Using the polyfructose, bacterial levan, as a model polysaccharide, we analyzed how V regions affect binding in anti-polysaccharide mAbs. Previously, panels of mAb were constructed from bacterial levan-immunized BALB/c and CBA/Ca mice. The BALB/c mAb were mostly germline VHJ606:Vκ11, and a subset contained presumed somatic mutations in the complementarity-determining regions (CDRs) that correlated with increases in avidity for the β(2→1) inulin linkage of levan. The CBA/Ca mAb were more heterogeneous in V gene usage, but a subset of inulin-nonreactive mAb were VHJ606:Vλ and had VH sequence differences in the CDRs from the VHJ606 regions of the BALB/c mAb. In this report, VHJ606 Abs containing various combinations of specifically mutated H and L chains were produced by engineered transfectants and tested for inulin avidity and levan binding. Two presumed somatic mutations seen in CDRs of the BALB/c hybridomas were shown to directly cause marked increases in avidity for inulin (VH N53H, 9-fold; VL N53I, 20-fold; together, 46-fold) but not for β(2→6) levan. Exchange of either positions 50 or 53 in VH or the H3 loop between the BALB/c and CBA/Ca mAb resulted in either fine specificity shift or total loss of bacterial levan binding. Three-dimensional models of the V regions suggested that residues that affect binding to inulin alone are near the edge of the CDR surface, while residues involved with binding both forms of levan and affecting fine specificity are in the VH:VL junctional area.

The polysaccharide (PS)4 capsule of encapsulated bacterial pathogens acts as a virulence factor for the organism, allowing it to escape the nonspecific host defense system (1, 2). Capsule-specific anti-PS Abs have been shown to play a critical role in host protection (Refs. 3, 4, 5, 6 and reviewed in Refs. 7 and 8) and passive administration of anti-PS Abs has demonstrated protection of high-risk populations from encapsulated pathogens (9). Given the emergence of multidrug-resistant encapsulated bacteria (10), protective anti-capsular Abs are increasingly important candidates for clinical use (11). Naturally occurring anti-PS Abs have low affinity (12), and use limited numbers of V genes (Refs. 13, 14, 15, 16, 17, 18, 19 and reviewed in Ref. 20). Improving anti-PS Ab affinity and targeting anti-PS specificity will require determining the residues in the V regions that are critical for Ig-PS binding and the conformation in solution of the PS. Understanding the contributions of individual mutations is important in the design and engineering of high-affinity mAb to PS.

Ab/Ag interactions are formed by complementarity between the Ag-binding complementarity-determining region (CDR) surface in the V region of the Ab and the structure of the Ag (reviewed in Ref. 21). Ag binding involves multiple noncovalent interactions between atoms in the Ag and the CDR surface, although only about one-third to one-fifth of the CDR surface participates directly in Ag contact. In the three-dimensional structure of the V regions, the residues in the CDR form loops, H1–3 in the H chain and L1-L3 in the L chain, between the strands of the β-pleated sheet of the Ig domain. Three-dimensional structural analysis has defined canonical structures for five of the six CDR loops, facilitating the prediction of the structure of Abs for which the three-dimensional structure has not been resolved (22). The H3 loop has eluded classification, although some rules delineating its structure have been described (23).

During the course of an Ab response, amino acid substitutions accumulate in the V regions in a process called somatic mutation or affinity maturation (24). Often, the amino acid substitutions that cause affinity changes are in CDR regions (25, 26), in many instances allowing for the formation of new contacts. However, the relationship between Ag contact and affinity is more complex in that some substitutions that eliminate contacts do not change affinity, possibly because interactions with water molecules form in the absence of direct contacts (27) and sites outside of the CDRs influence affinity (21). Additional mechanisms by which amino acid changes can increase affinity include the elimination or addition of glycosylation sites (28, 29) and perturbation of the H:L chain interaction (30) or nearby CDR loop structures (31).

Our lab has focused on Ab responses to levan (polyfructose) as a model PS. Levan can be classified based on the linkages of the repeating fructoses. Levan made by plants can be either polyfructose with predominantly β(2→6) linkages or polyfructose with predominantly β(2→1) linkages, also known as inulin (32). Bacterial levan (BL) consists of a backbone of β(2→6)-linked fructoses and β(2→1)-linked branches (33). Extensive studies have been performed on Abs raised in mice immunized with BL (18, 34) and levan-binding myeloma proteins that have arisen spontaneously (35, 36, 37). Like most anti-PS Abs, these Abs and myeloma proteins have modest affinity for levan (36). Analysis of the myeloma proteins has defined at least two fine specificities: inulin binding (e.g., J606, UPC61; Refs. 34 and 37, 38, 39) and inulin nonreactive (e.g., UPC10; Ref. 35).

Previously, our lab produced panels of anti-BL mAb from BALB/c and CBA/Ca mice (40) in which a fraction of the Abs contained VHJ606 H chains. The mAb from BALB/c were largely germline VHJ606:Vκ11 and similar to the myeloma protein J606 in that they bound inulin. Apparent somatic mutations and higher avidity for inulin were evident in some of the BALB/c mAb: N53H in the H chain CDR2; N53I in the Vκ CDR2; and additional mutations in the VH CDR2. The CBA/Ca mAb from the panel were more heterogeneous in VH usage, but a fraction were VHJ606:Vλ. The VHJ606:Vλ CBA/Ca Abs further differed from the BALB/c Abs in that they didn’t bind inulin and had H3 loops that were 2 aa longer. CDR2 sequence differences (E50Q and N53D) suggested that the CBA/Ca mice used a different J606 VH gene than BALB/c mice.

In this report, we have extended our studies by constructing engineered IgM Abs and V region models to determine how inulin avidity and levan fine specificity are affected by CDR substitutions.

RNA was isolated from hybridoma cells using a denaturation-based procedure (RNAgents; Promega, Madison, WI). VH and VL region segments were amplified by PCR with the following oligonucleotides: 5′-GGGACGACTAGTTTTACATTGGGATTCATCTCTGC-3′ and 5′-AGGTCCAACTTCTCGAGTCTGG-3′ for VHJ606, 5′-CCAGATGTGAGCTCGTGATGACCCAGACTCCA-3′ and 5′-GCGCCGTCTAGAATTAACACTCATTCCTGTTGAA-3′ for Vκ11, and 5′-CACTGTCACCACACCTGGGTAGAA-3′ and 5′-TGGCCTGGACTTCACTTATATCTC-3′ for Vλ. The PCR products were inserted into the pCR2.1-TOPO vector (Invitrogen, Carlsbad, CA) and sequenced on both strands (Bioserve, Laurel, MD). Each V gene sequence was confirmed by sequencing a second, independently derived PCR product from the same cell line. DNA sequence analysis was performed using the Wisconsin Package version 9.1 from Genetics Computer Group (Madison WI). Sequences were aligned using the clustal algorithm (41) and the blosum 62 (42) amino acid similarity scoring matrix.

RNA was isolated from BBLC44.1 hybridoma cells using RNAgents (Promega). A cDNA library from this RNA was constructed using the SuperScript system (Life Technologies, Gaithersburg, MD). The cDNA library was screened with a μ-chain DNA probe (pABμ-8; Ref. 43), and a full-length μ-chain cDNA was isolated and sequenced. The μ-chain cDNA was transferred from the pSPORT-1 library vector to the pCIneo CMV promoter-driven expression vector (Promega). Site-directed mutagenesis of this construct was performed with an oligonucleotide-based method (Clontech, Palo Alto, CA) and the following oligonucleotides: N53H, 5′-GATTGAAATCTCATAATTATGCAA-3′; N53D, 5′-GATTGAAATCTGATAATTATGCAA-3′; E50Q and N53D, 5′-GTGGGTTGCTCAAATTAGATTGAAATCTGATAATTATGC-3′; E50Q, 5′-GTGGGTTGCTCAAATTAGATTGAA-3′; T57I with A60V, 5′-TAATAATTATGCAATACATTATGTGGAGTCTGT3′. The following oligonucleotide, 5′-CTAGCCTCGAGAAGTCCCGGGTCG-3′, was used for selection during the mutagenesis procedure. The correct sequence of the V region for each mutated construct was verified by sequencing.

H chain loss mutants were produced from the hybridomas BBLC44.1, CBL166.10, and 2BBLC803.1 by selecting L chain-only producing variants from limiting dilution cultures (44). Approximately 500 hybridoma cells were seeded into 96-well plates at a density of 0.5 cells/well. Independent colonies were screened for L chain expression by a L chain-specific ELISA and for intact IgM production by an Ag-binding or a μ-chain-specific ELISA. About 1–2% of the colonies produced L chain, but not intact IgM. Northern analysis detected high levels of expression of L chain mRNA and greatly diminished levels of unproductive μ-chain mRNA in the H chain loss mutants (data not shown). H chain loss mutants from BBLC44.1 (Vκ11 germline), CBL166.10 (Vλ), and 2BBLC803.1 (Vκ11 N53I) were created as transfection substrates. The DNA sequences of the L chain V regions from the H chain loss mutants were determined and found to be identical with the L chain V regions of the original hybridomas (data not shown).

H chain loss mutant cell substrates were transfected with μ-chain constructs by electroporation and selected with 400 μg/ml G418 (Life Technologies). Optimal electroporation voltages for each H chain loss mutant were determined by transient expression of β-galactosidase by an adenovirus major late promoter-driven construct, pADβ (Clontech). Transfectants were screened for production of correct H and L chain-paired IgM by ELISA. Relatively high-producing transfectants (5–100 ng/ml IgM in cell-culture supernatants) were subcloned and used to produce IgM for binding studies. The mutants, as listed in Table I, are numbered and referred to as follows: 1, VHJ606:Vκ11 germline; 2, VHJ606 N53H; 3, VHJ606 N53D; 4, VHJ606 E50Q; 5, VHJ606 E50Q + N53D; 6, VHJ606:Vλ; 7, VHJ606 E50Q + N53D + Vλ; 8, Vκ11 N53I; 9, VHJ606 N53H + Vκ11 N53I; 10, VHJ606 T57I + A60V.

Table I.

Avidity of mAb for BL, GL, and inulin

mAbVH GeneAmino AcidVL GeneAmino AcidBLa Endpoint TiterGLb Endpoint TiterInulinc
50535760CDR353Endpoint titerRelative titer
Transfectants             
1 (germline) J606 GFAY κ11 22,124 × /÷ 1.4 19,463 × /÷ 1.5 4,390 × /÷ 2.0 
J606 H GFAY κ11 25,614 × /÷ 1.3 46,020 × /÷ 1.4 41,612 × /÷ 1.6 9.5 
J606 D GFAY κ11 32,498 × /÷ 1.5 25,150 × /÷ 1.1 232 × /÷ 2.6 0.053 
J606 Q GFAY κ11 <4 ND <6 
J606 Q D GFAY κ11 <4 ND <6 
J606 GFAY λCZ72  24 × /÷ 4.8 ND <6 
J606 Q D GFAY λCZ72 <4 ND <6  
J606 GFAY κ11 I 35,614 × /÷ 1.3 39,750 × /÷ 2.6 88,966 × /÷ 1.9 20.3 
J606 H GFAY κ11 I 34,333 × /÷ 1.6 25,614 × /÷ 2.4 202,797 × /÷ 1.6 46.2 
10 J606 I V GFAY κ11 N 14,257 × /÷ 2.4 ND 3,368 × /÷ 2.7 0.77 
Hybridomas/ myeloma             
BBLC44.1 J606 E N T A GFAY κ11 N 32,695 × /÷ 1.7 21,854 × /÷ 1.7 4,275 × /÷ 1.6 0.97 
UPC10 X24      κ10  60,566 × /÷ 1.3 41,994 × /÷ 4.7 <6 
CBL166.10 J606 Q D LWEFAY λCZ72  7,241 × /÷ 1.5 <8 <6 
mAbVH GeneAmino AcidVL GeneAmino AcidBLa Endpoint TiterGLb Endpoint TiterInulinc
50535760CDR353Endpoint titerRelative titer
Transfectants             
1 (germline) J606 GFAY κ11 22,124 × /÷ 1.4 19,463 × /÷ 1.5 4,390 × /÷ 2.0 
J606 H GFAY κ11 25,614 × /÷ 1.3 46,020 × /÷ 1.4 41,612 × /÷ 1.6 9.5 
J606 D GFAY κ11 32,498 × /÷ 1.5 25,150 × /÷ 1.1 232 × /÷ 2.6 0.053 
J606 Q GFAY κ11 <4 ND <6 
J606 Q D GFAY κ11 <4 ND <6 
J606 GFAY λCZ72  24 × /÷ 4.8 ND <6 
J606 Q D GFAY λCZ72 <4 ND <6  
J606 GFAY κ11 I 35,614 × /÷ 1.3 39,750 × /÷ 2.6 88,966 × /÷ 1.9 20.3 
J606 H GFAY κ11 I 34,333 × /÷ 1.6 25,614 × /÷ 2.4 202,797 × /÷ 1.6 46.2 
10 J606 I V GFAY κ11 N 14,257 × /÷ 2.4 ND 3,368 × /÷ 2.7 0.77 
Hybridomas/ myeloma             
BBLC44.1 J606 E N T A GFAY κ11 N 32,695 × /÷ 1.7 21,854 × /÷ 1.7 4,275 × /÷ 1.6 0.97 
UPC10 X24      κ10  60,566 × /÷ 1.3 41,994 × /÷ 4.7 <6 
CBL166.10 J606 Q D LWEFAY λCZ72  7,241 × /÷ 1.5 <8 <6 
a

Binding is expressed as the endpoint titer of BL binding per μg of Ab. The BL binding data presented in this table are the geometric mean of three separate measurements; the error is the geometric standard deviation.

b

Binding is expressed as the endpoint titer of GL binding per μg of Ab. The GL binding data presented in this table are the geometric mean of three separate measurements; the error is the geometric standard deviation.

c

Inulin binding is expressed as the absolute endpoint titer per μg of Ab of each hybridoma, myeloma, or transfectant mAb and as relative to that of the reference mAb (Ab number 1). Measurement of inulin binding strength was measured in duplicate in three independent experiments on separate days. Inulin binding measurements were made using IgM produced by three (mutants 8 and 9), two (reference 1 and mutants 2 and 3), or one (mutant 10) independent cell clones. Data are the geometric mean of all combined measurements for each mutant, and the error is the geometric standard deviation. Avidity changes induced by VH N53H, VH N53D, VL N531, and VH N53H + VL N53I (e.g., mutant 1 vs mutants 2, 3, 8, and 9) and differences in binding between single and double mutants (e.g., mutants 2 and 8 vs mutant 9) were found to be statistically significant (p < 0.01 in the Student’s t test). In contrast, the avidity change induced by VHJ606 T57I, A60V (e.g., mutant 1 vs mutant 10) was not statistically significant.

IgM-secreting transfectants were grown in either CL350 gas-permeable flasks (Integra Biosystems, Ijamsville, MD) or in T175 tissue culture flasks in RPMI 1640 supplemented with 2% FCS (HyClone, Logan, UT). IgM in the culture supernatants was concentrated using an Amicon 8200 concentration device (Beverly, MA) with 100 kDa nominal molecular weight cut-off membranes to a concentration of 50–1000 ng/ml.

Pentameric IgM was separated from monomeric μ-chains, free κ-chains, and most FCS proteins by size exclusion chromatography. A 200-μl sample from each concentrate was run on a Superose 6 HR 10/30 size exclusion chromatography column (Pharmacia, Uppsala, Sweden) using an fast protein liquid chromatography system (Pharmacia). The column was run at a flow rate of 0.4 ml/min in PBS; 2-min fractions were collected. Pentameric IgM (∼960 kDa) eluted after ∼28 min, while lower molecular mass IgM-related species eluted after ∼40 min (detected by ELISA). Fractions containing 2–100 ng/ml pentameric IgM were used to measure inulin binding activity. Between 40–80% of the IgM from each raw concentrate was pentameric (data not shown).

ELISAs were performed essentially as in Boswell and Stein (40). ELISA plates (Dynatech, Chantilly VA) were coated with inulin-BSA (1 μg/ml in PBS; a gift from Dr. J. Inman, National Institutes of Health, Bethesda, MD; Ref. 45), BL (2 μg/ml in PBS; a gift from Dr. Cornelis Glaudemans, National Institutes of Health), rye grass levan (GL; 50 μg/ml in PBS; a gift from Dr. Constantin Bona, Mt. Sinai Medical School, New York, NY) or goat anti-mouse κ- or λ-chain (2 μg/ml in 15 mM NaCO3, pH 9.6; Southern Biotechnology Associates, Birmingham, AL). Following washing and blocking with 1% BSA, a 1:3 dilution series of IgM starting between 2 and 100 ng/ml was allowed to bind overnight to the plate-bound Ags. The bound IgM was detected with goat anti-mouse IgM-alkaline phosphatase (Southern Biotechnology Associates) and methylumbelliferyl phosphate as a substrate (Sigma, St. Louis, MO). UPC10, a β(2→6) levan-specific IgG2a:κ myeloma protein, was detected with goat anti-mouse γ2a-alkaline phosphatase (Southern Biotechnology Associates). The endpoint titer for each Ab-Ag reaction was calculated by extrapolation of the linear portion of the binding curve to the assay baseline. The starting concentration of intact IgM Ab was determined by comparison of the end point titer to an IgM standard (PharMingen, San Diego, CA) in the anti-κ or anti-λ capture ELISA. The binding strength for each Ab was calculated by dividing the extinction Ab titer by the starting IgM concentration.

Measurement of inulin, GL, and BL binding strength was measured in duplicate in three independent experiments on separate days. Inulin binding strength measurements were made using IgM produced by three (VL N53I, mutant 8; VL N53I + VH N53H, mutant 9), two (VH N53H, mutant 2; VH N53D, mutant 3; germline; Ref. 1), or one (VH T57I + VH A60V, mutant 10) independent cell clones. Data in Table I are the geometric mean of all combined measurements for each mutant. The ELISA transfection-based system used in this report was validated in two ways. First, the inulin binding strengths of germline Abs produced by hybridomas and transfectants were compared and found to be equivalent (Table I). Second, the inulin binding strengths of hybridoma produced mAb were remeasured and found to be equivalent to the data produced several years ago (Ref. 40 and data not shown).

Modeling of V regions was performed using the Insight II version 97.2 (Molecular Simulations, San Diego, CA) package of software running on a Silicon Graphics O2 workstation (Mountain View, CA). Molecular models of the mutagenized mAb were produced using a modification of the canonical structures method (46). Reference structures for modeling protein database (PDB) structure 1IAII for VHJ606, Ref. 47 ; PDB structure 1FVDA for Vκ11, Ref. 31 ; PDB structure 1NGQ for Vλ, Ref. 48) were chosen based on sequence identity (85% between 1IAII and VHJ606; 73% between 1FVDA and Vκ11; 95% between 1NGQL and Vλ), the quality of the existing crystallographic structure (1IAI is 2.9 A; 1FVD is 2.2 A; 1NGQ is 2.4 A), and the likelihood of canonical structure based on an analysis of the key residues in the V regions of the anti-BL mAb and the published classifications for the resolved reference structures (22, 49) (1IAI and VHJ606 are classified as H11, H24; 1FVDA and Vκ11 are classified as L12, L21, L31; 1NGQL is L13, while the canonical structure of L1 of the anti-BL Vλ is indeterminate based on key residue analysis, the key residues defined by Chothia et al. (22) are 85% identical between 1NGQL and Vλ, the remaining loops of 1NGQL and Vλ are classified as L21 and L31). To construct the models, the L chains of 1FVD and 1NGQ were superimposed on the L chain of 1IAI creating VH/VL pairings with the VH of 1IAI. This was performed using the homology module of Insight II, so as to minimize the root mean square distance between framework residues of the three structures. Amino acid coordinates for the models were assigned from the reference structures, except for the H3 loops, for which the areas of correspondence between the 1IAI reference structure and the VHJ606 sequences were not obvious. To model the 9 and 11 amino acid H3 loops of the BALB/c and CBA/Ca Abs, H3 loops of the same length were identified in separate Ab structures (PDB structure 1NSN, Ref. 50 ; PDB structure 1BAF, Ref. 51) and used to provide amino acid coordinates. Using the discover module of Insight II, peptide bonds at the H3 loop splice junctions in the models were repaired, and then the side chains of all model residues were subjected to energy minimization.

In our previous analysis of a panel of anti-BL hybridomas, a correlation was noted in the BALB/c mAb between high avidity for inulin and presumed somatic mutations in the CDR regions, particularly VH N53H (40). BBLC44.1, with moderate avidity for inulin and an Asn at position 53, is a germline prototype (VHJ606:Vκ11) of these mAb. To strengthen this correlation, the VH and VL sequences of two previously unsequenced high-avidity anti-BL mAb from the panel were determined (2BBLC811.2 and 2BBLC854.5; Fig. 1). In each case, N53H was present, although additional L and H chain substitutions were noted. The V region sequence of mAb 2BBLC854.5 differs from the germline VHJ606:Vκ11 hybridoma BBLC44.1 only in two CDR residues, N53H in VH and S30N in VL CDR1. Thus, two CDR mutations together, one VH, one VL, directly cause a marked increase in inulin avidity (∼25-fold; Ref. 40). The V region of 2BBLC811.2 has framework region (FR) mutations in addition to N53H.

FIGURE 1.

Sequences of anti-BL mAb. The VH (A) and VL (B) amino acid sequences of high-avidity, inulin-specific BALB/c VHJ606:Vκ11 mAb (BBLC310.2, 2BBLC803.1, 2BBLC811.2, and 2BBLC854.5) and the inulin nonreactive CBA/Ca VHJ606:Vλ mAb (CBL166.10) are shown aligned to the germline BALB/c VHJ606:Vκ11 (BBLC44.1) amino acid sequence. The VH sequence of BBLC44.1, BBLC310.2, 2BBLC803.1, and CBL166.10 and the VL sequence of BBLC44.1, BBLC310.2, and 2BBLC803.1 were reproduced from Boswell and Stein (Ref. 40 ; with permission of The Journal of Immunology, copyright 1996, the American Association of Immunologists). The additional VH and VL exon nucleotide sequences determined for this report are in the GenBank database, accession numbers AF132844–8. Dots denote amino acid identities; dashes are gaps. CDR defined by sequence variability (52 ) are indicated by brackets; residues that correlate with increased avidity for inulin are boxed in red; residues that correlate with fine specificity changes are boxed in blue. Boswell and Stein (40 ) previously found that increases in avidity over germline for BBLC310.2, 2BBLC803.1, 2BBLC811.2, and 2BBLC854.5 were 13-, 65-, 33-, and 25-fold, respectively.

FIGURE 1.

Sequences of anti-BL mAb. The VH (A) and VL (B) amino acid sequences of high-avidity, inulin-specific BALB/c VHJ606:Vκ11 mAb (BBLC310.2, 2BBLC803.1, 2BBLC811.2, and 2BBLC854.5) and the inulin nonreactive CBA/Ca VHJ606:Vλ mAb (CBL166.10) are shown aligned to the germline BALB/c VHJ606:Vκ11 (BBLC44.1) amino acid sequence. The VH sequence of BBLC44.1, BBLC310.2, 2BBLC803.1, and CBL166.10 and the VL sequence of BBLC44.1, BBLC310.2, and 2BBLC803.1 were reproduced from Boswell and Stein (Ref. 40 ; with permission of The Journal of Immunology, copyright 1996, the American Association of Immunologists). The additional VH and VL exon nucleotide sequences determined for this report are in the GenBank database, accession numbers AF132844–8. Dots denote amino acid identities; dashes are gaps. CDR defined by sequence variability (52 ) are indicated by brackets; residues that correlate with increased avidity for inulin are boxed in red; residues that correlate with fine specificity changes are boxed in blue. Boswell and Stein (40 ) previously found that increases in avidity over germline for BBLC310.2, 2BBLC803.1, 2BBLC811.2, and 2BBLC854.5 were 13-, 65-, 33-, and 25-fold, respectively.

Close modal

CBL166.10, a mAb from the panel of CBA/Ca anti-BL hybridomas, has a VHJ606 H chain and a previously uncharacterized λ L chain. The Vλ exon from this hybridoma was isolated, sequenced, and identified as 86% identical with a Vλ from a wild mouse (Mus musculus musculus) isolated in Sladeckovce, Czechoslovakia, CZ72 (Fig. 1; Ref. 53).

Because almost all of the anti-BL mAb from the hybridoma panels contained presumed somatic mutations in the FR as well as the CDR, engineered IgM-secreting transfectants were constructed to isolate the affects of individual mutations on binding. The mutants were produced by transfection of μ-chain expression constructs into H chain loss variants of the original anti-BL hybridomas. The μ-chain expression constructs were produced by inserting a BALB/c germline VHJ606 cDNA (from BBLC44.1) into a mammalian expression vector and subsequent mutagenesis. A mAb with germline VHJ606 and Vκ11 V regions was constructed to serve as a control for the purification of transfectant IgM and as the germline reference. Fast protein liquid chromatography-purified pentameric IgM was used in ELISAs to determine inulin binding activity and concentrated harvests were used in ELISAs to determine BL and rye GL (pure β(2→6) levan) binding activity (Table I).

Two mutations predicted by the hybridoma panel to increase inulin avidity, N53H in VH (mutant 2) and N53I in VL (mutant 8), in fact do so (Table I). VH N53H increases inulin binding activity 9.5-fold. This increase is comparable to the ∼13-fold increase seen in the BBLC310.2 hybridoma (N53H and S30T in VH), arguing that the increased avidity of this Ab is mostly due to VH N53H (40). Vκ N53I in CDR2 increases avidity 20-fold. VH N53H and Vκ N53I together (mutant 9) have an additive effect, increasing avidity 46-fold. Thus, these two substitutions almost fully reconstitute the high avidity (65-fold of germline) seen in the very highest mAb from the hybridoma panel, 2BBLC803.1. Two additional VH CDR2 mutations seen in 2BBLC803.1, T57I and A60V (mutant 10), had a relatively small effect on inulin binding (0.8-fold change). Therefore, the very high avidity of 2BBLC803.1 appears to result mainly from two presumed somatic mutations; one in CDR2 of VL and one in CDR2 of VH. A negatively charged amino acid found in the CBA/Ca mAb at position 53, VH N53D (mutant 3), decreases avidity 15-fold.

Abs that bind to BL containing a β(2→6) backbone with β(2→1) branch linkages can be reactive with the β(2→1) inulin determinant or the β(2→6) levan determinant. Prototype mAb and myeloma proteins for different fine specificities have been described and classified based on binding studies (37). J606, with the same VH:Vκ pairing as the germline BBLC44.1, binds β(2→1) inulin, and has been described by some, but not all, investigators also to bind β(2→6) levan (34, 37, 38, 39). UPC10 has a β(2→6) levan fine specificity. Some Abs appear to bind BL, but neither pure β(2→1) inulin, nor pure β(2→6) levan (37). For our fine specificity analysis, BBLC44.1 was selected as a prototype VHJ606:Vκ11 mAb from the BALB/c hybridoma panel, and CBL166.10 was selected as a prototype VHJ606:Vλ mAb from the CBA/Ca hybridoma panel. The fine specificities of these two prototypes and the mAb from the transfectant panel were determined by inulin-, BL-, and rye GL (pure β(2→6) levan)-specific ELISAs (Table I). UPC10, a VHX24:Vκ10 myeloma protein specific for β(2→6) levan, was included as a reference.

It is evident that our VHJ606:Vκ11 inulin-reactive mAb are similar to the previously described mAb 2-1-3 (37) in that they also bind pure β(2→6) levan. Interestingly, mutations that increase or decrease inulin binding 9- to 46-fold (VHN53H, mutants 2 and 9; VHN53D, mutant 3; or VLN53I, mutants 8 and 9) have only a small effect (1.2- to 2.4-fold) on avidity for GL or BL. The VHJ606:Vλ CBA/Ca mAb, CBL166.10, appears to be similar to the previously described mAb 2-11-3 (37) in that it binds neither pure β(2→1) nor pure β(2→6) levan (GL), while binding BL.

As seen above, the BALB/c (BBLC44.1) and CBA/Ca (CBL166.10) Abs both bind BL and have VHJ606 H chains, yet they differ in fine specificity. This fine specificity difference results from different V region usage in the L chain (BBLC44.1 is Vκ11 vs CBL166.10 is Vλ) and/or a small number of V region sequence differences in the H chain (the CBL166.10 H3 loop is 2 aa longer; the CDR2 has 2 aa differences from BBLC44.1). To explore whether the VH differences between the BALB/c and CBA/Ca Abs influenced fine specificity, or whether fine specificity is determined solely by VL, engineered VHJ606:Vκ11 and VHJ606:Vλ mAb that essentially exchanged residues or short regions of the VH regions between them were constructed and tested for fine specificity by ELISA (Table I). A single mutation, VHE50Q (mutant 4) found in the CBA/Ca mAb CBL166.10, ablated all BL binding activity when engineered into a BALB/c germline VHJ606:Vκ11 mAb. BL binding was ablated in this engineered mAb with or without simultaneously grafting in the CBA/Ca VH residue 53D (mutant 5). Grafting the H3 loop and a VH FR residue, 84N, from BBLC44.1 into a VHJ606:Vλ mAb that otherwise had the same sequence as CBL166.10 also ablated all BL binding activity (mutant 7). BL binding also was ablated in this engineered Ab when simultaneously grafting in two additional BALB/c VH CDR2 residues, 50E and 53N (mutant 6). While these mutant Abs contain a BALB/c FR VH84 residue that was unavoidably introduced by the construction procedure, it is unlikely that VH84 is responsible for the BL binding loss in that this position has not been found to participate in Ag contact in the >70 currently resolved Ag:Ab complexes (E. Padlan, unpublished observations), and this residue is spatially removed from the CDR surface (Fig. 2). Furthermore, the total loss of BL binding in these mutant Abs cannot be explained by incorrect H and L chain pairing because intact IgM was detected in an anti-L chain capture ELISA developed with anti-μ. Therefore, Abs with VHJ606 H chains are capable of binding two determinants of levan, but binding to the two determinants is contingent on the correct VL paired with VHJ606 H chains with correct residues in the H3 loop and 50 in VH.

FIGURE 2.

Three dimensional models of anti-BL mAb. Models of anti-levan Abs are shown as solvent-accessible surfaces (54 ). CDR surfaces are shown colored light green (VL) or light lavender (VH), while FR surfaces are gray. The full H3 loop, defined by three-dimensional analysis as extending from 92C to 103W, were grafted from separate structures during the construction of the models and are colored based on this definition (55 ). A, The germline BALB/c VHJ606:Vκ11 Ab are shown with higher- and lower-avidity mutants. Residues changed in this study are shown colored yellow in the reference Ab. Residues that cause high inulin avidity (53H in VH and 53I and 30N in VL) are shown colored red, while residues shown to be not critical for inulin binding (VH 57T and 60V) are shown colored orange. Residues present in the CBA/Ca Ab that when grafted into the germline Ab cause decreased inulin binding are shown colored blue. B, The CBA/Ca VHJ606:Vλ Ab shown with a non-BL binding mutant constructed using BALB/c H chains and CBA/Ca L chains. BALB/c VH residues present in the non-BL binding, constructed VHJ606:Vλ Ab are shown colored yellow. VH 50Q and 53D present in the CBA/Ca mAb and engineered into the mutant mAb are shown colored blue. A single BALB/c FR residue, VH 84N, unavoidably introduced into the mutant mAb by the engineering procedure, is shown colored pale yellow.

FIGURE 2.

Three dimensional models of anti-BL mAb. Models of anti-levan Abs are shown as solvent-accessible surfaces (54 ). CDR surfaces are shown colored light green (VL) or light lavender (VH), while FR surfaces are gray. The full H3 loop, defined by three-dimensional analysis as extending from 92C to 103W, were grafted from separate structures during the construction of the models and are colored based on this definition (55 ). A, The germline BALB/c VHJ606:Vκ11 Ab are shown with higher- and lower-avidity mutants. Residues changed in this study are shown colored yellow in the reference Ab. Residues that cause high inulin avidity (53H in VH and 53I and 30N in VL) are shown colored red, while residues shown to be not critical for inulin binding (VH 57T and 60V) are shown colored orange. Residues present in the CBA/Ca Ab that when grafted into the germline Ab cause decreased inulin binding are shown colored blue. B, The CBA/Ca VHJ606:Vλ Ab shown with a non-BL binding mutant constructed using BALB/c H chains and CBA/Ca L chains. BALB/c VH residues present in the non-BL binding, constructed VHJ606:Vλ Ab are shown colored yellow. VH 50Q and 53D present in the CBA/Ca mAb and engineered into the mutant mAb are shown colored blue. A single BALB/c FR residue, VH 84N, unavoidably introduced into the mutant mAb by the engineering procedure, is shown colored pale yellow.

Close modal

To visualize the mutations associated with avidity and fine specificity changes, molecular models of the engineered V regions (Fig. 2) were constructed using the homology module of the Insight II software package. Reference structures (1IAI for VHJ606, Ref. 47 ; 1FVD for Vκ11, Ref. 31 ; and 1NGQ for Vλ, Ref. 48) were chosen based on sequence identity, likelihood of sharing of CDR loop canonical structure, and the quality of the existing crystallographic data. Residues critical for high inulin avidity, 53 in the VH and VL, are shown in the model to be near the edge of the CDR surface, spaced ∼28 A from each other. In contrast, residues involved with overall binding to levan and with fine specificity, 50 in VH and the H3 loop, reside near the VH:VL junction.

In this report, we have extended our previous studies (40) on a panel of BALB/c- and CBA/Ca BL-specific hybridomas by constructing engineered IgM Abs and V region models to determine how inulin avidity and fine specificity of anti-BL mAb are affected by CDR substitutions. We show that presumed somatic mutations evident in the VH and VL regions of the higher avidity mAb of the BALB/c hybridoma panel directly cause marked increases in inulin avidity. We also show that exchange of unique VH residues or regions between VHJ606:Vκ11 BALB/c and VHJ606:Vλ CBA/Ca mAb with different fine specificities results in either fine specificity shift or total loss of inulin binding. Our three-dimensional models of the V regions suggest that residues near the edge of the CDR surface are critical for binding to inulin, while the VH:VL junctional area is important for binding both β(2→1) inulin and β(2→6) levan and determining fine specificity.

The impact of individual residue changes on Ag affinity has been demonstrated in several Ag:Ab systems (double-stranded DNA, Refs. 56 and 57 ; galactogloboside, Ref. 58 ; cortisol, Ref. 26 ; digoxin, Ref. 25 and 59 ; 2-phenyl-5-oxazolone, Ref. 60 ; and phthalate, 61). In this report, we have identified three individual CDR substitutions, VH N53H, VL N53I, and VL S30N, that directly increase avidity for a polysaccharide, inulin. All of these substitutions were found in more than one independently derived high avidity mAb from panel of anti-BL hybridomas (Fig. 1; Ref. 40). Mutating these residues alone largely reconstitutes the avidity increases seen in mAb in the hybridoma panel, arguing that the CDR mutations alone contribute to the avidity increases. These data argue that the affinity maturation process for inulin favors the selection of a small set of substitutions that correlate with high avidity.

To gain insights into how these substitutions may affect avidity increases, they were analyzed in the context of existing three-dimensional structural data. Currently, the structurally resolved PS-Ig complexes are the Sel55-4 Ab complexed with the Salmonella oligosaccharide-Ag (62) and BR96 Ab complexed with the Lewis Y Ag (63). The Sel55-4:oligosaccharide-Ag interaction is dominated by Van der Waals contacts with hydrophobic residues and hydrogen bonds with the aromatic amino acids Trp and His in the Ab. In the second complex, BR96:Lewis Y Ag, contacts between aromatic residues and PS also predominate.

Our data with anti-inulin Abs correlate with these observations in that two substitutions that increased avidity for inulin were a change from a polar to a hydrophobic amino acid (VL N53I) and an the addition of an aromatic amino acid (VH N53H). Our data differ from that of Cygler et al. (62) in that we find that CDR2 in VH and VL play a major role in inulin binding, while they noted that these two CDRs in Sel55-4 had few contacts with the Salmonella oligosaccharide-Ag. Both positions 53 in VH and 53 in VL have been shown in other resolved Ab:Ag structures to participate in Ag contact (E. Padlan, unpublished observations). Thus, it is conceivable that these particular changes increase avidity by allowing for the creation of hydrogen bonds or hydrophobic contacts with inulin. Molecular models (Fig. 2) suggest that both positions 53 in VH and 53 in VL are near opposite edges of the CDR surface and spatially distant (∼28 A), yet play a role in inulin avidity. Our models are consistent with previous modeling data, predicting that inulin is a bulky structure relative to β(2→6) levan and likely to form contacts with CDR residues far from each other and the VH:VL junction (34, 38).

Our data demonstrate that fine specificity is determined in large part by the H3 loop and VH 50 and 53. Substitutions at position 53 in VH can either decrease or increase avidity for β(2→1) inulin over a ∼870-fold range, while at the same time leave avidity for β(2→6) levan largely unaffected. This observation is similar to the results of Schildbach et al. (64) who observed that certain V region single amino acid changes in anti-digoxin Abs lowered their specificity for digoxin and at the same time shifted their specificity toward digoxin analogues with steroid-12 hydroxyl groups.

In contrast, exchange of the H3 loop or position 50 in VH between the VHJ606:Vκ11 BALB/c and the VHJ606:Vλ CBA/Ca mAb, instead of modulating fine specificity, ablates all BL binding activity. Our models suggest that the H3 loop and position 50 in VH lie close to the VH:VL junction. Both position 50 in VH and the H3 loop have been shown in several resolved Ab:Ag structures to participate in contacts with Ag (E. Padlan, unpublished observations). Thus, it is conceivable that these residues in the VH:VL junctional area influence fine specificity by forming specific contacts with different determinants of levan, but only in the context of the correct VH:VL pairing. Interestingly, the BALB/c H3 loop that correlates with inulin binding in BBLC44.1 was also found in a CBA/Ca VHJ558:Vκ11 inulin-binding mAb from the same hybridoma panel (40), providing further evidence that a VH:VL junctional area formed partly by this H3 loop in the context of Vκ11 creates a bias toward this fine specificity.

The CBA/Ca mAb, CBL166.10, pairs VHJ606 with a Vλ gene of wild mouse origin, similar to the human Vλ subgroup VI. This gene was demonstrated to be absent in BALB/c mice, and only ∼40% identical with the existing BALB/c Vλ genes (53). Thus, the absence of this Vλ gene in BALB/c mice may preclude the development of this CBA/Ca specificity in BALB/c anti-BL responses. The observation that Vλ gene repertoire influences the fine specificity of anti-levan Ab responses is similar to our previous observation on VH gene repertoire influencing fine specificity to the same Ag (40). It is also similar to data from κ-chain-deficient mice where the absence of κ-chain expression leads to alterations in the VH gene usage in response to levan and other Ags (65).

In summary, we have extended our previous studies (40) on a panel of BALB/c- and CBA/Ca BL-specific hybridomas by constructing engineered IgM Abs and V region models to determine how inulin avidity and fine specificity of anti-BL mAb are affected by CDR substitutions. We also show that two presumed somatic mutations in the VH and VL regions of some of the higher-avidity mAb from the BALB/c hybridoma panel, VH N53H and VL N53I, directly cause marked increases in inulin avidity (between 9- and 46-fold). We show that exchange of BALB/c- or CBA/Ca-specific VH residues at 50 or 53 or exchange of the H3 loop between VHJ606:Vκ11 BALB/c and VHJ606:Vλ CBA/Ca mAb with different fine specificities results in either fine specificity shift or total loss of inulin binding. Our three-dimensional models of the V regions suggest that portions near the edge of the CDR surface are critical for binding to inulin, while the VH:VL junctional area is critical for binding both β(2→1) inulin and β(2→6) levan and determining fine specificity. Additional insights of how these specific mutations affect avidity and fine specificity can be gained by the crystallographic resolution of Ab:levan complexes of the BALB/c and CBA/Ca Abs.

We thank Dr. Sean Fitzsimmons for extensive advice on PCR cloning Ig V regions; Mrs. Elaine Lizzio for assistance with cell culture; Dr. Edwardo Padlan (National Institute of Diabetes and Digestive and Kidney Diseases/National Institutes of Health) for helpful discussions about Ab structure; and Drs. Suzanne Epstein and Edward Max and Mr. Richard Venable for critical review of this manuscript.

4

Abbreviations used in this paper: PS, polysaccharide; BL, bacterial levan; GL, grass levan; CDR, complementarity-determining region; FR, framework region.

1
Roberts, I..
1996
. The biochemistry and genetics of capsular polysaccharide production in bacteria.
Annu. Rev. Microbiol.
50
:
285
2
Moxon, E., J. Kroll.
1990
. The role of bacterial polysaccharide capsules as virulence factors.
Curr. Top. Micro. Immunol.
150
:
65
3
Austrian, R..
1977
. Prevention of pneumococcal infection by immunization with capsular polysaccharides of Streptococcus pneumoniae: current status of polyvalent vaccines.
J. Infect. Dis.
136
:
S38
4
Gold, R., M. Lepow, I. Goldschneider, T. Draper, E. Gotshlich.
1979
. Kinetics of antibody production to group A ad group C meningococcal polysaccharide vaccines administered during the first six years of life: prospects for routine immunization of infants and children.
J. Infect. Dis.
140
:
690
5
Smith, D., G. Peter, D. Ingram, P. Anderson.
1973
. Responses of children immunized with the capsular polysaccharide of Haemophilus influenzae type b.
Pediatric
52
:
637
6
Robbins, J., J. Parke, R. Schneerson, J. Whisnant.
1973
. Quantitative measurement of “natural” and immunization-induced Haemophilus influenzae type b capsular polysaccharide antibodies.
Pediat. Res.
7
:
103
7
Robbins, J..
1978
. Vaccines for the prevention of encapsulated bacterial diseases: current status, problems and prospects for the future.
Immunochemistry
15
:
839
8
Robbins, J., R. Schneerson, S. Szu.
1995
. Perspective: hypothesis: serum IgG antibody is sufficient to confer protection against infectious diseases by inactivating the inoculum.
J. Inf. Dis.
171
:
1387
9
Siber, G., C. Thompson, G. Reid, J. Almeido-Hill, B. Zacher, M. Wolff, M. Santosham.
1992
. Evaluation of bacterial polysaccharide immune globulin for the treatment or prevention of Haemophilus influenzae type b and pneumococal disease.
J. Infect. Dis.
165
:
S129
10
Butler, J. C., S. F. Dowell, R. F. Breiman.
1998
. Epidemiology of emerging pneumococcal drug resistance: implications for treatment and prevention.
Vaccine
16
:
1693
11
Casadevall, A., M. Scharff.
1994
. Serum therapy revisited: animal models of infection and development of passive antibody therapy.
Antimicrob. Agents Chemother.
38
:
1695
12
Heidelberger, M..
1953
. Persistence of antibodies in man after immunization. A. M. Pappenheimer, Jr, ed.
The Nature and Significance of the Antibody Response
90
Columbia University Press, New York.
13
Casadevall, A., M. Scharff.
1991
. The mouse antibody response to infection with Cryptococcus neoformans: VH and VL usage in polysaccharide binding antibodies.
J. Exp. Med.
174
:
151
14
Clevinger, B., J. Schilling, L. Hood, J. Davie.
1980
. Structural correlates of cross-reactive and individual idiotypic determinants on murine antibodies to α(1–2) dextran.
J. Exp. Med.
151
:
1059
15
Hartman, A., S. Rudikoff.
1984
. VH genes encoding the immune response to β(1, 6)-galactan: somatic mutation in IgM molecules.
EMBO J.
3
:
3023
16
Kimura, H., E. Buescher, E. Ball, D. Marcus.
1989
. Restricted usage of VH and Vκ genes by murine monoclonal antibodies against 3-fucosyllactosamine.
Eur. J. Immunol.
19
:
1741
17
Perlmutter, R. J., M. Klotz, M. Bond, J. Davie Nahn, L. Hood.
1984
. Multiple VH gene segments encode murine anti-streptococcal antibodies.
J. Exp. Med.
159
:
179
18
Boswell, C., D. Irwin, J. Goodnight, K. E. Stein.
1992
. Strain dependent restricted VH and VL usage by anti-bacterial levan monoclonal antibodies.
J. Immunol.
148
:
3864
19
Bonilla, F. H., M. Rubin Zaghouani, Bona C.
1990
. Vκ gene usage, idiotype expression and antigen binding among clones expressing the VHX24 gene family derived from naïve and anti-idiotype immune BALB/c mice.
J. Immunol.
145
:
616
20
Bona, C..
1993
. Molecular characteristics of anti-polysaccharode antibodies.
Springer Semin. Immunopathol.
15
:
103
21
Padlan, E..
1996
. X-ray crystallography of antibodies.
Adv. Protein Chem.
49
:
57
22
Chothia, C., A. Lesk, A. Tramontano, M. Levitt, S. Smith-Gill, G. Air, S. Sheriff, E. Padlan, D. Davies, W. Tulip, P. Colman, S. Spinelli, P. Alzari, R. Poljak.
1989
. Conformations of immunoglobulin hypervariable regions.
Nature
342
:
877
23
Morea, V., A. Tramontano, M. Rustici, C. Chothia, A. Lesk.
1998
. Conformations of the third hypervariable region in the VH domain of immunoglobulins.
J. Mol. Biol.
275
:
269
24
Griffiths, G., C. Berek, M. Kaartinen, C. Milstein.
1984
. Somatic mutation and maturation of the immune response to 2-phenyl oxazolone.
Nature
312
:
271
25
Schildbach, J., S-Y. Shaw, R. Bruccoleri, E. Haber, L. Herzenberg, G. Jager, P. Jeffrey, D. Panka, D. Parks, R. Near, J. Novotny, S. Sheriff, M. Margolies.
1994
. Contribution of a single heavy chain residue to specificity of an anti-digoxin monoclonal antibody.
Protein Sci.
3
:
737
26
Chames, P., S. Coulon, D. Baty.
1998
. Improving the affinity and the fine specificity of an anti-cortisol antibody by parsimonious mutagenesis and phage display.
J. Immunol.
161
:
5421
27
Dall’ Acqua, W., E. Goldman, W. Lin, C. Teng, D. Tsuchiya, H. Li, X. Ysern, B. Braden, Y. Li, S. Smith-Gill, R. Mariuzza.
1998
. A mutational analysis of binding interactions in an antigen-antibody protein-protein complex.
Biochemistry
37
:
7981
28
Wallick, S., E. Kabat, S. Morrison.
1988
. Glycosylation of a VH residue of a monoclonal antibody against α(1→6) dextran increases its affinity for antigen.
J. Exp. Med.
168
:
1099
29
Co, M. S., D. Scheinberg, N. Avdalovic, K. McGraw, M. Vasquez, P. Caron, C. Queen.
1993
. Genetically engineered deglycosylation of the variable domain increases the affinity of an anti-CD33 monoclonal antibody.
Mol. Immunol.
30
:
1361
30
Banfield, M., D. King, A. Mountain, R. Brady.
1997
. VL:VH domain rotations in engineered antibodies: crystal structures of the Fab fragments from two murine anti-tumor antibodies and their engineered human constructs.
Proteins
29
:
161
31
Eigenbrot, C., M. Randal. L. Presta, P. Carter, A. Kossiakoff.
1993
. X-ray structures of the antigen-binding domains from three variants of humanized anti-p185HER2 antibody 4D5 and comparison of molecular modeling.
J. Mol. Biol.
229
:
969
32
Carpita, N., T. Housley, J. Hendrix.
1991
. New features of plant-fructan structure revealed by methylation analysis and carbon-13 n.m.r. spectroscopy.
Carbohydrate Res.
217
:
127
33
Han, Y..
1990
. Microbial levan.
Adv. Appl. Micro.
35
:
171
34
Lieberman, R., C. Bona, C. Chien, K. Stein, W. Paul.
1979
. Genetic and cellular regulation of the expression of specific antibody idiotypes in the anti-polyfructosan immune response.
Ann. Immunol.
130C
:
247
35
Lieberman, R., M. Potter, W. Humphrey, E. Mushinski, M. Vrana.
1975
. Multiple individual and cross-specific idiotypes on 13 levan-binding myeloma proteins of BALB/c mice.
J. Exp. Med.
142
:
106
36
Streefkerk, D., C. Glaudemans.
1977
. Binding studies on anti-fructofuranan mouse myeloma immunoglobulins A47N, A4, U61, and E109.
Biochem.
16
:
3760
37
Victor-Kobrin, C., F. Bonilla, B. Bellon, C. Bona.
1985
. Immunochemical and molecular characterization of regulatory idiotopes expressed by monoclonal antibodies exhibiting or lacking β2–6 fructosan binding activity.
J. Exp. Med.
162
:
647
38
Streefkerk, D., B. Manjula, C. Glaudemans.
1979
. An interpretation of the apparent dual specificity of some murine myeloma immunoglobulins with inulin-binding activity.
J. Immunol.
122
:
537
39
Hall, B., C. Bona, C. Victor-Kobrin.
1990
. Binding specificities of inulin-binding immunoglobulins for sinistrin and oligosaccharides isolated from asparagus roots.
Mol. Immunol.
27
:
351
40
Boswell, C., K. Stein.
1996
. Avidity maturation, repertoire shift, and strain differences in antibodies to bacterial levan, a type 2 thymus-independent polysaccharide antigen.
J. Immunol.
157
:
1996
41
Higgins, D., P. Sharp.
1988
. Clustal: a package for performing multiple sequence alignment on a microcomputer.
Gene
73
:
237
42
Henikoff, S., J. Henikoff.
1992
. Amino acid substitution matrices from protein blocks.
Proc. Natl. Acad. Sci. USA
89
:
10915
43
Alt, F., A. Bothewell, M. Knapp, E. Siden, E. Mather, M. Koshland, D. Baltimore.
1980
. Synthesis of secreted and membrane bound immunoglobulin Mu heavy chains is directed by mRNA that differs at their 3′ ends.
Cell
20
:
293
44
Coffino, P., M. Scharff.
1971
. Rate of somatic mutation in immunoglobulin production by mouse myeloma cells.
Proc. Natl. Acad. Sci. USA
68
:
219
45
Chien, C., R. Lieberman, J. Inman.
1979
. Preparation of functionalized derivatives of inulin: conjugation of erythrocytes for hemagglutination and plaque-forming cell assays.
J. Immunol. Methods
26
:
39
46
Lesk, A., and A. Tramontano. 1992. Antibody structure and structural predictions useful in guiding antibody engineering. In Antibody Engineering, A Practical Guide, C. Borrebaeck, ed. Freeman, New York.
47
Ban, N., C. Escobar, R. Garcia, K. Hasel, J. Day, A. Greenwood, A. McPherson.
1994
. Crystal structure of an idiotype-anti-idiotype Fab complex.
Proc. Natl. Acad. Sci. USA
91
:
1604
48
Mizutani, R., K. Miura, T. Nakayama, I. Shimada, Y. Arata, Y. Satow.
1995
. Three-dimensional structures of the Fab fragment of murine N1G9 antibody from the primary immune response and of its complex with (4-hydroxy-3-nitrophenyl)-acetate.
J. Mol. Biol.
254
:
208
49
Al-Lazikani, B., A. Lesk, C. Chothia.
1997
. Standard conformations for the canonical structures of immunoglobulins.
J. Mol. Biol.
273
:
927
50
Chang, C., P. Whitaker, L. Tabernero, H. Einspahr, L. Workman, D. Benjamin, S. Sheriff.
1994
. Crystallization and preliminary X-ray analysis of an anti-staphylococcal nuclease-staphylococcal nuclease complex and of a second anti-staphylococcal nuclease antibody.
J. Mol. Biol.
239
:
154
51
Brunger, A., D. Leahy, T. Hynes, R. Fox.
1991
. 2.9 A resolution structure of an anti-dinitrophenyl-spin-label monoclonal antibody Fab fragment with bound hapten.
J. Mol. Biol.
221
:
239
52
Kabat E. A., T. Wu, H. Perry, K. Gottesman, and C. Foeller (eds.). 1991. Sequences of Proteins of Immunological Interest. U.S. Department of Health and Human Services, National Institutes of Health, Bethesda, MD.
53
Reidl, L., C. Kinoshita, L. Steiner.
1992
. Wild mice express an Ig Vλ gene that differs from any Vλ in BALB/c but resembles a human Vλ subgroup.
J. Immunol.
149
:
471
54
Connolly, M..
1983
. Solvent-accessible surfaces of proteins and nucleic acids.
Science
221
:
709
55
Morea, V., A. Tramontano, M. Rusticim C. Chothia, A. Lesk.
1997
. Antibody structure, prediction and redesign.
Biophys. Chem.
68
:
9
56
Wloch, M., S. Clarke, G. Gilikeson.
1997
. Influence of VH CDR3 arginine and light chain pairing on DNA reactivity of a bacterial DNA-induced anti-DNA antibody from a BALB/c mouse.
J. Immunol.
159
:
6083
57
Hande, S., T. Manser.
1998
. Single amino acid substitutions in VH CDR2 are sufficient to generate or enhance the specificity of two forms of an anti-arsonate antibody variable region for DNA.
Mol. Immunol.
34
:
1281
58
Snyder, J., Q. Dinh, S. Morrison, E. Padlan, M. Mitchell, L. Yu-Lee, D. Marcus.
1994
. Structure-function studies of anti-3fucosyllactosamine (Lex) and galactosylgloboside antibodies.
J. Immunol.
153
:
1161
59
Panka, D., M. Mudgett-Hunter, D. Parks, L. Peterson, L. Herzenberg, E. Haber, M. Margolies.
1988
. Variable region framework differences result in decreased or increased affinity of variant anti-digoxin antibodies.
Proc. Natl. Acad. Sci. USA
85
:
3080
60
Berek, C., C. Milstein.
1987
. Mutation drift and repertoire shift on the maturation of the immune response.
Immunol. Rev.
96
:
23
61
Lou, S., D. Winter, G. Mayers, R. Bankert.
1992
. Nucleotide sequence of messenger RNA encoding VHDJH and VκJκ of a highly conserved idiotype-defined primary response anti-hapten antibody.
J. Immunol.
149
:
3944
62
Cygler, M., D. Rose, D. Bundle.
1991
. Recognition of a cell-surface oligosaccharide of pathogenic Salmonella by an antibody Fab fragment.
Science
253
:
442
63
Jeffrey, P., J. Bajorath, C. Chang, D. Yelton, I. Hellstrom, K. Hellstrom, S. Sheriff.
1995
. The X-ray structure of an anti-tumour antibody in complex with antigen.
Nat. Struct. Biol.
2
:
466
64
Schildbach, J. D., D. Panka, G. Parks, J. Jager, L. Novotny, M. Herzenberg, R. Bruccolerim Mudgett-Hunter, E. Haber, M. Margolies.
1991
. Altered hapten recognition by two anti-digoxin hybridoma variants due to variable region point mutations.
J. Biol. Chem.
266
:
4640
65
Bot, A., A. Nangpal, L. Pricop, B. Bogen, A. Kaushik, C. Bona.
1996
. Vλ-light chain genes reconstitute immune responses to defined carbohydrate antigens or haptens by utilizing different VH genes.
Mol. Immunol.
33
:
1359