A comparison of the amino acid sequences of different IgG antibodies from human beings or mice shows that the carboxyl-terminal half of the L chains and the carboxyl-terminal three-quarters of the H chains are very similar in all of the antibodies. Importantly, the amino-terminal domain of each chain is more variable, including three stretches of approximately 7 to 12 amino acids within each chain that are hypervariable, as shown for the H chain in Figure 34.10. The amino-terminal immunoglobulin domain of each chain is thus referred to as the variable region, whereas the remaining immunoglobulin domains are much more similar in all antibodies and are referred to as constant regions (Figure 34.11).

An IgG molecule consists of a total of 12 immunoglobulin domains. These domains have many sequence features in common and adopt a common structure, the immunoglobulin fold (Figure 34.12). Remarkably, this same structural domain is found in many other proteins that play key roles in both immune and nonimmune functions.

The immunoglobulin fold consists of a pair of β sheets, each built of antiparallel β strands, that surround a central hydrophobic core. A single disulfide bond bridges the two sheets. Two aspects of this structure are particularly important for its function. First, three loops present at one end of the structure form a potential binding surface. These loops contain the hypervariable sequences present in antibodies and in T-cell receptors. Variation of the amino acid sequences of these loops provides the major mechanism for the generation of the vastly diverse set of antibodies and T-cell receptors expressed by the immune system. These loops are referred to as hypervariable loops or complementarity-determining regions (CDRs). Second, the amino terminus and the carboxyl terminus are at opposite ends of the structure, which allows structural domains to be strung together to form chains, as in the L and H chains of antibodies. Such chains are present in several other key molecules in the immune system.

The immunoglobulin fold is one of the most prevalent domains encoded by the human genome: more than 750 genes encode proteins with at least one immunoglobulin fold recognizable at the level of amino acid sequence. Such domains are also common in other multicellular animals such as flies and nematodes. However, from inspection of amino acid sequence alone, immunoglobulin-fold domains do not appear to be present in yeast or plants, although these organisms possess other structurally similar domains, including the key photosynthetic electron-transport protein plastocyanin in plants (Section 19.3). Thus, the immunoglobulin-fold family appears to have expanded greatly along evolutionary branches leading to animals—particularly, vertebrates.

989

For each class of antibody, the variable domains at the amino-terminal ends of the L and H chains (designated V_L and V_H) come together to form a binding surface. The positions of the CDRs are striking. These hypervariable sequences, present in three loops of each domain, come together so that all six loops form a single surface at the end of each arm (Figure 34.13). Because virtually any V_L can pair with any V_H, a very large number of different binding sites can be constructed by their combinatorial association.

The results of x-ray crystallographic studies of several hundred large and small antigens bound to F_ab molecules have been sources of much insight into the structural basis of antibody specificity. The binding of antigens to antibodies is governed by the same principles that govern the binding of substrates to enzymes. The interaction between complementary shapes results in numerous contacts between amino acids at the binding surfaces of both molecules. Many hydrogen bonds, electrostatic interactions, and van der Waals interactions, reinforced by hydrophobic interactions, combine to give specific and strong binding.

A few aspects of antibody binding merit specific attention, inasmuch as they relate directly to the structure of immunoglobulins. The binding site on the antibody incorporates some or all of the CDRs in the variable domains of the antibody. Small molecules are likely to make contact with fewer CDRs, with perhaps 15 residues of the antibody participating in the binding interaction. Macromolecules often make more extensive contact, sometimes interacting with all six CDRs and 20 or more residues of the antibody. Small molecules often bind in a cleft of the antigen-binding region. Macromolecules, such as globular proteins, tend to interact across larger, fairly flat apposed surfaces bearing complementary protrusions and depressions.

A well-studied case of small-molecule binding is seen in an example of phosphorylcholine bound to F_ab. Crystallographic analysis revealed phosphorylcholine bound to a cavity lined by residues from five CDRs—two from the L chain and three from the H chain (Figure 34.14). The positively charged trimethylammonium group of phosphorylcholine is buried inside the wedge-shaped cavity, where it interacts electrostatically with two negatively charged residues, a glutamate and an aspartate. The negatively charged phosphoryl group of phosphorylcholine binds to the positively charged guanidinium group of an arginine residue at the mouth of the crevice and is hydrogen bonded to the side chain of a nearby tyrosine residue. Numerous van der Waals interactions, such as those made by a tryptophan side chain, also stabilize this complex.

990

Residues from five CDRs participate in the binding of phosphorylcholine to human F_ab. This binding does not significantly change the structure of the antibody, yet induced fit plays a role in the formation of many antibody–antigen complexes. A malleable binding site can accommodate many more kinds of ligands than can a rigid one. Thus, induced fit increases the repertoire of antibody specificities.

How do large antigens interact with antibodies? A large collection of antibodies against hen egg-white lysozyme has been structurally characterized in great detail (Figure 34.15). Each different antibody binds to a distinct surface of lysozyme. Let us examine the interactions in one of these complexes (complex ii in Figure 34.15) in detail. This antibody binds two polypeptide segments of lysozyme that are widely separated in the primary structure (Figure 34.16).

All six CDRs of the antibody make contact with this epitope. The region of contact is quite extensive (about 30 × 20 Å). The apposed surfaces are rather flat. The only exception is the side chain of glutamine 121 of lysozyme, which penetrates deeply into the antibody’s binding site, where it forms a hydrogen bond with a main-chain carbonyl oxygen atom and is surrounded by three aromatic side chains. The formation of 12 hydrogen bonds and numerous van der Waals interactions contributes to the high affinity (K_d = 20 nM) of this antibody–antigen interaction. Examination of the F_ab molecule without bound protein reveals that the structures of the V_L and V_H domains change little on binding, although they slide 1 Å apart to allow more intimate contact with lysozyme.

991