Antibody–Antigen Interactions

 Binding of antigens has long been a central paradigm for molecular recognition. In addition, the biological nuances of antigen recognition have such profound ramifications for medicine that the study of antibody–antigen interactions remains a key branch of molecular immunology. In this entry we first describe the structural chemistry of antigen binding, then follow with aspects of antibody–antigen interaction that lead to unique biological phenomena.

1. Chemical Aspects of Antibody–Antigen Interaction

1.1.  General Properties 

The chemical interactions between antibody and antigen do not differ substantially from other protein–ligand interactions. Hydrogen bonds, van der Waals interactions, and sometimes salt bridges are used to form the antibody-antigen contact. Extremely close steric complementarity between antibody and antigen surfaces seems to be a common characteristic of interfaces (1). Gaps between the opposing antibody and antigen surfaces are sometimes filled by water molecules (2). The association constant for antibody–antigen interactions ranges from 10⁵ to 10¹² M^–1, with typical values for protein antigens around 10⁸–10⁹ M^–1 (3). Rate constants for binding low molecular weight haptens can be as high as 10⁸ M^–1s^–1. Reactions with macromolecular antigens are slower, ≤10⁶ M^–1s^–1, except for highly charged antigens, which sometimes enhance the rate through electrostatic interactions (4). In most cases, antigen binding does not cause easily observed changes in the binding site of an antibody. In some cases, conformational changes in the antibody are intrinsic to the binding mechanism. These changes can be caused subsequent to antigen binding, termed an “induced fit” mechanism (5); alternatively, antigen can bind selectively to one of several preexisting antibody conformations (6). Generalizations about the nature of the antibody–antigen interaction are hazardous, as is clear from the observation that the same antibody can bind one ligand through a few strong contacts and another ligand through many weak interactions (7).

1.2. Combining Site Structure and Function 

Comparative studies showed that antibody variability was concentrated in three short stretches of sequence within each variable domain (8). These regions were postulated to confer antigenic specificity on an antibody molecule, and were termed “complementarity-determining residues” (CDRs). X-ray crystallographic studies confirmed that the CDRs, although noncontiguous in the antibody primary structure, formed loops that were adjacent to each other in three-dimensional space (9). The position of CDRs within an IgG molecule is shown in Figure 1. Structural studies on antibody–antigen complexes have confirmed that the CDRs form the vast majority of intermolecular contacts. Non-CDR residues can also form contacts, however, and in some cases not all CDRs participate in the interaction with antigen (10). Of the six CDRs, heavy chain CDR3, formed at the genetic level by joining of a D segment to V_H and J_H, is the most structurally diverse, and usually the most energetically significant in binding antigen.

Figure 1. Position of CDRs (dark lines) within an IgG molecule.

1.3. Antigenicity 

Antibodies bind to structurally precise surfaces on protein antigens (11). These surface areas can be composed of a contiguous length of polypeptide chain or from several parts of a chain that are separate in primary structure but adjacent in three-dimensional space. The only true prerequisite for a portion of a protein to be recognized as an antigen is surface accessibility, although other factors, such as hydrophilicity and mobility, are often considered in predicting antigenicity from protein sequence (12). In the case of peptide–antibody complexes, about 7–10 peptide residues fit in the binding site of the antibody and form an ordered structure, even if the free peptide itself is not strongly ordered in solution (13). Sometimes a bound peptide structure resembles the conformation of the same sequence in an intact protein, a phenomenon that underlies the utility of peptide vaccines (14) .

 2. Biological Aspects of Antibody–Antigen Interaction

2.1. T-Dependent and T-Independent Immune Responses 

“Normal” immune responses depend on the participation of T cells. In outline, antibody-producing B cells capture antigen (protein or protein–hapten conjugate) on their surface, internalize and fragment the antigen, and represent short peptide fragments on the cell surface, embedded in the binding groove of major histocompatibility complex (MHC) molecules. T-cells recognize the peptide–MHC combination and produce signals that cause B cells to enter pathways of proliferation and differentiation (15). Molecular genetic processes activated at the immunoglobulin loci include heavy-chain class switching and somatic hypermutation, leading to production of soluble IgG, IgE, and IgA of high affinity. T-independent antigens include virtually all nonprotein macromolecules. B-cell proliferation and differentiation also occur in a T-independent response, but chain switching and somatic mutation are difficult to detect. Structural attributes of antibodies in a T-independent response are that they are of the IgM isotype, contain germline (unmutated) variable region sequences, and show low antigen affinity.

2.2. Maturation of the Immune Response 

Antibodies isolated soon after an initial antigen exposure, termed “primary response antibodies, ”differ from those obtained later in the response or after a second administration of antigen. This transformation of the antibody repertoire, which leads to progressive increases in the affinity for antigen (16), is termed “maturation of the immune response.” The structural basis of this phenomenon has been determined from studies of immune responses to haptens, and can be outlined as follows. Contact with antigen induces a process within lymphocytes that introduces point mutations in the variable regions of antibody genes (17). Most mutations have a neutral or deleterious effect on antigen affinity (18). However, some mutations improve the interaction with antigen. Higher affinity confers a selective advantage on lymphocytes that express this mutation, which may be competing with nearby lymphocytes for a limited amount of antigen. Selected lymphocytes proliferate and can undergo further rounds of mutation and selection. Maturation of the immune response leads to improved affinity universally in anti-hapten responses. The same progressive affinity increase is presumed to occur with protein antigens, but unequivocal evidence for this is lacking.

The diversity of the initial repertoire is determined by germline variable gene diversity and processes involved in gene rearrangement. At a structural level, antibody diversity is concentrated around the center of the antigen combining region. Somatic mutation acts to form the mature repertoire, which for protein antigens will show association constants in excess of 10⁸ M^–1 . Somatic point mutations can occur anywhere in the variable region, but those selected during maturation are in general located in a band immediately peripheral to the area of diversity expressed in the early repertoire (19) .

2.3. Anti-idiotype Antibodies 

“Idiotype” is an immunological word that corresponds in structural terms to the unique CDRs of an individual antibody. If an antibody (Ab1) is used to immunize an animal, its CDRs can be recognized as a foreign structure, even in an animal of the same species. An antibody that reacts with the CDRs of Ab1 will be made, termed an anti-idiotype antibody (Ab2). The anti-idiotype can be used in the same way to generate an anti-anti-idiotype (Ab3). In a theory of immune regulation (20), if Ab1 is complementary to an antigen and Ab2 is complementary to Ab1, then the binding site of

Ab2 should resemble the initiating epitope on the antigen. In vivo, Ab2 is then said to carry an “internal image” of the antigen. This chain of alternating complementary recognition properties is diagrammed in Figure 2). 

Figure 2. Chain of complementary binding activities within an idiotype network.

Structural studies have investigated the extent to which the internal image model is accurate. No gross structural similarity between an antigen and Ab2 is necessary for both molecules to bind to Ab1 with high affinity. However, the structure of an antilysozyme (Ab1) complexed with an Ab2 showed that many of the Ab1 residues used for binding lysozyme were also used to bind Ab2 (21). Furthermore, many Ab2 atoms contacting Ab1 occupied positions analogous to contact atoms in lysozyme, and intermolecular hydrogen bonds and even water molecules at the interface were conserved between the Ab1:lysozyme and Ab1:Ab2 complexes. Thus Ab2 can mimic an antigen through close chemical complementarity with Ab1, even in the absence of sequence or structural homology. In the case of Ab3 molecules, structural similarity to Ab1 is unequivocal, although not surprising. CDR sequences are often nearly identical between Ab1 and Ab3, and the structure of an antigen:Ab3 complex shows chemical complementarity as stringent as would be expected at the interface between antigen and an antibody raised directly (22).

2.4. Heterophile Antibodies and Molecular Mimicry 

An antibody that reacts with an antigen to which the host has not been exposed is termed a “heterophile antibody” (23). The origin of heterophile antibodies is obscure; no antigen has been unambiguously identified that in a natural situation causes heterophile antibodies to appear. Structurally, heterophile antibodies tend to be low affinity IgM that react with carbohydrate antigens. Some heterophile antibodies, termed “natural antibodies,” appear to arise spontaneously and persist for the lifetime of the organism. The heterophile antibodies that determine ABO blood group compatibility are of this type. In other cases, appearance of heterophile antibodies clearly results from infection. For example, infectious mononucleosis in humans reproducibly induces antibodies specific for an antigen on sheep red blood cells (24). The supposition in such cases is of a process of “molecular mimicry”: a pathogen-specified product is made during infection that possesses sufficient structural similarity to the heterophile antigen that heterophile antibodies are induced (25). When the reactivity induced is to a self-antigen normally subject to immune tolerance mechanisms, autoimmune disease may result. For example, streptococcal respiratory infection is the immediate antecedent of rheumatic fever, an autoimmune attack on the heart and vascular tissue. A carbohydrate structure on the bacterial surface is thought to mimic structural glycoproteins in heart valves and vessels (26). Experimental induction of autoimmune diseases by immunization supports this type of antibody–antigen interaction as an initiating mechanism for autoimmunity (27).

2.5. Polyreactive or Polyspecific Antibodies 

Polyreactive antibodies do not fit the standard paradigm of one antibody recognizing a single antigen. Instead, a single polyreactive antibody can recognize a panel of many antigens that bear no obvious chemical similarity to each other. Many polyreactive antibodies are IgM, and others are IgG and other isotypes and can bind antigens with high affinity. For example, a set of hybridomas screened for simultaneous reactivity with tubulin, actin, myosin, and single- and double-stranded DNA showed a median antibody affinity for all five antigens of 10⁷ M^–1 (28). The molecular basis of polyreactivity is unknown. Numerous V genes have appeared in polyreactive antibodies, and there is as yet no characteristic sequence motif that predicts polyreactivity. Nevertheless, sequence studies implicated heavy-chain CDR3 as a major determinant of polyreactivity (29). Structure–function studies comparing a polyreactive and monoreactive anti-insulin with similar sequences confirmed that polyreactivity was associated with heavy-chain CDR3 (30). This CDR loop is presumed to adopt multiple conformations to accommodate different antigens, but crystallographic data on the three-dimensional structure of polyreactive antibodies are lacking.

References

1. A. G. Amit, R. A. Mariuzza, S. E. V. Phillips, and R. J. Poljak (1986) Science 233, 747–753. 

2. T. N. Bhat, G. A. Bentley, G. Boulot, M. I. Green, D. Tello, W. Dall''Acqua, H. Souchon, F. P. Schwarz, R. A. Mariuzza, and R. J. Poljak (1994) Proc. Natl. Acad. Sci. USA 91, 1089–1093. 

3. I. Pecht (1982) "Dynamic aspects of antibody function", in The Antigens (M. Sela, ed.,( Academic Press, New York, Vol. 6, pp. 1–68. 

4. C. S. Raman, R. Jemmerson, B. T. Nall, and M. J. Allen (1992) Biochemistry 31, 10370–10379.

5. J. M. Rini, U. Schulze-Gahmen, and I. A. Wilson (1992) Science 255, 959–965. 

6. D. Lancet and I. Pecht (1976) Proc. Natl. Acad. Sci. USA 73, 3548–3553. 

7. W. Dall''Acqua, E. R. Goldman, E. Eisenstein, and R. A. Mariuzza (1996) Biochemistry 35 9667–9676- .

8. T. T. Wu and E. A. Kabat (1970) J. Exp. Med. 132, 211–250. 

9. R. J. Poljak, L. M. Amzel, H. P. Avey, B. L. Chen, R. P. Phizackerly, and F. Saul (1973) Proc. Natl. Acad. Sci. USA 70, 3305–3310. 

10. D. R. Davies, E. A. Padlan, and S. Sheriff (1990) Annu. Rev. Biochem. 59, 439–473. 

11. D. C. Benjamin, J. A. Berzofsky, I. J. East, F. R. N. Gurd, C. Hannum, S. J. Leach, E. Margoliash, J. G. Michael, A. Miller, E. M. Prager, M. Reichlin, E. E. Sercarz, S. J. Smith-Gill, P. E. Todd, and A. C. Wilson (1984) Annu. Rev. Immunol. 2, 67–101. 

12. J. A. Berzofsky (1985) Science 229, 932–940. 

13. R. Stanfield and I. A. Wilson (1995) Curr. Opin. Struct. Biol. 5, 103–113. 

14. R. A. Lerner (1982) Nature 299, 592–596. 

15. A. Lanzavecchia (1985) Nature 314, 537–539. 

16. H. N. Eisen and G. W. Siskind (1964) Biochemistry 3, 996–1008. 

17. M. G. Weigert, I. M. Cesari, S. J. Yonkovich, and M. Cohn (1970) Nature 228, 1045–1047. 

18. C. Chen, V. A. Roberts, and M. B. Rittenberg (1992) J. Exp. Med. 176, 855–866. 

19. I. M. Tomlinson, G. Walter, P. T. Jones, P. H. Dear, E. L. L. Sonnhammer, and G. Winter (1996) J. Mol. Biol. 256, 813–817. 

20. N. Jerne (1974) Ann. Inst. Pasteur (Immunol) 125C, 373–389. 

21. B. A. Fields, F. A. Goldbaum, X. Ysern, R. J. Poljak, and R. A. Mariuzza (1995) Nature 374 739–742-   .

22. K. C. Garcia, S. V. Desiderio, P. M. Ronco, P. J. Verroust, and L. M. Amzel (1992) Science 257,  528–531 .

23. J. Forssmann (1911) Biochem. Z. 37, 78–115. 

24. J. R. Paul and W. W. Bunnell (1932) Am. J. Med. Sci. 183, 90–104. 

25.  R. T. Damian (1964) Am. Naturalist 98, 129–149. 

26. J. B. Zabriskie, K. C. Hsu, and B. C. Seegal (1970) Clin. Exp. Immunol. 7, 147–159. 

27. R. S. Fujinami and M. B. A. Oldstone (1985) Science 230, 1043–1045. 

28. L. Diaw, C. Magnac, O. Pritsch, M. Buckle, P. M. Alzari, and G. Dighero (1997) J. Immunol. 158, 968–976.