Antigen-Binding Specificity of Isolated Cell-Surface Immunoglobulin from

Transcription

1 Proc. Nat. Acad. Sci. USA Vol. 69, No. 9, pp , September 1972 Antigen-Binding Specificity of Isolated Cell-Surface Immunoglobulin from Thymus Cells Activated to Histocompatibility Antigens (mouse T cells/l2i/precipitation) ROBERT E. CONE, JON SPRENT, AND JOHN J. MARCHALONIS Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, 3050, Australia Communicated by F. M. Burnet, June 30, 1972 ABSTRACT Lactoperoxidase-catalyzed radioiodination of cell-surface proteins was used in the isolation of cellsurface immunoglobulin from thymus-derived thoracic duct lymphocytes activated to histocompatibility-2 antigens. Immunoglobulin was identified by specific precipitation with antiserum to mouse immunoglobulin prepared in the rabbit and mouse immunoglobulin. Polyacrylamide gel electrophoresis of reduced and alkylated precipitates showed that the immunoglobulin molecules possessed I-type heavy chains and light chains. Cell-surface immunoglobulin isolated from thymus-derived cells activated to histocompatibility-2 antigens possessed binding specificity for the activating antigens. The ability of antisera to immunoglobulin to inhibit the immunocompetence of lymphocytes has been considered to provide evidence that specific interaction of these cells with antigen is mediated by immunoglobulin-like receptors located on the cell surface (1-7). Such studies, however, do not exclude the possibility that receptor sites are not immunoglobulin in nature but are situated sufficiently close to immunoglobulin molecules for antibodies directed against immunoglobulins to block, by steric hindrance, the access of antigen to receptor sites. This objection would be obviated if it were demonstrated that surface immunoglobulin extracted from lymphocytes specifically reactive to a certain antigen could bind specifically to that antigen. Such an approach has become feasible because lactoperoxidase (EC )- catalyzed radioiodination (8) of cell surface proteins (9-11) has enabled the isolation and partial characterization of surface immunoglobulin molecules from thymus-influenced (T) lymphocytes (12, 13) and bursa or bone-marrow-derived (B) lymphocytes (11-14). This sensitive technique provides a means of testing directly whether lymphocyte surface receptors for antigen are immunoglobulins. T lymphocytes serve as effector cells in cell-mediated immune responses and act as "helper" cells in enabling B lymphocytes, the precursors of antibody-secreting cells, to synthesize and secrete antibodies to certain antigens (15). Because T lymphocytes do not secrete antibody in the classical sense they are theoretically more suitable than B lymphocytes for determining whether cell-surface immunoglobulin functions as an antigen receptor. In addition, techniques are available for generating, in the absence of antibody-secreting cells, large numbers of T cells reactive to a specific antigen (16, 17). Abbreviations: T cells, thymus-derived lymphocytes; B cells, bone-marrow-derived lymphocytes; H, histocompatibility. In this report we describe the isolation of surface immunoglobulin molecules from a highly enriched population of T lymphocytes activated by histocompatibility (H)-2 antigens in vivo. These cells were obtained from thoracic duct lymph of heavily irradiated F1 hybrid mice injected with parental thymus lymphocytes. The immunoglobulin molecules were isolated under physiological conditions that permitted antigen-binding studies to be performed with the isolated molecules. The results indicate that the surface immunoglobulin of H-2-activated T cells possesses specificity for the activating antigens that parallels the functional activity of the cells. MATERIALS AND METHODS Cell Preparations. Populations of H-2-activated, thymusderived, thoracic duct lymphocytes were obtained by the the method of Sprent and Miller (17). Briefly, 3-month-old (CBA x C57BL)F1 or (CBA x BALB/c)F1 mice were injected intravenously with 2 X 108 CBA thymus cell's within 2 hr after receiving 750 R total body irradiation. 4 Days later the thoracic ducts of the mice were cannulated, and lymph was collected over the first hr of drainage % of thoracic-duct lymphocytes obtained in this way are cells of donor-thymus origin. Such preparations contain less than 0.3% B lymphocytes (17). Suspensions of normal CBA/H/Wehi, BALB/c, or C57BL thymus cells were prepared as described (11, 12). Preparation of Cell-Surface Immunoglobulin. T lymphocytes from thoracic ducts were 95-98% viable as judged by eosin dye exclusion. The cells were washed 5 times with Eisen's balanced salt solution and a total of 3 X X 108 cells were iodinated with [1251] iodide as described (9), except that the total reaction volume was 50 Al. We have previously shown that this technique iodinates cell-surface proteins exclusively but does not affect the viability of the cells (9). After iodination the cells were incubated under short-term cell-culture conditions for 2-6 hr as described (13, 18). Under these conditions cell-surface proteins, including immunoglobulin, are released into the cell-culture medium (13, 18). Aliquots were removed at intervals and centrifuged at 40 at 1500 rpm (500 X g) for 10 min, and the supernatants were retained. Radioactive cell-surface immunoglobulin was isolated from the supernatants by specific coprecipitation with antiserum against mouse immunoglobulin prepared in rabbits and purified mouse IgG as carrier (12, 13). The rabbit antiserum reacted with light chains and Iu heavy 2556

2 Proc. Nat. Acad. Sci. USA 69 (1972) chains. Controls for nonspecific precipitation included the use of normal rabbit serum in place of antiserum against mouse immunoglobulin prepared in rabbits or heterologous precipitation systems consisting of limulus hemocyanin and antiserum to limulus hemocyanin prepared in rabbits or bovine serum albumin and antiserum to bovine serum albumin prepared in rabbits. Conditions that specifically precipitated more than 80% of the antigen were used. Radioactivity was determined in a Packard autogamma scintillation counter with a deep well NaI crystal detector. Resolution of Cell-Surface Immunoglobulin into Polypeptide Chains. Precipitated immunoglobulin was dissolved in 9 M urea and reduced and alkylated to cleave interchain disulfide bonds (19). Reduced and alkylated samples and immunoglobulin markers were resolved into polypeptide chains by disc electrophoresis in acid urea (20). Antigen-Binding Assays with Cell-Surface Proteins. 100 JI of cell-surface protein supernatants were mixed with 1 X 107 allogeneic or syngeneic thymus cells suspended in 100 ml of Eisen's balanced salt solution and 10% fetal-calf serum (Commonwealth Serum Laboratories, Melbourne, Australia). The cells were then incubated for 3 hr at 40 and washed 5-6 times with the same solution. The amount of radioactivity in the cell pellet was determined. Statistics. P values were calculated by a modified Student's t-test for small samples (21). Isolation and Turnover Rate of T-Cell Surface Immunoglobulin. Surface proteins, including immunoglobulins, are continually synthesized and released by living lymphocytes (12, 13, 18, 21-24). Accordingly, to recover surface immunoglobulin molecules under gentle conditions in a state suitable for antigen-binding analysis, we isolated proteins released from the cell surface by membrane turnover (13). 108 Activated T cells were iodinated with 125I in aliquots of 107 cells. The radioiodinated cells were incubated for various intervals in tissue culture medium and then discarded after centrifugation. The supernatants were retained. Immunoglobulin was detected in the supernatants by specific precipitation. After at least four washes, counts in '25I-labeled surface proteins precipitated by antiserum to mouse immunoglobulin prepared in rabbits were 6- to 10-fold higher than counts precipitated by normal rabbit serum, and 4- to 7-fold higher than counts precipitated by limulus hemocyanin-anti-hemocyanin or bovine serum albumin-anti-bovine serum albumin, for cell-culture supernatants obtained after 2-6 hr incubation. The difference in the amount of radioactivity specifically and nonspecifically precipitated for 4-6 replicate samples was highly significant (P < 0.001). The quantity of radioactive material in the supernatant that was precipitable by antiserum to immunoglobulin increased linearly for 6 hr. The amount of immunoglobulin precipitated after 6 hr corresponds to about 8% of the iodinated macromolecular material released by the cells, as estimated by gel filtration of supernatants on Sephadex G-25. Polypeptide Chains of Activated T-Cell Surface Immunoglobulin. The type of polypeptide chains present in the specifically coprecipitated immunoglobulin obtained from T lymphocytes from thoracic ducts was ascertained by gel electrophoresis in acid urea. As may be seen in Fig. 1, surface immunoglobulin from T lymphocytes from thoracic ducts E 300,200 U 0 < 100 Activated T-Cell Surface Immunoglobulin A ' t x - * RELATIVE MOBILITY FIG. 1. Analysis of polypeptide chains of immunoglobulin isolated from the surfaces of BALB/c-activated T lymphocytes by disc electrophoresis in acid urea. Surface iodinated T lymphocytes were incubated in tissue culture medium for 3 hr at 37. The cells were centrifuged and the supernatants were retained for precipitation. The precipitates were dissolved in 9 M urea, reduced, and alkylated. The bars indicate the positions of polypeptide chains from purified immunoglobulins that were used as standards: MA, A-chain, ay, -y-chain, L, light chain. 5% acrylamide was used. Background radioactivity resulting from electrophoresis of material brought down by normal rabbit serum has been subtracted from each point. The electrophoretic pattern of counts precipitated by the heterologous precipitating system (limulus hemocyanin-anti-hemocyanin) was indistinguishable from background. activated by BALB/c antigens possessed polypeptide components that resembled light chains and heavy chains of standard immunoglobulin in mobility. The heavy chains resembled A-type heavy chains in their penetration of the gel. Similar patterns were obtained for T lymphocytes activated by C57BL antigens. Polyacrylamide gel electrophoresis of counts precipitated by normal rabbit serum or heterologous precipitations gave no components that did not fall within the range of background variation. For this reason, normal rabbit serum was used as a control in further studies. Antigen-Binding Specificity of Activated T-Cell Surface Proteins. To determine whether cell-surface proteins isolated from T lymphocytes from thoracic ducts possessed specificity for the activating antigens, ln2i-labeled cell-surface proteins obtained from T lymphocytes activated either by C57BL or BALB/c antigens were incubated separately with either CBA, C57BL, or BALB/c thymus lymphocytes. The cells were washed repeatedly after incubation, and radioactivity on the cells was monitored after each wash. A plot of the loss of radioactivity from the cells against number of washes yielded biphasic curves. The first phase consisted of rapid exponential attrition and was followed by a phase in which radioactivity bound to the cells decreased only slightly upon further washing. Specific binding of cell-surface proteins was L

3 2558 Immunology: Cone et al. Proc. Nat. Acad. Sci. USA 69 (1972) TABLE 1. Specificity of surface proteins of 125I-labeled T lymphocytes for activating antigens Cell-surface cpm Binding to thymus cells proteins Treatment C57BL BALB/c CBA BALB/c-activated* None 7900 d: ,200 i= i 550 CBA-T lymphocytes P < 0.05 P < 0.01 Precipitated with 1300 i ± antiserum to im- ns ns munoglobulin Centrifuged in nd 12,200 ± normal rabbit serum P < 0.01 C57BL-activatedt None t t ± 35 CBA T lymphocytes P < P < Precipitated with an- 46 i J± i 20 tiserum to immuno- ns ns globulin Centrifuged in 900 4± i i 40 normal rabbit serum P < ns "SI-labeled T lymphocytes were incubated for 3 hr in tissue culture medium. The cells were centrifuged, and the supernatants were retained. Untreated supernatants containing 125I-labeled cell-surface proteins or supernatants that had been precipitated with antisera to mouse immunoglobulin prepared in rabbits or normal rabbit serum were incubated with 1 X 107 allogeneic or syngeneic thymus cells. Background radioactivity has been subtracted from all values. Results represent the mean ±- standard error of data obtained in four separate experiments, 2 experiments with BALB-activated T lymphocytes and 2 experiments with C57BL-activated T lymphocytes; assays were done in triplicate. For determination of P values, all groups were compared to counts bound to CBA thymus cells. ns, not significant; nd, not done. * Number of counts added as high molecular weight protein = 5 X 106 cpm. t Number of counts added as high molecular weight protein = 7.5 X 104 cpm. observed in the latter phase. This was usually reached after 2-3 washes, although we routinely performed 5-6 washes to minimize counts that might be absorbed nonspecifically. As illustrated in Table 1, the ratio of 125J counts binding to allogeneic cells against syngeneic cells for surface proteins from BALB/c-activated CBA T lymphocytes was 4.6 when surface proteins were incubated with BALB/c cells and 2.1 when incubated with C57BL cells. Similarly, the ratio for surface proteins from C57BI-activated CBA T lymphocytes was 3.9 when tested with C57BL cells and 1.6 when tested with BALB/c cells. Although some crossreactivity between BALB/c and C57BL was observed, significantly more counts were bound to thymus cells obtained from the strain of mouse used to activate the T lymphocytes (P < 0.05). When cell-surface immunoglobulin was removed by specific precipitation before the assay, little or no radioactivity was bound to the target cells. Specificity was retained, however, if cell-surface proteins were treated with normal rabbit serum under conditions identical to those used above for depletion of immunoglobulins. Some decrease in the number of counts binding to the cells was observed with supernatants from material treated with normal rabbit serum. This decrease was probably due to nonspecific losses in surface immunoglobulin during the centrifugation procedure. In addition, normal rabbit serum may possess low titers of natural antibody to mouse immunoglobulin. Since the normal rabbit serum was not absorbed with mouse immunoglobulin, its presence during the binding assay may have interfered to some extent with the binding of immunoglobulin from T lymphocytes to the target cells. The number of counts specifically bound to the thymus cells (when corrected for losses due to washing) is equivalent to about 38-40% of counts precipitable as immunoglobulin. DISCUSSION Specifically activated T lymphocytes from thoracic ducts have a striking specificity in their ability to reject skin grafts, lyse target cells in vitro, and divide only in response to those antigens that originally provoked their formation (17). The proportion of these cells specifically reactive to the activating antigens is, as yet, unknown. Recent results have indicated that at least 20% of the cells will synthesize DNA specifically when exposed to the activating antigens for 24 hr in culture; although this figure is probably an underestimate (25). Thus it was not surprising that it was possible to demonstrate that a fraction of the surface proteins of activated T lymphocytes possessed antigen-binding specificity. The ability to remove the antigen-binding specificity of the surface proteins of T lymphocytes by precipitation with antisera to immunoglobulins indicates that the specific binding proteins were immunoglobulins. The fact that binding specificity is not abrogated by iodination of the molecule suggests that the tyrosine that is most accessible to our label may not reside in the active site of the immunoglobulin molecule. It is particularly significant that the antigen-binding specificity of cell-surface immunoglobulins from BALB/c-activated CBA T lymphocytes is crossreactive to a degree consistent with the number of H-2 antigens shared by BALB/c and C57BL mice (26). Considerable crossreactivity was also observed when activated T lymphocytes were tested for their capacity to lyse target cells in vitro (17). The isolation of IgM-type immunoglobulin molecules from the surface of thymus lymphocytes activated to H-2 antigens in vivo is consistent with previous studies that demonstrated the exclusive presence of monomeric IgM-type immunoglobulin on the surface of normal and in vitroactivated thymus cells (13). These results are also in accord

4 Proc. Nat. Acad. Sci. USA 69 (1972) with others who have demonstrated the presence of immunoglobulin determinants on the surface of T cells using radioactively labeled (27, 28) or virus-labeled (29) antisera to mouse immunoglobulin. The possibility that the IgM-type immunoglobulin was derived from a contaminating population of B lymphocytes must be considered in some detail. The population of activated thymus lymphocytes contained a maximum of 0.3% B lymphocytes (17). The amount of surface immunoglobulin isolated from T lymphocytes is similar to that amount isolated by metabolic turnover or acid-urea extraction from various T- and B-lymphocyte populations (12, 13). Therefore, the iodination conditions described here allow the isolation of comparable amounts of surface immunoglobulin from T cells and B cells (13). Since cell-surface immunoglobulin released from activated T cells is not cytophilic for syngeneic T or B lymphocytes (Table 1; and Feldmann, Cone, and Marchalonis, submitted for publication) the molecules could not have been absorbed from an exogenous source. In addition, secretion products of plasma cells would not be detected in this system (30). Taken together, these observations make it unlikely that the immunoglobulin studied was derived from a very small number (about 3 X 104) of B lymphocytes. The fact that the binding specificity of the isolated molecules parallels that shown by the activated cells (17) supports further the contention that these molecules are derived from T lymphocytes from thoracic ducts. Moreover, several quantitative arguments can be marshalled to discount the possibility that the specific immunoglobulin originated in contaminating B cells. The first relates to the level of sensitivity of the combined radioiodination-precipitation approach to detect immunoglobulin. Surface immunoglobulin from 107 B cells can readily be isolated and identified (13). Such cells possess about 105 molecules per cell (13), so the assay can detect 1012 molecules of immunoglobulin routinely. It might be possible to detect an order of magnitude fewer molecules with some difficulty, and we will take this level as our minimum threshold of detectability. Since it has been suggested that T cells contain fewer than 400 immunoglobulin molecules per cell (31), these cells would be negative in attempts to isolate surface immunoglobulin (32). The number of immunoglobulin molecules present in 107 thymus cells, as described here, would be the product of the number of cells (107) times the frequency of B cells (3 X 10-s) times the number of immunoglobulin molecules per B cell (105), which is 3 X 109 immunoglobulin molecules. This number falls nearly two orders of magnitude short of the minimum level of detectability. The second quantitative point to be discussed is that 40% of the detected immunoglobulin of T lymphocytes possessed binding specificity for the cells used in the activation process. This would require the highly unlikely situation in which close to 50% of the contaminating B cells showed the same specificity. In contrast, at least 20% of the T-lymphocyte populations used in this study can be specifically activated by the histocompatibility antigens used in the initial activation (25). The third factor stems from the possibility that B cells, in the absence of relatively large amounts of other B cells, may be stimulated to display inordinately large amounts of surface immunoglobulin. The above calculations indicate that such B cells, if they constitute only 0.3% of the cell population, must possess roughly 107 surface immunglobulin Activated T-Cell Surface Immunoglobulin 2559 molecules per cell. It is possible, by use of the dimensions of the IgG immunoglobulin molecules as determined by crystallographic methods (33), to calculate the maximum number of immunoglobulin molecules present on the surface of a lymphocyte. The IgG molecule is T-shaped with a horizontal length of 14.5 nm (145 X) (FAB2) and a height of 10.7 nm (107 A) (33). The immunoglobulin is probably bound to the cell surface via its Fc piece (12, 34). Therefore, in order for cell-surface immunoglobulin molecules not to collide with each other, each one would require a circle of area of diameter slightly greater than the FAB2 distance. This area is about 300 nm2 (3 X 104 A2). Assuming a spherical shape and a typical diameter of 15 um for a lymphocyte, the surface area of the cell computes to be about 6 X 108 nm2 (6 X 1010!k2). Thus, the maximum number of immunoglobulin molecules that will fit on one lymphocyte is in the range of 106. Furthermore, the calculation assumed that the lymphocyte was naked except for immunoglobulin molecules, whereas the fact of the matter is that lymphocytes possess large numbers of other surface molecules such as histocompatibility antigens and receptors for mitogens such as concanavalin A (>106 receptor per cell; ref 35). Based on analysis by Allan and Crumpton (36), lymphocyte membrane protein comprises % of the dry weight of the cell or about 2 to 3 X mg per cell. Assuming that IgM-type surface immunoglobulin has a molecular weight of 180,000 (12-14), 1 X 107 molecules would be about 3 X 10-9 mg of proteins, or 10-fold more protein than is present in the lymphocyte membrane. The approximate nature of these calculations must be admitted. However, based on spatial and mass considerations, it is evident that a super B cell would have to contain a number of surface immunoglobulin molecules that exceeds the theoretical limit. The mode of action of cell-surface immunoglobulin in cell-mediated immunity is unknown. It is probable that these molecules serve as the receptors that enable activated lymphocytes to recognize foreign antigenic determinants. We have suggested (18) that the union of antigen with receptor may retard or inhibit the release of cell-surface immunoglobulin. This event may cause conformational changes in membrane proteins that initiate cell differentiation. Antigen contact with activated cells, in contrast, might perpetuate the activated state and the synthesis and secretion of pharmacologically active molecules thought to be involved in cellular immunity (37). Moreover, contact between an activated cell and its target cell held in close juxtaposition by surface immunoglobulin receptors would facilitate the effect of pharmacologically active agents that might have a short radius of action. Since cell-surface immunoglobulin on activated and nonactivated cells is continually synthesized and released from the surface, it is reasonable to presume that the released immunoglobulin serves some immunological function. At the present time, this role is problematic. In 1900, Ehrlich postulated that antibody-forming cells recognized foreign antigen determinants by means of antibody-like receptors on the cell surface (38). The demonstration that antigen-specific immunoglobulin can be isolated from the surface of H-2 activated T lymphocytes, taken in conjunction with studies showing that the functional activity of T cells may be inhibited by anti-immunoglobulin reagents (4-7), provides experimental support for Ehrlich's prediction. Similar conclusions have been reached from studies on the

5 2560 Immunology: Cone et al. antigen-binding properties of immunoglobulin obtained from T cells activated to sheep erythrocyte antigens (Cone and Marchalonis, manuscript in preparation). It has yet to be proved that the antigen receptor site of nonactivated T lymphocytes is also immunoglobulin in nature. The fact that these cells, like activated cells, possess monomeric IgM-type immunoglobulin molecules on their surface makes this highly probable. It would seem necessary to obtain for study purified populations of nonactivated, specifically reactive T cells before nature of the receptor on resting T cells can be defined. NOTE ADDED IN PROOF Two recent studies [Dwyer, J. M., Warner, N. L. & Mackay, I. R. (1972) J. Immunol., 108, ; Hogg, N. M. & Greaves, M. F. (1972) Immunology, 22, ] provide evidence that the binding of antigen by mouse T lymphocytes is inhibited by treatment with antisera specific for, -chain antigenic determinants. We thank Miss Pat Smith, Miss Jenny Gamble, and Miss Helen Clark for excellent technical assistance. This work was supported by grants from the National Health and Medical Research Council of Australia, the Australian Research Grants Committee, the Damon Runyon Memorial Fund for Cancer Research, The British Heart Foundation, and the National Heart Foundation of Australia and the American Heart Association ( ). R. E. C. is a postdoctoral fellow of the Damon Runyon Memorial Fund for Cancer Research. 1. Warner, N. L., Byrt, P. & Ada, G. L. (1970) Nature 226, Feldmann, M. & Diener, E. (1971) Nature 281, Davie, J. M., Rosenthal, A. S. & Paul, W. E. (1971) J. Exp. Med. 134, Basten, A., Miller, J. F. A. P., Warner, N. L & Pye, J. (1941) Nature New Biol. 231, Mason, S. & Warner, N. L. (1970) J. Immunol. 104, Lesley, J. F., Kettman, R. & Dutton, R. W. (1971) J Exp. Med. 134, Greaves, M. F. & Hogg, N. W. (1971) "Cell Interactions and Receptor Antibodies in Immune Responses," in Third Sigrid Juselius Symposium, eds. Makela, O., Cross, A. & Kosunen, T. U. (Academic Press, Inc., New York), p Marchalonis, J. J. (1969) Biochem. J. 113, Marchalonis, J. J., Cone, R. E & Santer, V. (1971) Biochem. J. 124, Phillips, D. R. & Morrison, M. (1971) Biochemistry 10, Baur, S., Vitetta, E. S., Sherr, C. J., Schenkein, I. & Uhr, J. W. (1971) J. Immunol. 106, Marchalonis, J. J., Atwell, J. L. & Cone, R. E. (1972) Nature New Biol. 235, Proc. Nat. Acad. Sci. USA 69 (1972) 13. Marchalonis, J. J., Cone, R. E. & Atwell, J. L. (1972) J. Exp. Med. 135, Vitetta, E. S., Baur, S. & Uhr, J. W. (1971) J. Exp. Med. 134, Miller, J. F. A. P. & Mitchell, G. F. (1969) Transplant. Rev. 1, Mitchell, G. F. & Miller, J. F. A. P. (1968) Proc. Nat. Acad. Sci. USA 59, Sprent, J. & Miller, J. F. A. P. (1971) Nature New Biol. 234, Cone, R. E., Marchalonis, J. J. & Rolley, R. T. (1971) J. Exp. Med. 134, Edelman, G. M. & Marchalonis, J. J. (1967) in Methods in Immunology and Immunochemistry, eds. Williams, C. A. & Chase, M. W. (Academic Press, Inc., New York), Vol. 1, pp Parish, C R. & Marchalonis, J. J. (1970) Anal. Biochem. 34, Bailey, N. T. J. (1959) in Statistical Methods in Biology (English Universities Press, Ltd., London), pp Lerner, R. A., McConahey, P. J., Jansen, I. & Dixon, F. J. (1972) J. Exp. Med. 135, Vitetta, E & Uhr, J. W. (1972) J. Immunol. 108, Wilson, J. D., Nossal, G. J. V. & Lewis, H. (1972) Eur. J. Immunol., 2, Cheers, C. & Sprent, J. (1972) Fourth International Conference on Lymphatic Tissue and Germinal Centers in Immune Reactions, Dubrovnik, Yugoslavia, in press. 26. Snell, G. D. & Stimpfling, J. H. (1966) in Biology of the Laboratory Mouse, ed. Green, E. L. (McGraw Hill, New York), 2nd ed., pp Nossal, G. J. V., Warner, N. L., Lewis, H. & Sprent, J. (1972) J. Exp. Med. 135, Bankhurst, A. D., Warner, N. L. & Sprent, J. (1971) J. Exp. Med. 134, 100a Hammerling, V. & Rajewsky, K. (1971) Eur. J. Immunol. 1, Marchalonis, J. J., Cone, R. E., Atwell, J. L. & Rolley, R. T. (1972) in Biochemistry of Gene Expression in Higher Organisms, eds. Lee, J. W. & Pollak, J. K. (Australian and New Zealand Publishing Co. Sydney), in press. 31. Rabellino, E., Colon, S., Grey, H. M. & Unanue, E. R. (1971) J. Exp. Med. 133, llhr, J. W., in Genetic Control of Immune Responsiveness. Its Relationship to Disease Susceptibility, eds. 'I\cDevitt, H. 0. & Landy, M. (Academic Press, New York), in press. 33. Sarma, V. R., Silverton, E. W., Davies, D. R. & Terry, W. D. (1972) J. Biol. Chem. 246, Herd, Z. L. & Ada, G. L. (1969) Aust. J. Exp. Biol. Med. Sci. 47, Edelman, G. M. & Millette, C. F. (1971) Proc. Nat. Acad. Sci. USA 68, Allan, D. & Crumpton, M. J. (1970) Biochem. J. 120, Billingham, R. E. (1969) Mediators of Cellular Immunity. eds. Sherwood, H., Landy, L. & Landy, M. (Academic Press, New York), pp Ehrlich, P. (1900) Proc. Roy. Soc. Ser. B, 66,