抗原抗体结合

Antigen –Antibody Binding

Izumi Kumagai, Tohoku University, Sendai, Japan Kouhei Tsumoto, Tohoku University, Sendai, Japan

Antibodies are a family of glycoproteins that bind specifically to foreign molecules

(antigens).The binding between antibodies and antigens has high specificity and affinity resulting from various structural and energetic aspects.

Introductory article

. Introduction

. Antigen –Antibody Binding . Antigenic Epitope

. Forces Involved in Antigen-binding . Affinity and Avidity

. Biological Significance of Antibody Affinity and

Multivalency

Introduction

Antibodies (immunoglobulins)are produced by the immune system of vertebrates and are essential for the prevention and resolution of infection by foreign invaders such as viruses. Antibodies are a family of variable glycoproteins that bind specificallyto foreign molecules (antigens).The most striking feature of antigen–antibodyinteractions is their high specificityand affinity.In this article, the structural and energetic aspects of antigen–antibody binding are described, focusing in particular on howan antibody specificallyrecognizes its cognate antigen and binds to it tightly.

A binding strength between an antigenic determinant in an antigen (epitope)and an antigen-binding site in an antibody (paratope)is termed affinity.Each antibody unit has at least two antigen-binding sites, and is therefore bivalent, or multivalent, to its antigen. The functional combining strength of an antibody with its antigen, which is related to both the affinityof the reaction and the valencies of the antibody, is termed avidity. The signifi-cance of avidity to antigen–antibodybinding is also described.

Antigen –Antibody Binding

Antibody structure

The basic structure of an antibody (immunoglobulin)molecule comprises two identical light chains and two identical heavy chains linked together by disulfidebonds. There are fiveclasses of immunoglobulins (IgG,IgA, IgM, IgD and IgE), which differin amino acid sequence and number of domains in the constant regions of the heavy chains. There are two differentisotypes of light chains (l and k ). Immunoglobulin G (IgG)is the major type of immunoglobulin in normal serum and the most extensively investigated (Figure 1). The remarkable feature of the antibody molecule is revealed by comparison of amino acid sequences from various immunoglobulin molecules. This shows that immunoglobulins are composed of various copies of a folding unit of about 100amino acids, each of which forms an independent similar structure called the

immunoglobulin fold (Figure 2). The N-terminal domain of each polypeptide (heavyand light chains) is highly variable, while the remaining domains have constant sequences. The former domain is called the variable region (Vregion), while the latter is the constant region (Cregion). In addition, a comparison of V region sequences shows that variability is not uniformly distributed but concen-trated into three areas called the hypervariable regions. Investigations into the structures of various antigen–antibody complexes have demonstrated that the domain structure of the antibody molecule is a b barrel consisting of nine antiparallel b strands (Vregions) and seven C regions, and that the hypervariable regions are clustered at the end of the variable domain arms (Figure 3). The antigen-combining site of antibodies is formed almost entirely by six polypeptide segments, three from light variable domains and three from heavy variable domains. These segments showvariability in sequence as w ell as in number of residues, and it is this variability that provides the basis for the diversity found in the binding characteristics of the differentantibodies. These six hypervariable segments are often referred to as the complementarity-determining regions or CDRs.

It was realized that a more constant sequence of residues outside of the CDRs was required to maintain the essential immunoglobulin fold that results in the CDRs being brought into three-dimensional proximity. These residues are referred to as the framework regions. The framework residues do not usually form bonds with the antigen. However, they are essential for producing the folding of the V domains and maintaining the integrity of the binding site. Thus, the antibody-binding sites are formed by six segments of variable structure (CDRs)supported by a scaffoldingof essentially invariant architecture (frame-work regions). This characteristic structure has led to several approaches to the artificialdesign of novel antibodies by grafting newCDRs onto existing antibodies (Figure 3).

Canonical structure of CDRs

The antigen-binding specificityof an antibody is definedby the physical and chemical properties of its CDR surface. These in turn are determined by the conformation of

the

Figure 1Structure of immunoglobulin G (IgG).(a)Schematic representation of a typical IgG. The L and H chains are shown as solid lines, with the intramolecular disulfide linking Cys residues (S-S),that are characteristic of each immunoglobulin domain as shown in Figure 2. (b)Three-dimensional structure of IgG. The drawing is a composite of the crystal structures of human IgG Fab (PDBentry 21G2), and of human IgG Fc (PDBentry 1FC2). Some of the residues connecting Fab with Fc have not been visualized in the crystal structure; their probable location is indicated by the unconnected circles. Drawn using

MOLSCRIPT.

Figure 2The immunoglobulin fold. The variable domain of a light chain is shown. Drawn using

MOLSCRIPT.

individual CDRs, by the relative disposition of the CDRs, and by the nature and disposition of the side-chains of the amino acids in the CDRs. The structure of antibodies shows that antibodies can recognize infinitelyvariable

antigens by varying the amino acid sequences of the CDR loops (i.e.the surface structures of antigen-recognizing regions). While the structure of the CDR loops might vary randomly, there are certain preferred conformations. Such preferred conformations can be deduced from the lengths and sequences of these regions.

CDR loop conformations in the antigen–antibodycomplexes whose three-dimensional structures have been solved by X-ray crystallographic study have been exten-sively analysed. Examination of the sequences of variable domains of unknown structure has shown that many have CDR loops that are similar in size to those of one of the known structures and contain identical residues at the sites responsible for the observed conformation. For fiveof the six hypervariable regions of most immunoglobulins (theonly exception is CDR of heavy chain 3), there seems to be only a small repertoire of main-chain conformations, most of which are known from the set of immunoglobulin structures so far determined. These observations have interesting implications for the molecular mechanism involved in the generation of antibody diversity, since the combination of limited numbers of CDR loop structure creates infinitespecificitytoward foreign molecules. Sequence variations within the hypervariable regions modulate the surface that these canonical structures present to the antigens, altering the specificityand affinityof the antibodies. Sequence variations within both

the

residues (Tyr,Trp and Asn) seem to have a propensity for being in the CDRs, and for participating in antigen recognition. It seems that the aromatic side-chains (Tyr,Trp) are more exposed to the solvent than in usual water-soluble proteins, and they are frequently found to be involved in the interaction with the ligand. This is explained by their large size (hydrophobiceffect),large polarization (vander Waals interactions), ability to form hydrogen bonds, and rigidity (lessloss of conformational entropy upon complexation). Thus, the concentration of aromatic rings would give a certain ‘stickiness’to the CDRs and give diverse specificityto antibodies. Specificityfor a particular antigen would arise from the complemen-tarity of the shapes of the interacting surfaces created by the proper positioning of the aromatic rings and the correct location of polar and/orcharged groups.

Role of structural changes upon complex formation in antigen –antibody binding

On comparison of an antigen structure and the antibody with its complex, local conformational changes (oftencalled induced fitting)of both molecules have often been observed, creating high specificityand affinity.It can be supposed that induced fittingof an antibody to its antigen is critical for high specificityand affinity.Induced fittingcan be achieved (1)by small movements of side-chains, (2)by structural modificationssuch as deformation of CDR loops, or (3)by a change in the relative orientation of variable domains. All of these also seem to play a critical role in antigen–antibodyinteraction. In some antigen–antibody complexes, the significanceof (3)has been suggested. As for (1)and (2),although the flexibilityof the binding site leads to entropic loss upon complex formation, a greater interaction due to a more precise fittingmust result in an overall increase in binding energy. Favourable energy change is in part or completely compensated for by unfavourable entropic loss (seebelow). Therefore, the binding characteristics must be analysed structurally by both antigen-free and complex form.

It is known that diversity in the germline antibody population is generated by the association of V, D and J gene segments with additional diversity occurring at the joining regions (Figure 4). In addition, somatic hypermuta-tion, which alters the specificamino acid by specificmutations of the gene encoding V regions, provides further diversity and leads to increased affinityand specificityas the immune response proceeds, generating affinity-ma-tured antibody. A comparison of the structures of germline antibodies and affinity-maturedantibodies shows that the former display significantconformational changes upon complex formation, whereas the antigens bind to the mature antibodies by a lock-and-key mechanism. Thus, it can be speculated that germline antibodies may

adopt

Figure 3Structure of complementarity-determining regions (CDRs).Variable domains of a murine immunoglobulin G (composedof heavy chain and light chain) are shown (fromHyHEL10structure, Tsumoto and Kumagai, unpublished results). The hypervariable regions, CDRs

comprising the antigen-binding site are shown by thick lines. These are located at the one edge of the b -barrel structure. Drawn using

MOLSCRIPT.

framework and the hypervariable regions shift the canonical structures relative to each other by small but significantamounts.

Contrary to other biomolecular interactions, antibodies can recognize various foreign antigens (e.g.small mole-cules, DNA, soluble proteins, surface proteins on viruses) by varying the hypervariable regions. The six CDR loops from the two chains at the rim of the eight-strand barrel provide an ideal arrangement for generating antigen-binding sites of differentshapes depending on the size and sequence of the loops. DifferentCDR structures form according to the antigen structure. These CDR structures can create flat,extended binding surfaces for protein antigens, a specificgroove for a peptide, DNA and carbohydrate, or specificdeep binding cavities for small molecules called haptens. A hapten is a simple chemical molecule that has the ability to bind antibody and can induce specificantibody production when it is attached to a carrier molecule such as albumin or Ficoll.

It has been pointed out that the distribution of amino acids in variable domains seems to be biased, and certain

V H

D (1

20)

J H(1

6)

C H

Translation

Protein

(H chain polypeptide)

come together on the surface when the polypeptide chain folds to form the native protein.

Various researchers have suggested that any macro-molecule can be antigenic, and that all accessible areas of a protein can potentially be bound by antibodies. ‘Discon-tinuous’seems to be a more accurate description of nonlinear epitopes since they are assembled from residues from differentportions of the polypeptide chain. However, recent research has suggested that not all areas seem to contribute equally to the binding. Small sets of surface residues on antigens make a significantcontribution to the interaction (calledenergetic epitope), while the rest of the antibody-binding regions seem to make additional con-tributions to the binding energy, although with some exceptions (antibody–idiotypeantibody binding).

Limitation of antibody specificity –cross-reactivity

As mentioned above, the most striking feature of the antigen–antibodyinteraction is its high specificityand affinity.However, in some cases, antibodies can recognize other antigens. This is called cross-reactivity, i.e. the ability of a binding site to accommodate antigens other than the original immunogen. The structural basis of this cross-reactivity has been investigated in several antigen–anti-body-binding systems. For example, antisteroid (proges-terone)–antibody(DB3)cross-reacts with several steroids, and is seen, from X-ray crystallography, to bind to alternative binding subsites, i.e. differentbinding orienta-tions of the steroids can be formed in these binding sites. Anti-hen egg lysozyme antibody (D11.15)cross-reacts with several avian lysozymes, in some cases with a higher affinitythan that for the original immunogen, hen lysozyme (calledheteroclitic binding). In this case residues differentfrom the original immunogen were found to be located around the edge of the epitope. Thus, it was concluded that a stereochemically permissive environment for the variant antigen residues at the antibody–antigeninterface was required for cross-reactivity.

V H C

H

Figure 4Molecular basis for antibody diversity. Only the case of heavy chain has been shown. In the human genome, one of the about 80VH genes (inmouse, about 100) recombines with one of 30D segments (inmouse, about ten), and one of six J segments (inmouse, four) producing a functional V-D-J gene in the B cell. The recombined DNA is transcribed, spliced and translated into a polypeptide chain. Half of the VH genes in human B cell seem to be pseudogenes.

multiple configurationsupon antigen binding, and com-bined with somatic hypermutation, this could stabilize the configurationwith optimal complementary to antigen. The structural plasticity of CDRs, especially in the case of germline antibodies, may adopt many differentconforma-tions, enabling them to accommodate many differentantigenic structures, which might lead to polyreactivity (i.e.reactivity to various antigens). Thus, adequate stickiness and the plasticity of CDRs may create high specificityand affinityof antibodies toward target anti-gens.

Forces Involved in Antigen-binding

Antigenic Epitope

Each antibody binds to a particular part of the antigen called the antigenic determinant (orepitope). The term epitope was firstproposed by Jerne, to include surface configurations,haptenic groups, specificareas and so on. For protein antigens, it was later proposed that epitopes might be subdivided into sequential epitopes (involvinga single continuous length of the polypeptide chain) and conformational epitopes, in which several discrete amino acid sequences, widely separated in the primary structure,

The binding of an antigen to an antibody takes place by the formation of multiple noncovalent bonds between the antigen and the amino acids of the binding site. The strength of a single antigen–antibodybond is the antibody affinity.It is produced by summation of the attractive and repulsive forces (vander Waals interactions, hydrogen bonds, salt bridges and hydrophobic force). The interac-tion of the antibody-combining site with antigen can be investigated thermodynamically and kinetically by using monovalent antibody fragments (fragmentsof variable regions or Fab).

In principle, the increase in van der Waals contacts and/or varied surfaces upon complexation correlates well with the affinity(thestrength of a single antigen–antibodybond) for the antigen in the case of hapten–antibodybinding. However, hydrogen bond formation and/ora saltbridge link (alsocalled ion pairing, a noncovalent bond formed when a charged residue (e.g.aspartate) attracts its oppositely charged group (e.g.lysine)) seem to be required for specificrecognition. In protein antigen–antibodyinteractions, the buried surface is almost the same

3), and creating shape complementarity between ( 750A

proteins is probably needed. Hydrogen bond formation is more frequently observed in comparison with other protein–proteininteractions, and considered to be a critical specificity-determiningfactor. Saltbridge forma-tion (e.g.aspartate–lysine)is not always seen, and seems not to be necessary. Although no gross conformational change upon binding has been observed, local induced fitting(seeabove) has been observed.

If a monovalent antibody fragment is used for analysis, the equilibrium of antigen–antibodybinding is definedas:

Antibody Antigen () Complex

K a

1

where

K a =[Complex]/[Antibody][Antigen].

Association and dissociation rate constants are definedas

follows:

V ass 5k ass [Antibody][Antigen]V diss 5k diss [Complex]

[2]where V ass and V diss represent the rates of association and dissociation, respectively, and k ass and k diss represent the rate constants of association and dissociation, respec-tively. At equilibrium V ass is equal to V diss and from eqns [1]and [2],the following equation is obtained:

K a =k ass /k diss

[3]

The Gibbs’energy of formation (D G 0) of an antigen–antibody complex is given by:

D G 0=2RT ln K a

[4]

where R is the gas constant and T is temperature.

The free energy of complex formation represents a balance between enthalpic (D H 0) and entropic (D S 0) forces as definedby the equation:

D G 0=D H 02T D S 0

[5]

In general, antigens and antibodies in solution have to overcome large entropic barriers before they can form a tight binding. There is a loss of the entropy of free rotation and translation of the separate molecules as well as a loss of conformational entropy of mobile segments and of side-chains upon binding. On the other hand, entropy is gained

when water molecules are displaced from the surfaces that become the newinterface. This latter effectis quite significantand, in the structures observed by X-ray analysis, it appears that water molecules are almost totally excluded from the interface by the close contact between antibodies and antigens. Enthalpic contributions arise from van der Waals interactions and hydrogen bond formation.

It is believed that the driving force in antigen–antibodybinding originates from an increase in the entropy of solvent molecules displaced from the interface upon complexation (i.e.it is entropy-driven). On the other hand, hydrogen bond formation and van der Waals interactions make only a little contribution to the overall binding energy and act mainly to determine the specificityto the interaction. However, thermodynamic analyses have suggested that a considerable number of antigen–antibodyinteractions are enthalpy-driven, i.e. they make favourable enthalpy changes with some opposition from the negative entropy contribution to association.

As mentioned above, it has been suggested from crystal structures of antigen–antibodycomplexes that shape complementarity of binding surfaces (inthe case of protein antigens) or close contact with small antigens (hapten,peptide and others) are important. In particular, almost all solvent molecules have been observed to be excluded from the interfaces, and therefore hydrophobic interactions are supposed to make a significantcontribution to the interaction. However, a recent high-resolution crystal-lographic study shows that several water molecules remain in the interface and make hydrogen bonds with both antigen and antibody. The water molecule complements the imperfect structural complementarity between antigen and antibody and makes a significantcontribution to the binding (about1–2kcal mol 21). In addition to the direct antigen–antibodyhydrogen bonds, solvent-mediated hy-drogen bond formation should drive the interaction.

The structural basis of antigen–antibodybinding is fundamentally important for clarifying the binding me-chanism. However, for further discussion, a structural study using X-ray crystallographic study or nuclear magnetic resonance (NMR)should be combined with an energetic study using thermodynamics and kinetics. Recent advances in genetic engineering have enabled antibody fragments to be obtained more easily, and the mutants can be constructed more conveniently. Some antigen–antibodybinding systems have been investigated using mutants, and the role of contact residues in the binding has been discussed. Thus, biological specificityand affinityoften depend on very subtle structural parameters, and extensive research is in progress.

Kinetic analyses on several antigen–antibodybonds have been performed to investigate the mechanism of creating high specificity.Although K a is extremely high ( 1015L mol 21), in some protein–ligandinteractions (avidin–biotinreaction), the intrinsic affinitiesof antibodies do not exceed 1010L mol 21(anaffinityceiling). From

eqn

[3],the ceiling originates from the limits for association and dissociation rate constants. The maximum association rate constant for the binding of a monomeric protein antigen by its antibody is approximately 105–106, a value controlled by the diffusioncoefficientsof the reactant molecules and verifiedexperimentally. In contrast, no limitation for dissociation rate constant seems to exist, and affinitychanges appear mostly as variations of the dissociation rate constant. Nevertheless, the k diss of the naturally prepared antibody molecule is fixedat 1023–1024, and it is considered that dissociation rates which are too slow would not be selected in the immune system. Thus, the affinityceiling of an antibody for any antigen is around 1010L mol 21.

Biological Significance of Antibody Affinity and Multivalency

An antibody with high affinityfor its antigen can function most effectivelyin the immune system (e.g.in biological reactions such as haemagglutination, virus neutralization, enzyme inactivation, haemolysis, immune elimination of foreign antigens, etc.). However, an antibody molecule with high affinityfor target antigen does not usually exist in the primary naive antibody library, and antibody affinityusually increases during an immune response (calledaffinitymaturation) in vivo . Nature provides us with numerous examples of molecules with low-intrinsic affinitybinding sites that are capable of high-avidity interactions with their targets due to multivalent binding. For example, the lowintrinsic affinityof IgM produced during the primary immune response is compensated by its penta-meric structure, resulting in a high avidity toward repetitive antigenic determinants present on the surface of bacteria or viruses. Thus, one of the most efficientways to increase the binding activity of an antibody to a surface (e.g.cell surface) is to make use of the multivalency effect.Therefore, even if the intrinsic affinityof an antibody molecule toward various invaders (e.g.viruses, proteins) is relatively low, high avidity can overcome the low intrinsic affinity,leading to the production of antibody molecules with high intrinsic affinityfor the target antigen through affinitymaturation in the immune system.

Affinity and Avidity

The strength of a single antigen–antibodybond is termed the antibody affinity.It is produced by summation of the attractive and repulsive forces mentioned above. However, since each monoclonal antibody produced by hybridoma technology has two antigen-binding sites and antibodies obtained from serum contain polyclonal antibodies which can bind multiple antigenic determinants, antibodies are potentially multivalent in their reaction with antigen. When an antigen carrying multiple copies of the antigenic determinant (macromoleculesor microorganisms) com-bines with a multivalent antibody, the binding strength is greatly increased because all of the antigen–antibodybonds must be broken simultaneously before the antigen and antibody can dissociate. Thus, total binding energy between a multivalent antigen and more than one of the antigen-binding sites of an antibody is greater than the summation of the affinityof each binding site for an antigen.

The strength with which a multivalent antibody binds a multivalent antigen is termed avidity, to differentiateit from the affinityof the bond between a single antigenic determinant and an individual combining site. The avidity of an antibody for its antigen is determined by the sum of all of the individual interactions taking place between individual antigen-binding sites of antibodies and deter-minants on the antigens. The avidity of an antibody for its antigen strongly depends on the affinitiesof the individual combining sites for the determinants on the antigens. It is greater than the summation of these affinitiesif both antigen-binding sites of an antibody can combine with the antigen. The effectiverange of antibody valence is from 2(IgG)up to 10(IgM),and the advantage of multivalence to the functional affinity(asopposed to the affinityof monovalent interactions, termed intrinsic affinity)is estimated to be 103–7. Thus, even when each antigen-binding site has only a lowaffinity(e.g.IgM produced early in immune responses), antibodies can function effectivelyin the immune system.

Further Reading

Arevalo JH, Taussing MJ and Wilson IA (1993)Molecular basis of crossreactivity and the limits of antibody–antigencomplementarity. Nature 365:859–863.

Bhat TN, Mariuzza RA, Poljak RJ et al . (1994)Bound water molecules and conformational stabilization help mediate an antigen–antibodyassociation. Proceedings of the National Academy of Sciences of the USA 91:1089–1093.

Branden C and Tooze J (eds)(1991)Recognition of foreign molecules by the immune system. Introduction to Protein Structure , chap. 12. New York:Garland Publishing.

Chitarra V, Alzari PM, Poljak RJ et al . (1993)Three-dimensional structure of a heteroclitic antigen–antibodycross-reaction complex. Proceedings of the National Academy of Sciences of the USA 90:7711–7715.

Chothia C, Lesk AM, Tramontano A et al . (1989)Conformations of immunoglobulin hypervariable regions. Nature 342:877–883.

Davies DR and Cohen GH (1996)Interactions of protein antigens with antibodies. Proceedings of the National Academy of Sciences of the USA 93:7–12.

Davies DR, Padlan EA and SheriffS (1990)Antibody–antigencomplexes. An n ual Reviewof Biochemistry 59:439–473.

Kraulis PJ (1991)MOLSCRIPT. Journal of Applied Crystallography 24:946–950.

Novotny J, Bruccoleri RE and Saul FA (1989)On the attribution of binding energy in antigen–antibodycomplexes McPC603, D1.3and HyHEL-5. Biochemistry 28:

4735–4749.

Padlan EA (1990)On the nature of antibody combining sites:unusual structural features that may confer on these sites an enhanced capacity for binding ligands. Proteins:Structure, Function, and Genetics 7:112–124.Padlan EA (1994)Anatomy of the antibody molecule. Molecular Webster DM, Henry AH and Rees AR (1994)Antibody–antigeninteractions. Current Opinion in Structural Biology 4:123–129.

Wedmayer GJ, Patten PA, Wang LH, Schultz PG and Stevens RC (1997)Structural insights into the evolution of an antibody combining site. Immunology 31:169–217.

Science 276:

1665–1669.

Roitt I, BrostoffJ and Male D (1996)Immunology , 3rd edn. London:Mosby Year Book.