6) Dextran Antibody → Behavior of an Anti

Position Effects of Variable Region Carbohydrate on the Affinity and In Vivo Behavior of an Anti-(1→6) Dextran Antibody

This information is current as M. Josefina Coloma, Ryan K. Trinh, Alexander R. Martinez of September 27, 2021. and Sherie L. Morrison J Immunol 1999; 162:2162-2170; ; http://www.jimmunol.org/content/162/4/2162 Downloaded from

References This article cites 45 articles, 14 of which you can access for free at: http://www.jimmunol.org/content/162/4/2162.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 27, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Position Effects of Variable Region Carbohydrate on the Afﬁnity and In Vivo Behavior of an Anti-(136) Dextran Antibody1

M. Joseﬁna Coloma, Ryan K. Trinh, Alexander R. Martinez, and Sherie L. Morrison2

IgG is a glycoprotein with an N-linked carbohydrate structure attached to the CH2 domain of each of its heavy chains. In addition, the variable regions of IgG often contain potential N-linked carbohydrate addition sequences that frequently result in the attachment of V region carbohydrate. Nonetheless, the precise role of this V region glycan remains unclear. Studies from our laboratory have shown that a naturally occurring somatic mutant of an anti-dextran Ab that results in a carbohydrate addition 58 site at Asn of the VH has carbohydrate in the complementarity-determining region 2 (CDR2) of the VH, and the presence of

carbohydrate leads to an increase in affinity. However, carbohydrate attached to nearby positions within CDR2 had variable Downloaded from affects on affinity. In the present work we have extended these studies by adding carbohydrate addition sites close to or within all the CDRs of the same anti-dextran Ab. We find that carbohydrate is attached to all the novel addition sites, but the extent of glycosylation varies with the position of the site. In addition, we find that the position of the variable region carbohydrate influences some functional properties of the Ab, including those usually associated with the V region such as affinity for Ag as well as other characteristics typically attributed to the Fc such as half-life and organ targeting. These studies suggest that modification

of variable region glycosylation provides an alternate strategy for manipulating the functional attributes of the Ab molecule and http://www.jimmunol.org/ may shed light on how changes in carbohydrate structure affect protein conformation. The Journal of Immunology, 1999, 162: 2162–2170.

ntibodies are glycoproteins that have at least one showed that carbohydrate addition sites present within the CDR2

N-linked carbohydrate added to the constant regions of of the VH gene 19.22.1 were actually used, and N-linked carbo- A the heavy chains. Many Abs also have V region-associ- hydrate attached (7, 15). This was true both for a naturally occur- ated carbohydrate. In fact, it has been calculated that human serum ring somatic mutation at position 60 (Asn3Thr) resulting in gly- 58 IgG has, on the average, 2.8 N-glycoside-type sugar chains/protein cosylation of Asn in the VH CDR2 (hybridoma 14.6b.1) as well molecule (1). Two of these carbohydrate moieties represent the as for glycosylation sites introduced at residues 54 and 60 within by guest on September 27, 2021 conserved N-linked carbohydrate in the Fc region, while the re- the CDR2 using site-directed mutagenesis. Interestingly, glycosyl- 58 mainder reflect V region glycosylation. Analysis of human my- ation of Asn of the VH increased the affinity of the Ab for Ag 60 eloma proteins and Abs of other species has indeed verified the approximately 10-fold, while carbohydrate at Asn of the VH only 54 frequent presence of V region carbohydrate (2–7). About 18% of increased the affinity 3-fold, and carbohydrate at VH Asn actu- 3 the VH sequences in Kabat’s database (8) contain a potential N- ally blocked the binding of Ag (15). In the anti-dextran Ab one of linked glycosylation site, Asn-X-Ser/Thr (9). The presence of this the surprising observations was that the carbohydrate added at po- site, however, does not guarantee that it is used for carbohydrate sition 60 remains in the high mannose form, while position 58, attachment (10, 11). Although asymmetric utilization of the V re- only two amino acids apart in sequence, contains carbohydrate that gion glycosylation sites has been reported (6, 12, 13), both V re- is processed to complex glycans. While differential accessibility gions can be glycosylated (7). has been proposed as the cause of differential processing, it seems The role of the V region carbohydrate remains unclear, but sev- likely that the protein itself plays an important role in determining eral studies suggest that its presence may influence affinity and/or the specificity of processing. specificity (4, 9, 14). Previous studies of anti-(136) dextran Abs The structure of the V region carbohydrate is frequently different from that of the constant region carbohydrate, and the structure

Department of Microbiology, Immunology, and Molecular Genetics, The Molecular of the V region carbohydrate can be associated with differences in Biology Institute, University of California, Los Angeles, CA 90095 binding specificity and affinity (16, 17). For the anti-dextrans, car- 54 58 Received for publication June 11, 1998. Accepted for publication November 6, 1998. bohydrate attached at Asn and Asn was a biantennary complex ␣ 3 The costs of publication of this article were defrayed in part by the payment of page structure containing Gal 1 3Gal as a nonreducing terminus, charges. This article must therefore be hereby marked advertisement in accordance while the carbohydrate attached to the Fc of the same Ab lacked with 18 U.S.C. Section 1734 solely to indicate this fact. Gal␣133Gal at its terminus (18). Recent evidence suggests that 1 This work was supported by National Institutes of Health Grants CA16858, carbohydrate can influence characteristics such as organ localiza- AI39187, and AI29470. tion, clearance rates, and receptor binding (19, 20). Variation 2 Address correspondence and reprint requests to Dr. Sherie L. Morrison, Department of Microbiology, Immunology, and Molecular Genetics, University of California, in glycoform structure has been associated with some disease 1602 Molecular Sciences Building, 609 Circle Drive East, Los Angeles, CA 90095- states (21, 22). 1489. E-mail address: [email protected] In the present work we have extended our studies of V region 3 Abbreviations used in this paper: VH and VL, heavy and light chain variable regions; carbohydrate by adding carbohydrate addition sites close to or CDR, complementarity-determining region; IMDM, Iscove’s modified Dulbecco’s ␤ ␤ within CDR1 and CDR3 of the VH as well as the CDR1, CDR2, medium; Endo-H, endoglycosidase H; aka, apparent affinity constant; t1/2, phase half-life. and CDR3 of the VL of the same anti-dextran Ab. The sites chosen

Copyright © 1999 by The American Association of Immunologists 0022-1767/99/$02.00 The Journal of Immunology 2163 were predicted to be on the surface of the native protein based on Ab analysis the molecular model of the binding site of the 19.22.1 and were The Abs were analyzed by metabolic labeling and immunoprecipitation. introduced as single amino acid changes to minimize the effect of Between 3 and 5 ϫ 106 cells were washed, resuspended in 1 ml of labeling amino acid substitution on the binding site (23). We joined the medium (high glucose DMEM deficient in methionine: Life Technologies, ␬ Grand Island, NY) containing 25 ␮Ci of [35S]methionine (Amersham, Ar- mutant VL and VH to human IgG1 and constant regions and expressed different combinations of the mutations as chimeric Abs lington Heights, IL), and allowed to incorporate label for 3–4 h or overnight with the addition of 1% FCS or ␣-calf serum at 37°C under tissue in the cell line Sp2/0. culture conditions. The Abs were immunoprecipitated with 2.5 ␮lofa We found that the position of the V region carbohydrate can rabbit anti-human IgG Fc and anti-human Fab polyclonal antiserum and a influence many of the functional properties of the Ab. These in- 10% suspension of IgGSorb (Enzyme Center, Woburn, MA). The precip- clude not only properties usually associated with the V region, itated labeled Ab was then analyzed by SDS-PAGE on 5% sodium phosphate-buffered polyacrylamide gels. To examine heavy and light chains such as affinity for Ag, but other characteristics as well, such as separately, the labeled sample was reduced by treatment with 0.15 M 2-ME half-life and organ targeting, and may shed light on how changes at 37°C for 30 min and analyzed on 12.5% Tris-glycine buffered poly- in carbohydrate structure affect protein conformation and create acrylamide gels. potentially pathogenic molecules. These studies suggest that mod- Tunicamycin treatment of transfectomas ification of V region glycosylation provides an alternate strategy for manipulating the functional attributes of the Ab molecule. To prevent N-linked glycosylation of Abs during synthesis, transfectants were grown in the presence of tunicamycin before biosynthetic labeling (29). Between 3 and 5 ϫ 106 cells were washed twice with PBS and preincubated in 1 ml of tissue culture medium (IMDM with 5% calf serum) Materials and Methods in the presence of 8 ␮g/ml of tunicamycin (Boehringer Mannheim, Indi- Downloaded from Mutagenesis anapolis, IN) for 3–5 h under tissue culture conditions. Then the cells were metabolically labeled as described above in 1 ml of labeling medium con- The heavy chain sugar addition sites were introduced by site-directed mu- taining 25 ␮Ci of [35S]methionine and 8 ␮g/ml of tunicamycin. Immune tagenesis with a modified version of the Zoller and Smith method using as precipitations were performed, and the Abs were analyzed as before. template the cDNA from the 19.22.1 VH (7). The VHCDR1 oligonucleotide (ACA3AAC (Thr-Asn)) was 5Ј-ACTGGCTACAACTTCAGTAGC, Treatment with endoglycosidase H (Endo-H)

3 Ј http://www.jimmunol.org/ while the VHCDR3 oligonucleotide (TAC AAC (Tyr-Asn)) was 5 - To determine the degree of carbohydrate processing, we digested proteins GGCATTACAACGGTAGTAGC (mutagenic bases are in bold and the with Endo-H glycosidase, which cleaves oligomannose-type oligosaccha- Asn is underlined). Carbohydrate addition sites were introduced into the rides. Labeled immunoprecipitated Ab bound to IgGsorb was resuspended VLCDRs by overlapping PCR using standard or touch-down PCR (24) in 100 ␮l of reaction buffer (0.1 M citrate containing 1 ␮l of 2-ME and 1 using the cDNA from the 19.22.1 VL as template (7). To amplify the V ␮l of 250 mM PMSF), and 10 ␮l of Endo-H was added (Boehringer Mann- region from the plasmid DNA and introduce a complete leader region and heim). The digestion was allowed to proceed for 18–24 h at 37°C. After an EcoRV (bold) site for cloning, a 5Ј primer that includes a ribosome boiling for 2 min the Abs were analyzed by PAGE under reducing condi- recognition site (underlined) (25, 26) and a start codon and spans nine tions in a 12.5% Tris-glycine gel. A control Ab similarly treated but lack- amino acids of the leader sequence was used (27): 5Ј-GGGATATC ing Endo-H was included for each protein. CACCATGGATTTTCAAGTGCAGATTTTCAG. The 3Ј antisense exter- nal primer hybridizes to the J␬4 region in framework region 4 and contains Purification a splicing signal (underlined) and a SalI site (bold) for cloning: 5Ј-AGC by guest on September 27, 2021 GTCGACTTACGTTTTATTTCCAGCTGGCCC. The mutagenic primers Chimeric Abs were purified from 2–4 l of culture supernatant containing used were the sense and antisense oligomers of the following sequences: 1% calf serum by protein A-Sepharose affinity chromatography following 283 Ј the protocol suggested by the manufacturer (Sigma). They were then dia- VL CDR1 Ser Asn (AGT-AAT), 5 -GCCAGCTCAAATGTAAGT 503 Ј lyzed against 2–4 l of dialysis buffer (50 mM Tris and 150 mM NaCl, pH TAC; VL CDR2 Asp Asn (GAC-AAC), 5 -GATGGATTTATACC ,.؅ 7.8) and concentrated using Millipore Ultra free-15 filters (Millipore Corp 913 ACATCCAAAC; and VL CDR3 Trp Asn (TGG-AAC), 5 -ACTG CCAGCAGAACAGTAGTAACCCG (the mutagenic bases are in bold Bedford, MA), and the concentration of the Ab was determined by a bicin- and the Asn is underlined). The PCR products were cloned into a TA choninic acid protein assay (Pierce, Rockford, IL). Purified Abs were then plasmid (Invitrogen, Carlsbad, CA). All V region sequences were con- visualized by SDS-PAGE. The yield of the Abs varied from 1–20 mg firmed using the Sequenase 2.0 kit (U.S. Biochemical, Cleveland, OH) with depending on the transfectant cell line. the protocol described by the manufacturer. Quantitation of affinity Tissue culture supernatants were harvested from overnight cultures of 3 ϫ 6 Cloning in expression vectors and transfection 10 cells. To normalize the concentrations of all Abs tested, the supernatants were diluted to yield the same signal as that determined by an anti-

The VH genes containing the CDR1 and CDR3 glycosylation sites were IgG ELISA. The tissue culture supernatants were used to determine the excised from M13 RF with EcoRI and were used to replace the V region apparent association constant of the mutant Abs to dextran B512 using the in the human ␥1 pSV2⌬HgptHuG1 expression vector (28). The expression method of Nieto et al. (30). Flat-bottom Immulon I microtiter plates (Dy- 58 ␮ ␮ vector with the VH glycosylated in the CDR2 at Asn has been previously natech, Chantilly, VA) were coated with 100 lofa20 g/ml solution of described (15). The wild-type and the CDR mutant VL V regions were the Ag dextran B512 in borate-buffered saline. The plates were allowed to cloned as EcoRV-SalI fragments into the PCR expression vector dry overnight at 37°C, washed with PBS, and blocked with 3% BSA. pAG4622 (27). Dilutions ranging from 0–20 ␮g/ml of soluble Ag (dextran B512) in bo- Different combinations of the wild-type and mutant light and heavy rate-buffered saline containing 1% BSA were mixed with Ab supernatant, chains were cotransfected into Sp2/0-Ag-14 cells by electroporation. Ten and 50 ␮l of the mix was added to the plate in triplicate and allowed to micrograms of expression vector prepared using Qiagen Maxi-Plasmid reach equilibrium overnight at 4°C. Binding was detected with an anti- Prep (Qiagen, Valencia, CA) was linearized with PvuI and transfected into human ␬ Ab conjugated to alkaline phosphatase or horseradish peroxidase 107 cells using a Gene Pulser (Bio-Rad, Hercules, CA) and was plated in and the appropriate substrate. Dextran B512 (200 ␮g/ml) was also added to Falcon 96-well tissue culture plates (Becton Dickinson, Mountain View, one set of triplicates to determine background. Color intensity was read at CA) at 125 ␮l/well (2 ϫ 106 cells/plate) with Iscove’s modiﬁed Dulbecco’s 410 or 480 nm using a Dynatech MR700 plate reader. The apparent binding medium (IMDM) supplemented with 10% calf serum, 1% Nystatin sus- constant (aka; grams per milliliter) was calculated from the amount of pension (Life Technologies, Grand Island, NY), and 100 ␮g/ml gentamicin ligand (free dextran) necessary for 50% inhibition of Ab binding. The assay sulfate (Sigma, St. Louis, MO). Selection was performed as previously was performed three or four times, and averages and SDs were calculated described (27) with medium containing 3 ␮g/ml mycophenolic acid (plus to obtain the apparent binding afﬁnity constant. hypoxanthine at 125 ␮g/ml and xanthine at 7.5 ␮g/ml). Surviving clones The Abs that had no detectable binding were tested in the aglycosylated were screened for Ab production by ELISA with anti-human ␥-coated forms after treating the cell cultures with tunicamycin followed by a Con plates and detected with anti-human ␬ alkaline phosphatase-conjugated Ab. A-Sepharose incubation to remove all possible glycosylated proteins. One Positive clones were subcloned by limiting dilution, then screened again, milliliter of Con A-Sepharose (Sigma) was washed twice with PBS, and and the ones selected were maintained in IMDM containing 5% calf serum. 100 ␮l of a 50% slurry was added to 10 ml of supernatant obtained from 2164 EFFECTS OF VARIABLE REGION GLYCOSYLATION IN AN ANTI-DEXTRAN Ab

cells cultured overnight in the presence of tunicamycin. After an overnight incubation at 4°C the Con A-Sepharose was removed by centrifugation, and the supernatant was recovered. After a second treatment, the supernatants were used for the afﬁnity assay. An IAsys optical biosensor (Fisons Applied Sensor Technology, Cam-

bridge, U.K.) was also used to determine the on and off binding rates (ka, Ϫ1 kd) and to calculate the affinity (Kd, molar concentrations ) of the anti- dextrans for the carboxymethylated dextran matrix of commercially available cuvettes (Fisons Applied Sensor Technology). A cuvette was activated and blocked as recommended by the manufacturer. After washing with PBS-Tween (0.05%) different concentrations of purified Abs were added to the cuvette at room temperature in 200 ␮l of PBS-Tween, and association was measured for 10 min. After washing the cuvette, dissociation was followed for 7 min. Complete dissociation and regeneration of the surface were obtained by a quick wash with 10 mM HCl. Using the Fastfit program (Fisons Applied Sensor Technology), the apparent association and disso-

ciation rates were calculated for each concentration of ligand, and the Kd was calculated as kd/ka.

Half-life FIGURE 1. Positions of the asparagines introduced to create glycosylation addition sites. Shown is a molecular model of 19.1.2 Fv, which is Four- to eight-week-old female BALB/c mice (Taconic Farms, German- town, NY) were given 0.1 mg/ml potassium iodide in their drinking water similar to the combining site of 14.6b.1, the Ab used for the present studies Downloaded from at least 4 days before and throughout the experiment to saturate the thyroid (23). The positions at which the amino acid mutations were introduced are iodine receptors. Ten micrograms of Abs purified by affinity chromatog- labeled in black. The VL is dark gray, and the VH is light gray. The graphic raphy were radioiodinated with 125I using Pierce Iodobeads. Separation of was produced using the program MacImdad (Molecular applications group, free 125I from the labeled protein was achieved by size exclusion chroma- Stanford University, Stanford, CA). tography on a 10-ml G-50 column in PBS, pH 7.2–7.4. Two hundred mi- croliters of the labeled protein (5 ϫ 105 cpm) was injected i.p. into five mice for each Ab. Whole body radioactivity was then determined at various chain (15). Despite the close proximity of the addition sites, their http://www.jimmunol.org/ time points for up to 500 h using a sodium iodide crystal with a well presence had different affects on the affinity of the Ab for Ag. sufficiently large to accommodate a mouse (W. B. Johnson & Associates, Montville, NJ). For each protein analyzed, values obtained from at least three mice were Production of the mutant Abs ␤ ␤ averaged and used for subsequent analysis. Half-life ( -phase; t1/2) was To extend our investigation of the role of V region carbohydrate, calculated using an exponential regression analysis of the data obtained site-directed mutagenesis has now been used to generate carbohy- after 24 h. drate addition sites within or near the other two CDRs of the heavy Biodistribution and immune response chain and all three CDRs of the light chain. We selected the positions at which the amino acid changes were introduced using the Blood was obtained 48 h following injection by subaxillary bleeding of molecular model of the 19.22.1 V region (23) as a guide (Fig. 1). by guest on September 27, 2021 mice rendered unconscious by exposure to ether. The radioactive Ab was immunoprecipitated from the serum using protein A-Sepharose and was Residues predicted to be on the surface with a serine two positions analyzed by SDS-PAGE and autoradiography. The biodistribution of ra- downstream in the amino acid sequence were chosen to be mutated diolabeled protein was determined by dissecting out the relevant organs to an Asn, thus creating a canonical N-linked carbohydrate addi- and then determining the radioactivity left in the organs using the whole tion sequence Asn-X-Ser/Thr (Table I). While it is generally ac- body counter. An average weight was calculated for each organ from the values obtained from 22 mice. Data were corrected for the weight and cepted that the X amino acid cannot be a proline, there are no other blood volume of the excised organ and were plotted. The nature of the known constraints on this sequence. To minimize the possible del- immune response mounted against the injected recombinant Abs by some eterious effects of an amino acid change on binding, only one mice was determined by an ELISA. Immulon II plates were coated over- amino acid was mutated in each case. ␮ night with 1 g/ml of goat anti-human IgG (Zymed, South San Francisco, The anti-dextran V regions were expressed with human ␥1, ␬ CA) followed by 3-h incubation at room temperature with 100 ␮lof1 ␮g/ml dilutions of the different anti-dextran mutant Abs or with an anti- constant regions in the non-Ig-expressing mouse myeloma cell line dansyl IgG1 chimeric. For each protein tested sera from two mice were Sp2/0-Ag14. Various combinations of the wild-type and mutant V pooled and diluted 1/500 in PBS-Tween; 100 ␮l was added to each well regions were transfected to generate Abs potentially containing and allowed to bind overnight at 4°C. After washing, bound Ab was de- carbohydrate at different positions in the hypervariable loops of the tected with a rat anti-mouse IgG-alkaline phosphatase Ab (Sigma). VH, the VL, or both. All transfections resulted in secretion of proteins that could be detected by ELISA. To determine whether the Results new glycan addition sites in the other two VH CDRs and the VL The hybridoma 19.22.1 was originally isolated as a clone produc- CDRs were indeed used and to assess whether the presence of ing IgM-␬ antibodies specific for anti-(13 6) dextran (31). The carbohydrate interferes with either the assembly or the secretion of hybridoma 14.6b.1, isolated at the same time, exhibits a 10-fold the Ab, transfectants were biosynthetically labeled with [35S]me- higher affinity for anti-(13 6) dextran and differs at only a single thionine in the presence or the absence of tunicamycin, and the Ab

amino acid in VH, where Asn at position 60 was changed to Thr, was immunoprecipitated from the secretions and analyzed by presumably as a result of somatic mutation and selection. Asn58 SDS-PAGE. All the mutant Abs were secreted as fully assembled

within 14.6b.1 then constitutes a potential N-glycosylation site, H2L2 molecules (Fig. 2A), and all showed increased electro- and studies showed that the increased afﬁnity of 14.6b.1 for dex- phoretic mobility after the transfectants were grown in tunicamy- tran resulted from the presence of carbohydrate at this position cin to impede attachment of carbohydrate. (31). Subsequent studies with these V regions showed that novel glycosylation sites introduced by site-directed mutagenesis at po- Analysis of glycosylation

sitions 54 and 60 within the CDR2 of the VH gene had N-linked To further characterize the Abs and to determine whether sugar carbohydrate attached when the heavy chains were expressed by was present at the introduced glycosylation positions in VH or VL, transfection into a cell line producing the murine anti-dextran light the immunoprecipitated Ig was analyzed by reducing SDS-PAGE, The Journal of Immunology 2165

Table I. CDR sequences and position of the mutations introduced to create carbohydrate addition sites

Anti-dextran V H 28 ͓O¡ CDR 1: Thr Gly Tyr Thr Phe Ser Ser Tyr Trp Ile Glu Asn ء O¡ 58͓ CDR 2: Ile Leu Pro Gly Ser Gly Ser Thr Asn Tyr Asn Glu Lys Phe Lys Gly ͓O¡ 97 Thr CDR 3: His Tyr Tyr Gly Ser Ser Ser Asn

Anti-dextran VL ͓O¡ 28 CDR 1: Ser Ala Ser Ser Ser Val Ser Tyr Met His Asn ͓-50 O¡ CDR 2: Trp Ile Thr Asp Thr Ser Lys Leu Ala Ser Asn ͓O¡ 91 CDR 3: Gln Gln Trp Ser Ser Asn Pro Tyr Tyr Asn Downloaded from

*, Naturally occurring somatic mutation.

thus separating the heavy from the light chain. The shift in size of to be incomplete utilization of the other sites. For VH28VL0 and http://www.jimmunol.org/ the control Ab, which has no glycosylation sites in the V regions VH97VL0, a portion of the secreted heavy chains seemed to con- (VH0VL0), following tunicamycin treatment reﬂects the presence tain only one carbohydrate. VH0VL50, with only V region carbo- of the carbohydrate within CH2 (Fig. 2B). All the heavy and light hydrate in the L chain, showed incomplete VL glycosylation. All chains that contain an engineered glycosylation addition site in the Abs with a glycosylation site at VL28 showed incomplete utiliza- variable region showed a further shift in size upon treatment with tion of this site when it was present as the sole Fv carbohydrate in tunicamycin, suggesting that carbohydrate is actually attached to VH0VL28 or when it was expressed in combination with a carbo- these sites and that the sugar can be added to both chains when the hydrate containing VH in VH28VL28, VH58VL28, and VH97VL28. sites are available. Glycosylation sites with Asn at VH58 and VL91 The glycosylated VL28 in these Abs resolves as a compact doublet appeared to always be fully glycosylated. However, there seemed that may represent two glycoforms. VH58VL28 is particularly by guest on September 27, 2021

FIGURE 2. PAGE of biosynthetically labeled Abs. Transfectants were grown overnight in the presence of [35S]methionine, and the radiolabeled Ig was immunoprecipitated with rabbit anti-Fc and anti-Fab and protein A-Sepharose. A, Five percent phosphate-buffered gel showing assembly of Abs. B, Twelve percent Tris glycine-buffered gel of reduced Abs showing the light and heavy chains. ϩ/Ϫ, treatment of transfectants with tunicamycin to prevent carbohydrate attachment. 2166 EFFECTS OF VARIABLE REGION GLYCOSYLATION IN AN ANTI-DEXTRAN Ab

Table II. Apparent binding constants (aka) and afﬁnity constants (KD) of the anti-dextran Abs

ϩ Ab Asn aka TM KD IAsys Ϫ1 Name VH/VL aka (g/ml) (g/ml) M TAPH 0/0 6.71 ϫ 105 Ϯ 0.25 TAKZ 0/28 2.70 ϫ 105 Ϯ 0.77 TAJS 28/28 NDa 2.7 ϫ 105 TAJF 58/28 2.00 ϫ 106 Ϯ 0.35 6.3 ϫ 10Ϫ8 TAJI 97/28 ND 4.2 ϫ 105 TARA 28/0 2.51 ϫ 105b TARB 58/0 4.26 ϫ 106 Ϯ 0.15 9.6 ϫ 10Ϫ8 TAPV 97/0 6.91 ϫ 105 Ϯ 0.20 TAOC 0/50 ND ND TAOB 0/91 ND ND a ND, nondetectable binding. FIGURE 3. IAsys sensogram of the high affinity anti-dextran Abs. Dif- b Binding detected with horseradish peroxidase. ferent concentrations of Ab were allowed to bind to a carboxymethylated dextran cuvette surface, and dissociation was observed after washing. Only representative concentrations are shown to demonstrate the differences in binding between the two mutant Abs. noteworthy, in that while the heavy chain appears to be completely Downloaded from glycosylated, the light chain migrates as three bands, with the un- usual slowly migrating band presumably reflecting the attachment of different carbohydrate structures. In all cases the aglycosylated carbohydrate in the CDR1 VH28 or CDR3 VH97 resulted in Abs heavy and light chains recovered after treatment of the cells with for which no detectable Ag binding was seen. Interestingly, tunicamycin show identical mobility confirming that carbohydrate VH28VL28 and VH97VL28 did bind Ag when produced in the is responsible for the observed heterogeneity (Fig. 2B). presence of tunicamycin; however, their association constants http://www.jimmunol.org/ To determine whether any of the recombinant Abs have partially were 2.5- and 1.6-fold lower, respectively, than that of wild-type processed sugars attached, immunoprecipitated labeled Ab was VH0VL0, suggesting that the amino acid changes introduced in- treated with Endo-H glycosidase, which is specific for terminal fluence the interaction with Ag. In contrast, VH0VL50 and oligomannose glycans. Abs that were subjected to the same re- VH0VL91 showed no binding even after they were produced in the agents and incubation treatments but without Endo-H were in- presence of tunicamycin to avoid glycosylation, suggesting that the cluded for comparison. An anti-dextran Ab, glycosylated at posi- change in amino acid to generate the Asn is responsible for the tion 60 with high mannose sugars and sensitive to Endo-H, was lack of binding. 97 used as a control. Analysis of all Abs by reducing SDS-PAGE Glycosylation at Asn in the VH CDR3 appeared to have no by guest on September 27, 2021 revealed no shift in size in any of the heavy or light chains, sug- influence on binding, as VH97VL0 had the same apparent binding gesting that none of the Abs contain high mannose oligosaccha- constant as VH0VL0. This also confirms that it is the carbohydrate rides (data not shown). However, the structure of the attached car- in the VL that has a negative contribution leading to lack of binding 28 bohydrate remains to be determined. by VH97VL28. Carbohydrate at Asn of the CDR1 of the VH was not well tolerated, since VH28VL0 had a 2.7-fold decrease in af- Determination of the apparent binding constants and affinity finity, and VH28VL28 showed no binding. To assess the impact of carbohydrate present at different positions An IAsys biosensor (Fisons Applied Sensor Technology, Cam- within the combining site on the affinity of the anti-dextran Abs, bridge, U.K.) was used to determine the Kd (molar concentra- Ϫ we determined their apparent association constants using an inhi- tion 1) of the anti-dextran Abs through a direct measurement. A bition ELISA assay as previously described (15, 30) and compared cuvette with a carboxymethylated dextran surface was used as the them to that of an anti-dextran lacking carbohydrate in the V re- Ag, and different concentrations of Abs were tested for binding. gion produced in the same system. Supernatant harvested from Only the two Abs glycosylated in the VH CDR2, VH58VL0 and 6 overnight cultures of 10 cells was tested for binding to a dextran- VH58VL28, had a high enough affinity to be detected by the bio- coated plate in the presence of different concentrations of free dex- sensor (Fig. 3 and Table II). At the same concentration, VH58VL0 tran B512. The reciprocal value of the concentration of free dex- showed more extensive binding to the dextran cuvette than tran required to prevent 50% of binding is the apparent binding VH58VL28, consistent with its higher aka. The remaining Abs did constant, which is expressed in grams per milliliter. The Abs that not show detectable binding to the dextran cuvette. showed no binding were also tested in the aglycosylated form after production in tunicamycin-treated cells and incubation with Con In vivo behavior A-Sepharose to remove any remaining glycosylated species. The serum half-life in BALB/c mice was obtained by injecting

The Ab with an engineered Asn at VH58, VH58VL0, equivalent radioiodinated Ab into the peritoneum of 5- to 7-wk-old female to mAb 14.6b.1, had a 6.5-fold increase in the apparent association mice and counting the mice using a whole body counter to deter- constant compared with that of the Ab lacking V region carbohy- mine the residual radioactivity. For all Abs except VH58VL28, drate (VH0VL0; Table II). The presence of carbohydrate at VL28 40–50% rapidly cleared during the first 24 h; during the same time 28 always had a negative influence on binding. When Asn VL was 65% of the injected dose of VH58VL28 cleared. After 24 h the the only carbohydrate in the binding site, VH0VL28 bound Ag, remaining radioactivity was cleared at a slower rate, with 15% of albeit with a 2.5-fold reduced apparent affinity compared with Ab the injected dose of VH58VL28 and about 30–40% of the other ␤ lacking V region carbohydrate VH0VL0. When the heavy chain proteins remaining in the mice after 200 h. The calculated t1/2 for 58 28 with carbohydrate at Asn was paired with the light chain Asn , the majority of anti-dextran Abs was 7–8 days; VH58VL28 cleared ␤ the resulting Ab VH58VL28 showed a decrease in affinity com- more rapidly, with a t1/2 of 4 days, and VH0VL50 and VH0VL91 ␤ pared with that of VH58VL0, while VL28 in combination with cleared more slowly, with a t1/2 of 10 days (Fig. 4). Abs recov- The Journal of Immunology 2167

sylation. In contrast, much of the immune response against

VH28VL28 appeared directed against an epitope dependent on the site of V region glycosylation. Sera from mice injected with anti- dextran lacking V region carbohydrate showed no reactivity with the constant or V region, consistent with the failure to see an ac- celeration in clearance (Fig. 5). The biodistribution in mice of the injected Abs was determined by removing relevant organs from two mice for each protein 48 h after injection and determining the residual radioactivity. The radioactivity present in each organ was adjusted for the weight and blood volume, and the distribution of the counts in each mouse was compared for each protein. Only the organs that showed a significant accumulation of counts or targeting are shown in Fig. 6. The position of the V region carbohydrate was found to inﬂuence the in

vivo distribution of the Ab. For VH58VL28, which showed the FIGURE 4. In vivo clearance of the glycosylated Abs. Radioiodinated most rapid clearance, there was less in the blood and more in the Ab was injected i.p. in BALB/c mice, and the disappearance of whole body liver and kidney. Less VH97VL0 was also present in the blood, but radioactivity was monitored over time. The ␤t was calculated starting 1/2 this Ab was not rapidly cleared, and it did not accumulate in the 24 h postinjection. liver. Instead, an increase was seen in the lungs and skin, suggest- Downloaded from

ing that it was sequestered at these sites. VH97VL28 and to a lesser ered from the serum of two mice sacriﬁced at 48 h postinjection extent VH0VL28 were noteworthy in that very little was present in and analyzed by SDS-PAGE appeared intact. the skin but both persisted in the blood. Finally, VH0VL97 and

In some mice injected with VH28VL0, VH28VL28, VH58VL28, VH0VL50, which had the most extended half-lives, had at least 3 VH58VL0, and VH97VL28, all Abs with glycosylation in VH,an times more radioactivity present in the spleen than the other Abs. immune response was observed after 200 h as indicated by a sud- http://www.jimmunol.org/ den increase in the rate of clearance of the Ab (not shown). None of the mice injected with Abs glycosylated only in the light chain Discussion V region or the Ab devoid of V region carbohydrate showed ev- In studies performed several years ago we demonstrated that the 58 idence of an immune response even 400 h post injection. To de- presence of N-linked carbohydrate at Asn in the VH of the Ag termine the specificity of this immune response, an ELISA was binding site of an Ab specific for ␣(13 6) dextran (TKC3.2.2 or performed using anti-dextrans with V region glycosylation at dif- 14.6b.1) increases its affinity for dextran approximately 10-fold ferent positions and an anti dansyl-IgG1 chimeric Ab bound to a (15). Since carbohydrate plays a demonstrable role in their binding plate through an anti-human Fc region Ab. Sera from mice injected behavior, anti-dextrans provide a useful system for studying the by guest on September 27, 2021 with VH58VL0, VH58VL28, VH28VL28 or VH0VL0 collected role of V region glycosylation in Ab function. To investigate the 400 h postinjection diluted 1/500 were added to the Ag-coated effects of carbohydrate attached at Asn residues close to the Ag plates. Bound murine Abs were detected by an alkaline phospha- contact loops on the affinity and the in vivo behavior of Abs, site- tase-linked rat anti-mouse IgG. The immune response to VH58VL0 directed mutagenesis and gene transfection techniques were used appeared to be mostly specific for the constant regions, since to place carbohydrate addition sequences at novel sites in all the equivalent reactivity was seen with the chimeric anti-dansyl and CDRs of the anti-dextran Ab, choosing positions expected to be on anti-dextran Abs. Much of the immune response to VH58VL28 was the surface of the combining site and thus predicted to be acces- directed at the murine variable regions and did not appear to be sible to glycosyltransferases. Carbohydrate addition to the V re- directed against epitopes that were strongly influenced by glyco- gion of Abs can be a mechanism used by the immune system to

FIGURE 5. Immune response ELISA. Sera recovered after 400 h from mice injected with VH58VL0, VH58VL28, VH28VL28, and VH0VL0 were incubated with plates coated with anti-dextrans with different sites of V region glycosylation or an anti-dansyl chimeric IgG1. Bound murine Ab was detected using AP-labeled rat anti-mouse IgG. 2168 EFFECTS OF VARIABLE REGION GLYCOSYLATION IN AN ANTI-DEXTRAN Ab

FIGURE 6. Biodistribution of the glycosylated Abs. Radioiodinated Ab was injected i.p. into BALB/c mice, and the presence of radioactivity in organs harvested at 48 h was determined. All measurements were corrected for the blood volume of the organ.

modulate Ab specificity, affinity, and stability. Although sequence Although glycosylation takes place at all the V region carbohy- Downloaded from analysis shows that about one-third of Abs contain V region car- drate addition sites introduced in the current studies, we found that bohydrate addition sites, examination of the germline V sequences some sites were only partially used. It is unclear what causes this (V base, Medical Research Council Center for Protein Engineer- incomplete glycosylation. It is possible that the rate of folding of ing, Cambridge, U.K.) (32–34) shows an absence of potential N- the region might limit oligosaccharide attachment and processing. linked carbohydrate sites, suggesting that they mainly arise from Carbohydrate addition takes place on the nascent polypeptide somatic mutations. However, the presence of V region glycosyl- chain, and once folding of this region has occurred, sites can be- http://www.jimmunol.org/ ation does not always have a direct impact on binding. There are come inaccessible (37, 38). We found no evidence of terminal high Abs with glycosylation addition sites in which the attached oligomannose residues in any of our Abs, since all were Endo-H gly- saccharide does not affect the binding site (35). These oligosac- cosidase resistant. Therefore, the attached carbohydrate is avail- charides are usually present in the framework regions and could able to the processing enzymes as the protein transits the Golgi potentially be used for the attachment of drugs or markers in di- apparatus. However, the altered mobility of a portion of the light agnostic or therapeutic molecules without affecting the specificity chains in VH58VL28 suggests that different glycoforms are present of binding (36). In the present study we now find that only caron some of the light chains and that the presence of carbohydrate bohydrate added at position 58 within VH significantly increases at VH58 influences the processing of carbohydrate attached at VL28. Ab affinity compared with Abs without V region carbohydrate. The rules governing the use of a potential N-linked glycosyla- by guest on September 27, 2021 This increase is observed when the glycosylated heavy chain is tion site are not completely understood, although it is accepted that paired with a nonglycosylated light chain or a light chain with proline cannot be the middle amino acid in the canonical N-linked carbohydrate at position 28. Anti-carbohydrate Abs are intrinsi- carbohydrate addition site. However, analysis of V region se- cally of low affinity, and glycosylation at Asn VH 58 is present as quences that actually have N-linked carbohydrate suggests that the a naturally occurring somatic mutant. The combining site of the presence of hydrophobic amino acids such as Ala at the middle anti-dextran is a shallow groove, and all of the engineered glyco- residue can inhibit site utilization. In previous studies with an anti- sylation sites would be expected to surround this groove. The Asn dansyl Ab, mutation of Ala in the sequence Asn-Ala-Thr to Gly at VH 58 is at the very edge of the binding site, and the increased resulted in the use of this previously unused site (our unpublished affinity of the glycosylated Ab is thought to result from the hy- observation). It also has been observed that when proline is present drophilic interactions of carbohydrate with carbohydrate (Ag). in the fourth position after the Asn, the site is not used. In fact, 97 Glycosylation at Asn in the VH CDR3 appears to have no although the majority of murine heavy chain V regions belonging influence on binding, suggesting that the attached carbohydrate is to subgroup IIIb have Asn at position 58, they differ from the positioned so that it is not available to interact with the carbohy- germline VH used in these studies in that they have a proline in- drate Ag. In contrast, we find that carbohydrate introduced at other stead of a glutamic acid at position 61. sites has an adverse affect on the affinity of the Ab. Glycosylation The presence of carbohydrate in one V region of one chain does at Asn28 of the heavy chain decreases the affinity of the anti- not impede the attachment of carbohydrate to the other chain. 28 dextran by 2.5-fold. Glycosylation at Asn of the VL always re- There is no evidence of asymmetric glycosylation in any of the sulted in a decrease in binding affinity, reducing the affinity by half mutant Abs containing engineered sites in both VH and VL. The 3 when present by itself or in combination with VH58. Ag binding 1 6 linkages of the N-linked oligosaccharides are a major source was abolished when VL28 was paired with either VH28 or VH97. of intramolecular flexibility, and conformational heterogeneity Since Ag binding was observed for these Abs when they were may contribute to structural heterogeneity, since the accessibility produced in the presence of tunicamycin, the presence of the two of enzymes to the site can be altered by changing the orientation of bulky carbohydrate groups must impair access to the Ab binding the flexible linkages (39). The protein itself can also play an im- site. In contrast, the glycosylation sites introduced at VL50 and portant role in determining the specificity of processing, and even VL91 lead to Abs that do not bind Ag even if glycosylation is the quaternary structure of a protein has been found to influence prevented by growth in the presence of tunicamycin. Analysis of the structure of the carbohydrate added to different sites (40). contact loops predicts that VL50 and VL91 make important con- Different glycoforms on the surface of immune-related proteins tacts with the Ag, and it is therefore probable that the amino acids can give them altered or distinguishable functions (19, 20) and introduced at these sites to form the carbohydrate addition sites have been related to disease states (21, 22, 41). Although the in destroy the ability to bind Ag. vivo half-lives of the majority of mutant anti-dextrans were very The Journal of Immunology 2169 similar to the expected half-life of human IgG1 in mice, three Abs of human monoclonal antibody HB4C5 to lung tumor xenografts by N-deglyco- showed a marked difference in the rate of clearance; V 0V 50 and sylation. J. Nucl. Med. 35:289. H L 5. Meng, Y. G., A. B. Criss, and K. E. Georgiadis. 1990. J chain deficiency in VH0VL91 had a longer half-life of 10 days, while VH58VL28 had human IgM monoclonal antibodies produced by Epstein-Barr virus-transformed a faster clearance rate and a shorter half-life of 4 days. These data B lymphocytes. Eur. J. Immunol. 20:2505. ␥ are consistent with the radioactivity measured in the different or- 6. Prelli, F., and B. Frangione. 1992. Franklin’s disease: Ig 2 H chain mutant BUR. J. Immunol. 148:949. gans at 48 h postinjection, in that VH58VL28 showed a markedly 7. Wallick, S. C., E. A. Kabat, and S. L. Morrison. 1988. Glycosylation of a VH lower blood retention and larger accumulation in the liver and residue of a monoclonal antibody against ␣ (1–6) dextran increases its affinity for clearance through the kidneys. It is noteworthy that both V 0V 50 antigen. J. Exp. Med. 168:1099. H L 8. Kabat, E. A. 1991. Sequences of Proteins of Immunological Interest. U.S. De- and VH0VL91 were present in the spleen in increased amounts. partment of Health and Human Services, Public Health Service, National Insti- While mannose receptors abundant in the liver have been im- tutes of Health, Bethesda, MD, publication 91-3242:3 v. 9. Co, M. S., D. A. Scheinberg, N. M. Avdalovic, K. McGraw, M. Vasquez, plicated in the rapid clearance of Abs and other proteins containing P. C. Caron, and C. Queen. 1993. Genetically engineered deglycosylation of the unprocessed sugars with high mannose content (42), this does not variable domain increases the affinity of an anti-CD33 monoclonal antibody. Mol. Immunol. 30:1361. appear to be the case for VH58VL28, since its variable region carbohydrate is not sensitive to digestion with Endo-H glycosidase. 10. Kimura, H., R. Cook, K. Meek, M. Umeda, E. Ball, J. D. Capra, and D. M. Marcus. 1988. Sequences of the VH and VL regions of murine monoclonal Alternatively, it is possible that this Ab is clearing through the antibodies against 3-fucosyllactosamine. J. Immunol. 140:1212. asialoglycoprotein receptor present in the hepatocytes. This would 11. Snyder, J. G., N. Weng, L. Y. Yu-Lee, and D. M. Marcus. 1990. Heavy and light chain sequences of four monoclonal antibodies that bind galactosylgloboside imply that VH58VL28 is heavily galactosylated and could explain (GalGb4). Eur. J. Immunol. 20:2673. the presence of the third glycosylated species in the light chain (43, 12. Grebenau, R. C., D. M. Goldenberg, C. H. Chang, G. A. Koch, D. V. Gold, 44). The spleen uptake of V 0V 50 and V 0V 91 suggests that A. Kunz, and H. J. Hansen. 1992. Microheterogeneity of a purified IgG1 due to Downloaded from H L H L asymmetric Fab glycosylation. Mol. Immunol. 29:751. they are being processed by the reticuloendothelial system (45), 13. Malan Borel, I., T. Gentile, J. Angelucci, R. A. Margni, and R. A. Binaghi. 1990. but it is unclear how this would translate into a longer half-life. Asymmetric Fab glycosylation in guinea-pig IgG1 and IgG2. Immunology 70: The presence of carbohydrate in the V region of the heavy chain 281. 14. Kato, M., K. Mochizuki, S. Hashizume, H. Tachibana, S. Shirahata, and seems to increase the immunogenicity of the anti-dextrans. It is H. Murakami. 1993. Activity enhancement of a lung cancer-associated human interesting that carbohydrate at VH58 enhanced the response to the monoclonal antibody HB4C5 by N-deglycosylation. Hum. Antibodies Hybrid- omas 4:9. constant regions when paired with a light chain lacking carbohy- http://www.jimmunol.org/ 15. Wright, A., M. H. Tao, E. A. Kabat, and S. L. Morrison. 1991. Antibody variable drate, but when paired with a light chain with carbohydrate at region glycosylation: position effects on antigen binding and carbohydrate struc- position 28 it induced both an anti-V and an anti-constant region ture. EMBO J. 10:2717. response. A significant amount of the immune response to 16. Matsuuchi, L., J. Sharon, and S. L. Morrison. 1981. An analysis of heavy chain glycopeptides of hybridoma antibodies: correlation between antibody specificity VH28VL28 was anti-idiotypic. It is noteworthy that chimeric Abs and sialic acid content. J. Immunol. 127:2188. lacking V region carbohydrate did not elicit a detectable immune 17. Tachibana, H., S. Shirahata, and H. Murakami. 1992. Generation of specificity- variant antibodies by alteration of carbohydrate in light chain of human mono- response during the 10–15 days of the half-life studies despite the clonal antibodies. Biochem. Biophys. Res. Commun. 189:625. fact that they possessed human constant regions. 18. Endo, T., A. Wright, S. L. Morrison, and A. Kobata. 1995. Glycosylation of the Traditionally, changes in amino acid sequence have been used to variable region of immunoglobulin G: site specific maturation of the sugar chains.

Mol. Immunol. 32:931. by guest on September 27, 2021 alter the functional properties of Ab molecules. Our current studies 19. Cebon, J., N. Nicola, M. Ward, I. Gardner, P. Dempsey, J. Layton, U. Du¨hrsen, demonstrate that addition of V region carbohydrate can be used to A. W. Burgess, E. Nice, and G. Morstyn. 1990. Granulocyte-macrophage colony influence not only the ability of the Ab to interact with Ag but also stimulating factor from human lymphocytes: the effect of glycosylation on receptor binding and biological activity. J. Biol. Chem. 265:4483. in vivo properties such as half-life. Glycosylation at positions that 20. Donahue, R. E., E. A. Wang, R. J. Kaufman, L. Foutch, A. C. Leary, result in the reduction of the serum retention of an Ab could in- J. S. Witek-Giannetti, M. Metzger, R. M. Hewick, D. R. Steinbrink, G. Shaw, et crease its therapeutic value when rapid clearance is needed. Al- al. 1986. Effects of N-linked carbohydrate on the in vivo properties of human GM-CSF. Cold Spring Harb. Symp. Quant. Biol. 51:685. though the production of Ab fragments (46–48) or isotype switch- 21. Parekh, R. B., R. A. Dwek, J. R. Thomas, G. Opdenakker, T. W. Rademacher, ing (49) can be used to decrease half-life, when the effector A. J. Wittwer, S. C. Howard, R. Nelson, N. R. Siegel, M. G. Jennings, et al. 1989. functions of an Ab, such as Fc receptor binding and cytolysis, are Cell-type-specific and site-specific N-glycosylation of type I and type II human tissue plasminogen activator. Biochemistry 28:7644. important for the specific application of an Ab, conserving the 22. Delves, P. J. 1998. The role of glycosylation in autoimmune disease. Autoimmu- constant region of an active isotype and altering the half-life by nity 27:239. other means, such as carbohydrate engineering, are attractive op- 23. Padlan, E. A., and E. A. Kabat. 1988. Model-building study of the combining sites of two antibodies to alpha (1-Ͼ6)dextran. Proc. Natl. Acad. Sci. USA 85: tions. Additionally, the affinity of some Abs can be increased or 6885. decreased by the addition of carbohydrate at defined positions in 24. Roux, K. H., and K. H. Hecker. 1997. One-step optimization using touchdown the variable region. Carbohydrate accessible on the surface of the and stepdown PCR. Methods Mol. Biol. 67:39. 25. Kozak, M. 1987. At least six nucleotides preceding the AUG initiator codon Ab that does not alter the functional properties can provide sites enhance translation in mammalian cells. J. Mol. Biol. 196:947. for the attachment of drugs, radioactive tags, or even purification 26. Kozak, M. 1994. Determinants of translational fidelity and efficiency in vertebrate using lectins. mRNAs. Biochimie 76:815. 27. Coloma, M. J., A. Hastings, L. A. Wims, and S. L. Morrison. 1992. Novel vectors for the expression of antibody molecules using variable regions generated by References polymerase chain reaction. J. Immunol. Methods 152:89. 28. Dangl, J. L., T. G. Wensel, S. L. Morrison, L. Stryer, L. A. Herzenberg, and 1. Kinoshita, N., M. Ohno, T. Nishiura, S. Fujii, A. Nishikawa, Y. Kawakami, V. T. Oi. 1988. Segmental flexibility and complement fixation of genetically N. Uozumi, and N. Taniguchi. 1991. Glycosylation at the Fab portion of my- engineered chimeric human, rabbit and mouse antibodies. EMBO J. 7:1989. eloma immunoglobulin G and increased fucosylated biantennary sugar chains: 29. Kuo, S., and J. Lampen. 1974. Tunicamycin: an inhibitor of yeast glycoprotein structural analysis by high-performance liquid chromatography and antibody- synthesis. Biochem. Biophys. Res. Commun. 58:287. lectin enzyme immunoassay using Lens culinaris agglutinin. Cancer Res. 51: 30. Nieto, A., A. Gaya, M. Jansa, C. Moreno, and J. Vives. 1984. Direct measurement 5888. of antibody affinity distribution by hapten-inhibition enzyme immunoassay. Mol. 2. Milstein, C. P., and E. V. Deverson. 1974. Primary structure of ␬ light chain from Immunol. 21:537. a human myeloma protein. Eur. J. Biochem. 49:377. 31. Akolkar, P. N., S. K. Sikder, S. B. Bhattacharya, J. Liao, F. Gruezo, S. L. Morrison, 3. Arakawa, F., M. Haruno, M. Kuroki, H. Kanda, T. Watanabe, Y. Misumi, and and E. A. Kabat. 1987. Different VL and VH germ-line genes are used to produce Y. Matsuoka. 1993. Construction and expression of two mouse-human chimeric similar combining sites with specificity for alpha(136)dextrans [Published erratum antibodies with high specificity and affinity for carcinoembryonic antigen. Hy- appears in 1987 J. Immunol. 139:3911]. J. Immunol. 138:4472. bridoma 12:365. 32. Tomlinson, I. M., G. Walter, P. T. Jones, P. H. Dear, E. L. Sonnhammer, and 4. Kusakabe, K., K. Kanaya, S. Sato, M. Kato, K. Mochizuki, S. Hashizume, G. Winter. 1996. The imprint of somatic hypermutation on the repertoire of K. Yasumoto, H. Murakami, and K. Nomoto. 1994. Accumulation enhancement human germline V genes. J. Mol. Biol. 256:813. 2170 EFFECTS OF VARIABLE REGION GLYCOSYLATION IN AN ANTI-DEXTRAN Ab

33. Cox, J. P., I. M. Tomlinson, and G. Winter. 1994. A directory of human germ-line 43. Stein, R., D. M. Goldenberg, G. L. Ong, S. R. Thorpe, and M. J. Mattes. 1997. V kappa segments reveals a strong bias in their usage. Eur. J. Immunol. 24:827. Manipulation of blood clearance to optimize delivery of residualizing label-an- 34. Walter, G., I. M. Tomlinson, P. H. Dear, E. L. Sonnhammer, G. P. Cook, and tibody conjugates to tumor cells in vivo. J. Nucl. Med. 38:1392. G. Winter. 1995. Comparison of the human germline and rearranged VH reper- 44. Newkirk, M. M., J. Novick, M. M. Stevenson, M. J. Fournier, and toire reveals complementarity between germline variability and somatic muta- P. Apostolakos. 1996. Differential clearance of glycoforms of IgG in normal and tion. Ann. NY Acad. Sci. 764:180. autoimmune-prone mice. Clin. Exp. Immunol. 106:259. 35. Leung, S. O., A. S. Dion, M. C. Pellegrini, M. J. Losman, R. C. Grebenau, D. M. Goldenberg, and H. J. Hansen. 1995. Effect of VK framework-1 glycosyl- 45. Casey, J. L., D. J. King, L. C. Chaplin, A. M. Haines, R. B. Pedley, A. Mountain, ation on the binding affinity of lymphoma-specific murine and chimeric LL2 G. T. Yarranton, and R. H. Begent. 1996. Preparation, characterisation and tu- antibodies and its potential use as a novel conjugation site. Int. J. Cancer 60:534. mour targeting of cross-linked divalent and trivalent anti-tumour FabЈ fragments. 36. Leung, S., M. J. Losman, S. V. Govindan, G. L. Griffiths, D. M. Goldenberg, and Br. J. Cancer 74:1397. H. J. Hansen. 1995. Engineering a unique glycosylation site for site-specific 46. Murray, J. L., M. G. Rosenblum, H. Z. Zhang, D. A. Podoloff, L. P. Kasi, conjugation of haptens to antibody fragments. J. Immunol. 154:5919. S. A. Curley, J. C. Chan, M. Roh, D. C. Hohn, H. Brewer, et al. 1994. Compar- 37. Bergman, L. W., and W. M. Kuehl. 1978. Temporal relationship of translation and ative tumor localization of whole immunoglobulin G anticarcinoembryonic an- glycosylation of immunoglobulin heavy and light chains. Biochemistry 17:5174. tigen monoclonal antibodies IMMU-4 and IMMU-4 F(abЈ) in colorectal cancer 38. Bergman, L. W., and W. M. Kuehl. 1979. Co-translational modification of nas- 2 cent immunoglobulin heavy and light chains. J. Supramol. Struct. 11:9. patients. Cancer 73:850. 39. Wu, P. G., K. G. Rice, L. Brand, and Y. C. Lee. 1991. Differential flexibilities in 47. Koziol, J. A., P. P. Lee, R. O. Dillman, R. Fagnani, and S. E. Halpern. 1995. three branches of an N-linked triantennary glycopeptide. Proc. Natl. Acad. Sci. Pharmacokinetics of 111In-labelled monoclonal antibody ZCE-025 and frag- USA 88:9355. ments in tumour-bearing mice. Nucl. Med. Commun. 16:299. 40. Dahms, N. M., and G. W. Hart. 1986. Influence of quaternary structure on gly- 48. King, D. J., A. Turner, A. P. Farnsworth, J. R. Adair, R. J. Owens, R. B. Pedley, cosylation: differential subunit association affects the site-specific glycosylation D. Baldock, K. A. Proudfoot, A. D. Lawson, N. R. Beeley, et al. 1994. Improved of the common beta-chain from Mac-1 and LFA-1. J. Biol. Chem. 261:13186. tumor targeting with chemically cross-linked recombinant antibody fragments. 41. Youings, A., S. C. Chang, R. A. Dwek, and I. G. Scragg. 1996. Site-specific Cancer Res. 54:6176. glycosylation of human immunoglobulin G is altered in four rheumatoid arthritis patients. Biochem. J. 314:621. 49. Hutchins, J. T., F. C. Kull, Jr., J. Bynum, V. C. Knick, L. M. Thurmond, and Downloaded from 42. Narita, M., G. Bu, J. Herz, and A. L. Schwartz. 1995. Two receptor systems are P. Ray. 1995. Improved biodistribution, tumor targeting, and reduced immuno- involved in the plasma clearance of tissue-type plasminogen activator (t-PA) in genicity in mice with a ␥4 variant of Campath-1H. Proc. Natl. Acad. Sci. USA vivo. J. Clin. Invest. 96:1164. 92:11980. http://www.jimmunol.org/ by guest on September 27, 2021