`May 1969
`
`Number 1
`
`',1]
`
`~ . ·:r
`r rrr.
`.
`Proceedinns of the
`National Academy
`of Sciences
`of the United States of America
`
`* * * * *
`
`PFIZER EX. 1591
`Page 1
`
`
`
`The Proceedings of the National Academy of Sciences
`OFFICERS OF THE ACADEMY
`FREDERICK SEITZ
`President
`G. B. KISTIAKOWSKY
`Vice President
`HARRISON s. BROWN
`Foreign Secretary
`l\IERLE A. TuvE
`Home Secretary
`E. R. PIORE
`Treasurer
`
`c. B. ANI<'INSEN
`H. W. BoDE
`BERNARD D. DAVIS
`CARL DJERASSI
`
`EDITORIAL BOARD OF THE PROCEEDINGS
`JoHN T. EDSALL, Chairman
`WALTER KAUZMANN, Vice Chairman
`JAMES V. NEEL, Vice Chairman
`:\lERLE A. TuvE, Home Secretary
`HARRISON S. BROWN, Foreign Secretary
`E. R. PIORE, Treasurer
`KATITERINE EsAu
`THEODORE T. PucK
`H. L. SHAPIRO
`WILLIAM A. FOWLER
`EMIL L. SMITH
`HERBERT FRIEDMAN
`JOSEPH L. WALSH
`GORDON J. F. MACDONALD
`JoSEPHINE A. WILLIAMS, Editorial Associate
`Editorial correspondence should be addre::;sed to the PROCEEDINGS OF THE
`NATIONAL AcADEMY OF SciENCES, 2101 Constitution Avenue, Washington, D.C.
`20418.
`Business correspondence should be addressed to the Printing and Publishing
`Office, N"ational Academy of Sciences, 2101 Constitution Avenue, Washington,
`D.C. 2041 '.
`Subscriptions may be entered only for calendar years. The current prices are:
`$25.00 for one year, $47.50 for two years, and $69.00 for three years. The
`price of a single i::;sue (available from the XAS) beginning with volume 59 (1968)
`is $3.50; volumes 49 (1963) through 58 (1967), $2.50. Volumes 1 through 4R
`are available from Johnson Reprint Corporation, 111 Fifth Avenue, New York,
`New York 10003; prices are available from that organization upon request.
`Subscribers are requested to notify the Printing and Publi::;hing Office of the
`Academy and their local postmaster immediately of change of address. Notices
`to the Academy should provide both the old and the new address.
`Microfilms of complete volumes of this journal arc available to regular sub(cid:173)
`scribers only and may be obtained from University ~Iicrofilms, 313 N. First
`Street, Ann Arbor, .;\lichigan.
`
`Second-class postage paid at Easton, Pennsylvania
`
`PRlNTED JN THE U. 8. A.
`
`THE PnocEEDrNos OF TilE N.~TJONAL AC.\DEMY OF SctENCEs
`
`is published monthly by
`
`THE N ,\TJON.U, ACADEMY OF SCIENCES
`
`PFIZER EX. 1591
`Page 2
`
`
`
`THE COVALENT STRUCTURE OF
`AN ENTIRE 'YG IMMUNOGLOBULIN MOLECULE*
`
`BY GERALD M. EDELMAN, BRUCE A. CuNNINGHAM, W. EINAR GALL,
`PAUL D. GoTTLIEB, URs RuTISHAUSER, AND MYRoN J. WAXDAL
`
`THE ROCKEFELLER UNIVE RSITY
`
`Communicated by Theodore Shedlovsky, March 21, 1969
`Abstract.-The complete amino acid sequence of a human 'YGl immunoglobu(cid:173)
`lin (Eu) has been determined and the arrangement of all of the disulfide bonds
`has been established. Comparison of the sequence with that of another myeloma
`protein (He) suggests that the variable regions of heavy and light chains are
`homologous and similar in length. The constant portion of the heavy chain con(cid:173)
`tains three homology regions each of which is similar in size and homologous to
`the constant region of the light chain. Each variable region and each constant
`homology region contains one intrachain disulfide bond. The half-cystines
`participating in the interchain bonds are all clustered within a stretch of ten
`residues at the middle of the heavy chains.
`These data support the hypothesis that immunoglobulins evolved by gene
`duplication after early divergence of V genes, which specified antigen-binding
`functions, and C genes, which specified other functions of antibody molecules.
`Each polypeptide chain may therefore be specified by two genes, V and C, which
`are fused to form a single gene (translocation hypothesis). The internal homol(cid:173)
`ogies and symmetry of the molecule suggest that homology regions may have
`similar three-dimensional structures each consisting of a compact domain which
`contributes to at least one active site (domain hypothesis). Both hypotheses
`are in accord with the linear regional differentiation of function in antibody
`molecules.
`
`Antibodies or immunoglobulins can interact with a wide range of different
`antigenic determinants and, after specific binding to an antigen, they play a
`fundamental part in physiological functions of the immune response. The
`specificity of antigen binding depends ultimately upon amino acid sequences of
`the variable or V regions of antibody molecules. It is the diversity of these
`sequences which results in the range of specificities required for a selective im(cid:173)
`mune response.
`In contrast, other regions of the anti,body molecule have rela(cid:173)
`tively constant sequences and are responsible for physiological functions. Like
`enzymes, these C regions appear to have evolved for a restricted set of interac(cid:173)
`tions. This unusual picture of intramolecular differentiation has emerged from
`studies of the structure of immunoglobulins from different animal species. 1 To
`date, only portions of immunoglobulin molecules have been subjected to amino
`acid sequence determination.
`We now report the amino acid sequence of an entire human 'YGl immuno(cid:173)
`globulin (molecular weight 150,000), the location of all disulfide bonds, the
`arrangement of light and heavy chains, and the length of the heavy chain V
`region.
`
`78
`
`PFIZER EX. 1591
`Page 3
`
`
`
`VoL. 63, 1969
`
`BIOCHEMISTRY: EDELMAN ET A L.
`
`79
`
`Materials and Methods.- The isolation of the myeloma protein Eu 2 and the preparation
`of its CNBr fragments3
`• 4 have been described. Similar methods were used for the isola(cid:173)
`tion of the 'YG 1 myeloma protein He and for the preparation of its CNBr fragments.
`We have previously described the methods used for enzymatic digestion with trypsin,
`chymotrypsin, and pepsin, gel filtration, ion exchange chromatography, high voltage paper
`electrophoresis, determination of NH2-terminal and COOH-terminal residues, amino acid
`analysis, and determination of amino acid sequences by the dansyl-Edman procedure. 4- 8
`The positions of glutamine and asparagine were assigned 9 by determining the electro(cid:173)
`phoretic mobility of peptides and by amino acid analysis of the peptides after enzymatic
`hydrolysis. The half-cystinyl residues contributing to each intrachain disulfide bond
`were determined 10 using the diagonal electrophoresis method. 11
`Results.-The organization of the whole molecule is shown in Figure 1; an un(cid:173)
`equivocal proof of the arrangement of the two identical light chains and two
`identical heavy chains has already been given. 4 Each light chain is linked to its
`neighboring heavy chain by a disulfide bond between corresponding half-
`
`Heavy
`chain
`
`Ligh t
`chain
`
`Fab (t )
`
`-5-5
`VI
`5- 5
`VII
`VIII
`
`Fab (t )
`(":::~=~~~~~iij
`~·
`11 --s
`5111
`~v 51-v -
`5-
`1
`5--s
`I I
`5 5 -- 5
`&t+NH1
`I
`
`1 1
`IX , rr 5 5
`
`CHO
`
`5 5
`
`XI
`
`COOH
`
`COOH
`
`Fc( t )
`
`FIG. L- Over-all arrangement of chains and disulfide bonds of , Gl immunoglobulin
`Eu. Ha!f-cystinyl residues are numbered I- XI; numbers I- V designate corresponding
`residues in light and heavy chains. PCA: pyrollidonecarboxylic acid. CHO: carbo(cid:173)
`hydrate. "Fab(t)" and " Fc(t)" refer to fragments produced by trypsin, which cleaves
`the heavy chain as indicated by dashed lines above half-cystinyl residues VI. VH, VL:
`variable regions of heavy and light chains, CL: constant region of light chain. CHl,
`CH2, CH3: homology regions comprising CH or constant region of heavy chain.
`
`cystines V. Half-cystines VI and VII form bonds linking the half-molecules via
`the heavy chains. Trypsin cleaves the molecule at lysyl residue 222 to form two
`Fab(t) and one Fc(t) fragments. 2 • 5
`There are several strikingly linear arrangements in the primary structure.
`From their amino termini to half-cystines V, the light and heavy chains can be
`aligned or put in register. The intrachain disulfide bonds are linearly and period(cid:173)
`ically disposed. 12 • 13
`In accord with the alignment of light and heavy chains,
`corresponding intrachain disulfide bonds are in similar positions and the disulfide
`loops are of approximately the same size.
`Previous studies7 have suggested that V regions of light and heavy chains have
`similar lengths and begin at the NH2-termini; this will be confirmed below.
`The CL region of the light chain has the same length as V L, but the CH region of
`the heavy chain is about three times as long. CR may be divided into three
`
`PFIZER EX. 1591
`Page 4
`
`
`
`80
`
`BIOCHEMISTRY: EDELMAN ET A L .
`
`Pnoc. N. A. S.
`
`homologous regions of approximately equal length: CHI , Cu2, and Cn3 (Fig. 1).
`We have already reported the amino acid sequence of t he fir t 77 and the last
`224 residues8 of the heavy chain as well as the partial ·equence of the entire ligh t
`chain.6 The complete amino acid sequence of t,he light chain (214 residues) is
`shown in Figure 2. Po itions of the half-cystinyl residues may be compared
`with Figure 1 and the methionyl residue· may be correlated with previous studie
`on the CNBr fragments of Eu .3 • 4 The variable region extends through residue
`108. In accord with other studies/ valine 191 i related to the Inv specificity.2
`The complete sequence of the heavy chain (446 residues) is pre ented in Figure
`3 which may be compared with Figure 2 for alignment with the light chain
`sequence.
`I solation of a single glycopeptide8 indicated that the polysaccharide
`
`20
`10
`1
`AS P- I LE- GLN- MET- THR - GLN- SER- PRO- SER- THR - LEU - SER -ALA- SER- VA L- GLY -ASP-ARG- VAL- THR -
`
`30
`40
`ILE - THR-~- ARG - A LA-SER-G L N-SER- ILE-ASN-THR- TRP- LEU- ALA- TRP-TYR- GLN-GLN- LYS-PR0-
`
`60
`50
`GLY- LYS - ALA- PRO- LYS- LEU - LEU - MET - TYR- LYS-ALA- SER- SER - LEU- GLU- SER- GL Y- VAL- PRO- SER-
`
`80
`70
`AR G- PH E- I LE- GLY- SER- GLY- SER- GLY- TH R- GLU- PHE- THR- LEU- THR- I LE-SER- SER- LEU- GLN- PRO-
`
`100
`90
`ASP-ASP - PHE - ALA- THR-TYR- TYR-~- GLN - GL N -TYR-ASN-SER-ASP-SE R - L YS-M ET- PHE-G L Y-GLN-
`
`120
`110
`G L Y -THR ~ L YS -VAL- G L U-VAL- LYS-GLY -T HR-VAL - ALA-ALA-PRO-SER-VA L -PHE - IL E-PHE-PRO- PR0-
`
`140
`130
`SE R- AS P- G L U-G L N - LEU-LYS-SE R -GLY-THR-ALA - SER-VA L -VAL -~- L EU- L EU-ASN-ASN-PHE- T YR -
`
`160
`150
`PRO - ARG- GL U-ALA- LYS-VAL- GLN- TR P-LYS -VAL- ASP - ASN-ALA- LEU - GLN-SER - GLY-ASN-SER-G LN-
`
`180
`170
`GLU-SE R-VAL - THR - GLU-G LN-ASP-S ER - LYS - ASP - SER - THR - TYR-SER -L EU-SER - SER- TH R- LEU- THR-
`
`200
`190
`L EU-SER-LYS-ALA-ASP-TYR -G L U- L YS-H I S- L YS - VA L - TYR-ALA-~-G L U-VAL -THR-H I S-GLN - G L Y-
`
`214
`210
`LEU- SE R- SER- PRO-VAL- TH R- LYS- SER- PHE -ASN - ARG - GLY- GLU -~
`
`F1 0. 2.-Complete amino acid equence of the Eu light chain . H alf-cystinyl residues are in
`boxes and methionyl residues are underlined.
`
`In a previous tudy8 we
`portion of the molecule is attached at Asx residue 297. 14
`have . uggested that glutamyl residue 356 and methionyl residue 35 may be
`associ ated with Gm 1 specificit ies. The sequence of Eu (Gm 4+ ) between
`residues 211- 252 can be compared with the partial sequence of immunoglobulin
`D aw 17 (Gm 4- ). The presence of arginine in po ition 214 of Eu and lysine in a
`comparable position of Daw may be associat ed with their Gm 4 specificities.18
`Of particular ignificance is the determination of the point at which V n ends
`and Cn begins. A CNBr fragment comparable to fragment IL wa
`isolat ed
`from myeloma protein He which has the same Gm specificity a protein Eu. The
`sequence of t he amino terminal portion of the CNBr fragment from He differed
`from that of the H 4 fragment from Eu. 21
`
`PFIZER EX. 1591
`Page 5
`
`
`
`VoL. 63, 1969
`
`BIOCH EMI STRY: EDELMAN E 'J' AL.
`
`1
`
`20
`10
`1
`PCA-VAL-GLN-LEU-VAL-GLN-SER-G L Y-ALA-G L U-VAL-LYS- L YS-PRO-G L Y-SER-SER-VAL-LYS-VAL-
`
`40
`30
`-
`SER -~- LYS -ALA- SER -GLY- GLY- THR- PHE- SER - ARG- SER - ALA - I L E-I L E- TRP- VAL- ARG- GLN -A L A-
`so
`60
`PRO- GLY- GLN- GL Y- L EU- GL U -TRP- MET-GLY- GLY- I L E- VA L- PR O - MET - PHE -G L Y- PRO- PRO-ASN- TYR-
`
`80
`70
`ALA - GLN- LYS - PHE- GLN-GLY-ARG-VAL- THR- ILE- THR -ALA -A SP- GLU- SER- THR -ASN- THR -AL A- TYR-
`
`100
`90
`MET- GLU- LEU- SER- SER - LEU- ARG- SER- GLU-ASP- THR- ALA- PHE- TYR- PHE -~-ALA- GL Y- GLY - TYR-
`
`120
`110
`GLY- ILE- TYR- SER - PRO -GL U- GLU - TYR- ASN-GLY - GLY- LEU-VAL - THR -VAL- SE R- SER -ALA- SE R- THR-
`
`140
`130
`LY S - GLY- PRO- SER -VAL - PHE- PRO- LE U- ALA- PRO- SER- SE R -L YS- SER- THR- SER -GLY- GLY- THR- ALA -
`
`160
`150
`ALA- LE U- GLY -§LEU -VAL-L YS- ASP- TYR- PHE- PRO- GLU- PRO -VAL- THR- VA L - SE R- TRP- ASN- SER -
`
`130
`170
`GLY-ALA-LEU-THR-SER-GLY-VAL-H I S-THR-PHE-PRO-ALA-VAL - L EU-G L N-SER-SER-GL Y- L EU- TYR -
`
`,l.Q.ll.,
`190
`SER -L EU - SE R - SER -VAL -VAL- TIIR -VAL- PRO- SER- SER- SER- LE U- GLY- THR- GLN- THR- TYR- ILE-~
`
`220
`210
`A SN-VA L- ASN- HI 5-LYS- PRO- SER -ASN- THR -LYS- VAL-ASP- LYS-ARG-VAL- GL U- PRO- LYS- SER - §
`
`240
`230
`A SP- LYS - THR- HIS- TIIR -~-PRO- PRO-~- PRO-ALA- PRO- GL U- LE U - LE U - GLY- GLY- PRO- SE R-VA L-
`
`260
`250
`PHE-LEU- PHE- PRO- PRO- LY S - PRO -LY S- A SP- THR -LEU- MET- I L E-SER -ARG- THR- PRO- GLU-VAL- THR-
`
`280
`270
`§vAL -VAL- VA L-A SP'VAL- SE R-HIS-GL U- A SP-PRO-G L N-VA L - L YS-PHE-ASN-TRP-TYR-VAL-ASP-
`
`300
`290
`GL Y-VAL - GLN - VAL -Ill 5 -ASN -ALA- L YS- THR -LYS- PRO-ARG-G L U -G L N -G L N- T YR -ASX- SER- THR - TYR -
`
`320
`3 10
`ARG-VAL-VAL- S[R-VAL- L EU-THR-VAL-LEU-HIS-GL N-ASN- TR P- L EU-ASP - GLY- L YS-GL U-TYR- L YS-
`
`340
`330
`0}-L YS-VA L- SER -ASN -LY S -ALA -LEU- PRO-A L A- PRO- I LE - GL U -L YS- THR- I LE- SER-LYS-ALA- L YS-
`
`360
`350
`GLY-GLN -PRO -AR G-GLU -PRO-GLN-VAL-TYR-THR-LEU-PRO - PRO - SER-ARG-GLU-GLU-MET-THR- L YS-
`
`330
`370
`A SN- GLN- VAL- SER- LEU- THR -§L EU- VA L - L YS- GL Y- PHE- TYR- PRO- SER -ASP- I LE- !IL A- VAL - GL U-
`
`400
`390
`TRP-GLU-SER-ASN-ASP-GLY-GLU-PRO-GLU-ASN- TYR- L YS- THR - THR-PRO - PRO-VAL-LEU-ASP-SER-
`
`420
`410
`ASP- GL Y · SER- PHE- PHE- LE U- TYR- SER- L YS- LEU - THR -V AL - A SP - L YS - SE R - AR G- TRP - GL N- GL U- GL Y-
`
`44 0
`430
`A SN - VAL -PH E-SE R -~-SER-VAL-MET-HIS-GLU-ALA -LEU- H IS-ASN-HIS- T YR-THR-GLN-LYS-SE R-
`
`446
`LE U-SE R - LE U - SER-PRO-G L Y
`
`FIG. 3.- Complcte am ino acid seq uence of the Eu heavy chain. Hal f-cy~linyl residues are iu
`boxes and methionyl residues are underlined.
`
`A comparison of the seq uence of the two fragments from residue 101 to 121
`is given in Figure 4. The sequences become identical at residue 115 (Eu num(cid:173)
`bering). F urther studies 2 1 confi rmed that from residue 115 to residue 252 the
`sequence of the He fragment was identical to that of Eu H 1.
`In addition,
`tryptic fingerprints of the Fe fragments from Eu and He were identical. These
`
`PFIZER EX. 1591
`Page 6
`
`
`
`2
`
`EU
`
`HE
`
`BIOCHEMJS'l'RY: EDELMAN E'l' AL.
`
`PROC. N. A. S.
`
`1 20
`11 5
`110
`105
`10 1
`-GLY-I LE-TYR-SER-PRO-GLU-GLU-TYR-ASN-GLY-GLY-LEU-VAL-THR-VAL-SER-SER-ALA-SER- THR- LYS -
`
`- THR- LEU -ALA- PHE -ASX - VAL- TR P- GLY- GLX- GLY- TIIR- LYS- VAL -ALA- VAL- SER- SER -ALA- SER- THR- LYS-
`
`FIG. 4.-Comparison of the amino acid sequen ce of the Eu heavy chain from residue 101- 121
`\Yi th the corresponding sequence of the heavy chain of myeloma protein H e.
`
`data suggest that the transition between Va and Ca is located in the vicinity of
`residue 114 (Eu numbering). Studies on a number of additional proteins and a
`search for VH region subgroups7 will be required to locate this point definitively.
`Discussion.-The present tudies provide proof of the covalent structure and
`arrangement of chains in ')'01 immunoglobulin. The half-molecule of Eu is the
`largest protein unit (446 + 214 residues) for which a complete amino acid
`In a protein of this size, one cannot neglect the
`sequence has been determined.
`possibility of small errors in sequence assignment; for this reason we are carrying
`out a number of further checks using various methods of peptide cleavage and
`fractionation.
`One of the most striking features of the immunoglobulin molecule that emerges
`from the completed sequence is the sharp demarcation of its polypeptide chains
`into linearly connected regions that are associated with different functions.
`Variations in the sequences of paired V H and V L regions for the function of
`antigen binding in the selective immune response, and at the same time, conser(cid:173)
`vation of sequence inCH and CL regions for other immunological functions appear
`to require special genetic and evolutionary mechanisms. 22
`The amino acid sequence of Eu provides convincing evidence that the immuno(cid:173)
`globulin molecule evolved by successive duplication of precur or genes. 23 · 24 Our
`analysis of a complete heavy chain has revealed an additional homology region
`(CH1), the structure of which was not previously known. Earlier comparisons7· 8
`of polynucleotide sequences corresponding to both chains of Eu showed evidence
`of homology between V H and V L and homologies among CL, Ca2, and CH3. A
`complete comparison of the amino acid sequences of CL, CH1, CH2, and CH3 is
`given in Figure 5.
`In a stretch of 100 residues, any two regions are identical in
`29 to 34 positions. It is noteworthy that the stretch in the heavy chain from
`residue 221 to 233 which contains the interchain disulfide bonds5 has no homol(cid:173)
`ogous counterpart in other portions of heavy or light chains.
`In the data accumulated so far, little or no homology has been found between
`V and C regions. This prompts the speculation that V genes and C genes
`diverged early in the evolution of antibodies to serve two major groups of func(cid:173)
`tions: antigen recognition functions (ARF) and effector functions (EF) such as
`interaction with cells and complement. The order of emergence of CL genes or
`CH genes from a precursor gene is not apparent from the data. A comparison
`with sequences of Jl., a, o, orE chains may indicate similarities in their VH regions25
`and may show whether any of the homology regions are conserved in constant
`regions of heavy chains of these classes. 26
`Early evolutionary divergence of V and C genes is consistent with the evi(cid:173)
`dence27 · 28 that each chain is specified by two genes, V and C, and the hypothe(cid:173)
`sis7· 22· 29 that V gene episomes are translocated to C genes to form a single VC
`
`PFIZER EX. 1591
`Page 7
`
`
`
`VoL. 63, 1969
`
`lJIOCH EMISTRY: ED ELM AN ET AL.
`
`83
`
`EU CL CRESID.UES 109- 214 )
`EU CHl CRE SID UES 119 -220)
`EU CH2 C RESID UES 234 - 341 l
`EU CH3 C RESID UES 342- 446 l
`
`110
`THR VAL ALA
`S.ER THR LYS
`LE U LEU GLY
`GLN PRO ARG
`
`GL N -
`LEU LYS
`ASP
`SER LYS SER
`THR SER
`PRO LYS ASP THR LE U lilT ILE
`GL U -
`THR
`ARG
`
`ALA LYS
`VAL THR
`PRO
`ILE
`
`GLN
`SER
`
`LYS PHE
`
`LYS
`
`TYR -
`
`THR
`
`PRO VAL
`
`SER ASN ASP -
`
`GLY GL U
`
`GLU
`
`SER
`PRO
`LE U
`ASP
`
`SER SER
`ASN THR
`PRO ALA
`HIS ASN
`
`ASP TYR GLU LYS
`SER LE U GLY THR
`1RP LE U ASP
`GLN GLU
`
`ALA
`ILE
`LYS
`SER
`
`GLU
`ASN
`LYS
`SER
`
`THR
`ASN
`SER
`MET
`
`AS P
`GL U
`GLN
`
`PHE -
`-
`VAL
`ILE
`LEU
`
`210
`ASN
`GL U
`LYS ALA
`LEU SER
`
`GL U
`SER
`
`FIG. 5.-Comparison of the amino acid sequence of CL, CHl, CH2, and Ca3. Deletions indi-
`cated by dashes have been in troduced to maxim ize t he homology. Identical re idues are darkly
`shaded ; both dark and light shad ing are u~ed to indi cate identities whi l'h occur in pairR in the
`same positions.
`
`gene in lymphoid cell precursors. Dreyer and Bennett 30 have previou ·ly sug(cid:173)
`gested a translocation of C genes and, more recently, this has been abandoned in
`favor of a detailed "copy-splice" mechanism. 31 Translocation of genes may be the
`basis of the phenomena of clonal expression and allelic exclusion in antibody
`production .
`I rreversible differentiation and commitment of a lymphoid pre(cid:173)
`cursor cell may thus occur at the time of gene translocation.
`The alignment of disulfide bonds, the arrangement of symmetry axes, a nd the
`fact that proteolytic enzyme cleave the molecule to produce Fab, Fe and 'Fe'
`fragments32 suggest that each homology region may be folded in a compact
`domain 29 stabilized by a single intrachain disulfide bond and linked to neighbor(cid:173)
`ing regions by less tightly folded stretches of polypeptide chain. Such domains
`would have similar but not identical tertiary structures, and each domain would
`
`PFIZER EX. 1591
`Page 8
`
`
`
`84
`
`13JOCHEMISTRY: EDELMAN E7' AL.
`
`PROC. N. A. ~.
`
`contribute to at least one active site mediating a function of that class of im(cid:173)
`munoglobulin. This domain hypothesis is consistent with the hypothesis that
`the molecule evolved by gene duplication as well as with the translocation
`hypothesis. As mentioned above, comparison of the structure and function of
`CH regions in different immunoglobulin classes should reveal whether addition or
`deletion of homology regions and corresponding domains is a major mechanism
`in the evolution of these classe .
`Support for the domain hypothesis would come from finding that limited
`proteolysis of Fab fragments yields fragments containing halves of the Fd frag(cid:173)
`ments. Similar treatment of Bence-Jones proteins may produce V L and CL
`fragments. 33 Additional evidence may come from location in CH of sites for
`complement fixation and skin fixation. Final proof or disproof of this hypothesis
`obviously rests on the results of X-ray crystallographic analysis. It is clear
`that a rotation axis pas es through the disulfide bonds linking the heavy chains
`and there may be an axis of pseudosymmetry between the light and heavy chains.
`The locations of the interchain and intrachain disulfide bonds, the extensive
`homologies, and the alignment of the light and heavy chains with each other
`uggest that the overall relation ·hips described above will be conserved in the
`three-dimensional structure regardless of the details of folding.
`The exact contribution of the variable regions to the antigen combining site
`It is known that the
`must also await analysis of the three-dimensional structure.
`Fab fragment contains both V L and V H regions, and affinity-labeling experi(cid:173)
`ments24 indicate that tyrosyl residues34 · 35 in these regions are directly involved
`in antigen hindi ng. The constancy of the disulfide bonds in V H and V L regions
`and their coordinate location suggest the possibility that the site is fixed by these
`bonds and that variations in the branches of a chain connected by each bond
`may be sterically arranged around the bond as a center. The closely homologous
`CL and CHI regions may serve additionally to stabilize the structure in the face of
`the variation, so that both V Land V H can participate in the site.
`In contrast to the diversity of V regions, the origin of which is still unknown, 22
`the C regions appear to be quite stable in variou animal species. Recent
`studies8 show striking resemblances in the Fe portions of rabbit23 and human
`.,G immunoglobulin.
`In this respect, C regions, like enzymes, may have evolved
`to interact with specific molecules, e.g., those of the complement system. The
`presence of genetic differences in C regions has not so far been related to their
`function, but there is no reason to expect that the origin of variation inC regions
`will differ from other genetic polymorphisms.
`
`We are deeply indebted to Dr. Jack Brook for supplying Eu plasma. We are grateful
`to Miss Joan Low, Miss Catherine Volin, and Mrs. Helvi Hjelt for their expert technical
`assistance.
`
`• Supported by grant GB 6546 from the National Science Foundation and by grant AM
`04256 from the National Institutes of Health.
`1 For a general review, see Cold Spring Harbor Symposia on Quantitative Biology, vol. 32
`(1967), and Nobel Symposium 3: Gamma Glo/ntl-ins, Structure and Control of Biosynthesis, ed.
`J. Killander (Stockholm : Almqvist and Wiksell, 1967).
`2 Edelman, G. M., W. E. Gall, M. J . Waxdal, and W. H. Konigsberg, Biochemistry, 7, 1950
`(1968).
`
`PFIZER EX. 1591
`Page 9
`
`
`
`VoL. 63, 1969
`
`BIOCHEMISTRY: EDELMAN ET AL.
`
`5
`
`3 Waxdal, M. J ., W. H. K onigsberg, W. L. Henley, and G. M. Edelman, Biochemistry, 7, 19.i9
`(196 ).
`• Waxdal, M. J., W. H. Konigsberg, nnd G. M. Edelman, Biochem·istry, 7, 1967 ( HJ68).
`5 Gall, W. E., B. A. Cunningham, 1\l. J. Waxdal, W. H. Konig~berg, and G. M. Edelman,
`Biochemistry, 7, Hl73 (1968).
`6 Cunningham, B. A., P . D. Goltlieb, W. H. K onigsberg, and G. M. Edelman, Biochemist1·y,
`7' 1983 (1968).
`7 Gottlieb, P . D., B. A. Cunningham, M. J. Waxdal, W. H. Konig berg, and G. M. Edelman,
`these PROCEEDINGS, 61, 168 (1968).
`8 Rutishauser, U., B. A. Cunningham, C. Bennett, W. H. Konigsberg, and G. M. Edelman,
`these PRoCEFJDINGS, 61, 1414 (1969).
`9 Bennett, C., W. H. Konigsberg, and G. M. Edelman, manuscript in preparation.
`1o Gall, W. E ., and G. M. Edelman, manuscript in preparation.
`11 Brown, J . R., and B. S. Hartley, Biochern. J., 101 , 214 (1966).
`12 Waxdal, M. J ., W. H. Konigsberg, and G. M. Edelman, in Cold Spring Harbor Symposia on
`Quantitative Biology, vol. 32, 53 (1967).
`13 Pink, J. R. L., and C. Milstein, Nature, 216, 941 (1967).
`14 Inasmuch as the mechanism of attachment of the carbohydrate is unknown, we have not
`specified whether residue 297 is Asp or Asn. This residue begins the tripeptide sequence Asx(cid:173)
`Ser-Thr (residues 297- 299), which is the same as the Asx-X-Thr sequence found at the point
`of attachment of the carbohydrate portion of several other glycoprotein . 15. 1•
`15 Eylar, E. H., J . Theoret. Biol., 10, 89 (1966).
`16 Gatley, B. J., S. Moore, and W. H. Stein, J. Biol. Chern., 244, 933 (1969).
`17 Steiner, L. A., and R. R. Porter, Biochemistry, 6, 3957 (1967).
`tudie . 19 • 20 Additional
`18 The association with Gm 1 specificities is based on definitive
`evidence is required for the correlation of sequence variation with Gm 4 specificities.
`19 Thorpe, N. 0., and H. F. Deutsch,Irnrnunochernistry, 3, 329 (1966).
`20 Frangione, B., C. Milstein, and J. R. L. Pink, Nature, 221 , 145 (1969).
`21 Cunningham, B. A., M. Pflumm, U. Rutishauser, and G. M. Edelman, manuscript in
`preparation.
`22 Edelman, G. M., and J. A. Gaily, in Bmokluwen Symposia in Biology, No. 21, 328 (1968).
`23 Hill, R. L., R. Delaney, R. E. Fellows, J r., and H. E. Lebovitz, these PROCEEDINGS, 56,
`1762 (1966).
`24 Singer, S. J ., and R. F. Doolittle, Science, 153, 13 (1966).
`25 Wikler, M., H. Kohler, T. Shinoda, and F. W. Putnam, Science, 163, 75 (1969).
`If
`26 For generality, we have used CL and Cal, Ca2, etc., to designate homology regions.
`such regions are found in other classes, a more specific nomenclature, e.g. C", C'Y1, C'Y2, etc., and
`perhaps Cl'1, etc., may be required. The same obviously holds for V regions.
`27 Hood, L., and D. E in, Nature, 220, 764 (1968).
`28 Milstein, C., C. P. Milstein, and A. Feinstein, Nature, 22 1, 151 (1969).
`29 Edelman, G. M., and W. E. Gall, Ann. Rev. Biochem., 38, (1969), in press.
`30 Dreyer, W. J., and J . C. Bennett, these PROCEI!:DlNGS, 54, 64 (1965).
`31 Dreyer, W. J., W. R. Gray, and L. Hood, in Cold Spring Hm·bor Symposia on Quant?:tative
`Biology, vol. 32, p. 353 (1967).
`32 Turner, M. W., and H. Bennich, B iochem. J., 107, 171 (196 ).
`33 Berggll.rd, I., and P. Peterson, in Nobel Symposium 3: Gamma Globulins, Structure and
`Control of Biosynthesis, ed. J. Killander (Stockholm: Almqvist and Wiksell 1967), p. 71.
`34 Singer, S. J ., and N. 0. Thorpe, these PROCEEDINGS, 60, 1371 (196 ).
`3• Although there are no direct data on affinity labeling of human antibodies, by analogy (see
`ref. 24), we would suggest that tyrosyl residues 86 of the light chain and 94 of the heavy chain
`may have a function in the active site.
`
`PFIZER EX. 1591
`Page 10
`
`