Proc. Natl. Acad. Sci. USA Vol. 74, No. 8, pp. 3157-3161, August 1977 Biochemistry

Independent expression of the gene coding for the constant domain of immunoglobulin light chain: Evidence from sequence analyses of the precursor of the constant region polypeptide (gene translocation/transcription control/initiator methionine/mRNA translation/sequence of cell-free protein product)

YIGAL BURSTEIN*, RONALD ZEMELL1, FRIDA KANTORt, AND ISRAEL SCHECHTERt Departments of * Organic Chemistry and t Chemical Immunology, The Weizmann Institute of Science, Rehovot, Israel

Communicated by Edmond H. Fischer, May 2,1977

ABSTRACT The mRNA coding for the K-type constant region (C,,) was purified from two clones derived from the MPC-1I mouse myeloma. This mRNA directs the cell-free synthesis of a C, precursor (molecular weight, about 15,000) in which an extra piece, 17 residues long, precedes the NH1terminal residue (Ma'19) of the C. region. Te partial sequence of the extra piece is: Met-X-Thr-Asp-Thr-Leu-Leu-Leu-Trp ValLeu-Leu-Leu-Trp-Val-Pro-X- (X is unknown). Met1 was shown to be the initiator methionine. The sequence of the C, extra piece is completely different from any known sequence preceding residue Ma'°§ in whole light (L) chains, thus establishing that the CK-region mRNA could not have originated from mRNA coding for the whole L chain. The structural features of the C. extra piece (marked hydrophobicity, size, and a methionine at the NHrterminus) are identical to those characteristic of the NHrterminal extra piece linked to the variable (V) region of whole L-chain precursors. In addition, the C, extra iece and the extra piece linked to the V region of MOPC-321 L chain have 70% sequence homology. These findings can be explained by the two genes-one Ig chain hypothesis, if we assume that the DNA coding for the extra piece (xp-DNA) is a constitutive part of the V gene. According to this model, the C,-region mRNA could have originated from: (i) translocation of this V gene to the C gene, deletion of the entire mature V gene, and "end-toend" repair of the remaining xp-DNA to the C gene; (ii) translocation to the C gene only of the xp-DNA portion of the V gene. Alternatively, we may assume that the xpWDNA is not covalently linked to the mature V gene at all times, as might be the case for the DNA of hypervariable regions presumed to be in episomes. This raises the intriguing speculation that the xp-DNA represents a third distinct gene, designated xp-gene. The presumed xp-gene may be involved in the regulation of gene transcription: when linked to the mature V gene it initiates a chain of events leading to whole L-chain mRNA formation; when at-, tached to the C gene it leads to its transcription to provide the

C-region mRNA. Abundant evidence suggests that distinct genes code for the variable (V, NH2-terminal) and constant (C, COOH-terminal) parts of the immunoglobulin (Ig) chain, and it has been repeatedly proposed that these genes are joined to provide a singI6 mRNA for the entire polypeptide chain (1-3). However, it was suggested that in some mouse myelomas V- or C-region protein fragments may be synthesized per se (reviewed in refs. 4 and 5). It has been reported that three clones derived from MPG-il myeloma synthesize a C-region fragment of K light (L) chain of molecular weight (Mr) 11,600 (5). It was suggested that these clones contain mRNA that directs the cell-free synthesis of a protein somewhat larger than the C, region; this protein was tentatively identified as a precursor to the C, region (6). These studies, however, did not determine the structure and position The costs of publication of this article were defrayed in part by the payment of page charges from funds made available to support the research which is the subject of the article. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

(NH2- or COOH-terminal) of the extra peptide in the presumed

C, precursor, nor did they provide evidence whether or not a

single gene codes for both the fragment and the whole L chain (5, 6). To resolve this problem requires analysis of the 5' end of the mRNA, or sequence analysis of the immediate product of mRNA translation to determine if its NH2-terminus is similar to the V-region sequence preceding the C region. We here report purification of the C,,-region mRNA from two MPC-1 1 clones, demonstrate that this mRNA programs the synthesis of a precursor to the C,-region, and determine the partial amino acid sequence of this precursor. MATERIALS AND METHODS Myeloma Tumors. Two clones derived from the MPC-1 cultured mouse myeloma (7), kindly donated by R. Laskov (Hebrew University, Jerusalem), were maintained as solid tumors in BALB/c mice (8). Clone 66.2, designated MPC-11(L), has lost the ability to synthesize intact heavy (H) chain but continues to synthesize and secrete the MPC-11 K L chain; clone NP-2, designated MPC-11(NP), has lost the ability to synthesize both intact H and L chains (5, 9). Cell-Free Synthesis and Analysis of Precursors. Translation of the mRNAs was carried out in the wheat germ cell-free system (10) at 250 for 4 hr (11). Cell-free products that were labeled by one radioactive amino acid at a time were analyzed in the Beckman model 890C automatic sequencer as detailed elsewhere (12). Radioactive amino acids were: [3H]Val (15.3 Ci/mnmol), [3H]Leu (53 Ci/mmol), [3H]Pro (43 Ci/mmol), [3H]Ala (31 Ci/mmol), [3H]Phe (15.8 Ci/mmol), and [35S]Met (240 Ci/mmol) from the Radiochemical Centre, Amersham; [3H]Thr (2.1 Ci/mmol), [3H]Trp (20 Ci/mmol), [3H]Asp (23.7 Ci/mmol), and [3H]Arg (27.3 Ci/mmol) from New England Nuclear. [35SJMet-tRNAMet. The [3aS]Met-tRNAMet species that transfer Met to the NH2-terminal (initiator Met-tRNAMet) and internal (internal Met-tRNAMet) positions in proteins were prepared from wheat germ according to a published procedure (13) with some minor modifications. RESULTS AND DISCUSSION mRNA Preparations. The mRNAs were prepared from polysomes specifically precipitated with antibodies to the L chain (14, 15). Polysomes of the MPC-11(L) myeloma were reacted with antibodies to MOPC-321 L chain (about half of these antibodies are specific to the C, region, see ref. 15); polysomes of the MPC-11(NP) myeloma were reacted with Abbreviations: V and C, variable and constant regions of immunoglobulin; H and L, heavy and light chains of immunoglobulin; Mr, molecular weight. -

3157

3158

Biochemistry: Burstein et al.

A

B

C

D

E

Proc. Natl. Acad. Sci. USA 74

F

G

by the MPC-11(L) mRNA. Gel electrophoresis of the total cell-free products shows that this mRNA programmed the synthesis mainly of three proteins with Mrs of about 25,000, 18,000, and 15,000 (Fig. 1B). Discrete radioactive peaks were obtained from sequencer runs of the total cell-free products labeled with [3H]Leu, [3HJVal, and [3H]Pro (Fig. 2). To determine which protein yielded these peaks, the total cell-free products were subjected to preparative gel electrophoresis (19), and the proteins with electrophoretic mobilities corresponding to Mrs of about 25,000, 18,000, and 15,000 were eluted. The patterns of radioactivity obtained from sequencer runs of the 15,000 Mr protein were identical with those obtained from the total cell-free products. The fact that the radioactive peaks from the total cell-free products (not shown) and from the 15,000 Mr protein (insets in Fig. 2) fall on straight lines in the semi-logarithmic plots strongly indicates that the residues scored originate from one protein species (12, 18). When comparable amounts of radioactivity (cpm) of the 18,000 and 25,000 Mr proteins were sequenced, high and discrete radioactive peaks were not detected (Fig. 2). Thus, the results of sequence analyses of the total cell-free products faithfully represent the unique sequence of the 15,000 Mr protein, which is superimposed (but not perturbed) on a high background of radioactivity derived from the 18,000 and 25,000 Mr proteins (see Fig. 2). Accordingly, in subsequent experiments the total cell-free products were analyzed. The discrete radioactive peaks derived from sequencer runs of products labeled with radioactive Leu, Val, Pro (Fig. 2), Met (Fig. 3), Asp, Thr Trp, Ala, and Phe (data not shown) assign these residues in the following positions: Met', Thr3, Asp4, Thr5, Leu6, Leu7, Leu8, Trp9, Vallo, Leu11 Leu2, Leu3, Trp'4, Val'5, Pro'6, Ala'8, Asp'9, Ala20, Ala21, Pro22, Thr23, Val24, Phe27, Pro28, Pro29, Leu34, Val4l, Val42. No radioactive peaks were obtained from sequencer runs (24 cycles) of cell-free products labeled with [3H]Arg. The alignment given in Fig. 4 shows that after 17 degradation cycles of the cell-free product, all 13 residues identified (from Ala'8 to Val42) show perfect homology with the corresponding residues starting from Ala,19 of the mature C, region (numbering of residues in the mature

H

FIG. 1. Autoradiogram of sodium dodecyl sulfate/polyacrylamide gels of [35SjMet-labeled cell-free products. The mRNAs employed to direct protein synthesis were: MPC-11(L) mRNA (B and H); disaggregated MPC-11(L) mRNA resolved on a sucrose gradient (16) to fractions of 20 S (C), 13 S (D), 9 S (E), 6 S (F), and 4 S (G); MPC-11(NP) mRNA (I); none (A). Bars mark the positions (top to bottom) of proteins with Mrs of 25,000,18,000, and 15,000. Mr standards were: ovalbumin, MOPC-321 L chain, myoglobin, and hemoglobin. Gel analyses were done according to Maizel (17), using the discontinuous Tris-glycine system, 13% acrylamide, 13 V/cm, 2 hr.

antibodies specific mainly against the C, region that were isolated from the anti-MOPC-321 L-chain antibodies (15). The antibodies specifically precipitated 4.4% and 1.2% of the total polysome populations of MPC-11(L) and MPC-11(NP), respectively. RNAs extracted from the immunoprecipitated polysomes were chromatographed on oligo(dT)-cellulose to yield active mRNAs (15). Sequence Analysis of the Cell-Free Products Programmed

I

1000 Mr 15,000 lo000-

l

24

X20 a

(1977)

ooo

_

-5(00

X

0

0

500 -500-

10011

10

10 2

8 164-

1

0

20 30

0

20 3040

10

20 30

100Xr1800

10

8-

-Mr 18,000 Mr 25,000

8 0

10

20

30

40

0

10

20 30 40 Sequencer cycle

0

10

20

30

40

FIG. 2. Radioactivity recovered at each sequencer cycle from cell-free products programmed by MPC-11(L) mRNA. The total cell-free products and fractions resolved by gel electrophoresis at the indicated Mr were sequenced (numbers in parentheses represent cpm in the sample analyzed): [3H]Leu total (76,600), 15,000 Mr (62,700), 18,000 Mr (60,000), 25,000 Mr (60,200); [3H]Val total (61,200), 15,000 Mr (57,300), 18,000 Mr (64,800), 25,000 Mr (56,000); [3H]Pro total (73,200), 15,000 Mr (49,200), 18,000 Mr (48,000), 25,000 Mr (42,400). Insets, semi-logarithmic plots (12, 18) of radioactive peaks derived from cell-free product Of Mr 15,000. Cycle zero represents a blank cycle (without phenylisothiocyanate) that was used to wash out potential radioactive contaminants.

Proc. Nati. Acad. Sci. USA 74 (1977)

Biochemistry: Burstein et al.

mRNA preparation contains an mRNA species that directs the synthesis of a precursor to the C, region with Mr of about 15,000. In this precursor an extra peptide segment, 17 residues long, precedes residue Ala1'9 of the C, region (in the intact L chain Arg'08 precedes Ala109; the absence of Arg in the first 24 positions of the precursor establishes that residue 17 belongs to the extra piece). The partial sequence of the NHrterminal extra piece of the Q, precursor (Metl-X17) is given in Fig. 4. To ascertain that the C, precursor is the immediate product of mRNA translation (i.e., to rule out the possibility that Metl was preceded by a peptide that was rapidly cleaved) we demonstrated that Metl is the initiator residue. Experiments summarized in Fig. 3 show that insertion of [a5S]Met into the NH2-terminal position of the 15,000 Mr protein occurs with the initiator [a`S]Met-tRNAMet, but not with the internal [SSS]Met-tRNAMet. The yield data also support this conclusion. The precursor contains two methionines (the initiator residue and Met175 of the C, region), both of which should be labeled with [35S]Met, while the initiator [35S]Met-tRNAmet should label only Met'. Accordingly, it is expected that when comparable amounts (cpm) of radioactive precursors labeled with initiator [SSS]Met-tRNAMet or [3-S]Met are sequenced, the radioactivity recovered in cycle 1 from the former should be twice the amount recovered from the [35S]Met-labeled precursor. The results given in Fig. 3 are in complete agreement with this expectation. In addition to the CK-region mRNA, the mRNA preparation contains mRNAs coding for the proteins with Mrs of 18,000 and 25,000 which have not yet been characterized. To gain information on the O,-region mRNA, the MPC-11(L) mRNA was disaggregated with dimethyl sulfoxide and sedimented over a sucrose gradient, and RNA fractions of about 20 S, 13 S, 9 S, 6 S, and 4 S were collected (16). The gradient fractions were translated and the cell-free products were analyzed by gel electrophoresis. Although the mRNAs coding for the three proteins are poorly resolved, there is a correlation between the size of the mRNAs and the proteins they program (Fig. 1 C-G). Because the 9S mRNA fraction directs the synthesis of the C,region precursor (Mr 15,000) but not the 25,000 Mr protein (size of an L chain) which is programmed by the 13S mRNA fraction, it seems that the O,-region mRNA is smaller than the mRNA for whole L chain (6). We have not identified the precursor of the whole MPC-il L chain, yet it is possible that it is the 25,000 Mr protein. This

03 P-4 x Q

2

1

C~~~~~

0

10

20

Sequencer cycle FIG. 3. Radioactivity recovered at each sequencer

cycle from the 15,000 Mr protein programmed by MPC-11(L) mRNA and labeled with [35S]Met. The total cell-free products that were synthesized with [35S]Met (A), initiator [35S]Met-tRNAMet (B), or internal [35S]MettRNAMet (C) as sole source of label were resolved by gel electrophoresis, and the 15,000 Mr protein was sequenced. Samples analyzed contained: A, 5680 cpm; B, 5870 cpm; C, 6230 cpm. Cycle zero, see legend to Fig. 2. K L chain is according to ref. 22). The probability that the matching sequence, Ala-Asp-Ala-Ala-Pro-Thr-Val-X-X-PhePro-Pro-X-X-X-X-Leu-X-X-X-X-X-X-Val-Val, occurs by chance is practically zero (if X denotes amino acid residues other than those indicated and the probability of each amino acid is 0.05 as it would be if all amino acids occur randomly, then the probability that the above sequence occurs by chance is 2.5 X 10-24). This result establishes that the MPC-ll(L) 92

95

3159

105

100

110

115

M-321 mature

Asx-Glx-Asx-Pro-Trp-Thr-Phe-Gly-Ser-Gly-Thr-Lys-Leu-Glu-Ile-Lys-Arg-Ala-Asp-Ala-Ala-Pro-Thr-Val-Ser-

C,,

Met- X

precursor

-Thr-Asp-Thr-Leu-Leu-Leu-Trp-Val-Leu-Leu-Leu-Trp-Val-Pro-

1

Z)r,

1n 11)

1 I .CZ1

-Ala-Asp-Ala-Ala-Pro-Thr-Val-

X

f

ZonL

M-321 precursor Met- X -Thr- X -Thr-Leu-Leu-Leu-Trp-\'al-Leu-Leu-Leu-Trp-Val-Pro- X -Ser-Thr- X M-321 mature

CK,

precursor

-

1).' b

-

X -Ile-Val-Leu-Thr-

Asp-I le-Val-Leu-ThrI 5 130

125

120

M-321 mature

X

Ile-Phe-Pro-Pro-Ser-Ser-Glu-Gln-Leu-Thr-Ser-Gly-Gly-Ala-Ser-Val-ValX -Phe-Pro-Pro- X 30

-

X

X

X -Leu- X

35

-

X

X

X

X

X -Val-Val-

40

FIG. 4. Alignment of the C, precursor, MOPC-321 precursor, and MOPC-321 mature L-chain. Partial sequence of the C. precursor is based analyses of precursor molecules labeled with radioactive Leu, Val, Pro, Met, Asp, Thr, Trp, Ala, Phe, and Arg. Partial sequence of the MOPC-321 precursor is from ref. 20. Sequences of two regions (Asp'-Thr5, Asx92-Vall33) of the mature MOPC-321 L-chain are from ref. 21. X, an amino acid not yet identified. Arrows mark end of the NH2-terminal extra piece and beginning of the mature L-chain sequence. on sequence

3160

Biochemistry: Burstein et al.

suggestion is based on the finding that the 25,000 Mr protein

programmed by the mRNA from the MPC-11(NP) clone (Fig. 1 H and I), and Mr 25,000 corresponds to the size of other L-chain precursors (11, 14). If this is indeed the case, we should have detected the sequence of the mature MPC-li L chain (23) after 20 degradative steps (approximate size of the extra piece, ref. 24) of the 25,000 Mr protein, but we failed to detect it. This failure may be attributed to facile hydrolysis of two peptide bonds at the NH2-terminal end of the mature MPC-il L-chain under mild acidic conditions (pH 4.8, ref. 20). The radioactive samples were subjected to repeated precipitations with trichloroacetic acid (12, 21) that should have led to the cleavage of these susceptible bonds. As a result the 25,000 Mr protein would have ragged NH2-terminal ends that can not yield a discrete sequence for positive identification (see Fig. 2). Other procedures have to be applied to identify the 25,000 Mr was

not

protein.

Sequence Analyses of the Cell-Free Products Programmed by the MPG-i 1(NP) mRNA. The major translation product of this mRNA is a protein with Mr of about 15,000 (Fig. 11). The total cell-free products programmed by the MPC-ii(NP) mRNA were labeled with radioactive Leu, Val, Pro, Met, Asp, Thr, Trp, Ala, and Arg, and were subjected to sequence analyses. In every case, the patterns of radioactive peaks obtained (data not shown) were identical to those recovered from the sequencer runs of the MPC-ii(L) mRNA cell-free products labeled with the same radioactive amino acid. Here again, residues in the cell-free product from Ala'8 show perfect homology with the corresponding residues starting from Ala'°9 in the mature C.-region (Fig. 4). These findings establish that the MPCG-l(NP) mRNA directs the synthess of a precursor (Mr 15,000) to the CK region in which an extra peptide segment, 17 residues long, is linked to residue Ala1°9 of the CK region. The partial sequence of the NH2-terminal extra piece of the CKprecursor is identical to that present in the CK-precursor programmed by the MPC-ll(L) mRNA. Concluding Remarks. Two MPC-lI clones were found to contain an mRNA species that directs the cell-free synthesis of a CK precursor with Mr of about 15,000. In this precursor an extra piece, 17 residues long, precedes residue Ala'°9 of the mature G, region; subsequent residues identified in the precursor show perfect sequence homology, with the CK region (Alal8-Val42 in the precursor matches with Alal09-Vall33 in the mature L chain, Fig. 4). On the other hand, regions preceding the above homology region are quite different; i.e., the extra piece segment in the CK precursor (Met'-X'7) bears no resemblance to any known sequence preceding residue Ala'09 in whole L chains (21), one of which (Asx92-Arg'08) is given in Fig. 4. The segment Asx92-Arg'08 comprises a portion of the V region; nonetheless sequence data of several L chains show that residues 98-108 are fairly conserved in mouse K L chains (5, 21). The sequence data and the identification of Met' as the initiator residue (Fig. 3) establish that the mRNA coding for the CK region could not have originated from the mRNA coding for whole L chain. These findings strongly indicate that the CK_ region mRNA is the result of transcription of the CK gene that is not dependent on transcription of the V gene. The NH2-terminal extra piece of the CK precursor is highly enriched with hydrophobic residues (82%, 14/17), and it is 17 residues long and contains NH2-terminal methionine (Fig. 4). These structural features characterize the NH2-terminal extra pieces that are linked to the V regions of whole L-chain precursors, i.e., they are markedly hydrophobic (69-75% hydrophobic residues), are of comparable size (19-22 residues long), and all contain NH2-terminal methionine (24). In addition, the

Proc. Natl. Acad. Sci. USA 74 (1977)

CK extra piece and the extra piece linked to the V region in the MOPC-321 L chain precursor (20) show at least 70% homology (Fig. 4). These structural similarities suggest translocation of the DNA segment coding for the extra piece from the V gene to the C gene. The data reported here can be formally explained by the two genes-one Ig chain hypothesis (25), assuming that the extra piece DNA (xp-DNA) is a constitutive part of the V gene, i.e., the V gene may be larger than hitherto realized (11, 24). We favor this model because the pattern of variability in the extra piece is closely related to the V-region subgroups (11, 24). The CK-region mRNA could then have originated from translocation of the extended V gene (that contains the xp-DNA) to the C gene, deletion of the entire segment of the mature V gene [none of the known MPC-lI V-region peptides (23) are retained in the Ca,-region precursor], and "end-to-end" repair of the remaining xp-DNA to the C gene. A related process is believed to occur in H-chain disease, except that short segments of the mature V region are usually retained and are joined to the C region of the defective H chain (26). Another possibility is that only the xp-DNA portion of the extended V gene was translocated to the C gene, and that this event is sufficient to permit transcription of the C gene. This proposal contrasts with the view that transcription is made possible only after joining the V gene with the C gene (2). The similarity of variability pattern in the extra piece and the V region does not necessarily mean that in the genome the xp-DNA is covalently linked to the mature V gene at all times. Originally, the xp-DNA may be separated from the mature V gene, a possibility similar to the proposal on the storage of information for hypervariable regions in episomes (27). This raised the intriguing speculation that three genes may control the synthesis of one Ig chain; i.e., in addition to the mature V and C genes, the xp-DNA represents a third distinct gene designated xp gene. The presumed xp gene may be involved in the regulation of gene transcription: when linked to the mature V gene it initiates a chain of events leading to whole L-chain mRNA formation; when attached to the C gene it leads to its transcription to provide the C-region mRNA. The structure of the extra piece linked to the V region in the MPC-11 L-chain precursor is not yet known; it may be identical to or different from the extra piece of the CK-region precursor. Regardless of the structure that will be eventually found for this V-region extra piece, the result will not invalidate the proposed models for generating the C.,-region mRNA. The findings, however will enable investigators to focus on additional problems. For example, if the extra pieces linked to the V and C regions of the L chains are different, then the concept of one plasmacytoma cell-one V gene expression, and the question whether a homogeneous cell population comprises the MPC11(L) clone, will require reexamination. In the present report it is shown that Met' of the C, precursor is the initiator methionine (Fig. 3). This finding strongly supports early suggestions that the NH2-terminal methionine found at the extra piece of each L-chain precursor investigated (three K and one X L-chain precursors, ref. 24) is the initiator residue (12). Consequently, translation of the mRNAs coding for the CK region and for whole L chains should be contingent on the integrity of the nucleotide sequence coding for the NH2-terminal extra piece. In agreement, it was found that L-chain mRNA molecules that are deficient at the 5' end are untranslatable in a cell-free system (16). We found that 1.2% of the total polysome population of the MPC-11(NP) clone bears nascent chains of the CK region, in agreement with the finding that the CK-region fragment com-

Biochemistry:

Burstein et al.

Proc. Natl. Acad. Sci. USA 74 (1977)

3161

prises about 1% of the proteins newly synthesized by intfd X (5). In this connection, it is worthwhile to mention that the immune precipitation procedure (15) proved useful for pre-

167 oIberts,-B. & Paterson, B. (1973) Proc. Nat!. Acad. Sci. USA 70, 2330-2334. 11. Burstein, Y. & Schechter, I. (1976) Biochem. J. 157, 145-151. 12. Schechter, . & Burstein, Y. (1976) Biochem. J. 153,543-550. 13. Ghosh, K., Ghosh, H. P., Simsek, M. & RajBhandary, L. (1974) J. Biol. Chem. 249,4720-4729. 14. Schechter, I. (1973) Proc. Nat!. Acad. Sci. USA 70, 2256-

We thank Prof. David Papermaster for discussion of the three genes-one Ig chain hypothesis, Mrs. Ida Oren for help with sequencer analyses, and Mrs. Etty Ziv for excellent assistance. R.Z. is a recipient of a Fellowship from the Medical Research Council of Canada. This research was supported by Grant CA-20817 from the National Cancer Institute, U.S. Public Health Service.

2260. 15. Schechter, I. (1974) Biochemistry 13, 1875-1885. 16. Schechter, I. (1975) Biochem. Blophys. Res. Commun. 67,

paring reasonably pure mRNA (see Fig. 1I) that comprised only 1% of the total mRNA population.

1. Dreyer, W. J., Gray, W. R. & Hood, L. (1967) Cold Spring Harbor Symp. Quant. Blol. 32,353-367. 2. Gally, J. A. & Edelman, G. M. (1970) Nature 227,341-348. 3. Hozumi, N. & Tonegawa, S. (1976) Proc. Nat!. Acad. Sci. USA 4. 5.

6. 7. 8. 9.

73,3628-3632. Shubert, D. & Cohn, M. (1970) J. Mol. Biol. 53,305-320. Kuehl, W. M. & Scharff, M. D. (1974) J. Mol. Biol. 89, 409421. Kuehl, W. M., Kaplan, B. A., Scharff, M. D., Nau, M., Honjo, T. & Leder, P. (1975) Cell 5, 139-147. Laskov, R. & Scharff, M. D. (1970) J. Exp. Med. 131, 515541. Potter, M. (1967) Methods Cancer Res. 2, 105-157. Coffino, P. & Scharff, M. D. (1971) Proc. Nat!. Acad. Sci. USA 68,219-223.

228-235. 17. Maizel, J. V. (1972) Methods Virol, 5, 179-246. 18. Smithies, O., Gibson, D., Fanning, E. M., Goodfliesh, R. M., Gilman, R. J. & Ballantyne, D. L. (1971) Biochemistry 10, 4912-4921. 19. Schechter, I., McKean, D., Guyer, R. & Terry, W. (1975) Science

188, 160-162. 20. Schechter, I. & Burstein, Y. (1976) Proc. Nat!. Acad. Sci. USA

73,3273-3277.

21. McKean, D., Potter, M. & Hood, L. (1973) Biochemistry 12, 760-771. 22. Svasti, J. & Milstein, C. (1972) Biochem. J. 128,437-444. 23. Smith, G. P. (1973) Science 181, 941-943. 24. Burstein, Y. & Schechter, I. (1977) Proc. Nat!. Acad. Sci. USA

74,716-720.

25. Dreyer, W. J. & Bennet, J. C. (1965) Proc. Nat!. Acad. Sci. USA

54,864-869. 26. Frangione, B. (1976) Proc. Nat!. Acad. Sci. USA 73, 15221555. 27. Wu, T. T. & Kabat, E. A. (1970) J. Exp. Med. 132,211-250.