`Vol. 86, pp. 607-611, January 1989
`Immunology
`
`Three-dimensional structure of Fab R19.9, a monoclonal murine
`antibody specific for the p-azobenzenearsonate group
`M.-B. LASCOMBE* , P. M. ALZARI*, G. BOULOT*, P. SALUDJIAN*, P. TOUGARD*, C. BEREKt, S. HABAi,
`E. M. ROSENi, A. NISONOFFi, AND R. J. POLJAK*§
`•™panement d'lmmunologie, lnstitut Pasteur, 75015 Paris, France; t1nstitute of Genetics. University of Cologne. Cologne 41. Federal Republic of Gennany;
`and *Rosensttel Research Center. Department of Biology, Brandeis University, Waltham. MA 02254
`
`Contributed by A. Nisonoff, August 8, 1988
`
`The crystal structure of Fab Rl9.9, de.rived
`ABSTRACT
`from an anti-p-azobenzenearsonate monoclonal antibody, has
`been determined and refmed to 2.8-A resolution by x-ray
`crystallographic techniques. Monoclonal antibody Rl9.9
`(lgG2b1e) shares some idiotopes with a major idiotype (CRl.J
`associated with A/ J anti-p-azobenzenearsonate antibodies.
`The amino acid sequences or the variable (V) parts or the heavy
`(V H) and light (V L) polypeptide chains of monoclonal antibody
`Rl9.9 were determined through nucleotide sequencing of their
`mRNAs. The v L region is very similar to that or CRIA-positive
`anti-p-azol>enzenearsonate antibodies as is VH, except for its
`third complementarity-determining region, which is three
`amino acids longer; it makes a loop, unique to Rl9.9, that
`protrudes into the solvent. A large n0mber or tyrosine residues
`in the complementarity-determining region of V 8 and V L t with
`their side chains pointing towards the solvent, may have an
`important function in antigen binding.
`
`Murine antibodies to model antigens have provided valuable
`experimental systems to study the molecular bases of the
`specificity, diversity, and genetic control of immune re(cid:173)
`sponses. The hapten, p-azobenzenearsonate (Ar), has been
`used in several laboratories as a suitable probe for such studies
`(1-5), which have been facilitated by the presence of an
`intrastrain cross-reactive idiotype, designated CRIA, among
`the anti-Ar antibodies ofA/J mice or of closely related strains.
`The expression of CRIA is linked to genetic loci encoding
`heavy (H) chains (6) and light (L) chains (7). On the av~rage,
`about half of the anti-Ar antibodies induced by keyhole limpet
`hemocyanin-Ar in A/J mice share this idiotype. The variable
`(V) regions, Vu and V L• of CRI.A antibodies appear to be
`encoded by single germ-line genes (8, 9), and the diversity (D)
`region is encoded by a variant of the DFL16.1 gene (10). CRI.A
`molecul~s also utilize the V.,10, K chain joining (J) 1, and,
`almost invariably, Ju2 gene segments (4, 5, 11). Idiotype(cid:173)
`expressing antibodies from hyperimmunized mice display
`somatic variants of amino acid seq~ences in each of these gene
`segments (4, 5, 12), whereas the antibodies from an early
`primary response reflect few if any mutations (5). The Vu
`region appears to be somewhat more susceptible to somatic
`variation than V L (4). A disproportionate number of mutations
`in Yu and V L occurs in their complementarity-determining
`regions (CDR); this probably reflects selection by antigen of
`variants with higher affinity (5).
`Among the serum anti-Ar antibodies of immunized A/J
`mice are molecules that carry some but not all of the idiotopes
`associated with CRIA (13, 14). Such antibodies are bound by
`a~ti-CRIA antibodies, but they are unable to completely
`d1s~lac~ labele~ C~IA anti~odies from such anti-idiotype
`ant1bod1es. Ant1bod1es of this type were designated "minor
`idiotypes" (13 , 14). The subject of the present investigation,
`
`The publication costs of this article were defrayed in part by page charge
`payment. This article must therefore be hereby marked " advertisement"
`in accordance with 18 U.S.C. §1734 solely to indicate this fact.
`
`monoclonal antibody (mAb) R19.9 (lgG2bK), has these sero(cid:173)
`logical properties and is thus a member of a minor idiotypic
`anti-Ar family. The L chains of Rl9.9, when combined with
`H chains of a CRI.A mAb, yielded a CRIA product (14).
`However, the converse recombinant (H19.9LcR1) was CRIA
`by the criterion of inhibition in the standard assay for CRIA.
`Amino acid sequences (this paper) indicate that the V "10, J.,l,
`Yu, and Ju2 sequences of R19.9 are closely related to the
`putative germ-line sequences controlling CRIA but that the
`Du sequence (in CDR3 of Ytt) differs markedly ; it is 3
`residues longer than the characteristic Du sequence (11 vs. 8
`residues). There are also three amino acid substitutions in
`CDR2 of Vu that may contribute to the idiotypic variance of
`R19.9.
`The x-ray crystallographic study of Fab R19.9 presented
`here permits a correlation between amino acid sequences,
`idiotypic markers, and the three-dimensional structure of the
`A/J anti-Ar antibodies. Tentative conclusions can also be
`drawn about the conformation of the antigen-combining site
`of anti-Ar molecules.
`
`MATERIALS AND METHODS
`The monoclonal, A/J anti-Ar antibody R19.9 (lgG2bK) was
`prepared as described (15, 16). After papain digestion (17),
`the Fab fragment of Rl9.9 was extensively purified by three
`successive column chromatography steps with Sephadex
`G-100 (Pharmacia), DEAE-cellulose (0. 76 milliequivalent per
`g; Serva, Heidelberg) equilibrated in 0.04 M potassium
`phosphate buffer, and PBE 94 (Pharmacia) for chromatofo(cid:173)
`cusing. In this last step, the Fab was eluted with Polybuffer
`96 (Pharmacia) as a major peak at pH 7.5-7.6. The purified
`Fab ofR19.9 was crystallized at room temperature by vapor
`diffusion in hanging drops (18) or in capillaries against 20%
`(wt/vol) PEG 8000 (Sigma)/0.2 M sodium chloride/3 mM
`sodium azide/0.1 M potassium phosphate, pH 7.3 (19). The
`crystals grow to a size of up to 0.3 mm x 0.4 mm x 1.5 mm.
`They are monoclinic, space group ni. with unit cell di(cid:173)
`mensions a = 43.3 A, b = 80.8 A, c = 75.1 A, f3 = 96°. There
`is one Fab in the asymmetric unit.
`X-ray intensity data were measured on a diffractometer
`with CuKa radiation. Since the crystals of Fab R19.9 are
`polymorphic (19), they were carefully selected to conform to
`the unit cell dimensions given above. Crystals were replaced
`with the intensities of reference reflections decreased below
`70% of their starting values. Integrated intensities were
`obtained by profile fitting (20) and were further corrected for
`Lorentz-polarization factors , absorption (21), and radiation
`decay. For the native crystals, a complete data set to 2.8-A
`
`Abbreviations: Ar, p-azobenzenearsonate; CRIA, major cross(cid:173)
`reactive idiotype associated with anti-Ar antibodies of the A strain of
`mouse; H. heavy; L, light; V, variable; D, diversity; J, joining; C,
`cons!a~t ; ~b, monoclonal antibody; CDR, complementarity-de(cid:173)
`tenmmng region.
`§To whom reprint requests should be addressed.
`
`607
`
`1 of 5
`
`Celltrion, Inc., Exhibit 1060
`
`
`
`608
`
`Immunology: Lascombe et al.
`
`Proc. Natl. Acad. Sci. USA 86 (1989)
`
`Table 1. Crystallographic refinement of Fab Rl9.9
`Actual rms
`deviation, A
`
`0.011
`0.091
`0.085
`0.006
`
`Parameter
`Deviation from ideal distances
`Bond distances
`Angle distances
`Planar 1-4 distances
`Deviation from planarity
`Deviation from permitted constant
`distances
`Single torsion contacts
`0.349
`0.500
`Multiple torsion contacts
`0.465
`0.500
`Possible hydrogen bond
`0.469
`0.500
`R factor :; rrr llFobsl - IFca1cll/r>> IFobsl :; 0.296; number of
`hkf
`hk I
`structure factors F > 2u "" 12,145.
`
`Target u
`
`0.020
`0.060
`0.060
`0.020
`
`polypeptide chain. The amino acid sequence of Fab RI9.9
`was adjusted to the electron density map using the program
`FRODO (26) on a PS300 Evans and Sutherland interactive
`graphics system. The model was refined by the restrained
`least-squares program PROLSQ (27) by using reflections be(cid:173)
`tween IO- and 2.8-A resolution. After 33 cycles of refinement
`(R = 32.4%), the model was partially rebuilt based on
`difference Fourier (2Fca1c - Fobs• aca1c) and "OMIT" (28)
`maps, as well as inspection of the solvent-flattened electron
`density map. The structure was further improved by alter(cid:173)
`nating model building on difference Fourier and OMIT maps
`followed by PROLSQ refinement until reasonable convergence
`was achieved (see Table I). The current R factor is 0.296. No
`attempt was made to place solvent molecules in the model.
`Sequencing of polysomal mRNA for V Hand V" was carried
`out according to Griffiths and Milstein (29) by using primers
`described previously (30, 3I).
`
`RESULTS
`The 3.5-A resolution electron density map described above
`was used to fit the amino acid sequences of the variable
`regions [VH, D, and JH2 and also V" and J,, (see Figs. I and
`2)) and those of the constant regions, CHI of IgG2b and C,,
`(32). Since the sequence of V H was determined for residues
`I6-123 only, the first I5 residues were assumed to be the
`same as those in protein 36-65, which expresses the germ(cid:173)
`line-encoded V H sequence of CRI~ (8). This choice seemed
`reasonable because the same V H germ-line gene is expressed
`in 36-65 and Rl9.9 and because in 11 published N-terminal
`sequences ofCRIA mAbs there is a total of only seven amino
`acid substitutions in this region (33, 34). Although the
`electron density maps gave an indication for the trace of the
`po}ypeptide chain in CDR3 ofVH, the detailed conformation
`of this region was less clear than that of other parts of Fab
`Rl9.9. VH positions 34, 55, 58, 59, and 74 differ from those
`of the germ-line VH gene; however, the electron density map
`showed good agreement with the experimentally determined
`VH sequence (Fig. 1) at those positions. Although the
`a!llbiguity (His/ Asp) at V H position 105 cannot be resolved
`with certainty by the electron density map, aspartic acid
`
`resolution was obtained from 26,284 measurements that were
`merged to yield the intensities of 13,138 independent reflec(cid:173)
`tions (Rsym = 0.074). Three heavy atom derivatives were
`obtained by using phenylmercury(II) acetate, p-hydroxy(cid:173)
`methylbenzenesulfonate, and OsC~. Their intensities were
`measured to 3.5-A [phenylmercury(II) acetate and p(cid:173)
`hydroxymethylbenzenesuJfonate] and 6-A (OsC~) resolu(cid:173)
`tion. After refinement of he~vy atom occupancies, scale
`factors, and thermal (isotropic) and positional parameters, an
`electron density map was calculated to 3.5-A resolution
`(average figure of merit, m = 0.52). The interpretation of this
`map was facilitated by a six-dimensional real space search by
`using the model ofFab New (22) as a search object. A rotation
`function calculation (23) gave a clear indication of the
`orientation of the Fab in the unit cell, in agreem~nt with that
`found by the model search of the electron density map. The
`positions of the V "' V L• CHI, and CL domains were refined
`at 6-A resolution by a constrained-restrained least-square~
`procedure [CORELS (24)). At this s~ge the agreement
`factor, R, between observed (obs) and calcµlated (calc)
`structure factors (F), R = I(Fobs - FcaJc)/I(F0 bs), wa~ 42%.
`A solvent-flattened (25) electron density map at 3.5-A reso(cid:173)
`lution (m = 0.8I, Ac/>= 35°) allowed the tracing of most of the
`
`"'
`
`10
`B V 0 L 0 0 S G A B L V R A G S S V l
`GBlUILillB GAGGTTCAGC'lTCA
`
`119.9
`
`not sequenced
`
`-alll 1 - -
`30
`20
`" S C l A S G T T P T S T G I N
`TGTCCTGCAAGGCITTATACATTCACAAGCTACGGTATAAAC
`v
`
`-------~CDl 2--------
`ro
`ro
`~
`~
`V V l Q R P G Q G L B V I G T I N P G N G T T
`l T N B ~ F
`l G l T T L
`GBlUILillB TGGGTGAAACAGAGGCCI'GGACAGGGCCTGGAA'IGGATrGCATATATTAATCCTGGAAATGGTTATACTAAGTACAA~CCACACTG
`l
`L S
`- - - - - - - - - - - - - - - - - - - - - -G - - - -< :T -GT -T ' - - - - - - - - - - -
`
`R19.9
`
`~
`~
`~
`T V D l S S S T A T " 0 L R S L T S B D S A V T P C A R
`
`GERKLINB ACTGTAGACAAATCCTCCAGCACAGCCTACATGCAGCTCAGAAGCCTGACATCTGAGGACTC1'GCAGrCTAnTCTGTGCAAGA
`R
`
`R19.9
`
`-------------a>RJl-------
`1
`JB2 s ep en t - - - - -
`sepent
`s V T T G G S
`T T F D T V C 0 G T T L T V S S
`TACTACTTtGACTACr'GGGGCCAAGGCCCACTCTCACAGTCTCCTCA
`TCGGTCTACTATGGTCGTAGT
`S
`P
`B/D L A V
`~CCT~'-----c.-------------
`-<:T-
`
`36-65
`
`R19.9
`
`flG . 1. Nucleotide and deduced amino acid sequences of the VH regiqn of mAb Rl9.9. They are compared, between residues 1-98, with
`the VH germ-line sequence believed to encode CRIA molecules (8). The D and JH2 segments are compared with those of mAb 36-65, whose
`VH gene (codons 1-98) corresponds to that of an unmutated genn-line gene (8). A solid l!ne represe~ts identity with the prot.otrpe sequence.
`Amino acid substitutions are indicated above the nucleotide sequences. Gaps have been introduced m the D segment to max1m1ze homology.
`
`2 of 5
`
`Celltrion, Inc., Exhibit 1060
`
`
`
`Immunology: Lascombe et al.
`
`609
`
`VK10
`
`Proc. Natl. Acad. Sci. USA 86 (1989)
`-------<DI 1--------
`w
`w
`~
`I S C R A S Q 0
`I Q N T Q T T S S L S A S L C D R 9 T
`I S N T L N
`0
`CATATCCAGATGACACACAcrACATCCrCCCTCTCTCCCTCTC'roGCACACACAGTCACCA~CATTACCAATTATTTAMC
`?
`?
`1 : - - -
`R19 . 9 - - - - - - . e . - -- - - -- -- - - - - - - - - - ' I
`
`w
`~
`~
`V T Q Q
`I T T T S R L B S C V P S R P S C S C
`L L
`l
`l P 0 C T V
`TCCTATCACCACAAACCAGATCCMCTCTTAMcrCC'l'CATcrAcrACACATCAAGATTACACTCACCACTCCCA~
`?
`I - -1 - -1 -1 - - - - -
`
`-------CDll 2----
`
`ro
`~
`~
`I A T T P C Q Q C N T L P
`I S N L £ Q £ 0
`S C T 0 T S L T
`TCTCCMCACATT ATTcrcrCACCATT ACCMCCTGCACCAACAACATA '!TGCCACTT ACTTTTCCCAACACCCT AATAcccrrccr
`B
`S
`
`R
`CCG
`
`-------CDll 3-------
`
`VIClO
`
`Rl9.9
`
`VKlO
`
`Rl9. 9
`
`100
`l
`I
`l L £
`T F C C C T
`ACCTTCCCTCCACCCACCAACCTCCAAATCAAA
`?
`? R
`R19.9 -------~I-1-<CG
`
`JKl
`
`FtG. 2. Nucleotide and deduced amino acid sequences of the V K region of mAb R19.9. Comparison is made with sequences of the V "10-Ars-A
`and J"l genes that encode CRI,t mAb (9). See legend to Fig. 1. ?, undetermined nucleotide.
`
`seems to be the best choice. Electron density features in V L
`positions 31, 55, and 107 correspond reasonably well with the
`amino acid sequence encoded by the V "10-Ars-A germ-line
`gene that encodes the V L region of CRIA antibodies and to the
`J"l germ-line sequence (Fig. 2). At position 105, where an
`ambiguity in the nucleotide sequence is consistent with the
`presence of alanine, glutamic acid, glycine, or valine, the
`electron density map favors glutamic acid, in agreement with
`the germ-line J"l sequence.
`The domain structure of Fab R19.9 closely follows that
`observed for all Fabs whose three-dimensional structure has
`been determined by x-ray diffraction to date (reviewed in
`refs. 35 and 36). The quaternary structure of Fab R19.9 is that
`of an extended conformation. This conformation is usually
`described by reference to an "elbow" angle made by the
`pseudo 2-fold axes relating V H to V Land CHl to CL. For Fab
`Rl9.9 this angle is 178°, which is the largest (closest to 180°)
`so far observed in Fab structures. The rotation-translation
`operations necessary to superimpose selected a-carbon
`backbones of V H and V L are 175° and 0.45 A, respectively,
`indicating a nearly ideal symmetry relationship between
`those domains. The numerical values for the corresponding
`operations in the CHl-CL domains are 168° and 1.5 A, values
`similar to those of other known Fab structures.
`A measure of the closeness of the Fab R19.9 structure to
`that of several other Fab molecules can be obtained by
`analyzing the shifts in the relative positions of packed
`{3-sheets (37). Thus, if the spatial superposition of the V H or
`V L {31 sheets is optimized, the {32 sheets of different Fabs
`Qisted in Table 2) are no further away than an average of0.55
`A and a rotation angle of 9° from the corresponding sheet of
`Fab R19.9.
`The relative disposition of V Hand V Lin Fab R19.9 can be
`further compared to those of other Fabs by superposing the
`a-carbon coordinates of the residues that define the interface
`between the V Hand V L subunits, as described in Table 2. The
`largest deviations that were observed for Fab R19.9 are 1.5
`A and 12.8°. These deviations and others observed with
`different Fabs do not appear to be significant.
`
`DISCUSSION
`The three-dimensional structure ofFab R19.9 presented here
`provides a structural model of a specific hapten-binding mAb
`of predefined specificity. The calculation of electron density
`maps at 2.8-A resolution by x-ray crystallographic techniques
`and the determination of amino acid sequences (through
`nucleotide sequencing of the mRN As) of V H and V L allowed
`the tracing of most of the polypeptide chain. However,
`although the general position of V H CDR3 is clearly indicated
`
`in the different electron density maps that were calculated, its
`side chains could not be placed unambiguously. The ambi(cid:173)
`guity may arise from an intrinsic high mobility of V H CDR3.
`This possibility is supported by the fact that the long V H
`CDR3 loop protrudes into the solvent beyond other parts of
`the molecule (see Fig. 3).
`The three-dimensional structure of Fab Rl9.9 agrees well
`with those of other Fabs that have been determined. It
`displays the largest elbow angle that has been observed in
`Fabs, in agreement with the idea that an extended (or
`contracted) conformation is independent of ligand binding to
`the combining site. As shown in Table 2, the relative dispo(cid:173)
`sition of VH and VL in Fab R19.9 is well within a spectrum
`of angular and translational values observed in a number of
`Fabs, irrespective of their human or murine origin or of their
`liganded or unliganded state. The difference in relative
`disposition of V H and V L observed by Colman et al. (45) in
`a neuraminidase complex falls in the range of those observed
`in Table 2, indicating that formation of a complex with
`
`Table 2. The relative arrangements of VH and VL domains in
`different Fabs
`Hy5
`
`Hy5
`
`Kol
`
`New
`
`J539
`
`Rl9.9
`ou
`
`J539 McPC603 R19.9 01.3
`Kol New
`0.49 A 0.10 A 0.42 A o.53 A o.68 A o.69 A
`o.54 A o.69 A o.67 A o.39 A o.95 A o.62 A
`o.96 A
`0.78 A o.61 A o.56 A o.74 A o.73 A
`o.56 A 0.49 A o.30 A o.92 A o.55 A
`8.5°
`1.14 A o.36 A
`o.64 A o.80 A o.82 A o.87 A
`o.57 A o.62 A o.92 A o.56 A
`5.0°
`3.6°
`0.10 A 1.14 A 1.14 A
`o.62 A o.75 A 0.11 A
`o.54 A 1.03 A o.64 A
`1i.0°
`4.5°
`6.7°
`McPC603 0.68 A 0.74 A 0.12 A 0.44 A
`0.11 A o.57 A
`0.88 A 0.59 A
`4.9°
`5.4°
`3.8°
`7.2°
`1.93 A 1.19 A 1.08 A 2.13 A 1.69 A
`o.69 A
`0.93 A
`6.9°
`9.1°
`8.0° 12.8°
`6.9°
`o.91 A o.53 A o.38 A o.90 A 0.46 A i.22 A
`1.9°
`6.8°
`2.6°
`5.1°
`4.8°
`9.5°
`Atomic coordinates for Fab Kol (38), New (22). McPC603 (39),
`J539 (40), and Hy5 (41) were obtained from the Brookhaven Data
`Bank (42, 43), and those of Fab 01.3 (44) were obtained from the
`authors. The V L domain of the Fab corresponding to each horizontal
`row was mapped into the V L of the Fab in the column by using
`a-carbon coordinates of residues defining the interface between VH
`and VL domains (45). The cakulated transformation was applied to
`the Fab corresponding to the horizontal row, and the additional
`) to optimize overlap between the
`translation (A) and rotation (0
`corresponding pair of V H domains are given in the lower left triangle.
`The upper right triangle gives the corresponding root-mean-square
`distances for the pair of V L's (upper number of each pair) and V H's
`(lower number).
`
`3 of 5
`
`Celltrion, Inc., Exhibit 1060
`
`
`
`610
`
`Immunology: Lascombe et al.
`
`Proc. Natl. Acad. Sci. USA 86 (1989)
`
`by Jeske et al. (46), although orientations of these residues
`were not known.
`Site-directed mutagenesis (47) and chain recombination
`studies (48) have implicated V H Ser-99 and V L Arg-96, re(cid:173)
`spectively, as residues that are important in Ar binding.
`Inspection of the model of Fab R19.9 shows that the guani(cid:173)
`dino group of V L Arg-96 is at the bottom of the narrow cavity
`between V H and V L and that it could participate in direct
`contacts with an Ar ligand. Thus, it provides a positively
`charged group that could neutralize the negative charge of the
`Ar hapten. Ser-99 in VH appears less accessible to interac(cid:173)
`tions with an external ligand. However, it and Arg-96 may be
`important in hapten binding not only because they could
`contribute direct contacts with Ar but also through stabilizing
`interactions with other amino acid residues at the combining
`site of the antibody itself. A precise identification of residues
`contacting bound Ar will require the determination of the
`three-dimensional structure of a hapten-Fab complex.
`The presence of a long VH CDR3 loop in R19.9 has two
`major effects on the structure of the combining site: (i) it
`partially fills what otherwise would be a large cavity or
`depression surrounded by the other CD Rs of V Hand V L; and
`(ii) it provides a protruding structural feature that appears
`unique to R19.9 among other Fab structures that have been
`determined. It is interesting to compare this loop with that of
`human VH KOL (38). Although KOL's CDR3 is three amino
`acids longer, it is partially bent inwards, thus resulting in a
`less salient loop than that of R19.9 (Fig. 4).
`The anti-Ar mAb R19.9 expresses some but not all id(cid:173)
`iotopes associated with CRIA. The nucleotide and amino acid
`sequences of its V H and V" regions are consistent with the
`possibility that they are somatic variants of the germ-line V H
`and V" genes that encode CRIA antibodies (Figs. 1 and 2). In
`addition, Rl9.9 utilizes the canonical JH2 of CRIA. A major
`distinction in R19.9 is the contribution of a D gene segment
`giving rise to an atypical CDR3 that contains three more
`amino acid residues than are present in mAbs that fully
`express CRIA. The typical eight-residue D sequence appears
`to arise from the DFL16.1 genetic segment (10). The unusual,
`longer D segment as well as the three amino acid substitutions
`in V ff CDR2 are likely to account for the incomplete expres(cid:173)
`sion of CRIA idiotopes in R19.9. In agreement with this
`conclusion, the L chain ofR19.9, when combined with the H
`chain of a CRIA mAb, yields a recombinant molecule that
`expresses CRIM whereas the converse recombinant,
`HR19.9LcR1 is idiotypically inactive (14).
`The partial idiotypic cross-reactivity between R19.9 and
`CRI.t mAbs can be explained by the presence, in anti(cid:173)
`idiotype sera, of antibodies that recognize residues in V H and
`VL other than those in VH CDR3 . Small differences in the
`primary structure of Rl9.9 and the germ-line-encoded struc(cid:173)
`ture owing to somatic mutations, as well as the large
`difference in D-region structure, could also contribute to the
`lack of total identity in serological tests. Among the somatic
`variations, the change from Ile-34 to Val in VH R19.9 does not
`seem adequate to explain a change in antigenic properties
`since the amino acid side chain at that position is not exposed.
`The sequence variations observed at positions Lys-55, Tyr-
`57, Leu-58, and Ser-59 in CDR2 and Arg-74 in FR3 could
`readily affect recognition of idiotopes since they are exposed
`at the accessible surface of Vff.
`Characterization of two antigenic determinants occurring
`on a native protein (41, 44) indicates that the antigen(cid:173)
`antibody interface extends over an area of about 700 A2 and
`includes residues from each of the V ff and V L CD Rs.
`Contacts with an anti-idiotype antibody might similarly
`include residues from all CDRs. Thus, given the solvent(cid:173)
`exposed location and the spatial proximity of the CDR loops
`it is most reasonable to expect that the V Las well as V H CD Rs
`will contribute to many combining-site related idiotopes.
`
`FtG. 3. A superposition of a portion of the a-carbon backbones
`of Fabs Kol (38) and Rl9.9. The VH COR3 of the Fabs have very
`different conformations. That of R19.9 is labeled and indicated by a
`slightly thinner trace. The numbers above 300 represent position
`numbers in the VH sequence of R19.9 plus 300.
`
`antigen is not necessary to produce the observed variability
`in the relative disposition of V H and V L·
`The CDRs of Rl9.9 contain a striking number of aromatic
`side chains, mostly tyrosines, which are oriented in such a
`way that they expose their phenolic OH groups to the solvent
`(see Fig. 4). These include the VH tyrosines 27, 32, 57, 101,
`and 109. VH Phe-100 in this region is also exposed to the
`solvent. V H Phe-29, Tyr-50, and Tyr-110 are oriented towards
`the interior of the structure, thus precluding their participa(cid:173)
`tion in contacts with ligands. In VL, Tyr-32, Tyr-49, and
`Tyr-50 are oriented with their side chains pointing towards
`the solvent. Thus, a total of nine aromatic side chains
`belonging to the V H and V L CD Rs, mostly tyrosines, are
`positioned to provide possible contacts with antigen. The
`prevalence of tyrosine residues in these CDRs has been noted
`
`FIG. 4. A view of the combining siteofFab R19.9, which shows
`a high density of aromatic residues, mostly tyrosine V Lis at left; V H
`is at right. The numbers above 300 represent positions in the V H
`sequence plus 300 (i.e., CA 401 = VH residue 101, etc.). Lower
`numbers are positions in V L·
`
`4 of 5
`
`Celltrion, Inc., Exhibit 1060
`
`
`
`Immunology: Lascombe et al.
`
`Proc. Natl. Acad. Sci. USA 86 ( 1989)
`
`611
`
`Note Added in Proof. After this paper was submitted for publication,
`Stevens et al. (49) reported differences in low and high ionic strength
`crystalline forms of human immunoglobulin L-chain dimer Loe,
`which they take to indicate different potential conformations of an
`antibody. The different VL-VL contacts observed in Loe, as well as
`those between VH and VL domains in FabNC41 complexed to
`antigen (45), are postulated (49) to increase antigenic and idiotypic
`specificities ofantibodies. However, as shown in Table 2, the V H-V L
`contacts observed in NC41 are within the range observed in Fabs,
`independently of their liganded (Dl.3 and Hy5) or nonliganded
`states. Since Table 2 includes Fabs crystallized at low (Hy5, Dl.3,
`and R19.9) and high ionic strength (Kol, New, J539, and McPC603),
`the conformational variations observed in Loe may be unique to it or
`to L chains, which do not normally occur as dimers in nature. The
`lack of significant differences in the relative disposition of V Hand V L
`thus far observed in liganded or unliganded Fahs does not seem to
`suggest that such differences could contribute to increase the
`functional diversity of antibodies.
`
`We wish to thank Dr. Claude Riche (Centre National de la
`Recherche Scientifique, Gif sur Yvette, France) for the interpreta(cid:173)
`tion of the electron density map ofR19.9 at 6-A resolution by means
`of a six-dimensional real space search. This work was supported by
`grants from the lnstitut Pasteur, Centre National de la Recherche
`Scientifique, contract BAP-022l(DC) from the European Economic
`Community, and by National Institutes of Health grants Al-25369
`and Al-22068.
`
`1. Kuettner, M. G., Wang, A. L . & Nisonoff, A. (1972) J. Exp.
`Med. 135, 579-595.
`2. Nisonoff, A. , Ju, S.-T. &Owen, F. L. (1977) /mmunol. Rev. 34,
`89-118.
`3. Greene, M. I., Nelles, M. J., Sy, M.-S. & Nisonoff, A. (1982)
`Adv. lmmunol. 32, 253-300.
`4. Rathbun, G., Sanz, I., Meek, K ., Tucker, P. & Capra, J. D.
`(1988) Adv. lmmunol. 42, 95-164.
`5. Manser, T ., Wysocki, L. J., Margolies, M. N. & Gefter, M. L.
`(1987) Immunol. Rev. 96, 141-162.
`6. Pawlak, L . L., Mushinski, E . B., Nisonoff, A. & Potter, M.
`(1973) J . Exp. Med. 137, 22-31.
`7. Laskin, J. A. , Gray, A., Nisonoff, A., Klinman, N. R. &
`Gottlieb, P. G. (1977) Proc. Natl. Acad. Sci. USA 74, 4600-
`4604.
`8. Siekevitz, M. , Huang, S.-Y. & Gefter, M. L . (1983) Eur. J.
`lmmunol. 13, 123-132.
`9. Sanz, I. & Capra, J. D. (1987) Proc. Natl. Acad. Sci. USA 84,
`1085-1089.
`10. Landolfi, N. F .• Capra, J. D. & Tucker, P. W. (1986) J.
`lmmunol. 137, 362-365.
`11. Meek, K ., Sanz, I. , Rathbun, G., Nisonoff, A. & Capra, J. D.
`(1987) Proc. Natl. Acad. Sci. USA 84, 6244-6248.
`12. Estess, P. , Lamoyi, E. , Nisonoff, A. & Capra, J. D. (1980) J.
`Exp. Med. 151, 863-875.
`13. Gill-Pazaris, L . A., Brown, A. R. & Nisonoff, A. (1979) Ann.
`lmmunol. (Paris) 130C, 199-213.
`14. Gill-Pazaris, L. A. , Lamoyi, E., Brown, A. R. & Nisonoff, A.
`(1981) J. lmmuno/. 126, 75-79.
`15. Lamoyi, E .• Estess, P .• Capra, J. D. & Nisonoff, A. (1980) J.
`lmmunol. 124, 2834-2840.
`16. Amit, A. G ., Harper, M., Mariuzza, R. A., Saludjian, P.,
`Poljak, R. J., Lamoyi, E. & Nisonoff, A. (1983) J. Mo/. Biol.
`165, 415-417.
`17. Porter, R. R. (1959) Biochem. J . 73, 119-127.
`18. Wlodawer, A. & Hodgson, K. 0. (1975) Proc. Natl. Acad. Sci.
`USA 12, 398-399.
`19. Mariuzza, R. A., Amit, A. G. , Boulot, G., Saludjian, P., Saul,
`F. , Tougard, P. , Poljak, R. J .• Conger, J. , Lamoyi, E . &
`Nisonoff, A. (1984) J . Biol. Chem. 259, 5954-5958.
`
`20. Oatley, S. & French, S. (1982) Acta Crystallogr. Sect. A 38,
`537-549.
`21. North, A. C. T., Phillips, D. C . & Matthews, F. S . (1968) Acta
`Crystallogr. Sect. A 24, 351-358.
`22. Saul, F. A .• Amzel, L. M. & Poljak, R. J. (1978) J. Biol. Chem.
`253, 585-597.
`23. Crowther, R. A. (1972) The Molecular Replacement Method,
`ed. Rossmann, M. G. (Gordon & Breach, New York), pp. 173-
`178.
`24. Sussman, J. L. (1985) Methods Enzymol. 115, 271-303.
`25. Wang, B. C. (1985) Methods Enzymol. 115, 90-122.
`26. Jones, T . A. (1985) Methods Enzymol. 115, 157-171.
`27. Hendrickson, W. A. & Konnert, J. H. (1980) in Computing in
`Crystallography, eds. Diamond, R. , Ramaseshan, S. & Ven(cid:173)
`katesan, K. (Indian Acad. Sci., Bangalore, India), pp. 13.01-
`13.25.
`28. Bhat, T. N. (1988) J. Appl. Crystallogr. 21, 279-281.
`29. Griffiths, G. M. & Milstein, C. (1985) in Hybridoma Technol(cid:173)
`ogy in Biosciences and Medicine, ed. Springer, T. A. (Plenum,
`New York), pp. 103-115.
`30. Kaartinen, M., Griffiths, G . M. & Milstein, C. (1983) Nature
`(London) 304, 320-324.
`31. Robbins, P. F. , Rosen, E . M., Haba, S. & Nisonoff, A. (1986)
`Proc. Natl. Acad. Sci. USA 83, 1050-1054.
`32. Kabat, E. A., Wu, T. T., Reid-Miller, M., Perry, H . M. &
`Gottesman, K . S. (1987) Sequences of Proteins of Immunolog(cid:173)
`ical Interest (U.S. Public Health Service, National Institutes of
`Health, Bethesda, MD), 4th Ed.
`33. Slaughter, C . A. & Capra, J. D. (1984) in Biology of ldiotypes,
`eds. Greene, M. I. & Nisonoff, A. (Plenum, New York), pp.
`35-58.
`34. Huang, S. Y., Margolies, M. N., Marshak-Rothstein, A. ,
`Siekevitz, M. & Gefter, M. L. (1982) Pharmacol. Rev. 34, 43-
`49.
`35. Davies, D. R. & Metzger, H. A. (1983) Annu. Rev. lmmunol.
`1, 87-117.
`36. Alzari, P. M., Lascombe, M.-B. & Poljak, R. J. (1988) Annu.
`Rev. lmmunol. 6, 555-580.
`37. Lesk, A. M. & Chothia, C. (1982) J. Mo/. Biol. 160, 325-342.
`38. Marquart, M .• Deisenhofer, J., Huber. R. & Palm. W. (1980)J.
`Mo/. Biol. 141, 369-391.
`39. Satow, Y., Cohen, G. H., Padlan, E. A. & Davies, D. R. (1986)
`J . Mo/. Biol. 190, 593-604.
`40. Suh, S. W. , Bhat, T . N., Navia, M. A., Cohen, G. H. , Rao,
`D. N., Rudikoff, S. & Davies, D. R. (1986) Proteins 1, 74-80.
`41. Sheriff, S ., Silverton, E.W. , Padlan, E . A., Cohen, G. H .,
`Smith-Gill, S. J., Finzel, B. C. & Davies, D. R. (1987) Proc.
`Natl. Acad. Sci. USA 84, 8075-8079.
`42. Bernstein, F . C., Koetzle, T . F. , Williams, G. J. B., Jr. ,
`Meyer, E. F., Brice, M. D. , Rodgers, J. R .• Kennard, 0.,
`Shimanouchi, T. & Tasumi, M. (1977) J . Mol. Biol. 112, 535-
`542.
`43. Abola, E. E. , Bernstein, F. C ., Bryant, S. H., Koetzle, T . F.
`& Weng, I. (1987) In Crystallographic Databases: Information
`Content, Software Systems, Scientific Applications, eds. Allen,
`F . H., Bergerhoff, G. & Sievers, R. (Data Commission of the
`Int. Union of Crystallography, Bonn}, pp. 107-132.
`44. Amit, A.G., Mariuzza, R. A., Phillips, S . E. V. & Poljak,
`R. J. (1986) Science 233, 747-753.
`45. Colman, P. M., Laver, W. G., Varghese, J. N., Baker, A. T.,
`Tulloch, P. A., Air, G . M. & Webster, R. G. (1987) Nature
`(London) 326, 358-363.
`46. Jeske, D. , Milner, E. C. B. , Leo, 0. , Moser, M., Marvel, J. ,
`Urbain, J. & Capra, J. D. (1986) J. lmmunol. 136, 2568-2574.
`47. Sharon, J., Gefler, M. L ., Manser, T . & Ptashne, M. (1986)
`Proc. Natl. Acad. Sci. USA 83, 2628-2631.
`48. Jeske, D. J., Jarvis. J., Milstein, C. & Capra, J. D. (1984) J.
`lmmunol. 133, 1090-1092.
`49. Stevens, F. J., Chang, C. H . & Schiffer, M. (1988) Proc. Natl.
`Acad. Sci. USA 85, 6895-6899.
`
`5 of 5
`
`Celltrion, Inc., Exhibit 1060
`
`