IMMUNOLOGY
`
`IVAN ROITT
`
`IIn
`
`/% Blecgvielig
`A Scientific: , V
`
`J;/I
`I.
`
`Fublicail,
`BEQ1043
`A
`_
`
`J‘
`
`Page1‘
`
`.
`
`"
`
`BEQ 1043
`Page 1
`
`

`
`SIXTH EDITION
`
`LESSENTIAL
`IMMUNOLOGY
`
`Ivan M. Roitt
`
`MA, DSc(Oxon), FRCPath , Hon MRCP (Lond), FRS
`Professor and Head of
`Departments of Immunology and Rheumatology Research
`University Co llege and Middle sex School of Medicin e
`Unive~sity College
`London WlP 9PG
`
`BLACKWELL SCIENTIFIC PUBLICATIONS
`
`OXFORD LONDON EDINBURGH
`
`BOSTON PALO ALTO MELBOURNE
`
`BEQ 1043
`Page 2
`
`

`
`© 1971, 1974, 1977, 1980, 1984, 1988 by
`Blackwell Scientific Publications
`Editorial offices:
`Osney Mead, Oxford OX2 OEL
`(Orders : Tel. 0865 240201)
`8 John Street, London WC1N 2ES
`23 Ainslie Place, Edinburgh EH3 6AJ
`3 Cambridge Center, Suite 208, Cambridge
`Massachusetts 02142, USA
`667 Lytton Avenue, Palo Alto, California 94301, USA
`107 Barry Street, Carlton, Victoria 3053, Australia
`
`All rights reserved . No part of this publication may be
`reproduced, stored in a retrieval system, or transmitted, in
`any form or by any means, electronic, mechanical,
`photocopying, recording or otherwise without the prior
`permission of the copyright owner
`
`First published 1971
`Reprinted 1972 (twice), 1973 (twice)
`Second edition 1974, Reprinted 1975
`Third edition 1977, Reprinted 1978, 1979
`Fourth edition 1980, Reprinted, 1982, 1983
`Fifth edition 1984
`Sixth edition 1988
`Reprinted 1988
`
`Spanish editions 1972, 1975, 1978, 1982
`Italian editions 1973, 1975, 1979
`Portuguese editions 1973, 1976
`French editions 1975, 1979
`Dutch editions 1975, 1978, 1982
`Japanese editions 1976, 1978, 1982, 1986
`German editions 1977, 1984
`Polish edition 1977
`Greek edition 1978
`Slovak edition 1981
`Indonesian edition 1985
`ELBS editions 1978, 1982, 1988
`
`DISTRIBUTORS
`
`USA
`Year Book Medical Publishers
`35 East Wacker Drive
`Chicago, Illinois 60601
`(Orders: Tel. 312 726-9733)
`
`Canada
`The C.V. Mosby Company
`5240 Finch Avenue East,
`Scarborough, Ontario
`(Orders : Tel. 416-298-1588)
`
`Australia
`Blackwell Scientific Publications
`(Australia) Pty Ltd
`107 Barry Street
`Carlton, Victoria 3053
`(Orders: Tel. (03) 347 0300)
`
`British Library
`Cataloguing in Publication Data
`
`Roitt, Ivan M.
`Essential immunology.- 6th ed.
`1. Immunology
`I. Title
`616.07'9
`
`QR181
`
`ISBN 0-632-01994-8
`
`Set by Setrite Ltd, Hong Kong
`Printed and bound by
`Dah Hua Printing Press Co Ltd, Hong Kong
`
`BEQ 1043
`Page 3
`
`

`
`CHAPTER 3
`
`MOLECULES WHICH
`RECOGNIZE
`ANTIGEN
`
`THE IMMUNOGLOBULINS
`
`The basic structure is a four-peptide unit
`
`The antibody molecule is made up of two identical
`heavy and two identical light chains held together
`by interchain disulphide bonds (figure 3.1). These
`chains can be separated by reduction of the S-S
`bonds and acidification. The exposed hinge region
`is extended in structure due to the high proline
`content and is therefore vulnerable to proteolytic
`attack; thus the molecule is split by papain to yield
`two identical fragments, each with a single combin(cid:173)
`ing site for antigen (Fab; fragment antigen binding),
`and a third fragment which lacks the ability to bind
`antigen and is termed Fe (fragment crystallizable).
`Pepsin strikes at a different point and cleaves the Fe
`from the remainder of the molecule to leave a large
`55 fragment which is formulated as F(ab'h since it
`is still divalent with respect to antigen binding just
`like the parent antibody (figure 3.2).
`The location of the antigen combining sites was
`elegantly demonstrated by a study of purified anti(cid:173)
`bodies to the dinitrophenyl (DNP) group mixed with
`the compound:
`
`31
`
`The two DNP groups are far enough apart not to
`interfere with each other's combination with anti(cid:173)
`body so that they can bring the antigen combining
`sites on two different antibodies together end to
`end. When viewed by negative staining in the
`electron microscope, a series of geometric forms are
`observed which represent the different structures to
`be expected if a Y-shaped hinged molecule with a
`combining site at the end of each of the two arms of
`
`N
`
`s
`I s
`
`c
`heavy
`'------.~----------~
`s
`I s
`
`DNP
`
`DNP
`
`N02-<Q(--NH- CH2CH2CH2CH2CH2CH2CH2CH2 - NH---<J2>-N02
`
`N~
`
`N~
`
`Figure 3.1. Antibody model with two heavy and two light
`polypeptide chains held by interchain disulphide bonds. In
`the diagram the amino-terminal residue(N) is on the left for
`each chain.
`
`BEQ 1043
`Page 4
`
`

`
`-- ·--.--,-------.--
`
`.
`
`. . . _-:_~_:..· . -~-:c::::::::.-=.=-=:
`
`IMMUNOGLOBULIN
`
`IMMUNOGLOBULIN
`
`IMMUNOGLOBULIN
`
`Antigen
`combining~~
`
`s s
`
`PAPAIN
`
`s
`s
`
`s s
`
`PEPSIN
`
`Reduce and acidify
`
`s1te~ ~ ~
`~ ~ ~
`'
`'
`'
`~
`-~ ~ c:::J
`~ H
`
`~ .. .. ~ ~ c:::::::J .. ~ ~
`-
`
`s s
`
`1111
`
`~
`
`SH
`
`32
`
`PAPAIN FRAGMENTS
`
`PEPSIN FRAGMENTS
`
`ISOLATED CHAINS
`
`Figure 3.2. Degradation of immunoglobulin to constituent
`peptide chains and to proteolytic fragments showing
`divalence of pepsin F(ab')2 and univalence of the papain Fab.
`After pepsin digestion the pFc' fragment representing the
`
`C-terminal half of the Fe region is formed and is held together
`by non-covalent bonds. The portion of the heavy chain in the
`Fab fragment is given the symbol Fd.
`
`the Y were to complex with this divalent antigen.
`Triangular trimers, square tetramers and penta(cid:173)
`gonal pentamers may be readily discerned (figure
`3.3). The way in which these polymeric forms arise
`is indicated in figure 3.4. The position of the Fe
`fragment and its lack of involvement in the com(cid:173)
`bination with antigen are apparent from the shape
`of the polymers formed using the pepsin F(ab'h
`fragment (figure 3.3e).
`
`Amino acid sequences reveal variations
`in immunoglobulin structure
`
`For good reasons, the antibody population in any
`given individual is just incredibly heterogeneous,
`and this has meant that determination of amino
`acid sequences was utterly useless until it proved
`possible to obtain the homogeneous product of a
`single clone. The opportunity to do this first came
`from the study of myeloma proteins.
`In the human disease known as multiple mye(cid:173)
`loma, one cell making one particular individual im(cid:173)
`munoglobuhn-di:vides-overan-d-uveragarn-i.n-th
`
`uncontrolled way a cancer cell does, without regard
`for the overall requirement of the host. The patient
`then possesses enormous numbers of identical cells
`derived as a clone from the original cell and they all
`synthesize the same immunoglobulin-
`the mye(cid:173)
`loma or M-protein- which appears in the serum,
`sometimes in very high concentrations. By purifica(cid:173)
`tion of the myeloma protein we can obtain a prep(cid:173)
`aration of an immunoglobulin having a unique
`structure. Monoclonal antibodies can also be ob(cid:173)
`tained by fusing individual antibody-forming cells
`with a B-cell tumour to produce a constantly divid(cid:173)
`ing clone of cells dedicated to making the one
`antibody (cf. figures 2.10 and 7.10}.
`The sequencing of a number of such proteins has
`revealed that theN-terminal portions of both heavy
`and light chains show considerable variability where(cid:173)
`as the remaining parts of the chains are relatively
`constant, being grouped into a restricted number of
`structures. It is conventional to speak of variable
`and constant regions of both heavy and light chains
`(figure 3.5). ·
`Certain sequences in the variable regions show
`quite remarkable diversity and systematic analysis
`lmcatms-rhesetcypervnriable sequences to three
`
`CHAPTER 3
`
`BEQ 1043
`Page 5
`
`

`
`Figure 3.3. (a)-(d) Electron micrograph (X
`1 000 000) of complexes formed on mixing the
`divalent DNP hapten with rabbit anti-DNP
`antibodies. The 'negative stain' phosphotungstic
`acid is an electron-dense solution which penetrates
`in the spaces between the protein molecules. Thus
`the protein stands out as a 'light' structure in the
`electron beam. The hapten links together the ¥(cid:173)
`shaped antibody molecules to form (a) dimers, (b)
`trimers, (c) tetramers and (d) pen tamers (cf. figure
`3.4). The flexibility of the molecule at the hinge
`region is evident from the variation in angle of the
`arms of the 'Y'.
`(e) As in (b); trimers formed using the F(ab ')2
`antibody fragment from which the Fe structures
`have been digested by pepsin (X 500000). The
`trimers can be seen to lack the Fe projections at
`each corner evident in (b). (After Valentine R.C. &
`Green N.M. (1967) J.mol.Biol. 27, 615; courtesy of
`Dr Green and with the permission of Academic
`Press, New York.)
`
`33
`
`segments on the light chain (figure 3.6) and three on
`the heavy chain.
`
`Isotypes
`
`Based upon the structure of their heavy chain con(cid:173)
`stant regions, immunoglobulins are classed into
`major groups termed classes which may be further
`subdivided into subclasses. In the human, for ex(cid:173)
`ample, there are five classes: immunoglobulin G
`(IgG), IgA, IgM, IgD and IgE. They may be differen(cid:173)
`tiated not only by their sequences but also by the
`antigenic structures to which these sequences give
`rise. Thus, by injecting a human IgG myeloma pro(cid:173)
`tein into a rabbit, it is possible to raise an antiserum
`which can be absorbed by mixtures of myelomas of
`
`other classes to remove cross-reacting antibodies
`and which will then be capable of reacting with
`IgG, but not IgA, IgM, IgD or IgE (figure 3.7) .
`Since all the heavy chain constant region (CH)
`structures which give rise to classes and subclasses
`are expressed together in the serum of a normal
`subject, they are termed isotypic variants (table
`3.1). Likewise, the light chain constant regions (Cd
`exist in isotypic forms known as K and A which are
`associated with all heavy chain isotypes. Because
`the light chains in a given antibody are identical,
`immunoglobulins are either K or A but never mixed
`(unless specially engineered in the laboratory).
`Thus IgG exists as IgGK or IgGA, IgM as IgMK or
`IgMA, and so on.
`
`MOLECULES WHICH RECOGNIZE ANTIGEN
`
`BEQ 1043
`Page 6
`
`

`
`.
`
`.
`
`. '-
`
`•
`~- -
`
`-
`
`'
`
`-
`
`-
`
`• -
`
`-
`
`;-
`
`- ~ .. •,- ~ - ~ ""r
`
`\
`
`r;.-
`
`•
`
`' -
`
`•
`.
`
`a
`
`150
`
`HEAVY CHAINS
`
`Fe
`
`~ 100
`:c
`0 g
`
`50
`
`1~
`JU!J~ l[Vtr
`20
`0
`
`ltV ~ Mr'-hH,tli\j
`
`40
`
`60
`Position
`
`80
`
`f-- -
`
`~:-
`
`100
`
`120
`
`~
`
`LIGHT CHAINS
`
`Figure 3.4. Three DNP antibody molecules held together as a
`trimer by the divalent antigen (...._. ). Compare figure 3.3b.
`When the Fe fragments are first removed by pepsin, the
`corner pieces are no longer visible (figure 3.3e).
`
`34
`
`V:o. .
`'Ito.~.
`<~!e
`
`Constant
`
`-
`
`150
`
`~ 100
`:c
`0
`'§
`>
`
`50
`
`J vJI
`JYlnr~ru1J ~' lJutul ~~JV\flJL
`
`50
`Position
`
`75
`
`100
`
`0
`
`25
`
`1-
`
`s
`I s
`
`Figure 3.6. Wu and Kabat plot of amino acid variability in
`the variable region of immunoglobulin heavy and light
`chains. The sequences of chains from a large number of
`myeloma monoclonal proteins are compared and variability
`at each position is computed as the number of different amino
`acids found divided by the frequency of the most common
`amino acid; obviously, the higher the number the greater the
`variability. The three hypervariable regions (blue) in the
`heavy (a) and light chains (b), sometimes referred to as
`Complementarity Determining Regions (CDR), are clearly
`defined. The intervening peptide sequences are termed
`framework regions. (Courtesy of Prof. E.A. Kabat.)
`
`Figure 3.5. Amino acid sequence variability in the antibody
`molecule. • hypervariable; 0 variable; 0 relatively
`constant. The terms 'V region' and 'C region' are used to
`designate the variable and constant regions respectively, 'VL'
`and 'CL' are generic terms for these regions on the light chain
`and 'VH' and 'CH' specify variable and constant regions on the
`heavy chain. As stressed previously, each pair of heavy chains
`are identical, as are each pair of light chains.
`
`Allotypes
`
`This type of variation depends upon the existence
`of aHehc fm ms (encoded by alleles-oraltemativ
`
`genes at a single locus) which therefore provide gen(cid:173)
`etic markers (table 3.1). In somewhat the same way
`as the red cells in genetically different individuals
`can differ in terms of the blood group antigen system
`A, B, 0 , so the Ig heavy chains differ in the expression
`of their allotypic groups. Typical allotypes are the
`Gm specificities on IgG (Gm = marker on IgG)
`which are recognizable by the ability of the indivi(cid:173)
`dual's IgG to block agglutination of red cells coated
`with anti-rhesus 0 bearing the Gm allotype by sera
`from patients w ith rheumatoid arthritis con tainin g
`he appropriate -anti-Gm-rheumat(')id -f-actors (figure
`
`C HA PTER 3
`
`BEQ 1043
`Page 7
`
`

`
`!gG MYELOMA
`
`...
`
`Immunize
`
`... RABBIT ANTISERUM
`
`.,.
`
`ABSORBED BY lgA MYELOMA
`
`.,.
`
`SPECIFIC ANTIBODIES
`
`II
`Ill
`
`ANTI-Id
`
`ANTI-G
`
`ANTI-COMMON
`
`ANTI-Id
`
`ANTI-G
`
`Figure 3. 7. The use of monoclonal myeloma proteins to
`produce antibodies specific for different Ig structures. The
`rabbit makes antibodies directed to different parts of the IgG
`myeloma. Antibodies to those parts which are common to
`other classes can be absorbed out with myelomas of those
`classes leaving antibodies reacting with class-specific G and
`variable region-specific (idiotype; I d) structures on the
`
`· original molecule. By the same token, further absorption with
`other IgG myelomas will remove the common IgG specific
`antibodies leaving an antiserum directed to the idiotypic
`determinants alone. (In an attempt to simplify, I have ignored
`subclasses and allotypes, but the same principles can be
`extended to generating antisera specific for these variants.)
`
`3.8). Allotypic differences at a given Gm locus usual(cid:173)
`ly involve one or two amino acids in the peptide
`chain. Take, for example, the Glm(a) locus on IgGl
`(table 3.1). An individual with this allotype would
`have the peptide sequence: Asp.Glu.Leu.Thr.Lys.
`on each of his IgGl molecules. Another person
`whose IgGl was a-negative would have the se(cid:173)
`quence Glu.Glu.Met.Thr.Lys., i.e. two amino acids
`different. To date, 25 Gm groups have been found
`on they-heavy chains and a further three (the Km(cid:173)
`previously Inv groups) on the K constant region.
`Allotypic markers have also been found on the
`immunoglobulins of rabbits and of mice using re-
`
`agents prepared by immunizing one animal with
`an immune complex obtained with antibodies from
`another animal of the same species. As in other
`allelic systems, individuals may be homozygous or
`heterozygous for the genes encoding the markers;
`these are expressed co-dominantly and are inherit(cid:173)
`ed in simple Mendelian fashion. Take, for example,
`the b4, bS allotypes on rabbit light chains: an animal
`of b 4b 4 genotype would express the b4 allotype
`whereas a rabbit of b 4b 5 genotype derived from
`b 4b 4 and b 5b 5 parents would express the b4 marker
`on one fraction and bS on another fraction of its
`immunoglobulin molecules.
`
`35
`
`VHNL
`IDIOTYPE I
`'
`
`Hypervariable
`(Ag combining site)
`
`ISO TYPE
`
`CH/CL - ISO TYPE
`-'
`
`ALLOTYPE
`
`TYPE OF VARIATION
`
`ISOTYPIC
`
`DISTRIBUTION
`All variants present
`in serum of a normal
`individual
`
`ALLOTYPIC
`
`IDIOTYPIC
`
`Alternative forms:
`genetically controlled
`so not present in all
`individuals
`
`Individually specific
`to each immuno-
`globulin molecule
`
`VARIANT
`
`LOCATION
`
`Classes
`Subclasses
`Types
`Subgroups
`Subgroups
`
`Allotypes
`
`CH
`CH
`CL
`CL
`VHNL
`
`Mainly CH/CL
`sometimes
`VHNL
`
`ldiotypes
`
`Variable
`regions
`
`EXAMPLES
`lgM, lgE
`lgAl, lgA2
`K, >-
`wz +, >-Oz-
`v.r v.n v.m
`VHr VHn VHm
`
`Gm groups (human)
`b4,b5,b6,b9
`(rabbit light chains)
`lgh-1°, lgh-1 b
`(mouse -y20 heavy chains)
`Probably one or more
`hypervariable
`regions forming the
`antigen-combining site
`
`Table 3.1. Summary of immunoglobulin
`variants.
`
`MOLECULES WHICH RECOGNIZE ANTIGEN
`
`BEQ 1043
`Page 8
`
`

`
`..
`
`.
`
`.
`
`---~- -·-~~ ·-
`
`~
`
`Test lgG
`Gml~
`
`•
`
`Gml (a) - ve
`
`!\
`~nr 1'"1D~ n;::,INDING
`'
`'
`8 G G G
`\ l
`mno•~!\~
`
`Gml(a).,.
`
`AGGLUTINATION
`
`NO AGGLUTINATION
`
`'!\'
`
`AGGLUTINATION
`
`Figure 3.8. Demonstration of allotypic specificity on IgG by
`agglutination inhibition. RhD red cells coated with anti-D
`bearing the allotype, are agglutinated by a rheumatoid
`arthritis serum selected for the ability of the rheumatoid
`factor (RF; anti-IgG) to combine with the allotype. If the test
`
`36
`
`IgG added (drawn to show only one combining site and the
`backbone) bears the allotype in question, it will block the
`combining sites on RF which will no longer be able to
`agglutinate the red cells.
`
`Idiotypes
`
`We have seen that it is possible to obtain antibodies
`that recognize isotypic and allotypic variants; it is
`also possible to raise antisera which are specific for
`individual antibody molecules and discriminate
`between one monoclonal antibody and another in(cid:173)
`dependently of isotypic or allotypic structures
`(figure 3.7) . Such antisera define the individual de(cid:173)
`terminants characteristic of each antibody, collec(cid:173)
`tively termed the idiotype (Kunkel & Oudin) . Not
`surprisingly, it turns out that the idiotypic deter(cid:173)
`minants are located in the variable part of the anti(cid:173)
`body associated with the hypervariable regions
`(figure 3. 9) .
`Anti-idiotypes which react with one antibody
`and no other are said to recognize private idiotypes
`and provide further support for the idea that each .
`antibody has a unique structure. Frequently, anti(cid:173)
`body molecules of closely similar amino acid struc(cid:173)
`ture may, in addition, share idiotypes (e.g. MI04
`and Hdex2 in figure 3.9) and we then speak of public
`or cross-reacting idiotypes.
`Anti-idiotypic sera provide useful reagents for
`demonstrating the same V region on different heavy
`chains and on different cells, for identification of
`specific immune complexes in patients' sera, for
`recognition of VL type amyloid in subjects excreting
`Bence-J(')ftes' proteius, for-detectiurt of residual
`
`monoclonal protein after therapy and perhaps for
`selecting lymphocytes with certain surface receptors.
`The reader will (or should be) startled to learn that
`it is possible to raise autoanti-idiotypic sera since
`this means that individuals can make antibodies
`to their own idiotypes. This has quite momentous
`consequences as will become apparent when we
`discuss the Jerne network theory in Chapter 8.
`
`Immunoglobulins are folded into
`globular domains which subserve
`different functions
`
`Immunoglobulin domains have a characteristic
`structure
`
`In addition to the interchain disulphide bonds
`which bridge heavy and light chains, there are
`internal, intrachain disulphide links which form
`loops in the peptide chain. As Edelman predicted,
`the loops are compactly folded to form globular
`domains (figure 3.10) which have a characteristic
`~-pleated sheet protein structure .
`Significantly, the hypervariable sequences appear
`at one end of the variable domain where they form
`parts of the ~-turn loops and are clustered close to
`eadr-otherirr-space.
`
`-
`
`CHAPTER 3
`
`BEQ 1043
`Page 9
`
`

`
`VH_REGION
`Sequential
`numbering
`
`J558
`HDEX 9
`HDEX 31
`HDEX 1
`HDEX2
`
`HDEX3
`HDEX4
`HDEX5
`HDEX 6
`
`M104
`HDEX 8
`HDEX 11
`HDEX 7
`HDEX 14
`
`HDEX 10
`
`20
`
`40
`
`60
`
`80
`
`100
`
`-
`
`NN
`NN r-- -
`NN
`NN
`NN
`
`I
`
`R- - H- N- F
`
`T
`
`v HP-
`
`NN
`NN
`NN
`NN
`
`I
`
`T
`
`NN
`SN
`NN
`cJili_
`KK
`
`KK
`
`ldX I
`
`l
`
`G
`
`ldiotype correlations
`
`ldX
`
`++
`++
`++
`++
`++
`
`++
`++
`++
`++
`
`++
`·+
`++
`++
`
`Tv
`RY
`RY
`NY
`c!!'L
`
`RD = f=
`KD
`SN
`SH
`
`To
`YD
`YD
`AD
`~
`
`VN
`
`l
`
`ldl 1
`
`ldl
`(J558)
`
`ldl
`(M104)
`
`++
`++
`++
`+
`+
`
`++
`++
`++
`++
`++
`
`Figure 3.9. Structural correlates of idiotopes (individual
`determinants on an idiotype) on anti dextran antibodies.
`Amino acid sequences of variable heavy chain regions of
`mouse monoclonal antidextran antibodies are shown. All
`antibodies have ).1 L chains. Lines indicate identity to the
`sequence of the first protein, f558; letters (Dayhoff code) show
`differences or regions correlated with idiotopes (central
`boxed areas). The cross-reacting idiotope (IdX) is associated
`with second hypervariable region (hv2) structures while the
`
`private idiotopes (Idl) are features of the hv3 region in these
`antibodies. The presence of the idiotopes on each antibody
`molecule is assessed by reaction with antisera specific for
`ldX, f558 ldl and M104 ldl (on the right). Cross-reacting
`idiotopes may also be associated with the hv3 region in
`other systems (from J.M. Davie et al. (1986) Ann. Rev.
`Immunol. 4, 147, with permission. © by Annual Reviews
`Inc.).
`
`37
`
`VARIABLE DOMAIN (VL)
`
`CONSTANT DOMAIN (CL)
`
`Figure 3.10. Structure of the globular domains of a light
`chain (from X-ray crystallographic studies of a Bence-fones'
`protein by Schiffler et al. (1973) Biochemistry 12, 4620). One
`surface of each domain is composed essentially of four chains
`(grey arrows) arranged in an anti-parallel (J-pleated
`structure stabilized by interchain H bonds between the amide
`CO. and NH. groups running along the peptide backbone, and
`the other surface of three such chains (blue arrows); the dark
`bar represents the intra-chain disulphide bond. This structure
`is characteristic of all immunoglobulin domains. Of
`
`particular interest is the location of the hypervariable regions
`<• • • a ) in three separate loops which are closely
`disposed relative to each other and form the light chain
`contribution to the antigen binding site (cf. figure 3.11). One
`numbered residue from each complementarity determinant is
`identified. To generate a Fab fragment (cf. figure 3.13), imagine
`a V H- V H' segment just like the V L- CL in the diagram, rotate
`it 180° around the axis of the arrow on the right of the figure
`and lay it on top of VL- CL segment (Dr A. Feinstein).
`
`MOLECULES WHICH RECOGNIZE ANTIGEN
`
`BEQ 1043
`Page 10
`
`

`
`LIGHT
`CHAIN
`
`HEAVY
`CHAIN
`
`(a)
`
`(b)
`
`38
`
`The variable domain binds antigen
`
`The clustering of the hypervariable loops at the tips
`of the variable regions where the antigen binding
`site is localized (figures 3.3 and 3.4) makes them the
`obvious candidates to subserve the function of anti(cid:173)
`gen recognition (figure 3.11) and this has been con(cid:173)
`firmed by X-ray crystallographic analysis (cf. figure
`4.6). The sequence heterogeneity of the three heavy
`and three light chain hypervariable loops ensures
`tremendous diversity in combining specificity for
`antigen through variation in the shape and nature
`of the surface they create. Thus each hypervariable
`region may be looked upon as an independent
`structure contributing to the complementarity of
`the binding site for antigen and perhaps one can
`speak of complementarity determinants.
`That these variable regions of heavy and light
`chains both contribute to antibody specificity is
`suggested by experiments in which isolated chains
`were examined for their antigen combining power.
`In general, varying degrees of residual activity were
`associated with the heavy chains but relatively little
`with the light chains; on recombination, however,
`there was always a significant increase in antigen
`binding capacity.
`Amino acids associated with the combining site
`can be identified by 'affinity labelling'. In this tech(cid:173)
`nique, a hapten (a well-defined chemical grouping
`to which antibodies can be formed, e .g. DNP in
`figure 3.3) is equipped with a chemically reactive
`side-chain which will form covalent links with ad(cid:173)
`jacent amino acids after combination of the hapten
`with antibody, so labelling residues in the neigh(cid:173)
`bourhood of the combining site. A modification
`introduced by Porter and his colleagues utilizes a
`'flick-knife' principle. The hapten with an azide
`side-chain combines with its antibody and is then
`illuminated with ultraviolet light; this converts the
`azide to the reactive nitrene radical which will
`covalently link to almost any organic group with
`which it comes in contact (e.g. figure 3.12). The
`affinity label binds to both heavy and light chains in
`the hypervariable regions.
`
`Figure 3.11. (a) Idealized two-dimensional representation of
`an antigen binding site formed by spatial apposition of
`peptide loops containing the hypervariable regions (hot spots:
`D) on light and heavy chains. Numbers refer to amino acid
`residues. Glycine residues ( ®) are invariably present at the
`positions indicated whatever the specificity or animal species
`of the immunoglobulin. They are of importance in allowing
`peptide chains to fold back and form fJ-pleated sheet
`structures which enable the hypervariable regions to lie close
`to each other (figure 3.10). Wu and Kabat have suggested that
`the flexibility of bond angle in this amino acid contributes to
`the effective formation of a binding site. On this basis the
`greater frequency of invariant glycines on the light chain
`might indicate that coarse specificity for antigen binding
`was provided by the heavy chain and 'fine tuning' by the
`light chain. Through binding to different combinations of
`hypervariable regions and to different residues within each of
`these regions, each antibody molecule can form a complex
`with a variety of antigenic determinants (with a comparable
`variety of affinities). (b) A simulated combining site formed
`throughout the body' their ability to fix comple-
`by apposing the three middle fingers of each hand, each finger
`ment and their binding to cell surface Fe receptors.
`representing a hypervariable loop. (Photograph by B.N.A.
`Since lhe classes all have the same I< and A. ltght~-~ice; tnsptred vyA. Munro!)
`
`Constant region domains determine secondary
`biological function
`
`The classes of antibody differ from each other in
`many respects: in their half-life, their distribution
`
`CHAPTER 3
`
`BEQ 1043
`Page 11
`
`

`
`HAPTEN
`
`...
`
`_With u_v linht
`
`Antibody combining site
`co",
`//
`' cH
`//
`/
`' NH
`f
`CH3
`I
`
`/
`
`/
`
`"' /
`r
`
`9o
`I
`
`:
`I
`
`~N3
`e-J o
`
`NH-LYSINE
`
`'
`CH
`/
`'NH
`CH2
`1
`I
`
`/
`
`chains, and heavy and light variable region do(cid:173)
`mains, these differences must lie in the heavy chain
`constant regions.
`It has been possible to localize these biological
`activities to the various heavy chain domains by
`using myeloma proteins which have spontaneous
`domain deletions, or enzymic fragments produced
`by papain (Fe), pepsin (F(ab'h and pFc', the C(cid:173)
`terminal portion of Fe) and plasmin (Facb from
`rabbit IgG lacks pFc' but retains theN-terminal half
`of Fe). Nowadays, of course, it can all be done by
`genetically engineered proteins.
`A model of the IgG molecule is presented in
`figure 3.13 which indicates the spatial disposition
`and interaction of the domains in IgG and ascribes
`the various biological functions to the relevant
`structures. In principle, the V region domains form
`the recognition unit (cf. figure 2.1) and constant
`region domains mediate the secondary biological
`functions.
`
`co'""',
`
`NH-LYSINE
`
`"J ~ L
`r,,
`
`Figure 3.12. Affinity labelling: the hapten binds to its
`antibody and the azide group activated by ultraviolet light
`loses N 2 forming a reactive radical which combines with an
`in this hypothetical example an
`adjacent amino acid-
`alanine residue. Analysis of the protein after digestion would
`show the alanine to be labelled with the hapten and implicate
`this residue in the combining site. Studies by Fleet G. W.J.,
`Porter R.R. & Knowles J.R. (Nature 1969, 224, 511) indicate
`that the affinity label combines with heavy and light chains
`in a ratio of approximately 3.5:1 in their system.
`
`39
`
`Antigen binding
`
`C4b binding
`
`(a )
`
`(b)
`
`Figure 3.13. The disposition, interaction and biological
`properties of the Ig domains in IgG. (a) Computer generated
`model of IgG. One heavy chain is depicted in mid blue, the
`other in light blue and the light chains in grey. Carbohydrate
`bound to and separating the Cy2 domains is in dark blue.
`The structure was determined by Silverton E. W., Navia M.A.
`& Davies J.R. (1977) Proc. Nat. Acad. Sci. 74, 5140, and the
`figure generated by computer graphics using R.J. Feldmann's
`system (Nat. Inst. Health). (b) Diagram based on the model
`indicating the location of biological function and showing
`apposing domains making contact through hydrophobic
`regions (after Dr A. Feinstein). The structures of these contact
`frameworks are highly conserved, an essential feature if
`different VH and VL domains are to associate in order to
`
`generate a wide variety of antibody specificities. These
`hydrophobic regions on the two complement fixing CH2 (Cy2)
`domains are partly masked by carbohydrate and remain
`independent so allowing the formation of a hinge region
`which is extremely flexible both with respect to variation in
`the angle of the Fab fragments and their rotation about the
`hinge peptide chain. Thus combining sites in IgG can be
`readily adapted to spatial variations in the presentation of
`the antigenic epitopes. The combined Cy2 and Cy3 domains
`bind to Fe receptors on phagocytic cells, NK cells and
`placental syncytiotrophoblast; also to staphylococcal protein
`A. (Note the IgG heavy chain is designated y and the constant
`region domains Cy1, Cy2 and Cy3.)
`
`MOLECULES WH ICH RECOGN IZE ANT IG EN
`
`BEQ 1043
`Page 12
`
`

`
`Immunoglobulin classes and subclasses
`
`The physical and biological characteristics of the five
`major immunoglobulin classes in the human are
`summarized in tables 3.2 and 3.3. The following
`comments are intended to supplement this infor(cid:173)
`mation.
`
`Immunoglobulin G
`
`During the secondary response IgG is probably the
`major immunoglobulin to be synthesized. Through
`its ability to cross the placenta it provides a major
`line of defence against infection for the first few
`weeks of a baby's life which may be further rein(cid:173)
`forced by the transfer of colostral IgG across the gut
`mucosa in the neonate. IgG diffuses more readily
`than the other immunoglobulins into the extra(cid:173)
`vascular body spaces where, as the predominant
`species, it carries the major burden of neutralizing
`bacterial toxins and of binding to micro-organisms
`to enhance their phagocytosis. The complexes of
`bacteria with IgG antibody activate complement
`thereby chemotactically attracting polymorpho(cid:173)
`nuclear phagocytic cells ( cf. p . 11) which adhere to the
`bacteria through surface receptors for complement
`and the Fe portion of IgG (Fey); binding to the Fe
`receptor then stimulates ingestion of micro-organ(cid:173)
`isms through phagocytosis. In a similar way, the
`extracellular killing of target cells coated with IgG
`antibody is mediated largely through recognition of
`the surface Fey by NK cells bearing the appropriate
`
`receptors (cf. p . 26) . The interaction of IgG com(cid:173)
`plexes with platelet Fe receptors presumably leads
`to aggregation and vasoactive amine release but the
`physiological significance of Fey binding sites on
`other cell types, particularly lymphocytes, has not
`yet been clarified. Although unable to bind firmly
`to mast cells in human skin, IgG alone among the
`human immunoglobulins has the somewhat useless
`property of fixing to guinea-pig skin. The thesis
`that the biological individuality of different im(cid:173)
`munoglobulin classes is dependent on the heavy
`chain constant regions, particularly the Fe, is amply
`borne out in relationship to the activities we have
`discussed such as transplacental passage, comple(cid:173)
`ment fixation and binding to various cell types,
`where function has been shown to be mediated by
`the Fe part of the molecule.
`With respect to overall regulation of IgG levels in
`the body, the catabolic rate appears to depend
`directly upon the total IgG concentration whereas
`synthesis is largely governed by antigen stimula(cid:173)
`tion, so that in germ-free animals, for example, IgG
`levels are extremely low but rise rapidly on transfer
`to a normal environment.
`
`Immunoglobulin A
`
`IgA appears selectively in the sero-mucous secre(cid:173)
`tions such as saliva, tears, nasal fluids, sweat, colos(cid:173)
`trum and secretions of the lung, genito-urinary and
`gastro-intestinal tracts where it clearly has the job
`of defendin g the exposed external surfaces of the
`
`40
`
`lgG
`
`75
`
`lgA
`
`I
`75,95, 115*
`
`I
`
`lgM
`
`195
`
`I
`
`lgD
`
`75
`
`I
`
`lgE
`
`85
`
`150 000
`
`1
`
`160 000 and
`dimer
`1' 2*
`
`900 000
`
`185 000
`
`200 000
`
`5
`
`1
`
`1
`
`DESIGNATION
`
`Sedimentation
`coefficient
`Molecular weight
`
`Number of basic
`four-peptide units
`Heavy chains
`light chains K + 1\
`Molecular formulat
`
`'I
`K + A
`
`'12K2,'12A2
`
`2
`
`80
`
`3
`
`a
`K + A
`
`Ca2K2)1 - 2
`Ca 2A)1 - 2
`(a2K2)2 S •
`Ca2A2)2 S •
`2,4
`
`13
`
`8
`
`s
`K+A
`1i2K2(1i2A2 ?)
`
`E
`K+A
`E2K2,E2A2
`
`fJ.
`K+A
`
`CfJ.2K2)5
`CfJ.2A2)5
`
`5(1 0)