Isotype restriction of murine antibodies towards the loop region of hen's egg white lysozyme

Immunology Letters, 17 (1988) 21 -28 Elsevier IML 00983

Isotype restriction of murine antibodies towards the loop region of hen's egg white lysozyme Jan A d r i a n u s Verschoor 1, Karen J a n s e Van Vuuren ~, L e o n Visser 1, Israel Pecht 2 a n d Ruth Arnon 2 1Department of Biochemistry, University of Pretoria, Pretoria, Republic of South Africa, and 2Department of Chemical Immunology, The Weizmann Institute of Science, Rehovot, Israel (Received 4 September 1987; accepted 9 September 1987)

1. Summary Monoclonal antibodies specific for the loop determinant (residues 6 4 - 80) of hen's egg white lysozyme demonstrated an immunoglobulin class restriction. Only IgM response against the loop could be evoked in mice, irrespective of whether the immunogen was the intact native lysozyme as such, or the loop peptide covalently conjugated to a synthetic carrier, poly-DL-alanyl-poly-L-lysine ( A - L). Despite the fact that in polyclonal antisera from mice immunized with lysozyme, the ELISA-titre of antiloop reactivity was very low, 26°7o of the total antilysozyme response could be accounted for by the loop when expressed as the percentage of antilysozyme hybridoma colonies producing monoclonal antibodies reactive with the loop. The results can be interpreted either as determinant controlled isotype restriction, or alternatively, as an affinity restriction leading to the phenomenon that antibodies of isotypes other than IgM are formed, but are of too low avidity to be detected by the ELISA method.

2. Introduction The peptide corresponding to the "loop" region Key words." Murine monoclonal IgM; Lysozyme loop peptide; lsotype restriction; ELISA assay Correspondence to: J. A. Verschoor, Dept. of Biochemistry, University of Pretoria, Pretoria, Republic of South Africa.

of hen's egg white lysozyme (residues 6 4 - 8 0 ) provides an interesting model for the study of antigeninduced activation of antibody effector functions [1, 2]. This stems from the convenient features of this system, which allows both physicochemical (predominantly spectroscopic), and parallel biological studies. (a) It is a monovalent well-defined conformation-dependent determinant derived from a native protein molecule [3, 4]. (b) It lends itself readily to synthetic procedures allowing comparison among structural analogues [5]. (c) Being devoid of aromatic residues, the loop neither absorbs nor fluoresces in the range where protein emission is measured, hence allowing convenient monitoring of its interaction with antibodies [6]. We have indeed employed this loop system and the specific antibodies raised against it in correlating complement activation with spectroscopic changes induced in the antibody upon antigen binding [2]. These studies, which were performed with polyclonal antibodies, demonstrated interesting differences between the behaviour of molecules of the IgG and IgM classes, fractionated from the same sample of affinity purified polyclonal anti-loop antibodies [7]. In order to enable a more quantitative evaluation of the observed differences, we embarked upon the preparation of monoclonal anti-loop antibodies of different isotypes. In the course of these studies we discovered what appears to be an antigenic determinantspecific isotype restriction, for which evidence is presented in this paper. A restriction of some kind of the murine immune response against the loop was already indicated in an

0165-2478 / 88 / $ 3.50 c? 1988 Elsevier Science Publishers B.V. (Biomedical Division)

21

earlier study in our laboratory [8]. While only a negligible serum antibody titre was observed against the loop following immunization with lysozyme, sizable IgM and lgG responses were exhibited in a haemolytic plaque assay. Since the haemolytic plaque assay is less sensitive to affinity differences than antibody titres [9], this observation may be interpreted as an affinity restriction of the murine antibody response towards the loop. In the present study we observed that, irrespective of whether l o o p - ( A - L) or intact lysozyme was the immunizing agent, the only class of monoclonal anti-loop antibodies that could be obtained was IgM. This restriction prevailed in spite of attempts to evade or circumvent it by several different procedures.

3. Materials and Methods

3.1. Screening for anti-loop response by ELISA Bovine serum albumin (fraction 5) was purchased from Boehringer Mannheim, South Africa. Casein, Tween 20, O-phenylenediamine and p-nitrophenyl phosphate were purchased from Merck, South Africa. Urea, hydrogen peroxide and 2-amino-2-methylpropane-l,3-diol were products of BDH Laboratory Chemicals, U.K. and purchased from Merck, South Africa. Peroxidase-conjugated goat anti-mouse IgG (heavy and light chain specific) was a product of Cappel, Worthington and purchased from Sterilab Services, South Africa. Rabbit anti-mouse IgM, IgGl, IgG2b and lgG3 were products of Miles Research Products Division, IN, USA, and were purchased from Seravac, South Africa. Alkaline phosphatase conjugated goat antirabbit IgG (H + L) was a Bio-Yeda product, purchased from Seravac, South Africa. Chicken antimouse IgG, (H + L) (IgY) and peroxidase mouse monoclonai anti-peroxidase were purchased from Bioclones Labs (Pty) Ltd., Seravac, South Africa. Lysozyme (salt free, 9600 U/mg) was obtained from Worthington, Freehold, N J, USA. 3.2. Preparation of the native and synthetic loop

peptides Native loop The native loop fragment, comprising the sequence 6 0 - 8 4 , was obtained from lysozyme by digestion with pepsin, chromatography on 22

Sephadex G-25, reduction with dithiothreitol, oxidation in air and final purification on Sephadex G-25 according to the procedure described previously [10, 11].

Synthetic loop The synthetic loop peptide, corresponding to the sequence 6 4 - 82 of lysozyme, and differing from native loop in position 76 where cysteine is replaced by alanine was prepared by solid phase peptide synthesis according to Merrifield [12] as described previously by Arnon et al. [13]. Merrifield peptide resin (chloromethylated polystyrene, 1 - 2 % cross-linked with divinyl benzene) was obtained from Pierce Chemical Company, If., USA. Nt-butyloxycarbonyl (N-t-boc)-x.-amino acid derivatives were a gift from Mr. I. Jacobson, Department of Biophysics, Weizmann Institute of Science, or were purchased from Sigma Chemical Company, Haifa, Israel. Amino acid side chains were blocked as the benzyl ethers (serine and threonine), benzyl thioether (cysteine), nitroguanido group (arginine) and the t3-benzyl ester of aspartate. Other amino acids used were glycine, proline, leucine, alanine and isoleucine. Trifluoroacetic acid, dichloromethane, ethanol, chloroform, triethylamine, N, N'-dicyclohexylcarbodiimidc, 1-et hyl-3-(3-diet hylaminopropyl)-carbodiimide HCI and dimethylformamide were all of analytical grade. Sephadex gels were products of Pharmacia Fine Chemicals and purchased either from Pharmacia, or Seravac, South Africa. 3.3. Characterization of lysozyme loop preparations The experimental amino acid compositions of the native and synthetic loop peptides agreed satisfactorily with the expected theoretical values and with previously reported results [13]. Cysteine was not determined as its presence is indirectly indicated by the successful closure of the loop during oxidation.

3.4. Preparation of loop-(A - L) conjugates" of natural and synthetic loop Poly-DL-alanyl-po[y-L-lysine ( A - L ) of mean molecular weight 40 kDa, prepared according to Sela et al. [14] was used as a carrier for the loop. Loop peptide (9 mg) and A - L (20 mg) were dissolved in water (5 ml). Water soluble carbodiimide (10 rag) was added and the reaction allowed to proceed for 20 h at room temperature. A slight precipi-

tate was removed by centrifugation and the conjugate was then purified by gel permeation chromatography on Sephadex G 100 in water. Amino acid analysis of the purified products showed 2.2 mol of native loop bound per mol A - L and 19.4 mol of synthetic loop bound per mol A - L . This difference probably resulted from variation in the quality of reagents used for coupling. The native loop conjugate with A - L was used for immunization as well as for ELISA screening, while the synthetic loop conjugate was used only for ELISA screening. 3.5. Preparation of monoclonal

anti-loop im-

munoglobulins Mice were immunized intradermally on day 0 with either lysozyme or l o o p - ( A - L ) (50 p.g per mouse) in a 1:1 solution of saline in Freund's complete adjuvant (CFA). On day 14, the same dose of immunogen was administered subcutaneously and intramuscularly. At least 4 wk were allowed to elapse before the same dose of immunogen was administered intraperitoneally and intravenously, on days 4 and 3, respectively, before cell fusion. Myeloma fusion partners were either NSI/I Ag4.1 or NSO/I as shown in Table i. Cell fusion was effected using polyethylene glycol according to the method of K6hler and Milstein [15]. 3.6. Screening for anti-loop response by ELISA ELISA tests [16] were performed in antigencoated flat-bottom plates. Lysozyme (1 #g/well) dissolved in "Iris buffer pH 7.0, or synthetic loop peptide (1 ~g/well) dissolved in 70°/0 methanol [17] were used for coating the plates. After incubation with the respective antibodies, either polyclonal serum (1/30

dilution) or supernatant culture (undiluted), the plates were washed with BSA/PBS and 0.05°7o Tween 20 [18] for reducing the non-specific adsorption of IgM. A conjugate of either horseradish peroxidase or alkaline phosphatase with goat anti-mouse lgG (H + L) served for the assay, followed after washing by addition of the substrate, O-phenylendiamine or p-nitrophenyl phosphate, respectively. The developed colour was monitored at 450 and 405 nm, respectively, in a Titertek Multiskan automatic reader (Flow Laboratories). In amplified indirect ELISA for determination of the isotypes, polyclonal rabbit antibody specific to the various murine immunoglobulin classes was added, followed by addition of alkaline phosphatase conjugated to goat anti-rabbit IgG (H + L). Isotype determinations of the monoclonal antiloop antibodies were confirmed by Ouchterlony immunodiffusion of affinity purified antibodies from ascitic fluid. 3.7. Preparation of polyclonal anti-lysozyme an-

tisera Nine outbred laboratory mice (obtained from the H.A. Grov6 Animal Research Centre, Pretoria, South Africa) were immunized with lysozyme as described above. The animals were bled 6 days after the first booster had been administered. Immunization using antibody mediated suppression [19] was done on 27 NMRI mice divided into 3 groups of 9 each. The first group of mice was immunized with 100 txg of lysozyme each in saline: FCA (1:2). Two weeks later the mice were boosted with the same suspension and 6 days later, bled. A second group of mice was likewise immunized but subsequent boosting was done as follows: the mice

Table 1 Protocol and results of cell fusions for anti-loop hybridoma production.

Immunogen Immunization period (days) Parental mouse line Parental myeloma line Screening antigen Lysozyme positive clones Loop positive clones Isotype of anti-loop clones

Fusion I

Fusion 2

Fusion 3

Lysozyme 60 DBA/2 NSI/I Ag 4.1 koop-(A - L) 30/500 4 All lgM

Loop-(A - L) 76 DBA/2 NSO/I Lysozyme 29/552 29 All lgM

Lysozyme 55 C57BL/6 NSO/I Loop-(A -- L) 58/432 15 All lgM

23

were intravenously injected with the antisera of the first group and 2 rain later the immunogen was subcutaneously administered in saline/FCA. This procedure effects antibody mediated suppression o f i m m u n o d o m i n a n t epitopes [19]. Six days later these animals were bled, the antisera pooled, screened for anti-loop response and isotypically characterized. The third group of mice was immunized with lysozyme and immune serum and likewise boosted 2 wk later. After 6 days they were bled, their antisera pooled and screened for anti-loop response and isotypic character. With the percentage of passive immunization defined as the percentage ratio of number of times of administration of immune serum and immunogen to total number of administrations of immunogen, the second and third groups of mice were designated as manifesting 50 and 100% passive immunization, respectively.

3.8. Ultra-violet irradiation of hybridoma cells Anti-loop monoclonal IgM producing hybridoma ceils (see Table 1, fusion 1) were irradiated with UV light according to the method described by Rosen and Klein [201. Briefly, the cells were suspended in culture medium to a cell density of 2.5 × 10 6 cells per ml. The cell suspension was then placed in Petri dishes and irradiated with an UV tube generating a dosage of 3.3 J . m 2.s-i for periods of time varying from 30 to 150 s. The cells were harvested from the Petri dishes, placed in growth flasks and left in a humidified incubator at 5°70 CO2 and 37 °C for 2 days. The cells were harvested from the growth flasks, centrifuged and the pellet suspended in a small volume of culture medium. This suspension was layered on top of filtered sterile horse serum in 10 ml test tubes. The test tubes were capped and placed in an incubator for 2 h. The supernatant, containing mainly dead cells, was removed by aspira-

22

22

20

2O

18" 16 ¸

16"

14.

14.

o

6 ~12

g12 u

c o u

> "= 10 o c~ c 8 o

"6 ~ 6

o-6 "&;

... Y2 D

I

90

270

810

2630

I I i

7290 21870

D i l u t i o n -]

r

!

i

65610

90

270

810

2430

Ill

I1 i 7290 21870 65610

D i ! u t ' , o n -"

Fig. 1. Antigenic reactivity of monoclonal anti-loop IgM towards lysozyme (a) and synthetic loop (b), measured by indirect sand~ich ELISA.

24

tion and the pellet of living cells suspended in culture medium and returned to the incubator. Once the growing colonies were confluent, the culture supernatants were collected and screened for antibody isotype production using ELISA. The colonies of sets subjected to either relatively low UV dosage or high dosage were subcloned on soft agar. These clones were also screened for antibody production and characterized for their isotypic nature.

munized with i o o p - ( A - L ) , and the supernatants were screened with lysozyme. The antigenicity of the synthetic loop appeared identical to that of lysozyme when assayed by means of an indirect ELISA. Fig. 1 shows the similar ELISA results obtained when monoclonal anti-loop IgM derived from mice immunized with lysozyme is allowed to react with either lysozyme (Fig. la) or with synthetic loop (Fig. lb); the latter bound to the plate in 70°70 methanol.

4. Results

4.2. Monoclonal antibody response to the loop Irrespective of whether intact lysozyme or loop( A - L) was used as immunogen, whether DBA/2 or C57BL/6 mice were the parental mouse lines, or whether NSI/1 or NSO/I ceils were employed as parental myeloma lines in the preparation of antiloop imunoglobulin producing hybridoma ceils, only IgM producing colonies were found (Table 1). An increase in the duration of the immunization

4.1. Antigenic properties of the synthetic loop To select for loop-specific monoclonal antibodies, three different cell fusions were performed. In two fusions the spleen cells were derived from mice immunized with lysozyme and the hybridomas were screened with native l o o p - ( A - L ) , whereas in the third fusion the spleen ceils were taken from mice im~

_

a

,6

12

12

08

08

0

b

C o~

o

o+

04

04

02

02

I 0

30

60

90

Seconds

12G

IbO

0

| 30

I 60

I ~'0

120

150

Seconds

Fig. 2. lVlonoclonal anti-loop antibody secretion o f lgM (a) and IgG (b) classes, after exposure of anti-loop lgM producing hybridomas to LIV irradiation for 0 - 150 s. Absorbancc values represent the results of amplified indirect sandwich ELISA.

25

schedule from 50 and 55 days to 76 days had no effect on the isotype expression of anti-loop antibodies. The results of fusion 3 indicate that of the total monoclonal anti-lysozyme clones, 26°7'o were reactive towards the loop. 4.3. Response in UV irradiated hybridomas Hybridoma cells producing anti-loop IgM (Table 1, fusion 1) were exposed to UV light irradiation for periods ranging from 0 to 150 s and isotypes of their antibody response were screened by the amplified sandwich ELISA. The results, presented in Fig. 2, show that irradiation of the hybridomas with UV light had little effect on the class of antibody produced, but the higher UV dosage reduced the level of antibody production. This observation was confirmed by subcloning 2 sets of the irradiated hybridoma cells, one exposed to relatively low dosage (60 s) and the other to relatively high dosage (150 s). Of the 24 subclones of the 60 s irradiated cells, one had apparently lost its antibody producing ability; while the rest remained IgM producers. Of the 144 subclones of the 150 s irradiated cells, all had apparently lost their antibody producing ability. 4.4. Polyclonal antibody response to the loop Six outbred mice were immunized and boosted with lysozyme, their immune sera screened for antiloop and anti-lysozyme antibodies by indirect sandwich ELISA and the antibodies isotypically characterized by amplified indirect ELISA. The results in Fig. 3 show that whereas lysozyme was able to evoke a significant polyclonal antibody response, only a negligible response to the loop could be detected.

r

.,.:;2r

,r"

-

'

t__

-

m

.

~

r

__'~3-~

__

".......

.--i..

1

o

'Or

;

..... .

Ln,~nt

'.

" ---

i

~

. ;

-

-

: :'

} ....

I

:,

barite

Fig. 3. Polyclonal a n t i b o d y response o f o u t b r e d mice to lysoz y m e ( o p e n bars) a n d l y s o z y m e - l o o p (filled bars). A b s o r b a n c e values represent the results o f a m p l i f i e d indirect s a n d w i c h E L I SA.

26

i

->

I'ig. 4. T h e effect o f passive i m m u n i z a t i o n on the level of a n d loop a n t i b o d i e s in a n t i s e r u m , as m e a s u r e d by indirect sandv,'ich ELLS,",,.

The antibodies to lysozyme were mainly of the IgGl isotype, while the isotype of the anti-loop response could not be determined due to the low levels of antiloop in the antisera. Immunization of mice using antibody mediated suppression was attempted in order to increase the polyclonal anti-loop response to a level which would enable its isotypic characterization. Antisera were screened using the indirect sandwich ELISA with synthetic l o o p - ( A - L ) conjugate coupled to thc plate. When lysozyme was administered in FCA after passive immunization of the micc with antilysozyme immune serum and likewise boosted 2 wk later (100°70 passive immunization), the levels of anti-loop antibody increased significantly (Fig. 4). However, the decreased sensitivity experienced with our amplified indirect ELISA performed on these sera for isotype characterization, did not allow an unequivocal establishment of the restriction towards igM. An IgM value of only about 1.6-fold higher than the background was observed and no lgG could be detected.

5. Discussion

It was previously demonstrated that the immune response of mice towards the loop is genetically controlled, but this property is not linked to the major histocompatibility complex [21]. Thus, the DBA/2 (H2 ~) and C57BL/6 (H2 ~) strains which were used in this study were, respectively, medium and high responders to the loop, when immunized with the l o o p - ( A - L)conjugate [22]. Hence, cell fusions with immune spleen cells derived from these strains of mice resulted in hybridomas producing monoclonal anti-loop antibodies. The finding that these antibodies are of the IgM class exclusively, strongly suggests an IgM class restriction irrespective of the loop carrier. Attempts at inducing class switch in one anti-loop IgM producing hybridoma cell line by UV irradiation according to the method of Rosen and Klein [20], or subsequent cloning and selection [23] did not succeed either, implying that the lgM restriction may be constitutive. Switch in isotype production is Th dependent [24], except for the case of T cell independent antigens or polyclonal B cell stimulators like LPS [25]. The loop peptide represents a monovalent peptide determinant on lysozyme, and lysozyme should therefore constitute a T cell dependent antigen for the loop specificity. When coupled to a synthetic carrier such as multichain ( A - k ) , the antigen may become T cell independent if the loop density on the immunogen is high, as was demonstrated for a small hapten such as DNP. The loop conjugate used in this study for immunization of mice and leading to the anti-loop producing hybridomas was of low epitopc density (2.2 mol of loop per mol of ( A - L)). Hence, it would be anticipated to behave as a T cell dependent immunogen. However, if any T cell help occurred in the observed anti-loop antibody response, as was reported for another synthetic antigen [26], this did not effcct any subsequent isotype switch. A possible explanation for these observations may be provided by the work of Bottomly et al. [27], who demonstrated the existence of different populations of B cells which can be distinguished by their respective membrane la protein densities. They found that low la densities are associated with antigen specific B cells which are generally difficult to activate to secrete immunoglobulin in vitro, whereas cells with high Ia densities are easily activated to secrete im-

munoglobulins. It would be of interest to examine the density of Ia expression on the surfaces of B cells induced by the loop. If loop specific B cells do have low surface Ia densities, this would explain why the loop appears generally non-immunogenic when measured in terms of antibody titre whereas, in fact, a relatively large number of B cells with anti-loop surface immunoglobulin exists. The anti-loop IgM producing cells may be part of a sub-population ot" B cells which, if they do switch to IgG production, undergo a decrease in Ia density to such an extent that immunoglobulin secretion ceases. We have previously observed [4] that the heterogeneity of polyclonal loop specific antibodies is restricted, which would corroborate the view that the loop region of lysozyme is capable of selecting only a specific subpopulation of B cells. Non-H2 control of the anti-loop response may then exert its effect on the level of Ia expressed on the particular B cell subset which is selected by the loop. The interpretation of ELISA results where binding of monoclonal antibody to peptide derivatives and the corresponding natural protein is measured, is by no means simple, as was clearly illustrated by Hirayama et al. [28]. By screening colonies of hybridomas derived from mice immunized with a synthetic lysozyme peptide conjugate by ELISA, they found monoclonal IgM reactive against lysozyme but not significantly against the peptide used for immunization. In contrast, one lgG1 clone was found which did react with the peptide, but was almost undetectable on a lysozyme coated plate. The authors believe that the effect could mainly be ascribed to the avidity attained by multivalent interaction of antibody to hapten on the solid phase. The possibility cannot be overlooked that we may have missed IgGs of low intrinsic affinity to the loop by screening with ELISA, although this appears unlikely due to the fact that immunization and screening with lysozyme and l o o p - ( A - L) were reversed in 2 separate cell fusions. In conclusion, it appears that the routine antiloop antibody response constitutes a determinant controlled, carrier independent restriction of the immune response leading to exclusive IgM production. The limited antibody heterogeneity as well as the low level of immunoglobulin secretion by anti-loop antibody producing cells lend credence to the view that the loop is selected by a sub-population of B cells with low surface la expression. The constitutive 27

s e c r e t i o n o f I g M b y t h e s e cells d o e s n o t n e c e s s a r i l y e x c l u d e t h e p o s s i b i l i t y o f a c l a s s s w i t c h , b u t in s u c h a n e v e n t it w o u l d b e e x p e c t e d t o l e a d t o a n t i - l o o p I g G s y n t h e s i z i n g B cells w i t h e v e n l o w e r levels o f s u r f a c e la expression, making them effectively nonsecretory.

Acknowledgement T h i s w o r k w a s s u p p o r t e d in p a r t b y t h e S o u t h African Medical Research Council.

References [I] Schlessinger, J., Steinberg, 1. Z., Givol, D., Hochman, J. and Pecht, I. (1975) Proc. Natl. Acad. Sci. USA 72, 2775. [2] Pecht, I., Ehrenberg, B., Calef, E. and Arnon, A. (1977) Biochem. Biophys. Res. Commun. 74, 1302. [3] Arnon, R. and Sela, M. (1969) Proc. Natl. Acad. Sci. I,JSA 62, 163. [4] Maron, E., Shiozawa, C., Arnon, R. and Sela, M. (1971) Biochemistry 10, 763. 151 Teicher, E., Maron, E. and Arnon, R. (1973) Immunochemistry 10, 265. [61 Pccht, 1., Maron, E., Arnon, R. and Sela, M. (1971) Eur. J. Biochem. 19, 368. [7] Pecht, l., Gafni, A. and Arnon, R. (1977) Israel 3. Mcd. Sci. 13, 1053. [8] Geiger, B. and Arnon, R. (1974) Eur. 3. Immunol. 4, 634.

28

[9] Herbert, N. J. (1968) Immunology 14, 301. [10] Canfield, R. E. and l.iu, A. K. (1977) J. Biol. Chem. 240. [11] Maron, I:.., Shiozav~a, C., Arnon, R. and Sela, M. (1978) Mol. lmmunol. 7, 119. {12] Merrifield, R. B. (1965) Science 150, 178. [13] Arnon, R., Maron, E., Sela, M. and Anfinsen, C. B. (1971) Proc. Natl. Acad. Sci. USA 68, 1450. [14] Sela, M., Katchalski, I{. and Gehatia, M. (1967) J. Am. Chem. Soc. 78, 746. [15] Kohler, G. and Milstein, C. (1975) Nature 256, 495. [16] Engval, E. and Perlmann, P. (1972) J. Immunol. 109, 129. [17] Hammerling, G. J., Hammerling, U. and Kearny, J. F. (1981) in: Monoclonal Antibodies and Hybridomas: Perspectives and Technical Advances. Elsevier/North Holland, Amsterdam. [18] Vollcr, A., Bid',veil, D. t{. and Bartlett, A. (1979)in: The ['nzyme Linked Immunosorbent Assay. Dynatech, l{urope, G.B. [19] Thalhamer, J. and Freund, J. (1985) J. hnmunol. Methods 80, 7. [20] Rosen, A. and Klein, G. (1983) Nature 306, 189. [21] Maron, E., Scher, 1., Mozcs, E., Arnon, R. and Scla, M. (1973) J. lmmunol. 111, 101. [22] Mozcs, [-_., Maron, E., Arm)n, R. and Sela, M. (1971) J. hnmunol. 106, 862. 123] Oi, V. T., Vuong, T. M., Reidler, 3., Dangl, .1., lqert/enberg, L. A. and Strycr, I.. (1984) Nature 307, 136. [24] Torrigiani, G. (1972).1. lmmunol. 108, 161. [25] Shimizu, A. and Honjo, T. (1984) (?ell 36, 801. 126l Axclrod, O. and Mo.,'es, F. (1986) hnmunobiolo~y 172, 99. [27] Bottomly, K., .fonts, B., Kaye, .1. and .lones 111, F. (1983) .I. Exp. Mcd. 158, 265. 1281 Hirayama, A., Takagaki, Y. and Karush, F..I. (1985) hnmunology 134, 3241.