ALBERT EINSTEIN COLLEGE OF MEDICINE OF YESHIVA UNIVERSITY, BRONX, N.Y. DEPARTMENT OF MEDICINE (Acting Director: Prof. M. FuLoP) THE IMMUNOGLOBULIN H I N G E (INTERDOMAIN) R E G I O N JAY B. ADLERSBERG

In higher organisms, one of the first lines of defense is the antibody molecule, or immunoglobulin (Ig). Antibodies are serum proteins which protect against invasive microbes, participate in resistance to neoplasia, and after combination with their target molecules, or antigens, activate the humoral and cellular mechanisms of inflammation. Human immunoglobulins constitute approximately one-fifth of serum proteins, and are characterized electrophoretically by their cathodal (y) mobility. They can be divided into 5 classes: IgG, IgA, IgM, IgD and IgE. Although each class has distinctive chemical and physical properties, the basic structure of the antibodies in each class is very similar 2s. The immunoglobulin molecule can be visualized under the electron microscope as a Y-shaped structure u, 66. This shape is consistent with the known functional division of immunoglobulins. The arms of the Y are regions which bind specifically to the antigen and are termed the Fab fragments ~. Each Fab fragment contains an antigen-combining cleft, which must vary in configuration from antibody to antibody, to complement the shape of the antigen. The tail of the Y is the Fc fragment, the region which interacts with other serum proteins and cells to produce biologic responses. The Fc fragment contains sites which allow the molecule to bind and activate serum complement 34, fix to cell membrane receptors on lymphocytes, monocytes, granulocytes, platelets, and mast cells 2, 3.27.31, a7 and interact with receptors on cells necessary for transport of the molecule across the placenta m, or through the intestinal mucosa 32 The Fc and the 2 Fab fragments meet at a central point, which, under the electron microscope, alIows the Fab arms to swing out and lie at right angles to the Fc, giving the molecule a T-shaped appearance 66. This point, named the 'hinge', is one of the most fascinating regions of the immunoglobulin. To approach the molecular nature of the hinge region, it is essential to understand the basic immunoglobulin structure (fig. 1). The subunit of all immunoglobulin classes Key-words: Crossing-over; Heavy Chain Disease; Hinge (interdomain) region; Immunoglobulin; Partial duplications; Unique DNA fragment. Abbreviations used: nomenclature of immunoglobulins, their chains and fragments folIows the recommendation of the World Health Organization, Bull. Wid Hlth Org. 30, 477, 1964; 33, 721, 1965; 35, 953, 1966; 38, 151, 1968. Heavy Chain Disease and myelomaproteins are designated by the first 3 letters of the patient's name. Amino acid nomenclature conforms to that suggested by the IUPAC-IUB Commissionon BiochemicalNomenclature. Received, March 1, 1976. La Ricerca Clin. Lab. 6, 191, 1976. 191

THE I g HINGE REGION

t[ light chain

VL

110 J" •

CL

NH_~

/-

COOH

S Epapain r T

S heavy chain

HINGE

NH2

,

v.

" 51~

CHl-~'-[

216 ~I{ ' ~' --L~'----'--CH2 2al I I

COOH ~" '~

a41

CH3

>446

SS

..=

\

{,11]

coo.

S~ j papain S x~ COOH

NFI~ Fab

Fc

Fig. 1 - Model of the basic immunoglobulin subunit, in this case IgG. Noted are the variable and constant regions of the heaw and light chains, and the hinge region. Amino acid numbering is based on IgG1 Eu 11.One Fc and 2 Fab fragments are produced by papain cleavage. is a 4-chain polypeptide, which is a monomer in the IgG class, or a polymer of 2 to 5 subunits in the IgA and IgM classes 9. 28,42.49. The subunit consists of 2 types of polypeptide chains referred to as the light (L) chain, which has a molecular weight of 23,000, and the hea W (H) chain, with a molecular weight of 50,000-70,000. The basic subunit may be symbolized as H2L2,has a molecular weight of 150,000-180,000, and is composed of 2 identical chains of each type linked by disulfide bonds between cysteine amino acids. The heavy chains of all 5 classes are antigenically and immunochemically distinct from one another, and are referred to by the Greek letter denoting the class, i.e. there are 2 2", ~, b~, 8 or a chains per subunit of IgG, IgA, IgM, IgD and IgE, respectively. The hea W chains may be divided into a variable (Vu) region, from the N-terminus to approximately amino acid residue 119, and a constant (CH) region, from residue 120 to the carboxy-terminus is. The constant region is antigenicalIy and immunochemically characteristic for each H chain class; further, the T and 0~ chain constant regions can be divided into subclasses, based on antigenic and chemical differences 28. The VH region may be shared among all the H chain classes, and contains an antigen-binding site. Associated with the 2 heavy chains are 2 identical light chains, of either the kappa (x) or lambda (X) type, each about 220 residues long. The light chain contains a VL region of approximately 110 residues, and a CL region of the same length. The VL region contains an antigen-specific site which complements that of the VH region in forming the antigen-binding cleft. The GL regions are antigenically distinct for x and X, and different VL regions are associated with each Ci. type. The light chains are joined to the heavy chains by the terminal cysteine of the x chain, and the penultimate cysteine of the X chain. Both the H and L chains have intrachain disulfide bridges occurring once in each 110 amino acid stretch; the bridging cysteines are separated by approximately 60 residues, and their joining produces a loop. Thus, in the light chain, there is a V region loop between cysteine residues 23-88, and a C region loop between residues 192

J. B. ADLERSBERG

134-194. In the H chain, there is one V region, and, depending on the H class, 3 or 4 C region loops 4, 23, 57 The amino acid sequence data from an entire IgG1 heavy chain have allowed the division of the constant region into 3 ll0-residue 'domains', based on the regularity of these disulfide-linked loops. When the amino acid sequences of the 3 domains are aligned and compared, there is a 35 % identity among them and the same degree of homology with the CL region 10. The domains have thus also been termed the Cul, C~2 and Cu3 'homology regions'. This degree of similarity from one portion of the chain to another is too great to be a chance occurrence, and it thus appears that the IgG1 constant region is the result of duplications of an ancestral immunoglobulin gene, which produced a molecule about 110 residues in length 33. The looped configurations of the juxtaposed domains in the heavy and light chains confer a globular shape to each pairing along the molecule i2, 53.6o (~g. 2). This beaded effect is demonstrated by X-ray crystallographic analysis, where globular areas are noted for the variable regions s3, the CL and first CH region s3, and the terminal CH areas so, % The hinge lies between the first and second CH regions, hence its alternate name, 'interdomain' region; whether it actually is exposed between the globular domains formed by these regions has not been determined by crystallographic studies. However, this concept is probable because the hinge is extremely vulnerable to enzymatic attack. The hinge region links the 2 H chains together by one or more interchain disulfide bridges between its cysteine residues. The number of these cysteines and their position in the sequence of the region vary not onty from class to class, but within subdivisions of each class 22. In addition, if one attempts to align the T1 hinge, for example, with any of the CT homology regions, no similarity can be found; many prolines and cysteines distinguish this as an unusual amino acid sequence. Thus, lack of homology to any constant region domain is a feature of the hinge region.

Fig. 2 - Diagrammatic representation of the Ig domains. (From: FRANGIONEB. is, with permission of the author and publisher). 193

THE I g HINGE REGION

The variability in the number and position of the cysteine residues in the 4 IgG subclass hinges has been utilized to develop a simple method, called 'chemical typing' 21, for distinguishing the subclass of an IgG myeloma protein. This subclassification has been difficult to achieve in the past by immunologic techniques because of problems in developing antisera which could distinguish the minute differences in amino acid sequence in the constant regions of the 4 heavy chains. Chemical typing involves breaking the interchain cysteine bonds in the hinge under mild reducing conditions, then labelling them with 14C-alkylating agent. The molecule is then digested with proteolytic enzymes. The hinge peptides produced vary in composition from subclass to subclass and, by paper electrophoresis, migrate in distinctive patterns 17.23 (fig. 3). These patterns become visible by exposing the electrophoresed paper strip to photographic film, allowing the labelled cysteine-containing peptides to expose the ftlm in characteristic bands. In actual practice, this technique can be used to identify all the heavy chain classes and subclasses because the hinge region peptides are distinct from one subclass to another; the light chain types may also be demonstrated 15, 21. In addition to the variability in the number, position, and sequence around the cysteines, a large number of proline residues are found in the hinge. These amino acids impart a random coil structure and flexibility to this area of the polypeptide chain 64, presumably excluding it from the globular domains, and perhaps explaining its extreme susceptibility to cleavage by proteotytic enzymes 48,54 The hinge may thus be defined biochemically as a flexible amino acid stretch in the central part of the heavy chain, rich in cysteine and proline, extremely variable in amino acid sequence from class to class, and having no resemblance to any other region of the immunoglobulin heaw chain. This variability will assume greater importance in the discussion of the hinge region's evolution. T h e I g G hinge - Our ability to understand the nature of the hinge region is due to the surge of amino acid sequence information on immunoglobulins which has risen exponentially over the last 10 years. The elucidation of the complete amino acid sequence of an IgG1 ~r chain in 1969 i, demonstrated the "rl hinge to be a stretch of 15 amino acids lying between the CH1 and CH2 homology regions. The presence of the 2 cysteines uniting the ~rl chains, a third cysteine which forms the H-L bridge, and a high proline content distinguish this region from the Ca sequences. In the other subclasses of IgG, this distinction becomes even more dramatic, and not only does the size of the hinge region vary, but the number of cysteines and their position in the hinge change as well (fig. 4). The IgG2 and IgG4 molecules have hinges most similar in length to IgG1, and while the IgG4 hinge contains 2 H-H cysteines, the IgG2 hinge has a total of 4 47 The hinge region of the IgG3 molecule is currently the subject of intense investigation and some controversy. Until recently, it was postulated that this area contained 5 interchain cysteines, and was twice the size of the other ~" interdomains 19. When it was noted that the IgG3 H chain was about 10,000 daltons larger than the H chains of the other classes (60,000 versus 50,000) sg, ss, it was suggested 8, and later demonstrated 44, that the difference was due to an extended hinge region containing about 100 amino acids. The origin of this IgG3 interdomain which was 5 times longer than that of IgG1 was in part clarified by the observations on Heavy Chain Disease protein Omm-s 1, 26. The hinge region from this IgG3 molecule was easily isolated, and was found to contain approximately 100 amino acid residues. The enzymatic digestion of this region, however, demonstrated only the same small number of peptides obtained in the original work on the "r3 hinge 19. It thus appeared that the increased length of the IgG3 hinge was due to a series of duplications of a short amino acid stretch almost identical to the 15 residues of the IgG1 hinge (see legend, fig. 4).

194

~

I

x

m

m

m

m

f2

gG r3

r4.

"ilgM.

, .

i,

IgA,

:,

IgD gE

?ig. 3 - Chemical typing patterns of the major classes of human immunoglobulins. Bands represent areas of exposed X-ray film, Amino acid narkers: Gly = glycine; Ala -- alanine; Lys = lysine; Glu = glutamic acid; Asp = aspartic acid. ×, ),, light chain H-L cysteine peptides, From: FRANCIONEB. ~s, with permission of the author and publisher).

~sp

Glu

:-DNP-Lys

Gly-Ala

narkers

mbclass

:lass

J. B. ADLERSBERG

-¢'3 . . . . Lys - T h r - P r o - L e u - ~ A s p - T h r

- Thr - His - T h r - C y s - P r o ~

IAla- . . . 231 -fl --'y2 -.. ~4 .-.

I6 - Pro - Lys - Set - Cys - Asp ~

G l u ~ L y s ~

•

Thr - H~s - Thr -Cys - Pro--P ro - Cys - Pro -Ala . . . .

• -~}Cys-Pro-Pro-Cys-Pro

t+++ +++,c+

,o

+++ + + , +

-Ala . . . . Ata . . . .

Fig. 4 - The y chain hinge regions. The subclasses are ordered to maximize homology, and dots are introduced for the same purpose. Residues included in the boxes are identical. Numbering is based on IgG1 Eu n. The first 15 residues of y3 are arranged above the subsequent 15 to show homology within this hinge. The second 15 residues represent the duplicated segment; duplications are not shown because their number is not known. Note that the maximal homology is between y1 and the 2 series in y3. In addition to the repetitions of this segment is a piece 15 residues long at the N-terminal end of the hinge, whose sequence resembles the 3"1 hinge with only 50 % homology. Of the 7 differing residues, however, 6 can be explained on the basis of a single nucleotide substitution on the DNA level. It thus appears that the 3'3 and y1 hinges are closely related in evolution. At this point, it must be emphasized that though the data from protein Omm-s indicate multiple repetitions, their exact number is unknown i. In addition, Omm-s represents a Heavy Chain Disease protein, and though studies on one normal IgG3 protein suggest the presence of 15 hinge cysteines 43, both the number of repetitions of hinge sequence and the number and position of the cysteines await the complete sequence of a normal IgG3 interdomain region. The large number of 3"3 hinge cysteines in close proximity to one another may permit disulfide bond exchange to occur, and allow intermolecular binding. This phenomenon might account for the hyperviscosity syndrome occasionalXy seen in patients with IgG3 multiple myeloma 6. More knowledge of the 3" hinge region has come from studies of other IgG Heavy Chain Disease proteins 18. These are abnormal molecules, composed of 2 identical, shortened heavy chains with no light chains attached. After a short variable region stretch, there is an internal deletion, and normal amino acid sequence resumes at the N-terminal glutamic acid residue of the hinge 2o,z~, 7o (also the postulated initiation site of the 3"3 duplications 16). Another type of @-ICD protein, Mcg, has its entire hinge region deleted 14. The significance of these abnormal molecules in relation to the origin of the hinge will be considered in the last section. The I g A binge - Though it comprises only about one-tenth of the circulating serum immunoglobulin, IgA is the principal secretory antibody in the respiratory and intestinal tracts, as wetl as a major immunoglobulin product of the lactating breast 32. In these secretions, the molecule generally exists as a dimer of 2 subunits linked by a short polypeptide, the J chain 32. In addition, as the dimer passes through the intestinal epithelial cell, the 70,000-dalton secretory component is added to it a:. The function of the latter is not known with certainty, but perhaps serves to protect the proteolytically labile hinge region from the action of intestinal proteases 32, 35,36 There are 2 subclasses of 0: heaw chains, 0~1 and cz2, and thus 2 classes of serum and secretory IgA. The sequences of the ~1 and ~z2 hinge regions are shown in fig. 5.

197

THE I g HINGE REGION

Of major note is a size difference of 13 residues between the larger 0~1 and the ¢2. Also missing in o~2 are the carbohydrate moieties attached at each of the 5 serine residues in the o~1 sequence z4, 71. This size difference is of functional importance, because intestinal IgA protease breaks down secreted IgA1 by cleaving it at a point within the 13 extra-hinge residues, and thus has no effect on secreted IgA2 s2. It may have been the evolutionary advent of this protease that caused the concentration of IgA1 to decline from 90 % in serum to only 50 % in external secretions 30, allowing the more stable IgA2 molecule to assume a major protective role s2, Despite their different lengths, the 0~hinges are intimately related. A comparison reveals identical sequence for the first 10 of 12 residues, and the differing residues can be explained on the basis of a single nucleotide change 24. After the initial ctl valine, the next 24 residues can be divided into 3 series of 8 amino acids each. The second and third series are an exact duplication, as shown in fig. 5 24. The first 8 residues, which correspond to the 0~2 sequence, can now be aligned with the duplicated segment, forming a triplication of 8 residues with great homology among them. The relation of the cz2 and ~zl hinges may now be compared to that of the T1 and y3 hinges: first, the shorter hinge is represented near-identically in the longer; second, the longer hinges contain duplicated segments (only suspected in y3). Though there are 3 cysteines in the hinge of the ~ chains, only the central cysteine is involved in interchain binding 41,74. The first is involved in an intrachain bridge with a cysteine in the Ca1 region, and the last with a cysteine in the Fc region. This Fc cysteine is almost immediately adjacent to another cysteine which appears involved in either a second interchain bond, or in a bond linking 2 IgA monomers (an intersubunit bond) 74. The presence of these intrachain bonds within the hinges of IgA molecules involves an unusual degree of folding, the function of which is not understood 72. IgA proteins with internal deletions (~ Heavy Chain Disease) exist, and 2 have been biochemically characterized. Proteins Def and Ait are dimers of 2 czl H chains, each having an internal deletion of most of the Vu and Cul regions 73. Normal sequence resumes at the valine residue near the beginning of the hinge, which is postulated to be the initiation site of the true cz hinge, as is the glutamic acid residue at position 216 in the y hinge 73. Corroborating this finding is a valine which reinitiates normal sequence at the start of the hinge in a mouse IgG HCD protein with a Grl deletion 46.

The IgM, IgD, and IgE hinges - These 3 classes of immunoglobulins are present in relatively small amounts in the serum, IgM having the greatest concentration of approximately 100 rag/100 ml. IgM, though a polymer of 5 subunits when found in serum, is present as a monomer on the surface of unstimulated B-lymphocytes, and is thought to be a receptor molecule for antigen 69. When such a cell is stimulated by its specific antigen, the first molecule it produces is IgM in polymeric form. Though the evidence is not dear, an IgM-like molecule appears in the earliest antibody-producing vertebrates 39, and thus by ontogeny and perhaps phylogeny, IgM represents the earliest known antibody. IgD is present in very small concentrations in the serum, and is a major component of B-lymphocyte membranes 67. It is the immunoglobulin which appears on the surface of unstimulated B-ceils after IgM 6a and perhaps has an antigen-receptor function ss. IgE is the cytophilic reaginic antibody in man, which fixes to the surface of mast cells and causes histamine release and allergic phenomena on combination with its specific antigen 62,63. Perhaps because of its cytophiiic nature, it is present in serum in extremely low quantity 57 * A variant of the ~z2 hinge has been described which has 2 additional proline residues; their position in this hinge is not known 4s. 198

J. B. ADLERSBERG

The complete sequences of the ~ and ~ heavy chains have recently been published s. s~. 70. Both are over 550 residues in length, as compared to 446 for the 7 chain. In each case, there appears to be a variable region of approximately 120 residues, and a constant region which can be divided into 4 110-residue homology regions, one greater than in the y chain. Less data are known about the ~ chain sequence, but several cysteine-containing peptides have been isolated from it, including the inter-I-I chain peptide 40. When the sequence around the inter-heavy disulfide bridges of the gt, ~ or a chains is compared to the T and 0~ hinge region sequences, there is no evident homology. tn the central region of the former 3 H chains, there are only 1 to 2 inter-heavy disulfide bridges (the ~ chain has a cysteine bridge at the penultimate residue), and there is no proline-rich region surrounding them. These observations may indicate that when a fourth constant region domain exists, as in the bt and ~ chains, there is no discrete hinge region. This idea is supported by the finding that trypsin digestion cleaves the entire C~2 region into small peptides, suggesting that the C~2 is an exposed portion of the chain 56. Also the C~2, like the C~:2, is rich in proline residues in the region just before the first intrachain cysteine. By the criteria of enzymatic sensitivity and proline content, the C~2 domain may correspond to the T and 0~chain hinge regions. The C,2 region also contains the inter-heavy cysteine bonds. As in IgM, there is no proline-rich sequence around the cysteines, but there is some homology to the 8 chain hinge peptide 6~. As a result of these comparisons, it appears that the only 2 classes which have a discrete interdomain region enriched in proXine and cysteine are IgG and IgA.

The origin o] the hinge The homology among the amino acid sequences of the 1 10 constant residues of the light chain, the 3 constant region domains of the T heavy chain, and the 4 C region domains of the ~ and ~ heavy chains, suggests that the ancestral antibody molecule was about 1 10 residues in length 33. It is postulated that a very early duplication of its genetic material produced the 110-residue variable region, which diversified as a distinct cluster of genes. This ancestral gene then gave rise to the light chain constant regions, and duplicated in tandem fashion to produce the heavy chain constant regions. A functional difference may have arisen with structural duplication. The N-terminal portion of the H chain, for example, is involved with antigen binding, whereas the C-terminal portion initiates the biologic activities of the molecule: complement fixa-

~2 ¢1

F

T h r - P r o - Pro - T h r - Pro - - S e t - Pro - S e t l

t . . . .

Fig. 5 - The ~z chain hinge regions. The first part of the ~1 hinge is divided into 3 consecutive series of 8 residues to indicate duplicated sequences. Note the duplication present in the second and third series, and the great homologyof the first 8 residues with the duplicated segment.Also note the homologyof the tz2 hingewith the first 8 and the last 5 residues of the ~1 hinge. Identical residues are shown by boxes. Dots have been introduced to maximizehomology. 199

THE I g HINGE REGION

tion, placental transfer, binding to cell membranes. In the early T and 0~ classes, perhaps these 2 functional regions existed as isolated integrities, and the hinge was derived teleologically as a unit to join them into a molecule which could combine with an antigen and elicit a biologic response to it. How did this structurally distinct unit originate? Based on our knowledge of molecular genetics and on amino acid sequence data from a number of immunoglobulin heavy chains, 3 hypotheses can be advanced. i) The hinge material may have arisen from a preexisting immunoglobulin gene, by unequal crossing-over between 2 homologous chromosomes. Consider the IgM H chain with its extradomain, and its CH2 region, which like the T hinge, has increased susceptibility to enzymatic digestion. If 2 ancestral ~ chain genes were to align unevenly and cross-over in the CH2 domain during meiosis, i recombinant would have a partial duplication of this region, while the other would be left with only a smaU piece of the CH2. If there were some developmental advantage to having only 3 C region domains instead of 4, perhaps the latter I~ chain recombinant with its tiny CH2 remnant would survive, and over miUenia of genetic mutation, deletion, insertion, inversion and duplication, evolve into an ~ or T chain. The small CH2 remnant may have evolved its own specialized function by mutating to a high cysteine content, to become what we recognize today as an ¢ or T hinge region. As a demonstration that unequal crossing-over is a reasonable explanation for the evolution of the hinge, one has only to look at the sequence of the ~2 hinge. By unequal crossing-over involving the central Pro-Cys-Pro sequence, one can derive the central 6-residue sequence of the T4 hinge region (figs 4 and 5). ii) Because the hinge lies between 2 ll0-residue homology units, perhaps it arose from the transcription of normally untranscribed genetic material present at one end of the ancestral gene (cistron), which became an intracistronic piece as this gene duplicated to form the CH1 and Cu2 domains. This hypothesis would explain why the y hinge just precedes an 8-residue sequence in Cr2 which is nearly identical to the N-terminal of the ancestral-like, 110-residue CL region. Such a transcribed intracistronic region has been proposed to explain the joining of 2 closely-related enzymes in E. coli sl. In bacteriophage )., another situation which exemplifies this hypothesis exists. Just preceding a structural gene in the phage DNA is a series of reiterations of a 17 base pair sequence, which normally remains untranscribed into messenger RNA u' ss. It will be remembered that in the T3 and czl hinges, there are a series of repeated amino acid sequences i, 19,a. Could these hinges have originated from the transcription of such a reiterated piece which became centrally located between a duplicating ancestral immunoglobulin gene? iii) The third hypothesis for the origin of the hinge is that it represents genetic material which was inserted into the immunoglobulin constant region either during its ancestry, or in contemporary de novo synthesis. Why inserted? There are several reasons for suspecting that the genetic material coding for the hinge region of the y and ~z chains was not originally part of the constant region DNA. The first is the non-homology of the y and 0~ hinges with any of the CH homology regions. The second is evidence accumulated in the study of Heavy Chain Disease proteins of the y class. These molecules are dimers of 2 y H chains, each containing an internal deletion in the VH and CH1 regions. The resultant molecule has a normal N-terminal sequence for several to 100 residues, is missing the rest of VH and CH1, and is then followed by a normal hinge region, Cu2 and CH3. The sequence of these molecules has been established in 3 cases representing the y1, T2 and y3 subclasses 7, 2o.zs; after the CH1 deletion, normal sequence in each resumes at glutamic acid residue 216 (y1 numbering 1i). 200

J. B. A D L E R S B E R G

The simplest explanation for these internally deleted H chains is breaking and rejoining of V-C genes. However, residue 216 marks the beginning of the non-homologous hinge sequence (an analogous Val residue exists in 2 0~HCD proteins thus far studied, and one mouse IgG protein; see above). Even more provocative is the finding of y1 molecule Mcg which has deletion involving only the hinge 14. Normal amino acid sequence in Mcg stops at Val 215, and resumes at Ala residue 231, as if the hinge had simply been plucked from the heavy chain. Are G1u and Val thus sites for orderly recombination of genetic material, and did the ancestral H chain have its N and C terminal halves linked by site-specific recombination of hinge DNA with C~i DNA components? This event would parallel the insertion of viral genomes into host nucleic acids. It would have produced the intact y I-I chain as we know it today, with variation among subclass hinges arising as the subclasses diverged. On the other hand, perhaps the y and 0~hinge region DNA represents a cluster of distinct genes, and all our ~" and 0~chains synthesized de novo are the result of specific recombination among V, C, and 'H' (hinge) genes. This theory would allow for the independent evolution of the hinge regions by tandem duplication, just as the ancestral Ig gene is proposed to have produced the H chain. More of these HCD molecules, particularly those of the Mcg type, will have to be studied to further substantiate the insertion theory. For a moment, let us evaluate the known sequences of the "1"and ~ hinges in light of the preceding discussion. There is great homology among the hinges in the "y subclasses. In the y3 hinge, multiple repetitions of a ~'l-like sequence are thought to be present, but because the sequence of this hinge is not complete, the exact number of repetitions is unknown 1. These duplicated sequences can be explained by a tandem duplication of a "l'l-like hinge, or unequal crossing-over between 2 y1 hinge genes plus tandem duplication. However, there is a 15-residue stretch at the N-terminus of the hinge which has only 50 98 similarity to the y1 sequence. Possibly, the differing residues represent highly mutable points, as all but one of them can be derived from the ~'1 sequence by a single base change/codon. Alternatively, if a cluster of hinge region gene exists, these 15 residues may be the result of the N-terminal addition of an ancestral y hinge gene to a duplicating y1 hinge gene by unequal crossing-over. Similarly, the known 16-residue sequence duplication in the 0~1 hinge must be the result of tandem doubling of the hinge DNA 24. The preceding 8 residues are very similar to the duplicated segment, and though not identical, the differing residues can be explained as single-base interchanges in all but one position. The o~1 interdomain may thus represent a triplication of an original 8-residue piece, with a highly mutable N-terminal sequence, as is mentioned above for the y3 hinge. However, the 8 N-terminal residues are identical to the hinge of the 0~2 subclass, save for the last residue (explainable as a single base substitution). The 0~1 hinge thus may have arisen as the result of tandem duplication of an ancestral gene coding an 8-residue 0~1 hinge sequence, with N-terminal addition of an 0~2 hinge gene by unequal crossing-over * CONCLUSION The hinge region of the immunogiobulin thus appears to be a distinct unit in terms of amino acid sequence, structure and function. The relation of these hinge structures in the different immunoglobulin classes, particularly in IgG and IgA, supports the concept that this region is coded for by a unique small piece of DNA, which has evolved in parallel manner with the immunoglobulin genes, by partial duplication * The 8-residue segments mentioned above are comprisedof 2 4-residue segments, with remarkable homologybetween them, in both the =1 and =2 hinges. The ancestral = hinge might therefore have been only 4 residues in length, making the present-day =1 hinge the result of a sextuplication. 201

THE Ig HINGE REGION

a n d / o r u n e q u a l crossing-over. A n additional theory exists for its origin f r o m intracistronic D N A p r e s e n t as the result of duplication of the ancestral gene coding for a 110-residue chain. A third t h e o r y suggests its origin f r o m a f o r e s h o r t e n e d constant r e g i o n d o m a i n . A detailed structural comparison of the i n t e r d o m a i n s of i m m u n o g l o b u l i n molecules and i m m u n o g l o b u l i n m R N A from d i f f e r e n t species m a y p r o v i d e answers to t h e questions surrounding the e v o l u t i o n of this u n i q u e molecular region. SUMMARY The hinge region is a flexible amino acid stretch in the central part of the heavy chains of the IgG and IgA immunoglobulin classes, which links these 2 chains by disulfide bonds. It is rich in cysteine and proline amino acids, extremely variable in amino acid sequence, and has no resemblance to any other immunoglobulin region. The hinges in these 2 classes are compared and contrasted. Such a distinct molecular structure does not exist around the inter-heavy disulfide bonds of the other Ig classes, but a portion of the IgM heavy chain has similar properties to the y hinge. This similarity suggests one hypothesis for the genetic origin of the hinge region; data supporting 2 other hypotheses are also presented. ACKNOWLEDGEMENTS The author wishes to thank Drs. Bias Frangione and Edward C. Franklin for their thoughtful discussions and review of the manuscript. REFERENCES

1) ADLERSBERGJ. B., FRANKLIN E. C., FRANGIONEB.: Repetitive Hinge Region Sequences in 2) 3)

4) 5) 6) 7) 8) 9) 10)

11) 12) 13) 14) 15)

202

Human IgG3: Isolation of an ll,000-Dalton Fragment - Proc. nat. Acad. Sci. (Wash.) 72, 723, 1975. BASTENA., MILLER J. F., SPRENT J., PYE J.: A Receptor for Antibody on B Lymphocytes. I. Method of Detection and Functional Significance - J. exp. Med. 135, 610, 1972. BASTEN A., WARNER N. L., MANDEL T.: A Receptor for Antibody on B Lymphocytes. II. Immunochemical and Electron Microscopy Characteristics - J. exp. Med. 135, 627, 1972. BENmCH H., JOHANSSONS. G. O.: Immunoglobulin E and Immediate Hypersensitivity - Vox Sang. (Basel) I9, 1, 1970. BEm,aCH H., YONBAHR-LINDSTR6MH.: Structure of Immunoglobulin E (IgE) - In: BRENT L., HOLBOROW J. (Eds): Progress in Immunology II. American Elsevier Publishing Co., Inc., New York, 1974; vol. 1, p. 49. CAI'RAJ. D., KImKEL H. G.: Aggregation of TG3 Proteins: Relevance to the Hyperviscosity Syndrome - J. din. Invest. 49, 610, 1970. COOPERS. M., FRANKLINE. C., FXANGIONEB.: Molecular Defect in a Gamma-2 Heavy Chain Science 176, 187, 1972. DAMMACCOF., FRANKLINE. C., FRANGIONEB.: Studies of Two Papain Fragments (Fd, and FcL) from an IgG3 Myeloma Cryoglobulin - Ahstr. Commun. Meet. Fed. europ, biochem. Soc. 8, 680, 1972. Doum~cAslqxlXR. R., VIRELLAG., PARKHOUSER. M. E.: Electron Microscopy of Human and Mouse Myeloma Serum IgA - J. molec. Biol. 56, 207, 1971. EDELMANG. M.: The Covalent Structure of Human Gamma G-Immunoglohulin. XI. Functional Implications - Biochemistry 9, 3197, 1970. EDELMANG. M., CUNmNGHAMB. A., GALL W. E., GOTTLIEB P. D., RUTISHAUSERU., WAXDAL M. J.: The Covalent Structure of an Entire Gamma G Immunoglohulin Molecule - Proc. nat. Acad. Sci. (Wash.) 63, 78, 1969. EDMI_n'mSONA. B., SCHIFFER M., ELY K. R., WOOD M. K.: Structure of a Lambda-Like Bence-Jones Protein at 6-A Resolution - Biochemistry 11, 1822, 1972. FEINSTEINA., ROWEA. J.: Molecular Mechanism of Formation of an Antigen-Antibody Complex - Nature (Lond.) 205, 147, 1965. FETT J. W., DEUTSCH H. F., SMITHIES O.: Hinge-Region Deletion Localized in the IgG1Globulin Mcg - Immunochemistry 10, 115, 1973. FRANGIONEB.: Structure of Human Immunoglobulins and Their Variants - In: BENACERRAF B. (Ed.): Immunodeficiency and Immunogenetics. Medical and Technical Publishing Co., Ltd., Lancaster, England, 1975; p. 1.

J. B. ADLERSBERG

16) FRANGIONEB., ADLERSBERGJ. B.: Unpublished Observation. 17) FRANCIONEB., FRANKLINE. C.: Chemical Typing of the Immunoglobulins IgM, IgA1, and IgA2 - FEBS Letters 20, 321, 1972. 18) FRANGIONEB., FRAN~CLmE. C.: Heavy Chain Diseases: Clinical Features and Molecular Significance of the Disordered Immunoglobulin Structure - Semin. Haemat. 10, 53, 1973. 19) FRANGIONEB., MILSTmNC.: Variations in the S-S Bridges of Immunoglobulins G: Interchain Disulfide Bridges of 7G3 Myeloma Proteins - J. molec. Biol. 33, 893, 1968. 20) FRANCIONEB., MrLSTEm C.: Partial Deletion in the Heavy Chain Disease Protein ZUC Nature (Lond.) 224, 579, 1969. 21 ) FRANGIONEB., MILSTEm C., FRANKLINE. C.: Chemical Typing of Immunoglobuhns - Nature (Lond.) 221, 149, 1969. 22) FRANCIONEB., MILSTEIN C., PINKJ. R. L.: Structural Studies of ImmunoglobulinsG - Nature (Lond.) 221,145, 1969. 23) FRANGIONEB., PRELLI F., MIHAESCO C., WOLFENSTEIN-TODELC., MIHAESCOE., FRANKLIN E.C.: Structural Studies of Immunoglobulin G, M, and A Heavy Chains - Ann. N.Y. Acad. Sci. 190, 71, 1971. 24) FRANGIONEB., WOLFENSTEIN-TODELC.: Partial Duplication in the 'Hinge' Region of IgA1 Myeloma Proteins - Proc. nat. Acad. Sci. (Wash.) 69, 3673, 1972. 25) FRANKLINE. C., FRANGIONEB.: The Molecular Defect in a Protein (CRA) Found in Gamma-1 Heavy Chain Disease, and Its Genetic Implications - Proc. nat. Acad. Sci. (Wash.) 68, 187, 1971. 26) FRANKLII'~E. C., FRANGIONEB., ADLERSBERGJ. B.: An Antigenic Determinant Unique to the Hinge of T3 Heavy Chains - .1. Immunol. 115, 314, 1975. 27) FRANKLINE. C., OVARYZ.: On the Sensitizing Properties of Some Normal and Pathologic Human Immune Globulins and Fragments Obtained by Papain or Pepsin Digestion - Immunology 6, 434, 1963. 28) GALLYJ.: Structure of Immunoglobulins- In: 5ELAM. (Ed.): The Antigens. Academic Press, New York, 1973; vol. 1, p. 161. 29) GITLIN D., KUMATE J., URRUSTI J., MORALES C.: The Selectivity of the Human Placenta in the Transfer of Plasma Proteins from Mother to Fetus - J. clin. Invest. 43, 1938, 1964. 30) GREYH. M., ABEL C. A., YOImTW. J., KUNKELH. G.: A Subclass of Human yA Globulins (yA2) which Lacks the Disulfide Bonds Linking Heavy and Light Chains - J. exp. Med. 128, 1223, 1968. 31) HENSONP. M., SPIEGELBEgGH. L.: Release of Serotonin from Human P1atelets Induced by Aggregated Immunoglobulins of Different Classes and Subclasses - J. clin. Invest. 52, 1282, 1973. 32) HEREMANS J. F.: Immunoglobulin A - In: SELA M. (Ed.): The Antigens. Academic Press, New York, 1974; vol. 2, p. 365. 33) HILL R. L., DELANEYR., FELLOWSR. E., LEBOVITZH. E.: The Evolutionary Origins of the Immunoglobulins - Proc. nat. Acad. Sci. (Wash.) 56, 1762, 1966. 34) ISmZAKAK., ISHIZA~CAT., SUCAHARAT.: Biologic Activity of Soluble Antigen-Antibody Complexes. VIII. Complexes of Chicken Antibodies - J. Immunol. 91,257, 1963. 35) JEm~YL. M., KUNKELH. G.: Special Characteristics of the IRA2 Subclass - In: MESTECKYJ., LAWTONA. R. (Eds): The ImmunoglobulinA System. Plenum Press, New York and London, 1973; p. 151. 36) KOUAYASHIK., VAERMANJ.-P., I~REMANS J. F.: Studies on Human Secretory IRA. II. Comparative Studies on a Fragment of Secretory Component Derived from Secretory IgA and Fragments Obtained by Enzymatic Digestion of Free Secretory Component - Immmaochemistry 10, 73, 1973. 37) LAWRENCED. A., WEIGLV.W. O., SPIEGELUERCH. L.: ImmunoglobulinsCytophilic for Human Lymphocytes, Monocytes, and Neutrophils - J. clin. Invest. 55, 368, 1975. 38) M~NIATIS T., JEFFREY A., KLEID D.: Nucleotide Sequence of the Rightward Operator of Phage Lambda - Proc. nat. Acad. Sci. (Wash.) 72, 1184, 1975. 39) MA~CHALONISJ. J., EDELMANG. M.: Phylogenetic Origins of Antibody Structure. 3. Antibodies m the Primary Immune Response of the Sea Lamprey, Petromyzon marinus - J. exp. Med. 127, 891, 1968. 40) MENDEZE., FRANGIONEB.: Chemical Typing of Human Immunoglobulins E and D - FEBS Letters 33, 4, t973. 41) MENDF.ZE., FRAN~IONEB., FRANKLINE. C.: Structure of Immunoglobulin A. Cysteine-Contailing Peptides of the Alpha Chain of an Immunoglobulin A-1 Myeloma Protein - Biochemistry 12, 5186, 1973. 42) MEXLE~E., KARLINL., MATSUMOTOS.: The Valency of Human Gamma-M Immunoglobuhn Antibody - J. biol. Chem. 243, 386, 1968. 203

THE I g HINGE REGION

43) MICHAELSENT. E.: Evidence of 15 S-S Bridges in the Hinge Region of Human IgG3 - Scand. J. Immunol. 2, 523, 1973. 44) MICHAELSENT. E., NATVIGJ. B.: The Hinge Region of IgG3, an Extended Part of the Molecule - FEBS Letters 28, 121, 1972. 45) MIHKBSCOE., MIGLIERIN,~,R.: The Hinge Peptide of an IgA2m(2) Myeloma Protein - In: MESTECKY J., LAWTONA. R. (Eds): The Immunoglobulin A System. Plenum Press, New York and London, 1973; p. 211. 46) MILSTEINC., ADETUGBOK., COWANN. J., SECHERD. S.: Clonal Variants of Myeloma Cells In: BRENTL., HOLUOROWJ. (Eds): Progress in Immunology II. American Elsevier Publishing Co., Inc., New York, 1974; vol. 1, p. 157. 47) MILSTEINC., FRANGIONEB.: Disuifide Bridges of the Heavy Chain of Human Immunoglobulin G2 - Biochem. J. I21,217, 1971. 48) NISONOFFA., WISSLERF. C., LIPMANL. N., WOERNLEYD. L.: Separation of Univalent Fragments from the Bivalent Rabbit Antibody Molecule by Reduction of Disulfide Bonds - Arch. Biochem. 89, 230, 1960. 49) PARKHOUSER. M. E., ASKONASB. A., DOURMASHKINR. R.: Electron Microscopic Studies of Mouse Immunogtobutin M; Structure and 1KeconstimtionFollowing Reduction - Immunology 18, 575, 1970. 50) PINCTERICL., PAINTERR. H., CONNELLG. E.: Uitrastructure of the Fc Fragment of Human IgG - Immunochemistry 8, 1041, 1971. 51) PLATa"T., YANOFSKYC.: An Intercistronic Region and Ribosome-Binding Site in Bacterial Messenger RNA - Proc. nat. Acad. Sci. (Wash.) 72, 2399, 1975. 52) PLAUa"A. G., WISTAg R., Jr., CAPRAJ. D.: Differential Susceptibility of Human IgA Immunoglobulins to Streptococcal IgA Protease - J. din. Invest. 54, 1295, 1974. 53) POLJAKR. J., AMZELL. M., AvEy H. P., CHEN B. L., PHIZACKERLEYR. P., SAUL F.: ThreeDimensional Structure of the Fab' Fragment of Human Immunoglobulinat 2.8 A Resolution Proc. nat. Acad. Sci. (Wash.) 70, 3305, 1973. 54) PORTERR. R.: The Hydrolysis of Rabbit T-Globulin and Antibodies with Crystalline Papain Biochem. J. 73, 119, 1959. 55) Pa'ASHNEM.: Repressor, Operators, and Promotors in Bacteriophage Lambda - In: The Harvey Lectures, 1973-1974. Academic Press, New York, 1975; p. 143. 56) PUTNAM F. W., FLOgENT G., PAUL C., SmNODA T., SHIMIZU A.: Complete Amino Acid Sequence of the 1~Heavy Chain of a Human IgM Immunoglobulin - Science I82, 287, 1973. 57) PUTNAM F. W., SHIMIZU A., PAUL C., SHINODA T.: Variation and Homology in Immunoglobulin Heavy Chains - Fed. Proc. 31, 193, 1972. 58) Rowe D. S., HuG K., FORNI L., PEgNIS B.: Immunoglobulin D as a Lymphocyte Receptor J. exp. Med. 138, 965, 1973. 59) SALUKP. H., CLEM L. W.: The Unique Molecular Weight of the Heavy Chain from Human IgG3 - J. Immunol. I07, 298, 1971. 60) SARMAV. R., DAVIESD. R., LABAWL. W., SILVERTONn. W., TERRYW. D.: Crystal Structure of an Immunoglobulin Molecule by X-Ray Diffraction and Electron Microscopy - Cold Spr. Hark Syrup. quant. Biol. 36, 413, 1971. 61) SPIEGELBERGH. L.: NH2-Terminal Amino Acid Sequence of the Fc Fragment of IgD Resembles IgE and IgG Sequences - Nature (Lond.) 254, 723, 1975. 62) STaNWORTHD. R., HUMPHreY J. H., BENNICHH., JOHaNSSON S. G. O.: Specific Inhibition of the Prausnitz-K/istuer Reaction by an Atypical Human Myeloma Protein - Lancet 2, 330, 1967. 63) STaNWORTHD. R., HUMPHREY J. H., BENNICHH., Jo~SSON S. G. O.: Inhibition of the Prausnitz-Kfismer Reaction by Proteolytic-Cleavage Fragments of a Human Myeloma Protein of ImmunoglobuIin Class E - Lancet 2, 17, 1968. 64) TORCmAD. A., L~RLA J. R., Jr.: Mole~dar Mobility of Polypeptides Containing Proline as Determined by 13C Magnetic Resonance - Biopnlymers I3, 97, 1974. 65) TUr,NER M. W., BENNICHH., NATVIGJ. B.: Pepsin Digestion of Human G-Myeloma Proteins of Different Subclasses. I. The Characteristic Features of Pepsin Cleavage as a Function of Time - CAin. exp. ImmunoI. 7, 603, 1970. 66) VALENTINER. C., GREENN. M.: Electron Microscopy of an Antibody-Hapten Complex - J. moIec. Biol. 27, 615, 1967. 67) VAN BOXSLJ. A., PAULW. E., TEERYW. D., GREEN I.: IgD-Bearing Human Lymphocytes J. Immunol. I09, 648, 1972. 68) VITETTAE. S., MELCHER U., McWILLIaMS M., LaMM M. E., PHILLIPs-QuAGLIATAJ. M., UHR J. W.: Ceil Surface Immunoglobulin. XI. The Appearance of an IgD-Like Molecule on Murine Lymphoid Cells during Ontogeny - J. exp. Med. I41,206, 1975.

204

J.B. ADLERSBERG

69) WARNERN. L.: Membrane Immunoglobulins and Antigen Receptors on B and T Lymphocytes - Advanc. Immunol. 19, 67, 1974. 70) WATANABES., BARNIKOLH. U., HORN J., BERTRAM J., HILSCHMANN N.: Die Prim~rstruktur eines monoklonalen IgM-Immunoglobulins (Makroglobulin Gal.). II. Die Aminosiiuresequenz der H-Kette (t.t-Typ), Subgruppe H III. Struktur des gesamten IgM-Molekiils - Hoppe-Seyler's Z. physiol. Chem. 354, 1505, 1973. 71) WOLFENSTEIN-TODELC., FRANGIONEB., FRANKLINE. C.: Structure of Immunoglobulin A. II. Sequence around the Hinge Region and Labile Disulfide Bonds of an Immunoglobulin A2 Myeloma Protein - Biochemistry 11, 3971, 1972. 72) WOLFENSTEIN-TODELC., MENDEZ E., PRELLI F., FRANGIONEB., FRANKLINE. C.: Disulfide Bridges of Human Immunoglobulins, A1 and A2 - In: MESTECKYJ., LAWTONA. R. (Eds): The Immunoglobulin A System. Plenum Press, New York and London, 1973; p. 257. 73) WOLFENSTEIN-ToDELC., MIHAESCOE., FRANGIONEB.: 'Alpha Chain Disease' Protein Def: Internal Deletion of a Human Immunoglobulin A1 Heavy Chain - Proc. nat. Acad. Sci. (Wash.) 71,974, 1974. 74) WOLFENSTEIN-TODELC., PRELLI F., FRANGIONEB., FRANKLINE. C.: Immunoglobulin A. Arrangement of Disulfide Bridges in the 'Hinge' Region of an Immunoglobulin A1 Human Myeloma Protein - Biochemistry 12, 5195, 1973.

Requests/or reprints should be addressed to: JAY B. ADLERSBERG Department of Medicine Albert Einstein College of Medicine of Yeshiva University 1300 Morris Park Avenue, Bronx, N.Y. 10461 - USA

NOTE ADDED IN PROOFS Recently, the number of duplications of amino acid sequence in the IgG3 hinge region has been more dearly demonstrated [Fr,ANaIO~ B.: A New Immunoglobulin Variant: y3 Heavy Chain Disease Protein CHI - Proc. nat. Acad. Sci. (Wash.) 73, 1552, 1976]. Based on the molecular weights and amino acid compositions of a cyanogen bromide fragment and a peptic peptide from T3 Heavy Chain Disease protein CHI, it appears that the second series of 15 residues shown in the T3 sequence of fig. 4 (page 197) is repeated identically 3 times.

205