CELLULARIM?vIIrNOLOGY 30, 82-91 (1977) The Inhibition of Murine by Human Lymphocyte Mitotic and Mouse Sera Responses 1. Evidence for a Role of Antibody-Independent Activation of the Alternative Complement Pathway DAVID EIDINGER, IGAL GERY,~ AND CAROLE ELLEMAN Department of Microbiology and Iuwwzology, Queen’s U+tiver&, ICingston, Ontario, Canada Received November 12,1976 The responses of mouse lymphoid cell cultures to mitogens such as concanavalin A or antigen were inhibited by the addition of small amounts of fresh human serum. This inhibitory effect was reduced by specific decomplementation procedures such as heating at 50°C to inactivate factor B or absorption with zymosan at 17°C to deplete properdin from the serum. Human factor-B preparations reconstituted the inhibitory effect lost from human serum preparations heated at 50°C. These findings are interpreted to indicate a fundamental role of activation of the alternative complement pathway (ACP) as an underlying mechanism of inhibition. Additional experiments designed to demonstrate a role of natural antibodies activating the classical complement pathway, while successful in these respects, also provided confirmatory evidence for an antibody-independent role of the ACP. Furthermore, data derived from experiments utilizing mouse sera tested on syngeneic mouse lymphoid target cells, were qualitatively similar to the results obtained employing human sera. The data suggest that the functional activities of mouse lymphocytes in vitro are inhibited by antibody-independent activation of the ACP, implying that this pathway may exercise a role in regulating lymphocyte function. INTRODUCTION In previous papers, it was shown that human serum was cytotoxic for murine target cells.zf 3 The data indicated that this effect was mediated by a combination of antibody-independent activation of the ACP and by natural antibodies activating the classical complement pathway (CCP) . In initial experiments, it was also shown that human serum inhibited the mitogenic effect of concanavalin A (Con A) in cultures of mouse thymus and spleen cells. The capability of human serum to inhibit lymphocyte mitogenesis 1 On sabbatical leave from the Department of Medical Ecology, The Hebrew UniversityHadassah Medical School, Jerusalem, Israel. 2 Eidinger, D., Belle, E., and Mates, A., The heterocytotoxicity of human serum. I. Activation of the alternative complement pathway by heterologous target cells. Cell. ZmnzunoZ. 29, 1977, 174-186. s Eidinger, D., Gery, I., and Belle, E., The heterocytotoxicity of human serum. II. Role of natural antibody and of the classical and alternative complement pathways. Cell. Zmmzlnol. 29, 1977, 187-194. 82 Copyright 0 1977 by Academic Press, Inc. All rights of reproduction in any form reserved. ISSN 0008-8749 CELLULAR ACTIVATION OF ALTERNATIVE C PATHWAY 83 raised several questions which provided a basis for the present work. First, was human serum capable of influencing lymphocyte responses to other antigens and mitogens in vitro? Second, to what extent did the effects of human serum reflect homeostatic mechanisms, whereby mouse lymphoid cells bathed in an environment containing syngeneic mouse serum exhibit altered functional activity caused by one or both of the aforementioned mechanisms? MATERIALS AND METHODS Sera. Human blood samples were obtained from adult normal donors, hypoglobulinemic patients, or umbilical cords of newborn infants. Sera were prepared and treated by heat (50 or 56°C) or absorbed by zymosan (at 17’C) or mouse lymphoid cells (at 0’) as described elsewhere.” 3 Fetal calf serum (FCS) was purchased from Grand Island Biological Co., Grand Island, N.Y. and was heatinactivated before use. Mice and inzmunizations. All mice were supplied by Jackson Laboratory, Bar Harbor, Maine. Spleen and lymph node cells from CBA/J males were used throughout as the source of ‘untreated’ lymphoid cells. Immunization against keyhole limpet hemocyanin (KLH; Mann Research Laboratories, N.Y.) was carried out in CBA/J mice or in female B6DZ/J mice 12-14 weeks of age. Each mouse was injected in four footpads with a total of 0.05 ml of emulsion made of equal volumes of KLH, 1 mg/ml in saline, and complete Freund’s adjuvant, (Difco Laboratories, Detroit, Mich.) enriched with killed lyophilized Mycobacterium butyricunz (Difco Laboratories, 2 mg/ml of adjuvant). Axillary, popliteal, and inguinal lymph nodes were collected 1 week after immunization and were pooled before use. Mitogens. Stimulation of the cell cultures was induced by adding various concentrations of the following: (a) Con A (Pharmacia, Uppsala) ; (b) Phytohemagglutinin (PHA ; Difco Laboratories) ; (c) Lipopolysaccharide W from Eschcricia coli 055 : BS (LPS ; Difco Laboratories) ; and (d) KLH. Cell cultures. Spleen or lymph node cell suspensions were prepared by passing the organs through stainless steel mesh. The cell suspensions were spun down and washed once. Cultures were set in 12 X 75mm plastic tubes (Falcon Plastics, Oxnard, Calif.,) and consisted of 2 X lo6 nucleated cells in 1 ml of RPM1 1640 (Grand Island Biological CO.). Sera were added as described in each experiment. Stimulation of the cultures was induced by adding different concentrations of mitogen. The cultures were incubated at 37°C with 5% CO2 in air for 50 (responses to Con A) or 74 hr (responses to KLH) and were pulsed with 1 PC1 of [“HI thymidine (New England Nuclear, Boston, Mass. ; 6.7 Ci/mmol) during the last 6 hr of incubation. Harvesting was carried out by filtering the cultures through glass fiber filters (GF/A) as described by Ron et al. (1). The results are recorded as mean counts/minute values, which generally varied from the individual values by less than 10%. Reconstitution experiments with factor B. Functionally active preparations of human factor B were prepared according to the method of Gotze and MullerEberhard (2). The capability of these fractions to restore the hemolytic activity of 50°C human serum for rabbit erythrocytes suspended in EGTA-Mg++ (3) 84 EIDINGER, GERY HUMAN Con A dose Cells (Ng/ml). tested. 16 4 I 025 AND ELLEMAN 16 l--l 25 16 LYMPH SPLEEN MOUSE HUMAN MOUSE 1 02 NODES FIG. 1. The effect of 2% human and mouse sera on CBA mouse spleen and lymph node cultures stimulated with varying doses of Con A. The bars denote the percentage response of cultures containing fresh untreated serum versus equivalent cultures containing 56°C heatinactivated serum at the same concentration. The counts/minute of control cultures as a measure of incorporation of [“Hlthymidine was expressed as 100%. was employed as an assay of the activity of the isolated fractions. In practice, considerable variation was noted in the ability of heating at 50°C to abolish the lytic activity of numerous samples of adult human sera for nonsensitized rabbit erythrocytes. Only those sera with trace residual activity were utilized; in those instances, the addition of from 15-30 pg of the factor-B preparation (depending on the preparation) to 0.1 ml of 50°C serum reconstituted the activity of 50°C serum to 1 : 4 or 1 : 8, which represented 25-50s of the activity in fresh adult human serum. The same principle was employed in the present work, namely, the assay of the restoration of the inhibitory activity of 50°C human serum by the addition of 10-30 pg of factor-B preparation to each of the target cell cultures, The experience with the restorative effect of human factor B for human serum served as an example for similar experiments employing isolated mouse factor-B preparations added to 50°C mouse sera. Although the same technique of isolation was utilized (2), no further experiments were performed to characterize further the factor-B activity of the isolated mouse fractions. However, the restorative effect of both the human and mouse factor-B preparations was abolished by heating the isolated fractions at 50°C for 20 min. RESULTS Comparative Effect of Hzman and Mouse Sera on the Mitotic Response to Different Doses of Con A A study was undertaken of the effect of varying dilutions of human and mouse sera on the mitotic response of spleen and lymph node cell cultures containing varying doses of Con A. Figure 1 illustrates a representative experiment denoting mitotic responses in the presence of 2% fresh untreated serum versus responses in 2% serum heated to 56”C, representative of control cultures and expressed as 85 SERUM 70 1 adull hypogamma TESTED adult absof card II 8 4 FIG. 2. The effect of various dilutions of human serum naturally or experimentally deprived of antibodies on the Con A response. The bars denote the percentage (counts/minute) of cultures containing fresh serum versus the response measured in counts/minute of controls containing equivalent dilutions of 56°C heated serum. 100% response. It may be seen that the sera of both species inhibited the response in cultures containing the supraoptimal dose of Con A. The addition of 2% human serum also inhibited cultures containing suboptimal levels of the mitogen, an effect not observed in the presence of 25% mouse serum. The response of cultures containing increasing concentrations of human serum on the mitotic response of cells stimulated by various levels of Con A will be described. Under these conditions, cultures containing optimal as well as suboptimal levels of Con A were also inhibited, although not to the degree noted in cultures containing supraoptimal levies. Role of Natzml Antibodies and CCP in Hz~mn Scrzr~n In these experiments, the inhibitory effects of human sera, either naturally or experimentally deprived of antibody, on the Con A response of murine spleen cells were compared. Hypoglobulinemic serum samples, obtained from a patient and shown to contain less than 1% of the normal levels of immu~ioglobulins, and cord blood samples, possessing normal IgG levels, less than 10% of adult IgM levels, and totally deficient in other immunoglobulin classes (3), were studied. Normal adult human serum samples were absorbed with the murine lymphoid target cells prior to testing to remove antibodies. Figure 2 illustrates the data obtained by testing varying dilutions of the human samples. It may be seen that the hypoglobulinemic sera, as well as the cord sera, inhibited the mitotic response to Con A. particularly at the highest serum concentrations tested. In addition, fresh adult sera, absorbed on target cells prior to testing, exhibited inhibition approximately equivalent to that achieved with the naturally occurring antibody-deprived sera. However, the levels of inhibition at the higher serum dilutions were less severe than at equivalent dilutions of fresh unabsorbed normal adult sera. These data 86 EIDINGER, GERY AND ELLEMAN indicated a role of natural antibodies as well as a major role of antibody-independent activation of the ACP in human serum. The activity of the ACP was tested further in experiments which have been employed to deprive adult human sera of either properdin or factor B by zymosan absorption at 17°C (5, 6) or by heating at 50°C (2, 7, 8) as previously described.2* 3 Figure 3 illustrates that these decomplementation procedures markedly reduced the inhibition of the Con A response, particularly at serum dilutions of 4 and 2%, respectively. The persistent inhibition employing 8% ACP-deprived serum is compatible with the view that the natural antibodies in the human sera activating the CCP were responsible for the residual inhibition. Similar experiments (not shown) have also demonstrated the sensitivity of the inhibitory effect of hypoglobulinemic and cord blood sera to these decomplementation procedures. The role of the ACP was studied further in experiments designed to test the reconstitutive activity of functionally purified factor-B preparations on the inhibitory capability of 50°C serum. The data are presented in Table 1. Pretreatment of the serum at 50°C markedly reduced the inhibitory capability of fresh untreated serum, as expected, whereas the addition of the factor-B preparation to 50°C serum induced a marked restoration of the inhibitory effect. This reconstitutive effect was abolished by pretreatment of the factor-B preparation at 50°C for 20 min. Reconstitzttive Seruwh Efect of Mouse Factor B on the Inhibitory Activity of Mouse In preliminary experiments, it was shown that the inhibitory effect of mouse serum was markedly reduced by pretreatment of the serum samples at 50°C. These results implied that mouse behaved in a manner equivalent to that of human serum with respect to heat-lability, which, in the case of human serum, was shown to be due to the heat-labile factor B. Therefore, experiments designed to demonstrate the requirement for mouse factor B and for activation of the ACP in mouse serum were undertaken. These experiments were modeled on those performed with THEATMENT OF SERUM 50°C 1- - r 2 FIG. 3. The effect of decompletnentation procedures on the inhibition of the Con A response of CBA spleen cells. The results are expressed as a percentage of controls, as before. CELLULAR ACTIVATION OF ALTERNATIVE TABLE Reconstitutive 87 C PATHWAY 1 Effect of 16 rg of Human Factor B on the Inhibition of the Mitogenic Con A in Spleen Cell Cultures Containing 5% Human Serum Con A addeda Serum tested 20 e/ml None Active 5O’C 56’C 5O’C 50°C 56’C (fresh) (heated) (heated) + factor + factor + factor Response to B B* B 364 1,845 5,464 482 2,791 5,881 (7) (34) (100) (9) (51) (108) 160 23,452 71,410 9,350 27,487 74,889 (0) (33) (100) (13) (38) (105) Q Data are expressed in counts per minute, incorporation of C3H]thymidine. Numbers theses denote percentage of control cultures, incorporation in 56°C heated serum. * Factor B was heated at 50°C for 20 min prior to addition to cell cultures. in paren- B, employing mouse factor-B preparations prepared in a manner factor analogous to that for the human material (2). The data of a representative experiment are presented in Table 2. It may be seen that mouse serum appeared equivalent to human serum in that the inhibitory effect, lost by heating at 5O”C, was restored by the addition of small amounts of a presumptive mouse factor-B preparation. human The Efject of Human and Moztse Sera on the Mitotic Mitogens Resjonse to Alatigen and In order to confirm that the observations just presented did not represent an accidental occurrence of an unusual susceptibility of Con A-stimulated cells to activate the ACP, a study was undertaken of the response of mouse lymph node cells taken from donors immunized with the antigen KLH. The design of the experiment, which was to test the effect of human serum on the mitotic response of antigen, was similar to that employed in the Con A system. Table 3 illustrates the data employing various doses of KLH in the cell cultures. It may be seen that human serum inhibited the mitotic response in the presence of TABLE Reconstitutive 2 Effect of 20 pg of Mouse Factor B on the Inhibition of the Mitogenic Con A in Spleen Cell Cultures Containing 2y0 Mouse Serum Serum tested Con A added5 None Active 5O’C 56’C 50°C 56’C (fresh) (heated) (heated) + factor + factor Response to B Bb 1,675 2,053 1,873 1,020 3,893 (89) (110) (100) (54) (208) 10 rglml 2,521 13,361 14,299 2,762 21,757 (18) (93) (100) (19) (152) a Data are expressed in counts per minute, incorporation of [3H]thymidine. Numbers theses denote percentage of control cultures, incorporation in .56OC heated serum. * Mouse factor-B preparations frequently stimulated control cultures. in Ijareil- 88 EIDINGER, GERY AND TABLE Restorative to KLH Antigen None (fresh) (heated) (heated) + factor + factor B B 3 Effect of 16 /~g of Human Factor B on the Inhibition of Mitogenesis in Immune Lymph Node Cultures Containing 4oj, Human Serum Serum tested Active WC 56°C 50°C 56T ELLEMAN 244 3,164 10,237 1,037 9,068 concentration 10 (2.4)b (31) (100) (10) (86) 278 30,045 52,948 9,540 37,024 (pg/ml)a 1 (0.5) (57) (100) (18) (70) 275 24,066 58,741 3,153 43,335 a Data are expressed in counts per minute, incorporation * Numbers in parentheses denote percentage of control serum. 0.1 (0.5) (41) (100) (5) (74) 287 13,494 46,415 1,554 30,904 0.01 (0.6) (29) (100) (3) (67) 283 5,460 27,186 1,013 17,647 of [3H]thymidine. cultures, incorporation (1) (20) (100) (4) (65) in 56’C heated and that this inhibition was substantially reduced by pretreatment of the serum at 50°C and, furthermore, was restored by addition of purified factor-B preparations to the heated samples. Table 4 summarizes the results of experiments designed to test the effect of human and mouse sera on responses of cell cultures to PHA and LPS. Fresh sera inhibited the incorporation of 1”Hlthynncline by the target cells, an effect which was markedly reduced by heating the sera at 50 or 56°C. In other experiments (not shown), absorption of the sera with mouse target cells at 0°C reduced, but did not abolish, the inhibitory effect, confirming the antibody-independence of this activity. The results confirm that the effects denoted on the Con A system are not unique, but probably represent typical effects in cell culture employing bone marrow-derived (B) , as well as thymus-derived (T) , lymphoid cells as targets. antigen DISCUSSION The results of the present work indicated that the mitotic responses of lymphoid cells in vitro to Con A, employing incorporation of tritiated thymidine as a measure TABLE 4 Effect of Human and Mouse Serum on the Inhibition of the Mitogenic Responses of Mouse Spleen Cells to PHA and LPS Human Mitogen employed Serum concentration (70) PHA (1 Pgllnl) LPS (50 /%/~~~I) y Data 8 4 8 4 origin Mouse origin Serum tested Active 184a 4,196 8,733 4,633 5OT 56T 48,796 41,708 58,350 40,813 43,174 40,469 50,720 38,355 expressed as counts per minute of ~H]thymidine Serum concentration (%I 4 2 5 2 incorporated. Serum tested Active 1,951 5,194 4,150 11,641 50°C 56°C 3,222 6,168 18,848 18,776 4,202 6,510 20,257 20,195 CELLULAR ACTIVATION OF ALTERNATIVE C PATHWAY 89 of DNA synthesis, were inhibited in the presence of fresh untreated human sera. These inhibitory effects were markedly reduced by absorption of the sera with zymosan at 17°C or by heating at 50°C prior to testing. The loss of activity by pretreatment of human sera at 50°C was restored by the addition of functionally purified fractions of human factor B to the culture systems. Similar results were obtained in spleen cell cultures stimulated by the mitogens, PHA and LPS, and in cultures of immune lymph node cells containing the antigen KLH. These data are thus interpreted to indicate that a primary mechanism mediating the inhibitory effects of human sera is activation of the ACP in the serum by target cells. The persistent activity of normal adult human sera following prior absorption versus target cells, of naturally occurring hypoglobulinemic samples derived from an adult patient with marked hypoglobulinemia, and of numerous samples of cord sera pointed to the antibody-independence of this mechanism. However, a role of natural antibody in the human serum was also documented, since prior absorption of the test sera on mouse lymphoid target cells regularly reduced the level of inhibition. Since all inhibitory activity was abolished by heating the sera at 56”C, it is evident that these natural antibodies require an intact CCP to mediate their effects. These data confirm those reported for short-term cultures of murine tumor and lymphoid cells, employing incorporation of radiolabeled uridine as an indicator of metabolic activity.3 The data derived from experiments employing mouse sera tested on syngeneic target cells were similar to the data derived from the use of human serum reagents. Fresh mouse serum inhibited the mitogenic response in cultures containing supraoptimal levels of Con A. Pretreatment of the mouse serum by heating it at 50°C also reduced the inhibitory effect, which was restored by the addition of a presumptive mouse factor-B preparation. These data provide strong evidence for the activation of the ACP in mouse sera by stimulated syngeneic target cells in vitro. More formal proof must await the isolation of purified mouse factor B and the delineation of its role in the activation of the ACP in mouse serum, which, hitherto has not been carried out. No evidence for a role of autoantibodies in mouse sera versus T lymphocytes or versus hidden determinants on lymphoid cells of a type which has been described in the literature was demonstrable (9, 10). I n unpublished work, it was shown that absorption of the mouse sera on syngeneic target cells, syngeneic target cells stimulated to blast transformation with Con A, or normal cells pretreated with neuraminidase to expose hidden determinants (10) all failed to abolish the inhibitory effect of syngeneic mouse sera. Moreover, the complete susceptibility of the inhibitory effect of mouse sera to heating at 50°C (Table 2) suggests that the CCP and natural antibodies are not involved in this inhibitory system. Several groups of investigators have established that mouse serum contains an immunosuppressive factor(s) which has been variously determined to be present in either or both of the (Y- and y-globulin regions (11-13). However, the suppressive activity mediated by activation of the ACP has not been demonstrated previously. One reason for this may be that either heat-inactivated or fresh sera was used exclusively in the work of other investigators; comparative studies employing both kinds of sera do not appear to have been carried out. Much has been written about the role of coml)lement in humoral immunity. It has been suggested that complement activation via the C3 receptor on B lympho- 90 EIDINGER, GERY AND ELLEMAN cytes represents the second signal required for antibody production ( 14). However, several investigators have now confirmed that immunogenicity, mitogenicity for B cells, and C3 activation are not correlated properties for the thymus-independent family of antigens (15, 16). In addition, C3 has been shown not to be essential for humoral immunostimulation in vitro to the thymus-dependent antigen, sheep erythrocytes (17). However, depletion of complement in tivo by injections of cobra venom factor is associated with immunosuppression to sheep erythrocytes (IS), a result difficult to reconcile with in vitro data (17). Moreover, injections of cobra venom factor do not alter the induction of delayed hypersensitivity to a contact allergen ( 19). The presence of antibodies in a test system would appear to complicate matters further. Antibody versus a cell membrane receptor antigen in the presence of an active ACP acts as a growth stimulant, at least for mouse L cells (20). In addition, antibody of the IgG class to viral-induced membrane antigens appears to activate preferentially the ACP (21). These results imply that, in any analysis of the role of the activation of the ACP, it is essential to examine whether or not antibodies are present. However, the results of the present work demonstrate that activation of the ACP in human and mouse sera can take place via an antibodyindependent activation on target cell membranes of both T and B lymphocytes. It is also evident that this activation process is able to suppress cell metabolic activity and function.* These results imply that activation of the ACP occurs as a consequence and is an indicator of a fundamental change in the cell membrane in stimulated T and B lymphocytes. Numerous changes in lymphoid target cells have been demonstrated in vitro as a consequence of mitogenic activity or immune function. These changes include the appearance of new antigens (22, 23)) which may, in turn, trigger ACP activation. This notion is supported by the finding that transformed autologous lymphocytes may activate the ACP system of human serum (24, 25). If we extrapolate the findings observed in vitro to an in viva situation, it is reasonable to expect that cell homeostasis for resting cells does not induce activation of the ACP, whereas activated cells undergo the necessary membrane changes which will induce activation of this pathway. Thus, the results of the present work imply that the existence of this activity in sera may not only affect lymphocyte function but may also regulate it. ACKNOWLEDGMENTS The authors acknowledge the excellent technical assistanceof Mrs. Helen Derry and Miss Pauline Copleston. This work was supported by grants from the Medical Research Council, Ontario Heart Foundation, and National Cancer Institute of Canada. REFERENCES 1. Ron, N., Laufer, A., and Gery, I., Immunology 25, 433, 1974. 2. Gotze, O., and Muller-Eberhard, H. J., J. Exp. Med. 134, 9Os, 1971. 3. Platts-Mills, T. A. E., and Ishizaka, K., J. Zmmunol. 113, 348, 1974. 4. Dent, P. B., Finkel, A., Iturzaeta, N., Crent, M., and Bienenstock, J., Canad. Med. Asso. J. 106, 889, 1972. A. C., 5. Pillemer, L., Blum, L., Lepow, I. H., ROSS, 0. A., Todd, E. W., and Wardlaw, Science 120, 279, 1954. CELLULAR ACTIVATION OF ALTERNATIVE C PATHWAY 6. Pillemer, L., Schoenberg, M. D., Blum, L., and Wurz, L., Sciezcc 122, 545, 1955. 7. Boenisch, T., and Alper, C. A., Bioclzirlr. Biopky.~ i2ctn 221. 529, 1970. 8. Goodkofsky, I., and Lepow, I. H., J. IUIWW~~ 107, 1200, 1971. 9. Schlesinger, M., Nature (London) 207, 429, 1965. 10. Rogentine, G. N., J. Nat. Co~~rr Inst. 54, 1307, 1975. 11. Occhino, J. C., Glasgow, A. H., Cooperbrand, S. R., Mannick, J. A., and Schmid, J. Imwwaol. 91 D., 110, 68.5, 1973. 12. Veit, B., and Michael, J. G., J. I~r~nuzol. 111, 341, 1973. 13. Whitney, R. B., and Levy, J. G., J. Nat. Cancer Inst. S4, 733, 1975. 14. 15. 16. 17. Dukor, P., and Hartmann, K. U., Cell. ZI~~~III~~~OZ. 7, 349, 1973. Pryjma, J., Humphrey, J. H., and Klaus, G. G. B., Nature (Lolzdolz) 252, 505, 1974. Diamantstein, T., and Blitstein-Willinger, E., I?rzljlzlnologJ 29, 1087, 1975. Waldmann, H., and Lachmann, P. J., Ezfr. J. ~vIvI~~?~~. 5, 185, 1975. 18. Pepys, M. B., Nature (London) 237, 157, 1972. 19. Cochrane, C. G., Muller-Eberhard, H. J., and Aikin, B. S., J. Inr~m~m~l. 105, 55, 1970. 20. Shearer, W. T., Atkinson, J. P., Frank, M. M., and Parker, C. W., J. ExP. Med. 141, 736, 1975. 21. 22. 23. 24. 25. B. S., Cooper, N. R., and Oldstone, M. B. A., J. E.Q. Med. 141, 761, 1975. Weksler, M. E., J. E.Q. Med. 137, 799, 1973. Miller, J., Lifton, J., and Hattler, B. G., Cell. 1?~)1~t~no]. 19, 306, 1975. Budzko, D. B., Lachmann, P. J., and McConnell, I., Cell. I~r~~m~~ol. 22, 98, 1976. Okada, H., and Baba, T., Nature (London) 248, 521, 1974. Joseph,