Proc. Natl. Acad. Sci. USA
`Vol. 85, pp. 3080-3084, May 1988
`Immunology
`
`Variable region framework differences result in decreased or
`increased affinity of variant anti-digoxin antibodies
`(lmmuooglobullns/fluoresceoc:e-lldiva� cell sorting/ llOID8t(c: iqutatlon/ comp1'mentarity I digoxin)
`
`DAVID J. PANKA*t, MERE.DITH MUDGEIT-HUNTER*, DAVID R. PARKS*, LISA L. PETERSON*§,
`LEONARD A. HERZENBERG*, EDGAR HABER*, AND MICHAEL N. MARQ9LIES*�
`General Hospital and Harvard Medical School, Boston, MA 02114; and *Department of Genetics,
`*Departments of Medicine and Surgery, Massachusetts
`Stanford, CA 94305
`Stanford University,
`
`Contributed by Leonard A. Herzenberg, December 17, 1987
`
`Rare spontaneous variants of the anti­
`ABSTRACT
`digolin antibody-producing hybridoma 40-150 (K0 = S.4 x
`109 M"'"1) were selected for altered antigen binding by two­
`color Ouoresce�activated cell sorting. The parent antibody
`binds digolin 890-fold greater than digitolin. The variant
`40-150 A2.4 bas reduced aftlnity for digolin (K0 = 9.2 x 10'
`M-1) and binds digolin 33-fold greater than digitolin. A
`second-order variant, derivecJ from 40-150 A2.4 (designated
`40-150 A2.4 P.10), demonstrated partial regain of digolin
`binding (K0 = 4.4 x 10' M-1). The altered b"4fing of the
`variant 40-150 A2.4 was accounted for by a point mutation
`resulting in substitution of arginine ror serine at position 94 in
`the heavy chain variable region. Antibody 40-150 A2.4 P.10
`also contains this arginine but owes Its enhanced antigen
`binding to deletion or two amino acids from the heavy chain
`amino terminus. ThJs unusual sequence alteration in an Im·
`munoglobulin framework region confers increased atllnlty ror
`·
`antigen.
`
`J:i:l
`l ir�µ1 H
`
`»'£NIN
`
`HO� OIGITOXOSE
`HO H 3
`
`OIGOXIN
`
`Dl�ITOXIN
`
`FIG. 1. Structures of digoxin (digoxigenin tridigitoxose) and
`digitoxin.
`
`mutations in either hypervariable or framework regions,
`
`except for one instance where multiple amino acid differ­
`
`
`ences, perhaps due to gene co0version, resulted in loss of an
`idiotope
`(1�).
`Digoxin (digoxigenin tridigitoxose, Fig. 1) is a useful
`
`
`Although much is known about the genetic basis of antibody
`
`hapten for studies of antibody complementarity as it approx­
`(reviewed in refs. 1 and 2), knowledge of the
`diversity
`imates in size the antibody combining site, is uncharged,
`
`
`precise molecular interactions between specific antibodies
`
`a rigid steroid nucleus, and has a known three-di­
`contains
`
`
`
`and antigens remains limited. The combining site of antibody
`
`mensional structure (16); multiple congen!!rs of defined
`molecules
`
`is composed of the amino-terminal domains [vari­
`are known (17, 18). Anti-digoxin antibodies
`structure
`are
`
`
`The high able (V) regions] of two nonidentical polypeptides.
`
`
`characterized by 'high affinity and varying specificity for
`
`degree of amino acid sequence variability found in three
`
`
`
`related cardiac glycosides. The anti-digoxin antibody 40-150
`segments of the light (L) and heavy (ff)
`noncontiguous
`
`(18) was chosen for isolation of variants because of its
`
`
`polypeptide chains led to the prediction (3) that these "hy­
`to the presence of a hydroxyl group at the 12
`sensitivity
`
`
`pervariable" or "complementarity-determining regions"
`
`
`positic;m on the steroid C ring of digoxin (Fig. 1) as it binds t<;>
`
`
`(CDR) would fold together in three dimensions to form the
`
`digoxin more avidly than digitoxin, which lacks the hydroxyl
`
`combining site. This prediction was verified by x-ray crys­
`
`group. We report the isolation and characteriµtion from cell
`
`tallographic analyses of hapten binding Fab fragments (4-7).
`
`culture of a spontaneous antibody variant with reduced
`[The Fab fragment (Mr 50,000) obtained by limited proteol­
`
`
`affinity due to a point mutation. A second-order variant
`ysis (8) is composed of one L and the amino-terminal half of
`
`
`demonstrated partial regain of binding due to a "compen­
`the H chain and contains
`
`one of two identical antigen binding
`
`satory" deletion of two amino acids from the framework
`
`
`
`sites.] An examination of Fab crystal struc�ures indicated
`region.
`
`that hypervariable region sequ�nces are engraft�d onto
`
`"frameworks" for which three-dimensional folding is re­
`MATERIAL AND MEIBODS
`markably uniform (9).
`CeU Lines and Antibody Puriftcation. The production
`of
`The availability of monoclonal antibodies, whether se­
`
`
`
`
`the murine A/J anti-digoxin hybridoma cell line 40-150
`
`
`creted by myeloma tumors or by hybridomas resulting from
`(lgGl, K) has been reported (18). Production of 40.150 and
`
`
`somatic cell fusion (10), makes possible further correlation
`
`of primary amino acid sequences with antigen binding spec­
`
`variant hybridoma proteins in ascites was performed as
`(17). Antibodies
`
`
`
`ificity, through examination of structural variants obtained
`described
`were purified from ascites or
`
`
`spontaneously or by design through site-directed mut�gene­
`
`
`sis. Spontaneous variant antibodies with reduced or absent
`Abbreviations: CDR, complementarity-determining region(s);
`FACS, fluorescence-activated cell sorting; H, heavy; L, light; V,
`
`
`antigen binding wer� reported for the phosphocholine (11,
`12) and 4-hydroxy-3-nitro-5-iodophenylacetyl
`variable.
`(13-15) hap­
`tPresent address: Department of Biochemistry and Biophysics,
`
`tens. In these cases the structural chan.ges were due to point
`Center for Biochemical and Biophysical Sciences and Medicine,
`Harvard Medical School, 250 Longwood Avenue, Boston, MA
`02115.
`§Present address: Invitron Corp., Redwood City, CA 94063.
`tro whom reprint requests should be addressed.
`
`The publication costs of this article were defrayed in part by page charg�
`payment. This artide must therefore be hereby marked "advertisement"
`in accordance with 18 U.S.e. §1734 solely to indicate this fact.
`
`
`3080
`
`1 of 5
`
`BI Exhibit 1135
`
`

`

`Immunology: Panka et al.
`
`Proc. Natl. Acad. Sci. USA 85 ( 1988)
`
`3081
`
`Table 1. Binding of antibody 40-150* and variants to cardiac glycosides
`
`Antibody or
`fragment
`
`Concentration of inhibitort relative to digoxin giving 50% inhibition of binding to 1251-labeled
`digoxin-bovine serum albumin
`
`K0 x lo-'*
`Ouab�n
`Acetylstrophanthidin
`Gitoxin
`Digitoxigenin
`Digitoxin
`Oigoxigenin
`(L/M)
`540 ± 60
`400
`7300
`7300
`7.3
`890
`8000
`40-150
`0.9 ± 0.1
`33
`46
`39
`19
`36
`44
`40-150 A2.4
`40-150 A2.4 P.10
`44 ± 9
`1300
`1300
`260
`1200
`1100
`6.9
`0.8 ± 0.2
`Fab I
`44
`21
`43
`36
`38
`39
`23 ± 4
`7.5
`1200
`180
`830
`Fab II
`1500
`1300
`*Minor differences in specificity determined here for antibody 40-150 as compared to reported values (18) are attributed to differences in the
`assay systems used and the 1251-labeled digoxin-bovine serum albumin probe employed here (29).
`tsinding of digoxin was set at 1.0 for all antibodies.
`*Affinity for digoxin.
`
`RESULTS
`In the initial FACS experiment a variant of 40-150 with
`reduced binding to digoxin, designated 40-150 A2.4, was
`isolated. This clone was subjected to a second round of
`sorting for cells with enhanced binding to digoxin, resulting
`in the selection of 40-150 A2.4 P.10. The binding character­
`istics of antibody 40-150 and the two variants are summa­
`rized in Table 1 and Fig. 2. Although antibody 40-150 binds
`digoxin (K0 = 5.4 x 109 M-1) 890 times more avidly than
`digitoxin, 40-150 A2.4 has reduced �nity for digoxin (K0 =
`9.2 x 106 M -1) but now binds digoxin only 33 times more
`avidly than digitoxin. The second-order variant 40-150 A2.4
`P.10 binds digoxin with an increased affinity (K0 = 4.4 x
`108 M-1) compared to antibody 40-150 A2.4, from which it
`was derived, and distinguishes digoxin and digitoxin to an
`extent (1100-fold difference) similar to the parental antibody.
`This altered recognition of the steroid C ring 12-hydroxyl
`group among the three antibodies is also reflected in their
`relative binding to gitoxin (Table 1). Gitoxin, like digitoxin,
`lacks the 12-hydroxyl group but has an additional hydroxyl
`at position 16. The v¥Iant 40-150 A2.4 is relatively more
`sensitive to gitoxin inhibition of digoxin binding than the
`parent 40-150 and the second-order variant 40-150 A2.4 P.10.
`Unlike the results for digoxin binding, in a competition assay
`for the 40-150 idiotype, antibodies 40-150, 40-150 A2.4, and
`40-150 A2.4 P.10 were indistinguishable (data not shown).
`To determine the structural changes responsible for anti·
`gen binding differences among the three hybridoma proteins,
`the complete V region sequences were determined by nucle-
`
`culture medium by affinity chromatography on digoxin­
`bovine serum albumin-Sepharose or digitoxin-bovine serum
`albumin-Sepharose. Digoxin or digitoxin was coupled to
`bovine serum albumin through the terminal digitoxose moi­
`ety (17, 19). Antibodies were eluted by using 3 M KSCN;
`however, for separation of Fab fragments from variants,
`gradient affinity chromatography (0-3 M KSCN) was used.
`Fluorescence-Activated Cell Sorting (FACS). Rare sponta­
`neous variants of anti-digoxin hybridomas were isolated by
`two-color FACS as described (20). Parental cell populations
`were stained with digoxin-fluorescein to measure antigen
`binding by cell-surface antibody. The amount of cell-surface
`antibody was simultaneously estimated by using biotin­
`conjugated goat anti-mouse IgG and Texas red avidin. Upon
`measurement in a two-laser FACS system, cells whose
`antigen binding relative to surface lgG did not match that of
`the parental population were sorted as candidate variants.
`After a period of culture the sorted populations were
`restained and resorted. Typically on the third cycle, proba­
`ble variant cells were cloned directly by using the F ACS.
`Sequence Analysis. For mRNA sequencing, total RNA and
`poly(A) mRNA were prepared as described (21, 22). Com­
`plementary DNA synthesis was performed simultaneously
`on Hand L chains with poly(A)+ RNA (100-200 µ.g) (23) by
`using complementary oligonucleotides to the respective con­
`stant regions (24). The remainder of the sequences through
`the N terminus were obtained by using oligonucleotides
`complementary to the �hird L chain framework region (see
`Fig. 3, positions 67-73) and the third H chain framework
`region (positions 68-73). The isolated full-length 32P end­
`labeled cDNA transcripts were sequenced by base-specific
`chemical cleavage as described (24). Amino acid sequence
`analyses were performed on isolated chains and peptides
`obtained by cleavage with CNBr, o-iodosobenzoic acid, and
`tryptic digestion of completely reduced and alkylated citra­
`conylated chains; the fragments were purified by gel filtra­
`tion and high-pressure liquid chromatography (HPLC) (25,
`26). Peptides were sequenced by automated Edman degra­
`dation on a Beckman 890C sequenper or an Applied Biosys­
`tems (Foster City, CA) 470A gas-phase sequencer. At cycles
`where proline was N-terminal, o-phthalaldehyde was used to
`reduce background or selectively sequence prolyl peptides
`(27). Phenylthiohydantoin-amino acids were identified by
`HPLC (28).
`Binding Assays. A double-antibody precipitation assay
`using [3H)digoxin was employed to measure antibody affin­
`ity (K0) as described in detail elsewhere (29). A microtiter
`plate radioimmunoassay was used to assess relative anti­
`body specificity for rel�ted cardiac glycosides (29). This
`method is applicable to lower affinity antibodies than that
`used previously (17, 18). The production of a rabbit polyclo­
`nal anti-40-150 idiotypic serum and competition radio­
`immunoassay for the 40-150 idiotype were performed by
`methods analogous to those reported for antidigoxin anti­
`body 26-10 (29).
`
`-Digoxin
`100 --· Oigi1oxin
`040-150
`'" 40·150 A2 4
`•40-150
`A2.4P.t0
`
`BO
`
`�
`;:;; � 60
`i:::;
`�
`
`� 40�
`
`20
`
`0
`
`10·1 10"6 10-s
`(M)
`INHIBITOR
`
`FtG. 2.
`Inhibition of binding of monoclonal anti-digoxin anti­
`bodies 40-150, 40-150 A2.4, and 40-150 A2.4 P.10 to 1251-labeled
`digoxin-bovine serum albumin by unlabeled digoxin and digitoxin.
`
`2 of 5
`
`BI Exhibit 1135
`
`

`

`I
`-10
`• H P C L l L I P L Y L T L l C Y ·O
`C 0 Y l L Y E S
`
`20
`,.--- CORI --,
`40
`30
`10
`• C C C L Y K P C C S L l L S C A A S C P T P ft S T T M A Y Y R 0 I P 0 K R L £ Y Y A
`
`.DAVY CIWIOS
`
`ATCMCTT=crCACATTCATTTtccrn:tCCTTACTTTMMCGTCTCCAC"TCT
`
`ccrc.uc=ACTCTCCCCCACC
`
`CTTACTCMCC
`
`CTGCACCCTCC
`
`C'ICAMCTCTC
`
`CTCTCCACCCTCTCC'l"l'CACTTT
`
`CACMGCTATT
`
`ACATCCCTTCCCTTCCCCACATTCCA
`
`CACMCACCC'IC()ACTCCCT
`
`CCCA
`
`40.1)0
`40-1)0 A2.4
`40-150 A2.4 P.10
`40-1)0
`40.1)0 A2.4
`40-150 A2.4 P.10
`
`40-150
`40-1)0 A2.4
`40-1)0 A2.4 f.10
`•0-1)0
`•0-1)0 •2.4
`40-150 A2.4 P.10
`
`CDR 3 ----�
`�------- CD 2
`100 • b
`80 82 • b c
`60
`)Cl
`70
`110
`
`T I S l S D I T T T T P 0 N Y l C R P T I $ R 0 N A K H T L T L 0 K S S L K S £ 0 T A " J T C T S V C N T D T A K D T V C 0 C T $ Y T Y S S
`
`
`=========================::::::::::::::::=:=::-_::::::::::=:::::::::===--==========::::::==---======= : ====:=============::::::::::::::::::::::::::::::
`
`MX AnACTATTACTCAT ATTTACACCTA CTATCCAC ACMTCTC.MCCCCCCATTCACCATCTC
`
`CACACAC.UTGCCMC
`
`MCACCCTCTACcnCMA
`
`TCACCAGTCTCMCTCTC
`
`AGGACACACCC.ATCTATTACTCT ACAACT'TGGCGC.AATTACC
`
`---------------------------------------------------------------------------------------------------------------------------�---------------------------------------------------------
`----------------------------------------------------------------------------------------------------------------------------------------�------------------------------------------------
`
`ACf ATCCTA'f'GGACTACT'OOOC"f'(:MGMCCTCACTCACCCTCl'CCTCA
`
`90 Jn
`
`1
`30
`21. b cai::
`10
`•
`20
`
`
`rsoCO« z-
`40
`0 Y Y M T 0 T P L T L S Y T 1 C 0 P A S 1 S C l S S 0 S L L H S 0 C K t T L I V L L O R P c O S P K R L I Y L Y S
`
`ucn CaA1llS
`
`CATCTl'CtCA'!CACCCM:.\C'ltCACTCACTTTCTCCCTT
`
`ACCA TTCCACMCCACCCTCCATCT
`
`CTTCCMCT
`
`CMCTCACACCC
`
`TCTT AMT ACTCATCCAMCACATATTTCATTTccncnACACACCCCM:CCCACTCT
`
`CCAMCCCC
`
`CTAATCTATCTCCTCTCT
`
`- CO« 2 -...,
`60
`
`
`r--;;-- CO« 3 ---,
`70
`80
`
`l L D S C Y P 0 R P T C S C S C T 0 P T L l 1 S R V 6 A E 0 L C Y Y T C V 0 C T 8 P P T t P C C C T K L 6 1 l R
`
`100
`
`40-1.50
`40-1)0 A2.4
`40-1)0 A2.4 P.10
`
`40-1)0
`40-1)0 A2.4
`
`40-1)0 A2.4 P.10
`
`40-150
`40-150 A2.4
`40-150 A2.4 P.10
`
`AAA CTCCACT CTCCACT
`
`CCCTC1. CAGCITCACTCCCAC
`
`40-150
`40-1)0 A2.4
`40-150 A2.4 P.10
`Fro. 3. Amino acid and nucleotide sequences of the V regions of 40-150, 40-150 A2.4, and 40-150 A2.4 P.10 H (Upper) and and L (Lcwer) chains. Amino acid sequences given were derived
`from the mRNA sequence results shown as well as from direct protein sequence analysis. The portions of the V region sequences determined by Edman degradation include 40-150 L, all
`residues except 76 and 77; 40-150 A2.4 L. completely sequenced; 40-150 H, all residues except 67-71and80-89; 40-150 A2.4 H, all residues except 82-89. See Fig. 4 and text for40-150 A2.4 P.10
`protein sequence results. Amino acid sequences are given at the top in a one-letter code (30). The numbering of amino acid residues and the designation of CDRs are as defined by Kabat et al. (31).
`
`ATCACCCA
`
`CACATTT CACACTCAAAA TCACCACACTCCACC
`
`CACCATTTCCGACTTT
`
`ATTATTCTTCCC.W;CT
`
`ACACATTTTCCCTACACCTW;CA
`
`CCCCCC ACCMCCTCCAM
`
`TAAAA CCC
`
`w
`� N
`
`�
`�
`
`[ c ::a 0
`.� II>
`� s::. :-
`
`�
`
`�
`;::.
`;:....
`2
`�
`� !'"·
`s
`;:....
`
`e: ...... .... �
`
`3 of 5
`
`BI Exhibit 1135
`
`

`

`Panka et al.
`Immunology:
`
`Proc. Natl. Acad. Sci. USA 85 ( 1988)
`
`3083
`
`The L chain V
`to that A2.4 P.10. The fine specificity of Fab II was identical
`
`
`otide and amino acid sequence analyses.
`
`
`
`region sequences of all three antibodies are identical (Fig. 3).
`of the intact antibody 40-150 A2.4 P.10 and its unfractionated
`
`The H chain V region sequences of 40-150 and 40-150 A2.4
`Fab. Fab 11 contained
`
`a single sequence corresponding to the
`
`
`
`are identical except for a single nucleotide difference
`shortened H chain (Fig. 4). To confirm that the apparent
`
`
`
`(guanine/thymidine) resulting in the substitution of arginine
`enhanced affinity of 40-150 A2.4 P.10 was due to dominance
`for serine at position 94 (Fig. 3). This single amino acid
`
`
`of the higher affinity truncated species in the binding assay,
`difference in a "framework" residue (3) at the boundary
`
`Fabs I and II were mixed together in a molar ratio (40:60)
`
`
`with CDR 3, resulting in the substitution of a bulky polar side
`
`corresponding to that observed in the protein sequence
`
`
`chain, must be responsible for the decreased affinity and
`
`
`analysis of unfractionated 40-150 A2.4 P.10. The binding
`
`altered specificity of 40-150 A2.4.
`
`
`characteristics of this mixture duplicated the results for
`Since antibody 40-150 A2.4 P.10 was derived from 40-150
`
`unfractionated 40-150 A2.4 P.10 (data not shown).
`
`
`A2.4 we anticipated that the increased binding in the former
`
`antibody might be due to a compensatory mutation else­
`DISCUSSION
`where in the V region. However, the V region nucleotide
`
`sequences of these two antibodies are identical for both H
`The ability to isolate somatic mutants from antibody­
`
`the mutation at position 94 (Fig. 3).
`and L chains, including
`
`producing cell lines in culture provides an opportunity to
`
`All three antibody H and L chains were indistinguishable in
`assess the effects upon antigen binding and idiotypy of
`
`size by PAGE in NaDodS04 under reducing conditions (not
`
`structural alterations limited to one or a few V region amino
`
`
`shown), precluding a major deletion. Fab fragments of all
`
`acids (11-15). Two-color FACS proved useful in isolating
`
`
`
`three antibodies had antigen binding properties identical to
`
`
`
`
`spontaneous variants of high-affinity anti-digoxin antibodies
`
`the intact antibodies from which they were derived, thus
`
`with frequencies in the 10-6 range (20). The variant hybrid­
`
`
`
`precluding structural alterations in Fe that affect antigen
`oma protein 40-150 A2.4 binds digoxin with lower affinity
`binding. We therefore analyzed the protein sequence of
`
`
`than the parental antibody and is relatively less sensitive
`
`
`40-150 A2.4 P.10. The sequence obtained by Edman degra­
`than the parent antibody to the absence of a 12-hydroxyl
`
`dation of the L chain (50 cycles) was identical to that
`group on the steroid C ring of the hapten. The structural
`
`
`
`predicted from the nucleotide sequence analysis shown in
`
`
`change responsible for the binding differences is due to a
`
`Fig. 3. However, sequence analysis (40 cycles) of the H
`
`single amino acid substitution in the H chain framework
`
`chain revealed two amino acids at each position (Fig. 4). In
`region at position 94, at the edge of CDR 3 (Fig. 3). The
`
`
`addition to the full-length H chain corresponding to the
`finding that a framework mutation can alter binding to
`
`
`nucleotide sequence (Fig. 3), a second species with an amino
`
`antigen is not unexpected. Previous studies of phosphocho­
`
`
`terminus at position 3 relative to the intact chain was also
`
`
`line binding myeloma variants demonstrated that a mutation
`
`present. This result was obtained for antibody purified from
`
`in the J" region reduces affinity for phosphocholine only
`ascites, from culture medium, and
`from fresh subclones of
`
`when the hapten is bound to carrier protein. Although that
`
`40-150 A2.4 P.10, demonstrating that the mixed antibody is
`JH mutation is formally in a hypervariable
`region based on
`
`an inherent product of the clone.
`sequence (3), it is not located in the binding site based on
`
`
`To determine whether the truncated H chain species was
`
`
`x-ray crystallography (32). In addition, crystallographic
`
`
`indeed responsible for the observed enhancement of antigen
`
`
`analysis of a lysozyme--antilysozyme complex (33) demon­
`
`binding by 40-150 A2.4 P.10, Fab fragments were prepared
`
`
`strated that among 17 antibody residues that contact antigen,
`(8) and separated
`
`
`by gradient affinity chromatography (not
`2 were found in framework regions at the edge of CDR.
`
`shown). Two major peaks were identified. An early eluting
`
`
`The results for a second-order variant of antibody 40-150
`peak (I) contained Fab fragments with affinity for digoxin
`
`
`
`were, however, unexpected. Regardless of the mechanisms
`
`
`and specificity indistinguishable from that of 40-150 A2.4
`
`involved in H chain truncation of the digoxin antibody
`(Table 1). Sequence analysis of Pab I revealed a single
`demon­
`variant 40-150 A2.4 P.10 described here, the results
`
`
`sequence corresponding to the full-length H chain (Fig. 4).
`
`strate that a change in framework structure may significantly
`
`The later eluting peak (II) contained Fab fragments that
`bound digoxin with an affinity (K0 = 2.3 x 108 M-1) higher
`enhance antibody affinity.
`The marked effect of these struc­
`
`than peak I and close to that found for unfractionated 40-150
`tural changes on antigen binding was not accompanied
`by a
`
`AJl'l'IBOOY
`
`40-150
`
`40-150 A2.4
`
`'
`1
`10
`0 V K L V E S G G G L V
`
`40'}
`
`60,
`
`40-150 A2.4 P.10
`
`Pab I
`
`Pab II
`
`l
`D V K L V E S G G G L V
`
`K L V E S G G G L V
`
`F10. 4. Comparative amino-terminal sequences of H chains 40-150, 40-150 A2.4, and 40-150 A2.4 P.10. Antibody 40-150 A2.4 P.10 H chain
`contained two amino acids at each cycle for both the intact antibody and Fab. Amino acid sequences of 40-150 A2.4 P.10 Fab I and Fab II
`obtained following affinity chromatography are shown below. The arrow indicates a putative cleavage site in the 40-150 A2.4 P.10 H chain
`accounting for the two sequences. A line indicates identity with the topmost sequence.
`
`4 of 5
`
`BI Exhibit 1135
`
`

`

`3084
`
`Immunology: Panka et al.
`
`Proc. Natl. Acad. Sci. USA 85 (1988)
`
`change in idiotypy, as detected by using a rabbit antiserum
`prepared against 40-150 (data not shown).
`Inasmuch as (i) the nucleotide sequence of 40-150 A2.4
`P.10, using an internal V region oligonucleotide primer (Fig.
`3), revealed only a single sequence without ambiguity
`through the V" amino-terminal region and (ii) the H chain
`nucleotide leader sequences for 40-150 A2.4 and 40-150 A2.4
`P.10 were identical, the appearance of a truncated H chain in
`40-150 A2.4 P.10 was not due to a mutation encoded in V" or
`to a mutated leader signal sequence leading to an aberrant
`cleavage. It is probable that the (partial) H chain truncation
`in this ceU line occurs from an as yet unidentified mutation
`affecting posttranslational modification of the H chain. The
`Val-Lys bond (positions 2 and 3, Fig. 3) is not a conventional
`cleavage site for signal peptidases (34, 35), although cleavage
`by an endopeptidase specific for the amino-terminal side of
`lysine (36) and aminopeptidase (37, 38) are possible expla­
`nations.
`Models of the structures of 40-150 and its two mutants
`were constructed by computer (J. Novotny, R. E. Brucco­
`leri, and E.H., unpublished data) based on atomic coordi­
`nates of the anti-phosphocholine antibody McPC 603. The
`results indicate that when Arg(H)-94, is substituted for
`Ser(H)-94 (as in 40-150 A2.4), Arg(H)-94 can form a chain of
`hydrogen bonds to Asp(H)-101 and then to Arg(L)-46 and
`Asp(L)-55, aU in the vicinity of the combining site. The
`resultant change in the antigen combining site surface may
`account for the affinity and specificity changes observed in
`40-150 A2.4. Deletion of the two amino-terminal H chain
`residues (as in 40-150 A2.4 P.10) increases solvent accessi­
`bility to Arg(H)-94 by 100%. The solvation may result in the
`loss of a hydrogen bond between this residue and Asp(H)-
`101, restoring the structure to one similar to that of the
`parent 40-150 antibody.
`Full molecular dissection of antigen-antibody complemen­
`tarity requires production of new binding sites by site­
`directed mutagenesis, but the results presented here empha­
`size the risk of a narrow focus in designing such experi­
`ments. Information from an array of spontaneous mutants as
`well as from computer modeling studies and x-ray crystaUo­
`graphic analyses is needed to guide such work.
`
`This work was supported by grants from the National Institutes of
`Health (HL 19259, Al 19512, and GM 17367).
`1. Tonegawa, S. (1983) Nature (Lendon) 302, 575-581.
`2. Honjo, T. & Habu, S. (1985) Annu. Rev. Biochem. 54,
`803-830.
`3. Wu, T. T. & Kabat, E. A. (1970) J. Exp. Med. 132, 211-250.
`4. Saul, F. A., Amzel, L. M. & Poljak, R. J. (1978) J. Biol.
`Chem. 253, 585-597.
`5. Marquart, M., Deisenhofer, J., Huber, R. & Palm, W. (1978)J.
`Mo/. Biol. 141, 369-391.
`6. Satow, Y., Cohen, G. H., Padlan, E. A. & Davies, D. R.
`(1986) J. Mo/. Biol. 190, 593-604.
`7. Suh, S. W., Navia, M. A., Cohen, G. H., Rao, D. N., Rudi·
`koff, S. & Davies, D. R. (1986) Proteins I, 74-80.
`8. Porter, R. R. (1973) Science 180, 713-716.
`9. Padlan, E. A. & Davies, D. R. (1975) Proc. Natl. Acad. Sci.
`USA 72, 819-823.
`
`10. Kohler, G. & Milstein, C. (1975) Nature (Lendon) 256,
`495-497.
`11. Rudikoff, S., Giusti, A. M., Cook, W. D. & Scharff, M. D.
`(1982) Proc. Natl. Acad. Sci. USA 19, 1979-1983.
`12. Cook, W. D., Rudikoff, S., Giusti, A. M. & Scharff, M. D.
`(1982) Proc. Natl. Acad. Sci. USA 19, 1240-1244.
`13. BrOggemann, M., Radbruch, A. & Rajewsky, K. (1982) EMBO
`J. 5, 629-634.
`14. Radbruch, A., Zaiss, S., Kappen, C., BrOggemann, M., Bey­
`reuther, K. & Rajewsky, K. (1985) NaJure (Lendon) 315,
`506-508.
`15. Dildrop, R., BrOggemann, M., Radbruch, A., Rajewsky, K. &
`Beyreuther, K. (1982) EMBO J. 5, 635-640.
`16. Go, K. & Kartha, G. (1980) Acta Crystal/ogr. Sect. B 36,
`1811-1819.
`17. Mudgett-Hunter, M., Margolies, M. N., Ju, A. & Haber, E.
`(1982) J. Immuno/. 129, 1165-1172.
`18. Mudgett-Hunter, M., Anderson, M., Haber, E. & Margolies,
`M. N. (1985) Mo/. lmmunol. 22, 477-488.
`19. Smith, T. W., Butler, V. P. & Haber, E. (1970) Biochemistry
`9, 331-337.
`20. Herzenberg, L. A., Kipps, T. J., Peterson, L. & Parks, D.R.
`(1985) in Biotechnology in Diagnostics, eds. Koprowski, H.,
`Ferrone, S. & Albertini, A. (Elsevier, Amsterdam), pp. 3-16.
`21. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. & Rutter,
`W. J. (1979) Biochemistry 18, 5294-5299.
`22. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular
`Cloning: A Laboratory Manual (Cold Spring Harbor Lab.,
`Cold Spring Harbor, NY), pp. 197-198.
`23. Schlomchik, M. J., Nemazee, D. A., Sato, V. L., Van Snick,
`J., Carson, D. A. & Weigert, M. G. (1986) J. Exp. Med. 164,
`407-427.
`24. Panka, D. J. & Margolies, M. N. (1987) J. Jmmunol. 139,
`2385-2391.
`25. Juszczak, E. C. & Margolies, M. N. (1983) Biochemistry 22,
`4291-4296.
`26. Smith, J. A. & Margolies, M. N. (1984) Biochemistry 23,
`4726-4732.
`27. Brauer, A. W., Oman, C. L. & Margolies, M. N. (1984) Anal.
`Biochem. 137, 134-142.
`28. Smith, J. A. & Margolies, M. N. (1987) Biochemistry 26,
`604-612.
`29. Hudson, N. W., Mudgett-Hunter, M., Panka, D. J. & Margo­
`lies, M. N. (1987) J. lmmunol. 139, 2715-2723.
`30. IUPAC-IUB Commission on Biochemical Nomenclature
`(1968) J. Biol. Chem. 243, 3557-3559.
`31. Kabat, E. A., Wu, T. T., Bilofsky, H., Reid-Miller, M. &
`Perry, H. (1987) Sequences of Proteins of Immunological
`Interest (Natl. Inst. Health, Bethesda, MD).
`32. Davies, D. R., Pad.Ian, E. A. & Segal, D. M. (1975) in Con·
`temporary Topics in Molecular Immunology, eds. Inman,
`F. P. & Mandy, W. J. (Plenum, New York), Vol. 4, pp.
`127-155.
`33. Amit, A.G., Mariuzza, R. A., Phillips, S. E. V. & Poljak,
`R. J. (1986) Science 233, 747-753.
`34. von Heijne, G. J. (1984) J. Mo/. Biol. 173, 243-251.
`35. von Heijne, G. J. (1983) Eur. J. Biochem. 133, 17-21.
`36. Wingard, M., Matsueda, G. & Wolfe, R. S. (1972) J. Bacterio/.
`112, 940-949.
`37. Achstetter, T., Ehmann, C. & Wolf, D. H. (1983) Arch.
`Biochem. Biophys. 226, 292-305.
`38. Frey, J. & R6hm, K. (1978) Biochim. Biophys. Acta 521,
`31-41.
`
`5 of 5
`
`BI Exhibit 1135
`
`