Antigenic and structural study of homogeneous immunoglobulins produced in human disease has enabled the classification of the normally heterogeneous system of γ-globulins. The division of immunoglobulins into three major classes (γG, γA, and γM) based on the kind of heavy chain and into two antigenic types (κ and λ) determined by the light chain seems to be general for all species; however, the relative distribution of each of the six groups may vary with the species and is affected by the age and immune state of the animal. Isotypic differences in structure give rise to subclasses present in all normal animals of the same species, and allotypic differences among individuals produce a genetic polymorphism within a species. The Bence Jones proteins excreted by patients with multiple myeloma correspond to light chains but differ individually in their amino acid sequences (16, 21, 54). Both κ- and λ-chains in man and κ-chains in the mouse have many loci subject to variation in the NH2-terminal half of the chain but have only one or a few inherited variations in the COOH-terminal half. A similar structural division of heavy chains is suggested by partial sequence analysis of the normal γ-chain of the rabbit and of the γG1-chain of human myeloma patients. Because this variability in structure is limited to the portion of the molecule containing the antigen-combining site, it must give rise to antibody diversity and probably is related to antibody specificity. The nature of the mechanism which produces the variability in amino acid sequence is unknown. There are two principal, alternate theories, namely (i) there are many genes for the κ- and for the λ-types of light chains and likewise for the three main classes of heavy chains, and (ii) genes for immunoglobulin light (and heavy) chains undergo somatic hypermutation. In the first theory (the multiple germ line theory), some authors assume that many separate genes code for the variable part of the chain and only a few allelic genes code for the constant part (17, 18). In the alternative hypothesis, the somatic hypermutation has been attributed to an excision and faulty repair mechanism (55) or to several possible kinds of recombinational events between two genes [the master and the scrambler in the Smithies hypothesis (33)], or among a small number of genes that differ only in the variable part (56). Neither the germ line theory nor the somatic hypermutation theory are readily subject to experimental testing; neither successfully explains the genetic and evolutionary stability of the constant portion of the chains including the allotypic variations. Evolutionary relationships among light and heavy chains of the same and different species are manifested by the homology in primary structure which persists, despite the variability in amino acid sequence even within the same type of chain from one species. Large disulfide loops containing about 60 amino acids are characteristic features of both the light and heavy chains; these loops dominate the conformation of the chains, impart internal symmetry and seem to have important roles in antibody function. Comparative structural study suggests that the genes for both light and heavy chains evolved from a common ancestral gene coding for about 110 amino acid residues. Subsequent to the submission of this paper, Gottlieb et al. (59) reported the sequence of the first 96 residues of the human γG1 heavy chain Eu. The latter was identical to the γG1 heavy chain Daw in only 26 of the 82 positions compared. Of these, 24 are also identical in the Ou μ-chain. These results indicate that the amino acid sequence of the NH2-terminus of heavy chains is highly variable, like that of light chains, but may not be specific for the class of heavy chains. The latter conclusion is also supported by the report of Bennett (60) who found differences in the NH2-terminal pentapeptide sequences of four human μ-chains.
