cell-cell interactions that underlies the formation of tissues (2). A variety of mechanisms of CAM modulation have since been discovered. The most detailed analyses of these mechanisms have been carried out on the neural CAM (N-CAM), which is so far unique in showing modulation as a result of an RNA splicing event (3). N-CAM contains three glycosylated polypeptide chains with molecular weights of 170,000, 140,000, and 120,000, designated ld (large intracellular domain), sd (small intracelThe 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.
lular domain), and ssd (small surface domain), respectively (4-6). Sequencing of N-CAM protein and N-CAM cDNA clones has shown that each polypeptide has a different carboxyl-terminal region but shares a common amino-terminal region that includes both the homophilic binding region and the attachment sites for poly(sialic acid) (4-7). This portion of the molecule contains five contiguous segments of about 100 amino acids that are homologous in sequence to each other (ref. 5; J. J. Hemperly, G.M.E., and B.A.C., unpublished results). Each segment contains two halfcystines about 50 amino acids apart that are considered to form intrachain disulfide bonds (5-7), and each has some structural similarity to immunoglobulin domains (5). Consistent with the idea of CAM modulation, the relative expression of each polypeptide chain is markedly different during development (2). The sd chain is expressed at the earliest stages of embryogenesis, whereas the ld chain appears during neurogenesis and is restricted to the nervous system (3). In chickens, significant expression of the ssd chain is first seen post-hatching (4). Despite the multiple forms of the molecule, there appears to be a single gene for N-CAM in chicken, mouse, and human (8-10). Recent data have lent strong support to the idea that each chain arises by alternative RNA processing (4, 7). According to this notion, RNA splicing events modulate N-CAM by changing the cytoplasmic domain or the domain related to association with the cell membrane. In the present study, we wished to relate these modulation events explicitly to the exon organization of the N-CAM gene. We have isolated and characterized a series of overlapping genomic clones covering at least 50 kilobases (kb) of the chicken N-CAM gene and report here the organization of 19 exons encompassing sequences from the 5' end of the coding region to the 3' polyadenylylation site of the ld and sd mRNAs. We show that the exon organization can be strongly correlated with the domain structure of the N-CAM protein and propose a scheme of alternative exon usage that generates the ld, sd, and ssd polypeptides.
MATERIALS AND METHODS Isolation of N-CAM Genomic Clones. An EMBL3B bacteriophage library of chicken genomic DNA (11) was screened (12) with cDNA probes labeled by oligonucleotide priming (13). The cDNA probes used were pECO01, pEC201, pEC208, and pEC120 (refs. 4, 5, and 8; unpublished results). Phage DNA from positive plaques was prepared by the method of Yamamoto et al. (14). Abbreviations: CAM, cell adhesion molecule; N-CAM, neural CAM; Id, large-intracellular-domain polypeptide; sd, small-intracellulardomain polypeptide; ssd, small-surface-domain polypeptide; kb, kilobase(s).
Neurobiology: Owens et al. Restriction Enzyme Mapping and DNA Sequence Analysis. Genomic EcoRI fragments that hybridized to cDNA were subcloned into pBR328 and mapped by standard procedures (ref. 15, pp. 363-402). Exon/intron junctions were located by further subcloning of electrophoretically purified restriction fragments into M13 mpl8 or mpl9 (ref. 15, p. 51). Sequencing was carried out by the dideoxynucleotide chain-termination method (16, 17). Isolation of DNA and RNA. Chicken liver genomic DNA was prepared by the method of Blin and Stafford (18). Total RNA was extracted (19) and fractionated in 0.8% agarose gels containing 2.2 M formaldehyde as described (8).
RESULTS Previously, a genomic Southern hybridization using the N-CAM cDNA clone pECOO1 as a probe indicated that there is only a single N-CAM gene in the chicken (8). In the present study, a genomic Southern hybridization was carried out at high stringency using the N-CAM cDNA clones pEC208, pEC201, and pEC120, which together span almost all of the N-CAM ld chain mRNA (Fig. 1). pEC120 encodes the 3' untranslated sequences of the ld and sd mRNAs and was isolated from a cDNA library by using pECO01 as a probe (unpublished results). This clone hybridized to the same 6.6-kb EcoRI genomic fragment as pECO01 (Fig. 1, lane a) (8). pEC208, which contains most of the coding sequence of the ld polypeptide (5), gave the pattern of hybridization seen in lane b of Fig. 1. The 1.0-kb genomic fragment reactive with pEC208 is not evident in lane b but is clearly seen in lane c, where pEC201 was used as a probe. This clone contains sequences identical to 1.3 kb at the 3' end ofpEC208 (5). The 5.5-kb hybridization signal in lane b was later interpreted as a doublet of 5.7 kb and 5.5 kb, after the analysis of genomic phage clones. These EcoRI fragments were found in a series of seven overlapping genomic phage clones that were isolated by using either cDNA fragments or genomic phage subfragments as probes (Fig. 2). By accounting for all the EcoRI fragments reactive with N-CAM cDNA, we have further confirmed that there is a single N-CAM gene in the chicken. a
b
8.3_ 6.6. 5.5-
_
3.4..
_
c
2.8_
1.0_
FIG. 1. Hybridization of N-CAM cDNA to chicken genomic DNA. Embryonic chicken liver DNA was digested with EcoRI and analyzed by Southern blotting using the N-CAM cDNA probes pEC120 (lane a), pEC208 (lane b), and pEC201 (lane c). Sizes of the reactive EcoRI genomic fragments are indicated (in kb) at left.
Proc. Natl. Acad. Sci. USA 84 (1987)
295
In order to locate the particular exons that encode each of the N-CAM polypeptide chains (Fig. 3) within this continuous stretch of genomic DNA, two additional cDNA clones were used as hybridization probes: pEC254, which contains sequences corresponding to the amino terminus and overlaps the 5' end of pEC208 (J. J. Hemperly, G.M.E., and B.A.C., unpublished results), and XN151, which contains sequences unique to the ssd chain (4). Positive EcoRI subclones of the phages were mapped with several restriction endonucleases, and appropriate subfragments were cloned into M13 sequencing vectors. The exon/intron junctions were identified by comparing the sequences of these genomic subfragments with cDNA sequence. In Fig. 2, the position of each exon, commencing with the exon encoding the amino terminus, is diagramed. Additional exons that contain the sequences upstream of the exon encoding the amino terminus remain to be identified. Other exons that encode sequence variants of the N-CAM polypeptide chains may exist; additional forms of the molecule that have different electrophoretic mobilities have been observed (3). All of the junction sequences are listed in Table 1, starting with the exon encoding the amino terminus (designated exon 1) and ending with the polyadenylylation site for the ld and sd mRNAs (exon 19). Each donor and acceptor site is conventional; there are three type 0 junctions (those between codons), ten type I junctions (those interrupting codons after the first nucleotide), and four type II junctions (those interrupting codons after the second nucleotide) (20). The 3' ends of exons 15 and 19 are defined by a polyadenylylation signal followed by a terminal CA, which marks the site of addition of the poly(A) tail (21). The positions of the exon junctions in the N-CAM protein are summarized in Fig. 3, together with an inset showing the ld, sd, and ssd chains and their interaction with the cell membrane. The amino terminus is encoded in the smallest exon, corresponding to only 25 amino acids. This exon also contains the first half-cystine of the first homologous repeat. The remainder of the repeat is encoded by a second exon. Each of the other homologous regions are also encoded by two exons-namely, 3 and 4, 5 and 6, 7 and 8, and 9 and 10. While the sizes of exons 1 and 2 and of exons 9 and 10 are markedly different, each pair of exons adds up to about the same number of amino acids (see Table 1). The introns separating the exons that encode the five regions of homology vary in length; the fourth repeat is split by a 6.6-kb intron (Fig. 2). Between the regions of homology and the transmembrane domain common to the ld and sd chains there are five exons (Fig. 3). The largest intron separating N-CAM coding exons lies between exons 12 and 13, effectively separating the exons encoding the amino-terminal and carboxyl-terminal parts of N-CAM. This 13-kb intervening sequence was found to contain sequences homologous to a 2-kb transcript found in poly(A)+ RNA from 10-day embryonic brain tissues (Fig. 4). The 1.9-kb and 6.2-kb genomic EcoRI fragments contained within this intron (Fig. 2) were used as probes in an RNA transfer-blot experiment. These fragments hybridized to a 2-kb mRNA species, whereas the flanking EcoRI fragments, which contain N-CAM exons, hybridized to the ld and sd mRNAs. Whether this observation indicates that an unrelated gene is contained within the N-CAM gene awaits the isolation and sequencing of a cDNA corresponding to the 2-kb RNA. The transcript may be related to an RNA species of similar size, detected with an N-CAM cDNA probe in mouse (22). Exon 15 encodes the carboxyl-terminal 25 amino acids unique to the ssd chain and the 3' untranslated region of the ssd mRNA. In addition to these sequences, the cDNA XN151 contains the sequences encoded by exons 3-14 that are also found in pEC208 (4, 5). The divergence in sequence between
296
Neurobiology: Owens et al. (kb)
4.0
1.4 5.0
Proc. Natl. Acad. Sci. USA 84 (1987)
A-cNI 4.8 2.3 5.7
XcN2
XcN3
-
8.3
6.2
1.9
XcN4 5.5
XcN5 6.3
3.4
XcN6 2.8 1.0
6.6
41.5
..
10
89 1112
1 234567
[rUi
5' 0
1314
15
.
XcN7
19
16 17 18
3' 20
l0
50
40
30
FIG. 2. Physical map of the chicken N-CAM gene. The arrangement of the seven overlapping phage clones covering the N-CAM gene is shown above the map, and the size of each EcoRI fragment is indicated. The map of the N-CAM gene shows the positions and sizes of the exons encompassing the 5' end of the protein-coding sequence to the polyadenylylation site of the Id and sd mRNAs. Shaded areas represent exons; unshaded areas represent introns.
these two cDNA clones occurs at the 3' junction of exon 14, at an alanine codon. In the ssd chain, splicing of exon 14 to exon 15 generates the alanine codon GCG; in the ld and sd chains splicing exon 14 to exon 16 generates the alanine codon GCC (Table 1). Exon 16 contains the sequences encoding the membranespanning region, and exons 17 and 18 specify the cytoplasmic domains of the ld and sd polypeptides (5). The carboxylterminal 21 amino acids of the ld and sd polypeptides plus the entire 3' untranslated sequences of the corresponding mRNAs are encoded in the largest exon of the N-CAM gene. As previously shown (3), exon 18 is unique to the ld chain, and regulation of the splicing of this exon modulates the expression of the Id chain during development.
DISCUSSION Our results indicate that in the chicken, N-CAM is encoded in a single gene that is at least 50 kb long. The 19 exons identified thus far encompass sequences from the amino terminus to the poly(A)-addition site of the ld and sd mRNAs. Previous results indicated that the three polypeptides of N-CAM (the Id, sd, and ssd chains) are translated from separate mRNAs that arise by alternative RNA processing (3, 4). The data from the present analysis of the N-CAM gene are consistent with this conclusion and indicate that the ld, sd, and ssd chains are assembled from 18, 17, and 15 exons, respectively. The first 14 exons are common to all three 1,2
3,4
5,6
7,8
9,10
12 13
11
chains and appear to specify the extracellular domains of these chains (Fig. 3). Of the two intrinsic N-CAM polypeptides, the ld chain has a larger intracellular domain than the sd chain. This may result in different interactions with other cytoplasmic components in the cortex of the cell. The difference between these two chains is encoded in a single exon corresponding to 261 amino acids. Aside from the two exons that contain 3' untranslated sequences, the remaining N-CAM proteincoding exons are comparatively uniform in size and are considerably smaller than the id-specific exon 18 (see Table 1). Eukaryotic genes appear to evolve toward a uniform size for protein-coding exons (23): exon 18 may encode a distinct function acquired later in the evolution of the N-CAM gene. The ssd chain is attached to the lipid bilayer via phosphatidylinositol (4). The lipid-attachment site is presumably contained within a carboxyl-terminal sequence that is unique to the ssd chain (4). This carboxyl-terminal domain and the 3' untranslated sequences of the miRNA comprise a single exon of 750 base pairs (exon 15) located upstream from the exons encoding the transmnembrane and cytoplasmic domains of N-CAM (exons 16-19). Selection of the polyadenylylation site presumably regulates the formation of either the ld/sd or the ssd pre-mRNA. This could occur by early termination of RNA synthesis between exon 15 and exon 16 or by cleavage of a full-length pre-mRNA back to the poly(A)-addition site in exon 15. In studies of the calcitonin/calcitonin gene-related peptide gene, the data favor the latter mechanism (24). A scheme 16
14
17
19
18
NH2
COO H
15 COOH
NH2
NH2
sd
NH2
ssd
I
COOH
~~~~~~~~~~I
FIG. 3. Schematic representation of the N-CAM protein showing the position of the exon junctions in the id, sd, and ssd polypeptides. The exons common to all three chains are unshaded; the exons found in the Id and sd chains are shaded (the carboxyl terminus is encoded in exon 19). The Id-specific exon is hatched, and the carboxyl terminus of the ssd chain, encoded in exon 15, is shown below. All other exon numbers are listed above, and the position of each exon junction is indicated by a short line. Loops represent the disulfide-linked regions of homology. (Inset) Diagram of the three polypeptide chains and their association with the cell membrane; hydrophobic membrane-spanning regions are represented by shaded boxes, and the intracellular segment unique to the ld chain is represented by an unshaded box.
Proc. Natl. Acad. Sci. USA 84 (1987)
Neurobiology: Owens et al.
297
Table 1. Sequences around the exon/intron junctions of the N-CAM gene
Exon no.
Acceptor
Exon Leu.Cys CTA.TGT
tttcttttgcag
Ala CC
Ser
1 2
tattctctgcag
Val TG
Ala GCA
Gin
3
cattgctttcag
AA
Phe..
tctctcttctag
Val TT
Arg
4
CGG
TTT ..
5
cttctcttctag
Val TA
Pro CCT
Pro .. CCT ..
Asp
Gly .. Val
6
gtgcccctgcag
G
GAT
GGA.. Pro ..
ttttctttccag
Ala CA
Lys
7
AAA
CCC..
8
ctttcctgccag
Thr ACC
Leu CTG
GAT ..
Tyr cgtttatttcag
AT
Ala GCT
Pro ..
9
10
tttc ttccatag
Val GTG
Asp 11
t tcctcccccag
12
tttc tgcaacag
13
ccattg ttttcag
14
TCT
Gly.. Ile
Donor
Exon size, amino acids
Gln CAM
G
gtgagtgatggc
25
GGG ..
ATT
Phe TTC
C
g taagagaacaa
73
Lys
Leu ..
AAG
CTC..
Lys MAAA
Asp GAT
G
g taaggtgcaag
48
Val GTA
Asn AAT
G
g tgaag tggcac
46
Trp
Thr ACA
Lys AA
gtaagatgttgt
40
GTC
Phe TTT
G
gtgagccttttg
56
Glu GAG
Lys AAG
gtacaccttcca
48
Asp .. Val GTG
Gln CAG
g taag tggagag
50
CCC..
Leu CTG
Glu GAG
g ttgaggaaaag
62
Thr ACA
Pro.. CCA ..
Gln CAG
Al a G
g tgagcatccca
32
Thr ACT
Pro.. CCG ..
Lys
Gl u
AT
AAA
GAG
G
g taagatggtta
57
Ala CA
Asn AAT
Val ..
Pro CCA
Val C
g taag taaaac c
44
Arg
Glu
GAA
Pro.. CCC..
Lys
GG
MAAA
Al a GCT
gtaagtatgatt
42
ccatccctgcag
Lys AAA
His CAT
Ser .. TCA ..
Ile ATC
CCA
gtatggcatatg
60
15
gggcacctgcag
Ala Thr CG ACC
Ala
Ser
16
ggtc tc taaaag
CC
AGC
17
tctc tgttacag
G
AAA
18
tcttc ttcacag
G
CAC
19
attc ttc ttcag
Lys Thr G AAA ACC ...... TGTGTTCAAATAAAAATTACAA-
GTG..
TGG
GCG
GTT
T
Pro G
Leu
(0. 75 kb)
TTG.....TGTGAATACAAACATCTGTGGGAATTGCACTGAGACCATTCTT
Lys
His
Thr . Ala ACT . GCC
Phe TTC
Ser TC
Asp . Glu GAT . GAA
Pro CCA
GA
Thr . Lys ACC. MAAG
Thr
Glu
ACC
GA
gtagtacctcat
70
gtatgttgccat
39
gtatggc tgcct
261
Glu
(3.475 kb)
AAAAAAAAATACCAACAAAAACCTCATGCCT Exon/intron junction sequences were obtained by comparing the sequence of N-CAM genomic DNA subfragments with N-CAM cDNA Intron sequences are given in lowercase letters. The two polyadenylylation signals and sites of attachment of poly(A) are underlined.
sequences.
summarizing the alternative use of exons leading to the three N-CAM polypeptide chains is shown in Fig. 5. A significant feature of the N-CAM protein structure is the possession of five regions of homology. The half-cystines within each region are assumed to form a disulfide bond
because there are no detectable free sulfhydryl groups in isolated N-CAM and no disulfide bonds between N-CAM polypeptides (6, 7). Although sequence data indicate some similarity to immunoglobulin-like domains (5), each homologous repeat is encoded by two exons (Fig. 3). In the case of
Neurobiology: Owens et aL
298
a
b
c
Proc. Natl. Acad. Sci. USA 84 (1987)
required to define their properties and significance to N-CAM
d
expression. 7.4_
6.7-
Factors that regulate polyadenylylation and splicing during RNA processing must play a crucial role in modulating the expression of the three polypeptide chains. It is likely, however, that the factors which interact with sequences near the 5' end of the gene will regulate the transcriptional activity of the N-CAM gene. The relation of such control events with other events controlling the cell surface modulation of NCAM should provide a further basis for understanding N-CAM-mediated cell adhesion in development.
_ ,
2.0_
FIG. 4. Hybridization of genomic fragments to poly(A)+ RNA. Poly(A)+ RNA was extracted from embryonic brain tissue, fractionated in a formaldehyde/agarose gel, and blotted. Genomic EcoRI fragments spanning the region of the gene between exons 12 and 13 were used as probes (see Fig. 2). The probes were as follows: 8.3 kb (lane a), 1.9 kb (lane b), 6.2 kb (lane c), and 5.5 kb (lane d). The Id and sd mRNAs (7.4 kb and 6.7 kb, respectively) are indicated together with the 2-kb transcript.
the immunoglobulin genes and of other genes containing immunoglobulin-like domains, the immunoglobulin-related domains are encoded by single exons, supporting the hypothesis that formation of immunoglobulin-like genes occurred by duplication of an ancestral gene encoding a complete domain (25). Structural arguments have been put forward in favor of a half-domain progenitor gene (26, 27). To our knowledge, the results for the N-CAM gene provide the first example of such domains being specified by two exons. It should be noted that the positions and sizes of the exons in the regions of homology are different (see Figs. 2 and 3). This does not exclude the possibility that the present-day structure of this part of the N-CAM protein resulted from a tandem duplication event. If the homology between these segments in N-CAM and the immunoglobulin-related domains in other proteins is the outcome of divergent evolution, then it is likely that N-CAM constitutes an evolutionary branch point. The data presented here describe the structure ofthe exons specifying both the protein-coding sequences and the 3' untranslated sequences of the chicken N-CAM gene and show how the three N-CAM polypeptide chains may be formed by the alternative use of exons. Additional sequences upstream of the amino-terminal exon that react with N-CAM mRNA have been identified, but further analysis will be Id 5
3'
14
15AATACA
1617
18
15 AATACA
1617
18
15 AATACA
1617
18
19
AATAAA
sd
14
AATAAA
ssd
3 14
kb
19
AATAAA
FIG. 5. Schematic representation of the alternative patterns of RNA splicing that generate the three N-CAM polypeptide chains Id, sd, and ssd. Exon numbers correspond to Fig. 2; exons 1-14 are common to all three chains. AATAAA and AATACA are the alternative polyadenylylation signals.
We thank Dr. J. D. Engel (Northwestern University) for the generous gift of the chicken genomic library and thank Kim Magloire and Mark Petruzziello for excellent technical assistance. This work was supported by National Institutes of Health Grants HD16550, HD09635, and AM04256 and a Senator Jacob Javits Center for Excellence in Neuroscience Grant (NS-22789). 1. Edelman, G. M. (1985) Annu. Rev. Biochem. 54, 135-169. 2. Edelman, G. M. (1985) Exp. Cell Res. 161, 1-16. 3. Murray, B. A., Owens, G. C., Prediger, E. A., Crossin, K. L., Cunningham, B. A. & Edelman, G. M. (1986) J. Cell Biol. 103, 1431-1439. 4. Hemperly, J. J., Edelman, G. M. & Cunningham, B. A. (1986) Proc. Natl. Acad. Sci. USA 83, 9822-9826. 5. Hemperly, J. J., Murray, B. A., Edelman, G. M. & Cunningham, B. A. (1986) Proc. Natl. Acad. Sci. USA 83, 3037-3041. 6. Hoffman, S., Sorkin, B. C., White, P. C., Brackenbury, R., Mailhammer, R., Rutishauser, U., Cunningham, B. A. & Edelman, G. M. (1982) J. Biol. Chem. 257, 7720-7729. 7. Cunningham, B. A., Hoffman, S., Rutishauser, U., Hemperly, J. J. & Edelman, G. M. (1983) Proc. Natl. Acad. Sci. USA 80, 3116-3120. 8. Murray, B. A., Hemperly, J. J., Gallin, W. J., MacGregor, J. S., Edelman, G. M. & Cunningham, B. A. (1984) Proc. Natl. Acad. Sci. USA 81, 5584-5588. 9. D'Eustachio, P., Owens, G. C., Edelman, G. M. & Cunningham, B. A. (1985) Proc. Natl. Acad. Sci. USA 82, 76317635. 10. Nguyen, C., Mattei, M.-G., Mattei, J.-F., Santoni, M.-J., Goridis, C. & Jordan, B. R. (1986) J. Cell Biol. 102, 711-715. 11. Choi, O.-R., Trainor, C., Graf, T., Beug, H. & Engel, J. D. (1986) Mol. Cell. Biol. 6, 1751-1759. 12. Benton, W. D. & Davis, R. W. (1977) Science 196, 180-182. 13. Feinberg, A. P. & Vogelstein, B. (1983) Anal. Biochem. 132, 6-13. 14. Yamamoto, K. R., Alberts, B. M., Benzinger, R., Lawhorne, L. & Treibe, G. (1970) Virology 40, 734-744. 15. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY). 16. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA 74, 5463-5467. 17. Biggin, M. D., Gibson, T. J. & Hong, G. F. (1983) Proc. Natl. Acad. Sci. USA 80, 3963-3965. 18. Blin, N. & Stafford, D. W. (1976) Nucleic Acids Res. 3, 2303-2308. 19. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. & Rutter, W. J. (1979) Biochemistry 18, 5294-5299. 20. Mount, S. M. (1982) Nucleic Acids Res. 10, 459-472. 21. Birnsteil, M. L., Busslinger, M. & Strubik, K. (1985) Cell 41, 349-359. 22. Goridis, C., Him, M., Santoni, M.-J., Gennarini, G., Deagostini-Bazin, H., Jordan, B. R., Kefer, M. & Steinmetz, M. (1985) EMBO J. 4, 631-635. 23. Naora, H. & Deacon, N. J. (1982) Proc. Natl. Acad. Sci. USA 79, 6196-6200. 24. Amara, S. G., Evans, R. M. & Rosenfeld, M. G. (1984) Mol. Cell. Biol. 4, 2151-2160. 25. Barclay, A. N., Johnson, P., McCaughan, G. W. & Williams, A. F., in The T-Cell Receptors, ed. Mak, T. (Plenum, New York), in press. 26. McLachlan, A. D. (1980) in Protides of the Biological Fluids, ed. Peeter, H. (Pergamon, Oxford) Vol. 28, pp. 29-32. 27. Bourgois, A. (1975) Immunochemistry 12, 873-876.
